Human T cell immunosenescence and inflammation in aging

Review of the origin of the mild proinflammatory state that characterizes many older individuals, with a focus on changes in T cell function over time.

Abstract

The aging process is driven by a finite number of inter-related mechanisms that ultimately lead to the emergence of characteristic phenotypes, including increased susceptibility to multiple chronic diseases, disability, and death. New assays and analytical tools have become available that start to unravel some of these mechanisms. A prevailing view is that aging leads to an imbalance between stressors and stress-buffering mechanisms that causes loss of compensatory reserve and accumulation of unrepaired damage. Central to this paradigm are changes in the immune system and the chronic low-grade proinflammatory state that affect many older individuals, even when they are apparently healthy and free of risk factors. Independent of chronological age, high circulating levels of proinflammatory markers are associated with a high risk of multiple adverse health outcomes in older persons. In this review, we discuss current theories about causes and consequences of the proinflammatory state of aging, with a focus on changes in T cell function. We examine the role of NF-κB activation and its dysregulation and how NF-κB activity differs among subgroups of T cells. We explore emerging hypotheses about immunosenescence and changes in T cell behavior with age, including consideration of the T cell antigen receptor and regulatory T cells (T_regs). We conclude by illustrating how research using advanced technology is uncovering clues at the core of inflammation and aging. Some of the preliminary work in this field is already improving our understanding of the complex mechanisms by which immunosenescence of T cells is intertwined during human aging.

Introduction

Human aging is a complex dynamic of inter-related processes that ultimately leads to rising susceptibility to acute and chronic diseases, as well as increased risk of physical and cognitive disability. The understanding of biologic mechanisms that cause major aging phenotypes is an essential prerequisite to discovering interventions that can delay the adverse consequences of aging, thereby increasing both lifespan and healthspan. Theories of aging have historically pointed to single causative factors (such as mitochondria dysfunction, mTOR activation, oxidative stress, etc.) as the “most important determinant of the aging process.” However, longevity likely results from evolutionary-driven adaptive solutions to multishaped environmental forces. According to this view, aging results from the progressive loss of reserve and robustness of a variety of interconnected homeostatic mechanisms necessary to maintain life in spite of multiform environmental challenges. Such complex interplay between many major biologic mechanisms of aging is so profound that their overarching contribution to disease susceptibility can all be detected, even in a specific chronic disease, such as chronic obstructive pulmonary disease [1].

Longitudinal studies have shown that effects of aging on risk of disease development, healthspan, and lifespan are pervasive but characterized by extreme heterogeneity. The search for biomarkers of aging and disease has embraced such complexity by integrating knowledge from animal models and humans and looking at the combined effects of multiple biomarkers through novel statistical methodologies of systems biology and artificial intelligence [2, 3]. Regardless of the method and the system, most studies conclude that dysregulated immune function is at the hub of most biologic changes that occur with aging and age-related diseases [4]. Here, we review the recent literature on the possible causes and consequences of such dysregulation, with particular focus on the new knowledge on changes in T cell functions.

INFLAMM-AGING

Aging is accompanied by a nonspecific state of chronic inflammation [3, 5], reflected by increased blood concentrations of several inflammatory biomarkers, including CRP, IL-18, TNF-α, and IL-6 [6]. Studies have also observed a concurrent rise in other biomarkers, including IL-1R antagonist, sTNFRI, and sTNFRII [3]. Although the biologic activity of these soluble species at the site of production is anti-inflammatory, they always emerge in the context of inflammatory conditions, and therefore, their blood levels are generally considered as proxy markers of a proinflammatory state. With the confirmation of the notion that a proinflammatory state is a hallmark of aging, studies of blood gene expression in animal models, as well as in humans, have consistently found that activated immune function and inflammation consistently rank at the top of pathways that are up-regulated with aging [7–9]. Independent of age, the state of chronic inflammation, sometimes referred to as “inflamm-aging” [10], is a strong risk factor for occurrence and progression of many chronic diseases, such as obesity, cardiovascular diseases, and neurodegenerative diseases, and affects their progression and complications [11–13].

Obesity, especially central obesity in the context of the metabolic syndrome [a metabolic risk state characterized by the co-occurrence of metabolic risk factors for both type 2 diabetes and coronary vascular disease (abdominal obesity, hyperglycemia, dyslipidemia, and hypertension)], is probably the most studied cause of a proinflammatory state in humans. Enlarged and stressed adipocytes in obese individuals can behave like immune cells and secrete proinflammatory mediators, such as IL-6 and TNF-α [14]. This phenotypic shift is particularly evident in adipocytes outside the subcutaneous compartment, which is the natural location of fat accumulation. Visceral, pericardial, and intramuscular fat is often infiltrated by macrophages that independently contribute to inflammation [15]. In addition, adipocytes in these same compartments tend to acquire senescent characteristics and express a SASP that further contributes to inflammation [16, 17]. Consistent with these findings, the Health Aging and Body Composition Study [18] found that obese older persons who are metabolically healthy have less visceral fat and greater subcutaneous fat and a more favorable inflammatory profile compared with their metabolically unhealthy, obese counterparts.

