**The Chalcone Lonchocarpin Inhibits Wnt**/β**-Catenin Signaling and Suppresses Colorectal Cancer Proliferation**

**Danilo Predes 1,**†**, Luiz F. S. Oliveira 1,**†**, Laís S. S. Ferreira 1,**†**, Lorena A. Maia 1,**†**, João M. A. Delou 1, Anderson Faletti 1, Igor Oliveira 1, Nathalia G. Amado 1, Alice H. Reis 1, Carlos A. M. Fraga 1, Ricardo Kuster 2, Fabio A. Mendes 1, Helena L. Borges <sup>1</sup> and Jose G. Abreu 1,\***


Received: 9 October 2019; Accepted: 13 November 2019; Published: 7 December 2019

**Abstract:** The deregulation of the Wnt/β-catenin signaling pathway is a central event in colorectal cancer progression, thus a promising target for drug development. Many natural compounds, such as flavonoids, have been described as Wnt/β-catenin inhibitors and consequently modulate important biological processes like inflammation, redox balance, cancer promotion and progress, as well as cancer cell death. In this context, we identified the chalcone lonchocarpin isolated from *Lonchocarpus sericeus* as a Wnt/β-catenin pathway inhibitor, both in vitro and in vivo. Lonchocarpin impairs β-catenin nuclear localization and also inhibits the constitutively active form of TCF4, dnTCF4-VP16. *Xenopus laevis* embryology assays suggest that lonchocarpin acts at the transcriptional level. Additionally, we described lonchocarpin inhibitory effects on cell migration and cell proliferation on HCT116, SW480, and DLD-1 colorectal cancer cell lines, without any detectable effects on the non-tumoral intestinal cell line IEC-6. Moreover, lonchocarpin reduces tumor proliferation on the colorectal cancer AOM/DSS mice model. Taken together, our results support lonchocarpin as a novel Wnt/β-catenin inhibitor compound that impairs colorectal cancer cell growth in vitro and in vivo.

**Keywords:** anticancer drugs; flavonoids; natural compounds; *Xenopus laevis*; AOM/DSS model

#### **1. Introduction**

Colorectal cancer (CRC) is the third most commonly diagnosed cancer and the second most common cause of cancer death. According to World Health Organization (WHO), it is expected that there were 1.8 million cases and 862,000 deaths in 2018 [1]. Sporadic CRC initiation, promotion, and progression is mostly driven by a sequence of known genetic mutations in key signaling pathways, frequently related to DNA damage response and sustained proliferation in the absence of growth factors. In CRC, 93% of the cases have at least one mutation of one Wnt/β-catenin pathway component [2]. The most frequently mutated gene in colorectal cancer is *APC* (adenomatous polyposis coli) that is a β-catenin destruction complex component. *APC* mutation occurs in 81% of non-hypermutated colorectal cancers cases and in 51% of hypermutated colorectal cancer cases, triggering tumorigenesis in intestinal polyps of patients with familial adenomatous polyposis [3]. The Wnt/β-catenin signaling

pathway coordinates several cell behavior aspects, such as cell proliferation, differentiation, stemness, polarity, and migration [4,5]. In the absence of Wnt ligands, the destruction complex is active in the cytoplasm, phosphorylating β-catenin, a key component of the canonical Wnt pathway, leading to its degradation by the proteasome [6,7]. Wnt interaction with its receptors Frizzled (Fzd) and LDL receptor-related protein 5/6 (LRP5/6) disrupts the destruction complex assembly leading to β-catenin stabilization, cytoplasmic accumulation, translocation to the nucleus and binding to the T-cell factor/lymphoid enhancer factor (TCF/LEF), allowing Wnt target gene transcription [8]. Despite the crucial role of Wnt signaling on colorectal tumorigenesis, there is no Wnt/β-catenin inhibitor approved for clinical use [9]. Due to the importance of Wnt/β-catenin and its frequent mutations upstream to β-catenin translocation to the nucleus, it is crucial to find anticancer drugs that target the pathway downstream to this phenomenon [2].

Addressing normal and pathological Wnt/β-catenin signaling functioning requires multidisciplinary experiments combining in vitro and in vivo approaches. Among different models for studying Wnt/β-catenin signaling in vivo, *Xenopus laevis* stands out for its liability and efficiency. Wnt/β-catenin signaling plays a key role in two fundamental steps during the Xenopus early development that can be exploited for the screening of new drug candidates: the dorso-ventral and the antero-posterior axis patterning [10–12]. Indeed, the Xenopus model system has been explored to discover Pyrvinium, an FDA approved compound, as a Wnt signaling inhibitor that acts downstream of β-catenin. Pyrvinium impaired Xenopus embryo secondary axis induction in a dose-dependent manner and decreased colon cancer cells viability [13].

In addition, the AOM/DSS mouse model stands as a relevant preclinical inflammation-associated CRC model with histologic and phenotypic features that recapitulates the aberrant crypt foci-adenoma-carcinoma found in the human CRC [14]. Consistent with CRC development, in the AOM/DSS murine model, β-catenin nuclear translocation is observed in both flat and polypoid lesions likely due to β-catenin mutation [15]. In this context, the study of synthetic and natural compounds able to inhibit the Wnt/β-catenin signaling pathway have been explored as possible antitumor prototypes. Among the small natural molecules studied, the flavonoids, polyphenolic compounds found in many plants with a wide range of biological effects, stand out. Many flavonoids have been described as inhibitors of Wnt signaling and potential antitumor compounds, such as apigenin, EGCG, silibin, kaempferol, isorhamnetin, quercetin, isoquercitrin, derricin, and derricidin [16–25]. However, the specific mechanism by which some of these compounds affect Wnt/β-catenin signaling as well as its capacity to impair CRC growth is still not elucidated. Along the flavonoid biosynthesis pathway, the chalcones are well known as precursors of the flavonoids. Lonchocarpin is a chalcone first isolated from *Lonchocarpus sericeus* (as known as *Derris sericeu*) by Baudrenghien and colleagues in 1949 [26], and its correct structure was elucidated by the same researchers in 1953. The cytotoxic effects of the chalcone lonchocarpin have been previously described in neuroblastoma and leukemia cell lines [27], however, its role in CRC and the Wnt/β-catenin pathway is unknown.

In the present work, we describe lonchocarpin ability of inhibiting Wnt/β-catenin both in vitro, in colon cancer cell lines, and in vivo, with Wnt specific *Xenopus laevis* embryonic assays. In addition, acute administration of lonchocarpin in a preclinical CRC mouse model reduced cell proliferation in adenocarcinomas. Altogether, our data show lonchocarpin as a potent Wnt/β-catenin inhibitor that impairs cancer cell proliferation both in vitro and in vivo, and a promising compound for further antitumor clinical investigation and development.

#### **2. Results**

#### *2.1. Lonchocarpin Inhibits Wnt*/β*-Catenin Pathway and Reduces Nuclear* β*-Catenin Levels*

It has been shown that natural compounds, including chalcones, have growth-inhibitory properties in cancer cell lines by modulating Wnt/β-catenin signaling [17,18]. We employed an RKO pBAR/Renilla based screening of natural compounds and found lonchocarpin as a Wnt signaling modulator hit (data not shown). In this context, we evaluated whether lonchocarpin, a chalcone isolated from *Lonchocarpus sericeus*, is able to inhibit the Wnt signaling pathway in human colorectal cancer cell lines RKO pBAR/Renilla stimulated with Wnt3a CM (conditioned medium) for 24 h and non-stimulated SW480 pBAR/Renilla (Figure 1A). SW480 does not need Wnt stimulation, since it harbors an APC mutation that overactivates the canonical Wnt signaling. Cells were treated with lonchocarpin overnight, with or without the conditioned medium. Lonchocarpin decreased luciferase reporter activity in a concentration-dependent manner, starting at 3 μM in RKO pBAR/Renilla and 5 μM in SW480 pBAR/Renilla (Figure 1B,C), and presented the half maximal inhibitory concentration (IC50) of 4 μM in SW480 pBAR/Renilla (Figure 1D).

In order to validate the Wnt/β-catenin reporter gene assay inhibition, we performed immunocytochemistry on SW480 cells. In order to activate the pathway, we treated RKO cells with Wnt3a CM for 24 h and performed immunocytochemistry to assess nuclear β-catenin cell count. These cells were cotreated with DMSO (vehicle), 10 or 20 μM lonchocarpin. L-cell CM treated cells displayed 34% β-catenin positive nuclei (Figure 1E–E"), while Wnt3a CM treatment increased the mean to 86% (Figure 1F–F"). RKO cells cotreated with 10 μM (Figure 1G–G") or 20 μM Figure 1H–H") lonchocarpin showed 54% and 30% nuclear β-catenin positive cells, respectively (Figure 1J). This data show that lonchocarpin treatment decreased β-catenin nuclear translocation in a concentration-dependent manner (Figure 1E–H"). To assess whether lonchocarpin affects totalβ-catenin protein level, we performed immunoblotting assay of RKO cells stimulated overnight with Wnt3a CM or 1 μM BIO, cotreated with DMSO or 20 μM lonchocarpin (Figure 1K). Immunoblot assay showed that lonchocarpin did not expressively reduced β-catenin total levels (Figure 1K), suggesting that β-catenin degradation, or β-catenin stabilization impairment might not be a lonchocarpin mechanism of action. In order to further validate our immunocytochemistry analysis, we performed immunoblotting of RKO cytosolic and nuclear fractions (Figure 1L). These cells were treated accordingly to previous assay. Immunoblotting showed that lonchocarpin reduces β-catenin nuclear level, suggesting an inhibition of nuclear translocation, or, possibly, an inhibition of β-catenin nuclear interaction with other proteins or DNA. Thus, this data suggests that lonchocarpin inhibits Wnt/β-catenin signaling pathway and impairs β-catenin nuclear localization.

**Figure 1.** Lonchocarpin inhibits the Wnt/β-catenin pathway. Treatment with lonchocarpin inhibits Wnt reporter activity. (**A**) Lonchocarpin chemical structure, (**B**) RKO-pBAR/Renilla, and (**C**) SW480-pBAR/Renila gene reporter luciferase assay. (**D**) Lonchocarpin half maximal inhibitory concentration is 4 μM in the SW480-pBAR/Renilla cell lineage. Graph bars represent mean and SD. (**E**–**I**) SW480 β-catenin and DAPI immunocytochemistry staining showed that lonchocarpin decreases β-catenin translocation after lonchocarpin treatment. RKO cells were treated with (**E**–**E"**) L-cell control conditioned medium or Wnt3a conditioned medium with (**F**–**F"**) DMSO, (**G**–**G"**) 10 μM or (**H**–**H"**) 20 μM lonchocarpin. (**I**–**I"**) Immunocytochemistry negative control. (**J**) Quantification of nuclear β-catenin positive cell count ratio. Graph bars represent mean and SEM. (**K**) Immunoblotting depicting β-catenin and β-actin total levels of RKO cells treated with Wnt3a CM (conditioned medium) for 7 h. Densitometry is shown as the ratio of β-catenin/β-actin. (**K**) Immunoblot of cell lysate of RKO cells treated with 1-L-cell CM, 2-Wnt3a CM + DMSO, 3-Wnt3a CM + lonchocarpin 20 μM, 4-BIO 1 μM + DMSO, 5-BIO 1 μM + lonchocarpin 20 μM. (**L**) Immunoblot of cytosolic and nuclear fractions of RKO cells treated with 1,6-L-cell CM, 2,7-Wnt3a CM + DMSO, 3,8-Wnt3a CM + lonchocarpin 20 μM, 4,9-BIO 1 μM + DMSO, 5,10-BIO 1 μM + lonchocarpin 20 μM. Cytosolic densitometry was calculated considering α-Tubulin as the loading control, while nuclear densitometry considered Lamin A/C as the loading control. \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001. Scale bar represents 20 μm.

#### *2.2. Lonchocarpin Inhibits the Canonical Wnt Pathway Downstream of the Destruction Complex*

SW480 harbors a mutation in APC that deletes its carboxyl-terminus domain, preventing the destruction complex assembly [28,29]. Considering that lonchocarpin inhibited the Wnt reporter gene in SW480 pBAR/Renilla (Figure 1C,D), and also prevented β-catenin nuclear localization (Figure 1L), we speculated that the flavonoid could act downstream of the destruction complex. To test this hypothesis, we performed epistasis assay using Wnt/β-catenin-specific reporter (Super TOPFLASH) transfected cells induced by Wnt3a CM (Figure 2A), or co-transfected with wild type β-catenin (Figure 2B), constitutively active β-catenin S33A (Figure 2C), or dnTCF4 VP16 (Figure 2D), a constitutive active TCF4 form that does not rely on β-catenin for inducing the pathway.

**Figure 2.** Lonchocarpin inhibitsWnt/β-catenin pathway downstream to TCF4. Lonchocarpin suppresses Wnt/β-catenin pathway induced by Wnt3a CM treatment (**A**) or by transfection with (**B**) β-catenin, (**C**) β-catenin S33A, or (**D**) dnTCF4VP16 in HEK293T cells. (**E**) Proposed lonchocarpin mechanism of action. Graph bars represent mean and SD. \*\*\* *p* < 0.001.

Our results showed that lonchocarpin inhibits Wnt/β-catenin signaling activation when pathway activation is triggered by wild-type β-catenin, β-catenin S33A, or dnTCF4 VP16 overexpression (Figure 2A–D). Interestingly, lonchocarpin showed a more potent and efficient inhibitory effect than the previously published compound PNU-74654 (Figure 2A–D) [30]. In this scenario, we propose that lonchocarpin acts downstream of the destruction complex and impairs TCF4 mediated transcription.

Together, these data suggest that lonchocarpin inhibits Wnt/β-catenin signaling by impairing β-catenin nuclear localization while also hindering TCF activity.

#### *2.3. Lonchocarpin Treatment Disturbs Xenopus laevis Embryos Axial Patterning and Rescues Wnt Overactivation Phenotypes*

Our in vitro data suggest that lonchocarpin inhibits Wnt/β-catenin by impairing β-catenin nuclear localization (Figure 1E–L) while also suppressing TCF mediated transcription (Figure 2D). Next, we investigated if this inhibitory effect is also observed in vivo. Manipulation of *Xenopus laevis* embryonic development has been successfully used to validate compounds targeting Wnt/β-catenin signaling through the interpretation of embryonic phenotypes and Wnt signaling overactivation [11]. In this context, injection of 1.6 pmol lonchocarpin in the embryo animal dorsal blastomeres (Figure 3A) induced head defects, characterized by reduction of anterior structures in 23% of injected embryos, such as the cement gland and diminished eyes (Figure 3D) while uninjected or DMSO-injected embryos

developed normally (Figure 3B,C). Considering that Wnt/β-catenin is active in the dorsal side of the Xenopus embryo, which will organize the anterior-posterior axis, this result suggests an inhibition of the pathway by lonchocarpin. In order to confirm this result, we injected 10 pmol of lonchocarpin into the blastocoele space of stage 9 embryos (Figure 3E). Lonchocarpin injection induced anterior structures enlargement, such as the head and cement gland in 48% of the embryos (Figure 3H), while uninjected and DMSO injected embryos developed normally (Figure 3F,G). This lonchocarpin effect is consistent with Wnt/β-catenin inhibition in the embryo anterior region, since the signaling at this embryonic stage induces posterior structures.

**Figure 3.** Lonchocarpin alters axial patterning in *Xenopus laevis* embryos and inhibits Wnt-8 induced axis and S01234-luciferase reporter. The 4-cell stage injected embryos display a smaller head compared uninjected and DMSO-injected embryos (**A**–**D**). Stage 9 blastulae injected embryos display a larger head (arrow) and cement gland (arrow) compared to uninjected or DMSO-injected embryos (**E**–**H**). Injection of xWnt8 mRNA into ventral blastomere induced ectopic axis (arrowhead) compared to uninjected embryos (**I**–**K**,**P**). Lonchocarpin inhibited Wnt8-induced secondary axis (**M**,**P**), but not DMSO (**L**,**P**). Lonchocarpin injection inhibits Wnt8-induced S01234-luciferase specific reporter activation in *Xenopus laevis* embryo (**N**,**O**) \*\*\* *p* < 0.001. Graph bars represent mean and SD (see also Figure S1).

It is well established that Wnt/β-catenin signaling overactivation in the Xenopus embryo 4-cell stage ventral side induces ectopic axis containing head and dorsal structures [10,12,31]. We injected 10 pg of xWnt8 mRNA combined or not with 10 pmol lonchocarpin or DMSO into two ventral blastomeres at 4-cell stage embryo (Figure 3I). xWnt8 or xWnt8 + DMSO-injected embryos developed a secondary axis in 90% of the embryos, while only 60% of the lonchocarpin-injected embryos developed a secondary axis (Figure 3K–M,P).

Thus, these results strongly suggest lonchocarpin induces phenotypes consistent with Wnt/β-catenin inhibition in Xenopus embryos (see also Figure S1 for quantification). To confirm whether lonchocarpin inhibits Wnt signaling in Xenopus, we coinjected a Wnt/β-catenin specific gene reporter S01234-luciferase with xWnt8 mRNA and 2.4 pmol lonchocarpin or DMSO into 4-cell *stage Xenopus laevis* embryos (Figure 3N). Lonchocarpin suppressed 82% of Wnt/β-catenin signaling gene reporter activation (Figure 3O).

These data support that lonchocarpin inhibits Wnt/β-catenin signaling pathway activation both in vivo and in vitro.

#### *2.4. Lonchocarpin Reduces HCT116, SW480, and DLD-1 Cell Proliferation*

Since lonchocarpin inhibited Wnt/β-catenin downstream to the destruction complex both in vitro and in vivo, we asked whether lonchocarpin has antitumor effects on colorectal cancer cell lines where Wnt/β-catenin signaling has a critical role on tumorigenesis. Canonical Wnt signaling activation leads to proliferation in many cell types, including colorectal cancer cells, thus we asked whether lonchocarpin inhibits colon cancer cell proliferation. We treated HCT116, SW480, DLD-1, and IEC-6 cell lines with 5, 10, or 20 μM lonchocarpin for 24 h and performed the Click-it EdU proliferation assay.

Lonchocarpin inhibited 33% of HCT116 EdU positive cell count at 5 and 10 μM and reduced 75% of EdU positive cell count at 30 μM (Figure 4B–E). Lonchocarpin inhibited 50% and 85% of SW480 EdU positive cell count at 10 and 20 μM, respectively (Figure 4H–J). Lonchocarpin inhibited 40% and 75% of DLD-1 proliferative cells ratio at 10 and 20 μM, respectively (Figure 4M–O). However, lonchocarpin did not affect the non-tumoral intestinal cell line IEC-6 proliferating cells ratio (Figure 4Q–T). These data show that lonchocarpin suppresses colorectal cancer cell lines proliferation while not affecting the non-tumoral cell line IEC-6 proliferation. Next, we asked whether lonchocarpin suppresses proliferation through cell toxicity. In order to assess cellular viability, we performed MTT assay after treatment with 10, 20, 30, 40, and 50 μM lonchocarpin for 24, 48, or 72 h (Figure 4). We noticed that 20 to 50 μM lonchocarpin at all treatment intervals reduced relative 570 nm absorbance of HCT116, SW480, and DLD-1 colorectal cancer cell lines but 10 μM had no effect (Figure 4U–W). Curiously, 20 μM lonchocarpin did not decrease relative 570 nm absorbance, but maintained the same reading throughout the experiment, suggesting a proliferation inhibition.

However, in the non-tumoral intestinal cell line IEC-6, lonchocarpin decreased relative 570 nm absorbance only at 50 μM. Exclusively at 72 h of treatment there was noticeable absorbance reduction in 40 μM lonchocarpin (Figure 4X). These data show that lonchocarpin inhibits cell proliferation and viability preferentially in the colorectal cancer cell lines.

**Figure 4.** Lonchocarpin inhibits HCT116, SW480, and DLD-1 colorectal cancer cell lines proliferation. Proliferation assay shows that lonchocarpin suppresses proliferation of HCT116, SW480, and DLD-1 colorectal cancer cell lines, while not affecting IEC-6 non-tumor small intestine cell line proliferation. DAPI and EdU stained cells micrographs acquired 24 h post 5, 10, and 20 μM lonchocarpin treatment. (**A**–**E**) HCT116, (**F**–**J**) SW480, (**K**–**O**) DLD-1, (**P**–**T**) IEC-6. Graphs show the percentage of EdU positive cells. Scale bar represents 10 μm. Graph bars represent mean and SD. MTT assay shows that lower lonchocarpin concentrations are less cytotoxic to the non-tumor cell line IEC-6 compared to the tumor cell lines. MTT assay performed after treatment with 10, 20, 30, 40, and 50 μM lonchocarpin of (**U**) HCT116, (**V**) SW480, (**W**) DLD-1, and (**X**) IEC-6 cells during 24, 48, and 72 h. R.U. (Relative Units). \* *p* < 0.1, \*\* *p* < 0.01. Graphs show mean and SEM.

#### *2.5. Lonchocarpin Reduces Cell Migration in HCT116, SW480, and DLD-1 Colorectal Cancer Cell Lines*

Canonical Wnt signaling key protein β-catenin interacts with adhesion proteins in the membrane that may affect cell adhesion and migration. We evaluated whether lonchocarpin affects HCT116, SW480 and DLD-1 colorectal cancer cell lines and IEC-6 non-tumoral intestinal cell line migration by performing scratch assay. HCT116 lonchocarpin treatment reduced 55% of scratch closure at 20 μM (Figure 5A–E). Lonchocarpin treatment of SW480 reduced 40% and 55% of scratch closure at 10 and 20 μM, respectively (Figure 5F–J). Likewise, DLD-1 lonchocarpin treatment reduced 45% of scratch closure at 20 μM (Figure 5K–O). However, lonchocarpin treatment did not affect cell migration of the non-tumoral cell line IEC-6 (Figure 5P–T). These data show that lonchocarpin treatment impairs cell migration of colorectal cell lines, while not affecting the migration of the non-tumoral cell line.

**Figure 5.** Lonchocarpin inhibits cell migration in HCT116, SW480, and DLD-1 colorectal cancer cell lines. Scratch assay shows that lonchocarpin impairs migration of HCT116, SW480, and DLD-1 colorectal cancer cell lines while not interfering with the IEC-6 non-tumor intestine cell line migration. Images show cell migration through the scratch healing 24 h post treatment with 5, 10, and 20 μM lonchocarpin (**A**–**E**) HCT116, (**F**–**J**) SW480, (**K**–**O**) DLD-1, (**P**–**T**) IEC-6. Graph shows relative wound area closure relative to time 0 h. All conditions were normalized to DMSO. \* *p* < 0.1, \*\*\* *p* < 0.001. Scale bar represents 100 μm. Graph bars represent mean and SD.

