Stage-Specific Expression of Cytokine and Receptor Messenger ...

12 downloads 0 Views 894KB Size Report
cyst able to implant into receptive endometrium. Consid- erable data ..... from the inclusion of a novel exon (shown as a hatched box) between exons 3 and. 4. B) Complementary .... that elevated TNFa in the peritoneal fluid of some women.
BIOLOGY OF REPRODUCTION 53, 955-962 (1995)

Stage-Specific Expression of Cytokine and Receptor Messenger Ribonucleic Acids in Human Preimplantation Embryos' Andrew M. Sharkey,2 '3 Kim Dellow,4 Martyn Blayney, 4 Mike Macnamee, 4 Steve Charnock-Jones, 3 and Stephen K. Smith 3 Reproductive MolecularResearch Group,3 Departmentof Obstetricsand Gynaecology University of Cambridge, Rosie Maternity Hospital, Cambridge, United Kingdom Bourn Hall Clinic,4 Bourn, Cambridge, UnitedKingdom ABSTRACT There is considerable evidence to suggest that polypeptide growth factors from either the oviduct or the endometrium can control preimplantation development of the mammalian embryo. These act directly through receptors expressed on the embryo. In addition, embryos also produce growth factors. The reverse transcriptase-polymerase chain reaction (RT-PCR) was used to determine the pattern of expression of mRNAs encoding several growth factor ligand and receptor genes throughout preimplantation development of cryopreserved human embryos. Transcripts encoding the receptor for c-fins, the receptor for colony-stimulating factor-1 (CSF-1), and c-kit (the receptor for stem cell factor [SCFI) were expressed throughout preimplantation development. Other growth factor ligand and receptor transcripts were expressed in astage-specific manner; these included receptors for interleukin (IL)-6 (IL-6R), leukemia inhibitory factor (LIFR), tumor necrosis factor a (TN F) (TNFRp80 and TNFRp60O), and gp130. The transcripts for gp130 and the ligand SCF showed stage-specific splice variants. Blastocysts expressed a novel cDNA encoding gp130, which predicts a truncated form lacking the intracellular signaling domain. No expression of mRNAs encoding LIF, CSF-1, or the cloned receptor for platelet-activating factor was seen inany embryonic stage studied. We have shown that RT-PCR provides a sensitive and powerful method for identifying transcripts encoding growth factors and their receptors insingle human embryos. The method iseconomical, allowing the expression pattern of many genes to be determined from a single embryo. These data are important in defining which cytokines may be involved in regulating human preimplantation development and when they may act. INTRODUCTION Successful implantation requires correct development of the preimplantation embryo resulting in a hatched blastocyst able to implant into receptive endometrium. Considerable data now indicate that soluble growth factors secreted by the uterine epithelium act directly on the embryo to control this process [1, 2]. In turn, developing embryos have been shown to produce a variety of cytokines that may act in an autocrine fashion or may act on the endometrium to influence its receptivity. Growth factors known to be produced by human embryos include interleukin (IL)-1, IL-6, colony-stimulating factor-i (CSF-1), and tumor necrosis factor a (TNFa) [3, 4]. TNFa has been demonstrated in the medium of human embryos cultured up to the morula stage but not in that from the blastocyst [5]. Production of cytokines by the embryo may therefore be regulated in a stagespecific manner. Data on the possible direct effects of cytokines on embryos come primarily from experiments in mice; in these studies many cytokines have been shown to affect the development of preimplantation embryos in vitro. Interferon Accepted May 26, 1995. Received January 9, 1995. IA.M.S. is supported by the Medical Research Council, project grant no. G9403371PA. This work was supported by a grant from Ares Serono. 'Correspondence: Dr. A.M. Sharkey, Department of Obstetrics and Gynaecology, University of Cambridge, Rosie Maternity Hospital, Robinson Way, Cambridge CB2 2SW. FAX: 44 1223 215327.

y and CSF-1 at physiological concentrations inhibit the number of embryos developing to the blastocyst stage [6]. TNFa has recently been shown to have more subtle effects. Although TNFct has no apparent effect on rates of blastocyst formation, it appears to specifically inhibit proliferation of cells contributing to the inner cell mass (ICM), resulting in blastocysts with a reduced ICM [7]. Other growth factors also have specific effects on ICM cells. Insulin-like growth factors 1 and 2 stimulate ICM proliferation, whereas leukemia inhibitory factor (LIF) inhibits their differentiation [8, 91. In spite of considerable effort, however, the specific roles of growth factors in preimplantation development, particularly in the human, are obscure. Even in well-studied systems such as that of the mouse, embryos cultured in vitro lag in development compared to in vivo controls and exhibit lower pregnancy rates after embryo transfer [10]. Clearly a better understanding of the role of growth factors in development could lead to improved in vitro culture conditions and enhance the outcome in human IVF programs. The first step in establishing the possible role of a specific growth factor is to determine its pattern of expression during preimplantation development. In the mouse this has been done for many growth factor receptor mRNAs by reverse transcriptase-polymerase chain reaction (RTPCR). Functional peptides have been demonstrated in mammalian preimplantation embryos through the use of techniques such as radioligand binding and immunocyto974

CYTOKINES AND RECEPTORS IN HUMAN EMBRYOS TABLE 1. Human embryo cDNAs and controls.

