Main

The purpose of the placenta is to provide an optimal milieu for fetal development. As gestation progresses, the syncytiotrophoblast lies in close proximity to the fetal circulation as well as direct contact with maternal blood(1, 2). It has the potential, therefore, to respond to regulatory factors derived from both the maternal and fetoplacental compartments. Previous studies have shown that the syncytiotrophoblast membrane is functionally and structurally polarized, supporting the hypothesis that it may be differentially regulated by factors derived from the fetal and maternal compartments(3, 4).

The human placenta was the first tissue shown to possess specific receptors for IGF-I, suggesting its importance as an IGF-regulated tissue(5). Two types of IGF receptors have been described to date. The type I IGF receptor is a heterotetramer, homologous to the insulin receptor with respect to its amino acid sequence and quaternary structure. It is composed of two ligand binding α-subunits and two β-subunits linked by disulfide bonds. The intracellular domain of the β-subunit possesses tyrosine kinase activity and undergoes ligand-induced autophosphorylation as the first step in its signal transduction pathway(6) (for review, seeRefs. 7 and 8). The type II IGF receptor, by contrast, is a single polypeptide of approximately 240 kD without tyrosine kinase activity and has been shown to be identical to the CI-M6P receptor(9–11). Until recently, all known biologic effects of both IGF-I and IGF-II were thought to be mediated through the type I IGF receptor, although the type II receptor has been linked to cellular responses by G protein-coupled mechanisms(12, 13).

Recently, a variety of atypical IGF binding sites have been identified based upon their relative affinities for the IGFs and insulin, as well as variations in the molecular mass of their β-subunits. First, hybrid receptors composed of α/β dimers from both the IGF and insulin receptors have been described(14). In addition, various forms of the β-subunit have been described that appear to be expressed in a developmental and tissue specific manner, most likely related to posttranslational processing(15, 16). Lastly, receptors structurally analogous to the type I IGF receptor have been described that share high affinity for IGF-I, IGF-II, and insulin(17–20), whereas others have been described that exhibit high affinity for IGF-II, with little or no affinity for IGF-I and insulin(21). The biochemical bases for these various “atypical” type I IGF or insulin receptors have not been fully defined but, in some instances, involve alternative splicing or post translational modification. The generation of different classes of“atypical” receptors may be a mechanism by which cell-specific and developmental responsiveness to the IGFs is regulated.

Previous studies have shown that insulin receptors are expressed primarily on the MVM surface of the syncytiotrophoblast, suggesting they are subject to regulation by maternal rather than fetal insulin(22, 23). Because of the importance of the IGFs in fetal and placental development we have begun to study the expression of IGF receptors on specific membrane surfaces of the syncytiotrophoblast to determine whether it may be differentially responsive to IGFs derived from the fetal and maternal compartments.

METHODS

Human placentas were obtained from normal, term pregnancies delivered vaginally or by cesarean section. There were no maternal histories of diabetes, hypertension, or other significant illnesses. Human IGF-I and IGF-II were obtained from Bachem, Inc. (Torrance, CA). Gln3Ala4Tyr15Leu16-IGF-I (QAYL-IGF-1) was generously provided by M. Cascieri, Merck, Sharp, and Dohme (Rahway, N.J.), Des(1-3)IGF-1, des(1-6)IGF-II, and R3IGF-I were obtained from GroPep Ltd. (Adelaide, Australia). [125I]Insulin (2000 Ci/mmol) was obtained from Amersham Corp. (Arlington Heights, IL). Carrierfree Na[125I] and[3H]dihydroalprenolol (61.6 Ci/mmol) were obtained from Dupont NEN(Boston, Mass.). Human insulin, polyethylene glycol, and BSA (RIA grade) were purchased from Sigma Chemical Co. (St. Louis, MO). Rabbit polyclonal antisera specific for the β-subunit of the human type I IGF receptor was the generous gift of Dr. K. Siddle (Cambridge, England). Rabbit polyclonal antisera raised against the bovine CI-M6P/IGF-II receptor was generously provided by Dr. S. Korenfeld (St. Louis, MO).

