Abstract

   Biofluid proteomics is a sensitive and high throughput technique that
   provides vast amounts of molecular data for biomarker discovery. More
   recently, dried blood spots (DBS) have gained traction as a stable,
   noninvasive, and relatively cheap source of proteomic data for
   biomarker identification in disease and injury. Snake envenomation is
   responsible for significant morbidity and mortality worldwide; however,
   much remains unknown about the systemic molecular response to
   envenomation and acquiring biological samples for analysis is a major
   hurdle. In this study, we utilized DBS acquired from a case of lethal
   rattlesnake envenomation to determine the feasibility of discovering
   biomarkers associated with human envenomation. We identified proteins
   that were either unique or upregulated in envenomated blood compared to
   non‐envenomated blood and evaluated if physiological response pathways
   and protein markers that correspond to the observed syndromes triggered
   by envenomation could be detected. We demonstrate that DBS provide
   useful proteomic information on the systemic processes that resulted
   from envenomation in this case and find evidence for a massive and
   systemic inflammatory cascade, combined with coagulation dysregulation,
   complement system activation, hypoxia response activation, and
   apoptosis. We also detected potential markers indicative of lethal
   anaphylaxis, cardiac arrest, and brain death. Ultimately, DBS
   proteomics has the potential to provide stable and sensitive molecular
   data on envenomation syndromes and response pathways, which is
   particularly relevant in low‐resource areas which may lack the
   materials for biofluid processing and storage.
     __________________________________________________________________

Abbreviations

   DBS
          Dried blood spot

   DOC
          Sodium deoxycholate

   TCEP
          Tris (2‐carboxyethyl) phosphine

   LC‐MS/MS
          Liquid chromatography‐tandem mass spectrometry

   NDBS
          Nonenvenomated dried blood spot

   EDBS
          Envenomated dried blood spot

   NETs
          Neutrophil extracellular traps

   CSR
          Cytokine storm reaction

   DIC
          Disseminated intravascular coagulation

   HSPs
          Heat shock proteins

   SVMPs
          Snake venom metalloproteases

   CRP
          C‐reactive protein

   VWF
          Von Willebrand factor

   CSF
          Cerebrospinal fluid

   CAND1
          Cullin‐associated and neddylation‐dissociated 1 protein

   CNS
          Central nervous system

   KLKB1
          Plasma kallikrein

   TGF‐β
          Transforming growth factor β

1. INTRODUCTION

   Blood is a rich source of valuable information regarding health status,
   pathological syndrome, and disease progression. Blood and plasma
   proteomics methods represent sensitive and high throughput techniques
   that can provide vast amounts of molecular data which has been used for
   the discovery and quantification of a myriad of human disease
   biomarkers.[40] ^1 , [41]^2 , [42]^3 , [43]^4 , [44]^5 Utilizing dried
   blood spots (DBS) for proteomic identification of biomarkers
   circumvents many of the drawbacks of biofluid collection and handling
   while providing a relatively cheap, stable, and minimally invasive
   method for collecting blood, even in a non‐clinical setting.[45] ^6 ,
   [46]^7

   Snake envenomation remains a worldwide scourge that primarily affects
   tropical and subtropical developing countries and was re‐classified as
   a category A Neglected Tropical Disease by the
   World Health Organization  in 2017.[47] ^8 , [48]^9 Although antivenoms
   exist for many medically significant species, snakebite continues to be
   a significant cause of morbidity and mortality in these areas.[49] ^10
   , [50]^11 One major hurdle to improving outcomes is the production of
   effective antivenoms, the development of which should incorporate
   aspects of regionally relevant venom variation and an understanding of
   its relationship to the pathophysiology of envenomation.[51] ^10

   Snake venoms are complex toxic mixtures whose summative and synergistic
   effects trigger a cascade of physiological disruptions. Recent
   proteomic studies have investigated the effects of venoms or isolated
   toxins on mouse plasma,[52] ^12 plasma extracellular vesicles,[53] ^13
   wound exudate,[54] ^14 , [55]^15 , [56]^16 , [57]^17 and cardiac
   tissue,[58] ^18 each providing insight into the specific biological
   processes activated during envenomation. Bridging the gap between the
   data gathered in model organisms under experimental conditions and the
   information acquired in a clinical setting has the potential to provide
   a comprehensive picture of the mechanisms of pathophysiology triggered
   by envenomation as well as a suitable panel of potential biomarkers
   indicative of envenomation syndromes and outcomes.

   In this study, we utilized DBS acquired from an individual case of
   lethal rattlesnake envenomation to determine the feasibility of
   discovering biomarkers associated with human envenomation. We
   identified proteins that were either unique or upregulated in
   envenomated blood compared to non‐envenomated blood and evaluated if
   biological pathways and protein markers could be detected that
   correspond to the observed syndromes triggered by envenomation.

