| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Pregnancy |
Reproductive Technologies Group,3 AgResearch, Ruakura Research Centre, Hamilton 2001, New Zealand
Departments of Developmental and Molecular Biology and Obstetrics and Gynecology and Women's Health, 4 Albert Einstein College of Medicine, New York, New York 10461
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONNOTE ADDED IN PROOFREFERENCES |
|---|
cytokines, female reproductive tract, placenta, pregnancy, uterus
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONNOTE ADDED IN PROOFREFERENCES |
|---|
The expression of CSF-1 by mouse uterine epithelium is regulated synergistically by estradiol-17ß and progesterone [10]. Although the pattern of uterine epithelial expression of CSF-1 and trophoblastic expression of the CSF-1R in mice and humans is similar, the pattern of expression is more complex in humans because CSF-1 is also expressed in the cytotrophoblast, the extravillous trophoblast, and trophoblastic cells of the villous columns [11], whereas in mice CSF-1 is not expressed by placental cells. Trophoblastic expression of CSF-1 is also found in the pig [12]. The coexpression of the ligand and its receptor in trophoblastic cells suggests both a paracrine and autocrine mode of action for CSF-1 in these cells.
CSF-1 is expressed as a variety of isoforms that are derived from alternative mRNA splicing and posttranslational differential proteolytic processing [13, 14]. In the mouse, the predominant mRNA species in most if not all tissues is the 4.0-kilobase (kb) form that encodes a secreted proteoglycanated protein [15]. During pregnancy, the large increase in CSF-1 protein in the mouse uterus predominantly is due to a switch to and increase in expression of a 2.3-kb mRNA that uses the untranslated exon 9 instead of exon 10 while maintaining the same coding region as the 4.0-kb form [8, 10]. In addition, a cell-surface biologically active form with alternative splicing within exon 6 (3-kb mRNA) has also been found [16]. This form is present only at very low levels in the pregnant mouse uterus [17] but represents about 40% of the mRNA for CSF-1 in the uterine epithelium during the menstrual cycle in humans. This 3-kb form becomes less prominent but is still detectable in the pregnant human uterus.
The pattern of CSF-1 expression during pregnancy is less similar between mice and humans, which have a hemochorial invasive placenta, than between humans and pigs, which have an epitheliochorial noninvasive placentation. We investigated how the patterns of CSF-1 expression are conserved in species such as cattle, with a synepitheliochorial cotyledonary placenta [18], in which the extent of invasiveness of the maternal endometrium is intermediate between that in mice and that in humans. CSF-1R has been detected in trophoblastic cells of bovine placentomes, particularly in the binucleate cells, which play a key role in the formation of the syncytium [19]. We report here the molecular cloning of two alternatively spliced forms of bovine CSF-1 cDNA and its expression pattern in the nonpregnant cyclic uterus and the pregnant uteroplacental unit. The pattern of expression of mRNAs encoding the secreted and cell-surface forms of CSF-1 in both maternal and extraembryonic tissues more closely resembles the human pattern than the mouse pattern of expression and shows a dynamic regulation of alternatively spliced forms through the estrous cycles and during pregnancy.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONNOTE ADDED IN PROOFREFERENCES |
|---|
All animal treatments and handling were carried out in accordance with the guidelines of the Animal Welfare Act 1999 of New Zealand. Mature mixed-breed dairy cows were synchronized for estrus using an intravaginal controlled internal drug release (CIDR) device for progesterone (InterAg, Hamilton, New Zealand) inserted for 12 days. On Day 8 after insertion, the animals were injected with 1 ml of estrumate (Schering-Plough, Union, NJ). CIDRs were removed on Day 12, and estrus was usually detected within 48 h. Estrous animals were inseminated with thawed cryopreserved semen from a bull of proven fertility. At different times of gestation, animals were killed, and the reproductive tracts were recovered. Nonpregnant cycling cows, where Day 0 was the day of estrus, were killed on Day 1 (n = 3), Day 2 (n = 1), Day 3 (n = 3), Day 4 (n = 3), Day 14 (n = 2), Day 17 (n = 2), Day 19 (n = 2), Day 20 (n = 3), and Day 21 (n = 4). Animals determined to be pregnant, based on the recovery of a conceptus in the tract, were killed on Day 13 (n = 2), Day 14 (n = 4), Day 17 (n = 4), Day 19 (n = 2), Day 21 (n = 2), Day 24 (n = 2), Day 27 (n = 2), Day 29 (n = 1), Day 31 (n = 1), Day 50 (n = 2), and Day 61 (n = 3). Endometrial samples were collected from the midsection of each horn after opening up the section and cutting off the endometrium with sharp scissors. Cross sections of the uterine horn for in situ hybridization were obtained from the site adjacent to the 3- to 4-cm2 pieces of endometrium taken for RNA extraction. Several RNA samples from conceptuses were obtained from controls of another experiment. Tissue samples were either snap frozen for RNA isolation or fixed in 4% formalin in PBS for histology.
