Aging-associated immune system changes in multiple myeloma: the dark side of the moon

Abstract

Multiple myeloma (MM) is a disease of the elderly. Changes that occur in the immune system with aging, also known as immunosenescence, have been associated with decreased tumor immunosurveillance and are thought to contribute to the development of MM and other cancers in the elderly. Once MM establishes itself in the bone marrow, immunosenescence related changes have been observed in the immune tumor microenvironment (iTME) and are driven by the malignant cells. The efficacy of novel immunotherapies used to treat MM has been blunted by detrimental iTME changes that occur at later disease stages and are, to some extent, driven by prior therapies. In this review, we discuss general changes that occur in the immune system with aging as well as our current knowledge of immunosenescence in MM. We discuss the differences and overlap between T cell senescence and exhaustion as well as potential methods to prevent or reverse immunosenescence. We focus predominantly on T cell immunosenescence which has been better evaluated in this disease and is more pertinent to novel MM immunotherapies. Our lack of understanding the drivers of immunosenescence at each stage of the disease, from precursor stages to heavily pretreated MM, represents a major barrier to improving the efficacy of novel and existing therapies.

Introduction

Multiple myeloma (MM) is an incurable plasma cell malignancy most commonly affecting patients over 60. Due to significant progress in MM therapeutics, the overall survival of MM patients has been increasing over time [1]. However, the outcomes of patients with relapsed MM and especially triple-refractory MM remain poor [2, 3]. Therapy induced clonal selection and clonal evolution [4, 5] play a role in disease progression and refractoriness. In addition, alteration of the normal bone marrow (BM) immune microenvironment (iTME) can lead to tumor escape from immunosurveillance. The importance of leveraging the immune system to control MM progression is highlighted by the recent use of immune-based therapies, such as chimeric antigen receptor T (CAR-T) cells and monoclonal antibodies with novel mechanisms of action. However, immune therapies rely on a functional iTME to be effective. The impact of iTME immunosenescence on tumor immunosurveillance is an important aspect of iTME dysfunction that has not been explored well in MM. We use the term immunosenescence throughout this article to broadly define the overall dysfunction of the innate and adaptive immune systems with aging. The term encompasses several “hallmarks of aging” : genomic instability, telomere attrition, epigenetic alterations, loss of proteostasis, deregulated nutrient sensing, mitochondrial dysfunction, cellular senescence, stem cell exhaustion, and altered intercellular communication [6]. However, the term remains incompletely characterized since not all leukocyte subsets are equally negatively affected during aging. Immunosenescence is important to address in MM specifically, since it is a disease of the elderly and several existing therapies can promote immunosenescence. Here, we will review general concepts of immune system aging followed by our current knowledge of immunosenescence in MM. We will focus predominantly on T cell immunosenescence which has been better evaluated in this disease and is more pertinent to novel MM immunotherapies.

Immune system changes with aging in healthy patients

An aging immune system reflects the exposure of its host to prior exogenous immune challenges, such as microbial or, in the case of autoimmune diseases and cancer, self-antigens, and other environmental influences (diet, gut microbiome, physical activity, medications). It also reflects the intrinsic ability of a person’s immune system to adapt to these challenges with time. The latter represents a complicated network of genomic (inherited), epigenomic and metabolic factors that are in a complex interplay with one another. Hallmarks of immune system aging have been recently characterized with single-cell approaches and include increased expression and chromosomal accessibility of genes associated with oxidative stress, cell cycle regulation and inflammation across all immune subsets [7]. Epigenetic changes, such as chromatin modifications of immune cells, also increase with age and are predominantly driven by non-heritable influences [8]. In fact, increased cell-to-cell variability in chromatin modifications within each immune cell subtype is a molecular signature of aging [8]. The result is an overall increase in several proinflammatory mediators in older individuals (cytokines, chemokines, miRNAs, and other pro-inflammatory signals) collectively known as the senescence associated secretory phenotype ( SASP) [9], who result in a chronic, low-grade inflammatory state also known as “inflammaging” [10].

