Ketamine’s antidepressant effect is mediated by energy metabolism and antioxidant defense system

Mood disorders, particularly major depressive disorder (MDD), are highly prevalent and a leading cause of disability. Conventional antidepressants require weeks to take effect and achieve full remission in fewer than 50% of patients; around one third develop treatment-resistant depression (TRD). Ketamine, an N-methyl-D-aspartate receptor (NMDAR) antagonist, has been shown to produce a rapid antidepressant effect, including in TRD, but its psychotomimetic side-effects limit broad clinical use. Earlier mechanistic work has implicated mTORC1 activation, synaptogenesis and, more recently, active ketamine metabolites acting via AMPA receptors. The investigators' prior time-dependent metabolomics work suggested a role for mitochondrial energy metabolism pathways such as glycolysis and the citrate cycle in ketamine's effects. Weckmann and colleagues set out to extend those findings by profiling the hippocampal metabolome and proteome after a single low dose of ketamine in mice, correlating molecular changes with antidepressant-like behaviour in the forced swim test (FST). The study aimed to identify ketamine-affected pathways and biosignatures, with particular attention to mitochondrial oxidative phosphorylation (OXPHOS), cellular energy status (for example ATP/ADP ratio), and antioxidant defence systems that could mediate ketamine's rapid antidepressant-like action.

The study used adult male C57BL/6 mice (8 weeks old) treated with a single low dose of ketamine; the Discussion reports the dose as 3 mg kg-1. Behavioural assessment employed the forced swim test (FST), with floating time used as the readout for antidepressant-like effect. Hippocampal tissue was collected at multiple time points after ketamine administration, including 2 h, 14 h, 24 h and 72 h for some assays, and metabolomics and proteomics focused on the 2 h and 24 h time points. Hippocampal proteins were fractionated into cytoplasmic (CF) and membrane-associated (MF) fractions by differential homogenisation and extraction. Western blotting measured total and phosphorylated AMPK (pAMPK), and peroxiredoxins Prdx1 and Prdx3. Total antioxidant capacity (TAC) was assayed spectrophotometrically, and protein oxidation (carbonylation) was assessed using a commercial OxyBlot kit. Quantitative proteomics used metabolic labelling: hippocampal CF and MF proteins were mixed 1:1 with 14N/15N-labelled reference proteins derived from mice on a labelled diet. Fifty micrograms of protein mixtures were separated by SDS-PAGE, in-gel digested with trypsin, and peptides analysed by LC-MS/MS on a Q-Exactive Orbitrap. Peptide identification and quantification used an in-house pipeline (iSPY) with Mascot searches against the SwissProt Mouse database and Percolator to improve identifications. Only peptides meeting stringent error thresholds were retained. Polar metabolites were isolated and analysed as described in the authors' prior work; metabolite pair ratios calculated included ATP/ADP, NADH/NAD and GTP/GDP. For proteomic and metabolomic statistics, metabolite intensities and protein ratios were median-normalised and auto-scaled. Significant changes were identified by combining multivariate partial least squares-discriminant analysis (PLS-DA, VIP ≥1.0) with Significance Analysis of Microarrays (SAM, q ≤0.10 and FDR ≤0.10) to reduce false positives. Pathway enrichment used MetaboAnalyst (hypergeometric test and pathway topology) for metabolites and STRING for proteomic pathway analysis. Correlations between molecular measures and FST floating time were reported where available.

