**Antiangiogenic E**ff**ects of Coumarins against Cancer: From Chemistry to Medicine**

**Mohammad Bagher Majnooni 1, Sajad Fakhri 2, Antonella Smeriglio 3, Domenico Trombetta 3, Courtney R. Croley 4, Piyali Bhattacharyya 5, Eduardo Sobarzo-Sánchez 6,7, Mohammad Hosein Farzaei 2,\* and Anupam Bishayee 4,\***


Academic Editor: Kyoko Nakagawa-Goto Received: 27 October 2019; Accepted: 19 November 2019; Published: 24 November 2019

**Abstract:** Angiogenesis, the process of formation and recruitment of new blood vessels from pre-existing vessels, plays an important role in the development of cancer. Therefore, the use of antiangiogenic agents is one of the most critical strategies for the treatment of cancer. In addition, the complexity of cancer pathogenicity raises the need for multi-targeting agents. Coumarins are multi-targeting natural agents belonging to the class of benzopyrones. Coumarins have several biological and pharmacological effects, including antimicrobial, antioxidant, anti-inflammation, anticoagulant, anxiolytic, analgesic, and anticancer properties. Several reports have shown that the anticancer effect of coumarins and their derivatives are mediated through targeting angiogenesis by modulating the functions of vascular endothelial growth factor as well as vascular endothelial growth factor receptor 2, which are involved in cancer pathogenesis. In the present review, we focus on the antiangiogenic effects of coumarins and related structure-activity relationships with particular emphasis on cancer.

**Keywords:** coumarins; antiangiogenic; cancer; natural agents; chemistry; medicine

### **1. Introduction**

Angiogenesis (also known as neovascularization), the growth of blood vessels from the existing vasculature, has been shown to play a critical role in the development of various diseases, including rheumatoid arthritis, diabetic retinopathy, asthma, endometriosis, psoriasis, obesity, and cancer [1–3]. Inflammation, tissue ischemia, and hypoxia which cause the release of the angiogenesis factors, such as vascular endothelial growth factor (VEGF), cytokines, cell adhesion molecules, and nitric oxide (NO), are among the most important triggers of angiogenesis [4]. In 1971, Folkman reported that tumor metastasis occurs as a consequence of angiogenesis [5]. This was the starting point for the design and

use of bevacizumab, thalidomide, sunitinib, and axitinib as antiangiogenic drugs in the treatment of a variety of cancers [6–8]. Considering the crucial role of angiogenesis in the progression of cancer, investigating novel and potential antiangiogenic compounds is of great importance to combat cancer. Several naturally occurring compounds, including vinblastine, vincristine, paclitaxel, were reported as antiangiogenic and anticancer agents. Besides, other natural compounds with antiangiogenic activities, such as resveratrol, artemisinin, boswellic acid, and cannabidiol, have shown enormous potential for cancer prevention and therapy [9–13]. For instance, endocannabinoid 2-arachidonoyl-glycerol showed a promising anticancer effect in several cell lines [14]. Overall, cancer remains a clinical challenge, despite advancements in its treatment. This raises the need to investigate novel multi-target agents to attenuate multiple signaling pathways involved in tumor progression.

Growing evidence has introduced coumarins as potential multi-targeting agents with various pharmacological effects and medicinal uses [15]. Coumarins, with their 2H-1-benzopyran-2-one structure, are natural compounds that exist in various plant families, including Apiaceae, Asteraceae, Fabaceae, Rutaceae, Moraceae, Oleaceae, and Thymelaeaceae [16]. Apiaceae is the greatest family of plants containing coumarin compounds [16]. Also, due to the antioxidant [17], anti-inflammatory [18], anxiolytic [19], analgesic [20], neuroprotective [21], cardioprotective [22], antidiabetic [23], and anticancer [24] activities of coumarins [25], researchers have studied the synthesis of various coumarin derivatives, in addition to their purification from natural sources [26]. Both synthetic and natural coumarins have shown noticeable anticancer effects in vitro and in vivo through various mechanisms [27], including the inhibition of angiogenesis [28,29]. From a mechanistic point of view, some coumarins have shown promising antiangiogenic effects through the interaction with and repression of signaling mediators involved in angiogenesis [30,31].

In this review, we focus on the cellular signaling pathways of angiogenesis and recent pharmacological antiangiogenic agents, emphasizing natural and synthetic coumarins with antiangiogenic effects as well as their pharmacological mechanisms and structure-activity relationship in cancer.

### **2. Angiogenesis: Biology and Cellular Signaling**

Angiogenesis could be controlled by achieving a balance among activating cytokines and growth factors on one hand and inhibiting agents on the other, which stimulate or inhibit endothelial cells (ECs), respectively. Proangiogenic agents include growth factors, namely, VEGF, fibroblast growth factors, epidermal growth factor (EGF), transforming growth factor-β (TGF-β), platelet-derived growth factor (PDGF), placental growth factor (PGF), hepatocyte growth factor/scatter factor (HGF/SF), and cytokines, such as tumor necrosis factor-α (TNF-α), colony-stimulating factor-1 (CSF-1), and interleukin-8 (IL-8) [32–35]. ECs, fibroblasts, platelets, smooth muscle cells, inflammatory cells, and cancer cells are involved in producing angiogenic growth factors and cytokines [4].

VEGF and FGF have been considered as promising antiangiogenic targets [35]. VEGF mainly acts through tyrosine kinase VEGF receptor 2 (VEGFR2) [36]. Furthermore, bioactive lipids, such as prostaglandin E2 (PGE2), and sphingosine-1-phosphate (S1P), matrix degenerating enzymes, namely, matrix metalloproteinases (MMP) and heparinases, small mediators (e.g., NO, peroxynitrite, serotonin, and histamine), angiopoietins (Ang), and erythropoietin are among other activators of angiogenesis [35]. In order to attract other angiogenesis-stimulating factors, cancer cells induce a situation of hypoxia by increasing the demand for nutrients and oxygen. In the hypoxic condition, hypoxia-inducible factor-1α (HIF-1α), together with released anti-apoptotic factors, growth factors, and cytokines, provokes angiogenesis [4,37].

On the other hand, angiogenesis could be suppressed by inhibiting proteins, which are classified into either direct or indirect angiogenesis inhibitors. The first class of inhibitors directly suppress ECs in the growing vasculature, while the second class indirectly suppress either tumor cells or other tumor-associated stromal cells [35,38]. This direct inhibitory effect could also be mediated by integrin receptors through several intracellular signaling pathways [39]. Angiostatin, endostatin, arrestin, canstatin, and tumstatin are released by the proteolysis of distinct endothelial cell-matrix molecules and prevent vascular ECs from proliferating and migrating in response to angiogenesis inducers [40]. Interferons, retinoic acid, IL-1, IL-12, tissue inhibitor of metalloproteinases, and multimerin 2 are other angiogenesis inhibitors [41,42].

As previously mentioned, angiogenesis is controlled by a balance between activators and inhibitors of angiogenesis. Hypoxia, as a critical determinant, causes an imbalance between activators and inhibitors by inducing the upregulation of HIF-1α, which elevates the expression of pro-angiogenesis agents as well as suppresses the expression of angiogenesis inhibitors [43]. Therefore, all the mediators in these pathways could be therapeutic targets to inhibit angiogenesis.

### **3. Recent Advancement in Pharmacological Antiangiogenic Agents**

Several angiogenesis inhibitors have been found since Folkman first presented the concept of introducing angiogenesis inhibitors as anticancer drugs [5]. RNA interference (RNAi) therapy, chimeric antigen receptor T cell therapy, gene therapy, and pharmacological agents are auspicious antiangiogenic interventions [44]. According to the United States Food and Drug Administration (FDA), approved antiangiogenic agents are classified into two major groups, namely, monoclonal antibodies (mAbs) and small molecules [45].

VEGF receptors (VEGFRs) and related downstream signaling pathways are crucial targets of mAbs. Small molecules also target receptors, including PDGFR, VEGFR, Fms-like tyrosine kinase 3, and c-Kit receptor, and signaling proteins such as Raf, mitogen-activated protein kinase (MAPK), mammalian target of rapamycin (mTOR), and phosphoinositide 3-kinases (PI3K). Besides, the antiangiogenic/anticancer effects of FDA-approved herbal drugs, including vinca [46], taxan [47], camptothecins [48], podophyllotoxins [49], and homoharringtonine [50], related receptors, and downstream signaling pathways have now been confirmed [13].

VEGF (A–D), PDGF (A–D), HGF [51,52], and FGF [53,54] bind to VEGFR (1 and 2), PDGFR (α and β), MET [55,56], and FGFR (1-4) tyrosine kinase receptors, respectively, and activate downstream signaling pathways, thereby regulating cell growth, differentiation, and angiogenesis [54,57–59]. Their overactivation is attributed to several mutations promoting tumor vascularization in different types of cancers [60,61], while their inhibitors exert antitumor effects [62]. Bevacizumab, aflibercept, and ramucirumab have been developed as antiangiogenic agents to target the VEGF/VEGFR signaling pathway [63].

Angiopoietins (Ang1–4) bind to the Tie2 receptor. While Ang1 helps the vessels stabilize, Ang2 is secreted by ECs in response to proangiogenic factors, including hypoxia, cytokines, and inflammation [64]. Ang/Tie2-targeted therapy is challenging, since it could be either antitumor or protumor, depending on the context [65].

The rearranged during transfection (RET) protein binds receptor tyrosine kinases (RTKs) associated with normal development, maintenance, and maturation of cells and tissues [66]. However, its mutation is related to the growth and progression of tumors [66,67]. Therefore, RET inhibition could be of great importance in combating cancer.

Multi-targeting antiangiogenic drugs are shown in Figure 1. These drugs exert anticancer effects through simultaneously modulating several signaling pathways involved in angiogenesis.

**Figure 1.** Signaling pathways and therapeutic targets of antiangiogenic and anticancer drugs and agents. VEGF, vascular endothelial growth factor; FGF, fibroblast growth factor; EGF, epidermal growth factor; TGF-β, transforming growth factor-β; PDGF, platelet-derived growth factor; PGF, placental growth factor; HGF/SF, hepatocyte growth factor/scatter factor; TNF-α, tumor necrosis factor-α; CSF-1, colony-stimulating factor-1; IL, interleukin; MMP, matrix metalloproteinase; TIMPs, tissue inhibitors of metalloproteinases; S1PR, sphingosine-1-phosphate receptor; NO, nitric oxide; PI3K:,phosphatidylinositol-3-kinase; PLC, phospholipase C; PKC, protein kinase C; HIF, hypoxia-inducible factor; and m-TOR: mammalian target of rapamycin.

### **4. Coumarins**

### *4.1. Chemical Structure and Sources*

Coumarin (C9H6O2, 2H-1-benzopyran-2-one, 146.145 g/mol) and its derivatives (Figure 2) are a large class of natural compounds that are widely distributed in the plant kingdom and are biosynthesized from ortho-hydroxy-cinnamic acid in the shikimic acid pathways [68]. In terms of chemical structure, coumarins are subdivided into four main groups: (a) simple coumarins, such as heparin and scopoletin; (b) furanocoumarins (linear and angular), such as bergapten and imperatorin; (c) pyranocoumarins, such as grandivittin and agasyllin; (d) dicoumarins and pyrone-substituted coumarins, such as phenylcoumarins (Figure 2) [69–71].

Coumarins are isolated and purified from fruits, leaves, stems, roots, and flowers of more than 40 plant families. The Apiaceae represents a family of plants with the highest number of species producing coumarins, including *Anethum graveolens*, *Angelica dahurica*, *Apium graveolens*, *Petroselinum* *crispum* and *Heracleum mantegazzianum*. Other plant families producing coumarins are Rutaceae (*Citrus aurantium*, *Citrus sinensis*, and *Melicope glabra*), Asteraceae (*Matricaria recutita* and *Achillea millefolium*), Fabaceae (*Melilotus o*ffi*cinalis* and *Glycyrriza glabra*), and Moraceae (*Ficus carica*) [16,72–76].

**Figure 2.** *Cont.*

**Figure 2.** Chemical structures of coumarins with antiangiogenic effects.

### *4.2. Biological and Pharmacological E*ff*ects of Coumarins*

Coumarins have several biological and pharmacological effects. For example, coumarins isolated from the bark of *M. glabra* showed antioxidant properties [73]. In this line, antimicrobial effects of coumarins from the fruits of *H. mantegazzianum* Sommier & Levier as well as *Peucedanum luxurians* Tamamsch were reported [76,77]. Antiviral effects of coumarins isolated from *Prangos ferulacea* L. have been shown by Shokoohinia et al. [78]. In addition, anxiolytic effects of coumarin derivatives, purified from the root of *Biebersteinia multifida* DC, have been demonstrated [79]. Other coumarins, such as umbelliferone and pimpinellin, were isolated from the root of *Zosima absinthifolia* and these compounds showed anti-Alzheimer effects [80]. Kontogiorgis and co-workers [81] designed and synthesized coumarin derivatives based on azomethine, with anti-inflammatory activities. Synthesized coumarins based on 3,4-dihydro-2H-benzothiazines showed analgesic effects in formalin- and acetic acid-induced writhing tests [82]. Additionally, various coumarins have shown antiulcerogenic [83], spasmolytic [84], anticoagulant [85], vasorelaxant [86], cytotoxic, and anticancer activities [87].

On the other hand, hepatotoxicity, nausea, and diarrhea were reported as the side effects of coumarin derivatives [88,89].

#### *4.3. Coumarins as Anticancer Agents*

As the second leading cause of death worldwide, cancer is one of the most critical diseases that threaten public health and imposes a high cost on countries' health systems each year. Because of the resistance of cancer cells to conventional drugs used in chemotherapy as well as the side effects of these drugs, it is necessary to find new anticancer agents. Undoubtedly, medicinal plants are one of the richest sources of biologically active compounds and potential novel anticancer drugs. Coumarins are natural compounds with low to moderate side effects, which have been introduced by researchers as promising anticancer compounds [29,90–92]. Several coumarins also inhibit cytochrome P450, thereby affecting the blood concentration of various anticancer drugs. In this line, bergamottin inhibits cytochrome P450 and reduces the effects of various carcinogenic agents [93].

The anticancer and cytotoxic activities of synthetic and natural coumarins with different functional groups on their basic structure (Figure 2) have been reported by several investigators. These studies showed the anticancer activity of coumarins against breast cancer [89], colon cancer [94], lung cancer [24], ovarian cancer [95], hepatocellular carcinoma [96], bladder carcinoma [97], leukemia [98], and other types of cancer in vitro and in vivo, via different mechanisms, including free-radical scavenging, antioxidant activity [99], induction of cell cycle arrest [100], interaction with various signaling pathways with important role in cell differentiation and proliferation [101], telomerase and carbonic anhydrase inhibition [102,103], and antiangiogenic activity [104]. For example, Taniguchi and co-workers [105] isolated eight coumarins from the leaves of *Rhizophora mucronata* and reported the anticancer effects of methoxyinophyllum P, calocoumarin B, and calophyllolide against HeLa cells (cervical cancer), with IC50 equal to 3.8, 29.9, and 36.4 μM, respectively, and HL-60 cells (promyelocytic leukemia cells) with IC50 of 12.9, 2.6, and 2.2 μM, respectively. Also, three hemiterpene ether coumarins, isolated from *Artemisia armeniaca* Lam, showed cytotoxic effects in HL-60 and K562 cells (chronic myelogenous leukemia cells). In their study, armenin showed the highest cytotoxic effect through cycle arrest, with IC50 equal to 22.5 and 71.1 μM for K562 and HL-60 cells, respectively [106].

