Delay of T cell senescence by caloric restriction
in aged long-lived nonhuman primates
Ilhem Messaoudi*, Jessica Warner*, Miranda Fischer*, Buyng Park†, Brenna Hill‡, Julie Mattison§, Mark A. Lane§,
George S. Roth¶, Donald K. Ingram§?, Louis J. Picker*, Daniel C. Douek‡, Motomi Mori†, and Janko Nikolich-Zˇugich*,**
*Vaccine and Gene Therapy Institute and Oregon National Primate Research Center, Oregon Health and Science University, Beaverton, OR 97006;
†Biostatistics Shared Resource, Oregon Cancer Institute, Oregon Health and Science University, Portland, OR 97201;‡Vaccine Research Center, National
Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD 20892;§Laboratory of Experimental Gerontology, National
Institute on Aging, National Institutes of Health, Baltimore, MD 21224; and¶GeroScience, Inc., Pylesville, MD 21132
Edited by Michael Sela, Weizmann Institute of Science, Rehovot, Israel, and approved October 18, 2006 (received for review August 2, 2006)
Caloric restriction (CR) has long been known to increase median
and maximal lifespans and to decreases mortality and morbidity in
short-lived animal models, likely by altering fundamental biolog-
reported to delay the aging of the immune system (immune
senescence), which is believed to be largely responsible for a
dramatic increase in age-related susceptibility to infectious dis-
eases. However, it is unclear whether CR can exert similar effects
in long-lived organisms. Previous studies involving 2- to 4-year CR
treatment of long-lived primates failed to find a CR effect or
reported effects on the immune system opposite to those seen in
CR-treated rodents. Here we show that long-term CR delays the
adverse effects of aging on nonhuman primate T cells. CR effected
a marked improvement in the maintenance and/or production of
naı ¨ve T cells and the consequent preservation of T cell receptor
repertoire diversity. Furthermore, CR also improved T cell function
and reduced production of inflammatory cytokines by memory T
cells. Our results provide evidence that CR can delay immune
senescence in nonhuman primates, potentially contributing to an
extended lifespan by reducing susceptibility to infectious disease.
aging ? immunity ? T cell subsets ? nutrition
search is under way for interventions to reduce or delay the
occurrence, and lessen the severity, of age-related disease. Among
animal models of aging, the most promising intervention thus far is
caloric restriction (CR), which has long been known to extend both
lifespan and retards the onset of age-related dysfunction in other
short-lived organisms, including yeast, worms, flies, zebra fish, and
spiders (as reviewed in refs. 2 and 3).
Unlike many interventions that increase average life span by
reducing the negative effects of a specific age-related disease or
of individual organ-specific aspects of aging, CR is believed to
alter fundamental biological processes that regulate aging and
longevity (4). However, its mechanisms remain unclear and are
under intense study (5, 6). There is evidence in short-lived
organisms for the modification of numerous processes, including
energy metabolism, mitochondrial function, oxidative damage,
and neuroendocrine homeostasis (2), giving rise to diverse
hypotheses on how CR may extend lifespan by a variety of
mechanisms (2, 3, 6, 7). Aging also results in profound changes
in the immune system (immune senescence) (8, 9), resulting in
a dysregulation of both the innate and adaptive arms of the
immune response, which normally ensure optimal protection
against pathogens (10). Consequently, infectious diseases are
among the leading causes of morbidity and mortality in the
elderly (11, 12). The fact that CR can partially retard or reverse
some of the changes associated with immune senescence in
rodents provides the basis for the hypothesis that improvement
iven the demographic trend in industrialized societies toward
a marked increase in the elderly population, an intensive
of the immune function by CR contributes to extended lifespan
by affording higher resistance to pathogens (8, 9).
Within the adaptive immune system, T cells are probably the
component most affected by immune senescence. Diminished T
cell function (13); decreased production of new, naı ¨ve T cells by
the thymus (14); increased homeostatic turnover of naı ¨ve T cells
(15); and lifelong consumption of naı ¨ve cells due to pathogenic
assaults all contribute to a decline in T cell immunity. The loss
repertoire diversity, which is further compounded by the ap-
pearance of T cell clonal expansions (16–19) that additionally
constrict normal lymphocyte repertoire and can potentially
interfere with protective immunity (20). Rodent studies suggest
that CR can reduce the rate of immune senescence, preserve
naı ¨ve T cells into late age (21), and reduce the production of
proinflammatory cytokines such as IL-6 and TNF-? (22).
