Environmental change driven by anthropogenic activities is causing unprecedented rates of species extinction, presenting a major threat to global biodiversity [1]. Among vertebrates, all classes have suffered high extinction rates, but amphibians have been most severely impacted. Based on recent estimates, more than one-third of the worlds amphibians are threatened with extinction [2, 3], and almost one-half of the remaining species are in a state of decline [2]. In response to this crisis, the international Amphibian Conservation Action Plan (ACAP), devised in 2005, urged the establishment of captive assurance colonies for threatened species [4, 5]. In accordance with this recommendation, a large number of institutions worldwide have initiated captive breeding programmes for declining and endangered amphibians. However, in almost all cases, breeding attempts have failed due to the inherent difficulties associated with simulating the complex combination of social and environmental factors that trigger amphibians to breed [6, 5]. In reaction to this captive breeding crisis, there has been a growing interest in determining whether threatened amphibians can be propagated, and genetically managed, using assisted reproductive technologies (ART)[5, 7, 8].

One component of ART is the artificial manipulation of reproductive events using exogenous hormones. Specifically, males and females are administered hormones to stimulate the production and release of gametes (spermatozoa and oocytes), which are then used to generate embryos via in vitro fertilisation (IVF), also referred to as artificial fertilisation (AF) [5]. Among anurans (frogs and toads), it has been known for several decades that exogenous gonadotropins, and gonadotropin-releasing hormones, can be used to induce both sperm release (spermiation) and oocyte release (ovulation) [9–12]. Pituitary extracts are an effective source of amphibian gonadotropins and such preparations have been successfully used to induce gamete release in various anuran species, including Bufo arenarum, Rana pipiens, Hyla regilla and Eleutherodactylus coqui[9, 10, 13–16]. However, the use of pituitary extracts is now strongly discouraged due to a high risk of pathogen transmission [5, 7], and because production of pituitary preparations requires the euthanasia of large numbers of reproductively mature anurans [17]. An alternative approach is the use of synthetic hormones, in particular, commercially available analogues of luteinizing-hormone releasing hormone (LHRH) and human chorionic gonadotropin (hCG) [8, 17–19].

Luteinizing-hormone releasing hormone is a hypothalamic hormone that acts by stimulating the anterior pituitary to synthesize and release natural luteinizing hormone (LH), which in turn stimulates gonadal activity. In contrast, hCG acts by mimicking LH, due to identical alpha and shared beta subunits [20], bypassing the hypothalamic-pituitary-gonadal axis to exert a direct influence on the gonads [21]. Evidence that synthetic analogues of LHRH and hCG can successfully induce gamete release has been obtained for a broad diversity of anuran species [7, 17, 18, 22–26], but the efficacy of these hormones has been found to vary considerably. For example, LHRHa is significantly more effective than hCG at stimulating ovulation in Eleutherodactylus coqui[17], but hCG is highly effective at stimulating ovulation in Xenopus laevis[26].

Although there is still much to learn about the relative efficacy of hCG and LHRHa across species, baseline knowledge concerning their potency has permitted IVF to be attempted in a small, but growing number of anuran families, including the bufonidae [8], pipidae [26] and myobatrachidae [27]. Surprisingly, however, almost all anuran IVF studies conducted to date have refrained from artificially fertilizing oocytes using hormonally induced sperm, instead opting to use sperm obtained from testes macerates [13, 27–29]. The benefit of conducting IVF using testes macerates is that sperm can be obtained in high concentrations, and can also be acquired at the exact moment when females begin ovulating, eliminating the need for sperm storage [19]. Given these practical advantages, euthanizing males for the purpose of IVF might be useful in common species, but this approach cannot be justified in endangered species where individual animals are of high genetic value [5, 7].

To date, few studies have attempted IVF in anurans using hormonally induced gametes collected from live animals, and outcomes have been highly variable. For example, Waggener and Carrol [22] achieved 100% fertilisation success in the leptodactylid frogs Lepidobatrachus laevis and L. illanensis[22], but Browne et al. [8] reported a mean fertilisation success of 12.7% in the endangered toad Bufo baxteri. This extreme variance in IVF success, which is probably related to species-specific differences in reproductive mode and physiology, suggests that protocols for combining hormonally induced gametes are not readily transferable between species [19]. Given this inherent level of unpredictability, there is a need to develop techniques for collecting and combining gametes obtained from live individuals. In particular, there is an urgent need to develop these protocols for endangered species [7, 8].

