The first evidence of the relationship between an active immune response and systemic metabolic alterations dates from 1883-1884, when the association between meningococci meningitis infection and transient diabetic syndrome was described (8). These observations pointed at increased insulin resistance upon infection, that was further confirmed in both animal models (9) and acute infection in humans (10). In addition to this, evidence of the presence of immune infiltrates in metabolic tissue was already known, although a possible functional role of the infiltrate was not interrogated (11, 12). The cellular and molecular mechanism underlaying all these observations remained unknown until landmark works from Feingold and Lang described how Tumour Necrosis Factor alpha (TNFa), a potent pro-inflammatory cytokine, induced insulin resistance in vivo (13, 14). Further studies described increased levels of TNFa in obese adipose tissue in both animal models and humans, while TNFa neutralisation led to improvement in metabolic parameters such insulin resistance in mice (15–17). However, even up to this day, a causal relationship between inflammation and systemic metabolic changes is still not fully described in humans (18, 19). Despite this, the presence of an abundant immune cell resident population in metabolic tissue suggests their importance in maintaining tissue homeostasis as well as performing surveillance functions. Moreover, resident cells are key players in the mechanism of action of current immunotherapies, as we will discuss later.

At the tissue level and in homeostatic conditions, this is, in the absence of any danger, stress or inflammation, the metabolic tissue immune infiltrate is predominantly of regulatory nature, with abundant regulatory T cells (Tregs) and homeostatic macrophages (M2 type) that promote correct insulin signalling in non-immune cells such adipocytes or hepatocytes (20, 21). Immune cells adapt to their environment by even acquiring the expression of master transcription factors of metabolic cell function, like Peroxisome proliferator-activated receptor gamma (PPARγ) (22) ( Figure 2 ). This cellular landscape changes upon immunological challenge, like an infection. Infection by most pathogens with high replicative capacity such as virus and bacteria, requires a type 1 response, associated with T cell activation and the production of the cytokines interferon gamma (IFNγ) and TNFα, that will activate innate immune cells like macrophages (M1 type) to target the pathogen ( Figure 2 ). This type of response also requires a high proliferative rate, which is supported by glucose. In order to shift nutrient allocation towards the immune system, a transient state of insulin resistance is induced in metabolic tissue (23). This molecular crosstalk is mediated by pro-inflammatory cytokines and adipokines.

A very similar landscape is observed upon metabolic overload, where increased inflammation in metabolic tissue has been widely described. Akin to infection, the immune infiltration consists of T cells producing IFNγ (T-helper cells type 1, Th1) and interleukin (IL)-17 (T-helper cell type 17, Th17) pro-inflammatory cytokines and M1 macrophages, establishing a feedback loop where low grade chronic inflammation induced by nutrient excess leads to chronic insulin resistance in metabolic tissue (18, 19, 24, 25). Interestingly, nutrient excess in an aberrant environmental condition that is exclusive to humans. Moreover, it is a recent event in evolutionary terms, hence there has been no requirement for survival pressure to evolve mechanism to protect the body from caloric excess (as opposed to food deprivation) and as such there are limited natural options to return the body to homeostasis (18, 26).

Metabolic overload in metabolic tissue is generally the consequence of “Western lifestyle” patterns, linked to high caloric intake, lack of physical exercise and other factors that overstrain systemic metabolism. A key manifestation of metabolic overload is obesity, that is associated with reduced immune function. Obese individuals have increased susceptibility to infections and reduced vaccine responses - a similar phenotype observed in aging - possibly due to continuous immune activation over time (27). These alterations include a reduced function in primary and secondary lymphoid organs (28–31); and increased proportions of dysfunctional immune cells (29, 32–34).

The progression from tumoral transformation of normal cells into clinically relevant cancer depends on one hand on the capability of tumour cells to proliferate and migrate to other tissues and on the other hand, the ability of the immune system to mount an adequate anti-tumour response. Both processes depend on metabolic cues. As already explained, cancer and immune cells undergo metabolic reprograming, a requirement for their high proliferative rate and acquisition of effector function. As such, numerous cancer-related genetic mutations or tumour-suppressor genes have a direct effect on cellular metabolism, highlighting how important metabolism is as a target when thinking about cancer treatment or even prevention (35). Additionally, many cancers develop adjacent to adipose tissue, and they can manipulate adipocyte biology to provide metabolites that fuel cancer cell division, growth and metastasis. Moreover, cancer cells can also induce a dedifferentiation programme in the adipocytes that further support these processes (36, 37). As a consequence, calorie restriction had been proposed to inhibit cancer growth for the last 100 years (38, 39).

