|
|
||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Female Reproductive Tract |
Department of Reproductive Biology,3 German Primate Center, 37077 Göttingen, Germany
GSF National Research Center for Health and Environment,4 Institute for Experimental Genetics, 85764 Neuherberg, Germany
| ABSTRACT |
|---|
|
|
|---|
corpus luteum function, estradiol, female reproductive tract, ovulatory cycle, pregnancy
| INTRODUCTION |
|---|
|
|
|---|
17HSD1 was the first of the 17HSD enzymes identified and purified [4, 5]. It is a cytosolic enzyme and has a different substrate specificity in rodents and humans. It converts only estrogens in humans, but it can convert androgens equally well as estrogens in rodents. 17HSD1 is known to be expressed in granulosa cells of developing ovarian follicles of probably all mammals [68] and, in addition, in the syncytiotrophoblast of the placenta in primates and some other species [9, 10]. 17HSD1 is thought to be responsible for the preovulatory surge of estradiol arising from the granulosa cells of the dominant follicle and for fetal estradiol supply from the placenta during primate pregnancy [11, 12].
17HSD7 was first identified in the corpus luteum (CL) of the rat ovary as a microsomal prolactin receptor-associated protein, which is upregulated during pregnancy [13, 14]. Its enzymatic properties as an estrogenic 17HSD were discovered later in mice [15]. Results from experiments with mice and rats clearly indicated a specific role for the maintenance of pregnancy, but maintenance of pregnancy and control of CL function are very different in primates and rodents [12, 16]. In rodents, estradiol synthesis takes place exclusively in the CL throughout pregnancy, and the placenta is not a steroidogenic tissue. In humans and nonhuman primates, estradiol synthesis takes place in the CL only during a period at the beginning of pregnancy; later it takes place mainly in the placenta, i.e., there is a luteo-placental shift. According to these species-specific differences, a different regulation of 17HSD7 function might be expected.
The human 17HSD7 cDNA was recently cloned and the genomic structure was determined [17]. The human protein has not been examined yet. The common marmoset monkey has been proven as a suitable animal model to investigate the role of 17HSDs in primate reproductive biology [18, 19]. The aim of our study was to elucidate the role of 17HSD7 compared with that of 17HSD1 in primate reproductive tissues and especially during primate pregnancy where estradiol is known to be a prerequisite of placental and fetal development [20]. There is some phylogenetic evidence that 17HSD7 may have an ancient function in cholesterogenesis apart from its function in estradiol synthesis [21]. For this reason, the expression of aromatase (ARO) was monitored in parallel to 17HSD1 and 17HSD7 in this study to make sure that estrone as a substrate for estradiol synthesis can be provided in a particular tissue.
| MATERIALS AND METHODS |
|---|
|
|
|---|
Adult female and male marmoset monkeys (Callithrix jacchus) were housed pairwise in a controlled environment at the German Primate Center, Göttingen, Germany. Ovarian cycles were controlled by PGF2
-induced luteolysis, and cycling or state of pregnancy were monitored as described earlier by regular assessment of serum progesterone [22]. The presence of corpora lutea and initiation of pregnancy were additionally confirmed by ultrasound examination of the trained animals (Esaote ultrasound system, Esaote Biomedica, Munich, Germany) [23, 24]. CL were obtained by luteectomy and placental tissue from Cesarean sections. For the determination of tissue-specific expression, various organs were collected from healthy animals that were culled as part of the colony management program. Tissues were immediately frozen in liquid nitrogen and stored at -80°C.
All experimental protocols received approval of the local animal experiment committee.
Cloning and Sequencing of cDNA
Total RNA was extracted from marmoset placenta using an RNeasy Midikit (Qiagen, Hilden, Germany) and digested by DNase I (Promega, Mannheim, Germany) for 30 min at 37°C. Complementary DNA was synthesized by SuperscriptII reverse transcriptase with oligo(dT)1218 primers (Invitrogen Life Technologies, Karlsruhe, Germany) and 2 µg of template RNA. Marmoset-specific polymerase chain reaction (PCR) fragments were generated by PCR using Turbo Pfu polymerase (Stratagene, Heidelberg, Germany). Primers were derived from the following human sequences: For 17HSD1 [25], forward, 5'-CAG CTG GAC GTA AGG GAC TC-3'; reverse, oligo (dT) anchor primer (39 bp) from 5'/3'RACE Kit (Roche Molecular Biochemicals, Mannheim, Germany), product length 1219 bp; and forward, 5'-GGC CTG CAC TTG GCC GTA CG-3'; reverse, 5'-GGC CTG CAG CAT CCG CAC AG-3'; product length 330 bp. For 17HSD7 [17], forward, 5'-CTG CTG AGG TCA CCA TTG TA-3'; reverse, 5'-CCA GAT GAG CTG AGA TGG AT-3'; product length 344 bp. For ARO [38], forward, 5'-TAA CAT CAC CAG CAT CGT GCC-3'; reverse, 5'-CCG AAT CGA GAC CTG TAA TG-3'; product length 318 bp). The fragments were cloned into the vector PCR-Script (PCR-Script Amp Cloning Kit, Stratagene). Sequencing reactions were performed by Sequiserve, Vaterstetten, Germany.
