Horse Conceptuses Secrete Insulin-Like Growth Factor-Binding Protein 3¹

Abstract

Insulin-like growth factor-I (IGF-I) promotes early embryonic development in several species. In the rabbit, IGF-I binds to the embryonic coats from Day 3 of development onward by a 38-kDa protein that is probably insulin-like growth factor-binding protein 3 (IGFBP3). In the present study, ligand, Western, and Northern blot analyses were used to demonstrate the presence of IGF-I-binding activity, several immunoreactive IGFBP3 proteins, and IGFBP3 mRNA in horse conceptuses with particularly large amounts of immunoreactive IGFBP3 in the conceptus capsule. In addition, immunoprecipitation of radiolabeled proteins showed that cultured horse conceptuses secreted IGFBP3 into the culture medium. Endometrial samples from mares also contained IGFBP3 mRNA and protein; but there was no evidence of secretion of IGFBP3 into the uterine lumen by ligand blot analysis, and there was evidence of only very small amounts by Western blot analysis.

These results indicate that the horse conceptus secretes significant quantities of IGFBP3 toward the conceptus capsule from as early as Day 10 after ovulation. Thus, most of the IGFBP3 contained within the capsule, which binds IGF-I to this special extracellular matrix of the preimplantation horse conceptus, is likely to be embryonic in origin. IGFBP3 in the horse conceptus capsule may enhance or modulate the action of IGFs on the developing conceptus.

Introduction

Several growth factors play important roles in early development of the conceptus via their effects on cell proliferation and morphogenesis. Insulin-like growth factor (IGF)-I and IGF-II have both been implicated in early development [1–3], and in vitro studies have shown that IGF-I can stimulate glucose uptake, protein synthesis, cell proliferation, and conceptus development [4–7]. Furthermore, IGF-I appears to act as a survival factor by decreasing apoptosis [7]. IGF-I receptors are present in mouse embryos from the compacted 8-cell stage of development and in rabbits from the morula stage onward [8–10]. However, studies on the synthesis of IGF-I by the conceptus prior to implantation have led to conflicting results. While some investigators were unable to detect IGF-I mRNA in preimplantation embryos [9, 11, 12], others described its presence from the oocyte to the blastocyst stage [13, 14]. However, IGF-I protein has not been shown to be synthesized by the preimplantation conceptus [11, 15]. On the other hand, maternal IGF-I is present in oviductal and uterine secretions and therefore surrounds the conceptus at all times [15–17]. We demonstrated previously that IGF-I is bound to the external coat of the preimplantation rabbit conceptus by a 38-kDa protein that is probably insulin-like growth factor-binding protein 3 (IGFBP3) [10]. In addition, mRNAs for IGFBP 2, 3, and 4 are present in preimplantation mouse and bovine conceptuses [15, 18]. Together these findings suggest that preimplantation mammalian embryos synthesize IGFBPs that may regulate the bioavailability and action of IGFs on preimplantation development. IGFBP3 appears to be of special interest in this regard, as in coculture with Vero cells this binding protein has been implicated as being responsible for the observed enhanced development of embryos [19, 20].

The coats of conceptuses are extracellular matrices (ECM) that surround them during most of the time of preimplantation development [21, 22]. This ECM is formed initially by the oocyte and granulosa cells and is modified subsequently by oviductal, uterine, and, in some cases, conceptus secretions [21–24]. In the horse, the zona pellucida surrounds the oocyte and conceptus until Day 8 after ovulation [22]. On Day 6 a new coat, the capsule, forms beneath the zona pellucida [22, 25]. This is derived mainly from products of the conceptus [24] and is similar to the neozona in the rabbit [21]. The horse capsule increases approximately 20-fold in mass between Days 10 and 18 after ovulation, but it begins to disintegrate at about Day 20 and has disappeared by Day 23 [26, 27].

