The aging immune systemImmunodeficiency, inflammaging, and immunobiography

Age-related changes are present in every organ system. However, the impact of aging on immunometabolism has long been of particular interest. The term immunosenescence has been coined to summarize the changes occurring in the immune system with aging.6 Reciprocally, chronic inflammatory signaling is considered to be a major driver of aging-related physiological processes (inflammaging).7 Both processes fuel each other, that is, inflammation drives the aging process, while aging, in turn, influences inflammatory signaling dynamics.8

Aging affects both innate and adaptive immunity. Each organism encounters a distinct repertoire of immune cues during its life cycle, resulting in an individualized immunobiography. Repetitive immune recognition of the same stimulus enables a potent adaptive immune response, establishing immunological memory. More recently, the existence of immunological memory in innate immunity, mediated by the epigenetic inscription of previous immune encounters, has been demonstrated.9 This trained immune memory confers an adaptive advantage to the aging organism, however, when repetitive immune stimulation surpasses the adaptive threshold, the immune system will decrease its responsiveness, entering a state of “anergy”.10 11

T cell subsets and exhaustion during aging

Adaptive immunity relies on somatic recombination in order to generate a near-limitless receptor repertoire. Adaptive immune cells proliferate on antigen recognition, resulting in a unique receptor repertoire as the result of clonal selection.12 In cellular immunity, T cells differentiate into various subsets with distinct functions. While the amount of antigen stimuli inevitably increases with an organism’s age, substantial attrition of the T cell repertoire in aged individuals has been demonstrated. Furthermore, it has been shown that proportions of T cell subsets vary considerably across an organism’s lifespan.13

The proportion of naïve T cells, which account for approximately 90% of the T cell population in cord blood, sharply decreases with age, representing only 25% of peripheral blood T cells in older individuals.14 As naïve T cells have immense proliferative capacity and are able to differentiate into memory T cell subsets, their decreased abundance may result in impaired responsiveness to novel stimuli. This potentially explains the decreased receptor repertoire in aged individuals. However, it is thought that their absence can be partly compensated by homeostatic proliferation and stemness features of memory T cells.15 Conversely, the memory Tcell population expands in both the CD4+ and the CD8+compartment.16

T cell exhaustion describes a hyporesponsive state following prolonged antigen stimulation in which T cells generally reduce their biological activity including cytokine secretion and augment expression of inhibitory checkpoint molecules, such as programmed cell death protein (PD-) 1, cytotoxic T-lymphocyte-associated protein (CTLA-) 4, and T cell immunoglobulin and mucin domain (TIM-) 3; T cell exhaustion increases with age and presumably reflects the above-mentioned loss of versatility in the aging immune system.17 Correspondingly, an increased abundance of exhausted T cells in the tumor microenvironment has been associated with reduced survival in multiple cancer types.18 There is increasing evidence indicating that reinvigoration of exhausted T cells is an actionable feature of T cell-mediated immunotherapy.19 However, T cell exhaustion is now seen as a complex epigenetic program, and most approaches focusing on inhibiting immune checkpoints only partially reverse its phenotype.20

Cellular senescence

Cellular senescence is a stress- and aging-induced state resulting in a terminal cell cycle arrest.21 Especially T cell senescence is often induced by chronic inflammation.22 In addition, therapy-induced senescence by prior genotoxic therapies can occur in cancer and bystander cells which might have an impact, especially in heavily pretreated patients receiving immunotherapies.22 Senescent cells demonstrate metabolic and increased secretory activity, referred to as senescence-associated secretory phenotype (SASP).23 SASP describes the secretion of multiple pro-inflammatory molecules, fueling inflammaging.24 In turn, inflammaging stimulates the development of senescence in T cells. As such, the SASP is another example of a reciprocal relationship between aging-associated phenotypes. Of note, the loss of CD28, a co-receptor necessary for the full activation of T cells, as a surface marker is a prominent marker of T cell senescence.25

