ORIGINAL ARTICLE
Cortisol, Platelet Serotonin Content, and Platelet Activity in Patients With Major Depression and Type 2 Diabetes: An Exploratory Investigation Daniela Zahn, PhD, Frank Petrak, PhD, Leonora Franke, PhD, Anna-Karolina Hägele, MD, Georg Juckel, MD, PhD, Florian Lederbogen, MD, Horst Neubauer, MD, Christine Norra, MD, PhD, Idun Uhl, MD, and Stephan Herpertz, MD ABSTRACT Objective: Hypothalamic-pituitary-adrenal system dysfunction, serotonergic system alterations, and enhanced platelet activity may contribute to the increased cardiac risk in depression. This exploratory study examined associations between cortisol parameters, platelet serotonin (5-HT) content, and platelet activity markers in patients with newly diagnosed major depression (MD) and/or Type 2 diabetes (T2DM) compared with healthy controls. Methods: We compared cortisol awakening response (CAR), diurnal decrease in salivary cortisol concentrations (slope), platelet 5-HT, and platelet markers (CD40, CD40 ligand [CD40L], soluble CD40L, CD62P, β-thromboglobulin, and platelet factor-4) in 22 T2DM patients, 20 MD patients, 18 T2DM patients with MD, and 24 healthy controls. Results: Platelet markers were elevated in MD (F(6,60) = 11.14, p < .001) and T2DM (F(6,60) = 13.07, p < .001). Subgroups did not differ in 5-HT or cortisol slope, whereas T2DM patients without depression had significantly lower CAR than did healthy controls (F(1,61) = 7.46, p = .008). In healthy controls, cortisol slope correlated with platelet activity for CD40 (r = −0.43, p = .048) and 5-HT was correlated with CD40L (r = 0.53, p = .007). In patients with both T2DM and MD, 5-HT and CD62P were correlated (r = 0.52, p = .033). Conclusions: Increased platelet activity in T2DM and MD may play a role in the association between diabetes, depression, and coronary artery disease. The present data suggest that group differences in cortisol or 5-HT as well as group-specific associations of cortisol or 5-HT with platelet markers might be of limited importance in the shared pathways of T2DM and depression in the pathophysiology of coronary artery disease. Key words: major depression, Type 2 diabetes, platelet serotonin content, cortisol, platelet activity, CD40.
INTRODUCTION
Although the mechanisms linking depression, diabetes, and coronary artery disease (CAD) are still unclear, hyperactivity of the hypothalamic-pituitary-adrenal (HPA) axis might play an important role (for an overview, see Refs. (10,11). A recent meta-analysis showed that depression is related to elevated cortisol levels with an overall effect size
A
pproximately 350 million people worldwide have depression (1). Depression prevalence is approximately two times higher in patients with diabetes compared with those without (2,3). Evidence suggests a bidirectional relationship between Type 2 diabetes (T2DM) and depression: whereas patients with T2DM are at increased risk from depression, individuals with depression are also more likely to develop T2DM (4,5). Furthermore, depression is associated with the development of cardiovascular diseases in healthy populations (6,7) and an increased risk for diabetes complications and all-cause mortality in diabetes patients (8,9).
5-HT = 5-hydroxy-tryptamine (serotonin), BMI = body mass index, CAD = coronary artery disease, CAR = cortisol awakening response, CD40L = CD40 ligand, HbA1c = glycated hemoglobin, HPA axis = hypothalamic-pituitary-adrenal axis, MD = major depression, PF-4 = platelet factor-4, sCD40L = soluble CD40 ligand, Slope = diurnal decrease in cortisol levels, T2DM = Type 2 diabetes mellitus, β-TG = β-thromboglobulin
From the Department of Health Psychology (Zahn), Institute of Psychology, Johannes Gutenberg University Mainz, Germany; Departments of Psychosomatic Medicine and Psychotherapy (Petrak, Hägele, Herpertz) and Psychiatry, Psychotherapy, and Preventive Medicine (Juckel, Norra, Uhl), LWL University Hospital Bochum, Ruhr University Bochum, Germany; Department of Psychiatry and Psychotherapy (Franke), Charité-Universitätsmedizin Berlin, Germany; Central Institute of Mental Health (Lederbogen), Medical Faculty, University of Heidelberg, Mannheim, Germany; and Cardiovascular Centre (Neubauer), St. Josef-Hospital, Ruhr University Bochum, Bochum, Germany. Address correspondence and reprint requests to Daniela Zahn, PhD, Department of Health Psychology, Johannes Gutenberg University Mainz, Binger Str. 14-16, 55122 Mainz, Germany. E-mail: [email protected] Received for publication October 17, 2013; revision received August 28, 2014. DOI: 10.1097/PSY.0000000000000145 Copyright © 2015 by the American Psychosomatic Society
Psychosomatic Medicine, V 77 • 145-155
145
February/March 2015
Copyright © 2015 by the American Psychosomatic Society. Unauthorized reproduction of this article is prohibited.