If cancer and metabolism are deeply interlinked, then it is no surprise to observe that obesity has profound effects on cancer risk and development. At the population level, a “Western lifestyle” – associated with a high diet caloric intake – is widening across Asian and African countries and this has been associated with increased rates of cancer in these countries. Similarly, migration to a “Western lifestyle” country is accompanied by the acquisition of the indigenous population’s cancer risk (40–42). These epidemiological observations suggest that, at least in these cases, cancer should be preventable with interventions that target the “Western lifestyle”. This is a changing paradigm in the current cancer clinical arena, as described in this excellent review by Holly et al. (43).

Overall, obesity has been associated with both cancer risk and progression (36, 44). The mechanism(s) that link both are intense areas of research and implicate immune (chronic inflammation, increased systemic pro-inflammatory cytokines), metabolic (hyperinsulinemia, dysregulation of leptin and adiponectin levels) and endocrine (steroid hormones) factors. Moreover, many of the therapies used to treat cancer today specifically target the immune system, seeking to harness and amplify the body’s built-in anti-tumour response. Given the profound effect of obesity on the immune system and cancer risk, it is necessary to consider how obesity affects the efficacy and mechanism of these immunotherapeutic approaches. To explore this topic further, we will first consider mechanistic insights into the phenotypic and functional changes that occur in T cells, in the context of both cancer and obesity.

T cell exhaustion is a dysfunctional cell state, classically described as arising due to chronic antigenic stimulation (45), but hypoxia and glucose deprivation have also been shown to have important roles in driving the exhaustion process (46, 47). Exhaustion is characterised by increased expression of immune inhibitory receptors, weakened effector function, reduced self-renewal capacity, altered epigenetics, and a specific transcriptional programme and metabolism (48). T cell exhaustion is one of the major causes leading to immune escape of cancer, creating an environment that supports tumour development and metastatic spread (49). In addition, T cell exhaustion plays a pivotal role in the efficacy of current immunotherapies for cancer, hence we will next provide a comprehensive view of the role of T cell exhaustion in cancer development and progression.

The key drivers of exhaustion are present in most cancers. Tumour infiltrating lymphocytes will be consistently stimulated by tumour antigen presented by antigen-presenting cells, often over a prolonged period prior to diagnosis. Tumours are also highly hypoxic tissues which consume considerable amounts of nutrients compared to healthy tissues, so T cells in the vicinity of tumours will also experience hypoxia and nutrient deprivation (50). It has been found that tumour-infiltrating lymphocytes in patients with a range of cancers have high proportions of exhausted T cells (49, 51).

Interestingly, several studies also demonstrate an increase in T cell exhaustion as a result of obesity, possibly due to the continuous immune activation present in both conditions. In obese mouse models, T cell exhaustion is described not only in adipose tissue T cells (52), but also in peripheral blood, liver and spleen, particularly when combined with aging (34, 53). Though there is no definitive answer yet as to the exact mechanisms of obesity-driven T cell exhaustion, there is evidence that adipocytes in the obese state can upregulate major histocompatibility complex class II (MHC-II) and act as antigen-presenting cells to adipose tissue infiltrating T cells (54). Furthermore, excess adiposity is linked to increased leptin secretion ( Figure 2 ). Leptin is a hormone involved in regulating appetite and energy storage, but it has also been shown to affect T cell differentiation pathways. T cells (especially CD4⁺) exposed to leptin are more likely to differentiate into a pro-inflammatory phenotype, increasing production of pro-inflammatory cytokines (54). This pro-inflammatory adipose environment, with increased levels of cytokines such as TNFα and IL-6 (55) along with chronic T cell stimulation through MHC-II leads to T cell exhaustion.

Rather than proliferating, adipose cells undergo hypertrophy due to obesity, and there is evidence that adipose tissue becomes hypoxic due to increased tissue density (56). In obesity, adipose cells will upregulate genes associated with hypoxia, leading to activation of the hypoxia-inducible factor 1 (HIF-1) pathway. This creates an environment that drives persistent glycolysis and effector function in CD8⁺ T cells (55). This consistent stimulation may further drive the increase in exhausted phenotype seen in obesity models.

Furthermore, obesity leads to both systemic and tissue-specific inflammation. There are several mechanisms that contribute to this phenomenon, one of which is nutrient overload. Excessively high nutrient levels can cause oxidative stress through the release of reactive oxygen species and may also lead to endoplasmic reticulum (ER) stress. Increased oxidative and ER stress, combined with the activation of Toll-like receptors by fatty acids and glucose, will cause an inflammatory response which will further stimulate T cells (57).

