**Naturally Occurring Calanolides: Occurrence, Biosynthesis, and Pharmacological Properties Including Therapeutic Potential**

**Lutfun Nahar 1,\*, Anupam Das Talukdar <sup>2</sup> , Deepa Nath <sup>3</sup> , Sushmita Nath <sup>4</sup> , Aman Mehan <sup>5</sup> , Fyaz M. D. Ismail <sup>4</sup> and Satyajit D. Sarker 4,\***


Academic Editors: Maria João Matos and Pascal Richomme Received: 30 September 2020; Accepted: 26 October 2020; Published: 28 October 2020

**Abstract:** Calanolides are tetracyclic 4-substituted dipyranocoumarins. Calanolide A, isolated from the leaves and twigs of *Calophyllum lanigerum* var. *austrocoriaceum* (Whitmore) P. F. Stevens, is the first member of this group of compounds with anti-HIV-1 activity mediated by reverse transcriptase inhibition. Calanolides are classified pharmacologically as non-nucleoside reverse transcriptase inhibitors (NNRTI). There are at least 15 naturally occurring calanolides distributed mainly within the genus *Calophyllum,* but some of them are also present in the genus *Clausena*. Besides significant anti-HIV properties, which have been exploited towards potential development of new NNRTIs for anti-HIV therapy, calanolides have also been found to possess anticancer, antimicrobial and antiparasitic potential. This review article provides a comprehensive update on all aspects of naturally occurring calanolides, including their chemistry, natural occurrence, biosynthesis, pharmacological and toxicological aspects including mechanism of action and structure activity relationships, pharmacokinetics, therapeutic potentials and available patents.

**Keywords:** calanolides; pseudocalanolides; calanolide A; *Calophyllum*; Calophyllaceae; anti-HIV; reverse transcriptase; non-nucleoside reverse transcriptase inhibitors (NNRTIs)

#### **1. Introduction**

Calanolides are tetracyclic 4-substituted dipyranocoumarins, and their C-ring contains a *gem*-dimethyl group (Figure 1), e.g., (+)-calanolide A (**1**), (−)-calanolide B (costatolide) (**14**) (Figure 2). The discovery of calanolides from the leaves and twigs of the tree *Calophyllum lanigerum var. austrocoriaceum* (Whitmore) P. F. Stevens, collected from Sarawak, Malaysia in 1987 happened during one of the largest anti-HIV screening programs conducted by the National Cancer Institute (NCI) during 1987–1996. In that program, over 30,000 plant extracts were screened utilizing an in vitro cell-based anti-HIV screen that could determine the degree of HIV-1 replication in treated infected lymphoblastic cells versus that in treated uninfected control cells [1,2]. Calanolide A (**1**) (Figure 1), which can be described as a 11,12-dihydro-2*H*,6*H*,10*H*-dipyrano[2,3-f:2′ ,3′ -h]chromen-2-one substituted by a hydroxyl (–OH) group at C-12, methyl groups at positions 6, 6, 10 and 11 and a propyl group at C-4

(the 10*R*,11*S*,12*S* stereoisomer), was isolated as the first member of anti-HIV compounds, calanolides, as a potential novel therapeutic option for the treatment of HIV infections. However, a subsequent attempt to recollect this plant sample failed and the collection of other specimens of the same species (not necessarily the same variety), afforded only a negligible amount of calanolide A (**1**). In fact, calanolides are among the first plant-based compounds to demonstrate potential anti-HIV-1 activity. Later, an extract of the latex of *C. teysmanii* showed significant anti-HIV activity in the screening, but the major active compound was (−)-calanolide B (**14**, also known as costatolide), regrettably not calanolide A (**1**) (Figure 2). The anti-HIV activity of (−)-calanolide B (**14**) was less potent than that of calanolide A (**1**), possibly because of difference in stereochemistry at the chiral centers. To date calanolides A-F and some of their methyl, acetyl and dihydro derivatives have been reported mainly from various *Calophyllum* species (Figure 2). Among these, the structures of calanolides C (**6**) and D (**7**), as reported initially by Kashman et al. [1] from *C. lanigerum*, were revised and renamed as pseudocalanolides C (**8**) and D (**9**) [3] (Figure 2). However, the true calanolides C (**6**) and D (**7**) were later reported from *C. brasiliense* Cambess. [4–6].

**Figure 1.** Rings A, B, C and D, and carbon numbering in (+)-calanolide A (**1**).

The first isolation of calanolides from *C. lanigerum var. austrocoriaceum,* involved multiple steps, starting with the extraction of dried fruits and twigs of this plant with a 1:1 mixture of dichloromethane and methanol, followed by a sequential solvent partitioning process involving various solvents. The *n*-hexane and CCl<sup>4</sup> fractions emerged as the active fractions [1]. Repeated vacuum liquid chromatography (VLC) on silica gel, eluting with a mixture of *n*-hexane and ethyl acetate afforded crude calanolides, which were further purified by HPLC, employing normal phase for calanolide A (**1**), calanolide B (**4**) and pseudocalanolide D (**9**) [reported incorrectly as calanolide D (**7**)], while reversed-phase for 12-acetoxycalanolide A (**2**), 12-methoxycalanolide A (**3**), 12-methoxycalanolide B (**5**), pseudocalanolide C (**8**) [reported as calanolide C (**6**)] and calanolide E (**10**). The structures of these compounds were determined by a combination of UV, IR, NMR and MS spectroscopic methods, and all spectroscopic data were published [1]. The absolute stereochemistry of calanolides A (**1**) and B (**4**) was confirmed by a modified Mosher′ s method.

There is a review [7] and a book chapter on calanolides [8], published about six years ago, that mainly cover anti-HIV activity, and the literature published until early 2014. This present review is not on the genus *Calophyllum*, the family Calophyllaceae or pyranocoumarins a such, but it exclusively focuses on various aspects of naturally occurring calanolides. This review is significantly different from any other previous articles on calanolides in its approach and coverage, and is a comprehensive update on naturally occurring calanolides, encompassing their chemistry, natural occurrence, biosynthesis, pharmacological and toxicological aspects including mechanism of action and structure activity relationships, pharmacokinetics, therapeutic potentials and available patents.

**Figure 2.** *Cont.*

**Figure 2.** Naturally occurring calanolides

#### **2. Occurrence**

Calanolides, calanolide A (**1**) being the first member of these 4-substituted pyranocoumarins isolated from *C. lanigerum var. austrocoriaceum*, are almost exclusively distributed within the genus *Calophyllum* L., which comprises a large group of ca. 200 species of tropical trees distributed in the Indo-Pacific region, but was also reported from one species (*Clausena excavate* Brum. f.) of the closely related genus *Clausena* [9–12] (Table 1). Calanolide A (**1**) and other calanolides were subsequently isolated from other *Calophyllum* species, e.g., *C. brasiliense* Cambess. [4,13,14], *C. inophyllum* L. [6], *C. teysmanii* Miq. [2] and *C. wallichianum* Planch. & Triana [15]. In a chemotaxonomic study on the *Calophyllum* species, the presence of calanolides was detected in the extracts of *C. inophyllum*, *C. lanigerum* var. *austrocoriaceum*, *C. mole* King, *C. nodosum*, aff. *Pervillei* Vesque., *C. soulattri* Burm. f., *C. tacamahaca* Willd. and *C. teysmanii* [9] (Table 1).


**Table 1.**Naturally occurring calanolides, their sources and properties.


**Table 1.** *Cont.*

Bernabe-Antonio et al. [5] reported the production of calanolides in a callus culture of *C. brasiliense*, where different concentrations and combinations of plant growth regulators were tested in leaf and seed explants to establish callus cultures capable of producing calanolides. Higher calanolides B (**4**) and C (**6**) production was observed in calluses from seed explants than those developed from leaves. In continuation of the search for new natural anti-HIV compounds, and at the same time to find new botanical sources of calanolides, McKee et al. [20] purified calanolide E2 (**12**), and calanolide F (**13**) from the extracts of *C. lanigerum var. austrocoriaceum* and *C. teysmanii var. inophylloide* (King.) P. F. Stevens. Later, costatolide (**14**), also known as (−)-calanolide B, was reported as an anti-HIV compound present in *C. cerasiferum* Vesque and *C. inophyllum* L. [24]. Calanolides A (**1**), and C (**6**), and costatolide (**14**) were isolated from the leaves of *C. brasiliense*, and their anti-HIV potential was evaluated [4].

#### **3. Biosynthesis**

Calanolides are biosynthesized from the parent simple coumarin 7-hydroxycoumarin, also known as umbelliferone (Schemes 1–3). The biosynthesis of umbelliferone in plants starts from the amino acid L-phenylalanine, and proceeds through the formation of *trans*-cinnamic acid, *p*-coumaric acid, 2-hydroxy-*p*-coumaric acid, 2-glucosyloxy-*p*-coumaric acid, and 2-glucosyloxy-*p*-*cis*-coumaric acid with the help of various enzymes like cinnamate 4-hydroxylase, 4-coumarate-CoA ligase, 4-coumaroyl 2 ′ -hydroxylase and so on [27]. The biosynthesis of dipetalolactone, a pyranocoumarin, and subsequent conversion to the 3-propyl-intermediate for calanolides may proceed through two routes, one through conversion of umbelliferone to osthenol (Scheme 1), and the other via formation of 5,7-dihdroxycoumarin (Scheme 2). Reactions are generally mediated by p450 monooxygenase and other non-p450 enzymes [28]. 3-Propyl-intermediate is converted to the precursor compound for calanolides A–C (**1, 4** and **6**), utilizing the Wagner-Meerwein rearrangement reaction, and the precursor compound is believed to be converted to calanolides with the help of p450 monooxygenase enzyme (Scheme 3). Published studies on the biosynthesis of calanolides are rather limited and only two publications are available on this topic to date [28,29]. Therefore, detailed knowledge of specific enzymes involved in the biosynthesis of calanolides is still in its infancy.

**Scheme 1.** Plausible biosynthetic route to 2′ -hydroxydihydrodipetalolactone from umbelliferone *via* formation of 5-hydroxy-6-prenylseselin.

**Scheme 2.** Plausible biosynthetic route to 2′ -hydroxydihydrodipetalolactone from umbelliferone *via* formation of 8-prenylalloxanthoxyletol.

In a recent study, the influence of soil nutrients, e.g., Ca2<sup>+</sup> and K+, on the biosynthesis of pharmacologically active calanolides in the seedlings of *C. brasiliense* was studied [29]. It was observed that the use of K<sup>+</sup> deficient modified Hoagland solution (MHS) could induce a 15, 4.2 and 4.3-fold decrease of calanolides B (**4**), C (**6**), and apetalic acid concentrations in the leaves of the seedlings, respectively. On the other hand, Ca2<sup>+</sup> deficient MHS could lead to a decrease of 4.3 and 2.4-fold for calanolides B (**4**) and C (**6**), respectively. This study demonstrated that, like many other plant secondary metabolites, the biosynthesis of calanolides, albeit genetically controlled, may also be affected by environmental conditions, e.g., soil nutrients (minerals).

As genes dictate biosynthesis of secondary metabolites, a study was conducted to identify candidate genes that regulate to the biosynthesis of calanolides in *C. brasiliense* [28]. The unigene dataset constructed in this study could offer an insight for further molecular studies of *C. brasiliense*, particularly for characterizing candidate genes responsible for the biosynthesis of angular and linear pyranocoumarins. The candidate genes, e.g., UN36044, UN28345 and UN34582, identified in the transcriptome of the leaves, stem and roots of *C. brasiliense* might be involved in the biosynthesis of calanolides, which are essentially modified angular pyranocoumarins. Candidate unigenes in the transcriptome dataset were screened using mainly homology-based BLAST and phylogenetic analyses. It is worthy of mention that the BLAST programs are widely used for searching protein and DNA databases for optimizing sequence similarities [30]. For protein comparisons, several definitional, algorithmic and statistical refinements allow substantial decrease in the execution time of the BLAST programs and enhancement of their sensitivity to weak similarities.

**Scheme 3.** Plausible biosynthetic route to calanolides A (**1**), B (**4**) and C (**6**) from the intermediate, 2 ′ -hydroxydihydrodipetalolactone.

#### **4. Pharmacological Properties**

Although well-known for non-nucleoside reverse transcriptase inhibitory activity offering anti-HIV potential, calanolides have also been shown to possess various other pharmacological properties (Figure 3). The following sub-sections deal with anticancer, anti-HIV, antimycobacterial and antiparasitic activity of naturally occurring calanolides. As much of the published pharmacological studies, both in vitro and in vivo including human trials, on naturally occurring calanolides are about their anti-HIV property, over the years, significant amounts of information have become available on their mechanism of action, structure-activity-relationships, synergistic and/or additive property and their potential in anti-HIV combination therapy, which have been discussed adequately under individual headings within the anti-HIV sub-section. All other pharmacological properties of these compounds as outlined in different publications still require further investigations to establish their realistic therapeutic potential. Also, in silico pharmacological activity and toxicity studies on these pyranocoumarins have just begun to emerge in recent years.

**Figure 3.** A pictorial summary of pharmacological properties of naturally occurring calanolides.

#### *4.1. Anticancer Activity*

In the later part of 1980s, as a part of the initiative of the United States National Cancer Institute (NCI), plant samples from the Malaysian flora were collected for routine screening for potential cytotoxicity against a collection of cancer cell lines as well as for possible anti-HIV activity. One of the samples, the leaves and twigs of the tree *C. lanigerum* var. *austrocoriaceum*, despite not being active against any of the cancer cell lines tested, showed inhibitory activity of viral replication when tested against HIV-1 virus [31,32]. However, later, calanolide A (**1**) and calanolide C (**6**) were shown to possess antiproliferative or antitumor-promoting property through inhibition of TPA-induced EBV-EA activation in Raji cell lines [13]. The phorbol ester, 12-*O*-tetradecanoylphorbol-13-acetate (TPA) is a potent stimulator of differentiation and apoptosis in myeloid leukemia cells. Calanolide A (**1**) was found to be more active (IC<sup>50</sup> = 290 mol ratio/32 pmol TPA) than its 10,11-*cis*-isomer, calanolide C (**6**) (IC<sup>50</sup> = 351 mol ratio/32 pmol TPA). It was inferred that 4-substituted pyranocoumarins like calanolides might possess potential as cancer chemopreventive agents or antitumor-promoters. A recent study with the crude ethanolic extract of the leaves of *C. inophyllum* revealed its potential as a cytotoxic agent (IC<sup>50</sup> 120 µg/mL) against the breast cancer cell line MCF-7 [33]; it was also found to possess antiproliferative and apoptotic properties. However, no definitive proof was provided to establish which of the secondary metabolites biosynthesized by this plant, calanolides being one major class, were responsible for the putative anticancer activity. Although not calanolides, a few other 4-substituted coumarins, isolated from *C. brasiliense*, were tested against human leukemia HL-60 cells with some promising results [34], which might highlight the need for more comprehensive studies with all major 4-susbtitued coumarins, including calanolides, to find antileukemia lead compounds for new anticancer drug development. Calanolide A (**1**), isolated from a chloroform extract of *Clausena excavata*, was found to induce toxicity to the cells used in a syncytium assay for anti-HIV activity [10].

The efficacy of calanolide A (**1**) in AIDS-associated cancer was evaluated in silico utilizing an integrated approach combining network-based systems biology, molecular docking and molecular dynamics [35]. Molecular targets were screened and only the targets, e.g., HRAS, that are common to HIV and sarcoma, HIV and lymphoma, and HIV and cervical cancer, were utilized in this study. Calanolide A (**1**) was found to form a stable complex with the screened target HRAS, which is a small G protein in the RAS subfamily of the RAS superfamily of small GTPases, and is considered as a proto-oncogene; when mutated, this proto-oncogene has the potential to cause normal cells to become cancerous.

#### *4.2. Anti-HIV Activity*

Calanolide A (**1**), an anti-HIV non-nucleoside reverse transcriptase inhibitor (NNRTI), paved the way for the discovery and synthesis of a series of 4-substituted angular pyranocoumarins with potential anti-HIV property [1,36]. NNRTIs are a class of anti-HIV drugs that prevent healthy T-cells in the body from becoming infected with HIV. Kashman et al. [1] first reported this new class of anti-HIV agents from the tropical rainforest tree, *C. lanigerum*. Calanolide A (**1**), 12-acetoxycalanolide A (**2**), 12-methoxycalanolide A (**3**), calanolide B (**4**), 12-methoxycalanolide B (**5**), pseudocalanolide C (**8**), pseudocalanolide D (**9**) and calanolide E (**10**) (Figure 2) were isolated through an anti-HIV bioassay-guided isolation. Calanolides A (**1**) and B (**4**) were found to be protective against HIV-1 replication and cytopathicity with EC<sup>50</sup> values of 0.1 µM and 0.4 µM, respectively. However, both compounds were inactive against HIV-2, which is known as less pathogenic than HIV-1 and mainly found in West African countries. The other compounds showed a low level of anti-HIV-1 activity. This study involving purified bacterial recombinant reverse transcriptases established that the calanolides are indeed HIV-1 specific reverse transcriptase inhibitors. A comparative report on the anti-HIV potentials of calanolide A (**1**), costatolide (**14**) and dihydrocostatolide (**16**) against a series of HIV isolates of different cellular phenotypes was published by Buckheit et al. [26], which clearly demonstrated that calanolide A (**1**) was the best anti-HIV candidate among the three calanolides tested.

Two analogs of calanolide A (**1**), i.e., costatolide (**14**) and dihydrocostatolide (**16**), were shown to possess anti-HIV property similar to that of calanolide A (**1**) [26] and could be ascribed to the class of NNRTIs. In fresh human cells, costatolide (**14**) and dihydrocostatolide (**16**) could significantly inhibit the low-passage clinical virus strains, including those representative of the various HIV-1 clade strains, syncytium-inducing and non-syncytium-inducing isolates, and T-tropic and monocyte-tropic isolates [26,37]. In continuation of the search for new natural anti-HIV compounds, McKee et al. [20] purified calanolide E2 (**12**), and calanolide F (**13**) from extracts of *C. lanigerum* var. *austrocoriaceum* and *C. teysmanii* var. *inophylloide* (King.) P. F. Stevens, and calanolide E2 (**12**) emerged as one of the most active anti-HIV compounds. Later, costatolide (**14**) was reported as an anti-HIV compounds present in *C. cerasiferum* Vesque and *C. inophyllum* L. [24], while calanolides A (**1**), and C (**6**), and costatolide (**14**), isolated from the leaves of *C. brasiliense*, were shown to possess anti-HIV potential [4]. Comparative anti-HIV activities of some naturally occurring calanolides, e.g., calanolide A (**1**), costatolide (**14**) and dihydrocostatolide (**16**), against various strains of HIV are available in the article by Buckheit, et al. [26].

#### 4.2.1. Activity Against Drug Resistant Strains of HIV-1

Interestingly, calanolide A (**1**) was not only found to be active against standard strains of HIV-1, but it was also active against the resistant strains, eAZT-resistant G-9106 strain of HIV-1 and pyridinone-resistant A17 strain [1,38]. The activity against the pyridinone-resistant A17 strain was of interest as this strain is highly resistant to most of the HIV-1 specific NNRTIs, for example, TIBO, BI-RG-587 and L693,593. Later, it was established that pyranocoumarin **1** could interact with HIV-1 reverse transcriptase within the previously defined common binding site for nonnucleoside inhibitors [38]. An assessment of the inhibition patterns of the chimeric reverse transcriptases containing complementary segments of HIV-1 and HIV-2 reverse transcriptases established that there was a segment between residues 94 and 157 in HIV-1 reverse transcriptase that was crucial for inhibition by calanolide A (**1**) [39]. However, it was assumed that there might be a second

segment, essential for specifying susceptibility to the drug, between amino acids 225 and 427 in HIV-1 reverse transcriptase. A couple of years later, it was noted that calanolide A (**1**) was active against virus isolates resistant to 1-[(2-hydroxyethoxy)methyl]-6-(phenylthio)thymine and its derivative, [1-benzyloxymethyl-5-ethyl-6-(alpha-pyridylthio)uracil] [40]. Furthermore, this pyranocoumarin (**1**) showed activity against HIV with the two most common NNRTI-related mutations, K103N and Y181C, and was found to select for a mutation that does not cause cross-resistance with any other NNRTIs under investigation. It was postulated that substitution at codon Y188H of reverse transcriptase could be associated with 30-fold resistance to calanolide A (**1**) in vitro [41]. The compound is essentially inactive against all strains of the less common HIV type 2. It is necessary to carry out appropriate in vivo experimentations, either in animal models or in human clinical trials, to understand the true potential of any putative drug candidate. In vivo anti-HIV activity of (+)-calanolide A (**1**) was assessed in a hollow fibre mouse model [42], and it was observed that this compound could suppress virus replication in two unique, but separate physiologic compartments following oral or parenteral administration.

Calanolides were found to possess an enhanced antiviral activity against one of the most prevalent NNRTI-resistant viruses that is engendered by the Y181C amino acid change in reverse transcriptase as well as with reverse transcriptases that possess the Y181C change together with AZT-resistant mutations [26,37]. Calanolides could also be active against viruses containing Y181C and K103N dual mutations, which are generally highly resistant to other known non-nucleus reverse transcriptase inhibitors. Anti-HIV activity of naturally occurring calanolides against drug-resistant strains of HIV have made these compounds promising structural templates for new anti-HIV drug development.

#### 4.2.2. Calanolides in Anti-HIV-1 Combination Therapy

For the treatment of HIV infections, use of combination therapy comprising several anti-HIV drugs has become a common practice in recent years. The synergistic effects of calanolide A (**1**), costatolide (**14**) and dihydrocostatolide (**16**) [26] in combination with established anti-HIV drugs, e.g., azidothymidine (AZT), indinavir, nelfinavir and saquinavir, are available in the literature [26]. Synergistic effects were observed in both cultured cells and animal models when calanolides were used in combination with other anti-HIV agents [43]. Both calanolide A (**1**) and costatolide (**14**) were found to be effective in combination therapy for HIV infections [44]; in combination with NNRTIs, costatolide (**14**) could only synergistically inhibit HIV type 1 with UC38, whilst calanolide A (**1**) in combination with one of the NNRTIs helped this drug to retain activity against virus isolates with the single Y181C mutation [26,41,44,45].

A combination of (+)-calanolide A (**1**) and nevirapine (marketed under the trade name viramune among others for the treatment and prevention HIV-1 infection) was found to possess an additive to weakly synergistic effect in blocking replication of HIV-1 in an in vitro tissue culture assay [41], indicating the possibility of using (+)-calanolide A (**1**) in anti-HIV-1 combination therapy. In an in vivo study using a hollow fibre mouse model [42], the synergistic potential of (+)-calanolide A (**1**) in combination therapy with AZT, a well-known anti-retroviral medication, was further established. A more comprehensive study on the anti-HIV activity of (+)-calanolide A (**1**) and its analogs, e.g., costatolide (**14**), dihydrocostatolide (**16**) and (+)-12-oxo-calanoldie A, in combination with other inhibitors of HIV-1 replication was published about a decade ago [37,46]. Calanolides were found to display synergistic antiviral interactions with other nucleoside and non-nucleoside reverse transcriptase inhibitors and protease inhibitors. In addition, additive interactions were also observed with calanolides when used with other anti-HIV drugs. It was concluded that the utility of convergent and divergent combination therapies using reverse transcriptase inhibitors and protease inhibitors in combination with (+)-calanolide A (**1**) or one of its analogues could be clinically relevant. Budihas et al. [47] demonstrated significant synergy between β-thujaplicinol and calanolide A (**1**).

#### 4.2.3. Structure-Activity-Relationships (SAR)

Among the naturally occurring calanolides, calanolide A (**1**) is one of the most potent anti-HIV compounds and has been the focus of various studies including the study of its possible mechanism of action, structural modifications, pharmacokinetics and toxicity [9,48–51]. The structures of naturally occurring calanolides mainly differ in their stereochemistry at various chiral centers (C-10, C-11 and C-12) on the ring D (Figures 1 and 2). McKee et al. [20] reported that calanolide-type compounds with a 12β hydroxyl group (as in compound **1**) generally possess anti-HIV activity. While calanolide A (**1**) and costatolide (**14**) were found to be active, (+)-calanolide C (**6**) was inactive in the in vitro anti-HIV assay [4,24]. The inactivity of (+)-calanolide C (**6**) despite possessing the pharmacophoric ring D, as well as a propyl group on C-4, could be due to the β-*cis* orientation of methyl groups on C-10 and C-11.

Like any other optically active drug molecules, optical activity plays an important role in the anti-HIV activity of calanolides. It has long been established that (+)-calanolide A (**1**) and (−)-calanolide B (**14**) are potent HIV-1 inhibitors, whilst (−)-calanolide A and (+)-calanolide B (**4**) are inactive against the virus [52]. It should be mentioned here that (+)-calanolide A (**1**) is the natural product, but its enantiomer (−)-calanolide A was prepared from the naturally occurring (−)-costatolide (**14**), isolated from *C. costatum*. Similarly, to establish structure-activity-relationships of calanolides, several analogs of calanolides have been synthesized to date, and tested in anti-HIV assays [53]. Although the synthesis of calanolides and the anti-HIV activity of synthetic calanolide analogs are not within the scope of this review, a few examples are given here in the context of structure-activity-relationships. One of the first attempts in this area was from Galinis et al. [53], where ∆ 7,8 olefinic bonds within (+)-calanolide A (**1**) and (−)-calanolide B (**14**) were reduced, and C-12 hydroxyl group in (−)-calanolide B (**14**) was modified to investigate variations in anti-HIV activity compared to parent calanolides. In this study, none of the 14 derivatives was found to possess superior activity to parent calanolides but revealed some preliminary structure-activity requirements for anti-HIV potencies. Later, in order to identify the structural features of naturally occurring (+)-calanolide A (**1**) necessary for its anti-HIV activity and to prepare synthetic analogues, oxo-derivatives (+)-, (−)- and (±)-12-oxocalanolides, were synthesized and tested in vitro using a biochemical reverse transcriptase inhibition assay for determining anti-HIV activity with a promising outcome [48]. In a review article covering various aspects of anti-HIV 4-substitued coumarins with an alkyl or a phenyl group as the substituent, isolated from the genus *Calophyllum*, summarized that all *trans* configurations (10*R*, 11*S*, 12 *S*), as in (+)-calanolide A (**1**) and (+)-inophyllum B (a 4-phenyl-substituted pyranocoumarin), are essential for the best anti-HIV activity [54].

Most of the SAR studies involving calanolides for their anti-HIV activities concentrated on the three chiral centers at C-10, C-11 and C-12 of (+)-calanolide A (**1**) [55,56]. As the number of naturally occurring calanolides are rather limited (calanolides A–F) (Figure 2), the SAR studies were often carried out with natural calanolides as well as their synthetic analogs. Of the diastereomers, compounds containing 10,11¬-*trans*-methylation and 12-(*S*)-OH chirality (Figure 2) displayed the most potent activity with EC<sup>50</sup> values in between 0.18 and 2.0 µM [55]. It was also observed that either the enantiomers (12-*R*-OH) or epimeric alcohols, e.g., calanolide C (**6**) could not produce any noticeable anti-HIV effect. It could be concluded that the relative stereochemistry at C-10 and C-11 are essential structural features for potent anti-HIV activity of calanolides, and at the same time, the *S* configuration at C-12 as well as the presence of a heteroatom, e.g., O, at C-12 are necessary for anti-HIV effects.

In order to assess the importance of the presence of 11-methyl functionality on calanolide A for its anti-HIV activity, the activity of the semi-synthetic racemic mixture of 11-demethyl-calanolide A was compared with the anti-HIV activity of its parent compound, (±)-calanolide A [57]. The in vitro HIV-1 reverse transcriptase inhibitory activity of these compounds was determined with the isotope 3H assay, which is a thymidine incorporation assay that often utilizes a strategy wherein a radioactive nucleoside, 3H-thymidine, is incorporated into new strands of chromosomal DNA during mitotic cell division; a scintillation beta-counter is used to measure the radioactivity in DNA recovered from the cells in order to determine the extent of cell division that has occurred in response to a test agent. The cytotoxicity and inhibition of cytopathic effect of (±)-calanolide A and (±)-11-demethyl-calanolide A were studied in HIV-1 IIIB infected MT-4 cell cultures by the MTT staining method. Both compounds inhibited HIV-1 reverse transcriptase in vitro with IC<sup>50</sup> value of 3.028 µM/L and 3.965 µM/L, respectively, for (±)-11-demethyl-calanolide and (±)-calanolide A. They also inhibited cytopathic effect in HIV-1 IIIB infected MT-4 cell cultures with IC<sup>50</sup> values of 1.081 and 1.297 µM/L, respectively. The outcome form this study indicated that (±)-11-demethyl-calanolide had a slightly more potent anti-HIV activity than (±)-calanolide A, suggesting the methyl functionality at C-11 in calanolide A (**1**) might not be an essential structural feature for anti-HIV activity. With the help of synthetic analogues a few other structural features that could impact on the anti-HIV activity of calanolides could be identified. Some of those are summarized below:


With the advent of various modern computational tools and mathematical models, it is now possible to study quantitative structure activity relationships (QSAR) in silico, and to predict the potential of any drug candidates for any therapeutic application [58]. A Caco-2 cell permeability QSAR model has recently been used to study various HIV-1 reverse transcriptase inhibitors, including (+)-calanolides A (**1**) and B (**4**), both of which showed a high degree of permeability [59]. This parallel computational screening method incorporated approaches of intestinal absorption prediction, receptor affinity estimation, inhibitor shape similarity, lipophilicity, and index-based lipophilic efficiency analyses. Calanolide A (**1**), among a few other HIV-1 reverse transcriptase inhibitors, emerged as one of the prioritized hits, as a result of guided prioritization task by the better binding affinity, crystal ligand similarity, permissible log*P* value and top lipophilic ligand efficiency scores.

#### 4.2.4. Mechanism of Action

The evaluation of the activity of (+)-calanolide A (**1**) against reverse transcriptase and nonnucleoside reverse transcriptase inhibitor-resistant viruses and enzyme kinetic studies for reverse transcriptase inhibition suggest that this coumarin possibly interacts with the HIV-1 reverse transcriptase in a fashion mechanistically different from other known NNTRIs. The biochemical mechanism of inhibition of HIV-1 reverse transcriptases by calanolide A (**1**) was studied using two primer systems, ribosomal RNA and homopolymeric rA-dT(12-18) [60]. Calanolide A (**1**) was found to bind near the active site of the enzyme and interfered with dNTP binding; it inhibited HIV-1 reverse transcriptase in a synergistic fashion with nevirapine, further distinguishing it from the general class of NNRTIs. It was also observed that at certain concentrations, this compound could bind HIV-1 reverse transcriptase in a mutually exclusive manner with respect to both the pyrophosphate analog, phosphonoformic acid and the acyclic nucleoside analogue 1-ethoxymethyl-5-ethyl-6-phenylthio-2-thiouracil. It was concluded that calanolide A (**1**) could share some binding domains with both phosphonoformic acid and 1-ethoxymethyl-5-ethyl-6-phenylthio-2-thiouracil. It might interact with reverse transcriptase near both the pyrophosphate binding site and the active site of the enzyme. Later, the same group of researchers studied possible mechanism of action of action of calanolide A (**1**) against the HIV type 1 including a variety of laboratory strains, with EC<sup>50</sup> values of 0.10–0.17 µM [60]. Calanolide (**1**) could inhibit promonocytotropic and lymphocytotropic isolates from patients in various stages of HIV disease, and drug-resistant strains, and was found to act early in the infection process like the known HIV reverse transcriptase inhibitor 2′ ,3′ -dideoxycytidine. It could selectively inhibit recombinant HIV type 1 reverse transcriptase but not cellular DNA polymerases or HIV type 2 reverse transcriptase. Auwerx et al. [50] studied the possible role of Thr139 in the HIV-1 reverse transcriptase sensitivity to (+)-calanolide A (**1**). As T139I reverse transcriptase proved to be resistant to (+)-calanolide A (**1**), represents a catalytically efficient enzyme, and requires only a single transition point mutation (ACA→ATA) in codon 139 could provide some explanation as to why mutant T139I reverse transcriptase virus strains, but not the other strains containing other amino acid changes at this position, predominantly emerge in cell cultures under (+)-calanolide A (**1**) pressure.

