Cytokines in bipolar disorder vs. healthy control subjects: A systematic review and meta-analysis

Abstract

Background

Bipolar disorder may be associated with peripheral immune system dysfunction; however, results in individual studies are conflicting. Our aim was to systematically review evidence of peripheral cytokine alterations in bipolar disorder integrating findings from various affective states.

Methods

We conducted a meta-analysis of studies comparing peripheral cytokine concentrations in patients with bipolar disorder with healthy control subjects. Results were reported according to the PRISMA statement.

Results

Eighteen studies with a total of 761 bipolar disorder patients and 919 healthy controls were included. Overall, concentrations of soluble Interleukin (IL)-2 receptor (sIL-2R), tumor necrosis factor-α (TNF-α), soluble tumor necrosis factor receptor type 1 (sTNFR1) (p < 0.001 each), sIL-6R (p = 0.01) and IL-4 (p = 0.04) were significantly higher in bipolar patients compared with healthy controls. There were no significant differences between bipolar disorder patients and healthy control subjects for IL-1, IL-2, IL-5, IL-6, IL-8, IL-10, IL-12, IL-1β, IL-1 receptor antagonist (IL-1RA), interferon-γ (IFN-γ), transforming growth factor-β1 (TGF-β1) and sTNFR2.

Conclusions

Employing a global approach, incorporating evidence across affective states, this meta-analysis found some support for peripheral inflammatory alterations in bipolar disorder. Results were limited by heterogeneity between studies, insufficient standardization and lacking control for confounders in individual studies. Further research exploring the role of the peripheral inflammatory system in relation to neuroinflammation is warranted.

Introduction

Current hypotheses regarding the neurobiological background for bipolar disorder involves impairment in both cellular energy regulation, the immune system and defects in neuroprotective mechanisms along with epigenetic abberations (Gardner and Boles, 2011; Grande et al., 2012). These components have been suggested to be central to the neuroprogressive changes observed in bipolar disorder (Berk, 2009; Berk et al., 2011) and inflammatory system dysfunction could potentially be a key integrating element in the pathophysiological mechanism behind the disorder (Kapczinski et al., 2008; Stertz et al., 2013).

One theory behind the involvement of the inflammatory system in bipolar disorder, the macrophage-T-lymphocyte theory, was initially proposed for depression and schizophrenia (Leonard, 2001; Smith, 1991; Smith & Maes, 1995), and recently extended to bipolar disorder. This theory states a role for a chronically activated immune response system mediated by macrophages (and in the brain, microglia) and T-lymphocytes as a basis for the disease, in which their production of cytokines and inflammatory substances renders a vulnerability and destabilization of the brain function, resulting in susceptibility towards environmental stressors and, in turn, precipitating mood disturbances.

Another theory, the microglial hypothesis, postulates that activated microglia in the brain, with the release of proinflammatory cytokines, exerts negative effects on the neuroprotective system, thereby mediating further pathophysiological disturbances in bipolar disorder (Rajkowska, 2000; Schroeter et al., 2011).

Recently, evidence from clinical studies has emerged in support of these theories, indicating a role for immune system dysregulation and inflammation in bipolar disorder (Goldstein et al., 2009). Immune system abberations have been demonstrated in peripheral blood using both in-vitro techniques (Kim et al., 2007; Knijff et al., 2007) and in-vivo analysis, showing alterations of inflammatory markers (Brietzke et al., 2009b; Cunha et al., 2008; Dickerson et al., 2007) and alterations of inflammation related gene signatures (Drexhage et al., 2010b; Padmos et al., 2008). Additionally, evidence from preclinicial studies suggests that modulation of neuroinflammatory processes is among the mechanisms of action of mood stabilizers such as lithium (Kang et al., 2012; Nahman et al., 2012) and valproic acid (Xuan et al., 2012), further indicating a role for neuroinflammation in bipolar disorder. Along these lines, treatment augmentation with nonsteroidal anti-inflammatory drugs (NSAID) may be of potential (Torrey & Davis, 2012) but has yet been explored in only one randomized trial, showing some effect on depressive symptoms in depressive and mixed episodes in bipolar disorder patients (Nery et al., 2008).

Several lines of evidence pointing to dysregulation of cellular metabolism, impaired neuroprotection and inflammation as key pathophysiological processes in bipolar disorder indicate that peripheral markers of these pathways are related to illness activity (Kapczinski et al., 2011). Alterations of peripheral markers have most consistently been demonstrated during acute illness episodes. However, both functional deficits and biomarker abberations also appear to be present in euthymia, indicating a role for peripheral biomarkers as trait markers as well (Breunis et al., 2003; Hsiao et al., 2009; Jamrozinski et al., 2009; Kauer-Sant'Anna et al., 2009), possibly reflecting pathophysiological processes present during euthymic periods. Thus disturbances in the inflammatory system appear to be present in all phases of the illness.

While alterations related to the inflammatory system have been reported in individual studies, results have been inconsistent. In a recent systematic review and meta-analysis of state related alterations of cytokine-, cytokine receptor and cytokine antagonists in bipolar disorder by our group, we found support of elevated levels of some cytokines and cytokine receptors in both manic patients and euthymic patients (Munkholm et al., 2013), with overall findings limited by heterogeneity, small sample sizes and lack of control for confounding factors. Stratification of the analysis according to affective state allowed for assessment of state-related alterations of cytokines, but also resulted in low study numbers and a lack of replicability for a majority of the cytokines investigated. Because of our finding of a largely identical direction of cytokine alterations in bipolar disorder patients compared with healthy controls regardless of the affective state, it may be relevant and well justified to do a global analysis, integrating research results from both euthymic, depressive and manic bipolar patients, with the potential of increasing the number of include studies, the statistical power of the analyses and the strength of evidence in this important area of research, and in this way broadening conclusions beyond our initial work.

We therefore conducted a meta-analysis of peripheral blood cytokine-, cytokine receptor and cytokine antagonist levels in bipolar disorder, integrating results globally, across affective states. Our primary objective was to investigate peripheral alterations of endogenous immune activity in bipolar disorder patients in comparison with healthy control subjects. Second, we wished to assess the effect of symptom severity, medication status and various clinical characteristics on cytokine-, cytokine receptor and cytokine antagonist levels.