This study was performed at the College of Veterinary Medicine, Mississippi State University.
The Effects of Cyclosporine on Platelet Function and Cyclooxygenase Expression in Normal Dogs
Article first published online: 28 OCT 2012
Copyright © 2012 by the American College of Veterinary Internal Medicine
Journal of Veterinary Internal Medicine
Volume 26, Issue 6, pages 1389–1401, November/December 2012
How to Cite
Thomason, J., Lunsford, K., Stokes, J., Pinchuk, L., Wills, R., Langston, C., Pruett, S. and Mackin, A. (2012), The Effects of Cyclosporine on Platelet Function and Cyclooxygenase Expression in Normal Dogs. Journal of Veterinary Internal Medicine, 26: 1389–1401. doi: 10.1111/j.1939-1676.2012.01025.x
Abstract previously presented in part at the 2010 ACVIM Forum, Anaheim, CA, June 2010.
- Issue published online: 20 NOV 2012
- Article first published online: 28 OCT 2012
- Manuscript Accepted: 10 SEP 2012
- Manuscript Revised: 10 AUG 2012
- Manuscript Received: 18 OCT 2011
- Mississippi State University College of Veterinary Medicine Internal Competitive Research
Cyclosporine has been shown to alter platelet plasma membranes and have a hypercoagulable effect in humans, leading to thromboembolic complications.
Our hypothesis was that by modulating platelet reactivity, cyclosporine increases the risk of thromboembolic complications. The objective was to determine the effects of cyclosporine on primary hemostasis in normal dogs.
Eight healthy, intact female dogs.
A repeated-measures design utilized flow cytometry to evaluate platelet expression of platelet reactivity markers (P-selectin and phosphatidylserine) and COX-1 and COX-2 during the administration of 2 cyclosporine dosages (19 mg/kg q12h [immunosuppressive dosage] and 5 mg/kg q24h [atopy dosage]). Urine 11-dehydro-thromboxane-B2 (11-dTXB2) concentration was normalized to urine creatinine concentration, and platelet function was analyzed by PFA-100.
After a week of the immunosuppressive dosage, all platelet reactivity markers showed a significant decrease in mean fluorescent intensity (MFI). After the atopy dosage, only P-selectin and COX-2 MFI demonstrated a change from baseline, decreasing by 29% (P = .013) and 31% (P = .003), respectively. Urinary 11-dTXB2-to-creatinine ratio significantly increased at all time points during the immunosuppressive dosage, but no significant change occurred during administration of the atopy dosage. PFA-100 closure times using collagen/ADP cartridges increased by 62% (P = .008) with the immunosuppressive dosage and decreased by 45% with the atopy dosage (P = .035). No significant changes in closure times occurred with collagen/epinephrine cartridges.
Conclusions and Clinical Importance
Our study suggests that, similar to what is observed in humans, cyclosporine alters the platelet plasma membrane and increases thromboxane production in dogs, especially at immunosuppressive dosages.
analysis of variance
enzyme-linked immunosorbent assay
fluorescence-activated cell sorting–phosphate buffered saline
high-performance liquid chromatography
immune-mediated hemolytic anemia
mean fluorescence intensity
mitochondrial permeability transition pore
nonsteroidal anti-inflammatory drugs
platelet function analyzer
solid phase extraction
tumor necrosis factor-alpha
Cyclosporine has become a popular treatment for allergic dermatitis,[1, 2] anal furunculosis,[3-6] and several other immune-mediated and inflammatory diseases in dogs, including immune-mediated hemolytic anemia (IMHA). This is due partly to the perception that cyclosporine has a more favorable adverse effect profile when compared with other immunosuppressive agents. Recent studies in humans, however, have shown that cyclosporine can alter the platelet plasma membrane, increase the thrombogenic properties of platelets, and accelerate thrombin generation.[8-11] Approximately 50% of dogs with IMHA have been shown to be in a hypercoagulable state, predisposing these patients to complications such as pulmonary thromboembolism (PTE). The prevalence of PTE in dogs with IMHA has been determined to be over 30%.[5, 13] Owing to the difficulty in making an antemortem diagnosis, however, it is possible that the incidence of PTE in such patients may be as high as 80%. Minimizing risk factors for thromboembolic complications is important in the treatment of dogs with IMHA and other conditions that predispose to thrombus formation.
The exact mechanisms of PTE are not completely understood, but it is hypothesized that predisposing prothrombotic conditions involves both blood platelets and the clotting system.[12, 14] Results of several studies in humans have suggested that cyclosporine may enhance platelet reactivity,[9-11] a finding that necessitates a better understanding of how canine platelets interact with this medication, and if drug administration is a potential risk factor for PTE. A recent pilot study performed in our laboratory using a small number of dogs found increased platelet P-selectin expression (an adhesion protein with procoagulant and proinflammatory properties) after administration of cyclosporine. The administration of cyclosporine also has been associated with increased platelet expression of phosphatidylserine, which acts as a catalyst for thrombin generation and accelerates thrombus formation. Cyclosporine also has been shown to increase production of thromboxane A2 in humans. Thromboxane A2 is synthesized and released by activated platelets, and triggers vasoconstriction, enhanced platelet aggregation, and additional platelet activation. Platelet thromboxane A2 is generated by the enzyme cyclooxygenase (COX), which exists in 2 major forms, COX-1 and COX-2. Although it was long believed that platelets did not contain COX-2, recent studies have identified platelet COX-1 and COX-2 expression in both humans and dogs.[16-18]
This study was designed to evaluate the effects of cyclosporine on platelet function in normal dogs, specifically by evaluating PFA-100 closure time, platelet expression of P-selectin, phosphatidylserine, COX-1 and COX-2, and thromboxane production.