Clonal hematopoiesis and aging

Recently, clonal hematopoiesis of indeterminate potential (CHIP) was discovered as a major factor in aging-associated pathologies26 and is now seen as a central phenomenon of inflammaging. CHIP is defined as the presence of a hematopoietic clone carrying a hematological malignancy-associated mutation without an underlying hematological malignancy. Diagnosis of CHIP requires a variant allele frequency of ≥2%.27 28 Many CHIP mutations relate to a proinflammatory state.29–31 Besides a slightly increased risk for myeloid malignancies,32 33 CHIP is strongly associated with cardiovascular diseases.29 Recently, lymphoid CHIP, defined as CHIP with mutations otherwise known from lymphoid malignancies, gained new interest.34 CHIP in the lymphoid compartment can be acquired in the early hematopoietic stem cell department or in the mature peripheral B cells or T cells due to the proliferative capacities of lymphocytes. Thus, CHIP in the stem cell or early progenitor compartment will likely generate B cell/T cell receptor diverse CHIP clones; mutations in peripheral B cells and T cells receptor homogenous CHIP clones.34 Both potential pathways suggest different outcomes based on the effector cell. CHIP mutations affecting proliferation and tolerance checkpoints potentially result in lymphomagenesis, autoimmunity, or augmentation of a T-cell activation.34 So far, only a few exemplary presentations of such scenarios have been demonstrated.35 36 Thus, its clinical impact and underlying mechanism need to be fully elucidated.

Impact of immunosenescence on immune defense towards cancer and infections

Immunosenescence and inflammaging have been linked to carcinogenesis,37–39 an impaired pathogen defense and a decreased responsiveness to non-older subjects adapted vaccination40 due to a decreased versatility in reacting to endogenous and exogenous immune stimuli. Clinically overt cancer results from an immune escape.41 As the adaptive immune response to malignancy requires a sufficient pool of naïve T cells, the increased incidence of cancer with aging is a rational consequence of immunosenescence. Furthermore, the increase of immunosuppressive regulatory T cells (Treg) and exhaustion phenotypes contributes to impaired immunosurveillance.42 Of note, age-related alterations in the innate immune response, such as decreased activity of antigen-presenting cells and dysregulation of pattern recognition receptor signaling, also contribute to carcinogenesis.43

In this context, the concept of cancer immunoediting plays a major role. Cancer immunoediting refers to the ability of the immune system to constrain and promote cancer development and progression paradoxically in a triphasic process.44 In the elimination phase, innate and adaptive immunity eliminate a wide range of transformed cells. Selected tumor clones surviving this elimination enter the equilibrium phase characterized by limited net tumor growth. The ongoing selection pressure by the immune attacks finally induces an immune escape of tumor cells that acquire genetic resistance towards such an elimination, leading to clinically overt cancer.44 Of note, this process can also occur in response to treatment.

Nevertheless, the human body appears to adapt to age-related changes in the immune system, reflected by the observation that cancer incidence decreases after the eighth life decade.45 CD4+CD28− T cells have been suggested to play a protective role in older individuals.45 Furthermore, it has been hypothesized that some CD8+ memory T cells, which are increasingly abundant in older individuals, may acquire innate characteristics rendering them efficient to combat malignant transformation.46

Although data on age-related changes in the tumor microenvironment are scarce, an influence of aging is suggested for several entities.1 Furthermore, the tumor-promoting effect of immunosenescence at higher age suggests a therapeutic vulnerability towards the restauration of immune responses against cancer. Thus, treatment approaches targeting these vulnerabilities might be especially effective.

Age-related changes in the immune system are depicted in figure 1.

Figure 1

The interplay between aging and immunity. What is termed “biological age” largely reflects the degree of senescence-associated alterations of an organism’s physiology. Senescence of the immune system is a potential consequence of compensatory immunosuppression due to persistent inflammatory signaling, whereas, at the same time, senescent immune cells contribute to chronic inflammation. As such, inflammatory signaling and senescence fuel each other and constitute the molecular basis of aging. IFN-y, interferon gamma; IL-6, interleukin-6; IL-1, interleukin-1; MHC, major histocompatibility complex; NK, natural killer; TCR, T-cell receptor; TNF, tumor necrosis factor.

CAR T-cell therapy

For CAR-T therapy, autologous T cells are genetically modified to express a CAR targeting a tumor surface antigen, thus re-enabling an effective T cell response to malignant cells.47 The unprecedented efficacy of CAR-T therapy has transformed the treatment landscape of various relapsed or refractory (R/R) B cell lymphoma and multiple myeloma (MM). However, CAR-T therapy has the potential for class-intrinsic, severe complications such as cytokine release syndrome (CRS),48 neurotoxicity (immune effector cell-associated neurotoxicity syndrome, ICANS),48 and hematotoxicity,49 and remains largely ineffective in solid tumors.

Older adults with R/R lymphoma and MM are historically precluded from intensive therapy approaches such as high-dose chemotherapy with autologous stem cell transplantation (ASCT). The unprecedented success of CAR-T therapy holds the promise to overcome this limitation, yet there is concern that immunological aging in the T cell compartment may affect both the efficacy and safety of CAR-T therapy. This section will synthesize available preclinical and clinical evidence on CAR-T therapy in aged individuals.

