Graham Pawelec ¹, Ed Remarque ², Yvonne Barnett ³, Rafael Solana ⁴

1 University of Tübingen, Tübingen, FRG ², University of Leiden, Leiden, Holland ³, University of Ulster, Coleraine, Northern Ireland ⁴, University of Córdoba, Córdoba, Spain

Received 12/29/97 Accepted 1/5/97

11. POSSIBLE APPROACHES TO INTERVENTIONIST MANIPULATIONS

11.1 Vitamins and minerals

Immunogerontological parameters may be affected by many outside influences rather than aging per se, particularly if donors are not selected for perfect health using strict clinical criteria. Some of these factors may be subject to manipulation. For example, much attention has been paid recently to the effects of exercise on immune function in the elderly, although details of the mechanisms involved in any observed improvement in immune function are completely unknown (317). It is difficult to dissect the effects of the many different interacting and confounding factors, including health status, nutritional status, psychological status etc., which overlay a background genetic influence. However, some factors are easily manipulated and have encouraged interventionist approaches simply because these are feasible. Even in carefully selected donors, for example, nutritional status may play a significant role in exacerbating immunological differences (318). Dietary supplementation studies are relatively easy to perform and several including immunological data have been reported. For example, vitamin C supplementation was reported to enhance the mitogenic responses of lymphocytes from elderly people (319) and Vitamin E supplementation has been known for some time to enhance lymphocyte proliferative responses and IL 2 production in vitro and DTH in vivo in elderly people (320). This correlated with a decline in prostaglandin E₂ (PGE₂) synthesis, which is known to increase with aging. The immunosuppressive effects of PGE₂ are predominantly mediated by increasing cAMP levels; therefore agents which decrease cAMP levels might also be expected to enhance lymphocyte responses. These may include insulin and chromium (321). Unlike a number of other proposed factors, the benefits of vitamin E supplementation have been subjected to vigorous scientific testing. A recent substantial study concluded that vitamin E supplementation for 4 months improved a number of clinically-relevant indices of cell-mediated immunity in the healthy elderly, including DTH and antibody responses to hepatitis B and tetanus vaccines but without increasing autoantibody titers (322). In mice, some data indicate that certain senescence-associated alterations which can be measured on T cells are prevented by in vivo treatment of mice with the anti-oxidant vitamin E (323). Thus, vitamin E supplementation prevented the observed age-related decline of anion transport by lymphocytes in mice and inhibited the generation of the "senescent cell antigen" (SCA) from the anion transporter "band 3". Prevention or delaying of band 3 aging and subsequent generation of SCA has the consequence that the lymphocytes are not eliminated from the system via SCA-mediated interactions with the reticuloendothelial system. By analogy with the mouse, vitamin E supplementation might be expected to have a greater impact on the old than the young; thus, in a mouse influenza model, high-dose vitamin E significantly enhanced lung virus clearance in old mice, with little effect on young mice. However, it is not known whether these effects of vitamin E are attributable to its anti-oxidant or some other function.

Another supplement commonly believed to enhance immunity is often taken along with other factors, namely ß-carotene. However, results of two careful recent studies do not support immunoenhancing effects of either short (3 weeks) or long-term (up to 12 years) ß-carotene supplementation in randomized double-blind placebo-controlled longitudinal comparisons in human. There were no pre- to post-intervention changes measured in DTH, lymphocyte proliferation, IL 2 production, PGE2-production or lymphocyte subset composition (324).

