PDE2A Is Indispensable for Mouse Liver Development and Hematopoiesis.

Phosphodiesterase 2A (PDE2A) is a cAMP-cGMP hydrolyzing enzyme essential for mouse development and the PDE2A knockout model (PDE2A-/-) is embryonic lethal. Notably, livers of PDE2A-/- embryos at embryonic day 14.5 (E14.5) have extremely reduced size. Morphological, cellular and molecular analyses revealed loss of integrity in the PDE2A-/- liver niche that compromises the hematopoietic function and maturation. Hematopoietic cells isolated from PDE2A-/- livers are instead able to differentiate in in vitro assays, suggesting the absence of blood cell-autonomous defects. Apoptosis was revealed in hepatoblasts and at the endothelial and stromal compartments in livers of PDE2A-/- embryos. The increase of the intracellular cAMP level and of the inducible cAMP early repressor (ICER) in liver of PDE2A-/- embryos might explain the impairment of liver development by downregulating the expression of the anti-apoptotic gene Bcl2. In summary, we propose PDE2A as an essential gene for integrity maintenance of liver niche and the accomplishment of hematopoiesis.


Introduction
During most of prenatal development, the fetal liver (FL) provides the specific niche required for hematopoiesis. Its components, hepatoblasts, resident macrophages, fibroblasts, vascular smooth muscle, stromal and endothelial cells participate with extracellular matrix (ECM) and humoral factors to its generation [1]. In the mouse, at embryonic day 11.5 (E11.5), hematopoietic stem cells (HSC), after their formation in the yolk sac, aorta-gonad-mesonephros region, and placenta, migrate in FL attracted by chemical signals. In FL between E12 and E16, HSC dramatically increase in number and differentiate into mature hematopoietic cells, which are necessary to sustain the ongoing growth of the embryo [2]. Around E16, the hematopoietic function of FL starts to decline in favor of the bone marrow and the liver begins to acquire specific metabolic functions, which include carbohydrate and lipid metabolism [3].
A plethora of genetic alterations affecting liver development has been reported to compromise hematopoiesis. By examining in vitro proliferation and differentiation of the hematopoietic committed
To further characterize the gross phenotype, embryos were analyzed by computational tomography (micro-CT) (Figure 1a and Supplemental Videos).
Histological examination of the livers of PDE2A −/− E14.5 embryos displays loss of the normal liver structure as the central vein is not recognizable in the lobule and the cells are less densely packed with increased empty space among them. Erythroblasts, easily detected at E14.5 in normal livers thanks to their hyperchromatic nuclei, are poorly represented in PDE2A −/− embryos (Figure 1d,e).
The absence of PDE2A in the cells could impact on cellular cAMP concentration if not compensated by the activity of other PDEs. To verify whether PDE2A −/− embryos had an abnormal cAMP accumulation, we measured the level of cAMP in the homogenate of livers of PDE2A +/+ and PDE2A −/− . Figure 1f shows that the level of cAMP in liver of E14.5 PDE2A −/− embryos is significantly higher compared to the levels detected in the liver of PDE2A +/+ embryos. Moreover, in liver of E14.5 PDE2A −/− embryos the mRNA level of ICER, a cAMP-dependent transcription repressor, was up-regulated compared to heterozygous and wild type mice (Figure 1g), as previously observed in PDE2A deficient embryonic hearts [9]. Expression of ICER in livers of E14.5 PDE2A +/+ and PDE2A −/− mice evaluated by qRT-PCR. At least N=3 embryos/genotype were analyzed in each experiment. * P < 0.05; ** P < 0.01.

