The immunomodulatory capacity of mental stress is one of the basic concepts of psychoneuroimmunology. The current prospective longitudinal study was designed to evaluate the effect of acute mental stress on neurotransmitter precursor amino acid levels in individuals with depression at 2 time points. Ten physically healthy patients with a diagnosis of major depressive episode and Montgomery–Åsberg Depression Rating Scale scores (MADRAS) ⩾20 points at inclusion were assessed on 2 study days (once with higher MADRAS scores, once with lower MADRAS scores; median 34.5 days apart) and subjected to a standardized acute mental stress test on each study day. Blood was collected at 4 time points: once prior to and at 3 time points (0, 30 minutes, 60 minutes) following mental stress. Neurotransmitter precursor amino acid levels, that is kynurenine/tryptophan (KYN/TRP) and phenylalanine/tyrosine (PHE/TYR), as well as neopterin and nitrite were analyzed in a total of 80 individual blood samples. Regression and correlation analyses were performed. Regression analyses of PHE/TYR ( R² = .547) and KYN/TRP ( R² = .440) in relation to MADRAS depression severity showed a quadratic curve fit. This was reflected by a negative linear correlation between MADRAS scores and PHE/TYR as well as KYN/TRP in the lower score range ( r = −.805, P < .001 and r = −.586, P < .001 respectively) and a positive correlation in the higher MADRAS score range ( r = .713, P < .001 and r = .379, P = .016 respectively). No effect of acute mental stress was found. This analysis exemplifies the implications of sampling as well as data distributions on results. The crosstalk of biological mechanisms that orchestrate metabolic and immunological signaling may vary depending on depression severity resulting in non-linear associations that may explain the heterogeneity of results found in the literature.