UCSD

Anshuman Bhanja, PhD, Postdoc, Gutkind Lab. "In vivo CRISPR Screens to Reveal the Role of the Immune GPCRome in CD8 T Cell Anti-Cancer Immunity" Lack of efficient and long-lasting tumor control in response to immune checkpoint blockade (ICB) therapies. Over the last two decades, ICB therapies using antibodies targeting PD-1, its ligand PD-L1 and CTLA-4 have drastically expanded our capacity to combat cancer by reinvigorating the immune system in the tumor microenvironment (TME), specifically CD8 T cells1,2. However, the efficacy of these treatments is limited to certain tumor types. Furthermore, patients that initially respond to ICB often lack subsequent durable tumor control3. This suggests that there may be additional immune checkpoints that must be overcome to achieve efficient and long-lasting immunity against tumors. These immune checkpoints are likely activated via additional immunosuppressive signals deployed by tumors in the form of secreted molecules that might directly affect CD8 T cells or indirectly affect them through cDC1 dendritic cells that play a pivotal role in priming CD8 T cells leading to tumor-infiltration and tumor control. Some of these tumor-secreted molecules present in the TME include prostaglandin E2 (PGE2), adrenaline/noradrenaline, adenosine and H , which target G protein-coupled receptors (GPCRs), and have been shown to promote CD8 T cell dysfunction and consequently inhibit T cell-mediated tumor control4-6. However, elucidation of key GPCRs involved in T cell dysfunction is currently lacking and will likely help to identify potential targets for therapies, which in combination to ICB, might enhance its efficacy.

Nicole Mattson, PhD, Postdoc, Ideker Lab, and Jeffery Turner, Graduate Student, Fraley Lab. "Mechanisms of Adhesion Driven Cell State Homogenization" Targeted cancer therapeutics highjack mutations that most tumor cells depend on, serving as a highly effective form of precision medicine1,2. However, the ability of tumor cells to take on heterogeneous cell states where their functional dependencies on mutations differ, leads to the development of therapeutic resistance or recurrence in patients3–5. The underlying factors that drive tumor heterogeneity are less understood than how mutations drive cancer progression, thus leading to the clinical need to understand tumor heterogeneity and develop therapeutic strategies. Our overarching hypothesis is that cell state “homogenization” therapies could provide a new adjuvant approach to prevent therapeutic resistance. We define cell state homogenization as a strategy that seeks to push heterogeneous tumor cell populations into a known responsive state and hold them there during targeted therapy. A major challenge to understanding tumor cell state heterogeneity and counteracting its mechanisms is the lack of in-vitro model systems that are able to recapitulate it. Towards addressing this challenge, the Fraley lab, project lead of Project 2 of CCMI, has developed 3D culture models of breast cancer cells that exhibit multiple states and associated growth phenotypes (Fig. 1A)6. These in vitro cell states map to human tumor cell states (Fig. 1B-C) and are predictive of different outcomes in patient survival (Fig. 1C-D). Importantly, the Fraley lab has identified a mechanism driving this adhesion heterogeneity: signaling driven oxidative stress7. This oxidative stress induces adaptive metabolic states that are unique to in-vivo tumor progression and are defined by limitations on heme and iron metabolism8,9. Standard 2D models do not capture these cell states and metabolic dependencies8. It is important to note that this data was presented to the CCMI 2.0 EAC in 2023, and in their report, the committee strongly suggested that the team investigate the role of adhesion proteins in this system. To gain a deeper understanding of how adhesion signaling regulates such heterogeneity, we propose a systems biology approach in which we interrogate the function of adhesion proteins, integrins, on tumor cell states and cellular phenotypes. With an understanding of the mechanistic drivers of tumor cell state heterogeneity, we can develop effective adjuvant therapeutic approaches that homogenize cell state and thereby the functional dependencies of genetic mutations, improving the effectiveness of targeted therapeutics. Our focus on the integrin protein family is rooted in their fundamental role as adhesion proteins on the cell surface. These heterodimer, single pass membrane bound proteins act as the bridge between the external environment and a cell. Integrin heterodimers, 24 total, are traditionally understood to bind to certain extracellular matrix (ECM) proteins (shaded sub-groups Fig. 2)10 and early evidence suggested that the loss of certain integrin heterodimers impaired a cell's ability to bind to growth substrates11–13. This canonical adhesion function extended into cancer where integrins have been linked to invasion and metastasis in multiple cancer types14–16. Stressed cancer cells have also been shown to change their integrin expression17, and under such conditions collagens, to which some integrins bind, can be used as a nutrient source18. The limitation of all these findings, however, is that they are limited to 2D growth environments, thus falling short of a full mechanistic understanding of how an integrin heterodimer contributes to adhesion and metabolism in a physiologically relevant context. Yet given the important role of integrins in cancer it would be imperative to develop effective therapeutic strategies, but to date none are FDA approved19,20. To fill this clinical need there have been major advances in integrin inhibition, yet still lack strong foundational clinical use20. Taken together, integrin targeting is a promising approach to modulate tumor heterogeneity as an adjuvant therapeutic. As a model of tumor heterogeneity and homogenized tumor states, we will focus on studying an invasive breast cancer. Breast cancer is the most common cancer in women in the United States and women have a 1 in 8 chance of developing an invasive breast cancer in their lifetime21. This alarming statistic highlights the need to develop strong therapeutic strategies for breast cancer and in particular invasive breast cancers22,23. The Fraley lab has shown that breast cancer cells can be grown in a 3D system to model the invasive properties in addition to other phenotypes yielding a heterogeneous population6,7. Therefore, utilizing this model is the best suited approach to advance the goals of this proposal. The proposed hypothesis is that loss of key integrins will alter the proportion of heterogeneous cell populations grown in 3D and the transcriptional states of the perturbed cells will lend insight into the mechanisms driving the observed phenotypes. This work will be a collaborative effort between Dr. Nicole Mattson, a post doc of the Ideker group, and Dr. Stephanie Fraley’s lab. For the purpose of this funding opportunity, we are seeking funds to support the application of a CRISPR screen performed in cancer cells grown in 3D collagen. Dr. Mattson’s salary is supported by her own funding (K00CA274649); intended funding will fully support the activities required for the proposed project.