Though the nutrient deprivation seen in cancer is not present in obesity (and in fact, nutrient levels are generally chronically elevated), persistent stimulation and hypoxia are evident. These two factors may therefore drive the presence of elevated levels of exhausted T cells in obese individuals, as demonstrated in murine models. Much of the existing literature on nutrient metabolism in cancer and obesity focuses on ATP-generating molecules, such as glucose, glutamine and fatty acids (58–62). Here, we consider an alternative nutrient that is often overlooked yet has a critical role in cellular proliferation, which is central to both cancer and the resulting immune response: cholesterol.

Cholesterol metabolic reprogramming is also a hallmark of cancer (63). Cholesterol is required to facilitate cell division, and so to enable the rapid division of tumour cells, cancer cells employ tactics to circumvent regulation of cholesterol synthesis and uptake and maintain high levels of intracellular cholesterol. Cholesterol is different to other metabolites, as it cannot be catabolised and used for the production of ATP, and it is therefore not an energy source for cells. While critical for cell proliferation, cellular cholesterol accumulation is toxic, and so both the biosynthesis and uptake of cholesterol by cells is very tightly regulated.

Cells in a wide range of cancers have been shown to overexpress the low-density lipoprotein receptor (LDLR) involved in cellular uptake of cholesterol, and sterol regulatory element-binding protein 2 (SREBP2), the master transcription factor that promotes cholesterol biosynthesis and import (64). Crucially, it was also shown that levels of these proteins did not decrease in the presence of exogenous cholesterol as they did in normal cells (65, 66, p. 19; 67). This enables cancer cells to circumvent the feedback loop that would normally inhibit cholesterol synthesis and uptake in the presence of high cholesterol levels. As a result, solid tumours are characterised by high cholesterol levels (68), and this dysregulation also leads to raised cholesterol levels in the TME (69) ( Figure 3 ).

Together with SREBP-2, Liver-X-Receptors (LXR) are transcription factors critical to cholesterol homeostasis, activated by oxysterols (70). Dysregulation of LXR expression and activation is implicated in different cancers, though with variable mechanisms. For example, LXRs are known to be highly activated in triple-negative breast cancer tumour associated myeloid cells, which regulates macrophage activity and leads to an anti-inflammatory effect, reducing tumour killing (71). However, LXR activation may actually be beneficial in some cancers, with LXR activation leading to cell cycle arrest in prostate cell cancer lines (70). This may be due to the role of LXR in androgen production, which is closely linked with prostate cancer development. What is known is that during cell proliferation, crucial to both the immune response and cancer growth, the link between increased intracellular cholesterol and LXR activation is dysregulated, which means normal cholesterol efflux mechanisms are disrupted (72). While this pathway may therefore be a target to reduce proliferation of cancer, this may also interrupt the tumour immune response which heavily relies on proliferation of cancer-specific T cells.

As with all cells, cholesterol is essential to the healthy functioning and division of T cells. Upon activation, T cells will reprogram their metabolism to increase cholesterol synthesis and uptake, facilitating their rapid proliferation (73). However, as described above, excessive cholesterol accumulation is toxic and can lead to ER stress and cellular dysfunction (74). In the TME specifically, it has been found that elevated cholesterol leads to exhaustion of CD8⁺ T cells (69). Beyond the cancer context, hypercholesterolaemia has been shown to be linked to changes in T cells. In mice, hypercholesterolaemia has been associated with a T cell inflammatory response (75, 76). While it has not been shown that this progresses into T cell exhaustion, it is known that chronic stimulation contributes to the development of exhaustion. It is therefore possible that T cell activation induced by systemic hypercholesterolaemia could induce the same exhaustion phenotype seen in cancer.

Obesity is associated with an increased risk of hypercholesterolaemia (77), so this mechanism of T cell activation (and potentially exhaustion) may be present in a higher proportion of individuals with obesity compared to non-obese individuals.

Both obesity and cancer appear to lead to T cell exhaustion, through some similar mechanisms (e.g, hypoxia and chronic stimulation), and some differing pathways (e.g, nutrient deprivation vs. nutrient overload). Given that the function of exhausted T cells is rescued with checkpoint inhibitor treatment, it is possible that the paradoxical improvement in checkpoint inhibitor (CPI) outcomes in obese patients is influenced by an increase in the dysfunctional T cell compartment in obesity.

Though immunotherapy has experienced a renaissance in recent years, the concept of using the immune system against cancer originates in the late 19^th century. William B. Foley, a bone surgeon, found that when his sarcoma patients developed post-operative infections, their remaining tumours began to shrink. By injecting them with a cocktail of sepsis-inducing bacteria, Foley was able to induce anti-tumour immune responses and achieve durable remission in some of his patients. However, the risks associated with the treatment meant that immunotherapy was omitted from the arsenal of cancer treatment for almost a century in favour of radiotherapy and resection (78).

