HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Yashin, A. I. Articles by Novoseltsev, V. N. Search for Related Content PubMed PubMed Citation Articles by Yashin, A. I. Articles by Novoseltsev, V. N.

Heat Shock Changes the Heterogeneity Distribution in Populations of Caenorhabditis elegans

Does It Tell Us Anything About the Biological Mechanism of Stress Response?

^a Max Planck Institute for Demographic Research, Rostock, Germany
^b Center for Demographic Studies, Duke University, Durham, North Carolina
^c Institute for Behavioral Genetics, Boulder, Colorado
^d Institute of Control Sciences, Russian Academy of Sciences, Moscow

Anatoli I. Yashin, Max Planck Institute for Demographic Research, Doberaner Strasse 114, 18057 Rostock, Germany E-mail: Yashin{at}demogr.mpg.de.

Decision Editor: John A. Faulkner, PhD

	Abstract

Top Abstract Methods Results Discussion References

 
In this paper we analyze survival data of populations of sterilized nematodes, Caenorhabditis elegans, exposed to heat shocks of different duration at the beginning of their adult lives. There are clear hormesis effects after short exposure to heat and clear debilitation effects after long exposure. Intermediate durations result in a mixture of these two effects. In this latter case, the survival curves for the control and experimental populations intersect. We show that observed effects may be explained by using a model of discrete heterogeneity. According to this model, each population of worms in the experiment is a mixture of subcohorts of frail, normal, and robust individuals; exposure to heat changes the initial proportion of worms in the subcohorts (heterogeneity distribution); and these changes depend on the duration of exposure. In other words, exposure to heat does not influence mortality rates (survival functions) in the subcohorts but does cause individuals to move from one subcohort to another. In a biological interpretation of this finding we hypothesize that, when coping with stress, the organisms of worms use several lines of defense. Switching these lines on and off in response to stress in individual organisms generates the spectrum of observed survival effects at the population level. We discuss possible molecular biological mechanisms of stress response and directions for further research.

THE comparison of empirical survival curves obtained in stress experiments with nematode worms Caenorhabditis elegans shows three major effects of varying duration of heat shock on subsequent survival. In the first one, survival in the stress group is higher than that in the control for all ages. This effect often happens as a result of small exposure to heat shock and is called longevity hormesis. The second effect arises as a result of exposure to heat shock of moderate duration. In this case survival functions in the stress group are lower at young and higher at old ages than those in the control group; that is, the survival functions in the stressed and in the control populations intersect. We call this mixed effect incomplete hormesis. The third effect deals with cases in which survival in the stress group is lower than that in the control for all ages. These effects are the results of long exposure to heat shock, and they are referred to as debilitation. Although longevity hormesis has been the subject of many studies ^(1)(2)(3), the mechanism of incomplete hormesis has yet to be described and investigated in detail. To our knowledge there are also no studies that explain how all three effects (hormesis, incomplete hormesis, and debilitation) arise as a result of an increasing stress load. Such studies would provide important insights for better understanding of the mechanisms of individual response to stress and its manifestation at the population level. Molecular biological studies of stress response show that an organism's adaptation to stress includes, among other things, the induction of stress proteins and the activation of antioxidant production mechanisms ^{(4)(5)(6)(7)(8)(9)}.

In this paper we develop a model of stress response in a heterogeneous population. We assume discrete heterogeneity in mortality and perform an estimation of model parameters by using survival data in four basic populations of nematode worms exposed to varying durations of heat stress, that is, in the control, hormesis, incomplete hormesis, and debilitation groups. The results of the analysis allow us to hypothesize that adaptation to heat shock at the level of the worm's organism uses a finite number of strategies. For example, in response to stress, the organism may switch on and off a finite number of defense lines. Different exposures to heat shock generate a difference in the proportions of worms with a respective level of defense. This difference is responsible for a variety of survival curves observed in each stress experiment. We discuss the possibility of using the discrete heterogeneity model for the prediction of stress response.