chemistry. In humans, for both ethical and practical reasons, much less information is available. In this paper we describe the use of RT-PCR to determine the expression patterns of mRNA transcripts encoding the growth factors LIF, IL-6, TNFa, CSF-1, and stem cell factor (SCF) and their cognate receptors in single preimplantation human embryos. MATERIALS AND METHODS Embryo Culture and RNA Extraction All laboratory reagents were purchased from Sigma (Poole, UK) unless otherwise stated. Cryopreserved human embryos that had been fertilized as part of an IVF program at the Bourn Hall Clinic were used in this study. These embryos had been donated for research purposes by the parents, and this study complied with the requirements of the Human Embryology and Fertilisation Authority and the local ethical committee. Frozen embryos were thawed and cultured in Earle's balanced salts medium supplemented with 0.75% human serum albumin (Alpha Therapeutics, Los Angeles, CA) until attainment of the required developmental stage, and any remaining cumulus cells were removed during routine handling. Embryos at the appropriate stage were then flash-frozen in liquid nitrogen in 5 tl of culture fluid. An identical volume of culture fluid was frozen as a control. Total RNA from first-trimester trophoblast was isolated by a method previously described [11] in which frozen tissue is homogenized in 5 ml of buffer containing 4 M guanidinium thiocyanate (Gibco BRL, Livingston, Scotland), 25 mM sodium citrate (pH 7.0), 0.5% sarcosyl, and 0.1 M 2mercaptoethanol. The lysate was acidified by the addition of 0.5 ml of 2 M sodium acetate (pH 4) and was phenol chloroform-extracted using 5 ml of buffer-saturated phenol and 1 ml chloroform-isoamylalcohol (49:1 [v/v]). The suspension was placed on ice for 15 min and centrifuged at 10 000 X gfor 20 min at 4°C. The aqueous phase containing RNA was precipitated by addition of isopropanol and incubation at - 70 0C. RNA was recovered by further centrifugation, desiccated, and redissolved in homogenization buffer. Finally, RNA was precipitated, washed twice in 70% ethanol, dried, and resuspended in 10 M Tris-HCl (pH 7.4) and 1 mM EDTA. The concentration of RNA was determined spectrophotometrically at 260 nm. RNA was prepared from single human embryos using a scaled-down protocol based on the above procedure. To assist precipitation of the RNA, 100 gg of carrier yeast tRNA (Gibco BRL) was added at the homogenization step. The remaining details were as described above except that all the volumes were 50-fold less and the whole procedure was carried out in 0.5-ml Eppendorf tubes. Details of the RNA samples extracted from embryos are listed in Table 1. A total of sixteen embryos from four patients were analyzed. The

975

Sample

Stage of development

a b c d e f g h j k I m n o

2 cell* 3 cell* 4 cell* 6 cell* 8 cell* morula* blastocyst* culture supernatant for a to g* three pooled blastocysts culture supernatant for j two 6 cell and one 8 cell culture supernatant for I one 4 cell and one 6 cell culture supernatant for n

*Samples a to h were prepared from a set of zygotes from the same patient.

embryo samples labeled a to g in Table 1 came from a single patient. All embryos were of good quality, exhibiting 10% fragmentation or less. RT-PCR

Complementary DNA was synthesized from half the total RNA from each embryo using AMV reverse transcriptase (Super RT; HT Biotech, Cambridge, UK) as described previously [12]. Because of the small amount of material, two pairs of primers were used for each target cDNA in a nested PCR protocol. One thirtieth of the cDNA products were amplified by means of Amplitaq (Cetus, Emeryville, CA) in the manufacturer's recommended buffer. After 30 cycles of PCR using the external primer pair, 1/50th of the first-round reaction was transferred to a fresh tube containing the inner primer pair and subjected to a further 30 rounds of amplification. As negative control, an equal volume of the culture fluid in which the embryo was grown was extracted and subjected to RT-PCR in the same way. Also, 200 cells of the BeWo cell line (European Collection of Animal Cell Cultures No. 86082803) were extracted as positive control. The primers used in this study for each cDNA of interest are shown in Table 2, together with the size of the expected product. The identity of each product was confirmed by cloning and sequencing. PCR products were end-repaired by the addition of MgCl 2 to 10 mM and T4 DNA polymerase (Boehringer Mannheim, Lewes, Sussex, UK). The repaired fragments were electrophoresed through an agarose gel, excised, and purified with Geneclean (BiolOl, La Jolla, CA). The fragments were then blunt end-ligated into Bluescript KS + vector (Stratagene, Cambridge, UK) and sequenced by means of a Sequenase kit (U.S. Biochemicals, Cleveland, OH). To ensure that the product detected resulted from amplification of cDNA rather than contaminating genomic DNA, primers were chosen to cross intron/exon boundaries. Genomic DNA (10 ng) was also subjected to PCR at the same time as the cDNA to verify that no product of the

976

SHARKEY ET AL. TABLE 2. Primers used for RT-PCR, outer pair A and B, inner pair Cand D. Specificity HistRS

