Risk prediction model construction for post myocardial infarction heart failure by blood immune B cells

Hourong Sun, Xiangjin Kong, Kaiming Wei, Jie Hao, Yue Xi, Lingwei Meng, GuanNan Li, Xin Lv, Xin Zou, Xinghua GuFrontiers in Immunology 2023 · DOI

Abstract

Sub-cluster B cells may play a significant role in post-myocardial infarction (MI) heart failure (HF). This study identifies a specific "immune-activated" B cell subtype as a critical early warning system for post-MI complications. We found that the STING1, HSPB1, CCL5, ACTN1, and ITGB2 genes in patients with post-MI HF showed a consistent trend of increase compared to those without post-MI HF, forming the basis for a highly accurate predictive model.

Introduction

Myocardial infarction remains a leading cause of global mortality. While immediate survival rates have improved due to better acute care, a significant portion of patients develop heart failure (HF) in the subsequent months. Predicting which individuals are at risk for "post-MI HF" is challenging; current biomarkers like natriuretic peptides often lack early predictive power or show inconsistency.

Traditional research has focused on damaged cardiac tissue, but obtaining heart biopsies is invasive and carries high risk. This study explores the peripheral blood as a "liquid biopsy," investigating how the systemic immune response—specifically B cells—can forecast heart failure before clinical symptoms manifest.

Methods

The researchers employed a multi-step bioinformatic pipeline to bridge the gap between single-cell resolution and clinical utility. The process integrated single-cell RNA sequencing (scRNA-seq) with bulk RNA-seq validation and machine learning.

Study methodology flowchart illustrating the multi-step bioinformatic pipeline from single-cell discovery to experimental validation.
  1. Single-Cell Discovery: Re-analysis of scRNA-seq data from peripheral blood mononuclear cells (PBMCs) (GSE145154) identified immune cell types with significant changes in post-MI HF patients.
  2. Sub-cluster Analysis: Using the scCODE R package, the team identified five distinct B cell sub-clusters. Sub-cluster "C4" emerged as the most predictive of disease progression.
  3. Signature Construction: Utilizing a cyclic algorithm and the Cancerclass framework, hundreds of marker genes were refined into a 13-gene signature (CCL5, B2M, CD3D, CD6, CD79B, DGKZ, STING1, LYAR, KLRG1, PRKCA, HSPB1, ACTN1, and MYL12A).
  4. Validation: The model was tested against independent bulk RNA-seq datasets (GSE59867, GSE60993, GSE61144) and confirmed in a human cohort (7 HF vs. 7 non-HF patients) using quantitative real-time PCR (RT-qPCR).

Key Results and Insights

B Cells as Sentinels

While monocytes and T cells are often the focus of cardiac inflammation research, this study highlights a specific subtype of "immune-activated" B cells (C4 subcluster) as the most reliable predictors of HF. Screening across multiple cohorts demonstrated that B cells and monocytes showed consistent pathway alterations associated with MI.

Heatmap of adjusted p-values showing significant Gene Ontology Biological Process (GOBP) pathway enrichment across different immune cell types in two independent cohorts.

Predictive Performance and Hub Genes

The 13-gene prediction model achieved an Area Under the Curve (AUC) of 1.0 in the GSE59867 validation cohort, indicating perfect separation between post-MI HF and non-HF patients in that dataset. Five key hub genes—STING1, HSPB1, CCL5, ACTN1, and ITGB2—were confirmed via RT-qPCR to be significantly upregulated in the blood of HF patients.

Notably, the upregulation of STING1 suggests that the cGAS-STING pathway in B cells may be a primary indicator of the maladaptive inflammation leading to heart failure. Functional enrichment analysis further suggested these B cell markers are involved in complex cross-talk with other immune players, including pathways related to graft rejection and NK cell cycles.

Discussion and Future Directions

While the 13-gene signature shows exceptional predictive promise, several limitations remain. The initial scRNA-seq analysis was based on a small sample size, necessitating larger, more diverse cohorts to ensure universal applicability. Furthermore, while a strong correlation exists, the mechanistic role of these B cells—whether they are active participants in pathology or passive markers—requires further investigation through in vivo models. Future clinical integration should aim to combine this transcriptomic signature with traditional clinical features like age and comorbidities to create a holistic risk-scoring tool.

Summary updated

Abstract

Sub-cluster B cells may play a significant role in post-MI HF. We found that the <i>STING1, HSPB1, CCL5, ACTN1</i>, and <i>ITGB2</i> genes in patients with post-MI HF showed the same trend of increase as those without post-MI HF.
BibTeX
@inproceedings{sun2023risk,
  author = {Hourong Sun and Xiangjin Kong and Kaiming Wei and Jie Hao and Yue Xi and Lingwei Meng and GuanNan Li and Xin Lv and Xin Zou and Xinghua Gu},
  title = {{Risk prediction model construction for post myocardial infarction heart failure by blood immune B cells}},
  booktitle = {Frontiers in Immunology},
  year = {2023},
  doi = {10.3389/fimmu.2023.1163350},
}
Figure