ORIGINAL ARTICLE
compared with healthy controls (43,44). In depressed patients without T2DM, most studies observed decreased or unchanged rather than increased 5-HT-induced platelet aggregation (25). In sum, current research indicates that cortisol parameters and platelet 5-HT content may be altered in MD and T2DM compared with healthy controls, and suggests that platelet activity might be related to cortisol and platelet 5HT. These associations remain unclear for comorbid patients (with MD and T2DM). Thus, the current pilot study aimed at comparing cortisol parameters, platelet 5-HT content, and platelet activity markers in patients with MD and/ or T2DM compared with healthy controls, and at determining the association of cortisol and platelet 5-HT with platelet markers in these patients. To avoid confoundation with antidepressant treatment, only newly diagnosed untreated MD patients were examined. Based on current research, we expected a reduced CAR in T2DM patients compared with those without diabetes. Because very early studies reported increased plasma levels of 5-HT or reduced 5-HT uptake rates in platelets of T2DM patients, we assumed lower platelet 5-HT content in T2DM patients versus healthy controls and patients with MD. Furthermore, we expected negative correlations between platelet 5-HT content and platelet markers, specifically in T2DM patients (with and without MD).
of d = 0.60 (12). Elevated cortisol levels also contribute to insulin resistance and abdominal adiposity—two known risk factors for T2DM (13,14). For patients diagnosed as having T2DM, there is also evidence for HPA dysfunction: T2DM patients had higher plasma cortisol levels (15–17) or a lower cortisol awakening response (CAR) (18–20) than did healthy controls. For diabetes patients with major depression (MD), a first study found elevated 24-hour urinefree cortisol levels compared with nondepressed diabetes patients (21), suggesting an additive effect of depression on cortisol levels in diabetes. Furthermore, evidence links elevated cortisol levels to platelet activity (22), atherosclerosis of the carotid arteries (23), and progress in coronary artery calcification (24), even after controlling for cardiovascular risk factors. As increased platelet activity is a frequently discussed explanation for the link between depression and CAD (10,25), HPA dysfunction might contribute directly or indirectly via inflammatory processes (26) to increased platelet activity. Alterations in the serotonergic system might also explain the link between MD and CAD (10,27). In addition to a central serotonin deficit (28), alterations in peripheral serotonin levels have been suggested: most of peripheral serotonin (5-hydroxy-tryptamine; 5-HT) is stored in platelets and released during platelet activation (29). 5-HT itself promotes platelet aggregation directly and indirectly via potentiating the effects of collagen and ADP on platelet aggregation (30). Thus, changes in the serotonergic signaling systems in platelets have been assumed in depression (31), which might promote platelet aggregation. Previous research on peripheral 5-HT levels in depressed participants versus healthy controls revealed controversial results: earlier studies assessing peripheral serotonin as an indicator of central serotonergic activity found similar levels of platelet 5-HT (32,33) or increased platelet 5-HT (34) in depressed versus nondepressed participants. Other studies also revealed decreased platelet 5-HT (35–38) in depressed versus nondepressed individuals, but these results might be attributable to either current use or too short wash-out periods of serotonin reuptake inhibitors, which decrease platelet 5-HT uptake (37). Studies examining depressed CAD patients showed elevated whole blood 5-HT levels (39,40) and elevated platelet 5-HT (39) compared with nondepressed CAD patients, whereas another study found no differences (41). In sum, these findings indicate either unchanged or increased 5-HT levels in untreated depressed patients with or without CAD. In T2DM patients, to our best knowledge, the association of platelet 5-HT and depression has not been studied. However, comparisons of T2DM patients and healthy controls indicate decreased platelet 5-HT (42) and whole blood 5-HT (43–45), and increased plasma 5-HT levels (42–44) in diabetes. For platelet activity, enhanced 5-HT-induced platelet activity has been reported in diabetes patients Psychosomatic Medicine, V 77 • 145-155
METHODS Participants and Procedure The original study sample consisted of 19 patients with T2DM and untreated comorbid MD, 24 T2DM patients, 21 patients with untreated MD, and 25 age- and sex-matched healthy controls recruited between March 2008 and July 2009. The study procedure has been described elsewhere (46). Inclusion criteria for diabetes patients were a diagnosis of T2DM according to the World Health Organization criteria for at least 1 year, insulin treatment for more than 6 months, two glycated hemoglobin (HbA1c) values greater than 7% in the preceding 9 months, and current HbA1c greater than 7%. Diagnosis-related information was retrieved from patients' records, and actual HbA1c was confirmed by an HbA1c screening before study inclusion. The Patient Health Questionnaire (47) was used to screen for depression. A diagnosis of MD (International Classification of Diseases, Tenth Revision, F32, F33) was verified using the Structured Clinical Interview for the Diagnosis of Psychiatric Disorders according to DSM-IV-TR criteria (48). Severity of depression was assessed with the clinician rating form of the Quick Inventory of Depressive Symptomatology (49). Exclusion criteria were any history of CAD, cerebrovascular disease, or peripheral artery disease; current substance abuse; psychotic illness; suicidal tendencies; use of platelet aggregation inhibitors, anticoagulants, or antidepressant medication; cognitive impairment; and severe medical conditions other than hypertension and dyslipidemia. All participants gave informed written consent and were compensated for participation. The ethics committee of the Ruhr University Bochum approved the study protocol.
