Assessing reproductive status in elasmobranch fishes using steroid hormones extracted from skeletal muscle tissue

Understanding the life history of elasmobranch fishes (sharks, skates, and rays) is essential for their management. This study investigated the utilization of steroid hormone concentrations in non-lethally harvested elasmobranch skeletal muscle to assess reproductive status.


Introduction
The life history characteristics of many elasmobranchs, such as slow growth and late age at maturity, make these fishes particularly susceptible to overexploitation (Dulvy et al., 2003). As a result, anthropogenic threats, such as direct and indirect commercial fishing, have led to significant population declines in numerous elasmobranch species (Dulvy et al., 2008). To manage elasmobranchs effectively, a comprehensive understanding of their life history characteristics is needed. However, data on these characteristics are lacking for many species (Castro et al., 1999;Walker, 2004;Pinhal et al., 2008;IUCN, 2011). For instance, essential life history characteristics needed for proper management include age/ size at maturity, gestation length, and reproductive cyclicity (Walker, 2004(Walker, , 2005. When this information is incorporated into fisheries models and species assessments, insight can be gained into when, where, and how often populations are reproducing, which can ultimately aid in reducing their decline through the development of management protocols that augment fishery practices (Walker, 2004(Walker, , 2005. For example, the management of blacknose sharks within the USA was directly affected and improved by the inclusion of data detailing regionally distinct reproductive periodicity (NMFS, 2011).
Lethal sampling has historically been considered the most effective approach for collecting life history information from elasmobranchs, particularly because it is quick and provides a comprehensive set of data (Heupel and Simpfendorfer, 2010). More recently, in response to species declines and ethical concerns, elasmobranch research has moved towards conservation-minded sampling approaches Heupel and Simpfendorfer, 2010;Hammerschlag and Sulikowski, 2011).
Among the non-lethal approaches to studying the reproductive biology of elasmobranchs, such as intrauterine endoscopy (Carrier et al., 2003) and ultrasonography (Daly et al., 2007), analysis of plasma hormones is the most widely used (e.g. Kneebone et al., 2007;Sulikowski et al., 2007;Henningsen et al., 2008). In particular, plasma levels of steroid hormones are correlated directly with reproductive events in elasmobranchs, such as the onset of maturity (Gelsleichter et al., 2002;Sulikowski et al., 2006) and reproductive activity (Sulikowski et al., 2004;Kneebone et al., 2007). Although plasma steroid hormone analysis has been conducted on many species of elasmobranchs spanning multiple reproductive modes (e.g. Gelsleichter et al., 2002;Sulikowski et al., 2006;Awruch et al., 2008;Henningsen et al., 2008), obtaining blood can be stressful for the animal because of the significant amount of handling time and prolonged removal from the water (Skomal, 2007). Following the success of extracting steroid hormones from bodily materials other than plasma and correlating concentrations to reproductive events in non-elasmobranch vertebrates (Heppell and Sullivan, 2000;Mansour et al., 2002;Kellar et al., 2006;Barnett et al., 2009), a new non-lethal approach has recently been developed that is potentially less stressful, because it does not require the animal be restrained or removed from the water, and it can be conducted rapidly. Prohaska et al. (2013) investigated the efficacy of extracting steroid hormones from the skeletal muscle tissue of elasmobranchs using two reproductive modes, the yolk-sac placental viviparous Rhizoprionodon terraenovae and the yolk-dependent viviparous Squalus acanthias. Steroid hormones were successfully extracted and quantified from skeletal muscle tissue, but more importantly, fluctuations in these hormones are correlated with gestational stage (Prohaska et al., 2013).
It is critical that effective non-lethal protocols for studying elasmobranch reproductive biology are thoroughly tested and validated prior to the shift towards their strict use in research, similar to what occurred for marine mammals in the USA (NMFS, 1972;Heupel and Simpfendorfer, 2010;Hammerschlag and Sulikowski, 2011). As a result of the diversity of reproductive modes in elasmobranchs, oviparity (egg laying), yolk-dependent viviparity (live birth of yolkdependent embryos), and yolk-sac placental viviparity (live birth of embryos with an initial yolk-sac followed by a placental attachment; Wourms, 1977), coupled with the promising results of Prohaska et al. (2013), the objectives of the present study were as follows: (i) to attain a more comprehensive understanding of muscle steroid hormones in R. terraenovae and S. acanthias; (ii) to determine whether this approach is appropriate for studying the reproductive biology of oviparous elasmobranchs using Leucoraja erinacea; and (iii) to determine whether skeletal muscle steroid hormones can be used as non-lethal indicators of reproductive status in elasmobranchs.

