Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Jan 1.
Published in final edited form as: Oral Oncol. 2023 Nov 30;148:106582. doi: 10.1016/j.oraloncology.2023.106582

Genomic and transcriptomic analysis of cutaneous squamous cell carcinoma arising in immunocompetent and immunosuppressed patients

Kristin P Bibee 1,2,*, Aditi Kulkarni 2, Sanghoon Lee 2,3, Johan Ho 1, Hatice Ulku Osmanbeyoglu 2,3,4,5, Robert L Ferris 2, Dan P Zandberg 2
PMCID: PMC10917075  NIHMSID: NIHMS1950353  PMID: 38039877

Abstract

Background:

Cutaneous squamous cell carcinoma (cSCC) is the most common skin malignancy arising in immunocompromised patients such as solid organ transplant recipients. In addition to an abundance in number, the morbidity and mortality of these tumors in this patient population exceeds that of immune competent individuals. Here, we used whole exome and bulk RNA sequencing to analyze mutation profiles between tumors arising in immunocompetent and immunosuppressed patients.

Methods:

DNA and RNA extracted from twenty formalin-fixed, paraffin embedded tumors and adjacent skin was sequenced. Bioinformatic analysis revealed tumor mutational burden, mutational signatures, microsatellite instability, and aberrant signaling pathways.

Results:

Similar median tumor mutational burden was found in both the tumors from the immunocompetent and the immunosuppressed cohorts. Mutation signature analysis revealed UVR signatures and evidence of azathioprine exposure. 50% of tumors from the immunosuppressed patients have mutations consistent with microsatellite instability, yet mismatch repair protein expression was preserved in the samples analyzed. Additionally, frequently mutated genes in this cohort belong to the extracellular matrix receptor interaction and calcium signaling pathways, suggesting these may be targets for future treatments of this disease.

Conclusions :

This study utilizes whole exome and bulk RNA sequencing to identify difference between cSCC arising in immunosuppressed and immunocompetent patients using the patient’s photo exposed, but histologically normal appearing skin as the “germline” comparison. We demonstrate an enrichment in microsatellite instability in the tumors from immunosuppressed patients and differences in oxidative phosphorylation and epithelial-mesenchymal transition which may be targets for therapeutic intervention based on identification of mutations.

Keywords: cutaneous squamous cell carcinoma, immunosuppression, organ transplantation, genomics, transcriptomics

Introduction

Cutaneous squamous cell carcinoma (cSCC) is the second most common malignancy with upwards of 700,000 cases in the USA each year[1]. Immunosuppression increases the risk of cSCC with solid organ transplant recipients (SOTRs) having a 65-fold increased risk of developing cSCC [2,3]. Additionally, SOTRs also develop tumors which are more aggressive leading to increased morbidity and mortality when compared to immunocompetent individuals[4]. While inadequate surveillance and tumor elimination by the immune system is postulated to be a driving factor in immunosuppressed patients, genomic, using whole exome sequencing (WES), and transcriptomic differences in tumors arising in these patients have not been evaluated.

Ultraviolet radiation (UVR) is the established carcinogen for cSCC development with multiple prior studies showing UVR signature mutations in cSCC. Prior studies have investigated the mutational signature in gene panels of cSCC and found mutations in NOTCH1, NOTCH2, PTEN, and TP53. Additionally, even photodamaged skin that is not histologically consistent with a tumor shows mutations in known oncogenes such as TP53 [5]. Recent hotspot mutation analysis on tumors arising in immunocompromised patients showed fewer UVB related mutations when compared to those arising in immunocompetent patients [6]. Yet, genomically and transcriptomically dysregulated pathways leading to carcinogenesis of keratinocytes in this population remain unclear. Given the exposure to multi-modality immunosuppressing medications, including those known to induce specific mutation signatures, we hypothesize that there are differences in the DNA mutations and transcriptional programs between IC vs IS patients.

Here, we coupled WES data of tumor and adjacent, photo-exposed but normal-appearing skin with bulk RNA sequencing to determine genome-based program differences in cSCC arising in immunosuppressed (IS) and immunocompetent (IC) patients.