UCSF

Richa Tiwari, PhD, Research Specialist, Krogan Lab. "Integrating Peptide-Tiling Screen and large-scale PPI mapping to forge novel routes in therapeutic intervention against lung squamous cell carcinoma." Non-small-cell lung cancer (NSCLC) represents the largest subgroup of lung cancer accounting for 85% of cases and a significant cause of mortality worldwide. The major subtypes of NSCLC are lung adenocarcinoma (LUAD) (40%) and lung squamous cell carcinoma (LUSC) (30%). LUSC is particularly challenging to treat, due to its highly aggressive nature, late diagnoses, and limited treatment options 1,2. A primary contributing factor to the lack of effective treatments against LUSC is the limited availability of tractable oncogenes paired with scarce information regarding enzymatic active sites or druggable pockets on proteins. To address this issue, we propose shifting the focus to protein-protein interaction (PPI) interfaces, which represent a promising class of drug targets previously considered undruggable. Identifying therapeutically relevant inhibitory peptides targeting interaction interfaces exhibits substantial potential in the field of drug discovery3,4. Prior findings have successfully identified peptides/mini-proteins targeting interaction interfaces including D-peptide inhibitors for P53-MDM2 interaction or structurally engineered omoMYC for cMYC-MAX heterodimerization ultimately counteracting tumor growth, however, the approaches were very protein specific 5,6,7. The current goal of CCMI2.0 is, to generate high-confidence differential PPI maps for the key LUSC drivers and their mutants across the panel of immortalized and LUSC-derived cell lines to identify therapeutically actionable PPIs. While approaches such as affinity-purification mass spectrometry (AP-MS) serve as valuable tools to elucidate PPI networks, these approaches typically lack the resolution needed to understand the specific protein interaction interfaces. The recently established PepTile screening platform serves as a great tool for rapid de novo mapping of bioactive protein domains and associated interfering peptides for an array of proteins simultaneously8. In line with this, pairing PepTile screening with ongoing AP-MS efforts will be extremely purposeful in establishing druggable interaction interfaces and propelling forward the long-term CCMI goal of rapidly elevating translational cancer research. With this in view, the current proposal intends to integrate two powerful state-of-the-art approaches, peptide tiling screen, and PPI mapping approaches for key LUSC drivers, associated pathway genes, and their mutants. This will pave the way to define druggable protein domains and associated inhibitory dominant negative peptides targeting therapeutically relevant oncogenic PPI interfaces. Notably, peptides originating from mutant proteins might harbor differential inhibitory potential. Finally, incorporating cross-link coupled AP-MS (XL-MS) and AlphaFold2guided PPI mapping for novel drug-like peptide inhibitors will provide better insights into disrupted interactions and their therapeutic significance with precise mechanisms. Moreover, short-listed druglike molecules once validated in vitro and in LUSC-derived murine models will open new avenues in the LUSC drug discovery front.