Calanolides are non-nucleoside reverse transcriptase inhibitors and mediate their inhibitory effect in two different template primer systems: primed ribosomal RNA template, and homopolymeric poly rA-oligoT12-18 primer. Calanolide A (**1**) was found to inhibit reverse transcriptase by involving two binding sites, and the action is because of the bi-bi ordered mechanism of reverse transcriptase, requiring primer binding prior to polymerization [55]. Calanolide A (**1**) can bind HIV-1 reverse transcriptase in a mutually exclusive manner with the pyrophosphate analogues phosphoformic acid or 1-ethoxymethyl-5-ethyl-6-phenylthio-2-thiouracil. This indicates that calanolide A (**1**) can interact with reverse transcriptase near the pyrophosphate binding site as well as the active site. Unlike general non-nucleoside reverse transcriptase inhibitors, calanolide A (**1**) appears to be at least partially competitive inhibitor of dNTP binding. Clinical and laboratory assessment on viral load and CD4 count indicated that antiviral effects of calanolide A (**1**) appeared to be dose-dependent and maximized on day 14 or 16. Viral life-cycle studies indicated that calanolide A (**1**) could act early in the infection process, similar to the known HIV reverse transcriptase inhibitor 2′ ,3′ -dideoxycytidine. In enzyme inhibition assays, calanolide A (**1**) could potently and selectively inhibit recombinant HIV type 1 reverse transcriptase but not cellular DNA polymerases or HIV type 2 reverse transcriptase within the concentration range tested.

#### *4.3. Antimycobacterial Activity*

The antibacterial (against *Bacillus cereus, B. pumilius, B. subtilis*, *Escherichia coli*, *Pseudomonas aeruginosa*, *Salmonella typhi*, *Staphylocossus aureus and Vibrio cholerae*) and antifungal (against *Alternaria tenuissima*, *Aspergillus fumigatus*, *Aspergillus niger*, *Candida albicans and Candida tropicalis*) properties of *Calophyllum* species and their bioactive secondary metabolites, including calanolides, are already known [6,15,61–67]. Kudera et al. [66] reported in vitro growth inhibitory activity of *C. inophyllum* extract against diarrhea-causing microorganisms, e.g.,*Clostridium difficile infant, Clostridium perfringens, Enterococcus faecalis, Escherichia coli, Listeria monocytogenes* and *Salmonella enterica*. The extract was particularly active against *C. perfringens* and *L. monocytogenes* (MIC = 128 µg/mL). Later, calanolide E (**10**) was isolated from *C. wallichianum* and tested for its anti-*Bacillus* activity against *Bacillus cereus, B. megaterium, B. pumilus* and *B. subtilis* [20]. However, calanolide E (**10**) was not bactericidal on the tested Bacillus species, and at the tested concentration.

Based on the initial findings on promising antimicrobial properties of calanolides and *Calophyllum* extracts, efforts have recently been directed to the study on the effect of these compounds on the acid-fast bacillus *Mycobacterium tuberculosis*that causes tuberculosis [17,68,69]. As over the years several antibiotic resistant and multidrug-resistant *M. tuberculosis* strains have emerged, and complicated the existing treatment modalities for tuberculosis, and there has been a recent increase in incidents of tuberculosis globally observed, the need for new effective, safe and affordable antimycobacterial drugs has become paramount. *Calophyllum brasiliense* extract was reported to be active against *M. tuberculosis* (IC<sup>50</sup> 3.02–3.64 µg/mL), and a follow up HPLC analysis of the active extract provided evidence of presence of calanolides and the antimycobacterial activity induced by *C. brasilliense* was attributed mainly to calanolides A (**1**) and B (**4**) [17]. Earlier, Xu et al. [68] demonstrated that calanolide A (**1**), from Colombian *C. lanigerum*, was active against both drug-susceptible and drug-resistant strains of *Mycobacterium tuberculosis,* e.g., H37Ra (ATCC 25177), H37Rv (ATCC 27294), CSU 19, CSU 33, H37Rv-INH-R (ATCC 35822), CSU 36, CSU 38 and H37Rv-EMB-R (ATCC 35837). Efficacy evaluations in macrophages established that this pyranocoumarin could inhibit intracellular replication of *M. tuberculosis* at concentrations below the minimum inhibitory concentration (MIC) determined in vitro. It was postulated that calanolide A (**1**), like the antitubercular drug rifampicin, could rapidly inhibit RNA and DNA synthesis followed by an inhibition of protein synthesis, and could lead to the generation of a new class of pyranocoumarin-based antitubercular drugs. In this study, the natural calanolides A (**1**), B (**4**) and D (**7**), as well as their semisynthetic analogues were tested, and (+)-calanolide A (**1**) and the semisynthetic analogue, 7,8-dihydrocalanolide B emerged as most effective against tuberculosis with the MIC value of 3.13 µg/mL. While (–)-calanolide B (**14**) was moderately effective, calanolide D (**7**) was found inactive at the highest tested concentration of 12.5 µg/mL. In fact, calanolides, especially calanolide A (**1**), is unique in a sense that these compounds have anti-HIV property and were found to be active against *M. tuberculosis* (MIC = 3.1 µg/mL) and an array of drug-resistant strains (MIC = 8–16 µg/mL). The antimycobacterial activity of calanolide A (**1**) is comparable to that of the well-known anti-tubercular drug isoniazid, and effective against rifampicin- and streptomycin-resistant *M. tuberculosis* strains. A recent patent described potent antimycobacterial property of calanolides and their analogs and provided a method of using these compounds for the treatment and prevention of mycobacterial infections [70].

#### *4.4. Antiparasitic Activity*

Traditionally, natural products, especially in crude forms, have long been used to treat various parasitic diseases, like babesiosis, leishmaniasis, malaria, trypanosomiasis and so on. Recently, leishmaniasis and trypanosomiasis have been in research focus of natural products researchers, aiming at discovering new drug candidates to treat these neglected diseases [71,72]. Extracts of *C. brasiliense* and *C. inophyllum* and calanolides were shown effective against intracellular parasites causing American trypanosomiasis and leishmaniasis [6]. In a recent study, Silva et al. (2020) [14] demonstrated that the MeOH extract from stem bark of *C. brasiliense* was active against amastigote forms of *Trypanosoma cruzi* and *Leishmania infantum*. Bioactivity-guided purification of the extra afforded calanolides E1 (**11**) and E2 (**12**), which were found to be active against *T. cruzi* (EC<sup>50</sup> values of 12.1 and 8.2 µM, respectively) and *L. infantum*, (EC<sup>50</sup> values of 37.1 and 29.1 µM, respectively) in vitro. Calanolide E1 (**11**) displayed the best selectivity index (SI) with values >24.4 to *T. cruzi* and >6.9 to *L. infantum* in comparison to calanolide E2 (**12**). It was concluded that these coumarins could be utilized as scaffolds for the design and development of novel drug candidates to treat Leishmaniasis and Chagas diseases.

#### **5. Toxicological Aspects Including Pharmacokinetics**

Among the naturally occurring calanolides (Figure 2), calanolide A (**1**), a specific nonnucleoside inhibitor of human immunodeficiency virus type 1 (HIV-1) reverse transcriptase, first isolated from a tropical tree *C. lanigerum* that grows abundantly in the Malaysian rain forest, is the most-studied compound in terms of its pharmacology, toxicology and synthesis. A series of animal studies [43] involving mice, rats and dogs established that calanolide A (**1**) is generally well-tolerated at oral doses of up to 150 mg/kg in rats and 100 mg/kg in dogs, and possesses a good safety profile [73,74]. Calenolides A (**1**), B (**4**) and C (**6**) were found to be nontoxic in mice (LD<sup>50</sup> = 1.99 g/kg), and no alternation on hepatocytes could be observed during the histological study of the mice treated with the highest dose applied [74]. During a study looking at the anti-HIV efficacy and toxicity of calanolides when used in combination with other anti-HIV drugs, no noticeable toxicity could be detected [46].

In the very first study on the safety and pharmacokinetics of calanolide A (**1**) in healthy HIV-negative human volunteers revealed that the toxicity of calanolide A (**1**) was minimal in the majority of subjects treated with four successive single dose, 200, 400, 600 and 800 mg. While there were no acute serious or life-threatening adverse effects were observed, among the usual minor adverse effects, dizziness, oily taste, headache, eructation, and nausea were noticed, but were of minimal clinical significance. These adverse effects were non-dose-dependent [73]. In this study, it was found that calanolide A (**1**) was rapidly absorbed following administration, with time to maximum concentration of drug in plasma (*T*max) values, depending on the doses, occurring between 2.4 and 5.2 h. It was noted that the levels of calanolide A (**1**) in human plasma were higher than would have been predicted from animal studies, but the safety profile was benign. However, taking calanolide A (**1**) with food was found to generate significant variability in pharmacokinetics, but with no detectable interaction with food. Later, these findings were further confirmed by another similar study carried out by Eiznhamer et al. [75]. Calanolide A (**1**), the first member of the new family of NNRTIs, was found to have long elimination half-life, the benign toxicity profile, to achieve trough plasma levels approximating the EC<sup>90</sup> of calanolide A (**1**) for HIV-1, to have the potential for twice daily dosing, and to offer the unique HIV-1 resistance profile could make this compound an attractive candidate for further clinical studies. It was reported that after oral administration, (+)-calanolide A (**1**) was generally well tolerated and indication of any safety concern could be observed [48]. Its plasma concentrations in humans were higher than anticipated from animal data. The AUC and Cmax values increased with increasing dose, and it appeared that therapeutic levels could easily be achieved in humans.

A comparative study on the relative pharamacokinetic parameters and bioavailability of calanolide A (**1**) and its synthetic analogue dihydrocalanolide A (**15**) was reported [76]. This study compared the intravenous pharmacokinetics of the dihydro analog relative to the parent compound, calanolide A (**1**), and determined the relative oral bioavailability of each drug in CD2F1 mice. Both compounds displayed similar pharmacokinetic parameters, but the oral bioavailability of the dihydro analogue was considerably better (almost 3.5-fold) than calanolide A (**1**). Moreover, the relative ability of calanolide A (**1**) and its dihydro analog to change to their inactive epimer forms, (+)-calanolide B (**4**) and (+)-dihydrocalanolide B, respectively, was also determined; while conversion of active calanolides to inactive forms occurred in vitro especially under acidic conditions, no epimers of either compound were observed in plasma of mice after administration of either (+)-calanolide A (**1**) or (+)-dihydrocalanolide A (**15**). It was suggested that the selection of the dihydro derivative of calanolide A (**1**) could be a reasonable choice for further preclinical development and possible Phase I clinical evaluation as an oral drug candidate for the treatment of HIV infection. Calanolide A (**1**) was shown to be distributed readily into the brain and lymph [55]. The distribution and elimination pattern of calanolide A (**1**) and its 7,8-dihydro derivative were found to be similar, but the apparent volume of distribution (Vd) and oral clearance of these compounds were significantly different after oral administration. It was also demonstrated that calanolide A (**1**) is generally well tolerated in doses up to 600 mg. As evident from animal studies, the gastrointestinal intolerance for this compound is not severe, but the most common adverse events as observed in human trails of calanolide A (**1**) include an oily after taste and transient dizziness [55]. The calculated half-life of calanolide A (**1**) from 800 mg dosing was reported to be 20 h [55,73].

During the study directed to the evaluation of antitubercular property of calanolides and their semisynthetic analogues, the pharmacokinetic data indicated that the (+)-calanolide A (**1**) concentrations in plasma could be comparable to the observed in vitro MICs against *M. tuberculosis* [68]. Both calanolides A (**1**) and B (**4**) metabolized by cytochrome P450 CYP3A, and their blood levels could be enhanced if co-administered with ritonavir. Usach et al. [77] reported the safety, tolerability and pharmacokinetics profiles of calanolide A (**1**), as a result of a comprehensive Phase I clinical trial.

#### **6. Therapeutic Potential**

Naturally occurring calanolides and their synthetic or semi-synthetic analogs have undergone several pre-clinical and clinical trials for their anti-HIV activity, aiming at novel anti-HIV drug development [2,16,55,78]. In fact, calanolide A (**1**) was at an advanced stage of development as an anti-HIV drug about a decade ago [78]. Buckheit [79] reviewed therapeutic potential of non-nucleoside reverse transcriptase inhibitors like calanolides as anti-HIV and commented on strategies for the treatment modalities for HIV infections. In fact, NNRTIs opened a new avenue of treatment of HIV infections, as previously this therapeutic area was predominantly covered by nucleoside reverse transcriptase inhibitors and protease inhibitors. Soon after the discovery of calanolides as a potential ant-HIV agents by the NCI/NIH, Sarawak Medichem Pharmaceuticals (Sarawak, Malaysia) synthesized calanolide A (**1**) and started developing calanolide A (**1**) as a clinical drug for the treatment of HIV infections. It was a joint venture between the Sarawak State Government and Medichem Research Inc.

During 2001–2005, an interventional clinical trial was conducted on human volunteers [80], where patients were randomized to receive (+)-calanolide A (**1**) or placebo for 21 days. All patients could elect to receive an open-label, 3-month course of approved retroviral therapy (up to triple-drug therapy) to be selected by, and administered under the care of, the patients' physicians. If the patient had no insurance coverage or did not wish to utilize his/her insurance for anti-HIV medications, Sarawak MediChem Pharmaceuticals provided these medications at no charge for up to three months. The trial was primarily aimed at the assessment of the safety and effectiveness of (+)-calanolide A (**1**) in HIV-infected patients who had never taken anti-HIV drugs. In 2006, Craun Research (Kuching, Malaysia), a company established by the Sarawak Government, acquired Sarawak MediChem, and in 2016, Craun Research announced the completion of Phase I clinical trials for calanolide A (**1**) with doses of 200 to 800 mg, which initially started in 2013 [77]. In 2017, F18 (10-chloromethyl-11-demethyl-12-oxo-calanolide A), a synthetic structural analog of calanolide A (**1**) was shown to have more potent anti-HIV activity than original molecule, calanolide A (**1**) [81,82]. This compound showed better druggable profile with 32.7% oral bioavailability in rat, tolerable oral single-dose toxicity in mice, and suppressed both the wild type HIV-1 and Y181C mutant HIV-1 at an EC<sup>50</sup> of 7.4 nM and 0.46 nM, respectively [83]. Furthermore, it was shown that two enantiomers F18, (*R*)-F18 and (*S*)-F18, had quite similar anti-HIV property, but (*R*)-F18 was more potent than (*S*)-F18 against wild type virus, K101E mutation and P225H mutation pseudoviruses [81]. However, calanolides, particularly calanolide A (**1**) remains as an investigational anti-HIV drug and has not yet been approved by the FDA or any other drugs regulatory bodies for their commercial pharmaceutical production.

#### **7. Patents**

In 1999, calanolides and related antiviral compounds were patented by the Board of Trustees of the University of Illinois [84]. The patent covered novel antiviral compounds, calanolides, and their derivatives that could be isolated from plants of the genus *Calophyllum* in accordance with the specified method. The patent also included the uses of these compounds and their derivatives alone or in combination with other antiviral agents in compositions, such as pharmaceutical compositions, to inhibit the growth or replication of a virus, such as a retrovirus, in particular a human immunodeficiency virus, specifically HIV-1 or HIV-2. Later, another patent, owned by Parker Hughes Institute, was reported, which described the novel uses of calanolides as Tec family/BTK (Bruton's tyrosine kinase) inhibitors, methods for their identification, and pharmaceutical compositions [85]. It can be mentioned here that the BTK inhibitors inhibit the enzyme BTK, which is a crucial part of the B-cell receptor signaling

pathway, and these inhibitors have emerged as a new therapeutic target in a variety of malignancies, e.g., chronic lymphocytic leukemia and small lymphocytic lymphoma [86].

#### **8. Conclusions**

Non-nucleoside reverse transcriptase inhibitors (NNRTIs), efavirenz, nevirapine and delavirdine, have become one of the cornerstones of highly active anti-retroviral therapy for HIV infections. Calanolides, as they belong to this pharmacological class of NNRTIs, and because of their high safety margins and favorable pharmacokinetic profiles, are ideal candidates for novel anti-HIV drug development. While several analogues of the naturally occurring calanolides have been synthesized, a good number of preclinical and clinical trials have been conducted to date, and there are a few patents published, further work is still required to commercially bring any of the calanolide candidates, natural or synthetic, to anti-HIV drug market. As calanolides show an excellent synergistic and additive profile in combination with other anti-HIV drugs, it is assumed that calanolides can be considered for use in combination therapy for HIV infections.

**Author Contributions:** All authors contributed equally to the data collection and compilation of information. Additionally, L.N. and S.D.S. played the lead role in writing, editing and submission of this manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the European Regional Development Fund—Project ENOCH (No. CZ. 02.1.01/0.0/0.0/16\_019/0000868).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Review* **Chalepin and Chalepensin: Occurrence, Biosynthesis and Therapeutic Potential**

**Lutfun Nahar 1,\* , Shaymaa Al-Majmaie <sup>2</sup> , Afaf Al-Groshi <sup>2</sup> , Azhar Rasul <sup>3</sup> and Satyajit D. Sarker 2,\***


**Abstract:** Dihydrofuranocoumarin, chalepin (**1**) and furanocoumarin, chalepensin (**2**) are 3-prenylated bioactive coumarins, first isolated from the well-known medicinal plant *Ruta chalepensis* L. (Fam: Rutaceae) but also distributed in various species of the genera *Boenminghausenia*, *Clausena* and *Ruta*. The distribution of these compounds appears to be restricted to the plants of the family Rutaceae. To date, there have been a considerable number of bioactivity studies performed on coumarins **1** and **2**, which include their anticancer, antidiabetic, antifertility, antimicrobial, antiplatelet aggregation, antiprotozoal, antiviral and calcium antagonistic properties. This review article presents a critical appraisal of publications on bioactivity of these 3-prenylated coumarins in the light of their feasibility as novel therapeutic agents and investigate their natural distribution in the plant kingdom, as well as a plausible biosynthetic route.

**Keywords:** *Ruta chalepensis*; Rutaceae; chalepin; chalepensin; bioactivity; biosynthesis

#### **1. Introduction**

Chalepin (**1**; mol formula: C19H22O4; mol weight 314) and chalepensin (**2**; mol formula: C16H14O3; mol weight 254) (Figure 1) are, respectively, a dihydrofuranocoumarin and a furanocoumarin, with a prenylation at C-3 of the coumarin core structure. These coumarins, as the names imply, were first isolated from *Ruta chalepensis* L. (Fam: Rutaceae), but are also found in other *Ruta* species, e.g., *R. angustifolia* and a few other plants of the genus *Clausena* (Fam: Rutaceae), e.g., *Clausena anisata* (Willd.) Hook. F. ex Benth. [1–4]. While chalepin (**1**), also known as heliettin, is optically active, chalepensin (**2**), also known as xylotenin, does not possess any optical activity. Although these coumarins are rather rare in the sense that there are not many 3-prenylated naturally occurring furanocoumarins reported to date, there are quite a good number of bioactivity studies carried out on these compounds. The present review critically appraises publications on bioactivity of these 3-prenylated furanocoumarins in the light of their feasibility as novel therapeutic agents and covers their natural distribution in the plant kingdom, as well as a plausible biosynthetic route.

−

−

−

−

−

− −

− −

**Citation:** Nahar, L.; Al-Majmaie, S.; Al-Groshi, A.; Rasul, A.; Sarker, S.D. Chalepin and Chalepensin: Occurrence, Biosynthesis and Therapeutic Potential. *Molecules* **2021**, *26*, 1609. https://doi.org/10.3390/ molecules26061609

Academic Editor: Maria João Matos

Received: 24 February 2021 Accepted: 12 March 2021 Published: 14 March 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

#### **2. Distribution**

First isolated from *Ruta chalepensis* more than half a century ago, chalepin (**1**) and chalepensin (**2**) have been further reported mainly from various species of the genera *Clausena* and *Ruta* of the family Rutaceae [4,5]. It appears that these compounds exclusively occur in the family Rutaceae [1–20], and predominantly within these two genera. However, *Boenminghausenia albiflora* var. *japonica (Hook.)* Rchb. Ex Meisn and *B. sessilicarpa* H. Lev. also produce chalepensin (**2**) [6,20] and this genus is phylogenetically close to the genus *Ruta* [21]. Chalepensin (**2**) was further found in the leaves of *Esenbeckia alata* (Karst and Triana) Tr. and Pl. [9], while *E. grandiflora* Mart. was reported to produce chalepin (**1**) [10]. Interestingly, the genus *Esenbeckia* Kunth. is a part of a small group of phylogenetically distant Rutaceae including the genera *Clausena* and *Ruta,* where 3-prenylated coumarins like **1** and **2** are generally produced [9]. Thus, co-occurrence of these 3-prenylated furanocoumarins in these genera might have some chemotaxonomic implications, at least at the family level, within the family Rutaceae. The distribution of these two coumarins (**1** and **2**) is summarized in Table 1. Within the source plants these compounds are well distributed almost in all parts, leaves, stem, flowers and fruits. Although not chalepensin (**2**) itself, a series of 5-*O*-prenylated chalepensin derivatives were reported from *Dorstenia foetida* Schweinf., a medicinal plant from the family Moraceae, distributed in various countries in the Middle-East Asia [22].

**Table 1.** Distribution of chalepin (**1**) and chalepensin (**2**) in the plant kingdom.


+ = Found; − = Not found.

#### **3. Biosynthesis**

Like all other coumarins, the biosynthesis of chalepin (**1**) and chalepensin (**2**) begins from the simple coumarin umbelliferone, which is formed from the amino acid L-phenylalanine through the formation of *trans*-cinnamic acid, *p*-coumaric acid, 2-hydroxy*p*-coumaric acid, 2-glucosyloxy-*p*-coumaric acid and 2-glucosyloxy-*p*-*cis*-coumaric acid aided by different enzymes, e.g., cinnamate 4-hydroxylase and 4-coumarate-CoA ligase, 4-coumaroyl 2′ -hydroxylase (Figure 2) [23,24]. Sharma et al. [25] studied the biosynthesis of chalepin (**1**) in *Ruta graveolens*. They suggested that 3-(1,1-dimethylallyl)-umbelliferone could be the key intermediate for the biosynthesis of chalepin (**1**), and the dihydrofuran moiety in chalepin (**1**) is formed via prenylation, aided by dimethylallyldiphosphate, at C-6 of the core coumarin skeleton followed by oxidative cyclization with neighboring hydroxyl function at C-7. Generally, prenyltransferases (6-prenyltransferase was identified in *R. graveolens* as a plastidic enzyme) are considered the enzymes involved in the biosynthesis of furano-/dihydrofuranocoumarins through umbelliferone prenylation. Further oxidation of chalepin (**1**) could lead to the formation of the furanocoumarin chalepensin (**2**) in a similar fashion as observed in the conversion of marmesin to psoralen [26]. In fact, biosynthesis of chalepin (**1**) resembles that of 3-prenylated furanocoumarin, rutamarin

(acetyl-chalepin) [26]. At this moment, it is not clear from the literature if the prenylation at C-3 takes precedence over that on C-6. In fact, the published information on the biosynthesis of these coumarins **1** and **2** is rather extremely limited, and much work, especially using radioisotopes is much needed to explore other possible routes to the biosynthesis of these compounds.

### A = 3-Prenylation; B = 6-Prenylation; C = Oxidative cyclization; D = Oxidation

**Figure 2.** Putative biosynthetic route for the formation of chalepin (**1**) and chalepensin (**2**).

#### **4. Bioactivity**

′

The general, *Clausena* and *Ruta*, the main sources of chalepin (**2**) and chalepensin (**2**), are well known for their uses in traditional medicines, and different studies have established their bioactivities [27,28]. Chalepin (**1**) and chalepensin (**2**) have emerged as two major bioactive components in many of those plants through bioassay-guided isolation protocols, and their bioactivities include antimicrobial, anti-inflammatory, anticancer, antiviral and many more. In this section, using several subsections, a critical appraisal is presented on bioactivities of these two compounds (**1** and **2**) reported in the literature to date (Table 2) [29–51]. Most of the reported bioactivity studies on these compounds involved predominantly in vitro assays and only a handful of in vivo and in silico studies. However, there is no report on any systematic preclinical or clinical trial with these compounds involving human volunteers available in the literature to date.



NR = No report available; + = Active.

#### *4.1. Antidiabetic Activity*

Among the bioactive compounds isolated from the stem bark of *Clausena lansium* (Lour.) Skeels, chalepin (**1**) exhibited antidiabetic properties, exerted through dose-dependent stimulated (glucose-mediated) insulin release in vitro from INS-1 cells (rat insulinoma cell line) [8]. Chalepin (**1**) showed 138% insulin secretory response in vitro at the concentration of 0.1 mg/mL. INS-1 cells are widely used as rat islet β-cell models for screening for antidiabetic properties of plant extracts or purified compounds. They express muscarinic M1 and M3 receptors, which are activated by carbachol to promote insulin release. Chalepensin (**2**) does not appear to have gone through any antidiabetic screening yet. It is known that insulin secretion involves a sequence of events in β-cells that lead to fusion of secretory granules with the plasma membrane; it is secreted primarily in response to glucose, while other nutrients such as free fatty acids and amino acids can augment glucose-induced insulin secretion.

#### *4.2. Antifertility Activity*

During the assessment of the extracts of *R. chalepensis* var *latifolia* for antifertility activity in rodents, chalepin (**1**) and chalepensin (**2**) were discovered as the major active antifertility principles in the extracts [13]. Despite these compounds showing antifertility activity, most of the tested animals developed cystic and atretic follicles in their ovaries and glomerulocapsular adhesion and segmental fusion in the kidneys. However, no brain toxicity was observed with these compounds. Kong et al. [29] assessed the antifertility activity of the chloroform extracts of the roots, stem and leaves of *R. graveolens* L. in rats and fractionation of the extracts afforded coumarin **2** as the active component with moderate toxicity. Time-dosing experiments showed that this coumarin (**2**) could act at the early stages of pregnancy. The observed antifertility activity of **1** and **2** [13,29] could provide some scientific evidence in support of the traditional uses of *R. chalepensis* as an abortifacient.

#### *4.3. Antimicrobial Property*

Antimicrobial assay-guided analysis of a root extract of *Clausena anisata* (Willd.) Hook. f. Benth., a well-known medicinal plant used traditionally for the treatment of parasitic infections, influenza, abdominal pain and constipation, afforded chalepin (**1**) as an antibacterial agent, particularly effective against *Bacillus subtilis* with a zone of inhibition of 16 mm as opposed to 15 mm of the positive control cifrofloxacin [4]. This coumarin was also found active against two other pathogenic bacterial strains, *Staphylococcus aureus* and *Pseudomonas aeruginosa*. Chalepensin (**2**), on the other hand, was reported to possess antifungal property and was found to inhibit the growth of the fungal strains, *Candida albicans* and *Cryptococcus neoformans* [30]. However, interestingly, none of these coumarins showed any antimicrobial activity at tested concentrations (50-100 µg/mL) against a range of microorganisms, e.g., *Bacillus subtilis*, *Mycobacterium smegmatis*, *Staphylococcus aureus* and *Candida albicans*, using a modified microtitre-plate assay as reported by El Sayed et al. [52]. Chalepensin (**2**), isolated from *R. chalepensis*, was assessed for antibacterial activity against *Streptococcus mutans* using the method of colony forming units counts in solid medium culture and reduction of tetrazolium salt MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] in liquid medium [31] and was shown to significantly inhibit the growth of this bacterial strain with an MIC (minimum inhibitory concentration) of 7.8 µg/mL.

In the most recent study [32] on anti-MRSA (methicillin-resistant *Staphylococcus aureus*) activity of several compounds, mainly coumarins and flavonoids, isolated from *R. chalepensis* grown in Iraq, both chalepin (**2**) and chalepensin (**3**) showed significant antimicrobial activities against the MRSA strains, ATCC 25923, SA-1199B, XU212, MRSA-274819 and EMRSA-15 with MIC values ranging between 32 and 128 µg/mL. In that study, two other furanocoumarins, bergapten and isopimpineline, which do not have a 3-prenylation as in **1** and **2**, were found inactive at tested concentrations. Based on this finding, it was suggested that the prenylation at C-3 of the coumarin nucleus might be a key determi-

nant of anti-MRSA activity. Chalepensin (**2**) was found to be more active than chalepin (**1**) and was subjected to in silico studies to gain an insight into the extent at which this compound (**2**) is able to bind to MRSA proteins and also their drug−like physicochemical characters. In silico studies on compound **2** showed that this compound could have high GI absorption and no violation of the Lipinski rules. It was also shown that chalepensin (**2**) could bind with certain MRSA protein targets, predominantly through hydrogen bonding as well as van de Waals forces. It was suggested that this coumarin could be utilized as a structural template for generating structural analogs and developing potential anti-MRSA therapeutic agents.

#### *4.4. Antiprotozoal Activity*

One of the major traditional medicinal uses of *R. chalepensis* and other *Ruta* species is their efficacy as antiparasitic agents [28], particularly as an anthelmintic medication. This traditional medicinal use of *R. chalepensis* has prompted antiparasitic activity screening of its extracts and isolated major compounds, including chalepin (**1**) and chalepensin (**2**). Antiprotozoal activity of chalepin (**1**) and chalepensin (**2**), obtained from *R. chalepensis* following a bioassay-guided protocol, against *Entamoeba histolytica*, which is a causative organism of ameoebiasis, was reported as a meeting presentation, but no further full scientific report was published [33]. Both coumarins showed >90% growth inhibition against *E. histolytica*, an anaerobic parasitic amoebozoan, at a concentration of 150 µg/mL with IC<sup>50</sup> values of 28.67 and 38.71 µg/mL, respectively, for compounds **1** and **2**. However, in a previous study [5], conducted by the same group, evaluated antiprotozoal activity of plants used in northwest Mexican traditional medicine, particularly *Lippia graveolens* Kunth. and *R. chalepensis*, against *E. histolytica,* and chalepensin (**2**) was found to be the main antiprotozoal component in *R. chalepensis*. Earlier, Kundu and Roy [34] carried out in silico studies involving chalepin (**1**) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) of the pathogenic protozoa *E. histolytica.* It can be noted that GAPDH is a major glycolytic enzyme (~37 kDA), which catalyzes the sixth step of glycolysis, and an attractive drug target like *E. histolytica* lacks a functional citric acid cycle and exclusively depends on glycolysis for its energy needs. Chalepin (**1**) was predicted as a GAPDH inhibitor and structural modifications offering additional polar interactions were suggested to improve potency.

*Trypanosoma cruzi*, a species of parasitic euglenoids, characteristically can bore tissue in another organism and feed on blood and lymph, causing diseases like Chagas disease (also known as American trypanosomiasis) in humans, that affects more than 7 million people worldwide, with Latin American countries being most affected. In recent years, a renewed interest has been observed in the search for antitrypanosomal natural products, especially from higher plants. In many in vitro as well as in silico studies, glycosomal glyceraldehyde-3-phosphate dehydrogenase (gGAPDH) from *T. cruzi* has been used as a target molecule for screening compounds for potential antitrypanosomal activity [35]. In an in silico study with various natural products, chalepin (**1**) emerged as a hit molecule for antitrypanosomal drug discovery [36], and subsequently, a series of 3-piperonylcoumarins were synthesized and tested for their inhibitory activity against gGAPDH. Chalepin (**1**) was shown in silico to possess the highest binding affinity to gGAPDH (IC<sup>50</sup> = 55.5 µM) among the natural coumarins screened and the best inhibitor of gGAPDH [36,37]. Earlier, during an in vitro screening of natural coumarins for trypanocidal or antitrypanosomal activity, chalepin (**1**) was found to be the most active coumarin with an IC<sup>50</sup> value of 64 µM [38]. However, to the best of our knowledge, there is no report available to date on antitrypanosomal property of chalepensin (**2**).

#### *4.5. Antiviral Activity*

Chalepin (**1**), isolated from *R. graveolens*, along with its 28 synthetic analogs were tested for their inhibitory activity on the Epstein–Barr virus (EBV, also known as human herpes virus 4) lytic replication activity [41]. It was noted that most of the synthesized

analogs were more active than their parent or precursor, (-)-chalepin (**1**). EVP is a human gamma-herpes virus that infects more than 90% of the human population globally, and preferentially infects B lymphocytes and epithelial cells causing various diseases like Hodgkin's disease, Burkitt's lymphoma, nasopharyngeal carcinoma and gastric carcinoma in humans. Thus, inhibition of EBV lytic replication is considered as one of the pragmatic strategies for the treatment of some these diseases.

Chalepin (**1**), isolated from the leaves of *R. angustifolia*, displayed significant inhibitory activity (IC<sup>50</sup> = 1.7 µg/mL) against hepatitis C virus replication and was found to be more potent than the positive control ribavirin (IC<sup>50</sup> = 2.8 µg/mL), a well-known antiviral drug used for the treatment of hepatitis C and other viral diseases [39]. In continuation of their study, they have recently reported enhancement of antihepatitis C virus activity of chalepin (**1**) in combination with conventional antiviral drugs including cyclosporine A, daclatasvir, ribavirin, simeprevir and telaprevir [40]. It was found that chalepin (**1**) could enhance antihepatitis C activities of these conventional drugs with a synergistic combination index of <1. It could be considered as an excellent finding as the need for new and effective drugs for treating hepatitis C is of paramount importance. It can be mentioned that hepatitis C virus infects around 71 million people globally, causes severe liver disease, e.g., liver cancer and deaths; the WHO (World Health Organization) estimated that in 2016, about 400,000 people died from hepatitis C, mainly from liver cirrhosis and liver cancer [53,54].