146
February/March 2015
Copyright © 2015 by the American Psychosomatic Society. Unauthorized reproduction of this article is prohibited.
Cortisol, Serotonin, Depression, and Diabetes
Variables and Instruments Demographic and Medical Characteristics
Coulter, Krefeld, Germany). A total of 15,000 platelets were analyzed at a flow rate less than 250 platelets/s. The “Cellquest pro” software (Becton Dickinson) was used for analysis. Platelet markers are expressed as percentage of positive cells. The inter-assay and intra-assay coefficients of variation were less than 5% and less than 7%, respectively. Soluble CD40L (sCD40L), platelet factor-4 (PF-4), and β-thromboglobulin (β-TG) in serum were analyzed with commercial enzyme-linked immunoabsorbent assay kits (sCD40L [Alexis Biochemicals, Grünberg, Germany] and PF-4 and β-TG [Diagnostica Stago/Roche Diagnostics]), as described before (53). The detection limit was 0.01 ng/ml. Intra-assay variation among the triplicates for all samples was less than 6%.
Demographic and basic medical data were collected with selfreport questionnaires and structured checklists. Height and weight were measured to calculate body mass index (BMI). Peripheral venous blood was drawn from the forearm in the morning of the examination day after an overnight fast and a rest of at least 15 minutes after minimal stasis. Citrated blood samples were either centrifuged (190g for 10 minutes at room temperature) to obtain platelet-rich plasma or immediately fixed with 1% formaldehyde (1:1, vol/vol). A separate aliquot of blood without any additives was put into the refrigerator and allowed to clot for 1 hour before centrifugation (1500g for 10 minutes). The serum supernatant was stored at −80 °C until analysis. HbA1c, high-density lipoprotein cholesterol, lowdensity lipoprotein cholesterol, and triglycerides were measured using standard laboratory techniques. High-sensitivity C-reactive protein was determined with a latex-enhanced immunoturbometric method as described before (46).
Statistical Analysis The current analyses focused on n = 84 participants with complete assessment of serotonin (n = 3 excluded due to missing samples, n = 2 due to outliers). Excluded participants did not differ in demographic, basic medical data, outcome measures, or patient group. Sample size for cortisol parameters and platelet markers varies from n = 65 (CAR) to n = 82 (PF-4) because of technical difficulties in laboratory analysis, insufficient saliva volume, or cortisol samples not returned by participants. The amount of missing cortisol samples was comparable in all subgroups. Variables are displayed with mean and standard deviation or median and interquartile range (25th percentile; 75th percentile) for nonnormally distributed continuous variables, and absolute frequencies and percentage for categorical variables. Because of deviations from normality distribution, BMI, triglycerides, highsensitivity C-reactive protein, CD62P, and CD40L were logarithmically transformed; 5-HT, high-density lipoprotein cholesterol, CD40, and sCD40L were square-root-transformed for all analyses. Differences in demographic and medical data were evaluated with analysis of variance (ANOVA) with factor subgroup and Tukey B post hoc tests to adjust for multiple pairwise comparisons between groups with different group sizes, and χ2 tests. For cortisol parameters (CAR, slope) and platelet markers (CD40, sCD40L, CD40L, CD62P, β-TG, PF-4), multivariate analyses of variance (MANOVAs) with factors depression, diabetes, and the interaction term depression diabetes were performed. In case of significant overall effects, ANOVAs were conducted with all available observations per variable to maximize statistical power. Analyses were repeated controlling for age and BMI. For 5-HT, an ANOVA was computed with depression, diabetes, and the interaction term depression diabetes. Simple main effects analyses were used to compare differences in depressed and nondepressed patients in the T2DM and non-T2DM subgroups. Pearson correlation coefficients were computed over all participants and within the four subgroups to assess the relationships between 5-HT or cortisol parameters and platelet markers. A sensitivity power analysis showed that our sample had sufficient power (β = 0.80) at a significance level of α = .05 to detect medium to large effect sizes ( f = 0.35) in a two (depression) two (diabetes) ANOVA, medium effect sizes (ρ = 0.33) in bivariate correlations over all participants, and large effect sizes (ρ = 0.51–0.73) in correlations within the four subgroups. Power analysis was done with G-Power (54), and statistical analyses were performed with SPSS 19. p Values less than .05 were considered statistically significant.
