Organ-Specific Defects in Insulin-Like Growth Factor and Insulin Receptor Signaling in Late Gestational Asymmetric Intrauterine Growth Restriction in Cited1 Mutant Mice

Studies in Cited1 mutant mice identify organ-specific defects in IGF1R and IR signaling that may regulate asymmetric organ growth in late gestational placental insufficiency.

Abstract

Late gestational placental insufficiency resulting in asymmetric intrauterine organ growth restriction (IUGR) is associated with an increased incidence of diabetes, cardiovascular and renal disease in adults. The molecular mechanisms mediating these defects are poorly understood. To explore this, we investigated the mechanisms leading to IUGR in Cited1 knockout mice, a genetic model of late gestational placental insufficiency. We show that loss of placental Cited1 leads to asymmetric IUGR with decreased liver, lung, and kidney sizes and preservation of fetal brain weight. IGF and insulin signaling regulate embryonic organ growth. IGF-I and IGF-II protein and mRNA expression are reduced in livers, lungs, and kidneys of embryonic d 18.5 embryos with IUGR. Decreased IGF-I is associated with reduced activating phosphorylation of the type 1 IGF receptor (pIGF-IR) in the kidney, whereas reduced IGF-II is associated with decreased phosphorylation of the insulin receptor (pIR) in the lung. In contrast, decreased pIR is associated with reduced IGF-I but not IGF-II in the liver. However, pancreatic β-cell mass and serum insulin levels are also decreased in mice with IUGR, suggesting that hepatic IR signaling may be regulated by alterations in fetal insulin production. These findings contrast with observations in IUGR fetal brains in which there is no change in IGF-IR/IR phosphorylation, and IGF-I and IGF-II expression is actually increased. In conclusion, IUGR disrupts normal fetal IGF and insulin production and is associated with organ-specific defects in IGF-IR and IR signaling that may regulate asymmetric IUGR in late gestational placental insufficiency.

Late gestational placental insufficiency is the commonest cause of intrauterine growth restriction (IUGR) in the United States (1). This has profound effects on fetal growth that increase perinatal mortality and predisposes to diabetes, cardiovascular, and renal disease in adult life (2, 3). Understanding the fetal mechanisms mediating these effects could have a major impact on human health and disease. Late gestational IUGR leads to asymmetric organ growth with reduced fetal kidney, liver, and lung size associated with relative preservation of brain size. This is thought to result from compensatory changes in the fetal circulation with preferential shunting of blood toward the brain and away from other organs that are not essential for fetal growth (4). However, the molecular mechanisms by which the affected organs respond to these changes are unclear. Experimental models have established common cellular and molecular defects associated with IUGR (5). In particular, maternal malnutrition and uterine artery ligation promote organ-specific increases in cellular apoptosis (6–11). This is thought to result from local tissue responses to nutritional substrate and/or oxygen deficiency. However, there is also evidence of more generalized defects in fetal growth factor signaling associated with IUGR that could mediate the effects of substrate deficiency on organ growth (12–21).

IGF-I receptor (IGF-IR) and insulin receptor (IR) signaling pathways are major regulators of late gestational fetal organ growth (22, 23). IGF-IR is the receptor for IGF-I and IGF-II, whereas IR mediates both IGF-II and insulin signaling in target tissues. Targeted deletion of the Igf1r and Ir in the germ line induces IUGR in mice (24, 25). The severity of growth restriction in these models is different (55% for Igf1r, 10% for Ir knockout mice), suggesting that the majority of growth-promoting signals are mediated through IGF-IR. However, Igfr1/Ir double-knockout mice have an even more profound reduction in fetal weight (22), suggesting that IR signaling also plays a role in regulating late gestational fetal growth. These findings have led a number of investigators to study the regulation of IGF-IR and IR ligands and their inhibitors in IUGR. For example, umbilical cord blood IGF-I and insulin (but not IGF-II) levels are lower in human infants with IUGR (12–14). Decreased circulating levels of IGF-I and insulin have also been demonstrated in rodent and sheep models of late gestational placental insufficiency, and decreased circulating IGF-I is associated with reduced hepatic Igf1 mRNA expression (15–18). However, there has been no systematic analysis of IGF-I or IGF-II expression in other organs affected by IUGR. Moreover, complex interactions with a variety of IGF-binding proteins regulate bioavailability and net activity of IGF ligands in vivo (26). Because these factors are also abnormally regulated in different models of IUGR (18–21), the net effect of decreasing circulating and/or hepatic IGF-I on IGF-IR signaling are unknown. In addition, because there are no established mouse models of late gestational placental insufficiency and IUGR, the functional significance of these changes in IGF and insulin expression in the fetus cannot be established using standard genetic approaches in mice.

To this end, we have developed Cited1 knockout mice as a genetic model of late gestational placental insufficiency (27, 28). CITED1 is a transcriptional coactivator that is expressed in the placenta, heart, limb buds, liver, and kidney during embryogenesis (28–30). Cited1-null mice on a 129/sv or mixed strain background develop normally but on a C57BL/6 background develop late gestational placental insufficiency resulting from a reduction in the surface area available for gas and nutrient exchange across the placenta (28). This gives rise to IUGR associated with abnormal patterning of the renal medulla, both of which are caused by placental insufficiency independent of changes in embryonic CITED1 expression (27, 28). It is unknown whether these mice show classical features of asymmetric IUGR. However, those mice surviving the early postnatal period are otherwise phenotypically normal and show postnatal catch-up growth curves that are typically seen in patients with IUGR (28). Placental insufficiency in Cited1 mutant mice invariably affects homozygous null mutants of both genders but also affects a proportion of heterozygous females (28). This occurs because as an X-linked gene (31), Cited1 undergoes imprinting when the paternally inherited X chromosome is inactivated in the placenta (32); therefore, female Cited1 heterozygotes with a paternally inherited wild-type X chromosome are Cited1 null in the placenta and exhibit placental insufficiency, whereas female Cited1 heterozygotes with a maternally inherited wild-type X chromosome are Cited1 heterozygous in the placenta with no placental or fetal growth abnormalities (27, 28). This phenomenon provides a biological tool that can differentiate embryonic from placental effects of the Cited1 mutation on fetal growth and indicates that IUGR in these mice is caused by placental insufficiency independent of changes in embryonic CITED1 expression. The purpose of these studies, therefore, was to characterize the IUGR phenotype in Cited1 mutant mice and to provide a comprehensive analysis of defects in the IGF and insulin pathways associated with asymmetric organ growth restriction in this mouse model of late gestational placental insufficiency.

