aUniversity of Illinois, Department of Pathobiology, 2001 S. Lincoln Ave., Urbana, IL 61802, USA
bWashington University, Department of Mathematics, St. Louis, MO 63130, USA
cNestle Research Center, St. Louis, MO 63164, USA

Received 25 July 2005; accepted 9 February 2006. Available online 29 March 2006.
Abstract
Caloric restriction (CR) has been shown to retard immunosenescence and to extend median and maximum life span in rodent species. Longitudinal effects of CR on the canine immune system are presented in this report. A group of 48 Labrador Retrievers, divided at weaning into weight- and sex-matched pairs, were maintained on a diet restriction protocol from age 8 weeks until death. Each restricted dog received 75% of the total food consumed by its control-fed pair mate. Immune parameters were monitored from 4 to 13 years. CR retarded age-related declines in both lymphoproliferative responses and absolute numbers of lymphocytes and the T, CD4, and CD8-cell subsets. In females, CR attenuated the age-related increase in T-cell percentages and marginally retarded the age-related increase in memory cell percentages. Age-related changes in B-cell percentages and numbers were augmented by CR. No direct effect of CR on phagocytic activity of PMN, antibody production or NK cell activity, was observed. Lower lymphoproliferative responses, lower numbers of lymphocytes, T, CD4 and CD8 cells, lower CD8 percentages and higher B-cell percentages were all found to be significantly associated with a decreased likelihood of survival in these dogs.
Keywords: Aging; Caloric restriction; Immune response; Immunosenescence; Canine; longitudinal study
Abbreviations: CR, caloric restriction; CTAC, canine thyroid adenocarcinoma

Article Outline
1. Introduction
2. Materials and methods
2.1. Animals, diet and testing schedules
2.2. White blood cell analysis and isolation of peripheral blood lymphocytes (PBL)
2.3. Lymphoproliferative response to mitogens
2.4. Flow cytometric analysis of lymphocyte subset distributions
2.5. Natural Killer Cell (NK) activity
2.6. PMN isolation and evaluation of phagocytosis
2.7. In vivo antibody formation
2.8. Statistical analysis
3. Results
3.1. Lymphoproliferative responses to mitogens
3.2. White blood cell and lymphocyte subset analysis
3.3. Other immune measurements
3.4. Survival analysis
4. Discussion
Acknowledgements
References

1. Introduction
Over 60 years after its initial description as a life-extension strategy, caloric restriction (CR) or reduced energy intake remains the only well-documented non-genetic intervention capable of extending both median and maximum life span. While the precise mechanisms through which CR prolongs life and retards age-related physiological decline are not entirely understood, evidence from a number of species suggests that the attenuation of age-related changes in immune function is integral to this process. It remains to be determined whether retardation of immunosenescence represents a driving force leading to increased longevity.
Immunosenescence is associated with decreased survivability, presumably due to a decline in an organism's ability to resist both internal and external stresses, especially infectious agents and tumors. Investigation of the effects of CR on aging in laboratory rodents has generated a large body of evidence supporting benefits of CR in retarding immunosenescence (for review see Pahlavani, 2000). Early studies demonstrated positive effects of CR on ex vivo lymphoproliferative responses and cell cytolytic activity, while more recent studies have documented effects on age-related changes in lymphocyte subset distribution, cytokine production, apoptosis, cell signaling events and most recently, gene expression (Pahlavani, 2004, Lee et al., 1999 and Cao et al., 2001). CR has also been found to reduce tumor incidence in susceptible strains (Sheldon et al., 1995 and Weindruch, 1992), and retard the onset and reduce the severity of rodent autoimmune diseases (Fernandes et al., 1976 and Friend et al., 1978). Until recently, the question of whether the benefits of CR would apply to larger, longer-lived mammals as well has remained unanswered. In the late 1980s and early 1990s studies evaluating the effects of CR on age-associated physiological changes in primates were initiated (Ingram et al., 1990 and Kemnitz et al., 1993). While verification of beneficial effects of CR on primate life span will not be available for another decade, to date, the calorically restricted primates have demonstrated many of the age-related physiological responses observed in rodents (Lane et al., 1997 and Mattison et al., 2003). Some disparity exists, however, with regard to immune findings in primates as compared to rodents. In the Wisconsin colony, 2-4 years of CR of male Rhesus monkeys actually resulted in depression of lymphoproliferative responses to mitogens, as well as reduced NK activity and Ab responses; lymphocyte subset distributions and lymphocyte counts in peripheral blood were unaffected (Roecker et al., 1996). A benefit of CR in reducing IL-6 levels associated with oxidative stress has also been observed in this colony (Kim et al., 1997). In the NIH colony, lymphoproliferative responses of Rhesus males following 7 years of CR were lower than those of control-fed monkeys in those animals restricted from an early age, but not in those animals whose restriction was initiated at age 3-5 years; lymphopenia was observed in restricted-fed animals (Weindruch et al., 1997). In this same study CR had a beneficial effect on the age-related decline in IFN gamma production in response to PHA, but had no effect on age-related increases in IL-6 and IL-10 production (Mascarucci et al., 2002).
The present study was initiated in 1987 with the goal of investigating the effect of CR on the incidence and severity of canine hip dysplasia. Evaluation of the dogs at 2 years of age revealed marked benefits of the diet in retarding orthopedic disease (Kealy et al., 1992). At that point, the study was extended and expanded with the intention of determining the effects of CR on life span as well as on a number of age-associated physiological parameters. CR, from 8 weeks of age until death, was found to extend the median life span of the dogs by 15%, from 11.2 years in the controls to 13 years in the restricted group (Kealy et al., 2002).
A battery of immunological parameters known to undergo age-related changes in other species were identified and adapted for dogs; repeated measurements of these parameters were collected from age 4 years until death. Analysis of age-related changes in immunological parameters in the control-fed dogs from age 4-11 years have been previously reported (Greeley et al., 2001). The relationships between immune findings and diet group, age, gender and ultimate life span are reported herein.
2. Materials and methods
2.1. Animals, diet and testing schedules
Forty-eight Labrador Retrievers (30 females, 18 males) from a total of seven litters were housed, fed and managed at the Purina PetCare facility (Gray Summit, MO); dogs were divided at weaning into litter-, weight- and gender-matched pairs and randomly assigned to a feeding group. The composition of the diet and the original feeding regimen were as previously described (Kealy et al., 1992). Briefly, from age 8 weeks until death, each restricted-fed dog received 75% of the total food consumed by its control-fed pair mate on the previous day. At 3.25 years, the diet was adjusted from a growth formula suitable for younger dogs to an adult formula. At the same time, the ad libitum feeding protocol for the "control-fed" dogs was modified to prevent obesity and maintain an ideal body weight for each dog based on skeletal size; the restricted-fed dogs continued to receive 75% of the control-fed diet (Kealy et al., 1997). The feeding regimen resulted in a reduction in mean body weight of 26% and a significant extension of median life span in the restricted-fed dogs (Kealy et al., 2002).
A high rate of pseudopregnancy accompanied by reduced food intake was observed in the control-fed females; if ovariohysterectomy became necessary for therapeutic purposes, the corresponding pair mate underwent the same procedure in order to maintain the experimental design (Lawler et al., 1999). The same procedure was followed for those male pairs where orchidectomy became necessary.
Blood samples were obtained from each dog on three different dates each year from 1991 to 2000: once for analysis of PMN phagocytosis, once for analysis of lymphoproliferation, lymphocyte subset distribution and white blood cell counts, and once for analysis of NK cell activity and a repeated analysis of lymphoproliferation (Table 1). Each individual assay on each member of a litter was performed in the same month each year, in order to avoid possible effects of seasonal variation on the measurements taken in an individual dog. Both dogs of a diet pair were always evaluated on the same day, with an average of eight dogs tested on each test date. During the latter 6 years of the 10-year period reported herein, 47 clinically healthy, young Labrador Retrievers (30 females and 17 males ranging in age from 0.8 to 3.6 years) were tested along with the study subjects as age-reference controls; on a given test date, 2-4 young dogs were included. These young dogs were housed under identical conditions and fed nutritionally complete and balanced diets comparable to that of the study subjects.
Table 1.
Experimental design and testing schedule
Individual litters
Month of birth (1987) Number of littermates testeda
Age (years) at initial testingb Restricted Control Lymphocyte %'s and #'s/proliferation NK/proliferation Phagocytosisc Females Males Females Males
February 2 1 2 1 4.1 4.6 4.9
February 3 1 3 1 4.1 4.6 4.9
March 1 1 4.0 4.5 4.8
April 1 2 1 2 4.0 4.5 4.8
April 2 2 2 2 4.1 4.6 4.8
May 1 2 1 2 4.0 4.5 4.8
August 4 1 4 1 4.1 4.6 4.9
a Number of littermates of each diet group and gender at the onset of testing.
b Each assay was performed in the same month each year for a given litter.
c Assays: FLOW analysis of lymphocyte subsets, automated cell counts, in vitro proliferative responses to mitogens, NK activity, and PMN phagocytic activity.

