Ameliorating the loss of genetic diversity in captive wildlife populations

Abstract

Small wildlife populations, particularly those that are isolated, face a variety of demographic and genetic challenges. Stochastic variations in birth and death rates drive random fluctuations in population size, which leave small populations at higher risk of extinction. Genetic concerns include inbreeding and its associated negative effects on fitness, as well as the increased influence of genetic drift on genetic diversity compared to that of natural selection. Together, these demographic and genetic forces can drive a small population into an ‘extinction vortex’ (Gilpin & Soule, 1986; Fagan & Holmes, 2006). In this issue, Howard et al. (2016) provide the first example of a formal wildlife recovery program that utilized assisted reproductive techniques to ameliorate the loss of gene diversity in a small wildlife population. This work on the black-footed ferret represents a significant achievement and proof of concept, although noteworthy challenges must still be overcome for assisted reproduction methodologies to be more widely applied to wildlife breeding programs. The largest concentration of wildlife breeding programs can be found associated with zoos and their conservation partners. Over 1000 populations are cooperatively managed across regional or global scales; in 2011, the World Association of Zoos and Aquariums (WAZA) published 1540 studbooks, representing 1350 regional and 190 global cooperative species management programs (ISIS/WAZA, 2011). The majority of these programs represent small, largely closed populations that exhibit poor prospects for long-term viability. As the following examples demonstrate, estimates of average population size vary and are strongly influenced by the taxa considered. Analyses of population size for programs managed through the European Association of Zoos and Aquaria (EAZA) indicated that bird programs (N = 91) have an average size of 90 living individuals while mammal programs (N = 177) have an average size of 128 living individuals (Leus et al., 2011). Similar analyses of population size for programs managed in the US through the Association of Zoos and Aquariums (AZA) indicated that, across all taxonomic groups, the median size of populations is 66 living individuals, with 39% of populations comprised of 50 or fewer animals (N = 428; Long, Dorsey & Boyle, 2011). For AZA populations with reasonably complete pedigrees (N = 264), median gene diversity (pedigree-based expected heterozygosity; Lacy, 1995) was reported to be 92% of founding gene diversity, and 38% of populations had a gene diversity below the 90% benchmark suggested by Soulé et al. (1986) as the threshold for increased inbreeding risk (Long et al., 2011). Consider a hypothetical population of 100 randomly mating individuals, with an effective size to actual population size ratio of 0.25 (a commonly reported Ne/N value for cooperatively managed zoo populations; Ballou & Traylor-Holzer, 2011). As a referential benchmark, the decline in that population's gene diversity over time in a randomly breeding population can be estimated as GDt/GD0 = (1 − 1/2Ne)t, where the proportion of founding gene diversity retained (GDt/GD0) at generation t is estimated from its effective size [Ne; (100 × 0.25) = 25]. Thus, in 5 and 10 generations, respectively, 90% and 82% of founding gene diversity is predicted to be retained. Although these predications highly oversimplify the expected rate of gene diversity loss in captive populations, they clearly demonstrate that small, closed populations can be expected to lose notable amounts of gene diversity (>10% of founding diversity; Soulé et al., 1986) over relatively short periods of time. Howard et al. (2016) have empirically demonstrated that artificial insemination, coupled with the use of germplasm banks, can be used to ameliorate the loss of gene diversity in small wildlife populations. Over a 4 year period, the gene diversity returned to the captive black-footed ferret breeding program ranged from approximately 0.15% to 0.20%. Although these numbers seem small, they are not insignificant when one considers that as a population's gene diversity drops to just 93.75% of its founding diversity, animals are related to each other, on average, at the level of first cousins (expected average inbreeding in the next generation, assuming unrelated founders, is F = 1 − GDt/GD0; Templeton & Reed, 1994). Furthermore, the gains in diversity were due to just eight kits, some of which were full siblings, and descendants that were produced over a maximum of only 3 years. The positive genetic impacts of these additions to the population are expected to grow over time as the kits and their descendants continue to reproduce, thereby resulting in more equal representations of their rare lineages (proportional gene diversity is maximized when founder representations are equal; Ballou & Lacy, 1995). As Howard et al. (2016) acknowledge, reproductive technologies have been considered as plausible options for managing the genetic diversity of small wildlife populations for at least 35 years. However, although the authors refer to artificial insemination as the ‘least sophisticated’ of available techniques, they still state that it ‘is hardly simple and is much more complicated than merely randomly placing sperm into a female’. Two of the greatest challenges, as cited by the authors, are remarkable taxonomic differences in reproductive biology and gamete cryopreservation sensitivities. Although the authors give a comprehensive overview of the challenges related to biology and logistics, one area neglected in their discussion is the lack of software and analysis tools, as well as best practices, for establishing germplasm banks. For a given species, how many and which individuals should be banked? New captive breeding programs can bank founders and their close descendants, but the banking of individuals from established programs is a more complicated endeavor (Johnston & Lacy, 1995). When should germplasm banks be utilized? Should population managers consider boosting a program's gene diversity when it falls below a critical threshold (e.g. 90% of founding gene diversity) or should managers wait until obvious negative impacts of inbreeding depression are observed? These questions and more must be better investigated before even the simplest of reproductive technologies – artificial insemination utilizing frozen germplasm banks – can be deployed more widely across hundreds of wildlife breeding programs that are critically in need of new tools for managing genetic diversity. The success by Howard et al. (2016) with the black-footed ferret program is a notable accomplishment; hopefully, it will encourage other wildlife breeding programs to invest in the research and resources necessary to integrate reproductive technologies into genetic management.