Steroidogenic capacity of the placenta as a supplemental source of progesterone during pregnancy in domestic cats

Background Until recently, the corpus luteum (CL) was considered to be the main source of progesterone (P4) during pregnancy in the domestic cat (Felis catus). However, other possible sources of P4 have not been ruled out. Although feline placental homogenates were found to be capable of synthesizing P4, expression of the respective steroidogenic enzymes has not been investigated at the molecular level. Therefore, in the present study, expression of the two major factors involved in the synthesis of P4 - 3beta-hydroxysteroid dehydrogenase (3betaHSD) and steroidogenic acute regulatory protein (StAR) - was investigated in the feline CL and placenta during the course of pseudopregnancy and pregnancy. Methods The mRNA levels of StAR and 3betaHSD were determined using Real Time PCR and their localizations were determined by immunohistochemistry. Placental P4 concentrations, after ethyl extraction, were measured by EIA. Results Luteal 3betaHSD and StAR mRNA levels were strongly time-dependent, peaking during mid-pregnancy. The placental 3betaHSD mRNA level was significantly upregulated towards the end of pregnancy. In the CL, 3betaHSD and StAR protein were localized in the luteal cells whereas in the placenta they were localized to the maternal decidual cells. Placental P4 concentrations were low in early pregnant queens, but increased along with gestational age. Conclusions These results confirm that the placenta is an additional source of P4 in pregnant queens and can thereby be considered as an important endocrine organ supporting feline pregnancy.


Background
Female domestic cats (Felis catus) are traditionally classified as seasonally polyestrous with ovulation provoked by coitus [1,2]. However, in as many as 50% of domestic cats, ovulation occurs without cervical or vaginal stimulation [3,4]. If ovulation is not followed by pregnancy, the queen enters pseudopregnancy during which the corpora lutea (CL) remain active. Hence, the reproductive cycle in the domestic cat differs significantly from that in ungulates [5,6] and dogs [7,8], in which spontaneous ovulation is followed by the formation of CL in each estrous cycle. In the cat [9], rabbit [10] and rat [11], the luteal phase of non-pregnant animals lasts only one-half of a normal gestation. This means elevated progesterone (P 4 ) levels measured in circulating blood for 35-40 days in pseudopregnant and for approximately 60-65 days in pregnant queens. In contrast, in the dog [12], mink [13] and ferret [14], the length of the luteal phase is similar in the presence or absence of pregnancy. Progesterone levels in pregnant and non-pregnant dogs are similar within the first 60 days after ovulation [15]. Moreover, there are no indications of an active luteolytic principle in non-pregnant dogs, and, the canine CL appears to be devoid of PGF 2α -synthase (PGFS) activity [16,17]. Consequently, luteal regression in non-pregnant bitches seems to be a passive degenerative process in the absence of endogenous PGF 2α [16,17]. The domestic cat thus seems to have a reproductive advantage over some other carnivores because its shorter luteal phase allows for an earlier return to estrual cyclicity. In the cat, ovarian activity returns within 7 to 10 days following pseudopregnancy [10]. The histological and endocrine patterns of early CL formation are the same in pregnant and pseudopregnant cats until days 10-12 after coitus [18]. However, after days 12-13, which coincides with the time of implantation [19,20], plasma P 4 concentration slowly declines in pseudopregnant queens, whereas during pregnancy it remains on a plateau until day 30 and then gradually declines towards the end of gestation [21].
Until the 1970s, the CL was considered to be the main source of P 4 during feline gestation. However, the finding that ovariectomy in queens after day 45 either did not interrupt pregnancy [22] or only did in some animals [23] supported an assumption that the placenta is a source of steroidogenic enzymes [18]. Subsequently, feline placental homogenates were found to be capable of synthesizing P 4 [19]; however, the expression of steroidogenic factors or P 4 content has not been examined in the placenta to date.
Cholesterol is the precursor of all steroid hormones. Steroidogenesis is regulated by steroidogenic acute regulatory (StAR) protein, which controls cholesterol transport from the outer to the inner mitochondrial membranes of the steroidogenic cells [24]. Conversion of pregnenolone to P 4 is a step catalyzed by 3β-hydroxysteroid dehydrogenase/isomerase (3βHSD). In this study, expression of these two factors was investigated in the feline CL and placenta throughout pregnancy and pseudopregnancy. Privately owned queens (n = 36, mixed breed, proven health status), aged 18 ± 5 months, housed individually or in pairs, with or without contact with an intact male, were enrolled in this study. No pharmacological treatment was performed to provoke ovulation in the animals. In the case of pregnant queens, an ovulation was provoked with coitus. Queens were checked daily for behavioral signs of estrus (treading of the hind feet, lordosis and tail deflection). The first day of silencing of estrus behavior in the queens was considered as Day 1 of the luteal phase. Reproductive tracts were collected from 13 pseudopregnant cats at early (n = 4, days 3-5), mid (n = 4, days 10-15) and late phases of pseudopregnancy (n = 5, days [25][26][27][28][29][30][31][32][33][34][35], and from 12 pregnant cats at early (n = 5), mid (n = 4) and late (n = 3) gestation: 1.5-2, 3-4 and 6-8 weeks after mating, respectively. For extraction of P 4 from placentas, an additional 11 cats at different gestational ages were subjected to ovariohysterectomy (OHE), which was done with the owners' request and consent. The stage of the estrous cycle was further confirmed by endocrine analysis, according to the P 4 values reported in the literature [10,25], and macroscopic observation of the ovaries and uterus. Before surgery, blood samples were collected from the cephalic vein into EDTA-containing tubes (Tyco Healthcare Group LP, Mansfield, USA) and transported to the laboratory at 4°C. Plasma obtained after blood centrifugation (3500 x g, 10 min) was frozen at −20°C until P 4 measurement using direct enzyme immunoassay (EIA). Confirmation of gestational age was done according to the P 4 values reported in the literature [18]. In addition, measurements of the crown-rump length in fetuses and uterine ampullae diameter or length were performed [26,27].

