Mutually exclusive genetic interactions and gene essentiality shape the genomic landscape of primary melanoma

Abstract

Melanoma is a heterogenous malignancy with an unpredictable clinical course. Most patients who present in the clinic are diagnosed with primary melanoma, yet large-scale sequencing efforts have focused primarily on metastatic disease. In this study we sequence-profiled 524 American Joint Committee on Cancer Stage I-III primary tumours. Our analysis of these data reveals recurrent driver mutations, mutually exclusive genetic interactions, where two genes were never or rarely co-mutated, and an absence of co-occurring genetic events. Further, we intersected copy number calls from our primary melanoma data with whole-genome CRISPR screening data to identify the transcription factor interferon regulatory factor 4 (IRF4) as a melanoma-associated dependency. © 2022 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

Keywords: CRISPR screening; driver genes; gene essentiality; genetic epistasis; genome sequencing; genomic landscape; interferon regulatory factor 4 (IRF4); melanoma; primary cancer; somatic mutation.

Figure 1

Summary of coding single nucleotide variants (SNVs) and driver genes in human primary melanoma (n = 524). (A) Overview of coding SNVs in primary melanoma. Top panel shows the number of exonic mutations in the capture regions and the distribution of variant consequences. Bottom panel shows the proportion of each base‐change type across the sample collection. (B) Primary melanoma driver genes identified using the algorithm DNdScv [31] and their respective alteration rate in the tumour collection.

Figure 2

Mutational landscape of human primary melanoma (n = 524). (A) The most frequently mutated single nucleotide positions (exonic and promoter region). (B) Overview of the genetic landscape showing mutation load, non‐synonymous mutations in candidate driver genes, promoter mutations and copy number alterations. For SNVs, where there were multiple mutations in a gene in a case, the most pathogenic was plotted.

Figure 3

Mutually exclusive genetic interactions in human primary melanoma. (A) Gene pairs displaying a mutually exclusive alteration pattern. Samples with mutations in one gene in the pair are shown in red, the other gene in yellow, and samples with mutations in both genes are shown in orange. P values were calculated using the DISCOVER algorithm [32]. (B) Genetic alterations targeting important components of the CDKN2A (p16INK4A)‐associated regulatory pathway, including the newly discovered CDKN2A‐mutually exclusive gene PRDM2.

Figure 4

Whole genome copy number landscape and gene essentiality analysis. An overview of copy number alterations for (A) the human primary melanomas (Leeds melanoma cohort) [21] and (B) the TCGA SKCM dataset [17]. All segments with a copy number differing more than 0.6 from the sample average were used to generate the figures. Red illustrates gains and blue losses. (C) CRISPR lethality scores (higher scores correspond to a larger reduction in cell viability when the specific gene is silenced) of eight genes associated with lethality/reduced cellular fitness in skin cancer cell lines (Fisher's exact test, p‐adj < 0.01) versus all other cell lines in the Broad DepMap collection [34]. These data and the analysis approach used were described in detail in Christodoulou et al. [37]. Red indicates a negative effect of CRISPR on cancer cell line fitness. Note that grey does not indicate an effect on cell fitness because the Bayes algorithm used is not configured to identify effects that enhance cell growth (see Materials and methods). (D) Amplified regions in the TCGA SKCM cohort overlaid with the genomic location of the eight genes associated with lethality in skin cancer cell lines. (E) 25‐Mb regions of high‐level amplifications of chromosome 6, shown for all IRF4‐amplified samples. The location of IRF4 is shown in grey. (F) Expression of IRF4 in the Rahman et al. [57], reprocessed TCGA expression dataset. (G) Correlation between IRF4 expression and cell line CRISPR lethality scores.

Figure 5

Validation of the essentiality of IRF4 in melanoma cell lines. (A) siRNA‐mediated knockdown of IRF4. WM1799 and RVH421 are cell lines identified by CRISPR screening as being sensitive to IRF4 loss, whereas HT144 and WM983B were not scored in this way. WT, NC, PC and IRF KD refer to siRNA treatments and correspond to untransfected, negative control, positive control and IRF4 siRNAs, respectively. More details are provided in Materials and methods. These data were collected from three independent experiments. Analysis was performed using a Student's two‐tailed t‐test comparing the number of viable/live cells between the IRF4 siRNA transfected cells versus cells transfected with the negative control siRNA. *P < 0.05, **P < 001. (B) Analysis of protein lysates for IRF4 and c‐MYC and for expression of the loading controls GAPDH/VINCULIN. These experiments are representative of three independent experiments.