Considerations and approaches for cancer immunotherapy in the aging host

Abstract

Advances in cancer immunotherapy are improving treatment successes in many distinct cancer types. Nonetheless, most tumors fail to respond. Age is the biggest risk for most cancers, and the median population age is rising world-wide. Advancing age is associated with manifold alterations in immune cell types, abundance, and functions rather than simple declines in these metrics, the consequences of which remain incompletely defined. Our understanding of the effects of host age on immunotherapy mechanisms, efficacy, and adverse events remains incomplete. A deeper understanding of age effects in all these areas is required. Most cancer immunotherapy pre-clinical studies examine young subjects and fail to assess age contributions, a remarkable deficit given the known importance of age effects on immune cells and factors mediating cancer immune surveillance and immunotherapy efficacy. Notably, some cancer immunotherapies are more effective in aged versus young hosts, while others fail despite efficacy in the young. Here, we review our current understanding of age effects on immunity and associated non-immune cells, the tumor microenvironment, cancer immunotherapy, and related adverse effects. We highlight important knowledge gaps and suggest areas for deeper enquiries, including in cancer immune surveillance, treatment response, adverse event outcomes, and their mitigation.

INTRODUCTION

Although immune defenses are exquisitely able to detect cells bearing specific antigens and eliminate them efficiently, tumors grow with generally apparent impunity, and spontaneous immune rejection of clinically apparent cancer is rare. This immune escape is remarkable considering that cancer-driving or passenger mutations should be immunologically visible and actionable. Tumors have complex, diverse mechanisms to reduce their immunogenicity, to which aging introduces additional barriers and complexities that impede antitumor immunity and treatment efficacy. Age-associated changes in specific immune cell functions and the tumor microenvironment (TME) could increase potential for treatment-related immune adverse events (irAE) (1, 2, 3, 4, 5).

The original six fundamental cancer hallmarks (6), attributes specific to cancer cells, did not include immune effects. Current cancer hallmarks now include distinct immune and environmental properties (7) that aging can alter to influence immunotherapy efficacy or irAE. Age amplifies specific immune effects, as evidenced by a generalized inflammation with aging referred to as inflammaging (8). However, age can diminish others, such as T-cell fitness (9). We use “Age Related Immune Dysfunction” (ARID) to underscore the entirety of age-related immunity alterations (Box 1).

BOX 1. Age-related immune differences.

Immune aging is not a simple decline in numbers or functions of immune cells in young hosts. There is broadly increased detrimental inflammation (referred to by the term “inflammaging”), which can also reflect age-related senescence with production of inflammatory senescence-associated secretory products. “Immunosenescence” is an imprecise term usually referring to any distinct immune property with age or occurring in inflammatory conditions irrespective of age (19). Novel cell populations emerge with age, (e.g., PDL2-expressing T-cells (20) and CD4⁺CD25^loFoxP3⁺ regulatory T-cells (Tregs) (3)), the consequences of which are largely unreported. Most immune cells originate from hematopoietic stem cells, which also experience aging effects that alter immune outcomes (21, 22, 23). Age effects on T-cells include reduced effector functions, reduced T-cell receptor (TCR) repertoire, and reduced naïve pool, as well as increased exhaustion. Age effects on myeloid cells include increased pro-inflammatory cytokine secretion, increased numbers, and increased suppressive functions. These T-cell and myeloid-cell age effects alter cancer immunotherapy through distinct mechanisms, as detailed in the main text. B-cell contributions to antitumor immunity or immunotherapy efficacy are well established. For example, IgA contributes to ovarian cancer immune surveillance (24) and natural immunoglobulins improve antitumor immunity (25). B-cell–replete tertiary lymphoid structures (TLS) improve antitumor immunity (26), predict superior responses to anti-PD1 in human soft tissue sarcoma (27), and predict irAE risk in patients receiving anti-PD1 (28), with other notable age-associated consequences in other settings (29, 30). Plasma cells can predict response to anti-PDL1 (31), and suppressive regulatory B-cells are well-known (32, 33, 34). Aged B-cells can be inflammatory (35). Thus, B-cells likely play considerable roles in age effects on tumor immunity and immunotherapy, although little is reported. PDL2-expressing B-cells enhanced antitumor immunity in subcutaneous MC38 colon cancer or B16 melanoma via Th1 and Th17 induction (36). In a model of carcinogen-induced bladder cancer, increased age correlated with increased expression of B-cell–associated genes and greater TLS numbers (37). Many innate immune cells including NK cells, γδ T cells, NKT cells, innate lymphoid cells, granulocytes, basophils, eosinophils, and mast cells play roles in cancer immune surveillance and immunotherapy efficacy (38, 39, 40, 41), but there are few reports on their age effects in cancer (42, 43, 44, 45). The novel immune cell subsets that appear with age include those that could affect antitumor immunity, immunotherapy efficacy, or irAE (3, 46), warranting further studies.

Immunoediting is the process by which immunity promotes and prevents cancer, characterized by the three E’s: elimination, equilibrium, and escape (10, 11, 12). Equilibrium keeps tumor cells clinically quiescent but not eliminated, a process that is well-documented in mouse models (13, 14, 15) and human patients (16, 17). However, there is a finite limit on specific antigen receptors, and tumor cells under immune pressure can eventually display antigens that are no longer immunogenic, allowing escape. Tumors from older patients appear to have undergone less immunoediting versus those from younger patients (18), potentially as a result of ARID, although net consequences to antitumor immunity and treatment outcomes remain unclear.