Materials and Methods

Identification and Preparation of samples

After institutional IRB approval (HCC 18–191), cases were identified from the pathology database at UPMC with slides independently reviewed. Squamous cell carcinoma in situ and keratoacanthoma type cSCC were excluded. Immunosuppression was defined by the patient’s current medication regimen at the time of tumor diagnosis. Adjacent, photo-exposed skin was also collected to serve as a control. This skin showed histological evidence of photodamage, but no keratinocyte atypia. Tumor regions were mapped onto paraffin sections using adjacent hematoxylin and eosin-stained sections. Slide sections were macrodissected based on marked regions. Qiagen AllPrep FFPE Kit was used for DNA and RNA extraction following manufacturer’s protocol.

Whole Exome Sequencing

Extracted FFPE DNA was quantitated using Qubit™ dsDNA BR Assay Kit (Thermo Fisher Scientific) and KAPA hgDNA Quantification and QC Kit (KAPA). DNA libraries were prepared using the KAPA Hyper Plus Kit (KAPA Biosystems). 500 ng of genomic DNA was processed through fragmentation, enzymatic end-repair and A-tailing, ligation, followed by quality check using Fragment Analyzer (AATI). Exonic regions were captured using the xGen Exome Research Panel v1.0 with xGen Universal Blockers (IDT) and SeqCap EZ Hybridization and Wash Kit (Roche) according to manufacturer protocol. Libraries with an average size of 450 bp (range: 300–600bp) were quantified by qPCR on the LightCycler 480 (Roche) using the KAPA qPCR quantification kit (KAPA biosystem). The libraries were normalized and pooled as per manufacturer protocol (Illumina). Sequencing was performed using NovaSeq 6000 platform (Illumina) with 151 paired-end reads to an average target depth of 30–50X for germline and 250X coverage for somatic samples.

RNA Sequencing

Extracted FFPE RNA was quantitated using Qubit™ RNA BR Assay Kit (Thermo Fisher Scientific) followed by the RNA quality check using Fragment Analyzer (AATI). The RNA libraries were prepared using the TruSeq RNA Exome Kit (Illumina), starting from 100ng of RNA input, according to manufacturer’s protocol. Libraries with an average size of 450 bp (range: 300–600bp) were quantified by qPCR on the LightCycler 480 (Roche) using the KAPA qPCR quantification kit (KAPA biosystem). The final libraries were pooled and sequenced using NovaSeq 6000 platform (Illumina) to an average of 65M 100PE reads.

Data Analysis

Variant calling from WES

The nf-core Sarek pipeline (v2.5.1) was used for quality control, alignment, sorting and recalibrating BAM files[7]. Variant calling and filtration were performed on the paired samples using Mutect2 (v4.1.4.0) [8]. The variants that passed the filtering step were annotated using Funcotator and exported in the MAF format. The Maftools (v2.2.10) R package was used to analyze and visualize the MAF files [9].

Mutation signature analysis

The sigminer package (v.1.0.0) was used to derive mutational signatures representing the cohort. These signatures were compared to existing signatures in the COSMIC database to look for similarities.

Gene set overlap analysis

The “Investigate Gene Sets” module from MSigDB was used to identify overlapping gene sets between the top two hundred most frequently mutated genes in our dataset and existing gene sets[10,11]. The KEGG pathways gene set was chosen for this analysis.

Neoantigen discovery from RNA sequencing

The NeoFuse pipeline was used for predicting fusion neoantigens from paired end tumor RNA sequencing reads[12]. The default filtering strategy used in the pipeline was used to narrow down the list of predicted neoantigens for each sample.

Microsatellite Instability Calculation

Using exomes from paired tumor-normal samples, the MSIsensor package was used to evaluate the presence of somatic microsatellite instability sites[13]. A cut-off of 10 percent was used to define MSI high samples.

RNA sequencing analysis

The nf-core rnaseq pipeline (v.1.3) was used to perform QC, align reads, and quantify the readouts at the gene or transcript level. BAM files generated at this step were used for downstream analysis. We used Rsubread [14] to align sequence reads to reference human genome (GRCh38) and mapped the aligned sequences to Entrez Genes. Differential gene expression was estimated using DESeq2 package [15] (e.g. IC tumors (n=5) vs IS tumors (n=7) or IC normal (n=5) vs IS normal (n=6)).