2. METHODS AND MATERIALS

2.1. Blood collection

   Whole blood was collected from a 6‐year‐old patient envenomated by a
   Prairie Rattlesnake (Crotalus viridis viridis) who rapidly declined
   within minutes of being envenomated and ultimately died several days
   later from the sequelae of the envenomation. Written consent for the
   use of blood for this analysis was obtained from the patient's parents
   and was verbally reconfirmed with them just prior to manuscript
   submission. The family requested no further details of events or
   medical treatment be discussed. Only information regarding the
   envenomation syndrome that could be gleaned from public sources was
   discussed in this paper.

   DBS from this patient were generated using whole blood drawn from the
   patient into EDTA‐coated tubes during treatment. A small volume of
   blood was spotted on a Whatman 903 Proteinsaver Card (Tisch
   Scientific). Blood spots were dried at ambient temperature for 24 h
   before storage in a re‐sealable zipper bag containing a humidity sponge
   with an indicating desiccant. Blood from a single healthy donor adult
   male was collected and processed in the same manner. Blood from this
   individual was used as a control in order to compare the blood
   proteomic profile of an envenomated individual to a non‐envenomated
   individual. Samples were processed for mass spectrometry 10 days
   following collection and spotting onto Whatman 903 Proteinsaver Cards.

   Two DBS discs (3 mm each) were punched using an Electron Microscopy
   Sciences (Hatfield, PA) Aluminum Punch Kit. Discs were added to 500 µl
   of 50 mM ammonium bicarbonate with 2% sodium deoxycholate (DOC) as in
   Eshghi et al.[59] ^6 Discs were incubated for 45 min at 37°C at 1400
   RPM on an Eppendorf Thermomixer C (Hampton, NH). To pellet extracted
   discs and particulates, samples were centrifuged for 15 s and 200 µl of
   supernatant was transferred to a new Eppendorf tube.

   The volume of supernatant corresponding to 30 µg of protein was diluted
   in 8 M urea/0.1 M Tris (pH 8.5) and reduced with 5 mM TCEP (tris
   (2‐carboxyethyl) phosphine) for 20 min at room temperature. Samples
   were alkylated with 50 mM 2‐chloroacetamide for 15 min in the dark at
   room temperature and then diluted four times with 100 mM Tris‐HCl (pH
   8.5) and trypsin digested overnight at 37°C using an enzyme/substrate
   ratio of 1:20. To stop the reaction, samples were acidified with 10 µl
   of 10% formic acid (FA), and digested peptides were purified with
   Pierce C18 Spin Tips (Thermo Scientific) according to the
   manufacturer's protocol. Samples were dried in a SpeedVac and
   redissolved in 0.1% FA.

2.2. Liquid chromatography‐tandem mass spectrometry analysis

   Liquid chromatography‐tandem mass spectrometry (LC‐MS/MS) was performed
   using an Easy nLC 1000 instrument coupled with a Q Exactive HF Mass
   Spectrometer (both from ThermoFisher Scientific). Peptides were loaded
   on a C18 column (100 µM inner diameter x 20 cm) packed in‐house with
   2.7 µm Cortecs C18 resin, and separated at a flow rate of 0.4 µl/min
   with solution A (0.1% FA) and solution B (0.1% FA in ACN) under the
   following conditions: isocratic at 4% B for 3 min, followed by 4%–32% B
   for 102 min, 32%–55% B for 5 min, 55%–95% B for 1 min and isocratic at
   95% B for 9 min. The data‐dependent acquisition was performed, and the
   top 15 most abundant precursors were considered for MS/MS analysis.
   Samples from the patient and the healthy control were each analyzed by
   LC‐MS/MS in technical triplicate.

2.3. Peptide and protein identification

   Fragmentation spectra were interpreted using the MSFragger‐based
   FragPipe computational platform[60] ^19 , [61]^20 against the UniProt
   human proteome (ID: UP000005640) database. Reverse decoys and
   contaminants were included in the search database. Cysteine
   carbamidomethylation was selected as a fixed modification, oxidation of
   methionine was selected as a variable modification, and precursor‐ion
   mass tolerance and fragment‐ion mass tolerance were set at 20 ppm and
   0.4 Da, respectively. Up to 2 missed tryptic cleavages were allowed and
   the protein‐level false discovery rate (FDR) was set to <1%. The
   relative abundance of proteins was compared across samples using
   log‐transformed spectral intensity in Metaboanalyst 5.0.[62] ^21