Chronic inflammation accounts for most of the excess risk of developing cardiovascular events in patient with the metabolic syndrome. In animal models and probably also in humans (although here, the mechanism is still unclear), inflammation is associated with insulin resistance and predicts the development of diabetes mellitus and cardiovascular disease [19, 20]. Contrary to this view, some recent research suggests that inflammation may also be beneficial. For example, Bapat et al. [21] showed that the depletion of fT_regs improved fat remodeling, thereby preventing age-associated insulin resistance. As fT_regs are thought to attenuate local inflammation and assimilate normally in the body fat of aging mice, the authors concluded that increased local inflammation may, under some circumstances, be beneficial [22]. Beyond obesity, other mechanisms that may cause chronic, low-grade inflammation include dysfunctional mitochondria, leading to production of excess and unopposed ROS, impaired autophagy, accumulation of senescent cells, and loss of integrity in intestinal mucosa leading to leaking of endotoxin and other toxins to the circulation. It is interesting to note that many of these proposed pathways intersect directly with the transcription factor NF-κB (further discussed below). For example, ROS and endotoxin are known to be activators of NF-κB [23, 24], whereas the SASP, which comprises several proinflammatory mediators (such as IL-1 and IL-8), is mediated via NF-κB-dependent gene expression [25, 26]. We conjecture that many pathways contribute differentially in individuals, but their cumulative effects ultimately lead to NF-κB activation and associated hallmarks of aging (Fig. 1).

Figure 1. Possible mechanisms leading to the mild proinflammatory state of aging.

A number of stressors that can potentially trigger inflammation and cell senescence are normally buffered by dedicated mechanisms. If these compensatory mechanisms fail, then the resulting stress is detected by signaling pathways that directly or through cell senescence lead to the production of inflammatory mediators and perturbation of immune cells. Both the accumulation of senescent cells and inflammatory mediators contribute to lack of tissue maintenance and repair and ultimately, contribute to the emergence of aging phenotypes and age-related diseases. An alternative hypothesis is that the chronic immune activation is caused by a primary defect in immune cells and does not require the presence of unbuffered chronic stressors. COX, Cyclooxygenase.

Inflammation is increasingly recognized as an important etiologic factor in AD. Many of the genetic variants associated with AD are associated with genes that are important for modulation of inflammatory responses, such as the TREM2 (triggering receptor expressed on myeloid cell 2) and the myeloid cell-surface antigen CD33 [27, 28]. Activated microglia that overproduce proinflammatory cytokines and have an impaired ability to phagocytose Aβ are detected in AD brain studies, and elevated CRP is an established risk factor for AD [27–29]. Aβ primes T cells to be autoreactive and directly or through the secretion of IFN-γ, results in the destruction of neurons and formation of plaques [29]. Perturbations of adaptive immune responses to persistent antigenic challenge appear to contribute to the pathogenesis of AD through decreased percentages of naive cells, elevated memory cells, and increased numbers of CD4⁺ and CD8⁺ cells lacking the costimulatory receptor CD28 [29, 30]. In addition, activated CD8⁺ T cells in the blood and CSF have been correlated with structural markers of AD pathology, as well as neuropsychological defects [31].

The role of inflammation in the pathophysiology of atherosclerosis is well established, with immune cells directly participating in the formation of the atherosclerotic plaque [32]. Autoantigens, such as heat shock protein 65, modified LDLs, and infections, such as Chlamydia pneumoniae, all can contribute to a proinflammatory state in the arterial wall and trigger infiltration of immune cells in the arterial wall intima. Activated CD4⁺ cells, recruited to the artery and associated IFN-γ production [29, 33, 34], may cause plaque necrosis and arterial rupture that precipitate coronary and cerebrovascular events [29]. Infiltration of effector memory CD4⁺ T cells in the carotid and coronary arteries correlates with the severity of atherosclerosis, and an increase in senescent memory CD4⁺ cells in the circulation is associated with subclinical atherosclerosis [35]. In addition, senescent CD4⁺ CD28^null T cells are associated with unstable coronary artery disease and correlate with the severity of inflammatory processes in ischemic heart disease [35]. Interestingly, higher levels of proinflammatory markers with aging, such as CRP, IL-18, and IL-6, are also found in individuals who are free of overt cardiovascular risk factors, major diseases, and disabilities [36]. Although low-grade chronic inflammation is characteristic of those with extreme longevity, it appears that in the oldest old such state of inflammation is counteracted by high level of molecules with anti-inflammatory properties [37, 38]. In an attempt to ascertain whether there was a predictive “immune risk phenotype” that would correlate with mortality over time, several large longitudinal cohort studies of Northern European populations found that those with high CD8⁺, low CD4⁺ counts, inverted CD4/CD8 ratios, a poor T cell-proliferative response, and persistent CMV infection had increased mortality [39]. In one of these Northern European longitudinal cohort studies, centenarians were found to have an “inverted” immune risk profile (a high CD4/CD8 ratio and low numbers of CD8⁺CD28⁻ cells) that was stable over time [40].