Cortisol Parameters Salivary cortisol collection followed an ambulatory protocol as described elsewhere (50). After baseline examination, participants received cotton-based sampling devices and plastic storage syringes. Participants were instructed to perform cortisol sampling on a regular weekday, immediately after awakening (F0), after the following 30 minutes (F0.5), and 4 hours (F4), 8 hours (F8), 12 hours (F12), and 16 hours (F16) later. Participants were asked not to smoke, eat, drink, or engage in physical activity 30 minutes before sampling. After receipt by mail, samples were frozen (−80 °C) until analysis by time-resolved immunoassay with fluorescence detection. Intra-assay and inter-assay variations were less than 8%. CAR was computed by subtracting cortisol levels at F0 from levels at F0.5. Diurnal decrease in cortisol secretion (slope) was calculated by subtracting F16 from levels at F0 divided by the hours between the measurements ((F0 − F16)/hours).
Serotonin (5-HT) 5-HT concentration in platelets was measured in platelet-rich plasma by high-performance liquid chromatography with UV detection at 220 nm, using a reverse-phase cartridge LiChrosper-100 RP18 column (particle size 5 μm, 125 4 mm; Merck, Darmstadt, Germany) as described elsewhere (51). Intra-assay and inter-assay coefficients of variation for platelet 5-HT were 2.5% and 4.2%, respectively. Each sample was analyzed in duplicate. Platelet 5-HT concentration was expressed as ng per 109 platelets.
Markers of Platelet Activity Immediate analysis of platelet activation via flow cytometer (FACS Calibur cytometer; Becton Dickinson, San Jose, CA) was performed after incubation of citrated blood with monoclonal antibodies against CD40, CD40 ligand [CD40L], and CD62P on platelet surface without stimulation, as previously described (52). Phycoerythin-conjugated anti-CD40 and anti-CD40L with phycoerythin-conjugated IgG1 and fluorescein isothiocyanate– conjugated anti-CD62P served as isotype controls (Beckman
Psychosomatic Medicine, V 77 • 145-155
147
February/March 2015
Copyright © 2015 by the American Psychosomatic Society. Unauthorized reproduction of this article is prohibited.
ORIGINAL ARTICLE
RESULTS
Markers of Platelet Activity
Table 1 displays demographic and medical data for each group. Nondepressed T2DM patients were older than depressed patients without T2DM. T2DM patients had higher HbA1c, had higher BMI, and received antihypertensive treatment more frequently than did patients without T2DM. Severity of depression was similar in depressed patients with or without T2DM.
MANOVA revealed an overall main effect of depression (Pillai's trace = 0.527, F(6,60) = 11.14, p < .001, η2 = 0.527), diabetes (Pillai's trace = 0.566, F(6,60) = 13.07, p < .001, η2 = 0.566), and an interaction (Pillai's trace = 0.187, F(6,60) = 2.30, p = .046, η2 = 0.187) on platelet markers (CD40, sCD40, CD40L, CD62P, β-TG, PF-4). After adjustment for BMI and age, main effects of diabetes and depression remained, but the interaction was reduced to a trend (Pillai's trace = 0.174, F(6,58) = 2.03, p = .076, η2 = 0.174). Neither multivariate nor univariate effects of BMI and age were observed (η2 ranging from 0.001 to 0.041). ANOVAs showed a significant interaction of MD T2DM only for β-TG (F(1,76) = 10.72, p = .002, η2 = 0.124). As displayed in Table 2, analysis of simple main effects revealed significantly higher β-TG only in depressed patients compared with healthy controls. A trend for an interaction was observed for sCD40L (F(1,77) = 2.87, p = .094, η2 = 0.032), with simple main effects analysis showing significantly higher sCD40L levels only in comorbid patients compared with T2DM patients. For all other platelet markers except PF-4, ANOVAs yielded significant main effects for MD (CD40: F(1,70) = 33.22, p < .001, η2 = 0.322; CD40L: F(1,75) = 7.43, p = .008, η2 = 0.090; CD62P: F(1,74) = 13.39, p < .001, η2 = 0.153) and main effects for T2DM (CD40: F(1,70) = 15.52, p < .001, η2 = 0.182; CD40L: F(1,75) = 9.80, p = .002, η2 = 0.116; CD62P: F(1,74) = 16.35, p
