Cancer is a shapeshifter. It continuously changes and adapts over time and in response to new therapies. To gain the upper hand, we must be able to: 1. Track and analyze cancer in real time in our patients; and 2. Elucidate and block plasticity mechanisms that allow tumor cells to change and evade treatment. To address this challenge, our research activities have come to focus on two main areas: liquid biopsy and cancer plasticity. Click on the area below to learn more:
Most solid malignancies (e.g., prostate, lung, breast, colon, and others) typically present with a primary tumor that is biopsied to obtain histological and molecular information about the cancer. However, once cancer recurs and spreads, it is no longer feasible to biopsy multiple metastatic sites invasively and repeatedly over time. Consequently, oncologists have no up-to-date information to guide treatment decisions as tumors spread and become resistant over time. One solution to this problem is to capture and analyze DNA, RNA, and cells shed by tumors into the bloodstream. A blood sample can serve as a “liquid biopsy” containing circulating tumor cells (CTCs), circulating tumor DNA (ctDNA), and circulating tumor RNA (ctRNA). Liquid biopsies are noninvasive and can be repeated over time to track emergence of resistance, predict clinical outcomes, and optimize treatment strategies. However, as CTCs, ctDNA, and ctRNA circulate in minute quantities, their capture and analysis present major technical challenges.
Our efforts have focused on developing new liquid biopsy technologies and workflows and implementing them to discover clinically relevant biomarkers in prostate and bladder cancer. In our first NCI-funded study, we used the FDA-cleared CellSearch assay to analyze CTC numbers in a phase III NCI-sponsored North American Cooperative Group trial for men with metastatic castration resistant prostate cancer. We found that high numbers of CTCs predicted 50% shorter survival, the first such finding in a large cooperative group setting (Lancet Oncology 2013, Journal of Clinical Oncology 2014). In a second NCI-funded study, we showed for the first time that CTC counts also predict disease response and progression on hormonal therapy in men with castration sensitive prostate cancer, the first such finding in a large cooperative group phase III trial (Clinical Cancer Research 2021, Journal of Clinical Oncology 2022). Collectively, these studies in metastatic prostate cancer clinically validated CTC count as a new blood-based predictor of treatment response, progression, and survival, a biomarker which can help guide the choice of therapy at the outset of treatment.
Beyond established assays, we have pioneered novel technologies and workflows that enable new information to be gleaned from liquid biopsies. With collaborators at Caltech, we co-developed and patented a new parylene-C slot microfilter for capture of live CTCs (Cancer Research 2011) and used it in a phase III NCI-sponsored clinical trial to show that high CTC telomerase activity portents shorter survival (International Journal of Cancer 2014). More recently, we developed a technique that enriched live CTCs from background white blood cells by five orders of magnitude. This approach achieved sufficiently high CTC purity for downstream transcriptional profiling by RNA-seq, revealing cancer-specific actionable pathways (International Journal of Cancer 2020). Using these methods, gene expression programs that drive aggressive disease can be identified and targeted in patients with metastatic cancer. Beyond CTCs, we also developed novel methods to analyze ctDNA in the blood, for example measuring the patterns of plasma ctDNA methylation, a marker of gene silencing. Using samples collected in a national NCI-sponsored clinical trial in muscle-invasive bladder cancer (Clinical Cancer Research 2021), we measured ctDNA methylation patterns and applied machine learning to generate biomarker signatures predictive of response to chemotherapy. Such predictive signatures can help identify patients likely to benefit from neoadjuvant chemotherapy vs. those who should proceed straight to surgery (Oral presentation, ASCO 2022, European Urology Oncology 2023).
Combining multiple liquid biopsy assays can create synergies and yield more comprehensive and informative tumor profiles. We pioneered this multi-parametric approach in prostate cancer by concurrently analyzing CTC counts, CTC DNA alterations, ctDNA alterations, and ctRNA expression to generate complementary molecular profiles from a single tube of blood (JCI Insight 2019). Building on this approach, we recently used machine learning to analyze single CTCs, plasma ctDNA, and tumor images from CT scans, revealing novel correlations and signatures (International Journal of Molecular Science 2022). We are currently leveraging these capabilities in a new NCI-funded study, implementing multi-parametric liquid biopsy and radiomics to track disease progression and predict outcomes in SWOG S1802, an NCI cooperative group phase III clinical trial in men with metastatic castrate sensitive prostate cancer.