Preparation of placental membranes. Briefly, fresh placental tissue was dissected free of the basal and chorionic plates and washed in Earle's balanced salt solution equilibrated with 95% O2-5% CO2(24–26). The tissue was passed through a meat grinder and stirred vigorously in 1.5 vol of 150 mM NaCl. The tissue was removed by filtration. The filtrate was sequentially centrifuged at 800g for 10 min, 10 000 × g for 10 min, and 150 000 ×g for 25 min. The pellet was resuspended in 300 mM D-mannitol, 2 mM HEPES-Tris, pH 7.5. The suspension was adjusted to 10 mM MgCl2 for 10 min and centrifuged at 2200 × g for 12 min. The resulting MVM were enriched 19-20-fold as assessed by alkaline phosphatase activity.

Tissue remaining from filtration of the MVM described above were used to isolate BM(27, 28). Briefly, the tissue was sonicated and stirred in hypotonic medium to remove residual MVM and cytoplasmic contents. The remaining tissue with attached BM was incubated in 10 mM EDTA and sonicated to remove the BM from the basal lamina. Free BM were obtained by differential centrifugation, resuspended in mannitol-HEPES buffer, pH 7.4, and stored under liquid nitrogen until assayed. BM were enriched 42-fold as assessed by [3H]dihydroalprenolol binding(27), and 0.63-fold for alkaline phosphatase activity, a marker for MVMs.

Receptor binding assays. IGF-I, IGF-II, and QAYL-IGF-I (2μg) were iodinated by the chloramine T method as described by Hunter and Greenwood(29) and modified by Rechler et al.(30), to a specific activity of 100-200 μCi/μg. BM were incubated in the presence of 1% Triton X-100 for 1 h at 22°C to expose sequestered receptors caused by membrane vesicles having an“inside out” orientation. The final Triton X-100 concentration was always less than 0.015% in the binding assay. The presence of similar concentrations of Triton X-100 had no effect on receptor affinity or specificity in the MVM, which are normally oriented right side out. Furthermore, the presence of protease inhibitors (1 μg/mL leupeptin, 1 mM phenylmethanesulfonyl fluoride, and 30 μg/mL aprotinin) during solubilization of BM and MVM had no effect on receptor number or affinity of[125I]QAYL-IGF-I binding sites (data not shown). Approximately 30 000-50 000 cpm of radiolabeled peptide were incubated with 40 μg of membrane protein in the presence or absence of varying concentrations of unlabeled peptide in a total volume of 0.5 mL. The incubation was carried out overnight in binding buffer (0.1 M HEPES, 1.2 mM MgSO4, 0.12 M NaCl, 5 mM KCl, 8 mM glucose, 0.5% BSA, pH 8.0) at 4°C. Nonspecific binding was defined as binding in the presence of 200 ng of homologous, unlabeled peptide, and was typically less than 5% of the total binding. At the end of the incubation period 100 μL of 2% human IgG and 600 μL of 25% polyethylene glycol were added to each tube and incubated for 15 min at 4°C. The resulting precipitate containing ligand-receptor complexes was collected by centrifugation for 10 min at 10 000 × g at 4°C. The pellet was washed with 600 μL of 12.5% polyethylene glycol and gamma counted. KD and Bmax were determined by Scatchard analysis using the Ligand program(31). The data were analyzed using one-site, two-site, and three-site models. In all cases two-site models provided a significantly better fit of the binding data than did one-site or three-site models using the partial F test(p < 0.05).

Receptor cross-linking. Receptor cross-linking studies were performed using the bifunctional cross-linking reagent dissucinimidyl suberate. Binding was performed in binding buffer modified by the removal of BSA and the addition of 0.2% Tween-20. Aliquots (80 μg of protein) of membranes from either the BM or MVM surfaces of the syncytiotrophoblast were incubated in modified binding buffer with approximately 0.5 nM[125I]IGF-1, -IGF-2, or -QAYL-IGF-1 for 16 h at 4°C in a total volume of 250 μl. At the end of the incubation an equal volume of binding buffer, without BSA or Tween 20, containing 0.2 mM disuccinimidyl suberate was added to each tube (0.1 mM final concentration) and incubated for 15 min at room temperature. The reaction was quenched by the addition of 40% vol/vol 1 M Tris/10 mM EDTA for 5 min. The receptors were then precipitated by the addition of 2% IgG and 25% polyethylene glycol as described for the binding assay. Pellets containing the cross-linked receptors were resuspended in sample buffer with and without 2-mercaptoethanol and subjected to 5% SDS-PAGE. The gels were dried and exposed to x-ray film at -80°C for 21 d.

