Improving ovarian cancer treatment decision using a novel risk predictive tool

Identification of epithelial-related artificial neural network prognostic models for the prediction of bladder cancer prognosis through comprehensive analysis of single-cell and bulk RNA sequencing

Background

Bladder cancer (BLCA) presents as a heterogeneous epithelial malignancy. Progress in the early detection and effective treatment of BLCA relies heavily on the identification of novel biomarkers. Therefore, the primary goal of this study is to pinpoint potential biomarkers for BLCA through the fusion of single-cell RNA sequencing and RNA sequencing assessments. Furthermore, the aim is to establish practical clinical prognostic models that can facilitate accurate categorization and individualized therapy for patients.

Methods

In this research, training sets were acquired from the TCGA database, whereas validation sets (GSE32894) and single-cell datasets (GSE135337) were extracted from the GEO database. Single-cell analysis was utilized to obtain characteristic subpopulations along with their associated marker genes. Subsequently, a novel BLCA subtype was identified within TCGA-BLCA. Furthermore, an artificial neural network prognostic model was constructed within the TCGA-BLCA cohort and subsequently verified utilizing a validation set. Two machine learning algorithms were employed to screen hub genes. QRT-qPCR was performed to detect the gene expression levels utilized in the construction of prognostic models across various cell lines. Additionally, the cMAP database and molecular docking were utilized for searching small molecule drugs.

Results

The results of single-cell analysis revealed the presence of epithelial cells in multiple subpopulations, with 1579 marker genes selected for subsequent investigations. Subsequently, four epithelial cell subtypes were identified within the TCGA-BLCA cohort. Notably, cluster A exhibited a significant survival advantage. Concurrently, an artificial neural network prognostic model comprising 17 feature genes was constructed, accurately stratifying patient risk. Patients categorized in the low-risk group demonstrated a considerable survival advantage. The ROC analysis suggested that the model has strong prognostic ability. Furthermore, the findings of the validation group align consistently with those from the training group. Two types of machine learning algorithms screened NFIC as hub genes. Forskolin, a small molecule drug that binds to NFIC, was identified by employing a cMAP database and molecular docking.

Conclusion

The analysis results supplement the research on the role of epithelial cells in BLCA. An artificial neural network prognostic model containing 17 characteristic genes demonstrates the capability to accurately stratify patient risk, thereby potentially improving clinical decision-making and optimizing personalized therapeutic approaches.

Integration of multiomics analyses reveals unique insights into CD24-mediated immunosuppressive tumor microenvironment of breast cancer

BACKGROUND

Tumor immunotherapy brings new light and vitality to breast cancer patients, but low response rate and limitations of therapeutic targets become major obstacles to its clinical application. Recent studies have shown that CD24 is involved in an important process of tumor immune regulation in breast cancer and is a promising target for immunotherapy.

METHODS

In this study, singleR was used to annotate each cell subpopulation after t-distributed stochastic neighbor embedding (t-SNE) methods. Pseudo-time trace analysis and cell communication were analyzed by Monocle2 package and CellChat, respectively. A prognostic model based on CD24-related genes was constructed using several machine learning methods. Multiple quantitative immunofluorescence (MQIF) was used to evaluate the spatial relationship between CD24+PANCK+cells and exhausted CD8+T cells.

RESULTS

Based on the scRNA-seq analysis, 1488 CD24-related differential genes were identified, and a risk model consisting of 15 prognostic characteristic genes was constructed by combining the bulk RNA-seq data. Patients were divided into high- and low-risk groups based on the median risk score. Immune landscape analysis showed that the low-risk group showed higher infiltration of immune-promoting cells and stronger immune reactivity. The results of cell communication demonstrated a strong interaction between CD24+epithelial cells and CD8+T cells. Subsequent MQIF demonstrated a strong interaction between CD24+PANCK+ and exhausted CD8+T cells with FOXP3+ in breast cancer. Additionally, CD24+PANCK+ and CD8+FOXP3+T cells were positively associated with lower survival rates.

CONCLUSION

This study highlights the importance of CD24+breast cancer cells in clinical prognosis and immunosuppressive microenvironment, which may provide a new direction for improving patient outcomes.

GGT5: a potential immunotherapy response inhibitor in gastric cancer by modulating GSH metabolism and sustaining memory CD8+ T cell infiltration

PURPOSE

The variable responses to immunotherapy observed in gastric cancer (GC) patients can be attributed to the intricate nature of the tumor microenvironment. Glutathione (GSH) metabolism significantly influences the initiation and progression of gastric cancer. Consequently, targeting GSH metabolism holds promise for improving the effectiveness of Immune checkpoints inhibitors (ICIs).

METHODS

We investigated 16 genes related to GSH metabolism, sourced from the MSigDB database, using pan-cancer datasets from TCGA. The most representative prognosis-related gene was identified for further analysis. ScRNA-sequencing analysis was used to explore the tumor heterogeneity of GC, and the results were confirmed by  Multiplex immunohistochemistry (mIHC).

RESULTS

Through DEGs, LASSO, univariate and multivariate Cox regression analyses, and survival analysis, we identified GGT5 as the hub gene in GSH metabolism with the potential to promote GC. Combining CIBERSORT, ssGSEA, and scRNA analysis, we constructed the immune architecture of GC. The subpopulations of T cells were isolated, revealing a strong association between GGT5 and memory CD8+ T cells. Furthermore, specimens from 10 GC patients receiving immunotherapy were collected. mIHC was used to assess the expression levels of GGT5 and memory CD8+ T cell markers. Our results established a positive correlation between GGT5 expression, the enrichment of memory CD8+ T cells, and a suboptimal response to immunotherapy.