Weicheng Li, PhD, Postdoc, Ostrem Lab. "Understanding the mechanism of small molecule ubiquitination of BRD1732" I will use novel photo-affinity labeling (PAL) proteomics to identify the interactome of small molecule BRD1732, discovering the protein machineries for producing ubiquitin-BRD1732 conjugate, and will perform a series of experiments to validate the results, and reveal the molecular basis of small molecule ubiquitination of BRD1732. Post-translational modification (PTM) on small molecules: My previous postdoc work has led to the identification of a small molecule BRD1732, which could be directly ubiquitinated under the E3 ligases RNF19, resulting in the accumulation of unanchored poly-ubiquitin chains attached on BRD1732 and general inhibition of proteasome (Figure 1A,B)(1). Our finding of small molecule ubiquitination represents the very first case of PTM on a small molecule, suggesting many future possibilities, such as new chemical tools and new therapeutics. The detailed molecular understanding of BRD1732 ubiquitination could reveal how this small molecule imitates a protein to receive ubiquitination, inspiring people to design novel chemical probes to other PTMs and beyond. This information also benefits on designing novel E3 ligase binder towards RNF19, and developing novel degradation modality to directly install the ubiquitin, the degradation marker, to the cancer-driving proteins for proteasomal degradation through BRD1732-based bifunctional molecules.

Kevin Lou, Medical student, Shokat Lab. "A chemical strategy to target oncogenic protein-protein complexes of HER3" Large-scale mapping of protein-protein interaction (PPI) networks has shown how different cellular contexts, such as tissue of origin and genetic mutations, lead to the presence of unique protein complexes in cancer cells.1 These PPI networks offer a promising avenue for therapeutic development,2–8 complementing traditional approaches that target oncogenic drivers identified at the genetic level. By recognizing and targeting the differential co-localization of two proteins, complex-specific inhibitors could potentially provide a broader therapeutic window9 compared to compounds that rely solely on the differential essentiality of a single protein, which is the classic paradigm in targeted therapy. My previous work uncovered a novel mechanism by which linked chemotypes—composed of two drug-like small molecules ligands connected by a flexible linker—gain access to the cytosol.10 These bivalent molecules possess a pharmacologic logic, recognizing the binding sites for ligand A and ligand B in close proximity, which can potentially be applied to recognize cancer-associated protein complexes. The proposed study aims to develop bivalent molecules that target the oncogenic association of HER3 with known binding partners (e.g., PIK3CA, HER2, EGFR) and inhibit their signaling within distinct PPI network contexts such as head and neck squamous cell carcinoma (HNSCC)7 and breast adenocarcinoma8. If initial studies in wellcharacterized cell lines demonstrate selective PPI targeting, we can follow up with a HER3-focused affinity purificationmass spectrometry (AP-MS) profiling across a broad panel of cancer cell types. This will help identify additional therapeutic opportunities where our strategy could apply and provide a deeper understanding of other oncogenic signaling networks involving HER3. This work is significant as it will establish the feasibility of targeting cancer-associated protein complexes using bivalent small molecules, a novel modality that could harness the therapeutic potential of large-scale PPI networks generated by the Cancer Cell Map Initiative (CCMI). Furthermore, despite HER3’s well-validated role as an oncogenic signaling linchpin, there are no FDA-approved HER3-directed therapies, and those in clinical trials are exclusively antibody-based, failing to address the critical issue of targeting complexes with intracellular binding partners.1