Potential influence of immunological aging on CAR-T cell efficacy

A growing body of evidence suggests that immunological aging affects CAR-T therapy response and toxicity. Several studies indicate an association of hallmarks of immunosenescence, such as T cell exhaustion and the presence of Treg subsets, with adverse outcomes. Two landmark studies demonstrated an increased proportion of exhausted T cells (identified by single-cell RNA sequencing, where exhausted cells were defined by increased expression of exhaustion markers), in CAR-T infusion products of non-responders both in chronic lymphocytic leukemia and diffuse large B cell lymphoma (DLBCL).50 51 In the DLBCL study, co-expression of TIM-1 and lymphocyte-activation gene 3 (LAG-3) was most informative for the distinction of exhausted T cells, whereas only very few cells expressed PD-1.50 Interestingly, increased LAG-3 expression as well as increased expression of other exhaustion markers such as T cell immunoreceptor with Ig and ITIM domains (TIGIT) in the apheresis product prior to transduction with the CAR construct has also been demonstrated to be associated with inferior clinical response.52 The same study also detected higher proportions of exhausted CAR+ T cells in the infusion product of non-responders to CAR-T therapy with an investigational CD19/CD20-directed construct.52 Another study identified an association between an increased abundance of CAR+ Treg cells and the lack of response to CAR-T therapy in patients with DLBCL.51 53 Notably, this study also used single-cell transcriptomics to define T cell subsets associated with prognosis, but did not identify a distinct exhaustion T cell phenotype in non-responders.

Emerging evidence suggests a role of immunological aging in CAR-T therapy beyond the infusion product. Increased abundance of exhausted T cells (defined by positivity for PD-1, LAG-3, and TIM-3, or thymocyte selection-associated HMG BOX (TOX) positivity) in the pretreatment lymphoma microenvironment was correlated inversely with CAR-T expansion in a recent study.54 Interestingly, the same study demonstrated that increased infiltration of the tumor microenvironment by Treg cells prior to infusion was shown to be associated with reduced axi-cel neurotoxicity.54

As mentioned above, a reversal of the exhaustion phenotype has been proposed as a therapeutic strategy in order to improve CAR-T therapy outcomes. To this end, a multitude of approaches to enhance T cell fitness have been investigated. These approaches have recently been extensively reviewed elsewhere.55

Ablation of inhibitory immune checkpoints appears to be a promising strategy that holds the potential to be rapidly implemented in clinical practice. Its utility has so far mostly been shown in preclinical studies,56 however, there are reports on clinical responses and reversal of T cell exhaustion in patients with CAR-T-refractory DLBCL treated with ICI.57

CAR-T therapies in older patients with diffuse large B cell lymphoma

Currently, three CAR-T products targeting CD19 are approved for RR DLBCL in third line or beyond by the US Food and Drug Administration (FDA) and the European Medicines Agency (EMA), based on the results of the JULIET trial58 (tisagenlecleucel, tisa-cel), the ZUMA-1 trial59 (axicabtagene ciloleucel, axi-cel), and the TRANSCEND-NHL-00160 trial (lisocabtagene maraleucel, liso-cel). All three trials included patients up to age 76 (JULIET58 and ZUMA-159) and 86 years (TRANSCEND-NHL-00160), respectively. Results suggest non-inferior outcomes in patients ≥65 years as compared with their younger counterparts.58 60 Interestingly, progression-free survival (PFS) was even increased in patients aged ≥65 compared with patients <65 years (13.2 vs 5.6 months) in ZUMA-1 trial.59

Regarding safety, the incidence and severity of CRS seemed comparable across ages. However, the incidence and severity of ICANS and other grade 3–5 treatment-related adverse events (AE) appeared to increase with age in all trials.

Based on the ZUMA-7 trial61 (axi-cel) and the TRANSFORM trial62 (liso-cel), both CAR-T products were approved by the FDA and EMA for patients with primary refractory DLBCL or DLBCL relapsed relapse within 12 months after first-line therapy. In these trials randomized controlled phase 3 trials, both axi-cel and liso-cel compared favorably to high-dose chemotherapy followed by ASCT.61 62 Advanced age was not associated with inferior outcomes.61 62

Patients with RR DLBCL deemed unfit for ASCT face particularly poor outcomes.63 CAR-T therapies in this setting are potentially curative.2 Trials with approved CD19-directed CAR-T products and indications which included patients ineligible for ASCT due to advanced age or comorbidities demonstrated no impact of age on inferior outcomes.2 56.3% of patients ≥70 years treated with liso-cel in the PILOT trial had a complete response (vs 46.2% for patients <70 years).64 Results of the ALYCANTE trial65 for patients deemed unfit for ASCT who were treated with axi-cel are in line (CRR 72.7% vs 69% for ≥70 vs <70 years). Median overall survival (OS) was not reached for both cohorts after a median follow-up of ≥12 months. Of note, patients ≥70 years treated in the ALYCANTE trial had a significantly higher risk for non-relapse mortality (NRM, 12.1% vs 6.9%), mostly related to infections.65 Similar observations were also made in real-world data (see online supplemental table 1).

