Acta Psychiatr Scand 2015: 1–7 All rights reserved DOI: 10.1111/acps.12417
© 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd ACTA PSYCHIATRICA SCANDINAVICA
Damage-associated molecular patterns and immune activation in bipolar disorder Stertz L, Fries GR, Rosa AR, Kauer-Sant’anna M, Ferrari P, ^ Paz AVC, Green C, Cunha ABM, Dal-Pizzol F, Gottfried C, Kapczinski F. Damage-associated molecular patterns and immune activation in bipolar disorder. Objective: Immune activation in bipolar disorder (BD) has been frequently reported. Damage-associated molecular patterns (DAMPs) are key players in the immune activation reaction. The aim of this study was to assess DAMP levels in drug-free patients with BD during acute episodes. Method: Serum levels of a predetermined set of DAMPs were assessed in drug-free patients with BD (n = 20) during an acute mood episode. We also included two control groups: healthy subjects, used as a negative control (n = 20); and patients with sepsis, used as a positive control for severe immune activation (n = 20). Results: Multivariate analysis using generalized linear mixed model indicated that all DAMPs differed as a function of group membership after controlling for age and addressing multiplicity (P < 0.0006 for all comparisons). Follow-up analyses showed higher levels in BD subjects of circulating cell-free (ccf) nuclear (n)DNA (P = 0.02), HSP70 (P = 0.03) and HSP90a (P = 0.02) as compared to healthy subjects. Also, patients with BD showed lower levels of ccf nDNA (P = 0.04), HSP60 (P = 0.03), HSP70 (P = 0.01), and HSP90a (P = 0.002) as compared to patients with sepsis and higher levels of ccf mitochondrial DNA (P < 0.0001). Conclusion: The present findings may be linked to the inflammatory activity previously described among patients with BD and may help in the development of more targeted and personalized treatments for patients under acute episodes of BD.
L. Stertz1,2,3, G. R. Fries1,2,
A. R. Rosa2,4, M. KauerSant’anna1,2, P. Ferrari2, A. V. C. Paz2, C. Green5, ^ B. M. Cunha4,6, F. Dal-Pizzol7, A. C. Gottfried1, F. Kapczinski2,3 1 Programa de Pos-Graduacß~ao em Ci^encias Biologicas: Bioquímica, Universidade Federal do Rio Grande do Sul, Porto Alegre, BRA, 2Bipolar Disorder Program & INCT Translational Medicine, Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil, 3Department of Psychiatry and Behavioral Sciences, Center for Molecular Psychiatry, University of Texas Health Science Center, Houston, TX, USA, 4Department of Pharmacology, Basic Science Institute, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil, 5Department of Pediatrics, Center for Clinical Research and Evidence-Based Medicine, The University of Texas Health Science Center at Houston, Houston, TX, USA, 6Department of Neuropsychiatry, Centro de Ci^encias da Saude, Universidade Federal de Santa Maria, Santa Maria, Brazil and 7Laboratorio de FisiopatologiaExperimental, Unidade Academica de Ciencias da Saude, Universidade do Extremo Sul Catarinense, Criciuma, SC, USA
Key words: bipolar disorder; cell death; circulating cell-free DNA; heat-shock proteins; acute mood episode; damage-associated molecular patterns Flavio Kapczinski, Bipolar Disorder Program & INCT Translational Medicine, Hospital de Clínicas de Porto Alegre, 2350 Ramiro Barcelos, Centro de Pesquisa Experimental, Porto Alegre, Rio Grande do Sul, Brazil. E-mail: [email protected]
Accepted for publication March 9, 2015
Significant outcomes
• Low-grade immune activation has been reported in bipolar disorder during acute episodes and euthymia.
• The immune activation response may be generated and amplified by the actions of the damage-associated molecular patterns (DAMPs).
• Increased levels of circulating DAMPs occur during acute episodes of bipolar disorder and may be related with the pattern of immune activation reported in these patients.
Limitations
• Cross-sectional study, relatively small sample size, admixture of patients with type I and type II bipolar disorder.
