Abstract

Background

Snake venom is widely regarded as an advanced “biochemical weapon” that exists in nature. Venomous snakes rely on venom for survival, where it is employed for predation (primary function), digestion and defense [1]. Rapid evolution drives the emergence of diverse pharmacologically active snake venom proteins that are adapted to target the normal physiology of prey by exerting various toxic effects [2]. The prey of venomous snakes consists in parts of small mammals such as rodents but may also include birds, lizards and amphibians [3, 4]. It is noteworthy that the mammalian system-targeting property of snake venom proteins is often specific and selective, thus making them a natural repertoire of therapeutic molecules [5, 6].

At present, advanced cancers that are fast-growing and capable of metastasizing is a major cause of mortality globally [7, 8]. In this context, snake venom proteins have the potential for drug discovery in line with the search for novel peptide-based anticancer agents with high efficacy and selectivity in targeting cancer cells [9, 10]. Notable examples typically involved venom proteins from vipers and pit vipers (family Viperidae), e.g. disintegrin from Agkistrodon contortrix venom [11], phospholipases A₂ from Cerastes cerastes [12] and L-amino acid oxidases from Bothrops sp. [13].

A common pharmacological feature of the venoms of viperid snakes, including the Asian arboreal pit vipers from the Trimeresurus complex, is their ability to derange hemostasis and induce local tissue necrosis [14-16]. The tissue-necrotizing effect associated with Trimeresurus pit viper envenomation suggests cytotoxic activity in the venom that may be further explored for anticancer potential. A recent study with Thai Trimeresurus purpureomaculatus venom reported high cytotoxicity of the venom toward SHSY5Y cell line, and moderate cytotoxicity of BPP-related peptides, PLA₂ and peptide-rich fraction of the venom [17]. Another publication reported that L-amino acid oxidase (LAAO, 55−60 kDa) from Malaysian Trimeresurus purpureomaculatus venom induced cytotoxicity in three colon cell lines (SW480, SW620 and CCD-18Co), but no selectivity was observed between the cancerous cell lines (SW480, SW620) and the non-cancerous, normal cell line (CCD-18Co) [18].

Our recent quantitative proteomic analysis of the Malaysian T. purpureomaculatus venom (MTP) revealed that LAAO constituted about 3% of the total venom proteins, while disintegrin, a compound with potential inhibitory effect on cancer cell growth, was present at a higher abundance in the venom (approximately 14%) [19]. The LAAO content reported previously for the Malaysian T. purpureomaculatus venom could not be compared as it was estimated with a qualitative method [20]. In a more recent study, relative protein abundances of 0.91% LAAO and 0.10% disintegrin were identified in the Thai T. purpureomaculatus venom [17].

Disintegrin is a much smaller (4-15 kDa, monomer or dimer) cysteine-rich and RGD-containing polypeptide that targets specifically on integrin - a ubiquitously expressed cell surface receptor that regulates cell motility, survival, proliferation, angiogenesis and cell invasion [21, 22]. The antineoplastic potential of snake venom disintegrins have been shown in several studies, e.g. contortrostatin (disintegrin isolated from Agkistrodon contortrix venom) [23], saxatilin (disintegrin from Gloydius saxatilis venom) [24], PAIEM (disintegrin isolated from Echis multisquamatis venom) [25] and a disintegrin from Crotalus durissus collilineatus venom [26]. The disintegrins act by inhibiting angiogenesis, invasion and migration of cancer cells. The finding of disintegrin in the Malaysian T. purpureomaculatus venom is significant as the peptide could be a promising local source of anticancer candidate that has not been characterized previously.

In this study, the Malaysian T. purpureomaculatus venom was suggested to be cytotoxic to a wider range of cancer cell lines including those of the human breast, lung, colon and prostate. The selective anticancer activity, if any, could be partly contributed by the disintegrin present in the venom. Hence, this study aimed to investigate the cytotoxicity of Malaysian T. purpureomaculatus venom across four common human cancerous and the corresponding normal cell lines. In addition, the disintegrin was purified from the venom, and its amino acid sequence as well as anticancer properties were characterized.

