The insulin/IGF system plays a critical role in embryo growth and development. We have investigated the expression of insulin receptor (IR) and IGF-I receptor (IGF-IR) and the activation of their downstream pathways in rabbit 6-d-old blastocysts. IR was expressed in embryoblast (Em, inner cell mass) and trophoblast (Tr) cells, whereas IGF-IR was localized mainly in Em. Isoform A (IR-A) represents the main insulin isoform in blastocysts and was found in Em and Tr cells. IR-B was detectable only in Tr. IR/IGF-IR signaling pathways were analyzed after stimulation with insulin (17 nm) or IGF-I (1.3 nm) in cultured blastocysts. Insulin stimulated Erk1/2 in Em and Tr and Akt in Tr but not in Em. IGF-I activated both kinases exclusively in Em. The target genes c-fos (for MAPK kinase-1/Erk signaling) and phosphoenolpyruvate carboxykinase (PEPCK, for PI3K/Akt signaling) were also specifically regulated. Insulin down-regulated PEPCK RNA amounts in Tr by activation of the phosphatidylinositol 3-kinase/Akt pathway. Expression of c-fos by insulin and IGF-I was different with respect to time and fortitude of expression, mirroring again the specific IR and IGF-IR expression patterns in Em and Tr. Taken together, we show that IGF-I acts primarily mitogenic, an effect that is cell lineage-specifically restricted to the Em. By contrast, insulin is the growth factor of the Tr stimulating mitogenesis and down-regulating metabolic responses. As soon as blastocyst differentiation in Em and Tr has been accomplished, insulin and IGF-I signaling is different in both cell lineages, implying a different developmental impact of both growth factors.
THE INSULIN RECEPTOR (IR) and the IGF-I receptor (IGF-IR) are highly homologous tyrosine kinase receptors. Many steps in their signaling pathways are shared in common. The ligands insulin, IGF-I, and IG-II can bind to either receptor, although with different affinities (1, 2). Both receptors are heterotetrameric proteins composed of two α-subunits that confer the ability to bind insulin or IGF. The two β-subunits contain the membrane-spanning and the tyrosine kinase domains. Binding of insulin and IGF-I to IR and IGF-IR triggers receptor tyrosine autophosphorylation and creates docking sites for downstream adaptor proteins.
The mature IR exists in two isoforms, designated A and B, which result from alternative splicing of the primary transcript (3). IR-A is a short isoform, lacking exon 11, a small exon encoding for 12 amino acid residues at the carboxy terminus of the IR α-subunit (4). Although expressed ubiquitously, IR-A is the only isoform in lymphocytes, brain, and spleen. The B isoform contains exon 11 and is expressed predominantly in liver, muscle, adipocytes, and kidney (3). The relative abundance of the two IR isoforms is regulated by cell differentiation, stage of development, and tissue-specific factors (1, 5, 6). IR-A has a slightly higher binding affinity, and IR-B has a more efficient signaling activity as evaluated by its tyrosine kinase activity and phosphorylation of IR substrate 1 (IRS-1) (7). It has been demonstrated that IR-A is the predominant isoform in fetal tissues and binds IGF-II with high affinity (1, 2, 8). The expression of the two IR isoforms in early embryo development and their potential biological roles are unknown.
Insulin and IGF-I and -II mediate mitogenic, antiapoptotic, and anabolic effects in mammalian preimplantation embryos (for reviews see Refs. 9 and 10). Mutations in the IR and IGF-IR lead to fetal growth retardation and metabolic disorders (11). In mice, IGF-IR mediates IGF-I and IGF-II action on prenatal growth and IGF-II actions on postnatal growth. The IR mediates prenatal growth in response to IGF-II and postnatal metabolism in response to insulin. In rabbit embryos, addition of insulin to the culture media results in increased proliferation and a decrease in apoptosis at the blastocyst stage (12).
Early mammalian embryos express IR, IGF-IR, and IGF-IIR and the ligands IGF-I and IGF-II (for reviews see Refs. 9 and 10). In rat (13), sheep, and cattle (14,15), IGF-IR transcripts were detected at all embryonic stages from the fertilized oocyte to the blastocyst stage. In mice, IGF-IR is expressed from the eight-cell stage onwards (16,17). IR transcription in human (18) and mice (16) was first detected at the morula stage, in rabbits at the early blastocyst stage (19).
Although IR/IGF-IR RNA expression has been described in many mammalian species, studies on expression and localization of the receptor proteins in preimplantation embryos are limited. In mice, the IGF-IR is localized in both the embryoblast [Em; inner cell mass (ICM)] and the trophoblast (Tr) cells with stage-dependent changes in the subcellular localization (17). In eight-cell embryos, IGF-IR was detected in the outer membrane and in cortical cytoplasmic vesicles. This outer or apical localization persisted throughout compaction. In blastocysts, the receptor was mostly confined to cytoplasmic vesicles in trophectoderm cells. There was only a weak vesicular staining in the Em and some staining of plasma membranes adjacent to trophectoderm cells (17).
Insulin is not expressed by preimplantation embryos in any species studied to date. IGF-I expression was detected in bovine and ovine embryos produced in vitro (14, 20, 21) but was not observed in embryos from water buffalo (22), mice (16), rats (13), and humans (18). Controversial data are published for mice and cattle embryos (15, 23). Insulin and IGF-I are found in the oviduct and uterine lumen during the pre- and periimplantation period (24, 25). Thus, the embryo either expresses these growth factors itself or has access through maternal genital tract secretions, implying functional autocrine, paracrine, and/or endocrine insulin/IGF circuits operating during early development.