Changes in innate immunity

Several innate immune system components are affected by aging and, overall, the number of myeloid cells increases with age as do myeloid progenitors in the bone marrow [11]. In addition, their lifespan is decreased with age because of decreased responsiveness to growth factors that prevent apoptosis [12, 13]. The functions of phagocytic cells such neutrophils, monocytes and macrophages becomes impaired at multiple levels. For instance, effective neutrophil chemotaxis is impaired and is causally associated with increased constitutive phosphoinositide 3-kinase (PI3K) signaling [14]. Once on site, the ability of neutrophils to phagocytose is also limited compared to younger subjects, likely as a result of reduced oxidative burst [15]. Similarly, monocytes demonstrate decreased cytokine production in response to toll-like receptor stimulation [16] and interferon (IFN) [17, 18] likely as a result of decreased activation of the mitogen-activated protein kinase (MAPK) pathway [19] and JAK/STAT pathways [17]. While phenotypic or numeric changes in the circulating neutrophil pool with aging are less well defined, non-classical monocytes (CD16+) increase with age [7, 20, 21] but have lower expression of CX3CR1 receptors, which are implicated in promoting monocyte survival, and the activation molecule HLA class II [20]. Therefore, increased numbers of circulating CD14+CD16+ monocytes may not necessarily mean enhanced functionality in this setting. The increase in myeloid derived suppressor cells (MDSCs), a highly immunosuppressive myeloid subset [22], that occurs with aging [23], may be due to the increased age-related myeloid hematopoiesis and could possibly counteract the generalize inflammation in this setting. Finally, single-cell sequencing analyses have demonstrated that circulating myeloid cells upregulate several genes associated with cellular senescence and the SASP such as IL1, CXCL8, IL1B, TNF, p21 and these changes are most prominent in the classical monocyte population [7].

Natural Killer (NK) cells are the only cell of lymphoid origin that is part of the innate immune system. However, recent work has demonstrated that NK cells can form antigen-specific immunologic memory [24], suggesting that these cells have attributes of both innate and adaptive immunity. With aging, immature CD56^high subsets decrease and the more mature CD56^low subsets with effector function, increase [25]. The latter also express higher levels of CD57+, a well characterized marker of terminal differentiation and immune senescence. These CD56^low cells respond poorly to cytokines but maintain their cytolytic capacity. They are, unlike CD56^high NK cells, unable to secrete cytokines that can activate other parts of innate immunity such as dendritic cells and monocytes, which has been assumed to contribute to age-associated declines in immune competence. Other well-characterized phenotypic changes in NK cells, include the decrease of the activating surface receptor NKp30 [25]. Single-cell transcriptomics reveal that genes enriched in apoptotic signaling pathways are upregulated and genes involved in cellular responses to lipopolysaccharide and virus defense are decreased in NK cells of old versus young humans [7].

Finally, dendritic cells (DCs), the classic antigen presenting cells that bridge innate and adaptive immunity, are also affected by aging. Myeloid DCs have been better studied and exhibit reduced ability to migrate towards inflammatory stimuli, phagocytose and present antigens to T cells as well as stimulate T cells after antigen presentation (reviewed in [26]). Similarly, aged plasmacytoid DCs produce lower levels of interferons and cannot present antigens as effectively. Recently, single-cell transcriptomics have also shown an upregulation of genes involved in apoptotic and inflammatory gene pathways and a downregulation of genes involved in antigen-presentation [7].

Changes in adaptive immunity

Immunosenescence has been more extensively studied in cellular compartments of the adaptive immune system. Overall, B cell production decreases with age [27], which is thought to be secondary to decreased production of lymphoid progenitor cells [28] and decrease maturation of committed B cell progenitors. The latter is explained by the decreased expression of genes required for the maturation of pro- and pre-B cells in aged individuals, including the recombination activating gene (RAG) enzymes and lambda-5 [29, 30]. In addition to reduced numbers, B cell functionality also decreases with age. B cells of aged individuals are characterized by decreased B cell receptor diversity and defects in class switch recombination [31], both of which are necessary for effective antibody production. A recently described B cell subset known as age-associated B Cells, expresses the transcription factor T-bet, has a unique surface phenotype (CD11c, CD11b+, CD21−, CD23− [32], and progressively accumulates with age [33]. These cells are thought to arise via B cell receptor-mediated activation in the context of antigen stimulation of toll-like receptors. They also accumulate in autoimmune diseases and chronic infections and perhaps their age-associated increase serves as a surrogate of a hosts “immunobiography” [34]. Single-cell transcriptomics have shown that aging results in an upregulation of gene pathways implicated in memory formation and inflammation and a downregulation of pathways implicated in viral defense [7].

