Recurrent spontaneous abortion (RSA) is defined as the loss of two or more consecutive pregnancies before the 20th week of gestation. More than half of RSA cases have no identified cause, making it a significant challenge in reproductive medicine. Growing evidence suggests that RSA is associated with an increased risk of certain cancers, including uterine cancer.
Uterine corpus endometrial carcinoma (UCEC), also known as endometrial cancer, originates from the inner lining of the uterus. Repeated endometrial damage from miscarriages, hormonal fluctuations following pregnancy loss, and shared immune dysregulation may link RSA to higher endometrial cancer risk, though the molecular mechanisms were largely unknown before this study.
This study sought to bridge the gap between RSA research and cancer biology by applying advanced genomic techniques to identify shared molecular players. The researchers analyzed gene expression from RSA tissue samples and then investigated those same genes across multiple cancer types in the TCGA (The Cancer Genome Atlas) database.
The central discovery was SLC8A1, a gene encoding a sodium-calcium exchanger protein (NCX1). This protein pumps calcium out of cells while bringing sodium in, regulating intracellular calcium levels. Abnormal calcium signaling has been linked to cancer cell proliferation, making SLC8A1 a compelling new target to investigate.
The researchers used a multi-step approach combining three types of genomic data from RSA patient samples. Bulk RNA sequencing (RNA-seq) measured overall gene expression levels across thousands of genes. Single-cell RNA sequencing (scRNA-seq) analyzed individual cell types within the decidua (the uterine lining during pregnancy). Weighted gene coexpression network analysis (WGCNA) identified groups of genes that are regulated together.
These three analyses identified 103 candidate genes that were differentially expressed in RSA. To narrow this down to the most important candidates, four machine learning algorithms were applied simultaneously: LASSO regression, random forest, support vector machine (SVM), and XGBoost. Only genes identified as important by all four methods were considered hub genes.
This multi-method convergence approach is powerful because different machine learning algorithms have different strengths and weaknesses. A gene that consistently emerges as important across diverse algorithms is much less likely to be a false positive than one identified by just a single method.
The single gene that emerged from all four machine learning analyses was SLC8A1. The researchers then conducted an extensive follow-up investigation of this gene across 33 cancer types in the TCGA database, with particular focus on endometrial cancer (UCEC).
Analysis of TCGA data showed that SLC8A1 is significantly lower in UCEC tumor tissue compared to normal endometrial tissue. An ROC curve analysis demonstrated that SLC8A1 expression could distinguish UCEC patients with an AUC of 0.871, indicating strong diagnostic potential.
Patients were divided into high-risk and low-risk groups based on a SLC8A1-based risk score calculated using Cox regression analysis. The high-risk group had significantly worse overall survival than the low-risk group. This survival difference was consistent across analysis at 1, 3, and 5 years, with AUC values consistently above 0.85.
A predictive nomogram combining the SLC8A1 risk score with clinical factors such as tumor grade and stage provided a visual tool for estimating individual patient prognosis. Calibration curves confirmed the model's predictions closely matched actual patient outcomes.
Mutation analysis showed that SLC8A1 has a 6% somatic mutation rate in UCEC, primarily through missense single nucleotide polymorphisms. This rate is relatively modest compared to major drivers like PTEN or TP53, suggesting SLC8A1's significance may lie more in its expression level than its mutational status.
High-risk UCEC patients (those with low SLC8A1) showed significantly elevated immune, stromal, and ESTIMATE scores, suggesting that tumors with low SLC8A1 have greater immune and stromal cell infiltration. This may reflect an immunosuppressive tumor microenvironment that supports cancer progression.
Detailed immune cell analysis revealed differences in CD8 T cells, regulatory T cells (Tregs), and multiple macrophage subtypes between high and low SLC8A1 expression groups. Higher proportions of immunosuppressive cell types in the high-risk group could explain why these patients respond poorly to conventional therapies.
SLC8A1 expression was significantly correlated with 29 immune checkpoint genes, with the strongest correlation involving CD40, a co-stimulatory receptor important for T cell activation. This connection suggests that SLC8A1 may influence whether the immune system can effectively recognize and attack tumor cells.
Drug sensitivity analysis identified several agents to which low-risk (high SLC8A1) patients showed greater sensitivity, including osimertinib, dasatinib, and ibrutinib. This information could help guide treatment selection in clinical practice, personalizing therapy based on a patient's SLC8A1 expression profile.
To test SLC8A1's biological function directly, the researchers used siRNA (short interfering RNA) to silence SLC8A1 expression in HTR-8/SVneo cells (human trophoblast cells used as a model for the uterine lining). Three different siRNA sequences effectively reduced SLC8A1 levels, with siRNA-SLC8A1-1 showing the strongest effect.
Silencing SLC8A1 using siRNA significantly reduced cell proliferation in CCK-8 assays and colony formation tests. This confirms that SLC8A1 promotes cell growth: when SLC8A1 is active, cells divide more readily; when it is silenced, proliferation slows. This is consistent with its role in supporting tumor survival.
Flow cytometry analysis showed that silencing SLC8A1 increased cell apoptosis (programmed cell death) significantly. Western blot analysis confirmed this: silencing SLC8A1 increased levels of pro-apoptotic proteins (caspase-3 and Bax) while reducing levels of survival-promoting proteins (Bcl-2 and PCNA).
Finally, silencing SLC8A1 caused a significant increase in intracellular calcium concentration. This makes biological sense given SLC8A1 normally pumps calcium out of cells. The calcium buildup likely triggers the apoptotic cascade, connecting SLC8A1's ion transport function to its role in cell survival and cancer biology.
The identification of SLC8A1 as a prognostic biomarker opens several potential clinical applications. A simple immunohistochemical test on biopsy tissue or an expression analysis on surgical specimens could risk-stratify endometrial cancer patients into high and low risk groups, guiding the intensity of surveillance and treatment.
The correlation with immune checkpoint gene expression suggests that SLC8A1 status could predict immunotherapy response. Patients with low SLC8A1 and associated immune checkpoint dysregulation might benefit most from checkpoint inhibitors like PD-1/PD-L1 blockade, while those with high SLC8A1 may respond better to other targeted agents.
The connection between RSA and UCEC through SLC8A1 raises an intriguing preventive possibility: women with recurrent miscarriage might benefit from enhanced endometrial cancer surveillance, particularly if they are found to have low SLC8A1 expression in their endometrial tissue.
As with most discovery-phase studies, these findings require validation in independent patient cohorts through prospective clinical trials. The study's reliance on publicly available databases and a single cell line model are limitations that future studies using primary human endometrial cancer samples and xenograft animal models should address.