Among coumarins, clausarin, nordentatin, dentatin, and xanthoxyletin, isolated from dichloromethane extraction of root bark of *Clausena harmandiana*, showed high cytotoxic activities [107]. In this study, clausarin (Figure 2) showed the highest cytotoxic activity, which was superior to that of cisplatin used as a positive control, against hepatocellular carcinoma (HepG2, IC50 = 17.6 ± 2.1 μM), colorectal carcinoma (HCT116, IC50 = 44.9 ± 1.4 μM), and lung adenocarcinoma (SK-LU-1, IC50 = 6.9 ± 1.6 μM) cell lines. Clausarin also showed the highest antioxidant activity in the 2,2-diphenyl-1-picryl-hydrazyl-hydrate scavenging assay. From a mechanistic point of view, apoptosis induction was reported as an anticancer mechanism of coumarins [107].

On the other hand, several studies have been carried out on the synthesis of coumarins to produce coumarin derivatives with improved anticancer effects. Among them, synthetic scopoletin derivatives have shown promising antitumor activities. In their study, out of 20 synthesized derivatives of scopoletin, 5 compounds showed the greatest effects (IC50 < 2 μM) in MCF-7 and MDA-MB 231 cells (human breast adenocarcinoma cell line) as well as in HT29 cells (human colorectal adenocarcinoma cell line). The relationship between the increase in Log P value and the increase in cytotoxic activity was established in this investigation. Cell cycle arrest was also suggested as the anticancer mechanism of these compounds [108].

Besides, Zang et al. [105] synthesized novel anticancer analogs of geiparvarin using a bioisosteric transformation method. In their study, it was also shown that adding electron-withdrawing substituents to the benzene rings, such as 7-((1-(4-fluorobenzyl)-1*H*-1,2,3-triazol-4-yl)methoxy)-4Hchromen-4-one (Figure 2), increased their cytotoxic effects in a human hepatoma cell line (QGY-7701, IC50 = 14.37 ± 9.93) and a colon carcinoma cell line (SW480, IC50 = 11.18 ± 2.16) compared with geiparvarin (IC50 = 17.68 ± 0.40 and 20.34 ± 0.75, respectively) [109]. Additionally, Cui and co-workers [110] showed the anticancer effects of three synthesized coumarins derived from triphenylethylene, occurring through the inhibition of angiogenesis.

Considering the above investigations on the anticancer effects of coumarin derivatives, these compounds could be developed as anticancer drugs.

### **5. Coumarin and Angiogenesis Inhibition**

Inhibition of angiogenesis is one of the most critical anticancer mechanisms of secondary plant metabolites, including coumarins. Natural and synthetic coumarins with different structures can inhibit the factors involved in angiogenesis, migration, proliferation, and differentiation of endothelial cells in vitro and in vivo. Coumarins act by blocking various molecular signaling pathways, involving growth factors (e.g., VEGFs, TNF-α, and FGF-2), cytokines (e.g., IL-1 and IL-6), angiogenic enzymes (e.g., MMP), endothelial-specific receptor tyrosine kinases (e.g., Tie2), and adhesion molecules (e.g., intercellular adhesion molecule-1) [30,111,112].

### *5.1. Natural Coumarins with Antiangiogenic E*ff*ects*

The antiangiogenic effects of coumarins from natural sources, especially isolated and purified from plants, are reported in several studies. Scopoletin [113,114], esculetin [104], herniarin [115], decursin, and decursinol [116] with coumarin structure, as well as imperatorin [117] and psoralidine [118] with furanocoumarin structure (Figure 2), are among the natural antiangiogenic coumarins.

Pan and co-workers [119] showed the antiangiogenic effects of scopoletin, isolated from *Erycibe obtusifolia* Benth stem, in vitro and in vivo. In their study, scopoletin markedly reduced the number of blood vessel branch points after a 48-h treatment with the dose of 100 nmol/egg in the chick chorioallantoic membrane (CAM) model. Additionally, the inhibitory effects of scopoletin on migration, proliferation, and tube formation of human umbilical vein endothelial cells (HUVECs) induced by VEGF (10 ng/mL) were observed. These investigators showed that the proliferation of HUVECs was significantly inhibited by 100 μM scopoletin after 72 h, and tube formation and migration of HUVECs were inhibited following treatment with 100 μM scopoletin by 52.4% and 38.1%, respectively [119].

Decursin and decursinol angelate, isolated from *Angelica gigas* root, showed substantial antiangiogenic effects in vitro and in vivo [120]. These natural coumarins significantly decreased the development of blood vessels in transgenic zebrafish embryos at 20 μM concentration as well as in the CAM model at 6 μM/egg. In their study, the inhibition of VEGFR2 (one of the most important receptors of VEGF) and other angiogenesis signaling pathway related to VGEF, such as phosphorylated extracellular signal-regulated kinase (p-ERK) and MAPK as well as phosphorylated-c-Jun N-terminal kinase (JNK) in ECs were observed [116].

The antiangiogenic effect of marmesin (Figure 2), a furanocoumarin isolated from ethanolic extract of the twigs of *Broussonetia kazinoki*, was reported by Kim et al. [121]. They showed that marmesin at 10 μM significantly inhibited the expression and activity of MMP-2 in response to VEGF-A (10 ng/mL) in HUVECs. Besides, marmesin reduced EC proliferation, migration, invasion, tube formation and also induced cell cycle arrest in a concentration-dependent manner. Marmesin at 10 μM also inhibited the development of angiogenesis in the rat aortic ring model. VEGF-A stimulated various critical molecules of angiogenesis signaling pathways, such as focal adhesion kinase (FAK), Src kinase, MEK, ERK, Akt, and p70S6K. These pathways were inhibited by 10 μM marmesin [121]. In another study, osthol, columbianadin and columbianetin acetate, three coumarins isolated from *Angelicae Pubescentis* Radix, showed inhibitory effects on the secretion of monocyte chemoattractant protein-1, a pro-inflammation factor and one of the most important migration-regulating chemokines [122]. On the other hand, conferone, a sesquiterpene coumarin isolated from *Ferula szwitziana,* showed antiangiogenic and cytotoxic effects on a human colorectal adenocarcinoma cell line (HT-29) through reducing proangiogenic factors, including angiopoietin 1 and 2 [123]. Besides, daphnetin (7,8-hydroxy coumarin), another natural coumarin, inhibited the expression of MAPK, VEGFR2, ERK1/2, AKT, FAK, and cSrc, which are involved in angiogenesis [124] (Table 1). Also, inhibiting PI3K/AKT activity, another angiogenesis inducer pathway, is one of the antiangiogenic mechanisms of murrangatin purified from *Micromelum falcatum* [125]. Moreover, reducing and blocking VEGF and MMPs are among the most important antiangiogenic mechanisms of coumarins such as galbanic acid [126], umbelliprenin [127], imperatorin [117], auraptene [128], esculetin [31], osthole [129], and scopolin [130] (Table 1).



### *5.2. Synthetic Coumarins with Antiangiogenic E*ff*ects*

Angiogenesis is a critical process in the development and progression of cancer, as demonstrated by various pre-clinical and clinical evidence. Among natural-based entities, coumarins present little cytotoxicity, while demonstrating more powerful antiangiogenic effects than conventional cytotoxic drugs [131,132]. Semi-synthesized and synthesized products from natural coumarins, used as lead compounds, led to the discovery of interesting antiangiogenic and non-cytotoxic molecules. Coumarins with interesting antiangiogenic and non-cytotoxic properties almost entirely mimic the behavior of the physiological ligands of the main therapeutic targets [132]. Recently, several sulfonyl derivatives of coumarins have been studied as cytotoxic and antiangiogenic agents against HepG2 hepatocellular carcinoma cells in vitro. All synthesized coumarins showed no cytotoxic effect but exhibited a high antimigration activity through the inhibition of MMP-2. CD105 was over-expressed in all cases and, therefore, was not involved in the antimigration activity [132]. In other cases, no statistically significant difference in gene expression of CD44 was found. A synthesized coumarin, 2-oxo-2*H*-chromene-6-sulfonyl derivative, was found to be the most promising antiangiogenic agent, since it was able to inhibit the migratory activity mediated by MMP-2 and down-regulate CD105; however, it did not show any effect on CD44.

On the other hand, the antiangiogenic capacity of several sulfonyl derivatives of coumarins was evaluated using molecular docking studies. In these studies, it was also observed that the compounds showed better docking scores with respect to the I52 ligand, with the nitro derivatives being the best, due to the ability of the nitro group to better coordinate the Zn++ ion within the binding site. However, the in vitro antiangiogenic activity of sulfonyl derivatives of coumarins was not statistically significant. Only the 2-oxo-2H-chromene-6-sulfonyl derivative with *N*-acetylpyrazolone substitution at the 6-position showed a promising antiangiogenic activity, exhibiting better binding interactions with the active site and a docking score comparable to that of the inhibitor I52 (−16.22 vs. 18.18 kJmol-1, respectively) [132].

NF-κB is another protein factor, which plays a pivotal role in gene expression and, therefore, is involved in proliferation, angiogenesis, and metastasis, as well as in drug resistance in cancer. In light of this, the development of angiogenesis inhibitors is of significant importance in the treatment of many cancers. Recently, the effects of 26 new synthetic coumarins were tested against hepatocellular carcinoma cells [133]. The investigators identified (7-carbethoxyamino-2-oxo-2*H*-chromen-4-yl) methylpyrrolidine-1-carbodithioate as the most promising one, because it was cytotoxic in a time- and concentration-dependent manner and it was able to hinder the binding of NF-κB to DNA, therefore inhibiting the expression of several genes, such as *cyclin D1*, *Bcl-2*, *survivin*, *MMP12*, and c-*Myc*. Furthermore, it was able to reduce cell migration and invasion induced by CXCL12, a cytokine that plays a pivotal role in angiogenesis by recruiting endothelial progenitor cells from the bone marrow [133].

Analysis of data present in the existing literature shows that it is possible to identify, through structure-activity analysis, coumarin derivatives more suitable for a given type of tumor. The coumarin derivatives that possess an N-aryl carboxamide, a phenyl substitution at the C-3 position, and 1, 2, 3-triazolyl, trihydroxystilbene, and amino substitutions at the C-4 position were the most effective in targeting lung cancer [24].

Preliminary in vitro results revealed that some coumarin-tethered isoxazolines exhibited significant antiproliferative effect against a human melanoma cancer cell line (UACC 903). Only one derivative with a 3,4-dimethoxy substitution did not show any cytotoxicity against a normal fibroblast cell line (FF2441) in the same concentration range. These results were corroborated in the Ehrlich ascites carcinoma animal model, highlighting significantly lowered cell viability, body weight, ascites volume as well as a down-regulation of angiogenesis and tumor growth [134].

Histone deacetylase 1 (HDAC1), a key element in the control of cell proliferation and differentiation as well as in angiogenesis, represents an attractive therapeutic target for new inhibitors of angiogenesis. In this regard, the benzamidic derivatives of coumarins were found to be the most promising candidates. Four compounds of *N*-(4-((2-aminophenyl)carbamoyl) benzyl)-2-oxo-2*H*-chromene-3-carboxamide derivatives showed the most promising cytotoxic effect, calculated as IC50 in the range of 0.53–57.59 μM, on several cancer cells, including HCT116, A2780, MCF7, PC-3, HL60, and A549, without any effects on a human normal cell line (HUVEC, IC50 > 100 μM). Moreover, they showed a strong HDAC1-inhibitory activity (IC50 0.47–0.87 μM) with *N*-(4-((2-aminophenyl) carbamoyl) benzyl)-7-((3,4-dichlorobenzyl)oxy)-2-oxo-2*H*-chromene-3-carboxamide, showing an IC50 value similar to that of the reference drug entinostat (0.47 ± 0.02 μM vs. 0.41 ± 0.06 μM) [135].

Another interesting coumarin derivative investigated is (*E*)-2-(4-methoxybenzyloxy)-3-prenyl-4-methoxy-N-hydroxycinamide (BMX), a semisynthetic derivative of osthole and a coumarin found in several plant species, such as *Cnidium monnieri* L, *Angelica archangelica* L, and *Angelica pubescens* Maxim. BMX was found to inhibit VEGF-induced proliferation, migration, and endothelial tube formation in HUVECs. These activities were also corroborated by ex vivo and in vivo studies decreasing VEGF-induced microvessel sprouting from aortic rings and HCT116 colorectal cancer cells. Moreover, BMX inhibited HCT116 cell proliferation and the growth of xenografts of HCT116 cells in vivo [136].

Scopoletin is a well-known natural coumarin with antiangiogenic properties. To develop new and robust angiogenesis inhibitors, several scopoletin derivatives were designed and synthesized. According to the study of Tabana et al. [137], scopoletin inhibited VEGF-A, ERK1, and FGF-2, and is thereby considered a strong antiangiogenic agent. In another study, among several scopoletin analogs, three compounds, including 4-bromo-phenyl and 4-chloro-phenyl scopoletin derivatives and 2-hydroxy-3-(piperidin-1-yl)-propoxy)-6-methoxy-2*H*-chromen-2-one, were able to inhibit VEGF-stimulated proliferation, migration, and tube formation of HUVECs. These results showed a significant decrease in the VEGF-triggered phosphorylated forms of ERK1/2 and Akt, which was corroborated by in vivo observations on chick chorioallantoic membrane [138]. Luo and co-workers also showed that 3-aryl-4-anilino/aryloxy-2*H*-chromen-2-one analogues significantly affected breast cancer through the inhibition of estrogen receptor-α and VEGFR-2 [139]. In another study, coumarin-conjugated benzophenone analogs showed promising antitumor activity against Ehrlich ascites carcinoma and Dalton's lymphoma ascites cell lines. In this study, a compound with a bromo group in the benzophenone structure markedly showed antiangiogenic effects through the inhibition of VEGF [28]. In a more recent study by Makowska et al. [140], a series of 2-imino-2*H*-chromen-3-yl-1,3,5-triazine compounds were synthesized. Among them, 4-[7-(diethylamino)- 2-imino-2*H*-chromen-3-yl]-6-(4-phenylpiperazin-1-yl)-1,3,5-triazin-2-amine showed the greatest cytotoxic effects against several human cancer cell lines, which underscores the promising role of synthetic coumarins in combating cancer. Figure 1 illustrates how coumarins inhibit various angiogenic signaling pathways.

### **6. Conclusions**

Considering the crucial role of angiogenesis in cancer development, antiangiogenic agents have significant potential to fight cancer. Thus, investigating novel drugs to attenuate or prevent angiogenesis-associated complications in cancer is of great importance. The several clinical limitations and side effects related to the administration of current antiangiogenic agents for cancer treatment raise the need to find alternative treatments. Natural and synthetic coumarins have shown a variety of pharmacological properties. They have demonstrated prominent anticancer effects by targeting multiple signaling pathways involved in several types of cancer.

