Abstract
Compelling evidence suggests that inflammation contributes to the development of depression. Many depressed individuals have higher levels of proinflammatory mediators, which appear to interact with many of the pathophysiological domains of depression, including neuroendocrine function, neurotransmitter metabolism, and synaptic plasticity. This is further supported by observation that therapeutic administration of interferon-α (IFN-α) leads to depression in a significant proportion of patients. These findings suggest that targeting proinflammatory cytokines and their signaling pathways may represent a unique therapeutic opportunity to treat depression and related conditions, such as labile anger, irritability, and fatigue.
Major depression, according to the World Health Organization, is a leading cause of disability worldwide1. With an annual prevalence of approximately 7% and lifetime prevalence of 16%, depression carries a significant burden to patients and society2,3. Depression is also associated with an increased risk of premature mortality4,5. While many depressed patients (up to 1% of women and 7% of men) eventually commit suicide6, the condition is also highly comorbid with medical disorders that are associated with increased mortality, such as cancer and some cardiovascular and autoimmune diseases7,8,9.
Although current antidepressants lead to some response in many cases, most patients are partially or completely unresponsive to treatment10. This is partly because the pathophysiology of depression has not been elucidated and currently available treatments do not effectively target all the pathological processes that are responsible for the major symptoms of depression. Thus, there is an urgent need to identify novel pathological pathways relevant to depression upon which new conceptual frameworks and targets to improve treatment outcomes can be built.
In recent years, significant attention has been directed at the relationship between the brain and peripheral organs. One promising development in this area is the emergence of inflammation and an activated immune system as common underlying mechanisms in the development of many neuropsychiatric diseases, including major depression. The aim of this article is to provide an overview of recent evidence in support of this hypothesis.
EVIDENCE FOR INCREASED INFLAMMATION IN DEPRESSION: THE ROLE OF PROINFLAMMATORY CYTOKINES
Patients with major depression, with or without other medical illness, have been found to have activated inflammatory pathways11. This includes elevations in relevant proinflammatory cytokines and acute-phase proteins, as well as increased expression of their biomarkers, including chemokines and adhesion molecules12. It is important to note that inflammatory markers have been found in the peripheral blood circulation and in the cerebrospinal fluid (CSF), suggesting that the brain is directly influenced by peripherally derived mediators12,13.
The effect of cytokines from the periphery on the brain is somewhat surprising due to the lack of specific transporters for larger peptides and proteins, such as cytokines at the blood-brain barrier. However, recent experimental evidence suggests at least 3 possible pathways by which such molecules could access the brain13,14,15. The first pathway is a leaky blood-brain barrier that occurs in major depression. The second pathway involves activation of the endothelial cells that line the cerebral vasculature and produce cytokines inside the blood-brain barrier. Finally, inflammatory mediators can bind to receptors associated with the vagus nerve and thereby signal inflammatory changes in the brain. Once in the brain, proinflammatory cytokines activate both neuronal and non-neuronal cells, similar to their activity in the peripheral inflammatory response16.
According to a recent metaanalysis by Howren, et al17, increased serum and/or plasma concentrations of interleukin 6 (IL-6) and C-reactive protein (CRP) are the most frequently observed inflammatory markers in depressed patients. Other cytokines implicated in depression include IL-1ß and tumor necrosis factor-α (TNF-α)18,19,20,21. Recent literature suggests that functional variants of the IL-1ß and TNF-α genes may increase risk for depression and reduce response to antidepressant therapy22,23. Further, according to a prospective study by van den Biggelaar, et al24, elevated levels of CRP and an increased capacity of leukocytes to produce IL-1 predicted later onset of depression in elderly individuals without a history of depression. This suggests that inflammation precedes and potentially leads to depression. Several studies have also reported positive correlations between severity of depressive symptoms and plasma concentrations of various inflammatory mediators21,25,26.
In addition to data that link inflammatory mediators with depressive symptoms, it has also been shown that both acute and chronic administration of cytokines or cytokine inducers [i.e., lipopolysaccharide (LPS) or vaccination] can cause behavioral symptoms that are similar to those found in people who suffer from major depression27,28. For example, individuals injected with LPS displayed short-term increases in symptoms of depression and anxiety27, and those given a Salmonella typhi vaccine exhibited depressed mood, fatigue, mental confusion, and psychomotor slowing28. Severity of symptoms, in both cases, correlated with the rise in inflammatory mediators in peripheral blood.
MECHANISMS OF CYTOKINE-INDUCED DEPRESSION
It has been repeatedly demonstrated that cytokines access the brain and interact with virtually every pathophysiological domain relevant to depression, including neurotransmitter metabolism, neuroendocrine function, neural circuitry, and synaptic plasticity29. A wealth of data indicate that, during prolonged periods of activation, cytokine networks in the central nervous system can induce abnormalities relevant to the pathophysiology of depression, including dysregulation of glial/neuronal interactions and cognitive function11. Peripheral administration of LPS, for instance, results in cognitive impairment and increased hippocampal concentrations of TNF-α and IL-130. This is linked to a decrease in hippocampal expression of brain-derived neurotrophic factor (BDNF). BDNF and its signaling pathways are involved in neuronal survival, synaptic plasticity, and hippocampus-dependent learning and memory31.
