**1. Introduction**

Snakebite envenoming (SBE) is classified as a neglected tropical disease that mainly affects rural areas in developing countries [1]. SBE victims display a wide array of symptoms and pathophysiology, largely depending on the type of snake that causes the wound, as well as intrinsic factors, such as the site of the bite, amount of venom injected, the health status of the victim, and the time elapsed before antivenom administration [2]. In Brazil, the northern states, which include the Amazon Forest, have the highest SBE prevalence. *Bothrops atrox* is the species responsible for the majority of SBE in the region, causing a significant detrimental economic and public health impact on the rural communities [3,4]. In this region, the typical local symptoms observed in *B. atrox* envenomation patients include edema, erythema, pain, and effects of tissue damage at the site of the bite, such as necrosis, inflammation and occasionally blistering in the proximity of the wound [5–7].

Blister formation following SBE, particularly in the case of *B. atrox,* occurs well after the bite [8,9]. Generally, blisters are related to a poor local prognosis as they increase the chance of infection and necrosis [10–13]. What is unknown is why there is such a significant delay in the formation of blisters following envenoming and why it appears antivenom treatment does not seem to prevent the appearance of blisters. This leads to a further question of what role venom components play in the pathophysiology of blister formation. It would seem unlikely that the venom components play a direct role in causing a blister given the delay in blister appearance, but perhaps it is possible they play an indirect role, whereby the actions of venom on the host ultimately lead to the production of agents or the activation of pathways that are known to produce blisters in other pathological conditions.

Some studies have attributed blister formation to snake venom metalloproteinases (SVMPs) [14]. SVMPs represent a class of zinc-dependent enzymes with molecular masses ranging from 20 kDa to 110 kDa [15]. This family of venom proteinases has been shown to give rise to the signature of the local and systemic hemorrhage associated with viperid SBEs, as well as playing a role in the recruitment of an immune infiltrate and the local production of cytokines and chemokines [16–18]. These activities are primarily the result of the proteolytic abilities of the SVMPs to degrade a variety of extracellular matrix (ECM) components, including laminin, fibronectin, nidogen, and collagen, giving rise to both structural and functional contributions to the pathophysiology of SBE [19,20]. These toxins have also been implicated in inflammatory reactions associated with envenomation at the onset of local tissue damage [18]. Furthermore, it is relevant that viperid venom in general and its SVMP components, in particular, can produce a wound exudate rich in damage-associated molecular patterns (DAMPs), which contribute to chemokine and cytokine production and tissue permeability [12,13,21]. As such, it is possible that this is a contributing factor by which venom and/or SVMPs may contribute to blister formation.

This report presents clinical and laboratory data derived from five *B. atrox*-envenomed patients focusing on blister production in these patients. These data are discussed to provide insight into blister formation in humans and how this may be important in other features of post-acute envenomation disabilities, and how this information may be considered for application in the clinical care of snakebite patients.

## **2. Results**

#### *2.1. Clinical Observations*

The clinical descriptions of the patients included in this study are presented in Table 1. The severity of the envenomation was determined from the patient clinical data, which were primarily focused on the extent of local edema throughout the bitten limb and systemic manifestations of the patient [22]. Moderate envenomation is characterized by significant pain, apparent edema that goes beyond the envenomed anatomical site, and sometimes the presence of blistering. Severe envenoming is classified by the presence of intense pain and extensive edema, involving the entire envenomed limb, often accompanied by blistering, secondary infection, severe bleeding, hypotension, shock, and acute renal failure [22]. Based on these observations, three patients in the cohort (P1, P2, and P3) were clinically

classified as suffering from "moderate" envenoming and two patients (P4 and P5) from "severe" envenoming. All patients, except one, were envenomed in the foot and there was a wide range of times recorded for the period of envenomation to hospitalization (2 h–11 h). Notably, all patients received antivenom at the time of admission. The times from envenoming until blister formation ranged from 63 h to 137 h. Surprisingly, a review of the clinical data did not show any correlation between the time of envenoming, hospitalization, antivenom administration, or envenoming severity with the time until the formation of blisters. However, this could simply be a reflection of the small sample size in this preliminary study.


**Table 1.** Clinical data of patients.

\* In all patients, blisters were hemorrhagic and appeared in the perilesional area, next to the bite. \*\* Pain (0–10) scale of 0 to 10 with 10 as highest; \*\*\* Edema was classified according to its extension, in segments as recommended by the Brazilian Ministry of Health [22]: mild = 1 to 2 segments, moderate = 3 to 4 segments, and severe = 5 affected segments. # Lactic Dehydrogenase normal value 190 UI/l; ## C-reactive protein normal value 0.8 mg/dL.

The clinical data also provided insight into the systemic and local symptoms experienced in this patient group. Firstly, from a review of the systemic markers, an increased inflammatory state after envenomation was observed in all patients, as evidenced by increased levels of C-reactive protein (CRP). All patients experienced increased CRP levels from time 0 (immediately before antivenom treatment), which confirms that all patients had an acute inflammatory condition following envenoming (Table 1).

From the analysis of the local symptoms, all five patients reported intense pain on the day of hospital admission into the following day. All were noted to have the presence of moderate to severe edema during the time of hospitalization. Interestingly, in this particular patient cohort, blister formation was not evident until after a minimum of 58 h from the time of hospital admission (Table 1). Furthermore, at admission, all patients presented elevated levels of lactic dehydrogenase (LDH) (>190 UI/I), a marker indicative of tissue damage, suggesting local damage was well underway at the time of admission and subsequent antivenom administration.