Determination of Gene Structures
A genomic cosmid library with 1.5-fold coverage and an average insert length of 32 kb was obtained from the resource center Primary Database of the German Human Genome Project at the Max Planck Institute for Molecular Genetics (Berlin, Germany). The library was screened with randomly primed [32P]dCTP-labeled probes corresponding to the full length of the coding sequences of 17HSD1 or 17HSD7, respectively. After hybridization, filters were washed and exposed to Kodak X-Omat film (Kodak GmbH, Stuttgart, Germany) for 12 h at room temperature as described [17]. Four positive clones were detected for each gene. Two 17HSD1 clones and one 17HSD7 clone could be verified by PCR using primers in the 3' region of the cDNA.
Northern Hybridization
For Northern analysis total RNA was extracted from various marmoset tissues as indicated in the figure legends using the RNeasy Midikit (Qiagen). Marmoset-specific digoxigenin-labeled riboprobes were prepared by in vitro transcription with T3-polymerase (DIG RNA labeling kit, Roche) from specific cDNA fragments cloned as described above. Total RNA (5 µg/lane) was separated on 1% formaldehyde agarose gels and transferred to positively charged nylon membranes(according to the DIG User's Manual, Roche). Prehybridization and hybridization were conducted at 68°C. The membranes were subjected to stringent washing and chemiluminescent detection with the substrate CDP Star (Roche). Densitometric analysis was performed by digitalization of x-ray films using the software Quantity One 4.0 and the Gel Doc video documentation system (Biorad, Munich, Germany).
Antibody Preparation
Polyclonal antipeptide antibodies directed against 17HSD7 and ARO were generated by immunization of rabbits (Pineda Antikörper Service, Berlin, Germany). The synthetic peptides [NH2-C]KMDLDEDTAEKFYQK[-CONH2] corresponding to amino acids 301315 of marmoset and human 17HSD7 and [NH2-C]KPNEFTLENFAKNVPYR[-CONH2] corresponding to amino acids 409425 of human ARO were coupled to keyhole limpet hemocyanin and were used as antigens. All antisera were purified by affinity chromatography.
Specificity of the antisera was tested by immunoblot analysis. The antibody directed against ARO was validated by direct comparison with the specificity of a commercially available antibody (Biozol Diagnostika, Eching, Germany, originally described by [26]) on microsomal extracts of marmoset placenta.
Immunoblotting
Bacterial lysates were obtained from E. coli BL21 CodonPlus (DE3)-RP (Stratagene) expressing fusion proteins of glutathione S-transferase with human 17HSD7, human 17HSD1, mouse 17HSD7, or glutathione S-transferase alone, respectively. The cells were lysed by treatment with 7.5 U/µl rLysozyme (Novagen, Schwalbach/Taunus, Germany) in PBS/5% glycerol followed by three cycles of freezing and thawing and treatment with Benzonase (Merck Eurolabs, Darmstadt, Germany) according to the manufacturer's instructions.
Microsomal protein extracts were prepared by disruption of marmoset tissues in PBS using an Ultraturrax homogenizer. Homogenates were centrifuged at 600 x g to remove remaining tissue debris. The 600 x g supernatant was subjected to centrifugation at 40 000 x g for 2 h. The 40 000 x g pellet was resuspended in PBS and aliquots were used for protein determination [27] and immunoblots.
For immunoblots, denaturing/loading buffer (100 mM Tris, pH 6.8, 100 mM DTT, 2% SDS, 10% glycerol, 0.01% bromophenol blue) was added to the aliquots. Samples were subjected to denaturing PAGE on 10% gels and subsequent semidry blotting onto nitrocellulose membranes (Amersham Biosciences, Freiburg, Germany) according to standard procedures (Biorad Mini-Protean 3 Cell, Instruction Manual). Membranes were blocked with PBS/0.05% Tween20 and incubated overnight at 4°C with primary antibody (17HSD7 1:2000, ARO 1:500). In negative controls the primary antibody was blocked by 1 h preincubation at room temperature with an excess concentration of the peptide that had been used for immunization.