Recently, studies have focused on the antiangiogenic effects of coumarins according to their structure-activity relationships. The present review reports the currently available literature data on the signaling and regulatory pathways of angiogenesis, as well as on antiangiogenic and anticancer mechanisms of natural and synthetic coumarins, critically analyzing and highlighting their use as possible therapeutic strategies. These studies are essential to identify novel and effective anticancer agents with fewer side effects than conventional drugs. It is also critical to identify potential synergies that may allow reducing the side effects of cytotoxic medicines and increasing the quality of life of patients. Additional studies should focus on additional in vitro and in vivo experiments followed by well-controlled clinical trials to reveal the exact signaling pathways involved in cancer angiogenesis as well as the precise pharmacological mechanisms of coumarins. In addition, there is a need to investigate and adjust novel antiangiogenic coumarin lead compounds to develop more potent and efficient anticancer drugs with lower toxicity. In addition, an appropriate drug delivery system should be introduced to overcome the existing pharmacokinetic challenges of coumarin administration. Such research will unveil the potential of coumarins in the prevention, attenuation, and treatment of angiogenesis in cancer.

**Author Contributions:** Conceptualization, M.B.M., and M.H.F.; drafting the manuscript, M.B.M., S.F., A.S., D.T., and M.H.F.; review and editing the paper: M.B.M., S.F., E.S.-S., M.H.F., C.R.C., P.B., and A.B.; revising, M.B.M., S.F., M.H.F., and A.B.

**Funding:** This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

**Acknowledgments:** The authors thank Yalda Shokoohinia for her excellent and helpful suggestions.

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

### **References**


© 2019 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* **Novel Antiretroviral Structures from Marine Organisms**

### **Karlo Wittine, Lara Safti´c, Željka Peršuri´c and Sandra Kraljevi´c Paveli´c \***

University of Rijeka, Department of Biotechnology, Centre for high-throughput technologies, Radmile Matejˇci´c 2, 51000 Rijeka, Croatia

**\*** Correspondence: sandrakp@biotech.uniri.hr; Tel.: +385-51-584-550

Academic Editor: Kyoko Nakagawa-Goto

Received: 2 September 2019; Accepted: 19 September 2019; Published: 26 September 2019

**Abstract:** In spite of significant advancements and success in antiretroviral therapies directed against HIV infection, there is no cure for HIV, which scan persist in a human body in its latent form and become reactivated under favorable conditions. Therefore, novel antiretroviral drugs with different modes of actions are still a major focus for researchers. In particular, novel lead structures are being sought from natural sources. So far, a number of compounds from marine organisms have been identified as promising therapeutics for HIV infection. Therefore, in this paper, we provide an overview of marine natural products that were first identified in the period between 2013 and 2018 that could be potentially used, or further optimized, as novel antiretroviral agents. This pipeline includes the systematization of antiretroviral activities for several categories of marine structures including chitosan and its derivatives, sulfated polysaccharides, lectins, bromotyrosine derivatives, peptides, alkaloids, diterpenes, phlorotannins, and xanthones as well as adjuvants to the HAART therapy such as fish oil. We critically discuss the structures and activities of the most promising new marine anti-HIV compounds.

**Keywords:** antiretroviral agents; anti-HIV; marine metabolites; natural products; drug development

### **1. Introduction**

Human immunodeficiency virus (HIV) infections pose a global challenge given that in 2017, according to the World Health Organization data, 36.9 million people were living with HIV and additional 1.8 million people were becoming newly infected globally (Table 1)**.** HIV targets immune cells and impairs the human defense against pneumonia, tuberculosis, and shingles as well as certain types of cancer [1]. The most advanced stage of HIV infection is the Acquired Immunodeficiency Syndrome (AIDS), which can take from two to 15 years to develop, depending on the individual [2].


**Table 1.** Summary of the global human immunodeficiency virus (HIV) epidemic (2017) according to World Health Organization (WHO) data.

HIV has two viral forms: HIV-1 (the most common form that accounts for around 95% of all infections worldwide) and HIV-2 (relatively uncommon and less infectious). HIV-1 consists of groups M, N, O, and P with at least nine genetically distinct subtypes of HIV-1 within group M (A, B, C, D, F, G, H, J, and K). Additionally, different subtypes can combine genetic material to form a hybrid virus known as the 'circulating recombinant form' (CRFs) (Figure 1). HIV-2 consists of eight known groups (A to H). Of these, only groups A and B are pandemic. The HIV-2 mechanism is not clearly defined and neither is its difference from HIV-1. However, the transmission rate is much lower in HIV-2 than in HIV-1. HIV-2 is estimated to be more than 55% genetically distinct from HIV-1.

**Figure 1.** HIV types and strains classification.

The HIV-1 genome has reading frames coding for structural and regulatory proteins. The *gag* gene encodes the Pr55Gag precursor of inner structural proteins p24 (capsid protein, CA), p17 (matrix protein, MA), p7 (nucleoprotein, NC), and p6 involved in the virus particle release. The *pol* gene encodes the Pr160GagPol precursor of the viral enzymes p10 (protease, PR), p51 (reverse transcriptase, RT), p15 (RNase H), and p32 (integrase, IN). The *env* gene encodes the PrGp160 precursor of the gp120 (surface glycoprotein, SU) and gp41 (transmembrane protein, TM). Other genes include *tat,* encoding p14 (transactivator protein), *rev,* encoding p19 (RNA splicing regulator), *nef,* encoding p27 (negative regulating factor), *vif,* encoding p23 (viral infectivity protein), *vpr,* encoding p15 (virus protein r), *vpu,* encoding p16 (virus protein unique), *vpx* in HIV2, encoding p15 (virus protein x), and *tev,* encoding p26 (tat/rev protein) [3].

The HIV infections are extremely problematic as the virus targets the CD4+ memory T-cells population, which is essential for organism immunity. HIV can attach itself to the host cell through 1) a relatively nonspecific interaction with negatively charged cell-surface heparan sulfate proteoglycans [4], 2) specific interactions between the Env and α4β7 integrin [5,6], and/or 3) the interaction with pattern-recognition receptors, such as the dendritic cell-specific intercellular adhesion molecular 3-grabbing non-integrin (DC-SIGN) [7]. The attachment of HIV in any of the abovementioned ways can increase the efficacy of infection because it brings Env, a heavily glycosylated trimer of gp120 and gp41 heterodimers, into close proximity with the viral receptor CD4 and co-receptor [8]. Finally, in order for the viral entry to occur, Env needs to bind itself to the host protein CD4 [9,10].

The binding of the HIV glycoprotein gp120 to the host cell CD4 receptor causes conformational changes of the gp120 glycoprotein, which uncover additional binding sites that interact with distinct proteins on the host cell membrane, known as β-chemokine co-receptors (mainly CCR5 and CXCR4), which facilitate the virus entry into the cell [11].

After the infection, a progressive decline of CD4 + cells consequently leads to the failure of the immune system function and the development of opportunistic infections that usually lead to death [1]. In HIV-infected patients, immunodeficiency develops both as a result of the viral replication and the failure of the patients' homeostatic mechanisms. The continuous viral presence in the patients after the application of therapy is attributed to the CD4 + T-cell homeostasis owing to a pool of latently infected and resting CD4 + T-cells, macrophages, and follicular dendritic cells that remain in the organism. Indeed, the complex interactions of the patient's immune system with the virus, and vice versa after the viral suppression, are thought to be crucial for the control of disease progression [12,13].

Current therapeutic approaches mainly target proteins that are vital for the viral cycle. One of the prominent examples is the linear 36-amino acid synthetic peptide enfuviritide (T20, Fuzeon), developed by Hoffmann-La Roche, and the first FDA approved fusion inhibitor for the treatment of HIV-1/AIDS acting through the binding to the gp41 subunit of the HIV-1 envelope glycoprotein. This induces a conformational change that brings the viral and cellular membranes into close enough proximity for the fusion and the subsequent viral entry into the host-cell to occur. Nevertheless, several restrictions, such as a low genetic barrier for drug resistance and a short in vivo half-life, limit its clinical use [14–17].

Other various FDA-approved antiretroviral drugs from seven mechanistic classes of inhibitors of the HIV replication are also available for the treatment of infected patients, namely, the nucleoside reverse-transcriptase inhibitors NRTIs, non-nucleoside reverse-transcriptase (RT) inhibitors NNRTIs, protein inhibitors PIs, fusion inhibitors, entry inhibitors—CCR5 co-receptor antagonists, HIV integrase strand transfer inhibitors, and multi-class combinations. None of the mentioned drug classes alone or in combination, the latter being known as the highly active antiretroviral therapy (HAART), can eradicate the HIV infection, and effective vaccines remain unavailable.

The difficulties of HIV-1 vaccine research are, in part, a result of 1) the unavailability of a model for natural immunity related to HIV; 2) the existence of genetically distinct subtypes of HIV and frequent mutations; 3) unidentified correlates of specific immune response to HIV; 4) lack of a reliable, non-human animal model for HIV infection (SIV in monkeys vs. HIV in humans).

The established latent pro-viral reservoirs in the patient's body can stochastically begin to reproduce viral particles, which makes the HIV disease practically incurable. From over 160 compounds identified so far as latency-reversing agents (LRAs), none have led to a promising cure [18].

Several rare and long-term remissions of HIV cases are described in the literature. For example, Berlin, London, and Düsseldorf's patients underwent bone marrow transplantation with stem cells from a donor with a rare genetic mutation of the CCR5. The Mississippi baby received a very early antiretroviral therapy that extended the time of the viral rebound for more than 27 months. There undoubtedly remains a lot to be learned from these cases, and further investigation of stem-cell transplantation in people living with HIV is required [19–21].

The currently used antiretroviral treatment, alone or in combination, extends the quality and life expectancy of HIV-infected individuals but does not cure them. Drug resistance, along with the emergence of drug-resistant virus strains, a high-cost of the lifetime treatment regimen, cell toxicity, and serious side effects of currently used anti-HIV drugs [2] underlie the need for a synthetic development of new drugs or the search for active anti-HIV molecules in natural sources. Mother Nature has been perfecting its chemistry for three billion years, and most of it has been done in water. Intense competition and feeding pressure as well as non-static marine environmental conditions yield compounds with chemical and structural features generally not found in terrestrial natural products.

New efficient molecules directed against HIV should demonstrate better performance in comparison with the currently approved drugs and suppress the HIV virus and/or eliminate the latent HIV reservoirs present in the human body.

Around 60% of drugs currently available on the market are derived or inspired by nature [22]. Turning to nature for drug development holds great potential, especially when it comes to marine organisms. Only few marine-derived drugs have been approved on the market so far but many are in the preclinical or clinical stage of development [23]. Marine organisms make up to two-thirds of Earth's species and produce, as a consequence of living in a highly competitive environment, unique and structurally diverse metabolites. Over the last 40 years, bioprospecting efforts have resulted in over 20,000 compounds of marine origin. The highest share of marine metabolites (up to 70%) are obtained from marine sponges, corals, and microorganisms, while mollusks, ascidians, and algae metabolites form only a minor part [24]. Oceans are, indeed, still a rather underexploited habitat, and biodiversity appears to be higher in the oceans than on land, which might be relevant when focusing on the marine environment as an untapped reservoir of novel antiretroviral candidates. In the discovery of new antiviral marine-derived drugs, researchers usually implement two strategies. They either screen the extracts from different strains (e.g., cyanobacteria, microalgae) or search directly for bioactive molecules in organisms—extract and purify them for evaluation within the drug development pipeline. It is thought that the marine environment might yield more potent anti-HIV candidates characterized by a higher efficiency (lower effective dose) and a better selectivity and which do not induce resistance development. This could, of course, only be speculation based on some of the previous success stories in the discovery of drugs from natural sources such as, e.g., lovastatin and paclitaxel. However, nature generally does create more sophisticated and perfected systems with a complex mode of action.

An excellent example is protein lectin, derived from marine red algae *Gri*ffi*thsia* sp. named Griffithsin with mid-picomolar activities, which groups it among the most potent HIV entry inhibitors reported so far [25]. It inhibits the HIV infection by binding itself to high mannose glycan structures on the surface of gp120, altering the gp120 structure or its oligomeric state [26]. This interaction relies on the specific trimeric "sugar tower," including N295 and N448 [27]. Griffitshin can also prevent infections caused by other glycoprotein-enveloped viruses such as the Ebola virus, hepatitis C virus, and the severe acute respiratory syndrome coronavirus. It has been shown that the dimerization of Griffithsin is necessary for a high potency inhibition of HIV-1 [28]. However, the discrepancy between the HIV gp120 binding activity and the HIV inhibitory activity points to the presence of mechanism unrelated to a merely simple HIV gp120 binding [26]. The most promising application of Griffithsin

would be its incorporation into vaginal and rectal gels, creams, or suppositories acting as an antiviral microbicide to prevent the transmission of HIV.

Despite the vast number of structurally diverse and unique bioactive molecules from the marine environment, the global marine pharmaceutical pipeline includes only eight approved drugs: Adcetris®, Cytosar-U®, Halaven®, Yondelis®, Carragelose®, Vira-A®, Lovaza®, and Prialt® [29]. Overall, it has taken 20 to 30 years from their discovery to their entry into the market. A sustainable supply, structural complexity, optimization of formulation, and ADMET properties, and a scale-up issue have prevented further development of several highly promising marine compounds. It is by no means an easy task to identify a marine candidate that may be considered as a potential drug. Initial high costs of developing a natural product into a drug could be balanced out with careful long-term considerations (biodiversity, supply, and technical, market) [30].

This paper provides an overview of natural marine metabolites that were first identified in the period between 2013 and 2018 or the previously identified marine constituents with a recently confirmed anti-HIV activity that could be potentially used or further optimized as novel ant-HIV agents. We also comprehensively summarize anti-HIV activities for several categories of marine structures including chitosan and its derivatives, sulfated polysaccharides, lectins, bromotyrosine derivatives, peptides, alkaloids, diterpenes, phlorotannins, and xanthones as well as fish oil as an auxiliary to HAART therapy.

### **2. Marine Compounds in the Treatment of HIV**/**AIDS**

### *2.1. Chitosan and Its Derivatives*

Chitosan (**2**, Figure 2), a natural marine byproduct, is a poly-cationic linear polysaccharide derived from chitin (**1**, Figure 2) after partial deacetylation. Chitin is a structural element in the exoskeleton of mainly shrimps and crabs and is mainly composed of the randomly distributed β-(1-4)-linked D-glucosamine and *N*-acetyl-D-glucosamine. It has been previously shown that this compound can exhibit a large scale of different bioactivities and can also be used as a carrier for anti-HIV drugs [31]. Chitosan is loaded with saquinavir, an anti-HIV drug with a protease inhibitory activity, which showed better cell targeting efficiency than saquinavir alone [32]. Furthermore, trimethyl chitosan has improved Atripla, an anti-HIV drug consisting of efavirenz, emtricitabine, and tenofovir disoproxil fumarate, anti-HIV 1 activity, and has allowed it to be used in lower concentrations [33]. The antiretroviral activity is manifested in the chitosan-specific cationic nature that allows the formation of electrostatic complexes or multilayer structures with other negatively charged polymers [34]. Karagozlu et al. reported about new QMW-COS and WMQ-COS oligomers with anti-HIV activities. These oligomers are conjugates of chitosan and the Gln-Met-Trp peptide, which were constructed as a continuation of the authors' previous research, in which a high potency of synthetically constructed chitosan oligomers was confirmed in anti-HIV therapy. More specifically, it was shown that these oligomers suppress syncytium formation, which occurs as a fusion of infected cells with neighboring cells, induced by HIV in a dose-dependent manner. However, the authors also noticed that after a certain period, the number of syncytia once again increased, suggesting that the cells should be re-treated with QMW-COS and WMQ-COS oligomers to maintain the primary therapeutically-relevant effect. The inhibition of the HIV-1 induced lytic effect, determined by the cell viability assay, showed that IC50 for QMW-COS was 48.14 μg/mL and was almost identical for WMQ-COS, 48.01 μg/mL. These oligomers effectively reduced the HIV load but showed no effects on HIV-1 RT and protease in vitro. Higher dosages were also required for the reduction in the HIV-1IIIB p24 antigen production assessed by the ELISA assay and the HIV-1RTMDR p24 antigen production. The highest difference between the compounds was reflected in IC50 values obtained from studies on the virus-induced luciferase activity in infected cells, where QMW-COS had a higher potency in comparison with WMQ-COS. Lastly, the authors determined the effects of oligomers on the interaction between gp41 and CD4 by using the CD4-gp41 ELISA assay, whereby both oligomers showed high potency. The effect of these oligomers was highest when they

were applied immediately upon the HIV-1 infection of cells, indicating that they should be used as a potential treatment in the early stages of HIV infection, probably at the entry stage [31].