	Methods

Top Abstract Methods Results Discussion References

 
Biological Methods
Nematodes from the strain TJ1060 [spe-9(hc88ts) I; fer-15(b26ts) II] were raised on solid nematode growth medium (NGM) plates, spread with Escherichia coli, at 25.5°C for 3 days, by which time they developed into sterile but otherwise phenotypically wild-type adults. At 3 days of age, worms were divided into 10 groups and exposed to 35°C heat shocks for periods of 0, 1, 2, 4, 6, 8, 10, 12, 16, or 24 hours (synchronous starts and asynchronous stops). Immediately following the heat shock the animals were permitted to recover for up to 24 hours at 20°C on NGM agar. They were then transferred to liquid survival medium and maintained at 20°C for the remainder of the experiment. Beginning with the fifth day of life, the number of alive and dead worms was counted daily for all groups (Experiment 1). No survivors were observed after 16 and 24 hours of heat shock. Two other experiments were performed with the same strain of worms, at two different times, to replicate the results of Experiment 1. In Experiment 2, 3-day-old sterilized worms were divided into nine groups and exposed to heat shocks at 35°C for periods of 0, 0.5, 1, 2, 3, 4, 6, 8, and 10 hours. In Experiment 3, 3-day-old sterilized worms were divided into 10 groups and exposed to heat shock at 35°C for 0, 0.5, 1, 2, 3, 4, 5, 6, 7, and 8 hours. The initial number of worms in each group and other characteristics of the experimental populations are shown in Table 1 , Table 2 , and Table 3 for Experiments 1, 2, and 3, respectively.

Model of Discrete Heterogeneity
To investigate whether the data collected in the stress experiments may be described by a discrete heterogeneity model, we assumed that the population of worms consists of three subcohorts called "frail," "normal," and "robust." We assumed that mortality rates in these groups are ordered; that is, mortality is the highest in the frail, and it is the lowest in the robust subcohorts. Let

(1)

be the initial (after-stress) proportions of worms in the respective subcohorts in the experimental group subjected to h hours heat shock, respectively. The after-stress survival function in the population of worms is a mixture of three survival functions:

(2)

Here, S_i, (j), i = 1, 2, and 3, are survival functions in respective subcohorts, where 1 codes for frail, 2 codes for normal, and 3 codes for robust subcohorts, respectively, and P_i^h are initial proportions of individuals in each subcohort. In this model we assume that functions S_i(j) (i = 1, 2, and 3) do not depend on the duration of exposure to heat stress. It means that only after-stress proportions P_i^h (i = 1, 2, and 3) determine the shape of survival functions in S^h(j) populations exposed to heat shock.

Estimation Procedure
The model described above was used in a maximum likelihood estimation procedure to simultaneously fit survival data in the four populations of worms, which include the control, hormesis, incomplete hormesis, and debilitation groups. These groups differ in their exposure to heat shock. This procedure was performed for each of three experiments. To specify the likelihood function, we denoted by q_j^h the probability of death on the jth day in the population with h hours of heating, who survived j - 1 day after heating. Then the log likelihood function for one experiment can be expressed in the form

(3)

Here d_j^h is the number of deaths observed on the jth day in the experiment with h hours of heat shock, n_j^h is the number of such worms surviving j - 1 days after heating, and c is a constant term, which does not contain unknown parameters. Values q_j^h are related with survival functions by the relationships q_j^h=1-. Functions in S_i(j) in 2 were represented as

(4)

where q_iy is the probability of death on the yth day in the ith subcohort (i = 1,2,3) who survived y - 1 day after heating. Parameters q_iy and (i = 1, 2, and 3; h = 0, 1, 4, and 6 for the first and third experiment and h = 0, 1, 3, and 6 for the second experiment; y = 0, 1, 2,...,y_max) were estimated in each experiment. Two approaches to describe survival in the subcohorts were used. In the first one, Gompertz's mortality rates in the subcohorts were assumed. In this case

(5)

where µ_i (x) = a_ie^{bi x}, i = 1, 2, and 3. In the second approach, conditional probabilities of death q_iy, i = 1, 2, and 3 were estimated together with parameters P_i, i = 1, 2, and 3. This approach is referred to as semiparametric because it combines a nonparametric estimation of probabilities of death in the subcohorts q_iy, i = 1, 2, and 3 with an estimation of parameters P_i, i = 1, 2, and 3.