• Not all damage-associated molecular patterns were assessed in this study. 1
Stertz et al. Introduction
Material and methods
Bipolar disorder (BD) is a severe and chronic disorder that affects about 2.5% of the population (1). BD has been associated with significant disability, morbidity, and premature mortality (2, 3). Moreover, BD is regularly accompanied by clinical comorbidities. Patients are at higher risk for developing a wide range of medical conditions, including metabolic syndrome (diabetes, obesity, dyslipidemia, and insulin resistance), cardiovascular, and cerebrovascular disease (4). The Systematic Treatment Enhancement Program for Bipolar Disorder (STEP-BD) demonstrated that 50% of the patients with BD present one or more associated comorbidities (5). In addition, there is increasing evidence showing that BD is associated with systemic alterations including increased inflammatory markers, oxidative stress, decreased BDNF, and DNA damage (6). The term immune activation can be defined as ‘a large range of events involved in molecular and cellular processes related to cell activation, proliferation and death, as well as their consequences’ (7). Immune activation in BD has been related to the severity and recurrence of mood episodes (8), illness progression (9, 10), and higher rates of comorbidities (9). Furthermore, it has been shown that cellular death is associated with BD and probably involved in its pathophysiology (11–13). For instance, a recent report showed increased early apoptosis in peripheral blood mononuclear cells from patients with BD (14). In this sense, the presence of damageassociated molecular patterns (DAMPs), which are endogenous signals released during cell stress, damage, or death, could act as a trigger for the immune activation and systemic toxicity observed in BD (15, 16). Nonetheless, no study to date has examined DAMP levels in patients with BD.
Subjects
Aims of the study
The aim of this study was to assess serum levels of a previously defined set of damage-associated molecular patterns (circulating cell-free nuclear DNA, ccf mitochondrial (mt)DNA, HSP70, HSP90a, and HSP60) in drug-free patients with BD during acute episodes. We also included two control groups: one including healthy subjects, taken as a negative control; and another group including patients with sepsis, used as a positive control for immune activation. 2
Patients with BD (n = 20) were recruited from subjects treated at the Bipolar Disorders Program of Clinical Hospital of Porto Alegre and University Hospital of Santa Maria, both in southern Brazil. All patients underwent a comprehensive clinical interview by a psychiatrist, and diagnosis was based on the Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I). Inclusion criteria included a diagnosis of BD and presence of a manic, depressive, or mixed episode at present according to DSM-IV-TR criteria. Patients also had to be medication-free for at least 2 weeks (6 weeks in the case of fluoxetine or depot antipsychotics), except for benzodiazepines. Exclusion criteria included DSM-IV diagnoses other than bipolar disorder I or II, current abuse of illicit substance, and presence of metabolic disorders such as diabetes, chronic inflammatory diseases (e.g., inflammatory bowel disease or rheumatoid arthritis), cancer, or other severe medical conditions. The use of both a negative and a positive control groups in this case–control study was based on a previous work (15). Negative controls (n = 20) were individuals with no history of psychiatric illness, no current abuse of illicit substance, and presence of metabolic disorders or chronic inflammatory diseases, cancer, or other severe medical conditions. Subjects were mostly family members of people using other healthcare facilities at the same hospital and were matched by frequency to the BD group according to age and gender. Positive controls were patients with sepsis (n = 20) selected at an intensive care unit and diagnosed according to American College of Chest Physicians/Society of Critical Care Medicine criteria and were selected by convenience (17). The investigation was carried out in accordance with the latest version of the Declaration of Helsinki, and the study design was reviewed by the Research Ethics Committee of Clinical Hospital of Porto Alegre. All participants provided written informed consent (which was approved by the local ethics committee) after the nature of the procedures had been fully explained. Assessment
Sociodemographic and clinical data were collected in an extensive interview session conducted with the patients. All patients were evaluated using the same protocol. Manic and depressive symptoms were evaluated using the Young Mania Rating
DAMPs in bipolar disorder Scale (YMRS) and the Hamilton Depression Rating Scale, 21-item version (HDRS-21), respectively (18, 19). Functioning was evaluated using the Global Assessment of Functioning (GAF) scale (20). Blood samples
Ten milliliters of blood were withdrawn from each subject by venipuncture prior to initial treatment. The blood without anticoagulant was immediately centrifuged at 2000 9 g for 10 min, and serum was kept frozen at 80°C until assayed.