Like many other antiviral coumarins including some 3-substitued ones, it can be assumed that chalepin (**1**) might offer antiviral activity through inhibition of various proteins that are involved in the transcription/translation processes essential for viral life cycle at different stages, and via modulation of host cell signaling, NF-kB (nuclear factor κB), and inflammatory redox-sensitive pathways and thus blocking viral replication [54]. However, clearly, further research is necessary to understand and establish definite mode of antiviral action mechanism of chalepin (**1**). However, there is no data available on any antiviral property of chalepensin (**2**) to date.

#### *4.6. Cytotoxicity (Potential Anticancer and Antitumor Activity)*

Cancer is one of the major causes of human mortality and morbidity. Currently available cancer treatment options or modalities are rather limited, and often suffer from severe side effects. Therefore, the search for new, effective, safe and affordable anticancer drugs is a part of many major modern drug discovery initiatives worldwide. Natural products have long been considered one of the major contributors in the continuing search for new anticancer molecules for safer and more effective anticancer drug development, and evidently, have already provided several successful anticancer drugs, e.g., taxol, vincristine and vinblastine [55]. The most common starting point in the search for anticancer molecules is the screening compounds for cytotoxicity against various human cancer cell lines because cytotoxicity is regarded as one of the major characteristics of anticancer agents. In order to assess anticancer potential of chalepin (**1**) and chalepensin (**2**), cytotoxicity of these compounds has been assessed against different human cancer cell lines in vitro, and some mechanistic studies on how they kill the cancer cells have also been published, showing anticancer and antitumor potential of these compounds (Table 2).

Chalepin (**1**), isolated from *Clausena emarginata* C. C. Huang, has been found to possess significant cytotoxicity against five human cancer cell lines including human leukemia (HL-60), hepatocarcinoma (SMMC-7721), lung carcinoma (A-549), breast cancer (MCF-7) and colon adenocarcinoma (SW-480) with IC<sup>50</sup> values comparable to that of the positive control, doxorubicin [7]. Chalepin (**1**), isolated from *Ruta angustifolia* Pers., was demonstrated to induce apoptosis through phosphatidylserine externalizations and DNA fragmentation in breast cancer cell line, MCF-7 [12,42]; this compound was considerably cytotoxic to MCF-7 cells, moderately cytotoxic to the epithelial human breast cancer cells (MDA-MB231), but not cytotoxic to normal cells, MRC-5 (Medical Research Council cell strain 5) in the SRB (sulforhodamine B) assay [56]. MRC-5 is a diploid cell culture line comprising fibroblasts, first developed from the lung tissue of a 14-week-old aborted Caucasian male fetus. It can be

mentioned here that apoptosis is a process by which cell commit suicide and is eliminated from the system; induction of apoptosis, a cell toxicity pathway, is considered as one of the early-stage mechanism for compounds to exert anticancer activity. This differential cytotoxicity against cancer cells and noncancerous cells might make this compound an ideal candidate, or at least a structural template, for anticancer drug development.

Earlier, in order to understand how chalepin (**1**) could exert its anticancer potential, a study conducted by Richardson et al. [11], revealed that this compound could dosedependently exhibit cell cycle arrest at S phase, suppress nuclear factor kappa B (NFκB) pathway, signal transducer and activation of transcription 3 phosphorylation and extrinsic apoptotic pathway in human non-small cell lung cancer cell line A-549. Cell cycle analysis using the flow cytometry confirmed that chalepin (**1**) could inhibit cell cycle at S phase (synthesis phase), which is the phase of the cell cycle, where DNA is replicated and occurs between the G<sup>1</sup> and G<sup>2</sup> phases. Since accurate duplication of the genome is essential for successful cell division to take place, the processes involved in the S phase are tightly regulated and widely conserved. A significant accumulation of cells in the S phase was observed after chalepin (**1**) treatment (45 µg/mL) for 48 (accumulation 27.7%) and 72 h (accumulation 25.4%), whereas the accumulation was only about 4% for the untreated cells [11]. It is well known that there is a remarkable link between cell cycle and cancer, as cell cycle appears to be the machinery that controls cell proliferation, and uncontrolled cell proliferation happens in cancer. The suppression of the NF-κB pathway by chalepin (**1**) was shown to be through modulation of the p65 subunit of NF-κB, where the phosphorylation of p65 and the translocation of the p65 subunit to nucleus were inhibited [11]. It can be noted that the NF-κB pathway is generally induced by carcinogens and inflammatory agents. Thus, suppression of NF-κB pathway by chalepin (**1**) could suggest its potential as an anticancer agent.

Caspase 8 is implicated to the activation of the intrinsic apoptotic pathway, and enhancement of caspase 8 activity can be exploited to identify compounds with plausible anticancer activity. In chalepin (**1**) treated cells, a significantly increased level of caspase 8 activity was noticed, when compared to the control; after 48 and 72 h of incubations, chalepin (**1**) (45 µg/mL) enhanced caspase 8 activity, respectively, by 5-fold and 8.6-fold [11].

This group of researchers also demonstrated that chalepin (**1**) and chalepensin (**2**) could induce mitochondrial mediated apoptosis in lung carcinoma cells (A-549), with chalepin (**1**) being more cytotoxic than chalepensin (**2**) [3]; chalepin (**1**) exhibited selective cytotoxicity against A-549 cells with an IC<sup>50</sup> value of 8.69 µg/mL (27.64 µM). Chalepin (**1**) was mildly toxic to the normal cell line with an IC<sup>50</sup> value of 23.4 µg/mL. Chalepensin (**2**) exhibited considerable cytotoxic property against A-549 cell line with IC<sup>50</sup> value of 18.5 µg/mL, while the cytotoxicity (IC<sup>50</sup> = 23.4 µg/mL) of this coumarin against noncancerous MRC-5 human lung fibroblast cell line was of moderate level as was with chalepin (**1**). Chalepin (**1**) showed morphological changes, typical for apoptosis, e.g., plasma membrane blebbing, cell vacuolization, echinoid spiking, chromatin condensation, formation of apoptotic bodies, cell shrinkage and nuclear fragmentation. Both coumarins (**1** and **2**) were found to downregulate inhibitors of apoptosis such as Bcl-2, survivin, Bcl-xl and cFLIP. They also triggered release of cytochrome c and activated caspases 9 and 3 to induce apoptosis. Chalepensin (**2**) was shown to possess cytotoxicity against colon (H-T29), lung (A-549), breast (MCF-7), kidney (A-498), and pancreatic (PACA-2) cancer cell lines [3].

Wu et al. [17] screened 19 compounds isolated from *Ruta chalepensis*, including chalepensin (**2**), for their potential cytotoxicity against KB (keratin forming tumor), Hela, DLD (colorectal adenocarcinoma) and Hepa tumor cell lines, but chalepensin (**2**) was found to be inactive against any of these cell lines at tested concentrations. From the available literature data, it is obvious that chalepin (**1**) is more cytotoxic than chalepensin (**2**). However, considerably more work has been carried out with chalepin (**1**) than with chalepensin (**2**) to date, and further comparative work may be necessary to gain a better insight into their anticancer potential.

#### *4.7. Miscellaneous Activities*

Spasmolytic activities of chalepin (**1**) and a few other coumarins, isolated from *Boenninghausenia albiflora* (Hook.) Rchb. Ex Meisn., were reported by Rizvi et al. [43]. Effects of aqueous extracts of *R. graveolens* and its ingredients, chalepensin (**2**) being one of them, on major drug metabolizing enzymes, cytochrome P450, uridine diphosphate (UDP)-glucuronosyltransferase and reduced nicotinamide adenine dinucleotide (phosphate) (NAD(P)H)-quinone oxidoreductase, were evaluated in mice [15]. The repeated administration of *R. graveolens* extract, rich in rutin and chalepensin (**2**), could induce hepatic CYP1a and CYP2b activities in a dose-dependent fashion. It was observed that male mice were more responsive than female mice to the extract-medicated induction of UGT (uridine glucuronosyltransferase). Earlier, the same group of researchers [48] showed mechanism-based inhibition of CYP1a1 and CYP3A4 by chalepensin (**2**), while this compound was also found to inhibit human CYP1a2, CYP2a13, CYP2c9, CYP2d6 and CYP2e1.

In order to study the in vivo effect of chalepensin (**2**), Lo et al. [44] assessed its effect on multiple hepatic P450 enzymes in C57BL/6JNarl mice, and observed that this coumarin, after oral administration (10 mg/kg) in mice for 7 days, could decrease hepatic coumarin 7-hydroxylation by CYP2a, and increase 7-pentoxyresorufin *O*-dealkylation by CYP2b, without affecting the activities of other CYP enzymes. It was further observed that the suicidal inhibition of CYP2a5 and the constitutive androstane receptor (CAR) mediated CYP2b9/10 induction simultaneously happened in chalepensin (**2**)-treated mice. Previously, Ueng et al. [46,47] and Lo et al. [45] carried out related extensive studies on mechanismbased inhibition of CYP enzymes by chalepensin (**2**) in various in vitro and in vivo models. However, there is no report on such activities of chalepin (**1**) available in the published literature to date.

In a study conducted by Wu et al. [17], chalepensin (**2**) at 100 µg/mL concentration displayed significant antiplatelet aggregation activity, induced by arachidonic acid and collagen. This coumarin, isolated from *Boenninghausenia albiflora* var*. japonica,* was also reported to possess calcium antagonistic property [6].

#### **5. Mutagenicity and Other Toxicities**

The mutagenicity of chalepin (**1**) was assessed at the HGPRT locus (AzGr) in Chinese hamster V79 cells [50], and this compound was found to be mutagenic. Chalepin (**1**), isolated from *Clausena aniseta*, a well-known medicinal plant from West Africa, showed anticoagulant (blood-thinning) activity when administered to rats in a single dose [51], and it could depress aniline hydroxylase activity. Ethylmorphine demethylase, hepatic DNA, reduced glutathione and glucose-6-phosphatase were unaffected by chalepin (**1**) treatment at a dose of 50 mg/kg for 3 days prior to sacrifice. This coumarin also resulted in α-1-globulin increase and a decrease in β-globulin content of the serum. Intraperitoneal treatment with chalepin (100 mg/kg) for 2 days caused death of 4 rats out of 100 within 48 h of treatment. Livers of dead rats showed generalized necrosis of hepatocytes. Chalepin (**1**) induced alterations in the serum protein pattern within this period. Liver lesions were observed in chalepin treated animals and were characterized by mild necrosis of hepatocytes. However, no report on mutagenicity of chalepensin (**2**) is available to date.

#### **6. Drugability' of Chalepin (1) and Chalepensin (2)**

"Drugability" can simply be defined as the ability of a compound to be used as a pharmaceutical drug. In order for a molecule to be developed as a drug, it must have certain physicochemical characteristics, which can be measured or predicted by various experimental or mathematical models. The Lipinski rule of five, formulated in 1997 by Christopher A. Lipniski, can be used, albeit not conclusively, to predict whether a compound could be an ideal candidate as a drug molecule, i.e., whether a compound possesses "druglikeness" or not [57]. This rule states that an orally active drug does not have more than one violation of the following criteria: a molecular mass less than 500 Daltons, no more than five hydrogen

donors, no more than 10 hydrogen bonds and an octanol-water partition coefficient (log *P*) that does not exceed five. Sometimes an additional criterion, "molar refractivity should be between 40–130" is also added to the above rule. If we consider these criteria in relation to chalepin (**1**) and chalepensin (**2**), both compounds tend to follow Lipinski rule of five, and there is no violation of this rule whatsoever (Table 3), which suggests that these compounds possess "druglikeness" or "drugability" and have the potential for further development as commercial drugs. However, it must be noted that this rule of five was originally presented to aid the development of orally bioavailable drugs and was not intended for guiding the medicinal chemistry in the development of all small-molecule drugs. Moreover, there is hardly any reliable experimental bioavailability data available on these coumarins (**1** and **2**) to make any connections between bioavailability and the predicted values for the criteria shown in Table 3.

**Table 3. "**Druglikeness" of chalepin (**1**) and chalepensin (**2**) \*.


\* Data obtained from ChemSpider (www.chemspider.com, (accessed on 24 February 2021)) and DrugBank (https://go.drugbank.com/drugs/DB02205, (accessed on 24 February 2021)).

#### **7. Conclusions**

The present work generated the first comprehensive and critical review of published literature on chalepin (**1**) and chalepensin (**2**), revealing various bioactivities of these compounds and their potential as new therapeutic agents. Among the activities, it appeared that antiprotozoal, antiviral and particularly anticancer activities bear promises for these compounds for further consideration for development as therapeutic agents, when considered in the light of nonviolation of the Lipinski rule of five and low level of toxicities. However, there is no report on any systematic preclinical or clinical trial with these compounds involving human volunteers available in the literature to date. Therefore, further studies, including controlled preclinical and clinical trials, are still needed before we can comment on the true therapeutic potential of these compounds.

**Author Contributions:** All authors contributed equally to collation of relevant information from extensive literature search. Additionally, L.N. and S.D.S. prepared, edited and submitted the manuscript as corresponding authors. All authors have read and agreed to the published version of the manuscript.

**Funding:** Lutfun Nahar (L.N.) gratefully acknowledges the financial support of the European Regional Development Fund—Project ENOCH (No. CZ.02.1.01/0.0/0.0/16\_019/0000868).

**Data Availability Statement:** All relevant data have been presented as an integral part of this man.

**Acknowledgments:** Lutfun Nahar gratefully acknowledges the financial support of the European.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


### *Review* **Coumarin Derivatives in Inflammatory Bowel Disease**

**Luiz C. Di Stasi**

Laboratory of Phytomedicines, Pharmacology, and Biotechnology (PhytoPharmaTech), Department of Biophysics and Pharmacology, Institute of Biosciences, São Paulo State University (UNESP), 18618-689 Botucatu, SP, Brazil; luiz.stasi@unesp.br

**Abstract:** Inflammatory bowel disease (IBD) is a non-communicable disease characterized by a chronic inflammatory process of the gut and categorized into Crohn's disease and ulcerative colitis, both currently without definitive pharmacological treatment and cure. The unclear etiology of IBD is a limiting factor for the development of new drugs and explains the high frequency of refractory patients to current drugs, which are also related to various adverse effects, mainly after long-term use. Dissatisfaction with current therapies has promoted an increased interest in new pharmacological approaches using natural products. Coumarins comprise a large class of natural phenolic compounds found in fungi, bacteria, and plants. Coumarin and its derivatives have been reported as antioxidant and anti-inflammatory compounds, potentially useful as complementary therapy of the IBD. These compounds produce protective effects in intestinal inflammation through different mechanisms and signaling pathways, mainly modulating immune and inflammatory responses, and protecting against oxidative stress, a central factor for IBD development. In this review, we described the main coumarin derivatives reported as intestinal anti-inflammatory products and its available pharmacodynamic data that support the protective effects of these products in the acute and subchronic phase of intestinal inflammation.

**Keywords:** inflammatory bowel disease; coumarin; isocoumarin; Crohn's disease; ulcerative colitis; glutathione; oxidative stress; complementary therapies; intestinal inflammation

#### **1. Introduction**

Currently, the search and discovery of new drugs with efficacy, safety, and quality control to prevent and treat non-communicable diseases is a huge challenge for the chemical and pharmaceutical sciences as well as medicine. This task is not only a challenge in the present time but also to guarantee health quality for the next generations. Noncommunicable diseases, also known as chronic diseases, are persistent illnesses generally without a pharmacological cure, that tend to be of long duration, requiring long-term and systematic treatment approaches, and the result from multifactorial etiological factors. In general, patients with non-communicable diseases live throughout their lives with several symptoms, continuously using drugs to relieve them. Non-communicable diseases are a group of chronic disorders including cardiovascular diseases, diabetes, multiple sclerosis, obesity, arthritis, asthma, Parkinson's and Alzheimer's diseases, cancer, and inflammatory bowel disease (IBD), which have a high impact on the health system. According to a recent study, 56 million people died in 2017 and non-communicable diseases account for more than 73.4% of these global deaths, i.e., 41.1 million people [1]. Based on this, the search for new drug development to relieve symptoms and mainly to prevent non-communicable diseases is an important approach to improve several world health problems and patients' life quality.

The discovery of new drugs is based on synthetic chemistry, partial synthesis or modification of active molecules of synthetic or natural origin, and bioprospection of natural products, particularly from fungi and plant species. The research with natural products was overshadowed by the advent of the new technologies, synthesis of several

**Citation:** Di Stasi, L.C. Coumarin Derivatives in Inflammatory Bowel Disease. *Molecules* **2021**, *26*, 422. https://doi.org/10.3390/ molecules26020422

Academic Editor: Maria João Matos Received: 28 November 2020 Accepted: 23 December 2020 Published: 15 January 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

chemical active compounds, and international regulatory systems for biodiversity access established by the United Nations Convention on Biological Diversity. However, all current difficulties did not reduce the importance of the world biological biodiversity, particularly from tropical areas, as an inexhaustible and magnificent source of new medicines, which should be carefully and legally studied, respecting international regulatory systems and the traditional knowledge from local communities.

Natural products from plant and fungi origin are the source of several drugs with wide applications and pharmacological importance. Some of these compounds have defined the way of science and modern medicine as well as represented the basis of the treatment of several serious diseases and health problems affecting the world's people. The antibiotic penicillin discovered in *Penicillium* genus fungi; morphine, an opioid compound useful as a pain reliever, isolated of the opium plant (*Papaver somniferum*), and acetylsalicylic acid, a lead compound of the non-steroidal anti-inflammatory drugs, which is related to salicin, obtained from plants belonging genera *Salix* and *Populus*, are some emblematic examples of the natural products that have changed the history of medicine. Even with advanced modern medicine and biotechnology, the most recent discoveries of lead compounds include two products of plant origin, artemisinin and taxol, an antimalarial and an antineoplastic agent, respectively. The research related to the discovery of artemisinin from *Artemisia annua* received the 2015 Nobel Prize in Physiology or Medicine. Artemisinin completely changes the control of malaria and represents a new class of antimalarial drugs, whereas taxol, isolated from species belonging to the genus *Taxus*, particularly *Taxus brevifolia*, represents a new class of anti-cancer drugs. The number of lead compounds obtained from nature is high, showing that natural products play a key role in human health surveillance and represent the support basis of drug research and discovery.

Plant-based products are rich in several chemical classes of compounds, among which the alkaloids, terpenoids, tannins, and phenol and polyphenol compounds stand out, which are potentially useful to prevent and treat several disorders, particularly non-communicable diseases. Phenol and polyphenolic compounds, one of the most important classes of secondary metabolites from plants, include a plethora of different classes of molecules with high pharmacological value, among which the flavonols, flavanones, flavones, anthocyanidins, xanthones, stilbenes, catechins, quinones, and coumarins may be highlighted. These compounds represent an important source of new molecules with several pharmacological properties and are widespread in vegetables commonly consumed daily as dietary foods and spices. Dietary intake of several plants containing these compounds contributes to the plasma bioavailability of active molecules, which are useful both to improve immune response and act as preventative products for several non-communicable diseases. Nowadays, it has been considered that a properly used nutritional approach might be a part of the treatment of non-communicable diseases, particularly patients with Crohn's disease and ulcerative colitis, two chronic inflammatory disorders of the gut [2]. The pharmacological properties of phenol and polyphenol compounds against the inflammatory processes of the gut have been exhaustively reported, focusing on flavonoids [3,4], proanthocyanidins and anthocyanins [5,6], and catechins [7]. However, there is a lack of data and analysis of the potential use and application of coumarin and their derivatives as preventative and curative compounds in non-communicable diseases, particularly in inflammatory bowel diseases (IBDs).

In this review, we aim to update and systematize the available knowledge on the pharmacological activities of coumarin derivatives in the various in vivo experimental models of intestinal inflammation and in vitro studies to provide data and insights to further preclinical, clinical, and molecular studies, demonstrating the main and potential active coumarins useful to prevent or treat inflammatory bowel diseases as well as its main mechanisms of action and signaling pathways.

#### **2. Coumarin and the Main Coumarin Derivatives**

Coumarins, also known as benzopyrones, comprise a class of cinnamic acid-derived phenolic compounds found in fungi, bacteria, and plant species, particularly in edible, medicinal, and spice plants from different botanical families [8]. Coumarins are secondary plant heterocyclic metabolites composed of fused benzene and α-pyrone rings (Figure 1), and they occur widely in different parts of plants, such as roots, seeds, nuts, flowers, and fruits either as heterosides or in free form [8,9]. The term coumarin originated from de name "cumaru" the local name for the Brazilian teak plant (*Dipteryx odorata* Wild.) from the Fabaceae botanical family, in the traditional medicine of the Brazilian Amazon forest. *Dipteryx odorata* is an endemic plant of Central America and the North of South America, widespread in the Amazon Forest region, from which the coumarin was firstly isolated by Vogel in 1820 [10]. Its seeds, named tonka beans, are the natural source of coumarin, a compound widely used by the perfumery companies to replace vanilla, particularly as a fixative and enhancing agent in perfumes as well as added to toilet soap, detergents, toothpaste, tobacco, and alcoholic products [9]. Moreover, the isocoumarin, also recognized as 1*H*-benzopyran, is an isomer of the basic structure of coumarin, in which the orientation of the lactone ring is reversed (Figure 1). From the different substituted groups on the basic structure of isocoumarin, several subclasses and isocoumarin derivatives are also found in plant species such as paepalantine, capillarin, and thumberginol A (Figure 1). α‐ ‐

‐

**Figure 1.** Basic structures of coumarins and isocoumarins and the main isocoumarin derivatives. Chemical structures were drawn using ACD/ChemSketch software.

Coumarins are categorized into four main subtype classes of compounds: simple coumarins, furanocoumarins, pyranocoumarins, and the pyrone-substituted coumarins. Simple coumarins are composed of molecules with hydroxyl, alkoxyl, and alkyl substitution patterns on the basic structure and their glucosides [11]. Simple coumarin class represents the main class of coumarin derivatives with intestinal anti-inflammatory properties, particularly esculetin, esculin, 4-hydroxycoumarin, osthole, and 4-methylesculetin (Figure 2). ‐ ‐ ‐ ‐

Furanocoumarins are composed of coumarins derivatives in which a furan ring is fused with the basic structure of coumarin via C6-C7 or C7-C8 [10,11], generating linear furanocoumarins (fusion via C6-C7) such as psolaren, imperatorin, and xanthotoxin or angular furanocoumarins (fusion via C7-C8) such as isobergapten and angelicin (Figure 3). Similarly, in the pyranocoumarins, a six-membered pyran ring is fused with the benzene ring of the basic structure of coumarins via C6-C7 or C7-C8 [10,11]. Decursin (a linear pyranocoumarin) and seselin (an angular pyranocoumarin) are some examples of pyranocoumarins (Figure 3), which have no intestinal anti-inflammatory activity. ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐

‐ **Figure 2.** Chemical structures of the main simple coumarin derivatives with intestinal anti-inflammatory activity. Chemical structures were drawn using ACD/ChemSketch software.

Finally, pyrone-substituted coumarins are coumarin derivatives containing different chemical radicals fused with the pyran ring of coumarin [10,11]. Pyrone-substituted coumarins include natural and synthetic coumarins such as warfarin and dicoumarol (Figure 3). Pyrano-substituted coumarins have no intestinal anti-inflammatory activity but comprise some compounds with high pharmacological relevance. Warfarin is a derivative of dicoumarol, a pyrano-substituted coumarin isolated from hay species (*Melilotus alba* and *Melilotus officinalis*) after natural oxidation by several fungi, mainly *Penicillium nigricans* and *Penicillium jensi* found in moldy hay [12]. Dicoumarol and warfarin were first used as rodenticides due to their ability to promote internal hemorrhage in rodents [12]. The anticoagulant properties of dicoumarol and warfarin were the basis for the development of anticoagulant drugs to prevent stroke in patients with cardiovascular diseases, mainly atrial fibrillation and valvular heart disease, and to prevent and treat vein thrombosis and pulmonary embolism [13].

‐ **Figure 3.** Chemical structures of angular and linear furanocoumarins, angular and linear pyranocoumarins, and pyranosubstituted coumarins. Chemical structures were drawn using ACD/ChemSketch software.

#### ‐ **3. Inflammatory Bowel Diseases: General Aspects**

‐ ‐ ‐ ‐ Inflammatory bowel disease (IBD) consists of Crohn's disease (CD) and ulcerative colitis (UC), two relapsing chronic inflammatory processes of the gastrointestinal tract, which are part of a group of immune-mediated inflammatory diseases, without a definitive pharmacological treatment and cure [14]. Patients with CD or UC live with several harmful effects in their daily physical, social, and professional activities because these diseases produce limiting effects such as changes in intestinal habits with several evacuations, abdominal pain, diarrhea, bleeding, perianal fistulas and other extraintestinal manifestations.

‐ IBD is a disease that is increasing globally, affecting some 6 to 8 million people in the world and presenting a prevalence rate of 84.3 people (79.2 to 89.9) per 100,000 population in 2017 [15]. Although it is a disorder with low mortality and with a death rate of 0.51 per 100,000 population, IBD is growing exponentially around the world, showing the highest prevalence rates in North America and the United Kingdom and other European countries such as Norway, Poland, and Slovakia, whereas lower prevalence rate has been reported in several countries of Africa, South America, and Southeast Asia [15]. There is a direct relationship between the high prevalence rate of IBD and the industrialization level of a specific country, but the prevalence and incidence rates notably are also rising in newly industrialized countries [16]. The prevalence and incidence rate increment in developing or newly industrialized countries is associated with the industrialization process and migration of population from rural to urban areas, which promote changes in the lifestyle and the people choices related to diet, daily activities, and social behaviors. These changes suggest rates of IBD prevalence and incidence should thrive in parallel to those in the industrialized countries [16].

The IBD etiology is unclear, but several triggering factors have been related to its occurrence and development, including dysregulated immune response, dysfunctional intestinal barrier function, genetic, and environmental aspects. Currently, the pathogenesis of IBD involves a dysregulated autoimmune response and increased intestinal permeability related to gut dysbiosis, which is accelerated by exposure to environmental factors in individuals who have a pre-existing high-risk genotype [17]. The complex and inexact etiology of IBD is a limiting factor in the discovery of new pharmacological and complementary

therapies, and the development of preventive strategies useful to define a general protocol of treatment and definitive management of IBD patients. The multifactorial aspects of IBD also explain the high frequency of patients who are refractory to current pharmacological treatments, including conventional drugs such as aminosalicylates, glucocorticoids, immunosuppressants, and biological therapies based on monoclonal antibodies [18]. Current pharmacological treatment of IBD is based on the relief or to create a time of deep remission of symptoms. However, the long-term use of these drugs, produces serious side effects, reducing patient adherence to pharmacological treatment.

IBD also includes a lot of risk factors with an imbricated relationship among them. A series of interactions among risk factors, which do not act in isolation because none of the risk factors alone is sufficient for IBD development, have been suggested [19]. Risk factors for IBD development include intrinsic and extrinsic factors. The intrinsic factors involve genetic predisposition and familiar history as well as the intestinal microbiota, whereas extrinsic risk factors include smoking, appendectomy, hygiene, infections, use of antibiotics and other drugs such as NSAIDs (Non-steroidal anti-inflammatory drugs) and oral contraceptives, a diet with lower fiber and higher fat, vitamin D deficiency as well as lifestyle and social behavior, mainly high stress, sleep privation and lower physical activity [19]. Genetic and epigenetic studies have been extensively used as an important source of data, which are important for a better prediction of IBD course, identification of loci and candidate genes yielding valuable insights into the pathogenesis of IBD and disease pathways, which can be relevant in the clinical practice [20,21].

Intestinal microbiota, which has a key role in the pathogenesis of IBD, is an intrinsic factor that can be modulated by a series of products able to differentially affect distinct microorganisms, including functional food products, mainly probiotic, prebiotic, and symbiotic, natural products such as polyphenol compounds and standardized phytomedicines. In health conditions, intestinal microbiota via fermentation of dietary components produces a series of metabolites, mainly short-chain fatty acids (SCFAs), which are a source of energy for colonocytes and bacteria, and play several protective effects in the body after prompt absorption (Figure 4). On the other hand, the management of extrinsic factors, including changes in lifestyle, social behavior, and diet options as well as a lower exposition to other extrinsic factors can represent an important approach to reduce IBD development.

The combinatory action among genetic predisposition, external environmental factors, and intestinal microbiota is essential to the development of the dysregulated immune response and dysfunctional intestinal barrier [22], which are responsible for IBD development (Figure 4). Both CD and UC patients exhibit a dysfunctional intestinal epithelial barrier with increased permeability as well as an exaggerated immune response in the gastrointestinal tract towards the intestinal microbiota, which is not appropriately controlled, leading to intestinal inflammation [3,22]. The increase of intestinal permeability has been recognized as an early feature of the intestinal inflammatory process, which reduces the intestinal barrier function, a key factor to maintain intestinal homeostasis [23]. In this process, several factors and mediators are involved, particularly the zonulin pathway activation, a key process to control intestinal permeability, suggesting zonulin as a biomarker of gut permeability as well as a key target for the action of intestinal anti-inflammatory products [23–25]. The dysregulated immune response includes an innate immune response, the first-line defense against any damage promoted by pathogens, which is mediated by different cell types including macrophages, neutrophils, monocytes, dendritic, epithelial, and endothelial cells [3,22]. These cell types are responsible for phagocytosis, elimination of pathogens, production of several cytokines, and development of barrier and transport functions (Figure 4). The response to pathogens includes prompt participation of the antigen-presenting cells (APCs), which mediate the differentiation of T-cells into effector T helper (Th) cells, including Th1, Th2, and Th17 cell types, and regulator T-cells (Treg) (Figure 4), which are constituents of the adaptative immune response [3,22]. These different cell types are responsible for the synthesis and dysregulated release of a series of immunologic and inflammatory mediators with wide importance in the pathogenesis of

IBD, including a plethora of chemokines and cytokines. The pathophysiology of intestinal inflammation is very complex, including a wide number of signaling pathways and endogenous mediators as illustrated in Figure 4, where are particularly indicated the main targets for the action of coumarin derivatives.

‐

‐

‐

‐

‐

**Figure 4.** The main pathways of the intestinal inflammatory process or the action of coumarin derivatives. GS ‐ **Figure 4.** The main pathways of the intestinal inflammatory process or the action of coumarin derivatives. GSH, glutathione; IL, interleukin; iNOS, inducible nitric oxide synthase; MPO, myeloperoxidase; SCFAs, short-chain fatty acids; TGF, transforming growth factor; Th, T helper cells; Treg, T regulatory cells; TNF, tumor necrosis factor.

> Moreover, in the gastrointestinal tract, there is homeostasis of intestinal microbiota with immune cells responsible for a balance of host defense and immune tolerance, which can be shift leading to a dysbiosis process that plays a key role in the pathogenesis of IBD [17]. Although dysregulated immune response associated with both genetic predisposition and intestinal microbiota participates in the intestinal inflammatory process, oxidative stress characterized by an excessive release of reactive oxygen and nitrogen species (ROS/RNS) plays a key role in IBD pathogenesis [26,27]. The excess of oxidative mediators reacts with cell membrane fatty acids, proteins, and DNA impairing their functions (Figure 4). The production of superoxide, the main source of free radicals, is required to kill bacteria, a process that occurs especially in neutrophils and other cell types such as epithelial cells from the intestine. From superoxide production, several free radicals are produced via nitric oxide synthase and glutathione-related enzymes (Figure 4).

#### **4. Intestinal Anti-Inflammatory Activity of Coumarin Derivatives**

The dysregulated immune and oxidative response triggered by intestinal inflammation is an imbricated and complex interaction involving a series of endogenous mediators from different signaling pathways and receptors, such as nuclear factor-kappa b (NF-κB), nuclear factor erythroid 2 (NEF2)-related factor 2 (Nrf2), peroxisome proliferator-activated

receptor gamma (PPAR-γ), pregnane X receptor (PXR), hypoxia-inducible factor (HIF), several enzymes, especially cyclooxygenase 2 (COX-2), mitogen-activated protein kinases (MAPKs), and HIF-prolyl hydroxylases (PHDs) as well as mediators of intestinal epithelial barrier function such as zona occludens 1 (ZO-1), occludin, mucins (MUC1, MUC2, MUC 3, MUC4), and E-cadherin. These different pathways will be discussed in this review to explain the effects of some coumarin derivatives and their partially elucidated mechanisms of action. However, until now the majority of coumarin derivatives produce intestinal anti-inflammatory activity acting as modulators of oxidative stress and immune response, whereas a little number of coumarin derivatives were reported to act by other signaling pathways, which are also partially involved in oxidative stress modulation.