In this paper, we show that Cited1 mutant mice with placental insufficiency develop classic features of asymmetric IUGR and that this is associated with organ-specific defects in IGF-IR and IR signaling. Parallel changes in IGF-I and IGF-II expression in the brain, lung, and kidneys suggest that defects in IGF-IR and IR phosphorylation result from organ-specific alterations in ligand expression levels. However, decreased IR phosphorylation in the liver is more closely associated with decreased levels circulating insulin and reduced pancreatic β-cell mass, suggesting that decreased fetal insulin secretion into the portal circulation could play a primary role in reducing late gestational liver growth in IUGR. These findings establish for the first time organ-specific changes in IGF-IR and IR signaling that could account for asymmetric IUGR associated with placental insufficiency and set the stage to develop genetic tools that will enable us to explore the role of insulin, IGF-I, and IGF-II signaling in promoting organ-specific growth defects in IUGR.

Materials and Methods

Mice

Cited1 mutant mice (Cited1^tm1Dunw) (28) were maintained on a C57BL/6 background by backcrossing 15 generations with wild-type C57BL/6 mice. Because Cited1 is localized on the X-chromosome (31), female heterozygous mutants are referred to as Cited1^P+/− or Cited1^M+/−, depending on whether the wild-type allele was inherited from the paternal or maternal X-chromosome, respectively. Male null mutants are referred to as Cited1^Y/− mice. Timed pregnancies were performed with the morning of the vaginal plug counted as 0.5 d post coitus. Cited1^tm1Dunw genotyping and embryo sexing was performed using Cited1 and Sry primer sets, as previously described (27). All procedures performed were approved by the Vanderbilt Medical Center Institutional Animal Care and Use Committee.

Assessment of fetal morphometric parameters and tissue preparation

At embryonic 18.5 days post coitus (E18.5), fetuses were harvested into the ice-cold PBS, dried on absorbent tissue, and weighed on precision scales; crown-rump length was measured using a digital caliper. Fetuses were decapitated, and 20–30 μl blood was collected for glucose and insulin assays. The abdomen and thoracic cavity was opened by median incision and organs dissected under a dissecting microscope. Excess liquid was removed, and the organs were weighed and snap frozen for RNA and protein extraction or fixed in 4% paraformaldehyde for 1 h before processing and paraffin mounting.

Immunoblots

Tissue lysates were prepared by homogenizing snap-frozen whole kidneys, livers, lungs, and brain samples directly in RIPA buffer (50 mm Tris-HCl, 150 mm NaCl, 0.1% sodium dodecyl sulfate, 0.5% sodium deoxycholate, 1% and Nonidet P-40) to which we added protease (Thermo Scientific, Pittsburgh, PA) and phosphatase inhibitor cocktails (Sigma Chemical Co., St. Louis, MO). Immunoblots were performed in these tissue lysates using rabbit anti-phospho-IGF-IR/IR, anti-IGF-IR-β-chain, anti-IR-β-chain (Cell Signaling, Beverly, MA), and mouse anti-β-actin (Sigma). These were detected with horseradish peroxidase (HRP)-conjugated secondary antibodies (antirabbit HRP from Cell Signaling; antimouse HRP from Kirkegaard and Perry, Gaithersburg, MD). Blots were scanned using a UVP Epichem Darkroom scanner, and individual bands were identified and quantified using Labworks software. Results were normalized to total IGF-IR or IR, as indicated. For comparative analyses, seven Cited1 mutant and seven wild-type (or Cited1^M+/−) control samples were prepared and run on the same gels. Experiments were performed at least twice using different samples.

Immunoprecipitation and immunodepletion studies

For immunoprecipitation studies, 250 μg lysates from E18.5 wild-type kidneys, livers, lungs, and brains (pooled from two embryos) were prepared by homogenizing snap-frozen tissues directly in RIPA buffer with protease and phosphates inhibitors (as outlined above). Antibody bead complexes were generated by incubating 2 μg rabbit anti-IGF-1R β-chain, anti-IR β-chain antibodies (Santa Cruz Biotechnology, Santa Cruz, CA) or normal rabbit IgG (Sigma) with 25 μl protein G-conjugated Dynabeads for 15 min at room temperature. Antibody-bead complexes where then incubated with the tissue lysates for 30 min at 4 C on a rotator. Dynabeads were washed three times in RIPA buffer, precipitated proteins were eluted in sample buffer at 95 C and separated by 8% SDS-PAGE, and receptors were detected by immunoblot, as outlined above. Tissue lysates (50 μg) were also separated for comparison. For immunodepletion studies, we followed the same protocol using only 50 μg tissue lysates and incubated lysates with the antibody-Dynabead complexes overnight to increase depletion efficiency. Dynabeads were removed and remaining supernatants evaluated by immunoblot, as outlined above. Whole tissue lysates (50 μg) that had not been immunodepleted were separated in parallel lanes for comparison.

