- Email: [email protected]
In vivo investigations on luteotropic activity of prostaglandins during early diestrus in nonpregnant bitches
Theriogenology 82 (2014) 915–920
Contents lists available at ScienceDirect
Theriogenology journal homepage: www.theriojournal.com
In vivo investigations on luteotropic activity of prostaglandins during early diestrus in nonpregnant bitches Tomasz Janowski a, *, Julia Fingerhut b, Mariusz P. Kowalewski c, czyk a, Anna Domos1awska a, Andrzej Jurczak a, Alois Boos c, S1awomir Zdun Gerhard Schuler b, Bernd Hoffmann b a
Department of Animal Reproduction with Clinic, University of Warmia and Mazury in Olsztyn, Olsztyn, Poland Clinic for Veterinary Obstetrics, Gynecology and Andrology, Justus-Liebig-University Giessen, Giessen, Germany c Institute of Veterinary Anatomy, Vetsuisse Faculty, University of Zurich, Zurich, Switzerland b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 14 April 2014 Received in revised form 25 June 2014 Accepted 3 July 2014
The aim of this study was to test for the postulated luteotropic effect of prostaglandin E2 during early diestrus in the dog in an in vivo study. This study was performed on 30 bitches which were randomly assigned to a treatment group (TG) and a control group. Starting on the day of ovulation (Day 0), dogs of the TG were treated for 5, 10, 20, or 30 days with 10 mg firocoxib/kg body weight per day (Previcox, a selective PTGS2 inhibitor) and ovariohysterectomized for collection of corpora lutea on the last day of treatment. Similarly, dogs of the control group were ovariohysterectomized on Days 0, 5, 10, 20, and 30. Blood samples for progesterone measurement were collected every second day; additionally, the area of luteal cell nuclei and the expression of 3b-hydroxysteroid-dehydrogenase at the mRNA and the protein levels were assessed. Mean P4 concentrations were lower in TGs; however, a significant difference was only observed on Day 10. This observation is in line with the finding that treatment with firocoxib reduced expression of 3b-hydroxysteroiddehydrogenase mRNA and protein (P < 0.05) and the area of luteal cell nuclei (P < 0.05). The results of this study further point to the postulated luteotropic function of prostaglandin E2. Ó 2014 Elsevier Inc. All rights reserved.
Keywords: Bitch Corpus luteum PTGS2 inhibitor Progesterone Luteal cell nuclei area 3bHSD
1. Introduction The reproductive cycle of the dog differs in many aspects from that of other domestic animal species. Thus, luteal function is almost identical in nonpregnant and pregnant females, except that pregnant animals attain baseline levels of progesterone (P4) earlier, owing to the immediate prepartal decline of this hormone [1]. Luteal regression in nonpregnant dogs occurs independently of a uterine luteolysin as hysterectomy does not affect ovarian and pituitary functions [2,3]. Concerning luteotropic mechanisms, the
* Corresponding author. Tel.: þ48 89 5233896; fax: þ48 89 5233440. E-mail address: [email protected] (T. Janowski). 0093-691X/$ – see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.theriogenology.2014.07.005
early corpus luteum seems to be independent of pituitary support as was suggested from a study in which hypophysectomies performed on Day 4 after ovulation resulted only in a temporary decrease of P4 concentrations, with levels returning to normal between 6 and 10 days [4]. Other than that, hypophysectomies performed between Days 10 to 50 after ovulation were followed by a rapid and permanent decrease of P4 concentrations to basal levels [4,5]. Although LH and particularly prolactin have been identified as luteotropic agents during the latter 2/3 parts of luteal life span [6,7], only recently, and when testing for a local CL-bound expression of the prostaglandin system, information on likely luteotropic mechanisms during the early phase of corpus luteum formation in the bitch has become available. Thus, expression of prostaglandin-endoperoxide
916
T. Janowski et al. / Theriogenology 82 (2014) 915–920
synthase 2 (PTGS2, COX2) was cycle related and highest at the beginning of the CL phase, both at the mRNA and protein levels [8,9]. Similarly, prostaglandin E2 (PGE2) synthase showed cycle-related changes with the highest expression during the early CL phase. A similar pattern was observed for the prostaglandin transporter (PGT) and the prostaglandin E2 receptor EP2 (PTGER2), whereas the EP4 (PTGER4) receptor showed no cycle-related changes [10,11]. Other than that, expression of prostaglandin F2a synthase (PGFS/ AKR1C3), the only canine-specific PGFS known so far, was either absent or only very weak [12]. On the basis of these data and observations made in other species [13,14], Kowalewski et al. [10,11] speculated on the luteotropic role of PGE2 during the initial phase of corpus luteum formation. This hypothesis was supported by an in vitro study with cell cultures of luteal cells from early diestrus bitches showing that addition of PGE2 to the system stimulated expression of steroid acute regulatory protein [11]. To further verify the previously mentioned hypothesis in the present in vivo study, dogs were treated with a selective COX2 (PTGS2) inhibitor from the day of ovulation until mid-diestrus; it was postulated that reducing the availability of prostaglandins, including PGE2, by this treatment would negatively affect corpus luteum function. This was assessed by determining luteal expression of 3bhydroxysteroid-dehydrogenase (3bHSD), a key steroidogenic enzyme allowing for the synthesis of P4, and the course of peripheral P4 concentrations. As has been shown in a number of studies, the volume of cell nuclei is indicative of cellular activity [15–18]; in the case of Leydig cells, this parameter has also been used for assessment of cellular activity [19,20]. Thus in this study, determination of the area of luteal cell nuclei served as a morphological parameter to test for effects of treatment. 2. Materials and methods Animal experiments were approved by the responsible ethics committee with permit 54/2008, University of Warmia and Mazury in Olsztyn, Poland. 2.1. Experimental design The study comprised a total of 30 bitches of various breeds and age (2–7 years). All dogs were observed for onset of spontaneous estrus, which was monitored by vaginal cytology and P4 assay, using a commercially available chemiluminiscence kit (Progesteroene II, Roche Diagnostics, Mannheim, Germany); the day when P4 concentration had just exceeded 5 ng/mL plasma was considered the day of ovulation (Day 0). Dogs were randomly assigned to 4 treatment groups (TG) and 5 control groups (CG). Starting with Day 0, dogs of TG were treated for 5 (n ¼ 4), 10 (n ¼ 3), 20 (n ¼ 2; originally 3 dogs had been assigned to the group, but from the tissue sample of one dog designated for determination of the area of the lutein cell nuclei, the CL had been accidently cut off), or 30 (n ¼ 3) days with 10-mg firocoxib/kg body weight per day (firocoxib is a highly selective COX2 inhibitor marketed as Previcox by Merial Ltd.; the recommended daily dose is 5 mg/kg body weight per day).