Insulin and IGF-I induce different biological effects, although a certain degree of overlap in their actions is well recognized (26). Insulin is important for metabolic activities, whereas IGF-I mediates cell growth and survival in target tissues. The second messenger system involved in IR signaling diverges into two separate pathways (reviewed in Ref. 27). IRS-1 and IRS-2, as well as Shc, serve as docking molecules that bind to and activate cellular kinases. IRS-1/2 phosphorylation leads to downstream activation of the phosphatidylinositol 3-kinase (PI3K) pathway that regulates most of the metabolic effects of insulin. Shc phosphorylation activates the Ras/MAPK pathway, which is associated with cell growth and protein synthesis (reviewed in Ref. 28). Signaling through the Ras/MAPK pathway primarily stimulates mitogenic and catabolic processes and includes the phosphorylation of the cytosolic effectors ERK1 and ERK2 (p42 and p44 MAPK, respectively). Major targets of MAPK signaling are the activator protein-1 (AP-1) transcription factors. They are dimeric complexes of the members of Jun, Fos, and activating transcription factor-2-related protein families. The activities of the AP-1 components depend on the dimer-partner constellation. Phosphorylation by MAPKs can regulate the transcriptional activity, stability, and/or mRNA levels of the AP-1 components, for instance of the Fos family member c-Fos. This transcription factor is believed to control the expression of many genes involved in growth and differentiation (29, 30).
The aim of this study is to more closely examine IR and IGF-IR expression at one of the most sensitive ontogenetic stages, in blastocysts, and to analyze postreceptor signaling in response to insulin and IGF-I.
Materials and Methods
Embryo recovery and in vitro culture
Embryos were collected from sexually mature rabbits stimulated by 150 U pregnant mare serum gonadotropin (Intervet, Tönisvorst, Germany) and 75 IU human chorionic gonadotropin (Schering AG, Berlin, Germany). Mating, embryo recovery (31), and embryo culture (32) were performed as described before. On d 3, 4, and 6 postcoitum, embryos were flushed from oviducts or uteri, washed three times with PBS, pooled, and randomly divided among the experimental groups.
To study the effects of insulin and IGF-I, d-6 blastocysts were cultured in groups of four to 10 embryos in 4000 μl basal synthetic medium II at 37 C in a saturated atmosphere of 5% O2, 5% CO2, and 90% N2 (33) in a water-jacketed incubator (BB 6060; Heraeus, Hanau, Germany). After preculture for 2 h in serum- and insulin-free basal synthetic medium II, 17 nm insulin (Invitrogen, Karlsruhe, Germany) or 1.3 nm IGF-I (Sigma-Aldrich, Taufkirchen, Germany) was added to the culture medium for fixed time intervals. The protein kinase inhibitors LY94002 and PD98054 were added to the culture media 30 min before growth factor treatments in a final concentration of 25 μm.
Controls were cultured without insulin or IGF-I but otherwise handled identically as the treated embryos. After culture, embryos were washed three times in ice-cold PBS. To investigate spatial expression of IR isoforms and to discriminate the growth factor effects in Tr and Em, blastocyst coverings were mechanically removed and the embryonic disks were microdissected from the trophoblast under a stereomicroscope. Isolated Em and Tr of single blastocysts were stored separately at −80 C for gene expression analysis. For Western blotting analyses, 10 Em and 10 Tr were pooled in each group. Probes were homogenized in 50 μl (Em) or 100 μl (Tr) cold RIPA buffer (PBS, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS) with protease and phosphatase inhibitor cocktail (Sigma Chemical Co., St. Louis, MO) and stored at −80 C.
It cannot be excluded that a few adjacent Tr cells contaminated the Em specimens. Blastocysts for immunohistochemistry were fixed overnight in 4% (vol/vol) paraformaldehyde/PBS.
RNA extraction and RT reaction
Poly(A)+ RNA was isolated from single or pooled embryos using a Dynabeads mRNA DIRECT Kit (Dynal, Oslo, Norway) according to the manufacturer’s instructions, with minor modifications (as described in Ref. 19). The final volume of the cDNA reaction was adjusted to 50 or 100 μl with 30 and 80 μl water for separated Tr and Em or single blastocysts, respectively.
Preparation of total RNA from tissues was performed by using TRIzol reagent (Invitrogen) according to a previously described protocol (34). RNA was treated with DNase for 1 h. The amount of total RNA was determined spectrophotometrically at 260 nm. cDNA synthesis on total RNA was performed essentially as previously described (34).
Cloning of rabbit IR-B sequence
Rabbit IR-B sequence was determined by PCR amplification with specific exon-spanning primers rabIRex10fw 5′-TCCTGAAGGAGGTGGAGGAG-3′ and rabIRex12rev 5′-GAGAATCCTGGGACTGTGGG-3′ derived from rabbit sequences (rabIR accession no. AY339877) (18) (see Fig. 2A) on reverse transcribed rabbit liver mRNA. The localizations of the specific rabbit PCR primers (rabIR ex10fw and rabIR ex12rev) used in this study are underlined in Fig. 2A. The amplified PCR products of 150 and 186 bp for IR-A (Ex11−) or IR-B (Ex11+), respectively, were purified by separation in a preparative 1.8% agarose gel and extracted by a gel extraction kit (QIAGEN, Hilden, Germany) following the manufacturer’s protocol. Cloning and sequencing of the PCR fragment was performed as described previously (19). The sequenced cDNA was screened for homology in the GenBank EMBL using the BLASTN search modus and for amino acids the BLASTP search modus (35). The partial sequences were published in the EMBL Data Bank under accession no. EF625883.