2.2. White blood cell analysis and isolation of peripheral blood lymphocytes (PBL)
Venous blood was collected in acid citrate dextrose (ACD) tubes, packed in insulated containers to maintain the temperature at 24 ± 4 °C and shipped Next Flight Out to the University of Illinois for cell isolation and testing. Samples were generally received within five hours of collection, and were processed upon arrival. Absolute WBC counts and differential percentages were evaluated using a CELL-DYN 3500 automated hematology analyzer (Abbott Lab., North Chicago, IL). Lymphocytes were isolated from peripheral blood by density gradient centrifugation, as previously described (Greeley et al., 1996). Since both monocytes and eosinophils may separate with lymphocytes in the dog, the actual percentage of lymphocytes in the isolated cells was determined by doing differential counts of stained Cytospin smears. Cell suspensions for all assays of lymphocyte function were then prepared at actual concentrations of lymphocytes per milliliter.
2.3. Lymphoproliferative response to mitogens
The in vitro responses of lymphocytes to mitogens were evaluated as described previously (Greeley et al., 1996 and Greeley et al., 2001). Briefly, isolated PBL were cultured at a final concentration of 2.5 × 105 lymphocytes/ml in a volume of 0.2 ml, in the presence of the mitogens Con A, (Sigma), Difco PHA-P (Fisher, Itaska, IL), and PWM (Gibco/BRL, Grand Island, NY). After 72 h, cultures were pulsed for 18 h with 1 µCi 3H-thymidine and then harvested and counted. For individual dogs, the maximal response for each mitogen was determined at the mitogen concentration eliciting the highest ?cpm (mean cpm of triplicate wells in the presence of mitogen - mean cpm in the presence of medium). Proliferative responses were measured biannually (Table 1), and an average annual response for each dog was determined.
2.4. Flow cytometric analysis of lymphocyte subset distributions
Staining and analysis of canine B, T, CD4 and CD8 cells were performed as in our previous studies (Greeley et al., 1996) using 30 min. incubations. At the 8-year test point, a rat moAb specific for canine CD4, Clone YKIX32.9 (Harlan Serotec, Indianapolis, IN) was substituted for the original mouse moAb obtained from D. Gebhard (NC State Univ.) after establishing that the staining properties of the two moAbs did not differ significantly; likewise, for CD8 detection, a rat moAb, Clone YCATE 55.9 (Harlan Serotec, Indianapolis, IN) was substituted at the 12-year test point. For detection of expression of high levels of CD44, cryopreserved cells were used. Cells were rapidly thawed, fixed with 1% paraformaldehyde and immediately stained with mouse anti-CD44 using BAG40A monoclonal antibody @1:500 (VMRD, Inc. Pullman, WA) followed by 1:40 FITC goat anti-mouse Ig (Southern Biotechnologies Associates, Inc., Birmingham, AL). Following phycoerythrin (PE) staining for CD4 or CD8 using the rat moAbs described above, the cells were analyzed and the percentage of PE-positive cells staining brightly for the CD44 marker (Log FITC) was recorded as the percentage of putative memory cells.
2.5. Natural Killer Cell (NK) activity
NK activity was evaluated as described previously (Greeley et al., 1996) utilizing 51Cr-labeled canine thyroid adenocarcinoma (CTAC) target cells and four effector to target ratios.
2.6. PMN isolation and evaluation of phagocytosis
PMN were isolated from peripheral blood and their ability to phagocytose fluorescent latex beads was evaluated as previously described (Greeley et al., 2001).
2.7. In vivo antibody formation
Dogs were immunized with one of the following thymus-dependent antigens: 10 mg soluble KLH (Calbiochem) at age 5 and 8 years; 10 mg soluble OVA (Sigma) at age 6 years, and a 5% suspension of SRBC at 7 and 9 years. All immunogens were administered without adjuvants via the subcutaneous route in a volume of 3 ml. Antibody titers to the protein immunogens were assessed by ELISA (Greeley et al., 1996) using serum samples collected at previously determined peak response times: 21 days following immunization for primary responses and 7 days for secondary. Anti-SRBC responses were evaluated using a tube hemagglutination assay; primary responses were determined at 14 days, and secondary at 7 days.
2.8. Statistical analysis
All observations were analyzed and are presented as least squares mean values to adjust for the fact that, within a given year, numbers of male and female dogs remaining in the study were different. Each immunological parameter was analyzed using a mixed-effects ANOVA model (Littell et al., 1996); diet group, gender, age, and their interaction were considered to be the fixed effects of interest. The repeated-measures aspect of the data was addressed by assigning random block effects to individual dogs within pairs. Variations among litters, litter interactions with age and gender, the three-way interaction of litter with age and gender, and variation among individual dogs of the same gender belonging to the same litter were analyzed with random effects. PROC MIXED in SAS® (SAS Institute Inc., Cary, NC) was used to obtain the estimates via restricted maximum likelihood estimation. Logarithmic transformations were employed for variables with skewed distributions. Relationships of immune variables to survival were examined using Cox proportional hazards methodology with time-dependent covariates as implemented in PROC PHREG in SAS.
All of the recorded data for each dog from age 4 years through 13.5 years (or death) were used in the analyses with the exception of one dog diagnosed with diabetes at 10 years. For this dog, only data collected before the onset of illness were included in the analyses. Numbers of dogs of each gender that were analyzed at each age point are indicated in Table 2, with numbers that had been spayed or neutered prior to each time point indicated in parentheses. Although the overall reproductive status of the females changed dramatically from age 7 years onward, at any point in time, the proportion of ovariohysterectomized dogs was relatively constant in the two diet groups, since surgical procedures were always performed on both dogs in a diet pair when therapeutically indicated for one of the pair.
Table 2.
Numbers of dogs analyzed at each time pointa
Diet group Gender Age (years)
4 5 6 7 8 9 10 11 12 13
Number of dogs tested
Restricted Female 14 14 14 (5)b
13 (7) 12 (8) 12 (10) 12 (10) 10 (8) 9 (7) 8 (7)
Restricted Male 9 9 9 9 9 (1) 9 (1) 9 (1) 8 (3) 8 (3) 5 (1)
Control Female 14 14 14 (5) 14 (7) 12 (8) 12 (9) 11 (10) 8 (8) 5 (5) 1 (1)
Control Male 9 9 9 9 9 (1) 9 (1) 6 (0) 5 (2) 3 (1) 0