To investigate IGFBPs in the horse conceptus, early conceptuses (Days 10–16) were recovered nonsurgically from mares and cultured in serum-free medium. The conceptus, capsule, and culture medium were analyzed for the presence of IGFBPs by ligand blotting using ¹²⁵I-IGF-I and by Western blotting using an anti-IGFBP3 serum. In addition, tissues (liver, endometrium, conceptus) were analyzed by Northern blotting to detect mRNA for IGFBP3, and the supernatants of conceptuses cultured in the presence of [³⁵S]methionine were analyzed by immunoprecipitation with the anti-IGFBP3 serum.

Materials and Methods

Sample Collection

Twelve thoroughbred mares, aged 3–7 yr, were used, some of them on more than one occasion, for the collection of 20 early conceptuses (Days 10–16) and one Day 30 conceptus. Follicular development and ovulation were monitored by daily transrectal ultrasound examination of the ovaries, and the mares were inseminated once with diluted fresh stallion semen when they showed a follicle of >35-mm diameter. Ovulation was confirmed subsequently by measuring a rise in progesterone concentrations in peripheral serum recovered daily, during and after estrus [28]. Day 1 of pregnancy was taken as the first day on which the serum progesterone concentration exceeded 1 ng/ml. Pregnancy was confirmed, and the conceptus was assessed for regular development by transrectal ultrasound examination from Day 10 onward.

Immediately before the conceptus was removed, a sample of uterine secretion was obtained by infusion of 20 ml sterile PBS into the horn of the uterus that did not contain the conceptus via a catheter passed through the working channel of a strobe-lit videoendoscope, as described by Bracher and Allen [29]. The conceptus migrates from one horn to the other frequently during this early time of development, and there should therefore be no difference between the two horns. The uterus was massaged gently per rectum, and as much fluid as possible was removed by directing the tip of the catheter into the pool of fluid. The conceptus was then recovered intact on Days 10, 12, or 16 of gestation by uterine lavage using an endotracheal tube passed through the cervix and flushing with 500–1000 ml sterile PBS. In total, 20 conceptuses were collected (five Day 10, ten Day 12, five Day 16), washed several times in PBS, and either cultured for immunoprecipitation or processed for ligand, Western, or Northern blot analysis.

For immunoprecipitation, five Day 10 and four Day 12 conceptuses were cultured individually for 4 h in a mixture of 1 ml 1:1 Dulbecco's modified Eagle's medium (DMEM) and RPMI medium, without glutamine and methionine (ICN, Thame, UK), supplemented with 3 mmol glutamine, 30 μmol methionine, and 5.5 MBq [³⁵S]methionine per milliliter (0.13 μmol; 43.5 TBq/mmol; ICN). Before and after culture, the diameter of each conceptus was measured and its volume estimated in order to assess conceptus viability and growth during culture. A stable or increasing conceptus volume is a valid measure of viability and growth in the horse, as the conceptus fluid at this developmental stage is hypotonic compared to culture media [30] and if damaged, the conceptus shrinks inside the capsule due to loss of water from the conceptus cavity toward the culture medium. The medium was collected and stored at −70°C prior to analysis.

For ligand and Western blot analysis, six Day 12 and three Day 16 freshly harvested conceptuses were washed several times in PBS and once in sterile water. Each conceptus was placed on a dry glass microscope slide from which excess fluid was removed with a thin pipette before it was punctured with a glass needle and the conceptus fluid collected. The residual conceptus was then floated in a drop of water so that the conceptus membranes and capsule could be collected separately and stored at −70°C until they were analyzed.

For Northern blot analysis, two of the Day 16 freshly harvested conceptuses were used. In addition, a Day 30 conceptus was collected nonsurgically by being punctured with a sharpened catheter through the working channel of the videoendoscope before the collapsed conceptus was flushed from the uterus with PBS.

Samples of the endometrium were collected by taking biopsies on Day 0 (estrus) and from pregnant mares on Days 10, 12, and 18 after ovulation (2 samples for each day). Liver samples were collected from a slaughtered sheep and horse. All samples were frozen immediately in liquid nitrogen and stored at −70°C until required for RNA extraction.