Specimen collection
Female R. terraenovae and S. acanthias were captured using the same methods and in the same locations as previously described by Prohaska et al. (2013). Briefly, R. terraenovae were captured by bottom longline in the northern Gulf of Mexico in an area centred around 88.812°W and 27.887°N, while S. acanthias were captured by bottom trawl and gill net in the US Northwest Atlantic in an area centred around 70.115°W and 42.471°N. The aforementioned fishing methods were conducted for a maximum of 1 h, with blood sampling taking place immediately after capture, to reduce the potential for stress hormones to interfere with sex steroid hormone concentrations. Female L. erinacea were captured by bottom trawl in November 2012 in an area centred around 70.466°W and 42.615°N. After capture, L. erinacea were maintained in an insulated livewell containing ambient surface seawater. To ensure that water quality was maintained, frequent water changes occurred during the ~1.5 h return to the dock. All live-captured L. erinacea were transported from the dock in an aerated 833 l insulated livewell to the University of New England's Marine Science Center (~2 h) and housed in a 3785 l, 2.4 m diameter hexagonal tank with an open flow-through seawater system with a turnover rate of 38 l min −1 . Animal husbandry of L. erinacea followed the protocols of Palm et al. (2011). Additionally, skates were palpated daily for up to 3 weeks to assess presence or absence of egg cases. Immediately prior to obtaining internal morphological data, L. erinacea were killed by lethal pithing.

Sampling
Sampling of R. terraenovae, S. acanthias, and L. erinacea followed the same protocols as those described by Prohaska et al. (2013). Briefly, at the time of sampling, an 8 ml aliquot of blood was collected and stored at 4°C for up to 24 h. Blood was then analysed for haematocrit prior to being centrifuged at 1242 g for 5 min. Plasma was then removed and stored at −20°C until steroid hormone analysis. The following morphological parameters were recorded: mass (in kilograms), fork length (FL; in sharks), disc width (DW; in skates) and natural total length (TL), all of which were measured to the nearest centimetre over a straight line along the axis of the body. Additionally, ovary and oviducal gland mass (in grams), oviducal gland width (in millimetres), and follicle diameter (in millimetres) were recorded. From sharks, embryo sex and stretch total length (STL; in millimetres), measured to the nearest millimetre over a straight line along the axis of the body, were also recorded. A 5 g white skeletal muscle tissue sample was then collected from behind the second dorsal fin from all sharks, and from the mid-point of the right pectoral fin from all skates, and immediately stored at −20°C until analysis (Prohaska et al., 2013).

Plasma steroid hormone extraction
Progesterone (P 4 ), testosterone (T), and 17β-estradiol (E 2 ) were extracted from all plasma samples following the methods of Tsang and Callard (1987) and Sulikowski et al. (2004). Briefly, each plasma sample was extracted twice with 10 volumes (5 ml) of ethyl ether (ACS grade), and the organic phase was evaporated at 37°C under a stream of nitrogen. Extracts were reconstituted in phosphate-buffered saline (PBS) containing 0.1% gelatin. Prior to extraction, each sample was spiked with 1000 counts min −1 of tritiated P 4 , T, or E 2 (Perkin Elmer, Waltham, MA, USA) to account for procedural loss.

Skeletal muscle tissue steroid hormone extraction
The hormones P 4 , T, and E 2 were extracted from all white skeletal muscle tissue samples following the protocol of Prohaska et al. (2013). Briefly, 2 g of white skeletal muscle tissue from each individual were homogenized with 8 ml of cold PBS and divided into 500 μl quadruplicate aliquots. All replicates were incubated at 50°C for 15 min prior to extraction with 10 volumes (5 ml) of 2:1 chloroform/methanol (ACS grade; histology grade). The organic phase was evaporated at 37°C under a stream of nitrogen before reconstitution in 1 ml of 70% methanol (histology grade) and kept at −20°C for 24 h. Samples were then centrifuged at 962g for 10 min at 4°C, before decanting and evaporating the methanol phase at 37°C under a stream of nitrogen. Dried extracts were reconstituted in PBS containing 0.1% gelatin. To account for procedural loss, two of the four replicates for each muscle sample were cold-spiked with the corresponding non-radiolabelled steroid hormone (Steraloids, Inc., Newport, RI, USA) prior to extraction.