Total dogs 46 46 46 45 42 42 37 31 25 14
a FLOW, WBC counts, NK activity and PMN phagocytosis were evaluated annually in each dog; lymphoproliferation was assessed 2X/year.
b Numbers in parentheses indicate spayed or neutered animals in each analysis group.

Formal tests of significance were not performed to compare the young age-reference dogs to those in the longitudinal study, since differences between groups are confounded with age, diet, and other factors. An ANOVA model using the predictors of gender, age, gender × age interaction and dog within gender was used to generate the least squares means for the immune parameters of the young group; values are presented for qualitative comparisons only. (From 12 to 19 young dogs were tested as age-references each year from the age "8-year" time point on.)
3. Results
3.1. Lymphoproliferative responses to mitogens
An age-related decline in the maximum responses to all three mitogens was detected in both diet groups (p < 0.001; Fig. 1A), with the restricted-fed dogs demonstrating a significantly slower rate of decline compared to control-fed dogs (age × diet Con A: p < 0.001; PHA: p < 0.01; PWM: p = 0.08). While CR significantly retarded the rate of decline of lymphoproliferative responses to Con A in both females and males, (age × diet, p < 0.01 for females; p < 0.05 for males), the pattern of diet-related effects in the two genders differed as seen in Fig. 1B. In females, the restricted-fed dogs had responses to Con A equal to or exceeding those of the control-fed dogs at all time points (p = 0.01); a similar pattern was seen for PHA (p < 0.05) and PWM (p = 0.06). In males, the pattern was somewhat different: at 4-7 years, the control-fed dogs actually had higher responses to Con A than restricted-fed dogs with no measurable differences between the two diet groups from 8 years onward.


Fig. 1. The effect of caloric restriction on lymphoproliferative responses to mitogens. Peripheral blood lymphocytes were cultured in the presence of Con A, PHA and PWM. Mean log cpm ± S.E. following 3 days in culture vs. age are presented. Restricted-fed dogs are presented as closed symbols and control-fed as open symbols; squares indicate responses for females, diamonds for males, and circles for the two genders combined. (A) Responses to Con A (slashed lines), PHA (solid lines) and PWM (dotted lines) are presented. Responses for young reference control animals are indicated in the side panel. CR retarded the rate of decline for all three mitogens (Con A, PHA p < 0.01; PWM p = 0.08). (B) Responses to Con A are presented as a function of gender and age. Responses of control-fed females are less than restricted-fed females (p = 0.01).

