Depression has been associated with inflammation, and inflammation may both influence and interact with growth factors such as brainderived neurotrophic factor (BDNF). Both the functional Val66Met BDNF polymorphism (rs6265) and BDNF levels have been associated with depression. It is thus plausible that decreased BDNF could mediate and/or moderate cytokine-induced depression. We therefore prospectively employed the Beck Depression Inventory-II (BDI-II), the Hospital Anxiety and Depression Scale (HADS), and the Montgomery-Asberg Depression Rating Scale (MADRS) in 124 initially euthymic patients during treatment with interferon-alpha (IFN-a), assessing serum BDNF and rs6265. Using mixed-effect repeated measures, lower pretreatment BDNF was associated with higher depression symptoms during IFN-a treatment (F 144,17.2 ¼ 6.8; Po0.0001). However, although the Met allele was associated with lower BDNF levels (F 1,83.0 ¼ 5.0; P ¼ 0.03), it was only associated with increased MADRS scores (F 4,8.9 ¼ 20.3; Po0.001), and not the BDI-II or HADS. An exploratory comparison of individual BDI-II items indicated that the Met allele was associated with suicidal ideation, sadness, and worthlessness, but not neurovegetative symptoms. Conversely, the serotonin transporter promoter polymorphism (5-HTTLPR) short allele was associated with neurovegetative symptoms such as insomnia, poor appetite and fatigue, but not sadness, worthlessness, or suicidal ideation. IFN-a therapy further lowered BDNF serum levels (F 4,37.7 ¼ 5.0; P ¼ 0.003), but this decrease occurred regardless of depression development. The findings thus do not support the hypothesis that decreasing BDNF is the primary pathway by which IFN-a worsens depression. Nonetheless, the results support the hypothesis that BDNF levels influence resiliency against developing inflammatory cytokine-associated depression, and specifically to a subset of symptoms distinct from those influenced by 5-HTTLPR. Neuropsychopharmacology (2013) 38, 985-995;