#### *2.6. Lonchocarpin Decreases Cell Proliferation in Azoxymethane (AOM)*/*Dextran Sulfate Sodium (DSS) Induced Adenocarcinomas*

Lonchocarpin inhibits Wnt/β-catenin in vitro and in vivo, while also presenting antitumor effects in vitro. We asked whether lonchocarpin also has antitumor effects in vivo. We assessed the efficacy of lonchocarpin therapeutic administration in an azoxymethane (AOM)/dextran sulfate sodium (DSS)-induced model of colon cancer. After three cycles of DSS, when colon tumors were expected in most animals, lonchocarpin (50 or 100 mg·kg−1·day−1) was injected intraperitoneally for four days, and the mice were assessed 3 h after the last injection (Figure 6A). Colon tumors were macroscopically observed in almost all mice submitted to the protocol (77%), and histopathological analyses revealed several changes in the intestinal mucosa (Figure S2). The most frequent alterations included no presence of mononuclear and polymorphonuclear leukocyte infiltrates in the lamina propria and

submucosa, hyperplastic epithelium, in addition several adenomas and adenocarcinomas (Figure 6B). Lonchocarpin did not show any toxicity to the treated animals, but significantly reduced tumor proliferation (Figure 6C–F"). Lonchocarpin at 100 mg·kg−1·day−<sup>1</sup> significantly decreased 31% and 38% of proliferative Ki-67 and BrdU positive cells in adenocarcinomas of the treated mice compared either with vehicle groups, respectively (Figure 6G,H, Tables S1 and S2). However, lonchocarpin showed a more efficient antiproliferative effect when administered at the highest dosage (100 mg·kg−1·day−1) in comparison to the lower dosage (50 mg·kg−1·day−1), in which no statistical relevance was found (Figure 6G,H, Tables S1 and S2).

**Figure 6.** Lonchocarpin decreases cell proliferation in adenocarcinomas. (**A**) Timeline of inflammation-driven colon cancer tumorigenesis model (azoxymethane (AOM)/dextran sulfate sodium (DSS)) protocol in adult mice. Groups were divided as follows, vehicle 50 or 100: i.p. vehicle corresponding to volume for the respective dose in the treated group; treated groups: i.p. injections of 2.5 mg/mL lonchocarpin 50 or 100 mg·kg–1·day–1 for four days. Vehicle: PEG400 30% in sterile NaCl 0.9%. BrdU i.p. injection (100 <sup>μ</sup>g/kg) 1 h before euthanasia. (**B**) Representative H&E swiss-roll image of a colon section from a mouse subjected to AOM/DSS protocol. The zoom picture shows the corresponding area of an adenocarcinoma. Scale bar represents 200 μm. (**C**) Representative immunofluorescence photomicrographs of adenocarcinoma areas from colon sections stained for proliferation markers (**C**'–**F**') Ki-67 (red, Cy3) and (**C"**–**F**") BrdU (green, Alexa 488). (**C**–**F**) Nuclei stained with DAPI (blue). (**D**,**E**) Quantification of the percentage of (**G**) Ki-67 (**H**) or BrdU positive cells per indicated group. Scale bar represents 100 μm. Graphs represent mean and SEM. \* *p* < 0.05 \*\* *p* < 0.01 Student *t*-test of lonchocarpin treatment condition in comparison to vehicle.

Taken together, our data demonstrate that lonchocarpin suppress the colorectal cancer cell growth in vitro and in vivo.

#### **3. Discussion**

The Wnt/β-catenin pathway plays a key role in colorectal tumorigenesis Integrated analysis of 195 colorectal tumors revealed that Wnt signaling pathway components were mutated in 94% of all tumors, and these mutations occurred mostly downstream to APC10. Hence, describing the novel canonical Wnt signaling pathway inhibiting small molecules that act downstream to APC is a recurrent strategy to improve colorectal cancer treatability.

Several studies show the antitumor effect of natural compounds that act as inhibitors of multiple components of the Wnt/β-catenin pathway [17]. Quercetin has been described to disrupt TCF/β-catenin interaction [23,32]. Epigallocatechin-3-gallate (EGCG) has been shown as a Wnt/β-catenin inhibitor by promoting β-catenin degradation [33]. Isoquercitrin has been described as an inhibitor of the Wnt signaling both in vitro and in vivo, and impairing tumor growth in vitro [34]. The chalcones derricin and derricidin have also been reported as canonical Wnt signaling inhibitors impairing tumor growth in vitro [18].

The present work is the first to identify lonchocarpin as a negative modulator of the Wnt/β-catenin pathway in RKO and SW480 colon tumor cell lines and in the HEK293T embryonic kidney cell line. We further elucidate that lonchocarpin acts downstream to β-catenin stabilization, probably at the TCF level, since it inhibits the overactivation of the Wnt/β-catenin pathway induced by the transfection of wild-type β-catenin, constitutively active β-catenin S33A or the constitutively active dnTCF4 VP16 in the HEK293T epistasis assay, while also inhibiting the Wnt/β-catenin pathway reporter in SW480 pBAR/Renilla cells, which harbors an APC truncation. We also show that lonchocarpin inhibits proliferation, migration, and cell viability in most of the three colorectal cancer cell lines, HCT116, SW480, and DLD-1, while not altering any of these aspects of the non-tumoral intestinal rat cell line IEC-6.

Recent work has identified that 24 h treatment with 50 μM lonchocarpin of SK-N-SH neuroblastoma line induces AMPK phosphorylation, which results in increased glucose uptake and inhibits protein synthesis [27]. Although our data shows effects at lower concentrations, this work corroborates our findings that lonchocarpin decreases cell viability. Thus, the antitumor effects that we described may be also a consequence of modulation of other signaling pathways besides Wnt/β-catenin.

Previous work has also measured cell viability following lonchocarpin treatment through MTT assay, indicating that the IC50 for cell growth in the CEM leukemia cell line is 10.4 μg/mL, the same as 3.4 μM of lonchocarpin [35]. Although not explored by the authors, leukemia cell lines are known to have Wnt signaling activating mutations, and lonchocarpin growth inhibiting effect could be due to Wnt signaling inhibition [36]. In this same study, authors show that derricin also inhibits leukemia cell growth [35], a chalcone also described as a canonical Wnt signaling inhibitor [18]. Comparatively, our cell viability data in HCT116, SW480, and DLD-1 colorectal cancer cell lines show that lonchocarpin reduces cell viability starting at 20 μM. This disparity may be due to the different origin of the cell lines. Lonchocarpin has also been shown as inhibiting lung cancer cells H292 growth in vitro and murine sarcoma S180 graft growth in vivo by inducing Caspase-3 mediated cell death [37]. Curiously, increased cleaved Caspase-3 levels were found at 48 h [37], but not at 24 h, suggesting that canonical Wnt signaling is inhibited prior apoptosis induction. Thus, previous lonchocarpin biological effect descriptions confirm that the antitumor effect of this chalcone is not exclusive to colorectal cancer.

Additionally, we show that lonchocarpin inhibits Wnt/β-catenin signaling in vivo in the *Xenopus laevis* embryo model. Xenopus embryo is a robust and reliable system to approach Wnt/β-catenin signaling in vivo since it is critical for axis patterning [10]. In that sense, Wnt signaling plays two major roles in early Xenopus development: a prior role that is composed by maternal Wnt components, and a latter role that is regulated by genes transcribed by the embryo itself. We show that injection of lonchocarpin at 4-cell stage, a stage where the embryo has no functioning transcription machinery, induced microcephaly and impaired Wnt8-induced axis, although a not very strong effect, most likely due to low lonchocarpin concentration into the embryo blastomere (Figure 3). Consistently, lonchocarpin injection into the blastocoele as well as co-injection with Wnt8-specific reporter (S01234), both addressing a moment where the embryo transcribes genes by itself, induced head enlargement and suppressed 82% of Wnt/β-catenin signaling gene reporter activation (Figure 3). These results are consistent with previous epistasis assay showing that lonchocarpin inhibits Wnt pathway downstream to β-catenin stabilization level, by impairing TCF mediated transcription (Figure 2).

Considering the relevance of the lonchocarpin Wnt/β-catenin inhibition and its functional effects on CRC cell lines as well as in Xenopus embryo, we tested lonchocarpin therapeutic administration in an inflammation-associated CRC mouse model. The AOM/DSS model has been widely used for CRC studies since it is highly reproducible, and recapitulates human cancer histological features and the major driven mutations. Indeed, it has become a powerful platform for chemopreventive and anticancer drug discovery studies. AOM is a procarcinogen metabolized into alkylating agent methylazocymethanol (MAM) in the liver. After excretion into the bile, it induces mutagenesis of the colonic epithelium. Colonic tumors are accelerated by a heparin-like polysaccharide DSS, which causes colonic epithelial damage, mirroring some of the features of inflammatory bowel disease [38]. Acute lonchocarpin i.p. administration in mice containing fully developed carcinomas reduced Ki-67 positive and BrdU positive cell count (Figure 6). Therefore, indicating lonchocarpin as an acute anti-proliferative agent in AOM/DSS induced adenocarcinoma. Although very promising, only the lonchocarpin highest dose produced anti-proliferative effects suggesting that its biological activity should be further enhanced through the synthesis of novel optimized analogues, or by using alternative administration approaches. In the middle of the 1950s, it was first isolated from the stem wood of *Camptotheca acuminata* the precursor of one of the most used chemotherapeutic agents for colorectal cancer treatment, the alkaloid camptothecin [39]. First described as showing antileukemic and antitumor activities, for the following three decades many camptothecin analogues have been described in order to enhance its antitumor properties. One of these analogues was CPT-11, currently known as Irinotecan, that is widely used to treat colorectal cancer [40]. Together with Irinotecan, 5-FU is also widely used clinically. Intriguingly, Wnt signaling inhibition has been shown to decrease resistance of colorectal cells to these chemotherapy drugs' treatment [41]. Interestingly, MEK signaling inhibitors have been shown to increase canonical Wnt signaling in CRC [42], and the co-treatment of MEK inhibitors and Wnt signaling inhibitors resulted in a reduction of tumor growth [42]. Thus, MEK inhibitors should also be addressed as a cotreatment with lonchocarpin.

There are currently 55 clinical trials aiming to inhibit Wnt signaling pathway in cancers, in which 21 are CRC studies (clinicaltrials.gov). Among the 21 CRC clinical trials, only two tested compounds inhibit canonical Wnt signaling at the transcriptional level, PRI-724 and resveratrol. In this context, in comparison with these compounds, the in vitro IC50 of lonchocarpin Wnt signaling inhibition is noticeable. PRI-724 inhibited the Wnt signaling pathway at 25 μM in vitro [43] (authors used cyclin D1 Western blot to check Wnt signaling modulation), resveratrol inhibited at 20 μM [44] (authors used a Wnt signaling specific gene reporter assay), whereas 4 uM lonchocarpin reached the IC50. The PRI-724 clinical trial has been withdrawn due to supply issues (NCT02413853) and resveratrol results have not been published yet (NCT00256334). We believe that lonchocarpin anticancer effects should be further addressed in preclinical studies, so it can be a possible clinical trial candidate.

The use of natural compounds as drug candidates has been improved by the use of alternative delivery approaches such as controlled delivery systems. Indeed, nanostructuration has been used to circumvent solubility issues [45–47]. These strategies can solve instability and poor water solubility issues [48,49] and improve a drug candidate pharmacokinetic profile. These systems could be used to further enhance lonchocarpin anticancer properties in vivo.

At last, the similarity of lonchocarpin and derricin chemical structure deserves to be noticed. These two natural compounds inhibit Wnt signaling through similar mechanisms, while derricidin has

a different one [18]. This comparison paves a new way for structure-function studies, and the quest for new Wnt signaling inhibitor pharmacophores.

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

#### *4.1. Cell Lines and Chemical Compounds*

HEK293T, SW480, HCT116, DLD-1, and IEC-6 cell lines were purchased from ATCC and RKO-pBAR/Renilla and SW480-pBAR/Renilla were a gift from Professor Xi He (Harvard Medical School). All cell lines were maintained in DMEM-F-12 (Gibco, Life Technologies Limited, Paisley, UK) enriched with 10% fetal bovine serum (Gibco). The chalcone lonchocarpin was purified by Nascimento and Mors, 1972 [50] and kindly donated for this study by professor Ricardo Kuster (Federal University of Espirito Santo). PNU-74654 was synthesized at the Laboratory of Evaluation and Synthesis of Bioactives substances (Biomedical Sciences Institute, UFRJ). Both compounds were diluted in DMSO (Sigma-Aldrich, Saint Louis, MO, USA) at the concentration of 10 mM. PNU-74654 has been previously described as an inhibitor of Wnt/β-catenin pathway by blocking the interaction between β-catenin and TCF [30]. L-cell conditioned medium (CM) and Wnt3a CM were obtained according to the ATCC protocol. L-cells and L-Wnt3a cells were plated into 75 cm<sup>2</sup> flasks at 50% confluence with 10 mL DMEM medium containing 10% FBS. After 4 days, the first batch of medium was obtained and kept at 4 ◦C. Three days later, the last batch of medium was obtained and combined with the first one. Finally, L-cell CM and Wnt3a CM were passed through a 0.22 μm filter, and kept at 20 ◦C.

#### *4.2. Wnt*/β*-Catenin Luciferase Reporter Assay*

First, 1.2 <sup>×</sup> 104 cells/well RKO-pBAR/Renilla and SW480-pBAR/Renilla cells were cultured on 96-well plates in DMEM/F-12 containing 10% fetal bovine serum (Gibco). Then, 24 h later, cells were treated with lonchocarpin at the concentrations of 1, 3, 5, 10, 20, and 30 μM in the presence of Wnt3a conditioned medium for an additional 24 h. L-cell CM was used as negative control, and 0.3% DMSO was used as vehicle control. After 24 h of treatment, Firefly and Renilla luciferase activities were detected according to the manufacturer's protocol (Dual Luciferase Reporter Assay System, Promega, Madison, WI, USA).

The 1.2 <sup>×</sup> 104 cells/well HEK293T cells were cultured on 96-well plates in DMEM/F-12 containing 10% fetal bovine serum (Gibco). At 70% confluence, each well was transfected with 100 ng TOP-Flash plasmid, 10 ng Renilla-luciferase plasmid, and 100 ng wild type β-catenin or 100 ng β-catenin S33A using Lipofectamine 3000 (Invitrogen, Life Technologies Corporation, Carlsbad, CA, USA). Then, 18 h after transfection, cells were treated with 1, 5, and 10 μM of lonchocarpin. After 24 h, Firefly and Renilla luciferase activities were detected according to the manufacturer's protocol (Dual Luciferase Reporter Assay System, Promega).

#### *4.3. Immunocytochemistry*

SW480 cells were cultured in 24-well plates with 4.0 <sup>×</sup> 104 cells/well in DMEM-F12 media (Gibco) containing 10% fetal bovine serum (Gibco). Cells were fixed in 4% paraformaldehyde, washed with phosphate buffered saline, and permeabilized with 0.3% Triton X-100. Then, each well was blocked for 1 h with 5% bovine serum albumin. The rabbit anti-β-catenin primary antibody (Sigma Aldrich) was diluted in PBS containing 1% bovine serum albumin (1:200) and incubated overnight. The secondary antibody anti-rabbit Alexa Fluor 546 (Sigma Aldrich) was diluted (1:500) with 1% bovine serum albumin (1:500) and incubated for 2 h at room temperature. 4,6-diamidino-2-phenylindole staining (Cell Signaling) was performed for 15 min, and then slides were mounted with Fluoromount (Sigma Aldrich). Images were captured using the confocal microscope Leica TCS SP5.

#### *4.4. Immunoblotting and Cell Fractionation*

First, 2 <sup>×</sup> 105 cells/well RKO-pBAR/Renilla cells were cultured on 12-well plates and treated with 20 μM lonchocarpin for 24 h. Then, the cells were harvested in ice-cold PBS 1X, followed by cytosolic lysis in digitonin buffer with protease inhibitors (150 mM NaCl, 50 mM Tris, 25 μg/mL digitonin, pH 7.4) for 5 min. Then, the lysate was centrifugated for 10 min at 2000 G at 4 ◦C and the supernatant was collected and considered as the cytosolic fraction. Following digitonin extraction, the remaining cell pellets were washed in ice-cold PBS 1X and lysed in NP-40 buffer (150 mM NaCl, 50 mM Tris, 1%NP-40, pH 7.4) with protease inhibitors for 15 min and centrifuged for 10 min at 7000 G at 4 ◦C. The supernatant was collected and considered as the membrane fraction. After that, the cell pellets were washed again in ice-cold PBS 1X and lysed in RIPA buffer with protease inhibitors for 20 min and centrifugated after at 16,000 G at 4 ◦C. The supernatant was collected and considered as the nuclear fraction. Whole cell lysates were prepared using Triton buffer (150 mM NaCl, 50 mM Tris, 1% Triton X-100, 1 mM EDTA, 10% Glycerol, pH 7.5). Finally, cell lysates were denatured with Laemmli buffer at 95 for 5 min, and the protein samples were subjected to SDS-PAGE and transferred into PVDF membranes (Millipore, Merck KGaA, Darmstadt, Germany). The membrane was then blocked with 2% PVP (Sigma Aldrich) in TBS-Tween-20, incubated with primary monoclonal antibodies β-catenin (BD, 1:500), β-actin (SCBT, 1:2000); lamin A/C (CST, 1:500) and α-tubulin (Sigma, 1:2000) overnight at 4 ◦C. After three washes with TBS-T, the membranes were incubated for 1 h with HRP-conjugated secondary antibodies (CST). The immunoblots were visualized by chemiluminescence using SuperSignal West Pico and West Femto (ThermoFisher, Life Technologies Corporation, Carlsbad, CA, USA).

#### *4.5. Xenopus laevis Embryo Manipulations*

Frog experiments were carried out according to the guidelines granted by the Animal Care and Use Ethic Committee from the Federal University of Rio de Janeiro and were approved by this committee under the permission number 152/13. Female adult frogs (Nasco Inc., Fort Atkinson, WI, USA) were stimulated with 1000 IU human chorionic gonadotropin (Ferring Pharmaceuticals, Kiel, Germany). *Xenopus laevis* embryos were obtained through in vitro fertilization and staged according to Nieuwkoop and Farber [51]. All experiments were performed at 22 ◦C. For the Wnt/β-catenin signaling specific reporter assay, two transverse blastomeres of 4-cell stage embryos were injected with 4 nL containing 280 pg of S01234-luciferase plasmid, 50 pg of TK-Renilla plasmid, 1 pg of xWnt8 mRNA, and 1.2 pmol of lonchocarpin or DMSO each, for a total of 2.4 pmol of lonchocarpin per embryo. For synthetic xWnt8 mRNA, the plasmid was linearized with NotI and transcribed with SP6 RNA polymerase using the mMessage mMachine kit (Applied Biosystems, Austin, TX, USA). After microinjections, embryos were cultivated in 0.1× Barth (8.89 mM NaCl; 0.1 mM KCl; 0.24 mM NaHCO3; 0.08 mM MgSO4·7H2O; 1 mM Hepes; 0.03 mM Ca(NO3)2·4H2O; 0.04 mM CaCl2·2H2O; pH 7.7) until sibling control embryos reached stage 10. Triplicates of four embryos were lysed using 50 μL of 1× Passive Lysis Buffer (Promega). Then, 10 μL of embryo lysate was used for gene reporter activity detection. Firefly and Renilla luciferase activities were detected according to the manufacturer's protocol (Dual Luciferase Reporter Assay System, Promega).

In order to modulate the maternal wave, 4-cell stage embryos dorsal blastomeres were injected in the animal pole with 4 nL containing 200 μM of lonchocarpin (0.8 pmol/embryo) or DMSO each, for a total of 1.6 pmol of lonchocarpin per embryo. In order to modulate the zygotic wave, stage 9 embryos were injected with 50 nL containing 200 μM of lonchocarpin (10 pmol/embryo) or DMSO into the blastocoel. After injection, embryos were maintained in 0.1X Barth, until stage 35, when the phenotypes were analyzed.

#### *4.6. MTT Assay*

3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) was used to assay mitochondrial activity in viable cells. Cells were plated at a concentration of 1.2 <sup>×</sup> 104 cells/well in

96-well tissue culture plates in DMEM/F-12 containing 10% fetal bovine serum (Gibco) and cultured for 24 h. Cells were treated for 24 h with 10, 20, 30, 40, or 50 μM of lonchocarpin and 0.5% DMSO was used as the vehicle control. MTT was added to each well at a final concentration of 0.25 mg/mL for 1 h. The formazan reaction product was dissolved with 100% DMSO and quantified spectrophotometrically at 570 nm (Modulus™ II Microplate Multimode Reader, Promega).

#### *4.7. Scratch Assay*

HCT116, SW480, DLD-1, and IEC-6 were cultured on 12-well plates in DMEM/F-12 containing 10% fetal bovine serum (Gibco). Confluent wells were scratched in the center of each well with a pipette tip and treated with the proliferation inhibitor Ara-C at 10 μM. After scratch, cells were treated with 5, 10, or 20 μM lonchocarpin for 24 h. Images were taken at 0 h and the wound areas were measured at 0 and 24 h after treatment. Each experiment was carried out in triplicate, and three fields were counted per well.

#### *4.8. Cell Proliferation Assay*

HCT116, SW480, DLD-1, and IEC-6 were cultured on 24-well plates with 4.0 <sup>×</sup> 104 cells/well in DMEM/F-12 containing 10% fetal bovine serum (Gibco). Then, 24 h later, cells were treated with 20 or 30 μM lonchocarpin. DMSO 0.3% was used as vehicle control. Then, 18 h later, we added EdU to the cells, and 6 h later cells were fixed with paraformaldehyde 4% and the experiment was conducted according to Click-iT EdU (Life Technologies Corporation, Carlsbad, CA, USA) manufacturer protocol.