Bound antibodies were visualized using a goat anti-rabbit IgG coupled to alkaline phosphatase in a dilution of 1:10 000 and nitroblue tetrazolium/ bromo-chloro-indolyl phosphate (Sigma Aldrich, Taufkirchen, Germany) as substrate.
Immunohistochemistry
Tissues were fixed with 4% paraformaldehyde and embedded in paraffin. Five-micrometer sections were deparaffinized, hydrated, and stained according to the DAKO Envision protocol (DAKO Diagnostika, Hamburg, Germany) using a dilution of 1:500 or 1:1000 of the primary antibody and using diaminobenzidine as the substrate for visualization of the immune reaction. In negative controls the primary antibody was blocked by 1 h preincubation at room temperature with an excess concentration of the peptide that had been used for immunization.
Staining was visualized with an Axiophot microscope (Zeiss, Oberkochen, Germany) using Openlab 3.0 digital image analysis (Improvision, Coventry, U.K.).
HPLC Analysis of Enzyme Activity
Frozen CLs were thawed on ice and homogenized by ultraturrax in 0.5 ml phosphate buffered saline. Conversion of estradiol to estrone was measured by HPLC analysis as described by Adamski et al. [28]. Determination of protein content was performed according to Markwell et al. [27].
Statistical Analyses
For comparison of reductive and oxidative enzyme activities in CL, a paired Student t-test was performed.
| RESULTS |
|---|
|
|
|---|
The cDNAs from the complete coding region and the proximal promoter region of marmoset 17HSD1 (cj17HSD1) and a partial sequence of marmoset ARO (cjARO) were obtained by RT-PCR from marmoset placental RNA. They were cloned and the sequence homologies to other species were determined (Table 1). Marmoset 17HSD7 (cj17HSD7), which had been cloned earlier [18], was also included in the table.
|
Marmoset 17HSD1 is remarkably less conserved than cj17HSD7. The coding region of cj17HSD1 is 46 nucleotides shorter than the human cDNA owing to a deletion in exon 6. Consequently, the deduced amino acid sequence is shortened by 14 residues.
In the coding region of cj17HSD7 there is a deletion of only two nucleotides, which causes the generation of an earlier stop codon than in the human orthologue. The deduced amino acid sequence is 11 residues shorter than in the human.
The proximal promoter region of cj17HSD1 was determined from position -1 to -361 and contained essentially the same regulatory elements as the human orthologue (data not shown). Overall homology of the promoter region amounts to 93%.
Intron-exon organization of cj17HSD1 and cj17HSD7 genes were obtained by screening a genomic cosmid library. Introns of 17HSD7 were not sequenced, but the sizes were determined using intron-spanning PCR primers. There is again a close resemblance to the human orthologues, with the exception of intron 8 from cj17HSD7. It has a length of 8.0 kb in the human gene and it spans only 3.0 kb in the marmoset gene. Unfortunately, no clone was identified in the cosmid library, which contained the complete cj17HSD7 gene. Therefore, the promoter region and exons 1 and 2 of this gene could not be analyzed in this study.
Tissue Distribution of 17HSD7 Compared with 17HSD1 and Aromatase
The overall tissue distribution of 17HSD7 mRNA was compared with that of 17HSD1 and ARO and quantified by densitometric analysis relative to ß-actin expression in northern blots using marmoset-specific RNA-probes (Fig. 1). Apart from placenta, all tissues were obtained from adult, nonpregnant animals.
|
As shown earlier [18], cj17HSD1 is represented by two transcripts of 2.3 and 1.3 kb in size. Only the smaller transcript that corresponds to the expression of an active enzyme [29] was quantified and was found exclusively in placenta. The mRNA sizes of cj17HSD7 transcripts are approximately 2.8 and 1.4 kb in the marmoset [18]. The density of these two transcripts showing a broader tissue distribution than cj17HSD1 mRNA was calculated by the addition of both values. Marmoset 17HSD7 mRNA was expressed in steroidogenic tissues like placenta or testes, but the strongest signal was observed in liver. Marmoset ARO mRNA was coexpressed with 17HSD7 in placenta, uterus (proliferative phase), testis, liver, and kidney.
