Abstract

This study examined the effect of housing density on the longevity-extending and disease-delaying actions of calorie restriction (CR). Singly or multiply housed (four per cage) mice were either fed ad libitum (AL) or were on CR beginning at 2 months. All CR mice were fed 40% less food than were multiply housed AL mice. CR increased median longevity by 19%, and housing density had no effect on this increase. CR also reduced neoplastic lesions in both housing groups, but lymphoma, the most common neoplasm, was reduced more in singly than in multiply housed mice. Singly housed AL mice ate 40% more food than did multiply housed AL mice, but weighed the same and lived as long as multiply housed AL mice. These results indicate that CR can extend life span as effectively in multiply as in singly housed mice, even though housing density can differentially affect the cancer-reducing effect of CR.

THE anticarcinogenic and life-extending effects of calorie restriction (CR) were first identified in laboratory mice and rats in the early 1900s (1–3). Subsequently, studies have shown that CR delays the age-related decline in many physiological processes, and retards and/or suppresses the occurrence of various age-related diseases (4,5). Because of its broad anti-aging effects, CR has been considered to be the “gold standard” for life-span extending interventions in experimental gerontology, yet the underlying mechanisms mediating these effects remain incompletely understood.

Recent studies of long-lived mutant mice, such as Ames and Snell dwarf mice and growth hormone receptor/binding protein (GHR/BP) knockout (KO) mice demonstrated that these long-lived mutant mice exhibit similar physiological changes seen in CR animals. For example, Ames and Snell dwarf mice exhibit small body size; lower insulin-like growth factor-1, insulin, and plasma glucose levels; lower core body temperature; and delayed occurrence of puberty (6,7) [which are physiological characteristics of CR animals (3,8–11)]. Because each of the mutant mouse models mimics one or more presumptive anti-aging traits observed in the CR animals, comparison between the CR and mutant mice enables testing which of the many effects of CR contribute to its broad anti-aging actions.

In the course of our studies comparing the life-history traits of genetically modified mice to those of CR mice, we faced a practical problem. Whereas aging studies of transgenic mice are commonly conducted using mice housed multiply, CR studies have almost always used singly housed animals. The reason for housing animals singly in CR studies is to insure that each CR animal obtains the same amount of food. A widely held concern has been that one or more of the multiply housed animals would dominate the limited food supply in the CR group, resulting in a variable distribution of food (calorie intake) among the cage-mates. The vast majority of studies of the effect of CR on longevity have thus been conducted using singly housed animals. Indeed, in one of the few studies in which the effect of CR on longevity was assessed using multiply housed mice, median life span of C57BL/6 mice was not extended in females and was shortened in males, although another strain did show a modest longevity effect of CR (12,13). No studies have directly compared the effect of CR on longevity and pathology in singly and multiply housed animals. We reasoned that if direct experimentation demonstrated that CR mice housed multiply exhibited the same anti-aging traits as CR mice housed singly, we could more readily compare results of CR studies to those of other environmental and genetic interventions in which mice are more likely to be housed in groups. Moreover, the economic advantage of group housing is that the unit of cost at most academic institutions is usually the cage, rather than the number of animals in the cage. Thus, single housing, for a given sample size, is more costly than multiple housing in which animals can be housed at 4–5 animals per cage.

The purpose of this study therefore was to determine whether multiple housing alters the effects of CR on longevity and age-related diseases in C57BL/6J mice, a strain commonly used for aging research. We report that CR was as effective in extending longevity in multiply as in singly housed mice, but that the cancer-reducing effect of CR, though present in multiply housed mice, was greater in singly housed mice.

Methods

Animal Husbandry

Weanling male C57BL/6J mice (26–30 days of age, n = 126) were received in a single shipment from the Jackson Laboratories (Bar Harbor, ME). Mice were maintained pathogen-free in microisolater units on Tek FRESH ultra laboratory bedding (Harlan Teklad, Madison, WI) under the care and supervision of the technical staff of the Animal Core of the Nathan Shock Center of Excellence in Basic Biology of Aging. Four mice were killed during the first week for monitoring for the possible presence of infectious diseases. All procedures followed the guidelines approved by the Institutional Animal Care and Use Committee at the University of Texas Health Science Center at San Antonio and South Texas Veterans Health Care System, Audie L. Murphy Division, and are consistent with the National Institutes of Health (NIH) Principles for the Utilization and Care of Vertebrate Animal Used in Testing, Research and Education, the Guide for the Care and Use of Laboratory Animals and Animal Welfare Act (National Academy Press, Washington, DC).

Mice were fed a commercial chow (Teklad Diet LM485; Madison, WI) and acidified (pH = 2.6–2.7) filtered reverse osmosis water ad libitum (AL), until 6 weeks of age, when the mice were randomly distributed into four groups: ALM, fed AL and housed four mice per cage; ALS, fed AL and housed one mouse per cage; CRM, fed 60% of the food intake of ALM mice and housed four mice per cage; and CRS, fed 60% of the food intake of ALM mice and housed one mouse per cage. To measure the amount of food consumption, the amount of chow removed from the cage hopper and the spillage (the chow on the bottom of the cage) were weighed weekly. Actual food consumed was calculated by subtracting the spillage from the chow removed from the hopper. The CR mice were given each day 60% of the actual food consumed by the ALM mice approximately 1 hour before the start of the dark phase of a 12-hour light/dark cycle (lights on at 5 am).