Not only vitamin- but also trace element-deficiencies in the elderly may contribute to immunodeficiency. For example, levels of selenium decrease in rat lymphoid tissues with increasing age (325) and selenium supplementation has been reported to reverse low levels of proliferation and CTL generation in aged mice (326). In elderly people, selenium supplementation was reported to enhance lymphocyte proliferative responses to pokeweed mitogen (PWM) (327). Moreover, it is well established that the availability of certain essential micronutrients decreases with age; one very important such factor may be zinc (328,329). Zinc is necessary for the function of many hormones and enzymes, including those known to affect immune responses (eg. testosterone, ref. 330)). This has practical implications, because even if absorption is compromized in the elderly, sufficient supplementation might still overcome the problem (331). In rats, there is also a decrease in serum zinc, and the levels found in the thymus are lower compared to young animals as well (332). These observations on zinc levels may not be limited to rodents, because in humans, zinc supplementation studies have indeed suggested improvement of some parameters of immune function (333,334). One reason for this may be the zinc-enhancement of otherwise age-associated lowered levels of interferon-alpha production in the aged (335). Experimental zinc depletion and repletion of healthy humans revealed that secretion of the Th1-type cytokines IFN-g and IL 2 was decreased during zinc deficiency, whereas Th2-type cytokines (IL 4, IL 6, IL 10) were not affected (336). In animals, Mocchegiani et al . (337) confirmed that oral zinc supplementation resulted in a recovery of thymic function (and also demonstrated its influence on extrathymic T cell differentiation pathways (338)) and showed that thymic regrowth was associated with a partial reconstitution of peripheral immune function (as measured by mitogen stimulated proliferation and NK activity). Moreover, low levels of activity of the zinc-dependent hormone thymulin were not dependent on the state of the thymus itself, but on decreased zinc saturation of the synthesized hormone. The authors concluded that age-dependent thymic involution and compromized thymic hormone function were not preprogrammed but were caused by the decreased availability of zinc. In this context it is interesting to note the claim that the beneficial effects of melatonin supplementation or pineal grafting are associated with increased plasma zinc levels in old mice in the absence of exogenous zinc supplementation (339) (although melatonin may have direct effects on lymphocytes, which express mRNA for melatonin receptors (340)). It is argued by Fabris et al . that the common pathway of several life-extending endocrinological manipulations is in fact via zinc bioavailability (341). Thus, even such a well-established concept as the inevitability of the age-associated process of thymic involution and the resulting perturbation of T cell generation may not be immutable. However, the beneficial effects of zinc supplementation are controversial and others have found no benefit of zinc replacement even in elderly populations confirmed to show serum zinc deficiency (129,342). In some studies, even inhibitory effects of zinc supplementation have been reported (343). It has also been found that the degree of decrease of lymphoproliferative responses observed in the elderly does not correlate with decreased levels of plasma zinc (or vitamin E, retinol or ß-carotene) (344). The situation is therefore not yet clarified.

11.2 Hormones

Aging of the immune system will most likely affect other organ systems and eventually impact upon the lifespan of the individual (345). Manipulations said to increase lifespan of mice by injecting melatonin or by transplanting pineal glands are accompanied by maintenance of T cell immune responsiveness (as measured by DTH) and prevention of thymic involution (346). Melatonin may have direct effects on CD4 but not CD8 cells because of a direct effect on gene regulation via binding the putative nuclear melatonin receptor (347). In vivo treatment with melatonin is reported not to reconstitute age-associated impairment of NK activity or lymphoproliferative responses in mice (348). Consistent with these results in vitro supplementation also failed to reconstitute proliferation or IL 2 production in old rat cells (349). It is therefore unclear whether and how melatonin extends lifespan.

Age-associated changes in secretion of growth hormone (GH) and related hormones, releasing factors and binding factors may contribute to immunosenescence. Thus, GH substitution may reverse some immune defects in humans and primates, as reviewed in (350). Administration of low-dose GH to elderly adults for 6 months resulted in an increase in IGF-1 levels (which are reduced in aging) and an improvement in some physiological parameters, such as muscle strength. Immunological parameters were not reported. The same considerations may apply to other factors, eg. the native steroid dehydroepiandrosterone (DHEA), which, like most steroids, has immunomodulating activity. Whereas levels of cortisol increase with age in both men and women (351), in general the levels of DHEA decline with age (352). However, longitudinal studies suggest a great deal of inter-individual variation, and can even show age-associated increases in DHEA levels in a sizeable proportion of the population (353). It has been suggested that decreases of DHEA could be associated in some way with immunosenescence, because treatment of old mice with DHEA augments the otherwise decreased capacity of T cells to produce IL 2 and IFN-g . It also decreases the spontaneous secretion of IL 6 (354) and IL 10 observed in old mice and reverses their hypersensitivity to endotoxin-stimulated release of both IL 6 and IL 10 (355). Analogously, it prevents the retrovirus-induced increased IL 6 and IL 10 secretion seen in old mice, prevents decreases of IL 2 and IFN-g production and enhances their T and B cell proliferative responses (356). DHEA reverses the senescent phenotype (as defined by the pattern of cytokine secretion) in mice and enhances the effects of vaccination of old mice to hepatitis B (357). In man, DHEA also enhances IL 2 production (358) and supplementation trials have been carried out (eg. see ref. 359), but there are few immunological studies. Khorram et al . (360) found that DHEA administration resulted in a significant augmentation of serum IGF-1 and decreased IGFBP-1, which may contribute to immune enhancement. They also found an increase of monocytes during treatment, as well as increases in mitogenic responses of both T cells and B cells. The numbers and activity of NK cells were also enhanced.

