Fetal and trophoblast PI3Kp110α have distinct roles in regulating resource supply to the growing fetus

Previous studies suggest that the placental supply of nutrients to the fetus adapts according to fetal demand. However, the signaling events underlying placental adaptations remain largely unknown. Earlier work in mice has revealed that loss of the phosphoinositide 3-kinase p110α impairs feto-placental growth but placental nutrient supply is adaptively increased. Here we explore the role of p110α in the epiblast-derived (fetal) and trophoblast lineages of the conceptus in relation to feto-placental growth and placental development and transfer function. Using conditional gene manipulations to knock-down p110α either by ~50% or ~100% in the fetal lineages and/or trophoblast, this study shows that p110α in the fetus is essential for prenatal development and a major regulator of placental phenotype in mice. Complete loss of fetal p110α caused embryonic death, whilst heterozygous loss resulted in fetal growth restriction and impaired placental formation and nutrient transport. Loss of trophoblast p110α also resulted in abnormal placental development, although fetuses were viable. However, in response to complete loss of trophoblast p110α, the placenta failed to transport sufficient amino acid to match fetal demands for growth. Using RNA-seq, we identified several genes downstream of p110α in the trophoblast that are important in adapting placental phenotype to support fetal growth. Further work using CRISPR/Cas9 genome targeting showed that loss of p110α differentially affects the expression of genes in trophoblast and embryonic stem cells. Our findings thus reveal important, but distinct roles for p110α signaling in the different compartments of the conceptus, which control fetal resource acquisition and ultimately affect healthy growth. One Sentence Summary Fetal and trophoblast p110α modify resource allocation


51
Intrauterine growth is dictated by the genetically-determined fetal drive for growth and the supply of 52 nutrients and oxygen to the fetus. In turn, fetal substrate supply depends on the functional capacity of the 53 placenta to transfer nutrients and oxygen from the mother to the fetus. Insufficient fetal substrate supply 54 prevents the fetus from achieving its genetic growth potential and leads to intrauterine growth restriction, 55 which affects up to 10% of the population and is associated with perinatal morbidity and mortality 56 (Baschat et al., 2007;Baschat and Hecher, 2004). Studies have shown that the transport capacity of the 57 placenta is diminished in human pregnancies associated with fetal growth restriction (Glazier et al., 1997; 58 Jansson and Powell, 2006), suggesting that placental insufficiency may be the underlying cause of a fetus 59 failing to achieve its genetic growth potential. However, there is also evidence that placental transport 60 capacity may adapt to maintain the supply of resources appropriate for the growth potential of the fetus 61 (Sandovici et al., 2012;Sferruzzi-Perri and Camm, 2016). For instance, the placental capacity to supply 62 resources to the fetus is higher in the lightest compared to heaviest placentas supporting babies within the 63 normal birth weight range (Godfrey et al., 1998)

117
Fetal and trophoblast p110α interplay to regulate growth of the conceptus 118 We found PI3K p110α was highly expressed by the placental transport labyrinthine zone and the 119 endocrine junctional zone of the mouse placenta (Fig. S1). We then halved p110α expression in either the 120 trophoblast or the fetal (epiblast-derived) lineages of the conceptus by mating mice in which exons 18-19 121 of the Pik3ca gene were flanked by LoxP sites (Graupera et al., 2008), to mice transgenic for Cyp19Cre 122 (Wenzel and Leone, 2007) and Meox2Cre (Tallquist and Soriano, 2000) respectively, termed throughout 123 as Het-P and Het-F. These Cre lines are active in opposite compartments of the conceptus; for Het-P the 124 Cyp19Cre targets expression in the trophoblast lineages of the placenta but not fetus, labyrinthine fetal 125 capillaries or mesenchyme in the chorion ( Fig. 1A and C). In contrast, in Het-F the Meox2Cre deletes 126 p110α expression in the fetus and placental labyrinthine fetal capillaries and chorionic mesenchyme but 127 not trophoblast (Fig. 1B and C) (further information regarding the genetic crosses can be found in Table   128 S1). We then compared the fetal and placental growth phenotype of the conditional Het-P and Het-F on 129 day 19 of pregnancy (term ~20 days) to conceptuses with global heterozygous p110α deficiency, 130 achieved with the ubiquitous CMVCre line (Schwenk et al., 1995) (termed Het-U, Fig. 1C). We found that 131 compared to their wild-type (WT) littermates, there was no effect of heterozygous deficiency of p110α in 132 the trophoblast on fetal or placental weight in Het-P mutants ( Fig. 2A). However, fetal and placental 133 weights were 10-15% lighter for Het-F and Het-U conceptuses (Fig. 2B and C). The findings in Het-F and   134 Het-U mutants are consistent with the proliferation defects observed in embryos with a deficiency in 135 p110α (Bi et al., 1999;Foukas et al., 2006) and reveal for the first time that p110α in the embryo is 136 important for determining the size of the placenta.

138
Although fetal and not trophoblast p110α deficiency reduced conceptus weight, both affected the 139 structure of the placenta (Fig. 3A-C). In the labyrinthine region of Het-P conceptuses, the volume of 140 maternal blood spaces, fetal capillaries and surface area were reduced though trophoblast increased 141 compared to their WT controls (Fig. 3A). In Het-F conceptuses, the labyrinthine zone, fetal capillaries 142 and trophoblast volume were decreased versus WT littermates (Fig. 3B). In the constitute p110α 143 heterozygote (Het-U), the volume of the labyrinthine zone, fetal capillaries and trophoblast were reduced, 144 surface area decreased and barrier to diffusion was greater, relative to WT littermates (Fig. 3C).