**Figure 2.** Chemical structures of chitin (**1**) and chitosan (**2**).

### *2.2. Sulfated Polysaccharides*

Sulfated polysaccharides (SP) are the most studied class of antiviral polysaccharides that are structural components of the alga cell wall where they play both the storage and structural role. They are an important source of galactans, commercially known as agar and carrageenan in red alga (Rhodophyta), fucans (fucoidan, sargassan, ascophyllan, and glucuronoxylofucan) in brown alga (Phaeophyta), and ulvans-sulfated heteropolysaccharides that contain galactose, xylose, arabinose, mannose, glucuronic acid, or glucose [35–37]. Many studies indicate that, in marine algae, sulfated polysaccharides facilitate water and ion retention in extracellular matrices, which is an important mechanism for coping with desiccation and osmotic stress in a highly salted environment [38–40]. The antiviral activity of this group of compounds is mainly connected to the degree of sulfation, constituent sugars, molecular weight, conformation, and dynamic stereochemistry [41,42]. The effect of counter cation should also be considered as an important factor in observed biological activity.

The antagonizing effect of the negatively charged sulfated polysaccharides on the HIV-1 entry into cells may be due to 1) their binding onto the positively charged V3 domain of gp120, thereby preventing the virus attachment to the cell surface [43–45] or 2) the masking of the docking sites of gp120 for sCD4 on the surface of T lymphocytes, thereby disrupting the CD4-gp120 interaction [46–48] and subsequently inhibiting the expression of the viral antigen and the activity of the viral reverse transcriptase [49,50].

### 2.2.1. Heparan Sulfate

Heparinoid polysaccharides can interact with the positive-charge regions of cell-surface glycoproteins, leading to a shielding effect on these regions, which prevents the binding of viruses to the cell surface [51]. The sulfated polysaccharides content in marine mollusks is high in comparison with the bovine mucosal heparin (73.5%) and the porcine mucosal heparin (72.8%) [52]. The acidic sulfate groups on heparin (**3**, Figure 3), or heparin-like compounds, can inhibit HIV through electrostatic interactions with basic amino-acid residues of the transcriptional activator Tat protein [53].

**Figure 3.** Structure of heparan sulfate (**3**).

### 2.2.2. Fucose Containing SP

So far, the main anti-infectious activities documented for the fucose-containing SP are those against viruses [54]. More importantly, these polysaccharides are selective inhibitors of various enveloped viruses, including HIV [54–56]. FCSP acts during the early phase of infection by blocking the virus attachment and entry into the host cells, but may also inhibit subsequent replication stages in vitro [57].

### 2.2.3. Fucoidans

Three fucoidans extracted from three brown seaweeds (*Sargassum mcclurei*, *Sargassum polycystum*, *Turbinara ornata*) inhibit the early stages of HIV-1 entry into target cells, with IC50 ranging from 0.33 to 0.7 μM. Neither the sulfate content nor the position of sulfate groups are related to the anti-HIV activity of fucoidans, suggesting the involvement of other structural parameters such as the molecular weight, the type of glycosidic linkage, or even a unique fucoidan sequence [56]. Although the presence of sulfo-groups seems to be necessary for anti-HIV activity [58], these data do not support random sulfation as the main antiviral factor.

Sulfated fucan polysaccharides, ascophyllan (**4**, Figure 4)**,** and two fucoidans (S and A) (**5** and **6**, Table 2), derived from different sources, significantly inhibit (IC50 1.3; 0.3; 0.6 μg/mL) the early step of HIV-1 (R9 and JR-Fl) infection. They also inhibit the VSV-G-pseudotype HIV-1 infection in HeLa cells [59].

**Figure 4.** Structure of ascophyllan (**4**) and fucoidan unit.

**Table 2.** Chemical composition of polysccharides (Fuc, Fucose; Xyl, Xylose; Glu, Glucose; Man, Mannose; Gal, Galactose) in ascophyllan, S- and A-fucoidan.


Chondroitin sulfate with fucosylated branches (FuCS) (**7**, Figure 5) has also attracted attention as an HIV antiviral compound. Depolymerized fucosylated CS, extracted from the sea cucumber, has shown in vitro activity against a range of viral strains, including the resistant ones [60]. FuCS is effective in blocking the laboratory strain HIV-1IIIB entry and replication by inhibiting the p24 antigen production (4.26 and 0.73 μg/mL, respectively) and the infection of the clinic isolate HIV-1KM018 and HIV-1TC-2 (23.75 and 31.86 μg/mL, respectively) as well as suppressing the HIV-1 drug-resistant virus. Additionally, FuCS is also effective in T-20-resistant strains (EC50 values ranging from 0.76 to 1.13 μg/mL). The depolymerized fragments seem to maintain a similar anti-HIV action at the early stages of infection, apparently through interaction with an HIV envelope glycoprotein gp120. The sulfated fucose branches appear necessary for antiviral activity, which is also affected by molecular weight and carboxylation [61]. While the in vitro results of the fucosylated CS against HIV are promising, it is questionable whether the antiviral activity would be maintained in vivo. Other polyanionic HIV entry inhibitors, which advanced into clinical trials, failed to prove effective against the heterosexual HIV-1 transmission. This was related to factors not considered in previous development stages, such as the presence of seminal plasma and the concentration and retention of polyanionic inhibitors [62].

**Figure 5.** General chemical structure of fucosylated chondroitin sulfate (**7**).

The complex chemical architecture and the sulfate patterning of marine polysaccharides depends on numerous factors (species, tidal cycles, environmental variations (e.g., salinity), harvesting season, plant age, geographical location etc.) [39,63–69], making isolation, purification, and comprehensive chemical characterization a highly challenging task [70]. The development of many polysaccharides into clinical application is hindered by the still limited view of their sophisticated and diverse nature. Despite having good antiviral effects, the use of carbohydrate drugs is still in its infancy, and intensive structure-activity and in vivo studies are needed in the future.

A relatively new strategy in inducing immunity and developing an HIV vaccine is to use carbohydrates. The major difficulty of such an approach lies in mimicking the specific glycan protective epitope. Gp120 of HIV is a highly glycosylated envelope surface glycoprotein responsible for the receptor and co-receptor binding, which, together with gp41, comprises the heterodimeric envelope trimer spikes of HIV. *N*-linked glycans, mainly mannose and complex-type, cover much of the gp120 surface-accessible face of the HIV envelope spike forming the glycan shield. Inadequate mimicry of the glycan shield, tolerance mechanisms, and/or the inability to induce a domain-exchange are reflecting difficulties in creating the proper specificity of Abs [71]. Most of the vaccines for HIV-1 in preclinical trials are based on a Manα1-2Man oligomannosyl epitope (various conjugates, engineered yeast strains, and modified glycoproteins) [72–79]. Better specificity could potentially be gained using carbohydrates of marine origin.

### *2.3. Lectins*

Lectins are a group of proteins that specifically, but reversibly, bind glycosylated molecules on the cell surface. Precisely, this group of molecules can affect cell-cell interactions, protect cells from pathogens, influence cell adhesion, and affect the intracellular glycoprotein translocation [80]. Recently, lectins have become promising agents for antiretroviral therapy, and different researches have confirmed their anti-HIV properties. Their antiretroviral activity is manifested through an alteration of the interaction between HIV gp120 or gp41 and the corresponding receptors [81], which, in the end, inhibit the HIV cell function, HIV infectivity, and the formation of the syncytium, multi-nucleated cells [82–84].

Several published review papers describe the previously found marine lectins with antiretroviral action [85,86]. For example, Gogineni et al. reported about some new, unusual lectins, such as the β-galactose specific lectin (CVL), CGL, DTL, DTL-A, SVL-1, and SVL-2 [86]. Additionally, Akkouh et al. reported about some new algal lectins, such as *Boodlea Coacta* Lectin, Griffithsin and *Oscillatoria Agardhii* Agglutinin (OAA), and some cyanobacterial lectins, such as Cyanovirin-N, Scytovirin, Microcystis Viridis Lectin, and Microvirin.

However, in the last few years, there has not been as much research focused on anti-HIV lectins from marine sources. Only Hirayama et al. (2016) reported about the new high-mannose specific lectin and its recombinants that possess anti-HIV activity [87]. In their research performed on the red alga *Kappaphycus alvarezii,* authors confirmed KAA-1 and KAA-2, two KAA mannose-binding lectin isomers, as potent anti-HIV agents. The anti-HIV role of action of these two compounds includes a strong binding to the virus envelope glycoprotein gp120 and, consequently, the inhibition of HIV entry into the host cells. These KAA recombinants, as well as the native one, inhibited the HIV-1 entry at IC50s (neutralization assay in Jurkat cells) of 7.3–12.9 nM. Authors concluded in the end that KAAs, besides their strong inhibitory effect on HIV entry into the cells, have a potential as agents in treatments against other viruses possessing high mannose glycans on their envelope as well.

### *2.4. Peptides*

It has been shown that the majority of marine peptides have strong anti-HIV activity. They are usually isolated from marine organisms through the process of enzymatic hydrolysis [88]. The most common source of such constituents is marine sponges that are known for their unique metabolome [89] and are a source of more than 36% of all marine bioactive compounds [90]. Their bioactive peptides can be found in cyclic or linear forms and contain unusual amino acids that form unique structures rarely found in other species. Antiretroviral activity of such structures works on several different levels: blocking of virus entry, inhibition of the cytopathic viral activity, neutralization of viral particles, or inhibition of viral fusion and entry [89,91].

Recently, Shin et al. discovered two new depsipeptides from marine sponges *Stelletta sp*., stellettapeptin A (**8**, Figure 6), and stellettapeptin B (**9**, Figure 6), with the inhibition of the cytopathic effect of HIV-1 infection [92]. Confirming the mentioned theory about the unique metabolome of marine sponges, the authors revealed that these two compounds have previously undescribed nonproteinogenic amino-acid parts on peptides that are rarely found in nature. Namely, stellettapeptin A and stellettapeptin B have an unexpected polyketide subunit, 3-hydroxy-6,8-dimethylnon-4-enoic acid, 3-OHGln, and 3-OHAsn residues. Their high potency is witnessed through low EC50 values (inhibition of the cytotoxic effect upon HIV infection)—values of 23 nM for stellettapeptin A and 27 nM for stellettapeptin B.

**Figure 6.** Structures of stelletapeptin (**8**) A and stelletapeptin B (**9**).

Furthermore, newly discovered anti-HIV constituents derived from marine sponges *Verongula rigida* and *Aiolochoria crassa* with amino-acid structure were published by Gomez-Archila et al. (2014) [93]. In their paper, they evaluated and confirmed the anti-HIV effect of 11 bromotyrosine derivatives (Table 3), whereby aeroplysinin-1 (**10**), 19-deoxyfistularin 3 (**15**), purealidin B (**16**), fistularin 3 (**17**) and 3-bromo-5-hydroxy-O-methyltyrosine (**18**, Figure 7) were the most potent in their anti-HIV activity. Aeroplysinin 1 (**15**) and purealidin B (**16**), compounds found in *V. rigida* species inhibited the HIV-1 replication in a dose-dependent manner by more than 50%. Specifically, for aeroplysinin 1, HIV-a replication was inhibited by 74% at a concentration of 20 μM, whereas purealidin was less

potent with inhibitory power of 57% at a concentration of 80 μM. These two compounds had been previously isolated; however, their anti-HIV activity was proven in this research. The same was with 3-bromo-5-hydroxy-*O*-methyltyrosine (**18**) that has a relatively high percentage of inhibition of HIV activity (47%) in a dose-dependent manner. However, the exact mechanism of action remains unclear. In the same study, additional tests with these compounds on the HIV RT inhibition (qPCR of the early and late transcripts), nuclear import (qPCR analysis of 2-LTR transcript), and HIV entry inhibition (viral infectivity assay) were performed. The results showed that aeroplysinin-1 (**10**), 19-deoxyfistularin 3 (**15**), purealidin B (**16**), fistularin 3 (**17**), and 3-bromo-5-hydroxy-*O*-methyltyrosine (**18**) influenced the nuclear import of the HIV virus with around or more than 50% of inhibition: aeroplysinin-1 (**10**) showed 67% of inhibition at 10 μM, 19-deoxyfistularin 3 62% inhibition at 20 μM, purealidin B 66% of inhibition at 20 μM, fistularin 3 47% of inhibition at 10 μM, and 3-bromo-5-hydroxy-*O*-methyltyrosine 73% of inhibition at 80 μM. Viral RT inhibition was not high for all compounds, whereby the highest results were around 50% of inhibition. For example, purealidin B had 58% of inhibition at 20 μM in the qPCR analysis of early transcripts. As for the HIV entry inhibition, all compounds were active in a dose-depended manner, with the highest results of inhibition obtained for 3,5-dibromo-*N*,*N*,*N*,*O*-tetramethyltyraminium (**13**), from 14% to 30%. Finally, the authors stressed the structural similarity of these compounds with the HIV integrase and protease inhibitors, suggesting that these compounds can have a broader mode of antiviral action.

**Figure 7.** Structures of aeroplysinin-1 (**10**), dihydroxyaerothionin (**11**), 3,4-dibromo-*N,N,N*trimethyltyraminium (**12**), 3,5-dibromo-*N,N,N,O*-tetramethyltyraminium (**13**), purealidin R (**14**), 19 deohxyfistularin 3 (**15**), purealidin B (**16**), fistularin-3 (**17**), 3-bromo-5-hydroxy-*O*-methyltyrosine (**18**), 3-bromo-*N,N,N*-trimethyltyrosinium (**19**), and 3,5-dibromo-*N,N,N*-trimethyltyrosinium (**20**).


Summaryofanti-HIVcompoundsfrommarineorganisms.






