INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

AN AEROBIC ENDURANCE exercise test is often used to assess cardiovascular fitness and to estimate the magnitude of the cardiovascular reserve capacity (7, 9, 11). Aerobic capacity of an organism is likely polygenic in origin and is subsequently influenced to varying degrees by environmental factors. Traits such as endurance are often referred to as multifactorial to emphasize their determination by multiple genetic and environmental factors.

Simple additive models of heredity plus environment have been utilized to estimate the genetic contribution to the variance in endurance performance in monozygotic and dizygotic twins. Although the assumptions of twin studies are often not completely fulfilled or verifiable (10), Klissouras (14) estimated that the variability in maximal aerobic power, as measured by treadmill running to exhaustion, is 93.4% genetically determined in unconditioned humans. If indeed there is a large genetic component to a phenotype as fundamental to health and survival as endurance performance, the identification of the genetic substrate that dictates the difference between low- and high-endurance performance would be of major importance.

A starting point for identification of genes that dictate the differences in the phenotypic expression of traits of polygenic origin is to create strains that diverge to the extremes of the trait of interest by selective breeding (18). Selective breeding for a given phenotypic trait is most effectively achieved if 1) a quantitative, convenient, and repeatable measure of the trait can be devised and 2) the measure of the trait is markedly variant in an outbred population.

The purpose of this study was to evaluate the suitability of using treadmill running in rats as a phenotype for selective breeding for low and high aerobic endurance performance. Our results demonstrate that treadmill endurance running in rats has characteristics favorable for selective breeding for this trait. Under the conditions of high intensity of selection, in a relatively small population, the narrow-sense heritability (h²) was 0.39 over the first three generations. Development of these types of models will provide a critical substrate for evaluating the functional genomics of complex traits.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Animals. The starting population was 24 male and 24 female outbred Sprague-Dawley rats (Harlan Sprague Dawley, Indianapolis, IN). Each rat was of different parentage so that endurance performance was not compared between brothers and sisters, which would tend to narrow the genetic variance. The rats were 10 wk old on arrival at our institution and were housed for 1 wk to allow for stabilization from any stress accrued during shipping. Each rat was provided food and water ad libitum and was placed on a 12:12-h light-dark cycle with the light cycle occurring during daytime.

Experimental protocol. The initiation to the treadmill was performed Monday through Friday for 1 wk, and endurance testing was performed Monday through Friday on the subsequent week. The first week consisted of introducing each rat to the treadmill (model Exer-4; Columbus Instruments, Columbus, OH) for gradually increasing durations each day. The goal was to identify those rats that were trainable to the extent of being able to run for 5 min at a velocity of 10 m/min on a slope of 15°. The ability to achieve this minimal level of running ability at least one time constituted the threshold performance necessary to be included in the group that would be evaluated for endurance the following week. This amount of exposure to treadmill running is below that required to produce a measurable change in aerobic capacity (1, 4). All rats not achieving this minimal running ability were excluded from further consideration.

Selective breeding. By the single best day performance criterion, the two lowest- and two highest-performing rats of each gender were selected and paired for mating at each generation. The offspring (2 litters from the high-performance parents and 2 litters from the low-performance parents) were weaned at 30 days after birth. At 11 wk of age the offspring were introduced to the treadmill and subsequently tested for endurance as described above for the original population of rats.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Performance of original population. All 24 of the females from the original population were able to satisfy the minimal running performance to be included in the group that went on for endurance testing. The females ran for an average duration of 27.2 min, which equated to an average distance of 445 ± 18 m on their best day of performance (range = 346-695 m). Eighteen of the twenty-four males were able to meet the minimum criteria for evaluation in the endurance test. The males ran for an average duration run of 21.8 ± 1 min and an average distance of 331 ± 21 m on their best day of performance (range = 149-544 m). The female rats in the original population weighed less (205 ± 1 vs. 302 ± 3 g) and ran both significantly longer (+5.4 min) and further (+114 m) than the males (Table 1).

Figure 1A shows the daily performances for each of the females in the original population. The five daily runs are presented from the worst day of performance (day 5) to the best day of performance (day 1); these 120 runs in 24 female rats were found not to be different from a normal distribution as assessed by the Kolmogorov-Smirnov test (Lilliefors P value = 0.0159) (19). In general, for this group and all others, the relative ranking order of performance for a given rat was consistent. That is, superior performers retained their superior rank and inferior performers retained their inferior rank with each evaluation. Figure 1B shows the daily runs for the male rats ranked from the worst (day 5) to best (day 1) day of performance; these 90 runs in 18 male rats were found not to be different from a normal distribution (Lilliefors P value = 0.0275).