Material and Methods
Study Design, Animals
Eight healthy intact female Walker Hound dogs were used in this study. The dogs were not exposed to any medications or vaccines for at least 1 month before initiation of the study. The mean age of the dogs was 5.5 years (range, 1.5–6.5 years), and their mean body weight was 22.7 kg (range, 18.2–28.6 kg). Body weight was obtained at the beginning of the study and used to calculate all subsequent doses. Normal health status was established by detection of no abnormalities on physical examination, CBC (including manual platelet count), serum biochemistry, urinalysis, prothrombin time, partial thromboplastin time, plasma fibrinogen concentration, buccal mucosal bleeding time, von Willebrand factor (ELISA method), Babesia and rickettsial serology, and heartworm testing. One dog developed pyometra during the washout period between the 2 different dosing phases of the study, and only 7 dogs continued in the 2nd (atopy dose) phase of the study. Animal use was approved by the Mississippi State University Institutional Animal Care and Use Committee and was in compliance with the requirements at a facility accredited by the American Association for Accreditation of Laboratory Animal Care.
Two dosages of cyclosporine1 were administered PO for 7 days each, a high (immunosuppressive) dosage followed by a low (atopy) dosage, with a 2-week washout period between dosing. For the immunosuppressive dosage, cyclosporine was administered at 10 mg/kg PO twice daily, and blood samples were collected from each dog on day 4 of drug administration 12 hours after the previous dose to determine trough cyclosporine blood concentration by high-performance liquid chromatography (HPLC) analysis. A target immunosuppressive trough blood concentration was set at 600 ng/mL, and drug dosages were adjusted upward if needed to ensure that the trough blood concentrations exceed this target level by the time of sample collection on day 7. The mean administered immunosuppressive dosage of cyclosporine was 10.1 mg/kg ± 0.48 (mean ± SD) mg/kg, PO, q12h, at the time of drug commencement, but mean dosage had increased to 19 ± 9 mg/kg, PO, q12h, by day 7. The mean administered dosage for the atopy dose of cyclosporine was 5.1 ± 0.21 mg/kg, PO, q24h. Drug doses were not adjusted during administration of the atopy dosage of cyclosporine.
Blood and urine samples were collected at regular intervals for platelet function testing, flow cytometric evaluation of platelet P-selectin, phosphatidylserine, COX-1 and COX-2 expression, and urinary thromboxane analysis. For both drug dosages, blood and urine samples were collected at baseline before cyclosporine administration, and on days 1 and 7 of drug administration. Samples were collected at 2 time points on each day of collection, at the estimated times of peak and trough cyclosporine concentrations. Samples were collected at estimated peak blood drug concentrations 2 hours after drug administration for both dosages, and at estimated trough blood drug concentration 12 hours after drug administration for the immunosuppressive dosage, and 24 hours after drug administration for the atopy dosage. All flow cytometric assays, thromboxane-to-creatinine ratios, and PFA-100 closure times were measured both before and at all tested time points during drug administration and compared to the baseline value for all assays. The initial baseline platelet phosphatidylserine expression result before the immunosuppressive dosage trial was considerably higher than at all subsequent time points, including the atopy dosage trial baseline. To confirm that this increased initial baseline result was not erroneous, blood samples were obtained for repeat analysis of platelet phosphatidylserine expression on 2 additional days after a prolonged washout time period.
Blood was collected by jugular venipuncture with a 20-gauge needle directly into a siliconized glass vacutainer tube containing 3.8% sodium citrate2 that had an approximate draw of 1.8 mL. Each sample was collected until the tube was completely filled to standardize the degree of anticoagulation with citrate. Sample preparation for flow cytometry was initiated within 2 hours of collection.
Previously described protocols[14, 20, 21] were modified for the detection of platelet P-selectin (CD62). Briefly, 5 μL of citrated whole blood was added to 45 μL of FACS-PBS and incubated in the dark at room temperature with an affinity-purified, monoclonal, mouse anticanine CD62 antibody.3 After incubation, samples were washed with PBS, pelleted by centrifugation, and resuspended. Goat antimouse-PE4 conjugate was added to all samples and incubated in the dark at room temperature for 30 minutes. For platelet identification, samples were incubated with monoclonal FITC-conjugated mouse antiporcine CD61,5 and fixed with 1% paraformaldehyde6 for 10 minutes at 4°C in the dark (Fig 1A).