Motif activity analysis

To infer activities of transcription factor binding motifs (TFBM) from cSCC RNA-seq data, we used the Integrated System for Motif Activity Response Analysis (ISMARA) [16]. We used Wilcoxon rank sum test to compare TF activities and assess the association between TFs and patient clinical types. The distribution of gene expression changes was visualized for TF-target genes and the rest of all expressed genes. Two-tailed Kolmogorov–Smirnov test was used to determine the significance between the distributions of TF-target genes and the rest of expressed genes. We used all 16,714 genes as background genes after removing genes with low mean counts across samples.

Gene set enrichment analysis

Gene Set Variation Analysis (GSVA) [17] was performed using the GSVA R package. We obtained pathway annotations from the Molecular Signature Database (MSigDB) for the collection of hallmark of cancers. The distribution of gene expression changes (IC tumor vs IC normal or IS tumor vs IS normal) was calculated for signature genes and the rest of all expressed genes. Two-tailed Kolmogorov–Smirnov test is used to determine the significance between the distributions of signature genes and the rest of expressed genes.

Results

In total, DNA from 20 tumors and matched photo-exposed skin met quality assurance standards for WES. Of those cases, 14 tumors came from patients who were immunosuppressed. Two individuals had two independent tumors diagnosed at the same time but on different locations on their body. RNA from 12 tumors and matched photo-exposed skin met quality assurance standards for bulk RNA sequencing. The IS patients were all SOTRs except one patient with inflammatory bowel disease. All SOTRs were greater than one year post transplant (range 1.8–29.3 years). All the tumors were from sites on the head and neck, treated by either Mohs micrographic surgery by Dermatologic surgeons or wide local excision by ENT-Head and Neck surgeons. The majority of tumors were BWH staging T1(n=12). The most commonly occurring high risk factor was clinical size greater than 2cm, which occurred in 8 of the 20 tumors. The details of the patients and tumors are available in Supplemental Table 1.

cSCC TMB is similar between immunocompetent and immunosuppressed individuals

Somatic mutation counts found in tumors when compared to photo-exposed skin from the same individual is represented in Figure 1A. The median TMB was 34 mutations per megabase (mut/Mb) (range:1–136) with a similar median TMB based on immune status (40 mut/mb immunocompetent vs. 32 mut/Mb in immunosuppressed). The non-synonymous mutations were categorized by type with missense mutations being the most common in each tumor (Figure 1B). The number of predicted fusion neoantigens was also similar by immune status and the number of predicted fusion neoantigens did not correlate with the number of mutations (Figure 1C).

Figure 1. Tumor mutation landscape.

Figure 1.

Open in a new tab

A. Number of somatic mutations found in each tumor when compared to photoexposed skin, grouped by immune status. B. Genomic effect of non-synonymous mutations as a proportion of mutations. C. Fusion neoantigens predicted compared to tumor mutation burden by cohort.

Mutation Signature analysis reveals expected UV radiation signature

Utilizing the sigminer package (v.1.0.0), three mutation signatures were derived from the group(Figure 2A)[18]. Signature 1 resembled the UVR exposure signature (COSMIC 7). All tumors contributed to signature 1 (Fig. 2B). Signature 3 resembles the azathioprine exposure signature previously described[19]. Signature 2 is most like due to sequencing artifact, although it bears some resemblance to an indirect UV radiation signature (COSMIC 38).

Figure 2. Mutational signatures.

Figure 2.

Open in a new tab

A. Three mutation signatures were identified in the entire cohort. Further analysis reveals similarity between signature 1 and COSMIC-38, Signature 2 and COSMIC-7a and 7b, and Signature 3 and COSMIC-32. B. All tumors contributed to signature 1 whereas IS3a, IS3b, IS4, IS5, IS8, IS9, and IS11 contributed to signature 2. IS6 was the major contributor to Signature 3 with minor contributions from other tumors from both immunosuppressed and immunocompetent individuals.

Microsatellite instability was more common in immunosuppressed patients

To better understand the mechanism of carcinogenesis, we examined the samples for evidence of dysfunctional DNA repair. Utilizing the MSIsensor package, microsatellite instability (MSI) was more common in the tumors from IS (50%) than IC patients (17%)(Fig. 3A)[13]. Of those IS patients, the majority of the MSI high tumors (85.8%) came from patients on antimetabolite therapy. Immunohistochemistry for the four mismatch repair proteins accepted as the clinical standard for assessing MSI (MLH1, MSH2, MSH6, PMS2) was performed and showed preservation in all three of the MSI high tumors analyzed, two from the IS, one from the IC cohort (Fig. 3B).