Hepatic Markers Are Reduced Whereas Stromal and Endothelial Markers Are Increased in Liver Of PDE2A Mutant
We first investigated the impact of PDE2A absence by comparing mutant, wild type and heterozygous developing livers for the level of markers relevant to hepatocytes differentiation and function [3]. Figure 2a shows that livers of PDE2A −/− embryos have reduced levels of albumin and α-fetoprotein (α-FP) respect to livers of PDE2A +/-and PDE2A +/+ embryos. The same outcome results examining the expression of hepatocyte growth factor receptor (cMet), which reduction reaches 20% and of CCAAT/enhancer binding protein alpha (cEBPα) and some hepatocyte nuclear factors (HNF1α, and HNF4α) that are halved. The expression of these factors in PDE2A −/− livers starts to decline at E12.5 (Supplemental Figure S1b) and becomes statistically significant at E14.5.
To investigate whether PDE2A activity directly affects hepatic marker expression, isolated hepatic cells from E14.5 C57BL/6 embryos were treated for 48 h with 10 µM of the selective PDE2A inhibitor erythro-9-(2hydroxy-3-nonyl)adenine (EHNA). As shown in Supplemental Figure S2 no major differences were observed in gene expression analysis after PDE2A inhibition, indicating that PDE2A activity is dispensable for hepatoblast differentiation, at least in vitro.
Afterwards, the impact of PDE2A was evaluated on endothelial and stromal cells which contribute to hematopoietic development in concert with hepatic cells. Figure 2b shows a significant increase of CD31 endothelial marker and of the stromal markers α-smooth muscle actin (α-SMA) and vimentin in PDE2A −/− compared to PDE2A +/-and PDE2A +/+ livers. Altogether, these results suggest that PDE2A is a master gene that orchestrates liver development during embryogenesis, possibly controlling cAMP and ICER concentration.

Hepatic Markers Are Reduced Whereas Stromal and Endothelial Markers Are Increased in Liver Of PDE2A Mutant
We first investigated the impact of PDE2A absence by comparing mutant, wild type and heterozygous developing livers for the level of markers relevant to hepatocytes differentiation and function [3]. Figure 2a shows that livers of PDE2A −/− embryos have reduced levels of albumin and α-fetoprotein (α-FP) respect to livers of PDE2A +/− and PDE2A +/+ embryos. The same outcome results examining the expression of hepatocyte growth factor receptor (cMet), which reduction reaches 20% and of CCAAT/enhancer binding protein alpha (cEBPα) and some hepatocyte nuclear factors (HNF1α, and HNF4α) that are halved. The expression of these factors in PDE2A −/− livers starts to decline at E12.5 (Supplemental Figure S1b) and becomes statistically significant at E14.5. Overall, these results suggest that in vivo PDE2A contributes to hepatoblast differentiation/survival and inhibits stromal and endothelial cells differentiation.

Increased Apoptosis in PDE2A −/− Fetal Livers
The small dimension of the livers of PDE2A knockout embryos and the histological data indicate a reduced cellularity of the organ. In the livers of knockout embryos, the number of cells is 25 times lower compared to heterozygote or wild type animals. This implies an increased rate of cell death and/or a decreased rate of cell proliferation.
To investigate these two possibilities, we evaluated cells dissociated from livers of E14.5 wild type, heterozygous and mutant mice by flow cytometry for their phase in the cell cycle.
The liver of PDE2A −/− embryos had the same proportion of cells in G1, S and G2 phase ( Figure 3a). Moreover, the proportion of proliferating cells was similar as evaluated by Ki67 staining in histological sections (not shown).
On the contrary, TUNEL assay in sections of PDE2A −/− livers at E14.5 displayed a clear increment of apoptotic cells compared to PDE2A +/+ livers (Figure 3b). Apoptosis affected the liver while was not appreciable in other organs except for the cardiac septum [9]. In keeping with this result, western blots analysis of lysates from mutant livers at E14.5, but not from livers of wild type and heterozygous embryos, displayed a strong positive band of cleaved caspase-3 (Figure 3c), a marker of apoptosis. Furthermore, in the liver of mutant embryos the level of transcript of the anti-apoptotic factor Bcl2 was lower respect to wild type and heterozygous embryos (Figure 3d). To investigate whether PDE2A activity directly affects hepatic marker expression, isolated hepatic cells from E14.5 C57BL/6 embryos were treated for 48 h with 10 µM of the selective PDE2A inhibitor erythro-9-(2-hydroxy-3-nonyl)adenine (EHNA). As shown in Supplemental Figure S2 no major differences were observed in gene expression analysis after PDE2A inhibition, indicating that PDE2A activity is dispensable for hepatoblast differentiation, at least in vitro.
Afterwards, the impact of PDE2A was evaluated on endothelial and stromal cells which contribute to hematopoietic development in concert with hepatic cells. Figure 2b shows a significant increase of CD31 endothelial marker and of the stromal markers α-smooth muscle actin (α-SMA) and vimentin in PDE2A −/− compared to PDE2A +/− and PDE2A +/+ livers.
Overall, these results suggest that in vivo PDE2A contributes to hepatoblast differentiation/survival and inhibits stromal and endothelial cells differentiation.