Tumor immune surveillance and immunotherapy responses are governed by distinct immune cells, non-hematopoietic cells, non-cellular extra-cellular matrix components, and other tumor and host factors that can all be affected by age. For a full understanding of age effects on cancer immunotherapy, all of these factors influencing tumor progression must be understood. In this Review, we discuss current understanding of age effects on immunity and associated non-immune cells, the tumor microenvironment, cancer immunotherapy, and irAEs, highlighting areas in which additional knowledge is required.

EFECTS OF AGING ON IMMUNE CELLS

Effector T Cells

Diverse processes affecting T-cell numbers, diversity, phenotype, and function contribute to age-related reduced T-cell immunity (47). Notably for antitumor immunity, naïve T-cell production continually decreases with age owing to thymic involution (48, 49, 50), and T-cell precursor production in bone marrow declines (51). These changes also reduce the T-cell repertoire (52), accompanied by an age-related decline in T-cell receptor (TCR) signaling (53) that could inhibit function of antitumor T cells. Altogether, these age-associated changes in effector T cells could increase the probability and rapidity of immune escape. Immune space is gradually overtaken by age-related memory T-cell increase, including virus-reactive terminally differentiated memory T cells (54). Activating bystander virus-specific T cells in tumors can improve tumor immunotherapy (55) and might be especially useful in aged hosts.

T-cell priming declines with age, but agonist CD137-specific antibodies can improve priming (56). TNFα or IL6 treatment can improve age-related reduced CD4⁺ T-cell function in mice (57). An age-related reduction in OX40-mediated T-cell differentiation was overcome by agonist anti-OX40 to improve tumor control in mouse lymphoma (58).

Regulatory T Cells

Regulatory T cells (Tregs) impede antitumor immunity, thus, reducing their suppressive capacity has potential as a cancer immunotherapy (59, 60, 61, 62). Depleting Tregs boosts anti-tumor immunity and immunotherapy efficacy in the young (60, 62). However, age effects on Tregs (recently reviewed (63)) are contradictory, with some reports defining increased age-related Treg function and prevalence in mice and humans while others show the opposite (64, 65, 66, 67, 68, 69). Age-dependent Treg effects, including from specific subpopulations (70, 71, 72), could be tissue or disease context–dependent. For example, Tregs in aged mice can inhibit delayed type hypersensitivity (66), fail to suppress Th17 specifically in the context of chronic inflammation (73), and suppress T-cell proliferation more potently versus those in young mice (3, 74). Additionally, while Treg depletion with anti-CD25 improved antitumor immunity in aged mice in a lymphoma model (75), we reported that Treg depletion with denileukin diftitox in B16 melanoma–bearing aged mice did not inhibit tumor growth or boost tumor-specific immunity (76). Age effects on human Tregs are less reported but include increases in blood with age (77, 78), including in lung cancer patients (79). These data demonstrate that understanding ARID can develop strategies to overcome age-related immune deficiencies.

Dendritic cells

Dendritic cells (DCs) are antigen-presenting cells important in T-cell priming and providing post-antigen exposure instructions to T cells (80). However, the TME skews many DC subsets to be dysfunctional and thwart antitumor immunity and immunotherapy efficacy (81, 82).

Reduced or defective DCs can reduce T-cell function or accelerate inflammation, which could suppress tumor immune surveillance. Augmenting DC antigen presentation, which generally declines with age, can support tumor-specific T-cell surveillance. CD40/CD40L signaling declines with age, reducing DC activation in human and animal studies (83). An adenovirus vaccine expressing CD40L genetically fused to the MUC1 tumor antigen improved MUC1-specific immunity in aged mice versus controls lacking CD40L, reportedly through improved DC activation. This vaccine was to undergo clinical testing in cancer patients with breast or prostate cancers but did not advance clinically (84).

Macrophages

Macrophages are significant components of most tumor stroma and promote TME immune dysfunction affecting treatment outcomes (85, 86, 87). Tumor-associated macrophages encompass a broad spectrum of functional phenotypes ranging from pro-inflammatory M1-like to anti-inflammatory M2-like that exhibit substantial differentiation and functional plasticity. M1-like macrophages produce TNFα, IL12 and other factors that facilitate antitumor immunity. M2-like macrophages are generally anti-inflammatory, producing IL10, transforming growth factor (TGF)β, and other factors supporting tumor growth and angiogenesis (85, 86).

Myelopoiesis increases in age (88), increasing myeloid-cell numbers (89). M1-like macrophage functions can be either enhanced with age through increased reactive oxygen species or reduced (90), possibly as a result of age-related increases in M2-like macrophages making IL10 (85) and immunosuppressive TGFβ (89) in the TME. It is possible to push M2-like macrophages to a beneficial M1-like phenotype with anti-IL10 receptor plus CpG or IL12 (91). Further work is required to understand age effects on tumor-associated macrophages whose attributes likely are affected by tumor type and/or anatomic location, including strategies to push age-related M2-like macrophages to a beneficial M1-like phenotype.