The intestinal anti-inflammatory activity of different coumarin derivatives was described using different experimental models of intestinal inflammation, mainly trinitrobenzene sulphonic acid (TNBS) and dextran sodium sulfate (DSS) in rats or mice as well as several in vitro studies with distinct cell types. Although coumarin is a class of natural and synthetic compounds with high chemical diversity, the intestinal anti-inflammatory has been limited to three subclasses of coumarin derivative, i.e., isocoumarins having one active compound namely paepalantine, a lot of simple coumarin derivatives, and the furanocoumarin imperatorin.

#### *4.1. Effects of Coumarin Derivatives on Oxidative Stress*

A recent review reported that different antioxidant compounds act by six general mechanisms: inhibiting free radical production by activated oxygen metabolites, changing the structural organization of free radical, producing a local decrease of oxygen concentration, interacting with organic radicals, chelating metal ions, and converting peroxides to stable and inactive products [28]. These processes reduce the availability of free radical species, improving the response against oxidative stress, which are the main properties of the coumarin derivatives with intestinal anti-inflammatory activity.

Oxidative stress is considered an imbalance in which excessive levels of oxygen free radicals such as reactive oxygen species (ROS) and reactive nitrogen species (RNS) are present in the biological system in the face of inadequate availability of the antioxidants, which are capable to destroy these harmful products from metabolic processes [29]. A free radical is a molecule of the normal metabolism of oxygen with an unpaired or odd number of electrons, which is highly reactive and able to react with lipids from the cell membranes, proteins, and nucleic acids, affecting their structure and functions. In an inflammation process, ROS and RNS production is a prompt defense response of the body to kill several invading pathogens as well as to regulate the immune response via pro-inflammatory chemotaxis induction into the site of inflammatory processes as well as modulate the interactions and activation of the immune cell types [30]. However, when occurs an imbalance between free radical production and endogenous antioxidant response, the reaction of reactive oxygen and nitrogen species with host lipids, proteins, and nucleic acids generate oxidative stress, triggering a series of molecular and cellular events such as tissue damage and fibrosis [30], which are related to the origin and development of several chronic diseases, including IBD.

The reactive oxygen species and the endogenous mediators of antioxidative response are represented in Figure 5. The main reactive species in the biological system responsible for oxidative stress is the superoxide radical anion (O<sup>2</sup> •−), which is formed by the addition of a single electron in molecular oxygen (O2) by the action of reduced nicotinamide adenine dinucleotide phosphate (NADPH) oxidase (NOX) enzyme [26,27]. O<sup>2</sup> •− is the source of hydrogen peroxide (H2O2), formed by the action of superoxide dismutase (SOD) enzyme, which catalyzes the dismutation of O<sup>2</sup> •− into H2O2. Although H2O<sup>2</sup> is not a free radical, it is a molecule highly reactive [29,30] and represents a key substrate for free radical production. H2O<sup>2</sup> via Fenton reaction generates hydroxyl radical (OH−), which is the most reactive and harmful metabolite of oxygen metabolism. Simultaneously, H2O<sup>2</sup> by the action

of myeloperoxidase (MPO) produces hypochlorous acid (HOCl), a potent antioxidant with antimicrobial activity and useful as a defense against several infectious pathogens [30,31].

MPO is the most toxic enzyme found in the granules of neutrophils and monocytes, two important cell types that participate in the intestinal inflammatory process with a key role in the innate immune response to pathogens [30]. MPO generates reactive intermediates, inducing oxidative peroxidation of lipids, and proteins and DNA damage. During the inflammatory response, MPO is released from neutrophils and monocytes, catalyzing the formation of HOCl from H2O<sup>2</sup> (Figure 5). HOCl produced by the action of MPO influences the conversion of glutathione (GSH) to oxidized glutathione (GSSG), which disrupts the cellular redox balance, reducing the antioxidant GSH pool (Figure 5) and increasing the susceptibility to oxidative stress [30]. Generally, MPO activity is strongly increased in experimental models of intestinal inflammation, whereas simultaneously is possible to observe a depletion of GSH [32]. − − − −

‐ γ‐ γ‐ ‐ ‐ ‐ **Figure 5.** Free radical production and glutathione (GSH) antioxidant system. CAT, catalase; GPX, glutathione peroxidase, GR, glutathione reductase; GS, glutathione synthetase; GST, glutathione S-transferase; NOX, NADPH oxidase; MPO, myeloperoxidase; SOD, superoxide dismutase; γ-GLC, γ-glutamylcysteine ligase. In these oxidative processes, the following activities were reported: A. Counteraction of GSH depletion by paepalantine, coumarin, 4-hydroxycoumarin, esculetin, 4-methylesculetin, daphnetin, esculin, scopoletin, scoparone, and fraxetin; B. Inhibition of MPO activity by paepalantine, esculetin, 4-methylesculetin, daphnetin, and esculin; C. Scavenging activity of free radical by paepalantine, daphnetin, esculin, scopoletin, scoparone, and fraxetin; D. Inhibition of lipid peroxidation by esculetin, and daphnetin; E. Inhibition of GPX activity and expression by 4-methylesculetin; F. Increase of GST and GR activity and expression by 4methylesculetin. Chemical structures were drawn using ACD/ChemSketch software.

GSH, a tripeptide formed by glutamic acid, cysteine, and glycine, is the key endogenous antioxidant that participates in antioxidant response. GSH is produced by a reaction with two steps (Figure 5). Firstly, a residue of glutamic acid (Glu) binding with cysteine via γ-glutamylcysteine ligase (γ-GCL) action, producing γ-glutamylcysteine [33]. Secondly, γ-glutamylcysteine reacts with a residue of glycine under the action of glutathione synthetase (GS) enzyme to produce GSH [33]. GSH participates in antioxidant response acting as a free radical scavenger, reducing dehydroascorbate to ascorbate, which regenerates α-tocopherol from the α-tocopherol radical oxidation, and serving as a co-substrate for several antioxidant enzymes, mainly glutathione S-transferase (GST) and glutathione peroxidase (GPx) [34]. High pools of GSH is essential to modulate oxidative stress, and GSH also can be regenerated by oxidation of its disulfide-oxidized dimer (GSSG) by the action of glutathione reductase (GR) (Figure 5). This reaction occurs at expense of the reduction of NADPH from the pentose phosphate pathway [34].

The modulation of oxidative stress for the body's defense against free radical tissue damage and macromolecules oxidation is modulated by antioxidants, which are classified into nonenzymatic antioxidants consisting of micronutrient components, and enzymatic endogenous system [17,29]. The nonenzymatic antioxidant system includes several small molecules, mainly GSH, vitamin E, vitamin C, β-carotene, retinol, uric acid, and ubiquinol as well as several microelements like selenium, iron, zinc, copper, and manganese [29]. Vitamins act as donors and acceptors of ROS and the micronutrients act as cofactors, which regulate the activities of the antioxidant enzymes [29]. Endogenous enzymatic antioxidants involve superoxide dismutase (SOD), catalase (CAT), glutathione peroxidase (GPX), glutathione S-transferase (GST), glutathione reductase (GR), and γ-glutamyl transferase (γ-GT) enzymes [32], which act by different pathways to reduce the free radical availability and control oxidative stress.

In this process, CAT can break down two H2O<sup>2</sup> molecules generating one molecular oxygen and two molecules of water [35], reducing H2O<sup>2</sup> pool availability, which is used as a substrate for the production of OH<sup>−</sup> (Figure 5). In association with SOD and CAT, GPX, a dependent enzyme of micronutrient selenium, plays an important role in the reductions of H2O<sup>2</sup> and lipid peroxides (LOOH) to produce water and lipid alcohol (LOH), contributing to modulation of oxidative stress and avoiding direct tissue damage [36]. Moreover, GST can catalyze the transfer of a GSH group to organic and inorganic electrophiles, reducing these compounds into unreactive products [34].

In another reaction pathway (Figure 6), nitric oxide synthase (NOS) controls the reaction of O<sup>2</sup> •− with nitric oxide radical (NO• ) to produce free radical peroxynitrite (ONOO−), which is responsible for nitrosylation of proteins and oxidation of lipoproteins [31]. NO• is produced via enzymatic oxidation of L-arginine to L-citrulline by the action of constitutive and inducible NOS. Besides the mediator in blood pressure, NO• participates in the immune and inflammatory responses with biocidal activity against several microorganisms and induces damages on the proteins and DNA [31,37]. Toxic and oxidative effects of NO• results from its oxidation, generating highly reactive species, such as nitrite (NO<sup>2</sup> <sup>−</sup>) and peroxynitrite (ONOO−).

NO<sup>2</sup> <sup>−</sup> is produced by NO• autooxidation forming nitrous anhydride (N2O3), an intermediate in this conversion recognized as a potent nitrosating agent [31], which can also be used to produce nitrogen dioxide by the action of the MPO enzyme (Figure 6). Carbon dioxide (CO2) reacts catalytically with ONOO<sup>−</sup> to produce nitroperoxycarbonate (ONOOCO2), which via homolysis of the O-O bonds, carbon trioxide (CO<sup>3</sup> •−), and nitrite dioxide (NO<sup>2</sup> • ) radicals are produced [31]. Moreover, when ONOO- decomposes in the absence of CO2, the NO<sup>2</sup> • and OH• radicals production take place, whereas, in the presence of CO2, CO<sup>3</sup> •− and NO<sup>2</sup> • radicals are produced, and in this process (Figure 6). MPO also participate, affecting tyrosine nitration when NO<sup>2</sup> <sup>−</sup> is used as a co-substrate, with consequent production of NO<sup>2</sup> • , a reactive free radical [31]. Based on its diverse action as a marker of neutrophil infiltration, inflammatory process, and oxidative stress, MPO represents a potential target for the development of synthetic and natural compounds against several diseases, including atherosclerosis, acute coronary syndromes, ischemic heart disease, and IBD [30,38,39].

− −

**Figure 6.** Nitric oxide synthase (NOS) pathway of free radical production. NOX, NADPH oxidase; NOS, nitric oxide synthase; MPO, myeloperoxidase. In these oxidative processes, the following activities were reported: A. Inhibition of iNOS activity by paepalantine, esculetin, esculin, auraptene, and collinin; B. Reduction of NO release by isomeranzin and esculin.

> − − ‐ − ‐ − − ‐ The major regulator of the endogenous antioxidant system is the nuclear factor erythroid 2 (NEF2)-related factor 2 (Nrf2) that protects cells from several stressors agents, such as ROS, RNS, and environmental damage [40]. In physiological conditions, Nrf2 binds with cullin 3 (cul3) and Kelch-like ECH-associated protein 1 (keap1), a key repressor of the Nrf2 signaling pathway (Figure 7), preventing the translocation of Nrf2 to the nucleus [41]. This complex, after ubiquitination, promotes Nrf2 degradation via proteolysis. Under oxidative stress conditions, Nrf2-keap1 complex is uncoupled and a free Nrf2 is translocated into the nucleus, where binds with small Maf (sMaf) proteins [42]. The heterodimer binds with antioxidant response elements (ARE) target genes (Figure 7), regulating the expression of several antioxidant-related endogenous genes, including the enzymes CAT, GPX, SOD, GST, γ-GCL, GR, NADPH quinone oxidoreductase, and heme oxygenase [41]. Based on this, Nrf2 is a key mediator of the antioxidant defense system as well as an important target for the action of new synthetic and natural compounds, including coumarin derivatives such as esculetin.

> > ‐ ‐

‐

‐

‐ ‐ ‐ ‐ ‐ ‐ ‐ γ γ‐ ‐ **Figure 7.** The nuclear factor erythroid 2 (NEF2)-related factor 2 (Nrf2) signaling pathway of oxidative stress. ARE, antioxidant element of response; CAT, catalase; cul3, cullin 3; GST, glutathione S-transferase; GPX, glutathione peroxidase; GR, glutathione reductase; HO-1, heme-oxygenase 1; Keap1, Kelch-like ECH-associated protein 1; Nrf2, nuclear factor erythroid 2-related factor 2; sMaf, small Maf proteins; Ub, ubiquitin; γGCL, γ-glutamylcysteine ligase. In these oxidative processes, the following activities were reported: A. Upregulation of Nrf2 by esculetin, 4-methylesculetin, daphnetin, and esculin.

#### ‐ ‐ 4.1.1. The Isocoumarin Paepalantine

‐ ‐ ‐ ‐ ‐ ‐ ‐ μ Paepalantine (9,10-dihydroxy-5,7-dimethoxy-1*H*-naptho(2,3c)pyran-1-one), was the first plant-derived coumarin studied in an experimental model of intestinal inflammation [43]. Paepalantine (Figure 1) is an isocoumarin previously isolated from the capitula of the Brazilian endemic *Paepalanthus bromelioides* plant from the Eriocalulaceae botanical family [44], which produced protective effects in the acute and relapse phases of the intestinal inflammation induced by TNBS in rats [43]. The protective effects observed after oral administration of the 5 and 10 mg/kg were similar to those promoted by the 25 mg/kg of sulphasalazine, a 5-aminosalicylate currently used to treat human IBD, i.e., paepalantine produced intestinal anti-inflammatory activity at doses 2.5 and 5.0-times lower than a reference drug [43]. Intestinal anti-inflammatory activity of the paepalantine was related to prevention of the GSH depletion (Figure 5) and inhibition of the colonic NOS activity (Figure 6), which was upregulated by the inflammatory process, suggesting that intestinal anti-inflammatory activity is related to its antioxidant properties [43]. Paepalantine also inhibited HOCl production in rat neutrophils, reducing oxidative stress

by the inhibition of MPO activity and scavenging HOCl (Figure 5) [45]. Moreover, the antioxidant properties of paepalantine were evidenced by its potent scavenging properties in the 1,1-diphenyl-2-picrylhydrazyl (DPPH) and superoxide radicals as well as by its ability to protect mitochondria from hydroperoxide accumulation and mitochondrial membrane lipid peroxidation [46,47]. The inhibition of iNOS activity by paepalantine was recently corroborated through in vitro studies with LPS-stimulated macrophages [48]. In this study, paepalantine binds with the NOS enzyme through several structural amino acids just on the active site of the enzyme, reducing its enzymatical activity [48].

#### 4.1.2. Coumarin and 4-hydroxycoumarin

Intestinal anti-inflammatory activity of coumarin and its derivative 4-hydroxycoumarin (Figure 2) were evaluated in the acute and subchronic phases of the TNBS-induced intestinal inflammation model in rats [49]. In this study, damage score and extension of tissue lesion induced by TNBS were significantly reduced after oral administration of coumarin (25 mg/kg) and 4-hydroxycoumarin (10 and 25 mg/kg), and these protective effects were accompanied by a counteraction of GSH depletion and inhibitory MPO activity (Figure 5) [49]. Although 4-hydroxycoumarin produced effects at lower doses when compared with coumarin, is not possible to suggest that the OH substitution in C4 was directly related to the improvement of its effects in the acute or sub-chronicle protocols of intestinal inflammation induced by TNBS in rats.

#### 4.1.3. Esculetin (6,7-dihydroxycoumarin) and 4-methyl Esculetin

Esculetin and 4-methylesculetin (Figure 2) oral administration produced antioxidant protective effects in rats with TNBS-induced intestinal inflammation [50]. While esculetin counteracted GSH depletion at the dose of 10 mg/kg, with no effects on MPO activity, 4-methylesculetin produced significantly positive effects on the GSH levels (2.5 and 5 mg/kg) and MPO activity (5 and 10 mg/kg) in the acute phase of intestinal inflammation (Figure 5). In the sub-chronicle protocol, when coumarins were administered after induction of intestinal inflammation, the effects of 4-methylesculetin were evidenced on the GSH level and MPO activity, whereas esculetin was inactive to restore the basal value of these mediators [50]. Moreover, the inhibitory concentration of 4-methylesculetin on the lipid peroxidation membranes was approximately twice lower than esculetin [50]. A comparative analysis of data suggesting that 6,7-dihydroxylated coumarins when substituted at C4 with a methyl group had an improvement of effects on the MPO activity. Reduction of damage score and MPO activity was also demonstrated after the intrarectal administration of esculetin at 100 and 200 µM in rats with intestinal inflammation previously induced by TNBS [51]. The antioxidant property of esculetin was also corroborated by its action inhibiting iNOs activity (Figure 6) and modulating Nrf2 (Figure 7) signaling pathway [51,52].

Besides intestinal anti-inflammatory activity related to antioxidant properties counteracting GSH depletion, inhibiting MPO activity and lipid peroxidation [50], the effects of 4-methylesculetin (6,7-dihydroxy-4-methylcoumarin) in acute and subchronic phases of TNBS-induced intestinal inflammation was evaluated in comparison with effects of sulphasalazine and prednisolone in rats as well as in RAW264.7, Caco-2 and splenocytes culture cells [53]. Similar to previously reported, 4-methylesculetin improved clinical, histopathological, and biochemical parameters, such as GSH levels and MPO activity (Figure 5) in both acute and subchronic phases of the TNBS-induced intestinal inflammation model [53]. In a recent study in the DSS-induced intestinal inflammation in mice, 4-methylesculetin also improved histopathological indicators of intestinal inflammation, reduced MPO activity, and markedly counteracted GSH depletion [54].

Considering that the intestinal anti-inflammatory activity of 4-methylesculetin was closely related to several mediators of oxidative stress, an interesting study was carried out to investigate the molecular mechanisms involved in these antioxidant properties [32]. In the TNBS model of intestinal inflammation, treatments with 5 and 10 mg/kg methyles-

culetin significantly decreased damage score, lesion extension, and diarrhea incidence [32]. These protective effects were accompanied by an inhibition of the MPO and GPx activities and simultaneous increment of GST and GR activities, with no effects on the SOD and CAT activities [32]. Moreover, treatment with 4-methylesculetin was able to prevent downregulation of GR and Nrf2, with no effects on the GRX, GST gene expression, suggesting that GR is a target enzyme for the action of 4-methylesculetin. Molecular interaction between 4-methylesculetin and GR using UV-vis absorbance spectroscopy, fluorescence measurements, saturation transfer difference nuclear magnetic resonance, and computational modeling were performed to identify this interaction [32]. These analyses showed that 4-methylesculetin forms a complex with GR with more than one binding site close to the FAD cofactor, which was reduced by NADPH, whereas equivalents were transferred to a redox-active GSSG, stabilizing the 4-methylesculetin-GR complex with a consequent increment of the GR activity [32]. Based on this, authors demonstrate that 4-methylesculetin acts by different antioxidant mechanisms, i.e., controlling the imbalance between MPO activity and GSH production with an increment of GSH availability, upregulating the GST activity with consequent increase of electrophiles inactivation, upregulating GR activity via stabilization of its enzymatic activity, and upregulating Nrf2 expression that leads to a GR regeneration with consequent GSH maintenance levels [32].

#### 4.1.4. Daphnetin (7,8-dihydroxycoumarin)

A comparative study with several coumarin derivatives in TNBS-model of intestinal inflammation daphnetin (Figure 2) demonstrated a protective effect of daphnetin (Figure 2) in intestinal inflammation after oral administration of the lower doses (2.5 and 5.0 mg/kg) [55]. Daphentin counteracted GSH depletion and inhibited MPO activity as well as showing a potent ROS scavenging property (Figure 5) [55]. Among coumarin derivatives, daphnetin is one of the most studied compounds, with a series of pharmacological activities that corroborate its use in the inflammatory process, mainly acting on the oxidative stress and other signaling pathways of the intestinal inflammatory process, which it will bellow discussed. Antioxidant and anti-inflammatory activities have been reported by different studies, in which daphnetin was reported as a potent antioxidant compound inhibiting lipid peroxidation, scavenging free radical generation, and upregulating the Nrf2 signaling pathway [8,56,57].

#### 4.1.5. Esculin (7-hydroxy-6-O-glucosylcoumarin)

Esculin (Figure 2) promoted protective effects on the DSS- and TNBS-induced intestinal inflammation, counteracting GSH depletion, and inhibiting MPO activity [55,58]. Esculin relieved intestinal inflammatory clinical indicators and histopathological damage promoted by DSS, effects that were accompanied by a downregulation of iNOS expression [58]. In vitro studies with RAW264.7 cells stimulated by LPS demonstrated esculin reducing NO generation as well as the gene expression and protein level of iNOS [58]. Several studies corroborated the antioxidant properties of esculin and its use in different inflammatory processes, mainly acting as a potent scavenging agent [8,56], reducing MPO activity [56], NO production, and iNOS levels [59], as well as markedly activated the Nrf2 signaling pathway related to oxidative stress [60,61].

#### 4.1.6. Other Simple Antioxidant Coumarin Derivatives

A comparative and preliminary study using several simple coumarins with different substitutions in the basic ring of coumarins, including scopoletin, scoparone, fraxetin, 4-methylumbelliferone, esculin, and daphnetin (Figure 2) demonstrated differential intestinal anti-inflammatory and antioxidant properties in a TNBS-induced intestinal inflammation in rats [55]. Among these coumarin derivatives, 4-methylumbelliferone produced no effects on the clinical (damage score, extension of lesion, diarrhea, and length/weight colon ratio) and biochemical parameters such as GSH level and MPO activity. Oral administration of scopoletin (5 and 25 mg/kg), scoparone (5 and 10 mg/kg), and fraxetin (5 and

10 mg/kg) were able to counteract GSH depletion induced by intestinal inflammation with no effects on the MPO activity [55]. On the other hand, oral administration of 25 mg/kg of esculin and 2.5 and 5.0 mg/kg of daphnetin counteracted GSH depletion and inhibited MPO activity, showing daphnetin with a protective effect against intestinal inflammatory process at lower doses when compared with the other coumarin derivatives [55]. Although all coumarin derivatives acted as a radical scavenger, only fraxetin and daphnetin inhibited in vitro assay of lipid peroxidation in the cell membrane with lower inhibitory concentrations [55]. The results corroborate the hypothesis that dihydroxylated coumarins with vicinal diol functionality such as fraxetin, esculetin, and daphnetin exhibit potent ROS scavenging when compared to other coumarin derivatives [8,55].

Osthole (Figure 2) in dinitrobenzene sulphonic acid model of intestinal inflammation improved the histopathological damage and some clinical indicators of intestinal inflammation and acted as an antioxidant product, reducing malondialdehyde levels and MPO activity, increasing GPX, CAT, SOD, and GST levels, and counteracting GSH depletion [62]. In the DSS-model of intestinal inflammation in mice osthole showed protective effects on intestinal inflammation improving clinical parameters and histological damages as well as reducing MPO activity and downregulating colon TNF-α and serum TNF-α levels [63].

Antioxidant simple coumarin derivatives with intestinal anti-inflammatory activity also include isomeranzin, auraptene, and collinin (Figure 2). Isomeranzin was reported as able to inhibit NO release in RAW264.7 cells [64], whereas auraptene and collinin intestinal anti-inflammatory effects were associated with reduced iNOS levels [65]. The intestinal anti-inflammatory activities of isomeranzin, auraptene, and collinin were related to other signaling pathways of the inflammatory process [64,66].

#### *4.2. Effects of Coumarin Derivatives on Aarachidonic Acid Metabolism*

Several metabolites of the arachidonic acid metabolism pathway have pro-inflammatory properties, indicating that inhibitory action on these metabolites production can be an important target for the development of intestinal anti-inflammatory drugs. Arachidonic acid is produced from membrane phospholipids by the action of several phospholipases, manly phospholipase A2 (Figure 8). The metabolism of arachidonic acid includes several enzymes, mainly cyclooxygenase 1 and 2 (COX-1 and COX-2), and lipoxygenase 5 and 12 (LOX-5 and LOX-12), which play a relevant role in intestinal inflammation [67]. COX-1 is constitutively expressed in several cell types and produces diverse eicosanoids such as thromboxane A2 (TXA2) and prostaglandins I2 (PGI2), which have platelet and cytoprotective effects, respectively. COX-2 is a cyclooxygenase induced under inflammatory stimuli and the main source of pro-inflammatory prostaglandins, such as PGI<sup>2</sup> and PGE2, and its inhibition by different chemical agents has been considered beneficial to control the intestinal inflammatory process [3,67]. On the other hand, lipoxygenases, mainly LOX-5 is a key enzyme for the production of leukotriene B4 (LTB4), the major pro-inflammatory metabolite of arachidonic acid that contributes to the perpetuation of intestinal inflammation [68].

Some coumarin derivatives produced different inhibitory action on arachidonic acid metabolism, improving response against intestinal inflammation. Esculetin was able to reduce the COX-2 levels in the colon of rats with intestinal inflammation induced by TNBS [51] as well as inhibited LTB<sup>4</sup> and TXB<sup>2</sup> generation via an inhibitory action on the LOX-5 activity [8,69]. Similar effects were reported to esculin, daphnetin, osthole, imperatorin, auraptene, collinin, and fraxetin [8,63,65,69–72].

‐ ‐ ‐

‐

‐

‐

‐

‐

‐ ‐ ‐ ‐ ‐ **Figure 8.** Arachidonic acid metabolism and its main pro-inflammatory mediators. 5-LOX, 5 lipooxygenase; COX-1, cyclooxygenase 1; COX-2 cyclooxygenase 2; LTA, leukotriene A; LTB, leukotriene B; PGI<sup>2</sup> , prostaglandin I2, PGE<sup>2</sup> , prostaglandin E2, TXA<sup>2</sup> , thromboxane A2. Inhibitory action on the arachidonic acid metabolism was demonstrated by treatment with esculetin, esculin, daphnetin, osthole, imperatorin, auraptene, collinin, and fraxetin.

#### *4.3. Effects of Coumarin Derivatives on the Immune Response*

‐ ‐ The modulation of the immune response has been reported as an important action of coumarin derivatives to control intestinal inflammation in experimental models and in vitro studies. Generally, the available data indicated the ability of several coumarins produce effects on the production and release of immune mediators, but the general mechanism of action to produce these responses was not fully investigated. However, several intestinal anti-inflammatory coumarin derivatives described probably modulated the immune response acting on the other signaling pathways here described, particularly those including nuclear signaling pathways.

‐ Besides the antioxidant properties, paepalantine (Figure 1) was demonstrated to inhibit the production of pro-inflammatory cytokines TNF-α and IL-6 in human gastric carcinoma cells and murine macrophages RAW264.7 line [48]. Although in the acute phase of the TNBS model, 4-methylesculetin (Figure 2) produced no effects on the IL-1β, TNF-α, MMP-2, and MMP-9 protein levels, in vitro studies demonstrated that 4-methylesculetin inhibits the production of IL-1β in LPS-stimulated RAW264.7 cells, IL-8 in IL1-β-stimulated Caco-2 cells, and INF-γ and IL-2 in concanavalin a-stimulated splenocytes [53]. 4-methylesculetin treatment in DSS-model of intestinal inflammation reduced IL-6 colon levels, with no effects on the IL-17 and TNF-α colon levels [54]. Esculin (Figure 2) relieved intestinal inflammatory clinical indicators and histopathological damage promoted by DDS, effects that were accompanied by a downregulation of IL-1β, TNF-α, and reduction of IL-1β and TNF-α protein levels [59]. Esculetin (Figure 2) in a model of psoriasis-like skin diseasedramatically suppressed pro-inflammatory cytokine releases such as TNF-α, IL-6, IL-22, IL-23, IL-17α, and INF-γ [73]. A recent and interesting study in mice treated with daphnetin was carried out using different approaches to describe the protection of daphnetin (Figure 2) on the DSSinduced intestinal inflammation model [74]. The evaluation of immune and inflammatory response in this experimental model demonstrated daphnetin avoid intestinal inflammation

progression, which was related to an improvement of histopathological damage induced by DSS and modulation of pro-inflammatory mediators, downregulating colon TNFα, IL-6, IL-1β, IL-21, IL-23, CXCL1, and CXCL2 expression, and increasing IL-10 [74]. Osthole at 100 mg/kg by intraperitoneal route attenuated several clinical indicators of the intestinal inflammation as well as the histopathological lesions and alterations induced by TNBS [75]. The protective clinical effects were accompanied by a significant reduction of IL-1β, TNF-α, IL-6, CXCL10, and COX-2 gene expression as well as by an improvement of the intestinal barrier function, upregulating claudin-1 and ZO-1 mRNA [75]. In another set of evaluations using a model experimental of intestinal inflammation induced by dinitrobenzene sulphonic acid, osthole reduced TNF-α and increased IL-10, with no effects on the INF-γ levels [62]. Isomeranzin treatment with an oral dose of 30 mg/kg, isomeranzin attenuated several clinical and histopathological indicators of DSS- and TNBS intestinal inflammation as well as decreased serum IL-6 and TNF-α expression and colon IL1-β, IL-6, TNF-α, and iNOS mRNA expression [64].

#### *4.4. Effects of Coumarin Derivatives on the Nuclear Signaling Pathways*

Several drugs and natural products, including coumarin derivatives, produce intestinal anti-inflammatory activity acting on the transcription factors, nuclear receptors, and enzymes related to the inflammatory response, particularly nuclear factor-kappa b (NF-κB), peroxisome proliferator-activated receptor gamma (PPAR-γ), mitogen-activated protein kinases (MAPKs), pregnane X receptors (PXRs), rexinoid X receptors (RXRs). Other receptors such as glucocorticoid receptor (GR), farnesoid X receptor (FXR), estrogen receptor (ER), liver X receptor (LXR) regulate the inflammatory response in several diseases such as atherosclerosis, obesity, diabetes, multiple sclerosis, cancer, and IBD [76], showing that these nuclear signaling pathways are key targets for the action of new intestinal anti-inflammatory compounds.

#### 4.4.1. NF-κB and PPAR-γ Signaling Pathways

The transcription factor kappa B (NK-κB) has a central role in the intestinal inflammatory processes, triggering a high pro-inflammatory cytokines production. NK-κB signaling pathway (Figure 9) can be activated either canonical or noncanonical pathways, however, the majority of products and studies were focused on the canonical signaling pathway [77,78]. In the canonical NK-κB signaling pathway, the NK-κB heterodimer consists of the subunits p50 and p65/Rel A, which is inactive in the cytoplasm when binding with inhibitors of protein kappa B (IκB). The IκB inhibitory enzymatic complex (IKK) is composed of a regulatory IKK gamma (IKKγ) subunit and two enzymatically active subunits, IKK alpha (IKKα) and beta (IKKβ) [79]. In the canonical NK-κB signaling pathway, IKK activation occurs by specific membrane ligands such as cytokines, bacteria, bacteria metabolites, viruses, and growth factors [80]. Under this stimulation, IKKβ is activated leading to IκB phosphorylation with consequent ubiquitination and proteasome degradation [74–77], whereas IKKα is phosphorylated to activate noncanonical NK-κB pathway (Figure 9), causing p100 processing and formation of p52/RelB dimers instead of p50 and p65/Real [77,80]. The released NK-κB is promptly translocated into the nucleus to activate specific response elements in DNA, triggering a transcriptional activity with high production of diverse inflammatory mediators, mainly TNF-α, IL-1β, COX-2, IL-6, IL-8, IL-12, and IL-23 (Figure 9) [77–81].

‐κ

β κ

‐κ

‐α ‐ β ‐ ‐ ‐ ‐ ‐

γ

α

‐κ

α β

‐ ‐κ ‐ ‐γ ‐κ ‐γ **Figure 9.** The nuclear factor-kappa b (NF-κB) and peroxisome proliferator-activated receptor gamma (PPAR-γ) signaling pathway in the intestinal inflammatory process. Modulation of NF-κB and PPAR-γ signaling pathways was demonstrated by treatment with esculetin, esculin, osthole, and isomeranzin.

‐ ‐γ ‐κ ‐γ α β ‐κ ‐ γ ‐κ ‐κ ‐ ‐γ ‐ ‐γ α ‐ In several studies, the peroxisome proliferator-activated receptor gamma (PPAR-γ) has been associated with the inflammatory response coordinated by the NF-κB signaling pathway [82,83]. PPAR-γ and other PPARs, such as PPARα and PPARβ are a group of nuclear receptors that modulates glucose metabolism, adipogenesis, fatty acid synthesis as well as inhibit the NF-κB inflammatory response [83]. It has been reported that PPAR-γ deletion induces an increment of the inflammatory process in the DSS model of intestinal inflammation, whereas its activation represses the nuclear localization of NF-κB [81,83], showing NFκB-dependent response of PPAR-γ as a target for the action of intestinal anti-inflammatory compounds. PPAR-γ is a heterodimer complex with retinoid X receptor alpha (RXRα) generally binding with a co-repressor and expressed in several cells that participates in the intestinal inflammatory response such as dendritic cells, macrophages, and monocytes [83]. Under receptor activation by ligands, the co-repressor molecule displaced, whereas PPARγ/RXRα free complex binding with coactivator molecules [83].This activated complex binding to PPAR-γ response elements (PPRE) inducing transcription and protein synthesis (Figure 9). It has been reported that activated PPAR-γ/RXRα/coactivators complex can also bind with NF-κB repressing its transcriptional function with consequent reduction of pro-inflammatory cytokines production and consequent anti-inflammatory effects [81–85]. Several exogenous and endogenous PPAR-γ ligands have been reported, including fatty acids (linoleic, palmitoleic, and oleic acids), eicosanoids (eicosapentaenoic and docosahexaenoic acids, and prostaglandins), thiazolidinediones (rosiglitazone and pioglitazone), non-steroidal anti-inflammatory drugs (indomethacin and ibuprofen) as well as short-chain

fatty acids, mainly butyrate and propionate, which are produced from the fermentative process of dietary fiber and other food products by intestinal microbiota [83,86,87].