The French DESCAR-T registry,66 including 957 patients with DLBCL, observed an increased incidence of NRM within the first month in patients with higher age. Furthermore, 62.5% and 39.6% of all NRM events occurred in patients with ≥65 and ≥70 years, respectively, with 88.9% and 55.6% of all early NRM events occurring in patients with ≥65 and ≥70 years, respectively.66 More than half of NRM events occurred due to infection. These findings have been confirmed in a recent meta-analysis which attributes over 50% of NRM in CAR-T therapy across entities to infections.67 Unfortunately, this work did not address the relationship between age and NRM.67 Nonetheless, these findings raise the possibility that immunological aging and impaired pathogen clearance may contribute to fatal infections and increase NRM in older patients.

CAR-T therapies in older adults with multiple myeloma

Although MM is mostly a disease of older adults, pivotal studies on B cell maturation antigen (BCMA)-directed CAR-T therapy for MM provide little information focusing specifically on older patients. The CARTITUDE-168 trial that led to the FDA/EMA approval for ciltacabtagene autoleucel (cilta-cel) for RR MM observed an overall response rate (ORR) of 96.9% and 97.1% for patients with <65 and ≥65 years, respectively. Similarly, the outcomes observed in the KarMMA study69 that led to the FDA/EMA approval of idecabtagene vicleleucel (ide-cel) in patients with RR MM observed that both the response rate (ORR and complete response rate, CRR) and the quality of response (duration of response, DOR, and PFS) of the overall population seemed to be comparable to the ones observed in older adults.

Same as for DLBCL, comparable outcomes between younger and older patients were observed in post-authorization studies for RR MM.70 A recent meta-analysis of 14 clinical trials demonstrates a similar ORR between patients <65 and ≥65 years (93% vs 86 %).71 Importantly, CRS rates were equally comparable between age groups, whereas older patients had a higher risk of neurotoxicity.71

Concerning neurotoxicity, BCMA-directed CAR-T therapy appears to bear a class-specific risk for movement and neurocognitive treatment-emergent AEs (MNT), particularly reported in cilta-cel recipients.72 MNT may manifest as parkinsonism, cranial nerve palsy or other forms of neurological deficit. Although there is little age-specific data, it is reasonable to hypothesize that its manifestations may be a disproportionate encumbrance to older and frail patients.

Results of available CAR-T therapies trials for DLBCL and MM are summarized in online supplemental table 1.

Interplay between CAR-T and frailty

Although the term frailty lacks a generally accepted definition, two major concepts describe this syndrome:

The deficit accumulation model describes frailty as a consequence of accumulated health deficits.73

The Fried criteria define frailty as a syndrome with at least three of the following criteria: unintentional weight loss, self-reported exhaustion, muscle weakness, low physical activity, and slow gait speed.74

Because frailty is inextricably linked to low-grade inflammation,75 its influence on immunotherapy results appears reasonable, in addition to its potential impact on treatment tolerability. Furthermore, the hematological malignancy itself can augment frailty. A recent single-center analysis evaluated outcomes after CAR-T therapy in patients with DLBCL ≥65 years versus <65 years and found no impact of chronological age and geriatric impairments, such as functional limitation, multimorbidity, cognitive impairments, or weight loss, on OS and toxicities.76 However, another single-center analysis found a positive Glasgow-prognostic score, defined as C-reactive protein >10 mg/dL and serum albumin <3.5 g/dL, a common laboratory finding in frailty as well as in advanced MM, was associated with inferior OS in patients with RR MM, even after adjustment for other biological risk factors. In the same cohort, frail patients also had inferior OS.77 This clearly suggests that disease-associated and frailty-associated features are difficult to differentiate and likely interact with each other. In another matched control multicenter cohort study of patients with DLBCL undergoing CAR-T therapy, outcomes in patients <70 versus ≥70 years did not differ significantly but older adults showed worsening of disability after treatment.78 This suggests that the decreased physiological reserve related to frailty impaired the recovery. In addition, cachexia79 80 and sarcopenia,80–82 both major determinants of frailty, were shown to be associated with adverse outcomes after CAR-T therapies in patients of all ages with DLBCL. Of note, increased total adipose tissue was independently associated with improved PFS and OS in patients with DLBCL treated with CAR-T80. This is in line with the recently described obesity paradox which suggests better treatment responses towards ICI treatment in obese patients with solids tumors.83 84 Although there is currently no mechanistic evidence supporting this theory, the pro-inflammatory environment related to obesity appears to facilitate at least some effective immunotherapeutic approaches. Furthermore, the prognostic value of cachexia and malnutrition offers the potential to improve outcomes of CAR-T therapies by pre-habilitation strategies that improve nutrition.