Given the robust results reported in short-lived species, long-
term studies have been initiated in the Old World nonhuman
impact of CR on mortality, morbidity, and age-related changes
in the function of various cells, tissues, and organ systems (23,
24). Shortly after the initiation of CR in RMs in the National
Institute on Aging (NIA) study (summarized in ref. 25), it was
reported that CR-treated RMs displayed physiological changes
similar to those observed in CR rodents, specifically lower body
weight, fat mass, temperature, and circulating plasma insulin
levels. CR was also found to maintain dehydroepiandrosterone
sulfate (DHEAS) and melatonin, two of the biomarkers of
human aging, at youthful levels in RMs (25, 26). Moreover,
preliminary studies in nonhuman primates suggest that mortality
due to diabetes, cardiovascular disease, and cancer might be
lower in CR animals than in controls (23). Similar results were
reported in a cohort of CR monkeys at the University of
Wisconsin (24). By contrast, studies measuring immunological
parameters in unseparated peripheral blood mononuclear cells
(PBMC) of adult and aged monkeys on CR for 2–4 years (27–30)
Author contributions: I.M., M.A.L., G.S.R., D.K.I., L.J.P., and J.N.-Zˇ. designed research; I.M.,
analyzed data; and I.M. and J.N.-Zˇ. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS direct submission.
Abbreviations: CM, central memory T cells; CON, control; CR, caloric restriction; EM,
effector/effector memory T cells; N, naı ¨ve T cells; PBMC, peripheral blood mononuclear
?Present address: Nutritional Neuroscience and Aging Laboratory, Pennington Biomedical
Research Center, Louisiana State University System, 6400 Perkins Road, Baton Rouge, LA
**To whom correspondence should be addressed at: Vaccine and Gene Therapy Institute,
Oregon Health and Science University, West Campus, 505 Northwest 185th Avenue,
Beaverton, OR 97006. E-mail: email@example.com.
This article contains supporting information online at www.pnas.org/cgi/content/full/
© 2006 by The National Academy of Sciences of the USA
December 19, 2006 ?
vol. 103 ?
results compared with CR-treated rodents. Therefore, the exact
effects of long-term CR on defined cell subsets of the immune
system in aged RMs remain poorly understood. To test the
impact of CR on T cell senescence in RMs, we investigated T cell
subset distribution, TCR repertoire diversity, thymic production
male and female monkeys subjected to CR or control (CON)
treatments for 13–18 years. Given that the estimated median
lifespan of RMs is 25 years (40 years maximal), monkeys in this
study represent a human age equivalent of ?60–70 years. Our
results show remarkable conservation of naı ¨ve T cell phenotype
and function by CR in long-lived primates.
Overall Study Design and Animal Cohorts. Experimental cohorts
consisted of aged CON and CR RMs of both sexes that were
19–23 years of age during the course of sampling for this study.
The study animals are described in detail in Table 1 and in
Materials and Methods, and their diet and husbandry are de-
scribed in ref. 31. We obtained samples for analysis for all of the
available animals in the cohort at four time points over 42
months. [For a note on attrition, see supporting information (SI)
Discussion.] The time points and numbers of animals sampled
were as follows: males: 11/2002 (7 CR, 19 CON), 05/2003 (7 CR,
18 CON), 07/2004 (7 CR, 18 CON), and 05/2005 (7CR, 18CON);
females: 02/2003 (6 CR, 10 CON), 07/2004 (5 CR, 9 CON),
03/2005 (4 CR, 8 CON), and 09/2005 (4 CR, 8 CON). The main
reason for female attrition was endometriosis, a condition that
is difficult to diagnose and problematic to treat in female RMs
and that peaks between 20 and 25 years of age. Samples were
counted, frozen by controlled cryopreservation, and subse-
quently thawed for analysis, allowing simultaneous analysis of
samples collected at disparate time points. Time on the CR diet
was different for the male and female cohorts due to the
different lengths of time they were enrolled in the study (14–17
years for males and 10–13 years for females; spread due to the
longitudinal observation over the 42-month period of the study),
did not statistically influence any of the parameters analyzed
(data not shown), and was not considered separately. However,
female and male groups are shown separately to preserve power
for statistical considerations, because sex-related differences in
immunological parameters introduced additional variability. Re-
gardless of these baseline sex-related differences in CON ani-
mals, CR exhibited strong sex-independent effects on the im-
mune system of aging animals throughout these studies.
