The importance of placental circulation is exemplified by the correlation of placental size and blood flow with fetal weight and survival during normal and compromised human pregnancies in such conditions as preeclampsia and intrauterine growth restriction (IUGR). Using noninvasive magnetic resonance imaging, we evaluated the role of PKBalpha/AKT1, a major mediator of angiogenesis, on placental vascular function. PKBalpha/AKT1 deficiency reduced maternal blood volume fraction without affecting the integrity of the fetomaternal blood barrier. In addition to angiogenesis, PKBalpha/AKT1 regulates additional processes related to survival and growth. In accordance with reports in adult mice, we demonstrated a role for PKBalpha/AKT1 in regulating chondrocyte organization in fetal long bones. Using tetraploid complementation experiments with PKBalpha/AKT1-expressing placentas, we found that although placental PKBalpha/AKT1 restored fetal survival, fetal PKBalpha/AKT1 regulated fetal size, because tetraploid complementation did not prevent intrauterine growth retardation. Histological examination of rescued fetuses showed reduced liver blood vessel and renal glomeruli capillary density in PKBalpha/Akt1 null fetuses, both of which were restored by tetraploid complementation. However, bone development was still impaired in tetraploid-rescued PKBalpha/Akt1 null fetuses. Although PKBalpha/AKT1-expressing placentas restored chondrocyte cell number in the hypertrophic layer of humeri, fetal PKBalpha/AKT1 was found to be necessary for chondrocyte columnar organization. Remarkably, a dose-dependent phenotype was exhibited for PKBalpha/AKT1 when examining PKBalpha/Akt1 heterozygous fetuses as well as those complemented by tetraploid placentas. The differential role of PKBalpha/AKT1 on mouse fetal survival and growth may shed light on its roles in human IUGR.
INTRODUCTION
In the mammalian placenta, respiratory gases, nutrients, and waste are exchanged between the maternal and fetal vasculature. From Embryonic Day 10.5 (E10.5), mouse fetal growth depends on the umbilical flow directing blood via the placenta, which includes three parts: placental labyrinth, spongiotrophoblast layer, and decidua basalis. Placental circulation is a critical determinant for fetal and placental size [1]. Reduced maternal placental blood flow is associated with early embryonic mortality, fetal growth retardation [2], and impaired neonatal survival and growth [3]. Etiologies of preeclampsia and intrauterine growth restriction (IUGR) are frequently associated with abnormalities in placental growth, structure, and function, eventually giving rise to decelerated fetal growth and subsequent infant mortality and morbidity [4].
Angiogenesis refers to the formation of a new vascular bed, and it is a critical process for normal tissue growth and development [5]. Placental angiogenesis is a major determinant in the increase of fetal placental blood flow throughout gestation [1]. Vascular endothelial growth factors (VEGFs), representing a major class of placental angiogenic factors, stimulate angiogenic processes [6]. Trophoblasts are a rich source of angiogenic growth factors, such as VEGFA, which directs the growth of maternal blood vessels towards the embryonic implantation site [7]. The maternal circulation in the placenta involves vascular mimicry by fetal trophoblast cells, which respond to the angiogenic signals as well.
PKB/AKT1 acts downstream of VEGFA via the VEGF receptor 2/PIK3/PKB signaling cascade, known to mediate the formation of new blood vessels [8]. In addition to angiogenesis, the three isoforms of PKB/AKT (1, 2, and 3) regulate many other cellular and physiological processes, such as glucose metabolism, transcription, and cell cycle regulation and survival. PIK3 and PKB/AKT are expressed and functional from the one-cell stage of the mouse preimplantation embryo, and, specifically, PKB can be detected in the inner cell mass (ICM) and the trophoblast cells [9, 10]. PKBα/AKT1 (also known as AKT1; one of three known PKB isoforms) was found to be present in all types of trophoblast cells and vascular endothelial cells [11].
Placentas of PKBα/Akt1 null (−/−) fetuses were shown to display decreased vascularization and significant hypotrophy, with marked reduction of the decidua basalis [11]. PKBα/Akt1−/− mice were found to be smaller, with increased perinatal mortality and disordered fetal vasculature [12, 13]. In human pregnancies, placentas of IUGR exhibit signs of oxidative stress, with reduced AKT signaling [14]; therefore, the PKBα/Akt1−/− mice were suggested as a model for the human IUGR. The persistence of the reduced size in adult PKBα/Akt1−/− mice was further suggested to be associated with bone mineralization defects characterized by decreased length and mass of the long bones [15, 16]. We recently reported that impaired endochondral bone growth in these mice was associated with decreased bone vascularization, which was significant also for mice lacking a single copy of PKBα/Akt1 [17]. From the evidence so far, it is clear that the absence of PKBα/AKT1 from the placenta and embryo results in substantial embryonic defects mainly characterized by fetal mortality and reduced size, and the latter persist after birth, possibly due to bone developmental defects. However, it is still unclear whether the embryonic defects are secondary to placental hypovascularity caused by lack of PKBα/AKT1 or due to its absence from the embryo proper.