Increased Apoptosis in PDE2A −/− Fetal Livers
The small dimension of the livers of PDE2A knockout embryos and the histological data indicate a reduced cellularity of the organ. In the livers of knockout embryos, the number of cells is 25 times lower compared to heterozygote or wild type animals. This implies an increased rate of cell death and/or a decreased rate of cell proliferation.
To investigate these two possibilities, we evaluated cells dissociated from livers of E14.5 wild type, heterozygous and mutant mice by flow cytometry for their phase in the cell cycle. The liver of PDE2A −/− embryos had the same proportion of cells in G1, S and G2 phase ( Figure 3a). Moreover, the proportion of proliferating cells was similar as evaluated by Ki67 staining in histological sections (not shown).
To investigate whether PDE2A activity prevents apoptosis of liver cells, we used EHNA in cultures of fetal liver cells isolated from C57BL/6 embryos. Treatment for 48 h with EHNA (10 µM) did not increase apoptosis. Nevertheless, EHNA (10 µM) treated cells underwent apoptosis and displayed caspase-3 activity if the inhibitor was administered together with tumor necrosis factor-alpha (TNFα, 5 ng/mL) and low doses of cycloheximide (CHX, 25 µg/mL) (Figure 3e) for the last 24 h, suggesting that PDE2A inhibition increases the sensitivity of liver cells to apoptotic stimuli.
Taken together, these data indicate that liver hypoplasia in PDE2A −/− embryos is caused by apoptotic increase rather than decrease of cellular proliferation and that PDE2A by reducing the intracellular concentration of cAMP/ICER, may play a role in protecting liver cells from apoptotic factors during development. .5 liver cells isolated from C57BL/6 embryos enter apoptosis after TNFα (5 ng/mL) and CHX (25 µg/mL) treatments if pretreated with the PDE2A inhibitor EHNA (10 µM). Apoptosis was evaluated by cleaved caspase-3 in western blots. Densitometry analysis relative to tubulin is shown. N = 2 embryos. * P < 0.05.

Hepatoblasts, Endothelial and Stromal Cells Undergo Apotosis in Livers of PDE2A −/− Embryos -
To investigate which cell types undergo apoptosis in livers of PDE2A −/− embryos, flow cytometry and histological analyses were performed in E14.5 mutant and wild type embryos. Isolated CD71 + immature red blood cells and CD45 + leucocytes [12,13] co-stained with the apoptotic cell marker Annexin-V, displayed a similar pattern in wild type and mutant embryos (Figure 4a,b). On the contrary, apoptosis was significantly  On the contrary, TUNEL assay in sections of PDE2A −/− livers at E14.5 displayed a clear increment of apoptotic cells compared to PDE2A +/+ livers (Figure 3b). Apoptosis affected the liver while was not appreciable in other organs except for the cardiac septum [9]. In keeping with this result, western blots analysis of lysates from mutant livers at E14.5, but not from livers of wild type and heterozygous embryos, displayed a strong positive band of cleaved caspase-3 (Figure 3c), a marker of apoptosis. Furthermore, in the liver of mutant embryos the level of transcript of the anti-apoptotic factor Bcl2 was lower respect to wild type and heterozygous embryos (Figure 3d).
To investigate whether PDE2A activity prevents apoptosis of liver cells, we used EHNA in cultures of fetal liver cells isolated from C57BL/6 embryos. Treatment for 48 h with EHNA (10 µM) did not increase apoptosis. Nevertheless, EHNA (10 µM) treated cells underwent apoptosis and displayed caspase-3 activity if the inhibitor was administered together with tumor necrosis factor-alpha (TNFα, 5 ng/mL) and low doses of cycloheximide (CHX, 25 µg/mL) (Figure 3e) for the last 24 h, suggesting that PDE2A inhibition increases the sensitivity of liver cells to apoptotic stimuli.
Taken together, these data indicate that liver hypoplasia in PDE2A −/− embryos is caused by apoptotic increase rather than decrease of cellular proliferation and that PDE2A by reducing the intracellular concentration of cAMP/ICER, may play a role in protecting liver cells from apoptotic factors during development.