60°C, and 15 s at 95°C. Each sample was analyzed in duplicate. ccf DNA concentrations were estimated according to standard curves, using the known concentration of human genomic DNA with a dilution factor of 5 (starting with 5 ng/ll). To minimize errors, the sample obtained from each patient and their respective controls were always run in the same plate. Results were expressed in genome equivalents (GE) per millilitre of serum, based on the methodology described by Xia et al. (21). Statistical analysis
Assays
All assays were performed in serum samples according to the manufacturer’s instructions. Heat-shock proteins (HSP70, HSP90a, and HSP60) were measured in serum using commercially available ELISA kits in accordance with manufacturer’s instructions (Enzo Biosciences, Farmingdale, NY, USA). We optimized the dilution of the samples as follows: HSP70 – 1 : 4, HSP90a – 1 : 10, and HSP60 – no dilution. Values are expressed as ng/ml. Serum circulating cell-free DNA quantification
Serum circulating cell-free DNA was extracted from 200 ll of serum using a commercial kit (QIAmp DNA Mini Kit; Qiagen, Valencia, CA, USA) and eluted in a final volume of 100 ll. DNA was stored at 20°C until use. For circulating cellfree (ccf) nuclear (n) DNA analysis, the glyceraldehyde-3-phosphodehydrogenase (GAPDH) housekeeping gene was assessed with the following primers: 50 -CCCACTCCTGATTTCTGGAAAAGAG-30 (forward), 50 -GTCCCAGGGCTTTGAT TTGC-30 (reverse), and 50 -FAM-CAAGTTGC CTGTCCTTCC-MGB-30 (probe). For ccf mitochondrial (mt)DNA analysis, a sequence of the mtDNA-encoded ATPase (MTATP8) gene starting at locus 8446 was amplified with forward primer 50 AATATTAAACACAAACTACCACCTA CCTC-30 , reverse primer 50 -GTTCATTTTG GTTCTCAGGGTTTGTT-30 , and 50 -6-FAMCCTCACCAAAGCCC-MGB-30 as the probe. Singleplex real-time PCR was performed using the ABI PRISM 7500 Sequence Detection System (Applied Biosystems, Foster City, CA, USA). A total reaction volume of 12.5 ll was used, containing 2.5 ll DNA, 6.25 ll TaqManÒ Universal PCR Master Mix, 4 primers, 2 probes (0.625 ll), and 2.5 lL of sample. The mixture was incubated for 2 min at 50°C, followed by an initial denaturation step at 95°C for 10 min, 40 cycles of 1 min at
All biochemical results were Box–Cox transformed for parametric analysis (22). Although theoretically driven, selection of DAMPs still resulted in data with high enough dimensionality to complicate interpretation. Therefore, all dependent variables were transformed into z-scores to insure that they were all in the same metric. The data analytic strategy was to first evaluate an omnibus model (the generalized linear mixed model – GLMM) that permitted an overall test of the degree to which DAMP levels showed differential multivariate profiles as a function of group membership after controlling for age. We use the term ‘multivariate’ in the conventional statistical sense to refer to an equation with multiple dependent measures (the DAMPs), not multiple predictors. This omnibus test used an alpha = 0.05. Follow-up evaluation of the statistically reliable finding from the multivariate analysis required that we identify which dependent measures were responsible for the statistically reliable omnibus test. These analyses proceeded in univariate fashion where we use the term ‘univariate’ to refer to equations in which there was only one dependent variable. Age and group remained predictors in these univariate equations. The subsequent evaluation of group differences required multiple tests to evaluate which DAMPs showed differences as a function of group after controlling for age. These follow-up tests corrected for multiplicity using a Holm–Bonferroni step-down procedure that maintained the overall alpha for these follow-up tests at 0.05. The software used for the analysis was SAS v. 9.3 (SAS Institute Inc., Cary, NC, USA). The initial omnibus test was conducted using alpha = 0.05. Results
Groups were matched by frequency according to age and sex, except for septic patients (ages between 40 and 78 years old) (see Table 1 for sample characteristics and Table 2 for median DAMP 3
Stertz et al. Table 1. Demographic and clinical characteristics of participants
Age at enrollment (years)* Female sex† Self-reported clinical illness† Disorder duration* BMI* Type I BD/type II BD (N) YMRS* HAMD* GAF*
Univariate analyses
Healthy subjects (n = 20)
Sepsis (n = 20)
Bipolar Disorder (n = 20)
35 13 13 (65%) – – 25.15 3.62
59 13 4 (20%) – –
– – –
– – –
35 13 12 (60%) 6 (40%) 14 13 25.8 4.8 17/3 19 14 13 11 41 19
YMRS, Young Mania Rating Scale; HDRS, Hamilton Depression Rating Scale; GAF, Global Assessment of Functioning scale; BD, bipolar disorder. *Values are indicated as mean SD. †Values are indicated as N (percentage).
