COVID-19, in common with other severe respiratory conditions [2], is associated with greater morbidity and mortality in older adults [3]. One of the potential explanations is that the ageing of the immune system makes older adults more susceptible to these infections, less well able to control them and more prone to harmful responses such as hyperinflammation and autoimmunity. Equally likely is the possibility that the infection itself would increase features of immunesenescence which may be acute or persistent. In the acute phase of infection, studies have shown evidence of an exhausted as well as an aged immune phenotype in COVID-19 patients, such as CD8 T cells and NK cells with reduced IL-2 and IFN-γ expression, reduced granzyme expression and degranulation (CD107a) and an increased expression of the inhibitory receptor NKG2A [40.41]. Moreover, this phenotype was more prominent with increasing disease severity suggesting that it may have influenced the compromised response to infection. Here we have taken these observations of individual cell phenotype changes forward and used a composite score of immune ageing, IMM-AGE [24, 35], to determine any association of immunesenescence with COVID-19 disease severity. We also recruited patients 3–5 months post-diagnosis to try and eliminate the influence of changes to immune cell profile in the acute phase. Our data reveal a greater degree of immune ageing, demonstrated by a higher IMM-AGE score, which was only seen in those with severe disease, though a trend to a higher score was also seen with moderate disease.
A key question addressed in our study was whether this higher degree of immunesenescence was present prior to infection, or was driven by the infection. Supporting the argument in favour of patients with severe disease potentially having a more aged immune system prior to infection is a study of participants in UK Biobank. For 347,571 individuals it was possible to calculate how biologically old they were when they enrolled in the study between 2006 and 2010, as opposed to their chronological age, using blood biochemistry data to derive the PhenoAge score. The analysis revealed that those participants who went on to develop severe COVID-19 were 10–14 years older biologically [42]. Crucially, in our study the prevalence of patients with pre-existing multimorbidity was highest in the moderate (46%) and severe (58%) groups, with only one patient with multimorbidity in the mild group. As the IMM-AGE algorithm was developed from longitudinal data and mortality [24] we would predict a higher score in a group at higher risk of death, i.e. those with multimorbidity or a high BMI [43]. However our multiple linear regression model revealed that neither multimorbidity nor BMI contributed significantly to the IMM-AGE score, instead South Asian ethnicity contributed to 17% of the increase in the IMM-AGE score. Interestingly we have shown recently that South Asians adults develop a broad range of immune-mediated diseases much earlier than white adults, possibly suggesting that their immune systems age more rapidly [44]. Our data thus suggest that the SARS-CoV2 infection itself results in an increase in immune ageing. The antigenic stimulation occurring during viral infections will certainly lead to telomere shortening, the appearance of more highly differentiated EMRA T cells, as well as exhausted and senescent T cells [45]. A similar state of acceleration of immune ageing has been observed in our studies in a younger cohort of traumatic injury patients [35], suggesting a negative influence of an acute challenge to immunesenescence. Our regression analysis revealed that the severity of disease made a significant contribution to the IMM-AGE score, supporting a major association of the SARS-CoV2 infection with immunesenescence.
Whether or not this enhanced immunesenescence is a result or consequence of COVID-19, it does suggest that these patients will be more vulnerable to future infections, show compromised vaccine responses and be at a higher risk of autoimmune disease. Moreover, as the induction of an aged immune system, specifically senescent CD4 T cells, has been shown in mice to be sufficient to drive an aged phenotype, including frailty and multimorbidity [46], our data may also suggest broader implications for the health of COVID-19 survivors. Evidence from recent studies has suggested the persistence of a spectrum of COVID-19 symptoms for up to 12 months after diagnosis, termed Long COVID, including persistent fatigue, myalgia and respiratory complications [47, 48]. Studies of COVID-19 convalescents 3–5 months post-infection have revealed maintained high levels of IL-6 associated with persistence of symptoms [4] and a study of autoantibody levels in serum found a high frequency of antibodies against the skin, skeletal muscle and cardiac tissue [5]. The aged immune system may thus contribute to both the acute and chronic sequelae of COVID-19.
T cell lymphopenia has been widely reported during the acute phase of COVID-19 infection [49] and a small study investigating the T cell profile in a cohort of 13 convalescent patients, four weeks post-resolution of infection observed a loss of naïve CD4 T cells and accumulation of memory T cells [50]. Here we show that a numerical deficit of CD4 T cells persists in severe cases several months post-infection, particularly in the naïve T cells. Previous studies have reported that infections can result in thymic atrophy and changes in thymocyte development [51], a potential explanation for the reduction in naïve T cells. The potential consequence is a reduced ability to respond to new pathogens, including substantially different SARS-CoV-2 variants and reactivation of latent viruses (e.g. EBV and herpes). In contrast, we observed an expansion of CD8 T cells with a central memory phenotype, which could provide long-term effective memory responses. Whilst previous studies in SARS infection found that central memory T cell responses persist for up to 4 years post infection [52], a recent study of 188 patients has shown that memory CD4 and CD8 T cell numbers decline with a half-life of 3–5 months [53].