Immunoblotting of IGF receptors. Aliquots of MVM and BM (100μg) were solubilized in 1% Triton for 1 h at 22°C and subjected to SDS-PAGE under reducing (type I IGF receptor) and nonreducing (type II IGF receptor) conditions. The proteins were then transferred to nitrocellulose paper, blocked in 5% milk for 2 h at 22°C, and incubated at 4°C overnight with rabbit polyclonal antibodies specific for either theβ-subunit of the type I IGF receptor or the bovine type II IGF receptor which also recognizes the human type II receptor. The receptors were visualized using a chemiluminescence detection system (Amersham Corp.) per instructions. The relative abundance of receptor protein was determined by scanning laser densitometry (Molecular Dynamics, Sunnyvale, CA) and expressed in arbitrary units.

Ligand blotting of IGF-binding proteins. Aliquots of each membrane preparation (40 μg) were dispersed in electrophoresis sample buffer, boiled for 5 min, and subjected to SDS-PAGE under nonreducing conditions. IGF binding proteins were electroblotted onto nitrocellulose paper and incubated with 4 × 106 cpm [125I]IGF-II at 4°C overnight essentially as described by Hossenlopp et al.(32). [125I]IGF-II was used as a probe to maximize detection of all IGFBPs because some are relatively specific for IGF-II. The paper was washed, dried, and exposed to x-ray film for 4-24 h at-80°C in the presence of intensifying screens.

Statistical analysis. Differences between groups of membrane preparations were analyzed by the t test.

RESULTS

[125I]IGF-I and IGF-II binding to MVM and BM. High affinity binding sites for radiolabeled IGF-I were identified on both basal and apical, microvillous preparations of the syncytiotrophoblast cell membrane. As shown in Figure 1A, IGF-I specifically bound to the BM with high affinity. Binding sites on the BM exhibited a specificity characteristic of type I IGF receptors. Unlabeled IGF-I was the most effective competitor of binding, whereas IGF-II was 2-4-fold less effective. Insulin inhibited binding only at concentrations exceeding 26 nM. Scatchard analysis revealed a curvilinear plot consistent with the presence of two distinct binding components or, alternatively, one site exhibiting negative cooperativity. The high affinity binding component exhibited aKD of approximately 1.9 × 10-10 M, whereas the low affinity binding component exhibited a KD in the range of 2.9 × 10-8 M.

Figure 1
figure 1

Binding of [125I]IGF-I to BM and MVM. Trace amounts of [125I]IGF-I were incubated with 40 μg of membrane protein derived from (A) BM, or (B) MVM, in the presence or absence of varying concentrations of unlabeled IGF-I, IGF-II, and insulin. Each point in the competition studies represents the average of triplicate determinations and is representative of at least two separate experiments using different placentae. Scatchard plots are representative of at least three separate experiments performed using different placentae. Binding parameters are expressed as the mean ± SEM, where () = SEM.*Statistically different from the corresponding parameter(KD2) of the BM; p < 0.05.

Full size image

The binding of [125I]IGF-I to the MVM differed from its binding to the BM in two significant ways. First, competition studies revealed that insulin was partially effective at inhibiting binding at 0.2-3 nM (Fig. 1B), whereas much higher concentrations of insulin were required to observe any competition in the BM (>26 nM). Second, the low affinity binding site exhibited a KD that was 10-fold lower than the low affinity site observed in the BM (2.4 × 10-9 Mversus 2.9 × 10-8 M), consistent with the presence of a population of higher affinity binding sites in the MVM.

It is not known if the lower affinity binding component on the MVM represents a modified variant of the low affinity binding site observed in the BM or represents distinct receptors. Their appearance, however, correlates with the appearance of more receptors by cross-linking analysis. As seen in Figure 2A, [125I]IGF-I-labeled binding species migrated at approximately 130-140 kD, consistent with the α-subunit of the IGF-I receptor. More intense bands migrating at 270-280 kD and higher were also observed, most likely representing α/α dimers and higher molecular weight multimers of type I IGF receptor subunits, respectively. This interpretation is supported by the fact that IGF-I and IGF-II effectively competed for [125I]IGF-I labeling in all of the bands, whereas insulin was less effective. When equivalent amounts of membrane protein were subjected to cross-linking under identical conditions, substantially more receptors were apparent in the MVM compared with the BM (Fig. 2B), suggesting that more type I IGF receptors are expressed in MVM.