An important aspect common to the northern region of Brazil, and observed in some of these patients as well, is the long distances and difficulties in medical transport, hence the time from the accident to the patient being seen at the hospital is often longer than six hours [2]. The delays in patient care, along with the use of substances from traditional medicine and inappropriate practices, such as tourniquets, may aggravate the conditions at the bite site, leading to a high frequency of local complications resulting from the envenomings [9]. It is important to emphasize that one of our patients reported the use of tourniquets. Pardal and colleagues (2004) showed that tourniquet use is a common practice in *Bothrops* accidents [23]. Our data showed that the patient (Patient 4) who used a tourniquet before arrival at the hospital showed more severe inflammatory clinical parameters (Table 1). This observation highlights the importance of avoiding alternative methods of envenomation treatment, which may aggravate the patient's clinical condition [22].

In general, it is considered that antivenom administration is relatively modest in its effectiveness at preventing local tissue damage from SBEs. The reason for the lower local efficacy of antivenom has been attributed to the low likelihood of the antivenom reaching the tissues in time to neutralize the critical venom components involved in local damage. In this study, all patients received antivenom therapy well after envenomation and all had delayed blister formation. This suggests that the mechanism through which blisters are formed, regardless of their delayed appearance, happens very early in envenomation.

### *2.2. Laboratory Characterization for the Presence of Venom and Antivenom in Patient Serum and Blister Fluid*

Previously, we have shown the presence of venom proteins in human blister fluid resulting from snakebites [12], and therefore we investigated whether venom proteins and antivenom were present in this cohort of patients' blisters, as well as the relative amount of venom present in their serum. We used enzyme-linked immunosorbent assay (ELISA) to evaluate the venom protein concentration present in the serum of patients taken at the time of admission, before antivenom administration, and in their blisters that developed at much later times (Figure 1A). Not unexpectedly, the patients had varying venom concentrations in their serum, no doubt reflecting many different factors. Interestingly, patients 1, 3, and 5 who showed higher levels of serum concentration of venom were also the patients who had the shorter time intervals between envenomation and blister formation suggesting a higher venom amount injected at the time of the bite and thus potentially a more rapid and possibly higher production of the agents involved in subsequent blister formation. Another interesting observation from these data is that Patients 2 and 5 had higher venom concentrations in their blister fluid compared with their serum, suggesting that venom proteins were preferentially concentrated in the tissues adjacent to the wound.

**Figure 1.** Quantitative analysis of snake venom and antivenom by enzyme-linked immunosorbent assay (ELISA). (**A**) Venom concentration in the serum at the time of patient admission and in the blister fluid. (**B**) Antivenom concentration in the blister fluid. (**C**) Comparative analysis of *B. atrox* venom and antivenom in the blister fluid. Results are expressed as the mean ± sd of three independent readings. \*—*p*< 0.1; \*\*—*p* < 0.05, \*\*\*—*p* < 0.01.

Figure 1B shows the concentrations of antivenom measured in the contents of the blisters. The presence of antivenom was observed in blister fluids of all patients; however, the relative amounts of antivenom in the blister fluids were variable with Patients 1 and 4, being significantly higher than Patients 2, 3, and 5. Finally, we compared the protein concentrations of both venom and antivenom in the blister fluid (Figure 1C). In all cases, there appeared to be a higher protein concentration of antivenom compared with the venom in each blister, suggesting that the venom components are likely to be complexed with antivenom, with a significant amount of uncomplexed antivenom remaining in the blister. Hence, it is probable that most of the venom components found in the blister did not play a significant direct role in situ for blister formation, but acted to initiate a multi-factorial process well before the blister formation.

To determine whether uncomplexed antivenom was present in the blister fluids as well as to possibly get a sense of what venom components they could detect, we performed Western blots of the fluid against *B. atrox* venom. As seen in Figure 2, all patients' fluids collected from the blisters contained antibodies able to react with *B. atrox* venom components. A closer examination of the Western blots suggested the presence of antibodies in Patients 1, 2, and 4, which could recognize higher molecular mass targets consistent with PIII-class SVMPs (~50 kDa). Interestingly, none of the blister fluids had antibodies that recognized the targets in the region of the PI-class SVMPs (~20 kDa). One potential explanation for this is that the antibody population in circulation represents the antibodies that remain after neutralizing venom components early in antivenom infusion. Thus, it may be that much of the PI-SVMP neutralizing antibodies were depleted from the antivenom before fluid filling these blisters, which appeared well after envenomation and antivenom treatment. However, we also have to consider that the *Bothrops* antivenom presents a higher reactivity to PIII-class SVMPs, recognizing preferentially epitopes located at the Disintegrin-like/Cysteine Rich Domains [24]. Regardless, it is fascinating that both venom, likely in complex with antivenom, and uncomplexed antivenom were detected in the blister fluid. The role this may play, if any, in blister formation will be discussed below.

**Figure 2.** Recognition of *B. atrox* venom antigens by the antivenom present in the Blister fluids. *B. atrox* venom was electrophoresed under reducing conditions in 12% SDS-PAGE gels (Ven). Proteins contained in the gels were transferred to nitrocellulose membranes, which were incubated with the patients' blister fluids (1–5), and the capacity of the antivenom present in the blister content to bind to venom proteins was detected by chemiluminescence using peroxidase-labeled anti-horse antibody and substrate. The images were captured after 15 s of membrane exposure. The positive control (C+) was the commercial antivenom similar to the one used to treat the patients, and the negative control (C−) was the plasma from a volunteer who never received antivenom as treatment. The numbers at the left indicate the migration of molecular mass markers and at the right, the bands corresponding to PIII-class snake venom metalloproteinases (SVMPs), PI-class SVMPs, and Phospholipases A2 (PLA2).
