Evaluation of uttroside B, a saponin from Solanum nigrum Linn, as a promising chemotherapeutic agent against hepatocellular carcinoma

We report, for the first time, the remarkable efficacy of uttroside B, a potent saponin from Solanum nigrum Linn, against liver cancer. The compound has been isolated and characterized from the leaves of Solanum nigrum Linn, a plant widely used in traditional medicine and is a rich resource of several anticancer molecules. Uttroside B, that comprises of β-D-glucopyranosyl unit at C-26 of the furostanol and β-lycotetraosyl unit at C-3, is ten times more cytotoxic to the liver cancer cell line, HepG2 (IC50: 0.5 μM) than sorafenib (IC50: 5.8 μM), the only FDA-approved drug for liver cancer. Moreover, it induces cytotoxicity in all liver cancer cell lines, irrespective of their HBV status, while being non-toxic to normal immortalized hepatocytes. It induces apoptosis in HepG2 cells by down-regulating mainly the activation of MAPK and mTOR pathways. The drastic reduction in HepG2-xenograft tumor size achieved by uttroside B in NOD-SCID mice and substantiation of its biological safety through both acute and chronic toxicity studies in Swiss albino mice warrants clinical validation of the molecule against hepatic cancer, for which, the chemotherapeutic armamentarium currently has limited weapons.

Triterpene and steroid glycosides, commonly referred to as saponins, are isolated primarily from the plant kingdom and exert a wide range of pharmacological properties owing to their large structural diversity 1 . A vast array of saponins have been reported to exhibit antitumor effect against cancer cells originating in different anatomical sites. In natural product research, analysis of the chemotherapeutic efficacy of saponins against various cancer cells is often confined to in vitro analysis and structure elucidation 2 . Various plant species of Solanum genera have been reported to have considerable amounts of saponins, which exhibit potent anticancer activity against different cancer cell lines [2][3][4] . Solanum nigrum Linn, commonly known as black nightshade, is a medicinal plant member of Solanaceae family, widely used in many traditional systems of medicine 4 . Alcoholic extract of the whole plant has been reported to contain various steroidal saponins, which induce cytotoxicity in different cancer cell lines [5][6][7] . Two furostanol saponins, uttroside A and B have been reported from the stems and roots of S. nigrum Linn 8 . In the present study, we isolated and characterized uttroside B from the leaves of S. nigrum Linn and found that the compound exhibits maximum cytotoxicity against liver cancer cells and is ten times more potent than sorafenib, the only FDA-approved drug for liver cancer. Though the cytotoxicity of the compound has been reported in cancer cells of other origins 9,10 , this is the first study evaluating its chemotherapeutic efficacy and exploring the molecular mechanisms involved. We have also validated the anticancer potency of the compound in vivo using HepG2-xenograft model in NOD-SCID mice and have confirmed its biological safety, both by in vitro and in vivo studies. 1 Division of Cancer Research, Rajiv Gandhi Centre for Biotechnology, Thiruvananthapuram-695014, Kerala, India. 2 Chemical Sciences and Technology Division, CSIR-National Institute for Interdisciplinary Science and Technology, Thiruvananthapuram-695019, Kerala, India. 3 Department of Pathology, Government Medical College, Thiruvananthapuram-695011, Kerala, India. 4 Division of Organic Chemistry, CSIR-National Chemical Laboratory, Dr. Homi Bhabha Road, Pune-411008, India. * These authors contributed equally to this work. Correspondence and requests for materials should be addressed to R.S.L. (email: ravishankar@niist.res.in) or R.J.A. (email: rjanto@rgcb.res.in)