T cells changes with normal aging have also been extensively studied. Unlike B cells, the overall number of T cells remains relatively stable in aged individuals. However, their phenotypic composition undergoes changes the most significant of which is the contraction of the naïve T cell pool and expansion of late differentiated forms as well as Tregs [35]. Naïve T cells express markers such as CD45RA and CCR7, which allow them to home into the lymph nodes, along with CD27 and CD28, which are co-stimulatory molecules that are used to interact with B cells and antigen presenting cells and result in the activation of T cells to produce cytokines. Central memory T cells (CD45RA−, but CD27+, CD28+ and CCR7+) are capable of producing high levels of interleukin-2 and IFN γ but not effector molecules such as tumor necrosis factor α TNFα, or cytotoxic molecules (i.e. perforin and granzymes). Effector memory T cells lose CD27, CD28 and CCR7 expression and are less able to proliferate in response to stimuli but have the capacity to produce high levels of effector molecules. Lastly, effector T cells regain CD45RA expression along with potentially other markers of terminal differentiation such as CD57 and KLRG1, have very limited proliferative capacity but tend to secrete a wider range of cytokines following activation. The CD28− group of T cells also seem to be enriched in oligo/monoclonal populations compared to CD28+ cells that are polyclonal [36], consistent with their status as cells responding to antigen.

Thymic involution was considered the primary cause of decreased naïve T cell production in the elderly [37]. However, recent studies have challenged its importance in human immunosenescence since division of peripheral naïve T cells remains the main source of naïve T cells in aged humans [38]. The contraction of the naïve T cell pool in humans is thought to be secondary to peripheral selection of T cell clones under pressure from various antigens, predominantly cytomegalovirus (CMV) [39], rather than thymic involution [40]. Whereas, the decrease in naïve T cells with aging affects CD8 T cells to a larger extent, the impact of chronic CMV infections disproportionately affects CD4 T cells for reasons yet unknown [39]. Irrespective of the underlying mechanisms, the shift of the T cell poor towards more mature and terminally differentiated phenotypes is characterized by canonical phenotypic changes [41]. One of the hallmarks of T cell immunosenescence, is the loss of CD28, a major co-stimulatory molecule. CD28− T cells have high cytolytic activity [42] and are still capable of proliferation in response to antigenic stimulation but at rates much slower than CD28+ T cells [43]. Recent studies show that senescent T cells also harbor DNA damage, short telomeres, low telomerase activity and upregulate signaling pathways associated with cellular senescence and the SASP [44, 45]. Their T cell receptor signaling machinery is also impaired [45]. CD28− T cells are also prevalent in younger patients with chronic infections, autoimmune diseases [46] and cancer [47], which suggests that chronic inflammation drives CD28 downregulation and T cell immunosenescence. Therefore most immunosenescent T cells are late effector T cells that have escaped apoptosis. Immunosenescence may represent the far end of the spectrum in a T cell’s life and also a complementary mechanism to exhaustion to keep T cell responses “in check” while still maintaining a pool of antigen-specific cells in reserve [48].

Which specific components of chronic inflammation are responsible remains to be elucidated. Tumor necrosis factor alpha (TNFa), which has been shown to inhibit CD28 transcription [49–51], is a potential candidate. CD28 transcription regulation is also associated with the caspase pathway, which is downstream of the TNF-α receptor-1 [52]. Studies on Jurkat CD4+ T cell tumor line [52, 53] showed that elevated caspase-3, but not to the threshold level required for apoptosis, resulted in down-regulation of the CD28 promoter and the loss of CD28 cell surface expression. The same is true in primary human CD8+ T cells, where TNFa inhibition increased proliferative potential, delayed loss of CD28 expression and enhanced telomerase activity [54]. Differential methylation between CD28+ and CD28− cells has also been described [36] and in CD28− T cells, genes associated with cytotoxicity and cytokine signaling are demethylated whereas genes implicated in T cell receptor signaling are methylated. However, even within the CD28− T cell compartment significant heterogeneity exists in terms of replicative potential with KLRG1+ and CD57+ T cells, thought to represent the most senescent subsets. For this reason, these 2 markers are considered definitive for T cell replicative senescence [55–57]. The expression of NK cell-associated receptors (KLRG1, NKG2a, NKG2C and NKG2D) is of interest and some suggest that senescent T cells, rather than being dysfunctional, simply acquire NK cell-like function and can engage antigens in an MHC-I-independent manner [48].