RNA was extracted from tissues ground to a fine granular form in liquid nitrogen. The Trizol reagent (Life Technologies, Gaithersburg, MD) was used according to the protocol specified by the manufacturer.
Cloning of the Bovine CSF-1 Gene by Reverse Transcription Polymerase Chain Reaction
Primers for the initial cloning of the coding region by reverse transcription polymerase chain reaction (RT-PCR) were designed to hybridize to regions showing the greatest homology in the alignment of human, mouse, and rat DNA sequences. Primers used are shown in Table 1, and the positions to which they bind on the cDNA are shown in Figure 1. Two micrograms of total RNA from bovine endometrium at Day 31 of gestation was reverse transcribed with 0.5 pmol of reverse primer (ERS57), using Moloney murine leukemia virus reverse transcriptase (Life Technologies) in a total volume of 20 µl according to the protocol for the use of the Superscript II enzyme (Life Technologies). As a control to check that the primers were working, the RT-PCR was also carried out with RNA from Day 16 pregnant mouse uteri. After annealing of the primer with the RNA, reverse transcription was initiated at room temperature for 2 min to allow for some extension from the primer, and then the reaction was carried out at 42°C. One microliter of the reverse transcription reaction was used for each 12.5 µl of PCR; the optimal MgCl2 concentration was determined by titration. The PCR contained 0.5 µM of each primer (JP224 and ERS57), 0.2 mM deoxynucleotide triphosphate (Roche Diagnostics, Basel, Switzerland), MgCl2 concentrations of 0.5–2.0 mM, and 0.5 U Taq polymerase (Roche). The same concentrations of the above components were used for all PCRs unless specified otherwise. The PCRs were carried out as follows: 94°C for 1 min, followed by 30 cycles of 94°C for 30 sec, 56°C for 30 sec, and 72°C for 30 sec, and a final extension at 72°C for 7 min. The PCR product (amplicon A) was then sequenced.
|
|
After sequence verification of the 
450-bp partial cDNA, the rest of the gene was cloned (Fig. 1A). The upper primer (CSF6a) was designed to overlap with the 450-bp fragment and was placed upstream of the lower primer (ERS57). Reverse transcription was carried out with primer CSF8R, and the PCR was carried out with primer pairs CSF6a and CSF6R or CSF6a and CSF8R. Instead of Taq polymerase, the Elongase enzyme mix (Life Technologies) with proofreading properties was used according to the manufacturer's protocol. The PCR was carried out in a total volume of 50 µl, and the conditions used are as follows for both sets of primers: 5 cycles of 94°C for 45 sec, 61°C for 30 sec, and 70°C for 1 min, followed by 25 cycles of 94°C for 30 sec, 63°C for 30 sec, and 70°C for 1 min, and a final extension at 70°C for 7 min. PCR products (amplicons B and C) were then sequenced.
Cloning the "Short" Form of Bovine CSF-1
A reverse primer was designed from the bovine sequence corresponding to the most 3' end of exon 6 in the human gene. The template used was total RNA from Day 61 pregnant uterus reverse transcribed with the CSF8R primer. PCR was carried out with CSF2F and CSF6altR and 1 mM MgCl2. The PCR conditions were as follows: 94°C for 2 min, followed by 30 cycles of 94°C for 30 sec, 64°C for 30 sec, and 72°C for 45 sec, and a final extension at 72°C for 7 min. The smaller of the two PCR products (amplicon Fb) was then sequenced.
