Type I Interferon Emerges as the Reproducible Core of a Shared Mitochondrial-Immune Transcriptomic Signature in Pulmonary Arterial Hypertension and Bipolar Disorder.

Background/

Objectives

Mitochondrial dysfunction and immune activation are reported across many chronic diseases, yet whether such shared signals are reproducible across independent cohorts is rarely tested. We compared the peripheral blood transcriptomes of two clinically distinct disorders, pulmonary arterial hypertension (PAH) and bipolar disorder (BD), to determine which shared transcriptomic signals remain directionally concordant across independent cohorts.

Methods

Five public microarray cohorts (two PAH, three BD) were analyzed through differential expression, Hallmark gene set enrichment, and weighted gene co-expression network analysis. Instead of focusing on overlapping genes, we quantified the directional reproducibility of pathway-level signals using a stability index, evaluated against chance baselines, permutation testing, and complementary reproducibility metrics.

Results

Both disorders converged on an overlapping mitochondrial-immune transcriptional program, and the interferon-α response was the only Hallmark pathway with positive normalized enrichment score (NES) values across all five cohorts and was significantly enriched in four of them. Beyond the interferon response, the broader pathway signature was directionally concordant across the two PAH cohorts (stability index = 0.66; Spearman ρ = 0.61) but not across the three BD cohorts (0.22; ρ = 0.14).

Conclusions

PAH and BD share a peripheral mitochondrial-immune transcriptional program whose most reproducible element is type I interferon signaling, a signal directionally preserved across all cohorts despite differences in blood fraction. The broader transcriptomic program showed variable reproducibility that depended in part on the blood fraction sampled, highlighting the importance of evaluating both reproducibility and sample characteristics before inferring shared disease biology.