Metabolite ratios and energy status: The ATP/ADP ratio tended to decrease at both 2 h and 24 h after a single low dose of ketamine. Individual ATP and ADP levels showed strong, significant correlations with FST floating time at 24 h, consistent with a relation between hippocampal energy status and antidepressant-like behaviour. The NADH/NAD ratio did not change significantly and did not correlate with FST floating time. The GTP/GDP ratio was significantly elevated at 2 h and trended downward at 24 h; GTP concentrations significantly correlated with FST floating time at both time points. AMPK and western blot results: Total AMPK protein levels were unchanged, whereas phosphorylated AMPK (pAMPK) increased significantly at 24 h, indicating activation of the principal cellular energy sensor following ketamine. This finding aligns with the observed decrease in ATP/ADP ratio at 24 h. Proteomics profiling: Time-dependent quantitative proteomics was performed on hippocampal MF and CF fractions with five mice per group per time point. After processing and imputation, 889 MF and 1173 CF proteins were analysed at 2 h, and 929 MF and 1091 CF proteins at 24 h. Multivariate PLS-DA separated ketamine- and vehicle-treated mice at both time points, but model quality metrics (accuracy, R2, Q2) indicated robust models mainly at 2 h; the 24 h models were weaker and might include false positives. Combining PLS-DA VIP scores with SAM produced lists of significant proteins: 61 MF and 95 CF proteins altered at 2 h. Pathway enrichment of these proteins highlighted OXPHOS, purine metabolism, fructose and mannose metabolism, proteasome, protein processing in the endoplasmic reticulum and ribosome pathways. OXPHOS-related proteins showed notable changes at 2 h. Protein levels increased for complex I NADH dehydrogenase flavoprotein 2 (Ndufv2), cytochrome c1 (Cyc1) and a complex I assembly factor (Tmem126B), whereas Ndufaf7 (another complex I assembly factor) decreased. Levels of Ndufv2, Cyc1, Ndufaf7 and Tmem126B all correlated significantly with FST floating time, supporting a link between OXPHOS protein alterations and behavioural response. Oxidative stress and antioxidant defence: Given that elevated OXPHOS can increase reactive oxygen species (ROS), the investigators measured antioxidant proteins and oxidative damage. Cytoplasmic Prdx1 protein levels were lower at 2 h, and mitochondrial Prdx3 levels were lower at 72 h after ketamine. Total antioxidant capacity decreased at 72 h. Measurement of protein carbonylation showed a statistically significant reduction at 72 h, indicating less detectable protein oxidation in ketamine-treated mice at that time. The proteomics data also indicated upregulation of the proteasome pathway at 2 h, which the authors link to increased degradation of damaged proteins.

Weckmann and colleagues interpret their findings as supporting a model in which ketamine rapidly alters hippocampal energy metabolism and antioxidant defence systems, processes that may mediate its antidepressant-like action. The observed reduction in ATP/ADP ratio at 2 h and 24 h, together with increased pAMPK at 24 h, is taken as evidence of an initial energy deficit that follows ketamine-induced anabolic processes mediated by mTORC1; AMPK activation is proposed to follow to restore energy balance by promoting catabolic ATP-generating pathways. Proteomic and metabolite changes implicate the citrate cycle, glycolysis and OXPHOS, consistent with previous metabolomics work and with literature linking glutamatergic neurotransmission to increased glucose utilisation and mitochondrial metabolism. The authors emphasise OXPHOS involvement because several complex I-related proteins changed after ketamine and correlated with FST behaviour. They note that increased oxidative phosphorylation can elevate ROS, but their data show reduced protein carbonylation at 72 h alongside decreased peroxiredoxins and total antioxidant capacity, which they interpret as evidence for enhanced protein quality control and homeostatic redox adjustment; proteasome upregulation at 2 h may contribute to accelerated removal of damaged proteins. In placing their results in context, the investigators cite evidence of mitochondrial dysfunction in MDD and lower peripheral mitochondrial respiration correlating with depressive severity, proposing that ketamine's efficacy—especially in TRD—may relate to beneficial effects on mitochondrial energy metabolism in a subset of patients. They acknowledge limitations: behavioural assessment relied solely on FST floating time rather than multiple behavioural paradigms, and the 24 h proteomic models were weaker and might include false positives. The authors call for follow-up experiments to confirm causal roles, for example by blocking OXPHOS complex I or AMPK and testing different ketamine doses, to validate the mechanistic links suggested by their correlational data.