2.4. Gene ontology and network analysis

   Overlap between proteins identified in whole blood and DBS from the
   non‐envenomated individual (NDBS) or NDBS and envenomated DBS (EDBS)
   were determined using the Bioinformatics and Evolutionary Genomics Venn
   Diagram Tool (Figure [63]1;
   [64]http://bioinformatics.psb.ugent.be/webtools/Venn/). Proteins found
   to be unique between preparation methods were input for gene enrichment
   analysis into ShinyGO 0.76 with an FDR cutoff of 0.05, pathway size
   minimum of 2 and maximum of 2000 with redundancies removed using the
   gene ontology (GO) Biological Process Pathway and a KEGG database.[65]
   ^22 Unique proteins were also input into ELIXIR gProfiler
   ([66]https://biit.cs.ut.ee/gprofiler/gost) with Homo sapiens as the
   reference organism and statistical domain scope set to only annotated
   genes with a g:SCS significance threshold and user threshold of 0.05.
   Network analysis of upregulated proteins in EDBS was performed in
   Cytoscape 3.9.1[67] ^23 , [68]^24 using the plugin BiNGO 3.0.3[69] ^25
   with a hypergeometric statistical test, the Benjamini & Hochberg FDR
   correction at a significance level of 0.05 and using ontology terms for
   biological process. The output network was annotated to show broad
   groupings of similar processes.

FIGURE 1.

   FIGURE 1
   [70]Open in a new tab

   Workflow for the processing of dried blood spots from either
   envenomated or non‐envenomated blood for bottom‐up proteomics and
   bioinformatic analysis for envenomation biomarker identification and
   associated pathway analysis.

3. RESULTS

3.1. Fresh blood versus DBS

   We identified a total of 683 proteins in healthy human whole blood and
   DBS (Figure [71]2A). Of these proteins, 550 (80.5%) were able to be
   detected in DBS. One hundred thirty‐three proteins (19%) were
   identified in whole blood samples but not in DBS. The majority of the
   proteins unique to whole blood were assigned in Panther to the gene
   ontology molecular function terms associated with binding (53%;
   GO:0005488) and catalytic activity (35.7%; GO:0003824; Figure [72]2A).
   ShinyGO 0.76.2 identified GO terms associated with the biological
   process as proteasome protein deneddylation (n = 4, FC = 84.4),
   nucleotide‐excision repair/DNA damage recognition (n = 4; FC = 36.7),
   and protein activation cascade (n = 4, FC = 30.2) as the top 3 most
   enriched pathways.

FIGURE 2.

   FIGURE 2
   [73]Open in a new tab

   Shotgun proteomics from dried blood spots (DBS). (A) Comparison of
   proteomic coverage between DBS and whole blood from the non‐envenomated
   individual blood sample and identification and distribution of gene
   ontology (GO) molecular function terms in Panther of proteins unique to
   whole blood. (B) Comparison of proteomic profile between envenomated
   DBS (EDBS) and non‐envenomated individual (NDBS) and identification of
   top 20 pathways associated with proteins unique to EDBS using GO
   biological process terms in ShinyGO 0.76.2. (C) Heatmap cluster
   analysis of EDBS and NDBS run in triplicate showing clusters of
   overexpressed proteins in EDBS. (D) Volcano plot for identification of
   over‐ and underexpressed proteins in EDBS with a fold change > 2 and a
   significance threshold of 0.05. Significantly overexpressed proteins
   are indicated in red and significantly underexpressed proteins are
   indicated in blue. (E) Gene ontology and pathway enrichment analysis in
   ShinyGO 0.76.2 of the top 14 underexpressed proteins in DBS. (F) Gene
   ontology and pathway enrichment analysis of the top 20 overexpressed
   proteins in DBS. (G) Gene ontology analysis and abundance in Panther of
   all overexpressed proteins in DBS.

3.2. EDBS unique proteins

   We identified 480 proteins total from envenomated DBSs. In DBSs from
   both an EDBS and an NDBS, 436 proteins were common in both samples
   (73%; Figure [74]2B). Only 44 proteins (7.4%) were unique to
   envenomated DBS (Table [75]S1). Gene ontology analysis of unique
   proteins in EDBS showed significant enrichment for 92 pathways (Table
   [76]S2). The top 20 enriched pathways associated with unique proteins
   are primarily involved in an acute inflammatory response and immune
   system activation (Figure [77]2B). These include neutrophil activation
   and degranulation (n = 12, FC = 10), cell activation (n = 14, FC =
   4.2), and cell exocytosis (n = 16, FC = 7.7).

3.3. Significantly upregulated and downregulated pathways in EDBS

   Hierarchical clustering shows clear segregation of technical replicates
   of both NDBS and EDBS despite some variation between replicates
   (Figure [78]2C). There appear to be clear comprehensive differences
   between EDBS and NDBS including a cluster of highly upregulated
   proteins unique to the envenomated sample. Fold change analysis
   revealed the presence of 30 downregulated proteins and 118 upregulated
   proteins in EDBS compared to NDBS (Figure [79]2D and Table [80]S3).