**Table 3.**

*Cont.*




**Table 3.** *Cont.*

285


**Table**

**3.**

*Cont.*






Marine sponges are not the sole source of bioactive proteins. For example, Jang et al. reported about a new small hydroxyproline-rich peptide from Alaska Pollack collagen (APHCP, **21**, Figure 8) that exhibits a unique antiviral activity [94]. This peptide is a Gly-Pro-Hyp-Gly-Pro-Hyp-Gly-Pro-Hyp-Gly peptide, and the authors showed that the most important part of a peptide for anti-HIV activity is the hydroxyl group at hydroxyproline, whereas a peptide with only prolines does not exhibit antiviral activity. Its anti-HIV 1 mode of action is manifested through the inhibition of the induced syncytia formation by the interference of an HIV fusion, inhibition of cell lysis, RT activity, and the production of the p24 antigen. It was shown that APHCP can decrease the HIV-1 induced cell lysis at a potency around EC50 of 459 μM (EC50 against anti-HIV-1 induced cell lysis—MTT assay). Additionally, through the inhibition of the viral RT at EC50 at 374 μM, this peptide's crucial role in the inhibition of the conversion of viral RNA to DNA was also confirmed. With EC50 of 405 μM, this compound effectively suppressed the p24 production in viral cells, as determined by the Western blot analysis.

**Figure 8.** Structure of the Alaska Pollack collagen hydroxyl proline (APCHP) peptide (**21**).

Similarly, one new anti-HIV peptide was isolated from *Spirulina maxima* (SM-peptide) [95]–the Leu-Asp-Ala-Val-Asn-Arg peptide, and the authors showed its HIV-1 infection inhibition in a human T cell line MT4. The peptide inhibited cell lysis, p24 antigen production, and HIV-1 RT. Specifically, IC50 (obtained by a cell viability assay) against an anti-HIV 1 infection was determined as 0.691 mM, the inhibition of the HIV-1-induced RT activation (RT assay kit) in MT4 cells was at a high 90% at a concentration of 1.093 mM, and the p24 production (p24 antigen production assay) was inhibited at 95% at a concentration of 1.093 mM.

### *2.5. Alkaloids*

Marine organisms are well-established sources of natural alkaloids. Although the term 'alkaloid' seems puzzling and is prone to scientific controversy, alkaloids are generally defined as nitrogen-containing compounds derived from plants and animals. Relatively few alkaloids from marine sources have been found to possess antiretroviral properties and, so far, none have found their clinical use.

Aspernigrin C (**22**, Figure 9) and malformin C (**23,** Figure 9) have been isolated from marine-derived black aspergili, *Aspergillus niger* SCSIO Jcw6F30, and their inhibitory activity against the chemokine receptor subtype 5 (CCR5) tropic HIV-1 SF162 has been evaluated. They show potent inhibition of infection with IC50 values 4.7 ± 0.4 μM and 1.4 ± 0.06 μM, which is comparable to the nucleoside reverse transcriptase inhibitor—abacavir (IC50 = 0.8 ± 0.1 μM) and the HIV-1 entry inhibitor ADS-J1 (IC50 = 0.8 ± 0.1 μM). In comparison to other aspernigrins, it has been suggested that the 2-methylsuccinic moiety is responsible for the potency of aspernigrin C [96].

**Figure 9.** Structures of aspernigrin C (**22**) and malformin C (**23**).

Thiodiketopiperazine-type alkaloids, eutypellazines A-M, isolated from the EtOAc extract of the fermentation broth of deep-sea sediment fungus *Eutypella sp*. shows potent inhibitory effects against pNL4.3.Env-.Luc co-transfected 293T HIV model cells. Eutypellazine E (**24**, Figure 10) exerts activity in a low micromolar range (IC50 = 3.2 ± 0.4 μM), while eutypellazine J (**25**, Figure 10) shows a reactivating effect toward latent HIV-1 in J-Lat A2 cells. This could be used as a promising strategy to expunge the HIV-1 infection by activating latent virus cellular reservoirs in combination with HAART [97].

**Figure 10.** Structures of eutypellazine E (**24**) and eutypellazine J (**25**).

The *S. carteri* Red Sea sponge extract yields three previously characterized compounds: debromohymenialdisine (DBH) (**26**, Figure 11), hymenialdisine (HD) (**27**, Figure 11), and oroidin (**28**, Figure 11). DBH and HD exhibited a 30–40% inhibition of HIV-1 at 3.1 μM and 13 μM but with associated cytotoxicity. Conversely, oroidin displayed a 50% inhibition of viral replication at 50 μM without observed cytotoxicity. Also, it showed inhibition of HIV-1 reverse transcriptase up to 90% at 25 μMc [98].

**Figure 11.** Structures of debromohymenialdisine (**26**), 10Z-hymenialdisine (**27**), and oroidin (**28**).

The two known alkaloids of the aaptamine family containing 1*H*-benz[*de*]-1,6-naphthyridine skeleton, namely 3-(phenetylamino)demethyl(oxy)aaptamine (**29**, Figure 12) and 3-(isopentylamino)demethyl(oxy)aaptamine (**30**, Figure 11), were isolated from the sponge *A. aptos*. They exhibited anti-HIV activity, with inhibitory rates of 88.0% and 72.3%, respectively, at a concentration of 10 μM [99].

**Figure 12.** Structures of 3-(phenetylamino)dimethyl(oxo)aaptamine (**29**) and 3-(isopentylamino)dimethyl (oxo)aaptamine (**30**).

Bengamide A (**31**, Figure 13), haliclonycyclamine A+B (**32**, Figure 13) and keramamine C (**33**, Figure 13) inhibit HIV-1 with a 50% effective concentration of 3.8 μM or less. The most potent among them, bengamide A, blocked HIV-1 in a T cell line with an EC50 of 0.015 μM (which was comparable to control antiretrovirals indinavir 0.029 μM, efavirenz 0.0024 μM, and raltegravir 0.011 μM) and in peripheral blood mononuclear cells with EC50 of 0.032 μM. It was concluded that HIV-1 LTR NF-κB response elements are required for a bengamide A-mediated inhibition of LTR-dependent gene expression [100].

**Figure 13.** Structures of bengamide A (**31**), haliclonacyclamines A + B (**32**), keramamine C (**33**).

Phenylspirodrimane, stachybotrin D (**34**, Figure 14) isolated from the sponge-derived fungus *Stachybotrys chartarum* MXH-X73, was discovered to be a HIV-1 *RT* inhibitor, which showed inhibitory effects on the wild type (EC50 8.4 μM) and five NNRTI-resistant HIV-1 strains (EC50 7.0; 23.8; 13.3; 14.2; 6.2 μM) [101].

**Figure 14.** Structure of stachybotrin D (**34**).

### *2.6. Diterpenes*

Many terpenes from marine natural products demonstrated anti-HIV properties. Mechanisms of action involve blocking of different steps of the HIV-1 replicative cycle as reverse transcriptase inhibitors, protease inhibitors, or entry inhibitors. Among them, diterpenes from marine algae are nowadays in the spotlight due to their promising anti-HIV activities [102]. Dolabellane diterpenes are compounds from the diterpene group that have recently been extensively studied for their anti-HIV activity. Pardo-Vargas et al. characterized three new dolabellane diterpenes isolated from the marine brown alga *Dictyota pfa*ffi*i* from Northeast Brazil: (1R\*,2E,4R\*,7S,10S\*,11S\*,12R\*)10,18-diacetoxy-7-hydroxy-2,8(17)-dolabelladiene, (1R\*,2E,4R\*,7R\*,10S\*,11S\*,12R\*)10,18-diacetoxy-7-hydroxy-2,8(17)-dolabelladiene,

(1R\*,2E,4R\*,8E,10S\*,11S,12R\*)10,18-diacetoxy-7-hydroxy-2,8-dolabelladiene, named dolabelladienols A–C (**35**–**37**, Figure 15), respectively [102]. In particular, the new compounds, dolabelladienols A and B, showed potent anti-HIV-1 activities that can be confirmed with their low IC50 values of 2.9 and 4.1 μM and low cytotoxic activity against MT-2 lymphocyte tumor cells. These promising anti-HIV-1 agents were even more active than previously known 2,6-dolabelladienes series.

**Figure 15.** Structures of the new dolabellane diterpenoids dolabelladienols A–C (**35**–**37**).

De Souza Barros et al. tested marine dolastanes (**38**, **40**, Figure 16) and secodolastane diterpenes (**39**, Figure 16) isolated from the brown alga *Canistrocarpus cervicornis* for anti-HIV-1 activity [103]. They observed that the marine diterpenes **38**–**40** inhibit the HIV-1 replication in a dose-dependent manner (EC50 values of 0.35, 3.67, and 0.794 μM) without a cytotoxic effect (CC50 values ranging from 935 to 1910 μM). Additionally, they investigated the virucidal effect of these diterpenes and their potential use as microbicides. Dolastane-diterpenes **38** and **40** showed a potent effect on HIV-1 infectivity, whereas no virucidal effect was observed for secodolastane diterpene **39**, demonstrating another mechanism of antiretroviral activity. Therefore, the authors suggested a potential use of marine dolastanes **38** and **40** as microbicides that could directly inhibit virus infectivity and possibly act before the virus penetrates the target cells [103].

**Figure 16.** Structures of marine dolastanes (**38** and **40**) and secodolastane diterpene (**39**) derived from *Canistrocarpus cervicornis*.

Dolabelladienetriol from brown alga *Dictyota* spp has also been evaluated as a potential microbicide against HIV-1 in tissue explants. Namely, Stephens et al. examined the 8,10,18-trihydroxy-2,6-dolabelladiene (**41**, Figure 17) in pretreated peripheral blood cells (PBMC) and macrophages along with their protective effect in the ex vivo explant model of the uterine cervix [104]. Pre-treatment of peripheral PBMC and macrophages with dolabelladienotriol showed inhibitory effects on HIV-1 replication. Furthermore, in the explant model dolabelladienetriol inhibited viral replication in a dose-dependent manner from 20 to 99% in concentrations of 0.15 and 14.4 μM without a loss in the viability of the tissue. The authors concluded that this compound has great potential as a possible microbicide. The same compound was also theoretically analyzed as an inhibitor of the wild-type and mutants' HIV-1 reverse transcriptase [105]. Firstly, the structure-activity relationship studies revealed that a low dipole moment and high HOMO (highest occupied molecular orbital)-LUMO (lowest unoccupied molecular orbital) gap values are related to the antiviral activity. Secondly, molecular docking studies with RT wild-type and mutants showed a seahorse-like conformation of 8,10,18-trihydroxy-2,6-dolabelladiene, hydrophobic interactions, and hydrogen bonds with important residues of the binding pocket. Finally, the authors suggested a new derivative of the 8,10,18-trihydroxy-2,6-dolabelladiene with an aromatic moiety in the double bond to improve its biological activity.

**Figure 17.** Structure of (1*R*,2*E*,4*R*,6*E*,8*S*,10*S*,11*S*,12*R*)-8,10,18-trihydroxy-2,6-dolabelladiene (**41**).

Although dolabellane diterpenes of brown alga *Dictyota* spp showed a strong anti-HIV-1 activity, this was not confirmed for dolabellane diterpenes isolated from octocorals. Therefore, some chemical transformations have been conducted to improve the anti-HIV-1 potency of the main dolabellane 13-keto-1(*R*),11(*S*)-dolabella-3(*E*),7(*E*),12(18)-triene from Caribbean octocoral *Eunicea laciniata* [106]. Oxygenated dolabellanes derivatives (**42**–**44**, Figure 18), obtained by epoxidation, epoxide opening, and allylic oxidation of ketodolbellatriene have shown significantly improved antiviral activities and a low cytotoxicity to MT-2 cells, which makes them promising antiviral compounds.

**Figure 18.** Structures of semi-synthesized oxygenated dolabellanes (**42**–**44**) originally isolated from the Caribbean octocoral *Eunicea laciniata*.

### *2.7. Phlorotannins and Xanthones*

Phlorotannins are tannin derivatives made from several phloroglucinol units linked to each other in different ways. Phlorotannins contain phenyl linkage (fucols), ether linkage (fuhalols and phlorethols), phenyl and ether linkage (fucophloroethols), and dibenzodioxin linkage (eckols) [86,107]. So far, a series of phlorotannins have been identified with potent anti-HIV activity. For example, 8,8 -bieckol and 6,6 -bieckol from marine brown alga *Ecklonia cava* has shown an enhanced HIV-1 inhibitory effect [112,113]. Karadeniz et al. reported that 8,4-dieckol (**45**, Figure 19) is another phlorotannin derivative isolated from the same brown alga that could be used as a drug candidate for the development of new generation anti-HIV therapeutic agents [107]. The compound showed HIV-1 inhibitory activity at noncytotoxic concentrations. More precisely, the results indicated that 8,4----dieckol inhibited the cytopathic effects of HIV-1, including HIV-1 induced syncytia formation, lytic effects, and viral p24 antigen production. Furthermore, 8,4----dieckol inhibited an HIV-1 entry and RT enzyme with the inhibition ratio of 91% at a concentration of 50 μM.

**Figure 19.** Chemical structure of 8,4-dieckol (**45**) from *E. cava*.

Recently, for the first time, xanthone dimer was identified as a potential anti-HIV-1 agent [108]. Xanthones are secondary metabolites from higher plant families, fungi, and lichen [114,115]. Although structurally related to flavonoids, xanthones are not as frequently encountered in nature [9]. Penicillixanthone A (PXA) (**46**, Figure 20), a natural xanthone dimer, has been isolated from the jellyfish-derived fungus *Aspergillus fumigates* with fourteen other natural products [108]. However, only penicillixanthone A showed inhibitory activities in an HIV infection. Marine-derived PXA displayed potent anti-HIV-1 activity against CCR5-tropic HIV-1 SF162 and CXCR4-tropic HIV-1 NL4-3, with IC50 of 0.36 and 0.26 μM, respectively. A molecular docking study confirmed that PXA might bind

to either CCR5 or CXCR4 to prevent HIV entry into target cells. Therefore, PXA, as a CCR5/CXCR4 dual-coreceptor antagonist, may be seen as a new potential lead product type for the development of anti-HIV therapeutics.

**Figure 20.** Structure of penicillixanthone A (**46**).

### *2.8. Fish Oil as an Adjuvant to HAART Therapy*

HAART therapy can cause severe side effects, e.g., insulin resistance, lipoatrophy, dyslipidemia, and abnormalities of fat distribution. Therefore, finding an adequate diet and supplementation to lower the negative effects of the HAART combination therapy is desirable [116]. Fish oil contains omega-3 polyunsaturated fatty acids (PUFA), eicosapentaenoic (EPA, 20:5n-3) (**47**, Figure 21) and docosahexaenoic (DHA, 22:6n-3) (**48**, Figure 21) acids, which may have beneficial effects for HIV-infected patients. It has been shown that the addition of fish oil to the diet of HIV-infected individuals receiving usual antiretroviral therapy can significantly lower serum triglycerides levels [117], which is highly relevant knowing that HIV dyslipidemia is a serious problem related to an increased frequency of cardiovascular disease.

**Figure 21.** Structures of eicosapentanoic (**47**) and docosahexanoic acid (**48**).

Recently, He et al. analyzed the influence of DHA on the locomotor activity in ethanol-treated HIV-1 transgenic rats [109]. The prevalence of alcohol use and alcohol abuse in infected individuals is much higher, and numerous ethanol and HIV-1 viral proteins have synergistic effects on inflammation in the central nervous system [118–120]. HIV remains in the body in its latent form after HAART therapy and, as such, can induce neuroinflammation. DHA depletion has been found to be associated with various neurological abnormalities, and its administration can have a neuroprotective effect. DHA taken daily could reverse the effects of the ethanol negative effect on the locomotor activity in the presence of HIV viral proteins. An in vivo study, using real-time quantitative PCR, showed that the addition of DHA can reduce elevated levels of IL-6, IL-18, and increase the expression of NF-κB in the striatum. This proved the potential of this fish oil constituent as an adjuvant in HIV patients' treatment that can help in lowering the interactive effects of ethanol consumption during HIV infection.