In contrast to DHEA, dihydrotestosterone (DHT) downregulates IL 4, IL 5 and IFN-g production but does not affect IL 2 (361). DHT levels also decrease with age, and a recent cross-sectional study found that bioavailable testesterone correlated best with significantly age-associated cognitive and physical parameters (362). However, a recent longitudinal study found no correlation between entry-point testosterone levels and death rates over the 15 year follow-up period in 77 men (363). Together, DHEA and DHT supplementation alter the cytokine profile of old mice such that it again resembles that of young mice; such an activity could be measured in vivo as well as in vitro (361). Exogenous hormone supplementation might correct age-associated defects insofar as these are dependent upon cytokine profiles. This has been tested in a mouse model of influenza virus vaccination. Danenberg et al . (364) reported that DHEA supplementation resulted in a reversal of the age-associated decline in immune responsiveness in mice, reflected by increased humoral responses in treated mice and increased resistance to challenge with live virus. In another study, Ravaglia et al . (365) reported on the relationship between DHEA levels and health in free-living people over the age of 90. They found five-fold lower levels of DHEA in both males and females aged 90-106, compared to young controls. Thus, even "successfully" aged persons had greatly reduced levels, leading to the question of whether this matters. Ravaglia et al . demonstrated that it can matter, because within the old male group, the level of DHEA correlated with their health, as measured on the ADL scale. On the other hand, DHEA levels are clearly reduced in the aged although the degree of reduction fails to correlate with health status as assessed by the strict SENIEUR protocol (366). A supplementation trial to assess the effects of DHEA on responses to tetanus and influenza vaccination in man did not yield as dramatic results as seen in mice (367): there was a trend toward increased antibody titers to influenza but not tetanus, and even this failed to reach significance (367). Danenberg et al . even reported a decrease in attainment of protective antibody titer in elderly volunteers given DHEA in a prospective randomized placebo-controlled double-blind study of the effects of DHEA on influenza vaccination (368). Thus, the decreased ´flu response in elderly humans, unlike that in mice, could not be reversed by DHEA, and a higher baseline level of DHEA was also not found to be predictive of better ´flu vaccination outcome (368).

Given the known or suspected interactions between the endocrinological and immunological systems and the well-established impact of sex and other hormones on immune responses, it is perhaps surprising that few studies have addressed the question not only of gender differences but also the effects of pregnancies on immunosenescence. Some investigators have begun to approach this by surveying leukocyte subsets in mice of varied gynecological histories. One such study concluded that both gender and pregnancies affect the age-related distribution of lymphoid and macrophage populations in the spleens of C57Bl/6 mice, for example (369).