Methods

Results

Cytotoxicity of Trimeresurus purpureomaculatus Venom

MTP venom exhibited dose-dependent cytotoxic effects toward all cancer cell lines tested (median inhibition concentrations, IC₅₀ = 0.42−2.50 µg/mL). The effects were generally stronger in the cancer cells compared to the corresponding normal cells (IC₅₀ = 0.70-6.98 µg/mL) (Table 1). The venom cytotoxicity was most potent to the colorectal adenocarcinoma cell line (IC₅₀ = 0.42 ± 0.06 µg/mL), followed by the breast cancer cell line (IC₅₀ = 0.48 ± 0.02 µg/mL) and prostate cancer cell line (IC₅₀ = 1.60 ± 0.18 µg/mL). The venom was least cytotoxic to the lung cancer cell line (IC₅₀ = 2.50 ± 0.20 µg/mL). The venom was also significantly more potent compared to 5-FU in inhibiting the growth of breast, lung and prostate cancer cells (p < 0.05). In comparison to the IC₅₀ in normal cell lines, the cytotoxic effect was not selective toward colon, lung and prostate cancer cells as indicated by a selectivity index (SI) ≤ 10. The selectivity index of the venom was, however, much higher (SI = 14.54) for the breast cancer cell line (MCF7), implying that the venom was 15-fold more cytotoxic to MCF7 than it was to the corresponding normal cell line (184B5) (Table 1, Figure 1).

Table 1.

Median inhibition concentrations (IC₅₀) (µg/mL) and selectivity indices (SI) of Malaysian T. purpureomaculatus (MTP) venom for different cell lines after 72-hour treatment in comparison to 5-fluorouracil (5-FU).

	MTP		5-FU
Cell Lines	IC50 (µg/mL)	SI	IC50 (µg/mL)	SI
Breast
MCF7	0.48 ± 0.02	14.54	8.26 ± 0.77	0.16
184B5	6.98 ± 0.53		1.34 ± 0.14
Colon
HT-29	0.42 ± 0.06	1.67	0.12 ± 0.02	>258.33
CCD-18Co	0.70 ± 0.15		> 31.62
Lung
A549	2.50 ± 0.20	1.44	> 31.62	NA
MRC5	3.61 ± 0.14		> 31.62
Prostate
PC3	1.60 ± 0.18	1.19	12.83 ± 1.52	0.22
RWPE-1	1.91 ± 0.53		2.79 ± 0.66

Open in a separate window

IC₅₀ values were derived from triplicates ± S.E.M.

Selectivity index ≥ 10 is considered as the threshold for a compound having cancer-selective cytotoxic effect [28].

NA: not available, MCF7: human breast cancer cell line, 184B5: human breast normal cell line, HT-29: human colon cancer cell line, CCD-18Co: human colon normal cell line, A549: human lung cancer cell line, MRC5: human lung normal cell line, PC3: human prostate cancer cell line, RWPE-1: human prostate normal cell line.

Open in a separate window

Figure 1.

Dose-dependent growth inhibitory effect of Trimeresurus purpureomaculatus (MTP) venom in human breast cell lines. Median inhibition concentrations (IC₅₀) were determined from the dose-response curve. Values were presented as means ± S.E.M. (n = 3). MCF7: human breast cancer cell line, 184B5: human breast normal cell line.

Cytotoxicity of the Chromatographic Fractions of Trimeresurus purpureomaculatus Venom

MTP venom was resolved using C₁₈ reversed-phase high performance liquid chromatography (RP-HPLC) into 12 fractions as shown in Figure 2A. The cytotoxic effect of each fraction was further tested on MCF7 cell line in view of the high selectivity index (SI) of MTP venom in breast cancer cell line. MCF7 cells treated with protein from Fraction 1 (20 µg) showed the most potent inhibitory effect (cell viability reduced by 65%) compared to the non-treated control (**p < 0.01). Proteins from Fraction 2 and 12 exhibited moderate cytotoxic effect (cell viability reduced by 25−30%) whereas other fractions did not show significant cytotoxicity in the MCF7 cell line (Figure 2B). In comparison, 5-FU at IC₅₀ of 9 µg/mL induced a 50% reduction in the cell viability study.