145
Reassuringly, the same defects in Het-U placental structure were previously observed for α/+ mutants 146 near term (Sferruzzi-Perri et al., 2016). The volume of the endocrine junctional zone in Het-P, Het-F or 147 Het-U placentas was not significantly altered when compared to the respective WT controls (Fig. S2).

148
Taken together, these findings indicate that p110α in the fetal and trophoblast lineages of the conceptus 149 interplay to regulate the development of the transport region in the placenta.

152
To assess whether morphological alterations in the placenta affect placental resource allocation to the 153 fetus in Het-F, Het-U and Het-P, we measured the uni-directional maternal-fetal transfer of the non-154 metabolisable analogues of glucose ( 3 H-methyl-D glucose; MeG) and a neutral amino acid ( 14 C-methyl 155 amino-isobutyric acid; MeAIB) on day 19 of pregnancy. We assessed fetal counts in relation to the 156 estimated surface area for transport or to fetal weight, which respectively provided us with indices of the 157 placental capacity for nutrient transfer and fetal growth relative to supply. We found that in compensation

Fetal p110α is essential for embryonic development and trophoblast p110α is critical for its ability to
167 up-regulate amino acid transport to match fetal demands for growth 168 We wanted to know more about the regulation of placental resource allocation to the developing fetus 169 when there is a loss of fetal and trophoblast p110α. In particular, we wondered whether adaptation of 170 placental transport function would still occur in heterozygous mutants (Het-U) if the trophoblast or fetal 171 lineages were completely deficient in p110α. To do this, we selectively deleted the remaining p110α 172 from the trophoblast or the fetal compartment of Het-U mice, using Cyp19Cre and Meox2Cre, 173 respectively (termed Hom-P and Hom-F, respectively). We found that deleting the remaining p110α from 174 the fetal lineages was lethal between days 11 and 12 of pregnancy (Table 1). In contrast to the lethality of 175 Hom-F embryos, we found viable Hom-P fetuses in late gestation ( Fig. 1C and Table S2). The timing of 176 Hom-F lethality was identical to mutants with constitutive homozygous deficiency of p110α (α/α; 177 (Foukas et al., 2006)) and is consistent with the role of p110α in early murine embryonic development 178 (Xu et al., 2009). Taken together, these findings highlight that p110α in the fetal, but not the trophoblast 179 compartment of the conceptus, is obligatory for prenatal development.

181
When comparing the growth of the Hom-P conceptuses to the control Het-U littermates, we found that 182 fetal growth was restricted by a further 8% on day 19 of pregnancy (Table 2). However, despite the 183 reduction in fetal growth, there was no difference in placental weight and labyrinthine morphology in 184 Hom-P versus Het-U (Table 2). These observations suggest that the more severe reduction in fetal growth 185 in Hom-P (relative to Het-U), was not caused by additional defects in the formation of the placental 186 exchange region due to a complete loss of p110α in the trophoblast.

188
We wondered whether the greater reduction in fetal growth may have been caused by a defect in placental 189 transport function (a failure of the placenta to adapt its transfer capacity) in the Hom-P versus the Het-U 190 mutants. We found that Hom-P placentas transferred 30% less amino acid (MeAIB) for the surface area 191 available than Het-U littermates (Fig. 5A). Furthermore, Hom-P fetuses received less MeAIB solute for 192 their size, as well as overall ( Fig. 5B and Fig. S4). As the Het-U placenta up-regulated its transport 193 capacity (Fig. 4C), this suggests that the greater reduction in fetal growth was due to an inability of the 194 Hom-P placenta to adaptively increase amino acid transfer to the fetus. Placental glucose transfer capacity 195 however, was not affected by a lack of placental p110α; transfer of MeG by the Hom-P placenta was 196 equal to Het-U (Fig. 5A-B and Fig. S4). Data on feto-placental growth and placental transfer in Hom-P 197 compared to wild-type and Het-P, which in contrast, retain p110α in fetus, are shown in Tables S3 and   198 S4. Collectively, our data suggest that the demand signals of the compromised feto-placental unit for 199 more amino acids operate via p110α in the trophoblast. To identify genes responsible for the phenotypic differences observed between Het-U and Hom-P 203 placentas at day 19 of pregnancy we compared their transcriptome using RNA-seq. We identified 97 204 differentially expressed genes, with 61 up-and 36 down-regulated in Hom-P versus Het-U placenta 205 (Table S5) (Table   212 S5). Consistent with this, we found increased levels of apoptosis in the Hom-P relative to the Het-U 213 placenta, with greatest levels observed in the junctional zone ( Fig. 6B and C). We also found greater  (Table S5). Therefore, p110α operates via several genes in 224 the trophoblast to alter placental phenotype to support fetal growth.

227
We wanted to know whether the genes differentially expressed (DEGs) between Hom-P and Het-U

384
Collectively, our data indicate that p110α plays divergent roles in the different compartments of the 385 conceptus, which ultimately affect placental phenotype and fetal growth. In particular, p110α in the fetal 386 lineages of the conceptus is essential for prenatal development and a major regulator of placental 387 phenotype (Fig. 8). Moreover, trophoblast p110α signaling is critical for its ability to up-regulate amino 388 acid transport to match fetal demands for growth near term. However, the Hom-F and Hom-P have a  Table S1 for experimental crosses used. Note the wild-types for the Het-P 423 and Het-U are the same animals and the Het-U were the controls for Hom-P.