In patients who had severe COVID-19 the CD8 T cell profile features an increase in cells with a phenotype of senescence and functional exhaustion, in agreement with previous reports from the acute phase [40, 41], suggesting that this is not a transient phenomenon. This observation not only raises concerns about the cytotoxic function of memory CD8 cells, but has broader consequences for health as senescent T cells are characterised by the secretion of a range of pro-inflammatory cytokines, chemokines, proteases and growth factors, termed the senescence-associated secretory phenotype (SASP) [54]. Thus, we hypothesise that the expansion of senescent T cells could be contributing towards the persistence of a pro-inflammatory environment in convalescent patients and symptoms such as fatigue and myalgia [48]. Another potential contributor to this inflammatory environment is the Treg/Th17 imbalance and increase in senescent CD57 expressing NK cells that we found also persisted in severe COVID-19 infection survivors several months post-acute infection. In support of this proposal, a study comparing the circulating immune profile of COVID-19 patients found an accumulation of senescent NK cells, Th17 cells and senescent T cells to be predictors for residual lung lesions [55]. The aged profile in the severe COVID-19 convalescent patients might therefore be contributing to impaired lung function and pulmonary fibrosis seen in some convalescent patients [56].
In addition to T cells, humoral immunity also plays a critical role in responding to viral infections and immunological B cell memory generated after infection is fundamentally important for protecting the host from severe disease upon re-exposure. In this study, we found reduced B cell numbers in convalescent patients irrespective of disease severity, but an expansion in the proportion of memory B cells and plasmablasts in severe COVID-19 patients. Multiple studies have detected virus-specific antibodies several months post-recovery from COVID-19, potentially a result of an elevated frequency of antibody-secreting plasmablasts [57, 58]. Furthermore, we observed an expansion of regulatory B cells, a subtype that produces IL10, in patients who had severe disease, which might be a compensatory mechanism for the expansion of pro-inflammatory immune cell subsets. These findings agree with another study reporting higher levels of IL10+ ve B cells in convalescent patients [59].
Although the detailed mechanisms driving a relationship between an aged immune phenotype and COVID-19 severity remain poorly understood, our RNA expression analysis has identified elevated inflammatory signalling, cellular senescence pathways and defects in DNA damage repair and autophagy, which are key processes underlying immunesenescence [60]. As the clinical consequences of immunesenescence include an increased risk of bacterial infections, reactivation of latent viruses, poor vaccine responses, increased risk of chronic inflammatory conditions [61, 62] and organ functional decline [63], severe COVID-19 survivors can thus be considered as a vulnerable population. Finding ways to alleviate immunesenescence should therefore be a priority to improve the health outcomes of these patients. Focusing on restoring thymic function could be considered a potential holistic treatment for rejuvenating the adaptive immune system and restoring immune homeostasis. The TRIIM (Thymus Regeneration Immunorestoration and Insulin Mitigation) trial has shown it is possible to boost thymic regeneration in older males using three agents: metformin, growth hormone and dehydroepiandrosterone given for 12 months [64]. An alternative method using an injection of Thymosin alpha 1(Tα1), known to support T cell generation and survival, reversed T cell exhaustion by boosting thymic output and reducing mortality in severe COVID-19 patients [65]. Another drug with anti-immunesenescence properties is metformin [66], which has been shown recently to reduce mortality in hospitalised COVID-19 patients [67]. Autophagy-boosting therapies, such as spermidine supplementation, have yielded promising results in rejuvenating an aged immune system in older adults [68]. Exercise has also been shown to induce its beneficial effects on body systems via the stimulation of autophagy [69]. Both interventions may therefore be useful in COVID-19 convalescent patients.
Our study has some limitations which should be considered when interpreting the findings. Firstly, we have only assessed immunological phenotype in convalescent patients and do not have longitudinal data from during the acute phase of infection, or prior to infection. We cannot therefore determine the degree to which the enhanced immune ageing was a cause or consequence of infection, though our regression analysis only found one pre-existing variable to influence the IMM-AGE score, namely South Asian ethnicity. Second, due to the collection of a limited volume of blood from the participants, it has not been possible to assess immune cell function in convalescent patients. However, our data does hint toward proliferative defects, skewing towards an inflammatory phenotype and TCR signalling defects due to overexpression of dual-specific phosphatase DUSP6, a feature of aged T cells that attenuates ERK signalling after TCR activation [70]. Third, by the very definition of healthy, our uninfected controls did not have any chronic disease and so were not well matched for the moderate and severe disease groups which had a substantial number of multimorbid patients. However, our regression analysis was able to show that the presence of multimorbidity was not a significant influence on immunesenescence. Additionally, there was variation in ethnicity amongst our four cohorts and this will have affected the data as the regression analysis revealed that South Asian ethnicity was an influence on the degree of immunesenescence.