Sci. Aging Knowl. Environ., 19 May 2004
Vol. 2004, Issue 20, p. pe21
[DOI: 10.1126/sageke.2004.20.pe21]


Early Life Predictors of Old-Age Life Expectancy

Brad A. Rikke

The author is in the Institute for Behavioral Genetics, University of Colorado at Boulder, Boulder, CO 80309, USA. E-mail: rikke{at}

Key Words: biomarkers • immune system • mice • T lymphocytes • thyroxine • body weight

In an amazing tour de force, the laboratory of Richard Miller and numerous collaborators have been testing a variety of life-expectancy predictors and conducting genetic mapping studies on more than 1000 mice, following them for their entire life span. These mice are genetic full sibs from a four-way cross between BALB/c, C57BL/6, C3H, and DBA strains. In their most recent publication, Harper et al. (1) report on the utility of combining early-life measures of T cell subsets, body weight, and hormone concentrations to predict life span.

A major observation of this study is that the proportion of CD8 memory T lymphocytes (CD8M) (cells that divide after an initial immune response and retain the memory of the infection, which primes them to respond to a second exposure) in the peripheral blood, relative to the total number of peripheral CD8 T cells, significantly predicts life expectancy when measured at just 8 months of age. This result was first observed in an earlier cohort but was only suggestive because the earlier experiments tested multiple hypotheses (2); the current study replicates this result with high statistical confidence. They also found that the proportion of CD8M cells continues to be a significant predictor of life expectancy when measured at 18 months of age. Considering that the mice have a mean life span of 28 months, the 8-month result would be the equivalent in humans of predicting differences in the life expectancies of people in their early 20s!

The proportion of CD8M cells reflects age-related changes in the immune system that occur as a result of cumulative exposure to antigens and, to a limited extent, thymic involution. These changes lead to a correlated increase (R = 0.5, Pearson correlation coefficient) in the proportion of CD4 memory (CD4M) T cells, which can also predict life span (2). Combining the CD4M measure and four additional T cell subset measures with the CD8M measure at 8 months didn't explain more of the variance in life expectancy than the CD8M measure alone, but did give a more robust and more statistically significant result (1). Previous studies have shown that this ability to predict life span using T cell subsets is not associated with any particular pathology (3). The results are thus consistent with the T cell subset measures being linked very early to the process of aging.

Although statistically significant, the association between the composite T cell subset measure--the value derived from combining the CD8M and CD4M subsets and the four other T cell subsets--and life span is small; the linear regression correlation coefficient was 0.13 and thus explained only 2% of the variance. It's essential to note, however, that the full potential for associating life expectancy and immune system parameters is yet to be explored. Consider also that a pathway with a small effect on life span may still have large synergistic effects. For example, mutations in clk1 and the inhibition of other genes (by RNA interference) that affect mitochondrial function in the worm Caenorhabditis elegans (see Melov Perspective) don't seem to be associated with more than a 20% life span extension (4), but a clk1 mutation in conjunction with a daf-2 mutation (which alone produces a 100% increase in life span) can extend life span by nearly 400% (5).

An additional and slightly better predictor of life span in these mice is body weight measured at early ages (6) (see "The Shrimps Shall Inherit the Earth"). In the current study, Harper et al. (1) show that combining a body weight measure with their composite T cell measure was a significantly better predictor of life span than either measure alone. The two measures appear to be largely independent, additive, and linearly correlated with life span over their entire range of variation. The 28 mice in the lowest quartile for both body weight at 3 months and the composite T cell subset measure at 8 months had a mean life span that was significantly greater (P = 0.001, one-tailed t test) than that of the 37 mice in the highest quartile by nearly 20%, which is not too shabby. Maximum life span was similarly increased. By combining the body weight and composite T cell subset measure with serum concentrations of the thyroid hormone thyroxine (T4) measured at 4 months of age, predictive power was significantly increased a bit further.

Harper et al. (1) have also mapped quantitative trait loci that specify the variation in their composite T cell subset measures at 8 and 18 months of age. Three statistically significant loci have now been reproducibly mapped to chromosomes 4, 12, and 13. These loci demonstrate that a genetic basis underlies part of the T cell subset variation and could thus serve as the foundation for identifying mediators of life expectancy that correlate with this variation.

What are the caveats related to the findings of Harper et al.? One might be that the T cell subset measures are sensitive to environmental variations in the level of antigen exposure. Consequently, the utility of this measure as a predictor of life span might be limited to the relatively clean and controlled environment of a specific-pathogen free facility. On the other hand, the T cell subset measures may be an even better predictor of life expectancy in a dirtier environment. Another caveat is whether the predictors of life span in this study will be useful predictors of life expectancy in other strains or species. The use of a four-way cross suggests that these measures are likely to be more robust than previously identified predictors. A caveat of the genetic mapping is that some loci could affect the number or detectability of the surface markers (used in cell-sorting experiments to define the subsets) per cell. The use of a composite T cell measure mitigates this concern.