	Results

Top Abstract Methods Results Discussion References

 
Empirical Estimates and Hypotheses Testing
We calculated three families of Kaplan–Meier estimates of the survival functions ⁽¹¹⁾ corresponding to populations in three stress experiments. All three families of empirical survival curves have similar patterns of relative survival within each experiment. Because of this similarity and for the sake of brevity, only one family of Kaplan–Meier curves related to Experiment 1 is shown in Fig. 1 (left panel).

View larger version (15K): [in this window] [in a new window]   Figure 1. Kaplan–Meier estimates of survival functions in Experiment 1 (A) and their estimates obtained by using a semiparametric discrete heterogeneity model with three subcohorts (B). Populations were exposed to a 35°C heat shock for varying durations, as shown.

We tested hypotheses postulating hormesis or debilitation effects in populations of nematode worms using data from all three experiments; see Table 1 , Table 2 and Table 3 . One can see from these tables that the life expectancy observed in populations subjected to 0.5, 1, or 2 hours of heat shock is higher than that in the control group (line 2), and that this difference is statistically significant (line 5). This increase in life expectancy is associated with the longevity hormesis effect ^(1)(2)(3). These tables also show that life expectancy in populations subjected to 6, 8, 10, and 12 hours of heat shock is lower than that in the control group (line 2). This decline in life expectancy is associated with the debilitation effect. The p values in line 6 show that this effect is statistically significant.

Empirical Estimates of Conditional Probabilities of Death
We then estimated empirical conditional probabilities of death at jth day of age, given survival to age j - 1 days, using the relationships

(6)

Here d_j^h is the number of death cases observed during the jth day of life, and n_j-1^h is the number of worms alive at the end of the previous day. Fig. 2 (first five panels) shows the graphs of empirical probabilities of death together with their smoothed estimates calculated from data in Experiment 1. The kernel function K=0.75, if 1 and 0 otherwise, with b = 3, was used in the smoothing procedure. The graphs for respective conditional probabilities of death in Experiments 2 and 3 look similar to those shown in Fig. 2, and for this reason they are not shown in this article.

View larger version (33K): [in this window] [in a new window]   Figure 2. Empirical (——) and empirical smoothed (——) conditional probabilities of death observed in a stress experiment with C. elegans exposed to 0, 1, 2, 4, and 6 hours of heating (first five panels). The last panel shows conditional probabilities of death in a mixture of two subcohorts with Gompertz mortality rates: µi = ai exp (bix); a1 = 9.0 x 10-5; b1 = 0.5; a2 = 1.15 x 10-4; b2 = 0.3. The initial proportion of individuals in the first group is p1 = 0.9.

The last panel in Fig. 2 shows a typical pattern produced by a mixture of two populations with different Gompertz mortality rates. This nontraditional shape of empirical probabilities of death allows us to suggest a model of experimental data in which each population of worms is a mixture of three subcohorts with different mortality rates (survival functions). We assume that the heat shock changes the initial (after-stress) proportions of individuals in the subcohorts. The magnitude of such changes depends on the duration of heating.

Analysis of Data for Four Stress Populations With Basic Survival Curves.
For further analysis we selected data corresponding to four basic patterns of survival curve. This includes data on populations exposed to 0, 1, 4, or 6 hours of heat in Experiments 1 and 3, and to 0, 1, 3, and 6 hours in Experiment 2. The empirical survival curves corresponding to different heating regimes are similar in each of the three experiments. Using the maximum likelihood method described above, we estimated parameters of the discrete heterogeneity model in two situations. One is characterized by the Gompertz approximation to mortality rates in each of three subcohorts. The other deals with the more detailed specification of the model described above.