While some anticancer immunotherapies were developed in the 1980s (e.g, high-dose IL-2 treatment), a 1996 pre-clinical study published by Leach et al. (79) marked the start of the immunotherapy boom. This study concerned cytotoxic T-lymphocyte-associated protein 4 (CTLA-4), an immune checkpoint expressed on T cells which binds to ligands CD80 and CD86. These ligands normally bind to CD28, leading to co-stimulation which, alongside antigen presentation to the T cell receptor (TCR), facilitates T cell activation. CD28:CD80/86 binding also leads to T cell proliferation, increased longevity and differentiation. However, these ligands preferentially bind to CTLA-4, so when it is presented, these functions are inhibited. Overall, this leads to a reduction on the effector response of T cells. CD80 and CD86 are generally presented by antigen-presenting cells in secondary lymphoid organs (lymph nodes and spleen), and so CTLA-4 is thought to regulate the T cell response early in the immune process (80). The Leach study found that CTLA-4, which had been identified just a year prior, could be exploited to improve the anti-tumour immune response in mice. Leach found that by injecting mice with anti-CTLA-4 antibodies, tumour regression improved compared to untreated or anti-CD28 treated mice in a range of tumour types (79).

In 2002, the Honjo lab at Kyoto University discovered that another immune checkpoint, PD-1, could be a beneficial target for cancer immunotherapy along with its ligand, PD-L1 (81). PD-1 is present on the surface of a range of immune cells, including T cells (especially post-activation). When presented on T cells and bound to its ligand (PD-L1 or PD-L2), PD-1 causes downstream signalling that leads to both direct and indirect inhibition of TCR signal transduction, reducing activation of T cells. Additionally, PD-1 binding may reprogram T cells from glycolysis-led metabolism, indicative of an effector phenotype, to greater reliance on fatty acid oxidation, which is representative of memory phenotype (82). Together, these effects lead to reduced T cell effector function. This pathway is crucial to maintaining peripheral tolerance, thought there is increasing evidence for PD-1’s importance in the lymph node (83). The Honjo lab demonstrated that cytotoxic T cells were less effective at tumour killing & lysis when presented with a PD-L1 presenting tumour model in vitro. Furthermore, the same paper demonstrated that the growth of myeloma tumours in mice was inhibited with administration of PD-L1 blockade, and completely halted in PD-1 deficient mice. Later mouse studies showed that PD-1 was highly expressed on dysfunctional T cells following chronic infection (84, 85) and that PD-1 or PD-L1 blockade improved the function of these T cells (84). Together, this work showed promising evidence that PD-1/PD-L1 blockade could effectively inhibit tumour growth and help recover the function of chronically stimulated T cells.

In the clinic today, anti-PD-1/PD-L1 and anti-CTLA4 are commonly used checkpoint inhibitors for a range of cancers, though recently the first LAG-3 inhibitor has also been approved for clinical use (86). These drugs work by binding with the immune checkpoints on T cells without inducing the downstream signalling that would normally lead to suppression of effector function and killing ability. Studies have shown that these drugs are able to recover function in T cells that had previously been exhausted due to chronic stimulation and exposure to the tumour microenvironment (TME) (87, 88).

Surprisingly, there is currently little information regarding the effect of checkpoint inhibitors on systemic metabolism. A 2021 study of 374 cancer patients found that overweight patients with high metabolic risk were more prone to higher severity toxicities when taking immune checkpoint inhibitors, and that this additional risk was not present with either obesity or reduced metabolic risk (89). It is clear therefore that changes in systemic metabolism have an impact on the function of CPIs and other T cell therapies. However, the inverse, that is, the impact of CPIs on systemic metabolism, has not been explored, and would be an interesting topic for further investigation.

CPIs have been transformational in oncology treatment, and 2020 estimates suggest that 36-39% of all patients with cancer in the US are eligible for CPI treatment (90). Hundreds of thousands of patients worldwide have been treated with these drugs since the approval of anti-CTLA-4 agent Ipilimumab in 2011. With this large patient cohort, it is now possible to understand how the effects of these drugs change depending on patient characteristics.