Tissue IGF-I and IGF-II expression levels and serum glucose and insulin measurement

IGF-I and IGF-II protein concentrations were measured in tissue lysates using sandwich ELISA kits (R&D Systems, Minneapolis, MN) following the manufacturer's protocols. Results are expressed in picograms per milligram total protein (determined by Bradford assay). Because of the small volume of serum accessible from each embryo (5–10 μl), four to six samples from fetuses with same genotype were pooled for insulin RIA (Hormone Assay and Analytical Services Core, Vanderbilt Diabetes Research and Training Center) and glucose measurement by glucose oxidase assay (Mouse Metabolic Phenotyping Core, Vanderbilt University, Nashville, TN).

Quantitative RT-PCR

Total RNA was extracted from snap-frozen tissues using RNeasy mini kits (QIAGEN, Valencia, CA) and quantified using a NanoDrop UV spectrophotometer (Thermo Scientific). Reverse transcription of 1 μg RNA by Superscript III (Invitroge, Carlsbad, CA) and oligodT primers (Applied Biosystems, Foster City, CA) was used to generate cDNA for analysis by quantitative RT-PCR (7900 HT; Applied Biosystems) using SYBR Green Master Mix (Applied Biosystems). Primers were designed to cross exon boundaries and included for Igf1, forward AGCTGTGATCTGAGGAGACTGG and reverse TCTTGTTTGTCGATAGGGACGG (product size 100 bp); for Igf2, forward ACACGCTTCAGTTTGTCTGTTCGG and reverse AAGCAGCACTCTTCCACGATG (product size 100 bp); and for β-actin, forward AGTGTGACGTTGACATCCGTA and reverse GCCAGAGCAGTAATCTCCTTCT (product size 112 bp). Changes in mRNA expression were determined by comparison of sample cycle threshold values against a standard curve generated using pooled sample cDNA. Results were adjusted to β-actin level and expressed as degree of change of wild-type control.

Relative expression of IR-A and IR-B splice variants

Relative expression levels of Ir-A and Ir-B splice variant mRNA were evaluated by RT-PCR by comparing the 210-bp Ir-A and 246-bp Ir-B PCR products amplified in the same reaction with primers spanning the mouse Ir exon 11 (deleted in Ir-A), as described (33). PCR products were detected in ethidium bromide-stained gels and relative band densities quantified by scanning densitometry.

Assessment of β-cell mass

E18.5 pancreata were isolated, dried on absorbent tissue, and weighed on precision scales. Samples were fixed in 4% paraformaldehyde for 1 h and simultaneously flattened longitudinally by mild compression using a sponge insert in a mounting cassette before processing and embedding in paraffin. For quantitative analysis, every 10th 5-μm section was collected and stained using guinea pig anti-insulin polyclonal antibodies (Millipore, Billerica, MA), detected using species-specific HRP-conjugated secondary antibodies (Jackson Immunologicals, West Grove, PA), and counterstained with hematoxylin and eosin (Sigma). The section interval was determined after initial evaluation indicated islet diameters in wild-type and Cited1 mutant E18.5 pancreata of 30–40 μm. On this basis, analysis of pancreatic islets by sequential sampling of every 10th section provides a representative measure of islet surface area (34). Panoramic images of pancreatic sections were compiled for each section. Insulin staining and total pancreatic areas were determined from these images using Adobe Photoshop CS4 Extended software to measure defined surface areas in pixels. β-Cell mass = insulin staining/total pancreatic surface areas in pixels for all sections of each pancreas × total pancreatic weight in milligrams.

Statistical analyses

Group differences were determined using unpaired Student's t test. The threshold for significance was set at P < 0.05. For multiple, between-group comparisons (Supplemental Fig. 1, published on The Endocrine Society's Journals Online web site at http://endo.endojournals.org), we used one-way ANOVA with Bonferroni correction for post hoc, between-group comparisons. All values are expressed as mean ± sem unless otherwise indicated, and data were analyzed with GraphPad Prism version 5.0 (GraphPad Software Inc., San Diego, CA).

Results

Cited1-null mice have asymmetric IUGR

Cited1-null males on a C57BL/6 background have reduced body weight and crown rump length at E18.5 (Fig. 1, A and B). Moreover, although 80% of Cited1-null mice die in the early postnatal period (28), of those surviving, Cited1-null males show postnatal catch-up growth and are indistinguishable from wild-type littermates by 8 wk of age (Supplemental Fig. 1). Cited1-null females also show reduced body mass and crown-rump length at E18.5 (Fig. 1, D and E). There is a decrease in absolute liver, lung, and kidney weights in both male and female Cited1 knockout mice (Table 1), with a more profound reduction in liver and lung mass, as demonstrated by reduced organ to body weight ratios (Fig. 1, C and F). In contrast, although the absolute brain weight is lower in Cited1-null male (but not female) mice (Table 1), unlike the other organs, the brain to body weight ratios are increased in Cited1-null mice compared with wild-type littermates (Fig. 1, C and F). These findings indicate that both male and female Cited1-null mice have classical asymmetric IUGR associated with placental insufficiency.

Fig. 1.

Cited1 mutant mouse embryos have asymmetric IUGR. A–C, Anthropometric parameters (A and B) and organ to body weight ratios (C) in Cited1-null males (Cited1^Y/−, n = 55) vs. wild-type male littermates (Cited1^Y/+, n = 48) at E18.5; D–F, anthropometric parameters (D and E) and organ to body weight ratios (F) in Cited1-null females (Cited1^−/−, n = 18) vs. wild-type females (Cited1^+/+, n = 25) at E18.5. Organ to body weight ratios are expressed as percent change in Cited1-null vs. wild-type male or female embryos, respectively. Data are expressed as mean ± sem. Student's t test: *, P < 0.001 vs. wild-type littermates; **, P < 0.01 vs. wild-type littermates.

Table 1.

