Vampire Venom: Vasodilatory Mechanisms of Vampire Bat (Desmodus rotundus) Blood Feeding

Animals that specialise in blood feeding have particular challenges in obtaining their meal, whereby they impair blood hemostasis by promoting anticoagulation and vasodilation in order to facilitate feeding. These convergent selection pressures have been studied in a number of lineages, ranging from fleas to leeches. However, the vampire bat (Desmondus rotundus) is unstudied in regards to potential vasodilatory mechanisms of their feeding secretions (which are a type of venom). This is despite the intense investigations of their anticoagulant properties which have demonstrated that D. rotundus venom contains strong anticoagulant and proteolytic activities which delay the formation of blood clots and interfere with the blood coagulation cascade. In this study, we identified and tested a compound from D. rotundus venom that is similar in size and amino acid sequence to human calcitonin gene-related peptide (CGRP) which has potent vasodilatory properties. We found that the vampire bat-derived form of CGRP (i.e., vCGRP) selectively caused endothelium-independent relaxation of pre-contracted rat small mesenteric arteries. The vasorelaxant efficacy and potency of vCGRP were similar to that of CGRP, in activating CGRP receptors and Kv channels to relax arteriole smooth muscle, which would facilitate blood meal feeding by promoting continual blood flow. Our results provide, for the first time, a detailed investigation into the identification and function of a vasodilatory peptide found in D. rotundus venom, which provides a basis in understanding the convergent pathways and selectivity of hematophagous venoms. These unique peptides also show excellent drug design and development potential, thus highlighting the social and economic value of venomous animals.


Introduction
Common vampire bats (Desmondus rotundus) are found in Central and South America, and feed exclusively on mammalian blood [1,2]. They preferentially feed on livestock animals such as cattle [3] and produce venom components that disrupt the blood coagulation cascade, enabling a constant blood flow for feeding [4][5][6][7]. However, there are reports of rare incidents of human interactions which have led vampire bats to become more medically relevant to humans [8,9]. Outbreaks of rabies in human populations due to the vampire bats being vectors of the disease [10], have led to anti-vampire bat campaigns and culling of bat populations [11,12].
Previous studies have demonstrated that D. rotundus venom contains two important anticoagulant toxins: Draculin [6,7,13]; and DSPA (Desmodus rotundas salivary plasminogen activator) [14,15]. Draculin is a glycoprotein that irreversibly binds to factors IXa and X, and inhibits the conversion of prothrombin to thrombin [6,7,13]. This prevents fibrinogen being converted into fibrin and thus inhibits coagulation of blood during feeding [5]. DSPA components also aid in ensuring continuous blood flow by breaking up the fibrin mesh of any blood clots that are formed [16]. While there are relatively extensive studies on Draculin and DSPA, little is known about the other components of D. rotundus venom, with vasodilation a predicted but untested activity [15,16].
Other hematophagous animals induce anticoagulant and vasodilatory effects through the delivery of bioactive compounds, thus ensuring efficient blood flow for feeding. For example, mosquitos possess tachykinin-like peptides (sialokinins) [17,18], whilst bedbugs possess nitrosyl-hemoproteins (nitrophorins) [19,20]. In addition, sand flies contain a potent vasodilator (maxadilan) that acts via the PAC1 receptor [21,22], and horse fly disintegrins inhibit platelet aggregation like those from snake venoms [23]. Interestingly, tick prostaglandins constrict blood vessels [24]. The maintenance of blood flow during feeding is a major rate limiting step and challenge for blood feeders to overcome. Therefore, the longer they take to feed, the higher the chances the host or prey will notice, making them more vulnerable [25]. Thus, due to the similarities in feeding mechanisms between hematophagous animals, it has been postulated that vasodilators may play a key role in the venom of D. rotundus, targeting skin capillaries, to complement coagulation inhibition [15,16].
However, such actions have remained speculative until the current study which demonstrated selective and potent action for resistance-like arteries. Previously we showed that the transcriptome and proteinaceous products of the D. rotundus hematophagous secretion glands are rich in calcitonin gene related peptide variants [26], which are similar in size and amino acid sequences to CGRP but with modifications in key residues ( Figure 1). CGRP is a potent vasodilator that acts via activation of CGRP1 receptors on either endothelial or smooth muscle cells [27][28][29][30]. The significance of this peptide type in relation to the obtaining of blood-meals, and the impact of residues, was tested in order to ascertain the role in securing blood-meals by D. rotundus. In this study, we have demonstrated that vCGRP also causes vasodilation of resistance-like arteries via similar pathways to CGRP but with greater selectivity.

