Obesity, defined by excessive body fat, is a highly complex condition affecting numerous physiological processes, such as metabolism, proliferation, and cellular homeostasis. These multifaceted effects impact cells and tissues throughout the host, including immune cells as well as cancer biology. Because of the multifaceted nature of obesity, common parameters used to define it (such as body mass index in humans) can be problematic, and more nuanced methods are needed to characterize the pleiotropic metabolic effects of obesity. Obesity is well-accepted as an overall negative prognostic factor for cancer incidence, progression, and outcome. This is in part due to the meta-inflammatory and immunosuppressive effects of obesity. Immunotherapy is increasingly used in cancer therapy, and there are many different types of immunotherapy approaches. The effects of obesity on immunotherapy have only recently been studied with the demonstration of an obesity paradox, in which some immune therapies have been demonstrated to result in greater efficacy in obese subjects despite the direct adverse effects of obesity and excess body fat acting on the cancer itself. The multifactorial characteristics that influence the effects of obesity (age, sex, lean muscle mass, underlying metabolic conditions and drugs) further confound interpretation of clinical data and necessitate the use of more relevant preclinical models mirroring these variables in the human scenario. Such models will allow for more nuanced mechanistic assessment of how obesity can impact, both positively and negatively, cancer biology, host metabolism, immune regulation, and how these intersecting processes impact the delivery and outcome of cancer immunotherapy.

Obesity, a condition of excess adiposity usually defined by a BMI > 30, can have profound effects on both metabolism and immunity, connecting the condition with a broad range of diseases, including cancer and negative outcomes. Obesity and cancer have been associated with increased incidence, progression, and poorer outcomes of multiple cancer types in part due to the pro-inflammatory state that arises. Surprisingly, obesity has also recently been demonstrated in both preclinical models and clinical outcomes to be associated with improved response to immune checkpoint inhibition (ICI). These observations have laid the foundation for what has been termed the obesity paradox. The mechanisms underlying these augmented immunotherapy responses are still unclear given the pleiotropic effects obesity exerts on cells and tissues. Other important variables such as age and sex are being examined as further affecting the obesity effect. Sex-linked factors exert significant influences on obesity biology, metabolism as well as differential effects of different immune cell-types. Age can be another confounding factor contributing to the effects on both sex-linked changes, immune status, and obesity. This review aims to revisit the current body of literature describing the immune and metabolic changes mediated by obesity, the role of obesity on cancer immunotherapy, and to highlight questions on how sex-linked differences may influence obesity and immunotherapy outcome.

Immune checkpoint inhibition (ICI) has emerged as one of the most powerful tools to reverse cancer induced immune suppression. Monoclonal antibodies (mAbs) targeting programmed cell death 1/programmed cell death ligand 1(PD-1/PD-L1) are FDA-approved and their clinical use is rapidly expanding. As opposed to the clinical paradigm, which can result in significant responses and toxicities, it has been difficult to reproduce these effects preclinically using mouse models. In large part, this is due to models, which employ rapidly growing ex vivo cultured transplantable tumor cell lines engrafted into young naïve inbred laboratory mice. However, another issue concerns the use and repeated application of xenogeneic reagents in mice (i.e., rat or hamster mAbs directed against mouse antigens at variance with clinical use of human or humanized mAbs). Building on our previous studies demonstrating that repeated administration of commonly used xenogeneic anti-PD-1 mAbs derived from both rat and hamster can induce fatal hypersensitivity in some tumor-bearing mice, we sought to compare these result with the effects of a mouse anti-mouse PD-1 mAb. Application of a murine anti-mouse PD-1 (clone: MuDX400) did not result in lethal anaphylaxis in the 4T1 tumor model. It also displayed superior antitumor effects in this and other tumor models, as it did not induce neutralizing antibody responses against the anti-PD-1 mAb, such as were observed when using xenogeneic anti-PD1 mAbs. These results demonstrate that more accurate preclinical modeling necessitates the use of mouse reagents mirroring the clinical scenario to ascertain long-term effects or toxicities, while avoiding xenogeneic responses, which do not occur clinically. Furthermore, these studies suggest a direct mechanism, whereby preclinical murine studies have often failed to recapitulate the clinical efficacy and toxicity of single agent checkpoint inhibition.