The southern corroboree frog Pseudophryne corroboree, is one of Australia' s most critically endangered frog species [30]. Of the species known to be extant in Australia, P. corroboree was ranked by the Australasian Regional Association of Zoological Parks and Aquaria as the highest priority anuran requiring ex-situ conservation [31]. The distribution of P. corroboree is highly restricted, with the species confined to a linear distance of 51 km within subalpine regions of Koscuiszko National Park [32]. According to recent field surveys, there are currently less than 50 individuals remaining at natural breeding sites, and the species is predicted to go extinct in the wild within the next ten years [33]. Population declines in P. corroboree were first observed in the 1980's, and since 1996 the species has been the focus of an intensive management and recovery programme [33, 34]. Early management of the species concentrated on habitat protection, but more recently efforts have turned towards the establishment of captive assurance populations [33, 35]. At present, multiple populations of P. corroboree are being maintained in zoos and biological institutions throughout Australia, but successful breeding and recruitment in captivity has been limited. Despite this critical situation, there has been no attempt to bolster captive breeding activity using assisted reproductive technologies.

The aim of this study was twofold. First to test the efficacy of using LHRH and hCG to induce spermiation, and LHRH to induce ovulation, in captively reared P. corroboree, and second, to test whether hormonally induced sperm and oocytes can be used to generate embryo's via in-vitro fertilisation (IVF).

All research was conducted in compliance with the Monash University Animal Ethics Committee (AEC), permit number BSCI/2009/27.

Our study showed that hormone treatment induced gamete release in both male and female corroboree frogs, permitting us to test whether IVF can be used to augment captive breeding in this critically endangered species. Spermiation was induced following administration of either hCG or LHRHa, but these hormones were not equally effective. Both hormones led to approximately 80% of males releasing sperm, but there were significant differences in the speed and duration of response, and the number, concentration and proportion of live sperm released. Administration of hCG, led to a more rapid response, but LHRHa induced the release of a significantly higher number and concentration of viable sperm, and over a longer time period. This difference indicates that LHRHa was more effective at inducing spermiation in this species. That hCG and LHRHa elicited different responses was not unexpected because past research in anurans has shown that the relative efficacy of these hormones is highly species specific [5]. It is important to recognise, that as for the vast majority of anuran ART studies conducted to date, our study only tested each hormone at a single dose [5]. Therefore, to more thoroughly evaluate the relative efficacy of each hormone, it will be necessary to establish spermiation responses to hCG and LHRHa across a range of doses. Testing dose response relationships was beyond the scope of the current study, but this research will be incorporated into the recovery plan for the species.

Despite our finding that LHRHa was successful at inducing spermiation, hormone treated males did not release exceptionally high concentrations of sperm. On average, LHRHa treated males released approximately 4.5 × 10³ sperm per millilitre, a value that is several orders of magnitude lower than hormonally induced sperm concentrations (4.0 × 10⁵ to 4 × 10⁷ sperm per millilitre) previously reported for anurans [22, 24, 39]. There may be several explanations why sperm concentrations were comparatively low in P. corroboree. First, the doses used may have been too low, or too high, to induce an optimal response. Even though we used concentrations approximating those found to be effective in a broad range of frog species [18, 19], LHRHa may lack potency in P. corroboree. Another possibility is that males were not in prime physiological condition at the time of hormone treatment. In seasonally reproducing anurans such as P. corroboree, individuals typically coordinate their physiological state with environmental cues, which exert their effect by stimulating gonadotropin-releasing hormone neurons at the apex of the hypothalamus-pituitary-gonad axis [45]. In nature, spermatogenetic activity in P. corroboree commences several months prior to the onset of breeding [46], so if environmental changes that normally take place in spring (e.g. increasing photoperiod) were not suitably replicated in the captive environment, male investment in spermatogenic activity may have been limited prior to treatment. Another explanation for low sperm yield is that P. corroboree is not a species that invests heavily in spermatogenesis. Theoretically, anurans should only experience strong selection for high sperm production if male's either experience a high risk of sperm competition [47], are required to fertilise large egg clutches [48], or have exceptionally high mating rates [49], but all these conditions are absent in P. corroboree[50, 51]. In fact, testes size relative to body size in P. corroboree is amongst the smallest reported in the family Myobatrachidae [46, 51], so assuming that testis size reflects sperm production capacity in anurans [27], male P. corroboree might actually be incapable of producing high sperm yields.