In conclusion, when considering frailty as a determinant of therapy outcomes, not only the clinical deficits leading to a potentially decreased resilience but also the frailty-associated low-grade inflammation should be considered.

CHIP and CAR-T therapy

In patients undergoing CAR-T therapy, a CHIP frequency ranging from 34% to 48% prior to therapy has been reported.85–88 Although CHIP is seen as an age-related phenomenon, the high rate of CHIP even in younger patients can be explained by the usually multiple prior genotoxic treatment lines in CAR-T patients. There is currently no study describing a PFS difference of CHIP and non-CHIP patients treated with CAR-T therapy, but an increased rate of therapy-related myeloid neoplasms in CHIP patients has been reported.86 Furthermore, the pro-inflammatory state related to CHIP appears to be mediated largely by IL-689 90, the key mediator of CRS; thus, CHIP could potentially contribute to toxicity associated with CAR-T therapy.

Although there is a biological rationale for the role of CHIP in CAR-T toxicity, clinical data on this issue are conflicting. In two cohorts, no impact of CHIP on CRS or ICANS85 88 could be observed. However, another study found ICANS of grade ≥3 to be more frequent in patients undergoing CD19-directed CAR-T therapy who were CHIP-positive compared with CHIP-negative (45.2% vs 25.0%, p=0.038). The rate of higher toxicities occurred mainly in carriers of mutations in DNA methyltransferase 3A, tet-methylcytosin-dioxygenase 2, and additional sex combs-like 1 (so-called DTA mutations). The same accounted for CRS grade ≥3 (17.7% vs 4.2% for DTA-CHIP positive vs CHIP negative, p=0.08).86 Miller and colleagues described an increased rate of CRS grade ≥2 in patients undergoing CAR-T therapy for DLBCL and MM, but only in patients ≤60 years (77.8% CHIP vs 45.9% no CHIP, p=0.042/≤60 years; 62.0% CHIP vs 74.4% no CHIP, p=0.26 for ≥60 years).87 No difference in the incidence of ICANS was reported in this cohort.87 All of these studies are limited by sample size, indicating a need for collaborative efforts to decipher the relationship between CAR-T outcomes and CHIP.

CAR-T therapy: the great equalizer?

As mentioned in the previous sections, there is ample evidence for the influence of immunological aging on CAR-T therapy success. However, at the same time, there is considerable consistency in the clinical data reporting non-inferior—or even superior—outcomes of CAR-T therapy in older individuals. At the same time, there is an evident lack of data on the correlation between chronological age and hallmarks of biological age, such as T cell exhaustion, in CAR-T recipients, precluding definitive conclusions regarding the biological foundation of clinical outcomes. As a consequence, the chronological age cut-off applied in most clinical studies assessing CAR-T therapy in older adults remains somewhat arbitrary.

In our point of view, there are two most probable explanations bridging the apparent gap between impairment of CAR-T therapy by biological aging and sustained efficacy in chronologically older patients:

Discrepancies could be due to a potential lack of correlation between hallmarks of biological aging (such as a decrease of naïve T-cell populations, increase in Tregs, and T-cell exhaustion), that impair the success of CAR-T therapy, and biological age. This effect could be partially explained by disease-induced and therapy-induced immunometabolic alterations occurring regardless of chronological age.

Patient selection by application of a more restrictive “eligibility threshold” in older patients could drive the apparently non-inferior or superior efficacy of CAR-T therapy compared with their younger counterparts.

Current data do not allow for a definitive answer to this conundrum, however, the relatively consistent observation that features of biological age, both on the molecular level as well as regarding clinical parameters such as frailty, appear to impair CAR-T therapy efficacy, argue for an—at least partial—role of selection bias in clinical outcome data.