Discussion
D. rotundus venom is well known to contain anticoagulating properties in order to facilitate blood feeding [26]. Indeed, a glycoprotein, Draculin, which inhibits activated coagulation factors IX (IXa) and X (Xa) has been isolated from D. rotundus venom [5]. In the current study, we isolated and characterised a peptide (vCGRP) from the venom, which is similar in size and amino acid sequence to CGRP found in humans and rats. CGRP is a potent vasodilator that acts via activation of CGRP1

Contribution of Potassium Channels to D. rotundus vCGRP and rCGRP Mediated Relaxation
Raising the extracellular concentration of K + to 30 mM markedly attenuated the relaxant response to D. rotundus vCGRP ( Figure 4A). Blocking voltage-dependent K + channels with 4-aminopyridine (1 mM) markedly attenuated D. rotundus vCGRP-induced relaxation, reducing the potency by approximately 30-fold (p < 0.05) and reducing the response at 10 nM to 53.7 ± 17.3% (p < 0.01). However, vasorelaxation to D. rotundus vCGRP was unchanged in the presence of the ATP-sensitive K + channel inhibitor, glibenclamide (10 µM), or the Ca 2+ activated K + channel inhibitor, TEA (1 mM). Similarly, vasorelaxation to rCGRP was attenuated in the presence of 30 mM K + or 4-aminopyridine (1 mM) yet unchanged in the presence of TEA (1 mM) or glibenclamide (10 µM) ( Figure 4B). receptors on either endothelial or smooth muscle cells [27,31]. Therefore, the aim of this study was to determine whether vCGRP also causes vasodilation via similar pathways.   We identified vCGRP as a dilator of rat small mesenteric arteries with a potency and efficacy similar to rCGRP. Importantly, like rCGRP, the vasorelaxation was attenuated by the CGRP1 receptor antagonist, CGRP8-37, indicative of an ability of the peptide to target this receptor to mediate its response though direct activation of CGRP1 receptors can be further supported by radioactive ligand binding assays in the future. Next we examined the role of endothelial cells in vasorelaxation via vCGRP. Given the vasorelaxation to vCGRP was unchanged following endothelial denudation or inhibition of nitric oxide synthase (by L-NAME), it is likely that vCGRP targets CGRP1 on vascular smooth muscle cells (VSMC) to cause endothelium-independent relaxation. In contrast, relaxation to rCGRP appeared to be, in part, dependent on endothelial-derived nitric oxide (NO) as the potency was attenuated following NOS inhibition. These findings highlight a potential point of difference with regard to CGRP derived from distinct species. Thus whilst an endothelium-dependent component of vasorelaxation to rCGRP has been observed in mesenteric [32] and retinal [33] arteries, we have demonstrated that vCGRP, like human CGRP [34], mediates relaxation via endothelium-independent mechanisms. This similarity in mechanism of action between human CGRP and vCGRP supports the notion of vCGRP becoming a potential candidate for therapeutic drug discovery.
Previous studies have also demonstrated that activation of CGRP receptors can lead to the activation of the guanylyl cyclase pathway (endothelium-dependent) or adenylyl cyclase pathway (endothelium-independent) [33][34][35][36][37][38]. However, the presence of ODQ (guanylyl cyclase inhibitor) or SQ22536 (adenylyl cyclase inhibitor), had no significant effect on rCGRP or D. rotundus vCGRP relaxation curves. Differences between CGRP endothelium-independent and -dependent mechanisms are related to the region, size of the vessel tested and species of CGRP. For instance, human or rat CGRP tested in pig coronary leads to increased cAMP and causes vasorelaxation via endothelium-independent pathways [34]. However, human CGRP tested in human vessels are endothelium-dependent [28].
Therefore, we next sought to characterise the mechanism(s) via which vCGRP mediates endothelium-independent relaxation. Our finding that raising the extracellular K + concentration to 30 mM markedly attenuated the relaxation to vCGRP suggests that the peptide modulates relaxation of rat small mesenteric arteries in part via activation of K + channels. Indeed, we identified an ability