Results
The methanolic extract of the leaves of S. nigrum Linn contains a bioactive mixture of a saponin and proline (SP), which on further purification yields uttroside B. We conducted a polarity-graded successive extraction of the leaves of S. nigrum Linn using hexane, dichloromethane, ethyl acetate and methanol and the cytotoxic effect of the extracts were screened against a panel of human cancer cell lines of different origin by MTT assay. The methanolic extract emerged to be the most cytotoxic and the liver cancer cell line, HepG2 exhibited maximum sensitivity (IC50-37.5 μ g/ml) towards the extract followed by the cervical cancer cell line, HeLa (124.2 μ g/ml). Later on, the most active methanolic extract was selected for further purification and the most sensitive cell line to the extract, HepG2, was selected for further screening (Supplementary Figure S1A). The methanolic extract (6.3 g) was subjected to fractionation by column chromatography. Among the column fractions subjected to cytotoxicity analysis, 'fraction f ' turned out to be the most effective (Supplementary Figure S1B).
Characterization of the isolated saponin by Mass and NMR spectroscopy revealed that the bioactive compound is uttroside B, a furostanol glycoside. Further, we attempted for the structural elucidation of the saponin using 1 H and 13 C-NMR experiments, which were initially performed in CD 3 OD (deuterated methanol) solvent. The key information pertaining to steroidal furanose ring include H-21 methyl group at δ H 0.99 ppm (3H, d, J = 7 Hz), and hemiketal carbon C-22 at δ C 112.5 ppm. Owing to complex pattern of signals arising due to sugars in the region between 3 to 4 ppm, the isolated saponin was peracetylated ( Fig. 2A). 1 H and 13 C-NMR experiments of peracetylated product of saponin were performed in CDCl 3 (deuterated chloroform) solvent. Surprisingly, after acetylation the H-21 methyl group exhibited a downfield shift at δ H 1.57 ppm (3H, s), and H-17 at δ H 2.45 ppm (( Fig. 2B; 1H, d, J = 9.8 Hz, Supplementary Table S1). In 13 C-NMR, the hemiketal carbon peak disappeared and two additional peaks appeared at δ C 103.7 and 151.7 ppm indicating carbons C-20 and C-22, respectively (Fig. 2B). From NMR data, it is clear that a new olefinic bond in the furanose ring formed due to loss of a water molecule during acetylation (SI methods). On the basis of the aforementioned information and available literature, the isolated saponin was identified as uttroside B (Fig. 2C). The HRESIMS data of the isolated sapo- ]. Further, we evaluated the cytotoxicity of uttroside B in different hepatocellular carcinoma cells (HepG2, Hep3B, SK-Hep1, and Huh-7) and compared with that of normal immortalized hepatocytes (Chang liver) using MTT assay. While different liver cancer cell lines did not exhibit a drastic difference for their sensitivity towards uttroside B (IC50 400-600 nM), 70% of the normal immortalized hepatocytes (Chang Liver) were found viable even at concentrations as high as 1250 nM [ Fig. 3(B)]. For further studies, HepG2, the most sensitive among the panel of liver cancer cell lines was selected. The cytotoxicity of uttroside B was compared with that of sorafenib, the only FDA approved drug for liver cancer. Surprisingly, uttroside B was observed almost 12 fold efficacious than sorafenib in inducing cell death [ Fig. 3 HepG2 cells were examined for morphological changes by Phase contrast microscopy, 72 h after treatment with uttroside B. Nuclear condensation, membrane blebbing and formation of apoptotic bodies, which are characteristics of apoptosis was observed in a dose dependent manner in the uttroside B-treated HepG2 cells [ Fig. 3(D)]. Further we performed clonogenic assay, which is an in vitro assay routinely used as a technique for studying the effectiveness of specific agents on the survival and proliferation. Interestingly, uttroside B induced a drastic dose-dependent reduction in both number and size of the colonies formed, clearly demonstrating its anti-clonogenic potential [ Fig. 3E,F].

Uttroside B induces apoptosis in HepG2 cells, but does not influence any phase of cell cycle.
Uttroside B was found to induce the cleavage of procaspase 9, into its active fragments (p35/37 and p17) in a concentration-dependent manner [ Fig. 4(A)]. Similarly, the cleavage of procaspase 8 to its active fragments (p43/41) [ Fig. 4(B)], and procaspase 7 to its active fragment (p-20) were also enhanced due to uttroside B The error bars represent ± S.D. Two-way ANOVA followed by Tukey's post hoc t test analysis was used for statistical comparison between different groups. ***P ≤ 0.001; # P ≤ 0.001, ns non significant. Symbol asterisk (*) represents statistical significance between control and treatment groups where as hash (#) represents statistical significance between different treatment groups.