The study reported here aimed at differentiating between the role of PKBα-associated placental deficiency and the role of fetal PKBα/AKT1 in fetal survival and development. We used noninvasive macromolecular dynamic contrast enhanced (DCE) magnetic resonance imaging (MRI) to evaluate placental vascular functionality in PKBα/Akt1−/− fetuses. In contrast with histology, which allows quantification of vessel number and morphology, this MRI methodology allows the assessment of vessel functionality, as shown in our previous work on deciduas at embryo implantation sites [18]. Macromolecular DCE MRI was also applied to assess the maternal circulation in normal placental vasculature and in tetraploid placental complementation [19]. Tetraploid complementation/rescue is typically used to rescue embryonic lethality caused by defects in extraembryonic tissues like placenta. By doubling their ploidity at the two-cell stage, tetraploid embryos are able to contribute to the extraembryonic tissue and complement diploid embryos upon aggregation with their ICM [20]. Tetraploid-aggregated embryos will give rise to the placenta, whereas the ICM will give rise to the embryo proper. Tetraploid rescue was used here to complement PKBα/Akt1−/− embryos with PKBα-expressing placentas, allowing us to differentiate between the roles of placental versus fetal PKBα/AKT1-expressing placentas in determining fetal survival and size.
MATERIALS AND METHODS
Embryo Transfer and Tetraploid Complementation/Rescue
Embryo collection and culture.
All animal experiments were approved by the Weizmann Institutional Animal Care and Use Committee. Mice were maintained on a 12L:12D (light from 0600–1800 h) cycle. For superovulation, 3-wk-old ICR females were superovulated by i.p. injection of 5 IU of equine chorionic gonadotropin at 1300 h and 5 IU of human chorionic gonadotropin 46 h later. Females were then mated with B5/EGFP males [21], examined for vaginal plugs the following morning (defined as E0.5; 0.5 days postcoitum [dpc]), and killed at 46–48 h after human chorionic gonadotropin (1.5 dpc) to collect late two-cell embryos by oviduct flushing. These were later used to produce tetraploid embryos by electrofusion. B6(C57BL/6J)/PKBα/Akt1 heterozygote (+/−) females (5–6 wk old) were superovulated as described above, mated with B6/PKBα/Akt1 knockout (−/−) males, and killed at 2.5 dpc to isolate eight-cell embryos for aggregation with the tetraploid embryos.
Immunosurgery.
To minimize the contribution of the B6 origin (of the PKBα/Akt1−/− embryos) to the tetraploid placenta, the trophectoderm layer of the blastocysts was lysed and removed by immunosurgery as previously described [22], thus isolating the ICM. Briefly, blastocysts were exposed to rabbit anti-mouse lymphocyte serum (1:20 in Dulbecco modified Eagle medium) for 30 min at 37°C, washed with media, and transferred to guinea pig serum (1:2, as complement) for 10 min. Gentle pipettation allowed the final separation of the ICM from the lysed trophectoderm.
Production of tetraploid embryos.
Electrofusion was done using a CF-150 pulse generator equipped with a 250-μm electrode chamber (BLS Ltd.). Electrodes were covered by a drop of 3M mannitol (Sigma). Two-cell embryos were placed perpendicular to and between the electrodes, and two pulses of 30 V were delivered during 40 ms (AC field between 1.3 and 1.5 V). Following the pulse, embryos were transferred to the incubator for the fusion to take place within 30–40 min (with more than 90% efficiency). Fused embryos were separated and cultured overnight in KSO media [23], and after 24 h, compacting four-cell-stage embryos were used for aggregation.
Aggregation of tetraploid embryos and isolated ICM.
A single isolated (diploid) ICM was sandwiched between two tetraploid embryos (after removal of the zona pellucida with acidic Tyrode solution; Sigma) in an aggregation plate. After 24 h of incubation, blastocysts were transferred to the uteri of 2.5-dpc pseudopregnant ICR females, 8–10 blastocysts per uterine horn. As a control, embryo transfer only with no tetraploid complementation was used: E3.5 blastocysts derived from PKBα/Akt1+/− females mated with either PKBα/Akt1+/− or PKBα/Akt1−/− males (without mixing of blastocysts from the different matings prior to transfer) were transferred to 2.5-dpc pseudopregnant ICR females, 8–10 blastocysts per uterine horn [23].
In Vivo Contrast Enhanced MRI Studies
Female ICR mice (12 wk old) carrying transferred embryos (ICR; B6/PKBα/Akt1 wild type [PKBα/Akt1+/+]; PKBα/Akt1+/− or PKBα/Akt1−/− with native or tetraploid placentas) were analyzed by MRI on E18.5 of pregnancy. Images were taken before and sequentially until 13.5 min after i.v. tail vein injection of the macromolecular contrast agent, biotin-bovine serum albumin-GdDTPA [18, 19, 24, 25]. (See Supplemental Materials and Methods for full details; all Supplemental Data are available online at www.biolreprod.org.)
Histology, Histochemistry, and Fluorescence Microscopy
Bovine serum albumin labeled with rhodamine (BSA-ROX) was i.v. injected 3–5 min prior to animal killing for histological detection of functional blood vessels, as reported previously [24]. Placentas and fetuses were retrieved for further analysis. Macro images of ex vivo fetuses and placentas were achieved using regular and fluorescent light (Olympus SZX12 microscope equipped with DP50 camera), and then samples were fixed (fetuses in 4% paraformaldehyde and placentas in Carnoy mixture), embedded in paraffin, sectioned serially at 4-μm thickness, and stained with hematoxylin and eosin (H&E) and for biotin-BSA-GdDTPA (using fluorescein-labeled avidin [avidin-FITC]; Molecular Probes, San Francisco, CA), as previously described [26]. See details for the morphometric analysis in the Supplemental Materials and Methods.