Gompertz Model for Subcohorts
We assumed that the Gompertz parameters describing mortality rates in the frail, normal and robust subcohorts did not depend on the duration of heating within each experiment. They were, however, taken to be different for different experiments. The estimates of the Gompertz parameters a_i, b_i (i = 1, 2, and 3) and initial (after-stress) proportions P_i^h (i = 1, 2, and 3) were then used to construct the estimates of survival functions in populations exposed to heat shocks of different duration by using model ⁽²⁾. We estimated goodness of fit of this model by using the likelihood ratio test. We compared the value of likelihood functions of the Gompertz model (hypothesis H₀) with the saturated model associated with nonparametric (Kaplan–Meier) estimates of four empirical survival functions represented in Fig. 1. Respective p values are .56, .001, and .38 in Experiments 1, 2, and 3, respectively. This means that the Gompertz model does not contradict the data in Experiments 1 and 3. However, this analysis shows that Gompertz approximation does not fit the data in Experiment 2 well.

Semiparametric Model
To improve the goodness of fit we assumed that conditional probabilities of death 0 q_iy 1, i = 1, 2, and 3 and y = 1, 2,... y_max, in three subcohorts (frail, normal, and robust) are arbitrary nonnegative numbers, and that in each subcohort these probabilities do not depend on the duration of heating. These parameters are assumed to be different in Experiments 1, 2, and 3. (Here y_max is the maximum life span observed in the respective experiment.) This case corresponds to the semiparametric model described above. The estimates of q_iy were then used to construct survival functions

(7)

in the subcohorts, which together with estimates of P_i^h were used to construct survival functions S^h(j) by using 2. The estimated survival functions in populations from Experiment 1 are shown in Fig. 1 (right panel). One can see that the survival curves shown in the two panels of Fig. 1 look similar for each regime of heating. The graphs of initial proportions as functions of the duration of heat shocks for Experiment 1 are shown in Fig. 3.

View larger version (21K): [in this window] [in a new window]   Figure 3. Estimated initial proportions of C. elegans in robust, normal, and frail subcohorts as functions of duration of heat shock in the case of a semiparametric discrete heterogeneity model with three subcohorts for Experiment 1.

Fig. 4 shows how the semiparametric model fits survival data in populations exposed to heat shock during 0, 1, 2, 4, and 6 hours in Experiment 1. Note that the survival function for the 2 hours of heating was predicted by the model.

View larger version (27K): [in this window] [in a new window]   Figure 4. Empirical (triangles) and estimated (solid lines) survival functions for populations of C. elegans exposed to 0, 1, 2, 4, and 6 hours of heat shock. The estimates are obtained by using a semiparametric discrete heterogeneity model with three subcohorts. The third panel shows the survival function predicted by the model for a population exposed to 2 hours of heat shock. The survival functions for robust (dashed lines), normal (dotted lines), and frail (dashed–dotted lines) subcohorts are also shown for Experiment 1.

	Discussion

Top Abstract Methods Results Discussion References

What Are the Advantages of Statistical Modeling Compared With Traditional Statistical Methods?
Statistical calculations based on traditional methods are good at fitting the data by using parametric or nonparametric approximations. They are effective in testing hypotheses about the absence or presence of certain effects (e.g., longevity hormesis) in the data. Their strength is robustness and universality. The results obtained with such methods do not depend on somebody's opinion, findings from earlier experiments with the same or similar subjects, or theoretical laws and postulates, which summarize the experience of dealing with the phenomena under study. However, this strength becomes converted into weakness when one wants to use ancillary information and theory in the analysis of data. Standard statistical methods are not flexible enough to take such information into account in the estimation procedures. That is why they have limited applications in testing hypotheses about the mechanisms of observed phenomena. In such situations, statistical modeling is a better choice. For questions about the mechanism-generated experimental data to be addressed the respective hypothesis must be mathematically outlined and included into the model. Then it can be tested by using experimental data.

