**3. Discussion**

This study focused on analyzing the changes in the molecular pathways regulating HSPCs and erythroid differentiation in fetuses from PE pregnancies. Our results were in agreemen<sup>t</sup> with previous reports [27,48] demonstrating no significant di fferences in the frequency of UCB HSPCs or the expression of the SAMs on these cells, despite originating from PE or normotensive pregnancies. In addition, the absence of significant di fferences in the number of BFU-Es produced during in vitro di fferentiation of the UCB CD34+ cells from PE and normotensive groups in our samples was in accordance with previous reports [26,27]. Interestingly, both function and expression of SAMs on HSPCs are precisely regulated during embryonic development [49–51] as well as lineage di fferentiation [47,52,53]. Thus, the absence of significant di fferences in SAM expression and colony formation imply that PE may not cause significant changes in fetal HSPC migration/homing or intrinsic di fferentiation capacity.

Investigation of transcriptional changes demonstrated di fferences in gene expression in UCB HSPCs in PE vs. normotensive pregnancies. Many of the detected mRNAs were those of ribosomal proteins (RPs), important in proliferating cells [54,55] and primarily known to play a role in the maturation of ribosomal RNAs, ribosome biogenesis, and polysome formation [56,57]. However, ribosomal dysfunction can also trigger apoptosis, autophagy, cell cycle arrest as well as cellular senescence [58–62] and has been related to several hematological disorders, such as Diamond-Blackfan anemia [63,64]. These results prompted us to explore whether PE might be associated with changes in gene expression that adversely a ffect erythropoiesis and erythroid maturation. Analyzing the proteomics of the in vitro di fferentiated erythroid cells suggested changes in pathways important in the metabolism, immune system, protein processing, and export as well as phagosomes. Considering that efficient erythroid di fferentiation and maturation requires a synchronized regulation of iron, amino acid, and glucose metabolism [34], as well as various signaling pathways [65,66], alterations in these pathways could lead to intracellular changes and ine ffective erythroid maturation.

To investigate if the results obtained in our in vitro analyses translated into changes in vivo, we compared the UCB erythroid profiles in PE vs. normotensive pregnancies. No significant di fferences were observed in the erythroid populations from arterial or venous UCB of male or female fetuses in either group. This observation was in agreemen<sup>t</sup> with our earlier results and suggested that disruption of erythroid maturation may contribute to a higher erythroblast count in the UCB from PE pregnancies independent of fetal hypoxia-induced EPO-dependent erythropoiesis [18–21,24]. It is also important to consider that fetal hypoxia in pregnancy is chronic, rather than acute, and hemolysis is not commonly observed in the fetuses born to PE. Thus, the release of immature erythroblasts into the UCB may have explanations other than those suggested for acute hypoxia or hemolytic anemia [67].

To confirm any mechanisms that could a ffect fetal erythroid maturation and enucleation in PE, the transcriptome of UCB erythroid cells were analyzed. Clustering of the male and female erythroid cells based on transcriptome analysis indicated that the expression of some genes located on both autosomal and sex chromosomes varied between the sexes, specifically upregulating the RNA transcription pathway and several mitochondrial factors in the male fetuses. Comparing the male fetuses among the PE and the normotensive groups indicated a decrease in DNA repair, convergen<sup>t</sup> extension, protein ubiquitination, and vesicle tra fficking, as well as deregulation in the cell cycle. For instance, lower CDKN2D (p19) and RRAGA, two regulators of cell cycle G1 progression [68], and amino acid-dependent mTORC1 activity [69,70] were associated with increased BTG2 and WRINP1 that regulate G1/S transition [71] and G1/S or G2/M arrest [72], respectively. Along with these changes, inhibitors of RNA transcription and several genes important in cell cycle S and mitosis phases were upregulated (Supplementary Table S3). Interestingly, KLHDC8B, a factor that safeguards the cell against mitotic errors and nuclear abnormalities, was also increased [73,74]. Due to the absence of significant di fferences in UCB erythroid populations between PE and normotensive male fetuses, the changes in gene expression could not be explained by elevated late basophilic or polychromatic erythroblasts [75], which express high levels of mitosis-related genes. Thus, the gene expression alterations observed in PE samples from male fetuses may be related to regulation of the cell cycle via mammalian target of rapamycin (mTOR) and AMP-activated protein kinase (AMPK) pathways [76] (Figure 7). Considering their role in regulating cell division, growth, or autophagy [77], altered

phosphorylation and imbalance in AMPK/mTOR pathways can lead to a defective cell cycle and also erythroid maturation [78–81].