Figure 3. Microsatellite Instability.

Figure 3.

Open in a new tab

A. Tumor mutation burden as it relates to Microsatellite instability. Tumors with MSI score of >10 are considered unstable. 50% of the tumors from immunosuppressed patients were MSI high whereas only one of the six tumors from an immunocompetent individual was MSI high. B. Mismatch protein expression is preserved in tumors identified as MSI high.

Genomic mutation data predicts aberrant pathways

Prior studies have suggested a role for mutations in genes involved in cell cycle control pathways, differentiation pathways, Ras family members, and chromatin changes. [20–22] Here, we interrogated the commonly mutated genes in our samples (Supplemental Figure 1).

After setting a filter threshold of five reads supporting the alternate allele, the top two hundred mutated genes in the group were analyzed. Gene set overlap analysis was performed and demonstrated unique alterations in mutations in the extracellular matrix as well as in calcium signaling pathways. These pathways were then interrogated on all samples at the DNA level (Fig 4 A, B).

Figure 4. Mutated Genes.

Figure 4.

Open in a new tab

In addition to the previously published mutated genes, we find two unique pathways with an increase in mutations in cSCC: extracellular matrix genes (A) and calcium signaling genes (B).

Transcription factor and pathway activity differences between immunocompetent and immunosuppressed individuals

To analyze the expression changes associated with immunocompetent and immunosuppressed tumors, we performed RNA-seq analysis. From a gene expression perspective, the tumors and normal skin were easily distinguished (Figure 5A). However, there was not such a distinction between tumors from immunocompetent compared to tumors from immunosuppressed (Figure 5B). As seen in Figure 5C, two tumors in the immunosuppressed group stood out as outliers. Next, we performed sample-specific transcription factor (TF) activity analysis via the Integrated System for Motif Activity Response Analysis (ISMARA)[16] and sample-specific pathway enrichment analysis via GSVA [17]using the bulk RNA sequencing data to identify altered pathways and transcription factor activity differences. Inferred activities of 7 TF motifs were significantly associated with patient groups (immunosuppressed vs immunocompetent) by p-value < 0.05 and absolute mean activity difference > 0.01. We found that ATF4, PLAG1, SOX14, and RORA had significantly higher activities in IS than IC tumors. FOXN1 had significantly higher predicted activity in IC than IS tumors. Figure 6A shows the cumulative distribution of TF target gene expression changes between IC and IS tumors compared to background genes. TFs identified for lower activity in IC tumor than IS tumors were associated with the downregulation of their targets. A similar pattern of activity was seen for ATF4 when comparing normal skin from IC and IS (Figure 6B).

Figure 5. Transcriptional Differences.

Figure 5.

Open in a new tab

(A) Comparison of transcribed genes between normal skin and tumor shows an appreciable difference in program. (B) Tumors from immunosuppressed patients do not stratify in their gene expression profiling from those from immunocompetent individuals. (C) Two outliers are identified in gene expression profile within the tumors from immunosuppressed individuals (IS2, IS9).

Figure 6. Transcription Factor and Pathway Differences.

Figure 6.

Open in a new tab

(A) Expression changes of target genes for IS tumor- or IC tumor-specific TFs Targets of ATF4, PLAG1, RORA, and SOX14, suggest higher activity of these TFs in IS tumors. The blue lines are CDFs for background gene log2 fold changes between IC tumor and IS tumor. The p values are from the two-sided Kolmogorov–Smirnov (K–S) test between the target and the background distributions. (B) The association between TF activity and its target gene expression changes in normal skin from IS or IC patients differences in ATF4. (C) Genes involved in oxidative phosphorylation pathway showed significant upregulation in IC tumors compared to normal whereas genes involved in epithelial mesenchymal transition pathway showed significant upregulation in IS tumors compared to normal. The yellow lines are eCDFs for genes involved in the pathways log2 fold changes between IC tumor vs IC normal. The blue lines are between IS tumor and IS normal.

Further, the expression profiles were compared against the known cancer hallmark pathways and showed that gene expression patterns associated with oxidative phosphorylation was decreased in the immunosuppressed tumors compared to the normal (Fig 6C) and epithelial-mesenchymal transition pattern was increased in the tumors from immunosuppressed patients (Fig 6C).