Ligand Blot Analysis

Ligand blots were performed by the method of Hossenlopp et al. [31] using two uterine flushes from each stage (Days 10, 12, and 16); two endometrial biopsies from each stage (Days 10, 12, and 16); the capsules, conceptus membranes, and conceptus fluid samples from a Day 12 and a Day 16 conceptus; and two culture supernatants (Days 10 and 12; see below). The samples were subjected to SDS-PAGE using a 5% stacking gel and an 8–18% resolving gel under nonreducing conditions. The endometrial biopsies, capsules, and conceptus membranes were homogenized in approximately 100 μl PBS (0.01 M, pH 7.2)/mg tissue and diluted in loading buffer (0.5 M Tris, pH 6.8; 10% [v:v] glycerol; 10% [w:v] SDS; 0.05% [w:v] bromophenol blue), whereas uterine flushings, conceptus fluid, and culture supernatants were diluted in loading buffer. The samples were heated at 100°C for 5 min and microcentrifuged, and 20 μl of each sample was loaded onto the gel. After electrophoresis, the gel was blotted onto a polyvinylidene fluoride membrane (Sartorius, Epsom, UK) by wet blotting (Trans Blot Cell; BioRad, Hempstead, UK; transfer buffer: 48 mM Tris, 3 mM glycine, 20% [v:v] methanol, 1.3 mM SDS, pH 8.3) for 3 h at 200 mA, and the membrane was dried at room temperature for 10 min. The blot was then soaked for 30 min in 100 ml saline (65 mM Tris, pH 7.4; 0.05% Nonidet P-40) at 37°C and washed for 2 h in 65 mM Tris containing 3% BSA (insulin and IGF-I free) at 4°C, and subsequently for 20 min in 65 mM Tris, pH 7.4, containing 150 mM NaCl and 0.2% Tween 20 at 4°C. It was then incubated overnight at 4°C with approximately 500 000 cpm ¹²⁵I-human (h)IGF-I (iodinated using chloramine T to a specific activity of 50 μCi/μg) in 25 ml 65 mM Tris, pH 7.4, containing 3% BSA and 0.2% Tween 20; next it was washed twice, each for 15 min, in 100 ml 65 mM Tris, pH 7.4, supplemented with 150 mM NaCl and 0.1% Tween at room temperature and subsequently three times, each for 15 min, in 100 ml Tris containing 150 mM NaCl. After drying at room temperature, the blot was exposed to x-ray film at −70°C (X-Omat AR; IBI-Kodak Ltd., Cambridge, UK) for 6–9 days.

Western Blot Analysis

The IGFBP3 antiserum used for Western blotting and immunoprecipitation was raised in a sheep against a synthetic peptide corresponding to residues 67–79 of hIGFBP3 (⁶⁷CQPSPDEARPLOA⁷⁹), an antigenic region selected from computer-generated predictions of residue surface probability, hydrophilicity, flexibility, and Jameson-Wolf antigenic index. The peptide was synthesized and the antiserum raised by the Microchemical Facility at the Babraham Institute, Cambridge. Specific binding to recombinant-derived nonglycosylated hIGFBP3 (Celltrix Pharmaceuticals, Santa Clara, CA) and to horse serum was tested by Western blotting (see Fig. 2).

The procedure for the Western blot analysis was similar to that described previously for the ligand blots, except that the loading buffer contained 5% (v:v) β-mercaptoethanol; 30 mM Tris, pH 6.8; 1.5 M urea; 7.5% (v:v) glycerol; 0.5% (w:v) SDS; and 0.05% (w:v) bromophenol blue. After blotting, the membrane was blocked with 5% (w:v) milk powder in PBS-Tween 20 (0.1% w:v) for at least 1 h at room temperature. The primary antibody or preimmune serum was diluted 1:2000 in 5% milk powder in PBS and incubated with the blot at 4°C overnight. After three washes in PBS/Tween, the second antibody (rabbit anti-sheep IgG labeled with horseradish peroxidase diluted 1:5000 in PBS/Tween; Dako Ltd., Glostrup, Denmark) was added and left at room temperature for 1 h. After further rigorous washing, detection of the second antibody was performed using ECL reagents (Amersham, Little Chalfont, UK; ECL: horseradish peroxidase-induced chemiluminescent reaction) and exposure to x-ray film (X-Omat AR; IBI-Kodak Ltd.) for 1–3 min.