For our analysis, a group average (CON males, CR males,
and Methods, and was compared between time points. We
observed no significant differences between the time points for
the same group (data not shown), and the differences between
the CON and CR groups, as described below, were apparent at
all time points examined, leading us to conclude that, by the
earliest time point of sampling (10 years on CR in females) the
effects of long-term CR were fully established. This allowed us
to include all time points into our analysis, so as to average group
mean values between time points and then perform statistical
analysis on those values by using mixed-effect models, as de-
scribed in Materials and Methods.
Impact of CR on Blood Counts and T Cell Phenotype. In rodents, CR
was reported to cause a state of lymphopenia, and initial studies
conducted 2 years after the onset of CR in adult RMs and adult
humans suggested a slight decrease in circulating WBC (28, 32).
By contrast, complete blood cell count analysis of RMs in this
study (at six time points over 3 years) revealed that total WBC,
lymphocyte, and neutrophil counts in CR animals were similar
to those observed in CON animals, with no significant sex-
related differences (Table 1). Therefore, contrary to the results
in rodents, long-term CR did not induce a lifelong leuko- or
lymphopenic state in RMs. These results, however, do not rule
out the possibility that during the CR induction period and
thereafter, lymphopenia could occur in RMs adapting to a new
We next examined the phenotype distribution of T cell subsets
in PBMC from CON and CR animals. We classified
CD28intCD95locells as naı ¨ve (N), CD28hiCD95hicells as central
(resting) memory (CM), and CD28lo/negCD95hicells as effector/
effector memory (EM) for both CD4 and CD8 subsets, as
T cells decreases reproducibly with age in many mammals,
including RMs (33, 34, and our unpublished results), so that
naı ¨ve cells decline from ?40% in adult animals to ?20% of total
blood T cells in most aged RMs (Fig. 1). CR lessened this decline
and maintained a higher percentage of both CD8 and CD4 naı ¨ve
T cells independent of sex (Fig. 1). In the CD8 subset, this
increase occurred mostly at the expense of the EM T cells,
whereas in the CD4 subset naı ¨ve cells appeared to increase
perhaps more at the expense of CM cells. Of importance, the
above changes were absolute; total lymphocyte numbers did not
change significantly with CR (Table 1) or age (data not shown)
in the blood of RMs we have studied.
Analysis of Recent T Cell Emigrants and of T Cell Repertoire Under CR
Treatment. To independently confirm that the numbers of naı ¨ve
T cells were higher in CR-treated animals, we measured in the
PBMC of CON and CR animals the content of TCR excision
circles (TREC), as a marker of recently produced T cells (recent
thymic emigrants) or T cells that did not divide after being
produced by the thymus, as described previously (35). Both CD4
and CD8 T cells from male and female CR-treated RMs
exhibited higher TREC content per 103cells compared with
CON counterparts (borderline significance in males; significant
in females; Fig. 2A).
Increased representation of naı ¨ve T cells in peripheral blood
would be expected to increase TCR diversity in that population.
Consequently, TCR diversity was examined by measuring CDR3
length polymorphism (36) within peripheral blood T cells from
CR and CON RMs. Fig. 2B summarizes examples of polyclonal,
diverse CDR3 length profiles (typical Gaussian distribution, Fig.
oligo/monoclonal, Fig. 2B Right), which are characteristic of
expanded T cell clones. Such expanded clones frequently occur
in aged individuals (as reviewed in ref. 37), in which homeostatic
compensation and chronic or lifelong exposure to antigens likely
Table 1. Animal groups and blood cell counts
3.93 ? 0.31
4.1 ? 0.38
1.97 ? 0.15
2.13 ? 0.19
1.48 ? 0.26
1.38 ? 0.34
4.7 ? 0.44
4.24 ? 0.41
1.73 ? 0.20
2.14 ? 0.19
2.46 ? 0.36
1.73 ? 0.34
The number of white blood cells (WBC), lymphocytes (LYM), and neutro-
phils (NE) was determined by using the Beckman Coulter counter calibrated
aggregate mean values (? SEM) of all six time points. No significant sex-
related differences in blood cell counts were observed between the CON and
Messaoudi et al.
December 19, 2006 ?
vol. 103 ?
no. 51 ?
RMs, TCRV? CDR3 length polymorphism analysis revealed
that, of 22 successfully amplified V? families, only 4 had
polyclonal appearance (V?4, 9, 19, and 20), whereas 5 (V?1, 5,
10, 15, and 23) exhibited oligo- or monoclonal profiles. By
contrast, an age-matched CR animal exhibited Gaussian profiles
for at least 14 V? families (V?1–6, 9, 10, 13, 15, 16, 19, 20, and
24), with perhaps only one profile (V?14) being oligoclonal (Fig.