Several coumarin derivatives produced intestinal anti-inflammatory activity acting on the NF-κB and PPAR-γ signaling pathway in both in vivo and in vitro studies. Antioxidant esculetin (Figure 2) treatment of the human pancreatic cell lines resulted in a significant reduction of NF-κB levels via its binding with Keap1 regulator of the Nrf2 signaling pathway, attenuating the NF-κB activation [88]. Moreover, esculetin reduced the NF-κB p65 levels in the cell nucleus of human NB4 leukemic cell lines [52].

Esculin (Figure 2) was also able to decrease nuclear protein levels p65 from NF-κB signaling pathway both rectal tissue from the animal with DSS-induced intestinal inflammation and RAW264.7 cells [59]. Moreover, esculin suppressed the phosphorylation of IκBα, the major step of NFκB accumulation in the cell nucleus [58]. Finally, the authors elegantly demonstrated that inhibition of NFκB activation by esculin was partially mediated by the PPAR-γ stimulation, promoting nuclear localization of PPAR-γ (Figure 9) and the regulation on NFκB activation [58]. Osthole (Figure 2) was also evaluated in the DSS-model of intestinal inflammation in mice and its protective effects on intestinal inflammation were related to a downregulation of the NFκB p65 and IκB gene expression with a simultaneous effect increasing IκBα protein levels (Figure 9), suggesting that osthole at doses of 20 mg/kg inhibited NFκB activation [63]. Isomeranzin (Figure 2) treatment reduced the phosphorylation of ERK and p65 in DSS- and TNBS-induced intestinal inflammation models. In vitro studies was demonstrated isomeranzin inhibiting NF-κB activation via prevention of TRAF6 ubiquitination, a signal transductor of NF-Kb [64].

#### 4.4.2. MAPK Signaling Pathway

Mitogen-activated protein kinase (MAPK) signaling exerts several effects on cell function, including cell growth, proliferation, differentiation and survival, as well as is closely implicated in IBD, influencing the progression and perpetuation of intestinal inflammation [89,90]. MAPK activation is a response to several extracellular stimuli such as environmental stress, hormones, growth factors, and cytokines that via different kinase receptors, pathogen-associated molecular patterns, and danger-associated molecular patterns recruit pattern recognition receptors to induce a cell response [89,91]. The MAPK signaling pathway includes three groups of protein kinases, i.e., the extracellular signal-regulated kinases (ERKs), the c-Jun N-terminal kinases (JNKs), and the p38 MAPKs [89]. Phosphorylation, which occurs in a specific amino acid sequence of each group of MAPK, is pivotal for their activation [90]. Each group of MAPK is activated by different kinase pathways using distinct interlinked kinase components, as elegantly described [86]. After activation, MAPK is translocated to the nucleus to phosphorylate a series of transcription factors responsible for the expression of several genes and protein synthesis of mediators related to the inflammatory response [89–91].

MAPK signaling pathway has been related to the action of intestinal anti-inflammatory drugs, including aminosalicylates, glucocorticoids, and immunomodulators [92] as well as the target for the action of several coumarin derivatives. In mouse peritoneal macrophages, osthole (Figure 2) treatment significantly attenuated the production of the pro-inflammatory cytokines via suppressive effects on the p38 phosphorylation, suggesting its protective effects in TNBS-induced intestinal inflammation was related to the MAPK signaling pathway [75]. A similar evaluation of osthole was performed using the dinitrobenzene sulphonic acid model in rats, DSS-induced intestinal inflammation in mice, and murine macrophages [62,63]. Oral administration of 50 mg/kg of osthole reduced phosphorylation of the MAPK/p38 protein, promoting protective effects in the intestinal inflammatory process [62,63]. In vitro studies demonstrated osthole significantly reduced phosphorylation of p38/MAPK with no effects on the phosphorylation of the ERK and JNK [60], corroborating the data previously reported [62]. Differentially, isomeranzin (Figure 2) treatment in LPS-stimulated murine macrophages reduced phosphorylation of ERK with no effects of the JNK and p38 MAPKs [64].

#### 4.4.3. HIF-1α Signaling Pathway

The hypoxia-inducible factor 1 alpha (HIF-1α) is an innovative target for the action of new drugs with anti-inflammatory activity (Figure 10). Several studies with HIF-1α were performed in the last years as an attempt to explain how cells sense and to adapt to oxygen availability. These studies were recognized by the Nobel Prize of Physiology or Medicine in 2019 awarded to Kaelin, Ratcliffe, and Semenza.

‐ ‐ α **Figure 10.** The hypoxia-inducible factor 1 alpha (HIF-1α) signaling pathway of intestinal inflammation as the target for the action of esculetin.

‐ ‐ ‐κ ‐ Using human colon carcinoma HCT116 cells, esculetin was demonstrated to induce the hypoxia-inducible factor 1 alpha (HIF-1α), promote the secretion of vascular endothelial grown factor (VEGF), and inhibit HIF prolyl hydroxylases (PHD) activity [51]. This elegant study also suggested that catechol moiety in esculetin is required for HPH inhibition via competition with ascorbate and 2-ketoglutarate [51], given that several compounds containing catechol moieties such as quercetin and caffeic acid tend to activate HIF-1α [93,94]. In the intestinal inflammatory process, the epithelial cells provide barrier and transport functions, which are modulated by a series of physiological and morphological events such as mucus production, microvilli, and tight junctions. On the other hand, the high vascularization of intestinal tissue contributes to the counteraction of the high oxygen gradient from luminal anaerobic conditions to oxygenated tissue [95]. It has been considered that in acute and chronic inflammation, oxygen delivery, and oxygen availability or hypoxia is a key fac-

‐κ ‐ ‐

‐κ

‐

tor to trigger an inflammatory response [26,96]. The hypoxia signaling pathway is mainly coordinated by the HIF-1α stabilization, and in normoxia conditions, proline residues are hydroxylated by PHD action producing a complex with the Von Hippel-Landau (VHL) protein [97]. The complex HIF-VHL binds with ubiquitin, leading to proteasomal degradation of HIF-1α (Figure 10). Hypoxia signaling induces growth factors, such as transforming growth factor β (TGF-β) and VEGF binds with membrane-related tyrosine kinase receptors triggering a signaling pathway of phosphatidylinositol 3-kinase (PIP3K) with consequent serine/threonine-specific protein kinase 1(Akt1) phosphorylation (Figure 10) [98]. Under hypoxia, the activity of PHD is suppressed while phosphorylated Akt promotes the phosphorylation of mammalian target of rapamycin (mTOR) and FKBP-rapamycin associate protein (FRAP), regulating HIF-1α [98]. HIF-1α subunits translocate into the nucleus to bind with HIF-1β subunit and heterodimer HIF-α:HIF-β transcription factor complex then locate to the hypoxia-response elements (HRE) target genes (Figure 10), resulting in their transcriptional upregulation with the participation of coactivator p300/CREB binding protein (p300/CBP) [97].

#### 4.4.4. The Pregnane X Signaling Pathway

The pregnane X nuclear signaling pathway has been also reported as a target for the action of intestinal anti-inflammatory products, including coumarin derivatives such as imperatorin [99]. Nuclear pregnane X receptors (PXRs) are well-recognized for their function in the modulation of drug metabolism, acting as a flexible ligand for several products including drugs, natural and dietary products, hormones, and environmental pollutants [100]. Predominantly expressed in the intestine and liver, PXR after activation forms a heterodimer with the retinoid X receptor (RXR) [76]. This heterodimer binding to specific PXR response elements to control the gene expression of several proteins [76]. PXR agonists were demonstrated to attenuate intestinal inflammatory symptoms and to reduce intestinal permeability [101], improving epithelial barrier function via suppression of NF-κB expression that encoding pro-inflammatory cytokines [102]. PXR activation is a relevant antagonist of NF-κB transcriptional activity in the intestine during intestinal inflammation [103]. Imperatorin (Figure 2) mediated PXR activation suppressing the nuclear translocation of NF-κB and down-regulating pro-inflammatory production in DSS-induced intestinal inflammation in mice [99].

#### *4.5. Effects of Coumarin Derivatives Intestinal Microbiota*

Intestinal microbiota modulation by dietary products, mainly probiotic, prebiotic, and other natural products to improve SCFAs and other bacteria metabolites production from the fermentative process is an important approach to prevent IBD as well as to relieve symptoms of the intestinal inflammatory process. However, among all coumarin derivatives evaluated in several studies related to intestinal inflammation, only daphnetin (Figure 2) was demonstrated to act on the intestinal microbiota [74]. Daphnetin reversed DSS-induced gut dysbiosis, reducing *Bacteroides,* and increasing *Firmicutes*, which are the major SCFAs-producing bacteria [74]. Moreover, it was demonstrated that daphnetin was able to recovery zona occludens, occludin, mucin, and E-cadherin function compromised by DSS-induced intestinal inflammation, improving the intestinal epithelial integrity [74]. Using an elegant approach of the microbiota-transfer by cohousing untreated with daphnetin-treated mice, the authors reported an improvement of the clinical parameters, bacteria biodiversity, and immune response in the colon of cohousing DSS-untreated animals, when compared with DSS-inflamed mice singly housed [74]. Finally, to corroborate these data and intestinal microbiota importance in the maintenance of intestinal function, fecal microbiota from daphnetin-treated mice was transfer to mice depleted of intestinal microbiota, and the results demonstrated a remarkable improvement of disease manifestations, immune and inflammatory response when compared with the animal has received the vehicle, clearly showing that protective effects of daphnetin in intestinal inflammation, besides of its effects on the oxidative stress and immune response, were di-

rectly related of the regulation of intestinal integrity and tissue homeostasis modulated by intestinal microbiota [74]. Recently, daphnetin was also demonstrated to improve the altered intestinal microbiota composition of the glucocorticoid-induced osteoporosis rats, attenuating the intestinal barrier dysfunction [104]. Although daphnetin is the only coumarin whose intestinal anti-inflammatory activity has been directly associated with intestinal microbiota modulation, other natural and synthetic coumarin derivatives and plant extracts containing coumarins [105–111] were able to differentially modulate some pathogenic intestinal bacteria, but with no direct evidence and correlation with intestinal anti-inflammatory activity.

#### **5. Conclusions and Perspectives**

This review provided a general overview of the various coumarin derivatives with potential therapeutic applications on the intestinal inflammatory processes highlighting the ones for which the mechanism of action is at least partially defined and can serve for the design of further preclinical and clinical studies to support the use and application of coumarin derivatives as complementary therapies against IBD. In general, the mechanisms of action of coumarin derivatives observed in experimental models of intestinal inflammation and in vitro studies are similar to those described for other natural products such as flavonoids, anthocyanidins, and catechins. Although several coumarin derivatives such as paepalantine, 4-methylesculetin, daphnetin, esculetin, and osthole produce intestinal anti-inflammatory effects in lower doses when compared with other phenol compounds, it is no possible to attribute advantages in the use of these coumarins due to the lack of clinical trials and more detailed studies on efficacy and safety with these compounds. Protective effects of coumarin derivatives are related to antioxidant properties, similar to those produced by several phenolic compounds. However, some coumarins also interact with several endogenous macromolecules, different cell types, and signaling pathways as well as in innovative molecular targets. On the other hand, further studies are needed into the effects of some coumarin derivatives on the course of the disease, mechanisms of action, ability to modulated intestinal microbiota and intestinal permeability, and safety for use. Clinical trials in patients with IBD are very important to generate data for a potential application of coumarins derivatives as a complementary therapy for this chronic disease. There is scientific evidence here reported to support the suggestions of some coumarin derivatives as candidates for further pre-clinical studies and clinical trials, particularly those better studied, mechanism of action partially defined and with protective effects in lower doses, such as esculetin, 4-methylesculetin, osthole, and daphnetin.

**Funding:** The research in the Laboratory of Phytomedicines, Pharmacology, and Biotechnology (PhytoPharmaTech) has been supported by the São Paulo Research Foundation (FAPESP) and National Council for Scientific and Technological Development (CNPq).

**Conflicts of Interest:** The author declares no conflict of interest.

#### **References**


**Cristina Zaragozá \*, Francisco Zaragozá , Irene Gayo-Abeleira and Lucinda Villaescusa**

Pharmacology Unit, Biomedical Sciences Department, University of Alcalá, Alcalá de Henares, 28871 Madrid, Spain; francisco.zaragoza@uah.es (F.Z.); irene.gayo@uah.es (I.G.-A.); lucinda.villaescusa@uah.es (L.V.)

**\*** Correspondence: cristina.zaragoza@uah.es

**Abstract:** Atherosclerotic cardiovascular disease is the leading cause of death in developed countries. Therefore, there is an increasing interest in developing new potent and safe antiplatelet agents. Coumarins are a family of polyphenolic compounds with several pharmacological activities, including platelet aggregation inhibition. However, their antiplatelet mechanism of action needs to be further elucidated. The aim of this study is to provide insight into the biochemical mechanisms involved in this activity, as well as to establish a structure–activity relationship for these compounds. With this purpose, the antiplatelet aggregation activities of coumarin, esculetin and esculin were determined in vitro in human whole blood and platelet-rich plasma, to set the potential interference with the arachidonic acid cascade. Here, the platelet COX activity was evaluated from 0.75 mM to 6.5 mM concentration by measuring the levels of metabolites derived from its activity (MDA and TXB<sup>2</sup> ), together with colorimetric assays performed with the pure recombinant enzyme. Our results evidenced that the coumarin aglycones present the greatest antiplatelet activity at 5 mM and 6.5 mM on aggregometry experiments and inhibiting MDA levels.


**Citation:** Zaragozá, C.; Zaragozá, F.; Gayo-Abeleira, I.; Villaescusa, L. Antiplatelet Activity of Coumarins: In Vitro Assays on COX-1. *Molecules* **2021**, *26*, 3036. https://doi.org/ 10.3390/molecules26103036

Academic Editor: Maria João Matos

Received: 15 April 2021 Accepted: 14 May 2021 Published: 19 May 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

**Keywords:** coumarin; esculin; esculetin; antiplatelet activity; impedance aggregometry; COX; polyphenols

#### **1. Introduction**

Platelets present a wide variety of functions in the blood circulation, with a key role in the development of the atherosclerotic process and the subsequent physiopathology of the cardiovascular disease [1]. Once attached to the vascular endothelium and activated, the platelets release a broad range of molecules such as chemokines, proinflammatory agents and different substances able to modulate a biological response that will promote the interaction among platelets, endothelial cells and leukocytes [2]. These cell interactions trigger a local inflammatory response, which is mainly responsible for the atherosclerotic process [3]. Platelet adhesion to the luminal vascular surface occurs after exposure of the endothelium caused by a lesion or detachment of an atherosclerotic plaque. Platelet aggregation represents the initial stage in the formation of a blood clot that can lead, to a greater or lesser extent, to the vascular occlusion and eventually result in thromboembolic disease such as stroke or myocardial infarction [4].

An irreversible stage of platelet aggregation is mainly induced by the secretion of substances from the platelet granules content. This event has also been observed in vitro as a response to the addition of high concentrations of agonists. The process includes the formation of metabolites mostly derived from arachidonic acid (cyclic endoperoxides and TXA2) and the secretion of the content from lysosomes and dense and α-granules in platelets [5]. Coumarins (2*H*-1-benzopyran-2-ones), the lactones of the 2-hydroxy-Zcinnamic acids, are phenolic compounds with complex structures that differ substantially across the family [6] and are extensively distributed in the plant kingdom, especially in the families *Apiaceae, Asteraceae* and *Rutaceae* [7]. Naturally occurring coumarins, even though all of them contain the coumarin moiety, are structurally different and can be

classified according to their chemical structure in the following groups: simple coumarins, furanocoumarins, dihydro-furanocoumarins, phenylcoumarins, pyranocoumarins and dicoumarins [8].

Simple coumarins are usually substituted at position 7 (C-7) with a hydroxyl but can be also hydroxylated at positions 6 and 8. These hydroxyl groups can be sometimes methylated or substituted with sugar molecules, in which case they are referred to as glycosylated or heterosidic coumarins [9]. The presence of the different substituents in the main structure largely influences the biological activity of the resulting compound [10].

Coumarin (2*H*-1-benzopyran-2-one) (Figure 1) has been under research due to its interesting and wide-ranging bioactivities, inclusive of anti-inflammatory [11,12], antioxidant [7], antimicrobial [13–15], antiproliferative [16,17] and anticoagulant properties [18]. The vitamin K antagonists in clinical use are structurally derived from 4-hydroxycoumarin and share a common mechanism of action in that they noncompetitively inhibit the vitamin K epoxide reductase complex, which is essential in the recycling of vitamin K in the liver. As vitamin K serves as a cofactor in the activation of clotting factors II, VII, IX and X, the inhibition of its recycling results in strong anticoagulation activity [19].

**Figure 1.** Chemical structure of the different coumarins assayed: esculin (**a**), esculetin (**b**) and coumarin (**c**).

Esculin (7-hydroxy-6-[(2 S, 3 R, 4 S, 5 S, 6 R)-3,4,5-trihydroxy-6-(hydroxymethyl) oxane-2-yl] oxychromen-2-one) (Figure 1) is a coumarin derivative found in *Aesculus hippocastanum* L. (horse chestnut) [20] that has demonstrated promising anti-inflammatory, antioxidant and free radical scavenging properties. This compound was effective in diminishing the elevated blood creatinine levels in diabetic mice, which ameliorated diabetes-induced renal dysfunction through a reduction on the activation levels of caspase-3 in the mice kidney [21]. Likewise, esculin showed a protective effect against lipid metabolism disorders in diabetic rats in a dose-dependent manner. The authors of this study proposed that the possible mechanism might be associated with the inhibition of AGE (advanced glycation end products) formation [22].

Conversely, esculetin (6,7-dihydroxychromen-2-one) (Figure 1) is the aglycone of the heteroside esculin. This compound has been thoroughly investigated because of its anti-inflammatory activity, which is conducted through several mechanisms that include the inhibition of ICAM-1 release, the decrease of NO and PGE<sup>2</sup> levels in synovial fluid, myocardial protection or the inhibition of proinflammatory cytokines during the interaction between adipocytes and macrophages [23]. Some evidence for the potential of this aglycone to decrease oxidative stress has also been demonstrated [24], together with the presence of antidiabetic [25], antibacterial [26] and antitumor activities [27].

Despite the significant number of studies based on these types of chemical compounds and the diverse biological activities described for coumarin, esculetin and esculin, the mechanisms of action remain partially unknown. This research work focuses on demonstrating the antiplatelet activity of these coumarins and shedding some light on their mechanism

of action. Due to the important role of the cyclooxygenase (COX) enzyme in platelet aggregation, it has been hypothesized that the potential interaction with COX is a possible mechanism through which coumarins could exert their antiplatelet function.

#### **2. Results**

#### *2.1. Antiaggregant Effect of Coumarins by Impedance Platelet Aggregometry*

The percentage of platelet aggregation in whole blood (WB) and platelet-rich plasma (PRP) after activation by adenosine diphosphate (ADP) or arachidonic acid (AA) was calculated. Maximal aggregation (100%) was considered when ADP or AA were used in absence of any other compound. All the assayed phenolic compounds were tested at different concentrations: 0.75, 1.5, 3.0, 5.0, 6.5 mM. These concentrations are similar to the daily dose clinically used for the flavonoid diosmin (Daflon® 500 mg) [28]. The percentages of aggregation for coumarin, esculin and esculetin are shown in Figure 2.

**Figure 2.** Graphical representation of the percentage of platelet aggregation for coumarin, esculin and esculetin. Panels (**A**) (WB samples) and (**C**) (PRP samples) shows results in AA-induced platelet aggregation. Panels (**B**) (WB samples) and (**D**) (PRP samples) shows results in ADP-induced platelet aggregation. Results are expressed as the mean and standard deviation for 10 donors. Error bars represent the standard deviation. \* *p* < 0.05: statistically significant differences in platelet aggregation between samples with and without the tested phenolic compound.

In general, the antiplatelet effect of the screened phenolic compounds was observed at concentrations equal to or higher than 3 mM and was more potent in samples subjected to AA-induced platelet aggregation (Figure 2A,C) than in those subjected to ADP-induced platelet aggregation (Figure 2B,D), reaching on WB AA-induced experiments a IC50 for coumarin and esculetin of 2.45 mM and 3.07 mM respectively, and 5.12 mM and 5.82 mM on ADP experiments. The IC50 on PRP AA-induced samples were 1.12 mM for coumarin and 2.48 mM and 5.08 mM and 5.97 mM for ADP-induced assays.

The effects of coumarin and esculetin as antiaggregant agents were especially relevant in all the experiments performed. The complete inhibition of the platelet aggregation was achieved in AA-induced activated PRP samples after addition of 1.5 mM of coumarin (Figure 2C).

#### *2.2. Platelet MDA Levels and COX Activity*

Considering that the extent of inhibition was greater in the impedance aggregometry assays when AA was used as activator (Figure 2A,C) [29], the MDA levels were quantified

in AA-induced activated PRP samples. Maximal activity of COX (100%) was set as that obtained for AA-treated samples in the absence of phenolic compound. Indomethacin was used as a positive control at the doses shown below (Figure 3).

**Figure 3.** Graphical representation of the percentage of COX activity in AA-induced activated PRP samples after addition of increasing concentrations of indomethacin, as positive control, and assayed coumarins. Results are expressed as the mean and standard deviation for 10 donors. Error bars represent the standard deviation. \* *p* < 0.05: significant differences regarding COX activity with and without the examined substances.

> The inhibition of COX activity was confirmed in presence of indomethacin (Figure 3). Coumarin and esculetin showed their ability to inhibit COX-1 activity, while esculin did not affect it significantly (Figure 3). The IC50 for coumarin and esculetin were 5.93 mM and 2.76 mM for coumarin and esculetin, respectively.

#### *2.3. COX-1 Inhibitory Assay*

With the aim to determine the direct effect of the compounds here investigated on the COX-1 activity, analyses with the pure human recombinant enzyme (h-COX-1) were performed. Results were expressed as the percentage of COX-1 activity in the presence of indomethacin (as a positive control) or after incubation with the different tested molecules.

As it can be observed in Figure 3, indomethacin diminished almost completely the activity of the recombinant COX-1 enzyme. Coumarin produced a decrease of 49% in h-COX-1, while esculetin barely reached a drop of 42%. By contrast, esculin demonstrated the greatest COX-1 inhibitory effect, with a 74% of enzyme inhibition rate (Figure 4) and IC50 of 4.49 mM.

#### *2.4. TXB<sup>2</sup> Levels as COX-1 Activity Indicator*

TXB<sup>2</sup> quantification was conducted in WB and used as an indicator of the COX-1 activity since it is assumed that the administration of a COX inhibitor will decrease the levels of TXB2. The effect of indomethacin as a positive control was analysed at different concentrations. Meanwhile, the calcium ionophore (CI) (25 mM) was employed as a platelet aggregation inducer to evaluate the potential effect of the assayed phenolic compounds.

The results demonstrated the inhibition of the TXB<sup>2</sup> production by indomethacin at all the concentrations tested. On the contrary, any of the phenolic compounds investigated exerted any effect on the levels of this metabolite (Figure 5).

**Figure 4.** Graphical representation of the percentage of h-COX-1 activity after indomethacin and tested coumarins addition in AA-induced activated samples. Results are expressed as the mean and standard deviation for 10 donors. Error bars represent the standard deviation. \* *p* < 0.05: significant differences on h-COX-1 activity with and without the examined substances.

**Figure 5.** Graphical representation of the percentage of TXB<sup>2</sup> production at different concentrations of indomethacin and coumarins after CI-induced aggregation. Results are expressed as the mean and standard deviation for 10 donors. Error bars represent the standard deviation. \* *p* < 0.05: significant differences between the basal TXB<sup>2</sup> production with and without the examined substances.

#### **3. Discussion**

In this research study, the potency of the aglycones coumarin and esculetin as antiplatelet aggregation agents was evidenced by the impedance aggregometry assays. The effect observed was similar in WB and PRP, albeit the antiaggregant activity was superior in the latter. In both cases, the effect was more remarkable in AA-induced than in ADPinduced activated samples. Coumarin and esculetin showed an inhibition of 87% and 52%, respectively, in AA-induced samples at a concentration of 3 mM, whereas the inhibition was complete with highest concentrations.

With respect to coumarin, this compound was able to completely inhibit platelet agglutination in AA-induced PRP samples when a 1.5 mM solution was used. Regarding the effect of esculetin, the inhibition rate achieved a 16% at 3 mM, and was complete at 6.5 mM. The esculetin, in its heterosidic form as esculin, demonstrated a minimal decrease in platelet aggregation at 6.5 mM that was null at lower concentrations in WB and PRP, in either AA- or ADP-induced activated samples. Thus, our results revealed that the presence of the catechol group in esculetin favours the antiplatelet activity, while this activity is lost when the C-6 hydroxyl group is replaced by a sugar, such as in esculin.

The higher antiaggregant potency of aglycones vs. heterosides was confirmed in the COX-1 activity assay from the measurement of MDA when AA was employed as aggregant agent in PRP samples. Esculetin showed a 90% of inhibition of COX-1 activity at 6.5 mM, which was similar to the inhibition reached with the positive control indomethacin. Unlike esculetin, coumarin just showed a 60% of inhibition at the maximal concentration assayed and no significant effect was found for esculin.

However, the results obtained in the COX-1 inhibitory assay performed with the human pure recombinant enzyme were somewhat different. In this case, the heteroside esculin procured the greatest inhibitory effect on h-COX-1 since the enzyme activity decreased up to the 26%. Nevertheless, coumarin produced a 50% enzyme inhibition in a similar way to the inhibition showed by the MDA measurement assay. Esculetin, that had previously shown a 90% of COX-1 inhibition, returned a 57% of enzyme inhibition when tested with the pure enzyme. Regarding the indomethacin (positive control), the results with the pure enzyme seem to be more relevant than those related to the MDA production, since the enzyme inhibition reached a 95%.

Surprisingly, any of the coumarins investigated had an impact in TXB<sup>2</sup> levels. Indomethacin, for its part, prevented TXB<sup>2</sup> production in a 95%. Hence, it can be assumed that this molecule significantly inhibits COX-1 activity.

Coumarin and esculetin presented a dose-dependent effect on the platelet aggregation, the COX-1 activity and the pure enzyme. Higher concentrations of coumarins than those selected in this research work might exert a larger effect, but the low solubility of these compounds is a major limitation. Even though DMSO is an optimal solvent for these molecules, the use of higher amounts could affect the platelets integrity and is discouraged [30,31]. Previous experiments in our laboratory showed how a higher volume than 2 µL of DMSO could damage platelets contained in 1 mL of WB [32].

Polyphenols are naturally present in plants as O- and C-glycosides, while aglycones are not found in fresh plants but can occur after processing [33]. In general, the oral bioavailability of polyphenols is considerably limited [34,35]. As a consequence of enzyme hydrolysis, the heterosides lose the glycosidic moiety before reaching the bloodstream and can then pass through the cell membranes [36].

As previously indicated, our in vitro results show that in the specific case of coumarins, the aglycones present a greater antiplatelet effect than their heterosidic parent compounds.

Esculin did not show any activity in our in vitro experiments apart from those performed with the pure enzyme, which suggests the inability of this compound to access the platelet interior. This fact could be explained by the presence of a sugar ring in its chemical structure. Considering that glycoxidation favours the biological activity of coumarins [37], it could be hypothesized that esculin would present a similar activity to its aglycone in vivo.

The smaller size of the coumarin and esculetin structures could ease their transport across the platelet membrane, and hence, produce a higher effect on the COX activity. However, our results support the feasibility of and need for future studies on the interaction of the coumarins with blood platelet membrane. Notwithstanding, these two molecules were not able to inhibit TXB<sup>2</sup> production. Thereby, the results here presented point to a mechanism of action at a different level that would possibly involve TXA<sup>2</sup> receptors.

#### **4. Materials and Methods**

#### *4.1. Selected Compounds*

The selected phenolic compounds coumarin, esculin and esculetin were purchased from Sigma-Aldrich (Sigma-Aldrich Chemical, Madrid, Spain) and dissolved in DMSO (dimethylsulphoxide) (Dismadel, Madrid, Spain) to a final concentration of 0.5 mM. This concentration is similar to the daily dose clinically used for the flavonoid diosmin (Daflon® 500 mg) [28] and was established considering the structural similarity (presence of the benzo-α-pyrone core), the almost identical physicochemical properties and their comparable molecular weight between this drug and the compounds here investigated [38]. Further dilutions were performed to reach 0.75, 1.5, 3.0, 5.0, 6.5 mM for the different assays. To

avoid altering the platelet configuration, the lowest volume of DMSO (Dismadel S.L., Madrid, Spain) that could ensure the dissolution of the compounds (2 µL) was added to the blood samples [32].

#### *4.2. Study Cohort, Inclusion and Exclusion Criteria*

Ten healthy volunteers (seven women and three men; aged 22.2 ± 1.2 [mean ± SD] years), none of which had undergone platelet function or complement activation treatment during the previous year, were recruited to furnish blood for every assay of this work. Participants were not included if they were smokers or showed any sign of kidney, lung, heart, or autoimmune disease, any chronic or acute infection, diabetes mellitus, a history of tumours, immunodeficiency or thrombocytopathy, hypercholesterolemia or were undergoing immunosuppressant, steroids, or nonsteroidal anti-inflammatory drug (NSAID) treatment. They were excluded if they had undergone any other treatment that could affect the platelet activity during the six months prior to the assay, anovulants included.

Written informed consent was signed by every participant. The study protocol was carried out in strict accordance with the guidelines of the 1975 Declaration of Helsinki, under approval of the Biomedical Ethics Committee of the University of Alcalá.

#### *4.3. Peripheral Blood Extraction*

Peripheral blood was collected by an antecubital puncture in sodium citrate-containing (3.8% wt/vol) Vacutainer® tubes (Dismadel S.L., Madrid, Spain), discarding the first 2 mL. All extractions were performed at the Haematology Service of the Principe de Asturias Hospital, Alcalá de Henares (Madrid, Spain). Sodium citrate was selected as the anticoagulant instead of heparin, ethylenediaminetetraacetic acid (EDTA), or D-phenylalanyl-L-prolyl-Larginine chloromethyl ketone (PPACK) given its lesser impact on the complement activation pathways [39].

#### *4.4. Blood Samples Preparation*

Platelet aggregation assays were performed in both, WB and platelet-rich plasma PRP [40]. The removal of blood cell components in PRP allows to better evaluate the effect of a specific compound on the platelets. Two different platelet activity inducers were employed: AA (Sigma-Aldrich Chemical, Madrid, Spain) 0.5 mM [41] and ADP (Sigma-Aldrich Chemical, Madrid, Spain) 5 µM [42]. The use of two different substances (AA and ADP) that promote platelet aggregation through different pathways, together with their administration on the different types of samples (WB and PRP), provides with a better understanding on the level of action of the different compounds under research. It was considered that the screened compounds could present different activities depending on the medium or the aggregation inducer.

#### 4.4.1. WB Samples Preparation

Samples were kept at room temperature until use. They were homogenized in a plastic beaker and aliquots of 500 µL were distributed in aggregation Chronolog polyethylene cuvettes (Labmedics, Oxfordshire, UK) as soon as possible. After that, 500 µL of physiological saline solution (PSS) (Dismadel S.L., Madrid, Spain) were added to each cuvette. The diluted samples were incubated at 37 ◦C for 1 h with the selected compounds, or DMSO in the case of the control sample, in a thermostatic bath Unitronic 320 Selecta (Tecnylab, Madrid, Spain) to increase their solubility. Sample preparation and subsequent assays were performed in the first 3 h after blood extraction.

#### 4.4.2. PRP Samples Preparation

Blood samples were subjected to centrifugation for 10 min at 1200 rpm twice in a centrifuge Jouan B-3.11 (Tecnylab, Madrid, Spain) and PRP was obtained by collecting the supernatant. Platelet counts were normalized to 200,000 platelets/µL PRP. Briefly, PRP was dissolved in the hematologic solvent Diluid 601 (Biolab Diagnostics, Barcelona, Spain) and

counting was performed in a Neubauer cell chamber using a binocular microscope NIKON (Izasa S.L., Madrid, Spain). This method avoids some potential errors linked to automated cell counters, such as detection of bubbles as particles, the counting of cell components other than platelets or counting groups of platelets present in the sample. The refringence and morphology of platelets under the light microscope facilitated their unambiguous identification. Once the number of platelets in the original PRP samples was known, the calculated volume was transferred to the cuvettes and PSS was added until a final volume of 1 mL. Next, samples were incubated with the assayed compound or DMSO as a control for 5 min at 37 ◦C.