3.2. White blood cell and lymphocyte subset analysis
Commencing at age 7 years, total WBC and differential counts were monitored. While the numbers of WBC and PMN were not affected by age or diet (data not shown), the absolute numbers of lymphocytes and all subsets of lymphocytes declined with age (all p < 0.001; Table 3). Restricted-fed dogs had lower numbers of B cells than their control-fed pair mates (p = 0.02). Analysis of rate of change over time reveals that for numbers of total lymphocytes, T cells and CD8 cells, the control-fed dogs demonstrated significant age-related declines over time (age × diet, p = 0.001 for each cell type) while the restricted-fed dogs demonstrated no decline. For CD4, the decline over time in both diet groups was significant, but the rate of decline for the control-fed dogs was of significantly greater magnitude (age × diet, p < 0.001).
Table 3.
The effect of caloric restriction on cellularity
Cell type Diet Age (years)
7 8 9 10 11 12 13
Number of cells (×106/ml blood)
Lymphsa
Restricted 1.63 ± 0.30 1.70 ± 0.30 1.34 ± 0.30 1.44 ± 0.30 1.60 ± 0.30 1.54 ± 0.30 1.52 ± 0.31
Control 2.06 ± 0.30 1.76 ± 0.30 1.42 ± 0.30 1.63 ± 0.30 1.38 ± 0.31 1.42 ± 0.32

B Cells Restricted 0.13 ± 0.04 0.13 ± 0.04 0.14 ± 0.04 0.10 ± 0.04 0.10 ± 0.04 0.14 ± 0.04 0.07 ± 0.04
Control 0.20 ± 0.04 0.18 ± 0.04 0.16 ± 0.04 0.13 ± 0.04 0.14 ± 0.04 0.15 ± 0.04

T Cellsa
Restricted 1.40 ± 0.30 1.40 ± 0.30 1.15 ± 0.30 1.19 ± 0.30 1.41 ± 0.30 1.31 ± 0.30 1.35 ± 0.30
Control 1.80 ± 0.3 1.45 ± 0.3 1.17 ± 0.3 1.33 ± 0.3 1.20 ± 0.31 1.20 ± 0.32

CD4 Cellsa
Restricted 0.80 ± 0.17 0.74 ± 0.17 0.58 ± 0.17 0.62 ± 0.17 0.61 ± 0.17 0.60 ± 0.17 0.67 ± 0.17
Control 1.00 ± 0.17 0.78 ± 0.17 0.62 ± 0.17 0.69 ± 0.17 0.56 ± 0.17 0.60 ± 0.18

CD8 Cellsa
Restricted 0.53 ± 0.09 0.55 ± 0.09 0.40 ± 0.09 0.44 ± 0.09 0.58 ± 0.09 0.52 ± 0.09 0.47 ± 0.10
Control 0.67 ± 0.09 0.56 ± 0.09 0.41 ± 0.09 0.46 ± 0.09 0.43 ± 0.10 0.44 ± 0.10
Peripheral blood counts and differentials were performed using an automated hematology analyzer. Counts for lymphocyte subsets were obtained using flow cytometry-derived percentages and lymphocyte counts and are presented as mean numbers of cells ± S.E.
a A significant effect of CR on the age-related decline in numbers of lymphocytes, T, CD4 and CD8 cells was observed (age × diet, p = 0.001 for all cell types).

Age-related alterations of lymphocyte subset percentages in peripheral blood included decreases in CD4-cell and B-cell percentages (p < 0.001 for both), with restricted-fed dogs demonstrating lower B-cell percentages than control-fed dogs (p = 0.04; Fig. 2). No age-related changes in CD8 T cell percentages were observed for the 4-13 years period, although the values for the young reference controls appeared to be dramatically lower than those of the test dogs-even at the 4-year time point. Age-related increases in T-cell percentages were observed overall (p = 0.05), with females having higher percentages of T cells than males (p < 0.01). In females, a direct effect of CR on T-cell percentages was observed, with restricted-fed females demonstrating lower T-cell percentages over time compared to the control-fed females (p < 0.05).


Fig. 2. The effect of caloric restriction on lymphocyte subset distribution. Peripheral blood lymphocytes were stained with appropriate labeled antibodies and percentages (mean ± S.E.) of positively staining cells were determined using flow cytometry. Restricted-fed dogs are presented as closed symbols and control-fed as open symbols; squares indicate responses for females, diamonds for males, and circles for the two genders combined. Values for young reference control animals are indicated in the side panels.

While it was not feasible to monitor canine memory T cells at the initiation of the study, subsequent availability of a suitable CD44 reagent allowed us to examine this putative memory marker as a function of diet and age using cryopreserved cells that had been collected throughout the study. Frozen cells from three age categories (4-6 years, 7-8 years, and 10-13 years) were evaluated for each diet pair. Representative staining patterns for lymphocytes from young, middle-aged and old dogs of both genders are presented in Fig. 3. The percentages of CD8 memory cells (as defined by CD44 bright staining) increased markedly with age; in females, diet restriction was found to be marginally beneficial in retarding this increase (Table 4). (A similar pattern was observed for CD4 cells, but since slightly lower levels of CD4 were detected on cryopreserved cells in comparison to freshly stained cells, these data were not included.) Significant age-related increases in percentages of CD44hi-stained cells were corroborated by testing fresh cells from young dogs and the study dogs at age 12 years. (For CD8 cells, the percent of CD44hi in young dogs was 47.6 ± 2.2 S.E., while that of the 12-year-old was 72.4 ± 1.9; for CD4, the percent of CD44hi in young dogs was 15.8 ± 1.6 and that in 12-year-old was 34.2 ± 1.4.)


Fig. 3. Detection of CD44 expression in canine CD8 lymphocytes. Cryopreserved cells from both genders and each of three age groupings were rapidly thawed, fixed with 1% paraformaldehyde, and stained with mouse anti-CD44 followed by 1:40 FITC goat anti-mouse Ig. Rat MoAbs for CD8 were added, followed by phycoerythrin (PE)-labeled goat anti-rat Ig. The percentage of PE-positive cells staining brightly for the CD44 marker (Log FITC) was determined. Staining profiles are as follows: (a) young male; (b) young female; (c) middle-aged male; (d) middle-aged female; (e) old male; (f) old female.