Dogs of the CG received a placebo for 0 (n ¼ 3), 5 (n ¼ 5), 10 (n ¼ 3), 20 (n ¼ 3), or 30 (n ¼ 3) days. Blood samples were collected for P4 measurements at 2 -day intervals after ovulation until ovariohysterectomy. The blood plasma obtained was stored at 20 C until assay, which was within 4 to 5 months. The animals were ovariohysterectomized for collection of corpora lutea on the last day of treatment. The corpora lutea were trimmed off the connective tissue, divided into 2, with 1 part being preserved for mRNA analysis as described earlier [11] and the other part being fixed in 10% phosphate–buffered formalin and paraffin embedded, applying standard procedures. 2.2. Morphological assessment of luteal cell nuclei area Assessment was on 1 to 2 hematoxylin-eosin–stained slides (3–4 mm) and up to 10 sight fields, depending on the quality of the slide and the number of nuclei meeting the criteria set. Hundred nuclei per dog were evaluated using magnification 400 and the Leica Image Manager IM 1000 software program (Leica MIcrosysteme Vertrieb GmbH, Wetzlar, Germany). To be considered for evaluation, a luteal cell had to meet the following criteria: a clearly detectable excentrically located round and/or ellipsoid nucleus with sharp boundaries and 1 or 2 clearly visible nucleoli. The half axes “a” and “b” were determined and the area (A) was calculated using the equation A ¼ pab. 2.3. Progesterone assay Progesterone was determined by an in-house competitive radioimmunoassay following the procedure outlined by Hoffmann et al. [21]. Intra-assay coefficients of variation varied between 8.8% and 9.6% and interassay coefficients of variation between 10.2% and 14.7%. 2.4. Immunohistochemical detection of 3bHSD The immunoperoxidase immunohistochemistry protocol was applied following the previously published protocol [9,22]. The primary antibody was a rabbit polyclonal antiserum against human placental 3bHSD used at 1:3000 dilution (a gift from Dr. J.I. Mason, University of Edinburgh, UK [23]); serum from nonimmunized rabbit served as negative control. Additionally, sections omitting the primary antibody were included. Peroxidase-conjugated goat antirabbit IgG (Vector Laboratories, Burliname, CA) were used as a secondary antibody. Signals were enhanced using the avidin-biotin-peroxidase complex (Vectastain ABC kit, Vector Laboratories), and signals were detected using Liquid DAB þ substrate kit (Dako Schweiz AG, Baar, CH). Sections were counterstained with hematoxylin and embedded in Histokit (Assistant, Osterode, Germany). 2.5. Semiquantitative assessment of 3bHSD mRNA expression Total RNA was isolated using TRIzol reagent (Invitrogen, Carlsbad, CA) following the manufacturer’s instructions. The RNA content was measured with a NanoDrop 2000C
T. Janowski et al. / Theriogenology 82 (2014) 915–920
spectrophotometer (Thermo Fisher Scientific AG, Reinach, CH). All RNA samples were DNase-treated with RQ1 RNasefree DNase (Promega, Dübendorf, CH), according to the supplier’s protocol, to remove potential genomic DNA contamination, which may occur when applying TRIzol. The reverse transcriptase (RT) reaction with random hexamers used as primers and other reagents for cDNA synthesis from Applied Biosystems (Foster City, CA) was followed our previously published protocol [2,17]. The RT reactions were conducted in an Eppendorf Mastercycler (Vaudaux-Eppendrf AG, Basel, CH) according to the following protocol: 8 minutes at 21 C and 15 minutes at 42 C; the reaction was then stopped by incubation for 5 minutes at 99 C. Real-time (TaqMan) polymerase chain reaction (PCR) reactions were performed using an automated fluorometer ABI PRISM 7500 Sequence Detection System (Applied Biosystems), following our previously published protocol and the manufacturer’s instructions [9,24]. The following negative controls were used: water instead of cDNA and the socalled RT minus control (i.e., samples that were not reverse transcribed). Three independent endogenous reference genes (GAPDH, 18SrRNA, and cyclophilin A) were involved in the semiquantitation of 3bHSD expression in the comparative method (DDCT method) described previously [11,24] and as advised by Applied Biosystems, the manufacturer of the ABI 7500 Fast Real Time PCR System. Reactions were set up to achieve approximately 100% efficiency using the CT slope method. Complementary DNA corresponding to 100-ng DNase-treated RNA per sample was used in reactions with Fast Start Universal Probe Master (ROX; Roche Diagnostics AG, Schweiz). Selected PCR products were sent for commercial sequencing (Microsynth). The following TaqMan systems (primers and 6-carboxyflurescein–labeled and 6carboxytertamethyl rhodamine–labeled probes) were purchased from Microsynth: 3bHSD forward: 50 -GGG TAC TCA GCT CCT GTT GGA A-30 , 3bHSD reverse: 50 -GCC ACC TCT ATG GTG CTG GTA T-30 , 3bHSD TaqMan probe: 50 -TGC CCA GGC TAG TGT GCC GAT CTT-30 (gene bank accession number
917
AY739720); GAPDH forward: 30 -GCT GCC AAA TAT GAC GAC ATC A-30 , GAPDH reverse: 50 -GTA GCC CAG GAT GCC TTT GAG30 , TaqMan probe: 50 -TCC CTC CGA TGC CTG CTT CAC TAC CTT30 , TaqMan probe: 50 -TCC CTC CGA TGC CTG CTT CAC TAC CTT30 (gene bank accession number AB028142); 18SrRNA forward: 50 -GTC GCT TCC TCT CCT ACT-30 , 18SrRNA reverse: 50 GGC TGA CCG GGT TGG TTT-30 , and 18SrRNA TaqMan probe: 50 -ACA TGC CGA CGG GCG CTG AC-30 (gene bank accession number FJ797658). The canine-specific cyclophilin A TaqMan gene Expression Assay was commercially available from Applied Biosystems (Prod. No. Cf03986523_gH). 2.6. Statistical evaluation Progesterone data were skewed to the right and consequently the geometric mean (Xg) and deviation factor (DF) were calculated. To test for the effect of treatment and time and to account for the decreasing number of samples from Days 0 to 30, a stepwise ANOVA with repeated measures of time was performed, followed by a t test for independent samples. For the area of luteal cell nuclei, the arithmetic mean (X) and standard deviation were calculated; effects of treatment were assessed by a 2-factorial ANOVA with the factors being treatment and day of ovariohysterectomy. The progesterone concentrations and size of lutein cell nuclei were assessed by correlation analysis. As expression levels of 3bHSD mRNA were skewed to the right, the geometric mean (Xg) and DF were calculated; To test for effects of treatments an unpaired, 2-tailed Student t test was performed. 3. Results 3.1. Progesterone concentrations Progesterone concentrations showed a high variability as indicated by the DF varying between 1.3 and 4.1. As shown in Figure 1, from Day 4 onward, the calculated
Fig. 1. Course of progesterone concentrations in treated group and control group dogs. Values given as (Xg deviation factor1) a:b P < 0.05.