### *2.9. Others*

Resorcyclic acid lactones, namely radicicol (**49**, Figure 22) pochonin B (**50**, Figure 22) and C (**51**, Figure 22) isolated from *H. fuscoatra* exhibited a 92–98% reactivation efficiency of the latent HIV-1 relative to SAHA (subeoylanilide hydroxamic acid, vorinostat, HDAC inhibitor) and EC50 of 9.1, 39.6 and 6.3 μM [110]. The reactivation strategy is, indeed, a promising strategy to expunge the HIV-1 infection by reactivating latent viral loads, mainly in CD4 + T-cells, which quickly rebound when antiviral treatment is interrupted. It was noted that all active compounds contain Michael acceptor functionality. The PKC-independent mechanism of reactivation of the latent HIV-1 remains to be elucidated.

**Figure 22.** Structures of radicicol (**49**), pochonin B (**50**), pochonin C (**51**).

A team of researchers led by Zhao isolated new isoprenylated cyclohexanols from the sponge-associated fungus *Truncatella angustata* named truncateols O-V [111]. In vitro testing showed that truncateols O and P (**52** and **53**, Figure 23), analogues bearing the alkynyl group in the side chain, exhibit a significant inhibition toward the HIV-1 virus with IC50 values of 39.0 μM and 16.1 μM, respectively. These compounds could be considered as new anti-HIV lead compounds due to lower cytotoxicity (CC50 > 100 μM) in comparison with the positive control efavirenz (CC50 = 40.6 μM).

**Figure 23.** Structures of truncateols O (**52**) and P (**53**).

### **3. Future Directions in the Anti-HIV Marine Drug Development**

Marine organisms have been acknowledged as a precious source of bioactive compounds that may provide novel anti-HIV structures or lead structures for structural optimization. A large amount of evidence from scientific research confirmed a high biological potential of these compounds to treat serious diseases, including infective ones. Some of the marine-derived bioactive compounds discovered much earlier have emerged with novel properties and potential applications after a decade or two. Isolation and structural elucidation of compounds from marine organisms is not an easy task and still carries challenges. Identification of all the compounds is a daunting task, especially with regards to complex structural motifs that may be present in a single marine extract. Taxonomic knowledge is still insufficient to enable unambiguous species classification that can result in the false prediction of chemical constituents and hamper structural analysis. Furthermore, a temporal lag between the discovery, chemical characterization, and associated pharmacological activities is quite common, and the majority of marine metabolites are usually tested for anticancer activity, whereas anti-HIV and other possible biological effects are neglected or mostly not performed due to a lack of

funding. Targeted assays and in vivo analyses are similarly performed only for some of the potential candidates, while the translation into clinical trials remains very limited. Thus, the financial gap is certainly a relevant factor contributing to the slow drug development process in this area. In particular, the development of anti-HIV compounds, which act by mechanisms that differ from existing antivirals, requires a well-designed and focused approach to studying the mode of action. Libraries should be created for specifically defined crude extracts, their corresponding simplified fractions as well as for pure compounds for a well-balanced natural product discovery program. Additionally, there exist but few publications in which scientists have tried to modify known compounds of marine origin to improve their bioactivity. We are, however, continuously witnessing advancements in the deep-sea exploration technology, sampling strategies, genome sequencing, genome mining, genetic engineering, chemo-enzymatic synthesis, nanoscale NMR structure determination, and development and optimization of suitable fermentation strategies to ensure a continued supply of unique bioactive compounds from the oceans. Therefore, the grounds have been met for a broad, international effort based on scientific collaboration that would rely on well-equipped infrastructure and human resources as a prerequisite for a full advancement in the field and development of new drug candidates for the pharmaceutical market in the future.

**Author Contributions:** K.W. devised the main conceptual idea and, together with L.S. and Ž.P., wrote the manuscript parts related to medicinal chemistry. K.W. and Ž.P. performed literature searches and L.S. prepared the table. S.K.P. participated in the manuscript writing and wrote and discussed parts relevant for clinical applications, shaped the paper concept, and performed the final revision.

**Funding:** We want to thank the Croatian Government and the European Union (European Regional Development Fund—the Competitiveness and Cohesion Operational Programme—KK.01.1.1.01) for funding this research through project Bioprospecting of the Adriatic Sea (KK.01.1.1.01.0002) granted to The Scientific Centre of Excellence for Marine Bioprospecting—BioProCro. We also acknowledge the project "Research Infrastructure for Campus-based Laboratories at the University of Rijeka," co-financed by European Regional Development Fund (ERDF) and the University of Rijeka research grant uniri-biomed-18-133 (1277).

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

### **References**


© 2019 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* **Cytotoxic E**ff**ects of Diterpenoid Alkaloids Against Human Cancer Cells**

### **Koji Wada \* and Hiroshi Yamashita**

Department of Medicinal Chemistry, Faculty of Pharmaceutical Sciences, Hokkaido University of Science, 4-1, Maeda 7-jo 15-choume, Teine-ku, Sapporo 006-8590, Japan; yama@hus.ac.jp

**\*** Correspondence: kowada@hus.ac.jp; Tel.: +81-11-681-2161

Academic Editor: Kyoko Nakagawa-Goto Received: 23 May 2019; Accepted: 21 June 2019; Published: 22 June 2019

**Abstract:** Diterpenoid alkaloids are isolated from plants of the genera *Aconitum*, *Delphinium*, and *Garrya* (Ranunculaceae) and classified according to their chemical structures as C18-, C19- or C20-diterpenoid alkaloids. The extreme toxicity of certain compounds, e.g., aconitine, has prompted a thorough investigation of how structural features affect their bioactivities. Therefore, natural diterpenoid alkaloids and semi-synthetic alkaloid derivatives were evaluated for cytotoxic effects against human tumor cells [A549 (lung carcinoma), DU145 (prostate carcinoma), MDA-MB-231 (triple-negative breast cancer), MCF-7 (estrogen receptor-positive, HER2-negative breast cancer), KB (identical to cervical carcinoma HeLa derived AV-3 cell line), and multidrug-resistant (MDR) subline KB-VIN]. Among the tested alkaloids, C19-diterpenoid (e.g., lipojesaconitine, delcosine and delpheline derivatives) and C20-diterpenoid (e.g., kobusine and pseudokobusine derivatives) alkaloids exhibited significant cytotoxic activity and, thus, provide promising new leads for further development as antitumor agents. Notably, several diterpenoid alkaloids were more potent against MDR subline KB-VIN cells than the parental drug-sensitive KB cells.

**Keywords:** diterpenoid alkaloids; cytotoxicity; human tumor cells; lipojesaconitine; delcosine; delpheline; kobusine; pseudokobusine

### **1. Introduction**

Cancer therapy mainly involves surgery, chemotherapy, radiation therapy, immunotherapy, monoclonal antibody therapy, and hormone therapy. Chemotherapy generally refers to the use of cytotoxic drugs to treat cancer. Plant alkaloids are one major class of chemotherapeutic drugs [1–9]. Chemotherapeutic drugs that affect cell division by preventing the normal functioning of micro-tubules include the vinca alkaloids.

Numerous diterpenoid alkaloids have been isolated from various *Aconitum*, *Delphinium*, and *Garrya* (Family Ranunculaceae) species and are classified according to their chemical structures as C18-, C19 or C20-diterpenoid alkaloids (Figure 1) [10,11]. The C19-diterpenoid alkaloids may be divided into six types: aconitine, lycoctonine, pyro (C8=C15 or C15=O), lactone (δ-valerolactone rather than cyclopentyl C-ring), 7,17-*seco*, and rearranged ones [10,11]. Most of the isolated C19-diterpenoid alkaloids are aconitine- and lycoctonine-types and include aconitine, mesaconitine, hypaconitine and jesaconitine, all of which are extremely toxic. The C20-diterpenoid alkaloids may be divided into ten types: atisine, denudatine, hetidine, hetisine, vakognavine, napelline, kusnezoline, racemulosine, arcutine, and tricalysiamide [10,11]. Most of the isolated C20-diterpenoid alkaloids are atisine-, hetisine-, and napelline-types and include atisine, kobusine, pseudokobusine and lucidusculine, which are far less toxic [12].

**Figure 1.** Classifications, general structures and numbering systems for C18-, C19-, and C20-diterpenoid alkaloids.

The pharmacological properties of the C19-diterpenoid alkaloids have been studied extensively and reviewed [12]. Aconitine is a toxin that exhibits activity both centrally and peripherally, acting predominantly on the cardiovascular and respiratory systems by preventing the normal closing of sodium channels [12]. This extreme toxicity resulted in the use of *Aconitum* extracts as poisons in hunting and warfare [13], although extracts were also used as traditional medicines by oral and topical routes. For example, the roots of *Aconitum* plants have been used as "bushi", an herbal drug in some prescriptions of traditional Japanese medicine for the treatment of hypometabolism, dysuria, cardiac weakness, chills, neuralgia, gout, and certain rheumatic diseases [14]. However, proper processing is essential to reduce the content of toxic alkaloids and avoid inadvertent poisoning [15–17]. Such obstacles encourage a good understanding of the relationships between structure and cytotoxic activity of aconitine and related compounds before they can be considered for modification and development as chemotherapeutic agents.

Our previous study demonstrated the effects of various naturally occurring and semi-synthetic C19- and C20-diterpenoid alkaloids on the growth of the A172 human malignant glioma cell line [18]. Antitumor properties and radiation-sensitizing effects of various types of novel derivatives prepared from C19- and C20-diterpenoid alkaloids were also investigated [19]. Two novel hetisine-type C20-diterpenoid derivatives showed significant suppressive effects against the Raji non-Hodgkin's lymphoma cell line [20]. In addition, the effects of various semi-synthetic novel hetisine-type C20-diterpenoid alkaloids on the growth of the A549 human lung cancer cell line were examined and subsequent structure-activity relationships for the antiproliferative effects against A549 cells were considered [21]. Since 2012, several diterpenoid alkaloid components and their derivatives exhibited antiproliferative activity against human tumor cell lines, including A549 (lung carcinoma), DU145 (prostate carcinoma), MDA-MB-231 (estrogen and progesterone receptor-negative & HER2-negative triple-negative breast cancer), MCF-7 (estrogen receptor-positive, HER2-negative breast cancer), KB (identical to cervical carcinoma HeLa derived AV-3 cell line), and multidrug-resistant (MDR) subline KB-VIN [P-glycoprotein (P-gp) overexpressing vincristine-resistant KB subline]. Among such alkaloids, C19-diterpenoid (e.g., lipojesaconitine, delpheline, and delcosine derivative) and C20-diterpenoid (e.g., kobusine and pseudokobusine derivatives) alkaloids have shown significant antiproliferative activity, as well as provided promising new leads for further development as antitumor agents.

### **2. Antiproliferative Activity of C19-Diterpenoid Alkaloid Derivatives**

### *2.1. Aconitine-Type C19-Diterpenoid Alkaloids*

The tested aconitine-type C19-diterpenoid alkaloids included 21 natural alkaloids, aconitine (**1**), deoxyaconitine (**2**), jesaconitine (**3**), deoxyjesaconitine (**4**), aljesaconitine A (**5**), secojesaconitine (**6**), mesaconitine (**8**), hypaconitine (**9**), hokbusine A (**10**), 14-anisoyllasianine (**12**), *N*-deethylaljesaconitine A (**13**), aconine (**14**), lipomesaconitine (**15**), lipoaconitine (**16**), lipojesaconitine (**17**), neolinine (**18**), neoline (**19**), 14-benzoylneoline (**20**), isotalatizidine (**21**), karacoline (**22**), and 3-hydroxykaracoline (**23**), isolated from the rhizoma of *Aconitum japonicum* THUNB. subsp. *subcuneatum* (NAKAI) KADOTA [22–28] (Figure 2). Two synthetic aconitine-type C19-diterpenoid alkaloids, 3,15-diacetyljesaconitine (**7**) [26]

and 3-acetylmesaconitine (**11**) [29] prepared from secojesaconitine (**6**) and mesaconitine (**8**), respectively (Figure 2), were also tested.

\*: Natural alkaloid

As = COC6H4OCH3 (*p*)

lipo = linoleoyl, palmitoyl, oleoyl, stearoyl, linolenoyl

**Figure 2.** Chemical structures of aconitine-type C19-diterpenoid alkaloids **1**–**23**.

Eighteen of the 23 tested aconitine-type C19-diterpenoid alkaloids, both natural alkaloids (**1**~**6**, **8**~**10**, **12**~**14**, **18**~**23**) and synthetic analogs (**7** and **11**), were inactive (IC50 > 20 or 40 μM) [27,28,30] (Table 1). Three natural diterpenoid alkaloids (**15**~**17**) exhibited cytotoxic activity against five human tumor cell lines (A549, MDA-MB-231, MCF-7, KB, and MDR KB subline KB-VIN) (Table 1). Lipojesaconitine (**17**) showed significant cytotoxicity against four tested cell lines with IC50 values of 6.0 to 7.3 μM, but weak cytotoxicity against KB-VIN (IC50 = 18.6 μM) [28]. Lipomesaconitine (**15**) showed moderate cytotoxicity against the KB cell line (IC50 = 9.9 μM), but weak cytotoxicity against the other four human tumor cell lines (IC50 =17.2 ~ 21.5 μM) [27]. Lipoaconitine (**16**) was weakly cytotoxic (IC50 = 13.7 ~ 20.3 μM) against all five human tumor cell lines [28]. Based on the results, the fatty acid ester at C-8 and the anisoyl group at C-14 found in **17** may be important to the cytotoxic activity of aconitine-type C19-diterpenoid alkaloids.


**Table 1.** Cytotoxic activity data for aconitine-type C19-diterpenoid alkaloids and derivatives **1**–**23**.

<sup>1</sup> Values are means ± standard deviation; <sup>2</sup> Paclitaxel (PXL; nM) was used as an experimental control.