#### *4.9. AOM*/*DSS Protocol*

Animal procedures were approved and carried out according to the guidelines by the Animal Care and Use Ethic Committee from the Federal University of Rio de Janeiro under register 85/15. Male and female 129SvJxC57BL6 mixed mice (8–12 weeks) were housed in microisolator cages and maintained at 23 ◦C with a 12/12-h light/dark cycle and free access to food and water. A total of 26 experimental mice were pre-treated with vermifuge (vetmax plus 0.04% and ivermectin 0.4%) in quarantine for 5 days. All mice were intraperitoneally injected with AOM (12.5 mg/kg; Sigma-Aldrich) once. Five days later, all animals received the first cycle of DSS treatment composed by 2% DSS (MP Biochemicals, Solon, OH, USA) in drinking water for five days, followed by a 2-week rest period without DSS. The 2% DSS treatment cycle was repeated once and followed by a last cycle of 1.5% DSS for 4 days. Mice were monitored every day. Any mouse that lost greater than 20% body weight, demonstrate hunched posture, or moved in a limited fashion was euthanized. Four weeks after the last DSS cycle the animals were randomized into five groups as follows: no treatment group (only AOM/DSS); i.p. injections of 50 or 100 mg·kg−1·day−1, of vehicle (30% polyethylene glycol 400 (PEG400) with 0.9% saline); and i.p. injections of 2.5 mg/mL lonchocarpin 50 or 100 mg·kg−1·day−<sup>1</sup> for four days. Due to limitation of the lonchocarpin solubility, to reach 100 mg·kg−1·day−<sup>1</sup> i.p was performed every 12 h with 50 mg/kg. Euthanasia was performed after the treatment and the animals received an i.p. BrdU injection (100 mg/kg) 1 h before euthanasia.

#### *4.10. Tissue Processing, Histopathology, H&E, and Immunofluorescence*

After euthanasia, the colons were removed, longitudinally opened, cleaned with phosphate-buffered saline (PBS) and fixed in 4% buffered paraformaldehyde for 24 h at 4 ◦C. The swiss-rolls were processed in sequential ethanol and xilol for paraplast inclusion and the tissues were sectioned in Leica RM2125 RTS microtome and stained with hematoxylin and eosin (H&E) for microscopic identification of lesions, adenomas, and adenocarcinomas. Indirect immunofluorescence was performed after serial deparaffinization in xilol and ethanol. Heat induced epitope retrieval was performed in sodium citrate buffer (10 mM Sodium citrate, 0.05% Tween 20, pH 6.0) in a steamer and nonspecific binding sites were blocked with bovine serum albumin 3% in PBS. The sections were incubated with monoclonal antibodies rat anti-Ki-67 (Invitrogen #14569882; 1:100) and mouse anti-BrdU (GE Healthcare #RPN202; 1:3) overnight at 4 ◦C. Anti-rat biotinylated and anti-mouse Alexa 488-conjugated secondary antibodies (Invitrogen) were used to visualize Ki-67 and BrdU, respectively and the nuclei were stained with DAPI. Images were captured using the Olympus Light Microscope BX53 with a LUCPLFLN 20XPH objective and a SC50 color camera (Olympus Life Science Solutions America, Waltham, MA, USA). Only the adenocarcinoma areas confirmed in H&E staining were considered in immunofluorescence analysis. All images were manually and independently counted by at least two authors (LFSO, JMAD and AF). The proportion of positive-stained nuclei in the epithelial crypt cells found in the adenocarcinoma area were calculated and compared between the groups.

#### *4.11. Statistical Analysis*

For MTT assays statistical analysis we used a two-way ANOVA following a Bonferroni post-test (GraphPad Prism version 6.0), error bars represent standard error. For all other results, we used a one-way ANOVA (GraphPad Prism version 6.0). Figures show the mean of three replicates performed three times; standard deviation and statistical significance was set at \* *p* < 0.05 \*\* *p* < 0.01 \*\*\* *p* < 0.001.

#### **5. Conclusions**

In summary, our data describes lonchocarpin, a flavonoid from the chalcone class, as a potent inhibitor of the Wnt/β-catenin pathway that acts downstream to β-catenin stabilization level and impairs TCF mediated transcription. In addition, lonchocarpin presents anti-tumor growth properties in vitro and inhibits adenocarcinoma proliferation in an in vivo CRC model. Further studies should be conducted in order to improve its activity and perhaps propose lonchocarpin as an alternative in CRC treatment.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2072-6694/11/12/1968/s1, Figure S1: *Xenopus laevis* embryo assays quantification. Figure S2: AOM/DSS tumor induction protocol. Table S1: Ki67 positive cell means of each treatment group. Table S2: BrdU positive cell means of each treatment group.

**Author Contributions:** Conceptualization: D.P., L.F.S.O., L.S.S.F., L.A.M., J.M.A.D., N.G.A., A.H.R., F.A.M., H.L.B. and J.G.A. Methodology: D.P., L.F.S.O., L.A.M., J.M.A.D., N.G.A., A.H.R., F.A.M, H.L.B. and J.G.A. Investigation: D.P., L.F.S.O., L.S.S.F., L.A.M., J.M.A.D., A.F., I.O. Resources: C.A.M.F, R.K., F.A.M., H.L.B. and J.G.A. Writing: D.P., L.F.S.O., L.S.S.F., L.A.M., J.M.A.D., F.A.M., H.L.B. and J.G.A. Review and editing: D.P., L.F.S.O., L.S.S.F., L.A.M., J.M.A.D., N.G.A., A.H.R., C.A.M.F., R.K., F.A.M., H.L.B. and J.G.A. Visualization: D.P., L.F.S.O., L.S.S.F., L.A.M., F.A.M. and J.G.A. Supervision: J.G.A. Project administration: J.G.A. Funding acquisition: F.A.M., H.L.B., and J.G.A.

**Funding:** This research was funded by Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), by the Fundação Carlos Chagas Filho de Amparo à Pesquisa do Estado do Rio de Janeiro (FAPERJ), grant numbers 433522/2018-6-18/02/2019, E-26/203.009/2018 (239212) and by the Fundação do Câncer/Oncobiology Program.

**Acknowledgments:** We thank Walter B. Mors, in memoriam, one of the founders of the Research Institute of Natural Products at Federal University of Rio de Janeiro, for donation of lonchocapin purified compound. We thank Simone Rodrigues and Fabio Jorge da Silva for animal care. We thank Carolina Batista for providing technical assistance. This work was supported by the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) and the Fundação Carlos Chagas Filho de Amparo à Pesquisa do Estado do Rio de Janeiro (FAPERJ).

**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/).

#### *Article*

### **Isorhamnetin Induces Cell Cycle Arrest and Apoptosis Via Reactive Oxygen Species-Mediated AMP-Activated Protein Kinase Signaling Pathway Activation in Human Bladder Cancer Cells**

**Cheol Park 1,**†**, Hee-Jae Cha 2,**†**, Eun Ok Choi 3,4, Hyesook Lee 3,4, Hyun Hwang-Bo 3,5, Seon Yeong Ji 3,5, Min Yeong Kim 3,4, So Young Kim 3,4, Su Hyun Hong 3,4, JaeHun Cheong 5, Gi-Young Kim 6, Seok Joong Yun 7, Hye Jin Hwang 8, Wun-Jae Kim 7,\* and Yung Hyun Choi 3,4,\***


Received: 9 July 2019; Accepted: 5 September 2019; Published: 4 October 2019

**Abstract:** Isorhamnetin is an O-methylated flavonol that is predominantly found in the fruits and leaves of various plants, which have been used for traditional herbal remedies. Although several previous studies have reported that this flavonol has diverse health-promoting effects, evidence is still lacking for the underlying molecular mechanism of its anti-cancer efficacy. In this study, we examined the anti-proliferative effect of isorhamnetin on human bladder cancer cells and found that isorhamnetin triggered the gap 2/ mitosis (G2/M) phase cell arrest and apoptosis. Our data showed that isorhamnetin decreased the expression of Wee1 and cyclin B1, but increased the expression of cyclin-dependent kinase (Cdk) inhibitor p21WAF1/CIP1, and increased p21 was bound to Cdk1. In addition, isorhamnetin-induced apoptosis was associated with the increased expression of the Fas/Fas ligand, reduced ratio of B-cell lymphoma 2 (Bcl-2)/Bcl-2 associated X protein (Bax) expression, cytosolic release of cytochrome *c*, and activation of caspases. Moreover, isorhamnetin inactivated the adenosine 5- -monophosphate-activated protein kinase (AMPK) signaling pathway by diminishing the adenosine triphosphate (ATP) production due to impaired mitochondrial function. Furthermore, isorhamnetin stimulated production of intracellular reactive oxygen species (ROS); however, the interruption of ROS generation using a ROS scavenger led to an escape from isorhamnetin-mediated G2/M arrest and apoptosis. Collectively, this is the first report to show that isorhamnetin inhibited the proliferation of human bladder cancer cells by ROS-dependent arrest of the cell cycle at the G2/M phase and induction of apoptosis. Therefore, our results provide an important basis for the interpretation of the anti-cancer mechanism of isorhamnetin in bladder cancer cells and support the rationale for the need to evaluate more precise molecular mechanisms and in vivo anti-cancer properties.

**Keywords:** isorhamnetin; G2/M arrest; apoptosis; ROS; AMPK

#### **1. Introduction**

Although new therapies for treating cancer patients are being developed, chemotherapy is still the main approach for the treatment of cancer. However, some limitations, such as adverse side effects, drug resistance, and limited efficacy, remain to be solved [1,2]. Therefore, urgent new therapeutic strategies that minimize these limitations and have high therapeutic efficacy are required. In this respect, there is an increasing interest in the importance of compounds derived from natural resources that have been traditionally used for the prevention and treatment of various diseases [3–5]. In particular, numerous naturally occurring agents have been reported to cause cell cycle arrest and induce apoptosis, which are important strategies for the control of proliferation in cancer cells, without inducing toxicity in normal cells [6,7]. These agents have also emerged as an alternative to chemopreventive and chemotherapeutic agents because they can specifically regulate various cellular signaling pathways in cancer cells [8,9].

Isorhamnetin (3- -methoxy-3,4- ,5,7-tetrahydroxyflavone) is a flavonol aglycone found in some medicinal plants, such as *Hippophae rhamnoides* L., *Oenanthe javanica*, and *Ginkgo biloba* L., which are used as traditional medicines for the treatment of rheumatism, hemorrhage, cardiovascular disease, and cancer [10,11]. As one of the metabolites of quercetin, isorhamnetin is structurally similar to kaempferol, and is also called 3-O-methyl quercetin [12–14]. Isorhamnetin displays a number of biological properties due to its antioxidant, anti-inflammatory, and metabolic properties [15–19], and is also considered to have potential as an anti-cancer agent based on the results of various cancer cell models. For example, isorhamnetin has been reported to inhibit human leukemia, breast, colon, and cervical cancer cell proliferation through the gap 2/ mitosis (G2/M) phase arrest [20–23], and to induce mitotic block in non-small cell lung carcinoma cells, thus enhancing cisplatin- and carboplatin-induced G2/M arrest [24]. However, isorhamnetin induced S-phase arrest in some cancer cells [25,26], indicating that cell cycle arrest by isorhamnetin is dependent on the type of cancer cell line.

In addition, the anti-cancer effects of isorhamnetin in various cancer cell lines have been shown to involve the death receptor (DR)-dependent extrinsic and/or mitochondria-dependent intrinsic pathways [19,24,27–31], which are representative apoptosis inducing pathways. It was also found that the anti-cancer effect of isorhamnetin was accompanied by the disturbance of various cellular signaling pathways [20,25,32]. Furthermore, isorhamnetin showed a strong cytotoxic effect through a reactive oxygen species (ROS)-dependent apoptosis pathway in breast cancer cells [26]. In particular, isorhamnetin was able to induce high cytotoxicity at low doses compared to quercetin in cancer cells, including hepatocellular carcinoma and leukemia cells [33,34]. Although the possibility of the growth inhibitory activity of isorhamnetin in bladder cancer cells has recently been proposed [35], no molecular mechanism has been reported to support its effect. Therefore, in this study, we investigated the anti-cancer efficacy of isorhamnetin in human bladder cancer cells, focusing on the mechanisms associated with the induction of cell cycle arrest and apoptosis.

#### **2. Results**

#### *2.1. Isorhamnetin Inhibited Cell Viability in Bladder Cancer Cells*

To examine the cytotoxic effect of isorhamnetin, four bladder cancer T24 cell lines (T24, 5637, and 2531J) were treated with various concentrations of isorhamnetin, and then the 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyltetra-zolium bromide (MTT) assay was conducted. Although there are some differences depending on the cell line, the cell viability was significantly decreased in a concentration-dependent manner in isorhamnetin-treated cells (Figure 1A), without affecting normal cultured human keratinocyte HaCaT cells and Chang liver cells under the same conditions. In addition, the 50% inhibitory concentration (IC50) values of isorhamnetin on T24 and

5637 cells were 127.86 μM and 145.75 μM, respectively. The microscopic examination demonstrated that the phenotypic characteristics of isorhamnetin-treated T24 and 5637 cells showed irregular cell outlines, a decrease of cell density, shrinkage, and an increase of detached cells (Figure 1B, upper panel). In addition, 2531J cells showed similar results from the isorhamnetin treatment.

**Figure 1.** The inhibition of cell viability and induction of cell cycle arrest at gap 2/ mitosis (G2/M) phase using isorhamnetin in bladder cancer cells. T24, 5637, and 2531J cells were treated with the indicated concentrations of isorhamnetin for 48 h. (**A**) The cell viability was assessed using 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyltetra-zolium bromide (MTT) assay. Each bar represents the mean ± standard deviation (SD) of three independent experiments (\* *p* < 0.05 and \*\*\* *p* < 0.0001 compared to the control). (**B**, Upper panel) Morphological changes of T24 and 5637 cells were observed using phase-contrast microscopy. (B, Lower panel) The 4- ,6-diamidino-2-phenylindole (DAPI)-stained nuclei were pictured under a fluorescence microscope. Representative photographs of the morphological changes are presented. (**C**,**D**) The cells were stained with propidium iodide (PI) solution for flow cytometry analysis. Quantification of the cell population (in percent) in different cell cycle phases of viable cells is shown. (D) Sub-G 1% was calculated as the percentage of the number of cells in the sub-G1 population relative to the number of total cells. Data were expressed as the mean ± SD of three independent experiments (\* *p* < 0.05, \*\* *p* < 0.001, and \*\*\* *p* < 0.0001 compared to the control).

#### *2.2. Isorhamnetin Induced G2*/*M Phase Arrest and Apoptosis in Bladder Cancer Cells*

To examine the mechanism responsible for the isorhamnetin-induced anti-proliferative effect, the cell cycle distribution profile was examined. Flow cytometry data demonstrated that the percentage of cells arrested at G2/M phase was increased with increasing isorhamnetin treatment concentration, coupled with a decrease in the proportion of cells in the G1 and S phases (Figure 1C). Meanwhile, a significant increase of the cells in the sub-G1 phase, which was used as an index of apoptotic cells, was observed in isorhamnetin-treated cells (Figure 1D). Therefore, 4- ,6-diamidino-2-phenylindole (DAPI) staining was performed to investigate whether apoptosis was involved in cell the growth inhibition induced by isorhamnetin. Figure 1B (lower panel) indicates that morphological changes of the nuclei, which were observed in apoptosis-inducing cells, such as nuclear fragmentation and chromatin

condensation, were dominantly found in isorhamnetin-treated T24 and 5637 cells. Since 2531J cells also had the same results, the following experiments were performed on T24 and 5637 cells. To quantify the apoptosis triggered by isorhamnetin, an annexin V-fluorescein isothiocyanate (FITC)/propidium iodide (PI) double staining assay was conducted. As indicated in Figure 2A,B, after treatment with isorhamnetin, the populations of annexin V-staining positive cells were significantly increased, as compared to the control. On the other hand, T24 cells showed slightly increase in necrotic death upon 100 μM of isorhamnetin, but not 5637 cells (Figure 2B). Consistent with this, the results from agarose gel electrophoresis showed that as the isorhamnetin concentration increased and more fragmented DNA was observed (Figure 2C), indicating that an isorhamnetin-induced G2/M phase arrest was associated with the induction of apoptosis.

**Figure 2.** Induction of apoptosis using isorhamnetin in bladder cancer cells. (**A**,**B**) After treatment with different concentrations of isorhamnetin for 48 h, the cells were fixed and stained with annexin V-fluorescein isothiocyanate (FITC) and PI for flow cytometry analysis. (A) Representative profiles. The results show early apoptosis, defined as annexin V<sup>+</sup> and PI<sup>−</sup> cells (lower right quadrant), and late apoptosis, defined as annexin V<sup>+</sup> and PI<sup>+</sup> (upper right quadrant) cells. (**B**) The percentages of apoptotic cells (left) and necrotic cells (right) were determined by expressing the numbers of Annexin V<sup>+</sup> cells as percentages of all the present cells. The results are presented as the mean ± SD of three independent experiments (\*\* *p* < 0.001 and \*\*\* *p* < 0.0001 compared to the control). (**C**) DNA fragmentation in the cells cultured under the same conditions was analyzed via the extraction of genomic DNA, electrophoresis in agarose gel, and then visualization using ethidium bromide (EtBr) staining.

#### *2.3. Isorhamnetin Regulated the Expression of G2*/*M Phase-Associated Proteins in Bladder Cancer Cells*

To explore the biochemical event of the isorhamnetin-elicited cell cycle arrest, the levels of G2/M phase-associated proteins were analyzed. The immunoblotting results revealed that following isorhamnetin treatment, the levels of Wee1 and cyclin B1 were reduced, and the effect was concentration-dependent, while the expression of cyclin-dependent kinase (Cdk) 1 (also called cell division cycle 2, Cdc2) was maintained at the level of the control group (Figure 3A,B). However, the expression of Cdk inhibitor p21WAF1/CIP1 was markedly increased in response to isorhamnetin exposure. Next, we performed co-immunoprecipitation to investigate the role of isorhamnetin-induced p21, and found that this increased p21 via treatment with isorhamnetin complexed with Cdk1 (Figure 3C). These results suggest that increased p21 protein in isorhamnetin-treated cells contributed to G2/M phase arrest by inhibiting its activity through binding to Cdk1.

**Figure 3.** Effects of isorhamnetin on the levels of cell cycle regulatory proteins in bladder cancer cells. (**A**) T24 and 5637 cells were treated with the indicated concentrations of isorhamnetin for 48 h, and then total cell lysates were prepared. Western blotting was then performed using the indicated antibodies and an enhanced chemiluminescence (ECL) detection system. Actin was used as an internal control. (**B**) The expression of each protein was indicated as a fold change relative to the control. Quantitative analysis of mean pixel density was performed using the ImageJ® software. (**C**) Cells were incubated without or with 100 μM isorhamnetin for 48 h, and then equal amounts of proteins were immunoprecipitated with the anti-cyclin-dependent kinase (Cdk) 1 antibody. Western blotting using immunocomplexes was performed using anti-p21 or anti-Cdk1 antibodies and an ECL detection system (IP, immunoprecipitation).

#### *2.4. Isorhamnetin Modulated the Expression of Apoptosis-Regulatory Proteins in Bladder Cancer Cells*

To investigate the pathway of isorhamnetin-induced apoptosis, caspases activities were determined. Figure 4A,B shows that the protein levels of pro-caspase-8, -9, and -3 were concentration-dependently decreased, which was associated with the degradation of poly(ADP-ribose) polymerase (PARP). Therefore, we quantitatively assessed each caspase activity in the presence of isorhamnetin using fluorogenic substrates to determine whether these immunoblotting results were directly related to activation of the corresponding caspases and found that treatment with isorhamnetin significantly stimulated the activation of these caspases in a concentration-dependent manner in comparison with untreated control cells (Figure 4C). In addition, the effects of isorhamnetin on the expression of the Fas/Fas ligand (FasL) and B-cell lymphoma 2 (Bcl-2) family members were determined. Figure 4A,B shows that both Fas and FasL protein levels were up-regulated, and the level of Bcl-2-associated X protein (Bax), a pro-apoptotic protein, was also increased, while the level of Bcl-2, an anti-apoptotic protein, was reduced in isorhamnetin-treated cells. Furthermore, isorhamnetin promoted the release of cytochrome *c* from mitochondria into cytosol (Figure 4D).

**Figure 4.** Modulation of apoptosis-regulatory factors using isorhamnetin in bladder cancer cells. (**A**) After treatment with isorhamnetin for 48 h, Western blotting was performed using the indicated antibodies and an ECL detection system. Actin was used as an internal control. (**B**) The expression of each protein was indicated as a fold change relative to the control. Quantitative analysis of mean pixel density was performed using the ImageJ® software. (**C**) The activities of caspases were evaluated using caspase colorimetric assay kits. The data were expressed as the mean ± SD of three independent experiments (\* *p* < 0.05, \*\* *p* < 0.001, and \*\*\* *p* < 0.0001 compared to the control). (**D**) After treatment without or with 100 μM isorhamnetin for 48 h, cytosolic and mitochondrial proteins were prepared and analyzed for cytochrome *c* expression using Western blot analysis. Equal protein loading was confirmed via the analysis of actin and cytochrome oxidase subunit VI (COX VI) in each protein extract (C.F.—cytosolic fraction; M.F.—mitochondrial fraction).

#### *2.5. Isorhamnetin Increased ROS Generation but Decreased ATP Content in Cancer Cells*

To investigate the involvement of ROS on the cytotoxic effect of isorhamnetin, we performed flow cytometry analysis using a fluorescent probe, 5,6-carboxy-2- ,7- -dichlorodihydrofluorescein diacetate (DCF-DA). Our data indicated that the production of ROS showed a significant increase within 1 h of the isorhamnetin treatment, and then gradually decreased, while the antioxidant N-acetyl-L-cysteine (NAC) suppressed it to the control level (Figure 5A–C). In addition, an ATP colorimetric assay kit was used to measure the content of mitochondrial ATP in the cells. Figure 5D shows that the concentration of ATP in the isorhamnetin-treated cells decreased in a concentration-dependent manner. However, under the condition that NAC existed, it was markedly weakened, indicating that the decrease in ATP levels was associated with ROS production.

**Figure 5.** Accumulation of reactive oxygen species (ROS) and decrease of ATP content using isorhamnetin in bladder cancer cells. (**A**) Cells were treated with 100 μM isorhamnetin for the indicated times. (**B**) The cells were pre-treated with or without 10 mM N-acetyl-L-cysteine (NAC) for 1 h before isorhamnetin treatment for 1 h. (A,B) The medium was discarded and the cells were incubated for 20 min with medium containing 5,6-carboxy-2- ,7- -dichlorodihydrofluorescein diacetate (DCF-DA). ROS generation was measured using flow cytometry. (**C**) Each bar represents the mean ± SD of three independent experiments. (**D**) After treatment with the indicated concentrations of isorhamnetin in the presence or absence of NAC, the content of intracellular ATP was measured. Each point represents the mean ± SD of three independent experiments (\* *p* < 0.05, \*\* *p* < 0.001, and \*\*\* *p* < 0.0001 compared to control; ## *p* <0.001 and ### *p* < 0.0001 compared to isorhamnetin-treated cells). ISO—isorhamnetin.