In 2018, a particularly surprising story began to emerge. McQuade et al. published unexpected findings from a cohort of 1918 patients that obese male melanoma patients had improved progression-free survival after treatment with checkpoint inhibitors (91). Over the following years, this pattern was also seen in melanoma patients regardless of sex (92), and in non-small cell lung cancer (NSCLC) (93). Obesity is defined here as body mass index (BMI) over 30. Given that obesity generally increases the risk of cancer development, this was considered an unexpected finding and named the ‘obesity paradox’. Similar inverse relationships have also been reported in cancer beyond the CPI context, where patients undergoing chemotherapy who responded to treatment showed increases in serum total cholesterol and LDL levels from baseline and above healthy levels (94).

Several hypotheses have been suggested to explain the obesity paradox. The effect may be related to increased T cell exhaustion which is often seen in cases of chronic antigen presentation, such as in cancer, obesity and autoimmunity, or is possibly related to improved immune infiltration in obesity (95). As well as exhaustion, an increase in tissue-resident memory T cells (T_RMs) may also be implicated in the obesity paradox. T_RMs are a T cell subset which resides in tissues rather than circulating through the body. These cells are characterised by their high expression of immune checkpoints, and increased effector function (96). For these reasons, there are several studies that suggest a link between CPI treatment efficacy and heightened levels of T_RMs (97, 98). Interestingly, T_RMs also appear to be upregulated in the adipose tissue of individuals with obesity (99). It is therefore possible that the obesity paradox may be partly explained by an increase in the T_RM compartment in obese individuals. Despite these hypotheses, as yet there is no confirmed answer explaining the obesity paradox effect. However, CPI treatment in obesity is not the only context in which we see an interplay of metabolic dysregulation and immune response.

Interestingly, paradoxical associations between cholesterol levels and immune activation have also been reported in the context of chronic inflammatory diseases. The lipid paradox refers to an unexpected observation made nearly 15 years ago describing the paradoxical relationship between systemic lipid levels and disease activity in patients with Rheumatoid Arthritis (RA), a common chronic inflammatory autoimmune disorder (100). Cardiovascular disease is commonly associated with elevated lipid levels in serum. However, in RA decreased systemic levels of total cholesterol and LDL are associated with its hyperinflammatory state (100), and lipid concentration is inversely associated with inflammatory markers (101). Even more interesting is the fact that upon anti-inflammatory treatment, lipid levels increase (or normalise) and cardiovascular and metabolic health improves (100, 102, 103). This paradoxical relationship is thought to be due to the chronic inflammatory nature of RA, pointing out at the fact that chronic exposure to pro-inflammatory cytokines dissasociates plaque formation from hyperlipidaemia.

With the clear link between metabolic dysregulation (both in the case of obesity and dyslipidaemia) and the immune response, it is very likely that systemic metabolism is influential in the mechanism and efficacy of immune-modulating drugs that are a burgeoning part of both cancer and autoimmunity treatment. While overweight and obese patients respond better to CPIs, they also have a greater susceptibility to cancer, with obese people more likely to develop 13 types of cancer than those with a healthy BMI (104). It should be noted that BMI (which is generally used to classify obesity) is a non-specific measurement which does not take into account body composition (e.g, fat vs muscle mass). These additional factors can also have a significant effect on cancer treatment outcomes.

Although the “obesity paradox” in cancer immunotherapy is widely discussed, a major limitation of many supporting studies is their reliance on BMI, rather than direct measures of body composition. BMI fails to distinguish between fat types or lean mass and is thus confounded. In contrast, emerging evidence suggests that skeletal muscle and subcutaneous fat are critical predictors of immunotherapy response. Figure 4 gives an overview of the main factors relating to body composition and nutrition currently known to contribute to immunotherapy outcomes.

Skeletal muscle mass has consistently shown prognostic significance across multiple cancer types. In gastric cancer, low skeletal muscle index (SMI) or sarcopenia is associated with worse survival (105, 106) and reduced tumour regression in patients receiving immunotherapy alone or with chemotherapy (107). Similar associations are seen in non-small cell lung cancer (NSCLC) (108–110), small cell lung cancer (SCLC) (111), hepatocellular carcinoma (HCC) (112, 113), melanoma (108, 114), head and neck squamous cell carcinoma (HNSCC) (115), where low SMI, muscle density, or sarcopenia predict poorer progression-free survival (PFS) and overall survival (OS) following CPIs. In renal cell carcinoma (RCC), SMI was the only body composition metric linked to OS after CPI therapy, with transcriptomic data showing that tumours from low-SMI patients had increased angiogenic and inflammatory signatures (116). These findings seem to be supported by other studies where poor body composition risk score (117) or higher cachexia index (118) had worse OS and PFS. Together, these studies demonstrate that low muscle mass and sarcopenia are prognostic factors affecting immunotherapy efficacy and long-term survival of cancer patients. Retaining muscle mass during treatment is then imperative, as loss of skeletal muscle during systemic therapy, especially in male NSCLC patients, further predicts poor prognosis (119).