Hepatoblasts, Endothelial and Stromal Cells Undergo Apotosis in Livers
To investigate which cell types undergo apoptosis in livers of PDE2A −/− embryos, flow cytometry and histological analyses were performed in E14.5 mutant and wild type embryos. Isolated CD71 + immature red blood cells and CD45 + leucocytes [12,13] co-stained with the apoptotic cell marker Annexin-V, displayed a similar pattern in wild type and mutant embryos (Figure 4a,b). On the contrary, apoptosis was significantly higher in endothelial cells (CD31 + /Annexin-V + ) obtained from the liver of  In mouse embryos between E11.5 and E16.5, the liver is the primary hematopoietic organ [14,15]. Blood smears obtained from E14.5 normal peripheral blood display a great number of mature non-nucleated erythroid cells, which are mainly liver derived. In blood smears obtained from PDE2A −/− embryos, erythroid cells are still predominantly nucleated and hence immature. Most likely, they are remnants of yolk sac origin These results strongly indicate that in PDE2A −/− embryos, the absence of PDE2A induces apoptosis in cells of different lineages that overall forms the liver niche.

Impairment of Hematopoietic Differentation in PDE2A −/− Embryos
In mouse embryos between E11.5 and E16.5, the liver is the primary hematopoietic organ [14,15]. Blood smears obtained from E14.5 normal peripheral blood display a great number of mature non-nucleated erythroid cells, which are mainly liver derived. In blood smears obtained from PDE2A −/− embryos, erythroid cells are still predominantly nucleated and hence immature. Most likely, they are remnants of yolk sac origin suggesting an impairment of erythropoiesis in PDE2A −/− embryos (Figure 5a). Furthermore, the number of blood cells collected from the umbilical cord of PDE2A −/− embryos revealed a 4.5-fold reduction compared to PDE2A +/+ . blood cells collected from the umbilical cord of PDE2A −/− embryos revealed a 4.5-fold reduction compared to PDE2A +/+ .
The previous results prompted us to investigate the hematopoietic development analyzing with flow cytometry cells isolated from the liver of E14.5 wild type, heterozygous and mutant embryos stained with antibodies directed to specific hematopoietic lineages (Figure 5b−h).
The relative proportion of CD45 positive cells resulted similar to wild type in the liver of PDE2A −/− embryos (Figure 5b). Instead, the staining for Ter119, a cell surface erythroid lineage marker expressed in terminally differentiating erythroblasts [16,17] showed the fall of erythroblasts percentage from 90% in the liver of PDE2A +/+ to 75% in the liver of PDE2A −/− embryos (Figure 5c). Staining with F4/80, CD41/CD61 and Gr1, that label mature macrophages, megakaryoblasts and granulocytes respectively [13,18,19], showed a reduction of these cells in PDE2A −/− livers (Figure 5d-f). No major differences were observed in lymphoid lineage labeled with B220 and CD3 markers (Supplemental Figure S3). These results suggest that in fetal livers of PDE2A −/− embryos the number of mature blood cells of myeloid and erythroid lineages is lower.
The previous results prompted us to investigate the hematopoietic development analyzing with flow cytometry cells isolated from the liver of E14.5 wild type, heterozygous and mutant embryos stained with antibodies directed to specific hematopoietic lineages (Figure 5b-h).
The relative proportion of CD45 positive cells resulted similar to wild type in the liver of PDE2A −/− embryos (Figure 5b). Instead, the staining for Ter119, a cell surface erythroid lineage marker expressed in terminally differentiating erythroblasts [16,17] showed the fall of erythroblasts percentage from 90% in the liver of PDE2A +/+ to 75% in the liver of PDE2A −/− embryos (Figure 5c). Staining with F4/80, CD41/CD61 and Gr1, that label mature macrophages, megakaryoblasts and granulocytes respectively [13,18,19], showed a reduction of these cells in PDE2A −/− livers (Figure 5d-f). No major differences were observed in lymphoid lineage labeled with B220 and CD3 markers (Supplemental Figure S3). These results suggest that in fetal livers of PDE2A −/− embryos the number of mature blood cells of myeloid and erythroid lineages is lower.
On the other side, the percentage of the LSK population of stem/progenitor cells, identified by Lin − , Sca1 + , c-Kit + in fetal livers at E14.5 [20], increased in mutant embryos (Figure 5g), even if the absolute number of LSK was reduced proportionally to PDE2A −/− liver size (PDE2A +/+ 74473 ± 3476 vs PDE2A −/− 17548 ± 1168, P < 0.01). In agreement with this result, it was observed an increase in the percentage of CD11b positive cells (Figure 5h) that are also part of the progenitor population in fetal liver [12,21]. These results suggest that hematopoietic stem cells colonize and survive in PDE2A −/− fetal liver, but apoptosis of the cells in their supporting niche compromises their differentiation.
To further confirm this hypothesis, we compared the capacity of PDE2A −/− and PDE2A +/+ hematopoietic progenitor cells of livers from E14.5 to develop into the erythroid and myeloid lineages in an in vitro methylcellulose colony assay [22]. The number of erythroid and myeloid colonies obtained from the two genotypes after 7 days of culture was equivalent (Figure 6), indicating that PDE2A −/− hematopoietic stem cells are able to proliferate and differentiate when tested in vitro under proper conditions and even though originating from severely hypoplastic and immature livers, give rise to all canonical hematopoietic lineages. To further confirm this hypothesis, we compared the capacity of PDE2A −/− and PDE2A +/+ hematopoietic progenitor cells of livers from E14.5 to develop into the erythroid and myeloid lineages in an in vitro methylcellulose colony assay [22]. The number of erythroid and myeloid colonies obtained from the two genotypes after 7 days of culture was equivalent (Figure 6), indicating that PDE2A −/− hematopoietic stem cells are able to proliferate and differentiate when tested in vitro under proper conditions and even though originating from severely hypoplastic and immature livers, give rise to all canonical hematopoietic lineages.

Discussion
In this study, we show that the lack of PDE2A results in profound defects in early liver development. At the time of death, livers are hypocellular because of apoptosis and pale because the differentiation of mature blood cells from their progenitors is defective.
In parallel, we observed an increase of the stem/progenitor population, probably due to the lack of proper differentiation conditions. Nevertheless, the hematopoietic progenitors isolated from the livers of knockout