THE ROLE OF NF-κB

The specific molecular mechanisms that contribute to the mild proinflammatory state of aging are still not understood. The current lead hypothesis of a dysregulation of the NF-κB stems from the specific cytokines that appear to be dysregulated, namely TNF-α, IL-6, and CRP. NF-κB regulates inducible gene expression in response to a variety of stimuli in virtually every cell type that has been examined [41, 42]. The most important categories of putative NF-κB target genes include those involved in cell-cycle progression, control of cell death, regulation of cell adhesion, and genes that encode proinflammatory cytokines and chemokines, such as TNF-α, IL-1, and IL-6. Two important characteristics of the NF-κB system are especially pertinent to aging and aging-associated pathologies. First, NF-κB is activated by aging mediators, such as ROS, cellular senescence, and DNA damage. Thus, NF-κB is well positioned to mediate downstream effects of these aging stimuli. Second, a subset of proinflammatory NF-κB target genes, such as TNF-α and IL-1, also includes activators of NF-κB. Thus, NF-κB is poised as a critical feed-forward contributor to systemic inflammation. That is, environmental signals activate NF-κB, which leads to production of inflammatory cytokines, which in turn, leads to further production of NF-κB, thereby amplifying the homeostatic imbalance that could ultimately lead to the chronic inflammatory state of aging. Thus, it is appealing to hypothesize that interventions that block such NF-κB-mediated, self-sustaining inflammatory loops could possibly benefit age-associated chronic inflammatory diseases. However, it should be noted that NF-κB-dependent genes protect against cellular apoptosis. For example, TNF-α stimulation impedes apoptosis in an NF-κB-dependent production of FLIP_L, an inhibitor of caspase 8 [43]. The possible role of NF-κB during aging has been substantiated in several mouse models in which genetic manipulation of the NF-κB pathway has accelerated or attenuated aging phenotypes [44, 45]. Whereas these studies clearly demonstrate that changes in NF-κB activity can affect aging in a predictable manner, it is more difficult to provide direct evidence for the role of NF-κB during normal aging. This is especially true in humans.

NF-κB activity varies among subgroups of T cells. Huang et al. [46] showed a role for NF-κB in Th17 cytokine generation by using inhibitors of NF-κB activation to suppress generation of IL-17 and IL-21, as well as IFN-γ, by CD4 T cells of young and old mice. CD8⁺CD28⁻ T cells that accumulate in aged humans increased NF-κB following activation with TNF-α [47], and older subjects had decreased lymphocytic NF-κB activation following stimulation with TNF-α compared with young individuals [48]. In this latter case, reduced NF-κB activation was associated with increased cellular apoptosis. These observations suggest an NF-κB-dependent pathway to explain an apparent contradictory coexistence of heightened chronic inflammation and reduced immune responses in the elderly. Heightened NF-κB induction in senescent functionally low-responding CD8⁺CD28⁻ T cells could contribute to chronic inflammation, whereas an attenuated NF-κB response in immune-competent cells could lead to reduced responses. Our group has studied cell-intrinsic differences between older and younger individuals and has shown that in the absence of experimental stimulation, gene expression coordinated by NF-κB is up-regulated in CD4⁺ T cells from older compared with younger individuals. Genes that were up-regulated in T cells from older compared with younger individuals included proinflammatory cytokines, such as IL-1 and IL-6, and chemokines, including CCL2 and CXCL10 [49]. Similar to our observations in CD4⁺ T cells, NF-κB transcriptional signature has also been observed in other human tissues from the elderly, including skin, liver, muscle, cerebellum, cardiac muscle, gastric mucosa, and kidney [50, 51].

In addition, we found that NF-κB up-regulation was cell intrinsic and mediated, in part, by PI3K activity that was induced in response to metabolic activity [49]. Accordingly, expression of a subset of these genes was moderated by rapamycin treatment that ameliorated age-associated gene-expression patterns and demonstrated that the mTOR pathway was up-regulated in the cells of older individuals. This observation is particularly interesting, as inhibition of mTOR has been shown to extend the mammalian lifespan [52]. We have proposed that the reduced ability to maintain metabolic homeostasis in cells from older individuals is a result of basal dysregulation of NF-κB activity, which leads to basal up-regulation of inflammatory cytokines, thus contributing to age-associated chronic inflammation and its corresponding effects on health [49]. Our findings are consistent with M. V. Blagosklonny’s theory [52] that mTOR-driven hyperfunction causes aging through molecular damage accumulation and that PI3K activation drives many aging phenotypes via mTOR activity [53, 54]. However, we also showed that PI3K inhibition had a greater effect than rapamycin treatment, which suggested that PI3K hyperactivity contributes more than mTOR activation toward the aging phenotype [49]. At a functional level, a study of mTOR inhibition with RAD001 ameliorated the decline in immune function in elderly volunteers, as assessed by an increased response to the influenza vaccine of 20% [55]. RAD001 also reduced by >30% the percentage of CD4 and CD8 T cells that expressed the PD-1 (programmed death-1) receptor that inhibits T cell signaling and is increased with age [55].