918
T. Janowski et al. / Theriogenology 82 (2014) 915–920
Fig. 2. Development of area luteal cell nuclei; values given as (X standard deviation).
mean values (Xg) in the TG ranged below those of the CG. However, the effect of treatment was not significant (stepwise ANOVA, P > 0.05) and only on Day 10 was the observed difference between the TG and CG significant with P < 0.05. For both groups, the effect of time was highly significant (P < 0.0001).
Fig. 3. Expression of 3b-hydroxysteroid-dehydrogensae (3bHSD) as determined by semiquantitative (TaqMan) reverse transcription quantitative polymerase chain reaction in luteal tissue samples of control and treatment groups; values given as (Xg deviation factor1). Bars with (*) differ at P < 0.04 and bars with (**) differ at P < 0.05.
3.2. Area of luteal cell nuclei With P < 0.001 (ANOVA), there was a significant effect of treatment. After an initial increase from Day 0 to Day 5,
Fig. 4. Immunohistochemical localization of 3b-hydroxysteroid-dehydrogensae in canine luteal samples in control and treatment groups. Representative pictures are shown for Days 5, 20, and 30 after ovulation (oral). Immunohistochemical signals are localized in cytoplasm of lutein cells.
T. Janowski et al. / Theriogenology 82 (2014) 915–920
the values did not further change in the TG whereas there was a distinct increase in the CG on Days 10 and 20; similar values were again observed on Day 30 (Fig. 2). Progesterone concentrations and the area luteal cell nuclei showed a highly significant (P < 0.001) interaction indicating that the difference observed between the CG and TG depended on the day of castration. Progesterone values and area of luteal cell nuclei were positively correlated (P < 0.05) in the CG but not in the TG. 3.3. Expression of 3ßHSD The expression of 3bHSD was clearly detectable, both at the mRNA and at the protein levels in all luteal samples investigated, displaying high variability between individual animals (Fig. 3). Treatment with Previcox significantly reduced expression of 3bHSD-mRNA on Days 20 and 30 (P < 0.04) compared with the respective controls. Similar effects were observed at the protein level by immunohistochemistry; the evenly distributed signals, which were localized to the cytoplasm of luteal cells, appeared to be weaker after treatment (Fig. 4). 4. Discussion A likely role of PGE2 as a luteotropic factor has been addressed in a number of articles. Thus, Christenson et al. [25] concluded from their study in pigs that a conceptusassociated increase in uterine PGE2 secretion is consistent with the role of PGE2 in the local stimulation of luteal function during early pregnancy. In an in vitro assay, Weems et al. [26] showed that addition of PGE1 and PGE2 to Day 8 ovine corpora lutea linearly stimulated progesterone secretion, matching the effect of LH. Their observations of 88 to 90 days ovine corpora lutea of pregnancy led to the conclusion that these corpora lutea produce their own luteotropin which is PGE2. In the cow intraluteal implants containing PGE2 prevented or inhibited luteolysis when implanted on Day 13 post-estrus, with the effect apparently based on the prevention of the loss of luteal LH receptors, as delineated from their expression on the mRNA level [27]. In the dog, the present study has provided further evidence that prostaglandins, and presumably mostly PGE2, seem to act luteotropically in the early phase of diestrus in the dog. Thus, in addition to our preceding expression and in vitro studies [9–11], the present study clearly shows that inhibition of prostaglandin synthesis by applying a COX2 (PTGS2) inhibitor negatively affects the corpus luteum function in vivo. A reduced cellular activity of lutein cells is indicated by the reduced area of luteal cell nuclei after Day 5 of treatment. This situation was mirrored in the significantly reduced expression of 3bHSD on both the mRNA and the protein level. Also, the resulting decrease of progesterone mean levels observed in peripheral plasma of treated animals mirrors impaired luteal cell function. The fact that the latter effect was not significant on a group basis but only on Day 10 on a pairwise comparison may be explained by the large variation in individual progesterone concentrations and that the statistics applied (stepwise ANOVA) had to account for the continuous reduction of blood samples available from Day 0 to Day 30.