levels according to condition). Twenty patients with BD were included in the study. Of them, 50% had a current manic episode, 30% depressive episode, and 20% mixed episode. No differences were found in DAMP levels across different mood states. One patient with BD was using haloperidol at the time of assessment. Two patients with sepsis were currently using antipsychotics due to the presence of delirium. Multivariate analyses
Standard profile analysis using repeated measures MANOVA failed to satisfy the homogeneity of variance assumption (v2(30) = 57.86, P < 0.002). Although such tests are highly sensitive to departures from homogeneity of variance, the small sample with roughly equal cell sizes suggested that such a violation existed. After controlling for age (F(1,53) = 2.58, P < 0.114), group (F(2,53) = 21.87, P < 0.0001), and DAMPs (F(4,206) = 0.25, P < 0.910), analyses identified a group by DAMPs interaction (F(8,206) = 13.90, P < 0.0001) suggesting that DAMP levels demonstrated differential profiles as a function of group (Table 3). Analyses used standardized (z-score) values for analysis to enhance interpretation and graphing. Figure 1 depicts z-scored, least-square means estimates as a function of group and DAMPs.
Univariate GLMM analyses indicated that all DAMPs differed as a function of group membership after controlling for age and addressing multiplicity (Table 4). Follow-up analyses found that negative controls and BD participants showed higher levels of ccf mtDNA, and lower levels of ccf nDNA, HSP60, HSP70, and HSP90a compared to positive controls (Table 4). Patients with BD showed higher levels of ccf nDNA, HSP70, and HSP90a than negative controls (Table 4). Discussion
This is the first study to show increased DAMP levels in serum of patients with BD. DAMPs are cellular constituents that can be identified by the innate immune system once released from dying cells (23). They include sugars, metabolites, lipids, and nucleic acids such as RNA and DNA (24), which can bind to TLR, activating several signaling pathways and culminating in a sterile (nonpathogen-induced) inflammation response (23). The increase of those molecules corroborates the evidences of increased cell death in patients with BD (14), providing a link between BD, immune activation, and systemic toxicity (15). The classical concept that immune system provides the organisms the ability to protect themselves from exogenous pathogens, toxins, and to repair tissue damage that results from infection or trauma is changing. A broader understanding of the immune system role in central nervous system (CNS) suggests that it may play a central role in homeostasis, leading to adaptation (synaptic plasticity, neuroprotection, regeneration) and enabling the CNS to deal with the increase in metabolic demands (25). However, this homeostasis can become maladaptive and neurotoxic in states of enhanced neuronal activity, such as psychological stress, noxious stimuli, or epileptic seizures (25). The maladaptive immune activation can persist for longer than necessary in a state of low degree of inflammation or even spread to remote sites (25, 26).
Table 2. DAMP levels in patients and controls DAMPs ccf nDNA (GE/ml) ccf mtDNA (GE/ml) HSP90a (ng/ml) HSP70 (ng/ml) HSP60 (ng/ml)
Healthy subjects
Bipolar disorder
Sepsis
743 (567–1254) 158 481 (49 411–343 858) 13 (12–16) 1 (0.95–1.23) 1.54 (1–1.93)
1870 (1318–2553) 190 730 (48 283–517 042) 20 (14–34) 1.29 (1–1.64) 2.35 (1.51–5.47)
2242 (1708–2981) 18 213 (12 084–29 224) 45 (31–91) 2.61 (1.63–5.23) 3.82 (2.57–10.85)
DAMPs, damage-associated molecular patterns; ccf nDNA, circulating cell-free nuclear DNA; ccf mtDNA, circulating cell-free mitochondrial DNA, HSP, heat-shock protein. Data represent as median and interquartile range (IQR).