Rabbit IR-B partial cDNA sequence and the predicted protein sequence of alternative spliced exon 11. A and B, Comparison of the rabbit IR-A and IR-B partial cDNA sequences and the alignment of the predicted protein sequence of rabbit IR ex11 (rabbit IR-B) with human IR isoform B (human IR-B) (accession no. M32832), respectively. Alignments are displayed by the following symbols denoting the degree of conservation observed between sequences: |, nucleotides or residues are identical; −, sequences have a gap with no nucleotide equivalent; +, conserved substitutions have been observed and no symbol indicates that nonconserved substitutions have occurred. Arrows indicate the positions of the primers (rabIR ex10fw and rabIR ex12rev) used for RT-PCR.
PCR detection of IR isoforms
PCR amplification was carried out with 5 μl cDNA from rabbit tissues and 5 μl cDNA from blastocysts in a 50-μl volume containing 200 μm each dNTP and 2.5 U Taq polymerase, employing the primer combination rabIRex10fw, rabIRex12rev. Resulting PCR products of 150 and 186 bp for IR-A (Ex11−) or IR-B (Ex11+) were separated by electrophoresis on 2.0% agarose gel and stained with ethidium bromide. The intensity of the IR-A and IR-B product bands was densitometrically analyzed employing the software BIO-Profile 1D (LTF Labortechnik, Wasserburg, Germany). The relative amount of IR-A and IR-B mRNA of single blastocysts was calculated as the ratio of the PCR band intensity (OD) of each isoform (ODIR-A; ODIR-B) related to the whole IR amount (ODIR-A + ODIR-B = 100%). All PCRs were performed three times. The data are expressed as mean ± sem.
Cloning and sequencing of partial cDNA for rabbit phosphoenolpyruvate carboxykinase (PEPCK) and IGF-IR
Partial cDNAs of the rabbit PEPCK and IGF-IR gene were cloned and sequenced from rabbit kidney and liver tissues using PCR primer deduced from human and hamster gene sequences for PEPCK (accession no. NM_004563) and IGF-IR (accession no. NM_000875.2), respectively. Cloning and sequencing of the PCR fragment was performed as described previously (19). Sequences were compared for homology in the GenBank EMBL using BlastN (nucleotides) or BlastP (protein) search modus.
The rabbit PEPCK PCR fragment was homologous to the human nucleotide sequence (accession no. BC001454) for 95% and protein (accession no. AAH01454.1) for 97%. Homology alignment of rabbit IGF-IR resulted in 92% (accession no. NM_000875.2) and 96% (accession no. NP_000866.1) similarity with human cDNA. The partial sequences were published in the EMBL Data Bank under accession nos. EF616471 and EF616472 for rabbit PEPCK and IGF-IR, respectively.
Real-time PCR for IGF-IR, c-fos, PEPCK, and GAPDH
Samples were analyzed by real time RT-PCR using an Opticon 2 System (MJ Research, Waltham, MA) as described (36). Reactions were performed using SYBR green Master Mix (Applied Biosystems, Darmstadt, Germany) as a double-stranded DNA-specific fluorescent dye with the appropriate primer sets: rabGAPDH, forward 5′-GCCGCTTCTTCTCGTGCAG-3′ and reverse 5′-ATGGATCATTGATGGCGACAACAT-3′ (accession no. L23961); and IGF-IR, forward 5′-CCCAAGCTCACGGTCATCACTG-3′ and reverse 5′-ATGGGCTTCTCCTCCAAGGTCC-3′; rabFos, forward 5′-CAA CGA CCC GGA GCC TAA GCC-3′ and reverse 5′-TGCTGGGAACAGGAAGTCATCGAA-3′ (accession no. AB020214, EMBL Databank); and rabPEPCK2, forward 5′-TGCGGCCTCCAAAGATGATG-3′ and reverse 5′-CCCTGGAAACCTGGTGACAAGG-3′ (accession no. EF616471) (for rabbit GAPDH, IGF-IR, c-fos, and PEPCK, respectively). PCR was initiated for 2 min at 50 C followed by 10 min at 95 C. The program continued with 40 cycles of 15 sec at 95 C and 60 sec at 60 C. Each assay included duplicates of each cDNA sample and a no-template control for GAPDH and IGF-IR. The parameter cycle threshold (CT) is defined as the cycle number at which fluorescence intensity exceeds a fixed threshold. Relative mRNA expression for IGF-IR, PEPCK, and GAPDH were calculated using the ΔΔCT method (37). The expression of GAPDH RNA was used to normalize samples for the amount of cDNA used per reaction. To confirm the amplification, the resulting real-time PCR products were analyzed by dissociation curves, visualized in an agarose gel (GAPDH, 144 bp; IGF-IR, 347 bp; c-fos, 96 bp; and PEPCK, 143 bp), and sequenced.
In each real-time PCR run for c-fos and GAPDH, a calibration curve was included that was generated from serial dilutions (106,105, 104,103, 102, and 10 copies) of specific c-fos and GAPDH DNA probes generated from cDNA plasmid clones. Analysis of the individual data therefore yielded values relative to these standards. Data are presented as the ratio of the number of c-fos copies per 103 molecules of GAPDH. For each group, at least eight blastocysts were analyzed. Control blastocysts were cultured without growth factors with same time intervals and identical handling.
In the time curves, c-fos copies per 103 molecules GAPDH from control blastocyst cDNA was set to 1.0. Data are therefore expressed as values relative to control blastocysts. Levels of significance between groups were calculated using paired t test. P < 0.05 was considered significant.