Discussion

In order to better understand the genomic alterations in cSCC arising in immunosuppressed and immunocompetent patients, we analyzed 20 tumors and adjacent photo-exposed but histologically normal skin via WES and bulk RNA sequencing. Keratinocyte carcinomas are known to have some of the highest TMBs, with 61 mutations/Mb previously reported for cSCC [22]. While our TMB was lower at 34 mutations/Mb, likely resulting from using photodamaged skin as the reference, these TMBs are still considerably higher than other solid tumors including melanoma (13.2 mutations/Mb), and HNSCC (3.2 mutations/Mb)[22]. Higher TMB has been associated with increased efficacy with anti-PD-1 mAb immunotherapy in solid tumors, and the FDA approved anti-PD-1 mAb Cemiplimab has a very high response rate, near 50%, in advanced immunocompetent cSCC [23][24]. Importantly, in retrospective series, comparable response rates to IC patients were observed with anti-PD-1 mAb therapy in SOTR with various malignancies, including some cSCC patients, however graft rejection risk prevents its regular usage in cSCC with SOTRs. Notably, immunosuppressed patients in our study had a high TMB (32 mutations/Mb) which was similar to immunocompetent patients. Interestingly, no increase in fusion neoantigens was predicted with a higher TMB in the group thought to be subjected to lower immunoediting due to immunosuppression.

Traditionally, MSI is linked to defective mismatch repair genes. cSCC has a low incidence of gene mutations in these proteins [25]. However, cSCC is a tumor type seen in patients with inherited defects in mismatch repair proteins such as Lynch syndrome. In this data set, the tumors from immunosuppressed patients showed a higher percentage of tumors with MSI, via MSIsensor analysis, a previously validated testing method for MSI [13,26–28]. Interestingly, of the tumors in the IS cohort which were MSI-high, no mutations were found in the common mismatch repair genes and protein expression was preserved. The mismatch repair pathway is not only dependent on functioning proteins for repair but also on the building blocks for DNA synthesis. The mechanism of action of many immunosuppressing medications lies in DNA substrate synthesis, therefore it may be the case that the medications used for immunosuppression are also perpetuating DNA damage by limiting DNA synthesis.

In addition to the previously published mutations (Supp Figure 1), we found an increase in mutations in ECM genes as well as in calcium signaling genes. Interestingly, the gene products from these genes are often more recognized for their role in the stroma (fibroblasts, pericytes, VSMCs, endothelial cells) than in the keratinocytes. Yet, prior work by Lee and colleagues has also demonstrated collagen mutations in cSCC and show that mutations are related to invasion [29]. Notably, their work in cSCC suggests a role for mutant COL11A1 in beta 1 integrin signaling. Here, using paired RNA seq data, we see prediction of upregulation of HOXD1 specifically in tumors from immunosuppressed patients, which is a transcription factor known to play a role in extracellular matrix regulation through beta1 integrin expression [30].

Calcium signaling is an important pathway for terminal differentiation of keratinocytes. Therefore, it may be the case that alterations in these pathway members lead to disruption of differentiation further propagating the cancer phenotype. Eighteen of the twenty tumors analyzed had a mutation in at least one of these pathway members. Currently, systemic retinoids are used to prevent skin cancer development in high risk populations, including organ transplant recipients. One mechanism of action for systemic retinoids is the promotion of differentiation of keratinocytes. Given this data, it may be the case that drugging the calcium signaling pathway more specifically may be fruitful in chemoprevention. Finally, we found differential predicted transcription factor activity between our two tumor groups. ATF4 is related to the unfolded protein response. The predicted increase in transcription factor activity was found when we compared both tumor gene expression (Figure 6A upper left panel) and normal skin gene expression (Figure 6B) between the two groups. Therefore, there may be a role for the antirejection medications driving this phenotype seen across the normal and dysplastic skin. PLAG1 plays a role in adnexal tumor formation, but no known role for keratinocyte cancers of the skin[31]. SOX14 has been implicated in cervical cancer proliferation and invasion, but not in cSCC [32]. RORA loss promotes progression of cSCC[33]. Here we see an increase in the RORA transcription factor activity in the immunosuppressed tumors. As RORA is a tumor suppressor, this would suggest there is decreased RORA expression in the immunosuppressed tumors and may be a driver of the aggressive phenotype in this subpopulation. Further work to understand the role of these pathways is warranted as they may represent actionable pathways.

In evaluating known hallmark gene expression pathways, we found statistically significant differences in oxidative phosphorylation as well as epithelial-mesenchymal transition (EMT) between our two tumor groups. Interestingly, expression of genes associated with oxidative phosphorylation is decreased in the tumors from immunosuppressed patients whereas the genes associated with EMT are increased in this population. These may be therapeutic targets as more is understood about these contributors to the cancer phenotype.

While genomic studies often use DNA from the white blood cells as the germline control, our study is unique in that we used adjacent photo-damaged, but histologically normal skin which had also been formalin fixed and paraffin embedded from each patient for our control DNA. Photo-exposed skin is known to harbor mutations in many oncogenes. Therefore, our goal in using skin as the control DNA for comparison was to better tease out differences from photo-exposed skin and true tumor of the skin as well as to ensure the DNA has been processed in a similar manner to minimize the effect of processing artifacts. We hypothesize that using skin as control has the potential to identify mutations and transcription pathways truly unique to the tumor which may be more relevant therapeutically.

Our study has several limitations, the largest being the limited number of samples and heterogeneity of immunosuppression regimens. That being said, there were some differences observed that would be important to confirm in a larger sample size. Additionally, while all samples were able to undergo WES, only a subset had high enough quality RNA from FFPE extraction for RNA sequencing.

To our knowledge this is the first comparison of the mutational landscape of immunosuppressed and immunocompetent cSCC via WES and RNAseq with transcription factor activity prediction. Additionally, in our analysis we uniquely exclusively used photodamaged skin as our “germline” control, which we feel is important to identifying the most relevant mutations in cSCC. There is a great need to improve therapeutic options for immunosuppressed patients, both early by dermatologists and later in more advanced disease by medical oncologists. Despite a smaller sample size, we observed a difference in transcription programs that have the potential to be targeted to improve outcomes in immunosuppressed patients with cSCC.

Supplementary Material

1
2

Supplemental Figure 1. Previously reported genes with mutations found in cSCC. Ninety percent of our samples had at least one mutation in a gene previously reported to be mutated in cSCC with TP53 being the most common mutation in our cohort.

  • Increase in microsatellite instability in cSCC from immunosuppressed patients

  • ECM and calcium signaling pathway differences may be useful for therapeutic intervention

  • ATF4, PLAG1, SOX14, and RORA with higher predicted activity in cSCC from immunosuppressed patients

Acknowledgements

We thank Dr. Mary Collins, Dr. Michael Cardis, Carly Reeder, and Annie Roble for technical assistance. KPB was supported by T32 CA060397-20. DPZ was supported by CEP award of P30 CA047904. HUO was supported by R00CA207871 This research was supported in part by the University of Pittsburgh Center for Research Computing through the resources provided. This work was supported by the UPMC Genome Center with funding from UPMC’s Immunotherapy and Transplant Center.

This work was funded by the following:

KPB was supported by T32 CA060397-20.

HUO was supported by R00CA207871.

DPZ was supported by CEP award of P30 CA047904.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Conflict of Interest Statement:

None declared.

No conflicts of interest to disclose related to this work.

The data underlying this article will be available in a repository at the time of publication.

The University of Pittsburgh IRB approved this work HCC 18–191.