Cortisol, Platelet Serotonin Content, and Platelet Activity in Patients With Major Depression and Type 2 Diabetes: An Exploratory Investigation Daniela Zahn, PhD, Frank Petrak, PhD, Leonora Franke, PhD, Anna-Karolina Hägele, MD, Georg Juckel, MD, PhD, Florian Lederbogen, MD, Horst Neubauer, MD, Christine Norra, MD, PhD, Idun Uhl, MD, and Stephan Herpertz, MD ABSTRACT Objective: Hypothalamic-pituitary-adrenal system dysfunction, serotonergic system alterations, and enhanced platelet activity may contribute to the increased cardiac risk in depression. This exploratory study examined associations between cortisol parameters, platelet serotonin (5-HT) content, and platelet activity markers in patients with newly diagnosed major depression (MD) and/or Type 2 diabetes (T2DM) compared with healthy controls. Methods: We compared cortisol awakening response (CAR), diurnal decrease in salivary cortisol concentrations (slope), platelet 5-HT, and platelet markers (CD40, CD40 ligand [CD40L], soluble CD40L, CD62P, β-thromboglobulin, and platelet factor-4) in 22 T2DM patients, 20 MD patients, 18 T2DM patients with MD, and 24 healthy controls. Results: Platelet markers were elevated in MD (F(6,60) = 11.14, p < .001) and T2DM (F(6,60) = 13.07, p < .001). Subgroups did not differ in 5-HT or cortisol slope, whereas T2DM patients without depression had significantly lower CAR than did healthy controls (F(1,61) = 7.46, p = .008). In healthy controls, cortisol slope correlated with platelet activity for CD40 (r = −0.43, p = .048) and 5-HT was correlated with CD40L (r = 0.53, p = .007). In patients with both T2DM and MD, 5-HT and CD62P were correlated (r = 0.52, p = .033). Conclusions: Increased platelet activity in T2DM and MD may play a role in the association between diabetes, depression, and coronary artery disease. The present data suggest that group differences in cortisol or 5-HT as well as group-specific associations of cortisol or 5-HT with platelet markers might be of limited importance in the shared pathways of T2DM and depression in the pathophysiology of coronary artery disease. Key words: major depression, Type 2 diabetes, platelet serotonin content, cortisol, platelet activity, CD40.
INTRODUCTION
Although the mechanisms linking depression, diabetes, and coronary artery disease (CAD) are still unclear, hyperactivity of the hypothalamic-pituitary-adrenal (HPA) axis might play an important role (for an overview, see Refs. (10,11). A recent meta-analysis showed that depression is related to elevated cortisol levels with an overall effect size
A
pproximately 350 million people worldwide have depression (1). Depression prevalence is approximately two times higher in patients with diabetes compared with those without (2,3). Evidence suggests a bidirectional relationship between Type 2 diabetes (T2DM) and depression: whereas patients with T2DM are at increased risk from depression, individuals with depression are also more likely to develop T2DM (4,5). Furthermore, depression is associated with the development of cardiovascular diseases in healthy populations (6,7) and an increased risk for diabetes complications and all-cause mortality in diabetes patients (8,9).
5-HT = 5-hydroxy-tryptamine (serotonin), BMI = body mass index, CAD = coronary artery disease, CAR = cortisol awakening response, CD40L = CD40 ligand, HbA1c = glycated hemoglobin, HPA axis = hypothalamic-pituitary-adrenal axis, MD = major depression, PF-4 = platelet factor-4, sCD40L = soluble CD40 ligand, Slope = diurnal decrease in cortisol levels, T2DM = Type 2 diabetes mellitus, β-TG = β-thromboglobulin
From the Department of Health Psychology (Zahn), Institute of Psychology, Johannes Gutenberg University Mainz, Germany; Departments of Psychosomatic Medicine and Psychotherapy (Petrak, Hägele, Herpertz) and Psychiatry, Psychotherapy, and Preventive Medicine (Juckel, Norra, Uhl), LWL University Hospital Bochum, Ruhr University Bochum, Germany; Department of Psychiatry and Psychotherapy (Franke), Charité-Universitätsmedizin Berlin, Germany; Central Institute of Mental Health (Lederbogen), Medical Faculty, University of Heidelberg, Mannheim, Germany; and Cardiovascular Centre (Neubauer), St. Josef-Hospital, Ruhr University Bochum, Bochum, Germany. Address correspondence and reprint requests to Daniela Zahn, PhD, Department of Health Psychology, Johannes Gutenberg University Mainz, Binger Str. 14-16, 55122 Mainz, Germany. E-mail: [email protected] Received for publication October 17, 2013; revision received August 28, 2014. DOI: 10.1097/PSY.0000000000000145 Copyright © 2015 by the American Psychosomatic Society
Psychosomatic Medicine, V 77 • 145-155
145
February/March 2015
Copyright © 2015 by the American Psychosomatic Society. Unauthorized reproduction of this article is prohibited.