Cloning of the 3' Untranslated Region of the Bovine CSF-1 Gene
Because homology of the 3' untranslated region (UTR) sequences in the human and mouse genes is lower than that of the coding region, rapid amplification of cDNA ends (3' RACE) was used to isolate this part of the cDNA. Poly(A+) mRNA (0.5 µg) from Day 61 pregnant uterine tissue was reverse transcribed with 0.5 pmol of 3'-RACE-dT primer using Superscript II in a total volume of 20 µl. Reverse transcription was initiated at 42°C for 2 min, then the temperature was increased to 50°C for the next 1 h. Two microliters of the reverse transcription reaction was used as the template with Expand Long template polymerase enzymes (Roche) and buffer 3, according to the instructions of the manufacturer. The primers used were 3'-RACE-2 and CSF8F with the following PCR protocol: 94°C for 2 min, followed by 10 cycles of 94°C for 10 sec, 63°C for 30 sec, and 68°C for 2 min, 20 cycles of 94°C for 10 sec, 63°C for 30 sec, and 68°C for 2 min increasing the extension time by 5 sec/cycle, and a final extension at 68°C for 7 min.
PCR products were separated in 0.8% agarose gels, and the 1-kb and 2-kb (amplicon D) bands were excised and reamplified using the same primers and Expand polymerase. Because we could not stably propagate the entire 2-kb amplicon D in the pGEM-T Easy vector (Promega, Madison, WI) and DH5
Escherichia coli cells, the insert was subcloned as shorter overlapping restriction fragments in pBluescript II (Stratagene, La Jolla, CA) and then sequenced separately. The most 3' end of this amplicon (200–300 bp) was not sequenced because the DNA fragment could not be stably propagated in DH5 E. coli and because of technical difficulties in sequencing DNA with long stretches of repeated nucleotides. About 70 bp of this last 200–300 bp is the poly(A+) tail.
Cloning and Sequencing of Amplicons
All amplicons were gel purified in agarose gels and cloned into the pGEM-T Easy vector. Plasmids for sequencing were prepared using the High Pure plasmid purification kit (Roche). Sequencing was carried out using the dye-termination chemistry of either the ABI system (Applied Biosystems, Foster City, CA) or the LiCor system (Lincoln, NE). At least two independent clones were sequenced from both directions for each PCR product. Contigs were assembled using the Seqman program of Lazergene (DNAStar, Madison, WI).
Northern Analyses
Twenty micrograms of total RNA from each sample was electrophoresed in 1.2% SeaKem GTG agarose gels (FMC Bioproducts, Rockland, ME) in 20 mM 3(N-morpholino)propanesulfonic acid, 8 mM sodium acetate, 1 mM EDTA (pH 7) containing 1.8% (v/v) formaldelyde [20]. After transfer overnight to Nytran nylon membranes (Schleischer and Schuell, Dassel, Germany) and ultraviolet cross-linking, the membranes were stained with methylene blue [21] to visualize the RNA. The probe used for detection of CSF-1 transcripts was the [32P]-dCTP-labeled 405-bp bovine cDNA fragment (Fig. 1A, amplicon E) equivalent to exons 2–5 of the human or rodent genes. Blots were hybridized with the probe in NaHPO4/SDS buffer [22] at 65°C overnight and then washed twice with 1x sodium chloride/sodium citrate (SSC), pH 7.0, containing 0.1% SDS at 65°C. Hybridized blots were exposed to X-Omat film (Eastman Kodak, Rochester, NY) in the presence of intensifying screens at -70°C. Samples were analyzed by Northern blots two to four times, but only representative blots are shown in the figures. Band intensities were determined by scanning densitometry with the GS-800 densitometer (BioRad Laboratories, Hercules, CA).
RT-PCR of Alternatively Spliced CSF-1 Transcripts
RT-PCR was used to study the expression of alternatively spliced mRNA in different tissues. Primers CSF2F and CSF5R were used to detect the expression of all CSF-1 transcripts (Fig. 1A, amplicon E), whereas CSF2F and CSF6altR amplified both the long (amplicon Fa) and short (amplicon Fb) forms of CSF-1 mRNA. Primers CSF2F and CSF6R were used to amplify only the long form of CSF-1 (amplicon G) because previously when the same primer pair was used to detect both the short and long forms there was preferential competition by the short form for amplification [16], making it difficult to detect the long form. Two micrograms of total RNA was reverse transcribed with gene-specific primers CSF8R and b-actinR in a total reaction volume of 20 µl, as described above. For each sample, the negative control was the reaction carried out simultaneously in the absence of reverse transcriptase. One microliter of the reverse transcriptase reaction mix was used for each PCR volume of 12.5 µl. The same reverse transcriptase reaction mix was used for all primer pairs, including the actin control. The MgCl2 concentration for all the primer sets was 1 mM. After PCR, 5 µl of each reaction was run in 7% acrylamide gels (37.5:1 acrylamide:bis-acrylamide) containing 0.5x Tris-borate-EDTA (TBE) buffer, and the gels were stained with ethidium bromide. To verify that the amplicons were from the CSF-1 mRNA, 5 µl of each reaction was electrophoresed in a 1.8% agarose gel in TBE buffer, and the gels were soaked in 0.5 M NaOH and 1.5 M NaCl for 30 min. The DNA was then transferred to charged nylon membranes (Nytran+) and hybridized with labeled amplicon E (Fig. 1A). After hybridization, Southern blots were washed in 0.1x SSC with 0.1% SDS at 65°C and then exposed to X-Omat film.