It May Be Good When Not Only One Model Explains Data
Biologists who are not experienced in statistical modeling often express concern that many models that are complicated enough may describe available data equally well. In our opinion this fact should not discourage those who study the mechanisms of observed phenomena if the model is identifiable. (It means that the model's parameters must be uniquely determined from the model's output.) This property is usually tested either by analysis or by simulation. In our case, simulation studies showed that both the Gompertz and semiparametric models are identifiable. However, it could well happen that several identifiable models fit data equally well. If all such models are based on different but biologically plausible assumptions, then each has the right to exist, and each can pretend to interpret the data. Nothing is wrong with the fact that several hypotheses about the nature of the phenomenon do not contradict observed data. Such a situation just indicates the need for designing new experiments to obtain additional information, which help resolve the paradox. The difference in the structure and indices used in respective models may suggest alternative ideas about new experiment design.

How Can the Variety of Survival Curves Generated in Stress Experiments Be Explained?
In this article we developed a model of discrete heterogeneity in worms' survival that explains the spectrum of survival effects observed in stress experiments with the nematode worm C. elegans by changes in the parameters of heterogeneity distribution. Conclusions concerning the influence of different factors (genetic, environmental, etc.) on life span in such experiments are typically based on a comparison of average life spans (life expectancies) between respective populations. However, such a comparison ignores the shape of the life-span distributions (survival functions), which often contains important additional information about the mechanism of survival. For example, the mean life span in the population exposed to 4 hours of heat shock in Experiment 1 is approximately the same as in the control group. However, the respective survival functions look completely different. Insights concerning this difference can help to clarify the details related to the nature of the effects. We suggested a model that can use information from four basic survival functions to clarify the mechanism of response to stress.

We presented a discrete heterogeneity model with three states in which the effect of heat stress is to cause a transition of individuals from one state to another. Such transitions affect the heterogeneity of the overall population. Of major interest is the fact that increased survival (hormesis) and decreased survival (debilitation) as well as mixed effects (incomplete hormesis) can be seen to result from exposure to stress. Initially only hormetic effects are seen, but at higher doses, increasing evidence for disability is observed. The result may have important biological implications. The transitions between states in response to heat shock may indicate that the adaptation mechanism of a worm's organism consists of several lines of defense. The process of adaptation to stress involves switching these lines on and off. This generates a spectrum of population effects observed in stress experiments.

Hormesis
The term "hormesis," first coined by Southam and Erlich ⁽¹⁴⁾ in 1943, has been used to describe stimulatory responses to low doses of otherwise harmful agents ^{(15)(16)(17)(18)}. Longevity hormesis is associated with improved survival that can be observed at any age ⁽¹⁾. Sacher and Trucco ⁽¹⁹⁾ suggested that hormesis is a function of the state of an organism, not of the stimulus. According to these authors, the improved longevity is not due to real life extension but rather to elimination of some deleterious effects induced by sub-optimal environment.

Another view is that the hormesis effect involves mobilization or activation of reserves always available in the organism. Johnson and colleagues ⁽²⁰⁾ suggested that many species have a latent capability for increased life span that can be induced by relatively simple genetic or environmental manipulations. In the nematode, short exposure to either ionizing radiation ⁽²¹⁾ or elevated temperature ⁽⁷⁾⁽⁸⁾ has been shown to increase longevity. In Drosophila, short treatment at an elevated temperature also leads to decreased mortality in the survival population ^(2)(3)(22).

The molecular basis for increased thermotolerance and longevity after heat shock is presumably multiple. Heat shock can cause a severe inhibition of normal protein synthesis, and thermotolerant cells show an enhanced recovery of protein synthesis following heat shock ⁽⁵⁾. This increase in ability to recover can be induced by other stresses that also induce thermotolerance and heat shock proteins. Best studied is the fact that heat shock produces a rapid and coordinated increase in the expression of a group of heat shock proteins, some of which are essential for increased thermotolerance and survival ^(6)(23).

Different organisms use different heat shock proteins in response to similar levels of stress. For example, hsp70 is very important for tolerance for killing temperatures in mammals and Drosophila; however, it seems play a minor role in yeast and E. coli ⁽⁶⁾. Members of the hsp100 family are very important for surviving extreme stress in yeast ⁽²⁴⁾, but apparently they are not important in Drosophila. Other mechanisms of hormesis may involve activation of antioxidant defense systems (e.g., production of superoxide dismutase and catalase), DNA repair enzymes, stimulation of cytochrome P450 gene expression, and so on. Indeed, even the immune response is an example of hormesis.

