A preliminary study of microRNA expression in different types of primary melanoma

A deep learning approach based on multi-omics data integration to construct a risk stratification prediction model for skin cutaneous melanoma

PURPOSE

Skin cutaneous melanoma (SKCM) is a highly aggressive melanocytic carcinoma whose high heterogeneity and complex etiology make its prognosis difficult to predict. This study aimed to construct a risk subtype typing model for SKCM.

METHODS

The study proposes a deep learning framework combining early fusion feature autoencoder (AE) and late fusion feature AE for risk subtype prediction of SKCM. The deep learning framework integrates mRNA, miRNA, and DNA methylation data of SKCM patients from The Cancer Genome Atlas (TCGA), and clusters the screened multi-omics features associated with survival prognosis to identify risk subtypes. Differential expression analysis and functional enrichment analysis were performed between risk subtypes, while SVM classifiers were constructed between differentially expressed genes (DEGs) obtained by Least Absolute Shrinkage and Selection Operator (LASSO) logistic regression screening and risk subtype labels inferred from multi-omics data, and the predictive robustness of risk subtypes inferred from the risk subtype classification prediction model was validated using two independent datasets.

RESULTS

The deep learning framework that combined early fusion feature AE with late fusion feature AE distinguished the two best risk subtypes compared to the multi-omics integration approach with single strategy AE or PCA. A promising C-index (C-index = 0.748) and a significant difference in survival (log-rank P value = 4.61 × 10-9) were found between the identified risk subtypes. The DEGs with the top significance values together with differentially expressed miRNAs provided the biological interpretation of risk subtypes on SKCM. Finally, the framework was applied to predict risk subtypes in two independent test datasets of SKCM patients, all of which showed good predictive power (C-index > 0.680) and significant survival differences (log-rank P value < 0.01).

CONCLUSION

The SKCM risk subtypes identified by integrating multi-omics data based on deep learning can not only improve the understanding of the molecular mechanisms of SKCM, but also provide clinicians with assistance in treatment decisions.

Exosomes in the visual system: New avenues in ocular diseases

Exosomes are a subgroup of membrane-bound extracellular vesicles secreted by all cell types and present virtually in all biological fluids. The composition of exosomes in the same cell type varies in healthy and disease conditions. Hence, exosomes research is a prime focus area for clinical research in cancer and numerous age-related metabolic syndromes. Functions of exosomes include crucial cell-to-cell communication that mediates complex cellular processes, such as antigen presentation, stem cell differentiation, and angiogenesis. However, very few studies reported the presence and role of exosomes in normal physiological and pathological conditions of specialized ocular tissues of the eye and ocular cancers. The eye being a protected sense organ with unique connectivity with the rest of the body through the blood and natural passages, we believe that the role of exosomes in ocular tissues will significantly improve our understanding of ocular diseases and their interactions with the rest of the body. We present a review that highlights the existence and function of exosomes in various ocular tissues, their role in the progression of some of the neoplastic and non-neoplastic conditions of the eyes.

Investigating melanogenesis-related microRNAs as disease biomarkers in vitiligo

Vitiligo is considered a disabling disease that affects physical, social, psychological, and occupational aspects of an individual's quality of life. The search for non-invasive and reliable biomarkers for vitiligo's early diagnosis, prognosis, and treatment prediction is under intensive investigation. There is currently an emerging interest in employing miRNAs as biomarkers to predict vitiligo diagnosis and prognosis, inspired by the well-preserved nature of miRNAs in serum or plasma. In the current study, we assessed a panel of 20 melanogenesis pathway-related microRNAs (miRNAs) using quantitative real-time PCR technique in 85 non-segmental vitiligo (NSV) patients compared to 85 normal controls followed by function and pathway enrichment analysis for the miRNAs with significant results. Twelve out of the 20 circulating miRNAs showed significantly higher expression levels in vitiligo patients relative to controls where miR-423 show the highest expression level followed by miR-182, miR-106a, miR-23b, miR-9, miR-124, miR-130a, miR-203a, miR-181, miR-152, and miR-320a. While six miRNAs (miR-224, miR-148a, miR-137, and miR-7, miR-148b, miR-145, miR-374b, and miR-196b) didn’t show significant expression level. The analysis of the receiver operating curve indicated that miR-423, miR-106a, and miR-182 were outstanding biomarkers with the highest areas under the curve in vitiligo. This study is the first Egyptian study to investigate a panel of miRNAs expression profile in the plasma of patients with NSV. Our results suggest that specific circulating miRNAs signature might be implicated in vitiligo pathogenesis and could potentially be used as biomarkers in vitiligo.