In Situ Hybridization
Amplicon E was also used to generate RNA probes for in situ hybridization. The plasmid was linearized with HindIII for the generation of antisense RNA and with XhoI for the sense probe. Digoxigenin-labeled RNA probe was synthesized as described in the protocol from the manufacturer (Roche). Fixed paraffin-embedded tissue sections of placental and embryonic tissues were dewaxed, rehydrated, and treated with proteinase K and then acetylated essentially as previously described [23]. For endometrial tissue, sections were dewaxed, rehydrated, and then microwaved in 10 mM citrate buffer (pH 6.0) three times for 5 min each time instead of treatment with proteinase K. This treatment produced a better signal:noise ratio for endometrial tissue sections. Sections were hybridized with the probe at 60°C overnight. Detection of hybridized probes was carried out essentially as described in the manufacturer's protocol. Sections were photographed wet without counterstaining or after counterstaining with nuclear Fast Red and mounted.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONNOTE ADDED IN PROOFREFERENCES |
|---|
We cloned the cDNA of the bovine CSF-1 gene as a series of overlapping RT-PCR fragments using the strategy outlined in Figure 1, A and B. The cDNA contains the entire coding region of the gene with the exception of the first eight amino acid residues in the N-terminus (Fig. 2). The longest assembled contig of 3.6 kb (GenBank accession AY181987) is equivalent to the 4-kb mRNA in the mouse and the human, which is predicted to code for the secreted form of CSF-1. This contig had approximately 2 kb of the 3' UTR equivalent to exon 10 of the mouse gene. We also cloned a cDNA corresponding to the short form of the mouse and human sequence, which has 891 bp of the coding region spliced out from the long form. This splice variant is predicted to yield an 3-kb mRNA that in humans codes for a membrane-bound form of the protein.
|
Amino acid sequence alignment using the Clustal method revealed that the primary sequence of the protein is well conserved between the cow and other species, particularly in the first 180 residues of the N-terminus (Fig. 2). All 12 cysteine residues that are required for intra- and interdisulphide bonds, including those that are indispensable for biological activity [13], are completely conserved. The greatest degree of amino acid sequence variation is in the region that is spliced out in the membrane-bound form (between the arrows in Fig. 2), suggesting that this region is probably not critical for biological activity. Several of the amino acid substitutions highlighted in Figure 2, which have resulted in a change of charge, size, or hydrophobicity of the side chain, could have arisen as a result of single base substitutions in the codon. The nucleotide sequence in the 3' UTR was less well conserved, but at least one of the three AU-rich motifs (AUUUA) found in the human and mouse genes that are thought to determine mRNA stability [24] was also present in the bovine cDNA close to the expected position. We were unable to confirm the existence of the other two motifs because the bovine cDNA was not sequenced all the way to the 3' end, where the other two AUUUA motifs in the mouse gene are located.
CSF-1 mRNA Expression During the Estrous Cycle and Pregnancy
CSF-1 mRNA was expressed in both nonpregnant (Fig. 3) and pregnant bovine endometrial tissues (Fig. 5). Two predominant transcripts, 4 kb and 3 kb, were detected throughout the estrous cycle (which is approximately 21 days in the cow) and during early pregnancy. These two forms are equivalent to the cloned cDNAs (Fig. 1) that code for a secreted and a membrane-bound form, respectively. The ratio of the 4-kb and 3-kb forms changed during the estrous cycle (Fig. 3), with the 3-kb form more abundant on Day 14 (lanes 5–7) and Day 17 (lanes 8 and 9). By Days 19 and 20, there was a decline in the mRNA levels of both transcripts and a switch to expression of more of the 4-kb mRNA (lanes 10 and 11). Expression of both forms was upregulated on Day 21 (lane 12) during the proestrous estrogenic surge leading to ovulation.