2C). Results of cumulative analysis revealed significantly higher
diversity in both male and female CR animals (Fig. 2D),
demonstrating that the increase in naı ¨ve phenotype T cells with
CR, seen in Fig. 1, led to increased TCR diversity in peripheral
Naı ¨ve-phenotype cells, defined as in Fig. 1, did not produce
any cytokines upon TCR stimulation in the short-term intracel-
lular cytokine staining assay (see Fig. 4A and ref. 33) and
expressed other phenotypic characteristics of naı ¨ve cells, being
CCR7?and CD62Lhi(data not shown). Overall, we conclude
that CR countered the effects of aging on T cell subsets by
increasing or maintaining the youthful levels of bona fide naı ¨ve
T cells and that this could in part be due to increased thymic
production. However, because naı ¨ve T cells possess the capacity
for self-renewal without overt changes in membrane phenotype
(39), CR may also contribute to better maintenance by slowing
the turnover of these cells.
Impact of CR on T Cell Function in Aged Monkeys. Immune senes-
cence is accompanied by a decrease in the proliferative
potential of T cells, likely due to both an accumulation of
terminally senescent cells and an age-related reduction in the
proliferative ability of naı ¨ve cells (38, 40). Although CR was
reported to ameliorate these proliferative defects in response
to mitogens in aged rodents (41), initial reports in aged
CR-treated male RMs suggested unchanged or lower respon-
siveness to different mitogens (27, 28). We used fluorescein
label (CFSE) dilution to enumerate precisely the number of
divisions in response to TCR stimulation in male CR- or
CON-treated T cells in a 96-h stimulation assay. We confirmed
Example of four-color flow cytofluorometry analysis of PBMC from a repre-
sentative pair of age-matched CON and CR animals. PBMC were stained with
antibodies against CD8?, CD4, CD28, and CD95; were gated on CD8? or CD4
populations (Left); and were analyzed for the expression of CD28 and CD95
(Right). Circles denote N, CM, and EM subsets (encompassing ?94% of all
cells), as described in the text; numbers denote the percentage of cells in each
population. (B and C) Cumulative percentages of T cell subsets were analyzed
in all CON (open histograms) and CR (filled histograms) RM males (B) and
females (C) over the four different analysis time points over a span of 42
analysis was performed using the mixed-effects model approach, and signif-
icance is indicated by two-sided P values above the histograms.
Phenotypic changes in aged RM T cells as a consequence of CR. (A)
Estimation of recent thymic emigrants, using the TREC assay, was done as
described in ref. 32. Results are shown as number of TREC per 103T cells. (B)
Representative examples of polyclonal (Left), skewed (Center), and clonal
(Right) patterns of TCR CDR3 length polymorphism. (C) Example of CDR3
length analysis for all 24 TCRV? families for representative age-matched CON
and CR animals, performed as described in Materials and Methods. (D)
histograms) and CR (filled histograms) animals. For both A and D, statistical
analysis was performed using the mixed-effects model.
TCR repertoire analysis and TREC analysis of CON and CR T cells. (A)
www.pnas.org?cgi?doi?10.1073?pnas.0606661103Messaoudi et al.
that CD28?cells respond to stimulation by vigorous prolifer-
ation, whereas CD28?cells were far less active (Fig. 3A), as
expected from previous data on human CD28?T cells (42). CR
increased the percentage of male CD28?CD4?and
CD28?CD8?cells that divided at least once at every time point
analyzed (Fig. 3B), but this trend did not reach statistical
significance. However, when the average number of divisions
completed was analyzed (Fig. 3C), results were significant for
CD28?CD8?cells and borderline for CD28?CD4?cells. This
could be due to a higher content of naı ¨ve proliferating T cells,
better proliferation of CM T cells, removal of senescent cells
by CR, or a combination thereof. Regardless, these results
demonstrate that CR leads to improved proliferation of aged
male RM T cells. Female T cells analyzed at the first time point
of this study followed the trend of the male cells. Subsequently,
nearly half of the female CR group was lost to attrition,
severely reducing the power of the study. Consequently,
additional studies are needed to extend these conclusions to
Finally, we addressed the influence of CR on the immediate
cytokine secretion capacity of T cell subsets. Immune senescence
in mammals, including primates, is accompanied by an increase
in the levels of proinflammatory cytokines, including the inter-
ferons, IL-6, and TNF-? (34, 43, 44). Increased and poorly
controlled inflammatory reactions have been associated with
numerous organ-specific pathologies, including coronary vascu-
lar disease, dementias, neurodegenerative disorders, type II
diabetes, and arthritis (45). We performed brief stimulation of
T cells from CON and CR animals by means of the TCR/CD3
complex and measured production of intracellular IFN-? and
TNF-?. After such stimulation, only CM and EM cells produced
cytokines (Fig. 4A). However, fewer T cells from CR-treated
animals produced both cytokines compared with their CON
counterparts (Fig. 4B and SI Fig. 5). The effect showed a
consistent trend at all time points examined (four time points,
represented in aggregate in SI Fig. 5), but was not significant in
females, possibly because the age-related increase in T cell
cytokine production was less pronounced in females. Results
were of borderline significance in male CD4 cells for TNF-? and
were significant in male CD4 cells for IFN-? (SI Fig. 5). In light
of phenotype data (Fig. 1), this effect could be explained by a
higher proportion of naı ¨ve cells in CR-treated RMs. Because
these cells do not produce cytokines under short-term stimula-
tion (Fig. 4A), they would simply dilute memory cells that
produce cytokines. However, CR had a direct effect on memory
cells as well. Upon subdivision of male CD8 cells into memory
subsets, it became apparent that CR mediated a highly signifi-
cant reduction in CM cells producing IFN-?, whereas the
reduction in EM cells was consistent but did not reach statistical
significance (Fig. 4B). A similar trend was observed for TNF-?
