Delving into the Heterogeneity of Different Breast Cancer Subtypes and the Prognostic Models Utilizing scRNA-Seq and Bulk RNA-Seq

Abstract

Background: Breast cancer (BC) is the most common malignancy in women with high heterogeneity. The heterogeneity of cancer cells from different BC subtypes has not been thoroughly characterized and there is still no valid biomarker for predicting the prognosis of BC patients in clinical practice. Methods: Cancer cells were identified by calculating single cell copy number variation using the inferCNV algorithm. SCENIC was utilized to infer gene regulatory networks. CellPhoneDB software was used to analyze the intercellular communications in different cell types. Survival analysis, univariate Cox, least absolute shrinkage and selection operator (LASSO) regression and multivariate Cox analysis were used to construct subtype specific prognostic models. Results: Triple-negative breast cancer (TNBC) has a higher proportion of cancer cells than subtypes of HER2+ BC and luminal BC, and the specifically upregulated genes of the TNBC subtype are associated with antioxidant and chemical stress resistance. Key transcription factors (TFs) of tumor cells for three subtypes varied, and most of the TF-target genes are specifically upregulated in corresponding BC subtypes. The intercellular communications mediated by different receptor–ligand pairs lead to an inflammatory response with different degrees in the three BC subtypes. We establish a prognostic model containing 10 genes (risk genes: ATP6AP1, RNF139, BASP1, ESR1 and TSKU; protective genes: RPL31, PAK1, STARD10, TFPI2 and SIAH2) for luminal BC, seven genes (risk genes: ACTR6 and C2orf76; protective genes: DIO2, DCXR, NDUFA8, SULT1A2 and AQP3) for HER2+ BC, and seven genes (risk genes: HPGD, CDC42 and PGK1; protective genes: SMYD3, LMO4, FABP7 and PRKRA) for TNBC. Three prognostic models can distinguish high-risk patients from low-risk patients and accurately predict patient prognosis. Conclusions: Comparative analysis of the three BC subtypes based on cancer cell heterogeneity in this study will be of great clinical significance for the diagnosis, prognosis and targeted therapy for BC patients.

Benchmarking the translational potential of spatial gene expression prediction from histology

“…Subsequently, we binarised the risk model predictions into high and low-risk groups through using median prediction to split patients then constructed Kaplan-Meier (KM) survival curves. Following the analysis of Xu et al ( 29 ), we were able to capture the same distinct KM curves for the RNA-Seq models in each BC subset of patients ( p ≤ 5.1 × 10 −4 ) using the risk predictions of the patients the models were trained on ( Figure 4f-h ). However, we also showed that under cross-validation the distinction between high and low-risk groups were no longer significant except for Hist2ST ( p = 0.05) and ST-Net ( p = 6.9 × 10 −2 ) in the HER2+ subset ( Figure 4i ).…”

“…Out of the curated TCGA images, we split samples according to breast cancer (BC) subtypes as described in Xu et al 29 . The clinical information such as ER-, PR-, HER2- were defined according to the estrogen, progesterone and HER2 receptor status variables and these were used to define BC subsets.…”

“…For Figure 4a, the first principal component from a PCA using colour-related metrics was plotted against the patient-level correlation between TCGA RNA-Seq and predicted pseudobulk GE. Colourrelated metrics included: rms_contrast (the standard deviation of the pixel intensities across the pixels of interests), michelson_contrast (measurement of image contrast defined by luminance difference over average luminance), grayscale_brightness (mean pixel intensity of the image after converting the image to grayscale), chan1_brightness, chan2_brightness, chan3_brightness (mean pixel intensity of the image of the red, green and blue colour channels respectively), chan1_brightness_YUV, chan2_brightness_YUV and chan3_brightness_YUV (mean channel brightness of red, green and blue colour channels of image after converting to YUV colour space respectively) Clinical and/or translational impact: Out of the curated TCGA images, we split samples according to breast cancer (BC) subtypes as described in Xu et al 29 . The clinical information such as ER-, PR-, HER2-were defined according to the estrogen, progesterone and HER2 receptor status variables and these were used to define BC subsets.…”

Single-cell data analysis of malignant epithelial cell heterogeneity in lung adenocarcinoma for patient classification and prognosis prediction

“…To exclude the incorporation of normal epithelial cells into tumor samples due to sampling, we used a previously published method [ 16 ] to infer the exact malignant cells using copy number variation (CNV). According to previously published articles, epithelial cells with high CNV were considered as malignant cells [ [22] , [23] , [24] ]. In our result, almost all tumor tissues, most epithelial cells presented higher CNV compared with artificially inserted epithelial cells derived from normal lung tissues.…”

MCL1 Inhibition Overcomes the Aggressiveness Features of Triple-Negative Breast Cancer MDA-MB-231 Cells

“…Among all BC subtypes, luminal A and B account for 70%, HER2+ accounts for 15–20%, and TNBC accounts for 15–20%. Each BC subtype has specific molecular characteristics, clinical behavior, prognosis, and treatment modalities [ 3 ].…”