TNFa

CSF-1

SCF

LIF

IL-6

IL-6R

c-fms

TNFRp60O

TNFRp80

gp130

LIFR

c-kit

PAF-R

Primer A B C D A B C D A B C D A B C D A B C D A B C D A B C D A B C D A B C D A B C D A B C D A B C D A B C D A B C D

Sequence 5'-3' CCGCAGGTCGAGACAGC CAAACACCTTCTCGCGAA CTTCAGGGAGAGCGCGTGC TCATCAGGACCCAGCTGTGC ATGAGCACTGAAAGCATG TCACAGGGCAATGATCCC ATCCGGGACGTGGAGCTG AAAGTAGACCTGCCCAGAC CTCCTTrGACAAGGACTGGA CTCTGGTTGCTCCAAGGGA TGCTGAATGCTCCAGCCAA TCTGAGGCTCTTGATGGCT CAATGCGTGGACTATCTGCC GTTCTAAATGAGACCCAAGT AACAGCTAAACGGAGTCGCC ACAGTGTTGATACAAGCCAC AGATCAGGAGCCAACTGGCA ACCCAACTCCTGAGATCCCT CAGCTCAATGGCAGTGCCAA GTTCACAGCACACTTCAAGAC CCAGGAGCCCAGCTATG CATrGCCGAAGAGCC AACTCCTTCTCCACAAGCG TGGACTGCAGGAACTCCTT ATGGGAAGGAGGCTGCT TTGCTGAACTTGCTCCC TCCACGAACTCTGGAAACTAT ACTATGTAGAAAGAGCTGTC CCTGGTGAACAGCAAGTT AGGTCCTACTAGTTGGCA CTATAAACTGGTGAAGGA CAACACCATGAGAACAGTA CCTGGAGTGCACGAAGTT GCAAAGCGCCTCCTCGA TGCCTACCCCAGATTGAG CACCACGGCGTACAGCGT GCAGCACCTGCTGATCACA CTGCCTCTGCTCTTGGCC CGCCGAGCTCCAGCAGC AGGTCAGAGGGAGGCTGC TTGACTAGTGACACATTGTAC TGAAACTTGCMGACCTTT GGTACGAATGGCAGCATACA CTGGACTGGATTCATGCTGA GAAAACTGTAAAGCATTACA AGAGTCTGGAGACACTAA CAAAAGAGTGTCTGTGAG CCATGTATTTACATTGGC GAAGTACAGTGGAAGGTTGTT CATCGGCCACTAAAGTGTGCT GGTTGTTGAGGCAACTGCTTA GGTGACCCAAACACTGATTC CCAGGACCCAGACAGAGAC GAGTTCTGGAITFCCAACAGC ACACGGTCACTGCAGCTGAA CTGGCTCTGCATCATCCCT

expected size resulted from genomic DNA. The cycle profile was 30 sec at 95°C, 30 sec at X°C, 30 sec at 72°C for 30 cycles, where X is the annealing temperature for each pair of cytokine primers (external and internal primers, respectively) as follows: TNFa, 49 and 54°C; CSF-1, 55 and 54°C; LIF, 60 and 60°C; IL-6, 55 and 55°C; interleukin-6 receptor (IL-6R), 48 and 52°C; c-fms, 50 and 48°C; histidyl tRNA synthetase (HistRS), 52 and 59°C; TNFa receptor p60 (TNFRp60), 52

Fragment size (bp)

110

666 170-a 1064-b 716-c

966

477

609

306

581

517

580

713

458

1148

1150

Position on cDNA

Reference

518-534 791-773 595-613 704-685 86-103 784-770 104-121 769-751 620-638 1762-1744 670-688 1733-1715 9-28 1284-1264 61-80 1027-1008 187-206 703-684 207-226 681-661 73-89 722-707 90-108 699-681 666-682 1048-1032 700-720 1006-987 2877-2894 3509-3491 2895-2912 3476-3459 819-836 1581-1565 838-955 1344-1327 1037-1055 1660-1643 1059-1079 1639-1622 1744-1764 2514-2495 1767-1790 2480-2461 2807-2826 3310-3292 2831-2848 3289-3272 1681-1702 3105-3086 1827-1847 2975-2955 1-)96-(-)77 1211-1190 (-)76-(-)57 1189-1171

[13]

[14]

[161

[17]

[181

[19]

[20]

[211

[221

[23]

[24]

[251

[26]

and 56°C; TNFa receptor p80 (TNFRp80), 56 and 56°C; gp130, 50 and 54°C; LIF receptor (LIFR), 48 and 48°C; SCF, 54 and 54°C; c-kit, 56 and 56°C; and platelet-activating factor receptor (PAF-R), 57 and 58°C. Each RNA sample was transcribed to cDNA twice. Each of these cDNA samples was then analyzed by RT-PCR using primers for HistRS. On both occasions, amplification of HisRS was successful, indicating successful cDNA synthesis.

CYTOKINES AND RECEPTORS IN HUMAN EMBRYOS

977

Both cDNA samples from each embryo were then analyzed for expression of all the transcripts of interest. RESULTS The technique of RT-PCR was applied to total RNA extracted from human embryos produced by in vitro fertilization. Embryos were cultured to the appropriate stage, and total RNA was extracted. In order to produce detectable RTPCR product from total RNA extracted from a single embryo, a nested PCR protocol was employed in which the cDNA was subjected to two sets of PCR amplification with an external primer pair followed by an internal pair. Initially, cDNA from each embryo was tested with primers for HistRS to confirm successful RNA isolation and reverse transcription. The expected 110-bp amplification product for HistRS was detected in mRNA from embryos at all stages of development, as well as in decidua and the choriocarcinoma cell line BeWo, used as positive controls (Fig. 1, lanes p and q, respectively). No product was detected in an equal volume of embryo culture supernatant extracted and subjected to RT-PCR in the same way, indicating that there was no contamination of the culture with extraneous cDNA or RNA. Amplification with HistRS primers using genomic DNA yielded diffuse bands greater than 400 bp. Examples of similar RT-PCR analysis with primers for various growth factor receptor peptides and their cognate ligands are shown in Figure 1. The developmental expression of the following receptors and their ligands was investigated: c-fms (and CSF-1), IL-6R (and IL-6), LIFR (and LIF), TNFRp60 and TNFRp80 (and TNFa), c-kit (and SCF), gp130 (which is a component of the LIF and IL-6 receptors), and the receptor for PAF (PAF-R). Stocks of cDNA were reversetranscribed from each embryo RNA sample on two separate occasions, and the PCR assays were repeated twice on each cDNA stock. The results are shown in Figure 1, which displays the pattern of expression for each growth factor or receptor transcript during preimplantation development. The identity of the PCR fragment of the correct size was confirmed by cloning and sequencing of the PCR product. In cases in which novel-sized products were seen, these were also cloned and sequenced. The mRNA encoding receptors and ligands exhibited a variety of patterns of expression. For the ligands LIF and CSF-1 as well as the receptor for PAF, no expression was seen at any stage during preimplantation development (data not shown). On the other hand, c-fms and c-kitwere expressed throughout all stages analyzed (Fig. 1) except in the cDNA from the morula (lane f). This particular embryo may have been developmentally abnormal, since although it was positive for HistRS and gp130, indicating successful cDNA synthesis, it was negative for c-fms and c-kit when the stages before and after were positive for these receptors. Other receptor and ligand mRNAs showed more com-

