Significant heterogeneity exists in hormone receptor (HR)-positive/HER2-positive (HR⁺/HER2⁺) breast cancer, contributing to suboptimal pathological complete response rates with conventional neoadjuvant treatment regimens. Overcoming this challenge requires precise molecular classification, which is pivotal for the development of targeted therapies. We conducted molecular typing on a cohort of 211 patients with HR⁺/HER2⁺ breast cancer and performed a comprehensive analysis of the efficacy of various neoadjuvant treatment regimens. Our findings revealed four distinct molecular subtypes, each exhibiting unique characteristics and therapeutic implications. The HER2-enriched subtype, marked by activation of the HER2 signaling and hypoxia-inducible factor 1 (HIF-1) pathway, may benefit from intensified anti-HER2-targeted therapy. Estrogen receptor (ER)-activated subtype demonstrated potential sensitivity to combined therapeutic strategies targeting both ER and HER2 pathways. Characterized by high immune cell infiltration, the immunomodulatory subtype showed sensitivity to HER2-targeted antibody–drug conjugates (ADCs) and promise for immune checkpoint therapy. The highly heterogeneous subtype requires a multifaceted therapeutic approach. Organoid susceptibility assays suggested phosphoinositide 3-kinase inhibitors may be a potential treatment option. These findings underscore the importance of molecular subtyping in HR⁺/HER2⁺ breast cancer, offering a framework for developing precise and personalized treatment strategies. By addressing the heterogeneity of the disease, these approaches have the potential to optimize therapeutic outcomes and improve patient care.

HER2 gene amplification generates an enormous number of HER2 transcripts, but the global effects on endogenous miRNA targets including HER family members in breast cancer are unexplored. We generated a HER2–3’UTR expressing vector to test the tumor-promoting properties in HER2 low expressing T47D and MCF7 cells. Through microarray analysis and real-time PCR analysis we identified genes that were regulated by HER2–3’UTR. Positive and negative manipulation of miRNA expression, response element mutational studies and transcript reporter assays were performed to explore the mechanism of competitive sequestration of miR125a/miRNA125b by HER2 3’UTR. To investigate if trastuzumab-induced upregulation of HER3 is also mediated through miRNA de-repression, we used the CRISPR/cas9 to mutate the endogenous HER2 mRNA in HER2 over-expressing Au565 cells. Finally, we looked at cohorts of breast cancer samples of our own and the TCGA to show if HER2 and HER3 mRNAs correlate with each other. The HER2 3’UTR pronouncedly promoted cell proliferation, colony formation, and breast tumor growth. High-throughput sequencing revealed a significant increase in HER3 mRNA and protein levels by the HER2 3’untranslated region (3’UTR). The HER2 3’UTR harboring a shared miR-125a/b response element induced miR-125a/b sequestration and thus resulted in HER3 mRNA derepression. Trastuzumab treatment upregulated HER3 via elevated HER2 mRNA expression, leading to trastuzumab resistance. Depletion of miR-125a/b enhanced the antitumor activity of trastuzumab. Microarray data from HER2-overexpressing primary breast cancer showed significant elevation of mRNAs for predicted miR-125a/b targets compared to non-targets. These results suggest that HER2 3’UTR-mediated HER3 upregulation is involved in breast cell transformation, increased tumor growth, and resistance to anti-HER2 therapy. The combinatorial targeting of HER3 mRNA or miR-125a/b may offer an effective tool for breast cancer therapy.

OBJECTIVE To establish a prediction model of coronary heart disease (CHD) in elderly patients with diabetes mellitus (DM) based on machine learning (ML) algorithms. METHODS Based on the Medical Big Data Research Centre of Chinese PLA General Hospital in Beijing, China, we identified a cohort of elderly inpatients (>= 60 years), including 10,533 patients with DM complicated with CHD and 12,634 patients with DM without CHD, from January 2008 to December 2017. We collected demographic characteristics and clinical data. After selecting the important features, we established five ML models, including extreme gradient boosting (XGBoost), random forest (RF), decision tree (DT), adaptive boosting (Adaboost) and logistic regression (LR). We compared the receiver operating characteristic curves, area under the curve (AUC) and other relevant parameters of different models and determined the optimal classification model. The model was then applied to 7447 elderly patients with DM admitted from January 2018 to December 2019 to further validate the performance of the model. RESULTS Fifteen features were selected and included in the ML model. The classification precision in the test set of the XG-Boost, RF, DT, Adaboost and LR models was 0.778, 0.789, 0.753, 0.750 and 0.689, respectively; and the AUCs of the subjects were 0.851, 0.845, 0.823, 0.833 and 0.731, respectively. Applying the XGBoost model with optimal performance to a newly recruited dataset for validation, the diagnostic sensitivity, specificity, precision, and AUC were 0.792, 0.808, 0.748 and 0.880, respectively. CONCLUSIONS The XGBoost model established in the present study had certain predictive value for elderly patients with DM complicated with CHD.