Discussion
D. rotundus venom is well known to contain anticoagulating properties in order to facilitate blood feeding [26]. Indeed, a glycoprotein, Draculin, which inhibits activated coagulation factors IX (IXa) and X (Xa) has been isolated from D. rotundus venom [5]. In the current study, we isolated and characterised a peptide (vCGRP) from the venom, which is similar in size and amino acid sequence to CGRP found in humans and rats. CGRP is a potent vasodilator that acts via activation of CGRP1 receptors on either endothelial or smooth muscle cells [27,31]. Therefore, the aim of this study was to determine whether vCGRP also causes vasodilation via similar pathways.
We identified vCGRP as a dilator of rat small mesenteric arteries with a potency and efficacy similar to rCGRP. Importantly, like rCGRP, the vasorelaxation was attenuated by the CGRP1 receptor antagonist, CGRP8-37, indicative of an ability of the peptide to target this receptor to mediate its response though direct activation of CGRP1 receptors can be further supported by radioactive ligand binding assays in the future. Next we examined the role of endothelial cells in vasorelaxation via vCGRP. Given the vasorelaxation to vCGRP was unchanged following endothelial denudation or inhibition of nitric oxide synthase (by L-NAME), it is likely that vCGRP targets CGRP1 on vascular smooth muscle cells (VSMC) to cause endothelium-independent relaxation. In contrast, relaxation to rCGRP appeared to be, in part, dependent on endothelial-derived nitric oxide (NO) as the potency was attenuated following NOS inhibition. These findings highlight a potential point of difference with regard to CGRP derived from distinct species. Thus whilst an endothelium-dependent component of vasorelaxation to rCGRP has been observed in mesenteric [32] and retinal [33] arteries, we have demonstrated that vCGRP, like human CGRP [34], mediates relaxation via endothelium-independent mechanisms. This similarity in mechanism of action between human CGRP and vCGRP supports the notion of vCGRP becoming a potential candidate for therapeutic drug discovery.
Previous studies have also demonstrated that activation of CGRP receptors can lead to the activation of the guanylyl cyclase pathway (endothelium-dependent) or adenylyl cyclase pathway (endothelium-independent) [33][34][35][36][37][38]. However, the presence of ODQ (guanylyl cyclase inhibitor) or SQ22536 (adenylyl cyclase inhibitor), had no significant effect on rCGRP or D. rotundus vCGRP relaxation curves. Differences between CGRP endothelium-independent and -dependent mechanisms are related to the region, size of the vessel tested and species of CGRP. For instance, human or rat CGRP tested in pig coronary leads to increased cAMP and causes vasorelaxation via endothelium-independent pathways [34]. However, human CGRP tested in human vessels are endothelium-dependent [28].
Therefore, we next sought to characterise the mechanism(s) via which vCGRP mediates endothelium-independent relaxation. Our finding that raising the extracellular K + concentration to 30 mM markedly attenuated the relaxation to vCGRP suggests that the peptide modulates relaxation of rat small mesenteric arteries in part via activation of K + channels. Indeed, we identified an ability of vCGRP to activate voltage-dependent K + channels as relaxation responses were decreased by 4-AP. This was in agreement to findings with respect to rCGRP. Neither K ATP nor K Ca channels appeared to be involved in relaxation to vCGRP or rCGRP as glibenclamide and TEA were without effect. Indeed, there is evidence that activation of CGRP receptors could lead to direct opening of K + channels, in particular Kv channels [33]. There are conflicting reports on the involvement of K ATP and K Ca channels in vasorelaxation, which could be related to the type of vessel studied. For instance, studies using bovine retinal arteries and rabbit mesenteric arteries report that activation of K ATP channels, but not K Ca channels, leads to vasorelaxation [37,39,40]. However, studies in smooth muscle cells from rat mesenteric arteries have shown CGRP directly activates BK Ca channels [41]. These data further highlight that CGRP causes vasorelaxation through a variety of mechanisms which is dependent upon the species and vessel involved.
Considering the medical relevance to humans of D. rotundus and other vampire bat species as disease vectors for rabies [1], it is surprising that more in depth studies have not been conducted on the intricate mechanisms employed in their feeding behaviour, despite studies on other blood feeding animals such as fleas and leeches [13,26,42]. Such secretions fit within the definition of venom as 'A secretion produced in specialized cells in one animal, delivered to a target animal through the infliction of a wound and that disrupts endophysiological or biochemical processes in the receiving animal to facilitate feeding, defense or competition by/of the producing animal' [42]. As peptides used by venoms/hematophagous-secretions are modified versions of those routinely expressed in other tissues [43] future work including the other two species of vampire bat and non-hematophagous bats would be enlightening in regards to the timing of the recruitment for use in blood-feeding and the molecular diversification events. This study has opened the way for further research to investigate the pathways and intricate mechanisms of hematophagous venoms, in particular vampire bats. Therefore, we have made clear the ability of vCGRP to selectively mediate endothelium-independent vasorelaxation in part via activation of Kv channels.
This selectivity of vCGRP to target only vascular smooth cells (similar to that of human CGRP) highlights the interesting possibility that vCGRP may confer benefit in the context of cardiovascular diseases such as hypertension, heart failure and kidney diseases [44]. Further functional studies are required for vCGRP to become a therapeutic intervention with potential pharmacological applications. This research also paves the way for further evolutionary studies into hematophagous venoms.

Materials and Methods
Synthesis of vCGRP was accomplished using protocols previously described by us for other peptides [45].

Data Analysis and Statistical Procedures
Relaxation responses were expressed as a percentage reversal of the U46619 pre-contraction. Individual relaxation curves were fitted to a sigmoidal logistic equation and pEC 50 values (concentration of agonist resulting in a 50% relaxation) calculated and expressed as -Log mol.L −1 . Statistical comparisons between the experimental groups' mean pEC 50 and maximum relaxation (R max ) values were made using a Student's unpaired t-test or one-way ANOVA with Bonferroni's post hoc comparison. Where pEC 50 values could not be obtained, concentration-response curves were compared by means of a two-way ANOVA. n = number of artery segments from separate animals. Data represent the mean ± SEM (error bars on graph). Statistical significance was defined as * p < 0.05. All data analysis was performed using GraphPad Prism version 5.02 (GraphPad Software, San Diego, CA, USA, 2009) [46].

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