Figure 2
figure 2

Cross-linking of [125I]IGF-I to (A) MVM and (B) BM. Lane 1, total binding; lane 2, binding in the presence of 26 nM unlabeled IGF-I; lane 3, binding in the presence of 26 nM unlabeled IGF-II; lane 4, binding in the presence of 34 nM unlabeled insulin. Each lane represents binding to approximately 10 μg of membrane protein and is representative of two separate experiments using membranes derived from different placentas.

Full size image

[125I]IGF-II also exhibited significant binding to the BM. Competition studies suggested that 50-60% of the [125I]IGF-II binding sites are sensitive to IGF-I and high concentrations of insulin consistent with its binding to both the type I and type II IGF receptors (Fig. 3A). Insulin was able to inhibit IGF-II binding only at concentrations exceeding 30 nM. Similar to IGF-I binding, Scatchard analysis yielded a curvilinear plot suggesting the presence of two distinct binding sites with KD values of 2 × 10-10 M and 3.5 × 10-9 M (Fig. 3A).

Figure 3
figure 3

Binding of [125I]IGF-II to BM and MVM. Trace amounts of [125I]IGF-II were incubated with 40 μg of membrane protein derived from (A) BM, or (B) MVM, in the presence or absence of varying concentrations of unlabeled IGF-I, IGF-II, and insulin. Each point represents the average of triplicate determinations and is representative of at least two separate experiments using different placentas. Scatchard plots are representative of at least three separate experiments performed using different placentas. Binding parameters are expressed as the mean ± SEM of three separate experiments; () = SEM.

Full size image

[125I]IGF-II bound to receptors in the MVM with affinities similar to those observed in BM (Fig. 3B). Competition studies, however, revealed a significant difference in the specificity of these binding sites. Insulin, at concentrations of 0.2-3 nM, was able to inhibit a significant portion of the [125I]IGF-II binding, suggesting that a fraction of these binding sites had relatively high affinity for all three ligands.

Cross-linking analysis demonstrated the association of [125I]IGF-II with several binding species (Fig. 4A). First, a 240-260-kD species was labeled in MVM under reducing conditions. Binding was effectively inhibited by excess unlabeled IGF-II, but not by IGF-I or insulin, corresponding to type II IGF receptor specificity. Consistent with this interpretation is its migration at approximately 220 kD under nonreducing conditions, characteristic of the type II IGF receptor. In addition,[125I]IGF-II bound to a 130-140-kD species consistent with theα-subunit of the type I IGF or insulin receptor. Labeling of this binding species was inhibited most effectively by excess IGF-II, whereas IGF-I and insulin were partially effective at the concentrations tested. In contrast, there was much less apparent binding to both receptor types in the BM compared with MVM (Fig. 4B).

Figure 4
figure 4

Cross-linking of [125I]IGF-II to (A) MVM and (B) BM. Lane 1, total binding; lane 2, binding in the presence of 26 nM unlabeled IGF-I; lane 3, binding in the presence of 26 nM unlabeled IGF-II; lane 4, binding in the presence of 34 nM unlabeled insulin. Each lane represents binding to approximately 10 μg of membrane protein. Gels are representative of two separate experiments using different placentas.

Full size image

The structural correlates of the binding sites identified by Scatchard analysis have not been identified and may include the presence of one or more membrane-associated IGFBPs. The difference in type I and type II IGF receptors indicated by cross-linking analysis is in contrast to the similarities in binding sites indicated by Scatchard analysis, particularly for IGF-II, and suggests that nonreceptor binding sites, i.e. IGFBPs, contribute to the observed membrane binding. The presence of such IGFBPs could potentially compete with [125I]IGF-I and -II in receptor cross-linking studies, thereby masking the number of type I IGF receptors detected by this type of analysis. However, differences observed in the expression of type II receptors by cross-linking analysis cannot be explained by the presence of IGFBPs because that difference is seen in the presence of excess IGF-I which would mask any IGFBPs and type I IGF receptors that could compete for[125I]IGF-II. Significant binding of [125I]IGFs to lower molecular weight IGFBPs was not apparent when cross-linked samples were run under conditions permitting detection of lower molecular weight species (data not shown). Because cross-linking efficiency of IGFs to IGFBPs may be low, this may not reliably detect the presence of such competing binding sites.