All of the mice were weighed biweekly. Peak body weight (maximum midlife body weight) was measured for each mouse. Rate of gain of body weight was calculated as: [(peak body weight – initial body weight)/(age in weeks at peak body weight – age in weeks at initial body weight)]. Therefore, the rate of gain of body weight was expressed as “grams per week.”

Sentinel mice housed in the same room and exposed weekly to bedding collected from the cages of experimental mice were killed on receipt and every 6 months thereafter for monitoring of viral antibodies (Mouse Level II Complete Antibody Profile CARB, Ectro, EDIM, GDVII, LCM, M. Ad-FL, M. Ad-K87, MCMV, MHV, M. pul., MPV, MVM, Polyoma, PVM, Reo, Sendai; BioReliance, Rockville, MD). All tests were negative. All mice were inspected twice daily (between 7 and 8 am and between 3 and 4 pm) during weekdays and once daily (between 07 and 8 am) on weekends and holidays for morbidity and mortality. Date of death was recorded as the outcome measure when the mice died spontaneously. From these data, we determined the median, mean, 10th percentile, and maximum survival for each group. Mice found dead were removed from the cage and immediately necropsied for gross pathological lesions.

Necropsy and Histology

After mice were necropsied, organs and tissues were excised and preserved in 10% buffered formalin: brain, pituitary gland, heart, lung, trachea, thymus, aorta, esophagus, stomach, small intestine, colon, liver, pancreas, spleen, kidneys, urinary bladder, reproductive system (prostate, testes, epididymis, and seminal vesicles), thyroid gland, adrenal glands, parathyroid glands, psoas muscle, knee joint, sternum, and vertebrae. Any other tissue with gross lesions was also excised. The fixed tissues were processed conventionally, embedded in paraffin, sectioned at 5 μm, and stained with hematoxylin-eosin. Although autolysis of varying severity occurred, it did not prevent the histopathological evaluation of lesions, with the exception of three mice. Diagnosis of each histopathological change was made with histological classifications in aging mice described by Bronson and Lipman (14).

Pathological Assessment

A list of pathological lesions was constructed for each mouse that included both neoplastic and non-neoplastic diseases. Based on these histopathological data, the tumor burden, disease burden, and severity of each lesion in each mouse were assessed. The tumor burden was calculated as the sum of the different types of tumor in a mouse. For example, a mouse that had reticular sarcoma and pituitary adenoma had a tumor burden of 2. For this analysis, both fatal and incidental tumors were counted. A mouse that had no neoplastic lesion had a tumor burden of 0. The disease burden was similarly calculated as the sum of the histopathological changes in a mouse. The severity of neoplastic and nephrologic lesions was assessed with the grading system described below.

The percentage of tumor-bearing mice and overall and age-specific incidence of disease were calculated for each experimental group. The percentage of tumor-bearing mice was calculated as the percentage of mice that had one or more neoplastic lesions. For this assessment, all neoplastic lesions were counted, regardless of the severity of tumors, i.e., both incidental (not severe enough to be the cause of death) and fatal (severe enough to be the cause of death) tumors were counted.

Grading of Lesions

The severity of neoplastic lesions and glomerulonephritis were determined using grading systems. Severity of these lesions was determined because their high prevalence allowed a finer dissection of the effect of the treatments on the extent of neoplastic lesions and kidney pathology.

Glomerulonephritis was graded in order of increasing severity: Grade 0: no lesions; Grade 1: minimal change in glomeruli (minimal glomerulosclerosis); Grade 2: Grade 1 with a few (fewer than 10) casts in renal tubules; Grade 3: Grade 1 with more than 10 casts in renal tubules; and Grade 4: Grade 3 with interstitial fibrosis.

Grading of neoplastic lesions was based on a modification of previously reported criteria (15): Grade 1: primary site only; Grade 2: primary site and intra-organ or one other organ metastasis; Grade 3: metastasis to 2–3 organs; and Grade 4: metastasis to more than 4 organs or Grade 3+ additional pathology, e.g., pleural effusion, ascites, subcutaneous edema. Hydrothorax, ascites, and subcutaneous edema were the common complications associated with advanced neoplastic diseases.

Probable Cause of Death

The probable cause of death in each mouse was determined by the severity of diseases found by necropsy, and was assessed independently by two pathologists. For neoplastic diseases, cases which had Grades 3 and 4 lesions were categorized as death by neoplastic lesions. For non-neoplastic diseases, cases which had a severe lesion, e.g., Grade 4, associated with other histopathological changes (pleural effusion, ascites, congestion and edema in lung) were categorized as death by non-neoplastic lesion. In more than 90% of the cases, there was agreement by the two pathologists. In cases in which there was not agreement or in which no disease was considered severe enough, cause of death was categorized as unknown.

Statistical Analysis

The log-rank test (16) was used to determine the homogeneity of survival curves among the four treatment groups (ALM, ALS, CRM, and CRS). The effect of treatment group on life span at 75th percentile, median, 25th percentile, 20th percentile, and 10th percentile was determined by the quantile test (17). The effect of treatment on log-transformed body weight and peak body weight was tested by analysis of variance, adjusting for multiple comparisons with the ad hoc Tukey-Kramer test (18). Associations between the rate of weight gain and the rate of weight loss with life span was tested by Spearman's rank order correlation coefficient (ρ) (18) due to non-normality of life span. The effect of treatment on the frequencies and severities of the presumptive causes of death was determined by chi-square and Cochran-Armitage trend tests, respectively (19). When the expected frequencies were too small for the chi-square test, the data were analyzed with Fisher's exact test (19). The comparison of tumor burden among groups was examined by the Monte Carlo Estimates for Exact test. After controlling for group difference, adjusting for the effect of prognostic factors such as diet, housing, the rate of gain weight, and cause of death, the life span was tested by Cox proportional hazard regression with Efron for tie handling (16).

Results

Food Intake and Body Weight

The effect of housing density on food intake of AL mice is shown in Figure 1. Throughout life, ALS mice consumed more food than did ALM mice (p <.001, analysis of variance). Averaged across the life span, ALS mice consumed approximately 40% more food than did ALM mice (5.0 ± 0.16 g/day vs 3.5 ± 0.09 g/day, respectively). Month-to-month variation in food intake was considerable, but was similar in both housing groups. This variation was probably due to variation in temperature of the colony room, because food intake was inversely related to room temperature (data not shown). CR mice, whether housed singly or multiply, were fed the same amount of food (60% of the intake of ALM mice). Because the ALS mice consumed more food than did the ALM mice, CRS mice, in relation to ALS mice, were proportionately more restricted than were CRM mice in relation to ALM mice. Thus, when compared to their housing density-appropriate control group, CRS mice received only 42% of the food intake of ALS mice, whereas CRM mice received 60% of the food intake of ALM mice.

The mean lifetime body weight profiles of the four experimental groups are shown in Figure 2. Although ALS mice consumed 40% more food than did ALM mice, the body weights of the two groups were virtually identical across the life span until 140 weeks of age, when only 5–6 mice were alive in each group. In contrast to AL mice, CRS mice weighed about 20% less than CRM mice (p <.001, Tukey-Kramer, post hoc test, Table 1), despite the fact that CRS and CRM mice consumed the same amount of food. CR mice in both housing groups weighed significantly less than AL mice (Table 1; p <.001). CRS and CRM mice weighed 50% and 38% less, respectively, than AL mice (p <.001, Tukey-Kramer post hoc test, Table 1).

Although it was not possible to measure directly the food intake of individual mice housed multiply, we did observe the eating behavior of CRM mice when given their daily allotment of food. There was enough room at the hopper for all mice to have simultaneous access to the food and there was no evidence of mice interfering with the eating behavior of each other. However, it remained possible that different rates of eating could result in individual mice eating different amounts of the food. Indeed, as shown in Figure 2, CRM mice had more individual variability in their body weights than did CRS mice.

Longevity

The survival curves for the four experimental groups are presented in Figure 3. The median and various percentile life spans of the four experimental groups are presented in Table 2. The survival curves for ALS and ALM mice were nearly superimposable; the median survival and 10th percentile survival statistics of both groups were approximately 930 and 1050 days. Median life spans of CRS and CRM mice were 17% and 22% longer, respectively, than those of their AL counterparts (p <.001). Median survival time of CRM mice was 53 days longer than that of CRS mice, although this difference did not quite reach statistical significance (p =.07). The 10th percentile survival was approximately 1260 days for both multiply and singly housed CR groups, which was 20% and 29% longer than that of the ALS and ALM mice, respectively. Variance of individual life span was unaffected by housing density in AL and CR mice (p >.05; data not shown).

Although we could not measure the relationship between food intake and survival for each animal in the multiply housed groups, we could assess whether individual body weight or rate of weight gain, two measures related to food intake, were associated with differences in survival in the multiply as well as the singly housed mice. Within diet groups, there was no correlation between body weight and longevity in either the AL or CR mice. Moreover, there was no relationship between the rate of weight gain and longevity in the AL mice. However, there was a positive correlation (ρ = 0.40; p <.05) between life span and the rate of weight gain in CRS mice, and a negative correlation between rate of weight gain and life span (ρ = −0.47; p <.01) in CRM mice.

Probable Causes of Death

The probable causes of death in the four groups of mice are shown in Table 3. The causes of death in the ALS and ALM mice were similar. Approximately 70% of the ALS and ALM mice died of neoplastic diseases, about 85% of which were reticular sarcoma type B as described by Dunn (20), also termed, pleomorphic or mixed cell lymphomas (21,22) in C57BL6 mice. Currently, this type of lymphoma is diagnosed as follicular center lymphoma, which consists of the mixture of lymphoid cells, such as follicular center cells, centrocytes, centroblasts, small lymphocytes, lymphoblasts, plasmacytes, and plasmacytoid cells. The tumors originated in liver, spleen, and mesenteric lymph nodes. The presumptively fatal lymphoma was usually associated with multiple organ involvement or other pathological lesions, e.g., pleural effusion, ascites, or severe congestion and edema in lung. Other presumptively fatal neoplastic diseases observed in the AL mice were adenocarcinoma in lung, hemangioma in liver and spleen, and hepatocellular carcinoma. The major presumptively fatal non-neoplastic diseases observed in the AL mice were glomerulonephritis in kidney and thrombus in heart. In the CR groups, CRM mice had significantly lower incidence of fatal neoplasms (35%; p <.05) compared to the AL groups, and the major fatal neoplasm in this group was lymphoma. The incidence of presumptively fatal lymphoma was significantly lower in CRM mice (p <.05) compared to AL mice. Interestingly, occurrence of presumptively fatal neoplastic diseases in the CRS mice was dramatically reduced compared to CRM mice. Only one CRS mouse died presumptively of neoplastic disease. Both CRM and CRS mice had a higher incidence of non-neoplastic diseases compared to AL groups. The major non-neoplastic diseases in the CR mice were acidophilic macrophage pneumonia (AMP) in lung and rectal prolapse. CRS mice had a significantly higher incidence of acidophilic macrophage pneumonia in lung and rectal prolapse compared to the AL mice (p <.05), and CRM mice had a higher incidence of rectal prolapse compared to AL groups. The proportion of mice dying without pathology severe enough to be recorded as a probable cause of death did not differ between AL (11%–13%) and CR (25%–29%) groups. Housing density also did not affect the percentage of mice that died without a probable cause of death.