Table 4.
The effect of age and caloric restriction on the distribution of CD8 + CD44hi T cells
Age group (years) Gender Percentage of CD8 memory cellsa (least squares mean ± S.E.) Restricted diet Control diet
4-6 Female 64.2 ± 2.7b
68.1 ± 2.7
Male 66.1 ± 3.3 64.9 ± 3.5

7-8 Female 74.6 ± 2.8b
82.3 ± 2.8
Male 76.3 ± 3.3 76.7 ± 3.5

10-13 Female 79.6 ± 2.8b
87.2 ± 3.0
Male 83.7 ± 3.4 85.3 ± 3.7
a On a single test date, cryopreserved cells from each of the three age groupings for both the restricted and control dog in a given diet pair were tested. Cells were rapidly thawed, fixed with 1% paraformaldehyde, and stained with mouse anti-CD44 followed by 1:40 FITC goat anti-mouse Ig. Rat MoAbs for CD8 were added, followed by phycoerythrin (PE)-labeled goat anti-rat Ig. The percentage of PE-positive cells staining brightly for the CD44 marker (log FITC) was recorded as the percentage of memory cells.
b In females, development of memory cells was marginally retarded by CR (p = .07).

3.3. Other immune measurements
No significant effects of CR on NK activity or PMN phagocytic capacity were detected for the 9-year interval of testing. In addition, monitoring of antibody responses to thymus-dependent antigens from age 5-9 years revealed no consistent effect of diet group or gender (data not shown).
3.4. Survival analysis
When relationships of immune parameters to survival were examined using Cox proportional hazards methodology, an increased hazard of death was found to be significantly associated with: lower lymphoproliferative responses to PHA (p = 0.05) and PWM (p = 0.04), with a trend for Con A (p = 0.08). An increased hazard of death was also associated with lower cell counts for lymphocytes (p < 0.01), T cells (p = 0.04), CD4 cells (p = 0.07), and CD8 cells (p < 0.01). Additional risk factors included lower CD8 percentages and higher B-cell percentages. These associations with survival were calculated independent of the diet group; when diet groupings were taken into account, the PWM responses and the cell counts and percentages were still predictive.
4. Discussion
This study examining the effects of CR on immunosenescence was a segment of a comprehensive longitudinal study undertaken to examine the effects of CR on life span and age-related changes in numerous physiological parameters in the dog. The findings presented herein indicate that CR retards age-related changes in a number of immune parameters including lymphoproliferative responses to mitogens, and changes in lymphocyte subset distribution and numbers. Furthermore, several of these immune parameters positively affected by the diet were also predictive of an increased probability of survival, independent of diet.
Biannual evaluation of lymphoproliferative responses to mitogens from age 4-13 years revealed that CR significantly retarded the rate of age-related decline in Con A and PHA responses, an effect that was more pronounced in females than males.
Prevention or retardation of the age-related decline in in vitro lymphoproliferative responses to mitogens is one of the earliest and most consistently described benefits of CR on rodent immunosenescence (Gerbase-DeLima et al., 1975, Weindruch et al., 1982a, Weindruch et al., 1982b, Tian et al., 1995, Goonewardene and Murasko, 1995 and Fernandes et al., 1997). Both Gerbase-DeLima et al. (1975) and Fernandes et al. (1997) found that CR was more effective in potentiating the lymphoproliferative responses of older animals, and, in fact observed higher responses in the control-fed groups in younger rodents. A similar pattern of effects was observed in males in the present study. Whether gender consistently contributes to the effects observed following CR is not clear, since most studies have examined only one gender. Goonewardene and Murasko (1995) utilized both genders in their rat study, but do not address the issue of gender effects on lymphoproliferative responses, although they did find a gender difference in the affect of CR on life span. In two ongoing primate studies utilizing males, no beneficial effect of CR on lymphoproliferative responses of younger animals has been observed (Weindruch et al., 1997 and Roecker et al., 1996). There may well be gender differences in the level and/or timing of restriction needed to attain optimal biological effects at each stage of life. We recognize that the necessity for therapeutic ovariohysterectomy or orchidectomy in the present study may be an additional factor that could influence diet outcomes.
CR prevented the age-related decline in numbers of lymphocytes, T cells and CD8 cells and retarded the rate of decline of CD4 cells, while augmenting the decline in numbers of B cells with age. Although a borderline lymphopenia was observed in the restricted dogs at 7 years, this effect did not persist at later time points. Lymphopenia associated with caloric restriction has been observed in mice (Weindruch and Walford, 1988, Volk et al., 1994, Spaulding et al., 1997 and Chen et al., 1998) and in primates following 7 years of CR (Weindruch et al., 1997), but was not observed in primates following a shorter term of CR (Roecker et al., 1996). The reported disparities in the effects of CR on lymphocyte numbers may relate to strain and species differences, to the organ chosen for lymphocyte monitoring, and to the length of restriction relative to the total life span of the animal.
Age-related changes in B-cell percentages were the only lymphocyte subset distribution directly altered by CR, with control dogs demonstrating significantly higher percentages than CR dogs. This finding can be viewed as a beneficial effect of CR, since higher B-cell percentages were found to be associated with decreased survival potential in this study. While no significant effect of diet on T-cell percentages was observed for the males, in females CR prevented the age-related increase in T-cell percentages, again suggesting that gender may be an important variable in evaluating the beneficial effects of CR. Roecker et al. (1996) found no effect of CR on lymphocyte subset distributions in male Rhesus monkeys. While some beneficial effects of CR in retarding age-related changes in subset distributions in rodents have been described (Miller, 1997 and Chen et al., 1998), a clear pattern of effects of CR on lymphocyte subset distribution has not emerged.
A hallmark of aging in all species examined to date is the age-related shift from a naïve to memory phenotype in both CD4 and CD8 T-cell populations; CR appears to be extremely beneficial in retarding or preventing this shift. In F-344xBN rats, CR resulted in only a minimal shift to the memory phenotype (CD45RC/Ox-22low) in both CD4 and CD8 lymphocytes of 30-month old rats restricted from 16 weeks of age (Fernandes et al., 1997). Similarly in mice, the percentages of CD4 and CD8 memory cells (CD44hi, CD4+ and CD44hi, CD8+) in CR old animals were substantially lower than in the ad lib counterparts (Chen et al., 1998 and Miller, 1997). In the present study, the percentage of both CD4 and CD8 T cells expressing high levels of CD44 was found to increase dramatically with age; CR marginally retarded the rate of increase in CD44 expression by CD8 cells in females. While a correlation between CD44hi expression and memory has not been directly established in dogs, the parallel findings to those in rodents are intriguing.
A clear-cut effect of CR on canine NK activity was not observed in the present study. Previous studies have failed to establish a predictable relationship between CR and NK activity; Weindruch et al. (1983) and Roecker et al. (1996) have reported reduced levels of NK activity in CR mice and primates respectively. Riley et al. (1989) found no effect in rats. Gilman-Sachs et al. (1991) observed increased numbers of NK cells in CR rats; however, this observation may not have represented a beneficial effect of the diet, since an age-related increase in numbers of NK cells was also observed in these animals.
Establishment of reliable biomarkers of aging would greatly facilitate evaluation of life-extension strategies in a timely and cost-effective manner. In the present study, lower values of lymphoproliferative responses and lymphocyte, T, CD8 and CD4-cell numbers were associated with an increased hazard of death while lower B-cell percentages were predictive of a decreased hazard of death. Heller and colleagues monitored an extensive panel of potential biomarkers over time in a group of heterogeneous mice and found lower Con A lymphoproliferative responses and higher natural killer cell activity to be associated with decreased survival (Heller et al., 1998). In a study of 102 elderly Swedish individuals, immune parameters were evaluated and associations with survival were determined at time intervals thereafter; lower Con A responses and higher CD8 percentages were associated with decreased survival (Ferguson et al., 1995). Miller et al. (1997) found decreased survival was most closely associated with higher levels of CD4 memory cells. It is interesting to note that a shared hazard for survival in the three species examined in these studies is a diminished capacity of lymphocytes to respond to Con A. The biological mechanism linking robust lymphoproliferative responses with increased survival is not immediately apparent, since in vitro responsiveness to nonspecific activators such as Con A is dependent on a number of factors. Levels of cytokine production and cytokine receptor expression, the relative proportions of CD4 versus CD8 cells, as well as the naïve versus memory cell distributions of these two subsets all play a role in these responses. Identification of the true immune biomarker(s) of aging that is identified by proliferative responses awaits further characterization.
While the benefits of CR on the canine immune system observed in the present study are not as dramatic as those previously described in rodent systems, a significant role for CR in retarding immunosenescence in the dog has been demonstrated. Furthermore, several immune parameters that are both predictive of survival and enhanced by caloric restriction have been identified.