Organ and body weights in Cited1-null embryos at E18.5

Genotype	Sex	Body weight (g)	Brain weight (mg)	Liver weight (mg)	Lung weight (mg)	Kidney weight (mg)
Cited1Y/+	M	1.2 ± 0.07	74.6 ± 6.9	60.9 ± 11.5	36.9 ± 6.2	11.1 ± 2.4
Cited1Y/−	M	0.9 ± 0.09a	66.1 ± 8.6a	35.5 ± 8.9a	24.1 ± 4.6a	8.3 ± 2.0a
Cited1+/+	F	1.1 ± 0.08	69.5 ± 6.7	59.5 ± 10.1	36.2 ± 5.2	10.7 ± 1.5
Cited1−/−	F	0.9 ± 0.08a	68.8 ± 7.4	37.5 ± 9.2a	22.3 ± 4.5a	8.3 ± 1.7a

Male (M) Cited^Y/− embryos and wild-type male littermates and female (F) Cited1^−/− embryos and female controls at E18.5, as indicated. Numbers of embryos examined were as follows: for Cited1^Y/− males, n = 55 for body and brain weights, and n = 30 for other organ weights; for Cited1^Y/+ (wild-type) males, n = 48 for body and brain weights, and n = 25 for other organ weights; and for Cited^−/− females, n = 17, and for wild-type females, n = 29 for body weight and all organs. Results are expressed as means ± sd organ or body weight in grams or milligrams, as indicated.

^aStudent's t test: P < 0.0001 vs. male or female wild-type controls, respectively.

Loss of Cited1 expression in the placenta promotes asymmetric IUGR

Because CITED1 is expressed in a number of embryonic tissues (28–30), we wanted to determine whether asymmetric IUGR in Cited1-null mice is caused by placental insufficiency resulting from loss of placental CITED1 expression (28) or from loss of CITED1 expression in the affected fetal organs. To address this, we took advantage of the localization of the Cited1 gene on the X-chromosome (31). On this basis, comparative analysis of female Cited1 heterozygotes with paternally (Cited1^P+/−) vs. maternally (Cited1^M+/−) inherited wild-type X-chromosomes enables us to distinguish between placental and embryonic effects of the Cited1 deletion on embryonic growth. Female Cited1^P/+ heterozygotes with paternally inherited wild-type Cited1 alleles have placental insufficiency and have reduced body weight and crown-rump length at E18.5 when compared with Cited1^M/+ mice, which do not have placental insufficiency (Fig. 2, A and B). There is also a decrease in absolute liver, lung, and kidney weight in Cited1^P+/− female heterozygotes compared with Cited1^M+/− mutants (Table 2). In addition, there is reduced organ to body weight ratios in the liver and lungs, whereas brain to body weight ratios are significantly increased in Cited1^P+/− when compared with Cited1^M+/− embryos (Fig. 2C). These findings indicate that placental insufficiency resulting from loss of CITED1 expression in the placenta, and not in the embryo proper, gives rise to classic asymmetric IUGR in Cited1-null mice.

Fig. 2.

Cited1 heterozygous mutant embryos with placental insufficiency have asymmetric IUGR. A–C, Anthropometric parameters (A and B) and organ to body weight ratios (C) in female Cited1 heterozygotes with placental insufficiency (paternal wild-type Cited1 allele, Cited1^P+/−, n = 20) vs. those without placental insufficiency (maternal wild-type Cited1 allele, Cited1^M+/−, n = 27) at E18.5. Organ to body weight ratios expressed as percent Cited1^P+/− vs. Cited1^M+/− embryos. Data are expressed as mean ± sem. Student's t test: *, P < 0.01 vs. Cited1^M+/− embryos.

Table 2.

Organ and body weights in E18.5 female Cited1 heterozygotes with placental insufficiency

Genotype	Body weight (g)	Brain weight (mg)	Liver weight (mg)	Lung weight (mg)	Kidney weight (mg)
Cited1M+/−	1.2 ± 0.09	73.1 ± 7.1	63.9 ± 9.3	36.6 ± 6.3	11.3 ± 2.2
Cited1P+/−	1.0 ± 0.11a	68.2 ± 7.6	47.3 ± 10.7a	26.7 ± 5.6a	9.1 ± 1.8a

Numbers of embryos examined were as follows: Cited1^M+/− (without placental insufficiency), n = 30 for body and brain weights, and n = 20 for other organs weights; Cited1^P+/− (with placental insufficiency), n = 20 for body and brain weights, and n = 12 for other organ weights. Results are expressed as means ± sd organ or body weight in grams or milligrams, as indicated.

^aStudent's t test: P < 0.01 vs. Cited1^M+/− embryos.

Detecting distinct IGF-IR and IR phosphorylation events in mouse embryonic tissues

The IGF and insulin signaling regulate late gestational fetal growth (35). We therefore evaluated the levels of activating Y^1135/1136 phosphorylation of the type 1 IGF receptor β [phosphorylated IGF-IR (pIGF-IR)] and Y^1150/1151 phosphorylation of the IRβ (pIR) in affected organs. IGF-IR and IR show sequence similarity close to these tyrosine phosphorylation sites (36, 37). As a result, no specific antibodies against phosphorylated forms of these receptors have been developed that can distinguish between the two receptors. For our studies, we used commercially available antibodies that recognize residues surrounding phosphorylated Tyr1135/1136 of IGF-IR and phosphorylated Tyr1150/1151 of IR β-chains that are the earliest autophosphorylation sites necessary for receptor activation (38). We detect a 93- and 97-kDa doublet in embryonic and early postnatal kidneys using these antibodies (Fig. 3A, lanes 10–12). To identify these bands, we compared their mobility with total IGF-IR and IR in the same samples detected using specific IGF-IR and IR antibodies. The lower pIGF-IR/IR band comigrates with the 93-kDa IR band (Fig. 3A, lanes 1–4), whereas the upper pIGF-IR/IR band comigrates with the 97-kDa IGF-IR band in embryonic kidneys and early postnatal kidneys (Fig. 3A, lanes 5–8). Moreover, there is a reduction in total IGF-IR expression associated with a dominant increase in expression of the lower pIGF-IR/IR band in adult kidneys (Fig. 3A, lanes 1, 5, and 9). These findings are consistent with observations in cultured cells indicating that pIR migrates with a greater mobility than pIGF-IR by SDS-PAGE (39) and indicate that differential mobility of pIGF-IR/IR bands can be used to detect distinct IGF-IR and IR phosphorylation events in embryonic tissues.

