CPT2-Linked Fatty Acid Oxidation and Autophagy Nodes Connect NMN-Responsive Transcriptional Programs to Schizophrenia Biology

Background Schizophrenia is a highly heritable psychiatric disorder with substantial clinical heterogeneity, and treatment-resistant schizophrenia remains difficult to define biologically when clozapine exposure is used as a proxy. Nicotinamide mononucleotide, an NAD⁺ precursor, has been shown to mitigate age-associated physiological decline in mice, and a recent re-analysis identified 35 robust NMN-rescued genes with tissue-dependent effects. We asked whether these NMN-responsive transcriptional programs intersect with schizophrenia and clozapine-proxied treatment-resistant schizophrenia genetic liability. Methods Robust NMN-rescued genes from the Cheung re-analysis of the Mills et al. NMN dataset were mapped to 10 KEGG pathways, including fatty acid degradation, endocytosis, MAPK signaling, Ras signaling, PPAR signaling, autophagy, regulation of actin cytoskeleton, citrate cycle-related metabolism, neuroactive ligand-receptor interaction, and synaptic vesicle cycle. Gene-level schizophrenia and clozapine-proxied treatment-resistant schizophrenia TWAS profiles were evaluated using Stouffer Z enrichment, Wilcoxon tests, permutation testing, bootstrap confidence intervals, leave-one-out influence analysis, and profile-concordance metrics. Results Across all 10 NMN-derived pathways, schizophrenia and clozapine-proxied treatment-resistant schizophrenia showed strong directional concordance, with Pearson correlations ranging from 0.75 to 0.88 and high sign concordance and Lin’s concordance correlation coefficients. No pathway showed significant pathway-level separation between schizophrenia and treatment-resistant schizophrenia. Fatty acid degradation was the strongest shared signal, with the highest Stouffer Z in schizophrenia and treatment-resistant schizophrenia and a CPT2-centered gene-level profile. Autophagy showed a different pattern: although pathway-level effects did not define treatment resistance, leave-one-out analysis identified marked treatment-resistant schizophrenia influence from ATG13, RPTOR, TAX1BP1, CALCOCO2, PIK3R4, MAP1LC3A, AKT3, SNAP29, RHEB, AMBRA1, and GABARAP. MAPK3 emerged as the most recurrent high-influence driver across multiple pathways. Conclusions NMN-responsive transcriptional programs appear to overlap with core schizophrenia biology through shared metabolic and cellular homeostasis pathways. The strongest translational leads are CPT2-linked fatty acid oxidation and gene-level autophagy or mTOR-related nodes in treatment-resistant schizophrenia. These findings support future work on NAD⁺-linked metabolic stratification, autophagy burden, and adjunctive interventions that move beyond receptor-centered antipsychotic models.