A Hypoxia-Associated Prognostic Gene Signature Risk Model and Prognosis Predictors in Gliomas

Integration of graph neural networks and transcriptomics analysis identify key pathways and gene signature for immunotherapy response and prognosis of skin melanoma

OBJECTIVE

The assessment of immunotherapy plays a pivotal role in the clinical management of skin melanoma. Graph neural networks (GNNs), alongside other deep learning algorithms and bioinformatics approaches, have demonstrated substantial promise in advancing cancer diagnosis and treatment strategies.

METHODS

GNNs models were developed to predict the response to immunotherapy and to pinpoint key pathways. Utilizing the genes from these key pathways, multi-omics bioinformatics methods were employed to refine the construction of a gene signature, termed responseScore, aimed at enhancing the precision of immunotherapy response predictions. Subsequently, responseScore was explored from the perspectives of prognosis, genetic variation, pathway enrichment, and the tumor microenvironment. Concurrently, the association among 13 genes contributing to responseScore and factors such as immunotherapy response, prognosis, and the tumor microenvironment was investigated. Among these genes, PSMB6 was subjected to an in-depth analysis of its biological effect through experimental approaches like transfection and co-culture.

RESULTS

In the finalized model utilizing GNNs, it has revealed an AUC of 0.854 within the training dataset and 0.824 within the testing set, pinpointing key pathways such as R-HSA-70,268. The indicator named as responseScore excelled in its predictive accuracy regarding immunotherapy response and patient prognosis. Investigations into genetic variation, pathway enrichment, tumor microenvironment disclosed a profound association between responseScore and the enhancement of immune cell infiltration and anti-tumor immunity. A negative correlation was observed between the expression of PSMB6 and immune genes, with elevated PSMB6 expression correlating with poor prognosis. ELISA detection after co-cultivation experiments revealed significant reductions in the levels of cytokines IL-6 and IL-1β in specimens from the PCDH-PSMB6 group.

CONCLUSION

The GNNs prediction model and the responseScore developed in this research effectively indicate the immunotherapy response and prognosis for patients with skin melanoma. Additionally, responseScore provides insights into the tumor microenvironment and the characteristics of tumor immunity of melanoma. Thirteen genes identified in this study show promise as potential tumor markers or therapeutic targets. Notably, PSMB6 emerges as a potential therapeutic target for skin melanoma, where its elevated expression exhibits an inhibitory effect on the tumor immunity.

Defining hypoxia in cancer: A landmark evaluation of hypoxia gene expression signatures

Tumor hypoxia drives metabolic shifts, cancer progression, and therapeutic resistance. Challenges in quantifying hypoxia have hindered the exploitation of this potential "Achilles' heel." While gene expression signatures have shown promise as surrogate measures of hypoxia, signature usage is heterogeneous and debated. Here, we present a systematic pan-cancer evaluation of 70 hypoxia signatures and 14 summary scores in 104 cell lines and 5,407 tumor samples using 472 million length-matched random gene signatures. Signature and score choice strongly influenced the prediction of hypoxia in vitro and in vivo. In cell lines, the Tardon signature was highly accurate in both bulk and single-cell data (94% accuracy, interquartile mean). In tumors, the Buffa and Ragnum signatures demonstrated superior performance, with Buffa/mean and Ragnum/interquartile mean emerging as the most promising for prospective clinical trials. This work delivers recommendations for experimental hypoxia detection and patient stratification for hypoxia-targeting therapies, alongside a generalizable framework for signature evaluation.

Chronic hypoxia for the adaptation of extracellular vesicle phenotype

Variations in oxygen level affect the phenotype of cells and extracellular vesicles (EVs). Depending on the metabolic oxygen demand of cells, hypoxic cell culture can produce conditions more like those found in vivo, and with appropriate oxygen levels, mimic hypoxic tumours. However, most previous experiments studying both EVs and the effects of hypoxia on cells use periods of 72 h or less of hypoxia. We hypothesised that this was insufficient time for adaptation to hypoxic conditions both for EVs and cells which may skew the results of such studies. In this study, the effects of acute (72 h) and chronic hypoxia (> 2 weeks) on the phenotype of HepG2 and PC3 cells and their EVs were examined. Cells could be cultured normally under chronic hypoxic conditions and cryopreserved and recovered. The effects of hypoxia on EV phenotype are slow to establish and dependent on cell line. In PC3 cells, the greatest change in phenotype and increase in EV production occurred only with chronic hypoxic culture. In HepG2 cells, the number of EVs produced was insensitive to hypoxic culture and the greatest changes in protein expression were observed after acute hypoxic culture. Nonetheless, biphasic changes in EV phenotype were detected in both cell types in response to either acute or chronic hypoxia. These results indicate that for cells which do not induce consumptive oxygen depletion, prolonged hypoxic culture is required for complete adaptation.