A previously described protocol was used to detect platelet phosphatidylserine. Briefly, 5 μL of citrated whole blood was added to 250 μL of annexin-binding buffer.7 Samples were incubated with annexin-V-FITC8 and mouse antiporcine CD61-purified antibody9 for 30 minutes in the dark at room temperature. After incubation, goat antimouse IgG:RPE10 was added to the samples, followed by a 30-minute incubation. Samples were fixed with 1% paraformaldehyde for 10 minutes.
A previously described protocol was used to quantify platelet COX-1 expression. Briefly, 5 μL of citrated whole blood was added to 45 μL of FACS-PBS and incubated with ovine COX-1 specific monoclonal antibody.11 To identify platelets, samples were incubated with a monoclonal antibody to human CD912 (Fig 1B). Samples were fixed for 10 minutes at 4°C in the dark with 1% paraformaldehyde.
Analysis of canine platelet COX-2 expression was performed by a modification of a protocol used in humans that has been previously reported in dogs. Briefly, 5 μL of citrated whole blood was added to 45 uL of FACS-PBS and fixed at 4°C in the dark with 1% paraformaldehyde. Samples were washed, pelleted, resuspended, and incubated in 0.3% Triton X-10013 followed by another wash. Samples then were incubated with a monoclonal FITC-conjugated mouse anti-human-COX-2 antibody14 in the dark at room temperature. Platelets were identified with a monoclonal antibody to porcine CD61 and goat antimouse IgG:RPE antibody (Fig 1C).
All samples were stored in the dark at 4°C before flow cytometric analysis. Isotype-matched monoclonal antibodies were used for all marker specific antibodies. Flow cytometric analysis was performed with a flow cytometer15 with CellQuest Pro software.16 Platelet populations were displayed on log forward-scatter versus log side-angle light scatter plots. Gates were adjusted to baseline platelet populations, and 5,000 gated events were recorded for each labeling. Expression was quantified by the intensity of antibody fluorescence and expressed as mean fluorescence intensity (MFI). A histogram was created with MFI on the x-axis and platelet number on the y-axis.
Urinary 11-dehydro-thromboxane B2 (11-dTXB2) concentration was analyzed by a commercially available competitive enzyme immunoassay kit17 for a multiplex analyzer18 that has been previously validated in the dog. Urine was collected by cystocentesis with a 22-gauge, 1.5-inch needle with a 5-mL syringe. The volume of urine collected varied between 0.5 and 5 mL. All samples were batched, stored at −80°C, and thawed to room temperature for analysis. When the urine reached room temperature, and immediately before plating, each sample was vortexed thoroughly. The assay buffer was used to standardize the specific gravity of each sample to fit within the working range (1.003–1.012) of the analyzer, and a correction factor was applied to account for these dilutions. Samples were analyzed in duplicate according to the manufacturer's instructions, and reported in picograms per milliliter of urine. Briefly, a 96-well plate was prepared by adding 100 μL of each sample and 50 μL of both 11-dTXB2 phycoerythrin tracer and 11-dTXB2 XMAPR beads to each well. Before analysis, each plate was incubated at room temperature, in the dark and on an orbital shaker, for 4 hours. Urine creatinine concentration was measured with a biochemistry analyzer19 and the 11-dTXB2 concentration was normalized to create a urinary 11-TXB2-to-creatinine ratio.[22, 23]
Platelet Function Analysis
A commercial point-of-care platelet function analyzer (PFA-100)20 that has been previously evaluated for use in dogs was used to analyze platelet function.[24-27] The PFA-100 is an in vitro platelet function analyzer that mimics a high shear force over an area of vascular damage and stimulates platelet function with several platelet agonists to measure the time, in seconds, needed to form a platelet plug and inhibit blood flow. The cut-off time for the instrument is >300 seconds.
The instrument was used according to manufacturer's instructions. Briefly, samples were collected directly into 5-mL blood collection tubes containing 3.8% sodium citrate, and 800 μL of whole blood sample was transferred into PFA-100 cartridges (both collagen/ADP and collagen/epinephrine),2122 and analyzed. All samples were analyzed within 3 hours of collection. An automated hematologic analyzer23 was used to determine an accurate platelet count and hematocrit (HCT) on each sample. Cartridges were stored at 4°C and warmed to room temperature before analysis.
Measurement of peak and trough blood cyclosporine concentrations was performed by HPLC analysis based on a modification of the therapeutic drug monitoring assay used at the University of California at Davis.24 Briefly, blood was collected directly into tubes containing EDTA25 anticoagulant, batched, and stored at −80°C until analysis. Samples were thawed to room temperature and underwent an extraction procedure that mixed 2 mL of the whole blood sample with a protein precipitating solution that consisted of 5% zinc sulfate, 20% acetonitrile, 30% methanol, water, and an internal standard of 400 ng/mL of cyclosporine D.26 Samples were mixed, centrifuged, and the supernatant was added to a prepared C18 solid phase extraction (SPE) column.27 The SPE column, under vacuum, was washed with 5 mL of 50% acetonitrile and then the cyclosporine was eluted by 1 mL of 100% methanol. To the eluent, 200 μL of water followed by 300 μL of hexane was added. This sample was centrifuged at 14.5 revolutions per minute for 90 seconds and 200 μL of the aqueous layer was extracted and prepared for HPLC analysis. Only 100 μL of the sample was injected for HPLC analysis. Blank EDTA whole blood was used for a standard curve with cyclosporine28 at 0, 200, 400, 800, and 1600 ng/mL.
