Establishment of the Experiment:

An artificial selection experiment for high voluntary activity on wheels was begun in 1993 at the University of Wisconsin-Madison.  Briefly, 224 mice of the outbred (not inbred!), genetically variable Hsd:ICR strain (description in PDF file) were purchased from Harlan Sprague Dawley.  These founder mice were designated generation -2.  The generation -2 mice were randomly paired and their offspring were generation -1.  The generation -1 mice were again randomly paired and their offspring were then randomly assigned to one of 8 lines, 4 of which would be selectively bred High Runner (HR) lines and four of which would be non-selected Control (C) lines (generation 0).  After testing the generation 0 mice for voluntary wheel-running behavior (see below), offspring were selected in the HR lines or chosen without regard to their running in the C lines.  Thus, the offspring of generation 0 were the first "evolved" generation for the HR lines.

The Selection Protocol:

Each generation, when mice are ~6-9 weeks of age (young adults), they are individually given access to rat-sized wheels (1.1-m in circumference) attached to ordinary home cages for 6 days.  Food and water are provided ad libitum.  The number of revolutions on days 5 and 6 is used as the selection criterion.   Four banks of 50 wheels each allow testing of 200 mice, so it takes three weeks to test one generation.

Within-family selection is used, which means that the highest-running male and female from within each of 10 families in each HR line are selected as breeders.  A similar process occurs in each of C line, except that one male and one female are chosen randomly from each of 10 families.  Within each line, these mice are then randomly paired, except that sibling matings are not allowed.  

Direct Response to Selection:

The HR lines responded rapidly to selection.  The first publication, at generation 10,  reported an average 75% increase in daily running distance in the 4 selected lines, as compared with the 4 control lines.  Realized heritability averaged 0.19 (range = 0.12-0.24 for the HR lines), or 0.28 when adjusted for within-family selection.  Rev/day increased mainly through increases in average running speed, rather than the number of minutes per day of activity (especially for females).

Increased running distance in the High Runner mice is coupled with an increased intermittency of locomotion (they run in shorter, faster bouts: video on YouTube), which may reduce locomotor costs and/or increase endurance capacity.

All of the High Runner lines reached apparent selection limits at generations 17-27, depending on line and sex.  At these limits, mice from the HR lines ran approximately three-fold more than those from the C lines.

By generation 43, "multiple solutions" were clear for the sexes and among the replicate HR lines.   Differences in daily running distance (the trait under selection) were statistically significant among the C lines but not among the HR lines.  In contrast, average speed varied significantly among HR lines, but not among C, and showed a sex-by-linetype interaction, with the HR/C ratio being 2.02 for males and 2.45 for females.  Time spent running varied among both HR and C lines, and showed a sex-by-linetype interaction, with the HR/C ratio being 1.52 for males but only 1.17 for females. Thus, females (speed) and males (speed, but also time) evolved differently, as did the replicate selected lines. 

Also at generation 43, a trade-off between average running speed and duration of daily running had evolved among the HR lines but not among the C lines.  The basis for this trade-off might be related to ability (e.g., variation in muscle physiology) and/or motivation for running, and is currently under investigation.

Correlated Responses to Selection:

A response to selection on a particular trait always occurs in concert with changes in other traits.  For example, if a behavior evolves in response to selection, then changes in subordinate traits within the nervous system must also occur.  And so it is with the HR mice.  In the nervous system, for example, HR mice have evolved larger brains than the C mice, various regions within their brains respond differently to wheel running and to having wheels taken away, and their wheel running is suppressed ("treated") by Ritalin but increased in C mice.  These and other changes indicate differences in the motivation for wheel running and/or the reward received from running (e.g., their "runner's high").

In addition to changes in motivation, the High Runner mice have evolved higher ability for sustained, aerobically supported exercise.  When subjected to forced exercise, they have a higher maximal rate of oxygen consumption (VO2max) and a higher endurance capacity.  Various subordinate traits underlie the differences in organismal performance abilities, including larger hearts, smaller calf muscles, altered enzyme activities in the tibialis anterior muscle, larger joint surfaces, and distinctively shaped semicircular canals in their inner ears.

Aspects of the endocrine system have evolved in the HR mice.  For example, circulating concentrations of the corticosterone, a glucocorticoid hormone, are higher under baseline conditions during the day and higher still when mice are running on wheels at night.

Current Status & Future Directions:

As of June 2024, the experiment is at generation 105 and still running (pun intended).  Current projects include analysis of the genetic divergence between the High Runner and Control lines, skeletal adaptations, changes in the intestinal microbiome, trade-offs in muscle function that may underlie line differences in speed versus duration of daily wheel running, lifetime reproductive success of HR versus C lines, and effects of early-life exposures to diet and/or exercise.

... in progress ...