UC Berkeley

Our understanding of immune cell crosstalk in the tumor microenvironment (TME) has evolved significantly in recent years, revealing a complex network of interactions beyond the classical CD4+ T cell help to other adaptive immune cells such CD8+ cytotoxic T cells, B cells and macrophages. The TME is a highly dynamic and immunologically rich landscape, where both innate and adaptive immune cells, such as natural killer (NK) cells and CD4+ T cells, engage in complex, reciprocal signaling that shapes therapeutic outcomes. NK cells, as key components of the innate immune system, can detect and eliminate tumor cells that evade cytotoxic T lymphocytes by downregulating MHC class I molecules. Meanwhile, CD4+ T cells serve as potent cytokine producers and have emerged as critical amplifiers of antitumor immunity, not only by supporting CD8+ T cell function but also through direct effector activity and coordination of broader immune responses.Recent advances in immunotherapies such as checkpoint immunotherapies that are predicated on amplifying CD8+ T cell responses have led to vastly improved patient outcomes. While CD8+ T cells have long been the cornerstone of cancer immunotherapy, some tumors lack neoantigens and others acquire defects in antigen presentation or MHC I expression either before therapy or as a result of it, acquiring resistance to CD8+ T cell-dependent therapies. There is therefore a need for alternative regimens that can overcome tumor immune resistance. NK cells and CD4+ T cells represent two emerging immune cell populations that are favorable targets for novel and improved immunotherapy, as NK cells and CD4+ T cells have the potential to control MHC I-deficient tumors that can resist CD8+ T cell immune responses.Regulatory T cells (Tregs) play a pivotal role in maintaining immune homeostasis by suppressing autoimmune responses mediated by conventional T cells. Systemic depletion of Tregs is known to unleash lethal autoimmune responses. On the other hand, Tregs also suppress beneficial antitumor responses mediated by CD8+ T cells. Continuing Moreno et al.’s work on investigating CXCR3+ Tregs in suppressing cDC1 cross-presentation and CD8+ T cell control of cancer, I proposed to explore the hypothesis that Tregs also suppress antitumor immune responses mediated by NK cells and CD4+ T cells. If true, this raises the possibility that selective Treg depletion may provide an approach to generate responses against MHC I-deficient tumors, which are resistant to CD8+ T cell recognition and killing.The work I describe in this dissertation in Chapter 3 demonstrates that both systemic and intra-tumoral depletion of Tregs leads to effective control of B2m–/–, and thus MHC I-deficient, colon carcinoma, melanoma and lymphoma tumors implanted into mice. The findings reveal that Tregs suppress anti-tumor responses independent of CD8 T cell responses. Specifically, depletion of NK cells or CD4+ T cells restored tumor growth in Treg-depleted settings, while CD8+ T cell depletion had no effect. This indicates that NK and CD4+ T cells are the primary effectors responsible for tumor control under these conditions. Furthermore, NK cell and CD4+ T cells were found to act synergistically to mediate the clearance of MHC I-deficient tumors. Importantly, intra-tumoral—rather than systemic—Treg depletion elicited potent antitumor effects mediated by NK cells and CD4+ T cells without inducing significant autoimmunity. In Chapter 3, I detail how intra-tumoral depletion of Tregs unleashes anti-tumor function of tumor-infiltrating NK cells. I characterize this enhanced activity by documenting changes in NK cell surface marker expression, cytokine production, degranulation, and cytotoxic capacity. I further demonstrate that CD4+ T cell help is essential for optimal NK cell-mediated tumor control, and investigate the underlying mechanisms driving this interaction. My results identify a critical role for conventional dendritic cells (cDCs) in orchestrating the crosstalk between Tregs, CD4+ T cells, and NK cells within the tumor microenvironment via a key cytokine—IL-2. To dissect and validate these mechanisms of cellular crosstalk, I employ physiologically relevant tumor models and integrate high-resolution techniques, including immune cell sorting, transcriptomic profiling, and imaging. My approach combines in vivo tumor modeling, ex vivo functional assays, and spatial analysis to map the timing, location, and nature of NK–CD4–Treg interactions in the context of MHC I-deficient tumors. Collectively, this work offers new insights into alternative immune effector pathways in tumor control and provides a framework for enhancing cancer immunotherapy, particularly in tumors that escape CD8+ T cell recognition.From the experiments targeted at understanding NK cell-mediated tumor control following intra-tumoral Treg depletion, I identified that beyond providing essential help for robust NK cell tumor control, CD4+ T cells in parallel were also able to elicit tumor control independent of NK cells. In Chapter 4, I systematically characterize the CD4+ T effector cell response after intra-tumoral Treg depletion, and study the mechanism by which the CD4+ T cells contribute to anti-tumor immunity. We used a combination of bulk RNA sequencing data supplemented with spectral flow analyses, direct ex vivo killing assays, and the expression different activation markers