Uttroside B inhibits MAPK and mTOR signaling, but does not affect NF-κB in HepG2 cells.
To study the effect of uttroside B on some of the major signalling events associated with liver cancer progression, the cells were treated with uttroside B and the nuclear extracts and whole cell lysates (WCL) were prepared. Though there was a constitutive activation of NF-κ B in HepG2 cells, uttroside B did not produce any significant down-regulation of the same as assessed by the electrophoretic mobility shift assay (EMSA, Fig. 5A). Activation status of Akt was also assessed in these cells by Western Blot. However, no basal activation of Akt was observed in HepG2 and hence, no significant role could be expected for uttroside B (Fig. 5B). Interestingly, the basal activation of MAPK pathway was evident in these cells, among which the activation of p-JNK was significantly    2448 and 2481 phosphorylation sites, a read out of p-mTOR activation was observed in HepG2 cells. Interestingly, a time dependent decrease in p-mTOR activation was noticed on uttroside B exposure [ Fig. 5(E)].

Uttroside B is pharmacologically safe as validated by acute and chronic toxicity models.
To rule out the possibility of any toxic side effects due to uttroside B, a detailed toxicological evaluation of the compound was conducted in Swiss albino mice as shown in Fig. 6(A). Since the IC50 of uttroside B was 500 nM, the corresponding dose for animal studies was calculated as 10 mg/Kg using the formula, Dose in vivo = XY, where X = IC50 (in vitro) x molecular weight of the compound and Y = body weight x water content of the animal (0.6 for mice). Another group with 5 times the effective value (50 mg/kg) was also included. The group of mice, which received 10 mg/kg and 50 mg/kg doses of uttroside B did not exhibit any abnormal behaviour and did not show any deviation in serum AST (Aspartate aminotransferase), ALT (Alanine transferase) and ALP (Alkaline phosphatase) [ Fig. 6(B)] which are clear markers of abnormality in liver function and in the level of blood urea nitrogen (BUN), which is a marker of nephrotoxicity. In chronic cytotoxic study, the haematotoxicity, hepatotoxicity and nephrotoxcity due to uttroside B was assayed by analyzing the level of total and differential count of WBC and serum levels of AST, ALT, ALP and BUN, respectively in control and treated mice [Fig. 6C,D]. There was no significant difference in any of these parameters from their normal range, indicating that uttroside B is pharmacologically safe and non-toxic. The histopathological analysis of liver tissue isolated from mice administered with uttroside B at the dose same as that used for the tumor reduction studies (10 mg/kg), did not bring in observation of any morphological change characteristic of toxicity. In the liver tissues of acute toxicity study, where a five fold increase in dose compared to that of treatment was used (50 mg/kg), micro vesicular fatty changes were observed, which are reversible changes associated with any chemotherapy. These observations confirm that uttroside B can be safely used as a chemotherapeutic drug for being validated through clinical trials [ Fig. 6(E)].

Uttroside B inhibits development of hepatic xenograft tumor in NOD-SCID mice. Further, we
attempt to validate the anticancer efficacy of uttroside B against hepatic cancer, using an in vivo HepG2-xenograft model in NOD-SCID mice. The HepG2 cells suspended in matrigel were subcutaneously injected in the flank region of the mice. The experimental plan is schematically represented in Fig. 7(A). Uttroside B, dissolved in PBS, was administered after 15 days of tumor cell implantation when the tumor attained a size of 50-100 mm 3 approximately. Uttroside B (10 mg/kg bw) was injected intraperitoneally, thrice weekly for four weeks. Group 1 comprised of control animals, which did not receive any treatment. The size of the tumor was measured using Vernier callipers every week and the corresponding tumor volume was calculated. The volume of tumor developed in animals that received uttroside B was significantly low compared to that of the control mice which were injected with the vehicle. At the end of the treatment period, no tumor was visible externally in the group of mice which received uttroside B, whereas in the control group, measurable tumor was developed. However, upon sacrifice, very small tumors were observed beneath the skin of animals treated with uttroside B too, however the size was significantly less compared to that of control animals [ Fig. 7B,C]. The tumor mass developed was histopathologically analyzed using H&E staining, which indicated a massive destruction of cells in uttroside B-treated tumor tissue, which correlated with the drastic tumor reduction [ Fig. 7(D)]. IHC staining of the formalin fixed cryosections of ectopically implanted human liver xenografts in NOD-SCID mice against cleaved PARP specific antibody revealed the in vivo apoptotic effect of uttroside B. Significant up-regulation in the level of cleaved PARP was observed in tumor sections from mice treated with uttroside B [Fig. 7(E)].