Debilitation
Hormesis is frequently confounded by the presence of debilitation, which is a coexisting phenomenon. The manifestation of one or the other in survival experiments is dose dependent. Although the entire molecular biological mechanism involved in heat-induced injury is poorly understood, some key elements of this process are known. The cell membrane is considered a major target. Heat shock increases cellular generation of superoxide O₂^- by various enzymes, including xanthine oxidase, cyclooxygenase, and lypoxygenase, as well as H₂O₂, in proportion to the severity of heat shock ⁽⁴⁾. When the prooxidant load exceeds the ability of antioxidant enzymes to remove these toxic species, cell injury is observed. Subsequent lipid peroxidation results in disruption of the cytoskeleton and calcium accumulation ⁽²⁵⁾. Adenosine triphosphate (ATP) degradation following heat shock is another potential result of oxidant production. The flux of O₂ and H₂O₂, generated by heat shock, induces the synthesis of additional antioxidant enzymes. Inactivation of key enzymes and disruption of DNA binding proteins also result. Cells surviving single hyperthermic exposure are transiently resistant to the next heat shock ⁽⁴⁾.

Thus, considerable evidence links oxidant-induced and heat-induced cytotoxicity. ATP degradation, as well as uncoupling of mitochondrial respiration, may be the source of oxidant production after heat shock. Both heat and O₂^- have similar biochemical effects on cellular calcium levels and the cytoskeleton, suggesting that heat-induced toxicity may be mediated by O₂^-.

Incomplete Hormesis
The mixed pattern of debilitation and hormesis observed in the population exposed to 4 hours of heat shock clearly shows increased heterogeneity. We suggest here that this is an experimental model for the increased heterogeneity experienced in an aging population. Under stressful conditions, the probability of manifestation of hormesis versus debilitation varies not only with dose but also with other controlled and uncontrolled factors. As a result, the same dose of stress may produce debilitative effects in one group of individuals and hormetic effects in another group. The diagrams in Fig. 5 illustrate the situation.

View larger version (32K): [in this window] [in a new window]   Figure 5. The initial distribution of heterogeneity in the control population (Diagram 1), and changes in this distribution in the case of hormesis (Diagram 2), debilitation (Diagram 3), or mixed (incomplete hormesis) effects (Diagram 4).

Stress Resistance and Individual Frailty
The changes in heterogeneity distribution as a result of exposure to heat shock evaluated in our study may reveal a fundamental property of organisms in coping with stress, which was not discussed before. When stress is small, the organism regards it as an alarm signal, which is used to switch on additional defense mechanisms, perhaps in anticipation of larger stresses. When some key elements of the defense mechanism become exhausted or destroyed by a larger exposure to stress, the system is restructured ⁽²⁶⁾ to bring another (less efficient) mechanism into action to cope with continuing stresses of life (e.g., an oxidative stress). The ability of an organism to restructure and continue functioning in the conditions of stress characterizes an important capacity of an organism called "stress resistance." Many experimental studies of aging and longevity using laboratory subjects emphasize the positive correlation between stress resistance and longevity ^{(7)(8)(27)(28)(29)(30)}. It was found that long-lived mutants of nematode worms C. elegans show elevated resistance to many stresses ^{(7)(8)(31)(32)(33)}. The results of our analysis suggest that innate stress resistance differs among individuals in the population, and that this difference may be explained by several lines of defense, activated in worms' organisms in response to stress.

Induced stress resistance plays an important role in the adaptation of an organism to a changing environment. Several evolutionarily important adaptations to stress have also been reported. For example, exposure to high temperature stimulates an increase in recombination frequency during meiosis ⁽³⁴⁾. This results in an increased genetic variability of progeny. It has also been suggested that the appearance of "adaptive mutations" in genes allowing rescue from otherwise lethal conditions ⁽³⁵⁾ is induced by hard-times conditions. According to Foster ⁽³⁵⁾, "the DNA synthesis associated with recombination could be an important source of spontaneous mutation in cells that are not proliferating. The movement of insertion elements can be responsive to environmental conditions" (p. 57).