Cyclosporine analysis was performed by an Agilent 1100 HPLC system29 with degasser (G1322A), quaternary pump (G1311A), autoinjector (G1313A), and diode array detector (G1315). The reverse phase column was a Phenomenex Luna 5u C18 (2) (150 × 2.00 mm 5 μ), which was maintained at 75°C. A 1 mL per minute gradient mobile phase consisted of acetonitrile (A) and water adjusted to pH 3.1 (B). The gradient transitioned linearly from 65% A and 35% B to 70% A and 30% B over 5 minutes, and held for 15 minutes. After each injection, there was a 5-minute re-equilibration time period. Detection was at 200 nm. The retention time for cyclosporine and cyclosporine D was 4.2 and 5.6 minutes, respectively. The assay was linear over the standard curve range with an r2 of 0.98. The assay had an average coefficient of variation of 6.7% (range, 3.7–9.9%) and an average accuracy of 94.4% (range, 92–98%).
A repeated-measures design was utilized in this study. Visual assessment of the data using Q-Q plots and histograms with UNIVARIATE procedure of SAS for Windows 9.230 indicated that the data were not consistently normally distributed for all of the outcomes. Consequently, nonparametric methods for analysis of repeated measures were utilized. For all 4 flow cytometry assays, HCT, platelet count, PFA-100, and urine 11-dTBX2, midranks of the data were determined and analyzed by the methods outlined by Brunner and Madan and Shah and Madden.[28, 29] The midranked data were analyzed by the MIXED procedure of SAS for Windows 9.2 by the ANOVAF option and the MIVQUE0 estimation method for the covariance parameters and a REPEATED statement specifying an unstructured covariance structure. The ANOVAF option was used to provide analysis of variance type statistics (ATS), which were used to test the nonparametric hypothesis by contrasts of distribution functions.[28, 29] The ATS as opposed to Wald-type statistics were used because of the relatively small sample size. An unstructured covariance structure was assumed for the variances at each time point and between all pairs of time points. The MIVQUE0 estimation method was used because the data were not normally distributed and to prevent convergence issues. Each dosage of cyclosporine was analyzed separately from the other dosage, and the results were not compared. Sample time was included in the model as a fixed effect. For outcomes in which time had a significant effect (P ≤ .05), comparisons were made between baseline and the other time points by differences in least square means. Wilcoxon signed rank tests by the UNIVARIATE procedure of SAS for Windows 9.2 were conducted to compare baseline levels of outcomes before the immunosuppressive dosage and the atopydosage administration and between baseline levels before immunosuppressive dosage administration and levels after the washout period. A P value ≤.05 was considered significant for all analyses.
There were significant differences between time points in platelet P-selectin expression for both the immunosuppressive dosage (P < .001) and the atopy dosage (P < .001). All time points during the immunosuppressive dosage of cyclosporine, and both trough time points during the atopy dosage, were significantly lower than baseline levels (Fig 2).
There were significant changes over time in platelet phosphatidylserine expression during the immunosuppressive dosage (P < .001) of cyclosporine. Compared with baseline, phosphatidylserine expression decreased significantly at all time points during administration of the immunosuppressive dosage. The mean of the 2 additional repeat assays performed after washout still was greater than at all other time points, but the MFI was significantly (P = .008) less than the initial baseline result. A significant decrease in phosphatidylserine expression was observed at the peak time point and a significant increase occurred at the trough time point on day 1 (Fig 3).
There was a significant difference in platelet COX-1 expression during both the immunosuppressive and atopy dosage, P < .001 and P < .001, respectively. Compared with baseline levels, both trough time points and the peak time point on day 7 during administration of the immunosuppressive dosage of cyclosporine were significantly decreased. During the administration of the atopy dosage, there was a significant decrease in COX-1 expression at both time points on day 1 and day 7 peak (Fig 4).
There was a significant change in platelet COX-2 expression for both the immunosuppressive dosage (P < .001) and the atopy dosage (P < .001). Compared with baseline levels, there was an immediate significant increase in platelet COX-2 expression 2 hours after initiation of the immunosuppressive dosage of cyclosporine (day 1 peak time point). However, there then was a decline in platelet COX-2 expression until there was a significant decrease in COX-2 expression at both the peak and trough time points on day 7. During administration of the atopy dosage of cyclosporine, there was a significant decrease in platelet COX-2 expression 2 hours after drug initiation (day 1 peak time point). Platelet COX-2 expression then returned to levels comparable to baseline values 24 hours after administration of the first dose. Platelet COX-2 expression again was significantly decreased at the final time point (day 7 trough) (Fig 5).