(Fig. 4B) and was also seen at the first time point with female
animals, before power was lost to attrition (data not shown).
Overall, these results strongly suggest that CR down-regulates
baseline cytokine production by CM T cells.
agonistic antibody. (A) Illustration of flow cytofluorometry profiles of stimu-
lated fluorescein (CFSE)-labeled CD4 and CD8 cells (stained and gated as
described in Fig. 1) from a CON animal. Left shows CD4/CD8 profiles; Center
shows CD28/95 profiles of gated cells from left; and Right depicts the succes-
sive fluorescein dilution in each proliferating subset at 96 h after stimulation.
histograms) males. Results are shown as mean ? SEM from a representative
experiment out of three with comparable results. (C) Average number of
divisions in the CD8?CD28?and CD4?CD28?T cell subsets of CON (open
histograms) or CR (filled histograms) male monkeys. All male animals were
analyzed, and statistical analysis was as described in Fig. 1.
Proliferation of CON- and CR-treated T cells in response to an
CR animals. (A) Illustration of flow cytofluorometry assay to measure ex vivo
cytokine production by T cells. Blood T cells were stimulated for 6 h with an
agonistic anti-TCR antibody, as described in Materials and Methods, and
Selective gating is depicted by arrows and shows production of TNF-? by the
in CD8 CM and EM T cell subsets of CON (open histograms) and CR (filled
histograms) male monkeys. Statistical analysis was performed on all male
animals as described in Fig. 1; significance is indicated by two-sided P values.
Messaoudi et al.
December 19, 2006 ?
vol. 103 ?
no. 51 ?
In this study, we provide evidence for several beneficial effects
of long-term CR on the immune system of aged nonhuman
primates. T cell population balance and T cell subset function
animals. Thus, long-term CR initiated early in adulthood re-
sulted in higher percentages of naı ¨ve T cells and lower percent-
ages of memory T cells in the circulating PBMC of aged
nonhuman primates. These cells in CR animals were not only
phenotypically, but also functionally, naı ¨ve. This is notable
because a higher prevalence of naı ¨ve lymphocytes bearing a
diverse TCR repertoire should preserve immune reactivity to
new antigens encountered late in life. In that regard, it has been
demonstrated that TCR diversity is important in immune de-
fense (7) and that its decline with aging can lead to impaired
resistance to infection and even to death (20). Inasmuch as
susceptibility to infectious diseases is among the top five leading
causes of mortality and morbidity in elderly patients, a more
efficient immune response in aged individuals is likely to in-
crease lifespan. In that regard, Miller and colleagues (46, 47)
demonstrated that in genetically heterogeneous mice, the per-
centage of blood CD4 memory T cells, followed by the percent-
age of blood CD8 memory cells, measured at 18 months of life,
was the most reliable inverse predictor of longevity. Further
studies suggested that the pace and/or extent of memory T cell
accumulation might also be a reliable predictor of longevity in
rodents (47). Similar studies in humans have shown a strong
association between chronic pathogens that induce progressive
accumulation of memory T cells (particularly CMV) and shorter
lifespans in advanced aging (38). It should be noted that all
monkeys in this study were CMV?. The ability of CR to reduce
the numbers of memory cells and increase the numbers of naı ¨ve
T cells in nonhuman primates is consistent with the idea that,
even in the face of chronic infection, CR is capable of amelio-
rating immunological aging and inhibiting immune exhaustion.