Some studies are focused on the analysis of genetic aspects of stress resistance. It was found that heritability in stress resistance for Drosophila is approximately 60% ⁽³⁶⁾, which is close to the estimates of heritability in individual frailty (50%) obtained from survival data on twins using the correlated gamma-frailty model ⁽³⁷⁾. This interesting connection between individual stress resistance and individual frailty opens a new avenue for the application of heterogeneity modeling in stress studies.

To make sure that discrete heterogeneity is the key factor in determining the response to stress in survival experiments with C. elgans nematode worms, we need the identification of biological parameters responsible for individual differences in survival. Data on genetic markers or patterns of gene expression obtained for respective populations may help to address this problem.

	Acknowledgments

 
This research was partly supported by National Institutes of Health/National Institute on Aging Grants PO1 AG08761, RO1 AG16219, and KO2 AA00195, by gifts from the Ellison Medical Foundation and the Glenn Foundation for Medical Research (all to T. Johnson), and by a training grant for J. Cypser from the National Institute of Mental Health (MH-16880).

The authors thank Baerbel Splettstoesser and Karl Brehmer for help in preparing this paper for publication and two anonymous reviewers for valuable comments.

Received May 2, 2001

Accepted October 25, 2001

	References

Top Abstract Methods Results Discussion References

 

