Animals and Study Design
The study was a prospective observational case-control study. Clinical cats were enrolled in accordance with the Institutional Animal Care and Use Committee (IACUC) at the University of California, Davis. Those exposed to wildfires were treated according to the best veterinary standard of care and those that served as controls were enrolled as part of IACUC protocols #20095 and #21037. All experimental protocols were approved by the IACUC of the University of California, Davis and all methods were carried out in compliance with the ARRIVE guidelines (http://www.nc3rs.org.uk/page.asp?id=1357) and regulations for care and use of laboratory animals. Informed owner consent was obtained prior to enrolling all client-owned cats in this study.
Clinical cats with thermal and smoke inhalation injury
Twenty-nine cats presented to the Veterinary Medical Teaching Hospital of the University of California, Davis for treatment of thermal burn injury and smoke inhalation due to the 2018 California Camp Fire and were enrolled in this arm of the study. The severity of thermal burn injuries was retrospectively scored as 1 (mild), 2 (moderate) or 3 (severe) based on clinical records and photographic documentation as previously described1. Medical treatment of wildfire-exposed cats was left to the clinicians’ discretion and detailed as previously reported1. All samples were collected within 7 days after presentation.
Healthy cats with or without subclinical hypertrophic cardiomyopathy
Thirty-two student, staff or client-owned cats were recruited during the period of October 2018 to June 2019. All cats received a physical examination, complete blood count (Abaxis, Union City, CA), total thyroxine (Abaxis, Union City, CA) and blood pressure measurement via Doppler with sphygmomanometery prior to echocardiography. Cats were excluded if they had a previous history of CHF, uncooperative temperament, apparent systemic diseases including abnormal findings in physical examination, hematology, systemic hypertension (systolic blood pressure > 160 mmHg) or hyperthyroidism (> 4.8 µg/dl). All cats were screened and confirmed for the presence of HCM with echocardiography and assigned to the HCM or CC groups. Oral administration of gabapentin (50 to 100 mg) prior to enrollment was permitted.
Echocardiography
Transthoracic echocardiographic examinations were performed by board-certified veterinary cardiologists (JAS, CGH) or cardiology trainees (ANS, YU) under direct supervision of a board-certified cardiologist. Standard imaging planes were used with the patient in right and left lateral recumbency. One of two machines were used with a 12-4mHz sector array transducer (Philips iE33 Ultrasound, Philips Health Care, Andover, MA). All measurements were evaluated by the investigators using commercially available software (Syngo Dynamic Workspace 10.0.01, Siemens Medical Solutions, Malvern, PA).
Diagnosis of HCM was made based on identification of idiopathic left ventricular hypertrophy characterized by regional or global end-diastolic wall thickening ≥ 6mm, determined by 2-dimenional or m-mode echocardiography while avoiding inclusion of moderator band insertion sites52. Myocardial thickening was defined as a diagnostic wall thickness ≥ 6mm. LA enlargement was measured in the right parasternal short axis basilar view and defined as an LA/Ao of ≥ 1.6. The LAuV was measured in the oblique left apical parasternal long axis view with pulsed wave Doppler sample volume at the entrance to the left auricle44. A LAuV of > 47 cm/s was considered normal44. M-mode was used to acquire measurements of LV internal dimension at end-diastole (LVIDd) and LV internal dimension at end-systole (LVIDs) and fractional shortening FS%) was calculated using the equation:
CHF in the WF group was diagnosed by a combination of clinical signs, echocardiographic findings of LA enlargement with the presence of pericardial and/or pleural effusion and radiographic findings consistent of cardiogenic pulmonary edema.
Blood sample collection
Blood was collected from the jugular or medial saphenous vein with a 22G or 23G butterfly catheter and divided among lithium-heparin tubes and 3.2% trisodium citrate tubes immediately after collection. Citrated blood samples were processed within 60 to 90 minutes after collection.
Generation of platelet rich plasma
Citrated blood was transferred to polypropylene tubes and rested for 30 minutes at 37°C to facilitate red cell sedimentation. The samples were then centrifuged at room temperature at 200x g for 5 mins, no brakes). Platelet rich plasma (PRP) was then separated and rested for an additional 30 minutes at 37°C prior to analysis. Swirling characteristic of the PRP was noted to ensure that platelets maintained their discoid shape. A platelet count of the PRP was attained with the automated blood cell analyzer (HM5, Abaxis, Union City, CA) and confirmed by blood smear evaluation when indicated.