#### *2.6. Isorhamnetin Reduced Mitochondrial Membrane Potential (MMP,* ΔΨ*m) and Activated Adenosine 5'-Monophosphate-Activated Protein Kinase (AMPK) Signaling in Bladder Cancer Cells*

We assessed the level of MMP to investigate whether the inhibition of ROS-dependent ATP production by isorhamnetin was associated with impaired mitochondrial function. According to the results of flow cytometry using 5,5- ,6,6- -tetrachloro-1,1- ,3,3- -tetraethylimidacarbocyanine iodide (JC-1) dyes, the formation of JC-1 aggregates in mitochondria was maintained at a relatively high rate in cells not treated with isorhamnetin, while the ratio of JC-1 monomers increased with increasing isorhamnetin treatment concentration, indicating a remarkable depletion of MMP after isorhamnetin treatment (Figure 6A,B). Furthermore, isorhamnetin increased the phosphorylated level of AMPK, as well as its downstream factor acetyl-CoA carboxylase (ACC) (Figure 6C,F), indicating that the AMPK signaling pathway was activated as a result of the loss of ATP. Additionally, we evaluated the effects of isorhamnetin on the phosphorylation of the mechanistic target of rapamycin (mTOR), p70S6K, and Unc-51-like kinase (ULK1), which are AMPK downstream molecules that regulate cell proliferation, apoptosis, and autophagy [36]. Exposure of T24 and 5637 cells with isorhamnetin led to down-regulation in the phosphorylation of mTOR and p76S6K in a dose-dependent manner (Figure 6F,G). Interestingly, we found the isorhamnetin inhibited autophagy via down-regulation of the expression and phosphorylation of ULK1. In addition, our supplementary result showed that the expression of autophagy-related markers was down-regulated using the isorhamnetin treatment, similar to ULK1 (Supplementary Figure S1). However, the presence of NAC or compound C, an antagonist of AMPK, significantly prevented the isorhamnetin-induced loss of MMP (Figure 6A,B), and NAC or compound C also markedly abolished enhanced activation of the AMPK signaling by isorhamnetin (Figure 6E). These data indicate that isorhamnetin-promoted mitochondrial dysfunction associated with the disturbance of ATP production was mediated through an ROS-dependent pathway.

**Figure 6.** Mitochondrial dysfunction and activation of adenosine 5- -monophosphate-activated protein kinase (AMPK) signaling pathway using isorhamnetin in bladder cancer cells. (**A**,**B**) Cells were treated with 100 μM isorhamnetin for 48 h, or pre-treated with 20 μM compound C or 10 mM NAC for 1 h before isorhamnetin treatment for 48 h. (A) The cells were stained with 5,5- ,6,6- -tetrachloro-1,1- ,3,3- -tetraethylimidacarbocyanine iodide (JC-1) dye, and were then analyzed using flow cytometry in order to evaluate the changes in mitochondrial membrane potential (MMP). (**B**) Each bar represents the percentage of cells with JC-1 monomers (mean ± SD of triplicate determinations, \*\* *p* < 0.001 and \*\*\* *p* < 0.0001 compared to the control; ## *p* < 0.001 and ### *p* < 0.0001 compared to the isorhamnetin-treated cells). (**C**,**F**) After treatment with the indicated concentrations of isorhamnetin for 48 h, total cell lysates were prepared and Western blotting was then performed using the indicated antibodies and an ECL detection system. (**D**,**G**) The expression of each protein was indicated as a fold change relative to the control. Quantitative analysis of mean pixel density was performed using the ImageJ® software. (**E**) The cells cultured under the same conditions as A and B were collected, and Western blotting was then performed.

#### *2.7. ROS Acted as an Upstream Regulator of Isorhamnetin-Mediated Apoptosis and Cell Cycle Blockade in Bladder Cancer Cells*

The effect of ROS on isorhamnetin-mediated apoptosis and G2/M phase arrest was further investigated to determine the role of ROS in the anti-cancer activity of isorhamnetin. As depicted in the results of the DAPI staining and flow cytometry analysis, artificially blocking the production of ROS using NAC drastically attenuated isorhamnetin-induced apoptosis (Figure 7A–C). In parallel, pretreatment with NAC protected isorhamnetin-mediated G2/M arrest, which was related to a decrease in the number of sub-G1 phase cells (Figure 7E). Consistent with these results, inhibiting ROS production greatly restored reduced cell viability using isorhamnetin (Figure 7F), demonstrating that ROS generation was shown to be necessary for the contribution of apoptosis and G2/M arrest using isorhamnetin.

**Figure 7.** Roles of ROS in isorhamnetin-induced apoptosis and cell cycle arrest in bladder cancer cells. Cells were either treated with 100 μg/mL isorhamnetin for 48 h or pre-treated with 10 mM NAC for 1 h before isorhamnetin treatment, and were then collected. (**A**) The DAPI-stained nuclei were pictured under a fluorescence microscope. (**B**,**C**) The cells were stained with annexin V-FITC and PI for flow cytometry analysis. (B) Representative profiles. The results show early apoptosis, defined as annexin V<sup>+</sup> and PI<sup>−</sup> cells (lower right quadrant), and late apoptosis, defined as annexin V<sup>+</sup> and PI<sup>+</sup> (upper right quadrant) cells. (C) The percentages of apoptotic cells were determined by expressing the numbers of Annexin V<sup>+</sup> cells as percentages of all the present cells. (**D**) The cells were stained with PI solution for flow cytometry analysis. Quantification of the cell population (in percent) in different cell cycle phases of viable cells is shown. (**E**) The percentages of apoptotic sub-G1 were calculated as the percentage of the number of cells in the sub-G1 population relative to the number of total cells. (**F**) The cell viability was assessed using an MTT assay. Each bar represents the mean ± SD of three independent experiments (\*\*\* *p* < 0.0001 compared to the control; ### *p* < 0.0001 compared to the isorhamnetin-treated cells).

#### **3. Discussion**

In many previous studies, it is clear that the induction of apoptosis by many anti-cancer agents is associated with cell cycle arrest at specific checkpoints [6,37]. In particular, the deregulation of cell cycle control is clearly implicated in the development and progression of most tumors, and the interruption of this progression is considered to be an important strategy to inhibit the proliferation of cancer cells [6,37]. Therefore, we first investigated whether the suppression of bladder cancer cell proliferation by isorhamnetin was associated with cell cycle arrest. The results of flow cytometry analysis showed that isorhamnetin caused G2/M phase arrest, similar to the results of previous studies in several human cancer cell lines [20–23], suggesting that G2/M phase arrest is one of the mechanisms of the growth inhibitory effects of isorhamnetin in human bladder cancer cells. The progression of the cell cycle in eukaryotic cells is tightly controlled by the interaction of cyclins and Cdks with their inhibitory factors [38,39]. In this process, the transition from G2 to M phase is achieved through the increased activity of Cdk1 by cyclin B1 complexing with Cdk1. In addition, Wee1 is a tyrosine kinase that induces phosphorylation of Cdk1, resulting in inhibition of cyclin B-Cdk1 activity and preventing cell mitotic entry [40,41]. In the current study, exposure of bladder cancer cells to isorhamnetin markedly reduced the expression of cyclin B1 and Wee1, without significant changes in the expression of Cdk1.

p21, a typical Cdk inhibitor belonging to the kinase inhibitory protein/ CDK interacting protein (KIP/CIP) family, has a broad-spectrum of specificity in the cell cycle proteins [38,39]. p21 was first reported to be a major inducer of tumor suppressor p53-dependent cell cycle arrest induced by DNA damage, but it could act as a mediator of p53-independent cell arrest in various types of cancer cells [42,43]. As a Cdk inhibitor, when p21 expression increases, it forms complexes with Cdks, reducing their kinase activity and inhibiting cell cycle progression [42,44]. According to our data, isorhamnetin dramatically increased p21 levels with increasing treatment concentration, and increased p21 complexed with Cdk1, which might have contributed to the inhibition of Cdk1 kinase activity. In addition, since T24 and 5637 cells are mutant p53 gene-bearing cell lines [45], increased p21 expression using isorhamnetin was thought to contribute to G2/M arrest, regardless of p53 gene status. Collectively, our data suggest that isorhamnetin-triggered G2/M arrest was due to the decreased expression of Wee1 and cyclin B1, and inactivation of p53-independent p21-mediated Cdk1 kinase.

Because the induction of apoptosis in cancer cells along with cell cycle arrest is a promising approach to cancer therapy, we assessed whether G2/M arrest using isorhamnetin was associated with apoptosis induction. Based on the results of morphological changes, DNA fragmentation, and flow cytometry analysis, we found that the cytotoxic effect of isorhamnetin was achieved through the induction of apoptosis associated with G2/M arrest. As is well known, apoptosis can be largely categorized into extrinsic and intrinsic pathways in mammalian cells [37,46]. The extrinsic pathway is characterized by the activation of caspase-8 by the formation of the death-inducing signal complex through the binding of death ligands to the cell surface DRs. For example, when FasL, one of the typical death ligands, binds to the corresponding DR, Fas, caspase-8 is sequentially activated [46,47]. On the other hand, the intrinsic pathway begins via the activation of caspase-9 through the release of mitochondrial pro-apoptotic proteins, such as cytochrome *c*, from mitochondria to cytoplasm due to increased mitochondrial permeability. This pathway is tightly regulated by the Bcl-2 protein family that includes pro- and anti-apoptotic proteins, which guard mitochondrial integrity and control the release of cytochrome *c* through the mitochondrial transition pore [48,49]. Caspases-8 and -9, which correspond to the initiator caspases of each pathway, ultimately activate apoptosis through the cleavage of various cellular substrates, such as PARP, by activating downstream executioner caspases, including caspase-3 and -7 [37,50]. In addition, these pathways are strictly controlled by a variety of cellular signaling pathways and regulatory molecules [50,51]. Our results show that isorhamnetin increased the expression of Fas and FasL; activated caspase-8, -9, and -3; and induced the cleavage of PARP. In addition, consistent with previous studies in non-small cell lung cancer cells and Lewis lung cancer cells [31,52], mitochondrial dysfunction was induced, as confirmed by the loss of MMP in isorhamnetin-treated cells. Moreover, the loss of MMP was accompanied by a down-regulation in

the Bcl-2/Bax ratio and the promotion of cytosolic release of cytochrome *c*. Therefore, based on those observations, we speculated that the pro-apoptotic effect of isorhamnetin in bladder cancer cells could occur by simultaneously activating extrinsic and intrinsic pathways.

Growing evidence demonstrates that many anti-cancer agents induce apoptosis through pro-oxidant properties, such as increasing ROS accumulation or destroying cellular antioxidant systems [8,53]. In particular, mitochondria are the major subcellular organelles responsible for the production of ROS in the cells and are also a major target of ROS [54,55]. Therefore, elevating intracellular levels of ROS production is considered to be one of the ideal mechanisms for killing cancer cells through the activation of intrinsic pathways. Intriguingly, in various cell types, ROS are involved in activating the signaling system of AMPK, a key sensor that regulates energy balance and cell fate [56–58]. Mitochondrial dysfunction, due to excessive production of ROS, leads to a loss of function of the respiratory chain in the mitochondrial inner membrane, which can lower intracellular ATP levels and activate AMPK [56,57]. Choi et al. first reported that ROS induces concentration-dependent activation of AMPK [59]. More recently, it has been described that AMPK can be activated by ROS, thereby leading to an increase of glycolysis [60,61]. Furthermore, Corton et al. reported that hypoxic activation of AMPK was dependent on the levels of the mitochondrial ROS [62], and Tavazzi et al. demonstrated that AMPK activation was caused by ROS-mediated intracellular ATP depletion [63]. On the contrary, it has been reported that treatment of the potent ROS scavengers, including NAC and dimethyl sulfoxide (DMSO), significantly abolished oxidative stress-induced AMPK activation and ATP depletion [60,61,64]. Consistent with a previous study in breast cancer cells [22], our results show that isorhamnetin treatment markedly increased the levels of ROS production; however, the ROS scavenger, NAC, greatly weakened the accumulation of ROS by isorhamnetin. The quenching of ROS generation also significantly diminished the isorhamnetin-induced disruption of MMP to the control level, followed by significant ATP restoration, indicating that ROS act as upstream signaling molecules to enhance isorhamnetin-mediated mitochondrial dysfunction. Our results also demonstrate that the activation of the AMPK signaling pathway was increased in cells exposed to isorhamnetin, probably due to decreased ATP content. Furthermore, the presence of NAC markedly attenuated isorhamnetin-induced phosphorylation of AMPK, while their total protein levels were kept at an equivalent level, suggesting that the isorhamnetin-induced activation of AMPK signaling pathway is dependent on ROS production. Subsequently, NAC pretreatment also significantly reversed the enhanced apoptosis, G2/M phase arrest, and viability reduction induced by isorhamnetin, confirming that increasing ROS may serve as a key contributor to the anti-cancer effects of isorhamnetin. The AMPK acts as a metabolic mater switch that controls cell fate, such as cell survival, apoptosis, and autophagy [65]. Indeed, fatty acid synthesis is a critical energy-consuming process for the differentiation of tumor cells, and it has been demonstrated that AMPK inhibits lipid synthesis by the phosphorylation and inactivation of acetyl-CoA carboxylase 1 (ACC1) [65]. Furthermore, AMPK directly inhibits mTOR complex I, which regulates p70S6K, an enhancer of protein synthesis. In this sense, AMPK plays a critical role as a cell growth suppressor by inhibiting protein, rRNA, and lipid synthesis [65,66]. In the present study, we conjecture that the AMPK-mediated interruption of the mTOR/p70S6K/ACC1 signaling pathway may contribute to isorhamnetin-induced cell cycle arrest and apoptosis. On the other hand, there are conflicting opinions on the relationship between AMPK and autophagy. Although increasing evidence described that AMPK activation can induce the autophagy through the inhibition of mTOR and phosphorylation of ULK1 [67,68], a few studies reported that ROS attenuated autophagy by the down-regulation of ULK-1 [69,70]. Interestingly, based on our results, we found the isorhamnetin inhibited autophagy by down-regulation of the expression and phosphorylation of ULK1. Therefore, our data suggested that isorhamnetin-induced ROS activates AMPK, and subsequently down-regulates the mTOR/ACC1/ULK1 signaling pathway, which results in promoting cell apoptosis and inhibits autophagy at the same time.

The current results lead us to suggest that the production of ROS by isorhamnetin plays a critical role in the induction of G2/M arrest and apoptosis through simultaneous initiation of both extrinsic and intrinsic pathways in human bladder cancer cells. In addition, ROS act as an upstream signal related to the effect of isorhamnetin on the activation of the AMPK signaling pathway. However, further studies are warranted to identify the molecular mechanisms of isorhamnetin-mediated activation of AMPK signaling on autophagy and mitochondrial energy metabolism in bladder cancer cells. In addition, further studies are required to identify and understand the role of intracellular organelles involved in ROS generation by isorhamnetin, including in vivo animal experiments.

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

#### *4.1. Cell Culture and Isorhamnetin Treatment*

The human bladder cancer cell lines (T24, 5637, 2531J, and EJ) were purchased from the American Type Culture Collection (Manassas, MD, USA). The cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin (100 U/mL penicillin and 100 μg/mL streptomycin, all from WelGENE Inc., Daegu, Republic of Korea) at 37 ◦C under a humidified 5% CO2. The cells were sub-cultured every 3–4 days to maintain logarithmic growth, and were allowed to grow for 24 h before treatments were applied. Isorhamnetin was obtained from Sigma-Aldrich Chemical Co. (St. Louis, MO, USA), and was dissolved in dimethyl sulfoxide (DMSO, Sigma-Aldrich Chemical Co.) to a final concentration of 100 M. Prior to use, the stock solution was diluted with culture medium to the desired concentration.

#### *4.2. Cell Viability Assay*

Cell viability was determined using an MTT assay, as previously described [71]. Briefly, cells (1 <sup>×</sup> <sup>10</sup><sup>4</sup> cells/well) were seeded onto 96-well plates in 100 <sup>μ</sup>L medium. After overnight incubation, the cells were exposed to a series of concentrations of isorhamnetin for 48 h. Thereafter, the MTT reagent (Sigma-Aldrich Chemical Co.) at 50 μg/mL final concentration was added to each well and cells were incubated continuously at 37 ◦C for 2 h. The medium was then removed and 100 μL DMSO was added to each well to dissolve the formed blue formazan crystals, followed by measurement at 540 nm in a microplate reader (Molecular Device Co., Sunnyvale, CA, USA). All results were performed in three independent experiments and the cell survival rate was expressed as a percentage of the control. The morphological changes of cells were directly observed and photographed using phase-contrast microscopy (Carl Zeiss, Oberkochen, Germany).

#### *4.3. Determination of Cell Cycle Distribution Using Flow Cytometric Analysis*

PI staining was applied to analyze the DNA content and cell cycle distribution. In brief, cells were exposed to different concentrations of isorhamnetin for 48 h, and then the cells were harvested and fixed gently in 70% ice-cold ethanol (in phosphate-buffered saline, PBS, WelGENE Inc., Daegu, Korea) at 4 ◦C for 30 min. The cells were re-suspended in PBS containing 40 μg/mL PI, 100 μg/mL RNase A, and 0.1% triton X-100 (all from Sigma-Aldrich Chemical Co.) in a dark room at 37 ◦C for 30 min, and subjected to flow cytometry (BD Biosciences, San Jose, CA, USA), to determine the cell cycle distribution and apoptotic cells (sub-G1 phase).

#### *4.4. Determination of Apoptotic Cell Death by Flow Cytometric Analysis*

The Annexin V-FITC staining kit from BD Biosciences (San Jose, CA, USA) was used to determine and quantify the apoptotic cells using flow cytometry, according to the manufacturer's instruction. In brief, the collected cells were suspended in the supplied binding buffer, and then stained with FITC-conjugated annexin V and PI at room temperature (RT) for 20 min in the dark. The fluorescent intensities of the cells were detected using flow cytometry, and the annexin V+/PI<sup>−</sup> and annexin V+/PI<sup>+</sup> cell populations were considered indicators of apoptotic cells.

#### *4.5. Nuclear Staining and Deoxyribonucleic Acid (DAN) Fragmentation Assay*

The changes of nuclear morphology for assessing apoptosis were assessed using DAPI staining. Briefly, cells were cultured with or without different concentrations of isorhamnetin for 48 h, and were then fixed with 4% paraformaldehyde (Sigma-Aldrich Chemical Co.) for 10 min at RT. The cells were rinsed with PBS, and incubated with 1 μg/mL DAPI solution (Sigma-Aldrich Chemical Co.) at 37 ◦C for 10 min. Stained cells were visualized and photographed using fluorescence microscopy (Carl Zeiss, Oberkochen, Germany). For DNA fragmentation assay, the collected cells were lysed in a buffer containing 10 mM Tris-HCl (pH 7.4), 150 mM NaCl, 5 mM ethylenediaminetetraacetic acid, and 0.5% Triton X-100 for 30 min. The fragmented DNA in the supernatant was extracted using an equal volume of neutral phenol:chloroform:isoamyl alcohol (25:24:1, Sigma-Aldrich Chemical Co.), analyzed electrophoretically on 1% agarose gel containing EtBr (Sigma-Aldrich Chemical Co.), and photographed under a Fusion FX Image system (Vilber Lourmat, Torcy, France).

#### *4.6. Protein Extraction, Co-Immunoprecipitation, and Western Blot Analysis*

After treatment, both adherent and floating cells were harvested, and the whole cellular proteins were prepared using the Bradford protein assay kit (Bio-Rad Laboratories, Hercules, CA, USA), according to the manufacturer's protocol. For the preparation of mitochondrial and cytosolic proteins from the cells, NE-PER nuclear and cytoplasmic extraction reagents (Thermo Fisher Scientific Inc., Waltham, UT, USA) were applied. Protein concentration was measured using the Bio-Rad protein assay kit (Bio-Rad Laboratories, Hercules, CA, USA), according to the manufacturer's instructions. For the co-immunoprecipitation assay, the 500 μg of cell lysates from each sample was precleaned with normal rabbit immunoglobulin G (IgG) and a protein-A-sepharose bead slurry (Amersham, Arlington Heights, IL, USA), and immunoprecipitation was conducted using 1 μg of anti-Cdk1 antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA) and protein-A-sepharose (Sigma-Aldrich Chemical Co.). The protein complex was then prepared according to the previously described method [62]. For Western blot analysis, equal amounts of protein samples or immunoprecipitated proteins were separated using sodium dodecyl sulphate-polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride membranes (Millipore, Bedford, MA, USA) (whole blot figures can be found at the Supplementary). The membranes were blocked with Tris-buffered saline (10 mM Tris-Cl, pH 7.4) containing 0.5% Tween-20 and 5% nonfat dry milk for 1 h at RT, and then probed with the indicated primary antibodies (Santa Cruz Biotechnology, Inc., and Cell Signaling Technology, Danvers, MA, USA), to react with the blotted membranes at 4 ◦C overnight. Afterwards, the membranes were incubated with the corresponding horseradish peroxidase-conjugated secondary antibodies (Santa Cruz Biotechnology, Inc.), developed using an ECL detection kit (GE Healthcare Life Sciences, Little Chalfont, U.K.), and then visualized using a Fusion FX Image system. Densitometric analysis of the data was performed using the ImageJ® software (v1.48, NIH, Bethesda, MD, USA).

#### *4.7. Caspase Activity Assay*

The activity of caspases was measured according to the manufacturer's instructions for the Caspase colorimetric assay kits (R&D Systems, Minneapolis, MN, USA). Briefly, cells were harvested and lysed in the lysis buffer provided in the kit on ice for 10 min, and then centrifuged at 10,000× *g* for 1 min. The supernatants containing equal proteins were incubated with the supplied reaction mixtures, including the fluorogenic peptide substrate (Asp-Glu-Val-Asp specific for caspase-3, Ile-Glu-Thr-Asp for caspase-8, and Leu-Glu-His-Asp specific for caspase-9) labeled with p-nitroaniline (pNA) for 1 h at 37 ◦C in the dark. The amounts of released pNA was measured using a microplate reader using excitement at 405 nm and emitting at 510 nm.

#### *4.8. Measurement of ROS Production and MMP*

The production of ROS was measured using DCF-DA, as described previously [72]. At the end of the treatment with isorhamnetin for defined periods in the presence or absence of NAC (Sigma-Aldrich Chemical Co.), cells were washed with PBS and incubated with 10 μM DCF-DA (Invitrogen, Carlsbad, CA, USA) in the dark at 37 ◦C for 20 min. Subsequently, cells were analyzed for DCF fluorescence using flow cytometry at 480 nm/520 nm. To measure MMP, JC-1 staining was performed according to the manufacturer's instructions. After treatment with isorhamnetin for 48 h in the presence or absence of NAC or compound C, cells were exposed to 10 μM JC-1 (Invitrogen) for 30 min at 37 ◦C, and then analyzed using flow cytometry at 488 nm/575 nm, as previously described [73].