However, an alternative explanation that CR may reduce reac-
tivation or replication of persistent viruses cannot be ruled out
at present. Unfortunately, because RMs continuously shed CMV
in mucosal fluids (48), it is very difficult to measure viral
reactivation in this model. The final test of CR value to immu-
nity, of course, will lie in definitive experiments to demonstrate
whether CR animals are indeed more resistant to pathogenic
Long-term CR also reduced immediate, ex vivo cytokine
secretion by stimulated memory T cells. This contrasts with the
recent finding that 2 years of CR actually increased production
of IFN-? by PBMC (30) in another group of aged male RMs.
Differences in stimulation type and length, cell population,
duration of CR, or monkey cohort could, individually or to-
gether, account for the discrepancy between the two studies.
Regardless, our findings strongly suggest that lifelong CR may
have anti-inflammatory effects in primate T cells. Given the
association of the proinflammatory state with numerous chronic
diseases that are prevalent in old age, this effect would also be
likely to contribute to improved health and longevity in CR
animals. Clearly, a balance needs to exist between stimulating
adequate inflammatory and immune responses in the face of
immediate pathogen assault and dampening and controlling the
chronic responses, with potential to inflict immunopathology. In
that regard, our preliminary results on other RM cohorts kept on
CR for 8–9 years indicate that, in vivo, CR enables the mainte-
nance of robust primary immune responses. Meanwhile, in the
same cohort, the reactivity to a chronic pathogen, CMV, was not
appreciably different between CR and CON animals. If con-
firmed, this would suggest that reduction of baseline T cell
cytokine secretion does not adversely impact the ability to react
vigorously to new antigenic challenge, an issue critically impor-
tant to immune defense and longevity.
The key remaining challenge is to explain, at the molecular
and cellular levels, the mechanisms of CR action on diverse
organs and tissues. Our studies show that at least part of the CR
effect lies in the increased production and/or slower turnover of
naı ¨ve T cells, based on increased TREC levels. Because inflam-
matory cytokines tend to reduce thymic output by inducing
apoptosis in immature thymocytes (14), CR could increase
thymic output by reducing inflammatory cytokine levels. How-
ever, this effect could also occur at other steps of hematopoietic
and/or early T cell lineage development, and it may also involve
improved function of thymic stromal cells. As samples from the
animals studied here become available at necropsy, it may be
possible to address some of these questions. Moreover, molec-
ular analysis of gene and protein expression patterns from
primary and secondary lymphoid tissues, as well as from defined,
phenotypically and functionally homogenous lymphocyte sub-
sets, should provide additional clues regarding the impact of CR
on the immune system.
Materials and Methods
Animals. Monkeys in this study were part of an ongoing study of
aging and CR in RMs at the NIA and were housed continuously at
the Primate Unit of the National Institutes of Health (NIH)
Veterinary Research Program, Poolesville, MD. Diet composition
and animal husbandry procedures were as described in ref. 31.
Experimental cohorts (Table 1) included 13 CR (7 male and 6
female) and 28 CON (18 male and 10 female) rhesus monkeys
19–23 years of age. Over the 42 months of examination, attrition
reduced these numbers by one CON male, two CON females, and
two CR females (see SI Discussion). We obtained complete cohort
samples at four time points over the 42-month period and, unless
otherwise indicated, have used all four time points to perform
statistical analysis and draw our conclusions (see below).
Sample Collection, Flow Cytometry, and Functional Assays. After an
overnight fast, peripheral blood was collected within 30 min of
anesthesia into heparinized tubes, as described in refs. 33 and 34.
Whole blood was analyzed with a Beckman Coulter (Hialeah,
FL) complete blood count machine. PBMCs were isolated from
St Louis, MO). Cells were either analyzed immediately or were
frozen for subsequent analysis by using a controlled cyropreser-
Forma, Marietta, OH), with no significant differences seen
between the analyzed properties of fresh and frozen/thawed
cells. All antibodies were purchased from PharMingen (San
Diego, CA), Ebioscience (San Diego, CA), or Caltag (Burlin-
game, CA) and were used in accordance with manufacturer’s
recommendations. Samples were collected using FACSCalibur
or FACSLSRII (Becton Dickinson, San Jose, CA), and data
were analyzed using CellQuest (BD Biosciences, Mountain
View, CA) or FlowJo (Treestar, Ashland, OR), with a minimum
of 105events collected per sample. Proliferation assay and
intracellular cytokine staining were performed exactly as de-
scribed in ref. 34.
CDR3 Length and TREC Analysis. RNA was isolated from 3 ? 106to
5 ? 106PBMC by using RNAisolator (Sigma). cDNA reverse
transcription, TCRV?-specific PCRs, and runoff labeling with
fluorescent primers were performed as described previously
(49), using the primers described in ref. 50.
Quantification of TREC in DNA of PBMC was done by
quantitative–competitive PCR exactly as described in ref. 35.
DNA was extracted from the cells, and 1 ?g was added to each
PCR, which also contained 102, 103, or 104molecules of an
internal competitor standard. Primers and conditions were as
www.pnas.org?cgi?doi?10.1073?pnas.0606661103Messaoudi et al.
described in ref. 35. The reactions were run for 35 cycles and
incorporated32P-labeled dCTP. PCR products were separated
on nondenaturing 6% polyacrylamide gels. Bands were imaged
and analyzed with a Cyclone storage phosphor system and
Optiquant software (Packard, Meriden, CT).
Statistical Analysis. Because the parameters on all animals were
measured at multiple time points over time (four times over 42
measurement), mixed-effects models were used to analyze
whether the CR and CON groups had similar temporal changes
in WBC, percentage of naı ¨ve cells, percentage of cells divided,
and cytokine secretion. The covariance structure was assumed to
have a compound symmetry based on the Bayesian information
criterion (51). TREC data were analyzed by one-way ANOVA.
The data on the number of divisions were assumed to have
Poisson distribution and were analyzed by using the generalized
estimating equations (GEE) method (52). All statistical analyses
were performed using Statistical Analysis System software (SAS
version 9.0; SAS Institute, Cary, NC), and P ? 0.05 was
considered statistically significant.
We thank April Hobbs and Ed Tilmont (NIA Intramural Research
Program, Poolesville, MD) for their work and dedication, Drs. Doug
Powell and Rick Herbert (NIH Veterinary Research Program, Pooles-
ville, MD) for the veterinary care of the monkeys in this study and for
collection of the specimens, Dr. John Fanton [Oregon National Primate
Research Center (ONPRC)] for surgical assistance, and Dr. Steven
Kohama (ONPRC) for helpful discussion. This work was supported in
part by U.S. Public Health Service Awards AG21384 (from the NIA
to J.N.-Zˇ.), 5T32 AI007472-10 [from the National Institute of Allergy
and Infectious Diseases (NIAID) to I.M.], and RR0163 (from the
National Institute for Research Resources to ONPRC), as well as by the
NIA and NIAID intramural programs.
1. McCay CM, Crowell MF, Maynard LA (1989) Nutrition 5:155–171; discussion
2. Masoro EJ (2005) Mech Ageing Dev 126:913–922.
3. Heilbronn LK, Ravussin E (2003) Am J Clin Nutr 78:361–369.
4. Hursting SD, Lavigne JA, Berrigan D, Donehower LA, Davis BJ, Phang JM,
Barrett JC, Perkins SN (2004) J Nutr 134:2482S–2486S.
5. Guarente L, Picard F (2005) Cell 120:473–482.
6. Sinclair DA (2005) Mech Ageing Dev 126:987–1002.
7. Messaoudi I, Guevara Patino JA, Dyall R, LeMaoult J, Nikolich-Z ˇugich J
(2002) Science 298:1797–1801.
8. Walford RL (1982) J Am Geriatr Soc 30:617–625.
9. Miller RA (1996) Science 273:70–74.
10. Nikolich-Zˇugich J (2005) J Exp Med 201:837–840.
11. Gardiner I (1980) Rev Infect Dis 3:2801–2814.
12. Wick G, Jansen-Durr P, Berger P, Blasko I, Grubeck-Loebenstein B (2000)
13. Miller RA, Garcia G, Kirk CJ, Witkowski JM (1997) Immunol Rev 160:79–90.
14. Linton PJ, Dorshkind K (2004) Nat Immunol 5:133–139.
15. Naylor K, Li G, Vallejo AN, Lee WW, Koetz K, Bryl E, Witkowski J, Fulbright
J, Weyand CM, Goronzy JJ (2005) J Immunol 174:7446–7452.
16. Callahan JE, Kappler JW, Marrack P (1993) J Immunol 151:6657–6669.
17. Hingorani RC, Choi IH, Akolkar P, Gulwani-Akolkar B, Pergolizzi R, Silver
J, Gregersen PK (1993) J Immunol 151:5762–5769.