Membrane-associated IGF binding proteins. We subjected both the BM and MVM to Western ligand blot analysis as a more sensitive method to determine whether they contained IGFBPs that could interfere with ligand binding studies. As shown in Figure 5, BM contained multiple species of binding proteins migrating at approximately 38, 29, and 24 kD. By contrast, MVM contained only a doublet migrating at approximately 42/38 kD. These binding species have not been identified. The bands migrating at 38 and 42 kD, however, suggest that they represent IGFBP-3. The consistent appearance of additional, specific IGFBPs in the BM compared with the MVM may explain, in part, the disparity between Scatchard and cross-linking analysis of receptor binding using native IGF-I and IGF-II as radioligands. IGF analogs with reduced affinity for IGFBPs would therefore be necessary to study, more specifically, the expression of IGF receptors in the placental membranes. To identify an appropriate ligand we examined the ability of various IGF analogs(0.3 nM) to compete with [125I]IGF-II binding to the specific binding proteins contained in the BM. We specifically tested des(1-3)IGF-I, des(1-6)IGF-II, R3IGF-I, QAYL-IGF-I, and native IGF-I/IGF-II to inhibit[125I]IGF-II binding to the various binding species contained in our membrane preparations. These IGF analogs have been shown to possess reduced affinity for specific IGFBPs(33–35). As shown in Figure 6, both native IGF-I and -II significantly inhibit binding to all of the binding species, whereas all of the IGF analogs competed less effectively and appeared to be comparable to each other. We therefore selected QAYL-IGF-I as an appropriate analog to further study receptor binding in our system.

Figure 5
figure 5

Presence of IGFBPs in MVM and BM preparations. Aliquots of MVM and BM, 50 μg of protein, derived from three different placentas were subjected to Western ligand blot analysis as described in“Methods” using [125I]IGF-II as a probe. Each lane represents membranes derived from different placentas. B, basal membranes; M, MVMs; NS, 3 μl of normal human serum. Data are representative of three separate experiments.

Full size image
Figure 6
figure 6

Inhibition of [125I]IGF-II binding to membrane-associated IGFBPs by IGF analogs. Aliquots of BM (50 μg) were subjected to ligand blot analysis. The abilities of 3 nM IGF-I, IGF-II, and various analogs to inhibit [125I]IGF-II binding to BM-associated IGFBPs. TB, total binding in the absence of competition; Q-IGF-I = QAYL-IGF-I. Representative of two separate experiments.

Full size image

[125I]QAYL-IGF-I binding to MVMand BM.[125I]QAYL-IGF-I bound to both high affinity (KD = 2-4× 10-10 M) and low affinity (KD = 10-8-10-9 M) sites in the BM and MVM preparations (Fig. 7). The high affinity site most likely represents the type I IGF receptor because this is comparable to the KD of the type I IGF receptor reported for placental receptors in previous studies(36, 37). Furthermore, we have demonstrated that QAYL-IGF-1 exhibits little affinity for all of the IGFBPs associated with the BM (Fig. 6). Scatchard analysis of these binding sites (Fig. 7) revealed an approximately 5-fold increase in high affinity binding sites in the MVM compared with the BM suggesting that more type I IGF receptors are expressed on the maternal membrane surface. Cross-linking analysis supported this interpretation. As shown in Figure 7, significantly more type I IGF receptors are labeled with [125I]QAYL-IGF-1 in the MVM compared with the BM, consistent with the Scatchard analysis. When viewed collectively, these data suggest that the preponderance of type I IGF receptors expressed by the syncytiotrophoblast exist on the maternal-facing MVM surface.

Figure 7
figure 7

Scatchard and cross-linking analysis of[125I]QAYL-IGF-I binding to placental membranes. Binding parameters represent the average ± SEM obtained from three separate Scatchard plots using different placentas. Cross-linking experiments were performed in the presence and absence of 26 nM unlabeled insulin and IGF-1, and are representative of two separate experiments using different placentae. Cross-linking to both BM and MVM preparations were performed simultaneously and exposed on the same gel. (A) Binding to BM preparations as described above. (B) Binding to MVM preparations. *Statistically significant difference in the number of high affinity binding sites in the BM and MVM preparations.