Myeloid-derived suppressor cells

Myeloid-derived suppressor cells (MDSCs) are immature myeloid cells whose accumulation can be promoted by inflammation, including in cancers (92, 93, 94, 95), and potently suppress antitumor immunity through soluble mediators (e.g., arginase, IL10) or surface ligands (e.g., PDL1) (96). They can also promote suppressive M2-like macrophages and Tregs (97), which can be increased in aged hosts (98, 99). Circulating MDSCs increase with age in humans (100) and in mice in their lymphoid organs (3, 89). We have reported that denileukin diftitox–mediated Treg depletion in B16-bearing aged mice increases MDSC numbers. It did not increase MDSCs in young B16-bearing mice, suggesting that aged Tregs can control MDSCs (3). Lentinula edodes mycelia extract plus a cancer vaccine reduced MDSC numbers in aged CT26 tumor-bearing mice through IL6 and TNFα suppression and improved anti-tumor T cell cytotoxicity (101). Based on these and much additional data, suppressing MDSC numbers or functions are attractive strategies to mitigate age-related, cancer-associated immune dysfunction.

Monocytes also increase with age and can display anti-tumor properties. Their attributes are similar to tumor-associated macrophages (which derive from them) and MDSCs (102) and thus are not discussed separately here. Myeloid-targeting strategies in development (103) could be useful against monocytes.

EFFECTS OF AGING ON THE TUMOR MICROENVIRONMENT

Tumors are a complex mélange of tumor, immune and stromal cells, and matrix, collectively called TME (Fig. 1). Age effects on the TME show that it can be more immunosuppressive along several dimensions versus young TME. The TME attracts Tregs, neutrophils, and MDSCs that inhibit tumor-specific immunity, whose effects could differ in tumors in aged versus young hosts and could potentially be more immunosuppressive upon aging (89, 104). In prostate cancer, stromal fibroblasts are more inflammatory in aged versus young patients (105) and can attract CD4⁺ T cells that promote prostate epithelial cell proliferation (105). Age-related TGFβ1 increases in prostate epithelial cells in a mouse model of prostatic hyperplasia appeared to increase local inflammation (106) with similar effects noted in aged normal mouse prostates (107). An age-related decline in IFNγ signaling in the TME in murine and human triple-negative breast cancer tumors predicted reduced anti-PD1 and anti-CTLA4 efficacy (108).

Figure 1. The aged versus young tumor microenvironment.

The tumor microenvironment (TME) is comprised of tumor, immune cells, non-hematopoietic cells, soluble factors, and extracellular matrix (ECM) elements. Each cell and non-cellular element experiences age-related effects that can occur on independent trajectories. For example, long-lived immune memory cells might experience differential aging effects versus short-lived polymorphonuclear leukocytes. Each element can affect outcomes of treatments with immunotherapies, cytotoxic agents, targeted small molecules, or therapeutic irradiation. Developing effective cancer immunotherapies depends on understanding these age effects, how they can interact with other tumor elements, and how the totality of signals and effects can alter treatment efficacy. For example, aged blood vessels have differential capacity to arrest specific migratory immune cells or the efficiency of their penetration through endothelial walls (216). Once in ECM, age effects can alter how immune cells traverse the ECM or co-localize with other cells, although very little in these latter regards is reported. General reviews of aging and ECM are published (116), including various tissue-specific ECM reports. Aged blood vessels are known to affect nutrient delivery and drug penetration distinctly versus young (217). Age effects on cancer immunotherapies of cells with well-established phenotypes (e.g., innate lymphoid cells) or novel cells appearing with age (e.g., PDL2-expressing CD8⁺ T cells) are little reported. The aging TME alters metastases, reviewed in (218). Sex differences in any of these aging aspects is little reported in cancer.

The TME of aged hosts is enriched in specific soluble factors, including senescence-associated secretory products from fibroblasts that promote tumor progression, metastasis, and immunosuppressive cell infiltration, (e.g., Tregs) (109). Senescence in aged fibroblasts can promote tumor progression and angiogenesis through increased frizzled-related protein 2, ultimately increasing vemurafenib resistance (4). The aged lung microenvironment encourages the outgrowth of lung-metastatic, dormant melanoma cells through fibroblast WNT5A (109).

In addition, non-soluble factors from stromal cells in the TME can influence cancer progression and aggressiveness in the aged. FAM110B downregulation and CHI3L1 upregulation in aged cancer-associated fibroblasts predicted poor low grade glioma prognosis (110). Chitinase 3-like 1 promotes PDL1 expression in B16 melanoma pulmonary metastases, implicating aged fibroblasts in immunotherapy efficacy (111). Agonist anti-OX40 cancer immunotherapy efficacy was hindered in old mice owing to defective TME OX40 signals and not directly from aged T cell–intrinsic properties (112).

Metabolic changes in aged host TME can affect cancer treatment outcomes. For example, aged fibroblast fatty acid metabolism promotes BRAF/MEK inhibitor resistance in melanoma cells, which can be overcome through inhibiting the fatty acid transporter FATP2 (113). Increased TME angiogenesis but decreased vascular endothelial growth factor are associated with increased age in humans, contributing to bevacizumab failure (114). sFRP2 inhibition overcomes bevacizumab failure in aged versus young mice (114). Further insights into age effects on TME are merited based on these data, which are summarized in Table 1.

Table I.

Age-related effects on specific tumor microenvironments.