At a single-cell level, aging upregulates biological pathways in T cells included TNF signaling, IL-1 signaling and the apoptotic signaling pathway whereas mRNA processing is impaired [7]. Aging is associated with a decreased proportion of CD8 naïve T cells with increased apoptotic signaling and lymphocyte activation pathways and an expanded CD8 effector memory compartment with increased cytokine production as well as reduced chromatin remodeling and antiviral function. Finally, similar to the B cell receptor, T cell receptor repertoire diversity is decreased with age, which is more pronounced for CD8+ T cells [7].

Evidence of immunosenescence in myeloma

A summary of the factors that drive iTME immunosenescence in this disease as well as their impact on the iTME are summarized in the figure.

Figure.

Major drivers and impact of immune senescence in multiple myeloma. Intrinsic drivers include inflammaging, that occurs with normal aging and potentially senescence driven by the malignant clone itself. Extrinsic drivers include anti-myeloma therapies. The results is an increase in SASP factors, MDSCs and T cell senescence that result in immune suppression, tumor proliferation and decreased efficacy of immunotherapies. MDSCs: myeloid derived suppressor cells, SASP: senescence associated secretory phenotype, TME: tumor microenvironment. Created with biorender.com

SASP factors

Several components of the SASP are increased in MM and have a direct trophic effect on malignant plasma cells. Given the non-specific nature of SASP, the causal relationship between aging-related versus malignant cells- or microenvironment-related SASP and myelomagenesis is difficult to ascertain. Irrespective of this, several typical SASP proteins are also well-characterized molecules in myeloma immunobiology. Interleukin-6 is abundant in several malignancies and in the MM microenvironment it is produced by autocrine or paracrine mechanisms. It is a well characterized MM growth factor both by promoting proliferation [58] and inhibiting its apoptosis [59]. Similarly IL-8 [60], IL-1 [61], tumor growth factor beta [62], CCL5 [63], MIP-1a [63], MMP-1/3 [64], MCP-1/2 [65, 66] which are among the most robustly induced SASP factors [9] have all been implicated in MM pathogenesis and are produced by MM itself or components of its microenvironment. Many of these cytokines are increased throughout the course of the disease and do not return to normal after hematologic remission [67] which suggests that host related factors maintain a pro-inflammatory cellular iTME even after the malignant clone burden has been minimized.

Cellular immunosenescence

Our knowledge about how age-related immune changes impact MM progression remains limited. A shift towards myeloid hematopoiesis, manifested by higher myeloid to lymphoid ratio, a hallmark of an aging immune system as mentioned, is associated with worse prognosis in patients with newly diagnosed MM [67–70]. Interestingly, depletion of plasma cells from old mice reduced the number of myeloid-biased hematopoietic stem cells and mature myeloid cells to levels similar to young animals, but lymphopoiesis was not affected [71]. Therefore, bone marrow plasma cells appeared necessary for age-related increased myelopoiesis. These aged plasma cells also produced higher levels of inflammatory cytokines and blocking IL-1 and TNFa attenuated myelopoiesis in old mice. It is very possible that the same is true for malignant plasma cells and the worse prognosis in myelopoiesis reflects a more abundant and inflammatory malignant clone. MDSCs are increased in MM, similar to other cancers, and inhibit other immune cells, predominantly T cells [72]. In the bone marrow of MM, these cells reside in the mature granulocytic compartment [73] and are associated with worse patient outcomes. These mature subsets were characterized by the upregulation of several SASP-related proinflammatory genes (CXCL1, TGFB1, VEGFA, CCL4, IL-8, IL-6 and IL1B) and decreased epigenetic accessibility to genes related to normal neutrophil activation and metabolism. The impact of existing MM therapies on MDSCs is debated with some studies suggesting no effect [74] and others showing a reduction of MDSCs with immunomodulating drugs [75]. In terms of other innate subsets, CD16+ monocytes, which increase in normal aging, also increase as MM evolves from monoclonal gammopathy and smoldering myeloma to active disease [76] and with increasing disease burden in active myeloma [77] and promote osteoclastogenesis. Finally, NK cell composition in terms of mature and immature subsets appears similar to that of age-matched healthy donors [78, 79] with no increase of more mature forms but significant functional impairment [78].