919
To our knowledge, this is the first in vivo study substantiating a luteotropic activity of PGE2 in the dog during the early diestrus phase. On the basis of data concerning expression of the prostaglandin system in pregnant bitches [28], the same situation seems to apply for the early pregnancy corpus luteum. Clearly, inhibition of COX2 (PTGS2) inhibits the steroidogenesis in the CL as indicated by the decreased expression of 3bHSD. It should be noted that this effect was observed when only double the recommended therapeutic dose was used. However, additional information on the underlying mechanisms is necessary to allow further functional conclusions. Thus, treatment must have affected activity of COX2 (PTGS2) not only in the corpora lutea but also in other organs with the effect on the ovary itself still needing to be further defined. However, from a clinical point of view, treatment of bitches with a COX2 inhibitor may also impair luteal ovarian function depending on the dosage. Acknowledgments This study was supported by The Polish Ministry of Science and Higher Education as research project Number NN308225936. The technical assistance of Stefanie Ihle and Elisabeth Högger is highly appreciated. References [1] Concannon PW, Butler WR, Hansel W, Knight PJ, Hamilton JM. Parturition and lactation in the bitch: serum progesterone, cortisol and prolactin. Biol Reprod 1978;19:1113–8. [2] Hoffmann B, Höveler R, Hasan SH, Failing K. Ovarian and pituitary function in dogs after hysterectomy. J Reprod Fertil 1992;96:837–45. [3] Olson PN, Bowen RA, Behrendt MO, Olson JD, Nett TM. Concentrations of progesterone and luteinizing hormone in the serum of diestrus bitches before and after hysterectomy. Am J Vet Res 1984; 45:149–53. [4] Okkens AC, Dieleman DJ, Bevers MM, Lubberink AA, Willemse AH. Influence of hypophysectomy on life span of the corpus luteum in the cycling dog. J Reprod Fertil 1986;77:187–92. [5] Concannon PW. Effects of hypophysectomy and of LH administration on luteal phase plasma progesterone levels in the beagle bitch. J Reprod Fertil 1980;58:407–10. [6] Concannon PW, Weinstein P, Whaley S, Frank D. Suppression of luteal function in dogs by luteinizing hormone antiserum and by bromocryptine. J Reprod Fertil 1987;81:175–80. [7] Okkens AC, Bevers MM, Dieleman SJ, Willemse AH. Evidence for prolactin as the main luteotropic factor in the cyclic dog. Vet Q 1990;12:193–201. [8] Hoffmann B, Büsges F, Engel E, Kowalewski MP, Papa P. Regulation of corpus-luteum function in the bitch. Reprod Domest Anim 2004; 39:232–40. [9] Kowalewski MP, Schuler G, Taubert A, Engel E, Hoffmann B. Expression of cyclooxygenase 1 and 2 in the canine corpus luteum during diestrus. Theriogenology 2006;66:1423–30. [10] Kowalewski MP, Mutembei HM, Hoffmann B. Canine prostaglandin E2 synthase (PGES) and its receptors (EP2 and EP4): expression in the corpus luteum during dioestrus. Anim Reprod Sci 2008;109:319–29. [11] Kowalewski MP, Fox B, Gram A, Boos A, Reichler I. Prostaglandin E2 functions as a luteotrophic factor in the dog. Reproduction 2013; 145:213–26. [12] Kowalewski MP, Mutembei HM, Hoffmann B. Canine prostaglandin F2a receptor (FP) and prostaglandin F2a synthase (PGFS): molecular cloning and expression in the corpus luteum. Anim Reprod Sci 2008;107:161–75. [13] Diaz FJ, Anderson LE, Wu YL, Rabot A, Tsai SJ, Wiltbank MC. Regulation of progesterone and prostaglandin F2a production in the CL. Mol Cell Endocrinol 2002;191:65–80. [14] Wiltbank MC, Ottobre JS. Regulation of intraluteal production of prostaglandins. Reprod Biol Endocrinol 2003;1:91.
920
T. Janowski et al. / Theriogenology 82 (2014) 915–920
[15] Benninghoff A. Funktionelle Kernschwellung und Kernschrumpfung. Anatomische Nachrichten 1950;1:50–2. [16] Krantz H. Kern und Funktion I: Die Kerngröße und ihre Abhängigkeit von äußeren und inneren Faktoren. Z Zellforsch Mikrosk Anat 1951;35:425–57. [17] Polster C. Karyometrische und karyologische Untersuchungen an den Hauptstück- epithelien der Rattenniere bei experimenteller Eiweißnephrose. Virchows Arch Path Anat 1959;332:420–30. [18] Tietze KW, Inthraphuvasak J, Hiersche HD. Funktionelle Morphologie des Endometrium corporis uteri. Arch Gynakol 1970;209:331–8. [19] Gentil M, Hoffmann B, Spang A, Failing K, Goericke-Pesch S. Restart of steroidogenesis in dogs during recrudescence of testicular function following downregulation with a GnRH-agonist implant. Cell Tissue Res 2012;350:513–23. [20] Kronthaler O. Karyometrische Untersuchungen über den Einfluss einiger Wirkstoffe auf die Interstitialzellen von Gonaden. Diss med vet; Ludwig-Maximilians-Universität München; 1960. [21] Hoffmann B, Kyrein HJ, Ender ML. An efficient procedure for the determination of progesterone by radioimmunoassay applied to bovine peripheral plasma. Horm Res 1973;4:302–10. [22] Kowalewski MP, Mason JI, Howie AF, Morley SD, Schuler G, Hoffmann B. Characterization of the canine 3beta-hydroxysteroid dehydrogenase and its expression in the corpus luteum during diestrus. J Steroid Biochem Mol Biol 2006;101:254–62. [23] Lorence MC, Murry BA, Trant JM, Mason JI. Human 3ß-hydroxysteroid-dehydrogenase/D 5-4 isomerase from placenta: expression
[24]
[25]
[26]
[27]
[28]
in nonsteroidogenic cells of a protein that catalyzes the dehydrogenation/isomerisation of C21 and C19 steroids. Endocrinlogy 1990; 126:2493–8. Kowalewski MP, Meyer A, Hoffmann B, Aslan S, Boos A. Expression and functional implications of peroxisome proliferator-activated receptor gamma (PPARg) in canine reproductive tissues during normal pregnancy and parturition and at antiprogestin induced abortion. Theriogenology 2011;75:877–86. Christenson LK, Farley DB, Anderson LH, Ford SP. Luteal maintenance during early pregnancy in the pig: role for prostaglandin E2. Prostaglandins 1994;47:61–5. Weems YS, Bridges PJ, Tanaka Y, Sasser RG, LeaMaster BR, Vincent DL, et al. PGE1 or PGE2 not LH regulates secretion of progesterone in vitro by the 88-90 day ovine corpus luteum of pregnancy. Prostaglandins 1997;53:337–53. Weems YS, Alejandro Arreguin-Arevalo J, Nett TM, Vann RC, Ford SP, Bridges PJ, et al. In vivo intra-luteal implants of prostaglandin (PG) E1 or E2 (PGGE1, PGE2) prevent luteolysis in cows: I. Luteal weight, circulating progesterone, mRNA for luteal luteinizing hormone (LH) receptor, and occupied and unoccupied luteal receptor for LH. Prostaglandins and other Lipid Mediat 2011;95:34–44. Kowalewski MP, Beceriklisoy HB, Aslan S, Agaoglu AR, Hoffmann B. Time related changes in luteal prostaglandin synthesis and steroidogenic capacity during pregnancy, normal and antiprogestin induced luteolysis in the bitch. Anim Reprod Sci 2009;116:129–38.