Open in a separate window

Figure 2.

(A) C₁₈ reversed-phase high performance liquid chromatography of venom from the Malaysian T. purpureomaculatus. (B) Cell viability of human breast cancer cells (MCF7) after 72-hour treatment with the HPLC fractions of Malaysian T. purpureomaculatus venom. Positive and negative controls were 5-fluorouracil (5-FU) and treatment-free, respectively, in the assay.

Identification and Cytotoxicity of Disintegrin from Trimeresurus purpureomaculatus Venom

Figure 3 shows the SDS-PAGE profile of the cytotoxic protein isolated in Fraction 1. A homogenous protein band was observed and the molecular mass was estimated to be 7585 Da by its relative migration distance (Rf). LC-MS/MS identified Fraction 1 to contain only one protein, which is a disintegrin of snake venom and it is termed as “purpureomaculin” in the present study. The mass spectrometry data file of purpureomaculin is available in Additional file 1, and the primary data was deposited to the ProteomeXchange Consortium via the iProX partner repository [33] with the dataset identifier PXD018463.

Open in a separate window

Figure 3.

Protein content of Fraction 1 (F1) of Malaysian T. purpureomaculatus venom was validated under 15% reducing gel electrophoresis. Protein ladder PM2700 ExcelBand™ 3-color Broad Range Protein Marker was used for molecular weight standards (5−245 kDa).

Purpureomaculin at 20 µg significantly inhibited the growth of MCF7 cell line by 65% (**p < 0.01), while no significant cytotoxic effect was observed in the corresponding normal breast cells (184B5 cell line) (Figure 4). Microscopic examination revealed the presence of detached, clumped and rounded floater cells in the purpureomaculin-treated breast cancer cells following a treatment period of 72 hours. On the contrary, the normal cell line (184B5) treated with purpureomaculin was observed to remain healthy and viable, with slight reduced growth compared to the untreated control (Figure 5).

Open in a separate window

Figure 4.

Cell viability of human breast cell lines, MCF7 (cancerous) and 184B5 (normal) after 72-hour treatment with 20 µg of Fraction 1 (F1, purpureomaculin) purified from Trimeresurus purpureomaculatus venom. Positive control: 5-fluorouracil (5-FU); negative control: treatment-free well.

Open in a separate window

Figure 5.

Morphological changes in human breast cell lines, (A) MCF7 (cancerous) and (B) 184B5 (normal) after 72-hour treatment of Fraction 1 (purpureomaculin) of Trimeresurus purpureomaculatus venom. Scale bars = 30 µm.

Sequence Analysis and in silico Characterization of Purpureomaculin, the Cytotoxic Disintegrin from T. purpureomaculatus Venom

The tryptic peptides of purpureomaculin (K)LLPGAQCGEGLCCDQCSFMKK(G) and (R)ARGDDLDDYCNGISAGCPR(N) from Fraction 1 were matched to the disintegrin transcript “CL53.Contig5_CP” from the Malaysian T. purpureomaculatus venom-gland transcriptomic database. A disintegrin with full amino acid sequence was identified. It is a mature chain polypeptide of 73 amino acid residues: EAGEDCDCGSPANPCCNAATCKLLPGAQCGEGLCCDQCSFMKKGTICRRARGDDLDDYCNGISAGCPRNPLHA.

Multiple sequence alignment showed that purpureomaculin is highly homologous (98.6%) with albolabrin from Trimeresurus albolabris venom and trigramin-gamma from Trimeresurus gramineus venom, where purpureomaculin varies from albolabrin and trigramin-gamma by only one amino acid at the 17^th position of the sequence (substitution of Asp to Asn in purpureomaculin) (Figure 6). Six disulfide bonds could be formed within the sequence, pairing cysteine residues at the following positions: (6, 15), (8, 16), (21, 35), (29, 59), (34, 38), (47, 66). Purpureomaculin also harbored the tripeptide RGD, a short sequence motif of biological interest formed by amino acid residues 51−53. Compared with the disintegrin sequences of other viperids, high similarity (> 70%) was observed in phylogenetically related pit viper genera and species, e.g. Trimeresurus complex.