View larger version (35K): [in this window] [in a new window]   Fig. 1.   Each thin solid line connects five daily running performances ranked from lowest (day 5) to highest (day 1) distance run for each female (A) and male (B) rat of the original population. Thick line shows ranked average value within each group. In general, rats retained their relative ranking on each trial. Rats selected as breeders for the start of the low and high lines are enclosed by rectangles.

Figure 2 shows a ranking of the best performance by each rat (males and females combined) from the highest to lowest distance run for the original population and the first generation of selection. The original population contained 42 rats that ran for an average of 396 ± 16 m with a range from 695 to 149 m. The average distance run by the 27 low-performance offspring was 449 ± 24 m (range = 186-721 m), which was not significantly different (P = 0.131) from the performance of the original population. The average distance run by the 25 high-performance offspring was 581 ± 35 m, with a range from 271 to 1,095 m, which was significantly greater than the original population (P < 0.0001) and represents an average improvement of 185 m. Only 3 of the 25 high-performance offspring achieved running distances that were less than the average of the original population. On average, offspring from the low-performance parents were not different from the original population, whereas the high-performance offspring were significantly greater by 46.7%.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Distance run on the best day of performance for each rat for 1) original population (n = 24), 2) high-performance generation 1 offspring (n = 25), and 3) low-performance generation 1 offspring (n = 27). There was no significant difference between the distance run for the original population and the low-performance offspring. The high-performance offspring ran significantly further than the other two groups. Only 3 of the 25 high-performance offspring had running distances that were less than the average of the original population. Females are represented by open symbols and males are represented by closed symbols.

Response to selection. Figure 3 shows the response to selection at each of the three generations. In generation 1 there was a nonsignificant increase in the performance of the low-selected line and a large increase in the high-selected line. Presumably, this improvement in both lines over the original population represents an environmental response (the original population was born and raised for the first 10 wk in the colonies of Harlan Sprague Dawley). Despite this apparent environmental influence, the lines diverged significantly by 29% in the first generation (132 m difference). Response to selection continued in generations 2 and 3 (Table 2). By generation 3, the lines were separated in performance by 70% or different by 271 m of distance run.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Response to selection for endurance performance. At each generation there was a significant divergence between the low- and high-selected lines. By generation 3 the lines were separated by 70%.

Heritability of running endurance. Figure 4 shows the regression of the running performance for the individual offspring for each family on the mean performance of the parents from which those offspring were derived. Running performances for 12 parental pairs (midparent values) and their offspring are represented (i.e., 2 litters each for both high- and low-selected lines across 3 generations). In general, the running performance of the parents was a significant predictor of the performance of the offspring. Approximately 32% of the variance (r = 0.57) in performance in the parents was associated positively with the variance in the performance of the offspring. The regression of offspring performance on midparental performance, which is an estimate of narrow-sense heritability, was 0.39. 

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Endurance performance of individual offspring regressed on parental performance for each of the 12 families (2 lines, 2 mated pairs across 3 generations). Midparent values are the parental average for each family. Approximately 32% of the variation (r = 0.57) in the midparental values was associated positively with variation in the values of the offspring. Narrow-sense heritability, estimated from the regression of offspring values on midparent values, was 0.39.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

This study demonstrates that a phenotype related to aerobic endurance performance has characteristics that are suitable for selective breeding of this trait. Theoretically, artificial selection can alter the mean value of essentially any complex quantitative trait in both directions if the starting population is genetically heterogeneous (11). We used the following general criteria to guide the development of a suitable test of endurance performance: 1) simple to perform, 2) objectively interpretable, 3) gradable on a continuous numerical scale, 4) demonstrating a wide range in magnitude between the low and high values of the phenotype, 5) normally distributed, and 6) requiring relatively inexpensive equipment. Because endurance performance is a highly complex trait, it seemed particularly important to choose a simple yet sufficient test of this trait so that investigator-induced environmental factors are minimized.

Several human twin studies suggest that endurance capacity has substantial heritability. Klissouras (14) estimated the proportional contribution of heredity to the interindividual variance in performing an endurance run to exhaustion in pairs of monozygotic and dizygotic twins; subjects were evaluated at relatively young ages (7-13 yr) because it strengthened the assumption of a shared environment for both types of twins. He estimated that the variability in maximal aerobic power is 93.4% genetically determined. Bouchard et al. (2) estimated that a genetic component accounts for 70% of endurance performance in young adults when measured as the total work performed during a 90-min maximal ergocycle test. Other investigators, however, found only minimal (5) or no inheritable components related to measures of maximal aerobic performance (12). Although monozygotic twins are genetically identical and phenotypic variance between twins must be of environmental origin, twin studies should not be viewed as definitive. A finding of no significant genetic component could simply result from a large variability in the experimental measure of the phenotype of interest, especially for a complex measure such as maximal oxygen consumption (17). On the other hand, the effect of correlated environments would tend to inflate the apparent genetic contribution. For example, monozygotic twins may be treated more similarly by people than dizygotic twins because of their similar appearance.