References

  • [1].Karia PS, Han J, Schmults CD. Cutaneous squamous cell carcinoma: Estimated incidence of disease, nodal metastasis, and deaths from disease in the United States, 2012. J Am Acad Dermatol 2013. 10.1016/j.jaad.2012.11.037. [DOI] [PubMed] [Google Scholar]
  • [2].Hartevelt MM, Bouwes Bavinck JN, Kootte AMM, Vermeer BJ, Vandenbroucke JP. Incidence of skin cancer after renal transplantation in the netherlands. Transplantation 1990. 10.1097/00007890-199003000-00006. [DOI] [PubMed] [Google Scholar]
  • [3].Garrett GL, Blanc PD, Boscardin J, Lloyd AA, Ahmed RL, Anthony T, et al. Incidence of and risk factors for skin cancer in organ transplant recipients in the United States. JAMA Dermatology 2017;153. 10.1001/jamadermatol.2016.4920. [DOI] [PubMed] [Google Scholar]
  • [4].Chan AW, Fung K, Austin PC, Kim SJ, Singer LG, Baxter NN, et al. Improved keratinocyte carcinoma outcomes with annual dermatology assessment after solid organ transplantation: Population-based cohort study. Am J Transplant 2019. 10.1111/ajt.14966. [DOI] [PubMed] [Google Scholar]
  • [5].Martincorena I, Roshan A, Gerstung M, Ellis P, Loo P Van, Mclaren S, et al. High burden and pervasive positive selection of somatic mutations in normal human skin Europe PMC Funders Group. Science (80- ) 2015;348:880–6. 10.1126/science.aaa6806. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [6].Lobl MB, Clarey D, Higgins S, Thieman T, Wysong A. The correlation of immune status with ultraviolet radiation–associated mutations in cutaneous squamous cell carcinoma: A case-control study. J Am Acad Dermatol 2020. 10.1016/j.jaad.2019.10.069. [DOI] [PubMed] [Google Scholar]
  • [7].Garcia M, Juhos S, Larsson M, Olason PI, Martin M, Eisfeldt J, et al. Sarek: A portable workflow for whole-genome sequencing analysis of germline and somatic variants. F1000Research 2020. 10.12688/f1000research.16665.2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [8].McKenna A, Hanna M, Banks E, Sivachenko A, Cibulskis K, Kernytsky A, et al. The genome analysis toolkit: A MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res 2010. 10.1101/gr.107524.110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [9].Mayakonda A, Lin DC, Assenov Y, Plass C, Koeffler HP. Maftools: Efficient and comprehensive analysis of somatic variants in cancer. Genome Res 2018. 10.1101/gr.239244.118. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [10].Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A 2005;102:15545–50. 10.1073/pnas.0506580102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [11].Liberzon A, Subramanian A, Pinchback R, Thorvaldsdóttir H, Tamayo P, Mesirov JP. Molecular signatures database (MSigDB) 3.0. Bioinformatics 2011;27:1739–40. 10.1093/bioinformatics/btr260. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [12].Fotakis G, Rieder Di, Haider M, Trajanoski Z, Finotello F. NeoFuse: Predicting fusion neoantigens from RNA sequencing data. Bioinformatics 2020. 10.1093/bioinformatics/btz879. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [13].Niu B, Ye K, Zhang Q, Lu C, Xie M, McLellan MD, et al. MSIsensor: Microsatellite instability detection using paired tumor-normal sequence data. Bioinformatics 2014. 10.1093/bioinformatics/btt755. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [14].Liao Y, Smyth GK, Shi W. The Subread aligner: Fast, accurate and scalable read mapping by seed-and-vote. Nucleic Acids Res 2013. 10.1093/nar/gkt214. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [15].Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 2014. 10.1186/s13059-014-0550-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [16].Balwierz PJ, Pachkov M, Arnold P, Gruber AJ, Zavolan M, Van Nimwegen E. ISMARA: automated modeling of genomic signals as a democracy of regulatory motifs. Genome Res 2014. 10.1101/gr.169508.113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [17].Hänzelmann S, Castelo R, Guinney J. GSVA: Gene set variation analysis for microarray and RNA-Seq data. BMC Bioinformatics 2013. 10.1186/1471-2105-14-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [18].Wang S, Tao Z, Li H, Wu T, Liu X, Mayakoda A. Copy number signature analyses in prostate cancer reveal distinct etiologies and clinical outcomes. MedRxiv 2020. [Google Scholar]
