# Leveraging Diverse Cell-Death Patterns to Decipher the Interactive Relation of Unfavorable Outcome and Tumor Microenvironment in Breast Cancer

## TL;DR

This study explores how different cell death patterns in breast cancer interact with the tumor environment to affect patient outcomes.

## Contribution

A six-gene model is developed to predict breast cancer prognosis based on programmed cell death and tumor microenvironment interactions.

## Key findings

- High apoptosis and pyroptosis activity correlate with better prognosis due to enhanced anti-tumor immunity.
- A six-gene LASSO model predicts prognosis with significant accuracy in breast cancer patients.
- High-risk patients show lower immune infiltration and worse survival outcomes.

## Abstract

Background: Programmed cell death (PCD) dynamically influences breast cancer (BC) prognosis through interactions with the tumor microenvironment (TME). We investigated 13 PCD patterns to decipher their prognostic impact and mechanistic links to TME-driven outcomes. Our study aimed to explore the complex mechanisms underlying these interactions and establish a prognostic prediction model for breast cancer. Methods: Using TCGA and METABRIC datasets, we integrated single-sample gene set enrichment analysis (ssGSEA), weighted gene co-expression network analysis (WGCNA), and Least Absolute Shrinkage and Selection Operator (LASSO) to explore PCD-TME interactions. Multi-dimensional analyses included immune infiltration, genomic heterogeneity, and functional pathway enrichment. Results: Our results indicated that high apoptosis and pyroptosis activity, along with low autophagy, correlated with favorable prognosis, which was driven by enhanced anti-tumor immunity, including more M1 macrophage polarization and activated CD8+ T cells in TME. PCD-related genes could promote tumor metastasis and poor prognosis via VEGF/HIF-1/MAPK signaling and immune response, including Th1/Th2 cell differentiation, while new tumor event occurrences (metastasis/secondary cancers) were linked to specific clinical features and gene mutation spectrums, including TP53/CDH1 mutations and genomic instability. We constructed a six-gene LASSO model (BCAP31, BMF, GLUL, NFKBIA, PARP3, PROM2) to predict prognosis and identify high-risk BC patients (for five-year survival, AUC = 0.76 in TCGA; 0.74 in METABRIC). Therein, the high-risk subtype patients demonstrated a poorer prognosis, also characterized by lower microenvironment matrix and downregulated immunocyte infiltration. These six gene signatures also showed prognostic value with significant differential expression in gene and protein levels of BC samples. Conclusion: Our study provided a comprehensive landscape of the cancer survival difference and related PCD-TME interaction axis and highlighted that high-apoptosis/pyroptosis states caused favorable prognosis, underlying mechanisms closely related with the TME where anti-tumor immunity would be beneficial for patient prognosis. These findings highlighted the model’s potential for risk stratification in BC.