Protein preparation and immunoblotting
Protein isolation and Western blot analyses were performed as described (19). Gels were loaded with 20 μg protein from whole blastocysts and 10 μg separated Tr and Em. Protein samples of treated and untreated rabbit kidney cells (RK13) were added as positive controls to each blot. Cells were treated with 1.3 nm IGF-I for 10 min with subsequent protein isolation. The ERK and Akt phosphorylation assays were accomplished with the monoclonal phospho-ERK1/2 antibody (1:2000, phospho-p42/p44 MAPK Thr202/Tyr204) and phospho-Akt antibody (1:5000, phospho-Akt Ser473; Cell Signaling Technologies, Inc., Frankfurt, Germany) as described (38). After blotting, the nylon membrane was stained with Ponceau and cut into two parts at the position of approximately 55 kDa. Apparent molecular weights were determined by comparison with standard molecular weight markers (high range marker; Promega Corp., Mannheim, Germany). The upper membrane part (protein range > 55 kDA) was used for phospho-Akt detection, the lower (<55 kDa) for phospho-Erk and β-actin. Afterward, the immunoreactive signals were visualized by enhanced chemiluminescence detection (Millipore, Schwalbach, Germany).
For calculations of relative protein amounts in Em and Tr, the signals were related to β-actin in the same blot. The amounts of phospho-ERK and phospho-Akt proteins were evaluated by stripping the membranes and reblotting with β-actin antibody (mouse monoclonal anti-β-actin 1:40,000; Sigma-Aldrich). Protein amounts and phosphorylation was calculated as the ratio of band intensities (phospho-Akt or phospho-ERK1/2 vs. actin) in the same blot to correct for differences in protein loading. The relative stimulation of ERK1/2 and Akt phosphorylation by insulin or IGF-I was calculated as the ratio of phosphorylation signals of growth factor-treated and untreated controls at each time point. Western blot analyses were performed at least twice in three independent experiments.
Immunohistochemical localization of IR and IGF-IR
The IR and IGF-IR were localized on whole blastocysts. Paraformaldehyde-fixed blastocysts were rehydrated through a series of graded alcohols. The neozona was removed mechanically before serum blocking. Nonspecific antibody binding was blocked with 10% normal goat serum in PBS at room temperature for 1 h. The specimens were incubated with the primary antiserum overnight at 4 C. Both antibodies, the monoclonal mouse anti-IR α-subunit (Chemicon, Hampshire, UK) and the mouse monoclonal anti-IGF-IR (Santa Cruz Biotechnology, Heidelberg, Germany), were diluted 1:100 (anti-IR) and 1:50 (anti-IGF-IR) in 3% BSA/PBS. Whole blastocysts were washed with PBS/Tween 20 and incubated with the secondary antiserum (fluorescein isothiocyanate-conjugated AffinPure donkey antimouse-IgG, 1:600; Dianova, Hamburg, Germany). The nuclei were counterstained with 7-amino-actinomycin (Dako, Hamburg, Germany). Whole blastocysts were examined by confocal microscopy with Leica LMT (Leica, Heidelberg, Germany). The specificity of immunostaining was demonstrated by the absence of signals in sections incubated with control mouse IgG (Dako) or in sections processed after omission of the primary antibody. Only reactions with negative controls were included in the study.
Statistics
Statistical analysis of the relative stimulation of ERK1/2 phosphorylation was performed with the paired t test between the growth factor-treated and untreated specimens at the various time points studied (SigmaBlot). The data are expressed as mean ± sem.
Results
IR
IR protein is expressed in fully expanded rabbit blastocysts at d 6 postcoitum in both Em and Tr. It is mainly localized in the plasma membrane and to some extent in the cytoplasm of Tr and Em cells (Fig. 1).
Localization of IR protein in d-6 rabbit blastocysts. Whole-mount immunohistochemistry of 6-d-old blastocysts show positive staining for IR in Em (*; a) and Tr cells. The IR protein was visualized by peroxidase-diaminobenzidine reaction (brown in a). The margin of the embryonic disc (*) is marked with arrows. The subcellular localization was investigated by whole-mount confocal microscopy (b–d) with a fluorescence detection for IR (green) and nuclear counterstaining with 7-aminoactinomycin (in red; b–d). Em and Tr cells show an intense staining for IR in the plasma membrane (b and c) and sporadically also in the cytoplasm (d).
IR isoform A and B expression in Em and Tr
Specific PCR amplification of rabbit IR-B.
To determine the IR-A and -B expression pattern in embryos, we developed a sensitive PCR approach discriminating the IR transcripts by the absence or presence of exon 11. Specific oligonucleotides (rabIRex10fw and rabIRex12rev, Fig. 2A) were deduced from rabbit IR sequence (accession no. AY339877) (19) amplifying an exon 11-spanning product. The specific amplification of the two isoform transcripts was indicated by different PCR products of 150 and 186 bp for IR-A (Ex11−) and IR-B (Ex11+), respectively. Simultaneous detection of two DNA fragments indicated the presence of both isoforms on the mRNA level in the same tissue. The identity of the PCR products was confirmed by sequencing and sequence alignment. The 186-bp fragment was identified as partial cDNA of the rabbit IR. Sequence alignments were performed. At nucleotide position 2147 of the rabbit IR sequence (accession no. AY339877) (19), an additional 36-bp insertion was found (Fig. 2A). The deduced protein sequence of these 36 nucleotides showed a 75% identity with human IR-B sequence (Fig. 2B).
IR-A and IR-B expression in rabbit tissues and blastocysts.
IR isoforms were detected in 6-d-old blastocysts and in rabbit tissues. In blastocysts and in most analyzed tissues, both isoforms were found (Fig. 3). In the liver, isoform B was highly expressed, confirming the predominant expression of this isoform in the human and mouse liver. In heart muscle, only isoform A could be amplified.