### *2.2. Lycoctonine-Type (7,8-diol) C19-Diterpenoid Alkaloids*

The tested lycoctonine-type (7,8-diol) C19-diterpenoid alkaloid group included 12 natural alkaloids, namely nevadensine (**24**), *N*-deethylnevadensine (**25**), and virescenine (**27**), purified from rhizoma of *Aconitum japonicum* subsp. *subcuneatum* [27], and 18-methoxygadesine (**26**), delphinifoline (**28**), delcosine (**34**), 14-acetyldelcosine (**34–43**), and 14-acetylbrowniine (**35**), purified from root of *Aconitum yesoense* var. *macroyesoense* (NAKAI) TAMURA [31–34], and andersonidine (**30**), pacifiline (**31**), pacifinine (**32**), and pacifidine (**33**), purified from seeds of *Delphinium elatum* cv. Pacific Giant [35] (Figure 3). The remaining tested C19-diterpenoid alkaloids from this subtype were synthetic alkaloids, *N*-deethyldelsoline (**29**) [18], 1-acetyldelcosine (**34-1**) [36], 1,14-diacetyldelcosine (**34-2**) [37], 1-(4-trifluoromethylbenzoyl)delcosine (**34-24**) [30], delsoline (**34-42**) [37], 1,14-di-(4-nitrobenzoyl)-delcosine (**34-45**) [30], 14-acetyl-1-(4-nitrobenzoyl)delcosine (**34-46**) [30], and 1-acyl or 1,14-diacyldelcosine derivatives (**34-3**~**34-23**, **34-25**~**34-41**, **34-44**, and **34-47**) [38], prepared from delcosine (**34**) or delsoline (**34-42**) (Figure 3). These 42 C19-diterpenoid alkaloids were evaluated for antiproliferative activity against four to five human tumor cell lines (A549, DU145, MDA-MB-231, MCF-7, KB, and KB-VIN) [30,38] (Table 2). Several tested lycoctonine-type (7,8-diol) C19-diterpenoid alkaloids, both natural alkaloids (**24**~**28**, **30**~**33**) and a synthetic alkaloid (**29**), were inactive (IC50 > 20 or 40 μM). All tested delcosine derivatives that contain an acetyl or methoxy group, both natural alkaloids (**34**, **34-43**, **35**) and synthetic analogs (**34-1**, **34-2**, **34-42**), were inactive (IC50 > 20 μM). However, acylation, except with an acetyl group, of the C-1 and/or C-14 hydroxy group of **34** led to various degrees of antiproliferative activity. Among the C-1 esterified alkaloids, the synthetic derivatives **34-6**, **34-8**, **34-10**, and **34-18** exhibited significant potency against all cell lines (average IC50 9.3, 5.3, 5.0, and 6.9 μM, respectively). Also, alkaloids **34-3**, **34-16**, **34-17**, **34-21**, **34-25**, **34-27**, **34-31**, **34-32**, **34-38**, and **34-40** showed moderate potency toward all cell lines (average IC50 12.7−20.7 μM). While alkaloid **34-32** displayed good antiproliferative activity (IC50 8.7 μM) against KB cells, it was much less potent against A549, MDA-MB-231, and KB-VIN cells. Alkaloids **34-5**, **34-13**, **34-15**, **34-29**, **34-35**, **34-37**, and **34-41** exhibited only weak potency against all cell lines (average IC50 22.0−26.5 μM). Finally, alkaloids **34-24**, **34-30**, and **34-34** were inactive against all five human tumor cell lines, while **34-12**, **34-33**, and **34-39** showed limited potency.


**Figure 3.** *Cont.*


\*: Natural alkaloid.

**Figure 3.** Chemical structures of lycoctonine-type (7,8-diol) C19-diterpenoid alkaloids **24**–**35**.

Among the derivatives esterified at both C-1 and -14, alkaloids **34-19** and **34-20** exhibited significant potency against all five tested cell lines (average IC50 4.9 and 5.0 μM, respectively). Alkaloid **34-9** (average IC50 11.9 μM) showed significant antiproliferative activity against MDA-MB- 231 and KB cells (IC50 4.7 and 5.8 μM, respectively) comparable with **34-19** and **34-20**, but was less potent against MCF-7 and A549 (IC50 12.2 and 24.8 μM, respectively) and inactive against KB-VIN. Alkaloid **34-23** exhibited only weak potency toward all cell lines (average IC50 23.7 μM). Alkaloids **34-4**, **34-7**, **34-11**, **34-14**, **34-26**, **34-36**, **34-45**, **34-46**, and **34-47** were inactive against all five human tumor cell lines, while **34-22** and **34-28** showed limited potency.


**Table 2.** Cytotoxic activity data for lycoctonine-type (7,8-diol) C19-diterpenoid alkaloids and synthetic analogs of delcosine **24**~**35**.

<sup>1</sup> Values are means ± standard deviation; <sup>2</sup> Paclitaxel (PXL; nM) was used as an experimental control.

Particularly, C-1 monoacylated delcosine derivatives (**34-3**, **34-6**, **34-8**, **34-10**, **34-13**, **34-21**, **34-25**, **34-27**, and **34-35**) were significantly more potent compared with corresponding C-1,14 diacylated delcosine derivatives (**34-4**, **34-7**, **34-9**, **34-11**, **34-14**, **34-22**, **34-23**, **34-26**, **34-28** and **34-36**). Thus, a C-1 acyloxy group and C-14 hydroxy group are crucial for enhanced antiproliferative activity of **1**-derivatives. Regarding alkaloids **34-18** (pentafluorobenzoate at C-1, hydroxy at C-14), **34-19** (pentafluorobenzoate at C-1 and C-14), and **34-20** (pentafluorobenzoate at C-1, acetate at C-14), all three alkaloids were essentially equipotent against three of the five tumor cell lines, while **34-18** was somewhat less potent than the diacylated alkaloids against MCF-7 and KB-VIN cells.

Striking observations from the data in Table 2 were the consistent identities of the most potent alkaloids. Alkaloids **34-8**, **34-10**, **34-19**, and **34-20** exhibited the highest potency against all five tested tumor cell lines with IC50 values ranging from 4.3 to 6.0 μM. The same range of potency was found with alkaloid **34-18** against A549 cells, with alkaloids **34-9** and **34-18** against MDA-MB-231 cells, and with **34-6**, **34-9**, and **34-18** against KB cells. The potencies of **34-6** and **34-17** (IC50 5.6−11.8 μM) generally ranked somewhat below those of the most potent alkaloids, except against the MCF-7 cell line, where they were even less active.

The identity of the substituent(s) on the acyl group affected the antiproliferative potency. Notably, among the 1,14-diacyl and 1-acyl-14-acetyl derivatives, only alkaloids **34-19** and **34-20** with one or two pentafluorinated benzoyl esters, respectively, showed significant potency against all five tested cell lines. Alkaloid **34-9** with two 3-nitro-4-chlorobenzoyl groups showed good potency against certain cell lines. Similarly, the 1-monoacylated alkaloids with the highest potency against the five tumor cell lines contained 3-nitro-4-chloro- (**34-8**) and pentafluoro- (**34-18**) as well as 4-dichloro-methyl- (**34-10**) benzoyl esters. The chlorinated alkaloids **34-8** and **34-10** as well as **34-6**, which has 3,5-dichloro substitution on the benzoate ester, were more potent than **34-5** with only a single chloro group or **34-13** with chloro and fluoro groups. Similarly, alkaloid **34-18** showed increased antiproliferative activity against the five tumor cell lines compared with other fluorinated alkaloids **34-13**~**34-17**, **34-21**~**34-27**. Moreover, with some exceptions against certain cell lines, alkaloids with bromo (**34-3** and **34-4**), dimethylamino (**34-12**), dimethoxy (**34-29**), trimethoxy (**34-30**), diethoxy (**34-31**), benzyloxy (**34-32**), cyano (**34-33**), methylenedioxy (**34-34** and **34-35**), nitro (**34-45** and **34-46**), and ethoxy (**34-47**) substituted benzoate esters or phenylacetyl (**34-37**), cinnamoyl (**34-38** and **34-39**), 1-naphthoyl (**34-40**), and anthraquinone-2-carbonyl (**34-41**) esters were less potent or inactive.

Interestingly, the active alkaloids were generally effective against P-gp overexpressing MDR subline KB-VIN, while alkaloids such as vincristine and paclitaxel are ineffective due to excretion from the MDR cells by P-gp. These results suggest that these diterpenoids are not substrates for P-gp.

### *2.3. Lycoctonine-Type (7,8-methylenedioxy) C19-Diterpenoid Alkaloids*

The tested lycoctonine-type (7,8-methylenedioxy) C19-diterpenoid alkaloids included 19 natural alkaloids, delcorine (**36**), delpheline (**37**), pacinine (**38**), yunnadelphinine (**39**), melpheline (**40**), bonvalotidine C (**41**), *N*-deethyl-*N*-formylpaciline (**42**), *N*-deethyl-*N*-formylpacinine (**43**), isodel-pheline (**44**), pacidine (**45**), eladine (**46**), *N*-formyl-4,19-secopacinine (**47**), *N*-formyl-4,19-secoyunna-delphinine (**48**), iminoisodelpheline (**49**), iminodelpheline (**50**), laxicyminine (**51**), *N*-deethyl-19-oxo-isodelpheline (**52**), *N*-deethyl-19-oxodelpheline (**53**), and 19-oxoisodelpheline (**54**), purified from seeds of *Delphinium elatum* cv. Pacific Giant [35,39–42] (Figure 4). The remaining 22 tested C19-diterpenoids were synthetic derivatives (**37-1**~**37-22**) [43] prepared from **37** (Figure 4).

### **Figure 4.** *Cont.*

delpheline

**37-4** 6-(4-anisoyl)delpheline **37-9** 6-(3-trifluoromethylbenzoyl)-

**37-5** 6-(4-phenylbenzoyl)delpheline **37-10** 6-(3,5-dinitrobenzoyl)delpheline

\*: Natural alkaloid

**Figure 4.** Chemical structures of lycoctonine-type (7,8-methylenedioxy) C19-diterpenoid alkaloids **36**-**54**.

All tested lycoctonine-type (7,8-methylenedioxy) C19-diterpenoid alkaloids were evaluated for antiproliferative activity against human tumor cell lines [30,40–43] (Table 3). The lycoctonine-type (7,8-methylenedioxy) C19-diterpenoid alkaloids, both the natural alkaloids (**36**~**54**) and synthetic analogs that did not contain a C-6 ester group (**37-20** and **37-22**), were inactive (IC50 > 20 or 40 μM). Among the C-6 esterified alkaloids, **37-1**, **37-17**, and **37-18** exhibited the highest average potency toward four tested cell lines (A549, DU145, KB and KB-VIN; average IC50 9.83, 9.57, and 9.41 μM, respectively). Alkaloids **37-3**, **37-5**~**37-7**, **37-9**, **37-10**, **37-12**, **37-13**, **37-16**, and **37-19** showed moderate potency against all tested cell lines (average IC50 13.9−20.8 μM). However, alkaloid **37-13** showed significantly increased cytotoxic activity (IC50 10.2 μM) against A549 cells compared with **37-1**, **37-17**, and **37-18**, but was generally less potent against DU145 and KB cells. While alkaloids **37-12**, **37-13**, **37-16**, and **37-19** displayed good antiproliferative activity (IC50 6.8, 9.1, 6.5, and 4.7 μM, respectively) against KB-VIN cells, they were much less potent against A549, DU145, and KB cells. Alkaloids **37-4** and **37-21** were inactive against all tested cancer cell lines, while **37-2**, **37-8**, **37-11**, and **37-14** exhibited only weak potency toward all cell lines (average IC50 23.0−29.2 μM).

The most noticeable observations from the data in Table 3 were the degree and relative ratio of KB/KB-VIN potency. Among the four cancer cell lines tested, the highest potency was found against the KB-VIN cell line by alkaloids **37-17**~**37-19** (IC50 4.22, 4.40, and 4.71 μM, respectively), followed by alkaloids **37-16**, **37-12**, **37-1**, **37-13**, and **37-9** (IC50 6.50, 6.80, 8.27, 9.10, and 11.9 μM, respectively). Generally, all active alkaloids showed the highest potency against the KB-VIN cell line compared with the other three tested cancer cell lines. Moreover, alkaloids **37-12**, **37-16**, **37-13**, and **37-19** showed over two-fold selectivity between the two cell lines (ratio of IC50 KB/IC50 KB-VIN: 2.15, 2.28, 2.31, and 2.57, respectively). Alkaloids **37-2**, **37-5**, and **37-17** displayed weak selectivity between the KB and KB-VIN cell lines (ratio of IC50 KB/IC50 KB-VIN: 1.55, 1.36, and 1.62, respectively). Finally, alkaloids **37-1**, **37-3**, **37-6**~**37-9**, **37-11**, **37-14**, **37-15**, and **37-18** displayed similar potency against the KB and KB-VIN cell lines (ratio of IC50 KB/IC50 KB-VIN: 1.07, 1.17, 1.06, 1.21, 1.04, 1.25, 1.07, 1.07, 1.17, and 1.23, respectively).

The identity of the substituent on the C-6 acyl group affected the cytotoxic potency. For instance, the alkaloids with the highest potency against the KB-VIN cell line contained chloro (**37-1**), fluoro (**37-12**, **37-18**, and **37-19**), trifluoromethyl (**37-9**, **37-13**, and **37-18**), ethoxy (**37-16**), or benzyloxy (**37-17**) substituents on the acyl group. Against the KB-VIN cell line, alkaloids **37-18** and **37-19** with both fluoro and trifluoromethyl/methyl groups were more potent than **37-9** with only a single trifluoromethyl group and even more potent than **37-2** with a single fluoro group. Similarly, alkaloid **37-13** showed increased cytotoxic activity against most cell lines compared with the related fluorinated alkaloids **37-14**

and **37-15**. Moreover, alkaloids with nitro, methoxy, phenyl, trifluoromethoxy, trifluoromethythio, and methyl carboxylate groups on a C-6 benzoate ester were generally less potent.

**Table 3.** Cytotoxic activity data for lycoctonine-type (7,8-methylenedioxy) C19-diterpenoid alkaloids and synthetic analogs of delpheline **36**~**54**.


<sup>1</sup> Values are means ± standard deviation; <sup>2</sup> Paclitaxel (PXL; nM) was used as an experimental control.

### **3. Antiproliferative Activity of C20-Diterpenoid Alkaloid Derivatives**

### *3.1. Actaline and Napelline-Type C20-Diterpenoid Alkaloids*

One natural actaline-type C20-diterpenoid alkaloid [44], aconicarchamine A (**55**), isolated from rhizoma of *Aconitum japonicum* subsp. *subcuneatum* [30], (Figure 5) and seven natural napelline-type C20-diterpenoid alkaloids, lucidusculine (**57**), flavadine (**58**), 12-acetyllucidusculine (**59**), 1-acetyl-luciculine (**60**), dehydrolucidusculine (**61**), dehydroluciculine (**62**), and 12-acetyldehydroluciduscu-line (**63**), purified from roots of *Aconitum yesoense* var. *macroyesoense* [31–33], (Figure 5) were tested. Seven synthetic napelline-type C20-diterpenoid alkaloid derivatives (**56-1**~**56-7**) [18,32,45] were prepared from luciculine (**56**) (Figure 5) and tested also. All tested actaline- and napelline-type C20-diterpenoid alkaloids were evaluated for antiproliferative activity against four to five human tumor cell lines [28,30] (Table 4). Tested actaline- and napelline-type C20-diterpenoid alkaloids, both the natural alkaloids (**55** and **57**~**63**) and synthetic analogs (**56-1**~**56-4**, **56-6**, and **56-7**), were inactive (IC50 > 20 or 40 μM). Among the

synthetic alkaloids, alkaloid **56-5** exhibited only weak potency toward the tested cell lines (A549, DU145, KB and KB-VIN; average IC50 27.8 μM). Because the related alkaloids **57**, **60, 56-2**~**56-4**, **56-6**, and **56-7** were inactive against all tested cancer cell lines, a C-1 hydroxy group, C-12 acyloxy group, and C-15 acetoxy group found in **56-5** could be needed for antiproliferative activity of luciculine derivatives.