18. Posnett DN, Sinha S, Kabak S, Russo C (1994) J Exp Med 179:609–617.
19. LeMaoult J, Szabo P, Weksler ME (1997) Immunol Rev 160:115–126.
20. Messaoudi I, LeMaoult J, Guevara-Patino JA, Metzner BM, Nikolich-Zˇugich
J (2004) J Exp Med 200:1347–1358.
21. Chen J, Astle CM, Harrison DE (1998) J Gerontol A Biol Sci Med Sci
22. Jolly CA (2004) J Nutr 134:1853–1856.
23. Mattison J, Lane M, Roth G, Ingram D (2003) Exp Gerontol 38:35–46.
24. Ramsey JJ, Colman RJ, Binkley NC, Christensen JD, Gresl TA, Kemnitz JW,
Weindruch R (2000) Exp Gerontol 35:1131–1149.
25. Lane MA, Black A, Handy A, Tilmont EM, Ingram DK, Roth GS (2001) Ann
NY Acad Sci 928:287–295.
26. Roth GS, Lane MA, Ingram DK, Mattison JA, Elahi D, Tobin JD, Muller D,
Metter EJ (2002) Science 297:811.
27. Roecker EB, Kemnitz JW, Ershler WB, Weindruch R (1996) J Gerontol A Biol
Sci Med Sci 51:B276–B279.
28. Weindruch R, Lane M, Ingram D, Ershler WB, Roth G Aging Clin Exp Res
29. Grossmann A, Rabinovitch PS, Lane MA, Jinneman JC, Ingram DK, Wolf NS,
Cutler RG, Roth GS (1995) J Cell Physiol 162:298–303.
Ingram DK (2002) J Interferon Cytokine Res 22:565–571.
31. Ingram DK, Cutler RG, Weindruch R, Renquist DM, Knapka JJ, April M,
Belcher CT, Clark MA, Hatcherson CD, Marriott BM, Roth GS (1990) J
32. Walford RL, Mock D, Verdery R, MacCallum T (2002) J Gerontol A Biol Sci
Med Sci 57:B211–B224.
33. Pitcher C, Hagen S, Walker J, Lum R, Mitchell B, Maino V, Axthelm M, Picker
LJ (2002) J Immunol 168:29–43.
34. Jankovic ´ V, Messaoudi I, Nikolich-Zˇugich J (2003) Blood 102:3244–3251.
35. Sodora D, Douek DC, Silvestri G, Montgomery L, Rosenzweig M, Igarashi T,
Bernacky B, Johnson RP, Feinberg MB, Martin MA, Koup RA (2000) Eur
J Immunol 30:1145–1153.
36. Pannetier C, Even J, Kourilsky P (1995) Immunol Today 16:176–181.
37. Clambey ET, van Dyk LF, Kappler JW, Marrack P (2005) Immunol Rev
38. Pawelec G, Akbar A, Caruso C, Effros R, Grubeck-Loebenstein B, Wikby A
(2004) Trends Immunol 25:406–410.
39. Tough DF, Sprent J (1994) J Exp Med 179:1127–1135.
40. Effros RB (2004) Exp Gerontol 39:885–890.
41. Walford RL, Liu RK, Gerbase-Delima M, Mathies M, Smith GS (1973) Mech
Ageing Dev 2:447–454.
42. Effros RB (1997) Dev Comp Immunol 21:471–478.
43. Ershler WB, Sun WH, Binkley N (1994) Drugs Aging 5:358–365.
44. Zanni F, Vescovini R, Biasini C, Fagnoni F, Zanlari L, Telera A, Di Pede
P, Passeri G, Pedrazzoni M, Passeri M, et al. (2003) Exp Gerontol 38:981–
45. De Martinis M, Franceschi C, Monti D, Ginaldi L (2005) FEBS Lett 579:2035–
46. Miller RA (1997) Mech Ageing Dev 96:181–196.
47. Harper JM, Galecki AT, Burke DT, Miller RA (2004) Mech Ageing Dev
48. Huff JL, Eberle R, Capitanio J, Zhou SS, Barry PA (2003) J Gen Virol
49. Pannetier C, Cochet M, Darche S, Casrouge A, Zoller M, Kourilsky P (1993)
Proc Natl Acad Sci USA 90:4319–4323.
50. Chen Z, Li Y, Zeng X, Kuroda M, Schmitz J, Shen Y, Lai X, Shen L, Letvin
N (2001) J Immunol 166:4525–4533.
Control (Prentice–Hall, Englewood Cliffs, NJ).
52. Zeger SL, Liang KY (1986) Biometrics 42:121–130.
Messaoudi et al.
December 19, 2006 ?
vol. 103 ?
no. 51 ?