#### *4.9. Detection of ATP Levels*

The firefly luciferase-based ATP Bioluminescence assay kit (Roche Applied Science, Indianapolis, IN, USA) was used for the detection of intracellular ATP levels, according to the manufacturer's instructions. Briefly, cells treated with isorhamnetin for 48 h with or without NAC were lysed with the lysis buffer provided in the kit, and the supernatants were collected via centrifugation at 12,000× *g* for 10 min at 4 ◦C. Subsequently, an equal amount of supernatants and ATP detection reagent, which catalyzed the light production from ATP and luciferin, were mixed. Firefly luciferase activity was immediately measured using a luminometer and the ATP level was calculated according to the ATP standard curve. Intracellular ATP levels were calculated as a percentage of the untreated control.

#### *4.10. Statistical Analysis*

All experiments were performed at least three times. Data were analyzed using GraphPad Prism software (version 5.03; GraphPad Software, Inc., La Jolla, CA, USA), and expressed as the mean ± standard deviation (SD). Differences between groups were assessed using analysis of variance, followed by ANOVA-Tukey's post hoc test, and *p* < 0.05 was considered to indicate a statistically significant difference.

#### **5. Conclusions**

Our findings demonstrate that isorhamnetin exerted an anti-proliferative effect on human bladder cancer cells through the induction of cell cycle arrest during the G2/M phase and apoptosis. Isorhamnetin-induced G2/M arrest was attributed to the decrease in Wee1 and cyclin B1 expression and the upregulation of p21. Isorhamnetin also induced apoptosis by activating caspase-8 and -9, which belong to the initiator caspases of the extrinsic and intrinsic pathways, respectively, followed by the activation of effector caspase-3, leading to the degradation of PARP. In addition, isorhamnetin enhanced the mitochondrial dysfunction, which was associated with an increase in Bax/Bcl-2 expression ratio and cytochrome *c* release into the cytoplasm. Moreover, the induction of G2/M arrest and apoptosis by isorhamnetin was accompanied by activation of the AMPK signaling pathway, and excessive production of ROS. However, artificial interception of the AMPK signaling pathway attenuated isorhamnetin-induced apoptosis, and the interruption of ROS generation led cells to escape from G2/M arrest and apoptosis. Based on these finding, we suggest that isorhamnetin has chemopreventive potential by inducing G2/M arrest and apoptosis through ROS-dependent activation of the AMPK signaling pathway in bladder cancer cells.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2072-6694/11/10/1494/s1, Whole blot figures.

**Author Contributions:** Y.H.C., C.P., and W.-J.K. conceived and designed the experiments; H.-J.C., E.O.C., H.L., H.H.-B., S.Y.J., M.Y.K., S.Y.K., and S.H.H. performed the experiments; G.-Y.K., S.J.Y., J.H.C., and H.J.H. analyzed the data; C.P. wrote the paper; and Y.H.C. edited the paper.

**Funding:** This research was funded by the Basic Science Research Program through a National Research Foundation of Korea (NRF) grant funded by the Korea government (2017R1D1A1B03032689 and 2018R1A2B2005705) and a grant (0820050) from the National R&D Program for Cancer Control, Ministry of Health and Welfare, Republic of Korea.

**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/).

### *Article* **XIAP as a Target of New Small Organic Natural Molecules Inducing Human Cancer Cell Death**

**Diego Muñoz 1,2, Martina Brucoli 3, Silvia Zecchini 4, Adrian Sandoval-Hernandez 1, Gonzalo Arboleda 5, Fabian Lopez-Vallejo 1, Wilman Delgado 1, Matteo Giovarelli 4, Marco Coazzoli 4, Elisabetta Catalani 6, Clara De Palma 7, Cristiana Perrotta 4, Luis Cuca 1, Emilio Clementi 4,7,8 and Davide Cervia 6,\***


Received: 27 August 2019; Accepted: 4 September 2019; Published: 9 September 2019

**Abstract:** X-linked inhibitor of apoptosis protein (XIAP) is an emerging crucial therapeutic target in cancer. We report on the discovery and characterisation of small organic molecules from *Piper* genus plants exhibiting XIAP antagonism, namely erioquinol, a quinol substituted in the 4-position with an alkenyl group and the alkenylphenols eriopodols A–C. Another isolated compound was originally identified as gibbilimbol B. Erioquinol was the most potent inhibitor of human cancer cell viability when compared with gibbilimbol B and eriopodol A was listed as intermediate. Gibbilimbol B and eriopodol A induced apoptosis through mitochondrial permeabilisation and caspase activation while erioquinol acted on cell fate via caspase-independent/non-apoptotic mechanisms, likely involving mitochondrial dysfunctions and aberrant generation of reactive oxygen species. In silico modelling and molecular approaches suggested that all molecules inhibit XIAP by binding to XIAP-baculoviral IAP repeat domain. This demonstrates a novel aspect of XIAP as a key determinant of tumour control, at the molecular crossroad of caspase-dependent/independent cell death pathway and indicates molecular aspects to develop tumour-effective XIAP antagonists.

**Keywords:** phytochemicals; small organic agents; *Piper eriopodon*, alkenylphenols; human cancer cells; cell death; apoptosis; caspase-independent cell death; XIAP antagonists; XIAP-BIR3 domain

#### **1. Introduction**

The characterisation of small molecules (whose molecular weight does not exceed 900 Daltons) with well-defined chemical structures is a good approach to develop new therapeutic agents in proliferative, infectious, or neurodegenerative disorders [1–5]. Natural products possess enormous structural

and chemical diversity that cannot be matched by any synthetic libraries of small molecules and continue to show a great translational potential [6–10]. In some cases, the complex chemical composition of some natural products has made difficult their isolation, structure elucidation and characterisation, thus prompting the search of new efficient synthetic pathways. In recent years the interest in the fundamental understanding of natural products and their engineered variants has been strongly renewed [6].

The simple active chemical structures of phenolic compounds from plants make them optimal lead candidates because of their broad biological activity, especially the protective, anti-oxidant and anti-tumour effects [11–14]. Plants of the genus *Piper* (Piperaceae family), are a very common food resource in neotropical forests and are widely used to obtain culinary spices. *Piper* genus constitutes one major class of medicinal plants and contains a valuable resource of phenolic bioactive compounds [15–21]. Among them, piplartine, hydroxychavicol, 4-nerodlidylcatechol and gibbilimbols A–D displayed potent cytotoxic/anti-tumoural effects in a variety of human cancer cells in vitro and in vivo [19,22–29].

Apoptosis, a closely regulated programmed cell death mechanism, is an essential process to maintain tissue homeostasis and its escape it is one of the hallmarks of cancer [30]. Substantial advances have been made on apoptosis-based anti-cancer therapeutics [31]. The most potent human IAP currently identified is the X-linked inhibitor of apoptosis protein (XIAP), a 57 kDa protein with three zinc-binding baculovirus IAP repeat (BIR) domains (BIR 1–3) which may also have actions additional to regulation of apoptosis [32]. The anti-apoptotic function of XIAP is antagonised by the second mitochondria-derived activator of caspases or direct IAP binding protein with low pI (Smac/DIABLO), a mitochondria protein released during apoptosis. The key role of XIAP and its potential clinical relevance is well established in tumours and several XIAP inhibitors have been developed or discovered as cytotoxic agents [32–43]. Despite different small molecules that inhibit XIAP have been identified and are moving through the pipeline of clinical development, the need of new ones to refine further therapeutic approaches based on XIAP antagonism is undeniable in translational research [41].

Herein we wish to report the discovery and chemical/biological characterisation of novel natural small compounds from *Piper* genus. Furthermore, a deeper insight into their cell death mechanism in human cells provides a proof-of-concept study of their pharmaceutical potential as antagonists of XIAP that may open important insights on XIAP as a suitable turning point for multiple cellular pathways.

#### **2. Results and Discussion**

#### *2.1. Structural Identification of New Piper Genus-Derived Compounds*

The chemical structures of compounds isolated from leaves of *P. eriopodon* (Figure 1A) were identified by interpretation of their corresponding high resolution electrospray ionisation mass spectrometry (HRESIMS), 1H- and 13C-NMR (nuclear magnetic resonance) spectral data, including attached proton test (APT), correlated spectroscopy (COSY), heteronuclear multiple quantum coherence (HMQC) and heteronuclear multiple bond correlation (HMBC) experiments, as well as by comparison of the spectral data with those reported in the literature.

**Figure 1.** Identification of new *Piper* genus-derived compounds. (**A**) Structures of compounds 1–5. (**B**) Key correlated spectroscopy (COSY) (bold) and heteronuclear multiple bond correlation (HMBC) (H→C) for compounds **2–5**.

Compound 1 (Figure S1, Tables S1 and S2) was obtained as colorless oil and identified unequivocally as gibbilimbol B ((*E*)-4-(dec-3- -enyl)phenol) [19].

Compound 2 (Figure S2, Tables S1 and S2) was obtained as pale yellow oil. The molecular formula for compound **2** was established as C16H24O2 based on the HRESIMS peak at *m*/*z* 247.1706 [M-H]<sup>−</sup> (calcd. 247.1703). The 1H- NMR spectrum showed clear signals for a 1,2,4-trisubstituted aromatic ring δH 6.77 (1H, d, *J* = 7.6 Hz, H-6), 6.71 (1H, s, H-3), 6.60 (1H, d, *J* = 7.5 Hz, H-5) and an alkenyl fragment. The 13C-NMR spectrum showed ten signals, practically the same as the alkenyl chain of gibbilimbol B, including the double bond position in C-3- , which was confirmed by correlations observed in both COSY and HMBC experiments (Figure 1B). Based on the 13C-NMR chemical shifts of the allylic carbons δ<sup>C</sup> 34.6 (C-2- ) and δ<sup>C</sup> 32.6 (C-5- ), the configuration of the double bond for compound 2 was assigned as *E* [18], by comparison with the 13C-NMR chemical shift of the allylic carbons in the *E* analogue gibbilimbol B (δ<sup>C</sup> 34.6 (C-2- ) and δ<sup>C</sup> 32.6 (C-5- )), which differed significantly from the chemical shift values reported for the *Z* analogue climacostol [δ<sup>C</sup> 33.2 (C-1- ) and δ<sup>C</sup> 27.3 (C-4- )] [44]. Thus, the chemical structure of compound **2** was elucidated as (*E*)-4-(dec-3- -enyl)benzene-1,2-diol and it was given the common name of eriopodol A.

Compound **3** (Figure S3, Tables S1 and S2) was obtained as clear oil and its molecular formula was deduced as C16H24O2 from the HRESIMS spectrum, which exhibited a molecular ion peak at *m*/*z* 247.1706 [M-H]<sup>−</sup> (calcd. 247.1703). The 1H-NMR spectrum for compound 3 showed signals for an alkenyl chain and two signals in δH 6.11 (2H, d, *J* = 9.94 Hz) and 6.81(2H, d, *J* = 9.96 Hz). The 13C-NMR spectrum for compound **3** showed signals for an α-β unsaturated carbonyl in δ<sup>C</sup> 185.9, an oxygenated quaternary carbon in δ<sup>C</sup> 69.6 and ten signals for the typical side chain of the alkenyl fragment. Based on the correlations observed in COSY and HMBC experiments (Figure 1B), the structure of 3 was determined as a quinol derivative, substituted in the 4-position with an alkenyl group. The position and geometry of the double bond of compound 3 was assigned by comparing the chemical shift values of the allylic carbons δ<sup>C</sup> 32.4 (C-2- ) and δ<sup>C</sup> 26.6 (C-5- ) as explained above for eriopodol A. The geometry of compound *3* was determined as *Z* and its chemical structure was elucidated as (*Z*)-4-(dec-3- -enyl)-4-hydroxycyclohexa-2,5-dien-1-one. The common name of erioquinol was then assigned.

Compound 4 (Figure S4, Tables S1 and S2) was obtained as pale yellow oil. The molecular formula for compound 4 was confirmed to be C16H24O2 based on the HRESIMS peak at *m*/*z* 247.1715 [M-H]<sup>−</sup> (calcd. 247.1703). The 1H-NMR and 13C-NMR spectra of compound 4 showed almost the same chemical shifts as the alkenylphenol gibbilimbol B, but without the unsaturated signal in the 1H-NMR spectrum. Therefore, the carbons C-3 and C-4 showed chemical shifts in δ<sup>C</sup> 59.6 (C-3- ) and δ<sup>C</sup> 58.6 (C-4- ), corresponding two oxygenated methines from an epoxide group, which was confirmed with COSY and HMBC experiments (Figure 1B). The structure of compound **4** was elucidated as 4-(3- ,4- -epoxydecenyl)phenol and the common name of eriopodol B was assigned.

Compound 5 (Figure S5, Tables S1 and S2) was obtained as pale yellow amorphous solid (m.p. 138.5 ◦C). The molecular formula for compound **5** was established as C18H28O3 based on the HRESIMS peak at *m*/*z* 291.1973 [M-H]− (calcd. 291.1966). The NMR data for compound **5** were very close to those of eriopodol A, although it contains one additional hydroxyl group in the benzene ring and two additional carbons at the end of the alkenyl chain (Figure 1B). The position and geometry of the double bound for compound **5** was assigned as explained above. The structure of compound **5** was elucidated as (*E*)-5-(dodec-3- -enyl)benzene-1,2,4-triol and the common name of eriopodol C was assigned.

Taken together, phytochemical investigation of leaves from P. eriopodon yielded four new alkenyl derivatives and one known compound. In particular, erioquinol is a new quinol substituted in the 4-position with an alkenyl group and eriopodols A-C correspond to new alkenylphenols. The known isolated compound was originally identified as gibbilimbol B, from the medicinal plant P. gibbilimbum [19] and, more recently, from P. malacophyllum [21] and P. eriopodon [29]. The simple chemical structure of alkenylphenols are characterised by hydroxylated benzenes, substituted by side alkyl chains of different lengths with at least one double bond, generally with E geometry. Alkenylphenols with different reported biological properties, such as antibacterial, anti-parasitic, anti-inflammatory and cytotoxic activities, are widely found in the Piper genus [17,19–21,45]. Quinols are 4-hydroxycyclohexa-2,5-dien-1-ones which rarely occur as derivatives of some natural products [46–48]. An important feature of quinols substituted in the 4-position with aryl groups, is that they represent a class of potent anti-tumour molecules with activities against colon, renal, and breast cancer cells [49–51].

#### *2.2. Piper Genus-Derived Compounds Exhibit Cytotoxic E*ff*ects*

Several recent studies in glioblastoma and breast cancer cells have reported that extracts or active compounds isolated from *Piper* genus possess anti-tumoural/pro-apoptotic properties [52–61]. In order to assess whether the compounds we isolated could be developed further for therapeutic applications, we tested their cytotoxic action in the human cancer cells, U373 (glioblastoma astrocytoma) and MCF7 (breast adenocarcinoma) cell lines, since they are widely used as suitable in vitro models of cancer

research. We first examined the effects of gibbilimbol B, eriopodols A–C, and erioquinol on cell viability. Gibbilimbol B was used as a reference compound of *Piper* genus derivatives, since its cytotoxic action has been previously tested in various tumour cells, including MCF7 [19,29]. In our experiments, cell viability was analysed by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay after treatment with previously mentioned compounds at increasing concentrations for 24 h. As shown in Figure 2, a concentration-dependent inhibition of MTT absorbance was observed for all compounds with an IC50 (the concentration producing half the maximum inhibition) ranging from 1.78 to 31.91 μg/mL; the rank order of potencies was: erioquinol > eriopodol A > eriopodol C > gibbilimbol B > eriopodol B and erioquinol > eriopodol A > eriopodol C/gibbilimbol B > eriopodol B for U373 and MCF7 cells, respectively (Table 1). Their effects were maximal (Emax—concentration producing the maximum effect—nearly 100% inhibition) between 10–100 μg/mL.

**Figure 2.** *Piper* genus-derived compounds exhibit cytotoxic effects in human cancer cells. U373 and MCF7 cells were treated with increasing concentrations of gibbilimbol B, eriopodol A, eriopodol B, eriopodol C,

and erioquinol for 24 h before 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. Results are expressed by setting the absorbance of the reduced MTT in the respective control (vehicle-treated) samples, i.e., absence of compounds, as 100%. The data points are representative of four independent experiments.


**Table 1.** Inhibitory effects of *Piper* genus-derived compounds on human cancer cell viability.

3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay was performed treating cells for 24 h in the absence (vehicle) or in the presence of increasing concentrations of *Piper* genus-derived compounds. The results have been obtained in four independent experiments.

Eriopodol A and erioquinol were selected for further investigation, as they displayed the most potent inhibitory effects on cell viability. Gibbilimbol B (available in high quantity) was also included. When compared with gibbilimbol B [29], the higher cytotoxic effect of eriopodol A and erioquinol (24 h), was also shown by MTT assays using additional cell lines, like human A549 lung (IC50 of eriopodol A and erioquinol: 6.12 and 2.65 μg/mL, respectively) and PC-3 prostate (IC50 of eriopodol A and erioquinol: 1.84 and 2.21 μg/mL, respectively) cancer cells, further confirming enhanced pharmacological activity of these new *Piper* genus derivatives (Figure 3A).

**Figure 3.** *Piper* genus-derived compounds exhibit cytotoxic effects in cancer and non-transformed human cells. (**A**) PC-3/A549 cells were treated with increasing concentrations of eriopodol A and erioquinol while

(**B**) human umbilical vein endothelial cells (HUVEC)/MCF10 cells were treated with increasing concentrations of gibbilimbol B, eriopodol A, and erioquinol, for 24 h before 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. Results are expressed by setting the absorbance of the reduced MTT in the respective control (vehicle-treated) samples, i.e., absence of compounds, as 100%. The data points are representative of four independent experiments.

Similar results were obtained in human umbilical vein endothelial cells (HUVEC) (IC50 of 24 h gibbilimbol B, eriopodol A, and erioquinol: 11.49, 0.99, and 0.36 μg/mL, respectively) and the non-tumourigenic human breast MCF10 cells (IC50 of 24 h gibbilimbol B, eriopodol A, and erioquinol: 17.11, 4.27, and 1.70 μg/mL, respectively) (Figure 3B). The fact that the potency of the compounds was even slightly higher in these non-transformed/high proliferating cells suggests that their effects are not necessarily correlated to the cancerous origin of cells, in agreement with other small molecules we have recently characterised [62]. On the other hand, many cytotoxic compounds, including chemotherapy agents, are specifically designed to primarily affect rapidly proliferating cells, and many "normal" cells are also highly proliferative, such as cells in the bone marrow. The possibility that *Piper* genus-derived compounds preferentially affect high proliferating vs. low proliferating cells remains to be elucidated.

We then measured the concentration-dependent inhibition of MTT absorbance at increasing times of exposure in MCF7 cells, used as reference cell line. Our results indicated that the potency of gibbilimbol B did not substantially change (IC50-6 h: 20.31 μg/mL; 12 h: 27.36 μg/mL; 24 h: 16.44 μg/mL) while the potency of eriopodol A increased at 24 h (IC50 - 6 h: 31.19 μg/mL; 12 h: 32.75 μg/mL; 24 h: 11.13 μg/mL) (Figure 4). Of interest, the potency of erioquinol was greater than gibbilimbol B and eriopodol A at each time-point, even increasing over time (IC50-6 h: 14.72 μg/mL; 12 h: 4.25 μg/mL; 24 h: 1.93 μg/mL). These comparative data indicate that erioquinol is the most potent compound with faster kinetics when compared with gibbilimbol B; eriopodol A has a somewhat intermediate behavior.

**Figure 4.** Time-response of *Piper* genus-derived compounds on cell viability. MCF7 cells were treated with increasing concentrations of gibbilimbol B, eriopodol A, and erioquinol for 6, 12, and 24 h before 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. Results are expressed by setting the absorbance of the reduced MTT in the respective control (vehicle-treated) samples, i.e., absence of compounds, as 100%. The data points are representative of four independent experiments.

#### *2.3. Piper Genus-Derived Compounds Induce Cell Death*

MCF7 cells treated for 12 h with gibbilimbol B and eriopodol A (30 μg/mL) showed an inter-nucleosomal degradation of genomic DNA typical of late apoptotic cells, as determined by a terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay (Figure 5A), while DNA fragmented cells were few following erioquinol (10 μg/mL) treatment. Bright field microscopy demonstrated that cells exposed to increasing concentrations of gibbilimbol B and eriopodol A at 6 h (a temporal window sufficient to determine their cytotoxic effects) had morphological hallmarks of apoptosis, such as progressive roundness, shrunken cytoplasm and the formation of condensed nuclei (Figure 5B). In contrast, cells treated with erioquinol displayed a translucent cytoplasm and no overall nuclei condensation. Of interest, 4',6-diamidine-2'-phenylindole dihydrochloride (DAPI) staining clearly revealed the nuclei of cells undergoing apoptosis in the presence of gibbilimbol B

and eriopodol A (30 μg/mL) for 6 h, while erioquinol (10 μg/mL) treatment was associated with the appearence of multinucleated cells (Figure 5C). Accordingly, when analysed by flow cytometry using Annexin V and propidium iodide (PI) staining, MCF7 cells treated with erioquinol showed a progressive and marked increase of membrane disruption, as shown by early positivity to both Annexin V and PI staining, while the typical early apoptotic pattern, evidenced as Annexin V+/PI<sup>−</sup> was almost undetectable over time (Figure 5D).

**Figure 5.** *Piper* genus-derived compounds induce cell death. (**A**) terminal deoxynucleotidyl transferase

dUTP nick end labeling (TUNEL) staining of MCF7 cells treated for 12 h in the absence (CTRL, control) and in the presence of gibbilimbol B/eriopodol A (30 μg/mL) or erioquinol (10 μg/mL). 4',6-diamidine-2'-phenylindole dihydrochloride (DAPI) was used for nuclei detection. Scale bar = 50 μm. (**B**) Bright field microscopy of MCF7 cells treated for 6 h in the absence (CTRL) and in the presence of gibbilimbol B, eriopodol A, or erioquinol at increasing concentrations. Scale bar = 100 μm. (**C**) 4',6-diamidine-2'-phenylindole dihydrochloride (DAPI) staining of MCF7 cells treated for 6 h in the absence (CTRL) and in the presence of gibbilimbol B/eriopodol A (30 μg/mL) or erioquinol (10 μg/mL). Scale bar = 10 μm. Lower panels represent enlarged image details. (**D**) Evaluation by flow cytometry of Annexin V-fluorescein isothiocyanate(FITC)/propidium iodide (PI) staining in MCF7 cells treated in the absence (CTRL) and in the presence of 10 μg/mL erioquinol, for 3 and 6 h. Quadrants are drawn, and relative proportion of labelled cells is indicated. The events shown in the lower left-hand quadrant are unlabeled cells. Images and data are representative of four independent experiments.