4
DAMPs in bipolar disorder Table 3. Multivariate analysis results of DAMPs assessed as a function of group DAMPs
F value
NumDF
DenDF
P-value
Holm–Bonferroni P-value
cf nDNA cf mtDNA HSP60 HSP70 HSP90a
13.27 38.14 10.68 14.40 20.84
2 2 2 2 2
19 19 19 19 19
0.0002
© 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd ACTA PSYCHIATRICA SCANDINAVICA
Damage-associated molecular patterns and immune activation in bipolar disorder Stertz L, Fries GR, Rosa AR, Kauer-Sant’anna M, Ferrari P, ^ Paz AVC, Green C, Cunha ABM, Dal-Pizzol F, Gottfried C, Kapczinski F. Damage-associated molecular patterns and immune activation in bipolar disorder. Objective: Immune activation in bipolar disorder (BD) has been frequently reported. Damage-associated molecular patterns (DAMPs) are key players in the immune activation reaction. The aim of this study was to assess DAMP levels in drug-free patients with BD during acute episodes. Method: Serum levels of a predetermined set of DAMPs were assessed in drug-free patients with BD (n = 20) during an acute mood episode. We also included two control groups: healthy subjects, used as a negative control (n = 20); and patients with sepsis, used as a positive control for severe immune activation (n = 20). Results: Multivariate analysis using generalized linear mixed model indicated that all DAMPs differed as a function of group membership after controlling for age and addressing multiplicity (P < 0.0006 for all comparisons). Follow-up analyses showed higher levels in BD subjects of circulating cell-free (ccf) nuclear (n)DNA (P = 0.02), HSP70 (P = 0.03) and HSP90a (P = 0.02) as compared to healthy subjects. Also, patients with BD showed lower levels of ccf nDNA (P = 0.04), HSP60 (P = 0.03), HSP70 (P = 0.01), and HSP90a (P = 0.002) as compared to patients with sepsis and higher levels of ccf mitochondrial DNA (P < 0.0001). Conclusion: The present findings may be linked to the inflammatory activity previously described among patients with BD and may help in the development of more targeted and personalized treatments for patients under acute episodes of BD.
L. Stertz1,2,3, G. R. Fries1,2,
A. R. Rosa2,4, M. KauerSant’anna1,2, P. Ferrari2, A. V. C. Paz2, C. Green5, ^ B. M. Cunha4,6, F. Dal-Pizzol7, A. C. Gottfried1, F. Kapczinski2,3 1 Programa de Pos-Graduacß~ao em Ci^encias Biologicas: Bioquímica, Universidade Federal do Rio Grande do Sul, Porto Alegre, BRA, 2Bipolar Disorder Program & INCT Translational Medicine, Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil, 3Department of Psychiatry and Behavioral Sciences, Center for Molecular Psychiatry, University of Texas Health Science Center, Houston, TX, USA, 4Department of Pharmacology, Basic Science Institute, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil, 5Department of Pediatrics, Center for Clinical Research and Evidence-Based Medicine, The University of Texas Health Science Center at Houston, Houston, TX, USA, 6Department of Neuropsychiatry, Centro de Ci^encias da Saude, Universidade Federal de Santa Maria, Santa Maria, Brazil and 7Laboratorio de FisiopatologiaExperimental, Unidade Academica de Ciencias da Saude, Universidade do Extremo Sul Catarinense, Criciuma, SC, USA
Key words: bipolar disorder; cell death; circulating cell-free DNA; heat-shock proteins; acute mood episode; damage-associated molecular patterns Flavio Kapczinski, Bipolar Disorder Program & INCT Translational Medicine, Hospital de Clínicas de Porto Alegre, 2350 Ramiro Barcelos, Centro de Pesquisa Experimental, Porto Alegre, Rio Grande do Sul, Brazil. E-mail: [email protected]
Accepted for publication March 9, 2015
Significant outcomes
• Low-grade immune activation has been reported in bipolar disorder during acute episodes and euthymia.
• The immune activation response may be generated and amplified by the actions of the damage-associated molecular patterns (DAMPs).
• Increased levels of circulating DAMPs occur during acute episodes of bipolar disorder and may be related with the pattern of immune activation reported in these patients.
Limitations
• Cross-sectional study, relatively small sample size, admixture of patients with type I and type II bipolar disorder.