TME	Effects
Prostate cancer stromal fibroblasts	These are more inflammatory in aged versus young patients and can attract CD4+ T cells promoting prostate epithelial cell proliferation (104).
	Age-related TGFβ1 increase in prostate epithelial cells in a mouse model of prostatic hyperplasia appeared to increase local inflammation (105).
	Age-related TGFβ1 increase is seen in aged versus young prostate epithelial cells in normal mouse prostates (106).
Triple negative breast cancer	Age-related decline in IFNγ signaling in murine and human triple-negative breast cancer tumors predicted reduced anti-PD1 and anti-CTLA-4 efficacy (107).
Cutaneous melanoma	Aged dermal fibroblast HAPLN1 promoting melanoma metastasis including through inhibiting polymorphonuclear immune cell migration and decreasing Treg recruitment (116).
	Aged dermal fibroblast HAPLN1 inhibits melanoma lymphatic metastatic spread (117).
	Senescent aged fibroblasts can promote tumor progression and angiogenesis through increased frizzled-related protein 2 that promotes vemurafenib resistance (4).
	Aged fibroblast fatty acid metabolism promotes BRAF/MEK inhibitor resistance in melanoma cells (112).
	Increased angiogenesis and decreased vascular endothelial growth factor occur with age in human cutaneous melanomas that promote bevacizumab failure (113).
Lung metastatic melanoma	Chitinase 3-like 1 promotes PDL1 expression in lung to influence αPD1 checkpoint blockade immunotherapy of pulmonary B16 metastases (110).
	Fibroblast WNT5A supports outgrowth of lung-metastatic, dormant melanoma cells (108).
Glioblastoma multiforme	FAM110B downregulation and CHI3L1 upregulation in aged cancer associated fibroblasts predicted poor low grade glioma prognosis (109).
Sarcoma and colon carcinoma	Anti-OX40 agonist cancer immunotherapy efficacy was hindered in old mice owing to defective TME OX40 signals and not directly from aged T cell-intrinsic properties (111).
Gut microbiota	Gut microbial composition changes with age and has significant effects on systemic immunity (182–184) that could affect ICB efficacy.

The extracellular matrix (ECM) of the TME is composed of proteoglycans, hyaluronic acid, and cellular adhesion proteins such as collagen, fibronectin, and laminin and is an organ-specific critical component of tumor stroma that contributes to tumor development (115, 116). The dermal ECM is highly studied due to ease of tissue collection. Young skin has a collagen-rich, highly organized ECM while aged skin ECM exhibits poor structural integrity. Dermal fibroblast secretion of HAPLN1, a hyaluronic and proteoglycan link protein, is reduced in aged fibroblasts, ultimately promoting melanoma metastasis. Constitutive expression of fibroblast HAPLN1 inhibits primary melanoma tumor metastasis (117) and melanoma metastatic spread through lymphatics (118) in the aged TME, and is associated with better mononuclear immune cell migration and decreased polymorphonuclear immune cell and Treg recruitment in aged melanoma-bearing hosts (117). All these effects could affect immunotherapy strategy development and collectively suggest that age-specific ECM approaches could be applied in selected TME.

EFFECTS OF AGING ON IMMUNE CHECKPOINT BLOCKADE

Immune checkpoints are elevated in many cancers, and thus impair antitumor immunity (119, 120). Blocking negative signals from immune checkpoints with specific antibodies can improve antitumor immunity (119, 120) and is now FDA-approved standard of care and first line therapy for selected cancers. Nine distinct agents targeting four different immune checkpoints (PD1, PDL1, CTLA4, and LAG3) are now FDA-approved. Expression of immune checkpoint molecules associated with exhausted T cells, including PD1, Lag3, and Tim3, increases with age (3, 20, 121, 122). Other immune checkpoint receptors such as PDL1 are highly expressed on young myeloid cells and B-cells but can be expressed at much higher levels on tumor cells and aged immune cells (20, 123). PDL2 and CD80 (B7–1), an immune checkpoint ligand for CTLA4 and important factor in T-cell activation, also increase with age in a variety of immune cell populations (20). Aging reduces expression of human T cell CD28, the co-stimulatory molecule expressed by naïve and primed T cells that binds to CD80 and CD86 (B7–2) (124). CD80 and CD86 bind to both CD28 and CTLA4. Thus, loss of T cell CD28 in age could result in preferential binding of CD80/CD86 to CTLA4 resulting in inhibition and potentially stronger immunosuppression in age. The functional consequences of most of these observations in aged hosts remain unreported.

Age-specific strategies to mitigate inhibitory immune checkpoint signals and reduce hyporesponsiveness or exhaustion of aged T cells without compromising antitumor effects is an important goal in the development of these modalities. Table 2 summarizes notable treatment strategies aimed at improving antitumor immunotherapy by mitigating ARID features.

Table II.

Selected strategies to mitigate ARID for promotion of better cancer immunotherapy.

Strategy	Outcome
αPD1	Superior efficacy in aged versus young mice bearing YUMM melanoma (128) NR-S1K oral squamous cell carcinoma (141) presumably through increased T cell PD1 expression with age.
αPDL1	Superior efficacy in aged versus young mice bearing NR-S1K oral squamous cell carcinoma (141) presumably through increased T cell PDL1 expression with age.
αCTLA4	Superior efficacy in aged versus young mice bearing NR-S1K oral squamous cell carcinoma (141) presumably through increased T cell CTLA4 expression with age.
αGr1 + denileukin diftitox	αGr1 inhibited MDSC increase after denileukin diftitox-mediated Treg depletion aided Treg depletion in aged but not young hosts bearing B16 melanoma (3).
Lentinula edodes mycelium extract plus a cancer vaccine	Reduced MDSC numbers in aged CT26 tumor-bearing mice through IL6 and TNFα suppression and improved anti-tumor T cell cytotoxicity (100).
αCD137	Can improve priming of aged T cells (56).
αOX40	Improves T cell differentiation to improve tumor control in mouse lymphoma (58).
Adenovirus CD40L vaccine	CD40L genetically fused to MUC1 improved MUC1+ tumor control reportedly by improving aged DC activation (83).
αPDL2	Improved B16 and NCH1 melanoma control in aged, not young mice through IL17-dependent mechanism (130).