Fig. 3.

IGF-IR and IR phosphorylation in mouse tissues. A and B, Western blots for total IR, IGF-IR, and pIGF-IR/IR expression in wild-type E18.5 embryo, newborn (P0), 1-wk and 24-wk postnatal mouse kidneys (A) and in E18.5 wild-type mouse whole brain, lung, liver, and kidney specimens (B). C, Immunoprecipitation of IGF-IR, IR, or rabbit IgG controls from 250 μg of tissue lysates from wild-type E18.5 mouse whole brain, lung, liver, and kidney samples detected by immunoblot (IB) using pIGF-IR/pIR, total IGF-IR, and IR antibodies, as indicated. Tissue lysates (50 μg) were E18.5 separated for comparison. D, Immunodepletion of IGF-IR or IR by immunoprecipitation of 50 μg tissue lysates using receptor-specific or control IgG antibodies (IP, as indicated). Residual receptors in the depleted tissue lysates were detected by pIGF-IR/pIR, total IGF-IR, and IR antibodies, as indicated. Whole tissue lysates (50 μg) that had not been immunodepleted were separated in parallel lanes for comparison. Arrows indicate band identities referred to in the text.

Organ-specific regulation of IGF and IR phosphorylation in the mouse embryo

Comparative analysis of pIGF-IR/IR bands in E18.5 fetal organs also reveals organ-specific differences in the patterns of IGF-IR/IR phosphorylation (Fig. 3B, lanes 9–12). E18.5 kidneys show approximately equal levels of pIGF-IR and pIR, and there is dominant expression of the 93-kDa pIR band in lungs (Fig. 3B, lanes 10 and 12). Both organs express abundant IGF-IR and IR (Fig. 3B, lanes 2, 4, 6, and 8). Expression of IGF-IR in the liver is substantially lower than IR (Fig. 3B, lanes 3 and 7) and is associated with dominant expression of the 93-kDa pIR band (Fig. 3B, lane 11). The identity of an additional lower molecular mass band in the liver is uncertain because no corresponding 90-kDa band is detected using anti-IR antibodies (Fig. 3B, lane 3). A similar, less dominant 90-kDa pIGFIR/IR band is also detected in adult mouse kidneys (Fig. 3A, lane 9). The pIGF-IR/IR antibody detects an isolated 95-kDa band in E18.5 brain tissue distinct from the 97-kDA pIGF-IR band in lungs and kidneys (Fig. 3B, lanes 9, 10, and 12). Total IGF-IR also migrates at a slightly lower molecular mass in the brain (Fig. 3B, lanes 5–8), suggesting that this 95-kDA band is a pIGF-IR variant that is expressed only in the brain. This is consistent with the observation of a lower molecular mass IGF-IR variant in fetal (rat) brains compared with liver and placenta (40). For these studies, we refer to the 97-kDA pIGF-IR band as pIGF-IR(A) and the 95-kDA pIGF-IR band in the brain as pIGF-IR(B). Because the identity of the 90-kDA pIGF-IR/IR band in liver is unknown, we will evaluate expression only of the 93-kDA pIR band in liver samples.

To confirm identity of these pIGF-IR/pIR bands, we performed immunoprecipitation studies using IGF-IR- and IR-specific antibodies. The 95- and 97-kDA pIGF-IR/pIR bands in brains, kidneys, and lungs are detected only in the IGF-IR immunoprecipitates, whereas the 93-kDA pIGF-IR/pIR band in the kidney, lung, and liver is detected only in the IR immunoprecipitates (Fig. 3C). A faint 97-kDA pIGF-IR/pIR band is detected in IGF-IR immunoprecipitates from E18.5 livers (Fig. 3C, lane 11), indicating that there is in fact a low level of IGF-IR phosphorylation in E18.5 livers that is not detectable in tissue lysates (Fig. 3B, lane 11). Taken together, these findings support our initial observations indicating that the 95- and 97-kDA bands detected using the pIGF-IR/pIR antibodies represent phosphorylated variants of IGF-IR and that the 93-kDA pIGF-IR/pIR band is pIR. Additional evidence for this is provided from analysis of pIGF-IR/pIR bands in tissue lysates before and after depletion of IGF-IR or IR by immunoprecipitation (Fig. 3D). Notably, these studies also show that the 90-kDA pIGF-IR/pIR band that is detectable in E18.5 livers (Fig. 3B, lane 11) is not depleted by immunoprecipitation with either IGF-IR or IR antibodies (Fig. 3D, lanes 9–12), confirming our initial observation that the identity of this lower molecular mass pIGF-IR/pIR band is unknown.

Organ-specific regulation of IGF-IR and IR phosphorylation in Cited1 mutant mice

Having established the identity and expression of pIGF-IR and pIR bands in normal E18.5 mouse embryos, we determined whether activating IGF-IR/IR phosphorylation is modified in male Cited1-null embryos with asymmetric IUGR. There is a reduction in both pIGF-IR(A) and pIR expression in kidneys (Fig. 4, A and B), and reduced pIR but not pIGF-IR(A) in lungs and livers from Cited1-null mice (Fig. 4, C–F). In contrast, there is no change in expression of the pIGF-IR(B) band in brain lysates from Cited1-null mice (Fig. 4, G and H). These finding indicate that there are organ-specific changes in IGF-IR and IR phosphorylation in embryonic tissues subject to IUGR (kidney vs. liver and lung), whereas IGF-IR phosphorylation is unaffected in brains from IUGR embryos.