ORIGINAL ARTICLE
compared with healthy controls (43,44). In depressed patients without T2DM, most studies observed decreased or unchanged rather than increased 5-HT-induced platelet aggregation (25). In sum, current research indicates that cortisol parameters and platelet 5-HT content may be altered in MD and T2DM compared with healthy controls, and suggests that platelet activity might be related to cortisol and platelet 5HT. These associations remain unclear for comorbid patients (with MD and T2DM). Thus, the current pilot study aimed at comparing cortisol parameters, platelet 5-HT content, and platelet activity markers in patients with MD and/ or T2DM compared with healthy controls, and at determining the association of cortisol and platelet 5-HT with platelet markers in these patients. To avoid confoundation with antidepressant treatment, only newly diagnosed untreated MD patients were examined. Based on current research, we expected a reduced CAR in T2DM patients compared with those without diabetes. Because very early studies reported increased plasma levels of 5-HT or reduced 5-HT uptake rates in platelets of T2DM patients, we assumed lower platelet 5-HT content in T2DM patients versus healthy controls and patients with MD. Furthermore, we expected negative correlations between platelet 5-HT content and platelet markers, specifically in T2DM patients (with and without MD).
of d = 0.60 (12). Elevated cortisol levels also contribute to insulin resistance and abdominal adiposity—two known risk factors for T2DM (13,14). For patients diagnosed as having T2DM, there is also evidence for HPA dysfunction: T2DM patients had higher plasma cortisol levels (15–17) or a lower cortisol awakening response (CAR) (18–20) than did healthy controls. For diabetes patients with major depression (MD), a first study found elevated 24-hour urinefree cortisol levels compared with nondepressed diabetes patients (21), suggesting an additive effect of depression on cortisol levels in diabetes. Furthermore, evidence links elevated cortisol levels to platelet activity (22), atherosclerosis of the carotid arteries (23), and progress in coronary artery calcification (24), even after controlling for cardiovascular risk factors. As increased platelet activity is a frequently discussed explanation for the link between depression and CAD (10,25), HPA dysfunction might contribute directly or indirectly via inflammatory processes (26) to increased platelet activity. Alterations in the serotonergic system might also explain the link between MD and CAD (10,27). In addition to a central serotonin deficit (28), alterations in peripheral serotonin levels have been suggested: most of peripheral serotonin (5-hydroxy-tryptamine; 5-HT) is stored in platelets and released during platelet activation (29). 5-HT itself promotes platelet aggregation directly and indirectly via potentiating the effects of collagen and ADP on platelet aggregation (30). Thus, changes in the serotonergic signaling systems in platelets have been assumed in depression (31), which might promote platelet aggregation. Previous research on peripheral 5-HT levels in depressed participants versus healthy controls revealed controversial results: earlier studies assessing peripheral serotonin as an indicator of central serotonergic activity found similar levels of platelet 5-HT (32,33) or increased platelet 5-HT (34) in depressed versus nondepressed participants. Other studies also revealed decreased platelet 5-HT (35–38) in depressed versus nondepressed individuals, but these results might be attributable to either current use or too short wash-out periods of serotonin reuptake inhibitors, which decrease platelet 5-HT uptake (37). Studies examining depressed CAD patients showed elevated whole blood 5-HT levels (39,40) and elevated platelet 5-HT (39) compared with nondepressed CAD patients, whereas another study found no differences (41). In sum, these findings indicate either unchanged or increased 5-HT levels in untreated depressed patients with or without CAD. In T2DM patients, to our best knowledge, the association of platelet 5-HT and depression has not been studied. However, comparisons of T2DM patients and healthy controls indicate decreased platelet 5-HT (42) and whole blood 5-HT (43–45), and increased plasma 5-HT levels (42–44) in diabetes. For platelet activity, enhanced 5-HT-induced platelet activity has been reported in diabetes patients Psychosomatic Medicine, V 77 • 145-155
METHODS Participants and Procedure The original study sample consisted of 19 patients with T2DM and untreated comorbid MD, 24 T2DM patients, 21 patients with untreated MD, and 25 age- and sex-matched healthy controls recruited between March 2008 and July 2009. The study procedure has been described elsewhere (46). Inclusion criteria for diabetes patients were a diagnosis of T2DM according to the World Health Organization criteria for at least 1 year, insulin treatment for more than 6 months, two glycated hemoglobin (HbA1c) values greater than 7% in the preceding 9 months, and current HbA1c greater than 7%. Diagnosis-related information was retrieved from patients' records, and actual HbA1c was confirmed by an HbA1c screening before study inclusion. The Patient Health Questionnaire (47) was used to screen for depression. A diagnosis of MD (International Classification of Diseases, Tenth Revision, F32, F33) was verified using the Structured Clinical Interview for the Diagnosis of Psychiatric Disorders according to DSM-IV-TR criteria (48). Severity of depression was assessed with the clinician rating form of the Quick Inventory of Depressive Symptomatology (49). Exclusion criteria were any history of CAD, cerebrovascular disease, or peripheral artery disease; current substance abuse; psychotic illness; suicidal tendencies; use of platelet aggregation inhibitors, anticoagulants, or antidepressant medication; cognitive impairment; and severe medical conditions other than hypertension and dyslipidemia. All participants gave informed written consent and were compensated for participation. The ethics committee of the Ruhr University Bochum approved the study protocol.