#### *4.5. Impedance Platelet Aggregometry Assay*

The procedure was carried out in a Chrono-Log 500 Lumi-Aggregometer (Labmedics, Oxfordshire, UK) connected to an Omnioscribe II data-logger, according to the manufacturer's instructions. Only plastic material was used to be in contact with the samples. The experimental method is based on the measurement of the change in the electrical impedance (the resistance to the electric current) between two electrodes when platelet aggregation is induced by an agonist [40]. Thereby, the electrodes immersed in the WB or PRP samples continuously stirred at 1200 rpm become covered by a platelet monolayer. The impedance remains constant in the absence of an aggregation agent. On the contrary, the addition of an aggregant promotes the adhesion and agglutination of the platelets in the electrodes and produces an increase in the impedance that can be used as a measurement of the platelet aggregation.

#### *4.6. MDA Quantification and COX Activity Assessment*

Plasma MDA levels reflect COX activity and can be used as a qualitative test of platelet function or to quantify the effect of COX inhibitors. MDA is a product of the arachidonic acid metabolism in platelets that can be measured by spectrophotometric techniques. Absorbance is recorded at 532 nm to ensure that it is entirely due to the released MDA. The molar extinction coefficient of MDA at 532 nm is 1.56 × 10<sup>5</sup> [29].

MDA analysis was performed in AA-induced activated PRP samples [31], since the MDA absorbance levels in the ADP-induced activated samples were much lower than those corresponding to total COX activity.

Calibration curves were created by preparing a set of solutions with known concentrations of MDA and measured in a UV/VIS Philips PU 8700 spectrophotometer at 532 nm to later extrapolate the MDA levels in PRP. The curves were prepared in PSS and PRP without aggregant to establish the possible MDA release by nonactivated platelets.

Curves ranged from 100–1000 nM to 1–10 µM and had regression coefficients near 1 (0.997 y 0.994, respectively) (Figure 6).

**Figure 6.** Graphical representation of MDA absorbance in PSS y PRP. Panel (**A**): in the range from 100 nM to 1000 nM and linear fitting in PRP samples (r = 0.997). Panel (**B**): in the range from 1 µM to 10 µM and linear fitting in PRP samples (r = 0.994).

−

′ ′

Test samples were processed following a similar procedure to the samples used to obtain the calibration curves. A volume of 375 µL of 40% trichloroacetic acid (ATA) (Sigma-Aldrich Chemical, Madrid, Spain) was added to propylene tubes containing the AA-induced activated PRP samples and gently mixed to mediate the protein precipitation. The tubes were covered to prevent oxidation. PSS was added up to a final volume of 2 mL, like in the samples used in the calibration curves. After centrifugation at 3500 rpm for 10 min, the supernatant was filtered through glass wool. One more centrifugation step was performed in similar conditions and then 0.12 M tiobarbituric acid (TBA) (Sigma-Aldrich Chemical, Spain) was added in a relation of 0.2 volumes per volume of acid supernatant.

Next, the capped tubes were heated in a water bath at 100 ◦C for 15 min. Once cooled down, the spectrophotometric measures were obtained at 532 nm.

MDA concentrations in control samples were considered as 100% of COX-1 activity and the concentrations in the test samples were expressed as the percentage of COX-1 activity. Indomethacin (Sigma-Aldrich Chemical, Madrid, Spain) was employed as positive control at the following concentrations: 0.00257, 0.006, 0.013, 0.019, 0.02 mM. These concentrations were set considering that the therapeutic concentration is 0.004 mg/mL [43]. The assayed compounds were dissolved in DMSO (2 µL) and added to the reaction medium at the concentrations previously mentioned: 0.75, 1.5, 3.0, 5.0, 6.5 mM.

#### *4.7. Procedure for the COX-1 Inhibitory Assay*

Human purified COX-1 enzyme with a purity of 95% was purchased from Vitro S.A. (Madrid, Spain). The enzyme was supplied in 10KU vials prepared in Tris-HCl 80 Mm and 1% Tween 20. The enzyme unit (EU) is defined as the amount of enzyme required to produce a change of 0.001 mn−<sup>1</sup> in the optical density at a wavelength of 610 nm. According to the manufacturer, the vials were stored frozen and kept in the dark on ice during the assays. One hundred EU (4 µL) were added to the samples. AA was used as enzyme substrate to replicate the aggregation inducer employed in the impedance aggregometry and MDA quantification experiments.

The enzyme activity was determined by a chromogenic method based on the oxidation of N,N,N′ ,N′ -tetramethyl-p-phenylenediamine (TMPD) [44]. Among the substances produced during AA-induced platelet activation, the prostaglandin G<sup>2</sup> (PGG2) is quickly reduced to PGH<sup>2</sup> by the platelet enzyme COX-1. Because of this reduction, the TMPD is oxidized in a directly proportional amount to the enzyme activity.

The experimental studies were carried out in solutions containing the pure enzyme incubated with the test compounds and AA as the enzyme substrate. Absorbance produced by TMPD was measured at 610 nm in a Biotek ELx800 Absorbance Microplate Reader (Izasa Scientific, Madrid, Spain).

The screened compounds were prepared in DMSO at the concentrations abovementioned for the previous assays. Indomethacin, a selective COX-1 inhibitor, was used as control drug at the therapeutic concentrations of 0.001, 0.0025, 0.005, 0.0075, 0.01mg/mL [43].

#### *4.8. Enzyme Immunoassay for the Quantitative Determination of TXB<sup>2</sup>*

TXA<sup>2</sup> is produced from AA oxidation and physiologically active. However, it is rapidly hydrolysed (average life of 30 s) to form TXB2, a stable and biologically inactive metabolite [45]. TXB<sup>2</sup> concentration, as measured by immunoassay, is maximal at 20–30 min and declines thereafter [46], being considered as a measurement of the TXA<sup>2</sup> levels. For this reason, the samples were incubated in WB and calcium ionophore A23187 (CI) (Sigma Aldrich, Madrid, Spain) was added at a concentration of 25 mM [32] to trigger platelet activation that produces TXB2. In this way, it can be considered as an indirect measure of the COX-1 activity inhibition [46].

TXB<sup>2</sup> levels were determined by a specific enzyme immunoassay kit (TXB<sup>2</sup> Biotrak Enzymeimmunoassay System, Amersham Biosciences, Little Chalfont, Buckinghamshire, UK) according to the manufacturer´s protocol. This kit possesses a high sensitivity (0.2 pg) and the standard curve ranges from 0.5 to 64 pg. The absorbance values were obtained in a

Biotek ELx800 Absorbance Microplate Reader (Izasa Scientific, Madrid, Spain) coupled to an automatic microplate washer Biotek ELx50 (Izasa Scientific, Madrid, Spain).

Similarly to prior assays, indomethacin was used as control drug because of its ability to inhibit COX [43].

To perform the procedure, the total volume of WB from donors was distributed in equal aliquots of 1 mL and incubated in a thermostatic bath (Unitronic 320 Selecta) (Izasa Scientific, Madrid, Spain) at 37 ◦C for 1 h with 2 µL of the test solutions or DMSO as control. After that time, 2 µL of CI 25 mM were added and incubation maintained for a further 30 min cycle. The reaction was terminated by introducing the samples on dry ice. Then, the samples were centrifuged (centrifuge Jouan 3.11) (Tecnylab, Madrid, Spain) at 4000 rpm for 10 min and the supernatant collected and subjected to the enzyme-linked immunosorbent technique.

#### *4.9. Statistical Analysis*

All results are expressed as the mean ± standard deviation (SD) of values obtained in each experiment. Since most variables did not fulfil the normality hypothesis, the Wilcoxon test was used to analyse the variance of paired groups. The level of significance was set at *p* < 0.05. Statistical analysis was performed using SPSS-27.0 software (SPSS-IBM, Armonk, NY, USA).

**Author Contributions:** Conceptualization, C.Z. and F.Z.; methodology, C.Z. and L.V.; software, L.V.; formal analysis, F.Z., L.V. and C.Z.; investigation, L.V. and C.Z.; resources, L.V. and C.Z.; data curation, L.V.; writing—original draft preparation, F.Z.; writing—review and editing, C.Z., I.G.-A. and L.V.; supervision, F.Z.; funding acquisition, C.Z. and F.Z. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by University of Alcalá de Henares-Reig Jofré Art. 83 LOU, grant number (154/2018) titled: "Asesoramiento en materia de medicamentos y programas de investigación, desarrollo e innovación".

**Institutional Review Board Statement:** This assay was leaded following the guidelines of the 1975 Declaration of Helsinki, with approval of the Biomedical Ethics Committee of the University of Alcalá. General Ethic Code of the University of Alcalá approved by the Governing Council on 22 June 2017.

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study. Written informed consent has been obtained from the patients to publish this paper.

**Acknowledgments:** We thank the Haematology Service of the Principe de Asturias Hospital for their help with blood extractions. We appreciate the technical support.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

**Sample Availability:** Samples of the compounds are not available from the authors.

#### **References**


## *Article* **Antiglioma Potential of Coumarins Combined with Sorafenib**

#### **Joanna Sumorek-Wiadro <sup>1</sup> , Adrian Zaj ˛ac <sup>1</sup> , Ewa Langner <sup>2</sup> , Krystyna Skalicka-Wo ´zniak <sup>3</sup> , Aleksandra Maciejczyk <sup>1</sup> , Wojciech Rzeski 1,2 and Joanna Jakubowicz-Gil 1,\***


Academic Editor: Maria João Matos

Received: 28 September 2020; Accepted: 6 November 2020; Published: 8 November 2020

**Abstract:** Coumarins, which occur naturally in the plant kingdom, are diverse class of secondary metabolites. With their antiproliferative, chemopreventive and antiangiogenetic properties, they can be used in the treatment of cancer. Their therapeutic potential depends on the type and location of the attachment of substituents to the ring. Therefore, the aim of our study was to investigate the effect of simple coumarins (osthole, umbelliferone, esculin, and 4-hydroxycoumarin) combined with sorafenib (specific inhibitor of Raf (Rapidly Accelerated Fibrosarcoma) kinase) in programmed death induction in human glioblastoma multiforme (T98G) and anaplastic astrocytoma (MOGGCCM) cells lines. Osthole and umbelliferone were isolated from fruits: *Mutellina purpurea* L. and *Heracleum leskowii* L., respectively, while esculin and 4-hydroxycoumarin were purchased from Sigma Aldrich (St. Louis, MO, USA). Apoptosis, autophagy and necrosis were identified microscopically after straining with specific fluorochromes. The level of caspase 3, Beclin 1, PI3K (Phosphoinositide 3-kinase), and Raf kinases were estimated by immunoblotting. Transfection with specific siRNA (small interfering RNA) was used to block Bcl-2 (B-cell lymphoma 2), Raf, and PI3K expression. Cell migration was tested with the wound healing assay. The present study has shown that all the coumarins eliminated the MOGGCCM and T98G tumor cells mainly via apoptosis and, to a lesser extent, via autophagy. Osthole, which has an isoprenyl moiety, was shown to be the most effective compound. Sorafenib did not change the proapoptotic activity of this coumarin; however, it reduced the level of autophagy. At the molecular level, the induction of apoptosis was associated with a decrease in the expression of PI3K and Raf kinases, whereas an increase in the level of Beclin 1 was observed in the case of autophagy. Inhibition of the expression of this protein by specific siRNA eliminated autophagy. Moreover, the blocking of the expression of Bcl-2 and PI3K significantly increased the level of apoptosis. Osthole and sorafenib successfully inhibited the migration of the MOGGCCM and T98G cells.

**Keywords:** osthole; umbelliferone; esculin; 4-hydroxycoumarin; sorafenib; apoptosis; autophagy

#### **1. Introduction**

Coumarins, classified as secondary metabolites, constitute a large group of ubiquitous compounds in the plant world. Depending on the chemical structure, simple coumarins, pyranocoumarins, and furanocoumarins can be distinguished [1,2].

Coumarin derivatives exhibit a wide spectrum of biological activity. Research conducted to date indicates a beneficial effect of coumarins on the central nervous system (analgesic, anticonvulsant, antidepressant, and sedative) and the circulatory system (anticoagulant and antihypertensive effect). They show antioxidant, antibacterial, antifungal, anti-inflammatory, antiallergic, and antiviral activity. Their antitumor activity is particularly important as well. These compounds have been shown to act at various stages of carcinogenesis. They exhibit chemopreventive properties as well as cytotoxic and antiproliferative activity against cancer cells. In addition, they limit angiogenesis and prevent the formation of metastases to other tissues [1–5]. Clinical studies have shown that coumarins have promising activity against several types of cancer, such as breast cancer, lung cancer, malignant melanoma, prostate cancer and renal cell carcinoma [6,7]. Simple coumarin derivatives improved the health condition of patients and did not show any toxic properties. Renal cell carcinoma patients tolerated a wide spectrum of coumarin doses, and the most common side effect was nausea associated with the intense aroma of the compound [7,8]. Interestingly, previous studies have also shown that coumarins may be used not only in the treatment of cancer but also in the treatment of the side effects of radiation therapy, such as radiogenic sialadenitis and mucositis [9].

The cytotoxicity of coumarins towards cancer cells depends on their chemical structure; therefore, knowledge of the effect of various substituents on the antitumor properties of these compounds will ensure in more effective plans of therapeutic strategies. Special attention has been paid to simple coumarins, e.g., 4-hydroxycoumarin, umbelliferone, esculin, and osthole, differing in their location or the type of attached substituents (Figure 1). 4-hydroxycoumarin and umbelliferone (7-hydroxycoumarin) are isomers with a hydroxyl moiety located at the C4 and C7 positions of the coumarin ring, respectively. Esculin (6,7-dihydroxycoumarin 6-glucoside) is an analogue of umbelliferone with an additional glycosidic moiety at the C6 position. In turn, osthole (7-metoxy-8-isopentenyl-coumarin) has a methoxy moiety at the C7 position and an isoprenyl substituent at the C8 position [10].

**Figure 1.** Structure of 4-hydroxycoumarin (**A**), esculin (**B**), umbelliferone (**C**) and osthole (**D**).

It has been shown that all these hydroxycoumarins have the ability to reduce the proliferation, adhesion, and migration of cancer cells. Esculin interferes with the adhesion of U87 glioblastoma cells by modulating the function of integrins [11]. 4-hydroxycoumarin disorganizes the actin cytoskeleton in B16-F10 melanoma cells and reduces the potential of this tumor to metastasize to the lungs, as shown in mice [12–14]. Umbelliferone, in turn, reduces the migration of laryngeal cancer (RK33) and rat breast adenocarcinoma (RBA) cells [15,16]. 7-hydroxycoumarin has cytotoxic properties against many human cell lines such as leukemia (HL-60) and lung (A549 and H727), kidney (ACHN), and breast cancers (MCF-7) [17]. This compound inhibits the G1 phase cell cycle in human renal cell carcinoma

cells (786-O, OS-RC-2, and ACHN) by reducing the expression of proteins that positively regulate the cell cycle (CDK2, CyclinE1, CDK4, and CyclinD1). Moreover, it modulates the expression of proteins involved in apoptosis (Bax and Bcl-2) and in proliferation (Ki67) [18].

Similar to hydroxycoumarins, osthole inhibits the migration (MCF-7) and invasiveness of breast cancer cells (MDA-MB-231BO) [19]. Additionally, it inhibits proliferation (in lung cancer: A549, leukemia: P-388 D1, breast carcinoma: MCF-7 and MDA-MB 231, medulloblastoma: TE671 and larynx carcinoma: RK33) by inducing apoptosis and inhibiting the cell cycle in the G2/M phase. At the molecular level, it is associated with a decrease in the expression of proteins involved in the cell cycle (CDK2, CyclinB1) and activation of apoptotic proteins (caspase 3, caspase 9, caspase 8, p53 protein) [20–24].

Anaplastic astrocytoma (AA, grade III) and glioblastoma multiforme (GBM, grade IV) are malignant tumors of the central nervous system. At the molecular level, they are characterized by the presence of mutations within genes, the products of which are involved in enhancement of intracellular signal transmission from the cell membrane to the nucleus. This applies in particular to the prosurvival pathways responsible for the regulation of cell proliferation and differentiation: Ras/MEK/ERK (Ras-Ras protein, MEK—mitogen-activated protein kinase, ERK—extracellular signal-regulated kinase) and PI3K/Akt/mTOR (PI3K-phosphoinositide 3-kinase, Akt/PKB-protein kinase B, mTOR- mammalian target of rapamycin kinase). It has been described that blocked signal transmission may be beneficial in enhancement of glioma cell sensitivity. It is also known that combination therapy, especially with the use of natural compounds, can increase the anticancer potential of clinically used pharmacotherapy [25,26].

Therefore, in our research, the antiglioma effect of simple coumarins (4-hydroxycoumarin, umbelliferone, esculin, and osthole) in combination therapy with sorafenib (Raf kinase inhibitor) was evaluated in terms of programmed cell death induction and migratory potential. At the molecular level, these processes were confirmed by the level of caspase 3, Beclin 1, Raf, and PI3K expression. Direct involvement of these proteins in apoptosis, autophagy, and mobility was studied by blocking their expression by specific siRNA.

#### **2. Results**

#### *2.1. E*ff*ect of Simple Coumarins (Osthole, Esculin, Umbelliferone, or 4-Hydroxycoumarin) in Combination with Sorafenib on Apoptosis, Necrosis, and Autophagy Induction*

Our research shows that all the coumarins effectively eliminated tumor cells by apoptosis. Osthole was the most effective compound in both cell lines, as it induced this type of death in 40% and 30% in AA and GBM, respectively (Figure 2A,B). Moreover, the coumarin-induced autophagy (10%) in the MOGGCCM cell line. The application of umbelliferone, esculin, and 4-hydroxycoumarin led to apoptosis in approximately 15% of cells in the T98G cell line (Figure 2B). Slightly different results were obtained in the case of the MOGGCCM cell line (Figure 2A). It turned out to be less sensitive to the action of umbelliferone and 4-hydroxycoumarin, which initiated apoptosis at a level lower than 7%.

Sorafenib had no significant effect on the induction of apoptosis but initiated autophagy in approx. 15% of the T98G cells (Figure 2B). The simultaneous application with the coumarins did not potentiate such an anticancer activity effectively. In the MOGGCCM line, sorafenib diminished the antiglioma potential of osthole, inducing apoptosis in ca. 25% and autophagy in 1% of cells in comparison to the single application of the coumarin. Similar effects were obtained upon the application of esculin, which in combination with sorafenib also showed lower proapoptotic activity. Interestingly, the simultaneous treatment with the hydroxycoumarins and sorafenib was more effective, causing apoptosis in up to 17% of cells.

The experiments carried out on primary human skin fibroblasts (HSF) showed that osthole and esculin (alone and in combination with sorafenib) did not exert cytotoxic effects against normal cells. Different results were obtained for the hydroxycoumarins, as they showed a strong necrotic effect (nearly 20%), which was additionally enhanced by the addition of sorafenib. For this reason, osthole was chosen for further experiments.

**Figure 2.** Effect of sorafenib and osthole, esculin, umbelliferone or 4-hydroxycoumarin administered separately or simultaneously on apoptosis, necrosis, and autophagy induction in the MOGGCCM (**A**), T98G (**B**) and HSF (Human Skin Fibroblasts) (**C**) cell lines. C—control, S—sorafenib, O—osthole, E—esculin, U—umbelliferone, 4-hydroxycoumarin; \* *p* < 0.05.

### *2.2. E*ff*ect of Osthole and Sorafenib on the Migration Potential of Neoplastic Cells*

Inhibition of tumor cell migration plays an important role in anticancer therapy. The wound healing test showed that the treatment with osthole, sorafenib, and the combination of these compounds significantly decreased the migration potential of the AA and GBM cells (Figure 3). The combination therapy was the most effective, as it lowered this activity by approx. 70% compared to the control in both cell lines.

**Figure 3.** Migration potential of the MOGGCCM (**A**) and T98G (**B**) cells upon osthole and sorafenib treatment presented as the percent of cells within the wound. W—wound, C—control, O—osthole, S—sorafenib; \* *p* < 0.05.

#### *2.3. E*ff*ect of Osthole and Sorafenib on the Expression of Cell Death Marker Proteins*

#### 2.3.1. Expression of Caspase-3, PI3K, and Raf Kinases

Caspase 3 is a member of the cysteine-aspartic acid protease family playing a key role in the execution phase of apoptosis. Our studies showed that the use of osthole alone and in combination with sorafenib led to an increase in caspase 3 expression in both cell lines (Figure 4A,B). The best effects (a 14% increase) were obtained upon the administration of osthole in combination with sorafenib in the T98G line. Moreover, the treatment with sorafenib alone reduced the level of this protein by 25% in the MOGGCCM line and by 55% in the T98G line.

**Figure 4.** Effect of osthole, sorafenib and combined treatment with both drugs on the expression of caspase 3 (**A**,**B**), PI3K (Phosphoinositide 3-kinase) (**C**,**D**), Raf (Rapidly Accelerated Fibrosarcoma) (**E**,**F**), and Beclin 1 (**G**,**H**) in the MOGGCCM (**A**,**C**,**E**,**F**) and T98G (**B**,**D**,**G**,**H**) cell lines. C—control, O—osthole, S—sorafenib; \* *p* < 0.05.

PI3K and Raf kinases are also involved in the course of apoptosis and promote the survival of tumor cells. The Western blot analysis showed that osthole alone and in combination with sorafenib decreased the level of Raf kinase (Figure 4C,D). In both cell lines, the treatment with osthole exerted the greatest effect, as it reduced the expression of this protein by 20% in MOGGCCM and 35% in T98G. Moreover, the application of sorafenib increased the level of this protein by 15% in the GBM cells. Better effects were evident in the case of PI3K (Figure 4E,F). The treatment with osthole, sorafenib, and the combination of both compounds decreased the level of this protein. The simultaneous application of the drugs was the most effective, as it caused an over 70% decrease in PI3K expression in both cell lines. The worst effect was observed upon the application of sorafenib alone, which reduced the PI3K levels by 25% in MOGGCCM and 15% in T98G.

#### 2.3.2. Level of Beclin 1

Beclin 1 induces the formation of an autophagosome, thereby initiating the process of autophagy. In the MOGGCCM line (Figure 4G), overexpression of this protein was visible after the treatment with osthole alone (a 40% increase) and in combination with sorafenib (over a 10% increase). In turn, the use of sorafenib alone was associated with a slight decrease in the expression. Completely different results were obtained in the T98G cell line (Figure 4H). Sorafenib increased the level of Beclin 1 by ca. 15%. The coumarin (alone and in combination with sorafenib) inhibited the expression of this protein.

#### *2.4. Apoptosis, Autophagy, and Necrosis Induction Upon Inhibition of PI3, Beclin 1, and Bcl-2 Expression*

#### 2.4.1. Blocking PI3K Expression

The neoplastic transformation of gliomas is associated with excessive activation of the Ras/Raf/MEK/ERK and PI3K/Akt/mTOR pathways. Therefore, the studied cell lines were incubated with anti-PI3K siRNA and cells with blocked PI3K expression were additionally incubated with sorafenib. A significant increase in the sensitivity of the MOGGCCM and T98G cells to the induction of apoptosis was then observed (Figure 5A,B). The treatment with osthole and sorafenib, alone and in combination, induced apoptosis in at least 80% of cells. Sorafenib was the most effective agent leading to the death of almost all cancer cells (97%) in both cell lines. No autophagy or necrosis was observed in both cell lines.

#### 2.4.2. Inhibition of Beclin 1 and Bcl-2 Expression

The antiapoptotic protein Bcl-2 is responsible for the regulation of both apoptosis and autophagy. In a complex with Beclin 1, it inhibits autophagy and, after dissociation, disrupts apoptosis. Blocking the expression of Bcl-2 and Beclin 1 proteins in the AA cells inhibited autophagy and significantly increased apoptosis (Figure 5C,E). Osthole in combination with sorafenib exerted the most potent effect and induced apoptosis in over 70% of the cells. An increase in the apoptotic potential (up to 50%) was also observed in the GBM cells with blocked Bcl-2 expression upon the osthole treatment. In turn, the transfection had no effect on apoptosis induction in the treatment with sorafenib alone and in combination with the coumarin. However, it inhibited autophagy. The GBM cells with the blocked expression of Beclin 1 were less sensitive to induction of programmed death (Figure 5D,F). Apoptosis (30%), but not autophagy, was observed only after the sorafenib treatment.

#### *2.5. Chou-Talalay Method—E*ff*ect of Combination Therapy*

Drug interactions were determined using the isobologram, dose reduction, and combination index method derived from the median-effect principle proposed by Chou and Talalay [27]. As it turned out, the combination of osthole with sorafenib in the T98G line had a synergistic effect (Figure 6D–F). It was stronger when the higher dose was used, which is extremely important in anticancer therapy (for IC97, CI = 0.4). Moreover, the combination treatment significantly reduced the doses of both drugs (DRI > 1), which would have to be higher in a single application to yield the same effect. We observed different

results in the MOGGCCM line, where the doses used had an additive effect (CI ≈ 1), while the higher drug concentrations were already antagonistic (Figure 6A–C). The effectiveness of the combination therapy was estimated on the basis of the ability to induce apoptosis. We also observed autophagy in this cell line, which was inhibited by the simultaneous treatment with osthole. For this reason, the use of combinations of these compounds is also appropriate in AA cells.

**Figure 5.** Level of apoptosis, autophagy, and necrosis in the MOGGCCM (**A**,**C**,**E**) and T98G (**B**,**D**,**F**) cells with inhibited PI3K (**A**,**B**), Beclin 1 (**C**,**D**), and Bcl-2 (B-cell lymphoma 2) (**E**,**F**) expression by specific siRNA (small interfering RNA) upon the osthole and sorafenib treatment. C—control, O—osthole, S—sorafenib; \* *p* < 0.05.

**Figure 6.** Osthole (O) and sorafenib (S) combination treatment in the MOGGCCM (**A**–**C**) and T98G (**D**–**F**) cell line. (**A**,**D**) Combination index (CI) plot: the combination index is plotted as a function of Fa (fractional effect, line of blue color). (**B**,**E**) Isobologram for the combination: classic isobologram at IC50, IC75, and IC90. (**C**,**F**) Fa-DRI (dose-reduction index) plot (Chou-Martin plot).

#### **3. Discussion**

Gliomas, i.e., tumors of the central nervous system, account for approx. 70% of all brain tumors. Due to their infiltrative nature, they are practically impossible to remove surgically [1,25]. Many resistance mechanisms are activated at the molecular level. An example is the Ras/Raf/MEK/ERK pathway. It has been shown that inhibition of this pathway reduces the survival, proliferation, migration, and metabolism of cancer cells [28–30]. It also reduces cancer resistance to the chemotherapy. Our previous studies have shown that sorafenib, i.e., a Raf kinase inhibitor, increases the apoptotic activity of Temozolomide and quercetin [31,32]. It has also been proved that simple coumarins: 4-hydroxycoumarin, umbelliferone, esculin, and osthole can play an important role in cancer therapy [33]. Therefore, in our studies, we used a combination of sorafenib as the Raf kinase inhibitor and the coumarins.

Our experiments showed that the anticancer properties of coumarin derivatives are closely related to their chemical structure. Osthole had the strongest apoptotic activity, alone and in combination with sorafenib. This compound has a methoxy and isopentenyl moiety attached to the C7 and C8 positions, respectively (Figure 1D). The other coumarins (4-hydroxycoumarin, umbelliferone, and esculin) are monohydroxy derivatives (Figure 1A–C). Additionally, esculin has a glycosidic substituent. It has been shown that the cytotoxic effect of coumarins is stronger with the increase in the hydrophobicity of the substituent, which is ensured by the isoprenyl group in osthole [34]. It has also been reported that the length of the substituted aliphatic chain has a great influence on the antitumor activity of the compound [35,36]. The enhanced lipophilicity of the alkyl group contributes to improvement of the ability of compounds to penetrate the cell membrane [37]. In the present experiment, there were no significant differences in the antitumor activity of umbelliferone and 4-hydroxycoumarin in both cell lines. Research conducted by Budzisz et al. showed similar effects. It was found that both hydroxycoumarins inhibited cell proliferation in a gastric carcinoma cell line with similar effectiveness [34]. Thus, the site of attachment of the hydroxy moiety (at the C4 or C7 position) does not significantly affect the proapoptotic properties of the coumarins. Moreover, the presence of an additional glycosidic moiety (esculin) did not change the properties of the compounds in the T98G line and increased the proapoptotic activity in the AA cells. Our experiments confirm observations described by other authors who reported that the anticancer effect of coumarins depends on both the chemical structure and the cell line used. The human carcinoma KB cell line was more sensitive to esculin treatment, while umbelliferone was more effective in HL60 cells [37]. In addition, the combined treatment with sorafenib decreased the sensitivity of the AA cells to the hydroxycoumarins (4-hydroxycoumarin and umbelliferone) treatment and increased the exposure to esculin.

Osthole, alone and in combination with sorafenib, induced apoptosis in approx. 30% of the glioblastoma cells. Our previous research showed that Temozolomide (TMZ)—a drug currently used to treat gliomas, induced apoptosis in 12% of the GBM cells and 5% of the AA cells [38,39]. Interestingly, another chemotherapeutic agent, also used in the treatment of gliomas, Bevacizumab (BEV), similarly to TMZ, reduces the GBM viability by approx. 15% [40]. Thus, our results suggest that the efficacy of osthole with sorafenib may be much higher than that of TMZ and BEV.

In our experiments, at the molecular level, the induction of apoptosis by osthole and sorafenib was accompanied by an increase in caspase 3 expression. On the other hand, the coumarin induced programmed death type I and II in the MOGGCCM cells. Moreover, the combination of both compounds reduced the number of apoptotic cells and completely inhibited autophagy. In this case, the elimination of the process of autophagy is desirable, as it can inhibit cell death in conditions of nutrient deficiency. Cells in the center of tumors are metabolically stressed in this manner and therefore, they use autophagy as a survival mechanism [41]. Blocking this process significantly increases the effectiveness of anticancer therapies used [42]. We also observed that the treatment with osthole alone was associated with overexpression of Beclin 1, which enhanced apoptosis in addition to autophagy. This was accompanied by an increased level of caspase 3. These results suggesting that Beclin 1 has both proautophagous and proapoptotic functions are consistent with studies conducted by Fururya et al. [43] and Huang et al. [44]. They showed that an increase in the expression of Beclin 1 in human gastric (MKN28) and glioma (U87) cells led to apoptosis by increasing the activity of caspase 3, 7, and 8. Interestingly, the treatment with osthole (alone and in combination with sorafenib) in the T98G cells

with blocked Beclin 1 expression inhibited not only autophagy but also apoptosis. It has been shown that Beclin 1 can affect cell survival by interacting with Bcl-2 or Bcl-xL proteins. In turn, Bcl-2 may act as an antiautophagy protein by forming a complex with Beclin 1 [45]. Thus, after silencing the expression of the autophagy marker, the Bcl-2 protein inhibited apoptosis. Different effects were obtained in the AA cells, where instead of apoptosis reduction, we noted a significant increase in the percentage of apoptotic cells. Similar results were noticed after silencing the expression of the Bcl-2 protein. The percentage of apoptotic cells increased significantly after the combined treatment with sorafenib of the MOGGCCM cells. We did not observe autophagy. As with the Beclin 1 blocking, the T98G line was less sensitive to the treatment following the inhibition of Bcl-2 expression. The induction of apoptosis was also observed at that time, but the best effects were achieved only by the application of osthole.

Glioblastoma cells often have mutations in the PTEN and PI3K genes, resulting in continued Akt/PKB kinase activity. This enzyme performs antiapoptotic functions, reducing the susceptibility of cells to inducers of this process and thus enabling tumor growth [25]. We noticed that the use of the combination therapy decreased PI3K protein expression in both cell lines, which correlated with the induction of programmed death. Moreover, blocking the expression of this protein significantly increased the effectiveness of the drugs used. Then, the treatment with sorafenib eliminated almost 100% of cancer cells. The coumarin (alone and in combination with sorafenib) exerted slightly worse effects, inducing over 80% cell death. Interestingly, blocking only PI3K kinase did not decrease the survival rate of the cancer cells. As it turned out, the inhibitors of this protein were also not cytotoxic in the single application; however, when combined with mTOR inhibitors, they significantly increased their effectiveness in eliminating gliomas [46]. A possible explanation is that PI3K inhibition induces other pathways that promote cancer cell survival [47]. Therefore, blocking PI3K expression in combination with osthole or sorafenib gave such good results. Our results suggest that the Ras/Raf/MEK/ERK and PI3K/Akt/mTOR pathways play a key role in the pathogenesis of grade III and IV gliomas.