This assumption is underscored by data from a German registry analysis that reports superior outcomes of CAR-T therapy patients aged >65 years.91 Here, although no significant differences could be observed between older and younger patients regarding conventional risk factors such as the International Prognostic Index and prior lines of therapy, younger patients had a shorter interval from diagnosis to CAR-T infusion, and a higher proportion of younger patients had previously failed hematopoietic stem cell transplantation, arguing for an implicit accumulation of high-risk disease in the younger cohort.91 The same holds true for MM, where an international real-world analysis demonstrated a higher proportion of high-risk patients and a significantly shorter interval from MM diagnosis to CAR-T cell infusion in patients <65 years receiving BCMA-directed CAR-T therapy.92

Furthermore, it should be noted that the risk of CAR-T product manufacturing failure is increased in older patients.93 Given that, to our knowledge, all available analyses do not compare intention-to-treat cohorts, but only include patients who received a CAR-T infusion product, there is further potential for bias.

In summary, CAR-T cell therapy in older patients with lymphoid malignancies is unprecedentedly successful. Based on the available clinical data, there is reason to conclude that biological age should not be an exclusion criterion for therapy, and CAR-T therapy can be regarded as a standard of care in older patients.

However, the apparently increased vulnerability of older patients to ICANS, neurotoxicity, and potentially infection, remains concerning and raises the need for pre-emptive intervention strategies.65 91 94 Integrating biological, clinical, and functional data in older patients receiving CAR-T therapy remains a major challenge to the field, and will hopefully translate into improved outcomes.

Bispecific antibodies

BsAbs are dual-affinity antibodies typically directed against CD3 and a tumor antigen, redirecting adjacent T cells to malignant cells in order to induce a cytotoxic response.95 In contrast to CAR-T therapies, BsAbs do not require patient-specific manufacturing periods and are therefore immediately available, but repeated or continuous treatment is required for disease control. Excessive T-cell activation by BsAbs can also lead to CRS and ICANS.96

BsAbs in older adults with lymphoma

Currently, mosunetuzumab, glofitamab and epcoritamab are approved BsAbs for the treatment of various RR B cell Non-Hodgkin lymphoma (B-NHL). They all are IgG-like full-length CD20xCD3 BsAbs with Fc silencing mutations to prevent FcyR-mediated cross-binding.95 As summarized in online supplemental table 2, various BsAbs have been investigated in B-NHL and three already obtained approval.

After step-up dosing, mosunetuzumab for up to 17 cycles resulted in an ORR, CRR, and median DOR of 35%, 19%, and 7.6 months, respectively, in patients with aggressive B-NHL (aNHL) in a phase 1/2 study with 197 subjects, including 10% with prior CAR-T therapy. In the larger indolent NHL population 66% ORR, 48% CRR and 16.8 months median DOR were reported.97 Based on a single-arm phase 2 trial with 30 mg target dose in RR follicular lymphoma (FL), which reported an 80% ORR, 60% CRR and median DOR of 22.8 months, approval for RR FL after ≥2 prior lines of therapy was granted.98 Notably, efficacy outcomes did not differ between patients <65 or ≥65 years, and numerically lower CRS and serious AE rates indicate a favorable safety profile also in older patients.99 In aNHL, an expansion cohort recently confirmed ORR and CRR of 42% and 23.9% with a PFS of 3.2 months.100 An additional subgroup analysis of patients >80 years recently reported an ORR of 43%.101

Glofitamab, which is administered intravenous for up to 12 cycles after a pretreatment with obinutuzumab, showed an ORR of 52%, CRR of 39% and median DOR of 18.4 months in 155 patients with DLBCL treated in a single-arm phase 2 trial, irrespective of prior CAR-T therapy.102 This trial resulted in approval for RR DLBCL after ≥2 lines of prior treatment and included patients up to 90 years without reporting on older subgroups separately.

In its phase 2 registration trial including 157 patients with RR DLBCL, epcoritamab led to ORR and CRR of 63% and 39%, respectively, with a median DOR of 12.0 months, resulting in approval for the treatment of RR DLBCL after ≥2 prior lines of treatment.103 This study enrolled patients up to the age of 83 years and did not report any differences regarding safety or efficacy according to age groups.

Fueled by the favorable outcomes in the RR setting, these agents are investigated in the first-line setting, especially in frail patients, deemed unfit for standard approaches. Recently, a preliminary analysis of a phase 2 trial (NCT03677154) investigating mosunetuzumab plus polatuzumab vedotin in 108 patients with older/unfit treatment-naïve DLBCL with a median age of 81 years (range: 66–94 years) reported 80% ORR and 61% CRR.104 Grade ≥3 AEs were reported in 43% of patients and 15% of patients discontinued mosunetuzumab due to AEs. Fatal AEs occurred in 14% of patients and were mostly infectious complications such as COVID-19104.

