Annual Review of Cancer Biology - Volume 7, 2023

Ultrahigh–dose rate FLASH radiotherapy (FLASH-RT) is a potentially paradigm-shifting treatment modality that holds the promise of expanding the therapeutic index for nearly any cancer. At the heart of this exciting technology comes the capability to ameliorate major normal tissue complications without compromising the efficacy of tumor killing. This combination of benefits has now been termed the FLASH effect and relies on an in vivo validation to rigorously demonstrate the absence of normal tissue toxicity. The FLASH effect occurs when the overall irradiation time is extremely short (<500 ms), and in this review we attempt to understand how FLASH-RT can kill tumors but spare normal tissues—likely the single most pressing question confronting the field today.

Chimeric antigen receptor (CAR) T cell therapy has emerged as a new opportunity for cancer treatment; however, resistance can occur due to intrinsic (T cells), extrinsic (tumors), or acquired (tumors) factors. In many cases, the knowledge of these mechanisms comes from clinical observations of patients treated with CAR T cells. In addition, the structure of the CAR molecule and the manufacturing process can impact CAR T cell efficacy. Extrinsic factors such as the mutations in the tumor cell, or cells in the tumor microenvironment, can also play a role. Tumor cells may exhibit acquired antigen loss or heterogeneity that enables resistance to CAR T cell killing; additionally, myeloid cells, T regulatory cells, and fibroblasts can exert an immunosuppressive effect and abrogate CAR T cell antitumor efficacy. We will discuss these mechanisms of resistance and the novel approaches being used to overcome them to improve the widespread use of this promising cancer therapy.

Cancer-associated fibroblasts (CAFs) are present in all malignancies. Arguably, in none are they as prevalent as they are in pancreatic ductal adenocarcinoma (PDAC), where they often outnumber cancer cells. The origin and function of CAFs are still not completely understood, and attempts to target this cell population as a component of combination therapy have so far not succeeded. Our understanding of pancreatic CAFs is in rapid evolution. Heterogeneity of CAFs is the key concept to understand this cell population. We discuss heterogeneity of origin, with recent findings challenging the notion that CAFs uniformly derive from pancreatic stellate cells, and instead suggesting that multiple types of resident fibroblasts contribute to CAF expansion. Heterogeneity in gene expression divides CAFs in different subpopulations. Most importantly, heterogeneity in function underlies the complexity of CAFs. CAFs deposit components of the extracellular matrix, contributing to the high interstitial pressure in pancreatic cancer. CAFs serve as “feeder” cells for cancer cells by providing metabolites, thus mitigating the effect of the low-nutrient environment of PDAC. At the same time, CAFs regulate the function of the immune system, inhibiting antitumor immune responses. Understanding the functional role of different CAF populations and the drivers of each of their functional roles is key to devising new ways to target this cell population in PDAC.

Histopathology plays a fundamental role in the diagnosis and subtyping of solid tumors and has become a cornerstone of modern precision oncology. Histopathological evaluation is typically performed manually by expert pathologists due to the complexity of visual data. However, in the last ten years, new artificial intelligence (AI) methods have made it possible to train computers to perform visual tasks with high performance, reaching similar levels as experts in some applications. In cancer histopathology, these AI tools could help automate repetitive tasks, making more efficient use of pathologists’ time. In research studies, AI methods have been shown to have an astounding ability to predict genetic alterations and identify prognostic and predictive biomarkers directly from routine tissue slides. Here, we give an overview of these recent applications of AI in computational pathology, focusing on new tools for cancer research that could be pivotal in identifying clinical biomarkers for better treatment decisions.

Fusion oncoproteins (FOs) resulting from in-frame chromosomal translocations are associated with many aggressive cancers with poor patient outcomes. Several FOs are now understood to perform their oncogenic functions within biomolecular condensates formed through liquid-liquid phase separation (LLPS). Two classes of phase-separating FOs have emerged, those that form nuclear condensates and alter chromatin biology, including gene expression, and those that form cytoplasmic condensates and promote aberrant signaling, including RAS/MAPK signaling. The amino acid sequences of the FOs within these classes display LLPS-prone intrinsically disordered regions and folded domains that synergistically interact with themselves and other biomolecules to promote condensate formation. This review summarizes the roles of LLPS in the oncogenic functions of these two FO classes, provides examples of FOs that inhibit physiological LLPS in normal cells, and discusses the sequence features commonly associated with LLPS and their enrichment in many FOs.

With the significant successes of immune checkpoint blockade and adoptive cellular therapy, immunotherapy has now become an established treatment option to effectively treat cancer. However, the full potential of this treatment modality has yet to be realized, as there are many additional mechanisms whereby tumors continue to evade immune destruction. To this end, metabolic reprogramming by cancer cells serves not only to promote their own growth but also to create an immunosuppressive tumor microenvironment. The tumor metabolic microenvironment not only inhibits antitumor effector function but also supports the differentiation and function of suppressive immune cells. In this review, we delineate the major metabolic programs of cancer cells and immune cells. Furthermore, we discuss the role of so-called metabolic checkpoints that promote immune evasion and tumor growth. Finally, we review current and potential future strategies to target metabolism in order to not simply inhibit tumor growth but also enhance antitumor immune responses. Such strategies have the great potential to enhance the breadth and depth of immunotherapy for cancer by targeting metabolic checkpoints.

During tumor evolution, cancer cells can acquire the ability to proliferate, invade neighboring tissues, evade the immune system, and spread systemically. Tracking this process remains challenging, as many key events occur stochastically and over long times, which could be addressed by studying the phylogenetic relationships among cancer cells. Several lineage tracing approaches have been developed and employed in many tumor models and contexts, providing critical insights into tumor evolution. Recent advances in single-cell lineage tracing have greatly expanded the resolution, scale, and readout of lineage tracing toolkits. In this review, we provide an overview of static lineage tracing methods, and then focus on evolving lineage tracing technologies that enable reconstruction of tumor phylogenies at unprecedented resolution. We also discuss in vivo applications of these technologies to profile subclonal dynamics, quantify tumor plasticity, and track metastasis. Finally, we highlight outstanding questions and emerging technologies for building comprehensive cancer evolution roadmaps.

Intratumor heterogeneity (ITH) is associated with tumor progression in several clinical and experimental settings and contributes to therapeutic resistance. Its relation to cancer immunosurveillance is complex. Clonally heterogeneous tumors are associated with decreased immunosurveillance and are less responsive to immune checkpoint inhibition, but the mechanistic basis underlying these observations remains unclear. One possibility is that tumors that are under active immunosurveillance are relatively homogeneous because immunosurveillance prevents the outgrowth of immunogenic subclones. Alternatively, high ITH might directly impair immunosurveillance due to lower dosages of subclonal antigens, competition between antigens and immunodominance, the induction of detrimental T cell differentiation programs, or negative feedback loops. Here we review the evidence for these scenarios and outline hypotheses that could underlie the negative association between clonal heterogeneity and cancer immunosurveillance.

Immune checkpoint inhibitors induce significant and durable treatment responses in about 20% of all cancers, but many patients have natural resistance to current immunotherapies. The past decade of technological advances has resulted in large-scale profiling of many cancers and their tumor microenvironments, rapidly expanding our understanding of the mechanisms utilized by tumors to create immune-resistant microenvironments. In this review, we discuss key factors that create immune resistance and emerging concepts that are redefining how we view immune resistance, as well as highlight novel strategies that aim to convert immune-resistant into immune-sensitive tumors.

Organoid models have revolutionized cancer research through their ability to capture the cellular heterogeneity and spatial organization of a tumor in 3D culture. Patient-derived organoids can also mirror responses to therapy in vitro, opening the doors to personalized medicine that can direct clinical decision-making. As cancer immunotherapy has flourished and efforts to develop novel immunotherapies have increased, models that incorporate immune cells into organoid coculture to recapitulate the complexity of the tumor microenvironment faithfully are in high demand. To this end, a wide variety of organoid immune coculture methods have been developed, each differing in the source of immune cells used, types of immune cells maintained in culture, and their specific utility. This review aims to organize these methods into a framework that will aid researchers in choosing the appropriate system for their experimental needs. We also highlight several nonimmune cell types that have been successfully incorporated into organoid culture and the biology these coculture models are poised to interrogate.

Normal tissues progressively acquire mutations. Some mutations are positively selected, driving clonal expansions that may colonize the majority of a tissue by old age. In several cases mutant clonal expansion is due to biasing stem cell fate toward proliferation. However, the expansionary phase is transient and is followed by reversion toward wild-type behavior so that normal tissue integrity is retained. Here we consider the implications of these findings for carcinogenesis. We propose that to be considered a cancer driver, a mutant gene should be more prevalent in tumors than the normal lineage from which it emerged. Cancer risk is not dependent on mutational burden, but rather may reflect the relative frequency of pro- and anti-oncogenic mutants within a tissue. Understanding the basis of mutant clonal advantage over wild-type cells allows interventions to halt the expansion or even deplete oncogenic mutants from normal tissue, potentially lowering cancer risk.

Cellular senescence is a tumor-suppressive program that promotes tissue homeostasis by identifying damaged cells for immune-mediated clearance. Thus, the ability to evade senescence and the ensuing immune surveillance is a hallmark of cancer. Reactivation of senescence programs can result in profound immune-mediated tumor regressions or sensitize tumors to immunotherapy, although the aberrant persistence of senescent cells can promote tissue decline and contribute to the side effects of some cancer therapies. In this review, we first briefly describe the discovery of senescence as a tumor-suppressive program. Next, we highlight the dueling good and bad effects of the senescence-associated secretory program (SASP) in cancer, including SASP-dependent immune effects. We then summarize the beneficial and deleterious effects of senescence induction by cancer therapies and strategies in development to leverage senescence therapeutically. Finally, we highlight challenges and unmet needs in understanding senescence in cancer and developing senescence-modulating therapies.

Chimeric antigen receptor (CAR) T cell therapy has been a great success in CD19⁺ hematological diseases. Natural killer (NK) CAR cells offer an alternative to CAR T cells with an intrinsic potential for universal off-the-shelf cell therapeutics. The choice of cell type and the choice of CAR are both relevant for the feasibility, effectivity, engraftment, persistence, side effects, and safety of the cell therapy. Until recently CAR NK cells have proven difficult to develop into therapeutic products. Here, we give an overview of the source of CAR NK cells, gene transfer methods, and the manufacture of CAR NK cells for clinical application. We discuss improvements, as well as future options and problems that need to be addressed.

The principle of independent drug action proposes that responses to drug combinations result from responses to one or the other of two combining agents, but not both. Explorations of biological pathway interactions in signal transduction and immunobiology as synergy have not been connected to mathematical demonstrations of above–independent action activity, which would define pharmacologic synergy. We review independent action as the explanation for cancer drug combinations and find no evidence for pharmacologic synergy. Rather, a measure of correlation of response (ρ) when positive can explain below–independent action results, and negative correlation can explain above–independent action results. Anticorrelated responses may be a mathematical demonstration of collateral sensitivity, which can achieve above–independent action activity. Inappropriate use of biological concepts of synergy may be contributing to high failure rates for immuno-oncology clinical trials, indicating a need for more rigorous applications of independent action to the development of cancer drug combination therapy.

The blood-brain barrier is critically important for the treatment of both primary and metastatic cancers of the central nervous system (CNS). Clinical outcomes for patients with primary CNS tumors are poor and have not significantly improved in decades. As treatments for patients with extracranial solid tumors improve, the incidence of CNS metastases is on the rise due to suboptimal CNS exposure of otherwise systemically active agents. Despite state-of-the art surgical care and increasingly precise radiation therapy, clinical progress is limited by the ability to deliver an effective dose of a therapeutic agent to all cancerous cells. Given the tremendous heterogeneity of CNS cancers, both across cancer subtypes andwithin a single tumor, and the range of diverse therapies under investigation, a nuanced examination of CNS drug exposure is needed. With a shared goal, common vocabulary, and interdisciplinary collaboration, the field is poised for renewed progress in the treatment of CNS cancers.

Myeloid cells represent a dominant cellular compartment of tumor lesions and play key roles in tumor inception, progression, metastasis, and response to treatment. Mononuclear phagocytes (MNPs), which include dendritic cells and macrophages, are unique among myeloid cells, as they not only shape both the broader composition and state of the tumor microenvironment but can also specifically instruct cancer-specific, T cell–mediated tumor cell killing, making them especially attractive targets for cancer treatment. Although MNPs remain difficult to modulate therapeutically, our understanding of MNP biology in the antitumor immune response has expanded significantly, offering hope for new possibilities in cancer immunotherapy. Here, we review the recent advances in our study of the cellular identity, molecular diversity, and spatial organization of MNPs in tumors, and we discuss the importance of tailoring therapeutic strategies to incorporate these new insights into cancer treatment design.

Estrogen receptor (ER) α is expressed in the vast majority of breast cancers and is one of the most successfully prosecuted drug targets in oncology, with multiple classes of endocrine therapies approved for the treatment of ER⁺ breast cancer. These existing agents are highly active, both as single agents and as combination partners for other targeted therapies, and have significantly benefited patients. However, each of these standard-of-care (SOC) therapies has liabilities that allow for the reengagement of ER signaling as a mechanism of resistance. Data supporting the continued dependence of tumors on ER signaling following exposure to SOC agents have underpinned an extraordinary reenergizing of academic, biotechnology, and pharmaceutical groups pursuing next-generation ER-targeted therapies. The hypothesis that there remains an opportunity to bring further meaningful benefit to patients through fully optimized ER-targeted therapies is currently being investigated in the clinic.

T cell reactivity to tumor-specific neoantigens can drive endogenous and therapeutically induced antitumor immunity. However, most tumor-specific neoantigens are unique to each patient (private) and targeting them requires personalized therapy. A smaller subset of neoantigens includes epitopes that span recurrent mutation hotspots, translocations, or gene fusions in oncogenic drivers and tumor suppressors, as well as epitopes that arise from viral oncogenic proteins. Such antigens are likely to be shared across patients (public), uniformly expressed within a tumor, and required for cancer cell survival and fitness. Although a limited number of these public neoantigens are naturally immunogenic, recent studies affirm their clinical utility. In this review, we highlight efforts to target mutant KRAS, mutant p53, and epitopes derived from oncogenic viruses using T cells engineered with off-the-shelf T cell receptors. We also discuss the challenges and strategies to achieving more effective T cell therapies, particularly in the context of solid tumors.