146
February/March 2015
Copyright © 2015 by the American Psychosomatic Society. Unauthorized reproduction of this article is prohibited.
Cortisol, Serotonin, Depression, and Diabetes
Variables and Instruments Demographic and Medical Characteristics
Coulter, Krefeld, Germany). A total of 15,000 platelets were analyzed at a flow rate less than 250 platelets/s. The “Cellquest pro” software (Becton Dickinson) was used for analysis. Platelet markers are expressed as percentage of positive cells. The inter-assay and intra-assay coefficients of variation were less than 5% and less than 7%, respectively. Soluble CD40L (sCD40L), platelet factor-4 (PF-4), and β-thromboglobulin (β-TG) in serum were analyzed with commercial enzyme-linked immunoabsorbent assay kits (sCD40L [Alexis Biochemicals, Grünberg, Germany] and PF-4 and β-TG [Diagnostica Stago/Roche Diagnostics]), as described before (53). The detection limit was 0.01 ng/ml. Intra-assay variation among the triplicates for all samples was less than 6%.
Demographic and basic medical data were collected with selfreport questionnaires and structured checklists. Height and weight were measured to calculate body mass index (BMI). Peripheral venous blood was drawn from the forearm in the morning of the examination day after an overnight fast and a rest of at least 15 minutes after minimal stasis. Citrated blood samples were either centrifuged (190g for 10 minutes at room temperature) to obtain platelet-rich plasma or immediately fixed with 1% formaldehyde (1:1, vol/vol). A separate aliquot of blood without any additives was put into the refrigerator and allowed to clot for 1 hour before centrifugation (1500g for 10 minutes). The serum supernatant was stored at −80 °C until analysis. HbA1c, high-density lipoprotein cholesterol, lowdensity lipoprotein cholesterol, and triglycerides were measured using standard laboratory techniques. High-sensitivity C-reactive protein was determined with a latex-enhanced immunoturbometric method as described before (46).
Statistical Analysis The current analyses focused on n = 84 participants with complete assessment of serotonin (n = 3 excluded due to missing samples, n = 2 due to outliers). Excluded participants did not differ in demographic, basic medical data, outcome measures, or patient group. Sample size for cortisol parameters and platelet markers varies from n = 65 (CAR) to n = 82 (PF-4) because of technical difficulties in laboratory analysis, insufficient saliva volume, or cortisol samples not returned by participants. The amount of missing cortisol samples was comparable in all subgroups. Variables are displayed with mean and standard deviation or median and interquartile range (25th percentile; 75th percentile) for nonnormally distributed continuous variables, and absolute frequencies and percentage for categorical variables. Because of deviations from normality distribution, BMI, triglycerides, highsensitivity C-reactive protein, CD62P, and CD40L were logarithmically transformed; 5-HT, high-density lipoprotein cholesterol, CD40, and sCD40L were square-root-transformed for all analyses. Differences in demographic and medical data were evaluated with analysis of variance (ANOVA) with factor subgroup and Tukey B post hoc tests to adjust for multiple pairwise comparisons between groups with different group sizes, and χ2 tests. For cortisol parameters (CAR, slope) and platelet markers (CD40, sCD40L, CD40L, CD62P, β-TG, PF-4), multivariate analyses of variance (MANOVAs) with factors depression, diabetes, and the interaction term depression diabetes were performed. In case of significant overall effects, ANOVAs were conducted with all available observations per variable to maximize statistical power. Analyses were repeated controlling for age and BMI. For 5-HT, an ANOVA was computed with depression, diabetes, and the interaction term depression diabetes. Simple main effects analyses were used to compare differences in depressed and nondepressed patients in the T2DM and non-T2DM subgroups. Pearson correlation coefficients were computed over all participants and within the four subgroups to assess the relationships between 5-HT or cortisol parameters and platelet markers. A sensitivity power analysis showed that our sample had sufficient power (β = 0.80) at a significance level of α = .05 to detect medium to large effect sizes ( f = 0.35) in a two (depression) two (diabetes) ANOVA, medium effect sizes (ρ = 0.33) in bivariate correlations over all participants, and large effect sizes (ρ = 0.51–0.73) in correlations within the four subgroups. Power analysis was done with G-Power (54), and statistical analyses were performed with SPSS 19. p Values less than .05 were considered statistically significant.