Hsa-let-7c-5p, hsa-miR-130b-3p, and hsa-miR-142-3p as Novel miRNA Biomarkers for Melanoma Progression

This study aimed to screen miRNA biomarkers for melanoma progression. Raw melanoma data were downloaded from the Gene Expression Omnibus (GSE34460, GSE35579, GSE18509, and GSE24996) and the Cancer Genome Atlas (TCGA). Then, all differentially expressed miRNAs (DEmiRNAs) between benign vs. primary, metastatic vs. benign, and metastatic vs. primary groups were obtained in the GSE34460 and GSE35579 datasets, and the miRNAs related to disease progression were further screened. Then, the miRNA-gene network was constructed, followed by enrichment, survival, and cluster analyses. Differentially expressed genes (DEGs), tumor-infiltrating immune cells, and tumor mutation burden (TMB) between subtypes were analyzed. miRNAs were verified in the GSE18509 and GSE24996 datasets. A total of 132 and 209 DEmiRNAs were obtained in the GSE34460 and GSE35579 datasets, respectively, and 27 DEmiRNAs related to disease progression were screened. hsa-miR-106b-5p, hsa-miR-27b-3p, and hsa-miR-141-3p had a higher degree and were regulated by numerous genes in the miRNA-gene network. Moreover, four miRNAs were associated with prognosis: hsa-let-7c-5p, hsa-miR-130b-3p, hsa-miR-142-3p, and hsa-miR-509-3p. Furthermore, the bidirectional hierarchical clustering of 27 miRNAs was classified into three subtypes, and TMB and four types of immune cells, including activated dendritic cells, naïve CD4 T cells, M1 macrophages, and plasma cells, showed significant differences among the three subtypes. The expression levels of most miRNAs in the GSE18509 and GSE24996 datasets were consistent with those in the training dataset. These miRNAs, including hsa-let-7c-5p, hsa-miR-130b-3p, and hsa-miR-142-3p, and activated dendritic cells, naïve CD4 T cells, M1 macrophages, and plasma cells may play vital roles in the pathogenesis of melanoma.

Altitude Effect on Cutaneous Melanoma Epidemiology in the Veneto Region (Northern Italy): A Pilot Study

The incidence of cutaneous melanoma has been increasing in the last decades among the fair-skinned population. Despite its complex and multifactorial etiology, the exposure to ultraviolet radiation (UVR) is the most consistent modifiable risk factor for melanoma. Several factors influence the amount of UVR reaching the Earth’s surface. Our study aimed to explore the relationship between melanoma and altitude in an area with mixed geographic morphology, such as the Veneto region (Italy). We included 2752 melanoma patients who were referred to our centers between 1998 and 2014. Demographics, histological and clinical data, and survival information were extracted from a prospectively maintained local database. Head/neck and acral melanoma were more common in patients from the hills and the mountains, while limb and trunk melanoma were more common in patients living in plain and coastal areas. Breslow thickness, ulceration and mitotic rate impaired with increased altitude. However, the geographical area of origin was not associated with overall or disease-free survival. The geographical area of origin of melanoma patients and the “coast-plain-hill gradient” could help to estimate the influence of different sun exposure and to explain the importance of vitamin D levels in skin-cancer control.

Revisiting miRNA Association with Melanoma Recurrence and Metastasis from a Machine Learning Point of View

The diagnostic and prognostic value of miRNAs in cutaneous melanoma (CM) has been broadly studied and supported by advanced bioinformatics tools. From early studies using miRNA arrays with several limitations, to the recent NGS-derived miRNA expression profiles, an accurate diagnostic panel of a comprehensive pre-specified set of miRNAs that could aid timely identification of specific cancer stages is still elusive, mainly because of the heterogeneity of the approaches and the samples. Herein, we summarize the existing studies that report several miRNAs as important diagnostic and prognostic biomarkers in CM. Using publicly available NGS data, we analyzed the correlation of specific miRNA expression profiles with the expression signatures of known gene targets. Combining network analytics with machine learning, we developed specific non-linear classification models that could successfully predict CM recurrence and metastasis, based on two newly identified miRNA signatures. Subsequent unbiased analyses and independent test sets (i.e., a dataset not used for training, as a validation cohort) using our prediction models resulted in 73.85% and 82.09% accuracy in predicting CM recurrence and metastasis, respectively. Overall, our approach combines detailed analysis of miRNA profiles with heuristic optimization and machine learning, which facilitates dimensionality reduction and optimization of the prediction models. Our approach provides an improved prediction strategy that could serve as an auxiliary tool towards precision treatment.