There was a significant increase in urinary 11-dTXB2-to-creatinine ratio for the immunosuppressive dosage of cyclosporine (P < .001), but there was no significant change during the atopy dosage (P = .185). The median baseline urinary 11-dTXB2-to-creatinine ratio was 18.7 (range, 10.7–24.7) before the immunosuppressive dosage of cyclosporine. After administration of the immunosuppressive dosage, there was a significant increase in the urinary 11-dTXB2-to-creatinine ratio compared with baseline values at all time points. The peak time points on day 1 and 7 had the highest percentage increase in the 11-dTXB2-to-creatinine ratio, with increases of 1,027% (median ratio, 82.7; range, 15.6–623) (P = .004) and 1,461% (median ratio, 152.5; range, 47.8–799.3) (P < .001), respectively. There also was a significant increase in the baseline urinary 11-dTXB2-to-creatinine ratio at trough time points on days 1 and 7, with increases of 142.7% (median ratio, 41; range, 19.2–69.39) (P = .001) and 107.6% (median ratio, 28.1; range, 13.83–67.77) (P = .047), respectively (Fig 6). Urine was not available from 1 dog at the peak time period on day 1.
The median baseline urinary 11-dTXB2-to-creatinine ratio was 15 (range, 8.3–28.2) before the atopy dosage of cyclosporine. There was no significant change in the urinary 11-dTXB2-to-creatinine ratio compared with baseline values at any time point (Fig 6). Urine was not available from 1 dog at the peak time point on day 1, from 3 dogs at the peak time point on day 7, and from 2 dogs at each trough time point.
Platelet Function Analysis
The median baseline PFA-100 closure time using the collagen/ADP cartridge was 87 seconds (range, 69–172 seconds) before the immunosuppressive dosage of cyclosporine. There was a significant change in closure times with the collagen/ADP cartridge for both the immunosuppressive dosage (P = .0125) and the atopy dosage (P = .0436) of cyclosporine. There was an immediate significant (P = .006) increase in closure time 2 hours after the administration of the initial immunosuppressive dosage of cyclosporine (day 1 peak time point), on average by 100%. However, after this first time point, the mean closure time decreased to 30, 35, and 62% above baseline at the remaining 3 time points, respectively. The day 7 peak and trough time point closure times still were significantly increased (P = .005 and .008, respectively) (Fig 7).
The median baseline PFA-100 closure time using the collagen/ADP cartridge was 140 seconds (range, 67–234 seconds) before the atopy dosage of cyclosporine. There was a significant (P = .037) decrease in closure time 24 hours after administration of the initial atopy dosage of cyclosporine (day 1 trough time point), on average by 42% (Fig 7). Closure times remained significantly decreased for all remaining time points.
The median baseline PFA-100 closure time using the collagen/epinephrine cartridge was 164 seconds (range, 117–229 seconds) before the immunosuppressive dosage of cyclosporine, and 145 seconds (range, 130–254) before the atopy dosage. There was no significant association between closure times and sample time for either immunosuppressive dosage (P = .054) or atopy dosage (P = .472) of cyclosporine (Fig 8).
The median baseline HCT before the immunosuppressive dosage of cyclosporine was 44.4% (range, 41.1–52.6). There was a significant change in HCT for both the immunosuppressive and atopy dosages, (P < .001) and (P = .021), respectively. Significant differences in HCT during the immunosuppressive dosage occurred at the peak time points on day 1 and day 7, with relative percentage decreases from baseline of 4.8% (P = .020) and 9.2% (P = .028), respectively. The median baseline HCT before the atopy dosage of cyclosporine was 47.6% (range, 33.9–54.3). The only time point that demonstrated a significant difference in median HCT during the atopy dosage occurred at the peak time point on day 1, with a relative percentage increase of 11% (P = .019). There was no significant association (P ≥ 0.068) between platelet count and sampling time for either cyclosporine dosage.
Cyclosporine is a commonly used immunosuppressive agent in dogs, partly because of the perception of relatively minimal adverse effects.[7, 30] Several studies in humans, however, have suggested that cyclosporine can increase the risk of thromboembolic complications by altering the platelet surface membrane and increasing thromboxane production.[8-10, 31-33] We showed that cyclosporine does alter the platelet membrane, and is associated with a significant increase in thromboxane production.
We demonstrated a significant increase in the urinary 11-dTXB2-to-creatinine ratio during administration of an immunosuppressive dosage of cyclosporine, especially at peak cyclosporine blood concentrations. This effect appears to be dose-dependent, because a similar but lesser (but not statistically significant) effect was seen with the atopy dosage of cyclosporine. Our results are similar to those of previous studies in humans that demonstrate an increase in thromboxane production during cyclosporine administration in renal transplant and coronary artery disease patients.[9, 10, 31-33] When compared with renal transplant recipients treated with azathioprine, patients receiving cyclosporine had increased thromboxane production and increased rates of thrombus formation within the renal parenchyma. Unlike the normal dogs used in this study, most of the previous studies that demonstrated an increase in thromboxane concentration during cyclosporine administration took place in either renal transplant recipients or patients with coronary artery disease. With this difference in study populations, both vascular and systemic inflammation generated by the disease state in human patients could have influenced thromboxane concentrations. The exact mechanism for the increase in thromboxane is undetermined, and additional research would be needed to define the effects of cyclosporine on thromboxane production. After synthesis, thromboxane A2 is rapidly converted to several metabolites that then are eliminated in the urine, making it difficult to accurately measure plasma thromboxane concentrations. Measurement of urine concentrations of a stable thromboxane metabolite, 11-dTXB2, is considered to be a reliable marker of systemic thromboxane A2 production.[22, 35, 36] Our results therefore suggest that cyclosporine at a standard immunosuppressive dosage might predispose to platelet activation in dogs.