Open in a separate window

Figure 6.

Multiple sequence alignment of purpureomaculin (CL53.Contig5_CP) with other disintegrin sequences of Crotalinae. Purpureomaculin is a disintegrin isolated from the Malaysian Trimeresurus purpureomaculatus venom, while other disintegrin sequences were obtained from UniProtKB. The percentages of similarity (% similarity) of disintegrin sequences in comparison to purpureomaculin were calculated using Omega Clustal software. Zappo color scheme indicates residue properties: pink: aliphatic/hydrophobic; orange: aromatic; blue: positive; red: negative; green: hydrophilic; magenta: proline/glycine (conformationally special); yellow: cysteine. Disulfide bonds were illustrated in black connecting lines.

Table 2 shows the physicochemical properties of purpureomaculin analyzed in silico using the ProtParam Tool. Purpureomaculin has an estimated molecular weight of 7572.52 Da and is weakly acidic (pI = 4.79).

Table 2.

Physicochemical properties of purpureomaculin from the Malaysian Trimeresurus purpureomaculatus venom.

Properties	Values
Residues	73
Mr	7572.52
pI	4.79
Instability Index (II)a	25.23
Stability	Stable
Aliphatic Index (AI)b	48.36
GRAVYc	-0.390
Extinction coefficients*d	2240
Extinction coefficients* (reduced)e	1490

Open in a separate window

Mr: Molecular weight (Da); pI: Isoelectric point; GRAVY: Grand average of hydropathicity.

^aEstimate of protein stability (instability index of <40 is predicted as stable).

^bRelative volume occupied by aliphatic side chains (alanine, valine, isoleucine, and leucine). Positive factor denotes for the increase of thermostability.

^cSum of hydropathy values of all amino acid, divided by total number of residues in sequence.

*Amount of light a protein absorbs at a certain wavelength. Extinction coefficient unit is in M^-1cm^-1, measured at 280 nm in water.

^dAssuming all pairs of cysteine residues appears as half cystines.

^eAssuming all cysteine residues are reduced/no cysteine appears as half cystine.

Discussion

In this study, the potent cytotoxicity of Malaysian T. purpureomaculatus venom was shown in a panel of cancer (MCF7, HT-29, PC3, A549) and normal (184B5, CCD-18Co, RWPE-1, MRC5) cell lines. MTP venom was particularly toxic to breast cancer cell line (MCF7) and colon cell lines (HT-29 and CCD-18Co), indicating high specificity of the venom toward these tested cells. The venom also demonstrated selective cytotoxicity, particularly in the human breast cancer cells (MCF7). This suggests that the venom contains toxin(s) or protein(s) that could be cancer-specific in breast tumor, thus potentiating the venom cytotoxicity at a lower dose in the cancer cells compared to the corresponding normal cells. The cytotoxic effect of Malaysian T. purpureomaculatus venom also appeared to be more potent than other pit viper venoms, including venoms of Protobothrops flavoviridis [35], Trimeresurus macrops and Trimeresurus hageni [36], albeit different human cell lines were utilized in each study.

Further investigation revealed that most of the venom components (fractionated with HPLC) showed variable anticancer activity, implying that the overall venom cytotoxicity could be a result of synergistic interactions of multiple components in the venom. The most potent anti-neoplastic activity was found in the protein fraction containing snake venom disintegrin. This was a purified protein validated through SDS-PAGE and LC-MS/MS, and termed purpureomaculin in line with the naming of most snake venom-derived disintegrin protein (prefix “species-” + post-fix “-in”). The finding of purpureomaculin anticancer activity is consistent with earlier studies that reported the inhibitory activity of snake venom disintegrins on the proliferation, metastasis and adhesion of cancer cells [21, 25, 37]. The selective cytotoxic effect of purpureomaculin in breast cancer cells (MCF7) corroborates the selective cytotoxicity of MTP venom in the same cancer cell line observed in this study.