Contents lists available at ScienceDirect
Theriogenology journal homepage: www.theriojournal.com
In vivo investigations on luteotropic activity of prostaglandins during early diestrus in nonpregnant bitches Tomasz Janowski a, *, Julia Fingerhut b, Mariusz P. Kowalewski c, czyk a, Anna Domos1awska a, Andrzej Jurczak a, Alois Boos c, S1awomir Zdun Gerhard Schuler b, Bernd Hoffmann b a
Department of Animal Reproduction with Clinic, University of Warmia and Mazury in Olsztyn, Olsztyn, Poland Clinic for Veterinary Obstetrics, Gynecology and Andrology, Justus-Liebig-University Giessen, Giessen, Germany c Institute of Veterinary Anatomy, Vetsuisse Faculty, University of Zurich, Zurich, Switzerland b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 14 April 2014 Received in revised form 25 June 2014 Accepted 3 July 2014
The aim of this study was to test for the postulated luteotropic effect of prostaglandin E2 during early diestrus in the dog in an in vivo study. This study was performed on 30 bitches which were randomly assigned to a treatment group (TG) and a control group. Starting on the day of ovulation (Day 0), dogs of the TG were treated for 5, 10, 20, or 30 days with 10 mg firocoxib/kg body weight per day (Previcox, a selective PTGS2 inhibitor) and ovariohysterectomized for collection of corpora lutea on the last day of treatment. Similarly, dogs of the control group were ovariohysterectomized on Days 0, 5, 10, 20, and 30. Blood samples for progesterone measurement were collected every second day; additionally, the area of luteal cell nuclei and the expression of 3b-hydroxysteroid-dehydrogenase at the mRNA and the protein levels were assessed. Mean P4 concentrations were lower in TGs; however, a significant difference was only observed on Day 10. This observation is in line with the finding that treatment with firocoxib reduced expression of 3b-hydroxysteroiddehydrogenase mRNA and protein (P < 0.05) and the area of luteal cell nuclei (P < 0.05). The results of this study further point to the postulated luteotropic function of prostaglandin E2. Ó 2014 Elsevier Inc. All rights reserved.
Keywords: Bitch Corpus luteum PTGS2 inhibitor Progesterone Luteal cell nuclei area 3bHSD
1. Introduction The reproductive cycle of the dog differs in many aspects from that of other domestic animal species. Thus, luteal function is almost identical in nonpregnant and pregnant females, except that pregnant animals attain baseline levels of progesterone (P4) earlier, owing to the immediate prepartal decline of this hormone [1]. Luteal regression in nonpregnant dogs occurs independently of a uterine luteolysin as hysterectomy does not affect ovarian and pituitary functions [2,3]. Concerning luteotropic mechanisms, the
* Corresponding author. Tel.: þ48 89 5233896; fax: þ48 89 5233440. E-mail address: [email protected] (T. Janowski). 0093-691X/$ – see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.theriogenology.2014.07.005
early corpus luteum seems to be independent of pituitary support as was suggested from a study in which hypophysectomies performed on Day 4 after ovulation resulted only in a temporary decrease of P4 concentrations, with levels returning to normal between 6 and 10 days [4]. Other than that, hypophysectomies performed between Days 10 to 50 after ovulation were followed by a rapid and permanent decrease of P4 concentrations to basal levels [4,5]. Although LH and particularly prolactin have been identified as luteotropic agents during the latter 2/3 parts of luteal life span [6,7], only recently, and when testing for a local CL-bound expression of the prostaglandin system, information on likely luteotropic mechanisms during the early phase of corpus luteum formation in the bitch has become available. Thus, expression of prostaglandin-endoperoxide
916
T. Janowski et al. / Theriogenology 82 (2014) 915–920
synthase 2 (PTGS2, COX2) was cycle related and highest at the beginning of the CL phase, both at the mRNA and protein levels [8,9]. Similarly, prostaglandin E2 (PGE2) synthase showed cycle-related changes with the highest expression during the early CL phase. A similar pattern was observed for the prostaglandin transporter (PGT) and the prostaglandin E2 receptor EP2 (PTGER2), whereas the EP4 (PTGER4) receptor showed no cycle-related changes [10,11]. Other than that, expression of prostaglandin F2a synthase (PGFS/ AKR1C3), the only canine-specific PGFS known so far, was either absent or only very weak [12]. On the basis of these data and observations made in other species [13,14], Kowalewski et al. [10,11] speculated on the luteotropic role of PGE2 during the initial phase of corpus luteum formation. This hypothesis was supported by an in vitro study with cell cultures of luteal cells from early diestrus bitches showing that addition of PGE2 to the system stimulated expression of steroid acute regulatory protein [11]. To further verify the previously mentioned hypothesis in the present in vivo study, dogs were treated with a selective COX2 (PTGS2) inhibitor from the day of ovulation until mid-diestrus; it was postulated that reducing the availability of prostaglandins, including PGE2, by this treatment would negatively affect corpus luteum function. This was assessed by determining luteal expression of 3bhydroxysteroid-dehydrogenase (3bHSD), a key steroidogenic enzyme allowing for the synthesis of P4, and the course of peripheral P4 concentrations. As has been shown in a number of studies, the volume of cell nuclei is indicative of cellular activity [15–18]; in the case of Leydig cells, this parameter has also been used for assessment of cellular activity [19,20]. Thus in this study, determination of the area of luteal cell nuclei served as a morphological parameter to test for effects of treatment. 2. Materials and methods Animal experiments were approved by the responsible ethics committee with permit 54/2008, University of Warmia and Mazury in Olsztyn, Poland. 2.1. Experimental design The study comprised a total of 30 bitches of various breeds and age (2–7 years). All dogs were observed for onset of spontaneous estrus, which was monitored by vaginal cytology and P4 assay, using a commercially available chemiluminiscence kit (Progesteroene II, Roche Diagnostics, Mannheim, Germany); the day when P4 concentration had just exceeded 5 ng/mL plasma was considered the day of ovulation (Day 0). Dogs were randomly assigned to 4 treatment groups (TG) and 5 control groups (CG). Starting with Day 0, dogs of TG were treated for 5 (n ¼ 4), 10 (n ¼ 3), 20 (n ¼ 2; originally 3 dogs had been assigned to the group, but from the tissue sample of one dog designated for determination of the area of the lutein cell nuclei, the CL had been accidently cut off), or 30 (n ¼ 3) days with 10-mg firocoxib/kg body weight per day (firocoxib is a highly selective COX2 inhibitor marketed as Previcox by Merial Ltd.; the recommended daily dose is 5 mg/kg body weight per day).