In terms of the adaptive immune system, it is interesting to note that, like age-associated B cells, myeloma cells frequently express several myeloid-lineage related surface markers including CD11b and CD11c [80–82]. A tempting hypothesis would be that chronic antigen stimulation, that drives the increase of this late effector B cell subset, could also predispose them to malignant transformation as they transition to antibody secreting plasma cells. Little is known about senescence related changes in the malignant plasma cells themselves. Oncogene induced cellular senescence in malignant plasma cells has been shown to select for a cancer stem-cell-like population of malignant cells with very low CD138 expression that is present prior to therapy at lower numbers but gains clonal dominance after genotoxic stress inducing therapy and is maintained through the SASP [83]. It is unclear what gives rise to this senescent population to begin with.

T cell immunosenescence has been characterized much better in this disease. It is present at diagnosis and increases with therapy. It can be modulated with existing therapies and has significant implications on the use of novel immunotherapies. Indeed, circulating T cells and those in the immune microenvironment (bone marrow) of multiple myeloma show characteristics of immune senescence and exhaustion and are increased compared to age-matched healthy donors at diagnosis and after treatment [84]. The increase of exhausted subsets is more pronounced in the central memory compartment rather than in late differentiated subsets. Clonal T cells are exclusively located in the CD57+ pool of terminally differentiated T cells, and therefore display features of senescence with very little expression of the classic exhaustion markers PD-1 and CTLA4 [85]. Interestingly, this study [85] also found normal-for-age telomere lengths in this subset, indicating that senescence is telomere independent, as well as normal levels of p38-mitogen-activated protein kinase and p16, which have been implicated in T cell immunosenescence. Similar comprehensive studies of immunosenescence have not been performed in T cells from MM and therefore these results will need to be replicated since, some are not in agreement with what other published literature. Of note, patients with MGUS, an indolent MM precursor, had a T cell compartment resembling that of young healthy donors with higher levels of naïve T cells [86], which suggests that either MM is driving T cell senescence in the bone marrow or that poor MM control may be due to host-related immunosenescence. Others have shown that effector differentiation is seen in both MGUS and MM compared to healthy donors but MGUS is enriched in stem-like memory T cells (TCF1/7+) [87]. Finally, increases of some senescent T cell subsets at diagnosis is associated with inferior patient outcomes [86] similar to other lymphoid malignancies [88].

Therapy for MM also drives T cell differentiation and senescence. Autologous stem cell transplant (ASCT) has been shown to increase both senescent and exhausted populations [86, 89–91] and these subsets are associated with worse prognosis post-ASCT [89, 92] (also author’s unpublished observations). Increase in the cell cycle inhibitor p16 in the T cell compartment after ASCT has been described in MM and other diseases but likely only reflects these well described phenotypic changes [93] rather than true induction of the global hallmarks of cellular senescence at a more global level. For instance, we have not noted an increase in several classic SASP cytokines early post-ASCT nor have we seen an increase of SA-B-gal positive T cells (unpublished observations), a well characterized marker of cellular senescence. Non-cytotoxic therapies also increase the number of late differentiated T cells [90]. Among them, CD38 monoclonal antibodies that skew the T cell pool towards a more differentiated one an decrease the pool of naïve cells [94–96] likely because of the induction of clonal T cells with anti-MM activity. These alterations are clearly beneficial for disease control but leave the immune system at a disadvantage when subsequent immunotherapies are considered, such novel BCMA-based approaches, as is evident by the inferior outcomes of CD38 refractory patients with these approaches [97]. We have also shown that immune senescent T cells increase with subsequent lines of therapy and the naïve T cell pool shrinks [98].

Is there overlap between T cell exhaustion and senescence?