   The 30 downregulated proteins in EDBS were associated with 158 pathways
   with the top pathways involving very‐low‐density lipoprotein particle
   clearance (Figure [81]2E and Table [82]S4; n = 2, FC = 393),
   chylomicron remnant clearance (n = 2, FC = 197) remodelling (n = 2, FC
   = 174) and assembly (n = 2, FC = 143) and protein‐lipid complex
   remodelling (n = 3, FC = 59). The identified downregulated marker
   involved in the kinin cascade is plasma kallikrein (KLKB1; FC = 0.14, p
   = 0.02). The markers involved in lipoprotein particle clearance and
   protein‐lipid complex remodelling were apolipoprotein C‐II (APOC2; FC =
   0.16, p = 0.01) and C‐III APOC3 (FC = 0.22, p = 0.09), and lipoprotein
   a (LPA; FC = 0.04, p = 0.01).

   Gene ontology analysis of upregulated proteins in EDBS showed
   significant enrichment for a wide variety of pathways (Tables [83]1
   and [84]2 and Table [85]S5). We identified 512 pathways with an FDR
   under 0.05 associated with the 118 identified upregulated proteins in
   EDBS. The top 20 enriched pathways for upregulated proteins
   (Figure [86]2F) were significantly enriched for markers associated with
   an acute inflammatory response and immune system activation, including
   neutrophil activation and degranulation (n = 31 genes, FC = 10), cell
   activation (n = 45, FC = 5), and cell exocytosis (n = 43, FC = 8).

TABLE 1.

   Upregulated markers associated with immune system processes organized
   by gene ontology (GO) biological process or KEGG pathway terms
   identified in ShinyGO 0.76.2
   Immune‐related pathway GO term Gene IDs
   Activation of immune response GO:0002253 C6 CD5L CTB HSP90AA1 PSMA6
   ACTR3 ARPC1B ACTR2 FCN3 PSMB4 ARPC5 MNDA C1QA ACTG1
   Inflammatory response GO:0006954 AHSG ANXA1 CD5L ELANE FN1 GSTP1 KRT16
   LBP LYZ PPBP PRDX5 PSMA6 PSMB4 S100A12 S100A8 S100A9 SERPINA3 SOD1
   THBS1 MSB4X
   Complement activation GO:0006956 C1QA C6 CD5L FCN3
   Neutrophil‐mediated immunity GO:0002446 B2M BPI CAMP CAT CTSG DSG1
   EEF1A1 ELANE GCA GSN GSTP1 SP90AA1 LBP LCN2 LTF LYZ MNDA PGM1 PKM PPBP
   PRTN3 RAB5C S100A12 100A8 S100A9 SERPINA3 TUBB4B VCL WDR1
   NET formation KEGG hsa04613 CTSG CAMP ELANE H4C9 H2AC20 H2BC12 VWF
   Platelet activation GO:0030168 FLNA HSPB1 TLN1 VCL VWF YWHAZ
   Platelet degranulation GO:0002576 ACTN1 AHSG F13A1 FLNA FN1 PPBP
   SERPINA3 SOD1 TF THBS1 TLN1 TMSB4X VCL VWF WDR1
   Cytokine‐mediated signalling GO:0019221 ANXA1 B2M CAPZA1 CTSG F13A1 FN1
   GSTP1 HSPA5 HSPB1 HSP90AA1 LBP LCN2 LCP1 PPBP PRTN3 PSMA6 PSMB4 SOD1
   TALDO1 TMSB4X VIM YWHAZ
   Cytokine production GO:0001816 LTF HSP90AA1 GSTP1 BPI HSPB1 FN1 LBP
   ANXA1 THBS1 LUM SOD1 MNDA TMSB4X B2M ELANE
   Cellular response to cytokine stimulus GO:0071345 VIM HSPA5 HSP90AA1
   GSTP1 CTSG PSMA6 ACTR3 FN1 CAPZA1 F13A1 LBP ANXA1 LCP1 THBS1 ACTR2 ATIC
   SOD1 GSN LCN2 PSMB4 MNDA PPBP CAMP YWHAZ TALDO1 ACTG1 TMSB4X B2M PRTN3
   Interleukin‐8 production GO:0032637 ANXA1 BPI ELANE LBP TMSB4X
   Response to interferon‐γ GO:0071346 ACTR3 ACTR2 ACTG1 B2M GSN VIM
   Interleukin‐1 β production ANXA1 GSTP1 HSPB1 MNDA
   Response to interleukin‐1 β GO:0070555 ANXA1 CAMP LCN2 PSMA6 PSMB4
   Response to interleukin‐12 GO:0071349 CAPZA1 LCP1 SOD1 TALDO1
   Tumour necrosis factor production BPI GSTP1 HSPB1 LBP LTF THBS1
   Response to tumour necrosis factor GO:0034612 CAMP GSTP1 LCN2 PSMA6
   PSMB4 THBS1 MSB4X
   IL‐17 signalling pathway KEGG hsa04657 HSP90AA1 LCN2 S100A8 S100A9
   [87]Open in a new tab

TABLE 2.