Dogs of the CG received a placebo for 0 (n ¼ 3), 5 (n ¼ 5), 10 (n ¼ 3), 20 (n ¼ 3), or 30 (n ¼ 3) days. Blood samples were collected for P4 measurements at 2 -day intervals after ovulation until ovariohysterectomy. The blood plasma obtained was stored at 20 C until assay, which was within 4 to 5 months. The animals were ovariohysterectomized for collection of corpora lutea on the last day of treatment. The corpora lutea were trimmed off the connective tissue, divided into 2, with 1 part being preserved for mRNA analysis as described earlier [11] and the other part being fixed in 10% phosphate–buffered formalin and paraffin embedded, applying standard procedures. 2.2. Morphological assessment of luteal cell nuclei area Assessment was on 1 to 2 hematoxylin-eosin–stained slides (3–4 mm) and up to 10 sight fields, depending on the quality of the slide and the number of nuclei meeting the criteria set. Hundred nuclei per dog were evaluated using magnification 400 and the Leica Image Manager IM 1000 software program (Leica MIcrosysteme Vertrieb GmbH, Wetzlar, Germany). To be considered for evaluation, a luteal cell had to meet the following criteria: a clearly detectable excentrically located round and/or ellipsoid nucleus with sharp boundaries and 1 or 2 clearly visible nucleoli. The half axes “a” and “b” were determined and the area (A) was calculated using the equation A ¼ pab. 2.3. Progesterone assay Progesterone was determined by an in-house competitive radioimmunoassay following the procedure outlined by Hoffmann et al. [21]. Intra-assay coefficients of variation varied between 8.8% and 9.6% and interassay coefficients of variation between 10.2% and 14.7%. 2.4. Immunohistochemical detection of 3bHSD The immunoperoxidase immunohistochemistry protocol was applied following the previously published protocol [9,22]. The primary antibody was a rabbit polyclonal antiserum against human placental 3bHSD used at 1:3000 dilution (a gift from Dr. J.I. Mason, University of Edinburgh, UK [23]); serum from nonimmunized rabbit served as negative control. Additionally, sections omitting the primary antibody were included. Peroxidase-conjugated goat antirabbit IgG (Vector Laboratories, Burliname, CA) were used as a secondary antibody. Signals were enhanced using the avidin-biotin-peroxidase complex (Vectastain ABC kit, Vector Laboratories), and signals were detected using Liquid DAB þ substrate kit (Dako Schweiz AG, Baar, CH). Sections were counterstained with hematoxylin and embedded in Histokit (Assistant, Osterode, Germany). 2.5. Semiquantitative assessment of 3bHSD mRNA expression Total RNA was isolated using TRIzol reagent (Invitrogen, Carlsbad, CA) following the manufacturer’s instructions. The RNA content was measured with a NanoDrop 2000C
T. Janowski et al. / Theriogenology 82 (2014) 915–920
spectrophotometer (Thermo Fisher Scientific AG, Reinach, CH). All RNA samples were DNase-treated with RQ1 RNasefree DNase (Promega, Dübendorf, CH), according to the supplier’s protocol, to remove potential genomic DNA contamination, which may occur when applying TRIzol. The reverse transcriptase (RT) reaction with random hexamers used as primers and other reagents for cDNA synthesis from Applied Biosystems (Foster City, CA) was followed our previously published protocol [2,17]. The RT reactions were conducted in an Eppendorf Mastercycler (Vaudaux-Eppendrf AG, Basel, CH) according to the following protocol: 8 minutes at 21 C and 15 minutes at 42 C; the reaction was then stopped by incubation for 5 minutes at 99 C. Real-time (TaqMan) polymerase chain reaction (PCR) reactions were performed using an automated fluorometer ABI PRISM 7500 Sequence Detection System (Applied Biosystems), following our previously published protocol and the manufacturer’s instructions [9,24]. The following negative controls were used: water instead of cDNA and the socalled RT minus control (i.e., samples that were not reverse transcribed). Three independent endogenous reference genes (GAPDH, 18SrRNA, and cyclophilin A) were involved in the semiquantitation of 3bHSD expression in the comparative method (DDCT method) described previously [11,24] and as advised by Applied Biosystems, the manufacturer of the ABI 7500 Fast Real Time PCR System. Reactions were set up to achieve approximately 100% efficiency using the CT slope method. Complementary DNA corresponding to 100-ng DNase-treated RNA per sample was used in reactions with Fast Start Universal Probe Master (ROX; Roche Diagnostics AG, Schweiz). Selected PCR products were sent for commercial sequencing (Microsynth). The following TaqMan systems (primers and 6-carboxyflurescein–labeled and 6carboxytertamethyl rhodamine–labeled probes) were purchased from Microsynth: 3bHSD forward: 50 -GGG TAC TCA GCT CCT GTT GGA A-30 , 3bHSD reverse: 50 -GCC ACC TCT ATG GTG CTG GTA T-30 , 3bHSD TaqMan probe: 50 -TGC CCA GGC TAG TGT GCC GAT CTT-30 (gene bank accession number
917
AY739720); GAPDH forward: 30 -GCT GCC AAA TAT GAC GAC ATC A-30 , GAPDH reverse: 50 -GTA GCC CAG GAT GCC TTT GAG30 , TaqMan probe: 50 -TCC CTC CGA TGC CTG CTT CAC TAC CTT30 , TaqMan probe: 50 -TCC CTC CGA TGC CTG CTT CAC TAC CTT30 (gene bank accession number AB028142); 18SrRNA forward: 50 -GTC GCT TCC TCT CCT ACT-30 , 18SrRNA reverse: 50 GGC TGA CCG GGT TGG TTT-30 , and 18SrRNA TaqMan probe: 50 -ACA TGC CGA CGG GCG CTG AC-30 (gene bank accession number FJ797658). The canine-specific cyclophilin A TaqMan gene Expression Assay was commercially available from Applied Biosystems (Prod. No. Cf03986523_gH). 2.6. Statistical evaluation Progesterone data were skewed to the right and consequently the geometric mean (Xg) and deviation factor (DF) were calculated. To test for the effect of treatment and time and to account for the decreasing number of samples from Days 0 to 30, a stepwise ANOVA with repeated measures of time was performed, followed by a t test for independent samples. For the area of luteal cell nuclei, the arithmetic mean (X) and standard deviation were calculated; effects of treatment were assessed by a 2-factorial ANOVA with the factors being treatment and day of ovariohysterectomy. The progesterone concentrations and size of lutein cell nuclei were assessed by correlation analysis. As expression levels of 3bHSD mRNA were skewed to the right, the geometric mean (Xg) and DF were calculated; To test for effects of treatments an unpaired, 2-tailed Student t test was performed. 3. Results 3.1. Progesterone concentrations Progesterone concentrations showed a high variability as indicated by the DF varying between 1.3 and 4.1. As shown in Figure 1, from Day 4 onward, the calculated
Fig. 1. Course of progesterone concentrations in treated group and control group dogs. Values given as (Xg deviation factor1) a:b P < 0.05.
