T2D Research

Single-Cell PBMC Profiling in Type 2 Diabetes

Properties2
tagsconcept, type-2-diabetes, pbmc, transcriptomics, immunology
aliasesSingle-Cell PBMC Profiling in Type 2 Diabetes

3 min read

Single-Cell PBMC Profiling in Type 2 Diabetes

Single-cell PBMC profiling in type 2 diabetes can separate shifts in PBMC cell composition from changes in per-cell transcriptional state, immune-cell communication, and TCR/BCR repertoire structure.

Key Ideas

  • Gu et al. 2024 used scRNA-seq with paired TCR and BCR V(D)J profiling to study PBMC immune remodeling in T2D.
  • Gu et al. 2024 identified 11 major PBMC cell types, including CD4 T cells, CD8 T cells, NK cells, B cells, CD14 monocytes, CD16 monocytes, platelets, dendritic cells, erythrocytes, cycling cells, and HSPCs.
  • Gu et al. 2024 reported T2D-associated composition and cell-state changes, including altered monocyte subsets, cytotoxic T-cell phenotypes, and B-cell differentiation states.
  • For manuscript use, Gu et al. 2024 is a strong example of why bulk PBMC signals can mix cell-proportion effects with cell-state effects.
  • Li et al. 2025 used public PBMC scRNA-seq data from 37 T2D patients in GSE268210 and 11 healthy controls in GSE244515 to profile immunometabolic heterogeneity.
  • Li et al. 2025 extended composition and cell-state analysis by applying GSVA to 42 KEGG metabolic pathways in T-cell subpopulations and clustering samples into T2D subtypes A-C plus healthy-control group D.
  • Li et al. 2025 used CellChat, DoRothEA/VIPER, DGIdb, and machine-learning models to connect T-cell metabolic signatures with inferred cell communication, transcription-factor activity, and drug-enrichment hypotheses.
  • Zhao and Fang 2025 used 10x Genomics PBMC scRNA-seq on a small Chinese clinical-trial cohort of 3 T2DM patients and 3 healthy participants, retaining 13,591 cells and annotating B cells, T cells, monocytes, and NK cells.
  • Zhao and Fang 2025 analyzed T-cell and monocyte DEGs separately and linked T2DM PBMC signals to TNF/NF-kB, interferon-gamma response, T-cell receptor signaling, and chemokine signaling.
  • Zhao and Fang 2025 identified GSE255566 as the public sequence-data accession for the study.

Methods To Track

  • CellRanger produced expression matrices from 10x Genomics single-cell libraries.
  • Seurat was used for normalization, clustering, dimensionality reduction, marker detection, and module scores.
  • Harmony was used to integrate samples and reduce batch effects.
  • DESeq2 pseudobulk analysis tested pro-inflammatory gene expression across patients.
  • CellChat was used to infer cell-to-cell communication.
  • TCR and BCR analyses used productive paired chains and Gini coefficients for clonality.
  • GSVA can score metabolic pathway activity at single-cell resolution before sample-level consensus clustering.
  • DoRothEA/VIPER can infer transcription-factor activity from scRNA-seq expression data.
  • Machine-learning classification of cell-level metabolic signatures can be projected to patient-level subtype assignment by majority cell subtype.
  • Small-cohort PBMC scRNA-seq studies can still identify pathway hypotheses, but they require strict evidence weighting and external validation before being treated as replication-grade evidence.

Study Constraints

  • Gu et al. 2024 is not ancestry-stratified and should not be cited as direct evidence for ancestry-associated immune differences.
  • The cohort context still matters because a Korean single-country cohort may differ from cohorts used in multi-ancestry PBMC analyses.
  • Li et al. 2025 used T2D and healthy-control cells from different GEO accessions, making residual dataset effects a possible interpretation constraint even after Harmony integration.
  • Li et al. 2025 provides subtype and therapeutic-target hypotheses, but the subtype labels require independent validation before manuscript claims treat them as clinically stable categories.
  • Zhao and Fang 2025 has only 3 T2DM and 3 healthy participants from a single Chinese center, so its cell-proportion and DEG directions should be treated as exploratory.
  • Zhao and Fang 2025 does not test genetic ancestry or ancestry-stratified PBMC effects and should not be cited as direct ancestry evidence.

Sources

Graph View

Backlinks