From our experience with treadmill running in rats, we predicted that the selection process would more readily produce divergence of high-performance rats relative to the low-performance rats. This tenet is based on two ideas. 1) The environment can have an infinite negative influence on performance (i.e., take the distance run to zero). Factors such as subtle differences in housing or daily handling could cause a genetically superior rat to perform below its normal level on a given day. 2) The environment can have only a finite positive influence on endurance performance. These considerations also convinced us to use the single best performance of the five trials as being most exemplary of the endurance capacity associated with a genetic origin. In short, a negative environment could easily degrade a genetically fit rat into a poor performer, but we deemed it unlikely that a positive environment could transform a genetically below-normal performer into a top performing rat that would be selected for breeding.

We also considered whether the age of the rat at the time of testing could influence its performance. At 3 mo of age the separation in performance between the four low-performance and the four high-performance breeders from the original population was 215%. At 6 mo of age, these same rats were retested and the divide in their performance was 169%. Thus the divergence in performance for individual rats appears to be carried across a longer duration.

The absence of significant divergence for the offspring from the selected low-performance parents from the original population could originate from at least two sources of error. First, as described above, we think the accuracy of selection for low-performance rats is substantially less than for high-performance rats. Indeed, one or more of the low-performance parents may not have been as genetically low as our test estimated because of unknown environmental factors that compromised the running performance. This problem is of course more critical if it occurs in small populations and was perhaps amplified in the low-selected population because both litters had the same male parent; one of the low-selected males did not mate successfully. Second, subtle differences in housing conditions may have influenced the average performance for the original population versus the low and high offspring. The original population was born and housed for the first 10 wk at the Harlan Sprague Dawley Laboratories, whereas the low and high offspring were born and raised in the Animal Research Facility at the Medical College of Ohio in Toledo. The contribution from these two potential sources of error are unknown, but the wide divergence in performance between the offspring of the low- versus high-selected lines across three generations nevertheless demonstrates a substantial heritable component to endurance running.

Studies in other animal species also demonstrate that endurance performance has a heritable component that is measurable. Garland and his colleagues (3) found an h² of 0.17 for endurance swimming performance in mice and reported an h² of 0.70 for treadmill endurance running in garter snakes (6). Strict comparisons of heritability between studies are of limited value because single-generation heritability is influenced by nonstandardized factors that include 1) the magnitude of the breeding value and 2) the intensity of selection (16).

An ultimate goal is to develop and utilize selectively bred strains to identify the genetic substrate that dictates the difference between low-endurance performance and high-endurance performance. One view supports the hypothesis that endurance performance is determined largely by two factors: 1) the rate at which oxygen and nutrient substrates can be utilized to produce energy in the form of ATP and 2) the efficiency with which this energy is transferred into work (7, 15). Oxygen utilization is governed by the cardiac output and the ability of peripheral tissues to extract oxygen. Efficiency is the ratio of work performed to total energy expended and is determined by the complex interaction of all factors that translate energy into movement, such as skeletal mechanics, neuromuscular coordination, and heat dissipation. Given the complexity of endurance performance, it is impossible to predict the exact nature of the various individual traits that will be produced by artificial selection. Nevertheless, it seems reasonable to predict that major genetically mediated differences in the ability of the heart to create a cardiac output will be at least one of the factors selected for.

Joyner (13) has developed a more explicit model of endurance running performance that in its simplest form is the product of three physiological factors: 1) the maximal rate at which oxygen and nutrient substrates can be utilized (O_{2 max}) to produce energy in the form of ATP, 2) the percent of O_{2 max} at the threshold for lactate release, and 3) the efficiency of running. Each of these intermediates is polygenic in origin and highly complex in physiological expression. We hypothesize that allelic differences dictating cumulative changes in each of these three intermediate phenotypes will largely account for the magnitude of the difference in performance between selected lines for low- and high-endurance performance.

In summary, these results demonstrate that treadmill running endurance is a trait with a substantial heritable component. Based on a small outbred population of Sprague-Dawley rats, the treadmill running test we devised to assess endurance performance was normally distributed and demonstrated a wide range between low and high values for which to select for endurance performance. Selection was based on the one best day of performance, which seemed most appropriate because acute environmental factors can easily influence performance on any single day. The parents selected from this original population based on either low-endurance performance or high-endurance performance produced offspring that were divergent in performance over three generations. With appropriate divergent strains as a starting point, one will ultimately be able to identify genes that are causative of the difference between low- and high-endurance performance.