|
|
Using scanning densitometry of the autoradiograms to estimate the relative abundance of the two forms, in nonpregnant endometrium the 4-kb:3-kb ratio was between 1.0 and 1.3 in the first 4 days of the cycle but decreased to approximately 0.5 on Days 14 and 17 (Fig. 4). An ANOVA was used to compare mean log ratios between days of the cycle. On Day 4, the mean log ratio was significantly higher than that on either Day 14 or Day 17 in cyclic animals (P < 0.001). In the 3 days leading up to estrus (Days 19–21), the ratio increased significantly back to approximately 1.0 (P < 0.01 between Days 17 and 19 or Days 17 and 20; P < 0.001 between Days 17 and 21).
|
In pregnant animals, there was little or no change in the levels of CSF-1 mRNA between Days 13 and 17 (Fig. 5A, lanes 1–6) when compared with nonpregnant uteri. However, we detected a change in the 4-kb:3-kb ratio, which was verified by scanning densitometry (Fig. 4). In the example shown in Figure 5A, it was also possible to detect a difference in the 4-kb:3kb ratio in the Day 17 gravid (lane 6) compared with the nongravid (lane 7) horn. The 3-kb mRNA was more abundant in the nongravid horn, and the expression resembled that seen in nonpregnant endometrium (Fig. 3, lanes 8 and 9) at the equivalent day of the estrous cycle. In Day 17 pregnant uteri from two animals, the ratio was 1.04 and 1.07 in the gravid horn and 0.74 and 0.89 in the nongravid horn. The 4-kb mRNA increased progressively, relative to the 3 kb, once pregnancy was established and as it continued (Fig. 5A, lanes 10–13). We also detected an increase in expression of the 1.4- and 0.8-kb transcripts after the formation of the placentomes on Day 50 (lanes 12 and 13). The change in the 4-kb:3-kb ratio between pregnant and nonpregnant uteri could be detected between Days 14 and 17 (Fig. 4). Using an ANOVA, the mean log ratio on Day 17 in uteri of pregnant animals was significantly higher (P < 0.001) than that on Day 17 in nonpregnant animals; the difference between pregnant and nonpregnant animals was not significant on Days 14, 19, or 21.
After the formation of the placentomes, CSF-1 expression was detected in both the fetal cotyledonary and the maternal caruncular tissue (Fig. 5B) and in the intercaruncular endometrium on Day 61. There was no difference in CSF-1 expression level between caruncular tissues, which form the placentomes, and the intercaruncular endometrium, which contains the uterine secretory glands and only interacts with the chorioallantoic membrane in a superficial manner. Thus, even with the possibility that not all fetal cells were removed from the maternal caruncular tissues, the expression of CSF-1 in these two regions of the uterus is very similar and the level was higher than that in the fetal cotyledons. The predominant transcript in caruncular tissue was the 4-kb mRNA, with little of the 3-kb form detectable (Fig. 5B). The 1.4- and 0.8-kb transcripts, which were relatively abundant until around Day 70, were negligible by Day 100, when only the 4-kb form was detectable (results not shown).
CSF-1 expression was also detected in all of the extraembryonic membranes of the preimplantation conceptus (Fig. 6), such as the yolk sac, allantois, and trophoblast, although at lower levels than those observed in the endometrium. Approximately the same amount or slightly more of the 4-kb mRNA was expressed in these tissues (ratios between 1.0 and 1.6).
|
RT-PCR Analysis of Alternatively Spliced CSF-1 Transcripts
RT-PCR analysis was used to detect expression of CSF-1 in early embryos, where there was insufficient material to allow Northern analysis, and to investigate the expression of alternative splice variants in the different tissues (Fig. 7). Primer pairs were designed to detect both the long (secreted) and short (membrane-bound) forms (second panel down) or only the long form (third panel down). Using this approach, low levels of CSF-1 expression could be detected in elongated bovine conceptuses on Day 18 (Fig. 7A, lane 1). Although the RT-PCR results shown in Figure 7 are only semiquantitative, it is still possible to discern that on Day 20, there is more CSF-1 mRNA in the entire conceptus (lane 2) than in the trophoblast alone (lane 3). Similarly, CSF-1 was more abundant in the allantois (lane 5) than in the trophoblast (lane 6) on Day 27. Up until Day 24 of gestation (Fig. 7A, lanes 1–4), only the short form of CSF-1 was detected, whereas both the long and short forms were expressed on Day 27 in the allantois and yolk sac (lanes 5 and 8).