To confirm these results at the level of protein expression, a polyclonal antiserum directed against 17HSD7 was generated and its specificity was proven by immunoblot analysis (Fig. 2). Different fusion proteins with glutathione S-transferase were overexpressed in E. coli. The antiserum reacted with a fusion protein of human or murine 17HSD7, but not with a fusion protein of human 17HSD1 and glutathione S-transferase or glutathione S-transferase alone.
|
In microsomal extracts from different marmoset tissues, protein bands of approximately 30, 60, and 80 kDa were detected (Fig. 3). The immunoreaction with the 30- and 60-kDa protein could be inhibited by preabsorption of the antiserum with the purified peptide antigen. Interestingly, there was no specific protein signal in placenta and uterus, although a moderate signal for the respective mRNA was shown to be present in the Northern blot analysis (Fig. 1).
|
Expression of Estradiol Synthesizing Enzymes in Reproductive Tissues During Ovarian Cycle and Pregnancy
The time-related expression pattern of estradiol synthesizing enzymes was studied in three tissues that are involved in the maintenance of pregnancy, i.e., in CL, uterus, and placenta. Marmoset pregnancy lasts approximately 144 days [30]. Total RNA from late luteal phase (Day 13/14) and Days 28/29, 38, 40, 98, and 135 of pregnancy (near term) was subjected to Northern blot analysis (Fig. 4). In all tissues examined, cj17HSD1 mRNA was detected almost exclusively in the placenta, even as early as Day 38, and only faintly in one of the two CL from Day 38. In contrast to that, the strongest signal for cj17HSD7 mRNA was observed in the CL. It increased relative to ß-actin mRNA in the CL during the first part of pregnancy before the luteo-placental shift in marmoset monkeys [31]. After that, at Day 98, cj17HSD7 expression was reduced. In contrast to the expression pattern reported for the rat [14], cjHSD7 was additionally expressed at a lower level in the pregnant uterus and even lower in placenta. Marmoset ARO was expressed in CL and placenta at all times. Thus in both tissues, CL and placenta, the two steps of estradiol synthesis, aromatization and reduction, can be performed. ARO is expressed in all estradiol synthesizing tissues, CL and placenta, whereas cj17HSD1 and cj17HSD7 are never expressed at the same level in a given tissue. Compared with its expression level in uterus and placenta, cj17HSD7 mRNA is most prominent in the CL of pregnancy.
|
The coexpression of cj17HSD7 and cjARO was confirmed by immunoblots of CL samples from different time points of pregnancy (Fig. 5). The expression of both proteins increases up to Day 98 of pregnancy and fades after delivery.
|
Immunolocalization of cj17HSD7
Cell type-specific expression of cj17HSD7 was examined by immunohistochemistry on paraffin sections (Fig. 6). In the ovary the protein was assigned exclusively to luteal cells. No cj17HSD7 protein was detected in follicles (data not shown). A moderate immunoreaction was further observed in the placental villi, in Leydig cells, and in specific cells of the tubular compartment of the testis, presumably Sertoli cells, as well as in epithelial cells of the epididymis. Preabsorption with the immunizing peptide abolished the immunoreaction completely.
|
Enzymatic Interconversion of Estradiol and Estrone in Marmoset CL
As only 17HSD7 mRNA and protein, but not 17HSD1 mRNA, were shown to be present in marmoset CL (Fig. 4), any estradiol synthetic enzyme activity in this tissue should be due to 17HSD7. A limited amount of luteal tissue was available for the assessment of this enzymatic activity. Triplicate samples from two CLs of the midluteal phase and from Day 45 of pregnancy were examined. Interconversion of estradiol and estrone was readily detectable. Reductive activity (estradiol synthesis) was 1.5- to 3-fold higher than oxidative activity (30.9 ± 8.1 µU/mg protein reduction vs. 10.0 ± 1.5 µU/mg protein oxidation in the midluteal phase CL, and 18.9 ± 5.8 µU/mg protein reduction vs. 13.2 ± 3.7 µU/mg protein oxidation in the CL from Day 45 of pregnancy, P < 0.05).
| DISCUSSION |
|---|
|
|
|---|
The coding region of cj17HSD1 was cloned completely, and the gene structures of cj17HSD1 and cj17HSD7 were analyzed. Compared with the human enzymes, again no major differences were detected in the marmoset, indicating the same structure-function relationships in the two species. However, 17HSD7 proved to be more conserved than 17HSD1.
