*2.3. Proteomic Characterization of Patient Blister Fluids*

A total of 647 proteins were identified in the blister fluid from the five patients (Supplementary Table S1). Cluster analysis of these data did not provide any correlation to the clinical nor laboratory observations (data not shown). Proteins found in all of the blister fluids were compared to the control serum and included DAMPs, immunomodulators, complement, and ECM proteins (Table 2). Most of the complement proteins were of a lower abundance in the blister fluid than in the control serum, which is not surprising given that it is likely much of the complement system was activated and consumed by envenomation [25]. Interestingly, as observed in our previous studies [12,13,21], there was an increase in a variety of DAMPs in the fluids compared with the control serum, most notably the S100 protein family. In addition, several patients' fluids showed an increase in ECM components, such as the collagens. Another observation was the increase in matrix metalloproteinase-9 (MMP-9) in four of the five patients' blister fluids (Table 2). As MMP-9 has been implicated in blister formation, this was to be expected [26,27].


**Table 2.** Most abundant extracellular matrix proteins, immunomodulators, and damage-associated molecular patterns (DAMPs) identified in blister fluids.


**Table 2.** *Cont.*

The numbers in each cell are the abundance, according to the normalized spectra count. The red boxes represent values twice higher; blue boxes represent values twice lower; white boxes represent values without significant difference compared to the normal human serum.

> Following the molecular analysis of the endogenous factors, which likely contribute to local damage, we identified the presence of immunomodulators and DAMPs. These were generally observed to be at a higher abundance in the blister fluid than the normal serum (Table 2), reflecting the fact that DAMPs and immunomodulators are well-described markers associated with a pro-inflammatory effect. The presence of immunomodulators and DAMPs in blister fluid suggests that the local microenvironment could at some level be contributing to the blistering phenomenon.

> Thus, from the analyses of the local effects, it suggests that blistering is a molecular process that may involve some snake venom toxins, from which SVMPs are good candidates. SVMPs are responsible for ECM protein degradation and the subsequent release of endogenous pro-inflammatory molecules, which increase vessel permeability and contribute to tissue damage and blister formation [21,28,29]. In this context, in the blister contents we found some products associated with ECM degradation, DAMPs, and immunomodulators triggered by snake venom toxins. Rucavado and colleagues (2016) investigated the presence of proinflammatory molecules in the exudate generated from the action of snake venoms in experimental models early after venom injection. Proteomics studies showed the presence of cytokines and DAMPs released after 1 h of snake venom inoculation. In addition, the injection of these same exudates into mice skin increased vascular permeability. These findings suggest that DAMPs and cytokines present in the exudate may interfere with hemostasis through inflammatory pathways, such as Toll-like receptors (TLRs) [21].

> The molecular mechanism for tissue damage and blister formation in SBEs by activating the cascade of immunomodulators and TLRs can also be related to endogenous proteinases (MMPs). Endogenous proteinases can be released by infiltrating inflammatory cells triggering proteolytic and inflammatory cascades, which indicates a crucial role of a metalloprotease in wounds [27,30]. In bullous pemphigoid disease, an increasing perivascular inflammatory infiltration and activity of endogenous metalloproteinases, especially MMP-9 enhancing blisters, has been observed [26,27]. The presence of MMP-9 in the blister content from our patients suggests a combined action of endogenous and exogenous proteinases in pathophysiological events. MMPs can be up-regulated by the activation of Toll-like receptor pathways (TLR-4 and TLR-2) [31]. This leads to the speculation that MMP-9 in the blister may act on local ECM proteins and contribute to the blistering effect.

#### **3. Discussion**

In general, there is a paucity of studies on snake venom-induced blistering, although when observed in patients they can be quite dramatic, even though they do not significantly contribute to venom morbidity. In one study using an animal model of toxin-induced blister formation, a PI SVMP, BaP1, from *Bothrops asper* was shown to cause blister formation within an hour of injection of 80 μg of the toxin into mouse muscle [14]. In the wound exudate, endogenous MMPs and the toxin were identified along with several extracellular matrix fragments. This suggested to those authors that the SVMP and the endogenous MMPs likely were involved in blister formation via proteolytic degradation

of matrix proteins involved in epidermal/dermal association. Another investigation of blister formation utilizing the same toxin, but in a mouse ear model, also produced a rapid formation of blisters [32]. The authors attributed blister formation to the direct proteolytic action of the SVMP on known substrates at the dermal–epidermal junction, such as collagen type IV and laminin. In spite of the use of these elegant models, it is difficult to expect that they fully reflect the situation of blister formation in envenomed humans. In the mouse studies, the authors used a relatively high concentration of the toxin, which could be why blisters formed so rapidly in those models and thus could be attributed to a direct effect of the toxin on key proteins involved with dermal/epidermal association. Furthermore, mouse skin is quite different from human skin, primarily in that human skin is significantly thicker and firm, whereas mouse skin is thinner and loose, being comprised of only two or three cell layers in the epidermis compared to five to ten in humans [33]. Therefore, it seems unlikely that the mechanism of blister formation by the direct action of SVMPs at the dermal–epidermal junction is the same in humans, and thus this warrants different considerations.

In this study, we identified the presence of immunomodulators and DAMPs in blister fluid. These were generally observed to be at a higher abundance in the blister fluid in comparison with normal serum, reflecting the fact that DAMPs and immunomodulators are typically associated with a pro-inflammatory effect, contributing, at some level, to blistering phenomenon. Another important observation was that both antivenom and venom reach and remain in the envenomation region for a long period. However, the presence of antivenom in the blister did not prevent local tissue damage or blister formation. These observations highlight the immediacy of venom action at the site of envenomation activating endogenous pathways that promote tissue damage even after antivenom administration. In this study, it was evident that antivenom reaches the injured limb, and the activation of such endogenous pro-inflammatory pathways explains the general considerations that antivenom administration is relatively modest in its effectiveness in preventing local tissue damage in patients. Blister formation could be an independent process of the local protective effect of antivenom.

In this context, our results suggest that the pathophysiology of blister formation is more likely related to the generation of proinflammatory molecules (DAMPs and immunomodulators) by the toxins before antivenom administration. Although blisters do not play a significant role in morbidity, our investigation also illustrates how the initial action of the venom can contribute to pathophysiology that occurs at a relatively long period after envenoming, regardless of treatment with antivenom. The mechanism by which blisters are formed may be extended to other effects, leading to tissue damage in these patients, which may also be induced before antivenom administration. Finally, as has been long recognized, there remains an urgent need for therapeutic approaches that can rapidly be deployed at the envenomation site to attenuate the toxic activities that occur virtually immediately and have long-lasting effects, both locally and systemically. In this regard, our data strongly support the need for local treatments with anti-inflammatory drugs plus toxin inhibitors to prevent the severity of the wounds.

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