Antitumoral Potential of Tunisian Snake Venoms Secreted Phospholipases A2

Phospholipases type A2 (PLA2s) are the most abundant proteins found in Viperidae snake venom. They are quite fascinating from both a biological and structural point of view. Despite similarity in their structures and common catalytic properties, they exhibit a wide spectrum of pharmacological activities. Besides being hydrolases, secreted phospholipases A2 (sPLA2) are an important group of toxins, whose action at the molecular level is still a matter of debate. These proteins can display toxic effects by different mechanisms. In addition to neurotoxicity, myotoxicity, hemolytic activity, antibacterial, anticoagulant, and antiplatelet effects, some venom PLA2s show antitumor and antiangiogenic activities by mechanisms independent of their enzymatic activity. This paper aims to discuss original finding against anti-tumor and anti-angiogenic activities of sPLA2 isolated from Tunisian vipers: Cerastes cerastes and Macrovipera lebetina, representing new tools to target specific integrins, mainly, α5β1 and αv integrins.


Introduction
Snake venom is a natural biological resource, containing several neurotoxic, cardiotoxic, cytotoxic, and many other different active compounds [1,2]. Due to this broad range of biological functions, these biomolecules have been the sub�ect of hundreds of scienti�c articles in different research �elds, including biochemistry, biophysics, pharmacology, toxicology, and medicine [2][3][4][5]. Viperidae snake venoms contain class II PLA 2 s, which share structural features with secreted PLA 2 (sPLA 2 ) of the class II-A present in in�ammatory exudates in mammals. A number of venom PLA 2 s have been shown to induce a variety of pharmacological effects although comprehensive studies of the actions of venom PLA 2 s in the various events of toxicity are scarce [6,7].

Viperidae Snake Venom Phospholipase A 2 Enzymes: Secreted Phospholipases A 2
Secreted PLA 2 constitute a large superfamily of enzymes that are widely distributed in living organisms. e sPLA 2 from Viperidae snake venoms fall under group II. ey are generally Ca 2+ -dependant enzymes that catalyze the hydrolysis of the sn-2 fatty acid bond of phospholipids to release free fatty acids and lysophospholipids [7]. ese enzymes are small proteins (∼13-14 kDa), containing 120-125 aminoacid residues, 7 disul�de bridges, and have a partially conserved 2 BioMed Research International structure that de�ne the PLA 2 fold [8]. Group II snake venom PLA 2 enzymes can also be divided into different subgroups on the basis of the aminoacid residue in the forty-ninth position. Asp49 plays an important role in catalysis and it is conserved in most snake venom PLA 2 enzymes, and hence these are identi�ed as D49 enzymes [9]. However, in some of the group IIA PLA 2 enzymes this aminoacid residue is replaced by lysine, serine, asparagine, or arginine and they are identi�ed as �49 [10], S49 [11], N49 [12], or R49 [13] enzymes, respectively. Substitution of Asp in the forty-ninth position interrupts the binding of cofactor Ca 2+ to the Ca 2+ -binding loop, and hence "mutants" show low or no hydrolytic activity [10,14,15]. In addition, there are several substitutions in the Ca 2+ -binding loops of these mutant enzymes. Secreted phospholipases A 2 constitute major components of snake venoms and have been extensively investigated not only because they are very abundant in these venoms but mainly because they display a variety of relevant toxic actions such as neurotoxicity, myotoxicity, cytotoxicity, cardiotoxicity, edema-inducing, arti�cial membrane disrupting convulsant, hypotensive and proin�ammatory effects [7,[16][17][18][19]. Besides, they exert a wide range of biological effects, including anticoagulant, platelet aggregation inhibiting [7,20,21], bactericidal [22], anti-HIV [23], antimalarial and antiparasitic [24], antitumor [21,25,26], and recently antiangiogenic effect [27][28][29]. Due to this functional diversity, these structurally similar proteins aroused the interest of many researchers as molecular models for study of structurefunction relationships. One of the main experimental strategies used for the study of myotoxic PLA 2 s is the traditional chemical modi�cation of speci�c aminoacid residues and examination of the consequent effects upon the enzymatic, toxic, and pharmacological activities. Furthermore, some venom sPLA 2 have no catalytic activity while they exert various toxic and pharmacological effects [17,21,26]. e absence of direct correlation between catalytic activity and pharmacological effects has led to the hypothesis that speci�c actions of sPLA 2 are due to the presence of pharmacological sites on the sPLA 2 surface overlapping or distinct from the catalytic site. ese pharmacological sites would allow the sPLA 2 to bind speci�cally to soluble or membrane-bound proteins that participate to the sPLA 2 mechanism of action [30].
Since this hypothesis was proposed, a collection of binding proteins have been identi�ed using several toxic snake venom sPLA 2 [31]. Besides -bungarotoxin [32,33], early studies with the neurotoxic sPLA 2 OS2 from Australian Taipan snake Oxyuranus scutellatus scutellatus have led to the identi�cation of two families of binding proteins called Nand M-type receptors [31,34,35]. e N-type receptors are present in mammalian brain and other tissues. Neurotoxic sPLA 2 , such as OS2, bind with N-type receptors with high affinity, while nontoxic sPLA 2 including OS1 bind with much lower affinity, suggesting that these receptors are involved in neurotoxicity.
Conversely, the M-type receptors bind with high affinity both toxic and nontoxic sPLA 2 including OS1 and OS2 [31].
Importantly, the M-type receptors also bind with several mammalian sPLA 2 [31,36], suggesting that these proteins are the endogenous ligands for these receptors, and possibly for the collection of binding proteins initially identi�ed with venom sPLA 2 .