• Not all damage-associated molecular patterns were assessed in this study. 1
Stertz et al. Introduction
Material and methods
Bipolar disorder (BD) is a severe and chronic disorder that affects about 2.5% of the population (1). BD has been associated with significant disability, morbidity, and premature mortality (2, 3). Moreover, BD is regularly accompanied by clinical comorbidities. Patients are at higher risk for developing a wide range of medical conditions, including metabolic syndrome (diabetes, obesity, dyslipidemia, and insulin resistance), cardiovascular, and cerebrovascular disease (4). The Systematic Treatment Enhancement Program for Bipolar Disorder (STEP-BD) demonstrated that 50% of the patients with BD present one or more associated comorbidities (5). In addition, there is increasing evidence showing that BD is associated with systemic alterations including increased inflammatory markers, oxidative stress, decreased BDNF, and DNA damage (6). The term immune activation can be defined as ‘a large range of events involved in molecular and cellular processes related to cell activation, proliferation and death, as well as their consequences’ (7). Immune activation in BD has been related to the severity and recurrence of mood episodes (8), illness progression (9, 10), and higher rates of comorbidities (9). Furthermore, it has been shown that cellular death is associated with BD and probably involved in its pathophysiology (11–13). For instance, a recent report showed increased early apoptosis in peripheral blood mononuclear cells from patients with BD (14). In this sense, the presence of damageassociated molecular patterns (DAMPs), which are endogenous signals released during cell stress, damage, or death, could act as a trigger for the immune activation and systemic toxicity observed in BD (15, 16). Nonetheless, no study to date has examined DAMP levels in patients with BD.
Subjects
Aims of the study
The aim of this study was to assess serum levels of a previously defined set of damage-associated molecular patterns (circulating cell-free nuclear DNA, ccf mitochondrial (mt)DNA, HSP70, HSP90a, and HSP60) in drug-free patients with BD during acute episodes. We also included two control groups: one including healthy subjects, taken as a negative control; and another group including patients with sepsis, used as a positive control for immune activation. 2
Patients with BD (n = 20) were recruited from subjects treated at the Bipolar Disorders Program of Clinical Hospital of Porto Alegre and University Hospital of Santa Maria, both in southern Brazil. All patients underwent a comprehensive clinical interview by a psychiatrist, and diagnosis was based on the Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I). Inclusion criteria included a diagnosis of BD and presence of a manic, depressive, or mixed episode at present according to DSM-IV-TR criteria. Patients also had to be medication-free for at least 2 weeks (6 weeks in the case of fluoxetine or depot antipsychotics), except for benzodiazepines. Exclusion criteria included DSM-IV diagnoses other than bipolar disorder I or II, current abuse of illicit substance, and presence of metabolic disorders such as diabetes, chronic inflammatory diseases (e.g., inflammatory bowel disease or rheumatoid arthritis), cancer, or other severe medical conditions. The use of both a negative and a positive control groups in this case–control study was based on a previous work (15). Negative controls (n = 20) were individuals with no history of psychiatric illness, no current abuse of illicit substance, and presence of metabolic disorders or chronic inflammatory diseases, cancer, or other severe medical conditions. Subjects were mostly family members of people using other healthcare facilities at the same hospital and were matched by frequency to the BD group according to age and gender. Positive controls were patients with sepsis (n = 20) selected at an intensive care unit and diagnosed according to American College of Chest Physicians/Society of Critical Care Medicine criteria and were selected by convenience (17). The investigation was carried out in accordance with the latest version of the Declaration of Helsinki, and the study design was reviewed by the Research Ethics Committee of Clinical Hospital of Porto Alegre. All participants provided written informed consent (which was approved by the local ethics committee) after the nature of the procedures had been fully explained. Assessment
Sociodemographic and clinical data were collected in an extensive interview session conducted with the patients. All patients were evaluated using the same protocol. Manic and depressive symptoms were evaluated using the Young Mania Rating
DAMPs in bipolar disorder Scale (YMRS) and the Hamilton Depression Rating Scale, 21-item version (HDRS-21), respectively (18, 19). Functioning was evaluated using the Global Assessment of Functioning (GAF) scale (20). Blood samples
Ten milliliters of blood were withdrawn from each subject by venipuncture prior to initial treatment. The blood without anticoagulant was immediately centrifuged at 2000 9 g for 10 min, and serum was kept frozen at 80°C until assayed.