FIG. 1. Agarose gel showing the products of nested RT-PCR amplification on RNA from human embryos. Each panel shows the products of amplification with primers specific for different cDNA targets. Amplified cDNAs from different embryos were loaded ineach lane; lanes are labeled according to cDNA designations in Table 1. Additional samples were: lane p, first trimester trophoblast; lane q,cDNA from 200 BeWo cells; lane r, 10 ng human genomic DNA; and lane s, no input cDNA, as a negative control. DNA molecular weight markers were a 123-bp ladder loaded in lane i. The sizes of the expected PCR products are shown in base pairs.

plex stage-specific transcription. This was most marked in the case of IL-6 and IL-6R, which were preferentially transcribed in the blastocysts. The mRNA for LIFR was also detected only at the blastocyst stage, as we have reported previously [271. However, we have now shown that it is not expressed at the earlier stages of development that were tested in this study. RT-PCR with primers specific for the receptors for TNFa

978 G is base 1840

SHARKEY ET AL. A is base 1938

A signal

ACC CCA AAG TTT G AA TTA AAA AAC ACA TCT GGC CTA ATG TTC thr pro lys phe glu leu lys asn thr ser gly leu met phe

exons 1

3

2

4

5

= I

CAG ATC CTT CAA AGA GTC ATA TTG CCC AGT GGT CAC CTC ACA CTC gin ile leu gin arg val ile leu pro ser gly his leu thr leu

A

7

I

8

-I

c

D

106Obp -1

CTC CAA GGC ACA ATT TTA ATT CAA AAG ATC AAA TGT ATT CAG ATG leu gin gly thr ile leu ile gin lys ile ys cys ile gin met

6

I

I I

i

B

l

new exon

156 bp 821bp

I

I

= __

-

GCA ATT TCA CTG ATG TAA ala ile ser leu met END FIG. 2. The novel cDNA sequence and predicted amino acid sequence of the shorter splice variant of gp130 cDNA. This arises due to splicing of base 1840 to base 1938 and results in a frame shift. The new reading frame predicts 45 novel amino acids before a termination codon. The 98 bp missing from this splice variant include the region encoding the hydrophobic putative transmembrane region; hence this novel cDNA would be expected to encode a soluble truncated version of gp130. Numbering of bases is according to Hibi et al. 231.

did not reveal a simple pattern of expression. TNFRp60 appeared to be expressed in some but not all embryos from the 6-cell stage on (Fig. 1). This may reflect extremely low levels of expression of this mRNA or primers that do not amplify TNFRp60 with optimal efficiency although they work adequately on cDNA made from 200 BeWo cells (Fig. 1, lane q). On the other hand, TNFRp80 was detectable only up to the 4-cell stage (Fig. 1). Finally, for two cDNAs, products whose size differed from that predicted were observed following RT-PCR. In the case of gp130, the predicted fragment is 712 bp. However, during the morula-to-blastocyst transition, a novel, smaller transcript was detected of approximately 600 bp (Fig. 1, lane g). This result is consistent, since in sample j, which derived from cDNA made from three pooled blastocysts, the two products were detected simultaneously. Upon cloning and sequencing, the smaller product appeared to arise due to an alternative splicing event that removes the exon encoding the transmembrane domain. The sequences of the novel transcript and predicted peptide are shown in Figure 2. The novel splicing pattern also involves a frame shift, resulting in 45 new amino acids, before an in-frame stop codon. This species would be predicted to encode a truncated protein of 658 amino acids, which lacks both the transmembrane domain and cytoplasmic signal transduction apparatus of full-length gp130. A similar phenomenon was observed for the mRNA encoding SCF, with three alternative transcripts expressed. These were also cloned and sequenced. The sequences are shown in Figure 3. The SCF transcript expressed at the 8cell stage (Fig. 1, lane e) proved to be the full-length SCF transcript, which consists of 8 exons as described by Martin et al. [16]. This produces a PCR fragment of 905 bp with SCF primers C and D. The smaller fragment of 821 bp, expressed at the 2-cell and morula stages (Fig. 1, lanes a and f), was identical to a previously described splice variant of SCF that results from the loss of exon 6 [12]. The largest SCF species,

B

exon 3 new exon ATG GAT GTT TTG GAAATC TGT TCA TTG TTG ATA GGGCTG ACG GCC

met asp vol leu glu ile cys ser leu leu ile gly leu thr ala TAT AAG GAA TTA TCA CTC CCT AAA AGG AAA GAA ACT TGC AGA GCA tyr lys glu leu ser leu pro lys arg lys glu thr cys arg ala ATT CAG CAT CCA AGG AAA GAC TGA CAG CTT TGA AAGAGA CCT GAT ile gin his pro arg lys asp END new exon

exon 4

AAT GAT GCA AGT AGG AAC TTG CAT GTG CTT GAACCA AGT CAT TGT

FIG. 3. A) Diagram representing the structure of the three SCF splice variants. The top variant isthe full-length mRNA described by Martinet al. [16] and seen inFigure 1, lane e. The regions encoding the signal peptide and transmembrane domain are indicated by a black bar and a solid black box, respectively. Horizontal arrows indicate the hybridization sites of PCR primers A,B,C,D used for RT-PCR. The size of the PCR product from each variant using primers Cand D is shown in base pairs. The 821-bp variant is that seen in Figure 1, lanes a and f, and results from the excision of exon 6. The 1060-bp variant is that seen in Figure 1, lane d, and results from the inclusion of a novel exon (shown as a hatched box) between exons 3 and 4. B)Complementary DNA sequence of the novel exon inthe 1061-bp splice variant and predicted polypeptide sequence. This 1061-bp species predicts an SCF polypeptide with 33 amino acids following exon 3. The novel exon encodes 33 residues in a new reading frame, before an in-frame stop codon. The cDNA and predicted amino acid sequence are shown, with the novel exon indicated by vertical arrows.