BsAbs in older adults with MM

The EMA/FDA both approved the BsAbs teclistamab,105 talquetamab106 and elranatamab107 in RR MM after three (EMA) and four lines (FDA), respectively, of prior therapy. Teclistamab and elranatamab both target CD3 and BCMA, while talquetamab targets CD3 and the G-protein-coupled receptor 5D. None of their registration trials had an upper age limit. In MajesTEC-1,105 the phase 1/2 trial that led to the approval of teclistamab, the median age was 64 years (range: 33–84 years) and 14.5% were ≥75 years, in MonumenTAL-1106 that assessed talquetamab, median age was also 64 years (range: 39–84 years), while patients in MagnetisMM-3107 receiving elranatamab were slightly older (median age: 68 years, range: 36–89 years). Response rates are summarized in online supplemental table 2. In MajesTEC-1, ORR for patients aged 65–74 years as compared with ≥75 years were 64% versus 54% (N=35/55 vs N=13/24). This is in line with real-world data that demonstrated no difference in outcomes between patients ≥70 years versus <70 years.108 In MagnetisMM-3,107 ORR in patients <75 years and ≥75 years were 80% and 20% (N=99/123 vs N=24/123), respectively. Whether this major difference is due to a statistical bias of different group sizes and/or differences in biological risks remains elusive. While AEs were not separately reported for the older subpopulation, these patients are particularly vulnerable to fatigue and diarrhea, which occurred in about 25% of patients with teclistamab,105 with rates even higher in other trials (elranatamab107: 36% and 42%; talquetamab106: 33% and 30%). Dysgeusia appears to be a substance-specific side effect of talquetamab, with 63% of patients reporting taste changes in the MonumenTAL-1 trial.106 Dysgeusia is particularly concerning for older adults as these are at increased risk for weight loss and cachexia during cancer treatment,109 and 30% of patients indeed lost weight during treatment.106

Impact of immune aging on BsAbs efficacy

While the impact of immune aging on the efficacy and toxicity of BsAbs has not yet been dissected, recent data demonstrated an association of pretreatment CD8+T cell exhaustion and decreased clonal expansion capacity of T effector cells with non-response to BsAbs in patients with MM.110 Whether this also accounts for pre-existing exhausted T cells related to immune aging is currently unclear. In addition, a recent analysis of MajesTEC-1 study demonstrated more durable responses to teclistamab to be associated with lower Tregs and lower expression of inhibitory receptors (CD38, PD-1, and PD-1/TIM-3) on T cells at baseline.111 A real-world retrospective study of teclistamab found effector CD8+ T cell populations to be associated with response to therapy and a regulatory T cell population with non-response.112 In a mouse model of MM, high tumor burden was shown to induce T cell exhaustion and decreased efficacy of BsAbs that could be reverted by cytotoxic pretreatment reducing the tumor burden.113 In addition, continuous exposure to CD19xCD3 BsAbs induced T cell exhaustion leading to decreased efficacy in an experimental model of acute lymphoblastic leukemia. The introduction of treatment-free intervals counteracted T cell exhaustion and improved tumor control in this model.114 Although the results of these preclinical studies cannot be directly transferred and generalized to clinical use of BsAbs, it paves the way to design studies tailoring BsAbs dosing in order to overcome a possible loss of efficacy due to T cell exhaustion. Nonetheless, it needs to be clearly stated that there is currently no data that T cell exhaustion might negatively impact clinical outcome especially in older adults. Well-grounded translational substudies should further dissect T cell exhaustion as an influencing factor in dependency with age, chronic inflammation and prior treatment approaches.

In summary, most of the registration trials for BsAbs in patients with myeloma and lymphoma included patients >70 years although few subgroup analyses regarding efficacy and toxicities are available, some of which may be clinically relevant to this population. Based on real-world data, BsAbs appear to be tolerated and efficient in older adults.

Immune checkpoint inhibitor therapy

Despite the broad success and approval of ICI in solid tumors, their use in hematological malignancies is mostly restricted to classical Hodgkin’s lymphoma (HL). Although a first peak of HL incidence is found in early adulthood, a rising incidence and a large second peak is observed in adults ≥60 years.115 HL carries a dismal prognosis in older adults compared with the excellent long-term outcomes in younger patients, presumably due to inferior treatment tolerability. Biologically, HL is characterized by frequent 9p24.1 amplifications that, complemented by other mechanisms,116 results in relatively consistent programmed death-ligand 1 (PD-L1) expression on malignant Hodgkin’s and Reed-Sternberg (HRS) cells.117 118 In addition, a skewed PD-1/PD-L1 axis is also observed for various immune cells in the tumor microenvironment including macrophages and natural killer cells. Interestingly, major histocompatibility complex (MHC)-I and MHC-II expression is frequently lost in HRS cells117 119 120 and currently available evidence hints against a conventional adaptive cytotoxic T cell mediated mechanism as the primary correlate for anti-PD-1 efficacy in HL.121–124 Of note, neither PD-L1 nor MHC-I/MHC-II expression or 9p24.1 amplifications could be identified as reproducible biomarkers in HL.