918
T. Janowski et al. / Theriogenology 82 (2014) 915–920
Fig. 2. Development of area luteal cell nuclei; values given as (X standard deviation).
mean values (Xg) in the TG ranged below those of the CG. However, the effect of treatment was not significant (stepwise ANOVA, P > 0.05) and only on Day 10 was the observed difference between the TG and CG significant with P < 0.05. For both groups, the effect of time was highly significant (P < 0.0001).
Fig. 3. Expression of 3b-hydroxysteroid-dehydrogensae (3bHSD) as determined by semiquantitative (TaqMan) reverse transcription quantitative polymerase chain reaction in luteal tissue samples of control and treatment groups; values given as (Xg deviation factor1). Bars with (*) differ at P < 0.04 and bars with (**) differ at P < 0.05.
3.2. Area of luteal cell nuclei With P < 0.001 (ANOVA), there was a significant effect of treatment. After an initial increase from Day 0 to Day 5,
Fig. 4. Immunohistochemical localization of 3b-hydroxysteroid-dehydrogensae in canine luteal samples in control and treatment groups. Representative pictures are shown for Days 5, 20, and 30 after ovulation (oral). Immunohistochemical signals are localized in cytoplasm of lutein cells.
T. Janowski et al. / Theriogenology 82 (2014) 915–920
the values did not further change in the TG whereas there was a distinct increase in the CG on Days 10 and 20; similar values were again observed on Day 30 (Fig. 2). Progesterone concentrations and the area luteal cell nuclei showed a highly significant (P < 0.001) interaction indicating that the difference observed between the CG and TG depended on the day of castration. Progesterone values and area of luteal cell nuclei were positively correlated (P < 0.05) in the CG but not in the TG. 3.3. Expression of 3ßHSD The expression of 3bHSD was clearly detectable, both at the mRNA and at the protein levels in all luteal samples investigated, displaying high variability between individual animals (Fig. 3). Treatment with Previcox significantly reduced expression of 3bHSD-mRNA on Days 20 and 30 (P < 0.04) compared with the respective controls. Similar effects were observed at the protein level by immunohistochemistry; the evenly distributed signals, which were localized to the cytoplasm of luteal cells, appeared to be weaker after treatment (Fig. 4). 4. Discussion A likely role of PGE2 as a luteotropic factor has been addressed in a number of articles. Thus, Christenson et al. [25] concluded from their study in pigs that a conceptusassociated increase in uterine PGE2 secretion is consistent with the role of PGE2 in the local stimulation of luteal function during early pregnancy. In an in vitro assay, Weems et al. [26] showed that addition of PGE1 and PGE2 to Day 8 ovine corpora lutea linearly stimulated progesterone secretion, matching the effect of LH. Their observations of 88 to 90 days ovine corpora lutea of pregnancy led to the conclusion that these corpora lutea produce their own luteotropin which is PGE2. In the cow intraluteal implants containing PGE2 prevented or inhibited luteolysis when implanted on Day 13 post-estrus, with the effect apparently based on the prevention of the loss of luteal LH receptors, as delineated from their expression on the mRNA level [27]. In the dog, the present study has provided further evidence that prostaglandins, and presumably mostly PGE2, seem to act luteotropically in the early phase of diestrus in the dog. Thus, in addition to our preceding expression and in vitro studies [9–11], the present study clearly shows that inhibition of prostaglandin synthesis by applying a COX2 (PTGS2) inhibitor negatively affects the corpus luteum function in vivo. A reduced cellular activity of lutein cells is indicated by the reduced area of luteal cell nuclei after Day 5 of treatment. This situation was mirrored in the significantly reduced expression of 3bHSD on both the mRNA and the protein level. Also, the resulting decrease of progesterone mean levels observed in peripheral plasma of treated animals mirrors impaired luteal cell function. The fact that the latter effect was not significant on a group basis but only on Day 10 on a pairwise comparison may be explained by the large variation in individual progesterone concentrations and that the statistics applied (stepwise ANOVA) had to account for the continuous reduction of blood samples available from Day 0 to Day 30.