11.3 Anti-oxidants

Thus far, there are few studies with SENIEUR aged human T cells aimed at clarifying the mechanisms of reduced function. It is important that donors selected according to standardized health criteria should be entered into immunogerontological studies, in an attempt to exclude alterations in immune parameters caused by underlying pathology rather than aging per se. The SENIEUR protocol is a strict donor selection procedure whereby only a small fraction of the elderly are classified as perfectly healthy. It has been established that the reduced proliferative responses stimulated by phytohaemagglutinin (PHA) are still observed even in perfectly healthy SENIEUR donors, although it has been reported that the response to immobilized CD3 mAb may not be reduced (370). The reason for this discrepancy is unclear. However, direct biochemical signalling with phorbol esther and ionomycin may also result in reduced responses in elderly donors, suggesting that even if signal transduction and second messenger production proceed normally in aged T cells, downstream events may still be dysregulated (371). In mice, T cells from old animals stimulated by CD3 + CD4 ligation mobilize less calcium ions than T cells from young animals. They also perform less tyrosine phosphorylation of phospholipase C gamma 1 and other phosphoproteins. Moreover, these events appear to be sensitive to anti-oxidant levels, such that Grossmann et al . suggested that one reason for decreased PLC gamma-1-dependent signalling was the decrease in antioxidant levels in old cells in rats (372). The general importance of anti-oxidant systems is illustrated by the report that although resting young and old rat splenocytes did not differ in their content of the important anti-oxidant reduced glutathione, in proliferating cells from old animals, the expected increase in glutathione was delayed. This correlated with an increasing number of cells showing evidence of mitochondrial dysfunction in terms of depolarized membrane potential and decreased mitochondrial mass. Impairment was completely prevented by addition of extra glutathione to the medium (373). An in vivo relevance for these findings is suggested by the fact that reduced, but not oxidized, glutathione levels in the plasma are decreased in elderly compared to young donors (374). In mice, however, the age-associated reduction of GSH levels did not correlate with increased susceptibility of lymphocytes to oxidative damage. This was found to be due to a predominance of memory cells in the aged animals and the fact that memory cells, despite lower GSH, were more resistant (in young and old mice) to oxidative damage (375). Although it has proven difficult, as exemplifed above, to show age-related decreases in enzymatic anti-oxidant defences, most studies have concentrated on assaying levels of known anti-oxidants, rather than on measuring anti-oxidant function. A recent study investigating the latter in human plasma did provide evidence consistent with decrease with age, but only in males and then only over the age 74 (376). Another recent study confirmed decreased antioxidant activity in aged rat plasma but failed to show the same in man (377). Despite the above data, interventionist approaches with antioxidants remain attractive because of their cheapness and easiness.

11.4 Caloric restriction

Many studies have examined some aspects of the biochemical changes associated with T cell activation and their alteration with aging. In some animal models other than mouse, results comparable to those in humans have been obtained; thus, in rhesus monkeys, CD4⁺ cells from old donors respond less well to CD3-stimulation, and this is partly associated with a decreased frequency of responding cells and is reflected in lower calcium-mobilization in old cells (378). The same investigators also reported that one of the major strategies to prolong rodent life, which is associated with improved immune function, namely caloric restriction (CR), did not alter the depressed calcium-mobilization rates in old monkeys. It did, however, retard the marked age-associated decline of DHEA levels in rhesus monkeys. It also ameliorated the levels of lipid peroxidation of lymphocytes, supporting the view that CR effects are at least partly mediated through reduced free-radical damage (379). Moreover, CR reduced levels of a marker of oxidative DNA damage in old rats (380). Limited biochemical studies employing SENIEUR donors have indicated that membrane lipid alterations in the elderly may be important for altered immunological function (381). Rather than the membrane lipid constitution per se that was different between young and old, it was the changes observed upon blastic transformation of stimulated lymphocytes which correlated with decreased proliferative function. CR has multiple effects which are only now being elucidated, eg. some data show that old CR mice retain better GH receptor function than old ad libitum-fed mice (382). A major mechanism may be via the lowering of nutritionally-driven insulin exposure which lowers overall growth factor exposure (383). One relatively clear finding in monkeys is that CR reduced body temperature, as a result of decreasing energy expenditure, consistent with the "rate of living" theory of aging (384). Further evidence that any longevity enhancement by CR in rhesus monkeys is not correlated with improvement of immune responses stems from the study of Roecker et al . (385). They showed that mitogen-stimulated proliferation, NK activity, and antibody production were all reduced in CR monkeys compared to controls, with no effects discernible on cell number or surface markers. On the other hand, in a long-lived rat strain, CR clearly resulted in improved T cell proliferation after mitogenic stimulation. This, and the cytokine (higher IL 2, lower IL 6 and TNF-a ) and surface marker (higher OX-22) profiles of the T cells suggested that the CR animals had a higher fraction of "naive" cells compared to the controls with more "memory" cells (386). However, in two other rat strains, Konno (387) had shown accelerated thymic involution in CR animals, and either a slight decrease or no change in immune function. This suggests that the genetic background has a major impact on the effects of CR. In mice as well, CR results in decreases in the otherwise age-associated increased constitutive serum levels of IL 6 and TNF-a (388). Of course, CR may have many physiological effects which are nothing to do with immunity, for example, CR slows down the age-related increase in mutations (389), perhaps via effects on DNA polymerase-alpha (390).