We note that NF-κB target genes associated with human aging may also be activated by other transcription factors; therefore, up-regulation of these genes in the absence of overt cell stimulation should not be interpreted as indicating sufficiency of NF-κB in mediating the aging phenotype. There is also increasing evidence for post-transcriptional regulation of NF-κB target genes, including de-stabilization of target mRNAs and inducible RNA splicing [56–58]. Therefore, the increased steady-state levels of putative NF-κB target genes observed in tissues from the elderly could also arise from age-associated changes in mRNA stability and splicing. Future studies toward understanding the relative contributions of the multistep process of proinflammatory gene expression will greatly help to pinpoint critical aging-dependent changes and thereby, potential targets for therapeutic intervention.

IMMUNOSENESCENCE

Adequate functionality of physiologic systems that protect individuals from environmental stresses and attack of other organisms is critical for survival. Acute inflammatory responses and other responses by the immune system are an essential component of this defensive network. In normal conditions and in young and middle age, the immune system is quiescent but able to mount a strong but transient dynamic response promptly after detecting an “invasion.” However, during the aging process, the immune system appears to maintain a permanent state of mild activation, and in parallel, when stimulated, the amplitude of the dynamic response is compressed. The combination of a chronic proinflammatory state and reduced ability to mount an effective defense is often referred to as immunosenescence [59, 60]. As immune competence decreases with age, morbidity and mortality from infections increase [61], and response to immunologic challenges, such as flu vaccination, is lessened [59, 61, 62]. Furthermore, tumor surveillance decreases, which according to some researchers, is the reason why the incidence of malignancies increases with aging [63]. In addition, immunologic aging can impair tolerance mechanisms, leading to the risk of autoimmune disorders [59, 64].

The lack of effective response to vaccines can have serious consequences, as poor response to flu vaccine in the elderly is a major source of influenza associated mortality during hospitalization [65–67]. Likewise, the efficacy of live, attenuated herpes zoster vaccine (ZOSTAVAX) decreases with increasing age: 70% in persons 50–59 yr old, 64% in persons 60–69 yr old, 41% in persons 70–79 yr old, and 18% in persons ≥80 yr old [68]. Interestingly, a study (ZOE-50) of adults 50 yr of age or older receiving the herpes zoster subunit vaccine (HZ/su), which also included recombinant varicella-zoster glycoprotein E and the AS01_B, had an overall efficacy of 97.9% regardless of age [69]. As the study was not intended to assess definitively the vaccine efficacy against herpes zoster in those ≥70 yr old, a second trial (ZOE-70) was conducted in parallel that showed an efficacy of 89.8% in those ≥70 yr old, although there was no significant decrease when those 70–79 yr old (90.0%) were compared with participants ≥80 yr old (89.1%) [70]. Adjuvants thus appear to improve the efficacy of vaccines in older adults and perhaps overcome immunosenescence [69, 70].

Consequences of age-associated immune dysregulations can extend well beyond inadequate protection from pathogens. One of its most widespread consequences could be a direct contribution to initiation and maintenance of the low-grade proinflammatory state. As professional inflammatory cells, immune cells are developmentally programmed to mount short-lived acute inflammatory responses to fight pathogen. A key aspect of their normal function is the ability to re-establish rapidly a quiescent ground state that is no longer inflammatory. One possibility is that mechanisms that restore such homeostasis are diminished with age, resulting in temporal extension of the inflammatory response. Alternatively, as immune cells are programmed to induce inflammatory gene transcription rapidly and robustly, relatively small changes in cellular metabolism may result in low-level induction of these genes, even without overt cell stimulation. Our observations in CD4⁺ T cells are consistent with this idea. The cumulative effect of such dysregulation would be the establishment of low-grade inflammation.

A second area (beyond lack of immunity) where immune dysregulation may negatively impact the well-being of the elderly is in wound healing/tissue repair. Tissues damaged by trauma, chemical aggression, or organismal degenerative processes undergo repair through a series of well-coordinated steps, most of which require the participation of immune cells. Specialized macrophages remove debris of damaged tissue and through soluble factors, such as IL-10, IL-4, and TGF-β, and the proliferation and commitment of stem cells stimulate the repair process [71, 72]. Thus, an interesting possibility is that defective immunity not only reduces defense again invading micro-organisms and surveillance against cancer transformation but also contributes to degenerative diseases by reducing efficiency of tissue repair. Finally, in their role as danger sensors and responders, immune cells access most tissues and organs and during this journey, receive and convey signals from and to cells and tissues. These constant interactions between immune cell types and other organs may contribute to making the aging process uniform across the organism. Of note, cell-to-cell communication or lack thereof is considered one of the hallmarks of the aging process [73]. Immune cell interactions with other systems may imply that biologic signals detected in these cells may provide information on the physiologic and pathologic process that occurs in many different parts of the body and on the other hand, open the possibility that blood-borne cells adequately manipulated could be used in individualized treatments. Taken together, such considerations strongly indicate that the study of immune dysregulation and its mechanisms will broadly impact our understanding of the aging process, as well as identify means to enhance human healthspan.