919
To our knowledge, this is the first in vivo study substantiating a luteotropic activity of PGE2 in the dog during the early diestrus phase. On the basis of data concerning expression of the prostaglandin system in pregnant bitches [28], the same situation seems to apply for the early pregnancy corpus luteum. Clearly, inhibition of COX2 (PTGS2) inhibits the steroidogenesis in the CL as indicated by the decreased expression of 3bHSD. It should be noted that this effect was observed when only double the recommended therapeutic dose was used. However, additional information on the underlying mechanisms is necessary to allow further functional conclusions. Thus, treatment must have affected activity of COX2 (PTGS2) not only in the corpora lutea but also in other organs with the effect on the ovary itself still needing to be further defined. However, from a clinical point of view, treatment of bitches with a COX2 inhibitor may also impair luteal ovarian function depending on the dosage. Acknowledgments This study was supported by The Polish Ministry of Science and Higher Education as research project Number NN308225936. The technical assistance of Stefanie Ihle and Elisabeth Högger is highly appreciated. References [1] Concannon PW, Butler WR, Hansel W, Knight PJ, Hamilton JM. Parturition and lactation in the bitch: serum progesterone, cortisol and prolactin. Biol Reprod 1978;19:1113–8. [2] Hoffmann B, Höveler R, Hasan SH, Failing K. Ovarian and pituitary function in dogs after hysterectomy. J Reprod Fertil 1992;96:837–45. [3] Olson PN, Bowen RA, Behrendt MO, Olson JD, Nett TM. Concentrations of progesterone and luteinizing hormone in the serum of diestrus bitches before and after hysterectomy. Am J Vet Res 1984; 45:149–53. [4] Okkens AC, Dieleman DJ, Bevers MM, Lubberink AA, Willemse AH. Influence of hypophysectomy on life span of the corpus luteum in the cycling dog. J Reprod Fertil 1986;77:187–92. [5] Concannon PW. Effects of hypophysectomy and of LH administration on luteal phase plasma progesterone levels in the beagle bitch. J Reprod Fertil 1980;58:407–10. [6] Concannon PW, Weinstein P, Whaley S, Frank D. Suppression of luteal function in dogs by luteinizing hormone antiserum and by bromocryptine. J Reprod Fertil 1987;81:175–80. [7] Okkens AC, Bevers MM, Dieleman SJ, Willemse AH. Evidence for prolactin as the main luteotropic factor in the cyclic dog. Vet Q 1990;12:193–201. [8] Hoffmann B, Büsges F, Engel E, Kowalewski MP, Papa P. Regulation of corpus-luteum function in the bitch. Reprod Domest Anim 2004; 39:232–40. [9] Kowalewski MP, Schuler G, Taubert A, Engel E, Hoffmann B. Expression of cyclooxygenase 1 and 2 in the canine corpus luteum during diestrus. Theriogenology 2006;66:1423–30. [10] Kowalewski MP, Mutembei HM, Hoffmann B. Canine prostaglandin E2 synthase (PGES) and its receptors (EP2 and EP4): expression in the corpus luteum during dioestrus. Anim Reprod Sci 2008;109:319–29. [11] Kowalewski MP, Fox B, Gram A, Boos A, Reichler I. Prostaglandin E2 functions as a luteotrophic factor in the dog. Reproduction 2013; 145:213–26. [12] Kowalewski MP, Mutembei HM, Hoffmann B. Canine prostaglandin F2a receptor (FP) and prostaglandin F2a synthase (PGFS): molecular cloning and expression in the corpus luteum. Anim Reprod Sci 2008;107:161–75. [13] Diaz FJ, Anderson LE, Wu YL, Rabot A, Tsai SJ, Wiltbank MC. Regulation of progesterone and prostaglandin F2a production in the CL. Mol Cell Endocrinol 2002;191:65–80. [14] Wiltbank MC, Ottobre JS. Regulation of intraluteal production of prostaglandins. Reprod Biol Endocrinol 2003;1:91.
920
T. Janowski et al. / Theriogenology 82 (2014) 915–920
[15] Benninghoff A. Funktionelle Kernschwellung und Kernschrumpfung. Anatomische Nachrichten 1950;1:50–2. [16] Krantz H. Kern und Funktion I: Die Kerngröße und ihre Abhängigkeit von äußeren und inneren Faktoren. Z Zellforsch Mikrosk Anat 1951;35:425–57. [17] Polster C. Karyometrische und karyologische Untersuchungen an den Hauptstück- epithelien der Rattenniere bei experimenteller Eiweißnephrose. Virchows Arch Path Anat 1959;332:420–30. [18] Tietze KW, Inthraphuvasak J, Hiersche HD. Funktionelle Morphologie des Endometrium corporis uteri. Arch Gynakol 1970;209:331–8. [19] Gentil M, Hoffmann B, Spang A, Failing K, Goericke-Pesch S. Restart of steroidogenesis in dogs during recrudescence of testicular function following downregulation with a GnRH-agonist implant. Cell Tissue Res 2012;350:513–23. [20] Kronthaler O. Karyometrische Untersuchungen über den Einfluss einiger Wirkstoffe auf die Interstitialzellen von Gonaden. Diss med vet; Ludwig-Maximilians-Universität München; 1960. [21] Hoffmann B, Kyrein HJ, Ender ML. An efficient procedure for the determination of progesterone by radioimmunoassay applied to bovine peripheral plasma. Horm Res 1973;4:302–10. [22] Kowalewski MP, Mason JI, Howie AF, Morley SD, Schuler G, Hoffmann B. Characterization of the canine 3beta-hydroxysteroid dehydrogenase and its expression in the corpus luteum during diestrus. J Steroid Biochem Mol Biol 2006;101:254–62. [23] Lorence MC, Murry BA, Trant JM, Mason JI. Human 3ß-hydroxysteroid-dehydrogenase/D 5-4 isomerase from placenta: expression
[24]
[25]
[26]
[27]
[28]
in nonsteroidogenic cells of a protein that catalyzes the dehydrogenation/isomerisation of C21 and C19 steroids. Endocrinlogy 1990; 126:2493–8. Kowalewski MP, Meyer A, Hoffmann B, Aslan S, Boos A. Expression and functional implications of peroxisome proliferator-activated receptor gamma (PPARg) in canine reproductive tissues during normal pregnancy and parturition and at antiprogestin induced abortion. Theriogenology 2011;75:877–86. Christenson LK, Farley DB, Anderson LH, Ford SP. Luteal maintenance during early pregnancy in the pig: role for prostaglandin E2. Prostaglandins 1994;47:61–5. Weems YS, Bridges PJ, Tanaka Y, Sasser RG, LeaMaster BR, Vincent DL, et al. PGE1 or PGE2 not LH regulates secretion of progesterone in vitro by the 88-90 day ovine corpus luteum of pregnancy. Prostaglandins 1997;53:337–53. Weems YS, Alejandro Arreguin-Arevalo J, Nett TM, Vann RC, Ford SP, Bridges PJ, et al. In vivo intra-luteal implants of prostaglandin (PG) E1 or E2 (PGGE1, PGE2) prevent luteolysis in cows: I. Luteal weight, circulating progesterone, mRNA for luteal luteinizing hormone (LH) receptor, and occupied and unoccupied luteal receptor for LH. Prostaglandins and other Lipid Mediat 2011;95:34–44. Kowalewski MP, Beceriklisoy HB, Aslan S, Agaoglu AR, Hoffmann B. Time related changes in luteal prostaglandin synthesis and steroidogenic capacity during pregnancy, normal and antiprogestin induced luteolysis in the bitch. Anim Reprod Sci 2009;116:129–38.