Decoding peripartum cardiomyopathy through multi-omics: circulating and cellular signatures from the Italian multicenter registry

Abstract

Peripartum cardiomyopathy (PPCM) is a rare but potentially life-threatening form of heart failure occurring toward the end of pregnancy or in the months following delivery. The aim of this study was to investigate PPCM biology through a systems-level approach integrating multiple molecular layers within the same individuals to identify disease-associated signatures and inform precision medicine. We conducted a national, multicenter Italian registry study. Forty women aged ≥18 years with confirmed PPCM diagnosis and 40 healthy women were prospectively enrolled. At inclusion, all participants underwent detailed clinical assessment and standardized echocardiography, together with peripheral blood sampling for multi-omics profiling. Whole blood was used for whole-exome sequencing (WES)-based burden analysis and RNA sequencing to capture genomic variation and transcriptomic activity. Peripheral blood mononuclear cells (PBMCs) were isolated for cellular proteomics, while plasma samples were analyzed for circulating proteomics. This integrated design enabled simultaneous evaluation of genomic, transcriptomic, intracellular proteomic, and circulating proteomic layers in each participant. To extract biologically meaningful patterns from these high-dimensional data, we applied Multi-Omics Factor Analysis, a framework extending principal component analysis to multi-view datasets by identifying latent factors representing shared or modality-specific variation. We present preliminary results from 67 samples and four data modalities: RNA-seq (5,000 features), WES burden metrics (574 features), PBMC proteomics (4,528 features), and plasma proteomics (562 features), totaling 10,664 features. Twelve latent factors were initially specified, and eight retained after convergence. Correlations between latent factors and 45 clinical and biological covariates revealed a dominant immune-related axis of variation. Factor 1 emerged as the principal multi-view latent factor, explaining 30.7% of transcriptomic variance and 16.6% of plasma proteomic variance, and was strongly associated with PPCM status. This factor captured a coordinated molecular program spanning intracellular transcripts, PBMC proteins, and circulating plasma proteins. Pathway enrichment analysis identified 67 consensus pathways, with focal adhesion as the most significant signal, and 29 of 39 immune cell types significantly associated with this factor. Lower Factor 1 values, predominantly observed in PPCM patients, were linked to heightened adaptive immune activation, particularly T- and B-cell signatures, whereas higher values characterized healthy controls. Overall, the study defines integrated molecular determinants of PPCM and demonstrates that systemic immune dysregulation is a central system-level feature of the disease, providing a framework for mechanistic investigation and future precision-medicine strategies.