**Figure 5.** Chemical structures of actaline and napelline-type C20-diterpenoid alkaloids **55**~**63**.

**Table 4.** Cytotoxic activity data for actaline and napelline-type C20-diterpenoid alkaloids **55**~**63** and synthetic analogs **56-1**~**56-7** of luciculine.


<sup>1</sup> Values are means ± standard deviation; <sup>2</sup> Paclitaxel (PXL; nM) was used as an experimental control.

### *3.2. Hetisine-Type (Analogs of Kobusine) C20-Diterpenoid Alkaloids*

Tested hetisine-type (analogs of kobusine) C20-diterpenoid alkaloids included four natural alkaloids, ryosenamine (**64**), 9-hydroxynominine (**65**), and torokonine (**66**), isolated from rhizoma of *Aconitum japonicum* subsp. *subcuneatum* [27,28] (Figure 6) and kobusine (**67**), purified from roots of *Aconitum yesoense* var. *macroyesoense* [31] (Figure 6). Nineteen synthetic derivatives (**67-1**~**67-19**) [18,21, 30,46,47] (Figure 6) prepared from **67** were tested also.

All tested hetisine-type (kobusine analogs) C20-diterpenoid alkaloids were evaluated for antiproliferative activity against four human tumor cell lines [27,28,30] (Table 5). Fifteen of the 23 alkaloids, both natural (**64**~**67**) and synthetic (**67-1**~**67-4**, **67-6**, **67-9**, **67-11**, **67-12**, **67-15**~**67-17**), were inactive (IC50 > 20 or 40 μM). Kobusine derivatives **67-5**, **67-7**, **67-10**, **67-18**, and **67-19** exhibited the highest average potency over the four tested cell lines (A549, DU145, KB and KB-VIN; average IC50 7.8, 6.1, 6.2, 6.8, and 4.7 μM, respectively), and alkaloids **67-8**, **67-13**, and **67-14** showed moderate potency (average IC50 16.6, 14.3, and 11.6 μM, respectively). However, while alkaloid **67-14** showed good cytotoxic activity (IC50 9.6 μM) against DU145 cells, it was much less potent against A549, KB, and KB-VIN cells.


\*: Natural alkaloid

**Figure 6.** Chemical structures of hetisine-type (analogs of kobusine) C20-diterpenoid alkaloids **64**~**67-19**.


**Table 5.** Cytotoxic activity data for hetisine-type C20-diterpenoid alkaloids **64**~**67** and synthetic derivatives **67-1**~**67-19** of kobusine.

<sup>1</sup> Values are means ± standard deviation; <sup>2</sup> Paclitaxel (PXL; nM) was used as an experimental control.

Among these analogs of **67**, esterification of C-15 in addition to C-11 increased potency significantly (compare **67-8** to **67-10**) or even converted an inactive to an active alkaloid (compare **67-3** to **67-5**, **67-6** to **67-7**, **67-16** to **67-18**). Consequently, all of the most potent analogs (**67-5**, **67-7**, **67-10**, **67-18**, and **67-19**) of **67** were esterified at both C-11 and C-15.

Striking observations from the data in Table 5 were the degree and comparative ratio of KB/KB-VIN potency. Five alkaloids (**67-5**, **67-7**, **67-10**, **67-18**, and **67-19**) were quite potent (IC50 < 10 μM) against KB-VIN. Indeed, alkaloid **67-19** exhibited a significantly low IC50 value of 3.1 μM. The ratios of KB to KB-VIN (IC50 KB/IC50 KB-VIN) were greater than 0.73 for all active alkaloids, with many alkaloids displaying comparable potency against the two cell lines, in contrast with paclitaxel (ratio of 0.0067). Alkaloid **67-19** showed over 1.3-fold selectivity with the greatest cytotoxic activity against KB-VIN (IC50 KB/IC50 KB-VIN: 1.32).

### *3.3. Hetisine-Type (Analogs of Pseudokobusine) C20-Diterpenoid Alkaloids*

The two tested natural hetisine-type (analogs of pseudokobusine) C20-diterpenoid alkaloids pseudokobusine (**68**) and 15-veratroylpseudokobusine (**68-11**) were purified from the roots of *Aconitum yesoense* var. *macroyesoense* [31,32] (Figure 7). The 36 tested synthetic derivatives (**68-1**~**68-10**, **68-12**~**68-37**) [18,21,30,32,46–49] (Figure 7) were prepared from **68**.

All tested hetisine-type (**68** analogs) C20-diterpenoid alkaloids were evaluated for antiproliferative activity against four human tumor cell lines [30] (Table 6). Many alkaloids, both natural alkaloids (**68** and **68-11**) and synthetic analogs (**68-1**~**68-3**, **68-6**, **68-8**, **68-9**, **68-14**, **68-16**~**68-18**, **68-21**, **68-23**, **68-25**~**68-31**, **68-33**~**68-37**), were inactive (IC50 > 20 μM). The pseudokobusine derivatives **68-5**, **68-15**, **68-19**, **68-20**, **68-24**, and **68-32** exhibited the highest average potency over the tested cell lines (A549, DU145, KB and KB-VIN; average IC50 7.0, 5.2, 5.3, 7.4, 7.1, and 6.1 μM, respectively). Alkaloids **68-7**, **68-10**, **68-12**, **68-13**, and **68-22** showed moderate potency over all tested cell lines (average IC50 13.5-16.8 μM). However, although alkaloid **68-10** showed good cytotoxic activity (IC50 8.0 μM) against A549 cells, it was much less potent against DU145, KB, and KB-VIN cells.



**Figure 7.** *Cont.*



\*: Natural alkaloid

**Figure 7.** Chemical structures of hetisine-type (analogs of pseudokobusine) C20-diterpenoid alkaloids **68**~**68-37**.


**Table 6.** Cytotoxic activity data for hetisine-type C20-diterpenoid alkaloids pseudokobusine (**68**) and its synthetic analogs **68-1**~**68-37**.

<sup>1</sup> Values are means ± standard deviation; <sup>2</sup> Paclitaxel (PXL; nM) was used as an experimental control.

Among the analogs of **68**, four C-11 mono-substituted alkaloids (**68-15**, **68-20**, **68-24**, and **68-32**) and two C-11,15 di-esterified alkaloids (**68-5** and **68-19**) exhibited average IC50 values of less than 10 μM. Certain C-11 (**68-7**, **68-10**, and **68-22**), C-6,11 (**68-4** and **68-12**) and C-6,15 (**68-13**) esterified alkaloids were generally less potent, while all C-6 (**68-3**, **68-6**, **68-14**, and **68-23**) and C-15 (**68-1**, **68-11**, **68-16**, **68-25**, **68-28**, and **68-30**) mono-substituted alkaloids, as well as the tri-substituted analog (**68-18**), were inactive. Thus, all more active (IC50 < 10 μM) C20-diterpenoid alkaloids in this classification had an ester or ether group on the C-11 hydroxy and were 11-monoester/11,15-diester analogs of **68** (OH at C-6).

The data in Table 6 led to noticeable observations about the degree and comparative ratio of KB/KB-VIN potency. Six alkaloids (**68-5**, **68-15**, **68-19**, **68-20**, **68-24**, and **68-32**) were quite potent (IC50 < 10 μM) against KB-VIN. Indeed, alkaloid **68-32** exhibited a low IC50 value of 5.2 μM. The ratios of KB to KB-VIN (IC50 KB/IC50 KB-VIN) were greater than 0.70 for all active alkaloids, with many alkaloids displaying comparable potency against the two cell lines, in contrast with paclitaxel (ratio of 0.0067). Alkaloids **68-12**, **68-13**, and **68-20** showed over 1.3-fold selectivity with their greatest cytotoxic activity against KB-VIN (IC50 KB/IC50 KB-VIN: 1.34, 1.48, and 1.44, respectively).

In mechanism of action studies on selected diterpenoid alkaloids, the hetisine-type C20-diterpenoid alkaloid derivatives **68-7** and **68-22** showed important suppressive effects against Raji cells. Further study indicated that **68-22** inhibited extracellular signal-regulated kinase phosphorylation but induced enhanced phosphoinositide 3 kinase phosphorylation, leading to accumulation of Raji cells in the G1 or sub G1 phase [20]. More investigation is certainly warranted.

#### **4. Discussion**

We have synthesized acylated derivatives of various C19- and C20-diterpenoid alkaloids. Totally, 199 natural alkaloids and their derivatives were evaluated against four to five human tumor cell lines. Among all alkaloids, 128 alkaloids were non-toxic (IC50 > 20 or 40 μM) and 51 alkaloids showed moderate antiproliferative effects (average IC50 = 10–40 μM). General summaries are described briefly below, and the most active compounds are shown in Figure 8.

**Figure 8.** Most potent tested diterpenoid alkaloids & structure-activity correlations.

Among the aconitine-type C19-diterpenoid alkaloids, the fatty acid ester at C-8 and the anisoyl group at C-14 found in **17** may be important to the cytotoxic activity. Compounds without the fatty acid ester at C-8 were inactive, and compounds with an unsubstituted benzoyl group at C-14 were less potent.

Among the C19-diterpenoid alkaloids, the most active alkaloids were lycoctonine-type C19-diterpenoid alkaloids with two different substitution patterns, C-1 (delcosine derivatives) and C-6 (delpheline derivatives). Delcosine derivatives **34-6**, **34-8**, **34-10**, and **34-18**, which are acylated at the C-1 hydroxy, as well as delpheline derivatives **37-1**, **37-17**, and **37-18**, which are acylated at the C-6 hydroxy, exhibited the greatest potency over all tested cell lines, including MDR KB-VIN.

Among the lycoctonine-type (7,8-diol) C19-diterpenoid alkaloids, a C-1 acyloxy group and C-14 hydroxy group were important for improved antiproliferative activity. The C-1,14 diacylated delcosine derivatives were generally less potent than corresponding C-1 monoacylated delcosine derivatives. The 1-monoacylated alkaloids with the highest potency (IC50 4−6 μM) against five tested cell lines contained 3-nitro-4-chloro- (**34-8**) and pentafluoro- (**34-18**) as well as 4-dichloromethyl- (**34-10**) benzoyl esters. Two or one pentafluorinated benzoyl esters were also found in the two most consistently potent alkaloids (**34-19** and **34-20**) among the 1,14-diacyl and 1-acyl-14-acetyl derivatives.

Among the lycoctonine-type (7,8-methylenedioxy) C19-diterpenoid alkaloids, none of the tested compound reached the potency levels of the most active 7,8-diol compounds. However, three 6-acylated delpheline derivatives **37-17**~**37-19** did show significant potency against the KB-VIN cell line (IC50 4.22, 4.40, and 4.71 μM, respectively). Interestingly, the two latter compounds contained fluorinated benzoyl esters. In addition, among 19 tested delpheline derivatives, four compounds (**37-12**, **37-16**, **37-13**, and **37-19**) showed over two-fold selectivity between the MDR and parental cell lines (ratio of IC50 KB/IC50 KB-VIN: 2.15, 2.28, 2.31, and 2.57, respectively).

None of the 15 tested actaline- and napelline-type C20-diterpenoid alkaloids showed significant antiproliferative potency. Only 12-benzoyllucidsuculine (**56-5**) with C-1 hydroxy, C-12 acyloxy, and C-15 acetoxy groups showed even weak potency.

Among C20-diterpenoid alkaloids, the most active alkaloids were hetisine-type C20-diterpenoid alkaloids with two different substitution patterns, C-11,15 (kobusine) and C-6,11,15 (pseudo-kobusine). Hetisine-type C20-diterpenoid alkaloids **67-5**, **67-7**, **67-10**, **67-18**, **67-19**, **68-5**, **68-15**, **68-19**, **68-20**, **68-25**, and **68-32**, which are acylated or tritylated at the C-11 hydroxyl, exhibited the greatest potency over all tested cell lines, including MDR KB-VIN. All five most active kobusine derivatives (**67-5**, **67-7**, **67-10**, **67-18**, and **67-19**) are acylated at both C-11 and C-15. All tested derivatives with a hydroxy group at either C-11 or C-15 were inactive or much less active. All six most active pseudo-kobusine derivatives (**68-5**, **68-15**, **68-19**, **68-20**, **68-25**, and **68-32**) contain a free hydroxy group at C-6. The substituent at C-11 is either a benzoyl/cinnamoyl ester (**68-5**, **68-15**, **68-19**, **68-20**, and **68-25**) or a trityl ether (**68-32**). Finally, the moiety at C-15 is a hydroxy group (**68-15**, **68-20**, **68-25**, and **68-32**) or benzoyl ester (**68-5**, **68-19**).

Furthermore, previously our study, Antitumor properties and radiation-sensitizing effects of various types of novel derivatives prepared from C19- and C20-diterpenoid alkaloids were also investigated [19]. Two novel hetisine-type C20-diterpenoid derivatives (**68-7** and **68-20**) showed significant suppressive effects against the Raji non-Hodgkin's lymphoma cell line [20].

### **5. Conclusions**

We have synthesized acylated derivatives of various C19- and C20-diterpenoid alkaloids. All alkaloids and their derivatives were screened against four to five human tumor cell lines. Alkaloids **37-2**, **37-9**, **37-17**, **37-18**, **56-5**, **67-7**, **67-14**, **67-19**, **68-4**, **68-12**, **68-20**, **68-22**, **68-24**, and **68-32** showed comparable potency against KB and KB-VIN cancer cell lines, and some alkaloids showed tumorselective activity. Alkaloids **37-12**, **37-13**, **37-16**, and **37-19** exhibited greater inhibitory activity against drug-resistant KB-VIN cells (2.15~2.57-fold) than the parental KB cells. These results demonstrate that modified lycoctonine-type C19-diterpenoid alkaloids and hetisine-type C20-diterpenoid alkaloids are not substrates of P-gp and could be effective against P-gp overexpressing MDR tumors. These promising new lead alkaloids merit continued studies to evaluate their potential as antitumor agents, particularly with enhanced resistant tumor selectivity. In addition, our results from modification-based antitumor activity studies can be used for further development of anticancer drugs overcoming an MDR phenotype.

**Funding:** This study was supported in part by NIH grant CA177584 from the National Cancer Institute awarded to K.H.L. as well as the Eshelman Institute for Innovation, Chapel Hill, North Carolina, awarded to M.G.

**Acknowledgments:** The author gratefully acknowledges Lee, K.H., Goto, M., Morris-Natschke, S.L., Ohkoshi, E., Zhao, Yu., Li, K.P., Bastow, K.F., Natural Products Research Laboratories, UNC Eshelman School of Pharmacy, University of North Carolina at Chapel Hill; and Mizukami, M., Kaneda, K., Suzuki, Y., Shimizu, T., Kusanagi, N., Takeda, K., Haraguchi, M., Abe, Y., Kuwahara, N., Suzuki, S., Terui, A., Masaka, T., Munakata, N., Uchida, M., Nunokawa, M., Chiba, R., Kanazawa, R., Matsuoka, K., Suzuki, M., Ikuta, M., Asakawa, E., Tosho, Y., Nakata, A., Hasegawa, Y., Katoh, M., Kokubun, A., Uchimura, A., Mikami, S., Takeuchi, A., Department of Medicinal

Chemistry, Faculty of Pharmaceutical Sciences, Hokkaido University of Science, for their helpful advice and support throughout this work.

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

### **References**


© 2019 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/).

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Molecules* Editorial Office E-mail: molecules@mdpi.com www.mdpi.com/journal/molecules

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18