60°C, and 15 s at 95°C. Each sample was analyzed in duplicate. ccf DNA concentrations were estimated according to standard curves, using the known concentration of human genomic DNA with a dilution factor of 5 (starting with 5 ng/ll). To minimize errors, the sample obtained from each patient and their respective controls were always run in the same plate. Results were expressed in genome equivalents (GE) per millilitre of serum, based on the methodology described by Xia et al. (21). Statistical analysis
Assays
All assays were performed in serum samples according to the manufacturer’s instructions. Heat-shock proteins (HSP70, HSP90a, and HSP60) were measured in serum using commercially available ELISA kits in accordance with manufacturer’s instructions (Enzo Biosciences, Farmingdale, NY, USA). We optimized the dilution of the samples as follows: HSP70 – 1 : 4, HSP90a – 1 : 10, and HSP60 – no dilution. Values are expressed as ng/ml. Serum circulating cell-free DNA quantification
Serum circulating cell-free DNA was extracted from 200 ll of serum using a commercial kit (QIAmp DNA Mini Kit; Qiagen, Valencia, CA, USA) and eluted in a final volume of 100 ll. DNA was stored at 20°C until use. For circulating cellfree (ccf) nuclear (n) DNA analysis, the glyceraldehyde-3-phosphodehydrogenase (GAPDH) housekeeping gene was assessed with the following primers: 50 -CCCACTCCTGATTTCTGGAAAAGAG-30 (forward), 50 -GTCCCAGGGCTTTGAT TTGC-30 (reverse), and 50 -FAM-CAAGTTGC CTGTCCTTCC-MGB-30 (probe). For ccf mitochondrial (mt)DNA analysis, a sequence of the mtDNA-encoded ATPase (MTATP8) gene starting at locus 8446 was amplified with forward primer 50 AATATTAAACACAAACTACCACCTA CCTC-30 , reverse primer 50 -GTTCATTTTG GTTCTCAGGGTTTGTT-30 , and 50 -6-FAMCCTCACCAAAGCCC-MGB-30 as the probe. Singleplex real-time PCR was performed using the ABI PRISM 7500 Sequence Detection System (Applied Biosystems, Foster City, CA, USA). A total reaction volume of 12.5 ll was used, containing 2.5 ll DNA, 6.25 ll TaqManÒ Universal PCR Master Mix, 4 primers, 2 probes (0.625 ll), and 2.5 lL of sample. The mixture was incubated for 2 min at 50°C, followed by an initial denaturation step at 95°C for 10 min, 40 cycles of 1 min at
All biochemical results were Box–Cox transformed for parametric analysis (22). Although theoretically driven, selection of DAMPs still resulted in data with high enough dimensionality to complicate interpretation. Therefore, all dependent variables were transformed into z-scores to insure that they were all in the same metric. The data analytic strategy was to first evaluate an omnibus model (the generalized linear mixed model – GLMM) that permitted an overall test of the degree to which DAMP levels showed differential multivariate profiles as a function of group membership after controlling for age. We use the term ‘multivariate’ in the conventional statistical sense to refer to an equation with multiple dependent measures (the DAMPs), not multiple predictors. This omnibus test used an alpha = 0.05. Follow-up evaluation of the statistically reliable finding from the multivariate analysis required that we identify which dependent measures were responsible for the statistically reliable omnibus test. These analyses proceeded in univariate fashion where we use the term ‘univariate’ to refer to equations in which there was only one dependent variable. Age and group remained predictors in these univariate equations. The subsequent evaluation of group differences required multiple tests to evaluate which DAMPs showed differences as a function of group after controlling for age. These follow-up tests corrected for multiplicity using a Holm–Bonferroni step-down procedure that maintained the overall alpha for these follow-up tests at 0.05. The software used for the analysis was SAS v. 9.3 (SAS Institute Inc., Cary, NC, USA). The initial omnibus test was conducted using alpha = 0.05. Results
Groups were matched by frequency according to age and sex, except for septic patients (ages between 40 and 78 years old) (see Table 1 for sample characteristics and Table 2 for median DAMP 3
Stertz et al. Table 1. Demographic and clinical characteristics of participants
Age at enrollment (years)* Female sex† Self-reported clinical illness† Disorder duration* BMI* Type I BD/type II BD (N) YMRS* HAMD* GAF*
Univariate analyses
Healthy subjects (n = 20)
Sepsis (n = 20)
Bipolar Disorder (n = 20)
35 13 13 (65%) – – 25.15 3.62
59 13 4 (20%) – –
– – –
– – –
35 13 12 (60%) 6 (40%) 14 13 25.8 4.8 17/3 19 14 13 11 41 19
YMRS, Young Mania Rating Scale; HDRS, Hamilton Depression Rating Scale; GAF, Global Assessment of Functioning scale; BD, bipolar disorder. *Values are indicated as mean SD. †Values are indicated as N (percentage).