|
In nonpregnant cyclic uteri (Fig. 7B, lanes 1–5), the predominant form of CSF-1 expressed was also the short form. Some expression of the long form was detected after estrus on Day 1 (lane 1) and on Day 17 in the horn ipsilateral to the corpus luteum (lane 5) but not in the contralateral horn (lane 4) of the same uterus. No difference in expression of the two forms of mRNA between the two horns could be detected on Day 14 (lanes 2 and 3). On Day 61 of gestation, the long form was less abundant in fetal cotyledonary tissues (lane 6) than in the maternal caruncular tissues (lane 7).
Southern blot hybridization of the RT-PCR products revealed that in endometrial and uteroplacental tissues, primers CSF2F and CSF6r yielded low-molecular-weight bands that hybridized with the probe (Fig. 7B, right, third panel down) in addition to the predicted long form of the mRNA. Because primer CSF6r hybridizes to a region that is normally spliced out in the short form, these bands may represent splice variants in addition to the two we cloned; however, they were not visible with ethidium bromide staining and thus are not abundant. These additional bands most likely represent alternative spliced exon 6 transcripts, such as those described in the human [25]; PCR with primers CSF2F and CSF5R spanning the equivalent of exons 2–5 of the human and rodent genes only yielded one amplicon (top right). The relative abundance of these minor alternatively spliced transcripts differed between maternal and fetal tissues (Fig 7B, lanes 6 and 7), as indicated by the different banding patterns.
The relative abundance of the long and short amplicons in some cases did not reflect the relative abundance of the 4-kb and 3-kb mRNA in Northern blots. This finding may be the result of lower amplification efficiency for the longer PCR product when primers CSF2 and CSF6altr were used, because PCR conditions were optimized for the shorter product. In the Day 61 maternal caruncular tissues, the greater abundance of the short form despite an apparent lack of the 3-kb mRNA probably reflects the high abundance of the 1.4- and 0.8-kb mRNAs, which probably encode the membrane-bound form of CSF-1.
Histological Localization of CSF-1 Expression by In Situ Hybridization
The probe that was used to study the cellular localization of the CSF-1 mRNA contained only the coding region of the gene, which is found in all transcripts. Thus, all transcripts were detected without discrimination between the secreted or long form and membrane-bound or short form of CSF-1.
In the Day 14 pregnant uterus (Fig. 8A), CSF-1 was detected in the stroma and luminal epithelium of the endometrium; the expression levels in the luminal epithelium were not uniform. Most of the hybridization signal was, however, localized in the stroma surrounding the uterine glands, but the expression varied from one region to another. Little or no mRNA was detected in the glandular epithelium. On Day 18 (Fig. 8, B and C), when the conceptus is rapidly elongating but still unattached, CSF-1 was widely expressed in the stroma of the endometrium, in some cells of the glandular epithelium of the shallow glands, and in the luminal epithelium over both caruncular and glandular endometrium. In the stroma of the caruncular tissue, CSF-1 was expressed in a radial pattern away from an area with little expression adjacent to the epithelium (Fig. 8B). Figure 8D shows the lack of hybridization with the sense RNA probe. CSF-1 expression in the Day 24 trophoblast (Fig. 8E) was detected predominantly in the mesodermal cells, with little if any staining in the trophectodermal cells.