Statistical Analysis
Statistical analysis accounted for the mixed genotypes of the litters (resulting from matings of PKBα/Akt1+/− females mated with either PKBα/Akt1+/− or PKBα/Akt1−/− males). A two-sample, two-tail t-test ± SEM was applied for the analysis of significance of the MRI, histological, and ex vivo data. The data were considered significant for P < 0.05.
RESULTS
PKBα/AKT1-Deficient Placentas Exhibited Impaired Vascular Function with Maintained Integrity of the Fetal/Maternal Blood Barrier
The DCE MRI, using biotin-BSA-GdDTPA as a contrast agent (i.v.), was used for visualizing the maternal circulation of E18.5 placentas in utero (Fig. 1a). The distribution of the albumin-based contrast media was validated using fluorescently labeled albumin, BSA-ROX (Fig. 1b). Both biotin-BSA-GdDTPA and BSA-ROX did not extravasate to the fetal side, and thus the MRI contrast enhancement exclusively revealed the maternal circulatory bed in the placenta.
FIG. 1.
PKBα/AKT1-deficient placentas exhibited impaired vascular function with maintained integrity of the fetal-maternal blood barrier. a) Representative MR image (7.5 min after i.v. administration of biotin-BSA-GdDTPA) shows signal enhancement in E18.5 placentas (lighter) versus the adjacent nonenhanced embryos (darker). This is the result of the accumulation of the contrast material, biotin-BSA-GdDTPA, which was injected into the maternal circulation and was excluded from the fetal vasculature. The enhanced maternal-placental circulation can also be exhibited in the enlarged MR images of placentas. p1, cross section; p2, longitudinal section; VC, vena cava. Bar = 1 cm. b) The exclusion of the BSA-based contrast material from the fetal vasculature was verified using BSA-ROX. The BSA-ROX fluorescence is shown in macro views of placentas from the side of the umbilical cord connection (upper row) and the side of the decidual connection (lower row), both in bright field (left) and fluorescence (right). Arrowhead indicates embryo paw; uc (arrow), umbilical cord; and lab, maternal vessels with BSA-ROX in the placental labyrinth. Bar = 3 mm. c) Early-injected biotin-BSA-GdDTPA (stained with avidin-FITC, green; localized more to the edges of the labyrinth or vessel rim, arrowheads) and late-injected BSA-ROX (red; arrows indicate localized within vessels) are both confined to the maternal circulation in the placental labyrinth, as exhibited by histological analysis. Asterisks in the insets mark the fetal blood spaces empty of both early- and late-injected contrast materials. Colocalization of the early- and late-injected contrast materials appears in yellow. Hoechst (blue)-stained nuclei (lab, labyrinth layer). Bar = 100 μm; inset original magnification ×400. d) Representative sections of placental midtransverse histological sections showing the circulatory bed of the placental labyrinths (black arrows in insets indicate circulatory bed vessels). Bar = 25 μm; inset original magnification ×200. e) Placental initial enhancement of PKBα/Akt1−/− and PKBα/Akt1+/− calculated using the MRI data were significantly reduced versus enhancement of PKBα/Akt1+/+ placentas (PKBα/Akt1+/+: n = 11 in four dams; PKBα/Akt1+/−: n = 5 in three dams, *P = 0.007; PKBα/Akt1−/−: n = 11 in 4 dams, *P = 0.02). f and g) A morphometric analysis of the placental circulatory bed (n = 3 placentas each, three sections per placenta) verified the significant reduction in circulatory bed area (*P = 0.005) and length (*P = 9.6 × 10−6) of PKBα/Akt1−/− placentas. PKBα/Akt1+/− circulatory bed area was similar to PKBα/ Akt1−/− (*P = 0.01), whereas circulatory bed length was intermediate (*0.009 > P > 0.015).
As a central downstream target of VEGF receptor 2, PKBα/AKT1 plays an important role in mediating VEGFA-induced vascular permeability [27]. In the placenta, vascular permeability could be suppressed by heterodimers of VEGFA and placental growth factor [28]. Previous studies reported that vascular permeability was elevated for PKBα/Akt1 null mice upon induction of angiogenesis [28]. Therefore, we examined the impact of PKBα/AKT1 deficiency on the integrity of the fetomaternal blood barrier. Histological analysis of early-injected biotin-BSA-GdDTPA (stained with avidin-FITC; green) and late-injected BSA-ROX (red) revealed that they are both confined to the maternal circulatory bed of the placental labyrinth in PKBα/Akt1+/+, PKBα/Akt1+/−, and PKBα/Akt1−/− placentas, and did not extravasate to the fetoplacental vasculature (Fig. 1c; asterisks within insets illustrate the fetoplacental spaces void of either contrast material injected). Therefore, PKBα/Akt1−/− placentas, as well as PKBα/Akt1+/− placentas, maintain the integrity of the fetoplacental barrier (with no interstitial accumulation of the contrast material), similar to wild-type PKBα/AKT-expressing placentas [19]. Although the early- and late-injected contrast materials mostly colocalize (yellow; Fig. 1c), early-injected biotin-BSA-GdDTPA is localized more at the edges of the placental labyrinth or vessel rim (arrowheads; Fig. 1c), whereas the late-injected BSA-ROX filled most of the blood vessels within the labyrinth itself (white arrows; Fig. 1c). This spatial mismatch visualized by the dual-labeling approach revealed slow dynamics of inflow and clearance of maternal blood from the placental blood pool at E18.5, as well as accumulation of albumin toward the end of pregnancy, with possible uptake by various trophoblast cells [19].