Acknowledgements
This work was supported by Nestle Purina Company, St. Louis, MO 63164. The services of the Flow Cytometry Facility at the University of Illinois are gratefully acknowledged.

References
Cao et al., 2001 S.X. Cao, J.M. Dhahbi, P.L. Mote and S.R. Spindler, Genomic profiling of short- and long-term caloric restriction effects in the liver of aging mice, Proc. Natl. Acad. Sci. USA 98 (2001), pp. 10630-10635. Abstract-Elsevier BIOBASE | Abstract-EMBASE | Abstract-MEDLINE | Full Text via CrossRef | Abstract + References in Scopus | Cited By in Scopus
Chen et al., 1998 J. Chen, C.M. Astle and D.E. Harrison, Delayed immune aging in diet-restricted B6CBAT6 F1 mice is associated with preservation of naïve T cells, J. Gerontol. A Biol. Sci. Med. Sci. 53 (1998), pp. B330-B337. Abstract-EMBASE | Abstract + References in Scopus | Cited By in Scopus
Ferguson et al., 1995 F.G. Ferguson, A. Wikby, P. Maxson, J. Olsson and B. Johansson, Immune parameters in a longitudinal study of a very old population of Swedish people: a comparison between survivors and nonsurvivors, J. Gerontol. 50A (1995), pp. B378-B382. Abstract-MEDLINE | Abstract-EMBASE | Abstract + References in Scopus | Cited By in Scopus
Fernandes et al., 1997 G. Fernandes, J.T. Venkatraman, S. Turturro, V.G. Attwood and R.W. Hart, Effect of food restriction on life span and immune functions in long-lived Fischer-344 x Brown Norway F1 rats, J. Clin. Immunol. 17 (1997), pp. 85-95. Abstract-MEDLINE | Abstract-EMBASE | Full Text via CrossRef | Abstract + References in Scopus | Cited By in Scopus
Fernandes et al., 1976 G. Fernandes, E.J. Yunis and R.A. Good, Influence of diet on survival of mice, Proc. Nat. Acad. Sci. U.S.A. 73 (1976), pp. 1279-1283. Abstract-EMBASE | Abstract-MEDLINE | Full Text via CrossRef | Abstract + References in Scopus | Cited By in Scopus
Friend et al., 1978 P.S. Friend, G. Fernandes, R.A. Good, A.F. Michael and E.J. Yunis, Dietary restrictions early and late: effects on the nephropathy of the NZB X NZW mouse, Lab. Invest. 38 (1978), pp. 629-632. Abstract-EMBASE | Abstract-MEDLINE | Abstract + References in Scopus | Cited By in Scopus
Gerbase-DeLima et al., 1975 M. Gerbase-DeLima, R.K. Liu, K.E. Cheney, R. Mickey and R.L. Walford, Immune function and survival in a long-lived mouse strain subjected to undernutrition, Gerontologia 21 (1975), pp. 184-202. Abstract-MEDLINE | Abstract + References in Scopus | Cited By in Scopus
Gilman-Sachs et al., 1991 A. Gilman-Sachs, Y.B. Kim, M. Pollard and D.L. Snyder, Influence of aging, environmental antigens, and dietary restriction on expression of lymphocyte subsets in germ-free and conventional Lobund-Wistar rats, J. Gerontol. 46 (1991), pp. B101-B106. Abstract-MEDLINE
Goonewardene and Murasko, 1995 M. Goonewardene and D.M. Murasko, Age-associated changes in mitogen-induced lymphoproliferation and lymphokine production in the long-lived Brown-Norway rat: effect of caloric restriction, Mech. Ageing Dev. 83 (1995), pp. 103-116.
Greeley et al., 2001 E.H. Greeley, J.M. Ballam, J.M. Harrison, R.D. Kealy, D.F. Lawler and M. Segre, The influence of age and gender on the immune system: a longitudinal study in Labrador Retriever dogs, Vet. Immunol. Immunopathol. 82 (2001), pp. 57-71. SummaryPlus | Full Text + Links | PDF (149 K) | Abstract + References in Scopus | Cited By in Scopus
Greeley et al., 1996 E.H. Greeley, R.D. Kealy, J.M. Ballam, D.F. Lawler and M. Segre, The influence of age on the canine immune system, Vet. Immunol. Immunopathol. 55 (1996), pp. 1-10. SummaryPlus | Full Text + Links | PDF (716 K) | Abstract + References in Scopus | Cited By in Scopus
Heller et al., 1998 D.A. Heller, F.M. Ahern, J.T. Stout and G.E. McClearn, Mortality and biomarkers of aging in heterogeneous stock (HS) mice, J. Gerontol. 53A (1998), pp. B217-B230. Abstract-MEDLINE | Abstract-EMBASE | Abstract + References in Scopus | Cited By in Scopus
Ingram et al., 1990 D.K. Ingram, R.G. Cutler, R. Weindruch, D.M. Renquist, J.J. Knapka, M. April, C.T. Belcher, M.A. Clark, C.D. Hatcherson, B.M. Marriott and G.S. Roth, Dietary restriction and aging: the initiation of a primate study, J. Gerontol. 45 (1990), pp. B148-B163. Abstract-MEDLINE | Abstract-EMBASE | Abstract + References in Scopus | Cited By in Scopus