Full size image

Immunoblot analysis of IGF receptors in BM and MVM. To confirm the relative preponderance of IGF receptors suggested by the cross-linking and Scatchard analyses we subjected BM and MVM to immunoblot analysis using antibodies specific for the type I and type II IGF receptors. As seen in Figure 8, approximately 5-fold more type I IGF receptors are associated with the MVM compared with the BM, consistent with Scatchard analysis. However, approximately 50-100% more type II IGF receptors appear to be associated with BM compared with MVM in contrast to ligand cross-linking data. The discrepancy in type II receptors detected by bindingversus immunologic methods is probably due to the fact that the membrane preparations contain distinct pools of type II receptors. We suspected that receptors associated with BM may be occupied by lysosomal enzymes, containing phosphomannosyl groups, thereby masking IGF-II binding sites. To test this hypothesis we performed [125I]IGF-II cross-linking experiments in the presence or absence of 5 mM M6P. M6P has been shown to increase the binding of [125I]IGF-II to the type II receptor presumably by stripping endogenously bound lysosomal enzymes from the receptor and releasing steric constraints to binding(38, 39). As seen in Figure 9, incubating BM with 5 mM M6P resulted in significantly more [125I]IGF-II cross-linked to the type II receptor. A similar enhancement in binding and cross-linking was not apparent in MVM. Similar effects were observed in membranes derived from two separate placentae.

Figure 8
figure 8

Immunoblot analysis of type I and type II IGF receptors in placental membranes. A 100-μg sample of membrane protein was subjected to SDS-PAGE under reducing conditions (type I receptor) and nonreducing conditions (type II receptor) before incubation with the relevant primary antibody. Receptor-labeled bands were visualized by chemiluminescence detection system as described in “Methods.” Each lane represents membranes derived from different placentas. B = basal membranes,M = MVMs.

Full size image
Figure 9
figure 9

Effect of M6P on [125I]IGF-II binding to placental membranes. Aliquots of placental membranes (40 μg) were incubated with or without 5 mM M6P for 1 h at 22°C. Cross-linking to radiolabeled IGF-II was then carried out as previously described. Total [125I]IGF-II binding (lanes 1, 4, 7, and 10); binding in the presence of 13 nM IGF-I (lanes 2, 5, 8, and 11); binding in the presence of 13 nM IGF-II (lanes 3, 6, 9, and 12). CON = binding in the absence of M6P; + M6P = binding in the presence of 5 mM M6P.

Full size image

DISCUSSION

The syncytiotrophoblast membrane is not a homogeneous structure. Like other epithelial cell layers, it possesses an apical, microvillous surface in direct contact with maternal blood, and a basal surface in close proximity to the fetal vasculature. Integral membrane proteins exhibit specific topographical expression patterns providing portions of the syncytiotrophoblast cell membrane with functional specificity. This provides the structural basis by which the placenta effectively integrates signals from both maternal and fetal compartments to function in the best interests of the developing conceptus. Insulin and the IGFs are major hormonal regulators of fetal growth and likely important regulators of trophoblast growth, metabolism, and transport function. The responsiveness of the syncytiotrophoblast to insulin-related peptides is determined, in part, by the expression of receptors specific for these peptides. The specific portion of the syncytiotrophoblast membrane that expresses these receptors determines from which compartment, fetal or maternal, these regulatory peptides are likely to be derived if they are to influence trophoblast function.

These studies demonstrate that, like insulin, the expression of high affinity type I IGF receptors occurs predominantly on the MVM surface. In addition, differences in binding specificity were observed between the two membrane preparations. Specifically, 15-20% of the IGF-I binding to the MVM was inhibited by low concentrations of insulin whereas low insulin concentrations had no effect in BM. The insulin sensitivity of IGF-II binding to the MVM was even more pronounced. Cross-linking analysis suggested that the insulin sensitive IGF-II binding sites were structurally similar to the type I IGF and insulin receptors. These data provide evidence that type I IGF receptors, in addition to insulin receptors, are expressed primarily on the MVM surface of the syncytiotrophoblast. Furthermore, they suggest that the MVM, but not the BM, contain α2β2 receptors that possess relatively high affinity for both IGFs and insulin.

The presence of IGF/insulin receptors in human placenta that exhibit increased affinity for heterologous ligands have been described in several independent studies. First, Hintz et al.(17) described [125I]IGF-II binding sites in the placenta that possess high affinity for insulin and IGF-II, similar to the IGF binding sites on IM-9 lymphocytes. Structural analysis of these binding sites suggested anα2β2 quaternary structure similar to that of the type I IGF and insulin receptors. Second, Jonas and coworkers described“atypical” placental insulin receptors that copurified with“classical” insulin receptors on an insulin affinity column(18–20). These “atypical” receptors exhibited a higher affinity for IGF-I and rat IGF-II than the classical insulin receptor but were immunoprecipitated by a specific anti-insulin receptor antibody. Again, these binding sites possessed theα2β2 structure characteristic of type I IGF and insulin receptors. It is intriguing to speculate that the atypical binding sites observed in MVM represent those sites previously described by others. The structural basis for these sites is not known but may represent atypical insulin receptors, atypical IGF receptors, and/or IGF/Ins receptor hybrids(40–42).