Nutritional and metabolic status also influence outcomes. The ELY-2 study (120) in metastatic NSCLC patients under anti-PD1 therapy found that hypermetabolic patients who met ≥90% of their caloric needs had longer PFS, highlighting that adequate caloric intake may mitigate the negative impact of elevated resting energy expenditure. Nutritional indices like the Geriatric Nutritional Risk Index (GNRI) and Prognostic Nutritional Index (PNI) are also prognostically valuable. GNRI was associated with longer OS and PFS, particularly in older and early-stage NSCLC patients (121), while obesity was linked to poorer outcomes. Similarly, low PNI and low total adipose tissue predicted worse survival in NSCLC patients receiving CPIs and radiotherapy (122). Furthermore, Ding et al. found that malnutrition, sarcopenia, and low advanced-lung-cancer-inflammation-index scores were independently associated with poor survival in response to chemo-immunotherapy (123). Altogether, integrating GNRI, PNI, and body composition analysis enhances prognostic accuracy and supports more personalized treatment strategies.

Targeting amino acid metabolism through dietary manipulation or supplementation may be a strategy to simultaneously enhance immune-mediated tumour control and address cancer-associated metabolic dysfunctions. In a preclinical study, β-hydroxy-β-methylbutyrate (HMB), a leucine metabolite with muscle-sparing and immunomodulatory effects, reduced tumour growth, preserved muscle mass in obese mice, and improved anti-PD1 response in lean mice (124). These results underline HMB’s potential to not only support immune-based therapies but also ameliorate cancer cachexia, which makes it an attractive candidate for integrated cancer care.

Subcutaneous fat is another favourable prognostic factor. In gastric cancer, low subcutaneous fat area (SFA) or index (SFI) predicts poorer OS and treatment response in patients treated with CPIs (105, 106, 125) combined with PNI and SMI, high SFA further stratifies patients with better outcomes (105). However, fat and muscle loss during treatment still correlates with progression risk. Loss of subcutaneous adipose tissue and sarcopenia after beginning CPI treatment is correlated with higher risk of disease progression in metastatic cancer patients (126). In NSCLC, low SFI predicts shorter OS and PFS (109), while in RCC, high subcutaneous fat percentage (SAT%) correlates with improved PFS and increased intratumoral PD1⁺CD8⁺ T cell density (127), potentially explaining the better therapeutic outcome following CPI treatment. Yet, some studies found no significant associations between adiposity and outcomes in RCC (116) and melanoma (114). Overall, subcutaneous fat appears to be a reliable and favorable prognostic indicator in CPI-treated patients, although this association is influenced by cancer type and individual patient characteristics.

On the other end, myosteatosis and intramuscular adipose content (IMAC) tend to be associated with worse outcomes. In urothelial carcinoma (128) and NSCLC (110, 129), high levels of myosteatosis correlates with lower OS and PFS, while better muscle quality (low IMAC) is associated with improved outcomes.

Compared to subcutaneous fat and muscle, visceral fat shows inconsistent correlations. While low visceral adiposity has been linked to poor disease control in some cancers (HNSCC, HCC, colorectal, and urothelial) (112, 115, 128, 130), findings remain mixed. In one NSCLC study, higher visceral fat index (VFI) predicted worse survival (93), but results weren’t consistently replicated.

The emerging evidence strongly supports the notion that body composition, particularly skeletal muscle mass and subcutaneous fat, plays a critical role in determining outcomes for patients undergoing cancer immunotherapy. While BMI has traditionally been used as a surrogate for obesity, it remains a poor proxy for true body composition, failing to differentiate between adiposity and lean mass. Importantly, a high BMI may reflect not only increased fat mass but also greater muscle mass, both of which can confer prognostic value. This complexity challenges the oversimplified narrative of the obesity paradox. As such, reliance on BMI alone obscures the nuanced interplay between fat distribution, muscle integrity, and treatment response. Moreover, attention has to be paid in maintaining nutritional status during treatment, and routine nutritional assessment should be a key component of immunotherapy management. Moving forward, integrating direct body composition assessment with nutritional markers will be essential for optimizing patient stratification, guiding supportive care interventions. In the following sections, we will consider potential adjuvant interventions that modulate systemic metabolism, with the ultimate goal of improving efficacy of immunotherapies.

Much of the existing literature exploring the immunometabolic effects of checkpoint inhibitors relies on animal models, predominantly using rodents. While animal models provide a unique opportunity to study the systemic effects drugs in vivo, there are significant differences between human and rodent systems that should be considered.