CONCLUSIONS

Our study identifies GGT5, a hub gene in GSH metabolism, as a potential therapeutic target for inhibiting the response to immunotherapy in GC patients. These findings offer new insights into strategies for optimizing immunotherapy of GC.

Signatures of tumor-associated macrophages correlate with treatment response in ovarian cancer patients

Ovarian cancer (OC) ranks as the second leading cause of death among gynecological cancers. Numerous studies have indicated a correlation between the tumor microenvironment (TME) and the clinical response to treatment in OC patients. Tumor-associated macrophages (TAMs), a crucial component of the TME, exert influence on invasion, metastasis, and recurrence in OC patients. To delve deeper into the role of TAMs in OC, this study conducted an extensive analysis of single-cell data from OC patients. The aim is to develop a new risk score (RS) to characterize the response to treatment in OC patients to inform clinical treatment. We first identified TAM-associated genes (TAMGs) in OC patients and examined the protein and mRNA expression levels of TAMGs by Western blot and PCR experiments. Additionally, a scoring system for TAMGs was constructed, successfully categorizing patients into high and low RS subgroups. Remarkably, significant disparities were observed in immune cell infiltration and immunotherapy response between the high and low RS subgroups. The findings revealed that patients in the high RS group had a poorer prognosis but displayed greater sensitivity to immunotherapy. Another important finding was that patients in the high RS subgroup had a higher IC50 for chemotherapeutic agents. Furthermore, further experimental investigations led to the discovery that THEMIS2 could serve as a potential target in OC patients and is associated with EMT (epithelial-mesenchymal transition). Overall, the TAMGs-based scoring system holds promise for screening patients who would benefit from therapy and provides valuable information for the clinical treatment of OC.

Single-cell RNA sequencing integrated with bulk RNA sequencing analysis reveals diagnostic and prognostic signatures and immunoinfiltration in gastric cancer

BACKGROUND

Early diagnosis and prognostic predication of gastric cancer (GC) pose significant challenges in current clinical practice of GC treatments. Therefore, our aim was to explore relevant gene signatures that can predict the prognosis of GC patients.

METHODS

Here, we established a single-cell transcriptional atlas of GC, focusing on the expression of T-cell-related genes for cell-cell communication analysis, trajectory analysis, and transcription factor regulatory network analysis. Additionally, we conducted validation and prediction of immune-related prognostic gene signatures in GC patients using TCGA and GEO data. Based on these prognostic gene signatures, we predicted the immune infiltration status of GC patients by grouping the patient samples into high or low-risk groups.

RESULTS

Based on 10 tumor samples and corresponding normal samples from GC patients, we selected 18,416 cells for subsequent analysis using single-cell sequencing. From these, we identified 3,284 T-cells and obtained 641 differentially expressed genes related to T-cells from 5 different T-cell subtypes. By integrating bulk RNA sequencing data, we identified prognostic signatures associated with T-cells. Stratifying patients based on these prognostic signatures into high-risk or low-risk groups allowed us to effectively predict their survival rates and the immunoinfiltration status of the tumor microenvironment.

CONCLUSION

This study explored prognostic gene signatures associated with T-cells in GC patients, providing insights into predicting patients' survival rates and immunoinfiltration levels.

Comprehensive analysis of nicotinamide metabolism-related signature for predicting prognosis and immunotherapy response in breast cancer

Background

Breast cancer (BC) is the most common malignancy among women. Nicotinamide (NAM) metabolism regulates the development of multiple tumors. Herein, we sought to develop a NAM metabolism-related signature (NMRS) to make predictions of survival, tumor microenvironment (TME) and treatment efficacy in BC patients.

Methods

Transcriptional profiles and clinical data from The Cancer Genome Atlas (TCGA) were analyzed. NAM metabolism-related genes (NMRGs) were retrieved from the Molecular Signatures Database. Consensus clustering was performed on the NMRGs and the differentially expressed genes between different clusters were identified. Univariate Cox, Lasso, and multivariate Cox regression analyses were sequentially conducted to develop the NAM metabolism-related signature (NMRS), which was then validated in the International Cancer Genome Consortium (ICGC) database and Gene Expression Omnibus (GEO) single-cell RNA-seq data. Further studies, such as gene set enrichment analysis (GSEA), ESTIMATE, CIBERSORT, SubMap, and Immunophenoscore (IPS) algorithm, cancer-immunity cycle (CIC), tumor mutation burden (TMB), and drug sensitivity were performed to assess the TME and treatment response.

Results

We identified a 6-gene NMRS that was significantly associated with BC prognosis as an independent indicator. We performed risk stratification according to the NMRS and the low-risk group showed preferable clinical outcomes (P < 0.001). A comprehensive nomogram was developed and showed excellent predictive value for prognosis. GSEA demonstrated that the low-risk group was predominantly enriched in immune-associated pathways, whereas the high-risk group was enriched in cancer-related pathways. The ESTIMATE and CIBERSORT algorithms revealed that the low-risk group had a higher abundance of anti-tumor immunocyte infiltration (P < 0.05). Results of Submap, IPS, CIC, TMB, and external immunotherapy cohort (iMvigor210) analyses showed that the low-risk group were indicative of better immunotherapy response (P < 0.05).

Conclusions

The novel signature offers a promising way to evaluate the prognosis and treatment efficacy in BC patients, which may facilitate clinical practice and management.