The expression pattern of type II IGF receptors yielded unexpected results. Although total receptor protein was expressed in greater amounts in the BM, the preponderance of IGF-II binding activity was observed on the maternal facing MVM. We speculate that BM and MVM contain functionally distinct pools of receptor. This is supported by the observation that the IGF-II binding activity of receptors in BM is enhanced in the presence of M6P. M6P competes with lysosomal enzymes for binding to the type II receptor, suggesting that receptors associated with BM are already occupied by lysosomal enzymes(38, 39). M6P had no effect on the binding to MVM receptors. In addition, receptors associated with BM exhibited slightly faster mobility on SDS-PAGE compared with receptors derived from MVM. The basis for this difference has not been determined. The BM-associated receptor migrates at a rate identical to the truncated, soluble type II receptor (data not shown), but there is no evidence to suggest that the soluble receptor is tightly associated with cell membranes. It more likely reflects some degree of proteolytic degradation or perhaps a more compact structural conformation as a consequence of its association with endogenous phosphomannosyl groups. Only a small fraction of total cellular type II IGF receptors exist on the cell surface. The rest cycle intracellularly to target lysosomal enzymes for degradation. We, therefore, cannot be certain to what extent intracellular pools of receptor, closely associated with the BM, may be retained during its purification. Therefore, the extent to which the binding domains of BM-associated receptors exist on the cell surface, in situ, is not known. In any case, these data suggest that the type II IGF receptor may play functionally distinct roles on different sides of the trophoblast. Unoccupied type II IGF receptors are associated primarily with the maternal-facing MVM and may serve to regulate IGF-II levels at the fetal-maternal interface. By contrast, type II IGF receptors associated with fetal-facing BM appear to be occupied by endogenous phosphomannosyl compounds. This may reflect a relatively more important role in the trafficking of lysosomal enzymes involved in tissue remodeling within the stromal compartment of the villus. This is speculative, however, and requires further investigation.

Finally, the association of specific IGF binding proteins with the BM deserves comment. A 42/38-kD doublet, most likely representing IGFBP-3, is associated with both the BM and MVM. IGFBP-3 is produced by a variety of fetal and maternal cells including placental stromal cells and trophoblasts. In addition, two other binding species, 29 and 24 kD, were consistently associated with BM but not with the MVM. It is not known if this association is physiologically relevant or simply an artifact of the purification process. All six IGFBPs are known to be present in the maternal-fetal placental unit. The fact that two binding species are consistently associated with only the BM suggests that it is not artifact but reflects their association or proximity,in situ. If true, this could potentially serve as a mechanism for targeting IGFs to IGF receptors on the basal surface of the trophoblast or for providing a reservoir of IGFs that regulate stromal cell function via paracrine mechanisms. Alternatively, they may bind to specific sites on the trophoblast and regulate cell function in an IGF-independent manner. The identification of these IGFBPs, characterization of their association with the BM, and verification of their location, in situ, will be necessary to begin to answer these interesting questions.

In summary, type I IGF and unoccupied type II IGF receptors are expressed primarily on the maternal-facing MVM surface of the syncytiotrophoblast. This is similar to the expression pattern of insulin receptors, and suggests that insulin-related peptides derived from the maternal compartment play a major role in regulating syncytiotrophoblast cell function at term. Insulin and IGFs have been shown to exert diverse effects on trophoblasts, in culture, and exhibit some overlap in their regulatory activity. It is not known if the MVM receptors exhibiting relatively high affinity for insulin, IGF-I, and IGF-II regulate cellular processes distinct from those regulated by coexisting classical IGF and insulin receptors. Redundancy of available ligands may be a mechanism by which insulin regulated processes, important to the fetus, are buffered from acute fluctuations in circulating insulin levels. Their exclusive expression on the maternal surface of the syncytiotrophoblast is indicative of functional significance and suggests that they, along with classical type I IGF and insulin receptors, play an important role in providing the trophoblast with the ability to effectively balance the metabolic status of the mother with the nutritional requirements of the growing fetus.