We observed that the coumarin treatment led to a decrease in Raf kinase expression in both cell lines. At that time, the morphological analysis showed apoptosis, which is consistent with other literature reports. Raf kinase has antiapoptotic functions, hence a decrease in its activity promotes induction of apoptosis [48]. In turn, sorafenib led to the overexpression of this protein in the MOGGCCM line. The autophagy observed was a protective and adaptive response to the inhibition of the Raf/Raf/MEK/ERK pathway. This mechanism was described in many types of cancer, including brain cancer (B76, AM38, and BT40) [49], pancreas (MiaPaCa2 and BxPC3) [50], and melanoma (e.g., A375P, SKMEL5, 1205Lu, MEL624) [51]. It has also been shown that supporting targeted therapy through the use of autophagy inhibitors significantly increased the effectiveness of treatment [49,50]. Therefore, it can be assumed that, in addition to its proapoptotic properties, osthole has antiautophagy activity.

Gliomas have high migratory and invasive potential. Therefore, when planning new therapeutic strategies, drugs that inhibit the translocation of cancer cells should be considered. Our results indicated that the combination of osthole with sorafenib can be used for this purpose.

#### **4. Materials and Methods**

#### *4.1. Cells and Culture Conditions*

Human glioblastoma multiforme cells (T98G, European Collection of Cell Cultures) and human anaplastic astrocytoma cells (MOGGCCM, European Collection of Cell Cultures) were grown in a 1:3 mixture of DMEM (Dulbecco's modified Eagle medium) and nutrient mixture F-12 Ham (Ham's F-12, Sigma). Both cell lines were cultured in medium supplemented with 10% FBS (fetal bovine serum) (Sigma), 100 U/mL penicillin (Sigma), and 100 mg/mL streptomycin (Sigma) at 37 ◦C in a humidified atmosphere consisting of 5% CO<sup>2</sup> and 95% air.

A primary culture of human skin fibroblasts (HSF), which was carried out in the same conditions as the tumor lines, was used in the study as well.

#### *4.2. Coumarin Isolation*

Osthole was obtained for the experiments after isolation from a petroleum extract of *Mutellina purpurea* L. fruits, with a method described previously [22] in the Independent Laboratory of Natural Products Chemistry, Medical University, Lublin, Poland. Two-phase solvent systems made of n-heptane, ethyl acetate, methanol, and water (HEMWat) with a volume ratio 3:2:3:2 were chosen as the most proper system for purification of target compounds (K = 1.8). After injection of 600 mg of a crude oily extract, 2 mg of the target compound were obtained. The identification of the isolated compound was carried out by comparison of the retention time and UV-DAD spectra with those obtained by standards in the same conditions. The purity of osthole was 99% (established with the HPLC-DAD method).

Umbelliferone was purified from fruits of *Heracleum leskowii* L. (Apiaceae) collected in the Medicinal Plant Garden, Department of Pharmacognosy with Medicinal Plant Unit, Medical University, Lublin, Poland in summer 2009, in accordance with a method published previously [15]. Briefly, the fruits were air-dried at room temperature and powdered, and a batch (100 g) was extracted with 100 mL of methanol under reflux for 30 min. After filtration, the procedure was repeated twice. The filtrates were combined and concentrated with a rotary evaporator to remove the solvent. The dried crude extract (13 g) was stored in a refrigerator until further separation. The Spectrum High-Performance Countercurrent Chromatograph (HPCCC) apparatus delivered by Dynamic Extractions (Slough, UK) was employed in the present study. The integrated analytical column (22 mL) was first entirely filled with the upper stationary phase. Then the apparatus was rotated at 200× *g* and the lower mobile phase was pumped into the column at a flow rate of 1.0 mL/min. After hydrodynamic equilibrium was reached, each time 30 mg of the extract dissolved in 1 mL of the two-phase solvent system was loaded onto the column through a 1 mL injection valve. When optimal conditions were determined, the procedure was transferred to a semipreparative integrated column (137 mL volume). The mobile phase was pumped at a flow rate of 6.0 mL/min and 180 mg of the extract was dissolved in 6 mL of the two-phase solvent system and loaded onto the column through a 6 mL injection valve. The solid-phase retention was 70%. The effluent from the column was continuously monitored with a UV detector at 320 nm (Ecom, Prague, Czech Republic). A mixture of n-heptane, ethyl acetate, methanol and water at a ratio of 1:2:1:2 was chosen for further experiments. Umbelliferone was isolated after 20 min. After injection of 180 mg of the crude extract, 1.8 mg of umbelliferone with 99% purity (according to HPLC analysis) was purified. All solvents for HPCCC and HPLC analysis were delivered by Avantor Performance Materials Poland S.A. (Gliwice, Poland—formerly POCh).

Esculin and 4-hydroxycoumarin were delivered by Sigma Aldrich.

#### *4.3. Drug Treatment*

Sorafenib (Nexavar, BAY 43-9006) (1 µM), osthole (150 µM), and hydroxycoumarins: 4-hydroxycoumarin (Sigma), umbelliferone, and esculin (Sigma) at the final concentrations of 200 µM were used in the experiments. The drugs were dissolved in DMSO (Sigma) to the final concentration not exceeding 0.01%. The doses were chosen based on previous studies [32,33]. The cancer cells were treated with the coumarins or with sorafenib separately or in combination for 24 h. As controls, T98G and MOGGCCM cells were incubated only with 0.01% of DMSO.

#### *4.4. Fluorescence Microscopy (Apoptosis, Necrosis, Autophagy Identification)*

For identification of apoptosis and necrosis, a solution of propidium iodide (Sigma, St. Louis, MO, USA) and Hoechst 33342 (Sigma, St. Louis, MO, USA) in distilled water in a ratio of 3:2:5 were used. The cells were stained upon incubation with the appropriate combinations of the drug. After addition of 2.5 µL of the mixture to 1 mL of the medium, the tumor cells were incubated at 37 ◦C for 5 min. The morphological analysis was performed under a fluorescence microscope (Nikon E—800, Tokyo, Japan). It showed bright blue fluorescence characteristic of apoptotic cells. Necrotic cells emitted

red-pink fluorescence. Five percent acridine orange was used to identify autophagous cells incubated in the dark for 15 min. The dye induced the red glow of the autophagous bodies. At least 1000 cells in randomly selected microscopic fields were counted under the microscope. Each experiment was conducted in triplicate.

#### *4.5. Cell Migration Test*

Tumor cell migration was assessed by means of the wound assay model [31]. The cell lines were grown at 2.5 × 105 in standard conditions (37 ◦C, 95% humidity, 5% CO2) in 4 cm-diameter culture dishes (NuncTM, ThermoFisher, Rochester, NY, USA). The next day, the cell monolayer was scratched with the pipette tip (P300), the medium and dislodged cells were aspirated, and the plates were rinsed twice with PBS. Next, fresh culture medium was applied and the number of cells that had migrated into the wound area after 24 h was estimated in the control and drug-treated cultures. The plates were stained with the May–Grünwald–Giemsa method. The observation was performed with the use of a BX51 microscope (Olympus, UK), and the distances between the scratch fronts were estimated using the CellSans program. The results were presented as the migratory potential expressed as the percent of cells within the wound.

#### *4.6. Western Blotting Analysis*

The expression of cellular proteins was evaluated by Western blotting. After treatment for 24 h, cells grown in Falcon flasks (5 mL) were lysed in buffer (125 mM Tris-HCl pH 6.8; 4% SDS; 10% glycerol; 100 mM dithiothreitol). The cells prepared in this way were boiled for 10 min and then centrifuged at 12,000× *g* centrifugal force for 10 min; next, the supernatants were collected. The protein concentration in the cell-free extracts was determined with the Bradford method [52]. Equal amounts of protein (80 µg) from each sample were separated on SDS-PAGE (SDS poliacrylamide gel electrophoresis) [53] and transferred onto an Immmobilon P membrane (Sigma). After blocking with 3% low fat milk for 1 h, the membranes were incubated overnight with primary antibodies: rabbit anti-caspase 3 (Sigma, 1:1000), anti-Beclin 1 (Santa Cruz Biotechnology, 1:500), anti-Raf (Santa Cruz Biotechnology, 1:500), and anti-PI3K (Santa Cruz Biotechnology,1:500). After three washes with PBS enriched with 0.05% Triton X-100 (Sigma), the membranes were incubated with secondary antibodies conjugated with alkaline phosphatase (AP) for 2 h. Alkaline phosphatase substrates: 5-bromo-4-chloro-3-indolylphosphate (BCIP) and nitro-blue tetrazolium (NBT) (Sigma) in *N*,*N*-dimethylformamide (DMF, Sigma) were used for visualization of proteins (Bcl-2, Beclin 1, and caspase 3). The results were analyzed qualitatively on the basis of the band thickness, width, and color depth. The quantitative analysis of protein bands was performed using the Bio-Profil Bio-1D Windows Application V.99.03 program. The data were normalized relative to β-actin (Sigma, working dilution 1:2000). Three independent experiments were performed.

#### *4.7. Cell Transfection*

The cells at a density of 2 × 10<sup>5</sup> were incubated for 24 h at 37 ◦C in a CO<sup>2</sup> incubator to reach 60–80% of confluence. The cells were washed with a DMEM:Ham's F-12 (3;1) mixture without serum and antibiotics and aspirated. Next, a blocking mixture containing 2 µL of specific anti-PI3K, anti-Bcl-2, or anti-BCN1 siRNA (Santa Cruz Biotech Dallas, TX, USA), 2 µL of Transfection Reagent (Santa Cruz Biotech, Dallas, TX, USA), and 250 µL of Transfection Medium (Santa Cruz Biotech) was added. The cells were incubated for 5 h at 37 ◦C, 5% CO2, and 95% humidity. Next, the medium was supplemented with medium containing 20% of fetal bovine serum and 200 µg/mL of antibiotics. After 18 h, the medium was replaced with a fresh one (containing 10% FBS, 100 U/mL penicillin, and 100 mg/mL streptomycin) and the transfected cells were used for further studies (incubation with osthole and sorafenib alone and in combination as well as determination of cell death).

### *4.8. Statistical Analysis*

A one-way anova test followed by Dunnett's multiple comparison analysis was used for statistical evaluation. *p* < 0.05 of data presented as mean ± standard deviation (SD) was taken as the criterion of significance.

### *4.9. Chou-Talalay Method*

The combination index (CI) and the dose reduction index (DRI) were calculated with the method developed by Chou and Talalay [23] using the Compusyn software and the original data of programmed cell death induction in the MOGGCCM and T98G cells upon the sorafenib or osthole treatment [32,33]. CI < 1, CI = 1, and CI > 1 indicate a synergistic, additive, and antagonistic effect, respectively. The DRI represents the fold reduction of compounds as a result of the synergistic combination compared to the concentration of the drug alone required to reach the same effect.

#### **5. Conclusions**

Due to their widespread availability and the broad spectrum of biological activity, coumarins have enormous pharmacological potential. Their anticancer properties, which depend on the chemical structure of the compounds, deserve special attention. Our research has shown that the presence of an isoprenyl moiety (osthole) significantly increases this activity, compared to the other coumarins. It also sensitizes glioma cells with a decreased level of PI3K to apoptosis induction in combination therapy with sorafenib. Therefore, the present results may therefore constitute a basis for further research on the development of new anticancer therapies.

**Author Contributions:** Conceptualization, J.J.-G. and J.S.-W.; formal analysis, J.S.-W. and W.R.; investigation, J.S.-W., A.Z., E.L., K.S.-W., A.M. and W.R.; resources, A.Z., E.L.; writing—original draft preparation, J.S.-W.; writing—review and editing, J.J.-G.; visualization, A.M.; supervision, W.R. and J.J.-G.; All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Maria Curie-Sklodowska University.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


**Sample Availability:** Samples of the coumarins tested are available from the authors.

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Curcumin–Coumarin Hybrid Analogues as Multitarget Agents in Neurodegenerative Disorders**

**Elías Quezada <sup>1</sup> , Fernanda Rodríguez-Enríquez <sup>2</sup> , Reyes Laguna <sup>2</sup> , Elena Cutrín 3 , Francisco Otero <sup>3</sup> , Eugenio Uriarte 1,4 and Dolores Viña 2,\***


**Abstract:** Neurodegenerative diseases have a complex nature which highlights the need for multitarget ligands to address the complementary pathways involved in these diseases. Over the last decade, many innovative curcumin-based compounds have been designed and synthesized, searching for new derivatives having anti-amyloidogenic, inhibitory of tau formation, as well as anti-neuroinflammation, antioxidative, and AChE inhibitory activities. Regarding our experience studying 3-substituted coumarins with interesting properties for neurodegenerative diseases, our aim was to synthesize a new series of curcumin–coumarin hybrid analogues and evaluate their activity. Most of the 3-(7-phenyl-3,5-dioxohepta-1,6-dien-1-yl)coumarin derivatives **11**–**18** resulted in moderated inhibitors of hMAO isoforms and AChE and BuChE activity. Some of them are also capable of scavenger the free radical DPPH. Furthermore, compounds **14** and **16** showed neuroprotective activity against H2O<sup>2</sup> in SH-SY5Y cell line. Nanoparticles formulation of these derivatives improved this property increasing the neuroprotective activity to the nanomolar range. Results suggest that by modulating the substitution pattern on both coumarin moiety and phenyl ring, ChE and MAO-targeted derivatives or derivatives with activity in cell-based phenotypic assays can be obtained.

**Keywords:** curcumin; curcumin–coumarin hybrids; neuroprotection; monoamine oxidase inhibition; cholinesterase inhibition; scavenging activity

#### **1. Introduction**

In neurodegenerative diseases, a loss of nerve cells is observed in the brain and spinal cord, leading to sensory dysfunction (dementia) or loss of function (ataxia). Mitochondrial dysfunction, oxidative stress, protein misfolding, neuroinflammation, and finally apoptosis have been recognized by different studies as pathological causes of neurodegenerative diseases such as Parkinson's disease (PD), Alzheimer's disease (AD), sclerosis multiple (MS), and amyotrophic lateral sclerosis (ALS). Currently, commercially available and approved drugs for these disorders only temporarily relieve symptoms but do not significantly alter disease progression. The development of new treatment strategies remains in the preclinical and clinical stages. Due to the complex nature of neurodegenerative diseases, it seems necessary to design multitarget ligands to address the complementary pathways involved in these diseases [1,2].

Curcumin is a dietary polyphenol presented in the curry spice turmeric. Numerous studies describe its therapeutic potential for neurodegenerative diseases, including AD and

**Citation:** Quezada, E.; Rodríguez-Enríquez, F.; Laguna, R.; Cutrín, E.; Otero, F.; Uriarte, E.; Viña, D. Curcumin–Coumarin Hybrid Analogues as Multitarget Agents in Neurodegenerative Disorders. *Molecules* **2021**, *26*, 4550. https:// doi.org/10.3390/molecules26154550

Academic Editors: Luciana Mosca and Jose Luis Lavandera

Received: 4 June 2021 Accepted: 26 July 2021 Published: 28 July 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

PD, due to its powerful antioxidant, anti-protein aggregation, and anti-inflammatory properties [3,4]. However, curcumin exhibits instability, poor bioavailability, and low cellular uptake, which limits the interest of its use in these disorders [5]. To address this problem, new nanoformulations such as liposomes, solid-lipid nanoparticles, micelles, polymer nanoparticles, and polymer conjugates have been developed [6,7]. With the same objective and also to improve its activity, in recent years, many compounds derived from curcumin have been designed and synthesized. Some of them have shown anti-amyloidogenic activity, inhibitory of the formation of tau, as well as anti-neuroinflammatory, antioxidant, and inhibitory of acetylcholinesterase (AChE) [8,9]. ‐ ‐ ‐ ‐

‐ ‐

‐

‐

‐

‐ ‐

‐

‐

Coumarins are natural or synthetic compounds with diverse biological activities. Many synthetic coumarin derivatives have been designed to obtain new drugs with potential activity in neurodegenerative diseases. Coumarin moiety has the potential to achieve monoamine oxidase (MAO) inhibitory activity (e.g., MAO-A and MAO-B inhibitors), AChE, β- and γ-secretase inhibition. Some of these compounds display potent antioxidant activity and, therefore, could protect cells from neurodegeneration [10]. In the last few years, our group has described different series of 3-substituted coumarins displaying these properties [11–17]. ‐ ‐ ‐ β‐ γ‐ ‐ ‐

MAO inhibition by coumarins may also prevent oxidative stress, through inhibition of neurotransmitters degradation, leading a neuroprotective effect. It has been described that systemic injection of a MAO inhibitor decreases 6-hydroxydopamine-induced oxidative stress [18]. ‐ ‐ ‐ 

Considering the above described and with the aim to improve the properties of curcumin and coumarin for the treatment of neurodegenerative diseases such as AD or PD, we have synthesized a series of curcumin–coumarin hybrid analogues (Figure 1) to study their activity as MAO and AChE inhibitors, free radical scavengers as well as their neuroprotective activity against hydrogen peroxide (H2O2). In addition, to facilitate their passage through cell membranes and therefore improve their neuroprotective activity, some of the derivatives have been formulated into nanoparticles.

**Figure 1.** Overview of design of new curcumin–coumarin hybrid analogues.

#### **2. Results**

#### *2.1. Synthesis of Coumarins* **5** *and* **9**

‐ ‐ ‐ ‐ ‐ ‐ For the synthesis of coumarin **5**, firstly, pyrogallol was treated with K2CO<sup>3</sup> and dichlorodiphenylmethane in CH3CN, obtaining a protected catechol **2**. In a second step, the protected catechol was reacted with magnesium chloride, triethylamine, and *para*-formaldehyde to afford the protected *ortho*-hydroxybenzaldehyde **3**. Then, *ortho*hydroxybenzaldehyde **3** was reacted with sodium hydride and (trimethylsilyl)propioloyl chloride to obtain the silylated ester **4**. Silylated ester **4** was reacted with 1,4-dizabicyclo [2.2.2]octane (DABCO) in THF under reflux, resulting in the desired coumarin **5** (Scheme 1).

**Scheme 1.** (a) dichlorodiphenylmethane, (Ph)2O, 180 ◦C, 30 min; (b) MgCl<sup>2</sup> , Et3N, (CH2O)n, THF reflux, 4 h; (c) (trimethylsilyl)propioloyl chloride, NaH, THF, reflux, 10 h; (d) DABCO, THF, reflux, 12 h.

‐ ‐

Coumarin **9** was synthesized from 2,4,5-trihydroxybenzaldehyde (**6**) following a similar procedure to that described above (Scheme 1).

#### **‐** ‐ ‐ *2.2. Synthesis of a Series of Curcumin–Coumarin Hybrid Analogues* **11***–***18**

 − These compounds were obtained through direct coupling of an acethylacetone–B2O<sup>3</sup> complex with the corresponding formylcoumarin previously obtained (**5** or **9**) and the adequate substituted benzaldehyde (**10a**–**d**) in the presence of tributyl borate and *n*-butylamine (Scheme 2). The deprotection of the phenol groups was carried out in two steps, firstly acidium medium was used to hydrolyze tris(methoxymethoxy) groups (OMOM) followed by hydrolysis of the diphenylbenzodioxole group to obtain compounds **11**–**18**. Compounds **11**–**18** were stored at −20 ◦C and in the dark.

‐ **Scheme 2.** (a) i: 2,4-Pentanedione, B2O<sup>3</sup> , EtOAc, 40 ◦C, 2 h; (BuO)3B; nBuNH<sup>2</sup> , EtOAc, 25–40 ◦C 22 h; ii: HCl, 60 ◦C, 1 h; iii: H<sup>2</sup> , Pd/C, EtOH, rt, 48 h.

μ

μ

 **μ μ ‐ μ ‐ μ**

‐

− −

‐

#### *2.3. Nanoparticles Formulations*

Curcumin and curcumin–coumarin hybrid analogue loaded PLGA nanoparticles were prepared by an interfacial deposition method. All nanoparticles showed a narrow size distribution with mean diameters between 141–168 nm and PDI of 0.121–0.153, and a Zeta potential of −20 to −26 mV (dispersed in purified water). The encapsulation efficiency was similar for all the drugs assayed, obtaining percentages of encapsulation of 56, 53, and 55% for curcumin and curcumin–coumarin hybrid analogues **14** and **16,** respectively.

#### *2.4. In Vitro Activity*

#### 2.4.1. Cholinesterase Inhibition

As seen in Table 1, compounds **15** and **17** at 100 µM concentration inhibit the activity of both AChE and butyrylcholinesterase (BuChE) by approximately 50%. Therefore, their activity is lower than that presented by curcumin on AChE. Compound **12** resulted in the most selective derivative with activity only on BuChE.

**Table 1.** Percentage inhibition of human cholinesterases (hAChE and hBuChE) and human monoamine oxidases (hMAO-A and hMAO-B).


Results are expressed as the mean ± e.e.m (*n* = 3). nd: not determined. At concentration > 100 µM, compounds precipitate.

#### 2.4.2. Monoamine Oxidase Inhibition

Most of the curcumin–coumarin hybrid analogues herein evaluated were not selective, inhibiting both MAO isoforms in a similar percentage (Table 1). The most potent inhibitor was compound **16** with IC<sup>50</sup> (hMAO-B) = 26.18 ± 1.76 µM, which also exhibited greater selectivity over MAO-B. In any case, its inhibitory activity turned out to be lower than that shown by curcumin: IC<sup>50</sup> (hMAO-A) = 10.18 ± 0.68 µM and IC<sup>50</sup> (hMAO-B) = 1.78 ± 0.12 µM. For compounds **11**, **13**, **15**, **17**, their activity on the MAO isoforms could not be determined because they react with the Amplex Red reagent.

#### 2.4.3. Scavenging Activity

As seen in Figure 2, most of the curcumin–coumarin hybrid analogues **11**–**18** showed moderate activity as free radical scavengers. All of them resulted less active than curcumine or vitamin C. In general, among the curcumin–coumarin hybrid analogues, the compounds with the highest activity are those with three hydroxyl groups in contiguous positions of the phenyl substituent. Compounds **11**, **15,** and **16** that did not present this characteristic resulted in being the least active.

‐

μ

 μ μ **Figure 2.** Percentage of neutralization of radical DPPH. by curcumin, curcumin–coumarin hybrid analogues **11**–**18** (100 µM) and vitamin C used as reference (100 µM). Each value is the mean ± s.e.m. of 3 experiments (*n* = 3).

‐ μ

‐ μ ‐

‐

‐ ‐ ‐ ‐

‐

#### 2.4.4. Neuroprotective Activity against H2O<sup>2</sup>

‐ ‐ μ ‐ ‐ ‐ The neuroprotective activity of these curcumin–coumarin hybrid analogues **11**–**18** was evaluated in two different cell models, primary culture of rat motor cortex neurons and SH-SY5Y cell line. Neither curcumin nor any of the curcumin–coumarin hybrid analogues s **11**–**18**(10 µM) showed a protective effect against hydrogen peroxide (H2O2) in the primary culture of rat motor cortex (data not shown). However, **14** and **16** showed a significant increase of viability on the SH-SY5Y cell line treated with H2O<sup>2</sup> (Figure 3). Because of this neuroprotective effect, compounds **14** and **16** were formulated in biodegradable nanoparticles, and their neuroprotective activity against H2O<sup>2</sup> was also evaluated in the SH-SY5Y cell line. In this formulation, derivatives **14** and **16** at low concentration (10 nM) presented a statistically significant neuroprotective activity. As can be seen in Figure 4, the activity of compound **16** turned from a neurotoxic effect when the cultures were treated at 1 µM concentration to a neuroprotective effect at 10 nM concentration.

μ ‐ μ **Figure 3.** Neuroprotective effects of curcumin and curcumin–coumarin hybrid analogues **11**–**18** (10 µM) on SH-SY5Y cells. The results are expressed as % viability versus the control group (treated with DMSO 1%, or DMSO 1%, and H2O<sup>2</sup> 100 µM). Each value is the mean ± s.e.m of at least five experiments. # *p* < 0.0001 versus the control group without H2O<sup>2</sup> treatment. \* *p* < 0.05, \*\* *p* < 0.005 versus DMSO + H2O<sup>2</sup> treated group.

‐ ‐

‐ ‐ ‐

‐

‐ ‐ ‐ ‐ ‐ ‐

‐ ‐ ‐ ‐ ‐

μ ‐

μ

‐ ‐ **Figure 4.** Neuroprotective effects on SH-SY5Y cells of different concentrations of curcumin and curcumin–coumarin hybrid analogues **14** and **16** and their nanoparticle formulations (NC). Each value is the mean ± s.e.m of at least 5 experiments. # *p* < 0.0001 versus the control group (without H2O<sup>2</sup> treatment), \*\* *p* < 0.005, \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001 versus cells treated with H2O<sup>2</sup> .

#### **3. Discussion**

‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ Coumarins **5** and **9** were obtained by a similar route based on reactions described in the literature and both in good yield (68.3% and 81.3%, respectively). However, 3-(7-phenyl-3,5-dioxohepta-1,6-dien-1-yl)coumarin derivatives, namely as curcumin–coumarin hybrid analogues **11**–**18**, were obtained in low yields (5% approximately). This can be probably explained due to the low reactivity of the carbonyl of the formyl group at position 3 of coumarins. This fact was corroborated when obtaining in the same reaction the corresponding 1,7-biscoumarin-3,5-dioxohepta-1,6-dienyl derivatives in very low yield (data not shown). Furthermore, in the same reaction, we appreciated the formation of the corresponding 1,7-bisphenyl-3,5-dioxohepta-1,6-dienyl derivatives, previously described [19–22], and obtained in higher yields than the curcumin–coumarin hybrid analogues **11**–**18**. Despite the poor yield for curcumin–coumarin hybrid **11**–**18**, they could be easily detected and isolated because of their red color.

> Regarding curcumin, the curcumin–coumarin hybrid analogues **11**–**18** conserve two aromatic systems in their structure, replacing a phenyl ring with a coumarin and maintain the length and flexibility of the central link region. These characteristics have been identified as a key for the derivatives to maintain the interest of curcumin in neurodegenerative diseases [23]. Additionally, different substitution partners on both aromatic systems, coumarin moiety, and phenyl ring have been studied.

> MAO plays an important role in the homeostasis of neurotransmitters in the brain. MAO-B inhibitors are being used in combination with L-dopa to manage PD. However, the beneficial effects of MAO-B inhibitors in PD are not only associated with maintaining dopamine levels but also with their neuroprotective properties [24]. The occurrence of activated MAO-B in the brains of patients with AD has also been evidenced. Furthermore, MAO-A has a different appearance in different parts of the brains of patients with AD. MAO-A is increased in the hypothalamus and frontal pole, revealing that activated MAO-A in neurons is involved in the pathology of this disease as a predisposing factor. In addition, increased MAO-A activity appears more significant in the glia of patients with AD [25].

The above described demonstrates the interest in MAO inhibitors for the treatment of these diseases. Compounds **12**, **14**, **16,** and **18** showed moderate inhibitory activity on both MAO isoforms (Table 1). Hydroxyl substituents at positions 6 and 7 of the coumarin nucleus (compounds **16** and **18**) afforded more potent derivatives on the MAO-B isoform than substitution at positions 7 and 8 (compounds **12** and **14**). However, the position of the hydroxyl groups on the phenyl ring does not appear to significatively modify the activity of these compounds on MAO-B. The opposite behavior is observed in the activity of MAO-A. Compounds **12** and **16** resulted in the most potent derivatives, both bearing hydroxyl groups at positions 4, 6, and 7 of the phenyl ring. Only compound **16** exhibited moderate MAO-B selectivity [selectivity index (SI) = IC<sup>50</sup> hMAO-A/IC<sup>50</sup> hMAO-B; SI = 3.82].

Acetylcholine levels are regulated mainly by AChE but also by BuChE. Role of BuChE is less important than AChE in healthy brains. However, the AChE activity remains unchanged or even decreases in AD, while BuChE progressively increases, suggesting that inhibition of both enzymes may be considered a valid approach for AD therapy, increasing levels of AChE [26]. Curcumin–coumarin hybrid analogues **13**, **15,** and **17** showed similar activity on both AChE and BuChE, while compound **12** resulted in selectively inhibiting BuChE activity (Table 1).

Among the studied curcumin–coumarin hybrid analogues, only compound **12** showed potential to inhibit both degradation of acetylcholine (via BuChE inhibition) and monoamines (via non-selective MAO inhibition) (Table 1).

Curcumin can protect neurons against inflammation, oxidative stress, apoptosis, or mitochondrial dysfunction [27,28]. It has been described that curcumin concentrations up to 20 µM increase viability in different cell models treated with H2O<sup>2</sup> [29]. However, the effect of curcumin on SH-SY5Y cells is both dose and time-dependent. Approximately 40 µM concentration and 24 h exposure are the critical parameters at which the cell viability significantly decreases [30]. Other authors describe even lower concentrations (10 µM) to decreases SH-SY5Y proliferation and 20 µM to cause apoptosis [31]. Considering the controversies found in the literature, we studied the neuroprotective effects of low concentrations (≤10 µM) of curcumin and curcumin–coumarin hybrid analogues on two different neuronal models, primary culture of rat motor cortex neurons and SH-SY5Y cell line. While neither of the curcumin–coumarin hybrid analogues **11**–**18** nor curcumin protected rat motor cortex neurons against H2O<sup>2</sup> (data not shown), compounds **14** and **16** showed neuroprotective effects at 10 µM concentration on SH-SY5Y cells (Figure 3). Compound **14** also showed scavenger activity (Figure 2) which could justify, at least partially, its neuroprotective activity. However, compound **16** lacks this activity, but it is the most potent MAO-B inhibitor, indicating that different mechanisms may be implicated in this neuroprotective activity.

Based on the statistically significant increase in viability on SH-SY5Y cells treated with H2O<sup>2</sup> produced by compounds **14** and **16** at 10 µM concentration, both compounds were formulated in biodegradable nanoparticles. This formulation improves the neuroprotective effect at 10 nM concentration on SH-SY5Y cells treated with H2O<sup>2</sup> (Figure 4). Furthermore, compound **16** formulated in nanoparticles goes from having a neurotoxic effect at 1 µM concentration to a statistically significant neuroprotective effect at 10 nM concentration. To explain this neuroprotective effect, it is necessary to resort to the hormesis, which is shared by several phytochemical compounds, including curcumin [32]. Hormesis is defined as a stimulation of cellular protection at low doses while it is inhibited at high doses of the compound, resulting in an inverted J or U-shaped dose-response curve as it can be obtained for compounds **14** and **16**. However, at concentrations used in this work, neuroprotective activity was not found for coumarin.

#### **4. Materials and Methods**

#### *4.1. Materials and Instrumentation*

All reactions utilizing air- or moisture-sensitive reagents were carried out in flamedried glassware under an argon atmosphere, unless otherwise stated. Hexane, CH2Cl2, THF, Et2O, Et3N, and *n*-BuNH<sup>2</sup> were distilled prior to use according to the standard protocols. Other reagents were purchased and used as received without further purification unless otherwise stated. Reactions were magnetically stirred and monitored by thin-layer chromatography (TLC). Analytical TLC was performed on plates precoated with silica gel (Merck 60 F254, 0.25 mm). Compounds were visualized with UV light and/or by staining with ethanolic phosphomolybdic acid (PMA) followed by heating on a hot plate. Flash chromatography (FC) was performed with silica gel (35–60 mesh) under pressure. Melting points were determined in a Reichert Kofler thermopan or in capillary tubes in a Buchi 510 apparatus and are uncorrected. NMR spectra were recorded on Bruker AMX 250 (1H, 250 MHz; <sup>13</sup>C, 62.9 MHz) spectrometer in CDCl<sup>3</sup> or DMSO-*d*<sup>6</sup> with TMS as the internal standard. Chemical shifts (*δ*) are given in ppm and coupling constants (*J*) in Hz. Multiplicity is indicated as follows: s, singlet; d, doublet; m, multiplet; bs, broad singlet. Elemental analyses were performed on Thermo-Finnigan Flash 1112 CHNS/O analyzer (Supplementary Materials).

#### *4.2. Chemical Synthesis*

#### 4.2.1. Synthesis of 2,2-Diphenylbenzo[1,3]dioxol-4-ol (**2**)

Dichlorodiphenylmethane (9.65 mL, 50.31 mmol) was added to a stirred mixture of pyrogallol (1, 4.23 g, 33.54 mmol) in diphenyl ether (25 mL), and the reaction mixture was heated at 180 ◦C for 30 min. The mixture was cooled to room temperature, and petroleum ether (50 mL) was added to give a solid compound [33,34]. Then the solid was filtered and purified by column chromatography using CH2Cl<sup>2</sup> to yield **2** as a white solid (9.65 g, 99.2%). m.p.: 165 ◦C. <sup>1</sup>H-NMR (250 MHz, CDCl3, δ ppm): 7.54 (4H, m), 7.37 (6H, m), 6.71 (1H, t, *J* = 6.7 Hz), 6.53 (1H, d, *J* = 6.4 Hz), 6.46 (1H, d, *J* = 6.4 Hz), 4.98 (1H, br). <sup>13</sup>C-NMR (62.9 MHz, CDCl3, δ ppm): 148.2, 139.9, 139.3, 133.8, 129.1 (2C), 128.2 (4C), 126.3 (4C), 122.1, 116.1, 110.8, 101.9. Anal. Calcd. (%) for [C19H14O3]: C, 78.61; H, 4.86; found (%): C, 78.58; H, 4.83.