   Upregulated markers associated with various processes or pathologies
   organized by GO biological process or KEGG pathway terms identified in
   ShinyGO 0.76.2
   Pathway GO term Gene IDs
   Apoptosis GO:0042981 ACTN1 ANXA1 C1QA CAT CTSG ENO1 FCN3 FLNA GSN GSTP1
   HINT1 HSP90AA1 HSPA5 HSPB1 LTF MNDA PRDX3 PRDX5 S100A8 S100A9 SOD1
   THBS1
   YWHAZ
   Response to hypoxia/hypoxia‐inducible factor 1 signalling GO:0001666
   KEGG hsa04066 ALDOB CAT ENO1 PSMA6 PSMB4 THBS1 TF
   Response to oxidative stress GO:0006979 GSTP1 HSPB1 CAT PRDX5 ANXA1
   SOD1 LCN2 PRDX3 GPX3
   Blood coagulation GO:0007596 F13A1 FLNA FN1 HBG2 HSPB1 PRTN3 THBS1 TLN1
   VCL VWF YWHAZ
   Blood vessel development GO:0001568 ACTG1 ANXA1 ANXA3 CAMP FLNA FN1
   HSPB1 PKM THBS1 YWHAZ
   Brain development GO:0007420 ACTB ANXA3 ARPC5 ATIC B2M BASP1 FLNA
   HSPA5 UGP2
   Maintenance of the blood‐brain barrier GO:0035633 VCL ACTB ACTG1
   Pathways of neurodegeneration KEGG hsa05022 HSPA5 PSMA6 CAT SOD1 PSMB4
   TUBB4B GPX3
   Hypertrophic cardiomyopathy KEGG hsa05410 ACTB ACTC1 TPM4 ACTG1
   [88]Open in a new tab

   EDBS were also significantly enriched for proteins associated with the
   broad categories of an inflammatory response (n = 20, FC = 4.24;
   Table [89]1 and Table [90]S5), activation of immune response (n = 14,
   FC = 4.77), and complement activation (n = 4, FC = 8.5). Immune system
   pathways included platelet activation (n = 9, FC = 10.2; Table [91]1
   and Table [92]S5) and degranulation (n = 15, FC = 22), cytokine
   production (n = 15, FC = 3), signaling (n = 20, FC = 4.17), and
   response (n = 29, FC = 4.6). More specifically GO analysis in ShinyGo
   identified cytokine‐associated production and/or response pathways that
   were significantly enriched for interleukin‐1 beta, interleukin‐8,
   interleukin‐12, tumour necrosis factor and interferon‐gamma
   (Table [93]1 and Table [94]S4). ShinyGO KEGG pathway analysis (Table
   [95]S6) and network analysis in Cytoscape (Table [96]S7) identified
   interleukin‐17 and interleukin‐6 production, respectively. There was
   also significant enrichment in apoptosis pathways (Table [97]2; n = 20,
   FC = 2.5), blood coagulation (n = 14, FC = 7), cellular detoxification
   (n = 9, FC = 15), wound healing (n = 18, FC = 6), blood vessel
   development (n = 10, FC = 2.8), hemopoiesis (n = 13, FC = 2.5) and
   response to hypoxia (n = 5, FC = 2.7). Last, numerous nervous system
   and brain‐associated markers were significantly enriched, including
   astrocyte development (n = 4, FC = 14.6), blood‐brain barrier (n = 3,
   FC = 16.8), brain‐specific development (n = 9, FC = 2.2) and neuron
   development (n = 12, FC = 2).

   The markers with the highest positive fold change in EDBS are primarily
   associated with wound healing and blood coagulation, brain‐specific
   processes, apoptosis, and inflammatory processes including cytokine
   production and neutrophil activity (Table [98]S3 and [99]S5). The most
   upregulated markers of coagulation and brain‐specific processes were
   von Willebrand factor (VWF; FC = 625.16, p < 0.0001; Table [100]S3) and
   brain abundant membrane attached signal protein 1 (BASP1; FC = 95.8, p
   = 0.08), respectively. Numerous inflammatory markers associated with
   cytokine production and neutrophil activity had high fold enrichment.
   Markers of cytokine production included actin‐related protein 2 (ACTR2;
   FC = 39, p = 0.0012), bactericidal permeability‐increasing protein
   (BPI; FC = 75, p < 0.0001), lipopolysaccharide‐binding protein (LBP; FC
   = 50, p = 0.0026). Highly enriched markers of neutrophil activity
   included Ras‐related protein (RAB5C; FC = 59, p = 0.078), S100A8
   Protein (S100‐A8; FC = 46; p = 0.0004), S100A9 Protein (S100A9; FC =
   43, p = 0.0001), S100A12 Protein (S100A12; FC = 86, p < 0.001), and
   lysozyme C (LYZ; FC = 42; p < 0.0001). Markers of apoptosis included
   endoplasmic reticulum chaperone BiP (HSPA5; FC = 39; p < 0.0001).