|
In the placentomes (Fig. 8F), CSF-1 mRNA was most abundant in the trophoblast cells of the fetal cotyledonary villi and in the uterine epithelial cells that line the maternal crypts of the caruncles. Some staining was also observed in the stroma in the intercrypt columns. Strong signals were detected in the stroma located beneath the crypts of the placentome but not in the maternal blood vessels (Fig. 8G). However, part of this signal may be due to nonspecific hybridization, which was observed in the equivalent section hybridized with the sense RNA probes (Fig. 8H); such nonspecific staining was absent in other sections. In the fetal cotyledons, CSF-1 was expressed in most but not all trophoblast cells of the villi (Fig. 8I), in scattered cells of the mesenchyme, and in the endothelial cells lining the blood vessels that run through the fetal cotyledons (Fig. 8G). Staining in most trophoblast cells was localized to the cytoplasm, but certain cells (both mono- and binucleate) showed nuclear localization and negative cytoplasmic staining (Fig. 8K). These positively stained trophoblast cells are derived from the same trophectodermal cells that on Day 24 were negative for CSF-1 expression. The equivalent section hybridized with the sense probe is shown in Figure 8L.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONNOTE ADDED IN PROOFREFERENCES |
|---|
In this study, we cloned bovine CSF-1 cDNA using a strategy of overlapping RT-PCR. The bovine gene was expressed as at least four different transcripts, with the 4- and 3-kb mRNA, the predominant splice variants, accounting for two main classes of CSF-1: the secreted form and the membrane-bound form. RT-PCR analysis revealed at least two other splice variants at low abundances, involving alternative splicing in the region equivalent to exon 6 in the human and murine genes. The predicted amino acid sequence shows a high degree of conservation among rodent, human, and bovine species in the region conferring the biological activity and in all cysteine residues required for inter- and intra disulphide linkages, suggesting a conservation of function throughout diverse mammalian species. Also conserved is the stretch of hydrophobic residues in the C-terminus required for the anchoring of the protein to the lipid bilayer of the membrane. The recently cloned rat CSF-1 gene [26] contains a stretch of 25 amino acids that is repeated (underlined residues, Fig. 2). This repeat, which is located just upstream of the conserved hydrophobic C-terminus, is absent in other species.
CSF-1 was one of the first cytokines whose expression was shown to be responsive to steroid hormones in the mouse [10]. Our observations suggest that this is probably the case in cattle. Both the mRNA levels and the ratio of the 4-kb form to the 3-kb form show dynamic variation throughout the estrous cycle and with pregnancy in a manner that is reflected in the relative levels of progesterone and estrogen. Expression was upregulated around estrus in the cyclic uterus, with approximately equal amounts of the 4-kb and 3-kb mRNA being expressed. During diestrus, the 3-kb form was more abundant relative to the 4-kb form. A difference in the ratio of these forms was detected between the horn ipsilateral to the corpus luteum (CL) and the contralateral horn of a nonpregnant animal on Day 17, when the uterine environment is predominantly under the influence of progesterone. Progesterone levels have been reported to be higher in the ipsilateral horn than in the opposite horn further away from the CL [27]; such a progesterone gradient may account for the difference in CSF-1 expression between the two horns. The pattern of expression in cattle endometrium is more similar to that in humans because of the expression of the 3-kb transcript, which is present in very low levels in the mouse [17].
The presence of a conceptus in early pregnancy shifted the expression of CSF-1 in the uterus to more of the 4-kb form. This change in the 4-kb:3-kb ratio occurred within the window for maternal recognition of pregnancy, believed to be between Day 14 and Day 17 in cattle [28, 29]. The detectable difference in the 4-kb:3kb ratio between the gravid and empty horn on Day 17 is suggestive of a paracrine effect of the conceptus on CSF-1 expression. This suggestion is further supported by the observation that the ratio of the two forms differed between pregnant and nonpregnant uterine tissue on Days 14 and 17 even when the progesterone levels were expected to be similar. Factors secreted by the conceptus, such as interferon-
[30], may act locally to modulate the expression of CSF-1. However, the trophectoderm at this stage secretes many other proteins and metabolites, including prostaglandins; thus, other factors may mediate endometrial CSF-1 expression.
A large increase in the expression of the 4-kb mRNA occurs in both the uteroplacental and the intercaruncular tissues during pregnancy. However, expression of the 1.4- and 0.8-kb mRNAs is also upregulated; neither transcript has been detected in pregnant mouse or human uterine tissues, and the 0.8-kb mRNA has not been previously reported. The nature of the splicing in these two transcripts is currently unknown. In all cattle tissues we have examined, including some not reported here, we did not detect a 2.3-kb mRNA, which is the predominant transcript in pregnant mouse uteri. This lack of a 2.3-kb mRNA could explain our inability to isolate the equivalent of bovine exon 9. An amplicon of the expected size for the exon 9 homolog was obtained by 3' RACE, but when this amplicon was sequenced it turned out to be a partial cDNA for uterine milk protein, a highly abundant protein of the serpin superfamily that is secreted by the ruminant endometrium during pregnancy [31]. Thus, the expression of the 2.3-kb mRNA appears to be unique to pregnant mouse uteri.