expressed at the 6-cell stage (Fig. 1, lane d), represents a splice variant that has not been described previously. Sequencing shows that this variant of 1061 bp arises due to inclusion of a novel exon of 156 bp between exons 3 and 4. This results in a frame shift and predicts a species of SCF with 33 novel amino acids following exon 3 before it terminates at an in-frame stop codon. The SCF variants, including the sequence of the novel 156-bp exon, are shown in Figure 3. DISCUSSION Many growth factors have been shown to influence the development of cultured preimplantation mammalian embryos (for reviews see [28, 29]). However, there is good evidence for species differences in expression of growth factor receptors in preimplantation development. For instance, epidermal growth factor mRNA is expressed in the pig embryo but has not been found at any stage in mouse or ovine preimplantation embryos [30-32]. Therefore the usefulness of these animal studies to researchers interested in factors controlling human preimplantation development is limited.

CYTOKINES AND RECEPTORS IN HUMAN EMBRYOS

In addition, the specific growth factors and receptors investigated in such studies frequently have been chosen on an ad hoc basis. For both ethical and practical reasons, such an approach is not suitable for use with human embryos. We have therefore used a nested RT-PCR method that has allowed us to screen for the expression of many growth factor and receptor mRNAs in single human preimplantation embryos. For transcripts detected at early embryo stages, it is likely that these represent maternal transcripts inherited in the egg. RT-PCR is unable to distinguish between maternal and embryonic transcripts. Nonetheless, this method is ideal since it is reliable, sensitive, and economical in its use of embryo material. Transcripts of all the genes tested were detected at some stage in human preimplantation embryos except for LIF, CSF-1, and PAF-R. Transcripts for these genes were not detected at any stage examined; however, the primers effectively amplified cDNA obtained from BeWo cells, confirming their suitability for the study. The identity of the expected PCR products was verified by sequencing of the cloned product. RT-PCR with primers for HistRS was used on cDNA samples to confirm that cDNA had been successfully prepared from each embryo RNA sample. Complementary DNA specific for this housekeeping gene was successfully detected in cDNA samples made even from a single 2-cell embryo, indicating that the method was sufficiently sensitive for this study. No RT-PCR product was detected from any embryo sample with primers for CSF-1, although its receptor, c-fms, was expressed in 8 of 10 samples, throughout preimplantation development. The pattern of expression for this ligand/receptor pair is therefore similar to that described in the mouse, where c-fms mRNA was detected from the 2-cell stage to the blastocyst stage by both RT-PCR and in situ hybridization [1, 33]. CSF-1 peptide has been detected in the oviduct and uterine epithelium in mice and humans, and CSF-1 has been shown to enhance the development of 2cell mouse embryos to blastocysts [1]. In the mouse, therefore, it seems likely that the CSF-1 produced by the epithelium of the uterus and the oviduct can affect the development of the preimplantation embryo. The results in the present study suggest a similar paracrine role for CSF-1 in the human. There is one report of CSF-1 bioactivity detected in human embryos up to the 8-cell stage; however, it is possible that this activity derived from cumulus cells surrounding the embryo [3]. SCF is a growth factor related in structure to CSF-1, and it acts through the c-kit tyrosine kinase receptor. In bone marrow, SCF and CSF-1 act synergistically to promote proliferation and differentiation of stem cells into macrophage colonies. In the mouse, c-kit has been shown to be expressed throughout preimplantation development [33]. We have now shown that the same is true in human embryos. At certain stages human embryos also express SCF mRNA, suggesting that this growth factor may act in an autocrine

979

fashion. This is in contrast to findings for the mouse, in which no expression of SCF was detected in preimplantation embryos [33]. SCF was expressed at the 2-cell stage and then reappeared at the 6-cell stage. This is consistent with maternal expression followed by re-expression from the embryo's genome at the 6-cell stage [34]. SCF transcripts appeared to show stage-specific differences in the transcript size. On cloning and sequencing, these were found to be attributable to alternative splicing of the primary transcript. Two of these variants were similar to those previously published [16, 12], and one was a novel form that predicts a form of SCF with 33 new amino acids at the carboxy terminus. Several variants of SCF have now been reported in the literature; some are membrane-bound and bioactive, whereas the activity of others is untested and their biological role remains unresolved [12, 35]. However, SCF clearly shows stage-specific splicing in these embryos. The splice variants expressed by the 8-cell and morula stages would be expected to encode SCF species that are known to be bioactive and that could therefore act in an autocrine manner through c-kit expressed on the embryo. The growth factor LIF is of particular interest since expression of LIF by the maternal uterine epithelium appears to be obligatory for implantation in mice [36]. LIF maintains the in vitro growth of embryonic stem cells derived from the ICM of murine blastocysts, which express LIFR, and LIF itself is secreted at this stage [37]. LIF also blocks primitive ectoderm development in the murine ICM and enhances blastulation in vitro [9, 38]. However, its role in human implantation is unknown. We have previously shown that LIF polypeptide is expressed by human uterine glandular epithelium in the luteal phase. In this report we confirm that LIFR mRNA is not expressed in the preimplantation embryo prior to the blastocyst stage. This indicates that whatever role LIF plays in human implantation, it cannot act directly on the embryo prior to the blastocyst stage of development. There are considerable data regarding the expression and effects of TNFa and its two receptors, TNFRp60 and TNFRp80, in human and murine preimplantation embryos. Murine and human embryos are exposed to TNFa, since it is produced by uterine epithelial cells, and preimplantation human embryos have been shown to release TNFa between the 2-cell and the morula stages in vitro [4, 5, 39,40]. The results of the present study indicate that TNFa is expressed between the 4-cell stage and the morula, since no expression of TNFa mRNA was seen outside those time points. However, TNFa mRNA expression appears to be variable between embryos and was not seen in the 6- and 8-cell samples. In mice, only TNFRp60 mRNA and protein has been detected in murine blastocysts, and TNFa was shown to reduce the number of cells in the ICM without affecting the trophectoderm development [7]. The finding that cytokines such as TNFa and CSF-1 can have deleterious effects on murine embryos has led to the suggestion that altered