Cortisol Parameters Salivary cortisol collection followed an ambulatory protocol as described elsewhere (50). After baseline examination, participants received cotton-based sampling devices and plastic storage syringes. Participants were instructed to perform cortisol sampling on a regular weekday, immediately after awakening (F0), after the following 30 minutes (F0.5), and 4 hours (F4), 8 hours (F8), 12 hours (F12), and 16 hours (F16) later. Participants were asked not to smoke, eat, drink, or engage in physical activity 30 minutes before sampling. After receipt by mail, samples were frozen (−80 °C) until analysis by time-resolved immunoassay with fluorescence detection. Intra-assay and inter-assay variations were less than 8%. CAR was computed by subtracting cortisol levels at F0 from levels at F0.5. Diurnal decrease in cortisol secretion (slope) was calculated by subtracting F16 from levels at F0 divided by the hours between the measurements ((F0 − F16)/hours).
Serotonin (5-HT) 5-HT concentration in platelets was measured in platelet-rich plasma by high-performance liquid chromatography with UV detection at 220 nm, using a reverse-phase cartridge LiChrosper-100 RP18 column (particle size 5 μm, 125 4 mm; Merck, Darmstadt, Germany) as described elsewhere (51). Intra-assay and inter-assay coefficients of variation for platelet 5-HT were 2.5% and 4.2%, respectively. Each sample was analyzed in duplicate. Platelet 5-HT concentration was expressed as ng per 109 platelets.
Markers of Platelet Activity Immediate analysis of platelet activation via flow cytometer (FACS Calibur cytometer; Becton Dickinson, San Jose, CA) was performed after incubation of citrated blood with monoclonal antibodies against CD40, CD40 ligand [CD40L], and CD62P on platelet surface without stimulation, as previously described (52). Phycoerythin-conjugated anti-CD40 and anti-CD40L with phycoerythin-conjugated IgG1 and fluorescein isothiocyanate– conjugated anti-CD62P served as isotype controls (Beckman
Psychosomatic Medicine, V 77 • 145-155
147
February/March 2015
Copyright © 2015 by the American Psychosomatic Society. Unauthorized reproduction of this article is prohibited.
ORIGINAL ARTICLE
RESULTS
Markers of Platelet Activity
Table 1 displays demographic and medical data for each group. Nondepressed T2DM patients were older than depressed patients without T2DM. T2DM patients had higher HbA1c, had higher BMI, and received antihypertensive treatment more frequently than did patients without T2DM. Severity of depression was similar in depressed patients with or without T2DM.
MANOVA revealed an overall main effect of depression (Pillai's trace = 0.527, F(6,60) = 11.14, p < .001, η2 = 0.527), diabetes (Pillai's trace = 0.566, F(6,60) = 13.07, p < .001, η2 = 0.566), and an interaction (Pillai's trace = 0.187, F(6,60) = 2.30, p = .046, η2 = 0.187) on platelet markers (CD40, sCD40, CD40L, CD62P, β-TG, PF-4). After adjustment for BMI and age, main effects of diabetes and depression remained, but the interaction was reduced to a trend (Pillai's trace = 0.174, F(6,58) = 2.03, p = .076, η2 = 0.174). Neither multivariate nor univariate effects of BMI and age were observed (η2 ranging from 0.001 to 0.041). ANOVAs showed a significant interaction of MD T2DM only for β-TG (F(1,76) = 10.72, p = .002, η2 = 0.124). As displayed in Table 2, analysis of simple main effects revealed significantly higher β-TG only in depressed patients compared with healthy controls. A trend for an interaction was observed for sCD40L (F(1,77) = 2.87, p = .094, η2 = 0.032), with simple main effects analysis showing significantly higher sCD40L levels only in comorbid patients compared with T2DM patients. For all other platelet markers except PF-4, ANOVAs yielded significant main effects for MD (CD40: F(1,70) = 33.22, p < .001, η2 = 0.322; CD40L: F(1,75) = 7.43, p = .008, η2 = 0.090; CD62P: F(1,74) = 13.39, p < .001, η2 = 0.153) and main effects for T2DM (CD40: F(1,70) = 15.52, p < .001, η2 = 0.182; CD40L: F(1,75) = 9.80, p = .002, η2 = 0.116; CD62P: F(1,74) = 16.35, p