levels according to condition). Twenty patients with BD were included in the study. Of them, 50% had a current manic episode, 30% depressive episode, and 20% mixed episode. No differences were found in DAMP levels across different mood states. One patient with BD was using haloperidol at the time of assessment. Two patients with sepsis were currently using antipsychotics due to the presence of delirium. Multivariate analyses
Standard profile analysis using repeated measures MANOVA failed to satisfy the homogeneity of variance assumption (v2(30) = 57.86, P < 0.002). Although such tests are highly sensitive to departures from homogeneity of variance, the small sample with roughly equal cell sizes suggested that such a violation existed. After controlling for age (F(1,53) = 2.58, P < 0.114), group (F(2,53) = 21.87, P < 0.0001), and DAMPs (F(4,206) = 0.25, P < 0.910), analyses identified a group by DAMPs interaction (F(8,206) = 13.90, P < 0.0001) suggesting that DAMP levels demonstrated differential profiles as a function of group (Table 3). Analyses used standardized (z-score) values for analysis to enhance interpretation and graphing. Figure 1 depicts z-scored, least-square means estimates as a function of group and DAMPs.
Univariate GLMM analyses indicated that all DAMPs differed as a function of group membership after controlling for age and addressing multiplicity (Table 4). Follow-up analyses found that negative controls and BD participants showed higher levels of ccf mtDNA, and lower levels of ccf nDNA, HSP60, HSP70, and HSP90a compared to positive controls (Table 4). Patients with BD showed higher levels of ccf nDNA, HSP70, and HSP90a than negative controls (Table 4). Discussion
This is the first study to show increased DAMP levels in serum of patients with BD. DAMPs are cellular constituents that can be identified by the innate immune system once released from dying cells (23). They include sugars, metabolites, lipids, and nucleic acids such as RNA and DNA (24), which can bind to TLR, activating several signaling pathways and culminating in a sterile (nonpathogen-induced) inflammation response (23). The increase of those molecules corroborates the evidences of increased cell death in patients with BD (14), providing a link between BD, immune activation, and systemic toxicity (15). The classical concept that immune system provides the organisms the ability to protect themselves from exogenous pathogens, toxins, and to repair tissue damage that results from infection or trauma is changing. A broader understanding of the immune system role in central nervous system (CNS) suggests that it may play a central role in homeostasis, leading to adaptation (synaptic plasticity, neuroprotection, regeneration) and enabling the CNS to deal with the increase in metabolic demands (25). However, this homeostasis can become maladaptive and neurotoxic in states of enhanced neuronal activity, such as psychological stress, noxious stimuli, or epileptic seizures (25). The maladaptive immune activation can persist for longer than necessary in a state of low degree of inflammation or even spread to remote sites (25, 26).
Table 2. DAMP levels in patients and controls DAMPs ccf nDNA (GE/ml) ccf mtDNA (GE/ml) HSP90a (ng/ml) HSP70 (ng/ml) HSP60 (ng/ml)
Healthy subjects
Bipolar disorder
Sepsis
743 (567–1254) 158 481 (49 411–343 858) 13 (12–16) 1 (0.95–1.23) 1.54 (1–1.93)
1870 (1318–2553) 190 730 (48 283–517 042) 20 (14–34) 1.29 (1–1.64) 2.35 (1.51–5.47)
2242 (1708–2981) 18 213 (12 084–29 224) 45 (31–91) 2.61 (1.63–5.23) 3.82 (2.57–10.85)
DAMPs, damage-associated molecular patterns; ccf nDNA, circulating cell-free nuclear DNA; ccf mtDNA, circulating cell-free mitochondrial DNA, HSP, heat-shock protein. Data represent as median and interquartile range (IQR).
4
DAMPs in bipolar disorder Table 3. Multivariate analysis results of DAMPs assessed as a function of group DAMPs
F value
NumDF
DenDF
P-value
Holm–Bonferroni P-value
cf nDNA cf mtDNA HSP60 HSP70 HSP90a
13.27 38.14 10.68 14.40 20.84
2 2 2 2 2
19 19 19 19 19
0.0002