#### 4.2.2. Synthesis of 4-Hydroxy-2,2-diphenylbenzo[1,3]dioxol-5-carbaldehyde (**3**)

To a dry THF solution (300 mL) of the 2,2-diphenylbenzo[1,3]dioxol-4-ol (**2**) (6.2 g, 21.35 mmol), anhydrous magnesium chloride (4.065 g, 42.70 mmol), triethylamine (5.95 mL, 4.32 g, 42.70 mmol) and paraformaldehyde (1.923 g, 64.05 mmol) were added. The reaction mixture was heated to reflux under Ar atmosphere for 4 h, and monitored by TLC (hexane:ethyl acetate = 8:2). After complete consumption of the phenol, the reaction mixture was cooled and diluted with diethyl ether (100 mL). The organic layer was washed successively with HCl (1 M, 2 × 100 mL) and H2O (2 × 100 mL), and then dried (Na2SO4) [35]. The product was purified by column chromatography using hexane:ethyl acetate (98:2) to yield **3** as a white solid (5.6 g, 82.5%). m.p.: 159 ◦C. <sup>1</sup>H-NMR (250 MHz, CDCl3, δ ppm): 11.07 (1H, bs), 9.68 (1H, s), 7.59 (4H, m), 7.38 (6H, m), 7.12 (1H, d, *J* = 8.2 Hz), 6.62 (1H, d, *J* = 8.2 Hz). <sup>13</sup>C-NMR (62.9 MHz, CDCl3, δ ppm): 194.9, 154.2, 145.3, 139.1 (2C), 133.6, 130.2, 129.3 (2C), 128.2 (4C), 126.0 (4C), 119.2, 118.09, 101.9. Anal. Calcd. (%) for [C20H14O4]: C, 75.46; H, 4.43; found (%): C, 75.44; H, 4.40.

#### 4.2.3. Synthesis of 5-Formyl-2,2-diphenylbenzo[1,3]dioxol-4-yle (trimethylsilyl)propiolate (**4**)

Sodium hydride (0.942 g, 23.55 mmol, 60%, washed with hexane) was suspended in anhydrous THF (25 mL), then cooled to 0 ◦C. A solution of **3** (2.50 g, 7.85 mmol) in anhydrous THF (25 mL) was dropwise added and the suspension was stirring for 1 h. Then, trimethylsilylpropioloyl chloride (3.78 g, 23.55 mmol) [36] in THF (10 mL) was dropwise added. The mixture was refluxed for 10 h. The mixture was quenched with ice-cold water and extracted three times with EtOAc. The combined organic layers were washed with brine, dried over Na2SO4, and evaporated [37]. The residue was purified by column chromatography on SiO<sup>2</sup> hexane:ethyl acetate (2:98) to obtain silylated ester **4** as a colorless syrup (2.8 g, 80.7%). <sup>1</sup>H-NMR (250 MHz, CDCl3, δ ppm): 9.93 (1H, s), 7.53 (4H, m), 7.39 (7H, m), 6.87 (1H, d, *J* = 8.2 Hz), 0.27 (9H, s). <sup>13</sup>C-NMR (62.9 MHz, CDCl3, δ ppm): 186.8, 153.6, 149.1, 145.6, 139.5, 138.4 (2C), 129.5 (2C), 128.2 (4C), 127.1, 126.2 (4C), 123.5, 120.3, 106.8, 98.1, 92.8, −1.1. Anal. Calcd. (%) for [C26H22O5Si]: C, 70.57; H, 5.01; found (%): C, 70.56; H, 5.00.

#### 4.2.4. Synthesis of 2′ ,2′ -Diphenyl-1,3-dioxol[h]coumarin-3-carbaldehyde (**5**)

A mixture of silylated ester **4** (2.8 g, 6.33 mmol), and DABCO (1.42 g, 12.66 mmol) in THF (150 mL) was refluxed under Ar atmosphere. After 12 h, the mixture was diluted with CH2Cl2, washed with HCl (10%) and brine, and dried over Na2SO<sup>4</sup> [38]. The solvent was evaporated to leave a residue, which was purified by silica-gel chromatography (CH2Cl2) to afford the 3-formylcoumarin **5** as a yellow solid (1.6 g, 68.3%). m.p.: 181 ◦C. <sup>1</sup>H-NMR (250 MHz, CDCl3, δ ppm): 10.18 (1H, s), 8.31 (1H, s), 7.59 (4H, m), 7.40 (6H, m), 7.23 (1H, d, *J* = 8.3 Hz), 6.94 (1H, d, *J* = 8.3 Hz). <sup>13</sup>C-NMR (62.9 MHz, CDCl3, δ ppm): 187.4, 159.3, 153.7, 146.2, 139.0, 138.6 (2C), 133.5, 129.6 (2C), 128.4 (4C), 126.4, 126.0 (4C), 120.6, 118.5, 114.3, 107.0. Anal. Calcd. (%) for [C23H14O5]: C, 74.59; H, 3.81; found (%): C, 74.56; H, 3.80.

#### 4.2.5. Synthesis of 6-Hydroxy-2,2-diphenylbenzo[1,3]dioxol-5-carbaldehyde (**7**)

Following the procedure previously described to obtain compound **2**, dichlorodiphenylmethane (7.4 mL, 38.92 mmol) was reacted with **6** (4.0 g, 25.95 mmol) in diphenyl ether (25 mL), to yield **7** as a white solid (8.0 g, 96.8% yield) [33,39]. m.p.: 128 ◦C. <sup>1</sup>H-NMR (250 MHz, CDCl3, δ ppm): 11.79 (1H, br), 9.59 (1H, s), 7.58 (4H, m), 7.39 (6H, m), 6.90 (1H, s), 6.55 (1H, s). <sup>13</sup>C-NMR (62.9 MHz, CDCl3, δ ppm): 193.6, 161.4 (2C), 154.6, 141.0, 139.1 (2C), 129.4 (2C), 128.3 (4C), 126.1 (2C), 126.0 (2C), 109.4, 98.4, 90.1. Anal. Calcd. (%) for [C20H14O4]: C, 75.46; H, 4.43; found (%): C, 75.42; H, 4.42.

#### 4.2.6. Synthesis of 6-Formyl-2,2-diphenylbenzo[1,3]dioxol-5-yle(trimethylsilyl)propiolate (**8**)

Following the procedure previously described to obtain compound **4**, compound **7** (2.5 g, 7.85 mmol) was reacted with sodium hydride (0.942 g, 23.55 mmol, 60%, washed with hexane), and trimethylsilylpropioloyl chloride (3.78 g, 23.55 mmol) [36] to obtain **8** as a colorless syrup (2.5 g, 72.0%) [37]. <sup>1</sup>H-NMR (250 MHz, CDCl3, δ ppm): 9.96 (1H, s), 7.50 (5H, m), 7.34 (6H, m), 6.91 (1H, s), 0.20 (9H, s). <sup>13</sup>C-NMR (62.9 MHz, CDCl3, δ ppm): 187.0, 154.3, 151.0, 145.9, 140.9, 138.2 (2C), 129.2 (2C), 128.4 (4C), 126.9 (4C), 125.8, 122.5, 118.2, 107.1, 97.9, 85.8. Anal. Calcd. (%) for [C26H22O5Si]: C, 70.57; H, 5.01; found (%): C, 70.55; H, 4.99.

#### 4.2.7. Synthesis of 2′ ,2′ -Diphenyl-1,3-dioxol[g]coumarin-3-carbaldehyde (**9**)

Following the procedure previously described to obtain compound **5**, a mixture of the propionic ester **8** (2.5 g, 5.65 mmol), and DABCO (1.26 g, 11.30 mmol) afforded the 3-formylcoumarin **9** as a yellow solid (1.7 g, 81.3%) [38]. m.p.: 173 ◦C. <sup>1</sup>H-NMR (250 MHz, CDCl3, δ ppm): 9.97 (1H, s), 8.12 (1H, s), 7.44 (4H, m), 7.38 (6H, m), 6.92 (1H, s), 6.69 (1H, s). <sup>13</sup>C-NMR (62.9 MHz, CDCl3, δ ppm): 191.8, 160.6, 152.1, 151.2, 145.3, 143.5, 138.9 (2C), 129.5 (2C), 128.4 (4C), 126.0 (4C), 122.4, 119.2, 108.3, 103.8, 98.4. Anal. Calcd. (%) for [C23H14O5]: C, 74.59; H, 3.81; found (%): C, 74.55; H, 3.79.

#### 4.2.8. Synthesis of 7,8-Dihydroxy-3-(7-(2′ ,4′ ,6′ -trihydroxyphenyl)-3,5-dioxohepta-1,6-dien-1-yl)coumarin (**11**)

2,4-Pentanedione (0.25 g, 2.5 mmol) and boric anhydride (0.121 g, 1.75 mmol) were dissolved in EtOAc (5 mL) and stirred for 2 h at 40 ◦C. Coumarin **5** (0.925 g, 2.5 mmol), tris(methoxymethoxy)benzaldehyde **10a** [40,41] (0.715 g, 2.5 mmol) and tributyl borate (2.3 g, 10 mmol) were added and the reaction mixture was stirred for 0.5 h. Then a solution of *n*-butylamine (0.182 g, 2.5 mmol) in EtOAc (2.5 mL) was dropwise added over a period of 30 min, the mixture was stirred for a further 18 h at room temperature and 4 h at 40 ◦C. The mixture was hydrolyzed by the addition of 0.4 N HCl (10 mL) and heating to 60 ◦C for 1 h. The organic layer was separated, and the aqueous layer was extracted three times with EtOAc. The combined organic layers were washed with water and dried over Na2SO4.

Evaporation of the solvent left a red-brown powder [42]. The solid was dissolved in 80 mL of ethanol and treated with 10% Pd/C (0.6 g, 33 wt. % of starting material) [43]. The system was purged several times with hydrogen and stirred under hydrogen for 48 h. The reaction mixture was then purged with Ar and filtered through Celite washing with CH3OH. The filtrate was evaporated and the dark red solid was purified by column chromatography using CH2Cl2:CH3OH (9:1 and 8:2) to give **11** (44 mg, 4.15%). m.p.: 180 ◦C (dec.). <sup>1</sup>H-NMR (250 MHz, DMSO-*d*6, δ ppm): 10.10 (2H, bs), 9.70 (2H, bs), 9.51 (2H, bs), 7.72 (1H, s), 7.46 (1H, d, *J* = 15.5 Hz), 7.15 (1H, d, *J* = 8.8 Hz), 6.84 (1H, d, *J* = 15.4 Hz), 6.64 (3H, m), 5.96 (2H, s), 5.83 (1H, s). <sup>13</sup>C-NMR (62.9 MHz, DMSO-*d*6, δ ppm): 182.4, 169.7, 160.5, 158.7 (2C), 158.0, 148.7, 141.6, 139.5, 137.6, 134.0, 130.5, 126.9, 125.3, 120.7, 120.6, 114.3, 114.2, 109.9, 103.1, 95.7 (2C). Anal. Calcd. (%) for [C22H16O9]: C, 62.27; H, 3.80; found (%): C, 62.24; H, 3.78.

4.2.9. Synthesis of 7,8-Dihydroxy-3-(7-(2′ ,4′ ,5′ -trihydroxyphenyl)-3,5-dioxohepta-1,6-dien-1-yl)coumarin (**12**)

Following the procedure described above to obtain compound **11**, reaction of coumarin 5 (0.925 g, 2.5 mmol) and tris(methoxymethoxy)benzaldehyde **10b** [44] (0.715 g, 2.5 mmol) yielded compound **12** as a red solid (40 mg, 3.77%). m.p.: 177 ◦C (dec.). <sup>1</sup>H-NMR (250 MHz, DMSO-*d*6, δ ppm): 10.12 (2H, bs), 9.76 (2H, bs), 9.50 (2H, bs), 7.79 (1H, d, *J* = 16.1 Hz), 7.65 (1H, s), 7.07 (1H, d, *J* = 8.8 Hz), 6.78 (1H, d, *J* = 15.4 Hz), 6.65 (1H, d, *J* = 16.1 Hz), 6.59 (2H, m), 6.42 (1H, s), 6.19 (1H, s), 5.83 (1H, s). <sup>13</sup>C-NMR (62.9 MHz, DMSO-*d*6, δ ppm): 187.4, 179.5, 162.9, 152.5, 148.7, 148.6, 141.6, 140.4, 139.5, 137.6, 134.0, 129.0, 126.9, 124.2, 120.7, 120.6, 115.6, 114.4, 114.3, 114.2, 103.5, 96.9. Anal. Calcd. (%) for [C22H16O9]: C, 62.27; H, 3.80; found (%): C, 62.22; H, 3.77.

4.2.10. Synthesis of 7,8-Dihydroxy-3-(7-(2′ ,3′ ,4′ -trihydroxyphenyl)-3,5-dioxohepta-1,6 dien-1-yl)coumarin (**13**)

Following the procedure described above to obtain compound **11**, reaction of coumarin **5** (0.925 g, 2.5 mmol) and tris(methoxytrimethoxy)benzaldehyde **10c** [45,46] (0.715 g, 2.5 mmol) yielded compound **13** as a red solid (46 mg, 4.34%). m.p.: 185 ◦C (dec.). <sup>1</sup>H-NMR (250 MHz, DMSO-*d*6, δ ppm): 10.10 (2H, bs), 9.74 (2H, bs), 9.52 (2H, bs), 7.79 (1H, d, *J* = 16.0 Hz), 7.72 (1H, s), 7.08 (1H, d, *J* = 8.8 Hz), 6.86 (1H, d, *J* = 15.8 Hz), 6.78 (1H, d, *J* = 16.1 Hz), 6.58 (2H, m), 6.51 (1H, d, *J* = 8.4 Hz), 6.12 (1H, d, *J* = 8.4 Hz), 5.78 (1H, s). <sup>13</sup>C-NMR (62.9 MHz, DMSO-*d*6, δ ppm): 184.6, 162.4, 158.9, 148.8, 148.7, 147.4, 141.6, 139.5, 137.6, 134.0, 133.8, 128.4, 126.9, 124.4, 120.8, 120.7, 120.6, 114.7, 114.3, 114.2, 107.8, 101.6. Anal. Calcd. (%) for [C22H16O9]: C, 62.27; H, 3.80; found (%): C, 62.26; H, 3.79.

4.2.11. Synthesis of 7,8-Dihydroxy-3-(7-(3′ ,4′ ,5′ -trihydroxyphenyl)-3,5-dioxohepta-1,6 dien-1-yl)coumarin (**14**)

Following the procedure described above to obtain compound **11**, reaction of coumarin **5** (0.925 g, 2.5 mmol) and tris(methoxymethoxy)benzaldehyde **10d** [47–49] (0.715 g, 2.5 mmol) yielded compound **14** as a red solid (48 mg, 4.53%). m.p.: 183 ◦C (dec.). <sup>1</sup>H-NMR (250 MHz, DMSO-*d*6, δ ppm): 10.09 (2H, bs), 9.70 (2H, bs), 9.53 (2H, bs), 7.70 (2H, m), 7.08 (1H, d, *J* = 8.8 Hz), 6.91 (1H, d, *J* = 16.1 Hz), 6.83 (1H, d, *J* = 16.1 Hz), 6.74 (1H, d, *J* = 15.8 Hz), 6.58 (1H, d, *J* = 8.8 Hz), 6.25 (2H, s), 5.85 (1H, s). <sup>13</sup>C-NMR (62.9 MHz, DMSO-*d*6, δ ppm): 185.8, 180.1, 159.6, 148.7, 146.7 (2C), 141.6, 139.5, 137.6, 135.3, 135.1, 134.0, 128.9, 126.9, 123.0, 120.7, 120.6, 114.3, 114.2, 108.0 (2C), 101.9. Anal. Calcd. (%) for [C22H16O9]: C, 62.27; H, 3.80; found (%): C, 62.21; H, 3.78.

4.2.12. Synthesis of 6,7-Dihydroxy-3-(7-(2′ ,4′ ,6′ -trihydroxyphenyl)-3,5-dioxohepta-1,6 dien-1-yl)coumarin (**15**)

Following the procedure described above to obtain compound **11**, reaction of coumarin **9** (0.925 g, 2.5 mmol) and tris(methoxymethoxy)benzaldehyde **10a** (0.715 g, 2.5 mmol) yielded compound **15** as a red solid (50 mg, 4.72%). m.p.: 179 ◦C (dec). <sup>1</sup>H-NMR (250 MHz, DMSO-*d*6, δ ppm): 10.09 (2H, bs), 9.75 (2H, bs), 9.51 (2H, bs), 7.77 (1H, d, *J* = 15.3 Hz), 7.58

(1H, s), 6.87 (1H, s), 6.75 (1H, d, *J* = 15.3 Hz), 6.60 (2H, m), 6.45 (1H, s), 5.97 (2H, s), 5.84 (1H, s). <sup>13</sup>C-NMR (62.9 MHz, DMSO-*d*6, δ ppm): 185.2, 180.1, 160.5, 159.4, 158.7 (2C), 148.7, 148.1, 146.3, 139.2, 137.6, 130.5, 126.9, 125.3, 118.8, 113.7, 113.0, 104.9, 103.5, 102.1, 95.7 (2C). Anal. Calcd. (%) for [C22H16O9]: C, 62.27; H, 3.80; found (%): C, 62.25; H, 3.79.

4.2.13. Synthesis of 6,7-Dihydroxy-3-(7-(2′ ,4′ ,5′ -trihydroxyphenyl)-3,5-dioxohepta-1,6 dien-1-yl)coumarin (**16**)

Following the procedure described above to obtain compound **11**, reaction of coumarin **9** (0.925 g, 2.5 mmol) and tris(methoxytrimethoxy)benzaldehyde **10b** (0.715 g, 2.5 mmol) yielded compound **16** as a red solid (53 mg, 5.00%). m.p.: 188 ◦C (dec.). <sup>1</sup>H-NMR (250 MHz, DMSO-*d*6, δ ppm): 10.10 (2H, bs), 9.76 (2H, bs), 9.55 (2H, bs), 7.75 (1H, d, *J* = 16.1 Hz), 7.67 (1H, s), 6.95 (1H, d, *J* = 15.5 Hz), 6.86 (1H, s), 6.68 (1H, d, *J* = 16.1 Hz), 6.53 (1H, d, *J* = 15.5 Hz), 6.38 (1H, s), 6.30 (1H, s), 6.15 (1H, s), 5.84 (1H, s). <sup>13</sup>C-NMR (62.9 MHz, DMSO-*d*6, δ ppm): 183.8, 181.7, 159.1, 152.5, 148.7, 148.6, 148.1, 146.3, 140.4, 139.7, 137.6, 129.0, 126.9, 124.2, 118.8, 115.6, 114.3, 113.7, 113.0, 103.7, 102.2, 95.4. Anal. Calcd. (%) for [C22H16O9]: C, 62.27; H, 3.80; found (%): C, 62.20; H, 3.77.

4.2.14. Synthesis of 6,7-Dihydroxy-3-(7-(2′ ,3′ ,4′ -trihydroxyphenyl)-3,5-dioxohepta-1,6 dien-1-yl)coumarin (**17**)

Following the procedure described above to obtain compound **11**, reaction of coumarin **9** (0.925 g, 2.5 mmol) and tris(methoxytrimethoxy)benzaldehyde **10c** (0.715 g, 2.5 mmol) yielded compound **17** as a red solid (45 mg, 4.24%). m.p.: 190 ◦C (dec.). <sup>1</sup>H-NMR (250 MHz, DMSO-*d*6, δ ppm): 10.05 (2H, bs), 9.70 (2H, bs), 9.45 (2H, bs), 7.75 (2H, m), 7.04 (1H, s), 6.95 (1H, d, *J* = 15.8 Hz), 6.71 (1H, d, *J* = 15.8 Hz), 6.57 (2H, m), 6.43 (1H, d, *J* = 8.4 Hz), 6.14 (1H, d, *J* = 8.4 Hz), 5.85 (1H, s). <sup>13</sup>C-NMR (62.9 MHz, DMSO-*d*6, δ ppm): 186.4, 179.3, 159.0, 148.8, 148.7, 148.1, 147.4, 146.3, 139.2, 137.6, 133.8, 128.4, 126.9, 124.4, 120.8, 118.8, 114.7, 113.7, 113.0, 107.8, 102.1, 96.5. Anal. Calcd. (%) for [C22H16O9]: C, 62.27; H, 3.80; found (%): C, 62.24; H, 3.73.

4.2.15. Synthesis of 6,7-Dihydroxy-3-(7-(3′ ,4′ ,5′ -trihydroxyphenyl)-3,5-dioxohepta-1,6 dien-1-yl)coumarin (**18**)

Following the procedure described above to obtain compound **11**, reaction of coumarin **9** (0.925 g, 2.5 mmol) and tris(methoxytrimethoxy)benzaldehyde **10d** (0.715 g, 2.5 mmol) yielded compound **18** as a red solid (55 mg, 5.19%). m.p.: 185 ◦C (dec.). <sup>1</sup>H-NMR (250 MHz, DMSO-*d*6, δ ppm): 10.12 (2H, bs), 9.77 (2H, bs), 9.53 (2H, bs), 7.78 (1H, d, *J* = 15.7 Hz), 7.69 (1H, s), 7.15 (1H, s), 6.94 (1H, d, *J* = 16.0 Hz), 6.75 (1H, d, *J* = 16.0 Hz), 6.55 (1H, d, *J* = 15.7 Hz), 6.45 (1H, s), 6.25 (2H, s), 5.84 (1H, s). <sup>13</sup>C-NMR (62.9 MHz, DMSO-*d*6, δ ppm): 183.5, 181.8, 158.6, 148.7, 148.1, 146.7 (2C), 146.3, 139.2, 137.6, 135.3, 135.1, 128.9, 126.9, 123.0, 118.8, 113.7, 113.0, 108.0 (2C), 102.0, 94.6. Anal. Calcd. (%) for [C22H16O9]: C, 62.27; H, 3.80; found (%): C, 62.22; H, 3.73.

#### *4.3. Formulation of Biodegradable Nanoparticles*

Resomer® RG503H (Evonic) polymer, which is a 50:50 copolymer of polylactic and polyglycolic acid in its acid form following the nanoprecipitation technique, was used to make the biodegradable nanoparticles [50]. Briefly, a 1% aqueous solution of poloxamer 407 (Sigma Aldrich) was used as the aqueous phase. Acetone was used as a solvent for the polymer and the active principles, the concentration of Resomer® and drug being 0.4% and 0.1% (p/v), respectively. The organic phase was added slowly at room temperature, using a syringe, to the aqueous solution, under magnetic stirring. Finally, the acetone was removed by evaporation at 50 ◦C in a rotary evaporator (Buchi) until a final volume of 25 mL was obtained. Filtration was performed through 0.22 µm diameter polyamide membrane filters to remove the non-incorporated drug.

To determine the content of the active principle and the encapsulation efficiency, an aliquot of the nanosuspensions was diluted in ethanol to dissolve the Resomer® and release the drugs, determining its concentration by spectrophotometry.

Size distribution (mean diameter and polydispersity index) and zeta potential of nanoparticles were determined in purified water at 25 ◦C using a Malvern Zetasizer Nano ZS instrument (Malvern Instruments Ltd., Malvern, UK).

#### *4.4. Determination of hMAO-A and hMAO-B In Vitro Activity*

The in vitro activity of the synthesized curcumin–coumarin hybrid analogues **11**–**18** or curcumin on hMAO enzymatic activity was evaluated using an Amplex® Red MAO assay kit and following a fluorimetric method previously described by us [11]. Briefly, 50 µL of sodium phosphate buffer (0.05 M, pH 7.4) containing the test molecules (new compounds or reference inhibitors) in different concentrations and adequate amounts of recombinant hMAO-A or hMAO-B (adjusted to obtain in our experimental conditions the same reaction velocity (hMAO-A: 1.1 µg protein; specific activity: 150 nmol of *para*-tyramine oxidized to *para*-hydroxyphenylacetaldehyde/min/mg protein; hMAO-B: 7.5 µg protein; specific activity: 22 nmol of *para*-tyramine transformed/min/mg protein)) were incubated for 10 min at 37 ◦C in a flat-black bottom 96-well microtest plate, placed in the dark fluorimeter chamber. After this incubation period, the reaction was started by adding 50 µL of the mixture containing (final concentrations) 200 µM of the Amplex® Red reagent, 1 U/mL of horseradish peroxidase and 1 mM of *para*-tyramine. The production of H2O<sup>2</sup> and, consequently, of resorufin, was quantified at 37 ◦C in a multidetection microplate fluorescence reader (Fluo-star OptimaTM, BMG LABTECH, Offenburg, Germany) based on the fluorescence generated (excitation, 545 nm, emission, 590 nm) over a 10 min period, in which the fluorescence increased linearly. Control experiments were carried out simultaneously by replacing the tested molecules with appropriate dilutions of the vehicles. In addition, the possible capacity of these molecules to modify the fluorescence generated in the reaction mixture due to non-enzymatic inhibition (i.e., for directly reacting with Amplex® Red reagent) was determined by adding these molecules to solutions containing only the Amplex® Red reagent in sodium phosphate buffer. The specific fluorescence emission (used to obtain the final results) was calculated after subtraction of the background activity, which was determined from wells containing all components except the hMAO isoforms, which were replaced by sodium phosphate buffer solution. The IC<sup>50</sup> values for each compound were calculated by linear regression representing the logarithm of the concentration (M) of the studied compound (abscissa axis) against the percentage of inhibition of the control MAO activity (ordinate axis). This linear regression was performed with 4–6 concentrations of each evaluated compound capable of inhibiting the control enzymatic activity of the MAO isoenzymes between 20% and 80%.

#### *4.5. Determination of AChE and BuChE In Vitro Activity*

Ellman's method [51] was used to determine in vitro ChE activity. 0.01 U/mL human recombinant AChE expressed in HEK 293 cells or 0.0005 U/mL BuChE isolated from human serum were added to a 50 mM phosphate buffer solution (pH 7.2) containing different concentrations of curcumin–coumarin hybrid analogues **11**–**18** or curcumin. The mixture was preincubated at 37 ◦C for 5 min followed by the addition of 5 mM acetylthiocholine or butyrylthiocholine and 0.25 mM 5,5′ -dithio-bis(2-nitrobenzoic acid) (DNTB). The activity was measured by the absorbance increasing at λ 412 nm at 1 min intervals for 10 min at 37 ◦C (Fluo-Star OptimaTM, BMG LABTECH, Offenburg, Germany). Control experiments were performed simultaneously by replacing the test drugs with appropriate dilutions of the vehicles. The specific absorbance (used to obtain the final results) was calculated after subtraction of the background activity, which was determined in wells containing all components except the AChE or BuChE, which was replaced by a sodium phosphate buffer solution.

#### *4.6. DPPH Radical Scavenging Assay*

The DPPH was dissolved in methanol (50 µM), and 99 µL of the solution was transferred to each well of a 96-well microplate. Curcumin–coumarin hybrid analogues **11**–**18**,

coumarin or reference drug (vitamin C) were added to each well at a final concentration of 100 µM. Solutions were incubated at room temperature for 30 min. Absorbance was determined at λ 517 nm using a microplate reader (Fluo-star OptimaTM, BMG LABTECH, Offenburg, Germany). DPPH radical solution in methanol was used as a control, whereas a mixture of methanol and sample served as blank. The scavenging activity percentage (AA%) was determined according to the equation described by Mensor et al. [52]: AA% = 100 − ((Abssample − Absblank) × 100/Abscontrol)

#### *4.7. Cell Culture*

#### 4.7.1. Primary Culture of Rat Motor Cortex Neurons

Embryos were extracted by cesarean section from 18 days pregnant Wistar Kyoto rats which were euthanized by CO<sup>2</sup> inhalation. Brains were carefully dissected out, and after removing meninges, a portion of the motor cortex was isolated. Fragments obtained from several embryos were mechanically digested and cells were resuspended in Neurobasal medium. The Neurobasal medium was supplemented with 2% B-27 to obtain cortex neuronal cultures. Cells were seeded in 96-well plates at a density of 200,000 cells/mL. Cultures were grown for 7–8 days in an incubator (Form Direct Heat CO2, Thermo Electron Corporation, Madrid, Spain) under saturated humidity at a partial pressure of 5% CO<sup>2</sup> in air at 37 ◦C until a dense neuronal network could be observed [53].

#### 4.7.2. Human Neuroblastoma SH-SY5Y Cell Culture and Maintaining

The SH-SY5Y cells grew in a culture medium containing Ham's F12 and MEM (mixture 1:1) and supplemented with 15% FBS, 1% L-Glutamine, 1% non-essential amino acids and 1% of penicillin G/streptomycin sulfate (all of them from Sigma-Aldrich S.A.) [54]. The cells were grown in 75 cm<sup>2</sup> flasks in an incubator, under conditions of saturated humidity with a partial pressure of 5% CO<sup>2</sup> in the air, at 37 ◦C. Cell culture medium was replaced every 2 days, and, at 80–90% of confluence, the cells were sub-cultured. To carry out the viability assays, the cells were seeded in sterile 96-well plates, with a density of 2 × 10<sup>5</sup> cells/mL and grown distributed in aliquots of 100 µL for 24 h under the conditions described above.

#### 4.7.3. Cell Viability

Cells grown in 96-well plates were treated with H2O<sup>2</sup> (100 µM) and curcumin or test compounds **11**–**18** (10 µM). When cells were treated with curcumin or curcumin–coumarin hybrid analogues **14** and **16** formulated in nanoparticles, they were added in the 24 h prior to H2O<sup>2</sup> treatment. Then, cultures were incubated for 24 h. After this time, cell viability was determined using MTT (5 mg/mL in Hank's). 10 µL of MTT solution was added to each well containing 100 µL of culture medium and the cells were incubated for 2 h as described above. Then, the culture medium was removed, 100 µL DMSO/well was added to solve the formazan crystals formed by the viable cells and the absorbance (λ 540 nm) was quantified in a plate reader. The viability (percentage) was calculated as (Absorbance (treatment)/Absorbance (negative control))100% [55]. Statistical analysis was performed using one way ANOVA test followed by Dunnett's multiple comparison test by using GraphPad software.

#### **5. Conclusions**

A new series of curcumin–coumarin hybrid analogues **11**–**18** were synthesized in low yield starting from 2′ ,2′ -diphenyl-1,3-dioxol[h]coumarin-3-carbaldehyde (**5**), or 2 ′ ,2′ -diphenyl-1,3-dioxol[g]coumarin-3-carbaldehyde (**9**), and the corresponding tris (methoxymethoxy)benzaldehyde **10a**–**d**. Synthesized derivatives did not reach a potential as a multitarget drug. In general, they were either better at inhibiting MAO isoforms or AChE and BuChE activity. Only compound **12** inhibited BuChE and MAO isoforms with similar potency. In addition, compounds **14** and **16** resulted in being neuroprotective against H2O<sup>2</sup> in SH-SY5Y cells. The formulation of these compounds in nanoparticles improves their neuroprotective activity at low concentrations. Results suggest that by modulating the substitution pattern on both coumarin moiety and phenyl ring, ChEs and MAO-targeted derivatives or derivatives with activity in cell-based phenotypic assays can be obtained.

**Supplementary Materials:** The following are available online. 1H and 13C NMR of compounds **5**, **9**, **10a**–**d**, **11**–**18**.

**Author Contributions:** Conceptualization, D.V., R.L. and F.O.; methodology, E.Q., F.R.-E. and E.C.; validation, E.Q., F.R.-E. and E.C.; formal analysis, D.V. and E.U.; investigation, E.Q., F.R.-E. and E.C.; resources, D.V. and F.O.; data curation, D.V. and F.O.; writing—original draft preparation, D.V. and R.L.; writing—review and editing, D.V. and E.U.; visualization, D.V.; supervision, D.V., F.O.; project administration, D.V.; funding acquisition, D.V. and F.O. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Consellería de Cultura, Educación e Ordenación Universitaria (EM2014/016) and Centro Singular de Investigación de Galicia and the European Regional Development Fund (ERDF) (accreditation 2016–2019, ED431G/05).

**Institutional Review Board Statement:** The study was conducted according to the European regulations on the protection of animals (Directive 2010/63/UE), the Spanish Real Decreto 53/2013 (1 February) and the Guide for the Care and Use of Laboratory Animals as adopted and promulgated by the United States National Institutes of Health. In this context, the experimental protocol was approved by the Institutional Animal Care and Use Committee of the University of Santiago de Compostela and Xunta de Galicia, Spain (Register Number 15007DE/12/INVMED02/NERV02/B/MCT3).

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

**Sample Availability:** Samples of the compounds **5**, **9**, **10a**–**d**, **11**–**18** are available from the authors.

#### **References**


*Article*