Given that life span in these mice is inversely correlated with body weight, and body weight is significantly correlated with food intake [R = 0.8 (8)], one might also wonder whether the lower body weights and more youthful T cell subset measures (7) are associated with longer life span, because of an underlying dietary restriction effect that could result from a small subset of mice that are voluntarily undereating. In our own studies, however, mice that display body weights and food intake that are lower than those of their close counterparts do not have a significantly lower body temperature. Low body temperature is a hallmark of dietary restriction in mammals (especially mice) and would suggest a dietary restriction effect (8, 9). Specifically, if we look just within strain and cohort but combine data across many inbred strains to compare mice that are virtually identical, lower body weight is significantly associated with lower food intake (Fig. 1A; R = 0.4, P = 0.004, one-tailed test). However, lower body weight is not significantly associated with lower body temperature (Fig. 1B; R = –0.1, P = 0.5). Therefore, a significant dietary restriction effect in the Harper et al. findings seems unlikely as well.

View larger version (6K):
[in this window]
[in a new window]
Fig. 1. Scatter plots show that the lower body weights of genetically identical, ad libitum-fed mice are significantly associated with lower food intake (top panel; R = 0.4, P = 0.004, one-tailed test) but not lower body temperature (bottom panel, R = –0.1, P = 0.5). The black line across each graph is the linear regression line. The data are from 105 mice (between 2 and 9 months of age) that represent 29 inbred strains from which we collected all three measures (body weight, food intake, and body temperature) (8, 9). Within each strain and cohort, I've compared the mouse with the lowest body weight to a mouse with higher body weight to give 55 comparisons (typically just one comparison per strain and cohort). The difference in mean daily food intake and body temperature was then calculated for those same comparisons. Two comparisons were excluded as statistical outliers (outside the 99% confidence interval); excluding these did not affect whether the results were statistically significant or not.

A more direct test would be to ask if the body weight measure and the composite T cell subset measure in the Harper et al. data set are correlated, and, if so, whether removing the correlation by regression also removes the association with life span. If not, then the correlation between the two measures, whether due to a dietary restriction effect, early-life aging, or something else, cannot be responsible for the association with life span.

It is reasonable that variation in early life parameters could affect the course of aging. The results of Harper et al. suggest that early life measures of T cell subsets, body weight, and thyroxine are diagnostic of such variation. Understanding why they are diagnostic could thus reveal important clues about the aging process. In this regard and as previously suggested (2), future studies will need to address the extent to which these measures are associated with other age-sensitive traits. It will also be of interest to determine whether additional measures of immune, growth, and hormone status in early life will provide an even stronger association with life expectancy.

May 19, 2004
  1. J. M. Harper, A. T. Galecki, D. T. Burke, R. A. Miller, Body weight, hormones and T-cell subsets as predictors of lifespan in genetically heterogenous mice. Mech. Ageing Dev. 125, 381-390 (2004).[CrossRef][Medline]
  2. R. A. Miller, Biomarkers of aging: prediction of longevity by using age-sensitive T-cell subset determinations in a middle-aged, genetically heterogeneous mouse population. J. Gerontol. A Biol. Sci. Med. Sci. 56, B180-B186 (2001).[Abstract/Free Full Text]
  3. R. A. Miller, C. Chrisp, T cell subset patterns that predict resistance to spontaneous lymphoma, mammary adenocarcinoma, and fibrosarcoma in mice. J. Immunol. 169, 1619-1625 (2002).[Abstract/Free Full Text]
  4. S. S. Lee, R. Y. Lee, A. G. Fraser, R. S. Kamath, J. Ahringer, G. Ruvkun, A systematic RNAi screen identifies a critical role for mitochondria in C. elegans longevity. Nat. Genet. 33, 40-48 (2003).[CrossRef][Medline]
  5. B. Lakowski, S. Hekimi, Determination of life-span in Caenorhabditis elegans by four clock genes. Science 272, 1010-1013 (1996).[Abstract]
  6. R. A. Miller, J. M. Harper, A. Galecki, D. T. Burke, Big mice die young: early life body weight predicts longevity in genetically heterogeneous mice. Aging Cell 1, 22-29 (2002).[CrossRef][Medline]
  7. R. A. Miller, Age-related changes in T cell surface markers: a longitudinal analysis in genetically heterogeneous mice. Mech. Ageing Dev. 96, 181-196 (1997).[CrossRef][Medline]
  8. B. A. Rikke, J. E. Yerg 3rd, M. E. Battaglia, T. R. Nagy, D. B. Allison, T. E. Johnson, Strain variation in the response of body temperature to dietary restriction. Mech. Ageing Dev. 124, 663-678 (2003).[CrossRef][Medline]
  9. B. A. Rikke, J. E. Yerg 3rd, M. E. Battaglia, T. R. Nagy, D. B. Allison, T. E. Johnson, Quantitative trait loci specifying the response of body temperature to dietary restriction. J. Gerontol. A Biol. Sci. Med. Sci. 59, 118-125 (2004).
Citation: B. A. Rikke, Early Life Predictors of Old-Age Life Expectancy. Sci. Aging Knowl. Environ. 2004 (20), pe21 (2004).

Science of Aging Knowledge Environment. ISSN 1539-6150