Tunisian Viperidae Snake Venom Proteins
Snake venom is a natural source for molecules known as modulators of integrin-mediated functions [37]. Pharmacological study of snake venoms reveals structural and functional polymorphisms of proteins they contain. In our laboratory in Pasteur Institute of Tunis, we are interested in studying different pharmacological effects of Tunisian Viperidae venoms, mainly, the horned viper, Cerastes cerastes, Macrovipera lebetina transmediterranea, and Cerastes vipera [38]. Bazaa et al. showed that these venoms contain proteins belonging to a few protein families. However, each venom showed distinct degree of protein composition complexity. e three venoms shared a number of protein classes though the relative occurrence of these toxins was different in each snake species. On the other hand, the venoms of the Cerastes species and Macrovipera lebetina each contained unique components [38]. e comparative proteomic analysis of Tunisian snake venoms provides a comprehensible catalogue of secreted proteins, which may contribute to a deeper understanding of the biological effects of the venoms and may also serve as a starting point for studying structure-function correlations of individual toxins.
protagonists of the complex multistep process of angiogenesis, the major target for the development of anticancer therapies [21,27,28] (Figure 1).

Tunisian Viperidae sPLA 2 Effects on Haemostatic System.
Snake venom toxins are now regularly used in laboratories for assaying haemostatic parameters and as coagulation reagents [54,55]. PLA 2 enzymes are known to inhibit blood coagulation. Depending on the dose required to inhibit coagulation, they are classi�ed into strong, weak, and nonanticoagulant enzymes [56,57]. Strong anticoagulant PLA 2 enzymes inhibit the activation of FX to FXa by both enzymatic and nonenzymatic mechanisms and inhibit the activation of prothrombin to thrombin by nonenzymatic mechanism [58,59]. In our case, 0.14 M of both CC-PLA 2 s completely inhibited plasma coagulation. us, CC-PLA 2 s could be considered among the most anticoagulant yet described for PLA 2 s snake venom [21]. Lizaro and coworkers showed that myotoxin II, a basic PLA 2 from Bothrops nummifer, was unable to inhibit coagulation of the platelet-poor plasma until 3.57 M [60]. Moreover, it has been shown that BaspPLA(2)-II, an acidic, Asp49 PLA 2 from Bothrops asper venom lacks anticoagulant activity [61]. Platelet aggregation plays a role in clot retraction and wound healing. Any alteration in platelet aggregation could lead to debilitation or death. CC-PLA 2 -1 and CC-PLA 2 -2 showed high antiplatelet aggregation activities induced by arachidonic acid or ADP [21], contrary to b/D-PLA2 which displays high enzymatic and anticoagulant activities but has no platelet aggregation [62]. Moreover, Kashima and coworkers reported that BthA-I-PLA 2 , a nontoxic acidic PLA 2 from Bothrops jararacussu snake venom, inhibited ADP-induced platelet aggregation with moderate effect [63].
While, OHVA-PLA 2 , an acidic PLA 2 from Ophiophagus hannah, strongly inhibited platelet aggregation in the presence of ADP or arachidonic acid [64]. It thus appears that PLA2 platelet activity is not directly due to its acidic nature or its anticoagulation activity.
Adhesion and cell migration are two fundamental steps in numerous diseases, like cancer. CC-PLA 2 -1, CC-PLA 2 -2, and MVL-PLA 2 inhibit adhesion and migration of human HT1080 and IGR39 cells to �brinogen and �bronectin. is effect persists even aer complete blockage of the catalytic activity suggesting that, contrary to Bth-A-I-PLA 2 whose antitumoral effect appears to be linked to enzymatic site [63], the inhibitory and enzymatic activities are supported by different sites. RVV-7, a cytotoxic basic PLA 2 from Russsell's viper venom, inhibits also tumor development [65]. On the contrary, b/D-PLA 2 represents the exception of these enzymes as it stimulates tumor growth [62]. Since Tunisian phospholipases A 2 are not cytotoxic, it seems that their antitumoral activity is exerted by a different mechanism. Using different assays, such as a solid-phase binding assay and a panel of immobilized antibodies, we have proved that CC-PLA 2 -1, CC-PLA 2 -2, and MVL-PLA 2 inhibit cell adhesion and migration by interacting directly with v and 5 1 integrins [26,28].