Based on the results of the phase 2 CheckMate 205125 (nivolumab) and KEYNOTE-087126 (pembrolizumab) trials, both agents were approved for the use in relapsed HL after ASCT and/or as third line treatment, with comparable ORR (nivolumab: 69% and pembrolizumab: 71.9%), CRR (16% and 27.6%) and 5-year OS rates (71.4% and 70.7%). Both trials included <10% of patients ≥60 years and none reported on age-dependent differences in toxicities and/or efficacy.125 126

Recently, outstanding efficacy and a relatively favorable safety profile was reported for both concomitant and sequential anti-PD-1-based first-line treatments.127–130 An interim analysis of the randomized S1826 trial showed superior 1-year PFS with six cycles of nivolumab plus doxorubicin, vinblastine and dacarbazine (N-AVD, 94%) compared with six cycles of brentuximab vedotin plus AVD (BV-AVD, 86%).131 For patients ≥60 years (range, 60–83 years; N=97), a subgroup analysis showed superior 1-year PFS of 93% (95% CI, 79% to 98%) for N-AVD vs 64% (95% CI, 45% to 77%) for BV-AVD. N-AVD was associated with a reduced NRM (4% vs 14%) and reduced early treatment discontinuation (10% vs 33%).132 The toxicity profile of N-AVD was also favorable with significantly less sepsis and peripheral neuropathy.132 This data is in line with the results from a single-center single-arm trial (N=37 patients, age range: 60–78 years).133

The French phase 2 Niviniho trial evaluated an induction therapy with nivolumab in frail patients ≥61 years, followed by a combination of nivolumab±vinblastine after achieving a complete metabolic response. Toxicities were considerable and the overall outcomes were dismal (ORR 51.9%, median PFS 9.8 months).134 Another non-comparative phase 2 study evaluated nivolumab in combination with brentuximab vedotin in patients ≥60 years ineligible for combination therapy with doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD, N=21). CRR was 67% and median OS and median PFS were not reached at 51.1 months of median follow-up. 76% of patients developed treatment-related toxicities of grade 3 or higher according to the common terminology criteria for adverse events(CTCAE) with 19% motor polyneuropathy and 19% sensory neuropathy.5 These results are further supported by another single-arm trial.135

No valid comparative data for ICI efficacy in older patients with versus younger HL is available. However, HL is characterized by extensive non-neoplastic immune cells that constitute the majority of the tumor besides the few neoplastic HRS cells.136 Thus, age-related changes in immune cells are likely affecting the HL microenvironment although these effects and their relevance remain elusive so far. Preliminary data from older patients with melanoma and lung cancer treated with ICI suggested worse efficacy in older patients with T cell senescence, which was partly defined as loss of CD28 expression on CD8+T cells.137 138 As the efficacy of ICI in HL seems to be less dependent on a cytotoxic immune response,121 no conclusions can be drawn for ICI in older patients with HL yet. In light of relatively low response rates with ICI monotherapy in older patients, future research dissecting the effects of immune aging on ICI in HL is urgently needed. In addition, qualitative data suggest the presence of CHIP-positive immune cells in the microenvironment of HL.139 Given the proinflammatory nature of CHIP, an impact on disease biology appears reasonable.

Immune-related adverse events (irAE) induced by ICI can potentially affect every tissue or organ in the body. Presumably, irAE are mediated by the re-activation of autoreactive T cells.140 No data is available on whether rates of irAE are higher in older adults with HL. Data from solid cancers suggests some differences in irAE incidence between age groups depending on the affected tissue.141–146 Real-world data with limited case numbers suggest no difference in the incidence of irAE between frail and non-frail older adults,147 suggesting that the pro-inflammatory component of the frailty phenotype does not fuel the development of irAE. However, functional outcomes can be worse in frail patients, as demonstrated by a higher rate of hospitalizations despite a similar severity of irAE.148 149

In summary, ICI appears as attractive partners for lower-intensity treatments in frail adults with HL. However, its interplay with the aged immune system needs to be understood in greater details to improve outcomes.