Abstract B016: Recapitulation of gene expression-based immune subtypes of breast cancer using immunohistochemical staining
Mustapha Abubakar,Hela Koka,Priscilla Lee,Scott Lawrence,Karun Mutreja,Difei Wang,Bin Zhu,Shelly Tse,Xiaohong R Yang
DOI: https://doi.org/10.1158/1538-7445.advbc23-b016
IF: 11.2
2024-02-02
Cancer Research
Abstract:Using gene expression profiling data, we previously identified three immune subtypes (low-TIL, high-TIL, high-ISG) of breast cancer (BC) among Chinese BC patients in Hong Kong (HK) (Zhu et al., Breast Cancer Res, 2017). To investigate if standard in situ immunohistochemical (IHC) staining could capture the BC immune landscape, we analyzed archival tissue sections from 148 HKBC patients who were previously characterized based on gene expression. Dual IHC staining for 8 immune markers (CD3, CD8, CD20, FOXP3, CD68, CD163, PDL1, and IDO1) and PanCK was performed. Digitized whole slide images for each marker were analyzed using high-accuracy machine learning algorithms trained to quantify the percentage of positively stained regions. Strong correlations were found between IHC immune markers and gene expression-based immune and intrinsic subtypes. Compared to patients with lower level of tumor-infiltrating lymphocytes (low-TIL) classified based on gene expression data, those in the high-TIL subtype (characterized by high T cell and immune checkpoint gene expression) showed higher proportions of CD3, CD8, CD20, FOXP3, CD68, and PDL1 IHC positively stained cells. Similarly, the high-ISG subtype (with increased interferon-stimulated gene expression and TP53 somatic mutations) had higher proportions of CD3, CD8, CD20, FOXP3, and CD68 stained cells than the low-TIL subtype and in particular exhibited high CD163 staining, indicative of a proinflammatory tumor microenvironment. CD8, CD20, CD163, PDL1, and IDO1 expression varied significantly by PAM50 subtypes, with luminal (A and B) subtypes showing lower staining than non-luminal (HER2 and Basal) subtypes. A distinctive feature, however, of luminal B tumors was the higher infiltration by CD163-expressing cells when compared with the luminal A subtype. Except for FOXP3, all examined IHC markers showed significantly higher expression in patients with somatic TP53 mutations (CD3: P= 0.003, CD8: P= 0.005, CD20: P= 0.003, CD68: P= 0.002, CD163: P= 0.002, PDL1: P= 0.003, and IDO1: P= 0.013). On the other hand, expression of these markers did not significantly vary by PIK3CA mutation or germline APOBEC3A/B deletion polymorphism, which is more prevalent among East Asians than other populations and was previously linked to BC risk and immune environment. Of the BC risk factors assessed, only body mass index (BMI) demonstrated associations with immune cell densities, with higher BMI associated with lower expression of CD3, CD8, and CD20. This study sheds further light on the in situ immune landscape of gene expression-based BC subtypes, suggesting that standard IHC assessment of immune markers could refine BC subtype classification for clinical and epidemiological use. Citation Format: Mustapha Abubakar, Hela Koka, Priscilla Lee, Scott Lawrence, Karun Mutreja, Difei Wang, Bin Zhu, Shelly Tse, Xiaohong R Yang. Recapitulation of gene expression-based immune subtypes of breast cancer using immunohistochemical staining [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Advances in Breast Cancer Research; 2023 Oct 19-22; San Diego, California. Philadelphia (PA): AACR; Cancer Res 2024;84(3 Suppl_1) nr B016.
oncology