Tunisian Viperidae sPLA 2 Effects on Angiogenesis.
Angiogenesis is fundamental to normal healing, reproduction, and embryonic development. However, this process is also important in the pathogenesis of a broad range of disorders such as arthritis and cancer [66]. Angiogenesis is thus required to sustain malignant cells with nutrients and oxygen for tumors to grow beyond a microscopic size. us, the microvascular endothelial cell recruited by a tumor is an important target in cancer therapy and has the advantage of being genetically stable. erefore, treating both the cancer cell and the endothelial cell in a tumor may be more effective than treating the cancer cell alone. e role of v 3 integrin in the angiogenic process is well documented [67]. In the last decade, several clinical trials evaluating the efficacy of v 3 blockers have led to encouraging results in cancer therapy and diagnosis. Similarly, 5 1 integrin is involved in angiogenesis and more precisely in growing vessels, but its expression disappears in mature vessels [68]. ereby, when tested in vitro, the two CC-PLA 2 and MVL-PLA 2 impaired adhesion and migration of HBMEC (human brain microvascular endothelial cells) and HMEC-1 (human microvascular endothelial cell), respectively, by interfering with integrin function. Moreover, using the CAM assay, an ex vivo model, these sPLA 2 strongly reduced vasculature development. e treatment reduced the number of new capillaries and branching, without affecting the mature blood vessels, suggesting once again the implication of 5 1 integrin. Interestingly, CC-PLA 2 -1 and CC-PLA 2 -2 inhibit spontaneous angiogenesis as well as angiogenesis induced by growth factors such as VEGF or bFGF [28]. e antiangiogenic effect of PLA 2 can be due partly to the blockage of the v 3 and 5 1 integrins functions. However, inhibition of angiogenesis can also result from blockage of VEGF or its receptor. us, it has been reported that inactive PLA 2 homologues, such as KDR-bp isolated from Eastern cottonmouth venom, are common antagonists of KDR, a VEGF receptor [69].
Focal adhesions are specialized sites of attachment of cells where integrins receptors, such as v 3, link the extracellular matrix to the actin cytoskeleton, allowing migration [70]. Cell migration is a complex cellular behavior that results from the coordinated changes in the actin cytoskeleton and the controlled formation and dispersal of cell-substrate adhesion sites. While the actin cytoskeleton provides the driving force at the cell front, the microtubule network assumes a regulatory function in coordinating rear retraction. e polarity within migrating cells is further highlighted by the stationary behavior of focal adhesions in the front and their sliding in trailing ends [71].
Treatment of HMEC-1 cells with MVL-PLA 2 induced important changes in cell morphology. Treated cells have a circular shape and actin stress �bers are thinner or absent, with the actin mainly located at the cell periphery. Moreover, MVL-PLA 2 leads to drastic reduction in the size of focal adhesions and their distribution all over the ventral surface of cells, consistent with a decrease in v 3 integrin clustering and its absence from lamellipodia [27]. erefore, it appears that the inhibition of migration is associated with important reorganization of the actin cytoskeleton and focal adhesions. Again, there is a clear dissociation between the anti-angiogenic effect and the catalytic activity.
Furthermore, MVL-PLA 2 strongly increased MT dynamicity in HMEC-1 cells. Because the microtubule cytoskeleton is essential in the orchestration of endothelial cell motility [72,73], microtubule-targeting agents are known to have antiangiogenic effects through the modulation of cytoskeleton dynamicity [27]. us, microtubule-binding drugs are widely used in cancer chemotherapy and also have clinically relevant antiangiogenic and vascular-disrupting properties [74].