In addition, cells treated with 30 μg/mL gibbilimbol B and eriopodol A displayed activation of caspase 9 and 7 at 3 h and 6 h, as showed by western blot analysis (Figure 6A,B). On the other hand, erioquinol (10 μg/mL) treated cells did not display any sign of caspase 7 activity even at later time point (Figure 6C). These results were confirmed by immunofluorescence experiments. Indeed, a timedependent and intensive cleaved-caspase 7 staining was detected in the cytoplasm of MCF7 cells in the presence of gibbilimbol B and eriopodol A while positive cells were absent following the administration of erioquinol (Figure 6D). The activation of caspase 7 by 6 h gibbilimbol B and eriopodol A (30 μg/mL) but not erioquinol (10 μg/mL) was achieved also in U373 cells (Figure 7A). The fact that these cells displayed apoptotic and non-apoptotic features in the presence of gibbilimbol B/eriopodol A and erioquinol, respectively (Figure 7B), similarly to what obtained in MCF7 cells, indicate that cell death mechanisms of the compounds are comparable among cell lines. Accordingly, the activation of caspase 7 by gibbilimbol B and eriopodol A but not erioquinol was observed also in MCF10 cells (Figure S6A).

**Figure 6.** *Piper* genus-derived compounds induce cell death. Western blot analysis of cleaved-caspase 9 and 7 in MCF7 cells treated for increasing times in the absence (CTRL, control) and in the presence of (**A**) gibbilimbol B or (**B**) eriopodol A. Vinculin was used as internal standard. Right panels: densitometric analysis expressed as fold change of CTRL. Images and data are representative of three-five independent experiments. \* *p* < 0.01, \*\* *p* < 0.001, and \*\*\* *p* < 0.0001 relative to CTRL. (**C**) Western blot analysis of cleaved-caspase 7 in MCF7 cells treated for increasing times in the absence (CTRL) and in the presence of 10 μg/mL erioquinol. The stain-free gel was used as loading control. Images are representative of three independent experiments. PC: positive control. (**D**) Immunofluorescence imaging of cleaved-caspase 7 (punctate red pattern) in MCF7 cells treated for 3 and 6 h in the absence (CTRL) and in the presence of gibbilimbol B/eriopodol A (30 μg/mL) or erioquinol (10 μg/mL). 4',6-diamidine-2'-phenylindole dihydrochloride (DAPI) (blue) and phalloidin (green) were used for nuclei and cytoskeleton detection, respectively. Images are representative of four independent experiments. Scale bar = 25 μm.

**Figure 7.** *Piper* genus-derived compounds induce cell death and mitochondrial dysfunction. (**A**) Immunofluorescence (confocal) imaging of cleaved-caspase 7 (punctate red pattern) in U373 cells treated for 6 h in the absence (CTRL, control) and in the presence of gibbilimbol B/eriopodol A (30 μg/mL) or erioquinol (10 μg/mL). 4',6-diamidine-2'-phenylindole dihydrochloride (DAPI) (blue) and phalloidin (green) were used for nuclei and cytoskeleton detection, respectively. Scale bar = 25 μm. (**B**) Bright field microscopy (upper panels) and DAPI staining (lower panels) of U373 cells treated for 6 h in the absence (CTRL) and in the presence of of gibbilimbol B/eriopodol A (30 μg/mL) or erioquinol (10 μg/mL). Scale bars = 50 μm (bright field) and 10 μm (DAPI). Lower panels represent enlarged image details. (**C**) Quantitative analysis of tetramethylrhodamine methyl ester (TMRM) fluorescence changes over time in MCF7 (left panel) and U373 (righ panel) cells in the absence (CTRL) and in the presence of gibbilimbol B/eriopodol A (30 μg/mL) or erioquinol (10 μg/mL). Results are expressed by setting TMRM fluorescence in the respective control (vehicle-treated) samples, i.e., absence of compounds, as 1. \*\*\* *p* < 0.0001 relative to CTRL. Images and data are representative of four independent experiments.

In order to better describe the mechanism behind the activity of the compounds, we investigated mitochondria functionality with tetramethylrhodamine methyl ester (TMRM), a red fluorescent dye that is sequestered by active mitochondria. Of note, MCF7 and U373 cells treated for increasing time with 30 μg/mL gibbilimbol B/eriopodol A or 10 μg/mL erioquinol presented a comparable decrease in TMRM fluorescence vs. control, with MCF7 cells full responding within 1 h (Figure 7C). This indicates low mitochondria membrane potential likely associated to the destabilisation of the mitochondrial membrane systems.

The fact that the three compounds similarly induce mitochondria membrane permeabilisation both in MCF7 and U373 cells, was further confirmed by the subcellular location of cytochrome c. As shown in Figure 8A,B, 3 h administration of gibbilimbol B/eriopodol/erioquinol induced an alteration in the cytochrome c staining pattern from mitochondrial (co-localisation with COX IV, a marker for mitochondria), to a more cytosolic distribution (presence of many clusters which did not overlap with COX IV), indicating a release of cytochrome c from the dysfunctional mitochondria.

**Figure 8.** Confocal microscopy for co-localisation of cytochrome c with mitochondria. (**A**) MCF7 and

(**B**) U373 cells were treated for 3 h in the absence (CTRL, control) and in the presence of gibbilimbol B/eriopodol A (30 μg/mL) or erioquinol (10 μg/mL). Cells were then stained for cytochrome c (green) and the mitochondrial marker COX IV). 4',6-diamidine-2'-phenylindole dihydrochloride (DAPI) (blue) was used for nuclei detection. The images are representative of three independent experiments. Scale bars: 10 μm (MCF7) and 25 μm (U373). Panels on the right represent enlarged image details.

Damaged mitochondria are considered as the main source of reactive oxygen species (ROS) which play major roles in the fate of cancer cells [63]. Noteworthily, MCF7 and U373 cells staining with 2'-7'dichlorofluorescin diacetate (DCFH-DA), a permeant fluorogenic dye cell reagent that measures hydroxyl, peroxyl and other ROS activity, revealed that erioquinol effect (10 μg/mL, 6 h) is characterised by marked accumulation of ROS, which are absent in cells treated with gibbilimbol B and eriopodol A (30 μg/mL, 6 h) (Figure 9A,B). Together with lack of caspase activation, aberrant ROS production is another divergence between gibbilimbol B/eriopodol A and erioquinol-induced cell death. In this respect, erioquinol is likely inducing a robust mitochondrial stress which results in ROS production and release into the cytoplasm.

**Figure 9.** Confocal microscopy for reactive oxygen species (ROS) detection. (**A**) MCF7 and (**B**) U373 cells were treated for 6 h in the absence (CTRL, control) and in the presence of gibbilimbol B/eriopodol A (30 μg/mL) or erioquinol (10 μg/mL). Cells were then stained for ROS (2'-7'dichlorofluorescin diacetate - DCFH-DA, green). 4',6-diamidine-2'-phenylindole dihydrochloride (DAPI) (blue) and phalloidin (red) were used for nuclei and cytoskeleton detection, respectively. The images are representative of three independent experiments. Scale bar: 25 μm.

Finally we confirmed as apoptotic the effect of gibbilimbol B and eriopodol A by inhibiting their cytotoxic activity with the pan-caspase inhibitor Z-VAD-(OMe)-FMK. As displayed by MTT assays (Figure 10A), the loss of cell viability in MCF7 cells treated with 30 μg/mL gibbilimbol B and eriopodol A was significantly inhibited when 50 μM Z-VAD-(OMe)-FMK was simultaneously added to the 6 h treatment protocol, demonstrating the dependency on caspases of the two compounds. However, the simultaneous addition of Z-VAD-(OMe)-FMK did not affect the activity of 10 μg/mL erioquinol. Taken together our data data demonstrate that gibbilimbol B and eriopodol A effectively induced intrinsic apoptosis triggered by mitochondrial membrane permeabilisation, release of cytochrome c, an early induction of initiator caspase 9, and a consecutive activation of effector caspase 7. Erioquinol, although it affects comparably mitochondrial functions, appears to act in a different manner, i.e., involving mitochondrial ROS release and non-apoptotic/caspase-independent mechanisms. Caspase-independent cell death was first described to affect mitochondria potential, and eventually mitochondrial outer membrane permeabilisation [64], although not followed by caspase activation. Those features resemble the outcome of erioquinol treatment.

**Figure 10.** *Piper* genus-derived compounds induce caspase-dependent and independent loss of cell viability. (**A**) MCF7 cells were cultured in the absence (CTRL, control) and in the presence of 30 μg/mL gibbilimbol B/eriopodol A or 10 μg/mL erioquinol for 6 h, before 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. The pan-caspase inhibitor Z-VAD-(OMe)-FMK (50 μM) or its vehicle were used as well. Results are expressed by setting the absorbance of the reduced MTT in the CTRL, as 100%. Data are representative of four-twelve independent experiments. \*\*\* *p* < 0.0001 relative to the respective compound alone, i.e., + Z-VAD vehicle. (**B**) MCF7 (left panel) and U373 (right panel) cells were treated with increasing concentrations of erioquinol for 24 h before MTT assay. Erioquinol was administered both in the absence (vehicle) or in the presence of 50 μM necrostatin-1 and 10 μM ferrostatin-1 (2 h pre-treatment), a necroptosis and ferroptosis inhibitor, respectively. Results are expressed by setting the absorbance of the reduced MTT in the control samples (absence of erioquinol) as 100%. The data points are representative of four independent experiments.

Several forms of regulated cell death manifest with a morphology different from apoptosis [65,66], and many compounds from nature source can induce non-apoptotic programmed cell death in cancer cells [67]. Among them, necroptosis can be partially rescued by the receptor-interacting serine-threonine kinase 1 inhibitor necrostatin-1 and ferroptosis by ferrostatin-1, an inhibitor of lipid peroxidation. We thus treated MCF7 and U373 cells with increasing concentrations of erioquinol (24 h) with or without 50 μM necrostatin-1 and 10 μM ferrostatin-1 (2 h pre-treatment). As shown in Figure 10B, the concentration-dependent inhibition of MTT absorbance did not change, suggesting erioquinol-induced death was independent from necroptosis and ferroptosis, two cell death pathways known to be caspase-independent [65,66]. ROS were recently linked to a caspase-independent form of cell death, which cannot be rescue by necrostatin-1 or ferrostatin-1 treatment, and therefore not imputable to either necroptosis or ferroptosis [68]. Treatment with erioquinol might lead to a similar cascade of events, although additional work is required to fully characterise the role of ROS and the cell death process induced by this *Piper* genus-derived compound.

#### *2.4. XIAP as a Molecular Target of Piper Genus-Derived Compounds*

XIAP-mediated inhibition of apoptosis goes through its reversible binding to active caspase-9, via its BIR3 domain, and caspase-3/7 when stabilised to XIAP-BIR2 domain [69–71]. It has been also demonstrated that XIAP controls different pathways functionally uncoupled to caspases, leading to the possibility that XIAP system might control cell death/survival through multiple mechanisms [32,34,72–82].

Embelin, a natural benzoquinone with potential therapeutic interest, has been isolated from the fruit of the *Embelia* ribes and discovered through molecular docking analysis of over 8200 molecules as a potent small molecule XIAP inhibitor that binds to the XIAP-BIR3 domain [83–86]. It should be noted that embelin displays chemical features similar to those of erioquinol, eriopodol A, and gibbilimbol B [83]. We assessed if erioquinol, eriopodol A, and gibbilimbol B are able to bind to the XIAP-BIR3 domain in a similar way of embelin. Using molecular docking analysis and molecular dynamics simulations for embelin and isolated new compounds, it was found the structural basis of the predicted interactions with the BIR3 domain of XIAP. Figure 11A provides a general view of the docked conformations obtained for gibbilimbol B, eriopodol A, erioquinol, and embelin. Interestingly, the binding site for gibbilimbol B, eriopodol A, and erioquinol is the same binding site of embelin and with similar energy and binding mode. All docked compounds fits in to the P1, P2 and P3 of the P1–P4 pockets reported for the binding site of the XIAP-BIR3 domain in complex with Smac, the endogenous antagonist ligand of IAPs [35,87,88].

**Figure 11.** *Cont*.

**Figure 11.** X-linked inhibitor of apoptosis protein (XIAP) as a molecular target of *Piper* genus-derived compounds. (**A**) Molecular docking and (**C**) dynamics analysis of embelin (orange), erioquinol (green), eriopodol A (purple) and gibbilimbol B (blue) in complex with the baculovirus IAP repeat (BIR)-3 domain of XIAP (PDB code 5C83). Interacting residues are displayed in wireframe, hydrogen bonds are displayed in yellow dot lines and π-π stacking interactions are displayed in blue dot lines. (**B**) Protein-ligand root mean square deviation (RMSD) trajectory of the atomic positions for ligands (red, Lig fit Prot) and the receptor (blu, Cα positions) BIR-3 domain of XIAP, for the dynamics trajectory of 50 ns.

Experimental structures of the XIAP-BIR3 domain in different complexes with embelin, Smac or Smac mimetics and non-peptidomimetics small molecules, revealed that residues GLY306, THR308, GLU314, TRP323 and TYR324 are crucial residues involved in the interaction with the BIR3 domain of XIAP [35,41,86,89]. The results of the docking experiments show a possible binding mode for gibbilimbol B, eriopodol A, and erioquinol. Accordingly, the phenolic ring of gibbilimbol B and eriopodol A forms hydrogen bonds with LYS311 and GLU314 (Figure 11A), the quinol ring of erioquinol forms three hydrogen bonds with THR308, LYS322, and TRP323, while residues GLY306, LEU307, TRP323, and TYR324 of the XIAP-BIR3 domain forms hydrophobic interactions with the tail of the alkenyl derivatives.

In addition, molecular dynamics simulations for 50 ns were carried out to assess the stability of the protein-ligand complexes between the docked compounds and the BIR3 domain of XIAP. The stability of the modelled complex of alkenyl derivatives and embelin was confirmed during the period of simulation by little variations in the root mean square deviation (RMSD) trajectory (Figure 11B). Although some changes were observed in the interacting residues of XIAP BIR-3 domain after molecular dynamics simulations (Figure S7), the preferred location of the binding mode for all evaluated ligands were maintained in the pockets P1-P4 of BIR-3 domain of XIAP during the period of simulation (Figure 11C). Also, the binding mode obtained in the docking and dynamics simulations for embelin are according to the interactions pattern determined experimentally by NMR studies in the XIAP-embelin complex, which revealed that TRP323 of the BIR3 domain of XIAP are crucial in the binding of embelin [86]. These findings strongly suggest the highly stable complex formation between the BIR-3 domain of XIAP and the alkenyl derivatives.

XIAP is highly expressed in different human tumour cells and cancer specimens from patients and plays an important role in conferring chemoresistance [33,90]. Because XIAP blocks apoptosis at the downstream effector phase, where multiple signalling events may converge, it represents an attractive molecular target for the design of new anti-cancer drugs [32–41,43]. Two broad approaches have been taken to develop clinical inhibitors of XIAP—antisense oligonucleotides, targeting the entire protein, and small molecule inhibitors, binding a single domain. Small molecule inhibitors offer the potential of more rapid inhibition of their target in vivo and more predictable duration of action [34,41]. Among the small molecule phytochemicals, the XIAP inhibitor embelin exhibited cytotoxic activity in various human tumoural cells, including breast cancer [83–86,91]. In addition, the withaferin-A induced cytotoxicity in human breast cancer cells was associated with suppression of XIAP protein [92] and berberine was shown to induce apoptosis in tumours, likely through the inhibition of XIAP [93]. The just mentioned molecular modelling of our new molecules binding to XIAP-BIR3 domain drove us to examine if they shared a similar activity with already described XIAP inhibitors. With the aim of understanding the role of XIAP in the cell death phenotype, we first determined if our cellular model is anyhow affected by XIAP depletion. Using the Lipofectamine reagent, MCF7 cells were transiently transfected with a XIAP-specific or a scrambled targeting siRNA. When treated with 50 nM of siRNA for 24 h, the protein levels of XIAP markedly decreased to ca. 45% compared to control siRNA transfected samples (Figure 12A) indicating a partial depletion of XIAP. In agreement with previous indications [94–96], the outcome in viability of XIAP knockdown in MCF7 cells, which showed a significant reduction (ca. 40%) in MTT absorbance upon depletion of XIAP (Figure 12B), led us to the conclusion that MCF7 cells depend on XIAP for survival since death mechanisms are neutralised by physiological levels of XIAP. We then tried to add clues on the involvement of XIAP in the cytotoxic effect of *Piper* genus-derived compounds. As shown in Figure 12B, XIAP downregulation in MCF7 cells significantly enhanced the toxicity, as measured by MTT absorbance, of 6 h administration of gibbilimbol B (30 μg/mL), eriopodol A (30 μg/mL) and erioquinol (10 μg/mL) indicating their combined action with XIAP siRNA in inhibiting cell viability. Since 100% knockdown was never achieved with siRNA technique (the absence of detectable XIAP after siRNA transfection, i.e., by XIAP siRNA at 100 nM for 24 h, paralleled the increase of cleaved-caspase 7 levels and the complete loss of MCF7 cell viability (Figure S6D and data not shown)), it is reasonable to assume that that the effects of gibbilimbol B/eriopodol A/erioquinol on the residual XIAP protein in the siRNA-treated cells further induced MCF7 cell death. On the other hand, similar results (additive effect) would be achieved if the compounds target cytotoxic pathways other than XIAP. However, although this is not a formal biological evidence, the simplest explanation of the combined action is a XIAP-mediated mechanism accounting for, at least in part, the cytotoxicity of our new compounds. Accordingly, the positive effects of gibbilimbol B and eriopodol A on caspase 7 activity robustly increased after XIAP silencing (Figure 12C). Downregulation of XIAP by siRNA is known to sensitise human breast cancer cells to death mediated by different chemical agents [94,97]. Finally, using real-time PCR and western blot assays to measure XIAP expression, we found that cell exposure to gibbilimbol B, eriopodol A, and erioquinol at increasing times did not significantly modify mRNA (Figure 12D) and protein levels (Figure 12E) of XIAP. Overall, our data exclude a role of gibbilimbol B/eriopodol A/erioquinol on the regulation of XIAP expression but rather are consistent with the antagonism of XIAP activity through their binding to XIAP-BIR3 domain.

**Figure 12.** *Cont*.

**Figure 12.** X-linked inhibitor of apoptosis protein (XIAP) as a molecular target of *Piper* genus-derived compounds. (**A**) Western blot analysis of XIAP in MCF7 cells both untransfected (CTRL, control) or transfected for 24 h with a XIAP-specific and scrambled targeting (scr) siRNA (50 nM). Lactate dehydrogenase (LDH) was used as internal standard. Low panel: densitometric analysis expressed as fold change of scr siRNA. Images and data are representative of three independent experiments. \*\*P < 0.001 relative to scr siRNA. (**B**) MCF7 cells were transfected for 24 h with a XIAP-specific or scr siRNA (50 nM) and then cultured in the absence (CTRL) and in the presence of 30 μg/mL gibbilimbol B/eriopodol A or 10 μg/mL erioquinol for 6 h, before 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. Results are expressed by setting the absorbance of the reduced MTT in the scr siRNA CTRL, as 100%. Data are representative of three independent experiments. \* *p* < 0.0001 relative to the respective scr siRNA. (**C**) Western blot analysis of cleaved-caspase 7 in MCF7 cells transfected for 24 h with a XIAP-specific or scr siRNA (50 nM) and then cultured in the presence of 30 μg/mL gibbilimbol B and eriopodol A for 6 h. LDH was used as internal standard. Images are representative of three independent experiments. (**D**) Real-time PCR and (**E**) Western blot analysis of XIAP mRNA and protein expression, respectively, in MCF7 cells treated for increasing times in the absence (CTRL) and in the presence of 30 μg/mL gibbilimbol B/eriopodol A or 10 μg/mL erioquinol. β-actin (PCR) and LDH (Western blot) were used as internal standards. PCR results are expressed as fold change of respective CTRL, set as 1. Images and data are representative of three independent experiments.

Since escape from apoptosis is one of the preeminent features of cancer, pharmacological interest in targeting endogenous apoptosis inhibitors, such as B-cell lymphoma (BCL)-2 and IAPs family members, has been constant [32–43,82,98,99]. The efforts, including clinical trials, directed towards identifying small molecules inhibitors of the BCL-2 family of proteins and promote apoptosis with the so-called BH3 mimetics, that mimic the action of certain BH3-only proteins [98], proved the releasing of "apoptosis brakes" as a winning strategy to induce primary cell death in cancer or to sensitise tumour to chemotherapy. Differently to BCL-2 family members, IAPs, and in particular XIAP, have a late role in the apoptotic timeline, and they target already active caspases to prevent cell death. The structural data surrounding the interaction between the BIR3 domain of XIAP and caspases suggest that small molecules that bind the BIR3 pocket of XIAP could mimic the action of Smac and inhibit the interaction between XIAP and caspase [34,41]. Interestingly its multi-domain structure makes XIAP a component of multiple cellular pathways, not only the ones leading to apoptosis. XIAP versatility has been highlighted in inflammation and inflammatory cell death, such as necroptosis [32,79–82].

Even though these aspects are yet to be completely elucidated, we suggest here - in addition to the widely described activity of XIAP inhibitors in apoptosis induction (gibbilimbol B and eriopodol A) —an interesting example of how the pharmacological targeting of XIAP-BIR3 domain can go beyond the simple induction of apoptosis—and extends its influence in modulating cell death signalling events other than caspase-activation (erioquinol). The relevance of non-apoptotic cell death in cancer treatment has recently gained interest as a means to simultaneously targetting tumours and enhancing the inflammatory response [100]; XIAP, in this context, is an interesting crossroad of pathways involved in both cell death and inflammation.

#### **3. Materials and Methods**

#### *3.1. Extraction and Isolation of Natural Compounds*

*P. eriopodon* was collected in Fusagasuga, in the Department of Cundinamarca (Colombia). The plant material was identified by Dr. Adolfo Jara Muñoz at Herbario Nacional Colombiano and a voucher specimen (COL516757) was deposited at the Instituto de Ciencias Naturales, Universidad Nacional de Colombia.