11.5 Mutations and DNA repair

Mechanisms for recognition of DNA damage and its repair are important in maintaining cellular integrity. If these are reduced during aging, this would also contribute to failing function. Manipulations to enhance repair might therefore offer some benefits. Boerrigter et al . (391) found that the rate of disappearance of a particular kind of chemically-induced DNA damage was age-dependent in mice, and also varied between strains, with longer-lived strains having better damage repair capacity than shorter-lived strains of the same age. Cortopassi & Wang summarized various publications recently to survey agreement on rates of DNA repair in different species and the correlation between repair and maximal lifespan (392). They concluded that large differences in DNA repair capacity were found in different species and that the correlation between maximal lifespan and repair was indeed good, although not excellent. Moreover, DNA repair capacity within a particular species may correlate with age of the individual. Thus, there is an age-related decrease in post-UV-irradiation DNA repair capacity in cultured skin fibroblasts from normal human donors, estimated at -0.6% per year up to the age of > 90 years (393). Moreover, the same group estimated a corresponding increase in mutability of DNA in B cell lines from these donors of + 0.6% per year (393), suggesting that DNA repair decrease with age and this correlates with increased mutability. An underlying mechanism responsible for changes in DNA repair with aging may be decreased expression and function of DNA topoisomerase I, an enzyme that alters the superhelicity of DNA (394).

In T cells, studies of mutations revealed that background mutant frequency (MF) at the hprt locus increases with age up to advanced middle age (316). In parallel with the age-related increase in MF, there was an age-related decrease in DNA repair capacity against hydrogen peroxide-induced DNA damage (395). However, MF induced by X-rays was not increased with age, leading to the conclusion that in human T cells aging does not affect the cytotoxic or mutagenic effects of X-rays (396). If confirmed, this is significant, because it suggests that decreased effíciency of DNA repair may not be the sole reason for age-associated MF increases. However, when older aged individuals were examined, an increase in basal levels of DNA damage in lymphocytes from donors 75-80 years old compared to those 35-39 was no longer found. There was also no significant difference between frequency of mutation at the hprt locus in the young and more aged populations, nor was there any difference in DNA repair capacity after hydrogen peroxide-induced DNA damage (397). These findings, together with the increased levels of anti-oxidants glutathione peroxidase, catalase and ceruloplasmin in the elderly, suggest that those individuals with best retention of DNA repair mechanisms and anti-oxidant defences belong to the group with extended longevity.

It has been argued that one of the unifying factors shared by life-extension manipulations and mutations is the adjustment of the organism to low levels

of stress. It is hypothesized that activation of stress-protective mechanisms early in life may result in their better function later in life and that stress-resistance is a determining factor of longevity (315). In man, senescent T cells do show a reduced stress response as reflected by decreased production of hsp70 after heat shock, associated with decrease in binding of nuclear extracts to the consensus heat shock element. The progressive decline in hsp70 response with increasing age of T cells in culture was found to correlate with the percent of proliferative lifespan already completed (398). An age-dependent decrease of heat shock factor-1 (HSF-1) binding in isolated human lymphocytes ex vivo, as well as gradual loss of heat-iducible HSF-1 in cultured T cells as they age has also been observed by Jurivich et al . (399). A member of the hsp70 family, mortalin, has been proposed as a marker for cells committed to apoptosis. Some recent data implicate hsp70 as a protector against apoptosis (400), others show that overexpression of transfected hsp70 enhances AICD in T cells (401). The expression not only of the hsp 70 family, but also hsp90 family stress proteins, has been reported to be reduced in aged T cells examined directly ex vivo, suggesting that results with cultured cells are relevant to the in vivo situation (109). Interventions which would enhance the stress response may therefore also be directly relevant to delaying immunosenescence (315). Even low-level irradiation may fall into this category. For example, Hyun et al . (402) found that low level irradiation, followed by a period of adaptation and then high level irradiation protected mouse spleen cells against the latter (as measured in mitogen-induced proliferation assays and in irradiation-induced apoptosis assays).