Depression is one of the most common and serious side effects of IFN, which can limit its success as an antiviral or antitumor therapy. In this study, we used a simple depression model with definitive molecular targets, IFN-α and IFNAR. Rodents treated with mIFN-α were reported to show depressive behaviors and/or increased anxiety (Fahey et al., 2007; Makino et al., 1998, 2000b; Yamano et al., 2000), although several groups failed to reproduce those behavioral alterations (De La Garza et al., 2005; Loftis et al., 2006). This discrepancy might result from variations in experimental paradigms, including the types of IFN-α used, animal species or lines employed, treatment regimens, and behavioral tests performed. Because the interaction of IFN with IFNAR is highly species specific (Wang et al., 2008), we chose mouse IFN-α for our animal experiments. Clinical studies showed that patients frequently develop depressive symptoms after several weeks of IFN-α administration, but not within the first few weeks (Hauser et al., 2002; Raison et al., 2005). Similarly, 2-week mIFN-α treatments did not induce behavioral changes in our mice (Figure S1). Therefore, we used mice treated with mIFN-α for over a month to evaluate behavioral changes.
Systemic IFN treatment could affect various chemokine receptor functions other than emotional regulation, which might influence performance in tests assessing depressive-like behaviors. However, no studies have comprehensively investigated neurological and/or psychological alterations of IFN-treated animals. Using a battery of behavioral tests, we assessed a variety of brain functions. Whereas continuous mIFN-α treatment had no significant effect on general health or sensorimotor functions within the first 8 weeks, we noticed a gradual loss of body weight beginning in the ninth week (data not shown). Therefore, we included only the data obtained within the first 8 weeks in this report (Figures 4 and S4) and performed the depression-like behavioral tests (the tail-suspension test and the Porsolt forced swimming test) separately from the other tests, using mice immediately after the 5-week mIFN-α treatment. As a result, we found that mIFN-α treatment increased depression-like behaviors and impaired social interactions (Figures 4 and S4), consistent with the clinical symptoms of depression and independent of somatic conditions or sensorimotor functions. Taken together, we conclude that the IFN-α-treated mice are a reliable model for patients with IFN-induced depression, which is useful for analyzing the relationship between IFN-α’s effects on neural stem/progenitor cell function and on the induction of depressive behaviors.
IFN-α is reported to induce depression via upregulation of the hypothalamic-pituitary-adrenal (HPA) axis, alteration of monoamine neurotransmission, and induction of proinflammatory cytokines (Reyes-Vázquez et al., 2012; Schaefer et al., 2002). IFN-α can directly interact with opioid receptors (Jiang et al., 2000), which are also implicated in the induction of depressive behaviors (Makino et al., 2000a). Several proinflammatory cytokines, including interleukin (IL)-1 and IL-6, have been shown to modify the neurogenic behavior of NSCs (Gonzalez-Perez et al., 2012; Kohman and Rhodes, 2013). We previously found that IFN-α treatment suppresses cell proliferation in the hippocampal neurogenic region (Kaneko et al., 2006), which might mediate depression. Here, close examination revealed that chronic IFN-α treatment reduced the number of NSCs by nearly 40%, but not that of oligodendrocyte progenitors, another population that proliferates continuously in the adult brain (Figures 1 and S1M). Additionally, in vitro experiments showed that IFN-α significantly inhibited NSC proliferation, but did not affect their survival or neuronal differentiation, despite the presence of IFNAR on differentiated neurons, astrocytes, and oligodendrocytes (Figures 2, 3, and S2). Taken together, our findings indicate that NSCs in the adult brain may be a primary target of IFN-α. Indeed, the neurogenesis inhibition and depressive-behavior induction by chronic mIFN-α treatment were completely abrogated by CNS-specific and systemic IFNAR knockouts (Figure 5), suggesting that IFNAR in the brain mediates both of these effects of IFN-α.