Radioimmunoassay
Plasma and muscle steroid hormone concentrations were determined by radioimmunoassay, following a modified protocol from Tsang and Callard (1987). Non-radiolabelled P 4 , T, and E 2 (Steraloids, Inc.) were used to make stock concentrations of 80 μg ml −1 for P 4 and T, and 6.4 μg ml −1 for E 2 in absolute ethanol (ACS grade). The P 4 , T, and E 2 antibodies (Gordon D. Niswender, Colorado State University, Fort Collins, CO, USA) were diluted to final concentrations of 1:2500, 1:10 000, and 1:18 000, respectively. Tritiated hormones and antibodies were added to the reconstituted plasma and muscle samples using PBS containing 0.1% gelatin to bring the total assay volume to 400 μl. After incubation at 4°C for 24 h, free hormone was separated from bound hormone by the addition of a carbon (0.2%; Acros Organics, Fairlawn, NJ, USA) and dextran 70 (0.02%; Tokyo Chemical Industry Co., Ltd, Tokyo, Japan) suspension, and centrifuged at 1242g for 10 min at 4°C. The supernatant was combined with 3.5 ml of Ecolume (MPO Biomedicals, Solon, OH, USA), and the radioactivity was detected by a Perkin Elmer Tri-Carb 2900TR liquid scintillation analyzer (Waltham, MA, USA). Final concentrations were corrected for procedural loss using individual sample recoveries. When calculating the mean and standard error (±SEM) of plasma and muscle steroid hormone concentrations per stage, any value that was non-detectable was assigned the lowest possible concentration detectable by the assay for the aliquot utilized.

Statistical analysis
Data obtained from 10 R. terraenovae by Prohaska et al. (2013) were pooled with all R. terraenovae data collected in the present study for analysis. Additionally, statistical analyses were conducted on data obtained from 31 S. acanthias by Prohaska et al. (2013). Linear regressions were performed on plasma and skeletal muscle P 4 , T, and E 2 concentrations by species, R. terraenovae and L. erinacea. One-way ANOVAs were performed for R. terraenovae and S. acanthias on plasma and skeletal muscle P 4 , T, and E 2 concentrations by gestational stage, followed by Tukey's post hoc test. If variables failed tests of normality or homogeneity of variance, the data were Box-Cox transformed. If transformed variables still violated the assumptions, the non-parametric Kendall's τ rank correlation or a Kruskal-Wallis rank sum test was performed instead of linear regression or one-way ANOVA, respectively. Multiple regression analyses were conducted by species to generate mathematical equations that may be used to predict each of the following morphological parameters: maximal follicle diameter (MFD), ovary mass (OM), and oviducal gland mass (OGM) for both sharks and skates, and embryo STL for sharks, using skeletal muscle P 4 , T, and E 2 concentrations as explanatory variables. Prior to regression analyses, all skeletal muscle P 4 , T, and E 2 concentration data were log transformed to meet the assumptions of normality and homogeneity of variance. Multiple regression equations were generated by backwards selection of a fully interactive polynomial model, including all interactions between the first-order hormone terms and all first-order hormone terms squared, cubed, and to the fourth power. Any term that was found to be non-significant (P > 0.05) was removed from the model until all highestorder terms were significant. If co-linearity was detected in a model, all first-order terms within that model were centred, and the multiple regression equation was re-generated following the same backwards selection procedure. All data were analysed using R 2.15.2 (R-Core Development, 2012). All tests were considered significant at α ≤ 0.05.

Rhizoprionodon terraenovae
A total of 24 female R. terraenovae (FL, 69-92 cm; 1.6-7.9 kg) were sampled and assigned to the following discrete reproductive stages: five immature, four pre-ovulatory, six mid-gestation (containing embryos of STL 100-142 mm), and nine in late gestation (containing embryos of STL 314-364 mm; see Table 1 for gestational stage details). In addition, data from 10 female R. terraenovae collected by Prohaska et al. (2013) were combined with the present data set, boosting the overall sample size (n = 34). These data included female R. terraenovae assigned to the following discrete reproductive modes: one pre-ovulatory, two early gestation (containing embryos of STL 28-55 mm), four early to mid-gestation (containing embryos of STL 56-83 mm), and three mid-gestation (containing embryos of STL 85-139 mm). The overall mean recoveries of plasma and muscle P 4 , T, and E 2 were 74, 91, and 78% and 52, 47, 51%, respectively. The mean intra-assay coefficients of variation for R. terraenovae plasma and muscle P 4 , T, and E 2 assays were 8, 7, and 5% and 10, 8, and 7%, respectively. The mean inter-assay coefficients of variation for R. terraenovae plasma and muscle P 4 , T, and E 2 assays were 12, 10, and 12% and 12, 13, and 12%, respectively.