Severity of Diseases, Tumor and Disease Burdens

The effect of CR and housing density on severity of lymphoma and glomerulonephritis, two of the most common diseases observed in this strain of mice, was examined using the grading system described in the Methods section. As shown in Table 4, there was no difference in the severity of lymphoma among the four groups of mice (p >.05), but the severity of glomerulonephritis was less in the CR groups than in the AL groups (p <.05).

In addition to severity of diseases, tumor and disease burdens were compared. Tumor and disease burdens were defined as the number of different types of neoplastic lesions and total histological changes in a mouse, respectively. These indices of disease can be used as an overall index of age-related accumulation of tissue and cell injury (14). As shown in Table 4, tumor burden and disease burden were significantly higher in the AL groups compared to the CR groups (p <.05). Housing density had no effect on these indices in AL mice, but CRS mice had less tumor burden than did CRM mice (p <.001). Housing density did not affect disease burden in CR mice.

Discussion

The major aim of this study was to determine whether mice housed multiply would show the same life-span extending and disease-delaying effects of CR as mice housed singly. Using C57BL/6J male mice, we found that multiple housing is as effective as single housing in extending longevity. Also, the disease-delaying action of CR was present in multiply housed mice. However, the cancer-reducing effect of CR was more pronounced in singly housed than in multiply housed mice. As discussed below, whether the effect on cancer incidence was due to housing density or to the effectively greater reduction in food intake of the singly housed CR mice cannot be unraveled in this study. An unanticipated finding was that single housing of AL mice markedly increased food intake relative to that of multiply-housed mice. Strikingly, this increase in food intake had no effect on body weight, longevity, or age-related disease. Thus, whereas reducing caloric intake below an AL level extended life span, increasing calorie intake by singly housing AL mice had no life-shortening effect. The significance and implications of these findings are the subject of this discussion.

The finding that multiply-housed mice had the same life-span extension in response to CR as singly housed mice provides further evidence for the robustness of the CR effect. CR has been shown to be effective in extending longevity in numerous strains of laboratory rats and mice under many environmental conditions and diets (4,5). This study adds housing density to the factors that do not interfere with the longevity effect of CR. A limitation of this study is that it was conducted only in one strain and gender of rodents. However, CR-induced life-span extension has also been observed in pair-housed F344 rats (J. F. Nelson, Y. Ikeno, unpublished data) and in multiply housed Glut4 transgenic mice (R. McCarter, personal communication, 2005).

CR studies have usually been conducted with singly housed animals because of the concern that giving a restricted amount of food to multiply housed mice would result in different degrees of food restriction among the individuals. A consequent concern was that the different levels of CR could lead to different and variable effects on longevity. Although the present study could not allay the concern that individuals housed in groups eat different amounts of the restricted allotment of food provided them, it did indicate that the effect of CR on longevity was not affected by this variation. The finding that the rate of body weight gain of multiply housed CR mice was inversely related to longevity while there was a direct relation between these two parameters in singly housed mice indicates that housing density did differentially affect this relational parameter. The finding that the multiply housed CR mice that gained weight more slowly lived the longest is consistent with them ingesting fewer calories than their shorter lived cohabitants who gained weight more rapidly. In contrast, the finding that singly housed mice that gained weight the most rapidly lived longest would suggest that factors that enabled them to store a limited and fixed amount of energy more efficiently enabled them to outlive their counterparts who were less efficient at energy storage. Nevertheless, the fact that neither the median longevity nor its variance differed between the singly and the multiply housed CR groups indicates that, whatever the reason for the different relationships between weight gain and longevity in the two groups, this relationship did not impact the overall longevity-extending effect of CR.

Our survival curves and pathological data once again confirm the broad anti-aging effects of CR, i.e., extension of both median and maximum life spans associated with reduced incidence and retarded occurrence of various age-related diseases. A striking observation was that singly housed CR mice showed a greater reduction in the incidence of presumptively fatal neoplasms and in the severity of neoplasms than did CR mice housed multiply. There are several possible explanations for this difference. First, previous studies have shown a graded and proportionate reduction in induced cancers and the degree of CR (5,23). CRS mice were more restricted in relation to their singly housed AL counterparts than were CRM mice in relation to their AL counterparts. Furthermore, the fact that CRS mice weighed 20% less than CRM mice indicates that CRS mice were unable to store in body mass as many of the calories they consumed as were the CRM mice. Thus, proportionately more of the calories consumed by the CRS mice were expended in metabolism and maintenance of body temperature than were in CRM mice. Rats allowed to exercise voluntarily under CR weigh less than sedentary CR controls and also showed a greater reduction in cancer (24). Given the fact that the CRS mice weighed less than the CRM mice, it is possible that cell division was reduced more in the CRS group or that the partitioning of energy utilization between growth and heat generation favored heat generation more in the CRS group than in the CRM group. Either or both of these differences could plausibly contribute to the greater reduction in neoplastic disease in the CRS group. CR is known to reduce the rate of cell division in many tissues of mice (25,26). It may be possible that this reduction in cell division contributes to the anti-cancer effect of CR by reducing the chance of accumulation of multiple mutations that ultimately could give rise to lymphoma and other neoplastic disease.