Anti-PD1

The PD1 immune checkpoint molecule inhibits antitumor PDL1⁺ T cells and PDL2-expressing immune cells (125). Monoclonal PD1-specific antibodies are FDA-approved to treat a variety of carcinomas and certain lymphomas and can be remarkably effective.

PD1 is highly expressed by effector-memory (CD44^hiCD62L^lo) T cells, whose prevalence increases with age. Although aged mouse CD4⁺PD1⁺ T cells proliferate poorly (126, 127), anti-PD1 improves aged T-cell functions in vivo in mice (128). Aged mice bearing YUMM melanomas experienced improved anti-PD1–mediated tumor control versus young mice and lower tumor Treg infiltration (129). In contrast, in triple-negative breast cancer, diminished IFNγ signaling by tumors in older mice and humans correlated with reduced anti-PD1 and anti-CTLA4 response (108).

Chronic oral administration of rapamycin, an mTORC1 inhibitor, reduced age-related T cell PD1 expression and improved PD1⁺ T-cell function in mice (127). These and other data provide much pre-clinical and clinical research to suggest pharmacologic approaches to improve human immune performance in advancing age (130). PDL2 is a PD1 ligand that has received little cancer immunotherapy research attention until recently. We found that anti-PDL2 failed in young B16-challenged mice but, remarkably, was effective in aged mice (131), the first single-agent cancer immunotherapy that is effective in aged but not young hosts to our knowledge. Efficacy included IL17-dependent effects, which have been unreported for FDA-approved immune checkpoint inhibitors and are reminiscent of beneficial IL17 effects on aged lung-metastatic B16 (43) and human lung cancer (132), but require additional study.

Anti-PDL1

The PDL1 immune checkpoint is a PD1 ligand. As for anti-PD1, anti-PDL1 is generally considered to protect PD1⁺ antitumor immune cells (mostly CD8⁺PD1⁺ T cells) from inhibition through tumor PDL1 (120, 133, 134, 135, 136). Three distinct anti-PDL1 antibodies are FDA-approved for cancer treatment.

PDL1 expression on aged immune cells in mice, notably CD8⁺ T cells, is considerably higher than in young mice, including de novo expression on cells that are typically PDL1-negative in young mice (20). Aged CD8⁺ T cells exhibited higher levels of PDL1 versus those from young mice and relatively low proliferation, but anti-PDL1 increased proliferation in vitro to a level comparable to young CD8⁺ T cells (123). Anti-PDL1 boosted antilymphoma immunity in aged Balb/c mice to that observed in young lymphoma-bearing mice (123). We reported that anti-PDL1 immune checkpoint blockade (ICB) failed to treat B16 melanoma in aged BL6 mice, whereas it was effective in young mice (76). However, anti-CTLA4 modestly improved efficacy (137). A study on glioblastoma found that aged versus young mice had decreased response to anti-PDL1 and anti-CTLA4 combination therapy (138). These differences could be due to tumor (lymphoma versus melanoma), host genetic background, or hematologic versus non-hematologic malignancy among other considerations, and require further investigation. Moreover, these data suggest that single-agent efficacy might not predict combination efficacy in the aged (or young) and that specific agents can help overcome age-related reduction in ICB efficacy, as we also reported for the distinct approach of denileukin diftitox–mediated Treg depletion (3).

Anti-CTLA4

In 2011, the CTLA4-specific antibody ipilimumab became the first FDA-approved immune checkpoint inhibitor following demonstration that it improved advanced/metastatic melanoma disease-free and overall survival (139). We showed that anti-CTLA4 controlled B16 melanoma growth in aged mice but not as effectively as in young, which could partially be explained by reduced Treg depletion in aged mice (140). In NR-S1K oral squamous cell carcinoma, increased PD1 and CTLA4 expression on T cells of aged versus young mice correlated with increased anti-PD1, anti-PDL1, and anti-CTLA4 efficacy (141).

Anti-Lag3

In 2022, relatlimab became the first FDA-approved Lag3-specific antibody, indicated for use in advanced melanoma, but only in combination with nivolumab. Age-related immune outcomes of Lag3 are little reported, aside from a study showing that Lag3 expression increases with age (142), and require further investigation.

Immune Checkpoint Agonists

CD137-specific and OX40-specific agonist antibodies (56, 58) have undergone human cancer clinical trials and are described above. Age effects on OX40/OX40L signals have been reported (58) but not on CD137 to our knowledge.

EFFECTS OF AGING ON ADOPTIVE CELL TRANSFER

Adoptive transfer of tumor-infiltrating lymphocytes or DCs have generally not demonstrated significant antitumor efficacy. Sipuleucel-T, an FDA-approved cell product for certain prostate cancer patients, is called a DC vaccine, but the precise nature of adoptive cells mediating clinical efficacy is incompletely defined. As age greatly affects myeloid cells, we expect age effects on Sipuleucel-T and adoptive DC production, but little has been reported.