Northern Blot Analysis

Total RNA was isolated from two intact Day 16 conceptuses; one Day 30 conceptus; six endometrial biopsies taken at Days 10, 12, and 18 of gestation (2 at each stage); and the horse and sheep liver samples using the method of Chomczynski and Sacchi [32]. Approximately 20 μg total RNA per track was loaded onto a 1% (w:v) agarose formaldehyde gel, which was then run for 3 h at 50 V. After the gel was soaked in a denaturing solution of 0.05 M NaOH and 0.15 M NaCl for 15 min followed by rinsing in 1 M Tris, pH 7, containing 0.15 M NaCl for 15 min, the RNA was blotted onto a nylon membrane (Hybond N; Amersham) using positive pressure. The RNA was cross-linked to the membrane by UV light (Stratagene Crosslinker; Stratagene, Cambridge, UK), and the membrane was prehybridized in approximately 8 ml QuickHyb buffer (Stratagene) at 60°C for 1–2 h. The IGFBP3 probe was a cloned 285-base pair polymerase chain reaction product generated from nucleotides 400–685 of bovine IGFBP3 and checked by DNA sequence analysis. Approximately 50 ng of the cloned fragment was labeled with [³²P]dCTP (ICN) by random priming, separated from unincorporated [³²P]dCTP using a NucTrap column (Stratagene), and allowed to hybridize with the blot at 60°C for 2 h in QuickHyb buffer (Stratagene). The membrane was washed twice for 10 min at room temperature in double-strength SSC (single-strength SSC is 0.15 M sodium chloride and 0.015 M sodium citrate), 0.1% (w:v) SDS, and once in single-strength SSC, 0.1% (w:v) SDS for 15 min at 60°C; it was then exposed to x-ray film (X-Omat; IBI-Kodak Ltd.) for 16–48 h.

Immunoprecipitation

The culture medium supernatants in which the Day 10 or 12 conceptuses had been cultured for 4 h in the presence of [³⁵S]methionine were pooled to give two samples (Day 10 and Day 12). Aliquots (180 μl) were incubated with 20 μl of the sheep anti-hIGFBP3 serum or a control antiserum directed against a 19-kDa endometrial protein [33] that accumulates in the horse conceptus capsule and is released into the culture medium [34, 35]. After incubation on a rotating shelf overnight at 4°C, 50 μl of washed protein G sepharose beads (Pharmacia, St. Albans, UK) was added to the sample antiserum mixture, and incubation was continued on the rotating shelf for a further 24 h at 4°C. After centrifugation at 7000 × g for 2 min, the pellet was washed three times with Tris buffer (50 mM, pH 7.4) containing 0.5% Triton X-100, mixed with loading buffer (125 mM Tris, pH 6.8; 10% [w:v] SDS; 20% [v:v] glycerol, 0.05% [w:v] bromophenol blue), and heated to 100°C for 5 min. After brief centrifugation, 40 μl of the supernatant was loaded onto a 12% gel that was run for 1.3 h at 80 V. The proteins were then cross-linked overnight in 40% (v:v) methanol and 10% (v:v) acetic acid, and the gel was soaked for 20 min in a solution containing 30% (v:v) ethanol and 4% (v:v) glycerol. Finally, the gel was dried between two sheets of cellophane for 36 h and exposed to x-ray film (X-Omat AR; IBI-Kodak Ltd.) for 3 wk.

The range of proteins synthesized by the embryo and secreted into the culture medium was also analyzed by autoradiography. Aliquots (180 μl) of pooled culture media were freeze-dried on a Speedi Vac (Savant, Halbrook, NY), resuspended in 20 μl loading buffer as above, heated to 100°C for 5 min, and loaded onto a 12% gel. After electrophoresis, the gel was fixed and dried as described above and exposed to x-ray film (X-Omat AR; IBI-Kodak Ltd.) for 4 days.