   Analysis of upregulated proteins in Panther by protein class revealed
   that the majority of markers were associated with the cytoskeleton
   (Figure [101]2G; GO: PC00085; n = 21, 20%), defence and immunity (GO:
   PC00090; n = 23, 22%), and metabolite conversion (GO: PC00262; n = 22,
   21%). The majority of upregulated cytoskeletal elements were comprised
   of actin and actin‐binding cytoskeletal proteins (81%) including two
   myosin proteins (MYL12B and MYL6) and one actin specific to the cardiac
   muscle (ACTC1).

4. DISCUSSION

   In this study, we analyzed the proteomic profile of DBS in order to
   ascertain if biomarkers associated with the pathological events
   following a case of lethal human envenomation could be detected.
   Envenomation can result in the dysregulation of many systems in the
   body simultaneously, and human envenomations are a complex and evolving
   medical emergency. By compiling a panel of both upregulated and unique
   proteins compared to non‐envenomated blood, we identified a variety of
   affected pathways and flagged potential biomarkers indicative of severe
   envenomation in this case. We find multiple lines of evidence of a
   systemic and severe inflammatory response mediated by mast cell,
   neutrophil, and platelet activation and degranulation as well as a
   cytokine storm reaction (Figure [102]3). Further, we identified protein
   markers suggestive of lethal anaphylaxis, cardiac arrest, and
   neuroinflammation.

FIGURE 3.

   FIGURE 3
   [103]Open in a new tab

   Gene ontology (GO) term network analysis from Cytoscape. Node size
   indicates the number of markers associated with a given term and node
   colour indicates the p‐value. Nodes were annotated to show major
   biological systems and pathways affected.

   Gene ontology and network analysis illustrated a dramatic systemic
   effect triggered by envenomation with a dominant signal of immune
   system activation (Figures [104]2B,F and [105]3; Table [106]1). We also
   identified enrichment for pathways of apoptosis, blood coagulation, and
   those associated with a response to hypoxia along with markers that may
   be associated with brain damage and cardiac failure (Table [107]2). Our
   findings shed light on the specific mechanisms responsible for the
   syndromes observed following a lethal envenomation and demonstrate the
   utility of DBS to provide critical proteomic data on the
   pathophysiology of snakebites. This study is, to our knowledge, the
   first proteomic characterization of pathophysiology from envenomated
   human blood and the first investigation of envenomation biomarkers
   using DBS.

   In this specific case, envenomation by C. v. viridis triggered an
   apparent and lethal rapid anaphylactic reaction characterized by sudden
   collapse, cardiac arrest, and eventual brain death. It is unlikely that
   the initial inflammatory response was triggered by the specific action
   of venom toxins. Rarely, acute and severe anaphylactic reactions have
   occurred following venomous snake bites, including bites from
   rattlesnakes.[108] ^26 , [109]^27 Clinical manifestations of
   venom‐induced anaphylaxis include respiratory distress, hypotension,
   arrhythmia, shock, airway closure, and allergic myocardial
   infarction.[110] ^26 , [111]^28 , [112]^29 , [113]^30 Hypersensitivity
   in these cases occurred primarily in individuals with previous snake
   bites or long‐term exposure to venoms and is thought to be IgE‐mediated
   and therefore requiring previous sensitization.[114] ^31 , [115]^32
   However, cross‐reactivity with unrelated antigens could result in
   severe venom hypersensitivity in patients without prior exposure.[116]
   ^30 , [117]^33 , [118]^34

   Regardless of the mechanism of induction, immune system processes
   dominated the top 20 enriched pathways for both unique and upregulated
   proteins and we further identified multiple signatures associated with
   inflammatory signalling and response (Table [119]1 and Figure [120]2).
   Similar to previous studies on biofluids investigating the effects of
   envenomation, we identified differences in proteins associated with a
   thromboinflammatory reaction including activation of the complement
   system, coagulation, oxidative stress, neutrophil degranulation, and
   platelet degranulation.[121] ^12 , [122]^13 , [123]^35 Inflammation is
   a common response to envenomation. However, we did not detect the same
   proteins that were suggested to indicate the severity of inflammation,
   perhaps due to the development of anaphylaxis in this case as opposed
   to a less severe localized inflammatory reaction.[124] ^12 These
   authors attributed some of these inflammatory processes to the actions
   of venom toxins; however, it is also likely that early‐stage immune
   processes (i.e. platelet degranulation, neutrophil chemotaxis)
   triggered by envenomation are the result of a general defence response.
   The unravelling of the direct mechanisms of venom toxin action versus a
   general defence response to foreign antigens requires further study.