CAR T cells

CAR T-cell therapy is autologous adoptive cell transfer of peripherally circulating T cells that have been genetically manipulated to express a chimeric-antigen receptor (CAR) (143). Six CAR T-cell products are FDA-approved. Donor age (and T-cell fitness) affects the generation of sufficient cells of sufficient function for optimal treatment efficacy (9) but has not stopped adoption of this treatment approach. Major adverse events include cytokine storm and central nervous system depression.

Landmark clinical trials in diffuse large B-cell lymphoma (DLBCL) constitute the majority of clinical data supporting high response rates and efficacy of CAR T-cell therapies in older patients (144, 145, 146, 147). Despite these high response rates, some data suggest that older DLBCL patients treated with CAR T-cell therapies experience higher toxicity (148). One study found that IL37 augmented CAR T-cell efficacy in aged mice transplanted with leukemia cells, demonstrating the impact of age-effects on CAR T-cell therapy (149). Further investigation of the impact of aging on the efficacy of CAR T-cell therapy and novel ways to improve treatment responses in aged hosts is warranted.

Other CAR-bearing immune cells

NK cells, γδ T cells, and NKT cells are effective in a major histocompatibility-independent manner, although major histocompatibility antigens can affect their function (38, 39, 40, 150). CARs engineered into these allogeneic compatible cells are being investigated as off-the-shelf products free from donor constraints and could be especially useful in aged patients. Bone marrow chimeras between aged and young hosts showed that aged NK-cell function was restored in the non-hematopoietic environment of the young host (151), suggesting an environment-specific and cell-specific effect of aging on hematopoietic stem cell and immune cell function that could also be exploited.

EFFECTS OF AGING ON TOLL-LIKE RECEPTOR AGONISTS

Bacille Calmette-Guérin (BCG) is an attenuated Mycobacterium bovis and a potent Toll-like receptor (TLR) agonist that is FDA-approved to treat non-muscle invasive bladder cancer. Although its mechanism of action is not fully understood, it improves antitumor immunity, which we reported to include conventional and γδ T-cell contributions (152). A clinical trial of BCG plus IFNα found that aged versus younger bladder cancer patients relapse at higher rates (153). Retrospective studies show higher relapse rates with age after BCG treatment (154, 155). However, confounding effects of post-surgical management, associated morbidities, and age-related immune effects are insufficiently described in these reports to permit firm mechanistic conclusions. As the median age at diagnosis for bladder cancer patients is 73 (156), full understanding of the immune mechanisms behind poor BCG efficacy in aged patients is highly clinically relevant and warrants further investigation. Other TLR agonists have been shown to enhance immunotherapy efficacy in pre-clinical models of cancer and in clinical trials. However, their effects in aged hosts are largely unstudied (157).

MITIGATING ARID IN THE CONTEXT OF CANCER IMMUNOTHERAPY

mTORC1 Inhibitors

mTORC1 inhibition can have positive effects on T-cell functions, including in the aged and in cancer. mTORC1 controls T-cell differentiation fates, including suppressing Treg function and differentiation (127, 158). Thus, mTORC1 inhibition could improve antitumor immunity by improving antitumor T-cell numbers and/or functions or by reducing Treg suppression (60). Numerous studies now validate that low dose, small molecule first-generation mTORC1 inhibitors produce net antitumor immunity benefits without increasing Treg numbers or function significantly (127, 159, 160).

We have reported that the mTORC1 inhibitor rapamycin improves PD1⁺ T-cell functions, reduces T cell PD1 expression in aged mice (127), and improves γδ T cell–mediated antitumor immunity in an induced skin cancer model (161). Thus, mTOR inhibition could be a strategy to improve cancer immunotherapy efficacy in aged patients (127, 162) and could have additional health or longevity benefits (163). The mTORC1 inhibitor everolimus improved influenza vaccine–induced antibody production in aged subjects in a clinical trial (164). Thus, mTOR inhibition could be a strategy to improve cancer immunotherapy efficacy in aged patients (127, 162) and could have additional health or longevity benefits (163). Several biotechnology companies are developing next-generation mTORC1 inhibitors with more specificity for better beneficial immune outcomes that could enter clinical trials shortly.

Caloric Restriction

In some models, caloric restriction prolongs life and health span and suppresses mTORC1 (165), pharmacologically recapitulated by small molecule mTORC1 inhibitors. It enhances anti-CD40 responses in sarcoma and breast cancer in aged mice, potentially by improving antigen-specific CD4⁺ T-cell priming (166). Caloric restriction improves antitumor immunity in young hosts, but aged patients do not tolerate all caloric restriction regimens. Some caloric restriction studies have been found to have detrimental effects in the aged (167) and others have indicated that mTORC1 inhibition in aged hosts could reduce immunity (49). Caloric restriction is simple and inexpensive and merits investigations for appropriate and effective applications in aged hosts.

EFFECTS OF AGING ON IMMUNOTHERAPY TOXICITIES

The most serious toxicities associated with immunotherapies are irAEs, or inability to tolerate treatment, the development of which has been assessed to a limited extent in aged populations. irAEs were increased (47% versus 24% in historical controls) in advanced-stage melanoma patients aged 75–92 years receiving anti-PD1 ICB with either of pembrolizumab or nivolumab, and/or anti-CTLA4 ICB with ipilimumab (168). Aged (≥70 years old) patients experienced a significantly higher high-grade irAE rate (33%) versus 25% in young patients (169). A study of irAE in elderly melanoma patients found 35.1% receiving ICB developed ≥1 irAE regardless of pre-existing autoimmune disease (170). A large retrospective study of elderly patients with any solid tumor receiving any FDA-approved ICB found no clear change in overall irAE development but some small differences in subgroups (171). Another retrospective analysis found age was a significant risk for autoimmune irAE (172). Frailty in older patients did not correlate with increased number of irAEs but correlated with increased hospitalization for irAEs and longer hospital stays (173).