Dried and powdered leaves of *P. eriopodon* (1.14 Kg) were extracted exhaustively with ethanol 96% (3 × 5L) at room temperature. After filtration, the solvent was evaporated under reduced pressure, to yield 103.6 g of crude extract. The crude extract (100.0 g) was subjected to silica gel flash chromatography and eluted with a step gradient of toluene/ethyl acetate (0:100, 20:80, 40:60, 60:40, 80:20 and 0:100 (V/V)) to afford eight fractions. Fraction 1 (34.2 g) was chromatographed over silica gel, eluting with a mixture of a three-phase *n*-hexane/dichloromethane/ethyl acetate (25:70:5) solvent system to afford ten fractions (A to J). Fraction E (10.0 g) was chromatographed over Sephadex LH-20 (4.5 × 45 cm, *n*-hexane/chloroform/methanol, 2:2:1) to give six fractions (E1 to E6). In agreement with a previous report [29], compound **1** (7.93 g) was obtained from fraction E3, after column chromatography on Sephadex LH-20 (4.5 × 30 cm, *n*-hexane/acetone/methanol, 2:2:1). Fraction E4 (974.6 mg) was submitted to column chromatography on Sephadex LH-20 (4.0 × 20 cm, *n*-hexane/acetone/methanol, 2:2:1) to yield six fractions (E4.1 to E4.6). Compound **3** (33.2 mg) was obtained from fraction E4.4 (378.8 mg) through Sephadex LH-20 (2.0 × 25 cm, *n*-hexane/acetone/methanol, 2:2:1) and silica gel column chromatography eluted with *n*-hexane/acetone 8:2. Fraction E5 (2.19 g) was subjected to column chromatography on silica gel using a mixture of toluene/ethyl acetate (9:1) to afford ten fractions (E5.1 to E5.10). Fraction E5.3 (153.2 mg) was purified by flash chromatography to yield compound 4 (20.0 mg).

Fraction 2 (8.0 g) was submitted to silica gel column chromatography eluted with *n*-hexane/ethyl acetate 8:2, yielding seven fractions (K–Q). Fraction Q was subjected to flash chromatography eluted with dichloromethane/acetone (7:3) to yield six fractions (Q1–Q6). Fraction Q3 (1.44 g) was subjected to column chromatography over Sephadex LH-20 (4.0 x 20 cm, hexane/acetone/methanol, 2:2:1) to afford six fractions (Q3.1 to Q3.6). Fraction Q3.4 (239.4 mg) was chromatographed over Sephadex LH-20 (4.0 × 20 cm, hexane/chloroform/methanol, 2:2:1) and then purified by silica gel column chromatography eluted with *n*-hexane/acetone (7:3) to yield compound **5** (4.0 mg). Fraction Q6 (570.5 mg) was subjected to flash chromatography eluted with *n*-hexane/acetone 7:3 to afford seven fractions (Q6.1–Q6.7). Compound **2** (166.0 mg) was obtained from fraction Q6.1.

#### *3.2. General Chemical Methods*

Flash chromatography was carried out with silica gel (230–400 mesh; Merck, Darmastadt, Germany), column chromatography was performed using silica gel (70–230 mesh; Merck) and Shepadex® LH20 (Sigma-Aldrich, St. Louis, MO, USA), analytical thin layer chromatography was performed using precoated silica gel plates 60 F254 (0.25 mm, Merck). 1H and 13C NMR 1D and 2D (COSY, HMQC and HMBC) spectra, were recorded on an Avance 400 spectrometer (Bruker, Millerica, MA, USA) at 400 MHz for 1H and 100 MHz for 13C using the solvent peaks as internal references, the spectra were recorded

in CDCl3 and MeOD (Merck). High-resolution mass data were collected on an Accurate-Mass quadrupole Time-of-Flight (q-TOF) (Agilent Technologies, Santa Clara, CA, USA) mass spectrometer, ESI negative mode, Nebuliser 50 (psi), Gas Flow 10 L/min, Gas Temp 350 ◦C. Fragmentor 175 V, Skimmer 75 V, Vpp 750 V.

#### *3.3. Cell Culture and Chemicals*

Human U373 glioma, MCF7 breast cancer, A549 lung cancer and PC-3 prostate cancer cells were grown in Dulbecco's Modified Eagle Medium (DMEM), supplemented with 10% foetal bovine serum, 2 mM glutamine, 100 U/mL penicillin/streptomycin, at 37 ◦C in a humidified atmosphere containing 5% CO2 (logarithmic growth phase, routine passages every 3 days). The human breast epithelial cell line MCF10 was cultured in DMEM/F12 Ham's Mixture supplemented with 5% horse serum, epithelial growth factor 20 ng/mL, insulin 10 μg/mL, hydrocortisone 0.5 mg/mL, cholera toxin 100 ng/mL, and 100 U/mL penicillin/streptomycin. HUVEC were grown in EGM-2 Endothelial Cell Growth Medium-2 BulletKit (Lonza, Basel, Switzerland), according to the manufacturer's protocol.

Foetal bovine serum, horse serum, glutamine and penicillin/streptomycin were obtained from Euroclone (Milano, Italy). TMRM was purchased from ThermoFisher Scientific (Waltham, MA, USA) while necrostatin-1 and ferrostatin-1 were obtained from Santa Cruz Biotechnology (Dallas, TX, USA). Where not indicated, the reagents were purchased from Sigma-Aldrich.

#### *3.4. MTT Assay*

U373, MCF7, A549, PC-3, HUVEC, and MCF10 cell viability was determined by MTT assay using published protocols [101–105]. MTT absorbance was quantified spectrophotometrically using a Glomax Multi Detection System microplate reader (Promega, Milano, Italy).

#### *3.5. TUNEL Assay*

Using published protocols [106,107], MCF7 or U373 cells cultured in 120-mm coverslips were fixed in 4% paraformaldehyde in 0.1 M phosphate buffer (PB), pH 7.4, for 10 min. The TUNEL method (DeadEnd Fluorometric TUNEL System, Promega) was used to assay apoptosis, according to the manufacturer's protocol. DAPI (nuclei detection) staining was also performed.

#### *3.6. Immunofluorescence Microscopy Analysis*

Using published protocols [106,108], MCF7 or U373 cells cultured in 120-mm coverslips were fixed in 4% paraformaldehyde in 0.1 M PB, pH 7.4, for 10 min. Cells were pre-incubated for 1 h min with 5% of normal goat serum (Life Technologies, Monza, Italy) in 0.1 M PB (pH 7.4) containing 0.1% Triton X-100, before overnight incubation with the rabbit monoclonal anti-cleaved caspase 7 (Cell Signaling Technology, Danvers, MA, USA). In double-label immunofluorescence experiments, the mouse monoclonal anti-cytochrome c primary antibody (Cell Signaling Technology) was used in conjunction with the rabbit monoclonal primary antibody directed to COX IV (Cell Signaling Technology). For fluorescence detection, coverslips were stained with the appropriate Alexa Fluor secondary antibodies (Life Technologies) and mounted on glass slides in a ProLong Gold Antifade Mountant (Life Technologies). DAPI and/or fluorescein phalloidin (cytoskeleton detection) staining was also used. Cells were analysed with a DMI4000 B automated inverted microscope equipped with a DCF310 digital camera (Leica Microsystems, Wetzlar, Germany). When indicated, confocal imaging was performed with a TCS SP8 System (Leica Microsystems). Image acquisitions were controlled by the Leica Application Suite X.

#### *3.7. Annexin V Staining*

MCF7 cells were incubated with 5 μg/mL Annexin V-fluorescein isothiocyanate (FITC) to assess the phosphatidylserine exposure on the outer leaflet of the plasma membrane, and 5 μg/mL PI (DNA-binding probe) to exclude necrotic cells in binding buffer (10 mM HEPES, 140 mM NaCl, 2.5 mM CaCl2) [109]. Cell staining was analysed by Gallios Flow Cytometer (Beckman-Coulter, Brea, CA, USA) and the software FCS Express 4 (De Novo System, Portland, OR, USA).

#### *3.8. Western Blotting*

Using published protocols [107,110,111], MCF7 and MCF10 cells were homogenised in RIPA lysis buffer, supplemented with a cocktail of protease inhibitors (cOmplete; Roche Diagnostics, Milano, Italy). Equal amounts of proteins were separated by 4–20% SDS-polyacrylamide gel electrophoresis (Criterion TGX Stain-free precast gels and Criterion Cell system; Bio-Rad, Hercules, CA, USA) and transferred onto nitrocellulose membrane using a Bio-Rad Trans-Blot Turbo System. When indicated, the membranes were probed using the rabbit monoclonal anti-cleaved caspase 7 and anti-XIAP (Cell Signaling Technology) primary antibodies. After the incubation with the appropriate horseradish-peroxidase-conjugated secondary antibody (Cell Signaling Technology), bands were visualised using the Clarity Western ECL substrate with a ChemiDoc MP imaging system (Bio-Rad). To monitor for potential artefacts in loading and transfer among samples in different lanes, the blots were routinely treated with the Restore Western Blot Stripping Buffer (ThermoFisher Scientific) and re-probed with the goat anti-Lactate dehydrogenase (LDH)-A (Santa Cruz Biotechnology) and the mouse anti-vinculin primary antibodies. The stain-free gel was used as loading control as well. When appropriated, bands were quantified for densitometry using the Bio-Rad Image Lab software.

#### *3.9. Mitochondrial Membrane Potential Analysis*

Using published protocols [112], mitochondria of MCF7 and U373 cells were labeled using TMRM, a voltage-sensitive cationic lipophilic dye, partitioning and accumulating in the mitochondrial matrix based upon the Nernst equation. After treatments, cells were trypsinised, counted and incubated with 100 nM TMRM for 30 min at 37 ◦C. Fluorescence was measured by using a Glomax Multi Detection System microplate reader (Promega), excitation wavelength: 525 nm; emission wavelength: 580–640 nm). After background subtraction, the data were normalised on cell number.

#### *3.10. Measurement of ROS*

MCF7 or U373 cells cultured in 120-mm coverslips were exposed to 30 μM DCFH-DA (0.1 M PB, pH 7.4) and fixed in 4% paraformaldehyde for 20 min. For fluorescence detection, coverslips were mounted on glass slides and observed with a laser-scanning confocal microscope (TCS SP8 System and Application suite X, Leica Microsystems). DAPI and fluorescein phalloidin (nuclei and cytoskeleton detection, respectively) staining was also used.

#### *3.11. Molecular Modeling*

AutoDock4 was used to carry out the molecular docking. The Protein Data Bank crystallographic structure PDB 5C83 was considered as receptor model [87]. The preparation of the macromolecule was made with PyMOL (version 2.0, PyMol Molecular Graphics, Schrodinger, New York, NY, USA) System) and XIAP-BIR3 domain was selected as receptor [113]. Energy maps was established with Autogrid4 involving all atom types. After 25 million of energy evaluations in the binding pocket and using a grid of 50 × 50 × 50 points, all conformations of the ligand were clustered according to the energy and conformations. The docking results were visualised using the computational program Maestro 11.6. The molecular dynamics simulations were carried out with Desmond simulation package of Maestro (Desmond Molecular Dynamics System; D. E. Shaw Research, New York, NY, USA, 2016) using the OPLS 2005 force field parameters. A solvated system (TIP3P) and a predefined model for electrically neutral system (physiological concentrations of monovalent ions, NaCl 0.15 M) were used in an orthorhombic box and maintained at constant temperature of 300 K for all simulations. The dynamics simulations were analysed using the Simulation Interaction Diagram tool of Desmond

package, monitoring the behaviour and stability of simulations by RMSD of the ligand and protein atom positions in time.

#### *3.12. RNA Interference*

Gene silencing of XIAP inMCF7 cells was performed as previously published [106]. Briefly, according to the manufacturer's protocol, iBONI siRNA Pool (Riboxx, Radebeul, Germany) targeting human XIAP were mixed to Lipofectamine RNAiMax transfection reagent (Life Technologies). iBONI siRNA Pool negative control (Riboxx) (scrambled targeting siRNAs) was also used. The mix was added to cultured MCF7 cells at a siRNA concentration of 50 nM for 24 h.

#### *3.13. Real-Time PCR*

The analysis of mRNA expression was performed as previously described [106,114,115]. Briefly, total RNA from MCF7 cells was extracted with the High Pure RNA Isolation Kit (Roche Applied Science, Mannheim, Germany), according to the manufacturer's protocol. First-strand cDNA was generated from 1 μg of total RNA using iScript Reverse Transcription Supermix (Bio-Rad). Primer pairs (Eurofins Genomics, Milano, Italy) for XIAP (NM\_001167; forward ACCGTGCGGTGCTTTAGTT, reverse TGCGTGGCACTATTTTCAAGATA) and β-actin (NM\_001101; forward ATAGCACAGCCTGGATAGCAACGTAC, reverse CACCTTCTACAAT GAGCTGCGTGTG) were designed to hybridise to unique regions of the appropriate gene sequence. PCR was performed using SsoAdvanced Universal SYBR Green Supermix and the CFX96 Touch Real-Time PCR Detection System (Bio-Rad). The fold change was determined relative to the selected control sample after normalising to β-actin (internal standard) by the formula 2−ΔΔCT.

#### *3.14. Statistics*

Statistical significance of raw data between the groups in each experiment was evaluated using unpaired Student's *t*-test (single comparisons) or one-way ANOVA followed by the Newman-Keuls post-test (multiple comparisons). The IC50 and Emax concentration were determined by non-linear regression curve analysis of the concentration-effect responses. Potency values among concentration-response curves were compared with the F-test. Data belonging from different experiments were represented and averaged in the same graph. The GraphPad Prism software package (GraphPad Software, San Diego, CA, USA) was used. The results were expressed as means ± standard error of mean (SEM) of the indicated n values.

#### **4. Conclusions**

This study adds to the renewed biological interest in natural derived compounds, by presenting a chemical and biological characterisation of new small organic molecules derived from *Piper* genus plants. Following a recent preliminary report of gibbilimbol B as cytotoxic in breast cancer cell lines, we explored this observation by comparing it to similarly structured new molecules. Erioquinol that appeared to be the most potent compound versus gibbilimbol B and eriopodol A was listed as an intermediate. A more detailed investigation of the biological mechanism behind these molecules' activity in shaping cell viability revealed induction of caspase-dependent apoptosis following exposure of tumour cells to gibbilimbol B and eriopodol A and, interestingly, display of caspase-independent/non-apoptotic features in cell treated with erioquinol. In silico modelling and molecular approaches gave us a first preliminary insight into the molecular target of *Piper* genus compounds, the anti-apoptotic protein XIAP (Figure 13). Of note, an already identified XIAP inhibitor shared structural and binding similarities with them. The appeal of XIAP as a therapeutic target in cancer is not restricted to inhibition of apoptosis, but comprehends the regulation of other cellular physiological aspects, such as control of caspase-independent cell death. The molecular signature behind our observation opens important implications to further dissect the role of XIAP and for the development of novel XIAP antagonists for cancer treatment.

**Figure 13.** Schematic picture depicting cell death mechanisms of *Piper* genus-derived compounds. Escape of both intrinsic and extrinsic apoptosis is a common feature of cancer cells. (**A**) This hallmark is often carried out by overexpressing anti-apoptotic proteins, such as X-linked inhibitor of apoptosis protein (XIAP), which prevents the execution of apoptosis by binding of its baculovirus IAP repeat (BIR)

3 domain to already active initiator caspase 9. In order to counteract this resistance to cell death, several cancer pharmacological therapies have the aim of removing the 'molecular brakes' to apoptosis sensitising cancer cell to undergo loss of viability. The approach we described includes the use of three compounds from *Piper* genus plants which were predicted to bind XIAP-BIR3 domain. (**B**) Two of the compounds (gibbilimbol B and eriopodol A) were shown to induce a classical pro-apoptotic response, including mitochondrial outer membrane polarisation, release of cytochrome c, and subsequent activation of both initiator and effector caspases. (**C**) Despites triggering a similar response at the mitochondria level, erioquinol does not act through the apoptotic machinery, and results in a caspase-independent cell death characterised by cytoplasmic reactive oxygen species (ROS) accumulation.

*Cancers* **2019**, *11*, 1336

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2072-6694/11/9/1336/s1, Figure S1: NMR spectroscopy (400 MHz, CDCl3) and HRESIMS of compound 1, Figure S2: NMR spectroscopy (400 MHz, CDCl3) and HRESIMS of compound 2, Figure S3: NMR spectroscopy (400 MHz, CDCl3) and HRESIMS of compound 3, Figure S4: NMR spectroscopy (400 MHz, CDCl3) and HRESIMS of compound 4, Figure S5: NMR spectroscopy (400 MHz, CDCl3) and HRESIMS of compound 5, Figure S6: Western blot, Figure S7: Protein-ligand interactions fraction for evaluated ligands and X-linked inhibitor of apoptosis protein (XIAP) baculovirus IAP repeat (BIR)-3 domain during the molecular dynamics trajectory of 50 ns, Table S1: 1H NMR (400 MHz) data for compounds 1–4 in CDCl3 and compound 5 in MeOD, Table S2: 13C NMR (100 MHz) data for compounds 1–4 in CDCl3 and compound 5 in MeOD.

**Author Contributions:** D.M.: design and conception of the experimental plan and analysis work, isolation and characterisation of the compounds, acquisition analysis and interpretation of human cell data, in silico analysis, contribution to article writing. M.B., S.Z.: design, acquisition, analysis and interpretation of human cell data, contribution to the experimental plan and article writing. F.L.-V.: in silico analysis, contribution to the supervision of the experiments, interpretation and processing of the results. A.S-H., M.G., M.C., E.C. (Elisabetta Catalani): acquisition and processing of the results, contribution to the experimental plan. C.D.P., C.P., E.C. (Emilio Clementi): contribution to the design, conception and interpretation of human cell results, contribution to article writing. G.A., W.D., L.C.: supervision of the experiments and interpretation of the results, contribution to article writing. D.C.: conception and coordination of the experimental and analysis work, supervision of human cell experiments, analysis/interpretation and processing of data, article writing. All authors provided critical feedback, edited and approved the final manuscript.

**Funding:** The research has been supported by grants from the Italian Ministry of Education, University and Research: "PRIN2015" to E.Clem./D.C. and "Departments of Excellence-2018" Program (Dipartimenti di Eccellenza) to DIBAF (University of Tuscia, Viterbo, Italy) (Project "Landscape 4.0 - food, wellbeing and environment"). Authors also acknowledge the financial support provided by the Universidad Nacional de Colombia (Bogotá) and the Administrative Department of Science, Technology and Innovation from Colombia "COLCIENCIAS" (Grant No. 528-2011) to D.M.

**Acknowledgments:** We are grateful to Francesca Proietti Serafini (University of Tuscia, Viterbo, Italy) for the help with viability cell assays and to Debora Parolin (University of Milan, Milano, Italy) for HUVEC supply.

**Conflicts of Interest:** The authors declare no competing financial 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/).

*Article*

## **Molecular Mechanisms Underlying Yatein-Induced Cell-Cycle Arrest and Microtubule Destabilization in Human Lung Adenocarcinoma Cells**

#### **Shang-Tse Ho 1,2, Chi-Chen Lin 3, Yu-Tang Tung 4,5,6,\* and Jyh-Horng Wu 1,\***


Received: 30 July 2019; Accepted: 12 September 2019; Published: 17 September 2019

**Abstract:** Yatein is an antitumor agent isolated from *Calocedrus formosana* Florin leaves extract. In our previous study, we found that yatein inhibited the growth of human lung adenocarcinoma A549 and CL1-5 cells by inducing intrinsic and extrinsic apoptotic pathways. To further uncover the effects and mechanisms of yatein-induced inhibition on A549 and CL1-5 cell growth, we evaluated yatein-mediated antitumor activity in vivo and the regulatory effects of yatein on cell-cycle progression and microtubule dynamics. Flow cytometry and western blotting revealed that yatein induces G2/M arrest in A549 and CL1-5 cells. Yatein also destabilized microtubules and interfered with microtubule dynamics in the two cell lines. Furthermore, we evaluated the antitumor activity of yatein in vivo using a xenograft mouse model and found that yatein treatment altered cyclin B/Cdc2 complex expression and significantly inhibited tumor growth. Taken together, our results suggested that yatein effectively inhibited the growth of A549 and CL1-5 cells possibly by disrupting cell-cycle progression and microtubule dynamics.

**Keywords:** *Calocedrus formosana*; lung cancer; yatein; cell-cycle arrest; xenograft

#### **1. Introduction**

Natural products have been used for treating disease for thousands of years. More recently, natural products are being continuously developed for pharmaceutical applications, particularly for anticancer, antibacterial, and antiviral applications [1]. To date, explorative and mechanistic studies on new bioactive compounds are still ongoing.

Cells routinely come in contact with endogenous and exogenous stress stimuli, such as toxic chemicals, UV radiation, and reactive oxygen species. These stress stimuli damage chromosomes, affecting DNA replication and chromosome segregation [2]. Each phase of the cell cycle, including G0, G1, S, G2, and M phase, functions to maintain genetic material duplication and cell division. Abnormal expression of cell-cycle proteins disrupts the cell cycle and is thus closely associated with tumorigenesis [3]. Many natural products exhibit anticancer properties by interacting with cell-cycle proteins [4]. Thus, the identification of a new cell-cycle–targeting natural substance may provide new alternatives in cancer chemotherapy.

*Calocedrus formosana* is a valuable softwood species in Taiwan, which not only has high industrial economic value but also exhibits multiple bioactivities [5–14]. We previously found that the *C. formosana* extract and its active phytocompound, yatein, inhibited the growth of human lung adenocarcinoma A549 and CL1-5 cells by inducing caspase-related apoptosis [15]. However, whether yatein regulates the cell cycle in human lung adenocarcinoma remains unclear. To uncover the mechanisms of yatein-mediated human lung adenocarcinoma growth inhibition, we examined the effects of yatein on cell-cycle progression, tubulin dynamics, and in vivo tumor growth.

#### **2. Results**

#### *2.1. Yatein Induces Cell-Cycle Arrest at G2*/*M Phase and Enhances G2*/*M Phase-Related Protein Expression in Human A549 and CL1-5 Cells*

To elucidate the mechanism underlying the anti-lung adenocarcinoma effects of yatein, cell-cycle distribution was analyzed in the yatein-treated A549 and CL1-5 cells. We found that 5 μM yatein treatment induced cell-cycle arrest at G2/M phase in both cell lines (Figure 1). We further analyzed the kinetics of the effects of yatein on A549 and CL1-5 cells through flow cytometry (Figure 2). Compared with untreated cells, we found that more cells entered the G2/M phase at 6 and 12 h after yatein treatment in both cell types. Next, we evaluated the effects of yatein on G2/M arrest-related protein expression using western blot analysis (Figure 3). To this end, A549 and CL1-5 cells were treated with 5 μM yatein for 6 and 12 h, and the expression of Cdc2, Cdc25c, and cyclin B1 was analyzed (Figure 3). Cdc2, Cdc25c, and cyclin B1 are key regulators of the cell cycle (particularly in the G2/M phase). Our results revealed that 6 and 12 h of yatein treatment upregulated cyclin B1, but not Cdc2 and Cdc25c, expression in A549 and CL1-5 cells. However, yatein treatment showed an increasing trend of Cdc2 phosphorylation in both cell types. Notably, yatein-induced Cdc2 phosphorylation was higher at 6 h than at 12 h in both the cell types, indicating that Cdc2 was involved in G2/M phase regulation in A549 and CL1-5 cells at an early stage.