Statins are a class of drugs used to lower serum cholesterol levels. They work by inhibiting HMG-CoA reductase, an essential element of the mevalonate pathway which is upstream of cholesterol biosynthesis. As a result, people taking statins will produce less endogenous cholesterol (163). Statins are generally prescribed to people with hypercholesterolaemia to reduce their cholesterol levels.

Figures from 2008 state that approximately 39% of the global adult population have elevated cholesterol levels (164) and more recent estimates put this figure at around 53% in the UK (165). The prevalence of the condition gives us an opportunity to compare the outcomes of patients on CPI treatment with and without high cholesterol and associated medications. Analysis has found that hypercholesterolaemia is associated with improved CPI outcomes (such as overall survival and reduced all-cause mortality) (166–168), with a potentially increased benefit in those receiving PD-1 blockade (169). It has been suggested that this effect may be due to the proinflammatory effects of increased cholesterol supporting the mechanism of CPIs (167).

Interestingly, though statins reduce cholesterol levels, their use is also correlated with improved CPI outcomes in humans (170, 171). Similar results are seen with ezetimibe (another cholesterol lowering medication) (166). At present, statins are only indicated for lowering cholesterol levels in those with elevated cholesterol, so it is not clear whether statins themselves improve CPI outcomes, or whether elevated cholesterol levels lead to long-term effects on CPI efficacy that persist despite treatment.

There is some evidence that statins modulate the immune response. It has been shown that inhibiting cholesterol biosynthesis with either atorvastatin or 25-hydroxycholesterol prevents CD4⁺ T cells from switching from a pro-inflammatory to anti-inflammatory phenotype (172). As a result, statins may slow the resolution of the immune response, which is key in the cancer context where there is chronic tumour-associated antigen presentation. Additionally, statins are known to have an antiproliferative effect, which aligns with the necessity of cholesterol for cell division (173). This could be a double-edged sword for CPI treatment, as this may affect proliferation of both cancer cells and cytotoxic T cells. The impact of this mechanism on CPI efficacy is therefore unclear.

What is clear is that statins do not negate the positive effect on CPI efficacy conferred by hypercholesterolaemia. To understand whether cholesterol-lowering drugs alone benefit CPI efficacy, it would be necessary to assess their effects in humans outside the context of hypercholesterolaemia.

Glucagon-like peptide 1 (GLP-1) is an endogenous hormone that binds to the GLP-1 receptor. The GLP-1 receptor is expressed on a wide range of human cells, and the impact of receptor binding is tissue-specific. Notably, binding of GLP-1 to pancreatic β-cells facilitates insulin release. In the gut, GLP-1 binding inhibits gastric emptying and gut motility, which reduces appetite and slows the rate of glucose absorption after eating. These effects in particular have led to the development of GLP-1 receptor agonists (GLP-1RAs) which are used for the treatment of both obesity and diabetes (174).

In vitro, GLP-1RA treatment is associated with a reduction in the inflammatory response, particularly macrophage and monocyte mediated inflammation (175). Considering the T cell compartment, exhausted CD8⁺ T cells have been shown to express GLP-1R, and there is evidence that it behaves as a negative co-stimulatory molecule, reducing T cell effector function when bound to its ligand (176). These results present GLP-1RAs as anti-inflammatory treatments which affect a range of immune cells, and this may indicate a potential negative impact on the efficacy of CPIs.

There is a relative wealth of in vivo mouse studies investigating the impact of GLP-1RAs in the general cancer context, though it is conflicting. Growth of prostate cancer was reduced in both in vitro models using prostate cancer cell lines, and in in vivo mouse models with administration of exendin-4, a GLP-1RA (177). GLP1-RAs plus metformin treatment have also been shown to have antitumour effects in pancreatic cancer in vitro (178) and breast cancer both in vitro and with in vivo mouse models (179). Contrastingly, GLP-1 receptor agonism has been shown to increase proliferation of neuroendocrine tumour cells both in vitro and in mouse xenograft models (180). Furthermore, liraglutide (a GLP1-RA) has been shown to cause thyroid C-cell proliferation in rats, the cell type from which medullary thyroid cancer originates – however, this effect was not seen in non-human primates (181). GLP-1 receptor expression varies extensively between tumours, and it’s use as a prognostic marker also depends on cancer type (174). This may explain the differences in the antitumour (or pro-tumour) effects of GLP-1RAs in different cancer models.