Discussion
In this study, we show that the lack of PDE2A results in profound defects in early liver development. At the time of death, livers are hypocellular because of apoptosis and pale because the differentiation of mature blood cells from their progenitors is defective.
In parallel, we observed an increase of the stem/progenitor population, probably due to the lack of proper differentiation conditions. Nevertheless, the hematopoietic progenitors isolated from the livers of knockout embryos proliferate and differentiate normally in vitro giving rise to respective hematopoietic lineages. The increase in colony formation of PDE2A −/− embryos did not correspond to the increase observed in their stem/progenitor cells, probably because the latest lost their full competence in vivo since they were hosted in a compromised environment at the peri-lethal stage of development.
PDE2A deficient embryos share similarities with other mouse models in which gene disruption causes abnormal liver development. The most similar picture is offered by XBP-1 knockout mice [6]. XBP1 is a major component of the unfolded protein response (UPR) of the endoplasmic reticulum. Knockout mouse embryos die at mid-gestation because the hematopoietic progenitors do not generate mature blood cells for the inadequate environment in the hypoplastic liver, but they differentiate normally in vitro. However, XBP-1 is a cAMP-dependent factor and indeed we observed that XBP-1 mRNA is upregulated in PDE2A mutants (data not shown) implying that this factor is unlikely responsible for the phenotype of PDE2A −/− embryos. However, we cannot exclude that cAMP level in PDE2A −/− embryos affects the activity of other UPR components.
PDE2A and XBP-1 knockout embryos share a decreased expression of αFP, an early marker, which promotes the growth of liver cells and may contribute to the reduced liver size in both models.
Furthermore, PDE2A −/− embryos have reduced expression of several liver specific transcription factors, essential to establish liver function (such as cEBPα, cMet, HNF1α, HNF4α) and of albumin.
All of these genes are essential for liver development. Albumin expression starts at E13.5 and later increases to fulfill its function as serum protein [23]. The expression of αFP starts at E11.5 in the liver and increases until birth, when is silenced [24,25]. cMet codifies for the receptor of hepatocyte growth factor (HGFR), which is essential for the survival of hepatocytes during development [26]. cEBPα is a factor involved in the commitment of the hepatoblasts to the hepatocytic and cholangiocytic lineages [27,28] and HNFs regulate the expression of factors relevant for the function of the fetal liver [29,30].
None of these genes appears to contain CREB binding sites in the promoter; however, we cannot exclude that cAMP pathways affect indirectly their expression. Since there are several critical pathways which control hepatogenesis such as TGFβ, Wnt, FGF, BMP signaling [31][32][33], it is possible that the increase of cAMP in PDE2A mutant mice disturbs dosage and/or signaling of cytokines and growth factors severely compromising liver development.
In vitro experiments of PDE2A activity inhibition in wild type fetal liver cells did not replicate the downregulation of specific hepatic transcripts observed in mutant liver in vivo. These results may be interpreted in different ways. It is possible that the timing of mRNA analysis was not correct or that the reduction of expression in vivo reflects the loss of hepatic cells which does not occurs in vitro. In alternative, the lack of downregulation of gene expression in vitro could be due to the inadequate cellular cross talk in this environment.