Despite their greater reduction in cancer incidence, CRS mice did not live longer than CRM mice. There are at least two other examples of a dissociation between cancer incidence and longevity. Mice heterozygous for a null mutation of manganese superoxide dismutase have a higher incidence of cancer but virtually no difference in longevity, compared to wild-type littermates (27). Another example is the study mentioned earlier of rats allowed to exercise voluntarily under CR (24). The exercising rats weighed less than sedentary CR controls and also showed a greater reduction in cancer, but had no greater extension of life span than did sedentary CR rats. Together, these findings argue against the view that CR mainly extends longevity by reducing cancer incidence—a view that has been raised because cancer is the most predominant age-related disease associated with death in most strains of laboratory mice and rats. One possible explanation for these dissociations between cancer incidence and longevity is that cancer does not play a major role in the broader aging processes that limit longevity. It is noteworthy that CRS mice had a higher incidence of a presumptively fatal pneumonia than did CRM mice, indicating that the reduction of cancer was associated with a greater vulnerability to another life-threatening disease. Further study is needed to determine the basis for the reduced cancer incidence in these experimental designs (i.e., singly housed vs multiply housed and exercised vs sedentary models), and also to determine why reduction in cancer incidence has no effect on longevity.

It is noteworthy that CR did not reduce the incidence of all diseases. Indeed, CR increased the incidence of some pathophysiological changes. Overall, CR mice had a higher incidence of fatal non-neoplastic diseases than did AL mice. CRS mice had an increased incidence of acidophilic macrophage pneumonia, which could explain why, despite their markedly reduced incidence of lymphoma, they had a similar life span to CRM mice. CR mice also had a higher incidence of rectal prolapse than did AL groups. This lesion is usually associated with neoplasm, inflammatory bowel diseases, or irritation of the intestine (28), possibly caused by Helicobacter infection. Mice in our colony with rectal prolapse tested negative for Helicobactersp, and had no other apparent lesions, except a mild inflammatory change in the submucosa of the rectum. Understanding the etiologies behind the increased incidence of specific diseases in CR mice is an untapped and important area of study. One possibility is that the longer life of CR animals unmasks age-related vulnerabilities that AL animals never experience because of their earlier mortality. Alternatively, the altered physiology of CR animals may lead to different vulnerabilities. For example, CR animals exhibit marked changes in adrenocortical function (29–31) and in the insulin/insulin-like growth factor-1 signaling axes (32), which can profoundly influence pathophysiology (33). The most prevalent non-neoplastic disease in C57BL/6 mice is glomerulonephritis. None of the CR mice had presumptively fatal glomerulonephritis, and its severity was significantly less in CR than in AL mice—a result that has been shown repeatedly in rats (34–37).

Of particular note was the finding that ALS and ALM mice had the same longevity and pathology profiles, despite the fact that ALS mice consumed 40% more food than did ALM mice. The increased food intake of ALS mice was comparable in absolute magnitude to that which, when reduced in CR mice, extends longevity and reduces cancer. How can such an increase in food intake not have life-shortening consequences? Increased calorie intake induced by a more palatable diet can lead to insulin insensitivity and diabetes in mice (38), which presumably would be life-shortening. However, such diets result in increased body weight and adiposity. The fact that ALS mice did not weigh more than ALM mice could underlie their protection from an earlier mortality that might be expected on the basis of their increased food intake. Clearly, the increased energy utilized by ALS mice was not stored and most likely was used to generate the heat necessary to maintain a normal body temperature. ALS mice did not have the insulation and protection from heat loss that ALM mice had huddling together in their cages. Holloszy (39) also reported that increased food consumption induced by voluntary exercise or daily exposure to an energy-consuming swim in water below thermoneutrality did not shorten life span and even reduced the incidence of certain diseases in rats. In addition, Lipman and colleagues (40) showed that the increased calorie intake in adult F1 hybrid (Fisher 344 × Brown Norway) rats did not shorten the life span.

Summary

The major findings of this study were that: 1) multiple housing did not block the effect of CR on longevity; 2) singly housed CR mice showed a greater reduction of neoplasms and a higher incidence of acidophilic macrophage pneumonia than did multiply housed mice, but these effects may be more the consequence of the effect of greater food restriction in the singly housed group than in a difference in housing density; and 3) housing density changed neither longevity nor age-related pathology of AL mice despite causing one group to eat 40% more food than the other. Thus, our results indicate that CR can be as effective in group housed as in singly housed mice in terms of its life-span-extending effects.

Decision Editor: James R. Smith, PhD

Figure 1.

Effect of housing density on food intake of ad libitum-fed mice

Figure 1.

Effect of housing density on food intake of ad libitum-fed mice

Figure 2.

Effect of housing density on body weight of ad libitum-fed (AL) and calorie-restricted (CR) mice

Figure 2.

