Integrative Analysis of Membrane Proteome and MicroRNA Reveals Novel Lung Cancer Metastasis Biomarkers

A meta-learning approach to improving radiation response prediction in cancers

PURPOSE

Predicting the efficacy of radiotherapy in individual patients has drawn widespread attention, but the limited sample size remains a bottleneck for utilizing high-dimensional multi-omics data to guide personalized radiotherapy. We hypothesize the recently developed meta-learning framework could address this limitation.

METHODS AND MATERIALS

By combining gene expression, DNA methylation, and clinical data of 806 patients who had received radiotherapy from The Cancer Genome Atlas (TCGA), we applied the Model-Agnostic Meta-Learning (MAML) framework to tasks consisting of pan-cancer data, to obtain the best initial parameters of a neural network for a specific cancer with smaller number of samples. The performance of meta-learning framework was compared with four traditional machine learning methods based on two training schemes, and tested on Cancer Cell Line Encyclopedia (CCLE) and Chinese Glioma Genome Atlas (CGGA) datasets. Moreover, biological significance of the models was investigated by survival analysis and feature interpretation.

RESULTS

The mean AUC (Area under the ROC Curve) [95% confidence interval] of our models across nine cancer types was 0.702 [0.691-0.713], which improved by 0.166 on average over other the four machine learning methods on two training schemes. Our models performed significantly better (p < 0.05) in seven cancer types and performed comparable to the other predictors in the rest of two cancer types. The more pan-cancer samples were used to transfer meta-knowledge, the greater the performance improved (p < 0.05). The predicted response scores that our models generated were negatively correlated with cell radiosensitivity index in four cancer types (p < 0.05), while not statistically significant in the other three cancer types. Moreover, the predicted response scores were shown to be prognostic factors in seven cancer types and eight potential radiosensitivity-related genes were identified.

CONCLUSIONS

For the first time, we established the meta-learning approach to improving individual radiation response prediction by transferring common knowledge from pan-cancer data with MAML framework. The results demonstrated the superiority, generalizability, and biological significance of our approach.

The Fibrosis-Targeted Collagen/Integrins Gene Profile Predicts Risk of Metastasis in Pulmonary Neuroendocrine Neoplasms

Recently, collagen/integrin genes have shown promise as predictors of metastasis mainly in non-small cell lung cancer and breast cancer. However, it is unknown if these gene expression profiling differ in metastatic potential of pulmonary neuroendocrine neoplasms (PNENs). In this study, we sought to identify differentially expressed collagen/integrin genes in PNENs in order to understand the molecular mechanisms underlying the development of stroma-associated fibrosis for invasion and metastasis. We compared collagen/integrin gene expression profiling between PNE tumors (PNETs) and PNE carcinomas (PNECs) using a two-stage design. First, we used PCR Array System for 84 ECM-related genes, and among them, we found COL1A2, COL3A1, COL5A2, ITGA5, ITGAV, and ITGB1 functionally involved in the formation of the stroma-associated fibrosis among PNENs histological subtypes. Second, we examined the clinical association between the six collagen/integrin genes in tumor tissues from 24 patients with surgically excised PNENs. However, the pathological exam of their resected tissues demonstrated that 10 developed lymph node metastasis and 7 distant metastasis. We demonstrated and validated up regulation of the six fibrogenic genes in PNECs and down regulation in PNETs that were significantly associated with metastasis-free and overall survival (P<0.05). Our study implicates up regulation of fibrogenic genes as a critical molecular event leading to lymph node and distant metastasis in PNENs.

MicroRNA Interactome Multiomics Characterization for Cancer Research and Personalized Medicine: An Expert Review

MicroRNAs (miRNAs) that are mutually modulated by their interacting partners (interactome) are being increasingly noted for their significant role in pathogenesis and treatment of various human cancers. Recently, miRNA interactome dissected with multiomics approaches has been the subject of focus since individual tools or methods failed to provide the necessary comprehensive clues on the complete interactome. Even though single-omics technologies such as proteomics can uncover part of the interactome, the biological and clinical understanding still remain incomplete. In this study, we present an expert review of studies involving multiomics approaches to identification of miRNA interactome and its application in mechanistic characterization, classification, and therapeutic target identification in a variety of cancers, and with a focus on proteomics. We also discuss individual or multiple miRNA-based interactome identification in various pathological conditions of relevance to clinical medicine. Various new single-omics methods that can be integrated into multiomics cancer research and the computational approaches to analyze and predict miRNA interactome are also highlighted in this review. In all, we contextulize the power of multiomics approaches and the importance of the miRNA interactome to achieve the vision and practice of predictive, preventive, and personalized medicine in cancer research and clinical oncology.

Genes involved in the epithelial-mesenchymal transition in oral cancer: A systematic review

OBJECTIVE

Epithelial-mesenchymal transition (EMT) is considered the initial step in the invasion-metastasis cascade. The aim of this systematic review was to study the signature of genes involved in the EMT process in oral cancer (OC) confirmed by protein expression and its possible relationship with oral squamous cell carcinoma (OSCC) prognostic variables.

MATERIALS AND METHODS

A search of the scientific literature was carried out with no start date restriction until 17 September 2020 in the electronic databases Pubmed/MEDLINE, Web of Science, Cochrane Library and Scopus, following specific eligibility criteria. The methodological quality of the included studies was assessed using the Newcastle-Ottawa tool.

RESULTS

A total of 8 retrospective cohort studies were included, all of them performed in China and with low risk of bias. Overexpression of the genes HNRNPC, ITGA5, HMGA2 and SRSF3, and low expression of ALDH3A1 and ARID2 promote EMT in OC. The more advanced clinical stages of the TNM classification were significantly associated with overexpression of HNRNPC, ITGA5, HMGA2 and SRSF3, and low expression of ARID2.

CONCLUSIONS

HNRNPC, ITGA5, HMGA2, SRSF3, ALDH3A1 and ARID2 genes were associated with EMT process. Over- or under-expression of these genes is associated with worse stages of OSCC and/or worse prognosis of the tumor. Further studies on this topic are needed in different countries to be able to confirm these results, since the detection of these genes can help to know which tumors have a worse prognosis.