Fig. 4.

Organ-specific regulation of IGF-IR and IR phosphorylation in Cited1-null embryos. A, C, E, and G, Western blots for pIGF-IR/IR, total IGF-IR, and IR expression in E18.5 Cited1-null males (Cited1^Y/−, n = 7) and wild-type littermate (Cited1^Y/+, n = 7) kidneys (A), lungs (C), livers (E), and whole-brain lysates (G), with arrows indicating band identities (see text); B, D, F, and H, quantification of corresponding pIGF-IR/IR band densities relative to corresponding total IGF-IR and total IR expression in E18.5 in kidneys (B), lungs (D), livers (F), and brains (H). Results are expressed as mean ± sem. Student's t test: *, P < 0.01 vs. wild-type littermates.

To determine whether changes in IGF-IR/IR phosphorylation is caused by placental insufficiency or from loss of CITED1 expression in the affected fetal organs, we evaluated pIGF-IR/IR in E18.5 female Cited1 heterozygotes with placental insufficiency (Cited1^P+/−) vs. those with normal placental function (Cited1^M+/−). There is reduced pIR and pIGF-IR(A) in kidneys and a reduction in pIR expression in lungs and livers from E18.5 Cited1^P+/− embryos (Fig. 5, A–C) but no associated alteration in pIGF-IR(B) in Cited1^P+/− embryonic brains (Fig. 5D). These findings parallel our observations in Cited1-null mice and indicate that organ-specific defects in IGF-IR and IR signaling result from placental insufficiency and not from loss of CITED1 expression in the embryo.

Fig. 5.

Organ-specific regulation of IGF-IR and IR phosphorylation in female Cited1 heterozygotes with placental insufficiency (Cited1^P+/−, n = 7) and without placental defect (Cited1^M+/−, n = 7). A, C, E and G, Western blots for pIGF-IR, pIR, total IGF-IR and total IR expression in E18.5 Cited1^P+/− vs. Cited1^M+/− kidneys (A), lungs (C), livers (E), and whole-brain lysates (G), with arrows indicating band identities (see text). B, D, F and H, Quantification of corresponding pIGF-IR and pIR band densities relative to corresponding total IGF-IR and total IR expression in E18.5 Cited1^P+/− vs. Cited1^M+/− kidneys (B), lungs (D), livers (F), and whole-brain lysates (H). Results are expressed as mean ± sem. Student's t test: *, P < 0.01 vs. Cited1^M+/− embryos.

Organ-specific regulation of IGF-I and IGF-II in male Cited1-null mice

To determine the mechanism by which organ-specific defects in IGF-IR and IR signaling are regulated in mice with IUGR, we evaluated expression of IGF ligands in the affected organs. The IGF system includes two ligands, IGF-I and IGF-II, both of which activate IGF-IR signaling (23). IGF-II, but not IGF-I, also activate IR, predominantly via the dominant fetal IR splice variant, IR-A (41). Igf1 and Igf2 mRNA are selectively down-regulated in livers and lungs of Cited1-null males, respectively, whereas in the kidneys, both Igf1 and Igf2 mRNA expression is reduced (Fig. 6A). In contrast, there is no significant change in Igf1 or Igf2 mRNA expression in brains from Cited1-null embryos. These changes are associated with reduced IGF-I protein levels in kidneys and livers and decreased IGF-II in lungs of Cited1-null mice. In contrast both IGF-I and IGF-II protein levels are increased in Cited1-null embryonic brains (Fig. 6, B and C). These findings suggest that organ-specific defects in IGF-IR/IR phosphorylation in mice with asymmetric IUGR result from selective alterations in IGF-I and IGF-II ligand expression in kidneys, lungs, and brains. However, decreased liver size is associated with a reduction in pIR and not pIGF-IR expression in mice with IUGR (Figs. 4 and 5), indicating that alterations in hepatic IGF-I protein expression do not influence IGF-IR phosphorylation in the liver. This is not associated with reduced expression of IGF-II in the liver (Fig. 6B). Moreover, decreased expression of the IGF-II inhibitory receptor, IGF-IIR, in livers from mice with IUGR (Supplemental Fig. 2), suggests that decreased hepatic IR phosphorylation does not result from reduced IGF-II signaling in the liver.

Fig. 6.

Organ-specific regulation of IGF-I and IGF-II expression in E18.5 Cited1-null embryos. A, Igf1 and Igf2 mRNA levels determined by quantitative RT-PCR in Cited1-null (n = 8) and wild-type male (n = 8) brain, liver, lung, and kidney, as indicated, with mRNA levels corrected to β-actin mRNA, represented as fold change in Cited1-null vs. wild-type male embryos; B, IGF-I protein expression levels in Cited1-null (n = 16) and wild-type male (n = 14) embryonic organs; C, IGF-II protein expression levels in Cited1-null (n = 16) and wild-type male (n = 14) embryonic organs. IGF-I and IGF-II proteins are represented in picograms per milligram of protein in the tissue lysates. Results are expressed as means ± sem. Student's t test: *, P < 0.05 vs. wild-type controls; **, P < 0.0001 vs. wild-type controls.