   Gene ontology analysis revealed enrichment for cellular pathways
   involved in early immune response events including mast cell and
   neutrophil activation and degranulation, as well as pathways triggering
   degranulation and aggregation of platelets, which have been shown to
   play an integral role in inflammation via crosstalk with leukocytes and
   the vascular endothelium.[125] ^36 Both mast cells and neutrophils are
   recruited during anaphylactic reactions depending on the type of
   stimulus and pathway, and neutrophil activation has been correlated
   with the severity of anaphylaxis in IgE‐dependent reactions.[126] ^37
   We also find evidence for the occurrence of NETosis, a pro‐inflammatory
   neutrophil‐mediated reaction that can occur soon after the onset of
   anaphylaxis.[127] ^37 , [128]^38 Neutrophil extracellular traps (NETs)
   are cytotoxic to vascular endothelium and likely synergistically
   contribute to the vascular extravasation that occurs during
   anaphylaxis.[129] ^39 , [130]^40 , [131]^41

   EDBS‐enriched proteins were closely associated with pathways involved
   in the production, signalling, and response to numerous inflammatory
   molecules linked with severe immune reactions, including histamine,
   cytokines (IL‐1 β, IL‐6, IL‐8, IL‐12, IL‐17, IFN‐γ and TNF;
   Table [132]1), growth factors (TGF‐β), and C‐reactive protein (CRP).
   Histamine released from mast cell degranulation is a key mediator of
   anaphylactic reactions and, depending on its downstream targets, is
   involved in the generation of nitric oxide and increased vascular
   permeability, cytokine synthesis, blood‐brain barrier function,
   bronchospasm, and cardiac contraction.[133] ^42 , [134]^43 , [135]^44
   Further, allergic reaction severity and delayed deterioration have been
   linked to elevated levels of histamine, IL‐6, and TNF‐1 receptor.[136]
   ^45 Mast cell expression of IL‐8 acts as a magnifier of the
   inflammatory cascade via paracrine effects on other leukocytes involved
   in inflammation,[137] ^46 and CRP and IL‐6 may act as late‐phase
   mediators of hypersensitivity reactions.[138] ^47

   The enrichment for proteins involved in the production and signalling
   of numerous cytokines suggests the occurrence of a cytokine storm
   reaction (CSR), caused by the release of large quantities of cytokines
   including IL‐8, IFNγ, TNFα, IL‐1β and IL‐6. In severe cases, this rapid
   release of inflammatory signals triggers increased endothelial
   permeability and vascular leakage, activation of the coagulation system
   causing a systemic syndrome with organ failure, disseminated
   intravascular coagulation, hypotension, hypoxemia, and cardiovascular
   collapse.[139] ^48 Specifically, IL‐6 has been proposed as a biomarker
   of CSR because of its correlation with reaction severity.[140] ^48 ,
   [141]^49

   Another mediator of inflammation, heat shock proteins (HSPs), are a
   family of stress‐induced molecules with cytoprotective actions that act
   as danger‐signalling biomarkers.[142] ^50 EDBS showed upregulation of
   HSP90 and HSP27 (HSPB1) and uniquely expressed HSP70 (HSPA5). Increased
   levels of HSP27 in plasma have been found after acute ischemic stroke
   [143]^51 and are released by myocardial cells after ischemia.[144] ^52
   HSP70 has been shown to stimulate both arms of the immune system and
   trigger the release of cytokines[145] ^50 (IL‐1β, IL‐6, IL‐12 and
   IFNγ). There is also evidence that these HSPs play a protective role
   during neuroinflammation by suppressing astrocyte inflammatory signals
   or interacting with ASK1 and inhibiting the ASK1‐MKK4‐JNK pathway.[146]
   ^51 , [147]^53 , [148]^54 , [149]^55 Notably, serum levels of HSP70
   have been directly linked to mortality in patients with CSR and septic
   shock.[150] ^56 This suggests that the presence of HSP70 was a strong
   indicator of the outcome in this case.

   VWF was an upregulated protein with the highest fold increase. VWF is a
   mediator of coagulation and clot formation and exerts significant
   influence on inflammation via neutrophil recruitment, regulation of
   endothelial wall permeability, and altering leukocyte
   extravasation.[151] ^57 Another mediator of coagulation and
   inflammation, plasma kallikrein, KLKB1, was significantly
   downregulated. KLKB1 deficiency has been linked to delayed
   clotting,[152] ^58 while high levels of VWF have been linked to
   thrombotic disorders,[153] ^59 suggesting overall systemic hemostatic
   dysregulation in this case.