Effect of housing density on body weight of ad libitum-fed (AL) and calorie-restricted (CR) mice

Figure 3.

Effect of housing density on longevity of calorie-restricted (CR) and ad libitum-fed (AL) mice

Figure 3.

Effect of housing density on longevity of calorie-restricted (CR) and ad libitum-fed (AL) mice

Table 1.

Effect of Calorie Restriction and Housing Density Body Weight Averaged Across the Life Span.

Group N Body Weight Post Hoc Testinga 
ALS 28 40.2 ± 6.3 g 
ALM 30 40.3 ± 6.3 g 
CRS 28 20.1 ± 0.7 g 
CRM 32 25.5 ± 3.4 g 
Group N Body Weight Post Hoc Testinga 
ALS 28 40.2 ± 6.3 g 
ALM 30 40.3 ± 6.3 g 
CRS 28 20.1 ± 0.7 g 
CRM 32 25.5 ± 3.4 g 

Notes: Data are expressed as mean ± standard deviation. Log-transformed data revealed an effect of group p <.001.

aGroups sharing same alphabetic character are not different, Tukey-Kramer test.

ALS = ad libitum-fed, housed singly; ALM = ad libitum-fed, housed four per cage; CRS = calorie-restricted, housed singly; CRM = calorie-restricted, housed four per cage.

Table 2.

Effect of Housing Density and Calorie Restriction on Longevity Statistics.

Group N Median (50th Percentile) (95% CI) 25th Percentile (95% CI) 20th Percentile (95% CI) 10th Percentile (95% CI) 
ALS 28 925 (839, 973) 991 (949, 1034) 1001 (973, 1078) 1034 (1002, 1099) 
ALM 30 935 (861, 983) 990 (944, 1062) 992 (978, 1082) 1062 (992, 1135) 
CRS 28 1080 (957, 1135) 1197 (1088, 1252) 1201 (1135, 1349) 1252 (1209, 1362) 
CRM 32 1133 (931, 1284) 1309 (1225, 1360) 1339 (1277, 1370) 1368 (1326, 1452) 
Group N Median (50th Percentile) (95% CI) 25th Percentile (95% CI) 20th Percentile (95% CI) 10th Percentile (95% CI) 
ALS 28 925 (839, 973) 991 (949, 1034) 1001 (973, 1078) 1034 (1002, 1099) 
ALM 30 935 (861, 983) 990 (944, 1062) 992 (978, 1082) 1062 (992, 1135) 
CRS 28 1080 (957, 1135) 1197 (1088, 1252) 1201 (1135, 1349) 1252 (1209, 1362) 
CRM 32 1133 (931, 1284) 1309 (1225, 1360) 1339 (1277, 1370) 1368 (1326, 1452) 

Notes: Data are expressed as percentiles with 95% confidence intervals (CI).

ALS = ad libitum-fed, housed singly; ALM = ad libitum-fed, housed four per cage; CRS = calorie-restricted, housed singly; CRM = calorie-restricted, housed four per cage.

Table 3.

Effect of Housing Density and Calorie Restriction on Probable Cause of Death.

Cause of Death ALM ALS CRM CRS 
Neoplasm—all 21 (70%) 19 (68%) 11 (36%)* 1 (4%)** 
    Lymphoma 15 14 8* 
    Lymphoma + Other tumors 
    Hemangioma 
    Other 
Non-neoplasm—all 3 (10%) 3 (11%) 10 (32%) 20 (71%) 
    Thrombus 
    Glomerulonephritis 
    Rectal prolapse 5a* 
    AMP 10* 
    Rectal prolapse + AMP 4a* 
    Other 
Neoplasm and non-Neoplasm 1 (3%) 2 (7%) 0 (0%) 0 (0%) 
    Lymphoma + Glomerulonephritis 
Undetermined 4 (13%) 3 (11%) 9 (29%) 7 (25%) 
Autolyzed 1 (3%) 1 (4%) 1 (3%) 0 (0%) 
Total 30 28 31 28 
Cause of Death ALM ALS CRM CRS 
Neoplasm—all 21 (70%) 19 (68%) 11 (36%)* 1 (4%)** 
    Lymphoma 15 14 8* 
    Lymphoma + Other tumors 
    Hemangioma 
    Other 
Non-neoplasm—all 3 (10%) 3 (11%) 10 (32%) 20 (71%) 
    Thrombus 
    Glomerulonephritis 
    Rectal prolapse 5a* 
    AMP 10* 
    Rectal prolapse + AMP 4a* 
    Other 
Neoplasm and non-Neoplasm 1 (3%) 2 (7%) 0 (0%) 0 (0%) 
    Lymphoma + Glomerulonephritis 
Undetermined 4 (13%) 3 (11%) 9 (29%) 7 (25%) 
Autolyzed 1 (3%) 1 (4%) 1 (3%) 0 (0%) 
Total 30 28 31 28 

Notes: *The value was significantly different (p <.05) from AL groups (ALM and ALS).

**The value was significantly different (p <.05) from AL groups (ALM and ALS) and CRM.

ALS = ad libitum-fed, housed singly; ALM = ad libitum-fed, housed four per cage; CRS = calorie-restricted, housed singly; CRM = calorie-restricted, housed four per cage; AMP = acidophilic macrophage pneumonia.

Table 4.

Effect of Housing Density and Calorie Restriction on Severity of Diseases, and Tumor and Disease Burdens.