IR-A and -B mRNA expression in rabbit tissues and blastocysts. The expression of IR-A and IR-B was detected by RT-PCR using exon 11-spanning specific rabbit primers (A). Total RNA from rabbit liver (L), heart muscle (H), kidney (K), uterus (Ut), skeletal muscle (SM), and lung (Lu) and mRNA from separated Em and Tr of 6-d-old rabbit blastocysts were reverse transcribed and amplified with rabIRex10fw and rabIRex12rev. β-Actin PCR was performed as control of embryonic cDNA quantity. Resulting PCR fragments (IR-A 150 bp, IR-B 186 bp, and β-actin 450 bp) were resolved in 2.0% agarose gel. PCR control was performed without cDNA template (Ø). M, DNA ladder. B, Relative amounts of IR-A and -B were quantified by RT-PCR using scanning densitometry of IR-A and IR-B bands amplified with exon spanning primers in five in vivo d-6 blastocysts (1-5 ). Representative agarose gels are shown (IR-A, 150 bp; IR-B, 186 bp). PCR control was performed without cDNA template (Ø). For each blastocyst, band intensity was measured densitometrically and related to the total amount of both products (100%). The result of five individual measurements (mean ± sem; n = 5) is shown in the diagram. Expression of IR-A was significantly higher than IR-B expression (P < 0.05).
The intensity of the IR-A and IR-B product bands was quantified in individual blastocysts. The main isoform is IR-A with 76.2 ± 5.8% of total IR transcript amount. IR-B transcripts varied between 18 and 30% (average 23.8 ± 5.8%, Fig. 3B), i.e. amounting to maximally one third of the IR in blastocysts. In blastocysts separated into Em and Tr, a cell lineage-specific distribution pattern of the IR isoforms was found. Whereas IR-A was present in both Em and Tr cells, the IR-B isoform was detectable only in the trophoblast (Fig. 3A)
Expression of the IGF-IR in Em and Tr of rabbit blastocysts
Cloning and sequencing of rabbit partial IGF-IR cDNA.
A new partial IGF-IR cDNA (477 bp, accession no. EF616472) was amplified from rabbit liver cDNA using hamster primers. The sequence was identical with human IGF-IR to 91 and 96% at the mRNA and the protein level, respectively, using alignment BLASTN and BLASTX modus (35).
IGF-IR expression in rabbit embryos and tissues.
IGF-IR RNA was present in rabbit morulae and blastocysts as well as in all tested tissues (Fig. 4A). There were significant differences in IGF-IR RNA amounts between Em and Tr, as revealed by real-time PCR in single d-6 blastocysts (Fig. 4B). In the Em, there was 2-fold more IGF-IR RNA expressed than in the Tr.
IGF-IR expression in rabbit tissues and embryos. A, The expression of IGF-IR was detected by RT-PCR using specific rabbit primers. A representative agarose gel is shown with PCR products for IGF-IR (347 bp) and GAPDH (144 bp) in rabbit duodenum (D), kidney (K), ovary (Ov), liver (L), lung (Lu), heart (H), and preimplantation embryos of different age (Mo, morulae; d4, early blastocysts at d 4; d6, expanded blastocysts at d 6). Ø, Control without cDNA. B, The relative amount of IGF-IR RNA was measured in Em and Tr of 15 blastocysts by real-time PCR. For normalization of blastocyst sizes, real-time PCR was performed for the housekeeping gene GAPDH on the same probes. Relative transcript amounts were calculated by the ΔΔCT method. In A, the real-time PCR products for IGF-IR and GAPDH were controlled by gel electrophoresis. Representative gels of five individual blastocysts and relative IGF-IR amounts (mean ± sem; n = 15) in Em and Tr are shown. C, The localization of IGF-IR protein in Em and Tr of d-6 rabbit blastocysts was investigated by whole-mount confocal microscopy (a–c) with fluorescence detection for IGF-IR (green) and nuclear counterstaining with 7-aminoactinomycin (red) in Tr (a), margin of the embryonic disc (Em) and trophoblast (Tr) (b), and Em (c). The Tr shows a weak fluorescence signal, whereas intense positive staining for IGF-IR was detected in the plasma membrane and cytoplasm of Em cells. Scale bar, 20 μm.
The IGF-IR protein was detectable in the plasma membrane of Em cells with an intense immunofluorescence staining, whereas Tr cells exhibited only a weak fluorescence signal (Fig. 4C).
IGF-I-stimulated ERK and Akt phosphorylation.
ERK1 and -2 are activated by various cell surface receptors coupled to cytoplasmic tyrosine kinase including IR and IGF-IR. We have demonstrated before that insulin activates ERK phosphorylation in rabbit blastocysts. Addition of insulin to cultured d-6 blastocysts caused a rapid increase in the phosphorylation level of both ERK-1 and -2, reaching a maximum at 10 min after insulin exposure and returning to control levels by 60 min (38).
Here we show that also 1.3 nm IGF-I stimulates ERK phosphorylation in blastocysts after 10 min. This increase was not observed when the embryos had been cultured with MAPK kinase inhibitor PD98059 for 30 min before IGF-I stimulation (Fig. 5A). The amount of phospho-Akt protein increased more than 10-fold after 10 min of IGF-I stimulation, whereas the increase was approximately 5-fold after 20 min. An increase in phospho-Akt was not measurable after addition of the PI3K inhibitor LY294002 before IGF-I stimulation (Fig. 5B).
Akt and ERK1/2 phosphorylation in d-6 rabbit blastocysts. The blastocysts were cultured, in groups of 10, with or without 1.3 nm IGF-I for 10 or 20 min. The phosphorylation of ERK1/2 and Akt was analyzed by Western blotting with anti-phospho-Akt and anti-phospho-ERK1/2 antibodies, respectively, and set in relation to anti-actin antibody staining. Representative Western blots for ERK1/2 (A) and Akt (B) are shown. The reactivity of human IGF-I on the rabbit IGF system was proven in a rabbit kidney cell line. Cell lysates of treated and untreated rabbit kidney cells were added as positive controls to the blot (B). The specific inhibitors PD98058 and LY29002 for the MAPK/ERK pathway and PI3K/Akt pathway, respectively, were added to the culture medium 30 min before stimulation with IGF-I. The stimulation experiments were performed three times (n = 3) with 10 blastocysts per replicate.