Discussion
Traditional medicines comprising of plant bioactives are often used to treat different types of cancers [11][12][13] . Steroidal glycosides and triterpenes, commonly known as saponins are a group of compounds abundant in plants and have been reported to be cytotoxic to cancer cells of different origins [14][15][16] . Degalactotigonin, uttroside A and uttroside B are the cytotoxic saponins isolated from S. nigrum Linn 8,9 , which is a widely used herb in the Indian traditional systems of medicine. Uttroside B has been isolated from other plants too and has been shown to induce cytotoxicity in neuronal, colon and cervical cancer cells 9,10 . Nevertheless, no further information has been reported regarding its mechanism of action or biological safety. In this study, we demonstrate for the very first time, the promising anticancer potential of uttroside B against hepatocellular carcinoma, with mechanism-based evidence (Patent Application No. 201641018401).
It was much compelling to note that the compound triggers cytotoxicity in all liver cancer cells, irrespective of their HBV status, which is often regarded as a common risk factor in regulating hepatocellular carcinoma. In concordance with the currently acceptable dogma regarding apoptosis 17 , uttroside B induced all classical markers of caspase-dependent apoptosis, as evidenced by presence of apoptotic bodies and cleavage of caspases and PARP. However, Parvispinoside B, another saponin, which structurally differs from uttroside B by just one sugar moiety in the lycotetraosyl unit, exhibits strong cytotoxicity against the U937 leukemia cell line 18 , while being not effective against HepG2 cells (IC50 > 100 μ M), indicating the significance of the quantity and kind of sugar present in the saponin in determining its anticancer efficacy.
Induction of cell cycle arrest is a vital mechanism through which chemotherapeutic drugs elicit cytotoxicity in cancer cells. However, uttroside B did not exhibit any influence on the cell cycle. According to previous reports, some steroidal saponins such as diosgenin, and smilagenin arrest cell cycle in G0/G1 phase, while some others like tigogenin have no effect on cell cycle indicating that the difference in the spatial conformation of the A-and B-rings and the presence or lack of 5, 6-double bond are not the determinants of the mode of action of the saponin on the cell cycle 2,19 .
We also explored the most probable mechanisms behind the anti-cancer potential of uttroside B against liver cancer cells. The tumor survival in HCC involves, highly complex multiple signalling mechanisms that regulate its growth. NF-κ B, Akt, MAPK and mTOR are the most prevalent survival signals promoting the progression of hepatic cancer and most of the drugs targeting this cancer are inhibitors of these pathways 20 . Uttroside B could not produce any significant down-regulation in the constitutive activation of NF-κ B in HepG2 cells, eliminating Reports indicate that the mammalian target of rapamycin (mTOR) pathway is abnormally activated in a proportion of HCC patients and inhibition of mTOR can suppress liver tumor growth and metastasis [21][22][23] . Moreover, an up-regulation of mTOR is frequently observed in cholangiocarcinoma, the second most common primary cancer of the liver. A complex interplay between mTOR and MAPK pathways has also been demonstrated during hepatocarcinogenesis 24 . However, further studies are required in order to confirm whether both these pathways are dependent or independent of each other in regulating the anticancer potential of uttroside B.
The biological safety of uttroside B was evaluated in Swiss albino mice using acute and chronic toxicity models. Almost all currently available chemotherapy regimens for the treatment of cancer are associated with considerable systemic toxicities, thus limiting their clinical benefits. The major complication associated with chemotherapy is the reduction in the count of hematological parameters such as lymphocytes, neutrophils and monocytes. Uttroside B produced no significant difference in such hematological parameters confirming that it does not generate any significant toxicity or immunosuppression in animals. Following the onset of liver damage, ALP, AST and ALT are released from the damaged cells, elevating their levels in the serum. Uttroside B did not induce elevation of any of these enzymes in both acute and chronic toxicity studies illustrating its pharmacological safety. Moreover, the normal level of BUN confirmed that uttroside B does not produce any severe toxicological manifestations in the kidney. Additionally, uttroside B, up to 5 fold dose, failed to exhibit any sign of cumulative adverse response in experimental animals as concluded from gross measures such as loss of body weight, ruffling of fur and change in behaviour and food intake underscoring its biological safety. The drastic inhibition of tumor growth produced by uttroside B in NOD-SCID mice bearing human liver cancer xenografts, illustrates the chemotherapeutic efficacy of uttroside B, which was further authenticated by immunohistochemical analysis of the tumor sections for the expression of cleaved PARP, a classical marker of apoptosis.