The contribution of PKBα/AKT1 to the vascular functionality of the placenta was evaluated by MRI (Fig. 1e) and by histological morphometric analysis (Fig. 1, d, f, and g). The initial enhancement measured from the MRI data (analogous to the late-injected BSA-ROX; Fig. 1c) corresponds to the blood volume entering the maternal-placental blood spaces (Supplemental Fig. S1). The ability of the placenta to deliver sufficient blood supply is a prerequisite to its proper functionality. The initial enhancement of PKBα/Akt1−/− and, interestingly, PKBα/Akt1+/− placentas was significantly reduced compared with the enhancement of PKBα/Akt1+/+ placentas (Fig. 1e). Morphometric analysis of the placental circulatory bed (i.e., placental labyrinth, includes both maternal and fetal blood spaces) also revealed a significant reduction in the circulatory bed area and length of blood vessels in PKBα/Akt1−/− versus normal placentas (Fig. 1, f and g). PKBα/Akt1+/− revealed an overall intermediate phenotype: the circulatory bed area was similar to PKBα/Akt1−/− (corresponding to the initial enhancement measured by MRI), whereas the circulatory bed length was intermediate between PKBα/Akt1+/+ and PKBα/Akt1−/−.
Osteopenia in PKBα/Akt1-Deficient Mice Originates Prenatally and Persists after Birth
Determining the functional placental insufficiency caused by the absence of two copies, or even a single copy, of the PKBα/Akt1 gene led us to examine the effect that the absence of PKBα/AKT1 has on the fetal bone phenotype. The attempt was made to trace down bone defects reported for adult PKBα/Akt1−/− mice [15, 16] to E18.5 fetuses. Long bones were characterized in fetuses and neonates of various ages (E18.5–Postnatal Day 40; Fig. 2). Morphological abnormalities in the growth plate cartilage were observed in PKBα/Akt1−/− and PKBα/Akt1+/− mice, and were most profound at E18.5, and their severity diminished with age. The growth plate cartilage was narrowed because of a reduction in the width and disturbed organization of the proliferating and hypertrophic zones (Fig. 2a, between lines). A gradation in severity was observed between the PKBα/Akt1−/− and PKBα/Akt1+/− mice, where PKBα/Akt1+/− appeared to be intermediate between PKBα/Akt1−/− and PKBα/Akt1+/+ mice (Fig. 2, a–d). Furthermore, the columnar organization of the hypertrophic chondrocytes was less evident, and there was significant heterogeneity in their size (anisocytosis; Fig. 2c, box). Endochondral bone production was reduced in the PKBα/Akt1−/− and, to a lesser extent, in PKBα/Akt1+/− mice (Fig. 2d, asterisk). Therefore, it seemed that the bone defects found in the PKBα/Akt1−/− E18.5 fetuses were more severe than in the adult mice. Moreover, mild bone defects were also found in PKBα/Akt1+/− fetuses.
FIG. 2.
Osteopenia in PKBα/AKT1-deficient mice originates prenatally and persists after birth. Histological sections of humeri of (a) E18.5 fetuses as well as (b) Day 7 (c) Day 22, and (d) Day 40 neonates showing morphological abnormalities in the growth plate cartilage. The growth plate cartilage was narrowed because of disturbed organization of the proliferating and hypertrophic zones (a, between lines). Heterozygous PKBα/Akt1+/− mice appeared intermediate between PKBα/Akt1−/− and PKBα/Akt1+/+ mice (a–d). Furthermore, the columnar organization of the hypertrophic chondrocytes was less evident, and there was significant heterogeneity in their size (anisocytosis; boxed areas [c]). Endochondral bone production was reduced in PKBα/ Akt1−/− and, to a lesser extent, in PKBα/Akt1+/− mice (d, asterisks). e) Magnification of the hypertrophic layer of Postnatal Day 7 (arrowheads indicate hypertrophic chondrocytes; arrows, columnar organization of chondrocytes). Bars = 100 μm (a) and 50 μm (b–e).
