Melanoma remains one of the most aggressive cancers, with metastatic disease proving frequently fatal despite recent therapeutic advances. The Karreth laboratory focuses on understanding the molecular mechanisms underlying melanoma development, with particular emphasis on the late stages of progression and metastasis that remain poorly understood.
While early genetic alterations in melanoma—such as mutations in BRAF, NRAS, PTEN, and CDKN2A—are well characterized, the drivers of advanced disease are largely elusive. Importantly, work over the past decade has revealed that non-mutational mechanisms are the key drivers of melanoma progression and metastasis, enhancing the plasticity and adaptability of cancer cells. Our research integrates mouse modeling, functional genomics, and mechanistic studies to identify these novel non-mutational driver alterations and characterize how they promote melanoma development. Additionally, we develop sophisticated mouse modeling platforms to enable better understanding of melanoma formation in vivo, including rare melanoma subtypes.
Model Development
ESC-GEMM Platform: Accelerating Melanoma Research
We have developed a sophisticated embryonic stem cell-genetically engineered mouse modeling (ESC-GEMM) platform for melanoma research (Bok et al., Cancer Research 2020). This rapid and versatile platform consists of a comprehensive portfolio of driver gene combinations that enables selection of biologically and clinically relevant mutational backgrounds. The key advantage: we can generate multi-allelic experimental mice in 2-3 months without the need for breeding or genotyping, vastly accelerating our in vivo studies. From these models, we derived and characterized a panel of murine melanoma cell lines, including NRAS mutant lines, that are useful for syngeneic transplants into immune-competent recipients (Bok et al., JID Innovations 2021). The melanoma ESC-GEMM platform forms the cornerstone of many of our in vivo studies.
Building upon this platform, we developed a method to inducibly express ectopic circular RNAs (circRNAs) in mouse models and cell lines using a transposon delivery approach (Mecozzi et al., RNA Biology 2022). Our current model development efforts include novel approaches that enable: 1) inducible silencing of genes critical for tumor maintenance without the emergence of resistant cells, and 2) tumor suppressor knockout after oncogene activation to better recapitulate the sequence of genetic events observed in human melanoma.
Modeling Rare Melanoma Subtypes
We are pioneering mouse models of rare melanoma subtypes that have been historically difficult to study. We created a multi-step, immune-competent uveal melanoma mouse model that faithfully recapitulates the genetics, biology, and progression of human disease (Xu et al., bioRxiv 2025). This model identifies MYC as a potent driver of uveal melanoma formation and reveals phenotypic diversity associated with malignant progression. The model exhibits disseminated uveal melanoma cells to the liver, and we have established cell lines that form metastases with varying organ preferences. Using this system, we are investigating factors that control cellular states, drive liver metastasis, and influence metastatic dormancy.
Non-coding RNAs in Melanoma Progression
From ceRNAs to Antisense RNAs: Expanding the Non-coding Landscape
Our lab has a longstanding focus on non-coding RNAs in melanoma progression. Despite their critical roles as non-mutational drivers throughout melanoma development, these molecules remain understudied. Building upon foundational work showing that PTEN is regulated by competitive endogenous RNAs (ceRNAs) (Karreth et al., Cell 2011) and that the BRAF pseudogene enhances BRAF expression as a ceRNA (Karreth et al., Cell 2015), we have continued to elucidate ceRNA and miRNA functions in melanoma.
We identified several ceRNAs encoded by genes on chromosome 1q, which is frequently amplified in melanoma. These chromosome 1q ceRNAs promote melanoma metastasis by sequestering tumor suppressive microRNAs, thereby allowing expression of genes that drive metastatic behavior (Xu et al., Cancer Research 2022). We also investigated the complex regulation of miR-29 by MAPK pathway signaling, revealing a tumor suppression checkpoint that is overcome by partial miR-29 downregulation (Vera et al., Cancers 2021).
Novel RNA Species with Roles in Melanoma
Our current studies focus on antisense RNAs (asRNAs) and circular RNAs (circRNAs). We identified the antisense RNA CDH3-AS1 as strongly downregulated in melanoma. Notably, CDH3-AS1 functions as a SINEUP RNA—a class of regulatory RNAs that enhance translation of their target mRNAs. CDH3-AS1 promotes translation of CDH3, which encodes P-cadherin, a protein with tumor suppressive effects in melanoma (Chadourne et al., bioRxiv 2024). We are also characterizing a circRNA that promotes melanoma metastasis by influencing cytoskeletal rearrangements.
Pioneering RNA Glycosylation Research
We have become interested in an exciting new RNA modification: RNA glycosylation. Short, non-coding RNAs modified with sugar molecules (glycoRNAs) localize to the cell surface and influence cellular interactions with the microenvironment. We developed an innovative imaging mass spectrometry method to detect glycans on isolated RNA andin tissue sections in an unbiased manner. Using this approach, we have profiled glycoRNA changes during melanoma progression and are now investigating the role of surface RNA glycans and specific glycoRNAs in cancer development.
Signaling and Transcription Factors in Melanoma Progression
Interconnected Networks: From Signaling to Transcription
Several non-mutational mechanisms drive melanoma progression, including aberrant signaling pathway activation and transcription factor dysfunction. These mechanisms are often interconnected. We showed that the tumor suppressor PTEN, through its lipid phosphatase activity, opposes a signaling pathway involving AKT and mTORC1 that controls translation of the FRA1 transcription factor (Xu & Bok et al., Cancer Research 2024).
In follow-up studies, we found that FRA1 is a potent driver of melanoma metastasis, implicating this signaling axis in late-stage disease. FRA1 controls a transcriptional network that can be targeted with inhibitors to reduce metastatic burden (Xu et al., bioRxiv 2025). We also demonstrated that mutant BRAF activates mTORC1 to promote ATF4 translation, which transcriptionally induces PHGDH expression—the rate-limiting enzyme in serine synthesis. This creates a metabolic vulnerability where combining BRAF inhibition with dietary serine restriction reduces melanoma viability and tumor growth (Jasani et al., Cancer Research 2025).
Transcription Factors as Potent Drivers of Melanomagenesis
Our investigation of pro-tumorigenic transcription factors extends to MAFG and MYBL2, both regulated by miR-29 and frequently overexpressed in melanoma through multiple mechanisms. MAFG potently promotes melanoma formation through inducing a phenotypic switch mediated by direct interaction with MITF, altering its transcriptional output (Vera et al., bioRxiv 2024). Importantly, MAFG represents a critical vulnerability—its deletion completely prevents the transition from benign nevi to melanoma in mouse models.
MYBL2 also promotes melanoma growth, but through an unexpected mechanism. Rather than functioning as a traditional transcription factor, MYBL2 binds to RNA in the nucleus, which enhances its chromatin association. This promotes chromatin binding of other pro-tumorigenic transcription factors, including MAFG and FRA1, activating transcriptional programs associated with poor patient survival.
Future Directions
Over the coming years, we will further elucidate the molecular mechanisms underlying these oncogenic effects and devise strategies to target them therapeutically. Our integrated approach—combining cutting-edge mouse models, functional genomics, and mechanistic studies—positions us to make significant contributions to understanding melanoma biology and developing new therapeutic strategies for this challenging disease.