Introduction
Type 2 diabetes (T2D) is a global metabolic disorder characterized by chronic hyperglycemia arising from insulin resistance and/or inadequate insulin secretion. Beyond its canonical metabolic features, T2D is increasingly recognized as an immunometabolic disorder: systemic inflammation and dysregulated immune cell function are integral to its pathophysiology, contribute to disease progression, and drive its vascular complications 1. Large-scale genetic studies have identified over six hundred T2D-associated loci 2, yet the mechanisms through which these genetic factors translate into altered immune function remain incompletely understood.
Peripheral blood mononuclear cells (PBMCs) offer an accessible window into the systemic immune state. Comprising T cells, B cells, NK cells, monocytes, and dendritic cells, the PBMC compartment captures both cell-composition shifts and cell-state changes associated with disease 1. Single-cell profiling of PBMCs in T2D has revealed altered monocyte inflammatory states, expanded cytotoxic T-cell clonotypes, B-cell differentiation and isotype-switching shifts, and immunometabolic heterogeneity across T-cell subtypes 1 3. Parallel bulk blood transcriptome meta-analyses confirm that T2D carries a reproducible pathway-level immune signal — enriched for neutrophil effector function, ER stress, and mTOR signaling — despite poor cross-study concordance at the individual-gene level 4.
A critical dimension missing from this literature is genetic ancestry. The pathophysiology of T2D differs systematically across populations: East Asian T2D is predominantly driven by beta-cell dysfunction with lower insulin resistance, whereas European T2D is more strongly associated with obesity and insulin resistance 2. This distinction extends to the genetic architecture — East Asian T2D loci disproportionately implicate beta-cell function (e.g., KCNQ1, UBE2E2, C2CD4A/B), while European loci span insulin resistance, lipodystrophy, and obesity pathways 2. Critically, these ancestry-specific mechanisms are not purely continental: recent work within a single country (the Russian Federation) has demonstrated that partitioned polygenic score (pPGS) profiles differ markedly between Chechen, Tatar, and Yakut populations — Yakuts showing a beta-cell-dominant genetic pattern resembling East Asians, while Chechens and Tatars show obesity- and insulin-resistance-dominant patterns 5.
Because distinct T2D mechanisms produce different metabolic milieus — differing in the degree of hyperinsulinemia, lipotoxicity, inflammatory cytokine profiles, and incretin axis engagement — they may also produce distinct systemic immune environments 2. To date, however, no published study has directly tested whether T2D-associated PBMC immune features vary by genetic ancestry, nor whether ancestry-specific T2D genetic mechanisms correlate with specific PBMC immune phenotypes within the same individuals. The studies that characterize PBMC immune remodeling in T2D are single-cohort and unstratified by ancestry 1 3, and cross-study blood transcriptomic analyses highlight substantial technical and population heterogeneity that confounds naive comparison 4.
Here we address this gap through a multi-modal, ancestry-aware analysis of PBMCs from individuals representing multiple genetic ancestry groups within a single country. Our approach integrates three complementary data modalities from the same participants — genotyping (for ancestry determination and partitioned polygenic scoring), single-cell RNA sequencing (for cell-type-resolved transcriptomic profiling), and single-cell ATAC sequencing (for chromatin accessibility mapping). By correlating ancestry-specific pPGS profiles derived from our prior work Markelova et al. 2025 with PBMC immune features measured in the same individuals, we ask: do T2D-associated PBMC immune changes carry a genetic signal that varies with ancestry, and which of those differences can be resolved through ancestry-aware analysis rather than reflecting cohort composition, disease severity, medication exposure, or technical confounders?