Vascular Function in the Placentas of PKBα/AKT1-Deficient Fetuses Can Be Restored by Tetraploid Rescue
Previously reported perinatal mortality [12, 13], together with our conclusions concerning the placental functional insufficiency and the fetal bone defects caused by PKBα/Akt1 gene deficiency, emphasized the need to distinguish between the role of placental and fetal PKBα/AKT1 in determining fetal survival and size. Moreover, the greater severity of the bone phenotype in fetuses versus older mice deficient in PKBα/AKT1 also justified checking for a putative role for the placental functional insufficiency in this phenotype. Therefore, we decided to study survival and size in tetraploid-rescued fetuses, in which tetraploid placentas originated from GFP-expressing (PKBα/Akt1+/+) transgenic mice. We first wanted to examine the vascular functionality of tetraploid placentas serving either PKBα−/− or PKBα+/− fetuses versus normal, PKBα-expressing C57BL/6J+/+ placentas. The ICMs of PKBα−/− and PKBα+/− (C57BL/6J background) blastocysts were aggregated with the GFP-expressing tetraploid embryos and transplanted into 2.5-dpc ICR pseudopregnant mice. Embryos complemented with tetraploid placentas were identified by placental GFP fluorescence, with no fluorescence in the embryo (Fig. 3, a and b; control, nonaggregated diploid GFP embryos show fluorescence of both the placenta and embryo, shown in Fig. 3b, inset). The tetraploid placental vasculature was then visualized and studied by MRI (Fig. 3, c and d). Surface projections of the enhanced areas in the three-dimensional MRI data illustrate how this method resolved the functional blood vessels of the tetraploid placenta along with blood vessels in other maternal internal organs (Fig. 3, e and f). The tetraploid placentas serving either PKBα/Akt1−/− or PKBα/Akt1+/− PKBα/Akt1 fetuses were not significantly different from nontetraploid C57BL/6J+/+ placentas in both their initial enhancement (Fig. 3g) and size (Fig. 3h). Thus, tetraploid complementation restored normal placental vascular function.
FIG. 3.
Vascular function in the placentas of PKBα/AKT1-deficient fetuses can be restored by tetraploid rescue. a and c) E18.5 PKBα/Akt1+/− fetus (nonfluorescent) with tetraploid GFP placenta versus (b and d) PKBα/Akt1−/− fetus (nonfluorescent) with tetraploid GFP placenta transplanted into 2.5-dpc ICR pseudopregnant mice. a and b) Fluorescence microscopy depicting the GFP placentas (green), and BSA-ROX as a vascular marker (red), in contrast to the nonfluorescent aggregated embryos. Nonaggregated diploid GFP embryos (inset in b; GFP embryo and placenta). Chromatic light was also applied to show the nonfluorescent fetuses. c and d) Three-dimensional gradient echo maximal intensity projection of PKBα/Akt1+/− fetus (c) and PKBα/Akt1−/− fetus (d), both with GFP tetraploid placentas. Inset shows enlarged image of the maternoplacental vascular bed, with an arrow pointing at the uterine branch of ovarian artery and vein (ub Oa&v). Note the functional blood vessels of the tetraploid placenta in a and c, also in frontal (e) and dorsal (f) three-dimensional surface projections of enhanced biotin-BSA-GdDTPA blood vessels. p, Placenta; ov, ovary; k, kidney; vc, vena cava; b, bladder. g) Placental initial enhancement (arbitrary units) and (h) placental size (milliliters) of the tetraploid placentas serving either hetero (tetra PKBα/Akt1+/−) or null (tetra PKBα/Akt1−/−) fetuses were not different from C57BL/6J PKBα/Akt1+/+ placentas. Bars = 4 mm (a and b) and 1 cm (c–f).
Tetraploid Placentas Rescue PKBα/Akt1 Null Fetuses from Death In Utero but Do Not Prevent Fetal Growth Retardation
Normally functioning tetraploid placentas were used in an attempt to rescue the PKBα/Akt1−/− phenotype of reduced survival and size. The weight of PKBα/Akt1−/− placentas was found to be significantly reduced compared with that of PKBα/Akt1+/−, PKBα/Akt1+/+, and tetraploid placentas (Fig. 4a). Despite the improved placental vascular function, the weight of PKBα/Akt1−/− fetuses complemented with tetraploid placentas remained significantly lower than in PKBα/Akt1+/− and PKBα/Akt1+/+ fetuses and was not different from nonrescued PKBα/Akt1−/− fetuses (P = 0.8; Fig. 4b). Gene dosage was found to be important, because PKBα/Akt1+/− fetuses were also significantly smaller than PKBα/Akt1+/+ fetuses. The intermediate phenotype of PKBα/Akt1+/− placental vascularity (Fig. 1) was less pronounced by placental weight because, although slightly reduced, no significant difference in placental weight was found between PKBα/Akt1+/− and PKBα/Akt1+/+ placentas.
FIG. 4.
Tetraploid placentas rescue PKBα/Akt1 null fetuses from death in utero but do not prevent fetal growth retardation. a) The weight of PKBα/Akt1−/− placentas is significantly lower than that of PKBα/Akt1+/− (*P = 0.001), PKBα/Akt1+/+ (*P = 0.003), and tetraploid placentas (*P = 0.016; serving PKBα/Akt1 null fetuses, i.e., rescued PKBα/Akt1−/−; PKBα/Akt1−/−: n = 6 in four dams, PKBα/Akt1+/−: n = 22 in 14 dams, PKBα/Akt1+/+: n = 13 in four dams tetraploid (rescued PKBα/Akt1−/−): n = 5 in four dams). b) Fetal weight of PKBα/Akt1−/− fetuses is significantly lower than that of PKBα/Akt1+/− (*P = 0.046) and PKBα/Akt1+/+ (*P = 5.6 × 10−5) fetuses but not different from tetraploid rescued PKBα/Akt1−/− fetuses (P = 0.8). Also, PKBα/Akt1+/− fetuses are significantly smaller than PKBα/Akt1+/+ fetuses (*P = 5.6 × 10−5). c) Tetraploid placentas rescued PKBα/Akt1−/− fetuses from death in utero, because the fraction of PKBα/Akt1−/− nonrescued fetuses was significantly lower than that of rescued PKBα/Akt1−/− fetuses from total pups (PKBα/Akt1−/− and PKBα/Akt1+/− littermates) examined. Gray line indicates Mendelian frequency pups in natural heterozygous over homozygous mating (*P = 0.007). d) Of 296 mice genotyped at age 3 wk, 93 were PKBα/Akt1+/+ (31% instead of the expected 25%), 169 were PKBα/Akt1+/− (57%, but lower than the expected 186 relative to the PKBα/Akt1+/+ offspring), and only and 34 were PKBα/Akt1−/− (12%; significantly fewer than the expected 93).