Expression of CSF-1R in bovine conceptuses from Day 7 onward [19] and CSF-1 in all the extraembryonic membranes of preimplantation conceptuses suggests that the conceptus can respond to CSF-1 secreted from the endometrium and to its own CSF-1. Vascularized membranes such as the allantois and yolk sac, which are sites of embryonic hematopoiesis, expressed CSF-1, consistent with its role as a promotor of growth and differentiation of hematopoietic progenitor cells. The allantois in ruminants also plays a central role in the formation of the placentomes because its adhesion and fusion to the trophoblast is essential for the formation of the chorioallantoic membrane and its vascularization during the formation of the placentomes. CSF-1 expression by the allantois and mesodermal cells of bovine trophoblast may modulate the expression of cell adhesion molecules that are essential for chorioallantoic fusion and subsequent adhesion of the chorion to the endometrium. CSF-1 is known to modulate the expression of fibronectin and certain classes of integrins in human trophoblast [32].
Several roles for CSF-1 in female reproduction and pregnancy have been proposed. During both the estrous cycle and pregnancy in mice, CSF-1 regulates the uterine populations of macrophages [17]. Macrophages are also abundant in the human uterus and decidua, where CSF-1 is likely to play a similar role. However, these cells have not been studied extensively in ruminants. The rate of development and differentiation of mouse embryos [7] is enhanced by CSF-1, consistent with the detected expression of the receptor on mouse embryos [6]. Whether this cytokine has similar effects in other species, such as the cow, remains to be determined. CSF-1 enhanced bovine interferon- gene transcription in an in vitro system using the IFNT promoter linked to a reporter gene [33]. Thus, CSF-1 may promote the expression of interferon- during the critical period leading up to the maternal recognition of pregnancy. In humans, CSF-1 stimulated trophoblast differentiation and the secretion of both chorionic gonadotrophin [34] and placental lactogen [35]. Such a role may also be postulated in bovine preimplantation conceptuses, where CSF-1 from either the uterus or the conceptus could modulate the expression of placental lactogen by the trophectodermal cells, which themselves do not express CSF-1. After the formation of the placentomes, CSF-1 was also expressed in those binucleate cells that secrete placental lactogen. Despite all the observations suggesting a role for CSF-1 in placental development and function, there were no apparent differences in placental growth and differentiation in mice lacking CSF-1, although implantation rates were lower and fetal survival to term was significantly decreased [36]. CSF-1 does play an essential role in the regulation of maternal immunity against bacterial pathogens, acting through trophoblast cells [37]. Given the consistent pattern of expression in a range of mammalian species with very diverse modes of placentation, such a role for maternal immunity may be conserved.
The temporal and spatial patterns of CSF-1 expression in bovine uteroplacental tissues and the expression of CSF-1R in trophoblast cells suggest an important role for CSF-1 during bovine pregnancy. The expression of splice variants, which code for secreted and membrane-bound forms, similar to those observed in other mammalian species also argues for the conservation of the autocrine and paracrine mode of action in cattle.
|
NOTE ADDED IN PROOF |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONNOTE ADDED IN PROOFREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
2 Correspondence: Rita S.F. Lee, Reproductive Technologies Group, AgResearch, Ruakura Research Centre, East St., Private Bag 3123, Hamilton 2001, New Zealand. FAX: 64 7 838 5628; rita.lee{at}agresearch.co.nz
Received: 27 November 2002.
First decision: 18 December 2002.
Accepted: 24 March 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONNOTE ADDED IN PROOFREFERENCES |
|---|
5 integrin receptor in human trophoblasts. Endocrinology 1998 139:2190-2193This article has been cited by other articles:
|
|
 |
|
  P. Das, T. Ezashi, R. Gupta, and R. M. Roberts Combinatorial Roles of Protein Kinase A, Ets2, and 3',5'-Cyclic-Adenosine Monophosphate Response Element-Binding Protein-Binding Protein/p300 in the Transcriptional Control of Interferon-{tau} Expression in a Trophoblast Cell Line Mol. Endocrinol., February 1, 2008; 22(2): 331 - 343. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. G Lea and O. Sandra Immunoendocrine aspects of endometrial function and implantation Reproduction, September 1, 2007; 134(3): 389 - 404. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Y Lee and F. J DeMayo Animal models of implantation Reproduction, December 1, 2004; 128(6): 679 - 695. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Ezashi and R. M. Roberts Regulation of Interferon-{tau} (IFN-{tau}) Gene Promoters by Growth Factors that Target the Ets-2 Composite Enhancer: A Possible Model for Maternal Control of IFN-{tau} Production by the Conceptus during Early Pregnancy Endocrinology, October 1, 2004; 145(10): 4452 - 4460. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||