Once cytokine signals reach the brain, they can also influence the synthesis, release, and reuptake of mood-relevant neurotransmitters, including serotonin, norepinephrine, and dopamine16,32,33. It has also been observed that proinflammatory cytokines influence the hypothalamic-pituitary-adrenal axis34. Acute administration of cytokines has been shown to stimulate the expression and release of corticotropin-releasing hormone and adrenocorticotropic hormone, as well as cortisol, which have been found to be elevated in depressed patients34,35.
IFN-α AS A MODEL SYSTEM TO STUDY CYTOKINE-INDUCED DEPRESSION
Major depressive disorder during IFN-α therapy
IFN-α, an important cytokine in the early immune response to viral infection, has both antiproliferative and antiviral properties29. IFN plays a significant role in immune-mediated inflammatory diseases (IMID) such as systemic lupus and rheumatoid arthritis (RA)36. It has been used to treat a variety of malignancies and chronic viral infections, including malignant melanoma and hepatitis C. Despite its therapeutic benefits in these serious conditions, administration of IFN-α is often associated with high rates of neuropsychiatric symptoms, including depressed mood, anhedonia, fatigue, cognitive impairment, and sleep disturbances (Table 1)37,38,39. Suicidal ideation and completed suicides have also been reported37,40.
Clinical studies have confirmed that about 15%–45% of patients receiving IFN-α develop major depressive disorder (MDD) during the course of therapy41,42. In a double-blind, placebo-controlled study conducted by Musselman, et al42, 45% of patients (9 of 20) receiving IFN-α for malignant melanoma developed depression within 12 weeks of therapy (Figure 1). Severe depressive symptoms that required treatment discontinuation were noted in 7 (35%) patients.
Based on the Structured Clinical Interview for the Diagnostic and Statistical Manual of Mental Disorders and self-report questionnaire, Lotrich, et al43 identified 3 sets of symptoms that were worsened following IFN-α therapy. These include: (1) symptoms specific to depression, (2) hostility and labile anger, and (3) neurovegetative symptoms including fatigue and loss of appetite. In another study, Capuron, et al37 noted that neurovegetative and somatic symptoms such as anorexia, fatigue, and pain developed within 2 weeks of IFN-α therapy, and did not respond to antidepressant treatment. On the other hand, symptoms of depressed mood, anxiety, and cognitive impairment appeared later — i.e., between week 8 and 12 of IFN-α therapy — and were more responsive to paroxetine treatment.
Relationship between IFN-α, MDD, and circulating IL-6
Several studies have described increases in peripheral levels of IL-6 among patients undergoing IFN-α therapy44,45. There is also evidence that levels of IL-6 observed in the CSF of patients receiving IFN-α may negatively correlate with serotonin metabolism, which in turn negatively correlates with depression symptoms14.
According to Prather, et al46, patients who developed MDD during IFN-α treatment had higher Beck Depression Inventory (BDI) scores at baseline and were more likely to have a history of a mood disorder compared to those not affected with MDD. Further, patients who developed MDD displayed higher levels of IL-6 throughout treatment compared to those without MDD (p < 0.01; Figure 2). In patients who did not develop MDD, IL-6 concentrations stayed stable and low during the treatment. As in previous studies, MDD symptoms were apparent within 3 months. The study also found that higher levels of pretreatment IL-6 (> 1.25 pg/ml) predicted the development of MDD during IFN-α therapy (p < 0.05; Figure 3). In addition, poor sleep quality was associated with subsequent depression during IFN-α therapy, although the relationship was not bidirectional47. Similarly, increased levels of IL-6 were associated with subsequent poor sleep. Yet poor sleep did not lead to an increase in IL-6.
Risk of depression during IFN-α is affected by serotonin transporter polymorphism
Several publications indicate that polymorphisms in the promoter region (5-HTTLPR) of the serotonin transporter may interact with the inflammatory system and influence risk of depression48,49,50. According to a metaanalysis47 and large association studies48, the short/short (S/S) genotype in this region increases risk of MDD. Recently, Lotrich, et al49 examined 5-HTTLPR in 71 nondepressed patients with hepatitis C prior to initiation of IFN-α therapy. Patients were genotyped for the 5-HTTLPR (LG, LA, and S) and the variable number of tandem repeats polymorphism in the second intron. According to the study results, the LA allele was associated with lower rates of MDD and the S/S genotype appeared the most susceptible to depression while on IFN-α treatment (Figure 4).