A new specific polyclonal antibody was validated and enabled us to perform the first detailed study of 17HSD7 protein expression in primate tissues. Earlier Northern blot analysis of mouse and marmoset tissues as well as RT-PCR analysis of human tissues [15, 17, 18] had suggested a broader tissue distribution than originally assumed by immunological studies in rat [32]. Indeed, we detected cj17HSD7 protein in liver, kidney, reproductive tissues, adrenal, and intestine.
In this study the expression of ARO, which provides estrone as the substrate of 17HSD1 or 17HSD7, was examined in addition to the two 17HSDs. It was assumed that tissues expressing ARO together with 17HSD1 or 17HSD7 are indeed involved in estradiol synthesis. A look at the complete estrogenic pathway was considered as an important aspect in the elucidation of 17HSD7 function, because some phylogenetic evidence was presented that gave rise to speculations whether this enzyme might have an additional function in cholesterol synthesis [21]. Our study provides further arguments for this notion, showing the strongest expression of cj17HSD7 in liver accompanied only by a comparably weak expression of ARO. Moreover, exactly those gonadal cell types immunopositive for 17HSD7luteal cells and Leydig cellshave been identified as sites of de novo cholesterogenesis [33]. Nevertheless, cj17HSD7 is always coexpressed with ARO in the gonads, and its protein expression is tightly regulated like that of ARO in the CL of pregnant animals.
According to our results, estradiol synthesis in the CL of pregnant animals appears as one of the key functions of 17HSD7 not only in rodents, but also in primates in spite of different control mechanisms of CL function and steroidogenic function of the placenta in these species. Reductive estrogenic enzyme activity in CLs of marmoset monkeys must be assigned to cj17HSD7, because it is the only estradiol synthesizing enzyme detectable in their CL. Similarly, 17HSD1 was not found in the CL of cycling or pregnant rats [7, 34]. It is true that also in the human a weak expression of 17HSD1 was shown in luteinized granulosa cells originating from patients treated for in vitro fertilization [6, 35, 36], but these cells were retrieved shortly before ovulation and did not necessarily represent a physiological state of luteal cells. In immunohistochemistry on sections of human ovaries from the midluteal phase, only a weak immunoreaction specific for 17HSD1 could be detected [6]. The primate model examined in this work offered the possibility to examine CL from different and defined time points of the ovarian cycle and pregnancy. In the marmoset monkey only 17HSD1 of placental origin seems to complement luteal 17HSD7 in performing the last step in estradiol synthesis after the luteo-placental shift in steroidogenesis.
It was suggested that 17HSD7 may have an important function during implantation because it could be localized specifically at the implantation site in mouse uterus by in situ hybridization [37]. Marmoset uteri were processed in toto (endometrium and myometrium) in this study, thus diluting endometrium-specific gene products that are expressed at low levels or only locally at distinct areas like implantation sites. This may be the reason why only a moderate expression of cj17HSD7 mRNA and no expression of cj171HSD1 or cjARO was observed in all uteri examined, irrespective of the time point chosen. Further studies are needed to prove a role of 17HSD7 in primate implantation.
In view of its broad tissue distribution, 17HSD7 is also a candidate enzyme for extragonadal estrogen synthesis. In humans and most higher primates the zona reticularis of the adrenal cortex secretes androgen precursors into the circulation, which can be converted to estradiol in peripheral, extragonadal tissues [3842]. In New World monkeys such as C. jacchus, high circulating levels of dehydroepiandrosterone as an androgen precursor of extragonadal estradiol synthesis are maintained by the gonads [43]. This increased presence of circulating steroid precursors enables primates to take advantage of local tissue-specific estradiol synthesis.
In conclusion, the regulated coexpression pattern of cj17HSD7 and cjARO together with the estrogenic activity in CL tissue demonstrated in this study speak in favor of a genuine function of 17HSD7 in estradiol synthesis. On the other hand enzyme protein expression was also evident in nonreproductive tissues. Thus the proposed function in cholesterogenesis remains conceivable.
| ACKNOWLEDGMENTS |
|---|
| FOOTNOTES |
|---|
2 Correspondence: Bettina Husen, Solvay Pharmaceuticals GmbH, PH-REB, Hans-Böckler-Allee 20, 30173 Hannover, Germany. FAX: 49 511 857 3272; bettina.husen{at}solvay.com ![]()
Received: 17 October 2002.
First decision: 13 November 2002.
Accepted: 8 January 2003.
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
A. Einspanier, K. Lieder, A. Bruns, B. Husen, H. Thole, and C. Simon Induction of endometriosis in the marmoset monkey (Callithrix jacchus) Mol. Hum. Reprod., May 1, 2006; 12(5): 291 - 299. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |