Nutritional status has a profound influence on resistance to infection, which is exemplified by the vicious cycle between undernutrition and infection in developing countries( Reference Calder and Yaqoob 1 ). However, vulnerable groups in developed countries are also at risk of age- or disease-related malnutrition, which can impact on the immune response to infection and to vaccination. Thus, while decreased immune function due to malnutrition primarily affects children in developing countries, in the developed world, it is mainly a problem for older people( Reference Gavazzi and Krause 2 ). By 2050, approximately 25 % of the population will be older than 65 years( 3 ) and the impact of this on public health is a major global challenge. However, decreased immune function as a result of malnutrition should not be confused with immunosenescence; an obvious difference is that malnutrition and, to some extent its consequences, are treatable. Immunosenescence is irreversible and describes the biological ageing of the immune system, which is associated with a progressive decline in both innate and adaptive immunity, poor response to vaccination and increased prevalence of cancer, infections and autoimmune and chronic diseases. While nutritional interventions may delay this process, the evidence for this remains controversial, particularly in terms of the nature and potency of immunomodulatory activity and of translation into a corresponding change in clinical outcome( Reference Maijo, Clements and Ivory 4 , Reference Pae, Meydani and Wu 5 ). Furthermore, there is a fundamental lack of understanding as to how immunosenescence alters the response of cells of the immune system to dietary components. Most studies examining the effects of diet on immune function fail to adequately characterise target populations in terms of nutritional status, health status, genetic background and few, if any, characterise them in terms of the immunological status. This review focuses on two case studies, which demonstrate that failure to account for immunosenescence can significantly influence the outcome of a nutritional intervention. It also explores proposed mechanisms by which ageing alters metabolic regulation of immune cells and whether metabolic pathways could be targeted for immunoregulation.
Case study: ageing alters the immune response to n-3 PUFA
Fatty acids play diverse roles in all cells, serving as an important source of energy, as structural components of cell membranes, signalling molecules, bioactive mediators and regulators of gene expression. Human immune cell phospholipids contain about 1 % EPA and 2·5 % DHA in addition to 20 % arachidonic acid( Reference Calder 6 , Reference Fritsche 7 ). As the long-chain n-3 PUFA content of the diet increases, lymphocyte arachidonic acid decreases in a curvilinear fashion. In human studies, dietary n-3 PUFA never exceeds 3 % of total energy, whereas in animal studies, intake is often considerably higher, and this is thought to explain the discrepancies that exist between animal and human studies investigating the immunomodulatory effects of n-3 PUFA( Reference Fritsche 7 ). As a result, it remains unclear to what extent and at what dose n-3 PUFA have immunomodulatory effects in human subjects. Nevertheless, the literature suggests that fish oil has a greater impact on immune function in elderly compared with young subjects( Reference Meydani, Endres and Woods 8 – Reference Thies, Nebe-von-Caron and Powell 10 ) and that this may be related to the fact that older subjects appear to incorporate EPA into plasma and peripheral blood mononuclear cells more readily than younger subjects( Reference Rees, Miles and Banerjee 11 ) (Fig. 1). EPA resulted in a dose-dependent decrease in neutrophil respiratory burst in older, but not younger subjects( Reference Rees, Miles and Banerjee 11 ). However, prostaglandin E2 production by peripheral blood mononuclear cells was decreased in both groups and phagocytosis and cytokine production were not affected in either group( Reference Rees, Miles and Banerjee 11 ). This highlights the fact that age is likely to be an important factor when considering the impact of n-3 PUFA on immunity, not only because of the influence of immunosenescence, but also because immune cells from older subjects appear to be more responsive to the availability of n-3 PUFA. Recent work suggests that the cholesterol content of T lymphocytes from healthy elderly subjects is higher than that of young subjects, and that membrane fluidity is subsequently decreased( Reference Larbi, Douziech and Dupuis 12 ). Furthermore, the coalescence of lipid rafts at the site of T cell receptor engagement is impaired in elderly subjects( Reference Larbi, Douziech and Dupuis 12 , Reference Fulop, Larbi and Douziech 13 ). The impact of ageing on lipid raft composition and function appears to be most evident in the CD4+ T cell population and affects cytokine signalling( Reference Fulop, Larbi and Douziech 13 , Reference Larbi, Dupuis and Khalil 14 ). Thus, the greater responsiveness of T cell membranes to n-3 PUFA in older subjects could result in alteration of lipid raft structure, and subsequently of cell function, effects which are absent in younger subjects.
Fig. 1. Arachidonic acid (AA): EPA ratio in plasma phospholipids from young and older subjects. Mean (sem) ratios of AA to EPA in plasma phospholipids before (grey bars) and after (white bars) supplementation with placebo (0 g EPA) or low (1·35 g/d), moderate (2·7 g/d) or high (4·05 g/d) doses of an EPA-rich oil for 12 weeks in the young (upper panel) and older (lower panel) subjects. n 24, 23, 23 and 23 for the young subjects in the placebo, low-EPA, moderate-EPA and high-EPA groups, respectively. n 16, 16, 15 and 15 for the older subjects in the placebo, low-EPA, moderate-EPA and high-EPA groups, respectively. At baseline there was a significant effect of age (P < 0·001) but not of treatment group (i.e. EPA dose) and no age × treatment group interaction. At baseline the ratio was significantly higher in the young than in the older subjects (P < 0·05). Two-factor ANOVA showed a significant effect of treatment group (P < 0·001) but not of age and no age × treatment group interaction for the change in the ratio of AA to EPA. *Significantly different from baseline, P < 0·001 (paired Student's t test). Figure taken from( Reference Rees, Miles and Banerjee 11 ), with permission.
Case study: ageing alters the immune response to probiotics
Influenza is a major cause of death in older people and while vaccination offers a prophylactic solution for preventing infection and associated complications, immunosenescence significantly impairs vaccine efficacy( Reference Haq and McElhaney 15 ). Potential adjuvants and dietary strategies to improve the immune response to influenza vaccines are therefore of interest, particularly in older people. Emerging evidence suggests that the resident gut microbiota plays an influential role in shaping antiviral defences and modulating the outcome of viral infections through inflammasome-mediated cytokine release( Reference Pang and Iwasaki 16 ). Antibiotic-treated mice have reduced levels of IL-1β secretion in the lung during influenza infection, supporting the suggestion that gut-resident bacteria are involved in regulating cytokine production( Reference Pang and Iwasaki 16 ). It has been speculated that gut microbes release low levels of pattern recognition receptor ligands, which provide signals for inflammasome-mediated cytokine release (for example, in the lung during influenza infection). These in turn regulate the activity of respiratory dendritic cells during activation of adaptive immunity against the virus( Reference Pang and Iwasaki 16 ), and together, this forms the basis for the hypothesis that pre- and probiotics may modulate responses to infection or vaccination.
Trials investigating the use of probiotics in prevention of common respiratory illnesses have produced mixed results( Reference Pang and Iwasaki 17 ), although a recent systematic review concluded that they significantly reduce episodes of acute upper respiratory tract infection and antibiotic usage in infants and young to middle-aged adults( Reference Hao, Lu and Dong 18 ). Response to vaccination is increasingly being used as a surrogate for the response to infection( Reference MacDonald and Bell 19 ). The majority of studies investigating the impact of probiotics on responses to vaccination have been conducted in healthy adults, and some show borderline effects of probiotics on serum or salivary IgA titres, although the clinical relevance is not clear( Reference Maidens, Childs and Przemska 20 ). Studies in infants and in elderly subjects, particularly those examining the response to influenza vaccination, are very limited, as are studies on the effects of prebiotics on immune function( Reference Lomax and Calder 21 ) and vaccination( Reference Maidens, Childs and Przemska 20 ). Since ageing is associated with reduced biodiversity and compromised stability of the gut microbiota( Reference Biagi, Candela and Turroni 22 ), as well as immunosenescence, older individuals may derive particular benefit from intervention with pre- and/or probiotics.
Previous studies investigating the effects of probiotics on the response to vaccination have mainly focused on antibody production. While some studies have reported a modest effect of probiotics on the antibody response to vaccination in adults, trials in older subjects are largely inconsistent and data are limited( Reference Maidens, Childs and Przemska 20 ). In a recent study (the PRIMAGE (Probiotics, Immunity and Ageing) trial), we demonstrated that while there was marked impairment of the antibody response to influenza vaccination in older subjects, intervention with a novel synbiotic, Bifidobacterium longum bv. infantis CCUG 52486 combined with gluco-oligosaccharide (B. longum + Gl-OS) failed to reverse this impairment( Reference Przemska-Kosicka, Childs and Enani 23 ). Although there is general consensus that ageing impairs the response to influenza vaccination( Reference Derhovanessian and Pawelec 24 ), there are very few robust studies specifically comparing responses of young and older subjects, and there are no other studies directly comparing the efficacy of pre- and probiotics on the immune response of young and older subjects to vaccination. In the PRIMAGE trial, the response of the young and older subjects to the intervention differed to some degree. In older subjects consuming the synbiotic, there was a trend for reduced seroconversion to the Brisbane subunit of the vaccine, whereas in the young subjects, there were trends for enhanced production of vaccine-specific IgM and, to some extent, IgG( Reference Przemska-Kosicka, Childs and Enani 23 ). Increased production of vaccine-specific IgM and IgG following intervention with probiotics has been reported in several other studies( Reference Isolauri, Joensuu and Suomalainen 25 – Reference Rizzardini, Eskesen and Calder 29 ). The possibility that there is a differential immune response to probiotics in young v. older subjects has also been demonstrated in in vitro studies. You and Yaqoob( Reference You and Yaqoob 30 ) demonstrated that peripheral blood mononuclear cells from older subjects (60–85 years) were more responsive to the immunoregulatory effects (IL-10 induction) of two strains of bifidobacteria than young subjects (18–30 years), whereas peripheral blood mononuclear cells from young subjects were more responsive to the immunostimulatory effects (IL-12 induction) of two strains of lactobacilli. Further studies demonstrated that probiotics increased the responsiveness of dendritic cells in older subjects to a greater degree than young subjects, but this was not sufficient to overcome the impact of immunosenescence in a mixed leucocyte reaction( Reference You, Dong and Mann 31 ). The choice of probiotic, particularly for older individuals, is a matter of debate and it has been suggested that ‘successfully aged’ donors of probiotic strains might survive better in an older host and achieve a more suitable equilibrium with the resident microbiota( Reference Dominguez-Bello, Blaser and Ley 32 ). B. longum bv. infantis CCUG 52486 is an example of a strain present in particularly healthy subjects aged >90 years( Reference Silvi, Verdenelli and Orpianesi 33 ). It has subsequently been demonstrated to have particular ecological fitness and anti-pathogenic effects in vitro ( Reference Likotrafiti, Manderson and Fava 34 ) and, as described earlier, immunomodulatory effects, which are strongly influenced by the age of the host( Reference You and Yaqoob 30 , Reference You, Dong and Mann 31 ).
Further immunological characterisation in the PRIMAGE trial revealed that B and T cell profiles differed markedly between young and older subjects, and that vaccination increased numbers of specific memory subsets in young subjects, but failed to do so in older subjects (S Enani, A Przemska-Kosicka, CE Childs et al., unpublished results). A key finding was the observation that there was a greater degree of immunosenescence at baseline in older subjects randomised to the synbiotic, which occurred entirely by chance, but could explain the particularly poor response of these subjects to the vaccination( Reference Przemska-Kosicka, Childs and Enani 23 ). T cells are particularly susceptible to senescence, resulting in loss of CD28; repeated antigenic exposure, for example to cytomegalovirus, is suggested to play a major role in this( Reference Vallejo 35 , Reference Derhovanessian, Maier and Hahnel 36 ). Latent infection with cytomegalovirus has been demonstrated to result in a poor response to infection and vaccination( Reference Derhovanessian, Maier and Hahnel 36 ). In the PRIMAGE trial, not only did older subjects randomised to the synbiotic have a significantly higher number of senescent (CD28−CD57+) helper T cells at baseline compared with those randomised to the placebo, they also had significantly higher plasma levels of anti-cytomegalovirus IgG and a greater tendency for cytomegalovirus seropositivity. Moreover, higher numbers of CD28−CD57+ helper T cells were associated with failure to seroconvert to the Brisbane subunit of the vaccine, strongly suggesting that the subjects randomised to the synbiotic were already at a significant disadvantage in terms of likely ability to respond to the vaccine compared with those randomised to the placebo and that differences in immunosenescence between the randomised groups at baseline may have influenced the outcome of the intervention (Fig. 2). Future work therefore needs to consider prospective randomisation of subjects based on robust immunological markers; this is challenging given the wide range of potential markers and uncertainty regarding their predictive value.
Fig. 2. Baseline levels of anti-cytomegalovirus (CMV) IgG differ in older subjects randomised to B. longum + Gl-OS and placebo. Data are anti-CMV IgG (AU/ml)(2sem) for n 45 young and n 45 older subjects randomised to B. longum + Gl-OS or placebo. Data were analysed using Student's independent t tests for differences between young and older subjects. * Denotes significant difference between treatment groups within age cohort (P < 0·05). The difference in CMV status between the cohorts may have influenced the outcome of the subsequent intervention. Figure taken from( Reference Przemska-Kosicka, Childs and Enani 23 ), published by Springer.
Ageing alters metabolic regulation of T cells
Over the past few decades, our understanding of T cell activation has extended to exploration of integration between canonical T cell signalling pathways and metabolic signalling programmes( Reference Pollizzi and Powell 37 ), and it has been proposed that immunosenescence is linked to alterations or defects in that integration( Reference Lanna, Henson and Escors 38 ). Although several transcription factors and serine/threonine kinases are central to the integration of immunological and metabolic pathways( Reference Pollizzi and Powell 37 ), the energy sensor, AMPK, is of particular interest in the context of ageing. AMPK is a central regulator of metabolic stress and is activated by an increase in the AMP:ATP ratio, as well as by T cell receptor engagement. In fact, it has been suggested that AMPK activation in response to antigen anticipates ATP depletion even in the presence of adequate nutrients( Reference Przemska-Kosicka, Childs and Enani 23 ). In AMPK-deficient T cells, metabolic stress due to glucose deprivation induces enhanced cell death. Senescent T cells demonstrate spontaneous phosphorylation- and therefore activation- of AMP( Reference Lanna, Henson and Escors 38 ). However, contrary to expectation, senescent cells did not contains low levels of ATP( Reference Meydani, Endres and Woods 8 , Reference Lanna, Henson and Escors 38 ). Instead, it is suggested that AMPK activation triggered by glucose deprivation results in activation of the p38 pathway, which leads to DNA damage and immunosenescence( Reference Lanna, Henson and Escors 38 ). Conversely, AMPK silencing restores proliferation( Reference Pollizzi and Powell 37 ). This is a previously unrecognised mode of activation for p38 in T cells and the first demonstration of a pathway which integrates low nutrient sensing with DNA damage and senescence. The observation that nutrient deprivation triggers pathways linked with immunosenescence seems to contradict the widely held belief that energy restriction enhances lifespan, but data on energy restriction and infections is not clear cut and this remains an important area for future work.
Transcription factors and signalling proteins involved in regulatory and metabolic pathways represent novel targets for immune modulation. Indeed, it has been suggested that targeting AMPK and mechanistic target of rapamycin may be a strategy for suppressing immune responses and treating inflammatory diseases( Reference Pollizzi and Powell 37 ). However, the suggestion that this may allow more selective regulation of immune responses than ubiquitous signalling pathways should be interpreted with caution as there is no clear reason to believe that this is the case.
Conclusion
Ageing alters the immune response to dietary interventions; specific examples described in this review demonstrate that young and older subjects respond differently to interventions involving dietary fatty acids and probiotics. It is critical that baseline differences in immunosenescence in dietary studies involving older subjects are accounted for as they can directly influence the outcome of the intervention. Ageing also alters metabolic regulation of T cells; elucidation of alterations in metabolic regulation in ageing T cells may prove to be important in understanding the impact of diet on immune function in older people.
Acknowledgements
The author acknowledges Agnieszka Przemska-Kosicka, Caroline Childs, Sumia Enani, Catherine Maidens, Honglin Dong, Jialou You and Iman Bin Dayel, who contributed to some of the work described in this review.
Financial Support
Some of the work described in this review was supported by a grant (BB/H00470X/1) from the Biotechnology and Biological Sciences Research Council Diet and Health Research Industry Club (BBSRC-DRINC).
Conflicts of Interest
None.
Authorship
P. Y. is the sole author of this paper.