There is limited data available concerning the impact of GLP-1RAs on the efficacy of anticancer immunotherapies specifically. One mouse study found that liraglutide enhanced anti-tumour efficacy of PD-1 blockade via reduction of neutrophil extracellular traps, which are known tumour growth promoters (182). In humans, there is some evidence that cardiovascular IRAEs resulting from CPI treatment are reduced in those taking GLP-1RAs (183, p. 202), however, there is currently no evidence in humans regarding the efficacy of CPIs administered alongside GLP-1RAs.

It is clear that GLP-1RAs have systemic anti-inflammatory effects, and these may have an impact on CPI treatment by reducing cardiovascular IRAEs. However, it is currently unclear whether this would also attenuate the efficacy of CPIs in humans by inhibiting their proinflammatory mechanisms of action. Furthermore, GLP-1RAs appear to have varying impacts in different cancers. It remains to be seen whether this is also the case with differing CPI treatment regimes (e.g, PD-1 vs CTLA-4 blockade), or indeed whether the tissue-specificity of GLP-1 receptor binding influences the severity or location of IRAE development. With a growing cohort of patients taking GLP-1RAs for both diabetes and obesity, this presents an opportunity to assess their impact on CPI treatment and clarify the open questions in this area of research.

Beyond supportive care, specific dietary and micronutrient interventions are emerging as modulators of the tumour microenvironment and immune response. Strategies like ketogenic and fasting-mimicking diets, along with vitamin and bioactive compound supplementation, show promise in enhancing CPI efficacy, reducing toxicity, and overcoming resistance. This highlights the growing relevance of nutrition in personalized immuno-oncology.

The ketogenic diet (KD) and fasting-mimicking diets (FMD) have garnered growing interest as adjunctive interventions in cancer treatment due to their potential to modulate tumour metabolism and enhance anti-cancer immune responses. Preclinical studies demonstrate that KD and its metabolite, 3-hydroxybutyrate (3HB), enhance anti-PD-1 and anti-CTLA-4 therapies and these effects depended on T cell-mediated immunosurveillance and involved modulation of the tumour microenvironment (184). In prostate cancer models resistant to immunotherapy, cyclic KD and ketone body supplementation epigenetically reprogram both tumour and immune cells (185). In RCC, KD and ketone bodies improved mitochondrial metabolism and sensitized tumours to anti-PD-L1 therapy (186). FMD also enhanced immunotherapy efficacy and reduced immune-related adverse events in triple-negative breast cancer models (187), supporting the dual benefit of boosting response while minimizing toxicity.

Vitamins further modulate immune responses. Vitamin D supplementation corrected hypovitaminosis in cancer patients and was associated with improved OS, prolonged time to treatment failure, higher disease control rates and of reduced thyroid-related immune adverse events (188). Furthermore, vitamin C synergizes with anti-PD1 therapy by enhancing CD8⁺ T cell and macrophage function and shifting macrophage polarization from immunosuppressive M2 to pro-inflammatory M1, which is crucial for sustaining effective antitumor T cell responses (189, 190). Folate metabolism, essential for immune cell function, is another key target. Pemetrexed, a folate inhibitor, promotes immunogenic cell death and enhances T cell mitochondrial activity when combined with CPI (191), and elevated B12 has been associated with worse outcomes in CPI-treated patients, suggesting a need for targeted modulation (192). Of note, both folate and B12 are essential cofactors in DNA synthesis, methylation, and cellular energy metabolism, which are critical to the proliferation and function of immune cells. However, disruption of folate pathways may enhance anti-tumour immunity by impairing tumour cell proliferation, while anti-folates can simultaneously support T cell activation by increasing the availability of 3-phosphoglycerolphosphate and other key metabolic intermediates, thereby improving T cell metabolic fitness as suggested by Schaer et al. (191). Future studies should aim to dissect the mechanistic underpinnings of these associations and explore how targeted modulation of B12 and folate metabolism might enhance treatment outcomes.

Finally, phenolic diterpenes, such as carnosic acid and carnosol from rosemary extract (SFRE), show synergistic effects with pembrolizumab and chemotherapy in NSCLC. Clinical data further demonstrated that SFRE can reduce inflammatory and metabolic markers linked to immune suppression such as MAPK, NLRP3 inflammasome, and SREBF1 suggesting that phenolic diterpenes can remodel both tumour metabolism and systemic immune pathways, potentially amplifying the anti-tumour immune response and improving patient outcomes (193).

Together, these findings underscore the multifaceted and increasingly recognised role of ketogenic and fasting-mimicking diets, as well as vitamins and phenolic compounds, as modulators of tumour metabolism, immune microenvironment and response to CPIs. As the field advances, integrating nutritional and metabolic profiling into immuno-oncology could pave the way for novel adjunctive strategies aimed at optimizing patient outcomes through precision nutrition.