**Figure 1.** Effects of yatein treatment for 24 h with different concentrations on cell-cycle progression in A549 and CL1-5 cells. The results represent the mean ± SD (*n* = 3). Different letters indicate significant differences among each group in A549 and CL1-5 cells (*p* < 0.05).

**Figure 2.** Effect kinetics of 5 μM yatein treatment on cell-cycle progression in A549 and CL1-5 cells. The results represent the mean ± SD (*n* = 3). Different letters indicate significant differences among each group in A549 and CL1-5 cells (*p* < 0.05).

**Figure 3.** Expression of cell-cycle regulatory proteins in A549 and CL1-5 cells after yatein treatment (5 μM) for 6 h and 12 h. The bands were analyzed using the ImageJ software and normalized to β-actin expression. All data presented are representative of three independent experiments. The quantifications represent the mean ± SEM (*n* = 2-3). \* indicates a significant difference compared with the control group (*p* < 0.05).

#### *2.2. Yatein Induces DNA Damage through Activation of the ATM*/*ATR Pathway in Human A549 and CL1-5 Cells*

DNA damage induces cell-cycle arrest and apoptosis in cancer cells [16]. The ATM/ATR pathway is related to DNA damage process. To address whether yatein induced DNA damage in cells, we examined the effects of yatein treatment on the ATM/ATR pathway. We found that yatein treatment

showed an increasing trend of ATM and ATR phosphorylation level in A549 and CL1-5 cells for 6 h and 12 h treatments. However, ATM and ATR expression were not affected (Figure 4A). We next evaluated the expression and phosphorylation of Chk1 and Chk2 in A549 and CL1-5 cells after yatein treatment for 6 and 12 h. Our results revealed that yatein treatment showed an increasing trend of Chk1 and Chk2 phosphorylation level in A549 and CL1-5 cells. These results suggested that yatein induced DNA damage and altered cell-cycle distribution by activating the ATM/Chk2 and ATR/Chk1 DNA repair pathways in A549 and CL1-5 cells. In addition to the activation of the ATM/ATR pathway, we found that yatein treatment affected p53, Wee1, and 14-3-3 σ expression in the A549 cells (Figure 4B). Conversely, yatein did not upregulate the p53 and 14-3-3 σ expression in CL1-5 cells. However, Wee1 expression was increased after yatein treatment for 6 h in CL1-5 cells. These results indicated that the p53 pathway might play an additional role in yatein-induced growth inhibition in A549 cells but not in CL1-5 cells.

**Figure 4.** Expression of ATM, ATR, Chk1, and Chk2 (**A**), and p53 related proteins (**B**) in A549 and CL1-5 cells after yatein treatment (5 μM) for 6 and 12 h. All data presented are representative of two to three independent experiments. The quantifications represent the mean ± SEM (*n* = 2–3). \* indicates significant differences compared with the control group in A549 and CL1-5 cells (*p* < 0.05).

#### *2.3. Yatein Influences Microtubule Dynamics in Human A549 and CL1-5 Cells*

Apart from DNA damage, inhibition of microtubule dynamics is another mechanism that induces G2/M cell-cycle arrest. For instance, taxol, a well-known anticancer drug, arrests cells in the G2/M phase by disrupting microtubule polymerization [17]. We assessed microtubule dynamics in yatein-treated A549 and CL1-5 cells using confocal microscopy and western blot analysis. Our confocal microscopy results showed that 5 μM yatein treatment caused a diffusion of green fluorescence (tubulin) in the yatein-treated A549 and CL1-5 cells compared with control cells within 6 h (Figure 5A). In addition, as shown in Figure 5B, yatein decreased tubulin polymerization in a dose-dependent manner in both A549 and CL1-5 cells after 24 h treatment. These results revealed that the effects of yatein on microtubule dynamics were similar to those of colchicine, a microtubule-depolymerizing agent (Figure 5C). Consistently, yatein-treated cells exhibited a similar pattern of tubulin distribution to that of colchicine-treated cells. By contrast, taxol-treated cells demonstrated a higher level of tubulin polymerization compared with the untreated cells. Taken together, our results indicated that yatein and colchicine affected microtubule dynamics by inhibiting tubulin polymerization.

**Figure 5.** (**A**) Confocal images of tubulin expression of A549 (upper) and CL1-5 (lower) cells after yatein, colchicine, and taxol treatment for 6 h. Scale bar = 100 μm. (**B**) Western blot analysis of tubulin expression of A549 and CL1-5 cells after treatment with various concentrations of yatein for 24 h. (**C**) Western blot analysis of tubulin expression of A549 and CL1-5 cells after yatein and positive control treatment for 24 h. V: Vehicle control, C: 10 μM colchicine, T: 100 nM taxol, Y: 5 μM yatein. The quantitative values are the ratio of polymerization tubulin/free tubulin in different treatment groups.

#### *2.4. Yatein Exhibits In Vivo Antitumor E*ff*ects in a Xenograft Mouse Model*

To validate the growth inhibitory effects of yatein in an in vivo context, we applied a xenograft mouse model using A549 cells. To this end, NOD/SCID mice were inoculated with A549-luc cells for 10 days and were then given 20 mg/kg yatein (i.p.) five times per week for 42 days. Animal body weight, food intake, and tumor growth were monitored and quantified during the experiment. Our results showed that the body weight and food intake of the control and yatein-treated mice did not differ

(Figure 6A,B), suggesting that yatein was well tolerated in the mice. We found that the tumor volumes of the control mice (41.3 mm3) and yatein-treated mice (38.5 mm3) were similar in the first 14 days (Figure 6C). However, within the 21-42 days (end of experiment) time window, tumor growth in the yatein-treated mice was significantly slower compared with the control mice (*p* < 0.05). Consistently, In Vivo Imaging System (IVIS) analysis showed that the tumor tissue luminescence in the yatein-treated group (16.5 <sup>×</sup> 105 photons/s) was lower than that in the vehicle control group (24.0 <sup>×</sup> 10<sup>5</sup> photons/s). To confirm that yatein exhibited the same antitumor mechanisms in vivo as observed in vitro, we evaluated cyclin B1 expression and Cdc2 phosphorylation in the tumor tissue in the vehicle control and yatein-treated groups (Figure 6D). We found that both cyclin B1 expression and Cdc2 phosphorylation moderately increased in the yatein-treated group compared with the control group (Figure 6D). Taken together, these results suggested that the induction of cyclin B/Cdc2 complex expression and activation were associated with the antitumor effects of yatein in vivo.

**Figure 6.** In vivo antitumor effects of yatein. (**A**) Body weight, (**B**) daily diet intake, and (**C**) tumor size and IVIS images of A549-luc cell xenograft mice treated with vehicle or 20 mg/kg of yatein during the experimental period. The results represent the mean ± SEM (*n* = 5). (**D**) Western blot analysis of tumor tissue from the A549-luc xenograft mice (values were mean ± SEM, *n* = 3).

#### **3. Discussion**

According to our previous study [15], the IC50 values for the 24 h yatein treatment in the A549 and CL1-5 cells were 10.0 and 2.1 μM, respectively. In addition, yatein exhibited excellent growth inhibitory effect in the A549 (IC50 values = 3.5 μM) and CL1-5 cells (IC50 values = 1.9 μM) after 72 h treatment. Interestingly, yatein was unable to inhibit human nasopharyngeal carcinoma (HONE-1) and human gastric cancer (NUGC) cells growth at a concentration of 50 μg/mL (- 125 μM) [10]. In the present study, we found that yatein induced G2/M cell-cycle arrest in A549 and CL1-5 cells at the first 24 h treatment. Additionally, treatment with yatein for 48 h induced a dose-dependent increase in both early and late stage apoptosis in the A549 and CL1-5 cells [15]. As previously reported, some lignan compounds can inhibit cancer cell growth by inducing cell-cycle arrest. For example, vitexin compound-1 (VB1), a lignan isolated from a plant used in traditional Chinese medicine, namely *Vitex negundo*, inhibits the growth of MDA-MB-435 and SMMC-7721 cells by inducing G2/M phase cell-cycle arrest after 24 h treatment at a concentration of 10 μM [18]. Similarly, benozofuran lignan shows a dose- and time-dependent induction of G2/M cell-cycle arrest in Jurkat T-cells [19]. Additionally, the mechanism underlying the effects of yatein on cell-cycle regulation was examined in the present study. Cyclin-dependent kinases (CDKs) collaborate with specific cyclins to tightly regulate cell cycle and cell division. Thus, the CDK/cyclin complex plays a key role in cell-cycle progression [20]. Our results indicated that yatein treatment showed an increasing trend of cyclin B1 expression and Cdc2 phosphorylation level. Aberrant activation of cyclin B1/p-Cdc2 activity is closely associated with mitotic catastrophe. Mitotic catastrophe is a type of cell death that results from mitotic regulation dysfunctions and can be induced by chemical and physical stresses [21]. Liu et al. [22] reported that expression of cyclin B1 and level of Cdc2 phosphorylation were increased in malignant glioma cells treated with a synthetic quinazolinone analog (MJ-66), suggesting that MJ-66 induced G2/M arrest and mitotic catastrophe in malignant glioma cells. Subramaniam et al. [23] reported that curcumin-induced mitotic catastrophe coupled with increased expression of cyclin B1 and Cdc-2 in MiaPaCa-2 cells. Therefore, we can speculate the cytotoxic effects of yatein on A549 and CL1-5 cells might be partially due to the induction of mitotic catastrophe. The correlation between yatein-induced cell death and mitotic catastrophe still need to be investigated in a future study.

DNA damage is vital in the regulation of cell cycle and apoptosis. In this study, we found that yatein induced cell-cycle arrest by inducing DNA damage in A549 and CL1-5 cells. In general, ATM and ATR kinases are the initiating kinases of the DNA damage response (DDR) signaling pathway that is activated by DNA damage or DNA replication stress [24]. ATM is activated by DNA double-strand breaks, while ATR is activated by various factors related to DNA damage [25]. According to previous studies, several natural products, including berberine, curcumin, and sinularin, induced DNA damage and G2/M arrest, but not S phase arrest, in various cancer cells [26–28]. The results revealed that these natural products induced DNA damage and arrested cells at G2/M phase. At the same time, the ATM pathway was up-regulated [26–28]. After ATM activation, several proteins involved in the regulation of ATM on DNA repair, cell-cycle arrest, apoptosis, and other downstream processes, such as Brca1, Chk2, and p53, are phosphorylated [29]. Our results revealed that yatein induced the expression of ATR/Chk1 and ATM/Chk2 in A549 cells. Notably, only ATR/Chk1 expression was upregulated after yatein treatment in CL1-5 cells. We found that the expression and phosphorylation of p53, Wee1, and 14-3-3 σ expression were induced after yatein treatment in A549 cells, suggesting that yatein activated p53-mediated signaling pathway to inhibit growth in A549 cells. Yatein did not affect the expression of p53 and its related proteins in CL1-5 cells; this was because the CL1-5 cell line carries a p53 mutation, which has been implicated in more than 50% of all cancer cases [30]. The p53 mutation causes a defective G1 checkpoint in cancer cells that results in increased DNA damage at the G2 checkpoint compared with noncancer cells. Wee1 is a key protein that is closely associated with G2 checkpoint abrogation and mitotic catastrophe [30,31]. Accordingly, we found that yatein treatment increased Wee1 expression in CL1-5 cells, indicating that Wee1 played a crucial role in the regulation of G2/M cell-cycle distribution in CL1-5 cells in the context of yatein treatment. On the other hand, we also

found yatein is able to induce ROS production in A549 and CL1-5 cells in our previous study [15], suggesting that yatein may induce the oxidative DNA damage. Moreover, our previous results also revealed that the CL1-5 cells were more sensitive for the ROS production than the A549 cells after yatein treatment [15].

Additionally, we showed that yatein affected the microtubule assembly in both A549 and CL1-5 cells after 6 h treatment (Figure 5A) and 24 h treatment (Figure 5B and 5C). The induction of mitotic arrest is associated with dysfunctional microtubule dynamics [32]. Microtubules play vital roles in cell proliferation, trafficking, signaling, and migration [33]. Considering the importance of microtubule dynamics, several small microtubule-targeting molecules have been designed and used as anticancer drugs [30]. Tubulin has three major binding sites: taxane, vinca, and colchicine domains. In this study, we found that yatein exhibited similar effects to that of colchicine, implying that yatein inhibited tubulin polymerization. These results suggested that tubulin polymerization inhibition partially contributed to the ability of yatein to arrest A549 and CL1-5 cells at the G2/M phase.

The in vivo growth inhibitory effects of yatein on A549 cells were elucidated in this study. Our results revealed that yatein treatment significantly suppressed tumor growth in mice without affecting their body weight and food intake. Additionally, yatein treatment increased cyclin B1 expression and Cdc2 phosphorylation in vivo, indicating that the induction of mitotic catastrophe was involved in the anticancer mechanism of yatein in vivo. Taken together, we found that yatein affected cell-cycle progression and microtubule dynamics, induced DDR, and exhibited anticancer properties in vivo (Figure 7).

**Figure 7.** Proposed mechanism of the inhibitory effects of yatein on lung adenocarcinoma cells.

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

#### *4.1. Preparation of Yatein*

The phytocompound, yatein, was isolated from *C. formosana* leaves extracts. In brief, *C. formosana* leaves were extracted using methanol at room temperature (RT) for one week (twice) to obtain a methanolic extract. The dried samples were further divided to *n*-hexane, ethyl acetate (EtOAc), *n*-butanol, and H2O fractions using liquid–liquid partition. The *n*-hexane fraction was further

fractionated into ten subfractions using normal phase column chromatography (Geduran Si-60, Merck, Darmstadt, Germany). Yatein was isolated and purified from the subfraction 4 by semipreparative high-performance liquid chromatography using a PU-2080 pump (Jasco, Tokyo, Japan) equipped with an RI-2031 detector (Jasco) and a 5 μm Luna silica column (250 mm × 10.0 mm internal diameter; Phenomenex, Torrance, CA, USA). The mobile phase consisted of 30% of EtOAc and 70% of *n*-hexane (*v*/*v*), and the flow rate was 4 mL/min. The retention time of yatein in HPLC analysis was 18.0 min. The purity and the structure elucidation of yatein was conducted by 1H and 13C NMR, and all spectrum data were consistent with literature [34].

#### *4.2. Cell Culture*

Human A549 cell line was purchased from Bioresource Collection and Research Center (BCRC 60124) and cultured in RPMI-1640 (Gibco, Gran Island, NY, USA) supplemented with 10% (*v*/*v*) fetal bovine serum (Gibco) and 1% (*v*/*v*) antibiotic–antimycotic agent (Gibco). CL1-5 cell line was provided by Dr. Jeremy J.-W. Chen (National Chung Hsing University, Taichung, Taiwan) and cultured in Dulbecco's modified Eagle's medium (Gibco) supplemented with 10% (*v*/*v*) fetal bovine serum (Gibco) and 1% (*v*/*v*) antibiotic–antimycotic agent (Gibco). The cells were incubated in a 37 ◦C humidified incubator containing 5% CO2.

#### *4.3. Cell-Cycle Distribution Analysis*

The A549 and CL1-5 cells were seeded onto a 6-well plate at a density of 1 <sup>×</sup> 105 cells/well and incubated overnight. The cells were then treated with various concentrations of the test samples for the indicated durations. After treatment, the cells were collected and mixed with ice-cold 75% ethanol and then fixed overnight at −20 ◦C. Then, the cells were washed twice using phosphate-buffered saline (PBS), and then incubated with 200 μL of propidium iodide (PI, BD Bioscience, Franklin Lakes, NJ, USA) solution (containing 1 mg/mL PI, 2 mg/mL RNase A, and 0.5% Triton X-100) for 30 min at RT in the dark. After 30 min, the PI-stained cells were immediately analyzed using a flow cytometer (Accuri 5, Accuri Cytometers, Inc., Ann Arbor, MI, USA). The data were analyzed using the C6 Accuri system software (Accuri Cytometers, Inc.).

#### *4.4. Isolation of Microtubule Proteins*

The A549 and CL1-5 cells (1 <sup>×</sup> 106 cells) were seeded onto culture dishes (10 cm) overnight and treated with 5 μM yatein. The cells were harvested through trypsinization and washed with PBS. After centrifugation, the supernatant was removed and the cell pellets were mixed with 200 μL of microtubule stabilizing buffer (1 mM MgCl2, 2 mM Tris-HCl, 2 mM EGTA, and 0.5% Triton-100 in ddH2O) and incubated at RT for 20 min. Subsequently, the mixture was centrifuged at 12,000× *g* for 10 min (4 ◦C) to obtain the supernatant (monomer tubulin fraction). The remaining cell pellets were washed using the microtubule stabilizing buffer and lysed in a radioimmunoprecipitation assay buffer containing 10% proteinase inhibitor and 10% phosphatase inhibitor (Sigma-Aldrich, St. Louis, MO, USA) at 4 ◦C for 30 min to obtain the polymer tubulin fraction. The monomer tubulin and polymer tubulin fractions were transferred into microtubes (1.5 mL) and stored at −20 ◦C until further analysis.

#### *4.5. Western Blot Analysis*

The expression of proteins in cells were determined using western blot analysis as previously reported [15]. In this study, the primary antibodies were anti-14-3-3 σ (Santa Cruz, Dallas, TX, USA), anti-β-actin (Santa Cruz), anti-ATM (Santa Cruz), anti-ATR (Santa Cruz), anti-Cdc2 (Abcam, Cambridge, MA, USA), anti-Cdc25c (Santa Cruz), anti-Chk1 (Santa Cruz), anti-Chk2 (Santa Cruz), anti-cyclin B1 (Abcam), anti-p53 (Cell Signaling Technology Inc, Danvers, MA, USA), anti-phospho-p53 Ser 15 (Cell Signaling Technology Inc), anti-phospho-ATM (Santa Cruz), anti-phospho-ATR (Santa Cruz), anti-phospho-Cdc2 (Santa Cruz), anti-phospho-Chk1 (Santa Cruz), anti-phospho-Chk2 (Santa Cruz), anti-β-tubulin (Abcam), and anti-Wee1 (Santa Cruz) antibodies. An enhanced chemiluminescence

(ECL, Sigma-Aldrich) system was used for developing signals of the blots, which were analyzed using a LAS3000 system (Fujifilm, Tokyo, Japan).

#### *4.6. Immunofluorescence*

The A549 and CL1-5 cells (1 <sup>×</sup> 105 cells) were seeded on a chamber slide for 24 h and treated with colchicine (10 μM), taxol (100 nM), and yatein (5 μM) for 6 h. The cells were washed with PBS and fixed with 1% paraformaldehyde (Sigma-Aldrich) at RT for 30 min. The paraformaldehyde was removed, and the cells were washed thrice with PBS. Next, the cells were blocked using 1% BSA in PBS at 4 ◦C overnight. After they were washed with PBS, the cells were incubated with the anti-β-tubulin antibody (1:200 in 1% BSA solution) at 4 ◦C overnight. After washing with PBS, an anti-rabbit-FITC antibody was added to the cells and then incubated overnight. Then, the antibody was rinsed off and the cells were washed thrice using PBS. 4,6-Diamidino-2-phenylindole was added and the cells were incubated for 15 min at RT in the dark. A Leica TCS SP2 confocal spectral microscope (Buffalo, NY, USA) was used to observe immunofluorescence staining of microtubule dynamics.

#### *4.7. In Vivo Antitumor Activity*

The procedures involving animals were performed according to the guidelines of the Institutional Animal Care and Use Committee of National Chung Hsing University (IACUC no. 107-127). The A549-luc cells were mixed with Matrigel (Sigma-Aldrich) at a 1:1 ratio. The cells were injected subcutaneously into the back of nonobese diabetic and severe combined immunodeficiency (NOD/SCID) mice (male, 6–8 weeks old) at a density of 3.5 <sup>×</sup> 106 cells/mouse. Tumors were allowed to grow for 10 days and were then treated with an intraperitoneal (i.p.) injection of either 0.5% DMSO in ddH2O to the mice in the vehicle control group (*n* = 5) or 20 mg/kg of yatein (dissolved in 0.5% DMSO in ddH2O) to the mice in the yatein group (*n* = 5). The tumor-bearing mice were sacrificed after 42 days. Tumor volume was measured five times/week and calculated using the following formula: Length <sup>×</sup> width <sup>×</sup> thickness <sup>×</sup> 0.5 (mm3). An IVIS (Caliper lifescience IVIS Spectrum CT) was used to analyze the luminescence of tumor tissue.

#### *4.8. Statistical Analysis*

Data are expressed as mean ± standard deviation (SD) or mean ± standard error of the mean (SEM). Statistical analysis was performed using the shuffle test or non-parametric Kruskal-Wallis test with Dunn's post hoc tests. A *p* of < 0.05 was considered statistically significant.

#### **5. Conclusions**

The present study demonstrated that yatein suppressed lung adenocarcinoma cancer cells growth by inducing cell-cycle arrest, mitotic catastrophe, and microtubule depolymerization. Yatein induced G2/M arrest by upregulating the expression of cyclin B1 and Cdc2 phosphorylation in lung adenocarcinoma cancer cells. In addition, mitotic catastrophe and microtubule depolymerization were involved in yatein-mediated lung adenocarcinoma cancer cells growth inhibition. Furthermore, we confirmed the in vivo antitumor effects of yatein using a xenograft mouse model. These findings provide novel insights into the in vitro and in vivo antitumor efficacy of yatein and demonstrate the potential of this phytocompound as an anticancer lead compound for lung adenocarcinoma cancer treatment.

**Author Contributions:** Conceptualization, S.-T.H., Y.-T.T., and J.-H.W.; methodology, S.-T.H., Y.-T.T. and J.-H.W.; formal analysis, S.-T.H. and C.-C.L.; resources, Y.-T.T. and J.-H.W.; writing—original draft, S.-T.H.; writing—review and editing, C.-C.L., Y.-T.T. and J.-H.W.; supervision, Y.-T.T. and J.-H.W.

**Funding:** This research work was supported in part by the Ministry of Science and Technology of Taiwan and Taipei Medical University under Grant Nos. MOST 107-2628-H-038-001-MY3 and 106-6820-001-112, respectively.

**Conflicts of Interest:** The authors declare that there is no conflict of interest regarding the publication of this paper.

#### **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*