Multiple irAEs in an individual patient are unusual. A case report described an 85-year-old male who developed multiple autoimmune irAEs (type 1 diabetes, pneumonitis, hypothyroidism, and polymyalgia rheumatica) following pembrolizumab treatment for melanoma, which resolved with treatment cessation and steroids (174). The authors proposed age-related T-cell inflammation as contributing to this multiplicity of irAEs.

Generally, mice do not exhibit the significant autoimmune-related adverse events experienced by humans treated with ICB, but this difference may provide insight into irAE mechanisms in aged patients. Age-associated CD4⁺ T-cell production of CXCL13 is associated with tertiary lymphoid structure development and irAEs following anti-PD1 treatment in human patients, observations that have been recapitulated in aged mice (28). Aged macrophages in both mice and humans generally produce significantly more TNFα and IL6 versus younger macrophages (175), also seen in aged microglia in brain (176). Suppressing TNFα can improve survival and mitigate irAEs while preserving immunotherapy efficacy in aged mice (1, 2). Macrophages, increased in aged versus young (127), are often the most abundant immune cell in the TME and can potently drive cancer-associated inflammation through TNFα, IL1β, and IL6.

In humans, treatment with anti-TNFα can mitigate autoimmune irAEs related to ICB. The FDA-approved TNFα-specific antibody infliximab effectively mitigates irAE colitis when given concurrently with ICB and treats initial colitis in combination with corticosteroids more efficiently than corticosteroids alone, both without altering ICB efficacy (177, 178, 179). Likewise, IL6 blockade/neutralization alleviates immunotherapy toxicities and can be given concomitantly with ICB to maintain clinical benefit (180, 181).

ADDITIONAL AGE-EFFECTS REQUIRING CONSIDERATION FOR CANCER IMMUNOTHERAPY

Microbiome Effects

Gut microbial composition changes with age and has substantial effects on systemic immunity (182, 183, 184). These gut microbes influence antitumor immunotherapy efficacy in humans receiving anti-PDL1 (185), anti-PD1 (186), and anti-CTLA4 (187). Age effects on the gut microbiome in the context of ICB efficacy are little reported but are predictable given its influence on anti-cancer ICB through immune modulation (188), including altering myeloid cells in the TME (189), which could help explain known age effects (127).

Age-Related Tumor Differences

Age affects tumor-intrinsic properties, including the tumor genomic landscape and epigenetic modifications (190, 191). For example, specific gene expression differences were noted in younger (<40 years) versus older (>50 years) patients with colorectal cancer with both microsatellite stable and microsatellite instable disease (192). Age-related mutational signatures in human prostate cancer patients predicted disease progression (193), and age effects on the molecular background of tumors in pan-cancer analyses have also been shown (194, 195, 196). Further, aging diminished IFNγ signaling in mice and human tumors and predicted reduced anti-PD1 and anti-CTLA4 ICB efficacy (108). In addition, we showed increased tumor expression of selected immune checkpoints in aged mice and humans, including PDL1 and PDL2 (20), in part from age-related differences in TME cytokine content. These age-related effects on the tumor require additional studies to assess roles in treatment selection and response biomarkers.

Age-Related Sex Differences

Little has been reported regarding age-related sex differences in cancer immunotherapy, and sex-specific aging-related changes to the immune system have been incompletely characterized. However, generally males experience these changes at a more accelerated rate than females. We reported that female mice responded to anti-PDL1 immunotherapy against B16 melanoma better than males in a host PDL1-dependent manner, and this was associated with host PDL1-dependent impaired Treg function in young, but not aged PDL1^KO females (197). Older human males have reduced total B-cell and T-cell numbers versus females and older females have more NK cells versus males (198, 199). These observations can affect clinical outcomes, specifically in vaccine efficacy (200), but their role in cancer immunotherapy remains largely unstudied.

THE EFFECTS OF ARID IN HUMAN TRIALS OF CANCER IMMUNOTHERAPY

Clinical data regarding age effects on cancer immunotherapy are emerging but remain relatively underreported and largely descriptive without specific mechanistic insights, despite excellent data reporting. Bioinformatic analyses suggest that markers of improved ICB response increase with age (142), supporting the notion that the aged might respond better to (selected) immunotherapy approaches (201), but this concept requires further evaluation. Major challenges in developing age-appropriate cancer immunotherapies are outlined in Box 2.

BOX 2. Major challenges in developing age-appropriate immunotherapies.

Age affects tumor immune and stromal cells, non-cellular matric components, and soluble factors, any or which could affect the efficacy of immunotherapies, cytotoxic agents, targeted small molecules, or therapeutic irradiation. A deeper understanding of these mediators is needed as they are little reported. As most therapies work better in the presence of a well-functioning immune system, especially T cells, improving T-cell functions could augment response to cancer chemotherapy, small molecules, or irradiation responses in the aged (104) but is as-yet unreported. Immune checkpoints and expression of other immune regulatory molecules, donor cell fitness for cell infusion approaches, cell and host metabolism (202), immune exhaustion (203, 204), and many other factors can also affect distinct therapies (see Fig. 1), but all areas have knowledge gaps addressed in the main text. Cell-intrinsic properties or those altered by the aged TME are also areas requiring additional study. For example, inhibiting mitogen-activated protein kinases (205) or mTORC1 signals (206) could improve aged T-cell function and age effects on tumor-intrinsic signals (108), immune checkpoint expression (20), and genetic/epigenetic outcomes (190, 191) that could alter treatment efficacy. Attempts to develop age-related ICB response algorithms/predictive models have been reported (207), but none are well validated.