1. Neafsey PJ, 1990. Longevity hormesis. A review. Mech Aging Dev. 51:1-31.
2. Khazaeli AA, Curtsinger JW, Tatar U, Pletcher SD, 1997. Heat induced longevity extension in Drosophila. I. Heat treatment, mortality and thermotolerance. J Gerontol Biol Sci 52A:B48-B52. [Abstract]
3. Tatar M, Khazaeli A, Curtsinger JW, 1997. Chaperoning extended life. Nature. 390:30[Medline]
4. Loven DP, 1988. A role for reduced oxygen species in heat induced cell killing and the induction of thermotolerance. Med Hypoth. 26:39-50. [Medline]
5. Parsell DA, Lindquist S, 1993. The function of heat-shock proteins in stress tolerance: degradation and reactivation of damaged proteins. Ann Rev Genet. 27:437-496. [Medline]
6. Parsell DA, Lindquist S, 1994. Heat shock proteins and stress tolerance. Morimoto RI, Tissieres A, Georgopoulos C, , ed.The Biology of Heat Shock Protein and Molecular Chaperones 457-494. Cold Spring Harbor, New York.
7. Lithgow GJ, White TM, Hinerfeld DA, Johnson TE, 1994. Thermotolerance of a long-lived mutant of Caenorhabditis elegans. J Gerontol Biol Sci. 49:B270-B276.
8. Lithgow GJ, White TM, Melov S, Johnson TE, 1995. Thermotolerance and extended life span conferred by single-gene mutations and induced by thermal stress. Proc Natl Acad Sci USA. 92:7540-7544. [Abstract/Free Full Text]
9. Parsons P, 1996. The limit to human longevity: an approach through a stress theory of aging. Mech Aging Devel. 87:211-218.
10. Cox DR, Oakes D, 1988. Analysis of Survivival Data Chapman & Hall, London.
11. Kalbfleisch JD, Prentice RL, 198280. The Statistical Analysis of Failure Time Data Wiley, New York. 80.
12. Vaupel JW, Carey JR, Christensen K, Johnson TE, et al. 1998. Biodemographic trajectories of longevity. Science. 280:855-860. [Abstract/Free Full Text]
13. Vaupel JW, Yashin AI, 1985. Heterogeneity's ruses: some surprising effects of selection on population dynamics. Am Statist. 39:176-185. [Medline]
14. Southam CM, Erlich J, 1943. Effects of extract of western red-cedar heartwood on certain wood decaying fungi in culture. Phytopathology 33:517-524.
15. Foran J, 1998. Regulatory implications of hormesis. Hum Exp Toxicol. 17: (8) 441-443. [Free Full Text]
16. Paperiello CJ, 1998. Risk assessment and risk management implications of hormesis. Hum Exp Toxicol. 17: (8) 460-462. [Free Full Text]
17. Calabrese EJ, , ed.Biological Effects of Low Level Exposures to Chemicals and Radiation 199241–52Lewis, Chelsea. 41–52.
18. Calabrese EJ, Baldwin LA, 1998. Hormesis as a biological hypothesis. Environ Health Persp. 106:357-362.
19. Sacher GA, Trucco E, 1962. A theory of improved performance and survival produced by small doses of radiations and other poisons. Shock NW, , ed.Biological Aspects of Aging 244-251. Columbia University Press, New York.
20. Johnson TE, Lithgow GJ, Murakami S, 1996. Hypothesis: interventions that increase the response to stress offer the potential for effective life prolongation and increased health. J Gerontol Biol Sci. 51A:B392-B395. [Abstract]
21. Johnson TE, Hartman PS, 1988. Radiation effects on life span in Caenohabditis elegans. J Gerontol Biol Sci. 43:B137-B141.
22. Smith MJ, 1958. The effects of temperature and of egg-laying on the longevity of Drosophila subobscura. J Exp Biol. 35:832-841. [Abstract]
23. Holbrook NJ, Udelsman R, 1994. Heat shock protein genes expression in response to physiological stress and aging. Morimoto RI, Tissieres A, Georgopoulos C, , ed.The Biology of Heat Shock Protein and Molecular Chaperones 577-593. Cold Spring Harbor, New York.
24. Sanchez Y, Lindquist SL, 1990. HSP104 required for induced thermotolerance. Science 248:1112-1114. [Abstract/Free Full Text]
25. Wong RSL, Dewey WC, 1986. Effect of hyperthermia on DNA synthesis. Anghileri LJ, Robert J, , ed.Hyperthermia and Cancer Treatment. Vol. I 79CRC Press, Boca Raton, FL.
26. Franceschi C, Monti D, Barbier D, 1996. Successful immunosenescence and the remodeling of immune responses with aging. Nephrol Dial Transpl. 11:18-25.
27. Service PM, Hutchinson EW, Mackinley MD, Rose MR, 1985. Resistance to environmental stress in Drosophila melanogaster selected for postponed senescence. Physiolog Zool. 58:380-389.
28. Service PM, 1987. Physiological mechanism of increased stress resistance in Drosophila melanogaster selected for postponed senescence. Physiolog Zool. 60:321-326.
29. Rose MR, Vu LN, Park SU, Graves JL, Jr 1992. Selection on stress resistance increases longevity in Drosophila melanogaster. Exp Gerontol. 27:241-250. [Medline]
30. Vanfleteren JR, Braekman BP, 1999. Mechanism of life span determination in Caenorhabditis elegans. Neurobiol Aging. 20:487-502. [Medline]
31. Larsen PL, 1993. Aging and resistance to oxidative damage in Caenorhabditis elegans. Proc Natl Acad Sci USA 90:8905-8909. [Abstract/Free Full Text]
32. Johnson TE, 1997. Genetic influence on aging. Exp Gerontol. 32:11-22. [Medline]
33. Johnson TE, Bruunsgaard H, 1998. Implications of hormesis for biomedical aging research. Hum Exp Toxicol. 17:263-265. [Abstract/Free Full Text]
34. Hoffman AA, Parsons PA, 1991. Evolutionary Genetics and Environmental Stress Oxford University Press, Oxford.
35. Foster PL, 1999. Mechanisms of stationary phase mutation: a decade of adaptive mutation. Annu Rev Genet. 33:57-88. [Medline]
36. Parsons PA, 1989. Evolutionary rates: effects of stress upon recombination. Biol J Linnean Soc. 35:49-68.
37. Yashin AI, Iachine IA, 1997. How frailty models can be used for evaluating longevity limits: taking advantage of an interdisciplinary approach. Demography. 34: (1) 31-48. [Medline]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Yashin, A. I. Articles by Novoseltsev, V. N. Search for Related Content PubMed PubMed Citation Articles by Yashin, A. I. Articles by Novoseltsev, V. N.