Although they generate substantial amounts of thromboxane A2, platelets are not the only source of thromboxane synthesis. Neutrophils, monocytes or macrophages, smooth muscle cells, and endothelial cells are additional sources of thromboxane. Some of the increases in urinary thromboxane concentration seen in our study could have been associated with these additional cells, and not solely platelets. The kidney also is capable of synthesizing thromboxane, but this thromboxane usually is excreted into the urine as thromboxane B2 and not 11-dTXB2.[22, 34, 37] Regardless of whether the increase in urinary thromboxane in dogs receiving cyclosporine is solely derived from circulating platelets or other sources in the body, our study clearly demonstrates that cyclosporine administration is associated with increased thromboxane synthesis. Increased concentrations of thromboxane would be expected to increase platelet reactivity in at least some dogs. However, previous studies have suggested that in about 70% of dogs, platelets are insensitive to thromboxane stimulation,[38-40] possibly because of impaired platelet thromboxane A2 receptor-linked G proteins. Because only about 30% of dogs appear to have platelets that are sensitive to the effects of thromboxane, drug-induced increases in thromboxane may not reliably increase platelet reactivity in all dogs. Even if platelets in some dogs do not consistently respond to thromboxane stimulation, however, thromboxane still contributes to vasoconstriction and smooth muscle proliferation, both of which may contribute to the risk of thrombus formation.
Our study also demonstrated that the sustained administration of cyclosporine at both atopy and immunosuppressive dosages to dogs also leads to decreased platelet COX-2 expression. COX-1 is the primary cyclooxygenase isoform responsible for the production of thromboxane, whereas COX-2 is the main isoform that controls prostacyclin (prostaglandin I2) production.[42, 43] Compared with thromboxane, prostacyclin has the opposite physiologic effect on primary hemostasis, and these 2 prostaglandins work in concert to maintain a balanced, normal hemostatic system. COX-2-derived prostacyclin has been associated with a protective response to experimentally induced thrombus formation. Decreased COX-2 expression could lead to an imbalance between thromboxane and prostacyclin, resulting in relatively higher concentrations of thromboxane. Previous studies in rats have demonstrated that cyclosporine suppresses both COX-2 expression in several tissues[45, 46] and prostacyclin biosynthesis. We have demonstrated here that cyclosporine similarly decreases COX-2 expression in canine platelets.
Circulating platelets originally were thought to contain only the COX-1 isoform,[16, 47, 48] but more recent studies have identified the COX-2 isoform in both human and canine platelets.[16, 18, 49] Megakaryocytes contain both COX-1 and COX-2,[17, 49] and changes in COX expression that are induced at the level of these platelet precursors would be expected to result in similar changes in platelets derived from affected megakaryocytes. Although COX-2 is the primary isoform responsible for the generation of prostacyclins by endothelial cells, platelet COX-2 also potentially could provide an alternate source of platelet thromboxane A2 production. If platelet COX-2 is a source of thromboxane synthesis, than the decrease in COX-2 expression in face of increased thromboxane concentrations seen in our study could suggest a negative feedback mechanism. Additional research is needed to better define the precise role of platelet COX-2, and the effects of cyclosporine on this enzyme.
Our study demonstrated that there was a steady decrease in platelet COX-1 expression with sustained cyclosporine administration, and a similar but less consistent effect with lower drug dosages. As thromboxane production increased as measured by the urinary 11-dTXB2-to-creatinine ratio, platelet COX-1 expression tended to decrease. Interestingly, an opposite pattern recently was observed in an aspirin study in our laboratory. When an anti-inflammatory dosage of aspirin was administered to dogs, a decrease in platelet thromboxane production was associated with an increase in platelet COX-1 expression. Platelet function concurrently was inhibited in the dogs receiving aspirin, as demonstrated by prolonged PFA-100 collagen/epinephrine cartridge closure times, and platelet dysfunction was presumed to be caused by COX-1 enzyme inhibition. The results of our aspirin and cyclosporine studies, when considered in combination, suggest that there may be a negative feedback system between platelet thromboxane production and platelet COX-1 expression. Additional research would be required to determine the exact relationship between platelet COX expression and thromboxane production.
Platelet function was evaluated by the PFA-100 with 2 different cartridges containing 2 different platelet agonist combinations. In humans, the collagen/epinephrine cartridge has been shown to be more likely to be influenced by medications, whereas the collagen/ADP cartridge is considered to be a more general indicator of platelet function.[26, 50] In our study, cyclosporine had no significant impact on collagen/epinephrine cartridge closure times. However, although the collagen/epinephrine cartridge has been shown to be an excellent assay in humans for drug-induced decreases in platelet function, it is not known how useful the same cartridge would be at detecting drug-associated increases in platelet reactivity. Interestingly, with the collagen/ADP cartridge, the immunosuppressive dosage of cyclosporine was associated with an increase in closure times at most time points, whereas the atopy dosage was associated with a decrease in closure times at all time points. Closure times, however, did not fall outside of the reference range in any dog. The reason for these apparently contradictory results with the collagen/ADP cartridges is unknown.