Multiple regression analyses revealed that muscle E 2 concentrations were significant predictors of MFD, OGM, and OM, while a combination of muscle P 4 and E 2 concentrations were significant predictors of embryo STL for R. terraenovae (Tables 2 and 3).
The mean inter-assay coefficients of variation for plasma P 4 and E 2 assays (no inter-assay coefficient of variation was calculated for T because all samples were run in one assay) were 12 and 12% for L. erinacea plasma assays, respectively, and 13, 12, and 13% for P 4 , T, and E 2 muscle assays, respectively.
Plasma and muscle P 4 concentrations were elevated in mature, non-reproductively active females and decreased ~50% at the pre-ovulatory stage (Fig. 2a). Progesterone concentrations increased by 400 and 60% in plasma and muscle, respectively, during ovulation, which was followed by a 75% decrease in plasma and 25% decrease in muscle during pre-oviposition. Plasma and muscle P 4 concentrations continued to decrease by ~80% during post-oviposition (Fig. 2a). Similar to P 4 in R. terraenovae, L. erinacea muscle P 4 concentrations were significantly correlated with plasma P 4 (linear regression, r 2 = 0.35, P = 0.022).
In mature, non-reproductively active and pre-ovulatory females, T and E 2 concentrations were relatively low in plasma and muscle. An elevation in T and E 2 was noted during ovulation, increasing by ~1500% in plasma and 500% in muscle (Fig. 2b and c). During pre-oviposition, plasma T and E 2 concentrations decreased by ~50%, while muscle T and E 2 increased by ~30%. Muscle T and E 2 concentrations continued to decrease by ~80% during post-oviposition, while plasma T increased by 66%, and plasma E 2 decreased by 13% (Fig. 2b and c). Similar to P 4 , L. erinacea muscle T and E 2 concentrations were 5 Figure 1. Mean (±SEM) concentrations of plasma (in picograms per millilitre) and muscle steroid hormones (in picograms per gram). Progesterone (A), testosterone (B), and estradiol (C) in Rhizoprionodon terraenovae, plotted by reproductive stage [immature, pre-ovulatory (P/O), early gestation, early to mid-gestation, mid-gestation, and late gestation]. Numbers above the x-axis represent sample size. Uppercase letters denote statistically significant pairwise differences in muscle hormone concentrations between gestational stages, while lowercase letters denote statistically significant pairwise differences in plasma hormone concentrations between gestational stages (P < 0.05). Note the difference in scale between the plasma and muscle axes. The relationships between the steroid hormones progesterone (P 4 ), testosterone (T), and estradiol (E 2 ) and the response variables are indicated by the r 2 and P-values. For a more detailed account of the relationships between the hormones and the response variables, see the model definitions in Tables 3, 5, and 6.
Muscle T concentrations were significant indicators of MFD, while the combination of muscle P 4 and E 2 concentrations were significant indicators of OGM. In addition, the combination of muscle P 4 and T concentrations were significant indicators of OM for L. erinacea (Tables 2 and 5).

Squalus acanthias
A total of 31 female S. acanthias (FL, 78-90 cm; 2.7-4.9 kg) were sampled by Prohaska et al. (2013), and additional analyses, including one-way ANOVA and multiple regression analyses, were conducted on these individuals in the present study. The individuals sampled were previously assigned to the following gestational stages: six pre-ovulatory, six containing candles (fertilized follicles enveloped in a thin membrane within the uterus), five early gestation (containing embryos of STL 62-88 mm), eight mid-gestation (containing embryos of STL 190-240 mm), and six late gestation (containing embryos of STL 250-275 mm). The overall mean recoveries of plasma and muscle P 4 , T, and E 2 were 70, 87, and 80% and 20, 23, and 21%, respectively. The mean intra-assay coefficients of variation for S. acanthias plasma and muscle P 4 , T, and E 2 assays were 6, 7, and 7% and 8, 10, and 8%, respectively. The mean inter-assay coefficients of variation for S. acanthias plasma and muscle P 4 , T, and E 2 assays were 12, 11, and 10% and 12, 11, and 10%, respectively (Prohaska et al., 2013).