At this point it is important to note that the overlap between T cell immune senescence and exhaustion, if any, is not entirely clear [99, 100]. A major challenge in evaluating the 2 phenomena is the limitation of existing animal models, since mice [101] may not accurately reflect immune senescence in humans. Both senescence and exhaustion can be induced by persistent antigenic stimulation, but senescence can also be induced by the triggering of DNA damage response pathways [102]. Their functional characteristics are different. Exhausted T cells have poor cytolytic activity and impaired cytokine production [103] whereas senescent T cells are late effector subsets with high cytolytic activity and expressing a SASP-related array of molecules [104]. The molecular mechanisms driving each process are different. Whereas both exhausted and senescent T cells demonstrate loss of several components of the T cell receptor machinery [45, 105] senescent T cells upregulate cell cycle arrest genes such p16, p21 and p53 [104, 106]. Their metabolic profiles are distinct. Exhausted T cells display suppressed glycolysis and oxidative phosphorylation and dampened mitochondrial function [107] whereas senescent T cell preferentially depend on anaerobic glycolysis [108]. Their phenotypic features and ability to revert to an activated or proliferative state are also different. Senescent subsets are, invariably, differentiated subsets that lose CD28 expression and can sometimes re-express CD45RA. Exhausted subsets can be both CD28 positive or negative. Terminally differentiated T cells, at the far end of T cell senescence, such as those who express CD57, express classic markers of exhaustion, such as PD-1, at very low levels or not at all [85, 86]. This might not be true for novel immune checkpoints such as TIGIT where markers of terminal differentiation (CD57) can be co-expressed with TIGIT [98, 109]. The ability to revive the proliferative capacity of these cells with TIGIT blockade is not known. PD-1 blockade may be effective in the effector memory RA subset [110], whose telomere length is more preserved compared to CD57+ subsets. Interestingly, this effect was telomerase-independent. Checkpoint inhibition can clearly restore T cell functionality although recently it was shown that not all exhausted subsets are equally responsive to this approach [111]. Only progenitor exhausted T cells, with characteristics central memory T cells were responsive to checkpoint inhibition, whereas terminally exhausted T cells, which were transcriptomically similar to effector subsets i.e. cytotoxic but short-lived, were not and were associated with worse clinical outcomes after PD-1 inhibition. This also suggests that even exhausted T cells have the capacity to differentiate and eventually undergo senescence, which perhaps marks a “point-of-no-return” in terms of their ability to reverse to a non-exhausted state and underlines the significance of preventing/reversing immunosenescence.

Finally, the timeline of the 2 processes is different. PD-1 mediated T cell exhaustion is an early phenomenon, occurring rapidly after antigenic stimulation of some naïve T cells [112] and inhibits further T cell differentiation. PD-1 blockade allows for resumption of normal differentiation with recovery of early memory subsets. T cell senescence is a result of the gradual progression across the T cell differentiation spectrum and its inhibition, when possible, merely restores T cell proliferative capacity rather than reverse their phenotypic maturation [45].

Is T cell immunosenescence preventable or reversible?

Improving the detrimental, aging-related immune changes has the potential to prevent immune system dysfunction. The prevention of T cell immunosenescence should be prioritized since novel immunotherapies, such as CAR T cells and novel monoclonal antibodies rely on a functional T cell compartment to work. More specifically, harvesting less mature T cells prior to CAR T cell manufacturing has been associated with improved outcomes in several studies in MM and other lymphoid malignancies [113–116]. Therefore, the expected T cell immunosenescence in the mostly elderly MM population, which is further complicated by tumor-induced and then therapy related immunosenescence, is highly problematic and will limit the efficacy of novel therapies significantly.

Indeed, some MM therapies such immunomodulating drugs (e.g. lenalidomide) have been shown to increase the proportion of CD28+ T cells [79, 117, 118] but that this effect is more pronounced in patients with high IKZF T cell expression, a known target of the drug [119]. Studies led by the Akbar group have demonstrated that T cell senescence is potentially reversible. This group first noted that terminally differentiated effector memory cells that re-express CD45RA have significantly longer telomeres than less differentiated progenitors despite lower telomerase activity and thus may not differentiate similar to other T cell subsets [120]. They went on to show that telomerase is actively inhibited in these cells by TNFa- and IFNa- induced MAPK-p38 signaling and that total and phosphorylated p38 levels were increased in more senescent T cell subsets [120, 121]. These cells are unable to expand after TNF-induced activation due to increased levels of apoptosis, which was improved by p38 inhibition due, in part, to BCL-2 upregulation and reactivation of telomerase, suggesting that p38 signaling is involved in their capacity to proliferate and resist apoptosis. Signaling through the MAPK pathway activates p38 in response to environmental stress, inflammation, DNA damage [122] and the engagement of costimulatory receptors, in the case of T cells, the T cell receptor after antigenic stimulation [123]. In a set of experiments the same group showed that in CD27−CD28− CD4+ T cells, p38 was activated, and inhibited proliferation and telomerase activation, by an AMPK-TAB1 mediated mechanism, which was induced not only by endogenous DNA damage but also by a decrease in intracellular concentrations of glucose [45]. Similar to p38, AMPK or TAB1 silencing was able to restore proliferation in CD27−CD28− CD4 T cells and p38 activation induced senescence in CD27+CD28+ CD4+ T cells. These cells were able to reprogram their metabolic machinery by upregulating autophagy pathways after p38 inhibition [108].

This body of work provided “proof-of-concept” evidence that immune senescence in CD4 T cells is potentially reversible. p38 inhibition could be considered during manufacturing of T cell products for use in immunotherapies such as CAR T cells to improve the quality of the infused T cells although novel approaches such as harvesting these T cells sooner or using off-the -shelf allogeneic T cells is easier. However, there are several challenges in translating this into practice. For instance, the impact on the CD8+ compartment, which predominantly mediates anti-tumor responses is not known. The leukemogenic potential on T cells or other “on-target-off-T-cell” effects of targeting this common pathway in vivo are also not known.

TNF inhibition has been shown to decrease CD45RA+ effector CD4+ T cell subsets [124] and delay loss of CD28 on CD8+ T cells [54] but TNF activation of memory T cells did not increase CD45RA+ senescent T cells whereas IL-7 did [125]. Similar results have been reported for IL-15 in CD8+ T cells [126, 127]. IL-7 also reduced CD27/CD28 loss overall on T cells after exposure to tumor cells [21712448] and helped maintain the naïve T Cell pool in non-tumor models [128], underlying the pleiotropic effects of this cytokine. These observations may suggest that CD27−CD45RA+ cells may be derived directly from CD45RA−CD27+ naïve precursors, perhaps bypassing the memory state, due to cytokine-driven homeostatic proliferation. It is possible that only the CD45RA+CD27− effector subset expands in response to IL-7 and other senescent or pre-senescent/memory subsets do not. Human trials of IL-7 show that recombinant IL-7 preferentially expands the CD4+ naive and central memory T cells as well as CD8+ naive T cells resulting in a statistically significant broadening of T-cell repertoire diversity in both CD4+ and CD8+ T cells, within 3 weeks of treatment initiation [129]. These changes in the T cell pool reach their maximum levels quickly, possibly due to IL-7 receptor downregulation, but are sustained for weeks [130]. Even though this approach has not advanced further in clinical trials it might be time to reconsider it in the era of novel immunotherapies that drive clonal T cell expansion and T cell senescence, such as CD38 monoclonal antibodies [94]. Short term treatment with recombinant IL-7 prior to subsequent antibody-based therapy could theoretically benefit these patients. TNF or IFN inhibition might be less risky for heavily pretreated MM. We have shown [98] that MM cells after multiple lines of therapy, invariably including a CD38 monoclonal antibody, significantly downregulate IFN and TNF mediated signaling pathways, suggesting that their growth at this point may not be affected by these pathways. Finally, in vitro manipulation of T cell products with IL-7 and IL-15 for immunotherapy used is a promising approach [131] as are cytokine secreting CAR T products [132, 133]. More recently, ex vivo PI3K inhibition during CAR T cell manufacturing[134], allowed for persistence of early memory phenotypes which are associated with improved outcomes. The role of this pathway in normal T cell physiology is unclear.

Other approaches to “rejuvenate” the T cell compartment that have some limited evidence in humans include intense exercise [135] and hyperbaric oxygen therapy [136], although the mechanistic basis for these has not been described and their efficacy has not been validated. Finally, approaches that are implicated in replicative T cell senescence but might be difficult to bring to the clinic are restoration of p53 isoforms [137] and inhibition of KLRG1 signaling [57].

Concluding remarks

Immunosenescence may represent the next frontier in myeloma immunotherapy especially after the failure of PD-1 inhibitors and the advent of novel immunotherapies that rely on a functional T cell compartment to work. It increases with therapy and MM burden and identifying mechanisms that drive its development and ways to prevent it or reverse it may lead to improved efficacy of CAR T cell therapies or novel monoclonal antibodies for this disease. Studies in humans are essential given the major limitations of mouse models to study T cell immune senescence.

Highlights.

I thought highlights were not required for submission. Also, since this is not original research the abstract is better suited to summarize this descriptive review which has more than 1–2 keypoints.

Abbreviations:

ASCT

autologous stem cell transplant

BM

bone marrow

CAR-T cells

chimeric antigen receptor T cells

CMV

cytomegalovirus

DCs

dendritic cells

IFN

interferon

iTME

immune microenvironment

MAPK

mitogen-activated protein kinase

MDSCs

myeloid derived suppressor cells

MM

multiple myeloma

NK

natural killer

PI3K

phosphoinositide 3-kinase

RAG

recombination activating gene

SASP

senescence associated secretory phenotype

TNF

tumor necrosis factor