980

SHARKEY ET AL.

levels of cytokines in peritoneal fluid may impair embryo development and underlie some forms of infertility in humans [41]. Endometriosis, which is often associated with infertility, is also associated with increased levels of TNFa. The present study has shown that mRNAs encoding TNFa receptors are expressed at various stages in human preimplantation embryo development. Therefore it is possible that elevated TNFa in the peritoneal fluid of some women with endometriosis could act directly on embryos and be implicated in the decreased fecundity associated with this condition. The expression of IL-6 and IL-6R mRNAs was limited to the two blastocyst samples, indicating that no IL-6 or IL-6R protein expression is possible prior to these stages. IL-6 mRNA expression has been reported in porcine, ovine, bovine, and murine blastocysts, and IL-6 is secreted by murine blastocysts [37, 42]. Although we are unaware of any reports directly demonstrating IL-6R on embryos, IL-6 is reported to inhibit the rate of murine blastocyst attachment in vitro, indicating that IL-6R is expressed by trophectoderm at this stage [43]. Both IL-6 and LIF function through a complex receptor system in which IL-6 and LIF bind to corresponding receptors with low affinity. High-affinity binding arises when the ligand/receptor complex (IL-6/IL-6R or LIF/LIFR) interacts with an accessory signal transduction molecule, gp130 [44]. Since IL-6R and LIFR are unable to initiate intracellular signaling in the absence of gp130, we also undertook RT-PCR analysis to determine the expression of gp130. This showed that gp130 mRNA is expressed in the morula and blastocyst stages. An unexpected finding was that during the morula-to-blastocyst transition, the size of the gp130 PCR product decreased by about 100 bp. Sequencing of the smaller product indicated that it results from a novel splice variant of the gpl30 mRNA. This new splice variant would be predicted to encode a truncated form of gp130, with 45 novel amino acids at the carboxy terminus before an inframe stop codon. This species lacks the transmembrane domain and intracytoplasmic signaling apparatus of normal gp130, so is unlikely to be able to act as a signal transducer for LIF and IL-6. It is not clear whether this novel species encodes a soluble form of gp130 or a membrane boundform lacking signal transduction capability. Resolution of this question will require expression studies, which are underway. Soluble gp130 protein has recently been detected in human serum and has been shown to antagonize the action of IL-6 and LIF. When the cytokines associate with their respective receptors, soluble gp130 is able to bind to this complex, blocking association with membrane-bound gp130 [45]. If this novel gpl30 mRNA turns out to encode a membrane-bound but truncated form of gp130, this might also be expected to behave as an antagonist by acting as a "dummy" signal transducer. The biological function of such antagonists of the action of gp130 in human blastocysts is unclear. One possibility is that selective expression in the blastocoel cavity

in the late blastocyst stage would inactivate the effects of LIF in preventing differentiation of the ICM, allowing ICM differentiation to proceed at implantation. The finding that no transcript encoding the cloned receptor for PAF was found at any embryonic stage is of considerable interest to those concerned with improving IVF success rates in humans. PAF has been shown to be secreted by the preimplantation embryos of a number of species, including humans, and has been implicated in the establishment of pregnancy (reviewed in [461). When added to embryo culture medium in a blind trial, PAF appeared to increase the pregnancy rate [47]. Addition of PAF to defined culture medium has also been reported to enhance the development of human and murine embryos [48, 49]. This has led to the suggestion that PAF may act directly on the embryo. The results of our study suggest that PAF cannot act directly on the human embryo through the existing cloned PAF receptor. Supplementation of embryo culture media with PAF is now the subject of a multicenter IVF trial. The increased number of births may allow a clearer answer regarding whether PAF's mode of action is direct or indirect. In conclusion, we have used RT-PCR to screen human preimplantation embryos for the expression of mRNAs encoding a variety of growth factor and receptor cDNAs. This method is sensitive and economical and has allowed determination of the stages at which the embryo has the potential to express these proteins during in vitro development. It should be noted, however, that the pattern of expression of such genes may not be identical to that in vivo. This analysis has shown that several mRNAs are expressed in a stagespecific fashion, particularly in the blastocyst. Furthermore, the mRNAs encoding gp130 and SCF exhibit altered splicing patterns at different developmental stages. The novel form of gp130 predicts a truncated form that could antagonize the action of LIF and that may have a role in allowing ICM differentiation, at the time of implantation. To determine the functional significance of these normal and novel gene transcripts, the next stage will be to determine which are expressed as functional proteins. ACKNOWLEDGMENTS We would like to thank the Consultant and theatre staff of Addenbrookes' Hospital, Cambridge, and of Bourn Hall Clinic for their help with this study. Our thanks also to Sue Smith for invaluable tissue collection.