Similar to R. terraenovae and L. erinacea, multiple regression analyses showed that a combination of muscle P 4 , T, and E 2 concentrations were significant indicators of MFD, OM, OGM, and embryo STL in S. acanthias (Tables 2 and 6).

Discussion
In non-elasmobranch vertebrates, sex steroid hormones are successfully extracted from bodily materials, such as muscle, faeces, and urine, to study reproductive biology (e.g. Lasley and 6  Pre-ovulatory 5.9 ± 0.9 21 ± 1.6 5.5 ± 1.9 2 Ovulating 12 22 12 1 Pre-oviposition 8.6 ± 2.3 18 ± 0.9 6.7 ± 0.2 3 Post-oviposition 4.3 12 6.2 1 Values are means (±SEM), by reproductive stage; n represents sample size. Kirkpatrick, 1991;Heppell and Sullivan, 2000;Shimizu, 2005;Barnett et al., 2009). In addition, fluctuations of hormones present in these bodily materials mirror those in plasma (e.g. Hoffmann and Rattenberger, 1977;Heppell and Sullivan, 2000;Shimizu, 2005;Barnett et al., 2009). For example, studies have quantified P 4 from blubber of marine mammals to assess whether a female is mature and/or pregnant (Mansour et al., 2002;Kellar et al., 2006). Additionally, urinary and faecal E 2 , T, and P 4 have been used to track gestation in free-ranging terrestrial vertebrates for conservation purposes (Lasley and Kirkpatrick, 1991;Shimizu, 2005). In teleost fishes, fluctuations in skeletal muscle T and E 2 concentrations are associated with maturity, sex, and reproductive cycles (Heppell and Sullivan, 2000;Barnett et al., 2009). Previous research on a few species of elasmobranchs reported that circulating E 2 , T, and P 4 concentrations are correlated with morphological changes in the reproductive tract and with specific events that occur during sexual maturation and the reproductive cycle (e.g. Koob et al., 1986;Rasmussen and Murru, 1992;Sulikowski et al., 2007). More recently, Prohaska et al. (2013) successfully extracted E 2 , T, and P 4 from the skeletal muscle tissue of elasmobranchs and showed that the profiles of these hormones in the muscle are similar to those in the plasma.
Steroid hormones are essential for reproduction. Plasma E 2 concentrations are primarily linked to the growth and maturation of ovarian follicles in elasmobranchs (e.g. Sumpter and Dodd, 1979;Manire et al., 1995;Snelson et al., 1997, Heupel et al., 1999Tricas et al., 2000). Medium-sized follicles produce the highest concentrations of E 2 (Callard and Koob, 1993), which is transported to the liver, where binding to receptors stimulates vitellogenesis (e.g. Sumpter and Dodd, 1979;Manire et al., 1995;Snelson et al., 1997;Heupel et al., 1999;Tricas et al., 2000) and subsequent accumulation of yolk within developing follicles (Perez and Callard 1989;Koob and Callard, 1999). In addition, E 2 also plays a prominent role in the reproductive tract , such as enlargement of the oviducal gland (Koob et al., 1986) and vascularization of the uterus (Koob and Callard, 1999). Likewise, in the present study, we found that ovarian follicular growth accompanied increases in plasma and muscle E 2 concentrations. For example, R. terraenovae plasma and muscle E 2 increased significantly from mid-gestation to late gestation, which corresponded to a substantial increase in follicle diameter. A similar association was observed in the yolk-sac placental viviparous species, Sphyrna tiburo, in which serum E 2 was elevated during pre-ovulation, reduced throughout the majority of gestation, and increased again during the final stages of gestation when follicular growth was beginning (Manire et al., 1995). In S. acanthias, a yolkdependent viviparous species, plasma and muscle E 2 concentrations increased from early to mid-gestation, and remained elevated for the duration of gestation, paralleling follicular growth (Prohaska et al., 2013). These results are similar to those of Tsang and Callard (1987), who previously reported that plasma E 2 increased during the second half of gestation when cohorts of ovarian follicles with larger diameters began to develop. Thus, we anticipated that the highest concentrations of plasma and muscle E 2 would occur in L. erinacea at pre-ovulation, during the period of peak follicular development (Koob et al., 1986), but this was not observed in the present study. This E 2 peak might not have been detected because of the high individual variability of hormone concentrations in continuously reproducing oviparous species (Williams et al., 2013). Alternatively, the skates sampled could have been at the end of their follicular growth phase and/or at the beginning of their ovulatory phase. Furthermore, low sample size may account for our inability to detect this peak. 7 Figure 2. Mean (±SEM) concentrations of plasma (in picograms per millilitre) and muscle steroid hormones (in picograms per gram). Progesterone (A), testosterone (B), and estradiol (C) in Leucoraja erinacea, plotted by reproductive stage [immature, mature (nonreproductively active), pre-ovulatory, ovulating, pre-oviposition, and post-oviposition]. Numbers above the x-axis represent sample size. Note the difference in scale between the plasma and muscle axes.