Results

Ligand Blot Analysis

Representative examples of the ligand blot analysis of the endometrium, the uterine flushes, the capsule, the conceptuses, the conceptus fluid samples, and the culture medium supernatants from the incubation of conceptuses are shown in Figure 1. In all endometrial samples, i.e., a total of 4 sampled between 12 and 16 days of gestation, proteins at 46 and 49 kDa bound IGF-I. In contrast, there was no evidence of any IGFBPs in a total of 4 uterine flushes (20 μg total protein per slot) taken on Days 12 and 16 after ovulation (Day 12 shown in Fig. 1). The conceptus capsules showed a very high level of binding at 46 kDa as well as to a 36 kDa (Day 12 shown in Fig. 1). However, the Day 12 conceptuses themselves did not show any binding, although binding to a 32- and 46-kDa protein was detected in conceptus fluid. At Day 16, strong binding to a 46-kDa protein was detected in the conceptus itself and the conceptus fluid; and in the latter, binding to larger proteins (mainly around 72 kDa) was visible. The samples of culture medium in which the Day 10 and Day 12 conceptuses had been incubated for 4 h in the presence of [³⁵S]methionine showed evidence of a 46-kDa IGFBP (Day 12 shown in Fig. 1). To check whether ³⁵S present in these samples penetrated the cling film over 7 days, blots of the samples (prior to binding analysis) were exposed to x-ray film under identical conditions. The films were completely negative.

Fig. 1

This representative blot shows IGFBPs detected by ligand blot in lane A: endometrium, pregnant, Day (d)12; lane B: uterine flushing, pregnant, d12; lane C: capsule of conceptus, d12; lane D: conceptus, d12; lane E: conceptus, d16; lane F: conceptus fluid of d12; lane G: conceptus fluid of d16; lane H: culture medium in which a d12 conceptus had been incubated for 4 h. The major protein showing binding to ¹²⁵I-IGF-I was about 46 kDa in size and was strongest in the capsule from d12 conceptuses

Western Blot Analysis

The antiserum raised against a synthetic peptide based on a hIGFBP3 sequence, but not the serum collected before immunization, readily detected 200 ng nonglycosylated hIGFBP3 at 32 kDa (lanes A and B, Fig. 2). The same antiserum (but not the preimmune serum; lane D, Fig. 2) bound to proteins of 32, 49, 60, and 80 kDa in the serum of mares on Day 12 of gestation (lane C, Fig. 2). Different forms of IGFBP3 similar to these have been described previously in other species by Baxter and Martin [36], indicating that the anti-hIGFBP3 serum binds specifically to horse IGFBP3 isoforms.

Fig. 2

This representative Western blot shows proteins detected by an antiserum against a hIGFBP3 peptide (⁶⁷CQPSPDEARPCOA⁷⁹) generated in sheep. Lane A, 200 ng nonglycosylated recombinant hIGFBP3 and lane C, 2 μl horse serum from a Day (d)12 pregnant mare probed with the antiserum at a dilution of 1:2000. As a control, serum of the sheep collected before immunization was used: lanes B and D show the same samples (200 ng hIGFBP3 and 2 μl horse serum from a d12 mare) probed with the preimmune serum at a dilution of 1:2000. The antiserum reacted very well with the recombinant hIGFBP3 (lane A) and also with several proteins present in horse serum (lane C)