T CELL BEHAVIOR DURING IMMUNOSENESCENCE

T cell function undergoes profound and consistent change during aging [74]. With aging, T cells have been noted to have a progressive shrinking of the immune repertoire, with a reduced proportion of naive cells and increased proportion of memory cells in older animals [10, 38]. Studies in humans have shown an increased number of memory and effector cells during aging in both the CD4⁺ and CD8⁺ subsets [75]. Zanni et al. [76] noted that CD8⁺ cells (naive, memory, and effector) showed a continual increase with age (young, middle age, and old adults) of proinflammatory cytokines IFN-γ, IL-2, and TNF-α and that IL-4, IL-6, and IL-10 only increased with rising age in memory CD8⁺ cells. The CD8⁺CD28⁻ subset of cells (with a loss of the CD28 costimulatory molecule) in the expanded memory cell population has shortened telomeres, suggesting that they have a longer replicative history [77]. In humans, almost all T cells express CD28 at birth, and the proportion of CD28⁺ declines by the age of 80, to 10–15% of CD4⁺ and 50–60% of CD8⁺ cells [77, 78]. The increase of these cells has been observed consistently and is used as a prognostic indicator of immunosenescence in older populations [76].

As the ability to recognize diverse antigens is central to immune responsiveness, a reduction of the repertoire reduces immune efficiency and creates an imbalance of peripheral lymphocyte homeostasis with aging [79]. These changes are accompanied by progressive shrinking of the thymus—the primary site of T cell development. The progressive involution of the thymus and the decrease in thymopoiesis are the reasons that the number of naive T cells declines during aging [80]. In parallel, there is an increase of antigen-experienced memory and effector T cells, and hence, the absolute number of T cells stays roughly the same [81]. The adaptive immune system is unique in that the production of novel, naive T cells is entirely dependent on thymic function. However, unlike mice, repertoire maintenance is independent of the thymus in humans [82]. As other cell-intrinsic alterations occur with age, including impaired cytokine production and reduced proliferation of activated T cells, it is difficult to ascertain the relative contributions of a reduced antigenic-response repertoire to age-associated dysregulation.

TCR

TCR diversity contracts with age, and older persons are at increased risk of generating oligoclonal cell populations [83, 84]. One of the causes of increased oligoclonality in older individuals is that CD8⁺CD28⁻ T cells, which are enriched in the older population, appear to have receptor specificities for CMV antigens. These findings have raised the question of whether chronic antigenic stimulation is the primary cause of oligoclonality [74].

Although memory and effector T cells increase with age, response to vaccination in older persons is dampened [59, 61, 62], which is indicative of a gradual decline in functional response. As cytokines modulate the immune response, it has been proposed that alterations in cytokine production and response may explain T cell functional defects [74]. Classically, CD4⁺ subsets are grouped by their cytokine-secretion pattern, with Th1-secreting cytokines, such as IFN-γ; Th2, such as IL-4, IL-5, IL-10, and IL-13; Th17, such as IL-17 and IL-22; Th22, such as IL-22 and TNF-α; Tfh, such as IL-21 and ICOS; and T_regs, such as IL-10 and TGF-β [85, 86]. With aging, there is a shift in the induction profile of cytokines to predominately IL-4 and IL-10 (Th2-type profile). This change of the total T cell cytokine response to a Th2 type rather than a Th1 type (IL-2, IFN-γ) may be responsible for some of the age-related decline of Th1 activation [74, 87]. Other Th cells can have functional changes with age. Studies have demonstrated a T cell-intrinsic decline in somatic hypermutation and selection in germinal centers, which suggests that Tfh cell function is altered with age [88]. Tfh cells highly express the costimulatory molecule CD40L and are essential for germinal center function and response in humans and mice. As CD40L has decreased expression in CD4⁺ cells from older donors, this may affect proper functioning of the germinal center response [88]. Th17 cells provide protection against bacterial infection and are associated with the development of autoimmune diseases and chronic inflammatory diseases in humans [89]. Th17 cells, which are proinflammatory, are in homeostatic balance with T_regs, which are more anti-inflammatory and are derived from a common precursor [86]. The ratio of Th17 cells to T_regs appeared to increase with aging in humans and would favor a basal proinflammatory picture, whereas with stimulation, the ratio decreased in aged individuals [89]. The authors suggest that this alteration of Th17 cells to T_regs ratio may partially explain why autoimmune disease is enhanced with age, as T_regs recognize self-antigens, but reactions to infections are decreased in the elderly [89]. T_regs have several changes with age and will be discussed separately in a later section.

Studies have shown, mostly with murine lymphocytes, that the TCR is also functionally impaired with age through various mechanisms, including reduced IL-2 production, reduced phosphorylation of STATs, reduced inhibition of SHP-1 (Src homology region 2 domain-containing phosphatase 1) activation, and changes in membrane rafts of T cells [90, 91]. More recently, the level of the DUSP6 (dual-specificity phosphatase 6) was shown to increase with aging in human CD4⁺ T cells, causing a reduction of ERK activation in response to TCR signaling [92]. This impairment may be responsible for the reduced cytokine production and proliferation of cells from older individuals. We have found that activation of the TCR produces different patterns of gene expression in CD4⁺ T cells obtained from older compared with younger individuals. A subset of genes was underexpressed, presumably reflecting reduced TCR signaling; however, a subset of genes implicated in age-associated chronic inflammation, including IL6, IL1α, and IL-15, were overexpressed. These findings support the notion that CD4⁺ cells are a source of age-associated inflammatory cytokines. In addition, maintenance of the inducible transcription factor NF-κB after activation of CD4⁺ T cells was reduced in older individuals [93], providing a plausible explanation for reduced functional responses with age.

Gene-expression profiles show age-associated changes related to T cells and TCR signaling. CD8⁺ T cells from young compared with older individuals revealed many genes that were differentially up- or down-regulated between the two age groups [94]. Enhanced cellular pathways in the older group included apoptosis, the MAPK kinase signaling cascade, and response to oxidative stress and cytokines, whereas cellular pathways related to intracellular transport, RNA and DNA metabolism, RNA transcription regulation, and the ubiquitin cycle were down-regulated [94]. A study by Remondini et al. [95] revealed complex patterns of gene expression in peripheral T cells as a function of age. A subset of genes increased in middle age and then decreased, whereas others decreased in middle age and were followed by increased expression in older individuals. Yet, other genes decreased with age. A subset of genes whose basal expression increased with age had categories related to TCR and cytokine signaling. One of the interesting observations in this study was that the gene-expression profile of the very old (age >90 yr) closely matched the patterns seen in the younger groups (age 25–50). With the use of PBMCs from the InCHIANTI cohort (a large study of 733 individuals), Ferrucci and colleagues [96] examined gene-expression changes and found that mRNA splicing was significantly decreased in older compared with younger individuals. Taken together, these studies indicate that age has wide and varied effects on T cell and TCR function, and therefore, gene-expression profiles must be analyzed in context and in age cohorts beyond simply “young” vs. “old.”

T_regs

T_regs have a central role in immune homeostasis and function to keep aberrant and exaggerated immune reactions in balance [97]. The FoxP3 transcription factor specifies the T_reg lineage during development, and its continued expression is necessary for suppressive function [98]. As most T_regs recognize self-antigens, a decrease in FoxP3 expression and loss of suppressive function may result in autoreactive cells that are associated with autoimmunity [98]. T_regs are produced in the thymus (natural T_reg) or extrathymically in peripheral lymphoid tissue (induced T_reg) [99]. Although regulatory function has been ascribed to other cells, including NK T cells, CD4⁺ and CD8⁺ T_regs have been most studied in the context of aging [99]. CD4⁺ T_regs express high levels of CD25⁺ [100], and with age, CD25^lo T_regs accrue [101]. Aging also affects the number of circulating T_regs, subset distribution, and their functionality [99]. The frequency of Foxp3⁺ CD4⁺ T_regs significantly increases with age in mice and individuals [102, 103] and contributes to age-related immunosenescence. Examples of functional defects include aged T_regs that are unable to control Th17 expansion during intestinal inflammation secondary to impaired STAT3 activation [104] and in aged IL-6-deficient mice, a significant blunting of T_reg accrual, demonstrating a role for IL-6 in promoting the accumulation of T_regs with age [103]. Furthermore, when FoxP3⁺ CD4⁺ T_regs and CD4⁺ CD25⁻ T cells were cocultured in equal numbers, the production of the anti-inflammatory cytokine IL-10 from CD4⁺ CD25⁻ T cells was more suppressed in cells from older individuals [105]. Although CD8⁺ T_regs have been studied less than CD4⁺ T_regs, there is some evidence that CD8⁺ T_regs increase in older individuals [106], while still being able to maintain immunosuppressive function. CD8⁺ T_regs, as opposed to their CD8⁺ effector counterparts, lack cytotoxic activity and are hyporesponsive to secondary stimulation [99]. Induction of CD8⁺ T_regs in the periphery declines with age [99].

ROLE OF PATHOGENS

Pathogens are believed to have an important role in shaping T cell function with age and affecting inflammaging and immunosenescence. CMV, for example, infects >90% of individuals in the United States by the age of 80 yr old [107]. CMV, in general, leads to a latent infection that remains asymptomatic for decades but can lead to serious illness in those with a reduced immune protection [108]. Long-term control of the CMV infection requires considerable resources from the immune system. Sequelae of CMV infection include T cell senescence through a reduction of the TCR repertoire and clonally expanding T cell subsets. The latter phenomenon especially affects CMV-specific memory CD8⁺ T cells, including CD8⁺ CD28⁻ T cells [109] that are greatly increased in the elderly. This process has been called “memory inflation,” which is often associated with a proinflammatory profile [110]. Recent data suggest that such memory inflation is also associated with impaired immunity in the elderly [108]. Increased mortality in studies of Swedish longitudinal study cohorts [111, 112] has been correlated with CMV-related loss of naive T cells and an expansion of CD8⁺ effector-memory T cell populations, resulting in an immune risk phenotype [108, 111, 112], consisting of high CD8 and low CD4 numbers (an inverted CD4/CD8 ratio).