Efficacy of serotonin reuptake inhibitors in preventing IFN-α-induced MDD
Prevention studies evaluating the prophylactic use of selective serotonin reuptake inhibitors (SSRI) prior to initiation of IFN-α have produced conflicting results, indicating that SSRI may not be equally effective in this patient population42,50,51,52,53,54. In one study, only 11% of patients (2 of 18) treated with paroxetine developed MDD while taking IFN-α treatment compared to 45% of placebo-treated patients42. Similar results were obtained in 3 open-label trials of prophylactic SSRI in nondepressed patients with hepatitis C virus: only 9% (3 of 32) developed IFN-α-related MDD, despite the fact that all patients had a history of affective disorder50,51,52. However, these findings were not confirmed by 2 small studies also conducted in patients with hepatitis C53,54. The exploratory analysis of one of these studies indicated that subjects who already had high pretreatment baseline levels of depressive symptoms may benefit the most from the prophylactic use of SSRI54. This observation is consistent with findings that pretreatment baseline depression-rating scale scores are predictive of subsequent IFN-α-induced MDD41,43,49. Therefore, it is reasonable to assume that treating patients with histories of recurrent depressive symptoms might prevent further progression to IFN-α-related MDD, a possibility that requires further examination.
In an animal model of depression, colitis emerged; when the animal was treated with antidepressants, not only did the depression improve but so did the colitis, raising the possibility that antidepressants can reduce depression-triggered inflammation. Their influence on IMID is not fully known55.
IFN-α and labile anger
In addition to MDD, about 25% of patients receiving IFN-α develop labile anger, irritability, and/or hostility56,57. This mood disorder, distinct from depression, appears to be associated with TNF-α polymorphism58. In a study that involved 105 patients with hepatitis C treated with IFN-α, labile anger was significantly increased in those with the A allele compared to those with G/G genotype. The TNF-α A allele was also associated with fatigue during the IFN-α treatment but not with MDD58. Further, labile anger was not predicted by the serotonin transporter polymorphism. Thus, pretreatment with SSRI does not appear to be effective in preventing the occurrence of labile anger in individuals treated with IFN-α.
IL-28 polymorphism is linked to efficacy and side effects of IFN-α therapy
In a recent study, Lotrich, et al59 confirmed that IL-28B polymorphism was associated with differences in sustained viral response as well as with loss of appetite, energy, and sleep complaints during IFN-α treatment. However, the IL-28B genotype did not predict development of depression.
The finding that polymorphisms in different genes are responsible for different psychosocial symptoms carries significant implications for the management of patients treated with IFN-α and subsequently other individuals affected by psychosocial disorders.
ANTIINFLAMMATORY AGENTS AND THEIR ANTIDEPRESSANT EFFECTS
If depression is associated with excess inflammation, it is to be expected that inhibition or reduction in inflammation would diminish depressive symptoms. Several studies in humans support this hypothesis and suggest that immune-targeted therapies may have clinical benefits in depressed patients60,61,62,63. For example, agents that block production of prostaglandin E2 [i.e., acetylsalicylic acid and cyclooxygenase (COX)-2 inhibitors], an important mediator of inflammation, also display antidepressant effects60,61,62. Further, the COX-2 inhibitor celecoxib was also shown to have positive effects on cognition, which is frequently impaired with depression. Müller, et al60 demonstrated significantly greater improvements in symptoms of depression in healthy depressed patients treated with the norepinephrine reuptake inhibitor reboxetine in combination with celecoxib than in those who received reboxetine alone (Figure 5).
Antidepressant activity of antiinflammatory agents has also been observed in patients with IMID. For example, a large double-blind, placebo-controlled trial demonstrated that patients with psoriasis who received the TNF-α and lymphotoxin-α antagonist etanercept exhibited significant improvement in depressive symptoms compared with placebo-treated subjects (Figure 6). The effect of etanercept on depressive symptoms was independent of improvement in disease activity64. These results are in line with an immense body of evidence in laboratory animals indicating that anti-inflammatory agents or cytokine antagonists can block the development of behavioral changes following immune activation65. In addition, these data are in accord with observations that TNF-α receptor knockout mice exhibit an antidepressant phenotype66.
CONCLUSION
Increasing evidence indicates that cytokine-mediated inflammation may be an underlying mechanism leading to a variety of IMID, but also to symptoms of depression, anxiety, and fatigue often observed in these patient populations. This hypothesis may explain significant improvements in depression-related symptoms after anti-TNF exposure. However, additional research is needed to further determine which underlying factors are responsible for the cascade of events that lead to psychosocial comorbidities in patients with IMID. Further, a broader conceptualization of depression in the context of IMID will permit consideration of the usefulness of different pharmacological approaches in the management of these complex conditions.
Footnotes
-
Supported by an unrestricted grant from Abbott Canada. Dr. Lotrich is supported by NIMH and NARSAD. Dr. Guenther has acted as consultant for Abbott Laboratories, Amgen Canada, Galderma Canada, LEO Pharma, Janssen, Schering-Plough Canada, and Wyeth; and has received investigator-initiated study support from Astellas Pharma Canada; and contract research support from Abbott Laboratories, Amgen Canada, Astellas Pharma Canada, Celgene Corporation, Centocor Ortho Biotech, EMD Serono Canada, Galderma Canada, Isotechnika, Janssen-Ortho, LEO Pharma, Novartis Pharmaceuticals Canada, Pfizer, Schering-Plough Canada, and Stiefel Laboratories.