Cited1-null mice have reduced β-cell mass and decreased circulating insulin levels

An alternative explanation for these findings is that decreased IR phosphorylation in the fetal liver results from reduced hepatic insulin signaling in mice with IUGR. Moreover, there is an increased ratio of Ir-B vs. Ir-A IR α-chain splice variants in the E18.5 fetal livers vs. lungs (Supplemental Fig. 3). Because IR-B is the specific high-affinity receptor for insulin, whereas the IR-A variant is activated by insulin and IGF-II (41), this suggests that the fetal liver is primed to respond to changes in insulin and not IGF-II levels in mice with IUGR. Indeed, there is a profound reduction of circulating insulin levels associated with reduced serum glucose in Cited1 mutant mice with IUGR (Fig. 7, A and B). Because maternal insulin does not cross the placenta (42, 43), this indicates that reduced circulating insulin results from decreased insulin production by the fetal pancreas. Moreover, both absolute pancreatic weight and pancreas to body weight ratios are significantly reduced in E18.5 Cited1 mutant mice [mean ± sd was 5.93 ± 0.71 mg in wild-type males (n = 16) vs. 4.12 ± 1.07 mg Cited 1-null males (n = 11), t test, P < 0.0001, Fig. 7C]. This is associated with a significant reduction in pancreatic β-cell mass (Fig. 7, D and E). These findings suggest that decreased circulating insulin results from a deficiency in insulin-producing β-cells in fetal pancreata of mice with IUGR and that reduced IR signaling in livers could be regulated by decreased insulin production from pancreas.

Fig. 7.

Pancreatic β-cell mass and circulating insulin levels in Cited1-null embryos. A and B, Serum insulin and glucose levels were determined in multiple pooled samples from E18.5 wild-type males and Cited1^Y/− mutant embryos (n = 12 and 13 pools of four to six embryos, respectively). Insulin levels are presented in nanograms per milliliter, glucose in milligrams per deciliter. C, Pancreas to body weight ratio was determined in E18.5 wild-type males and Cited1^Y/− mutant embryos (n = 16 and 11 embryos, respectively). D, Representative images of insulin staining (black) in E18.5 wild-type and Cited1-null male pancreata as indicated. E, β-cell mass [calculated from total insulin staining area/total pancreatic surface area in pixels (derived from multiple sections through each pancreas) × pancreatic weight in milligrams] from E18.5 wild-type and Cited1-null embryos (n = 6 and 5, respectively). Results are expressed as mean ± sem. Student's t test: *, P < 0.0001; **, P < 0.05 vs. wild-type controls.

Discussion

We show that Cited1-null mice with placental insufficiency have classic features of asymmetric IUGR with reduced fetal liver and lung size, preservation of fetal brain mass, and accelerated postnatal catch-up growth. Absolute kidney weights are also reduced in these mice but mirror the global reduction in body mass, indicating that kidney growth is less severely affected in this model of IUGR. We use a genetic approach to demonstrate that these effects result from loss of CITED1 expression in the placenta and not in the embryo proper, and therefore establish Cited1 mutant mice as the first bona fide mouse model of asymmetric IUGR resulting from late gestational placental insufficiency. Using this model, we demonstrate for the first time that IUGR results in decreased activating phosphorylation of growth-promoting IGF and IR in livers, lungs, and kidneys but not embryonic brains. This suggests that preservation of IGF-IR/IR signaling may play a role in maintaining brain growth in asymmetric IUGR. Decreased IGF-IR and/or IR phosphorylation is associated with reduced expression of IGF-I and IGF-II ligands in kidneys and lungs, respectively, whereas decreased IR signaling in the liver is associated with reduced insulin production by the fetal pancreas. These findings suggest that different mechanisms regulate IGF-IR and IR signaling in organs subject to asymmetric IUGR and that hepatic growth may be regulated by alterations in pancreatic insulin production in IUGR.

A number of experimental models have been developed to study the effects of late gestational placental insufficiency on fetal growth and adult disease (44). However, fetal effects in these models are sensitive to subtle changes in the timing, duration, and the severity of placental insufficiency, and it has been challenging to reproduce the same level of IUGR in small animals using these models. Moreover, the effects of restricting uterine blood flow are dependent on the position of the fetus within the uterine horn (45), introducing inherent variability to the most commonly used models of placental insufficiency. Technical challenges have also restricted use of these models to larger rodents, sheep, and pigs, and none of the models have been established in mice. This has limited our ability to exploit the power of mouse genetics to investigate the functional mechanisms regulating fetal responses to late gestational placental insufficiency. Our studies establish the first mouse model of late gestational placental insufficiency with classic features of asymmetric IUGR. High perinatal mortality (∼80% Cited1-null mice die in the early postnatal period) (28) reflects the severity of IUGR in this model, although the degree of growth retardation is comparable to that described in other models of IUGR. For example, although Cited1 mutant mice with placental insufficiency show a 15–32% decrease in body weight, uterine artery ligation in rats leads to a comparable 15–30% decrease of body weight in affected offspring (45, 46). However, as a genetic model of placental insufficiency, technical challenges are limited to animal husbandry and genotyping. Because placental insufficiency in Cited1 mutant mice is dependent on the mice being maintained on a C57BL/6 background (28), careful attention to strain background is essential. However, when maintained on a pure C57BL/6 background, we found little evidence of phenotypic variability between pups, and we did not detect differences dependent on the pup position within the uterine horn (T.N., data not shown), indicating that this model has significant advantages over other commonly used models of placental insufficiency.