Interestingly, in the livers of mutant animals we found an increase of expression of endothelial and stromal markers. However, stromal cells more undergo to apoptosis and are displaced in the tissue, probably because apoptosis reduces the hepatic parenchyma, as previously reported [34].
Mice deficient for genes regulating the NF-kB pathway offer another example of phenotype with points in common with that of PDE2A mutants. During mid-fetal gestation, the embryos exhibit massive liver degeneration [5,[35][36][37][38] indicating that NF-kB protects embryonic liver from apoptosis induced by TNFα, which is produced by hematopoietic cells [38][39][40][41][42]. We did not investigate the NF-kB pathway in PDE2A −/− livers; however, hepatocytes from wild type livers undergo apoptosis when cultured with the PDE2A inhibitor EHNA and TNFα.
The absence of PDE2A is not compensated by the activity of other PDE members and this accounts for the high level of cAMP observed in liver of mutant embryos underlying the importance of this enzyme in balancing the cAMP signaling.
Agents that elevate cAMP stimulate apoptosis by activating PKA [42] in a variety of other systems, including thymocytes [41], immortalized primary granulosa cells [43,44], human mammary carcinoma cells [45] and various normal and transformed T and B cells [42]. In other cases, PKA activated by cAMP blocks apoptosis. Examples include aged neutrophils [46], macrophage cell lines exposed to exogenous nitric oxide [47], ovarian follicles [48] and T cells [49].
In PDE2A −/− embryos of E14.5, we found that apoptosis hits specific cell populations such as hepatoblasts, endothelial and stromal cells, that constitute the niche supporting the efficient blood cell differentiation at this stage of development. It is plausible that the moderate apoptosis of these cells we see in the liver at this time is just the tail end of events occurred earlier.
Moreover, we found that the number of F4/80 macrophages is severely reduced in the fetal livers of PDE2A −/− embryos. Erythroid cell differentiation in the erythroblastic island depends on macrophages in the fetal liver [18,50]. Thus, defects of macrophages may contribute to the impairment of red blood cell maturation in PDE2A −/− embryos.
A high level of cAMP could promote apoptosis not only by activating PKA, but also by directly deregulating the transcription of cAMP-responsive genes such as ICER that represses the transcription of several CRE (cAMP responsive regulatory element) containing genes [51]. We recently reported the upregulation of the cAMP/ICER system in the hearts of PDE2A −/− embryos or after PDE2A inhibition in vivo, in which like in the liver we observed increased apoptosis and modulation of several genes necessary for cardiogenesis [9]. Several examples indicate that ICER can induce apoptosis by downregulating the Bcl2 anti-apoptotic gene [52,53]. In PDE2A −/− livers, ICER is upregulated and Bcl2 is downregulated, suggesting a mechanism through which liver cells undergo apoptosis.
In conclusion, in the mouse, the lack of PDE2A through the increase of cAMP could activate apoptosis affecting directly both ICER and Bcl2 genes; however, we cannot exclude the involvement of alternative second messengers as for example the modulation of intracellular calcium concentration [54]. The widespread cell apoptosis causes the loss of the microenvironment necessary for the differentiation of hematopoietic stem cells. The consequent failure of erythrocytes production results in an anemia that is probably the principal cause of the death of PDE2A −/− embryos.