Single-Cell Analyses Offer Insights into the Different Remodeling Programs of Arteries and Veins

Arteries and veins develop different types of occlusive diseases and respond differently to injury. The biological reasons for this discrepancy are not well understood, which is a limiting factor for the development of vein-targeted therapies. This study contrasts human peripheral arteries and veins at the single-cell level, with a focus on cell populations with remodeling potential. Upper arm arteries (brachial) and veins (basilic/cephalic) from 30 organ donors were compared using a combination of bulk and single-cell RNA sequencing, proteomics, flow cytometry, and histology. The cellular atlases of six arteries and veins demonstrated a 7.8× higher proportion of contractile smooth muscle cells (SMCs) in arteries and a trend toward more modulated SMCs. In contrast, veins showed a higher abundance of endothelial cells, pericytes, and macrophages, as well as an increasing trend in fibroblasts. Activated fibroblasts had similar proportions in both types of vessels but with significant differences in gene expression. Modulated SMCs and activated fibroblasts were characterized by the upregulation of MYH10, FN1, COL8A1, and ITGA10. Activated fibroblasts also expressed F2R, POSTN, and COMP and were confirmed by F2R/CD90 flow cytometry. Activated fibroblasts from veins were the top producers of collagens among all fibroblast populations from both types of vessels. Venous fibroblasts were also highly angiogenic, proinflammatory, and hyper-responders to reactive oxygen species. Differences in wall structure further explain the significant contribution of fibroblast populations to remodeling in veins. Fibroblasts are almost exclusively located outside the external elastic lamina in arteries, while widely distributed throughout the venous wall. In line with the above, ECM-targeted proteomics confirmed a higher abundance of fibrillar collagens in veins vs. more basement ECM components in arteries. The distinct cellular compositions and transcriptional programs of reparative populations in arteries and veins may explain differences in acute and chronic wall remodeling between vessels. This information may be relevant for the development of antistenotic therapies.

Knockdown of liver cancer cell-secreted exosomal PSMA5 controls macrophage polarization to restrain cancer progression by blocking JAK2/STAT3 signaling

INTRODUCTION

Tumor-associated macrophages, a major component of the tumor microenvironment, undergo polarization into M2 macrophages (M2), and thereby exert an immunosuppressive effect to induce cancer metastasis. This study strives to uncover a molecular mechanism underlying this event in hepatocellular carcinoma (HCC).

METHODS

Proteasome subunit alpha 5 (PSMA5) expression in liver hepatocellular carcinoma (LIHC) tissues and its association with LIHC patients were predicted using StarBase. PSMA5 level in human HCC cells was manipulated via transfection. Exosomes were isolated from HCC cells, and internalized into macrophages which were cocultured with HCC cells. Exosome internalization was observed after fluorescence labeling. HCC cell migration and invasion were evaluated by wound healing and Transwell assays. Xenograft assay was performed to investigate the role of PSMA5 in in vitro tumorigenesis. M2 polarization was assessed by enzyme-linked immunosorbent assay, quantitative reverse transcription polymerase chain reaction, and immunohistochemistry. PSMA5 expression in exosomes and Janus Kinase 2 (JAK2)/signal transducer and activator of transcription 3 (STAT3) activation in macrophages and tumors were detected by Western blot analysis.

RESULTS

High PSMA5 expression was observed in LIHC tissues and associated with compromised survival of LIHC patients. PSMA5 knockdown inhibited HCC cell migration and invasion. PSMA5 knockdown in HCC cells downregulated PSMA5 level in exosomes from these HCC cells. HCC cell-isolated exosomes were successfully internalized into macrophages, and facilitated M2 polarization and JAK2/STAT3 pathway activation. HCC cell-secreted exosomal PSMA5 knockdown inhibited the exosome-induced effect on macrophages, and attenuated the promotion induced by exosome-treated macrophages on HCC cell migration/invasion and tumorigenesis along with in vivo M2 polarization and JAK2/STAT3 pathway activation.

CONCLUSION

HCC cell-secreted exosomal PSMA5 knockdown hinders M2 polarization to suppress cancer progression by restraining JAK2/STAT3 signaling.

