An Observational Study on the Molecular Profiling of Primary Melanomas Reveals a Progression Dependence on Mitochondrial Activation

Abstract

Melanoma in advanced stages is one of the most aggressive tumors and the deadliest of skin cancers. To date, the histopathological staging focuses on tumor thickness, and clinical staging is a major estimate of the clinical behavior of primary melanoma. Here we report on an observational study with in-depth molecular profiling at the protein level including post-translational modifications (PTMs) on eleven primary tumors from melanoma patients. Global proteomics, phosphoproteomics, and acetylomics were performed on each sample. We observed an up-regulation of key mitochondrial functions, including the mitochondrial translation machinery and the down-regulation of structural proteins involved in cell adhesion, the cytoskeleton organization, and epidermis development, which dictates the progression of the disease. Additionally, the PTM level pathways related to RNA processing and transport, as well as chromatin organization, were dysregulated in relation to the progression of melanoma. Most of the pathways dysregulated in this cohort were enriched in genes differentially expressed at the transcript level when similar groups are compared or metastasis to primary melanomas. At the genome level, we found significant differences in the mutation profiles between metastatic and primary melanomas. Our findings also highlighted sex-related differences in the molecular profiles. Remarkably, primary melanomas in women showed higher levels of antigen processing and presentation, and activation of the immune system response. Our results provide novel insights, relevant for developing personalized precision treatments for melanoma patients.

Keywords: MHC complex; PTMs; antigen presentation; disease progression; lysine acetylation stoichiometry; malignant melanoma; mitochondria; mitochondrial translation; primary melanoma; proteogenomics.

Figure 1

Multiomic analysis of primary melanomas. (A) Sample preparation workflow for global proteomics, phosphoproteomics and acetylomics. (B) Distribution of proteins identified in the primary melanoma cohort and the tissue composition according to the histopathological characterization (% of tumor cells and stromal compartment) (top panel), distribution of the PTMs identified (acetylation and phosphorylation) (middle and bottom panels), and relevant clinical data of the patients and samples in the cohort study.

Figure 2

Histopathological characterization of the cohort of primary melanomas. (A) Tissue sections before and after the scanner (Leica Biosystems, San Diego, CA, USA), and the digital images were submitted to histopathological characterization of the tissue composition. Histological images were analyzed under the QuPath platform using a cell-based profiling mode. Each section is presented as a H&E scanned image (left) and an annotated image as the output of the analysis platform (right). The summary of the histological characterization is outlined in Table S1. All images are arranged as top-down (from skin surfaces to inside of skin). (B) Representation of the histopathological characterization of the examined primary melanoma samples according to standardized dataset involving the growth, structural and cellular phenotypes, stromal niche variables (Fib: fibrotic/desmoplastic reaction; Me: melanophages; Ly: lymphocytic infiltration), regression, as well as pigmentation. For the extension of the horizontal spreading of the melanoma ‘+’ indicates focal superficial spreading counterpart which limited less than 25% of the total horizontal extension of the examined tumor sample without the dominancy of the superficial spreading component. The ‘++’ indicates well recognizable superficial counterpart occupying >25% of the total horizontal extension; however, with a dominant vertical spreading/regressive area. The ‘+++’ indicates a marked and dominant superficial spreading component >75% of the total horizontal extension of the examined tumor sample. Indeed ‘+++’ in the regression column indicates a dominant diffuse (>75%) regressive component within the tumor sample (PT46). Similarly, intratumoral pigmentation was semiquantitatively counted by ‘+++’ diffuse and marked pigmentation, ‘++’ scattered focal, but pronounced pigment accumulation, ‘+’ focal and moderate intratumoral pigmentation, ‘−/+’ focal mild pigmentation within the tumor cells.

Figure 3

Functional annotation enrichment analysis of the difference at the proteome level (A) at the acetylome (B), and at the phosphoproteome level (C), between the groups of primary melanomas from patients where the disease progressed and those without metastatic event. Highlighted in red and blue are relevant annotations enriched in the Progression and No progression groups respectively.

Figure 4

Molecular signature of the progression of primary melanomas toward metastatic disease. (A) Heatmap illustration of the significantly dysregulated proteins in primary melanomas from patient where the disease progressed. (B) Heatmap of the phosphorylated peptides dysregulated between the progression status groups, and their corresponding phosphoproteins. (C) Protein-biological annotation network of the clusters of proteins from differential abundance analysis of the proteins and PTMs dataset related to the progression of the disease. The analysis was performed using the ToppCluster function of the ToppGene Suite online version [20]. (D) Clustering analysis and heatmap illustration of the protein elements of the mitochondrial translation machinery. The corresponding abundance of the proteins in each sample is represented as a boxplot graph.

Figure 5

Functional biological annotations associated with the Breslow thickness of primary melanomas. (A) Functional enrichment analysis of the proteins with a significant correlation to the Breslow thickness of the tumors. (B) 2D functional enrichment analysis between the correlation at the proteome level with Breslow thickness of the tumors and the progression status of the patients. Highlighted in red are relevant annotations up-regulated in the progression group and with positive correlation with the Breslow thickness; in green, annotations down-regulated in the progression group and with negative correlation with Breslow thickness; in purple and brown are annotations oppositely correlate with progression and Breslow thickness.

Figure 6

Most frequently mutated genes in primary and metastatic melanoma are involved in key pathways for the development and progression of the disease. (A) Genes with the highest frequency alterations among primary and metastatic tumor samples using the project GENIE melanoma dataset. * Denotes significant differences (p < 0.05) in the frequency of alterations between primary and metastatic tumor samples. (B) Genes with significant differences in frequency alterations among primary and metastatic tumor samples (q-value < 0.05). (C) Network representation of the functional clustering analysis of genes with differences in frequency alterations between primary and metastatic melanomas. The genes were filtered based on a p-value 0.05 and the analysis was performed using the ToppCluster function of the ToppGene Suite online version [20].

Figure 7

Transcriptionally dysregulated biological pathways associated with progression in melanoma. Gene set enrichment analysis (GSEA) using the GO: Biological Process gene set database showed significant enrichment of pathways associated with progression (A, left) and non-progression (right) in primary tumors from TCGA. Enriched Pathways in Melanoma Metastases vs. Primary tumors from TCGA (B). GSEA using the GO:Biological Process gene set database showed significant enrichment of pathways associated with metastases (left) and primary tumors (right).

Figure 8

The sex-related differences in the melanoma proteome. (A) 1D functional annotation enrichment analysis of the differences at the proteome (left), acetylome (middle), and phosphoproteome (right) levels between women and men. (B) Functional annotation-protein network of identified elements dysregulated between women and men in the three datasets. the analysis was performed using the on the ToppCluster function of the ToppGene Suite online version [20].