Immune infiltration correlates with transcriptomic subtypes in primary estrogen receptor positive invasive lobular breast cancer

Abstract

Understanding interplay of breast cancer and microenvironment is critical. Here we identified two transcriptomic subtypes and five immune infiltration patterns from RNA-seq and multiplex immunohistochemistry from 21 ER + /HER2- ILCs. We found proliferative subtype associated with increased suppressive immune infiltration, and defined a signature associated with lower proliferative, pro-inflammatory TAM infiltration, and improved survival in ER+ breast cancer. Our work identified genes related to ILC immune microenvironment and prognosis.

Fig. 1. Study design and ILC subtype identification from RNA-seq.

A Study design and workflow. Treatment-naive ER + /HER2- primary ILC tumors were processed as FFPE blocks, from which FFPE sections underwent RNA-seq in tumor-rich regions selected by pathologist (N = 21), and multispectral IHC (mIHC) with immunophenotyping panel (N = 13). For mIHC, immune cell infiltration was quantified in regions of interest (ROIs) in both tumor and stromal area. B Expression of ILC classifier genes in the ER + /HER2- ILC cohort, showing the two ILC subtypes (proliferative in orange box, non-proliferative (immune + reactive) in blue box), TCGA ILC subtypes, PAM50 subtypes, and other demographic and clinical information. C Signature score (ssGSEA score) of reference ILC subtype pathways from RATHER study. Only pathways with significant signature score difference between non-proliferative and proliferative ILC subtypes were shown (Mann-Whitney-Wilcoxon test). D Estimated tumor purity, stromal score, and immune score from PUREE and ESTIMATE algorithm. E PAM50 subtype composition of non-proliferative and proliferative ILC tumors. F Estimated tumor purity in each PAM50 subtype.

Fig. 2. Immune cell infiltration landscape from mIHC.

A Immune cell infiltration in tumor and stromal regions among ROIs (log₂(number/mm²)), annotated with ILC cluster, PAM50 subtypes, and demographic or clinical information. B Histogram of immune cell type abundance (number/mm²) in stromal and tumor region ROIs. C Spearman correlation (FDR < 0.05) of immune cell infiltration (number/mm²) between cell types. D NMF consensus matrix from immune cell infiltration in both stromal and tumor regions among ROIs, showing 5 immune infiltration patterns. E Stromal region immune cell infiltration (number/mm²) among immune infiltration patterns, showing each pattern per panel. F Stromal region immune cell infiltration (number/mm²) among immune infiltration patterns, showing each immune cell type per panel. ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05, comparison (horizontal brackets) with no label, p ≥ 0.05. G, H Immune cell infiltration (number/mm²) in proliferative and non-proliferative ILC tumors in stromal (G) and tumor regions (H). I Number of ROIs in each immune infiltration pattern, stratified by ILC subtypes.

Fig. 3. Derivation and clinical associations of TAM-Low signature.

A Number of gene-cell pairs with significant (FDR < 0.05) positive or negative Spearman’s correlation between gene expression (TPM) and immune cell infiltration (median value among ROIs in number/mm²). B Pathway enrichment of TAM-Low signature genes (483 genes). C Spearman correlation (FDR < 0.05) between TAM-Low and other TAM subtype signatures in METABRIC ER+ tumors (N = 1,505). D, E Hazard ratio of relapse free survival (D) and overall survival (E) of TMA-Neg signature (deconvolution score) and demographic-clinical variables in METABRIC ER+ tumors (N = 1,096, excluding entries with missing data in any of the regression covariates) from multivariant Cox regression. F-H. Deconvolution score of TAM-Low signature and TAM subtype signatures in baseline (F), post-aromatase-inhibitor (AI) treatment tumors (G), and change after treatment (Δ(Post-AI – baseline)) (H) from POETIC trial cohort of 131 patients. IFN-TAMs, interferon-primed TAMs; Reg-TAMs, immune regulatory TAMs, Inflam-TAMs, inflammatory cytokine-enriched TAMs; Angio-TAMs, proangiogenic TAMs; LA-TAMs, lipid-associated TAMs; Prolif-TAMs, proliferating TAMs.