The genetic algorithm-aided three-stage ensemble learning method identified a robust survival risk score in patients with glioma

Ensemble learning is a kind of machine learning method which can integrate multiple basic learners together and achieve higher accuracy. Recently, single machine learning methods have been established to predict survival for patients with cancer. However, it still lacked a robust ensemble learning model with high accuracy to pick out patients with high risks. To achieve this, we proposed a novel genetic algorithm-aided three-stage ensemble learning method (3S score) for survival prediction. During the process of constructing the 3S score, double training sets were used to avoid over-fitting; the gene-pairing method was applied to reduce batch effect; a genetic algorithm was employed to select the best basic learner combination. When used to predict the survival state of glioma patients, this model achieved the highest C-index (0.697) as well as area under the receiver operating characteristic curve (ROC-AUCs) (first year = 0.705, third year = 0.825 and fifth year = 0.839) in the combined test set (n = 1191), compared with 12 other baseline models. Furthermore, the 3S score can distinguish survival significantly in eight cohorts among the total of nine independent test cohorts (P < 0.05), achieving significant improvement of ROC-AUCs. Notably, ablation experiments demonstrated that the gene-pairing method, double training sets and genetic algorithm make sure the robustness and effectiveness of the 3S score. The performance exploration on pan-cancer showed that the 3S score has excellent ability on survival prediction in five kinds of cancers, which was verified by Cox regression, survival curves and ROC curves together. To enable its clinical adoption, we implemented the 3S score and other two clinical factors as an easy-to-use web tool for risk scoring and therapy stratification in glioma patients.

A novel hypoxia-driven gene signature that can predict the prognosis and drug resistance of gliomas

Hypoxia spontaneously forms in the interior of glioma tissues and regulates the expression of various genes. However, the status of hypoxia-driven genes in glioma tissues is not completely known. In the current study, RNA-seq data of 695 glioma tissues in The Cancer Genome Atlas (TCGA) were set as a discovery cohort and were used to identify hypoxia-driven genes and construct a novel gene signature. The prognostic values of that signature were verified in data from the TCGA and the Chinese Glioma Genome Atlas (CGGA). The expression and diagnostic values of hypoxia-driven genes were analyzed using immunohistochemistry and receiver operator characteristic curves. Finally, the effects of hypoxia-driven genes on temozolomide (TMZ) resistance were analyzed by western blot, CCK-8 and colony formation assay. A total of 169 hypoxia-driven genes were identified, which were associated with a poor outcome in glioma patients. Among them, 22 genes had a degree score ≥10 and 6 genes (WT1, HOXA2, HOXC6, MMP9, SHOX2 and MYOD1) were selected to construct a signature to classify glioma patients into low- or high-risk groups. That signature had a remarkable prognostic value for glioma patients in TCGA and CGGA. The expression of HOXC6, MMP9, SHOX2 and MYOD1 was associated with hypoxia degree in glioma tissues and in recurrent cases, had a remarkable diagnostic value and a significant relationship with disease free survival in glioma patients. Moreover, SHOX2 was highly expressed in glioma tissues with O-6-methylguanine-DNA methyltransferase (MGMT)-unmethylation and temozolomide (TMZ) resistant glioma cell lines, and associated with MGMT expression. Knockdown the expression of SHOX2 significantly reduced the TMZ-resistance induced by hypoxia in glioma cells. Ultimately, we identified six novel hypoxia-driven genes for reliable prognostic prediction in gliomas and found that SHOX2 might be a potential target to overcome the TMZ resistance induced by hypoxia.

Construction and validation of an angiogenesis-related gene expression signature associated with clinical outcome and tumor immune microenvironment in glioma

Background: Glioma is the most prevalent malignant intracranial tumor. Many studies have shown that angiogenesis plays a crucial role in glioma tumorigenesis, metastasis, and prognosis. In this study, we conducted a comprehensive analysis of angiogenesis-related genes (ARGs) in glioma.

Methods: RNA-sequencing data of glioma patients were obtained from TCGA and CGGA databases. Via consensus clustering analysis, ARGs in the sequencing data were distinctly classified into two subgroups. We performed univariate Cox regression analysis to determine prognostic differentially expressed ARGs and least absolute shrinkage and selection operator Cox regression to construct a 14-ARG risk signature. The CIBERSORT algorithm was used to explore immune cell infiltration, and the ESTIMATE algorithm was applied to calculate immune and stromal scores.

Results: We found that the 14-ARG signature reflected the infiltration characteristics of different immune cells in the tumor immune microenvironment. Additionally, total tumor mutational burden increased significantly in the high-risk group. We combined the 14-ARG signature with patient clinicopathological data to construct a nomogram for predicting 1-, 3-, and 5-year overall survival with good accuracy. The predictive value of the prognostic model was verified in the CGGA cohort. SPP1 was a potential biomarker of glioma risk and was involved in the proliferation, invasion, and angiogenesis of glioma cells.

Conclusion: In conclusion, we established and validated a novel ARG risk signature that independently predicted the clinical outcomes of glioma patients and was associated with the tumor immune microenvironment.