**Figure 7.** Sex-specific disruption of fetal erythropoiesis in PE. The interplay of the intracellular pathways that could disrupt fetal erythroid maturation has been demonstrated. A magnified erythroid cell with various organelles is used to illustrate the altered pathways in detail. aa: amino acid; SLC7A5 (LAT1): solute carrier family 7 member 5; DNAJB1: DnaJ heat shock protein family (Hsp40) member B1; HSPA1A: heat shock protein family A (Hsp70) member 1A; HSPA1A: heat shock protein family A (Hsp70) member 1B; CAMK1D (CKLiK): calcium/calmodulin dependent protein kinase ID; LKB1 (STK11): serine/threonine kinase 11; AMPK: AMP-activated protein kinase; mTORC1: mammalian target of rapamycin complex1; WRNIP1: Werner helicase interacting protein 1; FAM83D: family with sequence similarity 83 member D; CENPF: centromere protein F; KLHDC8B: kelch domain containing 8B; RRAGA: Ras related GTP binding A; CDKN2D: cyclin dependent kinase inhibitor 2D; BTG2: BTG anti-proliferation factor 2.

Among the female fetuses, decreased metabolism and increased cellular response to stress were the major pathways altered in the PE UCB erythroid cells. Genes that were important in protein processing and calcium homeostasis in Golgi, as well as calcium/calmodulin-dependent protein kinase 1D (CAMK1D) were downregulated in PE. The CAMK1D has been shown to control calcium-induced apoptosis in serum-deprived erythroleukemia cells in vitro [82]. Also, an upstream inducer of AMPK (STK11/LKB1) and amino acid-transporter required for mTORC1 activity (SLC7A5) [83] were lower in UCB erythroid cells from female fetuses born to PE pregnancies. These results along with the upregulated heat shock proteins imply a disturbance in protein processing that might affect cell maturation and survival (Figure 7). Considering that the activity of mTOR and AMPK proteins are regulated by phosphorylation at protein levels, future experiments are required to confirm and evaluate possible phosphorylation changes in erythroblasts in both male and female fetuses from PE pregnancies to indicate any sex-specific associations.

Interestingly, a surface glycoprotein was differentially expressed in each sex group in PE vs. normotensive UCB erythroid cells. Among the male fetuses, the expression of CD99, a glycoprotein associated with the Xg blood group, was upregulated in PE cases. While CD99 is located on the

pseudoautosomal areas of sex chromosomes, its level of expression varies during development and based on sex [84]. On the other hand, the female fetuses from PE pregnancies indicated a significant decrease in expressing RP11-342M1.3, which is an antisense to erythroblast membrane associated protein (ERMAP), the surface glycoprotein known for the Scianna blood group [84]. Gene regulation by antisense expression can take place at di fferent layers [85]. Considering that the expression of ERMAP was not significantly di fferent between female PE vs. normotensive erythroid cells, it seems very likely that RP11-342M1.3 has a trans-regulatory e ffect on ERMAP mRNA. Other candidate genes for future studies include SLC25A6, MTRNR2L1, and CD36 that were determined in various analyses to be di fferentially expressed in PE and in a sex-specific manner. Further studies are required to confirm the possible sex-specific e ffects and outcomes of particular changes in fetal erythropoiesis in PE pregnancies.

There is a rising interest in evaluating the e ffect of various pregnancy complications on fetal development. Long-term analyses of PE pregnancies demonstrate several health problems among the o ffspring [86], with behavioral and cognitive dysfunction [87] having been a main focus. On the other hand, more studies are shedding light on the significance of fetal sex in regulating various aspects of a pregnancy, from maternal adaptation [41] to placental and fetal gene expression [88,89] and metabolism [90]. Sex-specificity of the placenta function and structure [91] as well as fetal and placental metabolism [90] have been suggested to underlie the higher vulnerability of male fetuses to pregnancy complications, such as obesity or PE [46,92]. The data in this study suggests that besides the nervous system, PE also a ffects the development of the fetal hematopoietic system. This is mainly through the triggering of transcriptional and proteomic changes in fetal HSPCs and erythroblasts that may disrupt erythroid maturation, explaining the higher UCB erythroblast count in these pregnancies. Our work also suggests that the changes observed in the molecular pathways are more severe among the male compared to the female fetuses. This is in line with higher adverse outcomes in pregnancies with male fetuses [39,40], as well as an increased risk of diseases of the blood and blood-forming organs, such as anemia, among the male children born to PE mothers [93].