In our study, there was an immediate and sustained decrease in platelet phosphatidylserine expression after administration of the immunosuppressive dosage of cyclosporine. Our findings are in contrast to the findings reported in a recent study of humans, which demonstrated that cyclosporine increased platelet phosphatidylserine expression. One possible explanation for the different results is that in our study, we evaluated the effects of cyclosporine administered in vivo, whereas the previous study in humans incubated platelets with cyclosporine in vitro. Although we cannot explain the difference in results, we believe that an in vivo study such as ours is more likely to reflect what actually is occurring in patients receiving cyclosporine. The change in phosphatidylserine expression seen in our study could indicate that cyclosporine decreases this aspect of platelet reactivity. However, the decrease in phosphatidylserine expression observed in our study could have been erroneous, because the initial baseline value before the immunosuppressive dosage of cyclosporine was considerably higher than all other time points, including the baseline for the atopy dosage, a result that potentially could be explained by a nonrepeatable complication with sample handling or preparation. The mean MFI of the 2 additional repeat assays performed after washout of the immunosuppressive dosage, however, although significantly less than the initial baseline results, still were significantly higher than all additional time points, suggesting that the observed effect was real rather than an artifact. We also observed slight but statistically significant changes in platelet phosphatidylserine expression (a decrease at the peak time point and an increase at the trough time point) during the 1st day of the atopy dosage of cyclosporine, findings that are unlikely to be of any clinical relevance.
Upon platelet activation, phosphatidylserine is exposed on platelet surface membranes and binds to prothrombin, converting prothrombin to thrombin, and enhancing secondary hemostasis. Phosphatidylserine also is expressed during apoptosis of nucleated cells.[51-53] Similar apoptotic pathways have been associated with platelet activation, primarily by opening the mitochondrial permeability transition pore (MPTP).[54-56] During apoptosis, the MPTP transiently opens and releases proapoptotic mitochondrial contents into the cell cytoplasm, mediating caspase-3 activation, and ultimately leading to increased phosphatidylserine expression on the cell surface.[53, 54] Cyclosporine has been shown to inhibit platelet MPTP function, thereby decreasing the release of mitochondrial proapoptoic contents and potentially limiting the amount of phosphatidylserine expressed on the platelet surface.[53, 54] Our findings in the immunosuppressive dosage trial, if real, would be consistent with cyclosporine-induced inhibition of MPTP function and a resultant reduction in platelet phosphatidylserine expression.
Our study demonstrated a significant, immediate, and sustained decrease in platelet P-selectin expression during administration of an immunosuppressive dosage of cyclosporine. This effect appears to be dose-dependent, because a similar but less consistent pattern was seen with the atopy dosage of cyclosporine. Several studies in humans, in contrast, have demonstrated an increase in soluble P-selectin in renal transplant recipients treated with cyclosporine.[11, 57] In 1 study comparing renal transplant patients with hypertensive control patients, P-selectin expression increased with all immunosuppressive agents, including cyclosporine, although platelet aggregation was not measurably affected. Another study also demonstrated an increase in P-selectin expression in renal transplant recipients treated with cyclosporine, but similarly found no measurable increase in platelet aggregation. As our study evaluated the effects of cyclosporine on platelet P-selectin expression in normal dogs, and the studies in humans evaluated P-selectin expression in renal transplant recipients, cyclosporine might have a different effect in diseased compared with healthy individuals.
Platelet P-selectin is a cell adhesion molecule that mediates platelet and leukocyte aggregation, induces tissue factor, and enhances fibrin deposition.[58-60] Therefore, like platelet phosphatidylserine, platelet P-selectin is believed to have procoagulant effects that enhance secondary hemostasis. Platelet P-selectin has been shown to be expressed in higher concentrations in dogs with primary IMHA, and could contribute to the high rate of thromboembolic complications seen in IMHA patients. Our previous pilot study, albeit performed in a small number of animals, demonstrated an increase in platelet P-selectin expression in dogs receiving cyclosporine, a result that raised concern because cyclosporine is increasingly used by veterinarians to treat canine IMHA. In this study, however, using a larger number of dogs and more sustained cyclosporine dosing, the effect reported in our pilot study was not reproducible and, in fact, a decrease in platelet P-selectin expression was observed.
One possible explanation for the decrease in platelet P-selectin expression seen in our study could be a drug-associated inhibition of the proinflammatory mediators TNF-α and IL-1β, which may act as stimulators of P-selectin expression.[61-63] Cyclosporine-induced inhibition of TNF-α and IL-1β has been associated with down-regulation of E-selectin, a similar adhesion protein found on vascular endothelium.[64, 65] Similar to phosphatidylserine, the decrease in P-selectin expression observed in our study could suggest that cyclosporine decreases this aspect of platelet reactivity. Additional research would be required to evaluate the effect of cyclosporine on platelet P-selectin expression, soluble plasma P-selectin concentration, and platelet-leukocyte aggregation in dogs with various diseases, particularly conditions such as IMHA that predispose to a hypercoagulable state.