Unlike E 2 , changes in P 4 are unique to reproductive mode and often linked to mode-specific reproductive events in elasmobranchs. For example, in S. tiburo, peaks in serum P 4 concentrations during early gestation are related to uterine compartmentalization and implantation (Schlernitzauer and Gilbert, 1966;Callard et al., 1992;Manire et al., 1995). In the present study, the peak in plasma and muscle P 4 during early to mid-gestation for R. terraenovae is also likely to be related to uterine compartmentalization and implantation, especially as implantation occurs when embryos are ~70-85 mm in length (Castro and Wourms, 1993). In S. acanthias, high concentrations of plasma P 4 during early gestation are linked to embryo retention  and suppression of vitellogenesis (Paolucci and Callard, 1998;Koob and Callard, 1999). Similar to those studies, Prohaska et al. (2013) showed that S. acanthias plasma and muscle P 4 peaked during early gestation when little to no follicular growth was occurring, and then decreased into late gestation when follicular growth was beginning. Despite the high individual variability in L. erinacea plasma P 4 concentrations (Williams et al., 2013), the patterns exhibited by plasma and muscle P 4 in the present study were similar to those found by Koob et al. (1986), who noted that plasma P 4 peaks during ovulation, and then decreases during pre-and post-oviposition. In addition, Koob et al. (1986) suggested that the rise in P 4 might be related to ovulation and the formation of egg cases, while the decrease in P 4 could be related to oviposition. Furthermore, decreases in P 4 after ovulation, like those found in the plasma and muscle in the present study, are suggested to inhibit the early release of egg cases by hormonally controlling and tightening cervix muscles to maintain them within the uterus to undergo sclerotization and tanning (Koob and Cox, 1988;Koob and Callard, 1999).
There is a close relationship between T and E 2 in yolkdependent viviparous and oviparous elasmobranchs. Plasma T 8 0.5 ± 0.3 4.1 ± 0.3 5.5 ± 1.9 12 6.7 ± 0.2 6.2 The characteristics are included in this table to be utilized as an index, and expressed as mean (±SEM) values of MFD (in millimetres), OM (in grams), and OGM (in grams) by reproductive stage. All hormone terms within the regression equations were log transformed. a All terms included in the model were centred. is primarily linked to the growth of follicles (Koob et al., 1986;Tsang and Callard, 1987), with larger follicles producing the highest concentrations, which serve as a substrate for E 2 synthesis to facilitate vitellogenesis  and the continued accumulation of yolk by developing follicles (Perez and Callard, 1989;Koob and Callard, 1999). In the present study, the patterns of plasma and muscle T concentrations in L. erinacea were similar to those reported in the aforementioned studies, i.e. increases in T are concurrent with increases in follicle diameter (Koob et al., 1986;Tsang and Callard, 1987). However, in the yolk-sac placental viviparous R. terraenovae, plasma and muscle T and P 4 had similar profiles, with both hormones peaking during early to mid-gestation. These profiles are exactly like the ones exhibited in S. tiburo serum, in which fluctuations of T were analogous to P 4 , peaking in early gestation when implantation and compartmentalization are suggested to occur (Manire et al., 1995).