�� �m�orta�ce o� the ��e�ti�catio� o�
Pharmacological Sites e pharmacological sites of PLA 2 enzymes determine the affinity between the PLA 2 and target proteins. e identi�cation of pharmacological sites helps in (1) understanding the structure-function relationships of PLA 2 enzymes, (2) developing strategies to neutralize the toxicity and pharmacological effects by targeting these sites, and (3) developing prototypes of novel research tools and pharmaceutical drugs [7,8]. In our studies, we showed that CC-PLA 2 -1, CC-PLA 2 -2, and MVL-PLA 2 target the 5 1 and v integrins, particularly v 3. Moreover, angiogenesis involves expression of the later, which binds to RGD-containing components of the interstitial matrix [75].
To further understand the mechanism of action, we report that endothelial cells are able to adhere on immobilized MVL-PLA 2 and that this adhesion is impaired by RGD peptides [27]. is suggests that interaction between MVL-PLA 2 , CC-PLA 2 -1, or CC-PLA 2 -2 and integrins involves RGD-like sequence which may be responsible for the inhibition of integrin function. is hypothesis is supported by Ramos and coworkers' study, showing that general folding of electrostatic potential is the main intervening of disintegrinintegrin interaction [76].
When MVL-PLA 2 contains a NGD sequence, which could be considered as an RGD-like motif, CC-PLA 2 -1 and CC-PLA 2 -2 present NQD and NQI, respectively, that may also be responsible for the inhibition of integrin function. erefore, bioinformatics study and structural criteria that would allow identifying biologically active RGD-sites on the base of a protein's spatial structure may become a helpful tool for analysis of cellular function of proteins [77]. Furthermore, conformation of the integrin-binding loop in a protein is de�ned not only by physicochemical properties and conformation of the sequence itself, but also by its structural environment and therefore of the potential biological activity. Besides the RGD-like sequence site should be placed on a loop or a beta-turn to be well exposed. We can cite disintegrin, like applied model, in which we can note a loop accessible stabilized by disul�de bridges [78]. Firstly, as shown in Figure 3(a), the three models are very similar. Interestingly, we can note the presence of very-wellexposed loop containing the suspected RGD-like motif. is loop is very similar to that of the disintegrins.
In the case of MVL-PLA 2 , we �nd the NGD motif, while for CC-PLA 2 -1 and CC-PLA 2 -2 there is NQD and NQI, respectively. According to our hypothesis, the residue R in RGD motif is replaced by the N which is hydrophilic and polar residue, it is even more hydrophilic than R residue, this leads to higher affinity towards the v 3 integrin [79]. Besides, the D residue favors recognition of v 3 and 5 1 integrins [79]. In addition, in CC-PLA 2 -1 and CC-PLA 2 -2 the RGD-like motif is �anked by two E residues, highly polarized which could enhance the inhibitory effect towards integrins that bind to ligands through RGD sites, including the �bronectin receptor, mainly, the 5 1 integrin [80].
On the other side, based on the study of disintegrins, it is known that integrin-binding ability is apparently more related to the Cys-rich domain. Similarly, CC-PLA 2 -1, CC-PLA 2 -2, and MVL-PLA 2 present 1� Cys forming 7 disul�des bridges. We can postulate that disul�de bonds, especially Cys50-Cys86 and Cys57-Cys79, stabilized the hypothetical integrin-binding loop. e superimposition of the structural models of CC-PLA 2 -1, CC-PLA 2 -2, and MVL-PLA 2 shows that they share similar conformational features (Figure 3(b)).
Nevertheless, further structure-function relationships study must be carried to verify this hypothesis.

Conclusion
Secreted phospholipase A 2 enzymes, especially from Viperidae Snake venom, exhibit a wide variety of pharmacological effects despite their structure similarity. ese enzymes provide a great challenge to protein chemists as subtle and complex puzzles in structure-function relationship. A better understanding will contribute to our knowledge of proteinprotein interactions, protein targeting, and protein engineering, and to the development of better-targeted delivery systems. Further research in identifying target proteins will bring details on the mechanisms of the pharmacological effects at the cellular and molecular levels. Studies in these areas will result in new, exciting, and innovative opportunities in the future, both in �nding answers to the toxicity of PLA 2 enzymes and could bring useful tools for developing proteins with novel functions.
Interestingly, we have demonstrated that two isoforms of PLA 2 (CC-PLA 2 -1 and -2), from horned Tunisian viper Cerastes cerastes and another from Macrovipera lebetina MVL-PLA 2 target integrins, a large and very important family of adhesion molecules that promote stable interactions between cells and their environment [26,28]. Indeed, these sPLA 2 exhibit a potent antitumor and antiangiogenic activities. We showed that their effect is likely due to the inhibition of 5 1-and v-containing integrins [26,28].
ese nontoxic secreted phospholipase A 2 could be new tools to disrupt different steps of tumor and angiogenesis progression through integrins. It is noteworthy that this effect is independent of the enzymatic activity. is �nding may serve, on the one hand, as a mean to discuss the molecular regions involved in recognition of tissue targets and, on the other hand, as starting point structure-function relationship studies leading to design a new generation of anticancer drugs.