Insulin- and IGF-I-stimulated ERK and Akt phosphorylation in Em and Tr.
To analyze the cell lineage-specific IGF signaling, ERK and Akt phosphorylation was determined in microsurgically isolated Em and Tr. Insulin increased the phospho-ERK amounts strongly in the Em (∼10-fold) and to a lower extent in the Tr (∼4-fold). Also, IGF-I enhanced the phospho-ERK amounts in the Em (approximately 5-fold) but not in the Tr (Fig. 6, A and B). Regarding Akt activation by IGF-I, an increase in phospho-Akt amounts was found in the Em and not in the Tr, indicating that IGF-I acts via its receptor only in the Em in rabbit blastocysts. Insulin significantly stimulated phospho-Akt protein amounts but only in the Tr (Fig. 6, A and C).
Akt and ERK1/2 phosphorylation in separated Em and Tr of d-6 rabbit blastocysts. The phosphorylation of Akt and ERK1/2 was analyzed by Western blotting on 10 μg protein from 10 separated blastocysts cultured for 10 min with 17 nm insulin (Ins) or 1.3 nm IGF-I. Control blastocysts were cultured without growth factors (–). Detections of phospho-ERK, phospho-Akt, and β-actin were performed on the same blotting membrane. Each Western blot was repeated twice. A, Representative blot; B and C, relative stimulation of Akt and ERK1/2 phosphorylation by insulin and IGF-I in Em and Tr (n = 3 experiments with 10 blastocysts per treatment per experiment; protein phosphorylation in untreated embryos was set at 1). Asterisks indicate statistically significant differences: *, P < 0.05; **, P < 0.005. Insulin treatment increased ERK1/2 and Akt phosphorylation in Em and Tr, whereas IGF-I increased ERK1/2 and Akt phosphorylation only in the Em.
Quantification of the MAPK/ERK target gene c-fos in d-6 blastocysts.
After culture with insulin or IGF-I for 10 and 30 min, c-fos was quantified in whole blastocysts and in separated Em and Tr by real-time PCR. A time-dependent increase was observed with maximal c-fos transcript numbers for insulin after 10 min and IGF-I after 30 min (Fig. 7A). The c-fos increase was inhibited by PD98059 (data not shown).
Amounts of c-fos RNA in d-6 rabbit blastocysts cultured with insulin or IGF-I. The amount of the MAPK/ERK target gene c-fos was quantified by real-time PCR, investigating different time intervals (10, 20, 30, 40, and 60 min) after growth factor stimulation (A). RNA amounts were quantified in pooled blastocysts (mean ± sem; n = 3 experiments with at least four blastocysts per treatment per experiment). A, Maxima of c-fos RNA were found after 10 and 30 min for insulin and IGF-I, respectively; B, stimulations for 10 and 30 min were used to analyze c-fos expression in individual blastocysts. After culture with (+insulin and +IGF1) or without (Contr) growth factor supplementation, blastocysts were separated into Em and Tr. Two experiments with at least five blastocysts were carried out per treatment per experiment. Real-time PCR was performed on individual Em and Tr. The amount of c-fos was calculated as number of c-fos copies per 103 molecules GAPDH. Compared with controls, a significant increase in c-fos copies was measured after supplementation for both growth factors (*, P < 0.05). Insulin led to higher c-fos RNA amounts in the Tr compared with the Em, whereas IGF-I stimulated the Em more than the Tr (**, P < 0.005).
In the two cell lineages, an approximately 14-fold increase of c-fos in Tr and a 3-fold higher transcript number in Em (Fig. 7B) was found. In contrast, the increase by IGF-I was observed mainly in the Em (∼12-fold) and only 2-fold in the Tr (Fig. 7B).
Quantification of the PI3K/Akt target gene PEPCK in d-6 blastocysts.
PEPCK RNA was quantified in whole blastocysts and in separated Em and Tr by real-time PCR. A time-dependent decrease was observed for insulin after 1 and 2 h, whereas IGF-I showed no effect (Fig. 8A). The PEPCK decrease was blocked by LY290002 (Fig. 8A). These results were confirmed in separated Em and Tr, where IGF-I did not regulate PEPCK transcription.
Phosphoenolpyruvate carboxykinase (PEPCK) RNA amounts in d-6 rabbit blastocysts cultured with insulin or IGF-I. The amount of the PI3K/Akt target PEPCK was quantified by real-time PCR (mean ± sem; n = 3 experiments). The number of pooled blastocysts used for analysis is inserted in the bars. Controls were cultured without growth factors. A, The growth factor stimulation was performed for 60 (insulin and IGF-I) and 120 min (insulin). Insulin supplementation significantly decreased PEPCK amounts (*, P < 0.05), whereas IGF-I had no effect. An inhibition of PI3K by LY294002 (LY) blocked the insulin-mediated reduction of PEPCK RNA. B, The time interval of 1 h was used to analyze PEPCK transcription in single blastocysts that had been separated into Em and Tr after culture. Real-time PCR was performed on individual probes of Em and Tr. The amounts of PEPCK RNA were calculated relative to GAPDH as arbitrary units with the unit for Em in controls set at 100%. A significant decrease in PEPCK copies after insulin treatment was measured in the Tr (*, P < 0.05) but not in the Em.
PEPCK amount was approximately 2-fold higher in the Tr than in the Em. The insulin effect was restricted to Tr. Only in this cell lineage were PEPCK RNA amounts significantly decreased by insulin, whereas the expression in the Em was unaffected (Fig. 8B).