For females, administration of LHRHa stimulated the release of multiple batches of oocytes over a 24 hour period, indicating that hormone treatment was effective at inducing ovulation, and that final egg maturation was asynchronous [52]. However, the average percentage of females responding was low (<30%), and the average clutch size (mean = approx 15 eggs) was at the lower end of the range (16-40 eggs) previously recorded for this species [50]. The suboptimal response might mean that LHRHa lacks potency in P. corroboree females, as was suggested for males (see above). Indeed, the ability of LHRHa to induce ovulation is known to vary considerably between anuran species. For example, doses required to reliably stimulate ovulation in the leptodactylid frog Eleutherodactylus coqui are 25 times higher than needed in the Wyoming toad Bufo baxteri[8, 17]. Quantifying dose response relationships for female P. corroboree, as well as for males (see above), would provide valuable insight into the potency of LHRHa in this species. An alternative explanation is that phenotypic differences between test females influenced their responsiveness to hormone treatment. For example, in a recent study on boreal toads (Bufo boreas boreas), Roth et al. [23] showed that female age, body size and condition were all important factors influencing the efficacy of LHRHa treatment. However, in our study, females were of similar age (4-5 yrs) and there was no relationship between female body size and number of eggs released, so it is unlikely that these factors underpinned the variable responses reported. A more plausible explanation is that unresponsive females failed to ovulate because their oocytes were immature at the time of hormone treatment [42]. This explanation seems likely because post experimental tactile inspection of unresponsive females revealed that most individuals were still carrying small to medium sized oocytes.

In nature, female P. corroboree undergo rapid oocyte growth 4-8 weeks before the onset of breeding [46], so assuming that similar changes take place in captivity, our treatments may not have perfectly coincided with the time when females were undergoing the final stages of oogenesis. If females were treated too early, oocytes might not have been competent for ovulation [53], but if they were treated too late, oocytes may have already commenced reabsorption [46]. Increasing egg yield in P. corroboree may therefore require developing techniques for reliably assessing changes in oocyte growth and development [54]. Furthermore, it may be necessary to artificially accelerate and synchronise ova maturation prior to any attempt to induce ovulation. Past work with anurans has shown that this might be achieved by repeatedly administering females with low-dose injections of gonadotropins. For example, Browne et al.[8] reported that two priming injections of hCG, in combination with LHRHa, significantly increased the percentage of spawning females, the number of oocytes released, and the survival of fertilised eggs in the Wyoming toad Bufo baxteri.

Oocyte maturation in frogs might also be enhanced via in vitro or in vivo treatment with steroid stimulants [53, 55]. The process of oocyte development in anurans has been extensively studied in Xenopus laevis and Rana pipiens and it is well established that the secretion of progesterone from late-stage follicles plays a fundamental role in germinal vesicle breakdown (GVBD) and oocyte maturation [55–63]. Furthermore, there is also experimental evidence to suggest that in vivo administration of progesterone can accelerate oocyte maturation and increase the effectiveness of hormone treatment. Specifically, Browne et al. [42] recently reported that administration of progesterone, in combination with LHRHa, significantly improved the number and quality of hormone-induced oocytes released by the toad Bufo fowleri. Based on these results, incorporating steroid treatment into future ART work with P. corroboree might be a valuable next step towards optimising ovulatory responses in this species.