Animals and tissue preservation
Tissues were washed immediately after surgery with sterile saline to remove blood contamination, then placed into fresh sterile saline at 4°C and transported to the laboratory within 1 h. The CL were removed from a randomly selected ovary among the pairs belonging to the same luteal stage. The second ovary was fixed in buffered 4% formaldehyde for 24 h, dehydrated and wax-embedded. Sections (2-3 μm thick) were used for evaluation of protein concentrations by immunohistochemistry. In pregnant animals, uterine horns were slit longitudinally and fragments of placenta were separated, washed in fresh saline to remove blood, preserved overnight at 4°C with RNAlater (Ambion Biotechnologie GmbH, Wiesbaden, Germany), and then stored at −80°C until total-RNA extraction. mRNA levels were determined using Real Time PCR, while localization of 3βHSD and StAR protein was examined by immunohistochemistry.
Cloning of feline 3βHSD by reverse transcription (RT) and rapid amplification of cDNA ends polymerase chain reaction (RACE PCR) Total RNA was isolated from feline CL using Trizol Reagent according to the manufacturer's instructions (Gibco-BRL, Life Technologies, Karlsruhe, Germany). The whole procedure was carried out as described before for canine 3βHSD [28]. For initial RT-PCR, 0.2 μg of total RNA was used. An alignment of the known canine 3βHSD sequence (GenBank accession number: AY739720) against the available online feline genomic sequence [29] was performed using BLAST W software to obtain feline-specific PCR product (primers 1-2; Table 1). Integrity of RNA was checked by amplification of the housekeeping gene β-actin (primers 3-4, Table 1).
First strand cDNA synthesis was performed with the PowerScript Reverse Transcriptase kit (BD Biosciences Clontech GmbH, Heidelberg, Germany) using 0.6 μg of total RNA. Subsequently, the SMART RACE cDNA Amplification kit (BD Biosciences Clontech GmbH) was used with gene-specific primers (GSP, primers 5-6; Table 1) in combination with universal primer mix (UPM, primers 7-8; Table 1) supplied by the manufacturer of the SMART RACE Kit. Overlapping products of the missing cDNA coding fragments of the 5 0 and 3 0 ends were amplified. After initial denaturation at 94°C for 1 min, the reactions were run for 35 cycles (94°C for 1 min, annealing at 65°C for 2 min, elongation at 72°C for 3 min), and the final extension was at 72°C for 10 min. Finally, RT-PCR for 40 cycles was performed at an annealing temperature of 57°C with specific primers (primers 9-10; Table 1) located at both ends of the open reading frame (ORF). All PCR products were visualized on a 1.5% ethidium bromide-stained gel, purified with a Qiaex II agarose gel extraction kit (Qiagen GmbH, Hilden, Germany), ligated into pGEM-T vector (Promega, Dübendorf, Switzerland), multiplied in XL1 BLUE competent cells (Stratagene, La Jolla, CA, USA) and sequenced (Microsynth, Balgach, Switzerland). Finally, the cloned cDNA sequence was submitted to GenBank with the accession number: JF794032, Felis catus 3β-hydroxysteroid dehydrogenase mRNA, complete cds.

Cloning of feline StAR cDNA
The procedure leading to characterization of feline StAR protein was carried out as previously described for canine StAR protein [30]. Total RNA was obtained from three feline CLs collected at early, mid and late phases of pseudopregnancy. The DNase-treatment was performed with RQ1 RNase free DNase (Promega), and the RT-PCR was done with the GeneAmp Gold RNA PCR kit (Perkin-Elmer Applied Biosystems GmbH, Weiterstadt, Germany), all as previously described [30]. Primers for qualitative PCR were obtained from the alignment of the canine sequence with GenBank accession number EF522840 using an online available feline genomic sequence. Using primer pairs (13-14; Table 1), a PCR product comprising 886 bp of feline StAR protein was amplified. PCR conditions were as follows: initial denaturation at 95°C for 10 min, then reactions were run for 40 cycles consisting of denaturation at 94°C for 1 min, annealing at 56°C for 1.5 min and elongation at 72°C for 1.5 min; the final extension was at 72°C for 10 min. PCR products were visualized on a 1.5% ethidium bromide-stained gel, purified with the Qiaex II agarose gel extraction kit (Qiagen GmbH, Hilden, Germany), ligated into the pGEM-T vector (Promega) multiplied in XL1 BLUE competent cells (Stratagene, 11. 3βHSD specific for Real Time PCR La Jolla, CA, USA) and sequenced (Microsynth, Balgach, Switzerland). The cloned cDNA sequence was submitted to GenBank with the accession number JF800676 (Felis catus Steroidogenic Acute Regulatory protein (StAR) mRNA cds).

Real Time PCR
The levels of mRNA expression of target genes were examined by Real-Time PCR using specific primers for 3βHSD, StAR and cyclophilin (Cyc). Among several different genes, including β-Actin, GAPDH and Cyc, the last one was employed as a reference because the differences in Cyc expression between different samples did not exceed two cycles. All primers were purchased from Microsynth (Balgach, Switzerland). The forward and reverse sequences used for quantitative Real-Time PCR and the GenBank accession numbers are given in The total reaction volume was 20 μl containing: 1 μl cDNA (200 ng), 250 nM each of forward and reverse primers, and 10 μl SYBR Green PCR Master Mix. Real time PCR was carried out as follows: initial denaturation (10 min at 95°C), followed by 40 cycles of denaturation (15 s at 95°C) and annealing (1 min at 60°C). After each PCR reaction, melting curves were obtained by stepwise increases in temperature from 60 to 95°C to ensure single product amplification. The presence of the product was also confirmed by electrophoresis on 2% agarose gel. Relative quantification was performed by normalizing the signals of target genes with the Cyc signal by the Miner method for quantifying qRT-PCR results using calculations based on the kinetics of individual PCR reactions [31].

Immunohistochemistry
Immunohistochemistry was done according to the procedure described by Kowalewski et al. [30]. Firstly, the sections were deparaffinized and rehydrated, and then incubated in citrate buffer ( Hematoxylin was used to counterstain the sections, then they were mounted in Histokit (Assistant, Osterode, Germany). Isotype control was done to avoid false positive results. The luteal and placental sections were incubated with serial dilutions of pre-immunized rabbit serum (starting at 1:1000) or pre-immunized mouse serum (starting at 1:50). No positive staining was observed at a 1:3000 dilution of pre-immunized rabbit serum or at a 1:100 dilution of pre-immunized mouse serum.

Placental progesterone extraction
Progesterone extraction was carried out using a modification of a methoddescribed previously [35]. Placenta samples were stored at −80°C. The tissues (200-300 mg each) were thawed and homogenized in glass vials using a tissue disruptor with 400 μl EIA buffer containing 45 μl 1 N HCl. Next, 3 ml of ethyl petroleum was added to each sampleand it was shaken for 10 min, then samples were incubated at −20°C for 4 h. Afterwards, the supernatant was collected and evaporated to dryness under nitrogen at 40°C. Finally, 400 μl of EIA buffer with 0.1% BSA was added. The suspension was frozen at −20°C until P 4 measurement by EIA.

Progesterone determination
Assessment of P 4 concentration in plasma followed the methodology previously described [36]. Horseradish peroxidase-labeled P 4 was used at a final dilution of 1:75.000. Anti-P 4 serum (final dilution of 1:100 000) was kindly donated by Dr. Stanisław Okrasa, University of Warmia and Mazury, Olsztyn, Poland. The ED 50 for P 4 was 4.2 ng/mL and the assay ranged from 0.39 to 100 ng/mL. The intra-and inter-assay CVs were 4.2% and 9.3%, respectively. Progesterone concentration in plasma was expressed as ng per mL (ng/mL) and in placenta as ng per g (ng/g).

Statistics
To test the effect of different stages of pregnancy or pseudopregnancy on mRNA levels, the Kruskal-Wallis Test (a nonparametric ANOVA) was used followed by the Newman-Keuls Multiple Comparison Test using the statistical software program GraphPad5 (GraphPad PRISM v 5.0; GraphPad Software Inc., San Diego, CA, USA). Progesterone concentrations in plasma are shown as a mean ± standard deviation. To test the effect of the presence or absence of pregnancy on circulating P 4 concentrations within each luteal phase, a nonparametric ANOVA was used followed by the Newman-Keuls Multiple Comparison Test. Significance was defined as values of P < 0.05.

Determination of pregnancy or pseudopregnancy stages
Queens were considered to be at the early luteal phase when the plasma P 4 was between 2 to 8 ng/mL and  Letters "a, b" indicate statistical differences among groups. Asterisks indicate statistical differences in P 4 levels between groups within pregnancy or pseudopregnancy conditions (*P < 0.05, **P < 0.01).
the corpora haemorrhagica were ≥ 2 mm in diameter. Mid-pseudopregnancy was determined when plasma P 4 was > 17 ng/mL and CL were reddish and 3-4 mm in diameter. Late pseudopregnancy was characterized by plasma P 4 < 5 ng/mL and the presence of pale CL. The queens were considered to be at early, mid and late pregnancy when the plasma P 4 was between 2 to 8, > 20 or < 12 ng/mL, respectively. Progesterone concentrations in plasma are shown in Figure 1. The serum P 4 concentration was 3.9 ± 1.4 ng/mL, 39.9 ± 32.2 ng/mL and 4.2 ± 1.2 ng/mL in queens at early, mid and late pseudopregnancy, respectively; and 5.4 ± 2.3 ng/mL, 46.5 ± 19.6 ng/mL and 5.3 ± 6.4 ng/mL in queens at early, mid and late pregnancy, respectively. No statistical differences were seen within each luteal phase examined with respect to the presence or absence of pregnancy (P > 0.05). Placental P 4 contents depended on the gestational age and are shown in Figure 2. The amount of extracted P 4 was 1.7 ± 0.77 ng/g, 5.5 ± 0.58 ng/g and 8.7 ± 3.01 ng/g of tissue from early, mid and late pregnant queens, respectively.

StAR-and 3βHSD-mRNA levels in the feline CL and placenta
Luteal StAR mRNA was strongly time-dependent, with significantly elevated mRNA-levels observed during mid-pregnancy (P < 0.001) (Figure 3). No statistically significant changes were observed among placental StAR-mRNAs detected at early, mid and late gestation (Figure 3).

Localization of StAR protein and 3βHSD in the feline CL and placenta
The spatial localizations of StAR and 3βHSD proteins in the gestational CL and placenta are shown in Figures 5  and 6, respectively. Within the CL, StAR and 3βHSD protein were localized in the luteal cells ( Figure 5A and 5B, respectively). Immunostaining against StAR protein and 3βHSD was observed in early, mid and late pseudopregnancy CL, as well as in early, mid and late pregnancy CL. Based on comparison with the pattern of staining with vimentin ( Figure 6B), the placental cells in which StAR ( Figure 6C) protein and 3βHSD ( Figure 6D) were immunolocalized were identified as cells of Figure 3 StAR mRNA transcription. StAR mRNA transcription in pseudopregnant and pregnant CLs (upper panel) and placenta (lower panel). Letter "a" indicates no statistical differences within pseudopregnancy. Letters "x, y" indicate statistical differences within pregnancy (P < 0.001). Asterisks indicate statistical differences between StAR mRNA levels during the course of pseudopregnancy and pregnancy (***P < 0.001). Letters "a, b" indicate statistical differences within pseudopregnancy (P < 0.05). Letters "x, y" indicate statistical differences within pregnancy (P < 0.01). Asterisks indicate statistical differences between 3β-HSD-mRNA levels during the course of pseudopregnancy and pregnancy (***P < 0.001). mesenchymal origin such as maternal decidual cells ( Figure 6B). Positive staining against both proteins examined was generally observed in mid and late pregnancy placentas. A slight positive staining against 3βHSD was also observed in the fetal component of the placentas.

Discussion
The present results confirm that the placenta is an additional source of P 4 in pregnant queens, possibly acting as an important endocrine organ during pregnancy. Both StAR and 3βHSD were immunolocalized in the placenta and CL in each luteal phase. In the present study, the luteal 3βHSD mRNA expression patterns in pregnant and pseudopregnant animals were similar, both peaking during the mid-luteal phase. However, luteal mRNA was 2.5-fold higher in pregnant than in pseudopregnant queens. Luteal StAR mRNA remained at a relatively constant low level in pseudopregnant animals, but followed the pattern of 3βHSD mRNA expression in pregnant animals. However, no quantitative assessment was performed for the expression of StAR and 3βHSD at the protein level.
A rapid development of the P 4 -producing CL is observed in both pregnant and pseudopregnant cats. Plasma P 4 levels are about the same in pregnant and pseudopregnant queens in the first 10-12 days after coitus [18] but their profiles diverge as early as on day 12 or day 13 after coitus, which is the period of implantation in the cat [18,20]. The luteal StARand 3βHSD-mRNA were highest during the mid-phase of pregnancy. Additionally, increased 3βHSD-levels were also observed at the mid-phase of pseudopregnancy. The high expression of both factors during both mid-luteal phases may be responsible for the strongly elevated P 4 levels in the circulating blood reported previously [2,18,21,37] and observed also in the present study. As reported previously, P 4 values rose rapidly in pseudopregnant queens between Day 4 and Day 9, peaking around Day 14 postovulation with a plateau between Day 9 and Day 23. The range of serum P 4 concentrations at that time was 30 ng/mL to more than 87 ng/mL [2]. The present results concur with data reported earlier, in which strong individual variations in plasma P 4 concentrations were observed. In one queen experiencing a nongravid luteal phase, P 4 peaked at 88 ng/mL, vs. a range of 20.1-28.7 ng/mL observed in other animals. The mean value of plasma P 4 observed in queens at the mid-luteal stage was 39.9 ng/mL and is consistent with the those reported in the literature [2]. However, in another study, P 4 values in pseudopregnant cats were shown to peak at levels of 17 ng/mL on Day 18 postcoitum, followed by a gradual decline to values < 1 ng/mL between Days 30-46 after coitus [25]. In pseudopregnant queens at the late luteal stage, we usually observed distinctly diminished P 4 values compared to the mid luteal phase. Wildt and coworkers [2] reported that the serum P 4 values were  During mid-pregnancy, the serum P 4 values varied significantly between individual cats (range 34 -74.4 ng/mL), with a mean P 4 concentration of 46.5 ng/ mL. Although P 4 concentrations in the blood of midpregnant queens were numerically higher than in non-pregnant animals in the mid-luteal phase, no statistical differences were seen in the present study between these two groups. It cannot be ruled out that these results might differ significantly if the number of females in each group were increased; however, it should also be mentioned that the present data are very similar to those reported by others (for review see: 2, 37). Distinct individual variations in the plasma P 4 levels collected from pregnant cats were previously reported [37], i.e., peak serum P 4 from 13.5 to 57 ng/mL extending over Days 11 to 60 in individual females. After peripheral P 4 peaked, a gradual decline began, usually at Day 44 in most of the queens that were examined. The P 4 levels were maintained at a relatively constant level until a few days before parturition, when they sharply decreased [37]. In the present study, serum P 4 decreasing to values lower than 10 ng/mL (range 3.8 -9.2 ng/mL) was observed in queens in the 7 th and 8 th week of pregnancy; however, in three individuals at the 6 th and 7 th week of pregnancy, the P 4 values were still near 30 ng/mL. Summarizing these data, it might be concluded that, besides visible differences existing in plasma P 4 levels between individuals, the distinct decrease in peripheral P 4 precedes onset of parturition. The successful extraction of P 4 from the placenta further supports its role as an additional source of P 4 during pregnancy in the cat. Placental P 4 concentrations seem to be dependent on gestational age. Thus, the higher P 4 values reported for pregnant queens may result from P 4 supplementation by placental tissue, as hypothesized previously [18]. Still remaining, however, is the question whether the placenta-originating P 4 would be alone able to maintain pregnancy in domestic cats ovariectomized after 45 days of pregnancy. In contrast to cats, in mice it was proven that ovariectomy at any time of pregnancy causes abortion [38]. It was thought that P 4 synthesis is not possible de novo in the rodent placenta. Nevertheless, trophoblast giant cells from mouse placenta were shown to produce P 4 from cholesterol [39]. Moreover, recent studies using molecular tools have shown that P 4 synthesis is possible in rodent maternal decidual cells and occurs upon decidualization, but is terminated at mid-gestation [40]. It is further assumed that this local synthesis of P 4 may act as an immunosuppressive factor at the implantation sites [41]. In contrast in cows placenta-derived progesterone and oestrone are supposed to be auto-or paracrine factors involved in placental growth and differentiation [42]. The synthesis of P 4 was recently demonstrated in uninucleate trophoblast cells and the trophoblast giant cells of the bovine placenta; however, these two kinds of cells were reported to have different steroidogenic capacities [43].
The morphological comparative study in the epitheliochorial and endotheliochorial placenta types published by Leiser and coworkers in 1998 [44] presents a photograph of the typical labyrinthine-like system in the cat placenta with very easily visible and differentiated decidual cells [44]. In the present study, anti-vimentin staining was applied in order to distinguish between cells of mesenchymal origin and other sources, as described and confirmed for the canine and feline placenta by Bezler [34]. Since the StAR-and 3βHSD-positive placental cells were identified as maternal decidual cells, and were found mostly in placentas from the second half of gestation, the present findings on the spatio-temporal expression of the key steroidogenic genes differ from data obtained for other species. Even though some 3βHSDpositive cells were observed in the trophoblast, the strongest signals were present in decidual cells suggesting that these cells may be the main source of steroid synthesis within the feline placenta. Although the reproductive biology characteristics of the dog and cat are often compared, especially on this point distinct differences are observed, because in dogs the CL is the only source of P 4 during both pseudopregnancy and pregnancy [7].

Conclusions
The cellular localization of the steroidogenic factors that are present in the feline placenta differs from their localizations in rodents and cows. The present data confirm previous observations that the feline placenta is a supplemental source of P 4 and could be important for the maintenance of pregnancy in this species.