Discussion
This study is the first report demonstrating not only a different expression pattern of IR isoforms and IGF-IR in Em and Tr but also a cell lineage-specific signaling of these growth factors in blastocysts. The IR-A is the main isoform in blastocysts with high expression levels in Em and Tr. IR-B expression is clearly lower and detectable only in Tr cells. Broken down per blastocyst, approximately 80% of IR RNA belong to the splice variant IR-A and approximately 20% to IR-B.
It is well known that the relative expression of the two isoforms varies in a tissue-specific manner. IR-A is expressed predominantly in central nervous system and hematopoietic cells, whereas IR-B is expressed predominantly in adipose tissue, liver, and muscle. The latter are the major target tissues for the metabolic effects of insulin (3, 40). Little functional differences in insulin binding and IR activation have been described for the two isoforms. IR-A has a slightly higher binding affinity, and IR-B has a more efficient signaling activity. The definite biological roles of the two isoforms in adult tissues are not yet known. Because of the high amounts of IR-A in embryonic tissues and tumor cells, a role in fetal growth and cancer biology has been assumed (1, 41, 42). In fetal fibroblasts, for example, the relative abundance of IR-A is as high as 72–84% in contrast to 20–39% in adult fibroblasts. Frasca and co-workers (1) suggested that during fetal development, most tissues predominantly express IR-A and that splicing of exon 11 of the IR is developmentally regulated. In the Em of rabbit blastocysts, i.e. in the precursor tissue of the embryo proper, the IR-A is the predominant isoform. In the Tr, the IR is represented by both isoforms, however, in relatively low amounts (∼20%). The receptor protein is localized in the membrane of Em and Tr cells indicating that the IR isoforms are functional.
The functional relevance of two IR splice variants in individual cells is not clear. IR-A binds insulin with slightly higher affinity than IR-B (43). Experiments by Frasca and co-workers (1) in IGF-IR-negative cells have shown that IGF-I binds only with low affinity to both isoforms. IGF-II is bound by IR-A with relatively high affinity (30–40% that of insulin). Only the IR-A isoform can be activated by IGF-II; IGF-II binding to IR-B does not result in autophosphorylation of IR-B (1). These differences led to the conclusion that IR-B is responsible only to insulin, whereas IR-A can react to both insulin and IGF-II. The blastocyst has access to both growth factors. Whether IGF-II exerts a specific IR-A-mediated function during blastocyst development and differentiation needs to be investigated in additional experiments.
To elucidate insulin and IGF-I effects, we have used insulin and IGF-I in physiological concentrations of 17 and 1.3 nm, respectively. These concentrations are in an optimal range for specific ligand binding (1) and induction of biological effects (11). In cell lines, the EC50 for IR-A autophosphorylation was estimated to be 0.6–0.8 and more than 30.0 nm for insulin and IGF-I, respectively. EC50 of IR-B for insulin and IGF-I range was at the same levels. For activation of IGF-IR, a higher concentration of insulin with EC50 higher than 30 nm and a lower concentration of IGF-I with EC50 of 1.6–2.1 nm are necessary (1). IGF-I has a low affinity to both IR isoforms. Implying homologous ligand-receptor interactions, we assume that the receptor expression pattern in Em and Tr defines the responsiveness to insulin and IGF-I in these cell lineages. The published data (44) report observations that IGF-I interacts exclusively with the IGF-IR, analyzing mouse phenotypes resulting from targeted mutagenesis of the IGF-I and IGF-II genes and the cognate IGF-IR gene. In mice blastocysts, the IGF-IR is localized in the cytoplasm of Tr and in the plasma membranes of those ICM cells adjacent to Tr cells (17, 45). These findings can be confirmed for the rabbit blastocyst with the IGF-IR predominantly confined to the plasma membrane and cytoplasm of Em cells and to lesser extent to the cytoplasm of Tr. The immunofluorescence intensities in both compartments correlated closely with the mRNA data, of which approximately 80% could be allocated to the Em. The main expression and plasma membrane localization of IGF-IR in the Em implicate a molecular link for IGF-I function in this cell lineage. Harvey and Kaye (46) showed that IGF-I stimulated exclusively the ICM in mice. The authors suggested that Tr receptors transcytose IGF-I to bind to the ICM receptors. Movement of labeled IGF-I across the Tr cells has been visualized by electron microscopy (46).
The IR and IGF-IR expression pattern, based on transcription and immunofluorescence analyses, is schematically shown in Fig. 9 and linked with the main downstream signaling pathways of insulin and IGF-I. Analyzing whole blastocysts, we have shown before that insulin activates the MAPK/ERK signaling pathway in rabbit blastocysts (38). Employing separated Em and Tr, we show now that ERK phosphorylation induced by insulin occurs in the Em and, to a lower amount, in the Tr. IGF-I exclusively activates the MAPK/ERK signaling in the Em. This finding is supported by c-fos transcription, which was significantly increased in the Em. Insulin also resulted in increased c-fos RNA amounts but, in contrast to IGF-I, mainly in the Tr and only to a lower extent in the Em. Surprisingly, the enhanced ERK phosphorylation by insulin did not result in equivalently increased c-fos stimulation in the Em. Therefore, insulin does not activate c-fos-mediated cell growth via the AP-1 complex in Em cells. Most likely, insulin activates other downstream target genes in the Em than c-fos.
Schematic sketch of insulin and IGF-IR expression and growth factor actions in 6-d-old rabbit blastocysts. Different cellular responses of the Em and Tr on insulin and IGF-I. The effect of IGF-I is mainly restricted on proliferation of the Em, whereas insulin activates proliferation in both cell lineages. In the Tr, insulin regulates PEPCK, a key enzyme of gluconeogenesis.