Since the culture medium samples, the capsules, the conceptus (Day 16), and the conceptus fluid all showed evidence of a 46-kDa IGFBP (see Fig. 1), Western blot analysis was employed to determine whether immunoreactive IGFBP3 was present in these samples. In all endometrial samples taken at Days 12 and 16 of pregnancy, a 28-kDa protein and two bands at 46 and 49 kDa were detectable (Day 12 sample shown in lane A, Fig. 3); and although no IGFBP could be detected by ligand blot in uterine fluid (see Fig. 1), two bands at 46 and 49 kDa and a single band at 20 kDa were visible by Western blot in uterine fluid from the mares at Days 12 and 16 of gestation (Day 12 sample shown in lane B, Fig. 3). In the Day 12 and Day 16 capsules, the major signal was in the area between 16 and 20 kDa, possibly representing cleaved forms of IGFBP3, with minor bands at 32, 38, 46, 49, and 60 kDa. The Day 12 and Day 16 conceptuses themselves gave a similar pattern (32, 38, 46 kDa), except that there was a weaker signal in the 22-kDa area and there were two additional very strong bands at approximately 80 and 150 kDa. The Day 12 conceptus fluid contained low amounts of the 49-kDa protein and two larger proteins (60 and 150 kDa). Supernatants from Days 10 and 12 conceptuses cultured for 4 h in the presence of [³⁵S]methionine showed only a single band at 46 kDa (Day 12 shown in lane F, Fig. 3), which corresponded very well with the 46-kDa protein detected by ligand blot (see Fig. 1).

Fig. 3

A representative Western blot using the same hIGFBP3 antiserum as in Figure 2. Lane A: endometrium, pregnant, Day (d)12; lane B: uterine flush, d12; lane C: capsule of conceptus, d12; lane D: conceptus, d12; lane E: conceptus fluid, d12; lane F: culture medium in which a d12 conceptus had been incubated for 4 h. All samples contained immunoreactive proteins, some of which were smaller than the nonglycosylated recombinant hIGFBP3 (see Fig. 2) and may therefore represent cleaved products of equine IGFBP3

Northern Blot Analysis

Since the Western blots indicated that IGFBPs in maternal and embryonic tissues detected by ligand blot were likely to be IGFBP3, it was important to determine where IGFBP3 might be expressed. Northern blot analysis demonstrated a single 2.7-kilobase (kb) transcript in sheep and horse liver RNA (Fig. 4), which was the expected size for IGFBP3 mRNA. The RNA from both Day 16 conceptuses (one shown in lane C, Fig. 4) and the Day 30 conceptus showed the same 2.7-kb transcript. Although loading of lane C and D was unequal (see the 18S ribosomal bands in Fig. 4), the Day 30 embryo appeared to contain more IGFBP3 mRNA than the Day 16 embryos. Likewise, all of the endometrial samples were positive; and although both Day 16 samples appeared to give a stronger signal than the other samples, it was not possible to statistically analyze this because of the small number of samples taken at each stage. These results indicate that both the conceptus and the endometrium express IGFBP3 mRNA.

Fig. 4

A representative Northern blot detecting IGFBP3 mRNA in liver (lane A/sheep; lane B/horse); lane C: conceptus, Day (d)16; lane D: conceptus, d30, and endometrium (lane E/estrus; lane F/pregnant, d10; lane G/pregnant, d18). A single band of the expected size for IGFBP3 mRNA (2.7 kB) was detected in all samples. The lower panel shows the 18S ribosomal RNA band stained with ethidium bromide (before transfer)

Immunoprecipitation

The conceptuses developed well during 4 h of culture as reflected by their increase in volume. The Day 10 conceptuses expanded from 49.8 ± 17.2 mm³ to 75 ± 21.5 mm³ (mean ± SEM; P < 0.001, n = 5), and the Day 12 conceptuses from 574.8 ± 22.7 mm³ to 609.7 ± 32.7 mm³ (mean ± SEM; P < 0.05, n = 4).

Since it was not possible to distinguish by Northern blot analysis where the IGFBP3 present in the conceptus capsule might be synthesized, immunoprecipitation was used to determine whether the conceptus secreted IGFBP3 into the culture medium, that is, toward the capsule. Proteins synthesized by the Day 10 and 12 conceptus were labeled by being cultured with [³⁵S]methionine, and those proteins secreted into the culture medium were detected by autoradiography (Day 10 and 12 supernatants shown in lanes A and B, Fig. 5). After the Day 10 and 12 culture supernatants were pooled, they were immunoprecipitated with the same IGFBP3 antibody used in the Western blotting experiments. Both showed a 48-kDa and a faint 23-kDa band (lanes C and D, Fig. 5), which were in the expected size range of IGFBP3 isoforms. The control serum raised against a maternal protein (P19) known to be present in the capsule of horse conceptuses at this stage [33] gave a negative result with both samples.

Fig. 5

Lanes A and B show representative autoradiographs of the culture supernatants in which intact horse conceptus had been incubated for 4 h in the presence of [³⁵S]methionine (lane A, Day (d)10 conceptus; lane B, d12 conceptus). Lanes C, D, and E show the immunoprecipitation results on the same samples using the hIGFBP3 peptide antiserum (lane C, d10 conceptus; lane D, d12 conceptus) and a control antiserum directed against P19, another equine protein (lane E, d12 conceptus)

Discussion

These results indicate that the IGFBPs detected by ligand blotting in early preimplantation horse conceptuses and their coats are most likely to be isoforms of IGFBP3. Furthermore, mRNA for IGFBP3 was detected in the conceptus, and immunoprecipitable IGFBP3 was shown to be synthesized and secreted by conceptuses as early as Day 10 of development. Although preimplantation mouse [18] and cattle conceptuses [15] transcribe IGFBP3, the actual synthesis of IGFBP3 protein by preimplantation conceptuses has not been shown before.

In a previous study we demonstrated by ligand blot analysis that a 38-kDa IGFBP, likely to be IGFBP3, is present in the coats of rabbit conceptuses and that the binding of IGF-I to the coats of rabbit conceptuses increases from the day of blastocyst formation onward [10]. Furthermore, IGF-I receptor mRNA is present from the 8-cell stage onward [9], and binding of IGF-I has been shown to occur from the morula stage onward [8, 10]. The question whether IGF-I is synthesized by preimplantation embryos is controversial. While some investigators were unable to detect IGF-I mRNA in preimplantation embryos [9, 11, 12], others described its presence from the oocyte to the blastocyst stage [13, 14]. However, IGF-I protein has not been shown to be synthesized by the preimplantation conceptus [11, 15]. Therefore, since preimplantation embryos in a number of species are bathed in maternal IGF-I [16, 17, 37, 38], we and others have suggested that maternally derived IGF-I may be regulated by embryonic IGFBPs to support the development of preattachment conceptuses [10, 19].

Whether IGFBP3 potentiates or inhibits IGF-I action has also led to controversy in the literature (see [39] for review). Several reports have described an inhibitory action for IGFBP3 [40, 41], but potentiation of IGF-I action has also been demonstrated [42], especially in in vivo studies [43, 44]. One reason for this controversy may be the higher affinity that IGFBP3 has for IGF-I in solution compared to IGFBP3 bound to extracellular matrices or cells. IGFBP3 in solution has a significantly higher affinity for IGF-I than does the IGF-I receptor and it is clearly capable of preventing IGF-I-receptor interaction. The affinity of IGFBP3 for IGF-I when it is associated with extracellular matrices or cells is significantly reduced [40, 45], and it is therefore capable of modulating the amount of IGF-I available for receptor binding [46]. Also, the local microenvironment is clearly an important modulator of IGFBP3 action. Conover et al. [47] demonstrated that, in a neutral environment, IGFBP3 is capable of blocking IGF-I-stimulated ³H-aminobutyric acid uptake, while in an acidic environment IGFBP3 potentiates the action of IGF-I. In this context it should be noted that conceptus coats are acidic because of their high amount of sialic acid [21, 24]. Therefore, IGFBP3 in this special ECM may sequester maternal IGF-I and potentiate its action as it passes it on to the embryo.

It has been suggested that IGFBP3 may contribute to the supportive effect of somatic cell coculture on early embryo development. Vero cells, which promote early embryonic development in coculture [19], release IGFBP3 into the culture medium [20], and the culture medium containing the most IGFBP3 was the most effective at promoting development. In addition, granulosa cells [48, 49], oviductal and endometrial cells [20, 50], and liver cells [51, 52] promote development of conceptuses in vitro, and these cells all secrete IGFBP3 [15, 20, 53–56]. In all these coculture systems, IGFBP3 may be one of the factors that supports development. Furthermore, IGFBP3 produced by the embryo itself may support early embryonic development. It has been established by Wiley et al. [57] and others [58, 59] that embryos should be cultured in small groups in small quantities of culture medium, suggesting the possible existence of autocrine factor(s) supporting development. IGFBP3 may be such a factor, as it is produced by the embryo and it supports early embryonic development.