Histological Procedures and TUNEL Assay
Uteri were removed from pregnant PDE2A +/− mice between E11 to E16 and washed with PBS. For successive analysis, individual embryos were processed whole or dissected to remove the liver. E11.5, E12.5 and E 14.5 embryos or isolated livers were fixed overnight at 4 • C in 4% paraformaldehyde in PBS and embedded in paraffin following standard procedures. Serial sagittal sections were examined after staining with hematoxylin and eosin or underwent to immunofluorescence analysis.
For TUNEL assay, paraffin sections were dewaxed in toluene and antigen retrieval was performed 3 times for 5 min each by microwave heating in 10 mM sodium citrate buffer. After further two washes of 5 min, the sections were incubated in TUNEL reaction mixture (in situ Cell Death Detection Kit, POD, Roche, Basel, Switzerland) for 1 h at 37 • C followed by staining of the nuclei with Hoechst 33342 (Sigma-Aldrich, St. Louis, MO, USA) for 5 min and mounting. Images were acquired by Nikon Eclipse Ti-S microscope.

Micro-CT Imaging and Volume Measurements
Imaging specimen preparation: embryos were washed twice in PBS for 10 min and soaked at room temperature for 1 week in a potassium iodine contrast agent, 0.1N (v/v) Lugol solution (Sigma-Aldrich), refreshing it twice during the week of exposure (Ermakova et al., 2018). Micro-CT scanning: Computed tomography images were acquired by a high-resolution 3D micro-CT imaging system (Skyscan 1172G Bruker, Kontich-Belgium), using a L7901-20 Microfocus X-ray Source (Hamamatsu). The acquisition of volumes was performed in 1.5 mL micro-tubes, with a camera pixel/size of 7.9 µm, camera binning 2 × 2, tube voltage peak of 39 kV, tube current of 240 µA, exposure time of 450 ms. Reconstructions of tomographic datasets were performed using built-in NRecon Skyscan Software (Version:1.6.6.0; Bruker). The 3D volumes were analyzed using 3D Visualization Software CTvox v. 2.5 (Bruker). Tissues segmentation: Manual image-by-image segmentation aimed at calculation of liver volume was applied, using Bruker micro-CT Analyzer Version 1.13 software. A histological atlas of mouse development [55] was used for guidance to accurately identify, demarcate and segment each liver (N = 3) in a specific VOI (Volume of Interest) for automated volume measurements [55,56].