REFERENCES 1. Pampfer S, Areci RJ, Pollard IW. Role of CSF-1 and other lymphatic factors in mouse preimplantation development. Bioessays 1991; 13:535-540. 2. Tartakovsy B, Ben Yair E. Cytokines modulate preimplantation development and pregnancy. Dev Biol 1991; 146:345-352. 3. Zolti M, Ben-Rafeal Z, Meironi K, Sheimesh M, Bider D, Mashiacla S, Apole R. Cytokine involvement in oocytes and early embryos. Fertil Steril 1991; 56:265-272. 4. Witkin S, Liu H-C, Davis OK, Rosenwatz Z. TNF is present in maleval sera in embryo culture fluids during IVF. J Reprod Immunol 1991; 19:85-93. 5. Lachappelle MH, Miron P, Hemmings R, Falcone T, Granger L, Bourque J, Langlais J.

CYTOKINES AND RECEPTORS IN HUMAN EMBRYOS Embryonic resistance to tumour necrosis factor-a mediated cytotoxicity: novel mechanism underlying maternal immunological tolerance to the fetal allograft. Hum Reprod 1993; 8:1032-1038. 6. HillJA, Haimovici F, Anderson DJ. Products of activated lymphocytes and macrophages inhibit mouse embryo development in vitro. J Immunol 1987; 139:2250-2254. 7. Pampfer S, Wuu Y-D, Vanderheyden I, De Hertogh R. Expression of tumour necrosis factor-a receptors and selective effect of TNF-a on the inner cell mass in mouse blastocysts. Endocrinology 1994; 134:206-212. 8. Harvey MB, Kaye PL. Insulin-like growth factor I stimulates growth of mouse preimplantation embryos in vitro. Mol Reprod Dev 1992; 31:195-199. 9. Shen MM, Leder P. Leukaemia inhibitory factor is expressed by the preimplantation uterus, and selectively blocks primitive ectoderm formation in vivo. Proc Natl Acad Sci USA 1992; 89:8240-8244. 10. Bowman P, McLaren A. Cleavage of mouse embryos in vivo and in vitro. J Embryol Exp Morphol 1970; 24:203-207. 11. Chomsczynski P, Sacchi E. Single-step method of RNA isolation by acid guanidium thiocyanate-phenol-chloroform extraction. Anal Biochem 1987; 162:156-159. 12. Sharkey AM, Charnock-Jones DS, Brown KD, Smith SK. Expression of the mRNA for kitligand in human placenta: localisation by in situ hybridization and identification of alternatively spliced variants. Mol Endocrinol 1992; 6:1235-1241. 13. Raben N, Borrelio F, AminJ, Horwitz R, Fraser D, Plotz P. Human histidyl-tRNA synthetase: recognition of amino acid signature regions in class 2a amino-acyl t-RNA synthetases. Nucleic Acids Res 1992; 20:1075-1081. 14. Wang AM, Creasey AA, Ladner MB, Lin LS,StricklerJ, van ArsdellJN, Yamamoto R,Mark DF Molecular cloning of the complementary DNA for human tumor necrosis factor. Science 1985; 228:149-154. 15. Wong GG, Temple PA, Leary AC, Witek-GiannottiJS, Yang Y-C, Ciarletta AG, Chung M, Murtha P, Kris R, Kaufman RJ, Ferenz CR, Sibley BS, Turner KJ, Hewick RM, Clark SC, Yanai N,Yokota H, Yamada M,Saito M, Motoyoshi K,Takaku F. Human CSF-1: molecular cloning and expression of 4-kb cDNA encoding the human urinary protein. Science 1987; 235:1504-1508. 16. Martin FH, Suggs SV, Langley KE, Lu HS, Ting J, Okino KH, Morris CF, McNeice IK, Jacobsen FW, Mendias EA, Birkitt NC, Smith KA, Johnson MJ, Parker VP, Flores JC, Patel AC, Fisher EF, Erjavec HO, Herrera CJ, Wypych J, Sachdev RK, Pope JA, Leslie I, Wen D, Lin C-H, Cupples RL, Zsebo KM. Primary structure and functional expression of rat and human stem cell factor DNAs. Cell 1990; 63:203-211. 17. Moreau JF, Donaldson DD, Bennett F, Witek-GiannottiJS, Clark SC, Wong GG. Leukaemia inhibitory factor is identical to the myeloid growth factor human interleukin for DA cells. Nature 1988; 336:690-692. 18. Zilberstein A, Ruggieri R, Korm JH, Revel M. Structure and expression of cDNA and genes for interferon-beta-2, a distinct species. EMBO J 1986; 5:2529-2537. 19. Yamasaki K, Taga T, Hirata Y, Yawata H, Kawanishi Y, Seed B, Taniguchi T, Hirano T, Kishimoto T. Cloning and expression of the human Interleukin-6 receptor. Science 1988; 241:825-828. 20. Coussens L,Van Beveren C, Smith D, Chen E,Mitchell RL,Isacke CM, Verma IM, Ullrich A. Structural alteration of viral homologue of receptor proto-oncogene fms at carboxyl terminus. Nature 1986; 320:277-280. 21. Nophar Y, Kemper O, Brakebusch C, Engelmann H, Zwang R, Aderka D, Holtmann H, Wallach D. Soluble forms of tumour necrosis factor receptors. EMBO J 1990; 10:32693278. 22. Smith CA, Davis T, Anderson D, Solam L, Beckmann MP, Jerzy R, Dower SK, Cosman D, Goodwin RG. A receptor for human tumour necrosis factor defines an unusual family of cellular and viral proteins. Science 1990; 248:1019-1023. 23. Hibi M, Murakami M, Saito M, Hirano T, Taga T, Kishimoto T. Molecular cloning and expression of an IL-6 signal transducer, gpl30. Cell 1990; 63:1149-1157. 24. Gearing DP, Thut CJ, Van den Bos T, Gimpel SD, Delaney PB, KingJ, Prince V,Cosman D, Beckmann MP. Leukaemia inhibitory factor receptor is structurally related to the IL6 transducer, gp130. EMBO J 1991; 10:2839-2848. 25. Yarden Y, Kuang W-J, Yang-Fen T, Coussens L, Munemitsu S, Dull TJ, Chen E, SchlessingerJ, Francke U,Ullrich A. Human proto-oncogene c-kit a new cell surface receptor tyrosine kinase for an unidentified ligand. EMBO J 1987; 6:3341-3351. 26. Nakamura M, Honda Z, Izumi T, Sakanaka C, Mutoh H, Minami M, Bito H, Seyama Y,