Chronic HIV infection also illustrates how T cells can be perturbed with aging in the presence of a pathogen. The discussion of the wide range of immunologic changes induced by HIV is outside the scope of this review, and we will limit this example to the aging field. Individuals with HIV have high incidence of age-related comorbidities, such as type II diabetes, cardiovascular disease, cognitive impairment, and certain cancers that occur earlier than in non-HIV-infected persons [113]. These data have been taken as suggesting that chronically infected HIV individuals experience a syndrome of accelerated aging, where deleterious features normally associated with age appear decades earlier. This may be, in part, caused by cellular senescence, which in HIV patients, may be induced by the virus and antiretroviral therapy [114]. In individuals with HIV, increased CD38 expression, a measure of disease progression, was associated with a decrease in CD8⁺ T cell telomerase activity [115]. In these HIV-infected persons, other biomarkers of immune status, such as adenosine deaminase, were found to predict CD8⁺ T cell activation. These changes in the CD8⁺ T cells of HIV⁺ individuals illustrated the progression of these cells to end-stage replicative senescence, a cellular state characterized by shortened telomeres, irreversible cell-cycle arrest, and genetic and functional cellular changes. The accumulation of senescent CD8⁺ T cells with HIV infection mirrors the accumulation of these cells in chronological aging [115]. Most HIV individuals are also infected with CMV, which exacerbates the immune risk phenotype [113]. In addition, HIV infection leads to a rapid depletion of memory CD4⁺ T cells in the gut [116] in disproportion to the peripheral blood and lymphoid tissues [117], which lead to intestinal epithelial deterioration and increased microbial translocation of LPS and resultant systemic immune activation [118]. Steele et al. [119] described a novel inflammaging phenotype in HIV-infected individuals based on T cell activation (sCD27), inflammation (high-sensitivity CRP), intestinal epithelial damage, and microbial translocation (sCD14) and a high risk of cardiovascular disease. The overall exhaustion of the T cell compartment with chronic HIV infection is similar to abnormalities that occur in aging, including a decreased T cell repertoire, reduction of naive T cells, involution of the thymus, and oligoclonal expansion of memory/effector cells directed toward infectious agents [116].

miRNA, TLR SIGNALING, AND NF-κB

The innate immune system is activated by recognition of conserved viral/microbial structures by cell receptors, including, importantly, the TLRs. Eleven TLRs (TLR-1–11) have been described in humans, and specificities range from bacterial lipoproteins, endotoxin LPS, flagella proteins, and uro-phatogenic bacteria. In addition, there is some evidence of specificity for viral dsRNAs and pathogenic nucleic acids, including both ssRNA and dsDNA. These pathogen-derived ligands engage with specific TLRs and transduce signals that lead to NF-κB activation and subsequent induction of antiviral and proinflammatory response genes [120]. miRNAs are ∼22 nt long, noncoding ssRNAs that post-transcriptionally regulate gene expression and are important in immune regulation, including T cell immunity [121]. In T cells, they help regulate proliferation, apoptosis, development, differentiation, and function. Examples of miRNAs affecting T cell function through NF-κB include miR-21 (which targets PTEN (phosphatase and tensin homolog), an inhibitor of AKT, which activates NF-κB) and miR-155 [that may control the IKKB (also called IKK2) and IKKe, which leads to repression of NF-κB activation] [120], and both regulate T cell activation. Another miRNA, miR-146a, serves as a negative-feedback regulator and suppresses the NF-κB signaling transducer TRAF6 (TNFR-associated factor 6) and IRAK1 (IL-1R-associated kinase 1) to reduce NF-κB activity; deficiency of miR-146a leads to chronic inflammation, similar to inflammaging, whereas TCR stimulation will activate NF-κB and up-regulate miR-146 expression [122, 123]. Regarding pathogen defense, lack of miR-155 prevents the accumulation of CD4⁺ Tfh cells, important in the adaptive immunity response against pathogens and inflammaging using a mouse model of age-dependent inflammation resulting from a deficiency of miR-146a [124]. It is believed that miRNAs can modulate TLR signaling by directly targeting components of the TLR signaling system and directly activating the RNA-sensing TLRs; conversely, miRNA expression can be directly regulated by TLRs [120]. Most miRNAs interact with the TLRs pathway molecules through a negative-feedback loop to prevent an excessive proinflammatory response by TLR signaling activation [120]. Hence, deregulation of these miRNAs may contribute to aberrations in the inflammatory/anti-inflammatory balance and can induce chronic TLR pathway activation through direct binding and consequently, contribute to the development of many age-related inflammatory diseases, including diabetes, cardiovascular disease, cancer, and AD [120].