The present findings varied when compared with previous studies on other cancer cells using T. purpureomaculatus venoms from Malaysia and Thailand [17, 18]. IC₅₀ values with wide discrepancy were observed between previous studies (Malaysian T. purpureomaculatus IC_{50 (72 hours)}: 15.99−29.43 µg/mL in human colon cell lines [18]; Thai T. purpureomaculatus (IC_{50 (48 hours)}: 0.25−5.24 µg/mL in human neuroblastoma, cervical, colon, breast, bladder, glioblastoma and kidney cell lines [17]) and the present work (Malaysian T. purpureomaculatus: IC_{50 (72 hours)}: 0.42-6.98 µg/mL in breast, lung, colon and prostate cell lines) (Table 3). Nevertheless, the cytotoxic activity of the present Malaysian T. purpureomaculatus venom in human colon cells appeared to be comparable to the Thai venom sample reported by Ozverel [17]. In contrast, much higher IC₅₀ values (by a difference of > 10 folds) in human colon cells were reported previously in the other Malaysian T. purpureomaculatus venom sample [18]. The discrepancy could be due to the variation in venom composition, the use of different colon cell lines, or the influence of the condition of venom preparation. In our experience, repeated freeze-thaw cycles of the MTP venom led to inconsistent and somewhat deteriorating venom cytotoxic activity (unpublished); hence, all venom samples used in the present study were freshly reconstituted from lyophilized stock sample prior to treating the cells in the MTT experiment. The additional measure, presumably, avoided any possible protein degradation [38] or inactivation of the venom bioactive components by unfavorable pH or temperature [39].

In addition, previous studies had demonstrated variable anticancer activities for venom disintegrin from pit vipers. For instance, lebein (disintegrin isolated from Macrovipera lebetina venom) has been shown to significantly reduce the cell viability of colon adenocarcinoma cell lines after 72-hour incubation [40]. A more recent study demonstrated that disintegrins isolated from Cerastes cerastes venom displayed varying IC₅₀ values (1.60−8.17 µg/mL) in SHSY5Y neuroblastoma cell line [17]. Tzabcanin (disintegrin isolated from Crotalus simus tzabcan venom) exhibited low cytotoxicity against Colo-205 cells and was not cytotoxic to MCF7 cells [41]. On the contrary, the disintegrin purpureomaculin of T. purpureomaculatus was found to be potent in inhibiting the growth of MCF7 cells (current study). The anticancer activities of disintegrin from many closely related species such as trigramin (disintegrin isolated from Trimeresurus gramineus venom) and albolabrin (disintegrin isolated from Trimeresurus albolabris venom) remain unknown to date, and warrant investigation for insights in the anticancer potential of the venom peptide from this large Asiatic pit viper complex.

The multiple sequence alignment revealed that the venom disintegrins of pit vipers (Crotalinae) are mainly of intermediate length with 12 conserved cysteine residues, forming 6 disulfide bonds that support the folding and stability of the protein [42]. Disintegrins from Trimeresurus, Bothrops, Gloydius, Crotalus, Protobothrops and Lachesis complexes in particular showed a higher rate of amino acid conservation, indicating the phylogenetic relationship among different genera within the Crotalinae subfamily. At the intra-generic level of Trimeresurus, purpureomaculin sequence was found almost identical (similarity: 98.6%) to albolabrin and trigramin-gamma (disintegrins from Trimeresurus albolabris and Trimeresurus gramineus). Like albolabrin and trigramin, purpureomaculin is structurally categorized as the medium-sized disintegrins which contain about 70 amino acids and 6 disulfide cysteine bonds, as described by Calvete [42]. The high similarity of purpureomaculin to these disintegrins indicates structural and functional resemblance between these disintegrins, and that purpureomaculin is likely to exert its cytotoxic activity similar to trigramin and albolabrin by inhibiting cell-matrix adhesion through the binding of RGD tripeptide domain to the cellular integrin receptors [43, 44]. One remarkable difference of purpureomaculin is the amino acid substitution at the 17^th position. It is noted that amino acid at this position is highly conserved throughout the pit viper genera, and the substitution of the aspartic acid (D) to the neutral asparagine (N) in purpureomaculin may have a unique evolutionary implication on the role of the protein expressed. The consequence of the amino acid substitution deserves further investigation.