Distinct meta-analyses of trials of various immune checkpoint inhibitors found that aged (>60 years) versus young (<60 years) melanoma patients experienced similar treatment efficacy (208, 209, 210, 211). Another meta-analysis defining aged cancer patients as >50 years and younger as <50 years strikingly found that older age predicted better efficacy outcomes and suggested that patient age could be a negative ICB response biomarker (201). In contrast, a separate meta-analysis defining aged patients as ≥75 years old showed that cancer progression-free survival from single-agent anti-PD1 or anti-PDL1, or combined anti-PD1 plus anti-CTLA4 was improved in younger (<75 years old) patients, compared with aged patients(212). A meta-analysis of eight randomized clinical trials of distinct immune checkpoint inhibitors (anti-PD1, anti-PDL1, and anti-CTLA4) in 5,487 patients with non-small cell lung cancer, of whom nearly half were >65 years old (“aged”), showed no difference in overall survival in young versus aged patients. In addition, progression-free survival studied in seven of those eight trials showed no age distinction (213).

These contradictory data could relate to differences in study populations, supportive care, or ICB use distribution, single-agent versus combination, and/or differences in how “aged” is defined. Lack of differences among studies could relate to underpowered cohorts, which itself could reflect relatively low trial enrollment of older patients (214). It is unlikely that age does not affect ICB efficacy given the well-documented age effects on many other immune-related outcomes such as anti-pathogen immunity and vaccine responsiveness (215). Thus, differences in ICB efficacy might not appear until very advanced age, in which case varying nutritional status and co-morbidity burden will likely influence data.

CONCLUSIONS

Many insights into antitumor immunity and cancer immunotherapy response were developed through studies of young hosts. Knowledge gaps on age effects on immunity, immunotherapy efficacy, and irAE are being filled, but much remains to be defined in detail. Nonetheless, cancer immunotherapies with improved efficacy in aged hosts, and treatment strategies that aid older but not younger subjects are now reported, with new approaches sure to follow.

Important areas for further study include more detailed assessments of age effects on the TME, including specific immune, non-immune, and non-cellular components. Differences in tumors arising in aged versus young hosts such as mutations, epigenetic changes, and signal differences are described, but much remains to be understood regarding the effects of these changes on tumor progression and aggresiveness, as well as potential strategies to target these differences to improve therapies or mitigate irAEs in aged patients.

As combination treatments are generally superior to single agents, insights into better combinations for aged patients are required. Adjuncts to consider include agents increasing tumor immunogenicity, adoptive transfers of allogeneic cells with better fitness versus aged donor cells or means to improve fitness of autologous cells, means to improve antigen presentation, thymic transplants, gene therapy aimed at specific age-associated defects, microbial alterations, novel molecules, and caloric restriction among other approaches. Regarding novel agents, most studies of aging effects on cancer immunotherapy relate to immune checkpoint blockade. We must await data on age effects on the myriad of new agents in clinical trials.

“Immune decline” of an aged immune system, “immunosenescence,” and “inflammaging” refer to specific subsets of alterations in age. We use “Age Related Immune Dysfunction” (ARID) to encompass the totality of age-related immune changes that include appearance of novel immune cell subsets, increased inflammation or immunosuppression, changes in immune cell content, reduced TCR repertoire, increased memory cells, and the numerous other age-related changes. It is now clear these challenges can be overcome using approaches based on understanding the underlying immunopathologic effects.

Finally, a better understanding of age-related changes to tumor cells, the genes that tumors express, and alterations in specific immune environments will help improve tailored treatments with enhanced efficacy. Ethical and cost considerations require examination as we use ever more potent treatments in an aging patient population. Initially counterintuitively, bioinformatic, pre-clinical, and clinical data suggest that in some cases, the aged will respond better to selected immunotherapies versus the young. As we gain insight into ARID, the mechanistic basis for this apparent paradox is becoming clearer, underscoring the fact that the aged can also teach us about how to improve cancer immunotherapies in younger patients.

Abbreviations used:

ARID

Age Related Immune Dysfunction

BCG

Bacille Calmette-Guérin

CAR

chimeric antigen receptor

CDMRP

Congressionally Directed Medical Research Program

CD40L

CD40 ligand

CpG

poly-(cysteine 5’ to guanine)

CRISPR

clustered regularly interspaced short palindromic repeats

CTLA4

cytotoxic T lymphocyte antigen-4

DC

dendritic cell

DNA

deoxyribonucleic acid

IL

interleukin

Lag3

lymphocyte activation gene 3

M1

type 1 (macrophage)

M2

type 2 (macrophage)

MDSC

myeloid derived suppressor cell

mTORC1

mammalian target of rapamycin complex 1

PD1

programmed death-1

PDL1

programmed death ligand-1

PDL2

programmed death ligand-2

Th17

T helper cell 17

TGFβ

transforming growth factor-β

Tim3

T-cell immunoglobulin and mucin-domain containing-3

TNFα

tumor necrosis factor-α

Treg

regulatory T cell