THE FUTURE; NEW TECHNOLOGIES AND STATISTICAL TECHNIQUES FOR OMICS AND SYSTEMS BIOLOGY APPROACHES

A fast-growing literature has outlined profound changes that occur in the immune system with aging. Some of these changes are degenerative in nature, whereas some others are probably adaptive responses to external or internal stresses. To approach the complexity of the inter-relatedness of immunosenescence, inflammation, and aging, the field is moving toward an integration of omics research and system biology approaches. Molecular profiling of the changes that occur with age is increasingly being examined by high-throughput omics technologies, including genomics, metagenomics, transcriptomics, and metabolomics [125], and has been used to examine the role of T cells in aging. Outcomes, such as mitochondrial DNA heteroplasmy, DNA methylation, miRNA expression, serum metabolite levels, and metagenomic diversity and composition of the gut microbiome, are being applied to aging research in combination with environmental factors, such as diet and lifestyle, in cross-sectional, longitudinal and longevity studies [125]. Large profiling projects have been launched to examine the underlying networks that relate human immune cells to aging with regard to T cells. Reynolds et al. [126], for example, have examined the transcriptome in >1000 individuals, ages 55–94 yr. They detected a decline in the expression of ribosomal synthesis genes with aging in CD14⁺ monocytes and CD4⁺ T cells. In another study, they [127] examined age-related variations in the methylome of CD14⁺ monocytes and CD4⁺ T cells looking at age-dMS. They also examined potentially functional age-dMS (age-eMS) by integrating genome-wide CpG methylation and gene-expression profiles from circulating T cells and monocytes. They found that age-eMS were hypomethylated with increasing age, were located in predictable enhancer regions, and were linked to expression of genes in antigen processing and presentation. Weng and colleagues [128] have reviewed transcriptional changes from a genome-wide analysis of T cell aging, both of thymocytes and peripheral mature T cells, as well as subsets of T cells accumulated with age, and assembled gene-clustering relationships. Searching for a biologic signature of the proinflammatory state of aging, Ferrucci and colleagues [9] examined gene-expression markers of age-related inflammation for 17,000 genes in a large cohort of patients (>2000) and used a second cohort to replicate results related to IL-6 protein levels in blood. Among mediators that correlated with age-related protein levels and were replicated, the largest effect gene was SCL4A10 (solute carrier family 4 member A10), encoding for NCBE, a sodium bicarbonate transporter; others included PRF1 (perforin, a cytolytic protein in cytotoxic T cells and NK cells) and IL1B (IL-1β).

Researchers are increasingly moving toward a systems biology approach integrating these data, using advanced bioinformatics approaches to understand better the effect in overall context and dynamic of the individual. This integration requires both technological and bioinformatics advances to handle “big data.” High-throughput technological advances are improving our ability to embrace the complexity required for integrating omics studies in a systems biology approach. The examination of “omics,” in general, has been limited by the speed and throughput of analytical technology. Montenegro-Burke et al. [129] have addressed this issue in a study of T cell metabolism. They developed a data-streaming platform, XCMS stream, which mimics just-in-time production strategies from the manufacturing industry. In their metabolic liquid chromatography–mass spectrometry experiments on T cells, data transfer time was reduced from 7.3 to 1.6 d when batch streaming was used instead of traditional manual uploading. Advances in high-throughput technologies for systems biology approaches necessitate the use of new, sophisticated statistical techniques and access to large sets of data. Examples of research based on this complex approach are starting to appear in the literature, which encompasses immunosenescence as one of the regulatory networks that modulate physiologic processes in aging, and try to elucidate both functional and structural inferences [2]. Although straightforward multivariate statistics can be used, there will be a time when high-dimensional analyses, such as those used in systems biology, may become necessary to understand the panoply of interactions between the multivariate processes related to aging. Ferrucci and colleagues [3] have also applied principal component analysis in examining how 19 inflammatory biomarkers interact to enable more robust interpretation of the various relationships among markers. Principal component analysis identifies “key” axes through a multivariate approach to create linear combinations of original variables and coefficients to deduce a biologic interpretation of each axis. With the use of this technique, Ferrucci and colleagues [3] revealed a novel biologic structure in the relationship among inflammatory markers that illustrated important roles of revealed axes in chronic disease. Specifically, the key axes—inflammatory activation and innate immune response—were both strongly predictive of mortality, whereas the inflammatory activation axis (both pro- and anti-inflammatory markers) was strongly correlated with age.

CONCLUDING REMARKS

Despite intense research in this field of immunosenescence, the causal mechanism that connects aging and inflammation remains elusive. We believe that enhanced data analysis modalities, combined with precision-based techniques, such as gene-expression arrays, will continue to lead to major changes in our understanding of how inflammation and the immunosenescence of T cells are intertwined during human aging. At the National Institute on Aging, NIH, we are combining these approaches to examine—through a study of genetic, epigenetic, and protein—age-related changes in blood (including T cell populations) compared with other tissues that correspond to hormonal and inflammatory processes that lead to age-related diseases (Genetic and Epigenetic Signatures of Translational Aging Laboratory Testing study). We are implementing many of the aforementioned technological and analytical modalities to understand better how the inflammatory process interacts with immune system senescence to lead to a decrease of function and increase of disease with age. We hope to discern further how the important roles of human T cell immunosenescence and inflammation interact with aging to lead to increased morbidity and mortality and target key pathways to improve the healthspan in the elderly.

AUTHORSHIP

A.B., S.H.S., R.S., and L.F. contributed to the written body of the review. L.F. created the figure.

Glossary

Aβ

amyloid β peptide

AD

Alzheimer’s disease

age-dMS

differentially methylated CpG sites by age

age-eMS

age- and cis-gene expression-associated methylation sites

CD40L

cluster of differentiation 40 ligand

CRP

C-reactive protein

fT_reg

fat-resident regulatory T cell

IKK

IκB kinase

miR/miRNA

microRNA

mTOR

mechanistic target of rapamycin

ROS

reactive oxygen species

SASP

senescent-associated secretory phenotype

sCD

soluble cluster of differentiation

sTNFRI/II

soluble TNFR I/II

Tfh

Th follicular cell

T_reg

regulatory T cell

DISCLOSURES

The authors declare no conflicts of interest.