Of the few studies so far that determined plasma sex steroid hormones in immature elasmobranchs, all have reported little to no detectable concentrations of these hormones in fishes classified as immature, based on length as well as the underdeveloped condition of reproductive organs (e.g. Rasmussen and Gruber, 1990;Rasmussen and Murru, 1992;Cicia et al., 2009). In the present study, we attempted to measure plasma and muscle P 4 , E 2 , and T concentrations in viviparous and oviparous elasmobranchs that were identified as immature, based on the presence of underdeveloped reproductive tracts. Similar to previous reports, plasma P 4 , E 2 , and T were low compared with those individuals that were mature. In the present study, we also found that the same is true for muscle hormone concentrations. These results provided further support for the direct link between sex steroid hormones, the readiness of the reproductive tract, and reproductive status in elasmobranchs (Rasmussen and Murru, 1992;Gelsleichter et al., 2002;Sulikowski et al., 2006Sulikowski et al., , 2007, as well as the potential use of the muscle hormones to assess maturity.
In addition to examining the relationships between muscle and plasma steroid hormones, multiple regression analyses were conducted in the present study to determine whether skeletal muscle P 4 , T, and E 2 concentrations relate to OGM, OM, and MFD in R. terraenovae, L. erinacea, and S. acanthias, as well as to embryo STL in the two species of sharks. So far, such analyses have led to the creation of several mathematical models, although the sample sizes used to generate them were low. Nonetheless, these preliminary models are a first step towards developing the use of P 4 , T, and E 2 concentrations in non-lethally obtained skeletal muscle as indicators of reproductive status and gestational stage.

Conclusions
The present study reported the successful detection of steroid hormones in the skeletal muscle of female R. terraenovae, L. erinacea, and S. acanthias during specific stages of their reproductive cycles. More importantly, the profiles of plasma 9 Figure 3. Mean (±SEM) concentrations of plasma (in picograms per millilitre) and muscle steroid hormones (in picograms per gram). Progesterone (A), testosterone (B), and estradiol (C) in Squalus acanthias, plotted by reproductive stage [pre-ovulatory (P/O), candle, early gestation, mid-gestation, and late gestation]. Numbers above the x-axis represent sample size. Uppercase letters denote statistically significant pairwise differences in muscle hormone concentrations between gestational stages, while lowercase letters denote statistically significant pairwise differences in plasma hormone concentrations between gestational stages (P < 0.05). Note the difference in scale between the plasma and muscle axes. This figure is reprinted from Prohaska et al. (2013) Development of a non-lethal and minimally invasive protocol to study elasmobranch reproduction. Mar Coast Fish 5: 181-188, by permission of Taylor and Francis (http://www.tandfonline.com); however, the significant pairwise differences in hormone concentrations between gestational stages, denoted by letters, have been added, because this was a new analysis in the present study. and skeletal muscle concentrations of P 4 , T, and E 2 in R. terraenovae and L. erinacea were significantly associated, indicating that skeletal muscle tissue is an appropriate substitute for plasma. The present results also affirm the efficacy of using steroid hormones in non-lethally obtained skeletal muscle to assess reproductive status in elasmobranchs (Prohaska et al., 2013). The primary advantage of utilizing muscle vs. blood is that obtaining a small muscle tissue sample may be less stressful, because it does not require that the animal be restrained or removed from the water, and can be obtained more rapidly than blood. While there are caveats in using this method on smaller species, in the long run, skeletal muscle can be used as an alternative when blood cannot be obtained, which will facilitate its use on large species, as well as threatened species of elasmobranchs. Our previous work on S. acanthias and R. terraenovae suggests that P 4 , T, and E 2 can be extracted from white skeletal muscle tissue (Prohaska et al., 2013), and the present study further suggests that white skeletal muscle tissue of the oviparous L. erinacea can also be used to extract the same three hormones. The analysis of L. erinacea skeletal muscle steroid hormones, so far, suggests that they may relate well to reproductive status. However, increasing sample size will strengthen and provide a more accurate depiction of these skeletal muscle hormones throughout the reproductive cycle of an oviparous species. Additionally, the findings of the present study provide a more complete understanding of fluctuations in P 4 , T, and E 2 during the gestation of R. terraenovae, reinforcing our claim that skeletal muscle steroid hormones relate well to gestation in a yolk-sac placental species.
Finally, multiple regression analyses conducted on R. terraenovae, L. erinacea, and S. acanthias suggest that skeletal muscle P 4 , T, and E 2 are significantly related to changes in reproductive tract morphologies during specific stages of the reproductive cycles including gestation. The robustness of these mathematical models will be strengthened by increasing sample size before they become tools for fisheries managers.