Immunofluorescence
Histological sections were permeabilized in 0.1% Triton v/v, 3 times for 5 min each and blocked in PBS-3% BSA w/v for 1 h and incubated with rabbit anti-Ki67 primary antibody (Abcam, Cambridge, UK), rabbit anti-α-Fetoprotein (α-FP) or rabbit α-smooth muscle actin (α-SMA) (all purchased from DAKO-Agilent, Santa Clara, CA, USA). After overnight incubation at 4 • C, sections were washed in PBS and incubated with Alexa Fluor 568 or Alexa Fluor 488 secondary antibody (ThermoFisher scientific, Waltham, MA, USA) for 1 h at room temperature and when specified stained also for TUNEL. Following extensive washes in PBS, the sections were counterstained with DAPI and mounted with PBS-50% v/v Glycerol. Images were acquired by a Nikon Eclipse Ti-S microscope equipped with a Photometric CoolSNAP EZ turbo 1394 camera.

cAMP Assay
Briefly, E14.5 livers were homogenized and sonicated in cAMP-Glo™ Lysis Buffer (Promega, Madison, WI, USA) containing 0.5 mM IBMX and 100 µM Rolipram (Sigma-Aldrich) following the manufacture's instruction to prevent cAMP hydrolysis and centrifuged to eliminate cellular debris. Samples were heated for 7-10 min at 70 • C and chilled 5 min in ice. 5 µg from each sample were then mixed with 40 µL of cAMP-Glo Detection Solution containing PKA and incubated for 20 min at room temperature. After the addiction of 80 µL of Kinase-Glo Reagent, the samples were incubated for 10 min and read to the luminometer. Results are referred to a cAMP Standard Curve conducted together with the cAMP-Glo Assay.

Liver Cell Culture and Treatments
Fetal livers were isolated from C57BL/6 E14.5 embryos, mechanically dissociated in phosphate-buffered saline without calcium and magnesium (CMF) (Gibco-ThermoFisher scientific, Waltham, MA, USA) into single cells suspension and centrifuged at 1200 rpm for 5 min at 4 • C. The cellular pellet was re-suspended in DMEM supplemented with 10% FBS and plated in 6 well plates. After 48 h the medium was changed with DMEM 1% FBS with or without EHNA (10 µM, Calbiochem-Sigma-Aldrich, St. Louis, MO, USA) for 24 h. The medium was then changed and supplemented with EHNA (10 µM), or with TNFα (5 ng/mL, Sigma-Aldrich) plus the protein synthesis inhibitor CHX (2.5 µg/mL, Sigma-Aldrich) or with EHNA, TNFα and CHX together. At the end of further 24 h of incubation, the cells were homogenized with lysis buffer and analyzed by western blot or with TRIzol ® (Life Technologies, Waltham, MA, USA) reagent for mRNA extraction.
After cDNA synthesis step, qRT-PCR reaction was carried out in duplicate for each gene for each liver by using GoTaq qPCR Master Mix (Promega, Madison, WI, USA) using GAPDH for normalization and the reaction carried out on Applied 7500 sequence detection system (Applied Biosystems, Foster city, CA, USA).
Primers pairs used for qRT-PCR are:  Dead cells were stained with 7Amino-Actinomycin D dye (Abcam, Cambridge, UK) and excluded from the analysis.

Cell Cycle Analysis
For cell cycle analysis, 3 × 10 5 cells for each point were re-suspended in DMEM 50% FBS and fixed in 70% ethanol for 24 h. The cells were then incubated with 50 µg/mL propidium iodide (Sigma-Aldrich) and 50 units/mL Rnase-free DnaseA (Sigma-Aldrich) and analyzed after 3 h (1 × 10 4 events were acquired) using an Epics XL Cytometer (Beckman Coulter, USA).

Statistical Analyses
All data are expressed as mean ± SEM and analyzed with one-way ANOVA with Bonferroni correction (Bonferroni Post Hoc Test) or with Student's t-Test. Differences were considered significant if * P < 00.5.