981

Matsumoto T, Noma M, Shimuzu T. Molecular cloning and expression of platelet activating factor receptor from human leukocytes. J Biol Chem 1991; 266:20400-20405. 27. Charnock-Jones DS, Sharkey AM, Fenwick P, Smith SK. Leukaemia inhibitory factor mRNA concentration peaks in human endometrium at the time of implantation and the blastocyst contains mRNA for the receptor at this time. J Reprod Fertil 1994; 101:421426. 28. Schultz GA, Heyner S. Growth factors in preimplantation embryos. Oxf Rev Reprod Biol 1992; 15:43-81. 29. Anderson ED. Activities of growth factors in preimplantation embryos. J Cell Biochem 1993; 53:280-287. 30. Vaughan TJ, James PS, Pascall JC, Brown KD. Expression of the genes for TGF-a, EGF and EGF receptor during early pig development. Development 1992; 116:663-669. 31. Rappolee DA, Brenner CA, Schultz R, Mark D, Werb Z. Developmental regulation of PDGF, TGF-a and TGF-Pj genes in preimplantation mouse embryos. Science 1988; 241:1823-1825. 32. Watson AJ, Watson PH, Arcellana-Panlilio, Warnes D, Walker SK, Schultz GA, Armstrong DT, Seamark RF. A growth factor phenotype map for ovine preimplantation development. Biol Reprod 1993; 50:725-733. 33. Arceci RJ, Pampfer S, PollardJW. Expression of CSF-1/c-fms and SF/c-kit mRNA during preimplantation mouse development. Dev Biol 1992; 151:1-8. 34. Braude P Bolton V, Moore S. Human gene expression first occurs between the four and eight cell stages of preimplantation development. Nature 1988; 332:459-461. 35. FlanaganJG, Chan DC, Leder P. Transmembrane form of the kit ligand growth factor is determined by alternative splicing and is missing in the Sld mutant. Cell 1991; 64:10251035. 36. Stewart CL, Kaspar P Brunet LJ,Bhatt H, Gadi I, Kontgen F, Abbondanzo SJ. Blastocyst implantation depends on maternal expression of leukaemia inhibitory factor. Nature 1992; 359:76-79. 37. Murray R, Lee F, Chin C-Y. The genes for leukaemia inhibitory factor and interleukin-6 are expressed in mouse blastocysts prior to the outset of haemopoiesis. Mol Cell Biol 1990; 10:4953-3956. 38. Fry RC, Batt PA, Fairclough RJ, Parr RA. Human leukemia inhibitory factor improves the viability of cultured ovine embryos. Biol Reprod 1992; 46:470-474. 39. Hunt JS, Chen HL, Hu XL, Pollard JW. Normal distribution of TNF-a mRNA and protein in the uteri, placentas and embryos of osteopetrotic (op/op) mice lacking CSF-I. Biol Reprod 1993; 49:441-452. 40. Chen HL, Yang XL, Yelavarthi KK, Fishback JL, Hunt JS. TNF-a mRNA and protein are present in human placental and uterine cells of early and late stages of gestation. Am J Pathol 1991; 139:327-335. 41. Syrop CH, Halme J. Peritonealfluid environment and infertility. Fertil Steril 1987; 35:696698. 42. Mathialagan N, Bixby JA, Roberts RM. Expression of IL-6 in porcine, ovine and bovine preimplantation conceptuses. Mol Reprod Dev 1992; 32:324-330. 43. Haimovici F, Hill JA, Anderson DJ. The effects of soluble products of activated lymphocytes and macrophages on blastocyst implantation events in vitro. Biol Reprod 1991; 44:69-75. 44. Taga T, Kishimoto T. Cytokine receptors and signal transduction. FASEB J 1993; 7:33873396. 45. Narazaki M, Yasukawa K, Saito T, Ohsugi Y, Fukui H, Koishihara Y, Yancopoulos GD, Taga T, Kishimoto T. Soluble forms of the interleukin-6 signal-transducing receptor component gp130 in human serum possessing a potential to inhibit signals through membrane-anchored gp130. Blood 1993; 82:1120-1126. 46. O'Neill C. A consideration of the factors which influence and control the viability and developmental potential of the preimplantation embryo. Ballieres Clin Obstet Gynaecol 1991; 5:159-178. 47. O'Neill C, RyanJP, Collier M, Saunders DM, Ammit AJ, Pike IL. Outcome of pregnancies resulting from a trial of supplementing human IVF culture media with platelet activating factor. Reprod Fertil Dev 1992; 4:109-112. 48. O'Neill C, Ryan JP, Collier M, Saunders DM, Ammit AJ, Pike IL.Supplementation of IVF culture media with platelet activating factor increased the pregnancy rate following embryo transfer. Lancet 1989; ii:769-772. 49. Ryan JP, Spinks NR, O'Neill C, Wales RG. Platelet activating factor production by mouse embryos in vitro and its effect on embryonic metabolism. J Cell Biochem 1989; 40:1-9.