The efficacy of allogeneic hematopoietic stem cell transplantation (allo-HSCT) is limited by the occurrence of acute and chronic graft-versus-host disease (GVHD). We have recently demonstrated that obesity results in exacerbated acute gastrointestinal GVHD in both mouse models and clinical outcomes due to increased pro-inflammatory cytokine responses and microbiota alterations. We therefore wanted to delineate the role of the various parameters in obesity, adiposity, effects of high-fat (HF) diet, and the role of microbiome on GVHD pathogenesis, by taking advantage of a mouse strain resistant to diet-induced obesity (DIO). Female BALB/c mice are resistant to DIO phenotype with approximately 50% becoming DIO under HF diets. The DIO-susceptible recipients rapidly succumb to acute gut GVHD, whereas the DIO-resistant recipient littermates, which do not become obese, are partially protected from GVHD, indicating that being on HF diet alone contributes to but is not the primary driver of GVHD. Microbiome assessment revealed restricted diversity in both cohorts of mice, but coprophagy normalizes the microbiota in mice housed together. We then individually housed DIO-resistant, DIO-susceptible, and lean control mice. Notably, each of the individually housed groups demonstrates marked restricted diversity that has been shown to occur from the stress of single housing. Despite the restricted microbiome diversity, the GVHD pathogenesis profile remains consistent in the group-housed mice, with the lean control single-housed mice exhibiting no acute GVHD and DIO-resistant recipients showing again partial protection. These results demonstrate that the deleterious effects of obesity on acute gut GVHD are critically dependent on adiposity with the HF diet also playing a lesser role, and the microbiome alterations with obesity instead appear to fuel ongoing acute GVHD processes.

Despite obesity reaching pandemic proportions, its impact on antigen-specific T cell responses is still unclear. We have recently demonstrated that obesity results in increased expression of PD-1 on T cells, and checkpoint blockade targeting PD-1/PD-L1 surprisingly resulted in greater clinical efficacy in cancer therapy. Adverse events associated with this therapy center around autoimmune reactions. In this study, we examined the impact of obesity on T cell priming and on autoimmune pathogenesis using the mouse model experimental autoimmune encephalomyelitis (EAE), which is mediated by autoreactive myelin-specific T cells generated after immunization. We observed that diet-induced obese (DIO) mice had a markedly delayed EAE onset and developed milder clinical symptoms compared to mice on control diet (CD). This delay was associated with impaired generation of myelin-specific T cell numbers and concurrently correlated with increased PD-L1 upregulation on antigen-presenting cells in secondary lymphoid organs. PD-1 blockade during the priming stage of EAE restored disease onset and severity and increased numbers of pathogenic CD4+ T cells in the central nervous system (CNS) of DIO mice to similar levels to those of CD mice. Administration of anti-PD-1 after onset of clinical symptoms did not increase EAE pathogenesis demonstrating that initial priming is the critical juncture affected by obesity. These findings demonstrate that obesity impairs antigen-specific T cell priming, but this can be reversed with PD-1 blockade. Our results further suggest that PD-1 blockade may increase the risk of autoimmune toxicities, particularly in obesity.

Purpose

Clinical successes using current T-cell based immunotherapies have been limited in soft tissue sarcomas (STS), while pre-clinical studies have shown evidence of natural killer (NK) cell activity. Since tumor immune infiltration, especially tumor-infiltrating lymphocytes, is associated with improved survival in most solid tumors, we sought to evaluate the gene expression profile of tumor and blood NK and T cells, as well as tumor cells, with the goal of identifying potential novel immune targets in STS.

Experimental design

Using fluorescence-activated cell sorting, we isolated blood and tumor-infiltrating CD3^-CD56⁺ NK and CD3⁺ T cells and CD45^- viable tumor cells from STS patients undergoing surgery. We then evaluated differential gene expression (DGE) of these purified populations with RNA sequencing analysis. To evaluate survival differences and validate primary DGE results, we also queried The Cancer Genome Atlas (TCGA) database to compare outcomes stratified by bulk gene expression.

Results

Sorted intra-tumoral CD3⁺ T cells showed significant upregulation of established activating (CD137) and inhibitory genes (TIM-3) compared to circulating T cells. In contrast, intra-tumoral NK cells did not exhibit upregulation of canonical cytotoxic genes (IFNG, GZMB), but rather significant DGE in mitogen signaling (DUSP4) and metabolic function (SMPD3, SLC7A5). Tumors with higher NK and T cell infiltration exhibited significantly increased expression of the pro-inflammatory receptor TLR4 in sorted CD45^- tumor cells. TCGA analysis revealed that tumors with high TLR4 expression (P = 0.03) and low expression of STMN1 involved in microtubule polymerization (P < 0.001) were associated with significantly improved survival.

Conclusions

Unlike T cells, which demonstrate significant DGE consistent with upregulation of both activating and inhibiting receptors in tumor-infiltrating subsets, NK cells appear to have more stable gene expression between blood and tumor subsets, with alterations restricted primarily to metabolic pathways. Increased immune cell infiltration and improved survival were positively correlated with TLR4 expression and inversely correlated with STMN1 expression within tumors, suggesting possible novel therapeutic targets for immunotherapy in STS.

Introduction

The incidence of obesity, a condition characterized by systemic chronic inflammation, has reached pandemic proportions and is a poor prognostic factor in many pathologic states. However, its role on immune parameters has been diverse and at times contradictory. We have previously demonstrated that obesity can result in what has been called the "obesity paradox" which results in increased T cell exhaustion, but also greater efficacy of immune checkpoint blockade in cancer treatment.

Methods

The role of obesity, particularly in the context of aging, has not been robustly explored using preclinical models. We therefore evaluated how age impacts the immune environment on T cell development and function using diet-induced obese (DIO) mice.

Results

We observed that DIO mice initially displayed greater thymopoiesis but then developed greater thymic involution over time compared to their lean counterparts. Both aging and obesity resulted in increased T cell memory conversion combined with increased expression of T cell exhaustion markers and Treg expansion. This increased T cell immunosuppression with age then resulted in a loss of anti-tumor efficacy by immune checkpoint inhibitors (ICIs) in older DIO mice compared to the younger DIO counterparts.

Discussion

These results suggest that both aging and obesity contribute to T cell dysfunction resulting in increased thymic involution. This combined with increased T cell exhaustion and immunosuppressive parameters affects immunotherapy efficacy reducing the advantage of obesity in cancer immunotherapy responses.

Stereotactic ablative radiotherapy (SABR) is a standard-of-care for medically-inoperable-early-stage non-small cell lung cancer (NSCLC). One third of patients progress and chemotherapy is rarely used in this population. We questioned if addition of the immune-checkpoint-inhibitor (ICI) atezolizumab to standard-of-care SABR can improve outcomes. We initiated a multi-institutional single-arm phase I study (NCT02599454) enrolling twenty patients with the primary endpoint of maximum tolerated dose (MTD); secondary endpoints of safety and efficacy; and exploratory mechanistic correlatives. Treatment is well tolerated and full dose atezolizumab (1200 mg) is the MTD. Efficacy signals include early responses (after 2 cycles of ICI, before initiation of SABR) in 17% of patients. Biomarkers of functional adaptive immunity, including T cell activation in the tumor and response to ex-vivo stimulation by circulating T cells, are highly predictive of benefit. These results require validation and are being tested in a phase III randomized trial.

Bystander activation of memory T cells occurs via cytokine signaling alone in the absence of T cell receptor (TCR) signaling and provides a means of amplifying T cell effector responses in an antigen-nonspecific manner. While the role of Programmed Cell Death Protein 1 (PD-1) on antigen-specific T cell responses is extensively characterized, its role in bystander T cell responses is less clear. We examined the role of the PD-1 pathway during human and mouse non-antigen-specific memory T cell bystander activation and observed that PD-1+ T cells demonstrated less activation and proliferation than activated PD-1- populations in vitro. Higher activation and proliferative responses were also observed in the PD-1- memory population in both mice and patients with cancer receiving high-dose IL-2, mirroring the in vitro phenotypes. This inhibitory effect of PD-1 could be reversed by PD-1 blockade in vivo or observed using memory T cells from PD-1-/- mice. Interestingly, increased activation through abrogation of PD-1 signaling in bystander-activated T cells also resulted in increased apoptosis due to activation-induced cell death (AICD) and eventual T cell loss in vivo. These results demonstrate that the PD-1/PD-Ligand 1 (PD-L1) pathway inhibited bystander-activated memory T cell responses but also protected cells from AICD.

Natural killer (NK) cells are involved in innate defense against viral infection and cancer. NK cells can be divided into subsets based on the ability of different receptors to bind to major histocompatibility (MHC) class 1 molecules, resulting in differential responses upon activation in a process called "licensing" or "arming." NK cells expressing receptors that bind self-MHC are considered licensed due to an augmented effector lytic function capability compared with unlicensed subsets. However, we demonstrated that unlicensed NK subsets instead positively regulate the adaptive T-cell response during viral infections that are related to localization and cytokine production. In this study, the differential effects of the two types of NK subsets were contingent on the environment in viral infection and hematopoietic stem cell transplantation (HSCT) models. Infection of mice with high-dose (HD) murine cytomegalovirus (MCMC) led to a loss of licensing-associated differences, as compared with mice with low-dose (LD) infection: the unlicensed NK subset no longer localized in lymph nodes (LNs), but instead remained at the site of infection. Similarly, the patterns observed during HD infection paralleled the phenotypes of both human and mouse NK cells in an HSCT setting where NK cells exhibit an activated phenotype. However, in contrast to the effects of subset depletion in T-cell replete models, the licensed NK cell subsets still dominated antiviral responses after HSCT. Overall, our results highlight the intricate tuning of NK cells and how it affects overall immune responses with regard to licensing patterns and their dependency on the level of stimulation and activation status.