The sequence similarities between purpureomaculin and other disintegrin sequences are lower (69−86%) when comparing with venom disintegrins of different genera other than Trimeresurus. All pit viper disintegrin sequences (including purpureomaculin) contains RGD-flanking motif, a tripeptide domain known to bind to a wide range of integrin receptors. In particular, disintegrins with RGD tripeptides were known to bind key integrin receptors such as α2β1 [45], α4β1 [46], αvβ3 [47, 48] and αvβ5 [49], leading to inhibition of integrin-mediated functions in cancer cells. Among these, the high expression of integrins αvβ3 and αvβ5 are commonly associated with tumorigenesis [50]. A distinct feature of MCF7 cell line is that it lacks integrin αvβ3 receptor, but showed hyperexpression of integrins αvβ5 receptor [51]. We hypothesized that the selective cytotoxicity of purpureomaculin is at least in part mediated through the binding of integrin receptors (e.g. αvβ5 receptor) that leads to the inhibition of cancer cell adhesion, angiogenesis and metastasis. This explains the microscopic observation in the present study, where purpureomaculin caused significant detachment of the treated breast cancer cells (MCF7), forming round-shape floaters while sparing the architecture of the normal breast cells (184B5).

In addition, the primary sequence of purpureomaculin was analyzed in silico to further understand the physicochemical properties of the identified disintegrin. Purpureomaculin was computed to have a molecular weight of 7572.5 Da, similar to that determined from the reducing SDS-PAGE of the protein (7585 Da). The molecular mass determined, however, may be slightly varied from that examined by intact mass profiling using a top-down mass spectrometry analysis. The intact mass profiling of purpureomaculin hence should be carried out in the future to establish the exact mass. On the other hand, the computed instability index (II) that falls below 40 indicates that purpureomaculin is theoretically stable in vitro. The instability index could be applied to estimate the in vivo half-life of a protein [52]: Proteins whose II < 40 have an in vivo half-life beyond 16 h, whereas proteins whose II > 40 are usually eliminated faster from the body with an in vivo half-life of less than 5 h [53]. The II value obtained for purpureomaculin in this work was about 25, predicting a long in vivo half-life beyond 16 h for this disintegrin. The aliphatic index (AI) represents the relative volume of aliphatic side chains (alanine, valine, isoleucine and leucine) in the amino acid sequence of a protein. The AI value computed for purpureomaculin reflects high protein thermostability. On the other hand, the GRAVY parameter predicts for the feature hydrophobic (positive values) or hydrophilic (negative values) of a protein. Purpureomaculin has a negative value of GRAVY, implying that it is more of a hydrophilic protein - this is also consistent with the early elution of this protein from the reversed-phase HPLC. Together, the physicochemical parameters predict that purpureomaculin is a relatively stable protein with an extended in vivo half-life, and readily dissolves in a polar solvent. Moreover, the small molecular size of the disintegrin peptide (7572.52 Da) implies that it is likely less antigenic, more accessible to cancer cell environment and amenable to structural modification e.g. nano-carrier conjugation in the advanced development of peptide-based anticancer agent [54, 55].

Conclusion

The present study demonstrated the potent cytotoxicity of the Malaysian T. purpureomaculatus venom and unveiled its selective anticancer activity in the human breast cancer cell line (MCF7). Purpureomaculin, a disintegrin identified from the Malaysian T. purpureomaculatus venom, was found to be most cytotoxic and selective among several protein fractions of the venom, presumably due to its inhibitory action on cell-cell adhesion. In silico characterization of purpureomaculin sequence predicted a stable, hydrophilic small polypeptide molecule, with features that may favorably unleash its pharmaceutical potential in the development of a peptide-based anticancer therapeutic. Nevertheless, further studies are required to investigate the peptide delivery system and immunogenicity of the peptide, which are the main obstacles of most current peptide-based therapeutics.

Supplementary material

Footnotes

References