Regarding potential effects of insulin on the Tr, various thoughts merit consideration. The stimulation of glucose transport and the metabolic effects of insulin are transduced by the PI3K/Akt pathway. This pathway is functional in the Tr cells of rabbit blastocysts. Presence and functionality of the PI3K/Akt pathway during preimplantation embryo development has also been shown in mice preimplantation embryos (47). The ultimate effector system for regulating glucose uptake is the translocation of glucose transporter 4 (GLUT4) to the plasma membrane. In previous studies (19, 38), however, we could not find GLUT4 translocation and found no increase in glucose uptake after insulin stimulation in cultured rabbit blastocysts. In the present study, we show that insulin inhibits gluconeogenesis by suppressing the expression of the key gluconeogenic enzyme PEPCK. PEPCK catalyzes one of the rate-limiting steps of gluconeogenesis, the reaction of oxaloacetic acid to phosphoenolpyruvate. As known from other cells and confirmed for blastocysts in present study, the inhibition of PEPCK by insulin is regulated transcriptionally (48) via the activation of the PI3K pathway (49). The PEPCK gene is expressed in both cell lineages of the blastocyst but with a significantly higher amount in the Tr. Insulin down-regulated PEPCK mRNA by activation of the PI3K/Akt pathway in the Tr but not in the Em. The cell lineage-specific differences in PEPCK regulation by insulin correlate closely with IR isoform expression. The IR-B isoform, mainly associated with metabolic activity, is transcribed only in the Tr. The down-regulation of PEPCK mRNA is insulin and Tr specific because IGF-I induces PI3K/AKT signaling in the Em with no regulatory effect on PEPCK expression. Therefore, the question arises as to which insulin effects are exerted on blastocysts during early ontogenesis. Our findings point toward stimulation of mitogenesis and inhibition of gluconeogenesis. Inhibition of gluconeogenesis is a classical metabolic effect of insulin. Although stimulation of Tr cell proliferation shortly before implantation is biologically meaningful, the down-regulation of the availability of the blastocyst’s main energy substrate glucose (no GLUT4 translocation, no stimulation of glucose uptake, down-regulation of gluconeogenesis by insulin) is difficult to understand and needs further investigation. Provision of precursors for cell proliferation by down-regulation of gluconeogenesis could be one possible regulatory mechanism but does not answer the question of how Tr cells secure their energy supply at a time of high cell proliferation. Whether Tr cells satisfy their energy needs by other metabolic pathways and/or substrates is currently under investigation. Finally, other downstream targets of PI3K/AKT signaling beside PEPCK have to be considered. Potential targets of this pathway are apoptotic processes. Akt is a primary target for mediating antiapoptotic signals (50). Apoptosis is a common and physiological feature in the implanting blastocyst (12, 47, 51).
Besides IR/IGF-IR expression and activation, the IRS proteins are sites of diversification of insulin and IGF-I outcomes (reviewed in Ref. 52). IRSs are responsible for transducing insulin and IGF-I signals from receptors to intracellular effectors. IRS1 has robust interaction with Grb2, leading to the activation of the MAPK pathway, whereas IRS2 interacts only weakly with Grb2, suggesting that IRS1 corresponds best with mitogenic and IRS2 with metabolic effects. In mice preimplantation embryos, IRS1 expression has been shown (53), suggesting that IRS1 is expressed in all cell lineages of the periimplantation mouse embryo and mediates effects of insulin and IGFs. In rabbit blastocysts, the expression of IRSs is still not known. Analyses of IRS1 (54) and IRS2 (55) knockout mice support the concept that each of these proteins has unique functions. Mice homozygous for targeted disruption of the IRS1 gene were retarded in embryonic and postnatal growth (54). In contrast to the IRS1 null mice, IRS2 null mice showed minimal growth retardation but developed peripheral insulin resistance (55). Potential variation in IRS expression and docking protein activation may result in differential action of insulin and IGF-I in Em and Tr and needs further investigation.
Taken together, we have shown that IGF-I binding to its receptor activates a mitogenic response in the Em. Insulin signaling is both mitogenic and metabolic and is restricted mainly to the Tr. The metabolic effects result in a down-regulation of gluconeogenesis. The differences in hormone action in blastocysts can be attributed to a cell lineage-specific receptor expression pattern (Fig. 9). The classical downstream pathways known from adult cells, PI3K/Akt and the MAPK/ERK, are functional in embryonic cells as early as during preimplantation embryo development. The actions of insulin and IGF-I in blastocysts deserve special attention because elevated levels of insulin are associated with common diseases such as diabetes, and increased levels of free IGF-I occur in women with polycystic ovary syndrome (39). In contrast, low levels of IGFs are associated with intrauterine growth restriction (11). Unphysiological IGF and insulin levels represent a potential risk for early embryo development and pregnancy outcome.
Acknowledgments
We thank Michaela Kirstein, Ivonne Peters, and Elisabeth Schlüter for excellent technical assistance and Evelyn Axmann for help in the preparation of the manuscript.
This work was supported by grants from the Deutsche Forschungsgemeinschaft (DFG Grant FI 306/13-1) and the Wilhelm Roux Programme of the MLU Faculty of Medicine.
Disclosure Statement: The authors have nothing to disclose.
Abbreviations:
- AP-1
Activator protein-1
- CT
cycle threshold
- Em
embryoblast
- GLUT4
glucose transporter 4
- ICM
inner cell mass
- IGF-IR
IGF-I receptor
- IR
insulin receptor
- IRS
insulin receptor substrate
- PEPCK
phosphoenolpyruvate carboxykinase
- PI3K
phosphatidylinositol 3-kinase
- Tr
trophoblast.
Author notes
A.N.S. and N.R. contributed equally to this work.