Ongoing Studies: Currently, we are conducting the new R01-funded S1802 study described above, which will clinically validate the multiparametric liquid biopsy methods we developed and identify new candidate composite biomarkers in metastatic prostate cancer. We are also further developing cfDNA methylation, single cell highly multiplexed protein analysis, and other novel liquid biopsy techniques. These will be implemented in a large new prostate cancer initiative that Dr. Goldkorn leads in the Norris Cancer Center – the Longitudinal Advanced Prostate Cancer Cohort (LAPCC) – which will track and analyze disease progression in men treated at our GU oncology clinics.
Although systemic therapies for cancer are effective at first, resistance frequently develops, leading to disease progression. The traditional model of resistance is genetic: Under selection pressure from a given drug, rare DNA mutations provide a survival advantage to a few tumor cells, allowing them to survive and expand clonally. However, recent work suggests that more rapid, non-genetic mechanisms are also at play. Cancer cells exhibit phenotypic plasticity, the capacity to shift in and out of a drug resistant state. This phenomenon, sometimes termed “bet-hedging”, allows the overall tumor cell population to survive and persist in the face of environmental stressors, much as infectious pathogens can adapt to antibiotic treatment.
We were among the first groups to report this phenomenon. In early work, we used flow cytometry in conjunction with Hoechst dye exclusion to sort cancer cell lines into two subpopulations: a “side-population” (SP) and a “non-side population” (NSP). Although these cell populations were genetically identical, SP cells expressed high levels of stemness genes and were more tumorigenic and drug resistant than NSP cells. Importantly, we demonstrated that cells could rapidly interconvert between these phenotypes (Molecular Cancer Therapeutics 2011), and that this plasticity involved key cancer and stemness signaling mediators like AKT and B-catenin (International Journal of Cancer 2013, Prostate 2016). Recent developments in cancer epigenetics prompted us explore the role of chromatin states in cancer plasticity. We found that when bladder cancer cells shifted to a drug-resistant SP phenotype, they exhibited dramatically altered DNA methylation, nucleosome positioning, and chromatin accessibility at gene promoters, allowing transcription factors like E2F3 to bind and drive resistance to cisplatin chemotherapy (International Journal of Cancer 2020). We found similar shifts in chromatin accessibility at enhancer elements; in particular, we discovered that cisplatin-treated bladder cancer cells overexpress the transcription factor FOXC1, which binds accessible enhancers and drives cisplatin resistance (Cancers 2022).
During these plasticity studies in bladder cancer, we used RNA-seq to profile differential gene expression between the drug-resistant SP cells and drug-sensitive NSP cells. We were intrigued to discover that the most highly upregulated genes in cells that shift to a drug-resistant SP phenotype were associated with mitochondrial oxidative phosphorylation (OxPhos), a finding that we confirmed with metabolomic analysis. We leveraged this metabolic shift to track cancer cell plasticity in real time using fluorescence lifetime microscopy (FLIM), a technique advanced by Dr. Scott Fraser at USC that images the redox state of single cells. Using FLIM, we reported, for the first time, bladder cancer cells shifting rapidly and spontaneously – within minutes to hours – to an OxPhos drug-resistant state (Scientific Reports 2022).
Collectively, our work in this field has shown that bladder cancer cells exhibit non-genetic plasticity: rapid adaptive transition to a chemotherapy resistant phenotype involving broad networks of epigenetic, transcriptional, and metabolic signals. We identified key mediators of these transitions (e.g., E2F3, FOXC1, mitochondrial OxPhos), which may constitute new therapeutic targets to forestall chemotherapy resistance.
Ongoing Studies: With support from a recent NIH F30 grant and a foundation grant, graduate students in our group have been generating exciting new data focused on m6A epitranscriptomic regulation and cancer plasticity. We are expanding these studies with additional functional assays in patient-derived models, including new bladder cancer patient-derived organoids (PDOs) recently established in our lab. These new data and models – along with our recent publications – will form the basis for expanded bladder cancer plasticity studies in the coming year. Ultimately, our goals will be to further elucidate plasticity mechanisms and to identify potent therapeutic inhibitors that block cancer’s escape routes.