Altered peripheral immune profiles in treatment-resistant depression: response to ketamine and prediction of treatment outcome

Major depressive disorder (MDD) is heterogeneous and a substantial subgroup of patients do not respond to standard monoaminergic antidepressants; these cases are commonly labelled treatment-resistant depression (TRD). Prior research has repeatedly identified alterations in peripheral inflammatory markers in MDD and some work suggests that inflammation may be particularly relevant in TRD, yet most studies have measured only a small number of cytokines or chemokines. In parallel, glutamatergic dysfunction has been implicated in depression and the N-methyl-D-aspartate receptor antagonist ketamine has emerged as a rapidly acting antidepressant with reported anti-inflammatory properties in preclinical and clinical studies, although findings on ketamine’s effects on peripheral cytokines in depressed humans have been mixed. Kiraly and colleagues set out to characterise a broad peripheral immune profile in medication-free adults with TRD compared with matched healthy controls, and to examine how these markers change acutely after a single intravenous ketamine infusion (0.5 mg kg-1). Their a priori hypothesis focused on canonical pro-inflammatory cytokines (IL-6, IL-1α/β and TNF-α) being elevated in TRD, with additional hypothesis-generating exploratory analyses across a wider panel of cytokines, chemokines and growth factors. The study also sought baseline predictors of clinical response to ketamine, using change in depressive severity at 24 h as the primary clinical outcome.

This was an observational study of peripheral blood markers embedded within concurrent clinical ketamine trials at Mount Sinai. Adults with TRD and healthy control (HC) volunteers were recruited; inclusion for TRD required a primary diagnosis of MDD by Structured Clinical Interview for DSM-IV, a current moderate-or-greater depressive episode, and non-response to at least two adequate antidepressant trials per the Antidepressant Treatment History Form. All participants were free of antidepressant medications for at least 2 weeks before participation and none had recent fluoxetine use. Exclusion criteria for TRD included lifetime psychotic or bipolar disorder, alcohol or substance abuse in the past 6 months, unstable medical illness, active nicotine use and active infections. Healthy controls had no lifetime psychiatric disorder or significant medical illness. TRD participants were concurrently enrolled in registered ketamine trials (ClinicalTrials.gov IDs provided in the text). Sample size planning used G*Power to detect a moderate effect with 80% power at α = 0.05. Blood sampling and ketamine administration: baseline samples for both groups were obtained between 08:00 and 09:00 after a period described in the extracted text as overnight fasting (48 h), a detail that appears inconsistent in the extraction. TRD subjects received a single intravenous ketamine infusion of 0.5 mg kg-1 in a clinical research unit and provided additional blood samples at +4 h and +24 h. The primary clinical endpoint was change in Montgomery-Åsberg Depression Rating Scale (MADRS) score from baseline to 24 h post-infusion. Laboratory assays and quality control: serum was assayed using a Millipore human cytokine/chemokine magnetic bead multiplex panel on a Luminex 200 system at the Human Immune Monitoring Core. The assay covered a broad set of analytes (the paper reports multiplex analysis of cytokines, chemokines and growth factors), with samples run across three plates and groups balanced across plates. Assays passed quality-control checks for standard-curve fit and coefficient of variation; IL-6 values were cross-checked by ELISA. Four samples (two HC, two TRD) with aberrant values across analytes were excluded. Statistical approach: analyte values below the limit of detection were imputed as half the lower limit. Thirteen analytes had more than 50% of values below detection and were therefore not analysable; remaining data were log-transformed for analysis. Baseline group comparisons used linear regression, conducted with and without covariates (age, sex, body mass index and race). Changes over time after ketamine were assessed with paired t-tests. Prediction of ketamine response was first tested with independent t-tests and then confirmed using logistic regression accounting for the same covariates. Correction for multiple comparisons was applied only to the a priori inflammatory cytokine analysis; other analyses reported uncorrected P-values unless stated otherwise. The extracted text indicates that covariate adjustment did not alter significance for reported findings.

