Abstract PR08: Noninvasive epidermal sampling as a means of genomic UV dosimetry

Abstract

Abstract It is now well accepted that normal-appearing, but chronically UV-exposed skin, can harbor tremendous mutational loads while maintaining homeostatic function. Even so, clones of keratinocytes harboring tumorigenic mutations can be readily observed using targeted and whole-exome sequencing. Therefore, it may be possible to objectively measure chronic, cumulative UV exposure as a means of predicting skin cancer risk and measuring molecular responses to chemoprevention by identifying UV-driven mutations or damage in epidermal keratinocytes. However, such a methodology requires the ability to noninvasively sample normal, nonlesional skin. We have previously optimized a novel surfactant mixture that, when coupled to ultrasound or mechanical abrasion, can noninvasively sample epidermis without scarring. This process can be tuned to selectively remove epidermis only. Within hours of application, a fibrinous exudate appears over the wound and this area heals completely within 2 weeks without scarring. Using a Hairless UV-driven model of squamous cell carcinoma, we have been able to sample and perform both exome and RNA sequencing with recovery rates of about 2.5-3.5 μg /cm2 of DNA and RNA. Using this model, we demonstrated a dose-dependent increase in mutational load in normal UV-exposed skin that reflects what is seen in humans. Mice were sampled and the DNA exome was sequenced (150X) following exposure to UV. Among the 28 most mutated genes across the exome, mice sampled after 1 month of UV exposure harbored variants in an average of 12 genes. After 3 months of UV exposure, followed by 2 months of no exposure, variants were observed in an average of 22 genes (80%) of the 28, suggesting continuing evolution of clones harboring mutations in exposed epidermis. Interestingly, at this relatively low coverage, none of the genes is strongly associated with skin cancer; instead, many are not expressed, reflecting the expansion of mutational burdens without strong selection on these alterations. We then extended our efforts into humans by sampling immunosuppressed solid organ transplant recipients with markedly elevated risk for skin cancer. Sun-exposed areas had an average of 4.5 times the number of variants than sun-protected areas across targeted exome sequencing of 348 cancer-related genes at a depth of ~5000X. Over 24 genes (including TP53, ATM, NOTCH, and KMT2 family members) were identified with nonsense C-T transitions, suggesting the presence of detectable (and potentially selected) mutational patterns in cancer-related genes that correlate with clinically relevant exposure. Our results collectively suggest that noninvasive sampling and genomic interrogation of normal-appearing epidermis is feasible and potentially useful as a risk prediction and chemoprevention assessment tool. This approach is easily extended to other surface epithelia, and we are currently enhancing the sensitivity of this assay based on identification of hotspots and nontranscribed genes as identified in our model. This abstract is also being presented as Poster A32. Citation Format: Kenneth Y. Tsai, Elnara Muradova, Sandy S. Ojeda, Charles H. Adelmann, Kimberly Nguyen, Samir Mitragotri, Nishit Patel. Noninvasive epidermal sampling as a means of genomic UV dosimetry [abstract]. In: Proceedings of the AACR Special Conference on Environmental Carcinogenesis: Potential Pathway to Cancer Prevention; 2019 Jun 22-24; Charlotte, NC. Philadelphia (PA): AACR; Can Prev Res 2020;13(7 Suppl): Abstract nr PR08.