 Severity
 
 Burden
 
 
Group Lymphoma Glomerulonephritis Tumor Disease 
ALS 2.37 ± 0.21 2.03 ± 0.18 1.55 ± 0.14 3.79 ± 0.17 
ALM 2.47 ± 0.21 2.42 ± 0.20 1.53 ± 0.15 3.77 ± 0.16 
CRS 2.00 ± 1.00 1.31 ± 0.17* 0.19 ± 0.07* 2.61 ± 0.13* 
CRM 2.25 ± 0.40 1.28 ± 0.21* 0.72 ± 0.12** 3.00 ± 0.19* 
 Severity
 
 Burden
 
 
Group Lymphoma Glomerulonephritis Tumor Disease 
ALS 2.37 ± 0.21 2.03 ± 0.18 1.55 ± 0.14 3.79 ± 0.17 
ALM 2.47 ± 0.21 2.42 ± 0.20 1.53 ± 0.15 3.77 ± 0.16 
CRS 2.00 ± 1.00 1.31 ± 0.17* 0.19 ± 0.07* 2.61 ± 0.13* 
CRM 2.25 ± 0.40 1.28 ± 0.21* 0.72 ± 0.12** 3.00 ± 0.19* 

Notes: See Methods section for explanation of numerical scoring of disease severity and burden, both of which are greater with a higher value.

The values represent the mean ± standard error of the mean.

*The value was significantly different (p <.05) from AL groups (ALM and ALS).

**The value was significantly different (p <.05) from AL groups (ALM and ALS) and CRM.

Tumor burden: sum of the different types tumor in a mouse.

Disease burden: sum of the histopathological changes in a mouse.

ALS = ad libitum-fed, housed singly; ALM = ad libitum-fed, housed four per cage; CRS = calorie-restricted, housed singly; CRM = calorie-restricted, housed four per cage.

This research was supported by the San Antonio Nathan Shock Aging Center grant 1P30-AG13319 and by National Institutes of Health grant AG19899.

We thank Corinne Price for her excellent editorial work.

References

1
Osborne TB, Mendel LB. The resumption of growth after a long continued failure of growth.
J Biol Chem.
 
1915
;
23
:
435
-454.
2
Osborne TB, Mendel LB, Ferry EL. The effect of retardation of growth upon the breeding period and duration of life in rats.
Science.
 
1917
;
45
:
294
-295.
3
McCay CM, Crowell MF, Maynard LA. The effect of retarded growth upon the length of life and upon the ultimate body size.
J Nutrition.
 
1935
;
10
:
63
-79.
4
Masoro EJ. Food restriction in rodents: an evaluation of its role in the study of aging.
J Gerontol Biol Sci.
 
1988
;
43
:
B57
-B64.
5
Weindruch R, Walford RL. The Retardation of Aging and Disease by Dietary Restriction. Springfield, IL: Charles C. Thomas; 1988.
6
Bartke A, Coschigano K, Kopchick J, et al. Genes that prolong life: relationships of growth hormone and growth to aging and life span.
J Gerontol Biol Sci.
 
2001
;
56A
:
B340
-B349.
7
Miller RA. Kleemeier Award Lecture: are there genes for aging?
J Gerontol Biol Sci.
 
1999
;
54A
:
B297
-B307.
8
Breese CR, Ingram RL, Sonntag WE. Influence of age and long-term dietary restriction on plasma insulin-like growth factor-1 (IGF-1), IGF-1 gene expression and IGF-1 binding proteins.
J Gerontol.
 
1991
;
46
:
B180
-B187.
9
Duffy PH, Feuers R, Nakamura KD, Leakey J, Hart RW. Effect of chronic caloric restriction on synchronization of various physiological measures in old female Fischer 344 rats.
Chronobiol Int.
 
1990
;
7
:
113
-124.
10
Duffy PH, Feuers R, Leakey J, Hart RW. Chronic caloric restriction in old female mice: changes in circadian rhythms of physiological and behavioral variables. In: Fishbein L, ed. Biological Effects of Dietary Restriction. Berlin: Springer-Verlag; 1991:245–263.
11
Masoro EJ, McCarter RJ, Katz MS, McMahan CA. Dietary restriction alters characteristics of glucose fuel use.
J Gerontol.
 
1992
;
47
:
B202
-B208.
12
Harrison DE, Archer JR, Astle CM. Effects of food restriction on aging: separation of food intake and adiposity.
Proc Natl Acad Sci U S A.
 
1984
;
81
:
1835
-1838.
13
Harrison DE, Archer JR. Genetic differences in effects of food restriction on aging in mice.
J Nutr.
 
1987
;
117
:
376
-382.
14
Bronson RT, Lipman RD. Reduction in rate of occurrence of age related lesions in dietary restricted laboratory mice.
Growth Dev Aging.
 
1991
;
55
:
169
-184.
15
Ikeno Y, Bronson RT, Hubbard GB, Lee S, Bartke A. The delayed occurrence of fatal neoplastic diseases in Ames dwarf mice: correlation to the extended longevity.
J Gerontol.
 
2003
;
58A
:
291
-296.
16
Hosmer Jr DW, Lemeshow S. Applied Survival Analysis. New York: John Wiley & Sons; 1999.
17
Conover WJ. Practical Nonparametric Statistics. New York: John Wiley & Sons; 1980.
18
Kirk RE. Experimental Design. Pacific Grove: Brooks/Cole Publishing Co; 1995.
19
Agresti A. Categorical Data Analysis. New York: John Wiley & Sons; 1990.
20
Dunn TB. Normal and pathologic anatomy of the reticular tissue in laboratory mice.
Natl Cancer Inst Monograph.
 