Normally functioning tetraploid placentas were able to rescue PKBα/Akt1−/− embryos from fetal death in utero (Fig. 4c). The average number of total tetraploid-rescued fetuses per dam was two times lower than the average number of nonrescued fetuses that were subjected to embryo transfer only, because of the general low yield of surviving embryos after tetraploid rescue. However, all fetuses but one that were recovered from the uterus at E18.5 (after tetraploid rescue or embryo transfer only) were viable. Two E18.5 pups (PKBα/Akt1−/− and PKBα/Akt1+/−) were given to a foster mother, developed normally, and were fertile.
The frequency of PKBα/Akt1+/− and PKBα/Akt1−/− versus the PKBα/Akt1+/+ mice derived from routine PKBα/Akt1+/− mating provided further evidence to the mortality associated with PKBα/Akt1 deficiency (Fig. 4d). Moreover, because the PKBα/Akt1−/− fetuses suffer from death in utero, we had to use PKBα/Akt1+/− and PKBα/Akt1−/− matings for tetraploid rescue to increase the finite number of PKBα/Akt1−/− embryos for tetraploid aggregation. Therefore, the survival of PKBα/Akt1+/− mice, which served as our control in this experiment, had to be evaluated. Of 296 mice genotyped at age 3 wk, 93 were PKBα/Akt1+/+ (31% instead of the expected 25%), 169 were PKBα/Akt1+/− (57%, but lower than the expected 186 relative to the PKBα/Akt1+/+ offspring), and only 34 were PKBα/Akt1−/− (12%; significantly fewer than the expected 93). These results complement the study, because the intermediate phenotype of PKBα/Akt1+/− placentas correlated with the mild reduction in live PKBα/Akt1+/− pups. In PKBα/Akt1-deficient embryos, a significant reduction in placental vascularity and weight correlated with a severe reduction in live PKBα/Akt1−/− pups (with no rescue). These results are consistent with the rescue results, because a healthy placenta could rescue PKBα/Akt1−/− pups from death in utero but did not prevent their growth retardation.
Phenotypic Changes in Various Organs in PKBα/Akt1 Null Fetuses Versus Rescued PKBα/Akt1 Null Fetuses
To determine the reason for the yet smaller size of the rescued PKBα/Akt1−/− E18.5 fetuses, vital internal organs, such as liver and kidney from these fetuses, were histopathologically examined and subjected to morphometric analysis. As major size determinants, long bones (humeri) were examined as well. Blood vessel volume fraction in the liver (Fig. 5a) was significantly reduced in both PKBα/Akt1+/− and PKBα/Akt1−/− livers but normal in rescued PKBα/Akt1−/− and PKBα/Akt1+/− livers (Fig. 5b). The observation that the livers of PKBαAkt1−− and PKBα/Akt1+/− fetuses had reduced percentage of blood vessels is consistent with the observation made for these genotypes in the placenta, and both were rescued by the tetraploid placentas. However, examination of the hematopoietic islet volume fraction in the liver revealed that PKBαAkt1−− fetuses had a significantly reduced percentage of hematopoietic islets that were not rescued by the tetraploid placentas. The PKBα/Akt1+/− livers showed an intermediate phenotype (Fig. 5c).
FIG. 5.
Tetraploid normally functioning placentas rescue the vascular features of the liver and kidney of PKBα/AKT1-deficient fetuses. a) The H&E staining of liver from E18.5 PKBα/Akt1+/+ (n = 2 dams), nonrescued PKBα/Akt1+/− (n = 3 dams, one fetus each), nonrescued PKBα/Akt1−/− (n = 2 dams), rescued PKBα/Akt1+/− (n = 2 dams), and rescued PKBα/Akt1−/− (n = 3 dams; for all: one fetus from each dam). Green arrowhead in the first inset indicates blood vessel; yellow arrowhead, hematopoietic islet. Bar = 50 μm; inset original magnification ×200. b) Morphometric analysis of blood vessel volume fraction. *0.0067 > P > 5.1 × 10−8. c) Morphometric analysis of hematopoietic islet volume fraction. *0.05 > P > 0.003. d) The H&E staining of kidney from E18.5 PKBα/Akt1+/+, nonrescued PKBα/Akt1+/−, nonrescued PKBα/Akt1−/−, rescued PKBα/Akt1+/−, and rescued PKBα/Akt1−/−. (The number of fetuses in each category is the same as in a. Red arrowheads in first inset indicate glomeruli.) Bar = 100 μm; inset original magnification ×200. e) Morphometric analysis of glomeruli density. *0.05 > P > 3.9 × 10−5. f) Morphometric analysis of glomeruli size *1.3 × 10−5 > P > 4.3 × 10−8.
The renal glomeruli (capillaries that perform the first step of blood filtration to form urine) were also examined. Similar to the liver blood vessel phenotype, glomerular density was significantly reduced in both PKBα/Akt1−/− and PKBα/Akt1+/− fetuses (Fig. 5d) that were both rescued by the tetraploid placentas (Fig. 5e). Surprisingly, the size of the glomeruli in rescued PKBα/Akt1−/− was reduced compared with PKBα/Akt1+/+, PKBα/Akt1+/−, rescued PKBα/Akt1+/−, and PKBα/Akt1−/− fetuses (Fig. 5f). The latter is probably a secondary effect, because total kidney size was still small in the rescued PKBα/Akt1−/−, restricting the ability of the glomeruli to regain normal size as their density returned to normal.
The yet reduced size of the rescued PKBα/Akt1−/− fetuses can probably be mostly attributed to the phenotype exhibited in the bone. The hypertrophic layer of the growth plate demonstrated significantly reduced size in PKBα/Akt1−/− fetuses (Fig. 6, a and c). In the rescued fetuses, the hypertrophic layer was rescued in size and even significantly enlarged compared with its PKBα/Akt1+/+ equivalents. The rescue effect in the rescued PKBα/Akt1+/− was more subtle in accordance with the initial, more subtle effect in the nonrescued PKBα/Akt1+/−. In the rescued PKBα/Akt1−/− and PKBα/Akt1+/−, the columnar organization of the chondrocytes in the proliferative and hypertrophic zones remained disorganized (see also Fig. 2). Interestingly, examination of the mineralized bone revealed impaired bone mineralization in PKBα/Akt1+/− and PKBα/Akt1−/− fetuses and even further impaired bone mineralization in rescued PKBα/Akt1−/− (Fig. 6, a and d). The overshortened mineralized bone in the rescued PKBα/Akt1−/− directly depends on the normal development (proliferation and differentiation) of the chondrocytes. In the PKBα/Akt1−/− fetuses, the overgrown hypertrophic layer with highly disorganized chondrocytes may explain the overshortened mineralized bone phenotype. In PKBα/Akt1+/− mice, the effect on the hypertrophic layer is more subtle, and so is the effect on the mineralized bone. Overall, although normally functioning (PKBα-expressing) placenta governed chondrocyte cell number, fetal PKBα/AKT1 controlled chondrocyte columnar organization (all the phenotypical changes are summarized in Supplemental Table S1).
FIG. 6.
Tetraploid normally functioning placentas rescue fetal PKBα/AKT1 deficiency-related impaired proliferation but not columnar organization of chondrocytes. a) Alcian blue staining of humerus from E18.5 PKBα/Akt1+/+, nonrescued PKBα/Akt1−/+, nonrescued PKBα/Akt1−/−, rescued PKBα/Akt1+/−, and rescued PKBα/Akt1−/−. b) Calcium staining of humerus from E18.5 PKBα/Akt1+/+, nonrescued PKBα/Akt1−/+, nonrescued PKBα/Akt1−/−, rescued PKBα/Akt1+/−, and rescued PKBα/Akt1−/−. (The number of fetuses in each category in a and b is the same as in Fig. 5a) c) Morphometric analysis of growth plate length. *0.016 > P > 4.3 × 10−7. d) Morphometric analysis of bone mineralization length. *0.05 > P > 2 × 10−9. Bars = 100 μm (a) and 500 μm (b).
DISCUSSION
Tetraploid complementation allowed us to differentiate between the roles of placental versus fetal expression of PKBα/AKT1, suggesting differential contributions of PKBα/AKT1 to in utero embryonic development according to its physiological compartment. Specifically, we demonstrated that a normally functioning, PKBα/AKT1-expressing placenta is sufficient for regulating fetal survival in utero, whereas fetal expression of PKBα/AKT1 regulates fetal size, in a gene dosage-dependent manner. Despite these differences in the effects of placental versus fetal expression of PKBα/AKT1, it is clear that extensive fetal-placental interactive processes contribute to the outcome of any specific alteration of gene expression in either the placenta or the fetus. Thus, these results represent the integrated effects of direct as well as indirect consequences of the manipulation of placental expression of PKBα/AKT1 by tetraploid complementation.
Using DCE MRI, we were able to detect a significant contribution of placental PKBα/AKT1 not only to the vascular content but also to the functionality of the maternal blood flow in a PKBα/AKT1 dose-dependent manner. Although the integrity of the fetomaternal blood barrier was not compromised by the PKBα/AKT1 deficiency, PKBα/Akt1−/− and PKBα/Akt1+/− placentas were characterized by a marked reduction in the placental initial enhancement. Besides playing a role in compromised fetal survival, placental hypovascularity in the absence of maternal PKBα/AKT1 was previously suggested to influence fetal phenotype, because placentas and embryos from homozygous mating were found to be smaller than those from heterozygous mating [11]. On the other hand, growth retardation of PKBα/Akt1−/− mice was attributed to a direct effect on bone, which was recently reported in adult mice [15–17] and was also addressed here with morphological abnormalities in the growth plate cartilage in PKBα/Akt1−/− and PKBα/Akt1+/− E18.5 fetuses. Therefore, segregating between the role of placental and fetal PKBα/AKT1 deficiency in determining the phenotype was important. The DCE MRI showed that tetraploid placentas exhibited functional vascular characteristics similar to normal, healthy placentas, enabling us to monitor the impact of tetraploid rescue of PKBα/Akt1−/− embryos from placental insufficiency. Tetraploid placentas serving PKBα/Akt1−/− fetuses were sufficient to significantly reduce PKBα/Akt1−/− fetal mortality. However, the tetraploid placentas did not prevent PKBα- associated IUGR. Thus, expression of PKBα/AKT1 by the fetus is critical to normal fetal size, suggesting that growth retardation related to PKBα/AKT1 deficiency may be a direct effect and is not secondary to placenta hypovascularity. Transplantation of the tetraploid aggregates into pseudopregnant wild-type dams also allowed isolation of the fetoplacental and embryo proper components from the maternoplacental component. Rescuing PKBα/Akt1−/− embryos from death in utero by tetraploid complementation suggests that defective placental formation causing the phenotype of impaired survival is a primary trophoblast defect and not due to defective allantoic blood vessels (the latter are ICM derived). Disruption of genes involved in extraembryonic allantoic-vitelline vasculature development frequently present a similar phenotype but cannot be rescued by tetraploid complementation [29]. Moreover, the yolk sac vasculature in the case of PKBα/AKT1 deficiency is apparently normal, because in our study, the yolk sac, which is both ICM derived and trophectoderm derived, was occasionally GFP positive after tetraploid rescue, but the umbilical cord and yolk sac vessels were never GFP positive.
The placental complementation experiments elucidated a role for placental PKBα/AKT1 in fetal hepatic blood vessel development and renal glomerular capillary density, in a PKBα/AKT1 dose-dependent manner. One hypothesis could be that by restoration of the impaired blood pressure and/or growth factors, transport caused by the underdeveloped placental circulatory bed in the absence of PKBα/AKT1 affected the development of the vascular components in those major organs. Liver is also a major site for hematopoiesis in embryos, and it was previously shown that knockdown of PKB/AKT1 activity significantly inhibits fetal liver-derived erythroid cell colony formation and gene expression [30]. The allantoic mesoderm in the placenta, which is a putative source for hematopoietic stem cells in the liver [31], does not originate from the PKBα-expressing tetraploid trophectoderm, but rather from PKBα/AKT1−/− ICM [23]. This may explain why hematopoiesis in the liver was not rescued. The most prominent feature underlying the yet reduced size of the rescued PKBα/AKT1 fetuses is the impaired development of long bones. Interestingly, although normally functioning (PKBα-expressing) placentas governed chondrocyte cell number in fetal humeri, fetal PKBα/AKT1 controlled chondrocyte columnar organization, a prerequisite for chondrocyte differentiation and ossification. The rescue of chondrocyte cell number could have resulted from an adequate transport of growth factors by the tetraploid placenta, such as parathyroid hormone-related protein (PTHrP, official symbol PTHLH). PTHLH has profound actions on the growth and differentiation of bones [32]. Within the placenta, PTHLH is expressed by placental syncytial trophoblasts and will be transported to the fetus [33]. In the absence of PTHLH, chondrocytes do not proliferate normally, giving rise to shortened bone [32]. Such a mechanism could possibly explain the rescue of chondrocyte cell number in PKBα/Akt1 null rescued fetuses. Terminal chondrocyte differentiation featuring adequate columnar organization of chondrocytes that was not rescued by the tetraploid placenta, was previously shown to be directly regulated by AKT, both in embryonic and adult chondrogenesis. In transgenic embryos that expressed constitutively active or dominant-negative AKT in chondrocytes, AKT was found to positively regulate all processes of chondrocyte maturation [34]. Among the AKT isoforms, PKBα/AKT1 was the most highly expressed in chondrocytes [35], and a delayed secondary ossification in the long bones of PKBα/Akt1 null mice was previously observed at 1 wk of age [15, 16]. Surprisingly, gene dosage of PKBα/AKT1 was also apparent when comparing the rescued PKBα/Akt1+/− to the effects exhibited in PKBα/Akt1−/−, overall exhibiting an intermediate, milder but significant phenotype looking at the parameters mentioned above for liver, kidney, and bone.
In summary, in utero death of PKBα/Akt1−/− fetuses was attributed to insufficient placental function of PKBα/Akt1−/− placentas and could be prevented by tetraploid complementation. Fetal growth retardation, on the other hand, was associated with the absence of PKBα/AKT1 in the embryo proper. Because AKT deficiency is relevant to IUGR in humans [14], this study provides substantial insights into the roles that PKBα/AKT1 may play in this pathology of human pregnancies.
Acknowledgements
We would like to thank Dr. Sylvain Provot for helpful discussions on bone developmental biology. We would also like to thank Itzhak Ino from the animal facility for technical assistance, as well as Idan Aharon and Naama Cirkin for genotyping. Michal Neeman is incumbent of the Helen and Morris Mauerberger Chair in Biological Sciences.