Detection of platelet P-selectin by flow cytometry
PRP was diluted with Tyrodes HEPES buffer (pH 7.2, 5 mM dextrose, without divalent cations) to a final concentration of 1x107 platelets/ml. Platelets were either unstimulated (resting) or activated in the presence of 20µM ADP (Sigma-Aldrich, St. Louis, MO), or 0.01U/ml bovine α-thrombin (Haematologic Technologies, Essex Junction, VT) for 15 minutes at 37°C. Platelet P-selectin was detected by fluorescein isothiocyanate-conjugated rat anti-mouse monoclonal antibody to CD62P (1:200, clone:RB40.34, BD Pharmingen, San Jose, CA). Platelet integrin-β3 was identified using allophycocyanin-conjugated mouse anti-human monoclonal antibodies to CD61 (1:500, Clone: VI-PL2, eBioscience, San Diego, CA). Samples were subsequently incubated in the dark at 37°C degrees for 45 minutes before fixation with 1% paraformaldehyde at room temperature for 30 minutes. Samples were then analyzed using a 5-color flow cytometer (FC500, Beckmann Coulter, Miami, FL).
Platelets were identified by CD61-positive events as well as forward- and side-scatter characteristics using 0.9µm and 3.0µm calibration beats as previously described 37. Gating was established using fluorescence minus one controls consisting of unstimulated or activated platelets labelled with either 1 of the antibodies (Fig. 1A,B). Fluorescence compensation was applied using anti-mouse isotype controls conjugated to matched fluorophores (BD Biosciences, San Diego, CA) in identical experimental conditions for calculations of compensation matrixes. Flow cytometry data were analyzed using commercially available software (Flowjo, TreeStar Inc, Ashland, OR). Surface P-selectin expression on platelets was quantified as percentage of CD62P-positive events out of 10,000 platelets or median fluorescence intensity (MFI), an indicator of protein density.
In WF cats, platelet activation in response to 1µM AA (Roche) was assessed by treating PRP with 1µM AA for 15 minutes at 37°C. Unexpectedly, AA led to platelet inhibition as indicated by a significant decrease in P-selectin positive platelets compared to resting platelets (3.87%; IQR: 0.41–32.40 vs. 63.80%: IQR: 28.75–71.35; p < 0.0001).
Detection of platelet-derived microvesicles by flow cytometry
Platelet poor plasma (PPP) from 28 WF and 8 cats from the CC group was generated from PRP by further centrifugation at 5000 x g for 15 minutes. Supernatant was flash frozen in liquid nitrogen and stored at -80°C until further analysis. PPP was thawed at room temperature and only the top 90% of PPP was used for analysis. Positive controls were generated by treating PRP (1x107/ml) from healthy controls with 2mM calcium chloride (CaCl2) over a 60-minute interval (1mM CaCl2 every 30 minutes) and 2.5µM A23187 (Millipore Sigma, Burlington, MA) for 15 minutes at 37°C (Fig. 5B,C). To detect PMV, PPP was incubated in the dark at room temperature for 30 minutes with phycoerythrin-conjugated monoclonal mouse IgG1 antibodies to CD61 (1:10, Clone: VI-PL2, eBioscience, San Diego, CA) and fluorescein isothiocyanate-conjugated Annexin-V (1:32), diluted in binding buffer (pH 7.4, 100 mM HEPES, 140 mM NaCl, 25 mM CaCl2 (BD Biosciences, San Jose, CA). Prior to analysis by flow cytometry (FC-500, Beckman Coulter, Brea, CA), samples were further diluted in binding buffer (1:10) and agitated on a vortex for 5 seconds. Gating to quantify PMV were set by CD61-positive events and Annexin-V positive scatter characteristics using 0.5µm and 3.0µm calibration beads (Megamix, Biocytex) as previously described (Fig. 4)20. PMV was quantified as a percentage of CD61- and Annexin-V positive events and also by absolute counts per µl of PPP. Flow cytometry data were analyzed by commercially available software (Beckman Coulter, Brea, CA).
Statistics
Assuming biologic variability of 20%, calculations from our preliminary data and standard deviation of data gathered from the 29 WF cats showed that a sample size of at least 8 normal cats and 20 HCM cats would be required to demonstrate significant changes with an alpha priori of 0.05 and a power of 80%. Data was assessed for normality with Shapiro-Wilk normality test. Categorical data between groups were compared using a Fisher’s exact test or Chi-square analysis. Continuous data between groups were analyzed using either one-way ANOVA or Kruskal-Wallis test, followed by Dunnett’s or Dunn’s multiple comparisons test as appropriate. Unpaired continuous data were analyzed using student t-test for normally distributed data or Mann-Whitney test for non-parametric data. Spearman’s rank correlation analysis was performed between resting and activated platelets. Univariate and multiple logistic regression analyses were performed to predict the development of clot formation and survival by burn severity (1, 2 or 3), platelet count, neutrophil count and degree of platelet activation (P-selectin MFI) and the magnitude of response to ADP, AA and thrombin based on MFI fold change (log10 MFI activated – log10 MFI rest) on a log10 scale. A forward selection technique was performed by including the variable with P values < 0.1. To test the adequacy of the multivariable regression model, a Hosmer-Lemeshow goodness of fit test was performed. Normally distributed data were presented as mean ± standard deviation while non-parametric data were presented as median and interquartile range (IQR). Data were analyzed using commercially available software (Prism 8, Graphpad, San Diego, CA and STATA v15.1, College Station, Texas).