1954
;
14
:
1281
-1433.
21
Frith CH, Wiley LD. Morphologic classification and correlation of incidence of hyperplastic and neoplastic hematopoietic lesions in mice and age.
J Gerontol.
 
1981
;
36
:
534
-545.
22
Harleman JH, Schuurman HJ, Kuper CF. Carcinogenesis of the hematopoietic system. In: Waalkes MP, Ward JM, eds. Carcinogenesis. New York: Raven Press; 1994:403–428.
23
Ruggeri BA, Klurfeld DM, Kritchevsky D, Furlanetto RW. Caloric restriction and 7,12-dimethylbenz(a)anthracene-induced mammary tumor growth in rats: alterations in circulating insulin, insulin-like growth factors I and II, and epidermal growth factor.
Cancer Res.
 
1989
;
49
:
4130
-4134.
24
McCarter RJM, Shimokawa I, Ikeno Y, et al. Physical activity as a factor in the action of dietary restriction on aging: effects in Fischer 344 rats.
Aging Clin Exp Res.
 
1997
;
9
:
73
-79.
25
Lok F, Scott FW, Margeau R, Malcolm S, Clayson DB. Caloric restriction and cellular proliferation in various tissues of the female Swiss Webster mouse.
Cancer Letts.
 
1990
;
51
:
67
-73.
26
Wolf NS, Penn PE, Jiang D, Fei RC, Pendergrass WR. Caloric restriction: conservation of in vivo cellular replicative capacity accompanies life span extension in mice.
Exp Cell Res.
 
1995
;
217
:
317
-323.
27
Van Remmen H, Ikeno Y, Hamilton M, et al. Increased oxidative damage to DNA increases cancer but does not accelerate aging.
Physiol Genomics.
 
2003
;
16
:
29
-37.
28
Shackelford CC, Elwell MR. Small and large intestine, and mesentery. In: Maronport RR, ed. Pathology of the Mouse. Vienna, IL: Cache River Press; 1999:81–118.
29
Sabatino F, Masoro EJ, McMahan CA, Kuhn RW. Assessment of the role of the glucocorticoid system in aging processes and in the action of food restriction.
J Gerontol.
 
1991
;
46
:
B171
-B179.
30
Klebanov S, Diais S, Stavinoha WB, Suh Y, Nelson JF. Hyperadrenocorticism, attenuated inflammation, and the life-prolonging action of food restriction in mice.
J Gerontol A Biol Sci Med Sci.
 
1995
;
50A
:
B79
-B82.
31
Han ES, Evans T, Shu JH, Lee S, Nelson JF. Food restriction enhances endogenous and ACTH-induced plasma elevations of free by not total corticosterone throughout life in rats.
J Gerontol A Biol Sci Med Sci.
 
2001
;
56
:
B391
-B397.
32
Richardson A, Liu F, Adamo ML, Van Remmen H, Nelson JF. The role of insulin and insulin-like growth factor-I in mammalian ageing.
Best Pract Res Clin Endocrinol Metab.
 
2004
;
18
:
393
-406.
33
Mobbs CV, Bray GA, Atkinson RL, et al. Neuroendocrine and pharmacological manipulations to assess how caloric restriction increases life span.
J Gerontol A Biol Sci Med Sci.
 
2001
;
56A
:
34
-44.
34
Bras G, Ross MH. Kidney disease and nutrition in the rats.
Toxicol Pharmacol.
 
1964
;
6
:
246
-262.
35
Yu BP, Masoro EJ, Murata I, Bertrand HA, Lynd FT. Life span study for SPF Fischer 344 male rats fed ad libitum or restricted diets: longevity, growth, lean body mass and disease.
J Gerontol.
 
1982
;
37
:
130
-141.
36
Maeda H, Gleiser CA, Masoro EJ, Murata I, McMahan CA, Yu BP. Nutritional influences on aging of Fischer 344 rats: II. Pathology.
J Gerontol.
 
1985
;
40
:
671
-688.
37
Gumprecht LA, Long CR, Soper KA, Smith PF, Haschek-Hock WM, Keenan KP. The early effect of dietary restriction on the pathogenesis of chronic renal disease in Sprague-Dawley rats at 12 months.
Toxicol Pathol.
 
1993
;
21
:
528
-537.
38
Sone H, Kagawa Y. Pancreatic beta cell senescence contributes to the pathogenesis of type 2 diabetes in high-fat diet-induced diabetic mice.
Diabetologia.
 
2005
;
48
:
58
-67.
39
Holloszy JO. Exercise increases average longevity of female rats despite increased food intake and no growth retardation.
J Gerontol.
 
1993
;
48
:
B97
-B100.
40
Lipman RD, Smith DE, Blumberg JB, Bronson RT. Effects of caloric restriction or augmentation in adult rats: longevity and lesion biomarkers of aging.
Aging Clin Exp Res.
 
1998
;
10
:
463
-470.

Author notes

Departments of 2Cellular & Structural Biology, 3Physiology, 4Pharmacology, and the 5Barshop Center for Longevity and Aging Studies at The University of Texas Health Science Center at San Antonio.