We have exploited this model to provide the first comprehensive analysis of organ-specific changes in IGF-I, IGF-II, and insulin expression levels associated with IUGR and have linked these changes to alterations in IGF-IR and IR signaling in organs affected by asymmetric IUGR. This is of importance because the activity of secreted IGF ligands is regulated by a variety of IGF-binding proteins in vivo (26), and IGF-binding proteins are also abnormally regulated in different models of IUGR (18–21). Therefore, our studies establish that alterations in tissue expression of IGF and insulin are associated with decreased activation of their cognate receptors in fetal organs affected by IUGR. Our studies show that whereas there is reduced activating IGF-IR and/or IR phosphorylation in organs affected by IUGR, there is no significant change in IGF-IR or IR signaling in the brain. Because fetal brain growth is unaffected in asymmetric IUGR, these findings are consistent with the hypothesis that alterations in IGF-IR/IR signaling mediate fetal effects of placental insufficiency on organ growth. Our studies also provide the first evidence that alterations in IGF-IR and IR signaling associated with IUGR are organ specific and associated with selective changes in IGF-IR and/or IR ligand expression. For example, there is IGF-IR phosphorylation in the kidney and brain, whereas the embryonic lung shows dominant activating phosphorylation of IR. Reduced pIGF-IR is associated with decreased IGF-I expression in kidneys, whereas decreased pIR is associated with a marked reduction in IGF-II expression in IUGR lungs. In contrast, IGF-IR phosphorylation in the brain is unaffected by IUGR and is associated with increased expression of IGF-I. Taken together, these findings suggest that reduced IGF-I in kidneys and IGF-II in lungs account for decreased IGF-IR and IR signaling in the respective organs subject to IUGR. These findings are consistent with earlier studies showing that fetal lungs express high levels of IGF-II (47, 48). Moreover, Igf2-null mice have defects in late gestational fetal lung maturation (49), suggesting that reduced autocrine IGF-II/IR signaling could play a role in regulating late gestational lung growth in IUGR. However, the pattern of IGF-IR/IR signaling in kidneys and lungs is complicated because we also saw evidence of activated IGF-IR signaling in lungs and IR signaling in kidneys. Moreover, although there is no significant reduction in pIGF-IR expression in IUGR lungs, there is reduced IR phosphorylation in IUGR kidneys. Because this is not associated with reduced renal IGF-II expression, the mechanisms regulating IR phosphorylation in the IUGR kidney remain to be established. However, reduced circulating insulin in Cited1 mutant mice could account for the effect on IR signaling in the IUGR kidney.

In contrast to the changes seen in IUGR lungs, decreased pIR expression in the livers of mice with placental insufficiency is associated with reduced IGF-I but not IGF-II expression. This is associated with a relative decrease in the nonsignaling IGF-II clearance receptor, IGFR2, indicating that if anything, autocrine IGF-II/IR signaling ought to be increased in livers of mice with IUGR. Therefore, although alterations in hepatic IGF-I secretion may regulate IGF-IR signaling in other organs, because IGF-I does not activate IR (23), IR signaling is unlikely to be regulated by alterations in IGF-II-dependent IR signaling in IUGR livers. However, the liver is uniquely positioned to respond to fetal pancreatic insulin secreted into the portal circulation, so that reduced embryonic insulin production could have a dominant influence on hepatic IR signaling. This is consistent with our finding that fetal livers express the specific high-affinity IR isoform IR-B, which is not activated by IGF-II (41).

To determine whether insulin signaling was decreased in mice with IUGR, we evaluated serum insulin and pancreatic β-cell mass in Cited1-null embryos. Consistent with previous reports in sheep and rats with late gestational placental insufficiency (42, 50), our studies identified a 3-fold reduction in serum insulin and 40% reduction in serum glucose in Cited1 mutant mice with IUGR. Because maternal insulin does not cross the placenta (42, 43), these findings suggest that reduced circulating insulin results from decreased insulin production by the fetal pancreas in response to reduced glucose delivery across the placenta. However, chronic fetal hypoglycemia that occurs in other models of IUGR is associated with reduced β-cell mass (50, 51). Moreover, hypoglycemia decreases β-cell differentiation in ex vivo rat embryonic pancreatic cultures (52), suggesting that β-cell mass could be directly regulated by decreased maternal glucose delivery to fetuses subject to late gestational placental insufficiency. This suggests that there may be additional effects on fetal β-cell development in Cited1 mutant mice with IUGR. Indeed, our studies show that Cited1-null mice have significantly reduced pancreatic β-cell mass, indicating that persistent down-regulation of insulin production by the fetal pancreas is associated with decreased β-cell numbers. This is consistent with previous reports indicating that IUGR results in decreased β-cell mass and insulin secretion (50) and that IUGR is associated with increased risk of developing diabetes in adult life (53). In addition, fetal insulin production increases dramatically during late gestation (54, 55), and mice with combined knockout of the two nonallelic insulin genes (Ins1 and Ins2) are also subject to late gestational IUGR (56). This indicates that fetal insulin regulates fetal growth and, along with our findings on IR signaling in the fetal liver, suggests there is a unique relationship between pancreatic and liver growth defects in IUGR.

Our studies show that asymmetric IUGR is associated with organ-specific alterations of IGF-IR and IR signaling that may be regulated by selective changes in IGF-I, IGF-II, and/or insulin expression. However, apart from pancreatic insulin production, which has been directly linked to the reduced maternal substrate delivery and fetal β-cell mass in IUGR (50–52), the mechanisms by which IGF-I and IGF-II ligands are differentially regulated by IUGR in liver, lung, and kidney are unknown. These alterations may be regulated directly by alterations in substrate delivery (nutrients and/or oxygen) resulting from late gestational placental insufficiency. However, although preferential shunting of blood toward the brain is thought to account for relative preservation of brain growth during late gestational IUGR (4), there is no evidence that blood supply to other organs is differentially regulated in IUGR fetuses. Therefore, because both Igf1 and Igf2 genes are complex and regulated by multiple, often distant and tissue-specific regulatory elements (57–59), it is likely that organ-specific regulation of IGF-I and IGF-II in late gestational placental insufficiency occurs because these genes have tissue-specific regulatory elements that respond differently in each organ to nutrient and/or oxygen depletion.

In conclusion, we have shown that placental insufficiency caused by deletion of Cited1 in placenta leads to asymmetric IUGR with growth restriction of liver, kidneys, and lungs and preservation of brain growth. This is associated with organ-specific alterations of IGF-IR and IR signaling along with selective changes in IGF-I, IGF-II, and insulin expression that could account for the effects of IUGR on organ growth. We propose that Cited1 mutant mice serve as a new genetic model that will enable us to explore the role of IGF and insulin signaling defects in the genesis of late gestational IUGR.