Kealy et al., 1992 R.D. Kealy, S.E. Olsson, K.L. Monti, D.F. Lawler, D.N. Biery, R.W. Helms, G. Lust and G.K. Smith, Effects of limited food consumption on the incidence of hip dysplasia in growing dogs, J. Am. Vet. Med. Assoc. 201 (1992), pp. 857-863. Abstract-MEDLINE | Abstract + References in Scopus | Cited By in Scopus
Kealy et al., 1997 R.D. Kealy, D.F. Lawler, J.M. Ballam, G. Lust, G.K. Smith, D.N. Biery and S.E. Olsson, Five-year longitudinal study on limited food consumption and development of osteoarthritis in coxofemoral joints of dogs, J. Am. Vet. Med. Assoc. 210 (1997), pp. 222-225. Abstract-MEDLINE | Abstract + References in Scopus | Cited By in Scopus
Kealy et al., 2002 R.D. Kealy, D.F. Lawler, J.M. Ballam, S.L. Mantz, D.N. Biery, E.H. Greeley, G. Lust, M. Segre, G.K. Smith and H.D. Stowe, Effects of diet restriction on life span and age-related changes in dogs, J. Am. Vet. Med. Assoc. 220 (2002), pp. 1315-1320. Abstract-MEDLINE | Full Text via CrossRef | Abstract + References in Scopus | Cited By in Scopus
Kemnitz et al., 1993 J.W. Kemnitz, R. Weindruch, E.B. Roecker, K. Crawford, P.L. Kaufman and W.B. Ershler, Dietary restriction of adult male rhesus monkeys: design, methodology and preliminary findings, J. Gerontol. 48 (1993), pp. B17-B26. Abstract-EMBASE | Abstract-MEDLINE | Abstract + References in Scopus | Cited By in Scopus
Kim et al., 1997 M. Kim, J.M. Aiken, T. Havighurst, J. Hollander, M.O. Ripple and R. Weindruch, Adult-onset energy restriction of Rhesus monkeys attenuates oxidative stress-induced cytokine expression by peripheral blood mononuclear cells, J. Nutr. 127 (1997), pp. 2293-2301. Abstract-EMBASE | Abstract-MEDLINE | Abstract + References in Scopus | Cited By in Scopus
Lane et al., 1997 M.A. Lane, D.K. Ingram and G.S. Roth, Beyond the rodent model: calorie restriction in Rhesus monkeys, Age 20 (1997), pp. 45-56. Abstract-EMBASE | Full Text via CrossRef | Abstract + References in Scopus | Cited By in Scopus
Lawler et al., 1999 D.F. Lawler, S.D. Johnston, D.G. Keltner, J.M. Ballam, R.D. Kealy, T. Bunte, G. Lust, S.L. Mantz and R.C. Nie, Influence of restricted food intake on estrous cycles and pseudopregnancies in dogs, Am. J. Vet. Res. 60 (1999), pp. 820-825. Abstract-MEDLINE | Abstract + References in Scopus | Cited By in Scopus
Lee et al., 1999 C.K. Lee, R.G. Klopp, R. Weindruch and T.A. Prolla, Gene expression profile of aging and its retardation by caloric restriction, Science 285 (1999), pp. 1390-1393. Abstract-MEDLINE | Abstract-EMBASE | Abstract-Elsevier BIOBASE | Full Text via CrossRef | Abstract + References in Scopus | Cited By in Scopus
Littell et al., 1996 R.C. Littell, G.A. Milliken, W.W. Stroup and R. Wolfinger, SAS System for Mixed Models, SAS Institute, Cary, NC (1996).
Mascarucci et al., 2002 P. Mascarucci, D. Taub, S. Saccani, M.A. Paloma, H. Dawson, G.S. Roth, M.A. Lane and D.K. Ingram, Cytokine responses in young and old Rhesus monkeys: effect of caloric restriction, J. Interferon Cytokine Res. 22 (2002), pp. 565-571. Abstract-Elsevier BIOBASE | Abstract-MEDLINE | Abstract-EMBASE | Abstract + References in Scopus | Cited By in Scopus
Mattison et al., 2003 J.A. Mattison, M.A. Lane, G.S. Roth and D.K. Ingram, Caloric restriction in rhesus monkeys, Exp. Gerontol. 38 (2003), pp. 35-46. SummaryPlus | Full Text + Links | PDF (304 K) | Abstract + References in Scopus | Cited By in Scopus
Miller, 1997 R.A. Miller, Age-related changes in T cell surface markers: a longitudinal analysis in genetically heterogeneous mice, Mech. Ageing Dev. 96 (1997), pp. 181-196. SummaryPlus | Full Text + Links | PDF (486 K) | Abstract + References in Scopus | Cited By in Scopus
Miller et al., 1997 R.A. Miller, C. Chrisp and A. Galecki, CD4 memory T cell levels predict life span in genetically heterogeneous mice, FASEB J. 11 (1997), pp. 775-783. Abstract-EMBASE | Abstract-MEDLINE | Abstract + References in Scopus | Cited By in Scopus