Recently, IGFBP3 fragments, especially those <20 kDa in size, have engendered special interest [47]. They lose their ability to bind IGF-I, but nevertheless continue to be able to influence the mitogenic effect of IGF-I [60]. This action independent of IGF-I binding is explained by the ability of IGFBP3 to bind to cell membranes [61]. A receptor for IGFBP3 has been postulated for some time, and there is now good evidence for such a receptor [62, 63]. In this study we provide evidence that the horse capsule contains more of the cleaved forms of IGFBP3 than the native forms. Lee et al. [64] recently described the presence of IGFBP3 protease activity within porcine uterine fluid on Days 11 and 12 of pregnancy. Such a protease could be responsible for the large amount of cleaved IGFBP3 within the embryonic coats, which would result in the release of large amounts of IGFs, thus promoting development of the conceptus. As the immunoprecipitation showed, only a very small amount of cleaved IGFBP3 appears to pass through the capsule and into the culture medium, suggesting that it may be bound to the capsule. IGFBP3 cleaved by matrix metalloproteases (MMP1 to MMP3) results in a 16-kDa fragment (residues 1–176) containing an amino acid sequence (¹⁴⁹KKGHA¹⁵³) that makes it capable of binding to glycosaminoglycans [65, 66]. Thus it is possible that cleaved forms of IGFBP3 accumulate within the coats. As the cleaved forms induce cell death [60], their presence in the coats could create a defensive wall against attacking cells.

In addition to the smaller, cleaved fragments, several high molecular weight IGFBP3-immunoreactive proteins were detected in the conceptus membranes and conceptus fluid samples (see Fig. 3). These may represent SDS-stable multimers of IGFBP3 and/or its cleaved fragments, which have been shown to form under certain conditions in both IGFBP3 and IGFBP5, but not IGFBP1 [67]. Furthermore, the Western blots appear to be much more sensitive than the ligand blots, particularly for the Day 12 uterine fluid and conceptus membrane samples, which gave negative results on the ligand blots but several strong bands on the Western blots. This could be due to a genuine difference in sensitivity (e.g., differences in gel running conditions, binding constants, and detection methods) but also due to lack of binding of the labeled IGF-I to some of the immunoreactive fragments and higher molecular weight forms of IGFBP3. Significantly, only the 46- to 49-kDa protein was detected via both ligand and Western blotting in the culture medium samples and also in the immunoprecipitation experiments, apart from a very faint band at 23 kDa. This represents the protein synthesized over 4 h by the embryo in culture and indicates that neither the cleavage nor large complexes of IGFBP3 were able to form under these conditions. In addition, since the size range agrees with the size range of IGFBP3 in other species, the results provide very convincing evidence that the equine conceptus secretes IGFBP3.

In summary, the horse conceptus capsule contains IGFBP3, which, since it is secreted in relatively large amounts by conceptuses from Day 10 onward, is likely to be largely embryonic in origin. This might regulate the influence of IGFs on the conceptus. Furthermore, a large amount of IGFBP3 within the capsule is in the form of cleaved fragments, possibly the result of proteases, which would release large amounts of IGFs, thus promoting the development of the conceptus. The cleaved IGFBP3 forms a defensive wall against attacking cells, whereas the intact IGFBP3 might concentrate maternal IGF-I in the capsule and present it to the embryo.

Acknowledgments

We thank W. John Coadwell for the protein sequence analysis and design of the peptide used to raise the IGFBP3 antiserum. We also thank Tom A.E. Stout and Cornelia Gerstenberg for technical advice and support during embryo and tissue collections.

Author notes

¹

This work was supported by the Deutsche Forschungsgemeinschaft (DFG-grant, He 2688/1-1, to A.H.) and the Thoroughbred Breeders' Association of Great Britain.