Importantly, the hormone treatment protocols we employed permitted gametes to be collected and IVF to be attempted, indicating that there is real potential for ART to assist with the captive breeding of P. corroboree. However, the IVF trials resulted in variable levels of fertilisation success. Variable fertilisation success may have occurred because sperm concentrations were suboptimal. In previous anuran ART studies, concentrations of sperm resulting in high fertilisation success have been within the range of 5 × 10⁵ to 1 × 10⁶ sperm per mL [5], but we were restricted to using concentrations of less than 2.8 × 10² sperm per mL, which may have greatly reduced the probability of gamete fusion. Alternatively, variable fertilisation success may have resulted because the osmolality of the fertilisation medium was inappropriate, as has been reported in the Australian myobatractid frog Limnodynastes tasmaniensis[27]. During the study all males were kept well hydrated, so if their spermic urine was too dilute, this may have significantly reduced the osmolality of the fertilisation medium. If so, sperm may have been activated soon after urine collection and lost viability before IVF was attempted. Furthermore, if osmolality was too low this could have resulted in the large egg capsules swelling too rapidly, making it impossible for individual sperm to penetrate an egg cortex. Clearly, further work will be needed to identify the primary cause of low fertilisation success. It may also be necessary to investigate alternative IVF techniques, such as intra-cytoplasmic sperm injection (ICSI). This more sophisticated IVF approach has been trialled in Bufo arenarum[64] and Xenopus laevis[65], and provides a promising solution for achieving high fertilisation success when numbers of spermatozoa are limiting [7].

Of greater immediate concern than suboptimal fertilisation, is the result that all embryos failed during early development. Early embryo failure may have occurred because hormonal induction impaired gamete viability [5]. In most studies testing the feasibility of IVF in anurans, sperm has been obtained from testes macerates, and embryo survival has been high [5, 13, 27–29]. However, in a recent study in which hormonally induced sperm was used to conduct IVF in the endangered toad Bufo baxteri, levels of embryo survival were also low [8]. These results suggest that there might be problems associated with using spermic urine to fertilize hormonally induced oocytes. Potential reasons for this are not obvious, so this is an area that may require research attention. An alternative explanation for embryo failure is that the incubation conditions employed were inappropriate for this species. Terrestrial breeding Pseudophryne species have extremely large and gelatinous eggs whose capsule size and surface area is primarily determined by hydration state [66]. Consequently, the amount of water present during incubation will significantly affect rates of gas exchange, and subsequently, embryo growth and survival [67]. If eggs were kept too hydrated, or were hydrated too early in development, this may have resulted in oxygen limitation within egg capsules, which in turn could have resulted in embryo mortality [67]. However, this seems unlikely because our incubation conditions were similar to those previously used to successfully rear Pseudophryne guentheri embryos (Silla unpublished data). As such, we suspect that there may be a more intrinsic explanation for embryo failure.

In a recent study that used a cross classified breeding design to examine genetic compatibility in Pseudophryne bibronii, a sister species to P. corroboree, we discovered that crosses made between individuals from the same population had high embryo survival, but those made between populations experienced complete early embryo failure (Byrne and Silla, unpublished data). These results indicate that terrestrial toadlets may be susceptible to high levels of genetic incompatibility, as has been reported in other Australian anurans [68, 69]. Critically, the frogs used in our study were all sourced from mixed populations, so there is a real possibility that embryo failure was indeed linked to developmental problems arising from genetic incompatibility. To address this potential problem, it will be necessary to conduct future IVF trials in P. corroboree using individuals derived from the same source populations.

In conclusion, this study demonstrated that exogenous hormones induced spermiation and ovulation in the southern corroboree frog P. corroboree. The hormone-treatment protocols used did not lead to exceptionally large numbers of gametes being released, but did permit the conduction of IVF, which resulted in moderate fertilisation success. Critically, however, all embryos failed during early stages of development. Embryo developmental failure may have occurred either because hormone treatment compromised gamete viability or because the incubation conditions employed were inappropriate. Alternatively, failure may have been linked to genetic incompatibility resulting from crosses being made between frogs sourced from several different populations. Additional work will be required to increase gamete yield for IVF and identify the causation of embryo failure. The findings are an important first step towards developing artificial reproductive technologies for assisting with the captive breeding of Australia's most critically endangered anuran.