In the absence of a pharmacodynamics assay to evaluate biomarkers of immunosuppression during cyclosporine treatment, it is difficult to determine if the immune system of an individual patient is appropriately suppressed during treatment. Unpredictable responses to cyclosporine treatment have led to the recommendation of monitoring blood drug concentrations to ensure immunosuppression. Unfortunately, there are considerable variations in the reference ranges for target blood cyclosporine concentrations, and there also is a highly variable relationship between drug blood concentrations and clinical response to the medication. After 4 days of administrating the immunosuppressive dosage of cyclosporine, only 2 dogs in our study had reached target immunosuppressive trough blood concentrations of 600 ng/mL, a concentration that has been shown to be reliably immunosuppressive in dogs based on pharmacodynamic assays. To be certain that all of the dogs in the study were reliably immunosuppressed, we chose to increase the cyclosporine dose to ensure that all dogs maintained a blood concentration that was greater than the recommended target immunosuppressive blood concentration.[19, 66] This approach is similar to that used by clinicians when treating dogs with life-threatening immune-mediated diseases such as IMHA.
Recent studies in humans have suggested that the use of cyclosporine could lead to an increase in thromboembolic complications. Our study indicates that cyclosporine has multiple effects on canine platelets, causing altered expression of proteins on the platelet surface membrane, altered platelet COX expression, and increased thromboxane production. It is particularly concerning that the administration of cyclosporine was associated with marked increases in thromboxane production, suggesting that exposure to the drug predisposes to increased platelet activation. On the other hand, our study also indicated that cyclosporine decreases platelet expression of the procoagulant proteins P-selectin and phosphatidylserine. Our study was performed in normal dogs, and our results may not be directly applicable to diseased dogs. However, these results strongly suggest that there is a pressing need to evaluate the effects of cyclosporine on platelet thromboxane production and expression of COX enzyme isoforms, P-selectin, and phosphatidylserine in dogs that have diseases that are associated with a hypercoaguable state and a predisposition to thromboembolic complications.
Funded by the Mississippi State University College of Veterinary Medicine Internal Competitive Research Grant and Dr. Hugh G. Ward Endowment. The authors thank Jesse Grady, Mandy Wallace, and Marla Waldrop for their assistance.
Conflict of Interest Declaration: Authors disclose no conflict of interest.
Atopica, Novartis Animal Health, Greensboro, NC
3.8% sodium citrate, Vacutainer tube, Becton Dickinson, Franklin Lakes, NJ
Clone MD6, IgG1, generously provided by Dr C. Wayne Smith, Baylor College of Medicine, Houston, TX
Goat, antimouse-PE, Lot Number LXP07, R&D Systems, Minneapolis, MN
FITC-conjugated monoclonal CD61, Clone JM2E5, AbDSerotec, Raleigh, NC
Paraformaldehyde, Biolegend Inc., San Diego, CA
Annexin-V Binding Buffer, Invitrogen, Camarillo, CA
Annexin-V FITC Conjugate (Recombinant), Invitrogen
Monoclonal CD61-purified, Clone JM2E5, Accurate Chemical, Westbury, NY
Goat antimouse IgG:RPE, AbDSerotec, Raleigh, NC
FITC-conjugated monoclonal COX-1, Clone CX111, Cayman Chemical Co, Ann Arbor, MI
Monoclonal anti-human CD9:RPE, Clone MM2/57, AbDSerotec
Triton X-100, Sigma-Aldrich, St. Louis, MO
FITC-conjugated monoclonal COX-2, Clone CX299, Cayman Chemical Co
FACS Calibur, BD Biosciences, San Jose, CA
CellQuest software, BD Biosciences
Luminex 11-dehydro Thromboxane B2 Kit, Cayman Chemical Co
Luminex 200 System xMAP Technology, Luminex Corporation, Austin, TX
ACE Alera Clinical Chemistry System, Alfa Wasserman, Inc, West Caldwell, NJ
PFA-100, Siemens Healthcare Diagnostics, Deerfield, IL
PFA Collagen/ADP Test Cartridge, Siemens Healthcare Diagnostics, Duluth, GA
PFA Collagen/EPI Test Cartridge, Siemens Healthcare Diagnostics
Abbott Cell-Dyn 3700, Abbott Laboratories, Abbott Park, IL
Personal communication, John D Patz, 2008
7.5% EDTA Blood Collection Tubes, Tyco Healthcare, Mansfield MA
Generous gift from Novartis Pharmaceuticals, East Hanover, NJ
Varian Bond Elut 100 mg; Agilent Technologies, Santa Clara, CA
SAS for Windows version 9.2, SAS Institute, Cary, NC
- 15Effects of cyclosporine on canine platelet procoagulant activity. J Vet Intern Med 2009;23:692., , , et al.