The analysed cohort comprised 33 medication-free TRD patients and 26 healthy volunteers after exclusion of four aberrant samples. Baseline comparisons showed a robust elevation of interleukin-6 (IL-6) in TRD versus HC; the raw P-value reported was 0.004 and remained significant after Bonferroni correction (corrected P = 0.01). IL-1α, IL-1β and TNF-α were not significantly different between groups, though IL-1 measures trended upwards with high variability. Exploratory analyses across the broader panel identified four additional analytes that were significantly elevated in TRD: the chemokine MCP-1 (CCL2; P = 0.02) and three colony-stimulating/growth factors—G-CSF (P = 0.037), GM-CSF (P = 0.02) and PDGF-BB (P = 0.005). Brain-derived neurotrophic factor (BDNF) did not differ with treatment and did not predict ketamine response. Effects of ketamine on peripheral markers were generally transient. At 4 h post-infusion, modest decreases were seen in IL-6 and IL-1α (both P < 0.05). Among exploratory markers, G-CSF (P = 0.038), IL-13 (P = 0.038) and IP-10 (P < 0.0001) decreased at 4 h, but these returned to baseline by 24 h. At 24 h, three analytes showed modest changes relative to baseline: IL-7 increased (P < 0.0001), IL-8 decreased (P < 0.0001) and PDGF-AA decreased (P = 0.024). Importantly, none of the ketamine-induced peripheral cytokine changes reported correlated with clinical antidepressant response. Clinical response and biomarkers: the cohort’s 24-h response rate to a single ketamine infusion was reported as 60%. Baseline comparisons between eventual responders and non-responders identified significantly lower pretreatment fibroblast growth factor 2 (FGF-2) in responders (unadjusted P = 0.0001; P = 0.03 after covariate adjustment) and lower IL-1 receptor antagonist (IL-1ra) (unadjusted P = 0.0035; P = 0.033 adjusted). FGF-2 was highlighted for further evaluation: using a cutoff of <34 pg ml-1 the receiver-operating characteristic analysis yielded a likelihood ratio of 9.47, sensitivity 79% and specificity 92% (P = 0.0008). Baseline FGF-2 correlated with change in MADRS at 24 h (Pearson r = -0.565; P = 0.0009), indicating lower baseline FGF-2 associated with greater symptom improvement.

Kiraly and colleagues interpret their findings as evidence that a subset of medication-free TRD patients exhibit a distinct peripheral immune profile, characterised by elevated IL-6 and upregulation of chemokines, colony-stimulating factors and certain growth factors. They regard the IL-6 elevation as consistent with prior literature linking inflammation and depression, and note that the broader pattern—including increased MCP-1, G-CSF and GM-CSF—could reflect enhanced mobilisation and chemotaxis of monocytes from bone marrow, a hypothesis supported by preclinical work showing stress-related monocyte recruitment to the brain. The authors emphasise that their data are associative and do not establish mechanistic peripheral-to-central pathways. Regarding ketamine, the researchers observed transient reductions in some peripheral inflammatory markers at 4 h that returned to baseline by 24 h; these short-lived changes did not correlate with antidepressant response. From this they conclude that although ketamine has modest peripheral anti-inflammatory effects, their results do not support a direct anti-inflammatory mechanism mediating clinical improvement in this cohort. They also highlight FGF-2 as a promising peripheral predictor of rapid ketamine response: patients with lower baseline FGF-2 were more likely to respond, with an empirically derived cutoff showing high specificity and reasonable sensitivity. The authors acknowledge limitations and uncertainties. They note that peripheral cytokine changes do not necessarily mirror central nervous system inflammation and that cerebrospinal fluid or post-mortem studies would be needed to address central mechanisms. They also recognise that their findings in TRD cannot be assumed specific to treatment resistance versus treatment-responsive MDD without direct comparison groups. Sample size and variability limit definitive conclusions, and they call for larger studies in well-characterised populations to clarify the role of immune dysregulation in TRD and its relationship to ketamine response. Overall, the study is presented as adding to evidence for immune dysregulation in depression and suggesting candidate biomarkers and pathways for further translational research.