Conclusion
Chemotherapeutic options for liver cancer are limited and the prognosis of HCC patients remains dismal. Sorafenib, derived from a de novo combinatorial approach by high-throughput screening and approved by US-FDA in 2007, is the only drug currently available for the treatment of hepatocellular carcinoma. It is a multi kinase inhibitor, which can prolong the survival rate up to 20% and the only systemic agent approved for treating advanced, unresectable HCC on the basis of two phase III clinical trials. However, it has been reported to have severe side effects. In the present study, we have isolated uttroside B from the leaves of Solanum nigrum Linn and have demonstrated that it is several times more potent than Sorafenib, and does not cause noticeable side-effects in vitro and in vivo. Drastic inhibition of tumor growth produced by uttroside B in NOD-SCID mice bearing human liver cancer xenografts demonstrates the chemotherapeutic efficacy of uttroside B. These results warrant further clinical evaluation of uttroside B, so that it can be used as an effective chemotherapeutic against liver cancer, which currently has minimal therapeutic options. Extraction, Isolation and Identification of active principle. The powdered leaves were extracted using solvents of increasing polarity and extracts were concentrated under Rotavap (SI methods). The most active methanolic extract was subjected to column chromatography to obtain the active principle. The pure active fraction was subjected to spectral analysis for its identification (SI methods).

Methods
Cell lines. All cell lines were procured from NCCS, Pune (SI methods).
Chemicals and Antibodies. Antibodies against Caspases, β -actin, p-p42/44, p-JNK, p-p38, p-Akt, p-mTOR and Vinculin were obtained from Cell Signaling Technologies (Beverly, MA, USA) and the antibody against PARP was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). All the chemicals used for extraction, column chromatography and high performance liquid chromatography (HPLC) were obtained from Merck Ltd, Mumbai, India. All other chemicals were purchased from Sigma Chemicals (St. Louis, MO, USA) unless otherwise mentioned.
MTT assay. Antiproliferative activity of the extracts and the isolated compound were evaluated using MTT assay as described earlier 25 (SI methods).
Clonogenic assay. The clonogenic cell survival assay was performed as described previously 26 (SI methods).
Western blot analysis. The whole cell lysate was prepared from the cells treated with or without drug and subjected to Western blot analysis as described before 27 (SI methods).
Flow cytometry. Cell cycle analysis was performed 28 as described in SI methods.
EMSA. Nuclear extracts of the cells treated with or without the compound were prepared and EMSA was conducted to evaluate DNA-binding activity of NF-κ B as described earlier 27,29 (SI methods).
In vivo studies. The methods were carried out in accordance with the CPCSEA guidelines and all experimental protocols were approved by Institutional Animal Ethical Committee of Rajiv Gandhi Centre for Biotechnology (IAEC No: 151 (a and b)/RUBY/2012). Toxicological evaluation were conducted in Swiss albino mice as described earlier 28 (SI methods). HepG2 xenografts models were established in NOD-SCID mice as described in literature 30 . Male NOD-SCID (NOD.CB17-Prkdc scid/J ) mice of age 6-8 weeks were used for the experiment.
Immunohistochemical analysis of various proteins in the xenograft tumor tissue sections was performed using the detection kit, as per manufacturer's protocol Super Sensitive Polymer-HRP IHC Detection System (Biogenex, CA, USA) (SI methods).
Statistical Analysis. The error bars represent ± S.D., taken from independent experiments. Statistical significance was analysed by one way ANOVA or two-way ANOVA followed by Turkey's post test. Significance level was set at P < 0.05.