Pahlavani, 2000 M.A. Pahlavani, Caloric restriction and immunosenescence: A current perspective, Front. Biosci. 5 (2000), pp. D580-D587. Abstract-MEDLINE
Pahlavani, 2004 M.A. Pahlavani, Influence of caloric restriction on aging immune system, J. Nutr. Health Aging 8 (2004), pp. 38-47. Abstract-Elsevier BIOBASE | Abstract-EMBASE | Abstract-MEDLINE | Abstract + References in Scopus | Cited By in Scopus
Riley et al., 1989 M.L. Riley, R.J. Turner, P.M. Evans and B.J. Merry, Failure of dietary restriction to influence natural killer activity in old rats, Mech. Ageing Dev. 50 (1989), pp. 81-93. Abstract | Abstract + References | PDF (583 K) | Abstract + References in Scopus | Cited By in Scopus
Roecker et al., 1996 E.B. Roecker, J.W. Kemnitz, W.B. Ershler and R. Weindruch, Reduced immune responses in Rhesus monkeys subjected to dietary restriction, J. Gerontol. Biol. Sci. 51A (1996), pp. B276-B279. Abstract-MEDLINE | Abstract-EMBASE | Abstract + References in Scopus | Cited By in Scopus
Sheldon et al., 1995 W.G. Sheldon, T.J. Bucci, R.W. Hart and A. Turturro, Age-related neoplasia in a lifetime study of ad libitum-fed and food restricted B6C3F1 mice, Toxicol. Pathol. 23 (1995), pp. 458-476. Abstract-EMBASE | Abstract-Elsevier BIOBASE | Abstract-MEDLINE | Abstract + References in Scopus | Cited By in Scopus
Spaulding et al., 1997 C.C. Spaulding, R.L. Walford and R.B. Effros, The accumulation of non-replicative, non-functional, senescent T cells with age is avoided in calorically restricted mice by an enhancement of T cell apoptosis, Mech. Ageing Dev. 93 (1997), pp. 25-33. SummaryPlus | Full Text + Links | PDF (306 K) | Abstract + References in Scopus | Cited By in Scopus
Tian et al., 1995 L. Tian, Q. Cai, R. Bowen and H. Wei, Effects of caloric restriction on age-related oxidative modifications of macromolecules and lymphocyte proliferation in rats, Free Radic. Biol. Med. 19 (1995), pp. 859-865. SummaryPlus | Full Text + Links | PDF (775 K) | Abstract + References in Scopus | Cited By in Scopus
Volk et al., 1994 M.J. Volk, T.D. Pugh, M.J. Kim, C.H. Frith, R.A. Daynes, W.B. Ershler and R. Weindruch, Dietary restriction from middle age attenuates age-associated lymphoma development and interleukin 6 dysregulation in C57BL/6 mice, Cancer Res. 54 (1994), pp. 3054-3061. Abstract-EMBASE | Abstract-MEDLINE | Abstract-Elsevier BIOBASE | Abstract + References in Scopus | Cited By in Scopus
Weindruch, 1992 R. Weindruch, Effect of caloric restriction on age-associated cancers, Exp. Gerontol. 27 (1992), pp. 575-581. Abstract | Abstract + References | PDF (440 K) | Abstract + References in Scopus | Cited By in Scopus
Weindruch et al., 1982a R. Weindruch, S.R. Gottesman and R.L. Walford, Modification of age-related immune decline in mice dietarily restricted from or after midadulthood, Proc. Natl. Acad. Sci. U.S.A. 79 (1982), pp. 898-902. Abstract-MEDLINE | Abstract-EMBASE | Full Text via CrossRef | Abstract + References in Scopus | Cited By in Scopus
Weindruch et al., 1982b R.H. Weindruch, J.A. Kristie, F. Naeim, B.G. Mullen and R.L. Walford, Influence of weaning-initiated dietary restriction on responses to T cell mitogens and on splenic T cell levels in a long-lived F1-hybrid mouse strain, Exp. Gerontol. 17 (1982), pp. 49-64. Abstract | Abstract + References | PDF (1015 K) | Abstract + References in Scopus | Cited By in Scopus
Weindruch et al., 1983 R. Weindruch, B.H. Devens, H.V. Raff and R.L. Walford, Influence of dietary restriction and aging on natural killer cell activity in mice, J. Immunol. 130 (1983), pp. 993-996. Abstract-MEDLINE | Abstract-EMBASE | Abstract + References in Scopus | Cited By in Scopus
Weindruch and Walford, 1988 R. Weindruch and R.L. Walford, The Retardation of Aging and Disease by Dietary Restriction, Charles C. Thomas, Springfield, IL (1988).
Weindruch et al., 1997 R. Weindruch, M.A. Lane, D.K. Ingram, W.B. Ershler and G.S. Roth, Dietary restriction in Rhesus monkeys: lymphopenia and reduced mitogen-induced proliferation in peripheral blood mononuclear cells, Aging Clin. Exp. Res. 9 (1997), pp. 304-308. Abstract-MEDLINE | Abstract + References in Scopus | Cited By in Scopus