  • [19].Inman GJ, Wang J, Nagano A, Alexandrov LB, Purdie KJ, Taylor RG, et al. The genomic landscape of cutaneous SCC reveals drivers and a novel azathioprine associated mutational signature. Nat Commun 2018. 10.1038/s41467-018-06027-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [20].South AP, Purdie KJ, Watt SA, Haldenby S, den Breems N, Dimon M, et al. NOTCH1 mutations occur early during cutaneous squamous cell carcinogenesis. J Invest Dermatol 2014;134:2630–8. 10.1038/jid.2014.154. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [21].Li YY, Hanna GJ, Laga AC, Haddad RI, Lorch JH, Hammerman PS. Genomic analysis of metastatic cutaneous squamous cell carcinoma. Clin Cancer Res 2015;21:1447–56. 10.1158/1078-0432.CCR-14-1773. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [22].Pickering CR, Zhou JH, Lee JJ, Drummond JA, Peng SA, Saade RE, et al. Mutational landscape of aggressive cutaneous squamous cell carcinoma. Clin Cancer Res 2014;20:6582–92. 10.1158/1078-0432.CCR-14-1768. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [23].Migden MR, Rischin D, Schmults CD, Guminski A, Hauschild A, Lewis KD, et al. PD-1 blockade with cemiplimab in advanced cutaneous squamous-cell carcinoma. N Engl J Med 2018. 10.1056/NEJMoa1805131. [DOI] [PubMed] [Google Scholar]
  • [24].Rizvi NA, Hellmann MD, Snyder A, Kvistborg P, Makarov V, Havel JJ, et al. Cancer immunology. Mutational landscape determines sensitivity to PD-1 blockade in non-small cell lung cancer. Science 2015;348:124–8. 10.1126/science.aaa1348. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [25].Gray SE, Kay EW, Leader M, Mabruk MJEMF. Enhanced detection of microsatellite instability and mismatch repair gene expression in cutaneous squamous cell carcinomas. Mol Diagn Ther 2006;10:327–34. 10.1007/BF03256208. [DOI] [PubMed] [Google Scholar]
  • [26].Janjigian YY, Maron SB, Chatila WK, Millang B, Chavan SS, Alterman C, et al. First-line pembrolizumab and trastuzumab in HER2-positive oesophageal, gastric, or gastro-oesophageal junction cancer: an open-label, single-arm, phase 2 trial. Lancet Oncol 2020;21:821–31. 10.1016/S1470-2045(20)30169-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [27].Heft Neal ME, Birkeland AC, Bhangale AD, Zhai J, Kulkarni A, Foltin SK, et al. Genetic analysis of sinonasal undifferentiated carcinoma discovers recurrent SWI/SNF alterations and a novel PGAP3-SRPK1 fusion gene. BMC Cancer 2021;21:636. 10.1186/s12885-021-08370-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [28].Dedeurwaerdere F, Claes KB, Van Dorpe J, Rottiers I, Van der Meulen J, Breyne J, et al. Comparison of microsatellite instability detection by immunohistochemistry and molecular techniques in colorectal and endometrial cancer. Sci Rep 2021;11:12880. 10.1038/s41598-021-91974-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [29].Lee CS, Siprashvili Z, Mah A, Bencomo T, Elcavage LE, Che Y, et al. Mutant collagen COL11A1 enhances cancerous invasion. Oncogene 2021;40:6299–307. 10.1038/s41388-021-02013-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [30].Park H, Choi H-J, Kim J, Kim M, Rho S-S, Hwang D, et al. Homeobox D1 regulates angiogenic functions of endothelial cells via integrin β1 expression. Biochem Biophys Res Commun 2011;408:186–92. 10.1016/j.bbrc.2011.04.017. [DOI] [PubMed] [Google Scholar]
  • [31].Russell-Goldman E, Dubuc A, Hanna J. Differential Expression of PLAG1 in Apocrine and Eccrine Cutaneous Mixed Tumors: Evidence for Distinct Molecular Pathogenesis. Am J Dermatopathol 2020;42:251–7. 10.1097/DAD.0000000000001393. [DOI] [PubMed] [Google Scholar]
  • [32].Li F, Wang T, Tang S. SOX14 promotes proliferation and invasion of cervical cancer cells through Wnt/β-catenin pathway. Int J Clin Exp Pathol 2015;8:1698–704. [PMC free article] [PubMed] [Google Scholar]
  • [33].Zhang G, Yan G, Fu Z, Wu Y, Wu F, Zheng Z, et al. Loss of retinoic acid receptor-related receptor alpha (Rorα) promotes the progression of UV-induced cSCC. Cell Death Dis 2021;12:247. 10.1038/s41419-021-03525-x. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1950353-supplement-1.xlsx (11KB, xlsx)
2

Supplemental Figure 1. Previously reported genes with mutations found in cSCC. Ninety percent of our samples had at least one mutation in a gene previously reported to be mutated in cSCC with TP53 being the most common mutation in our cohort.

NIHMS1950353-supplement-2.pdf (56.9KB, pdf)

RESOURCES