and cytokine of these CD4+ cytotoxic T cells to careful phenotype these cells. I also noted that these CD4+ T cells upregulated expression of canonical NK cell receptors. I examined whether the cytotoxic CD4+ T cell population is induced locally within the tumor microenvironment, or if a more systemic response is also observed following Treg depletion. I also generated evidence to suggest that tumor cell killing occurs independently of MHC II-recognition or death receptor-engagement on tumor cells, but is highly dependent on granzyme and perforin activity. Taken together, these data uncover a potent population of cytotoxic CD4+ T cells which have high anti-tumor potential, further expanding on the immune arsenal for cancer immunotherapy.While it is widely accepted that Treg depletion enhances conventional T cell effector functions by removing suppressive constraints, emerging evidence suggests a more complex picture in inflamed tissues (e.g. autoimmune sites or cancer). Specifically, Tregs that have lost Foxp3 expression, termed “ex-Tregs”, may themselves contribute to the effector immune response under these conditions. My work in Chapter 5 tests the hypothesis that ex-Tregs play a central role in shaping the immune response following Treg depletion both in autoimmunity and tumor contexts. Demonstrating that the inflammatory effector responses originate primarily from ex-Tregs will significantly shift our understanding of the immunological consequence of Treg depletion. Such insights could inform the development of novel therapeutic strategies to selectively target ex-Treg function to modulate immune responses in diseases. To investigate this, we are currently optimizing two new genetic mouse models to study ex-Tregs that will allow, for the first time, ex-Treg ablation in vivo and ex-Treg tracing in situ by fluorescence microscopy.To explore immune modulation in clinically relevant settings, I also employed and further developed multiple spontaneous models of cancer using genetically engineered mouse models (GEMMs) harboring K-ras and Tp53 mutations (KP), including sarcomas, lung adenocarcinomas and glioblastomas. These models more accurately reflect the complexity and immune resistance of human cancers. In combination with CRISPR-Cas9–mediated gene editing—such as B2m–/– to model MHC class I deficiency—these systems provide a robust platform to investigate mechanisms of immune evasion and response to therapy. In Chapter 6, I will detail a series of experiments setting up and optimizing these KP-Cas9 models.Finally, in Chapter 7, I will present preliminary findings on a novel pathway of tumor control uncovered through the use of tetracycline derivatives 9-TB (9-tert-butyl doxycycline) and AD (7-ethylphenyl 4’-methoxysancycline). These compounds have been chemically modified to eliminate antimicrobial activity while preserving their ability to inhibit mitochondrial protein translation and remain well tolerated in vivo. Treatment with 9-TB and AD activates a mitochondrial unfolded protein response (UPRmt), a stress response pathway that remains poorly understood in comparison to the more extensively characterized endoplasmic reticulum UPR (UPRer). In collaboration with Dr. Andrew Dillin, a leading expert on mitochondrial stress responses, as well as postdoctoral fellows in the Raulet lab (Hannah Ghasemi and Harrison Sudholz), this project seeks to elucidate the molecular mechanisms driving this UPRmt-mediated anti-tumor effect and to identify key regulators within this pathway that may be exploited for therapeutic benefit.Collectively, the studies in this thesis uncover previously unrecognized mechanisms of immune cooperation within the tumor microenvironment and expand our understanding of how tumor control can occur independently of classical CD8+ T cell responses. Through targeted depletion of regulatory T cells (Tregs), I demonstrate that both NK cells and CD4+ T cells can act as potent effectors capable of mediating robust antitumor immunity. I show that intra-tumoral Treg depletion restores NK cell cytotoxicity through IL-2–dependent crosstalk with CD4+ T cells and dendritic cells. A population of cytotoxic CD4+ T cells capable of directly eliminating MHC I–deficient tumors via granzyme–perforin activity also arises after IT Treg ablation, revealing an alternative effector pathway with therapeutic promise. Further, ex-Tregs have the potential to influence both immune activation and tissue inflammation. With tools such as genetically engineered mouse models of cancer combined with CRISPR editing platforms, we can now test interventions such as localized Treg depletion, cytokine delivery, cyclic dinucleotides, and other therapies with more physiologically relevant tumor models. Finally, I describe a novel line of investigation demonstrating that non-antimicrobial tetracycline derivatives activate the mitochondrial unfolded protein response (UPRmt), revealing a new link between mitochondrial stress and antitumor immunity. Together, these findings delineate a broader and more flexible immune network that can be harnessed to target tumors resistant to CD8+ T cell recognition. By integrating mechanistic studies with translationally relevant models, this work establishes a foundation for therapies that mobilize NK and CD4+ T cell immunity, modulate Treg and ex-Treg dynamics, and exploit stress-response pathways to achieve durable tumor control.