   Von Willebrand factor is also a regulator of the response to
   endothelial cell damage, platelet activation, and platelet plug
   formation at the site of injury.[154] ^60 , [155]^61 , [156]^62 ,
   [157]^63 C. v. viridis venom in southern Colorado contains high levels
   of snake venom metalloproteases (SVMPs; Smith et al., in review), which
   are known to attack microvessel structure by proteolytically degrading
   the basement membrane and disrupting endothelial cell adhesion leading
   to vessel damage and fluid extravasation.[158] ^64 , [159]^65 ,
   [160]^66 This ultimately can result in local and systemic bleeding,
   tissue destruction, and ischemia.[161] ^67 , [162]^68 SVMPs also have
   been shown to play a role in blood coagulation activation,[163] ^69 ,
   [164]^70 apoptosis,[165] ^71 , [166]^72 , [167]^73 inhibition of
   platelet aggregation,[168] ^74 , [169]^75 and inflammation.[170] ^68 ,
   [171]^74 It is possible that the high levels of VWF are indicative of a
   response to vascular damage caused by the abundance of SVMPs in C. v.
   viridis venom and/or that SVMPs had a synergistic effect on
   inflammation, blood coagulation, and apoptosis with the anaphylactic
   reaction observed.

   There were multiple brain‐derived proteins identified in EDBS that have
   previously been investigated in CSF as biomarkers of neuropathies
   associated with Alzheimer's Disease (BASP1, YWHAZ, SOD1 and PKM), and
   YWHAZ specifically was identified as a strong candidate for the
   identification of neurodegeneration.[172] ^76 The high abundance of
   brain‐associated proteins in blood as opposed to CSF in addition to the
   presence of neuroprotective HSPs, may have resulted from the
   significant inflammation and damage to the blood‐brain barrier and CNS
   that occurred in this case.

   The downregulated protein with the highest degree of connectivity in
   NetworkAnalyst, CAND1 (Figure [173]S1; Cullin‐associated and
   neddylation‐dissociated 1 protein), plays a central role in cardiac
   functioning by preventing the accumulation of hypertrophic proteins,
   thereby protecting against stress‐induced cardiac hypertrophy and heart
   failure.[174] ^77 The direct role of this protein in cardiac function
   merits its investigation as a biomarker of anaphylaxis‐induced heart
   failure.

   While there is significant evidence for the pathways and markers
   elucidated here, there remain inherent limitations to the current
   study. First, this study only includes data from a single case of
   envenomation and a single non‐envenomated control; therefore,
   individual or age‐related variation could be responsible for some of
   the differences in protein presence or abundance observed when
   comparing blood profiles. While this is likely the case for a number of
   identified proteins, we detected numerous signatures in EDBS unlikely
   to result from individual differences in baseline healthy physiological
   processes (i.e. the abundance of HSP). Further, the sheer number of
   activated pathways and upregulated markers associated with inflammation
   and immune system activation is indicative of an overall trend, despite
   the potential for false positives. It is also unclear if the other
   significantly enriched pathways, including apoptosis, response to
   hypoxia, blood coagulation, and blood vessel development, are
   downstream effects of anaphylaxis or if they are the result of the
   action of venom toxins.[175] ^31 It is likely that the broad systemic
   effects of anaphylaxis masked the direct action of venom toxins, and
   ultimately, this limits our ability to correlate specific venom toxin
   action with clinical presentation in this instance.

   Biofluid proteomics is a robust tool for the investigation of the
   pathological events of envenomation and has promise for the development
   of biomarkers associated with envenomation syndrome. Here we
   demonstrate that DBS provide useful proteomic information on the
   systemic processes that are triggered by envenomation. We find evidence
   for a massive and systemic inflammatory cascade, combined with
   coagulation dysregulation, complement system activation, hypoxia
   response activation, and apoptosis. We also identified potential
   markers indicative of lethal anaphylaxis, cardiac arrest, and brain
   death. The utilization of proteomic data from DBS has significant
   implications for the investigation of envenomation syndromes in
   limited‐resource and rural settings, which bear the highest costs of
   snakebite envenoming worldwide. An enduring challenge will be to bring
   the benefits indicated by MS‐based analysis of DBS to those areas that
   could benefit most from this technology.

CONFLICT OF INTEREST STATEMENT

   The authors declare no conflict of interest.

Supporting information

   Supporting Information
   [176]ANSA-4-26-s008.png^ (9.2MB, png)

   Supporting Information
   [177]ANSA-4-26-s007.xlsx^ (17.4KB, xlsx)

   Supporting Information
   [178]ANSA-4-26-s004.xlsx^ (52KB, xlsx)

   Supporting Information
   [179]ANSA-4-26-s005.csv^ (5.4KB, csv)

   Supporting Information
   [180]ANSA-4-26-s001.csv^ (72.1KB, csv)

   Supporting Information
   [181]ANSA-4-26-s006.xlsx^ (84.1KB, xlsx)

   Supporting Information
   [182]ANSA-4-26-s003.csv^ (10.8KB, csv)

   Supporting Information
   [183]ANSA-4-26-s002.csv^ (74.4KB, csv)

ACKNOWLEDGEMENTS