**2. Results**

#### *2.1. Preeclampsia Does Not Alter Migration*/*Homing or Di*ff*erentiation Capacity of UCB HSPCs*

To study the effect of PE on migration/homing of fetal HSPCs, the expression of known surface adhesion molecules (SAMs) was analyzed on the UCB CD34+ CD45+ cells (Figure 1A). The expression of CD49d, CD49e, CD184 (CXCR4), and CD11a (the upper panel Figure 1B) as well as CD44 and CD62L (L-selectin) (the lower panel Figure 1B) were not significantly different between the PE and normotensive groups (flow cytograms demonstrated in Supplementary Figure S1). In addition, the frequency of viable hematopoietic stem cell-enriched cells (HSCs) expressing specific markers (Table 1) was determined using flow cytometry (Figure 1C). The number of CD34+ cells or the HSCs per mL of UCB obtained from each sample was not significantly different between the PE and normotensive samples.

**Figure 1.** Flow cytometry analysis and isolation of UCB HSPCs and HSCs as well as assessment of surface adhesion molecules (SAMs) and in vitro erythroid differentiation capacity of the cells isolated from PE vs. normotensive (NO) pregnancies. (**A**) Flow cytometry analysis showing the UCB HSPC population gated based on size and granularity (FSC-A and SSC-A) and CD34+ CD45+ expression. (**B**) Demonstrating the median fluorescent intensity (MFI) for various SAMs in red (PE, *n* = 5) and black (NO, *n* = 10); despite large differences in some MFI values, the differences were not statistically significant. (**C**) Flow cytometry analysis of the HSC population from UCB samples; the population was gated (from left to right) based on size and granularity followed by CD34<sup>+</sup>, CD38lo, and CD45RA<sup>−</sup>, CD90+ expression. As previously reported by others, the CD34+ CD38lo population was very small in the majority of our samples. This specific individual sample with a large CD34+ CD38lo population was particularly chosen for specifically visualizing a clearly distinct CD34+ CD38lo CD45RA− and CD90+ population in the figure. (**D**) Example of BFU-Es in culture (10× magnification) from normotensive (*n* = 8) and PE (*n* = 7) samples after the UCB CD34+ cells were cultured for 14 days. No significant difference was observed BFU-E count comparison between PE and normotensive groups.



To investigate whether PE altered the erythroid differentiation capacity of fetal HSPCs, resulting in enhanced erythropoiesis and UCB erythroblast count, CD34+ cells from normotensive and PE UCB samples were isolated and used in a colony formation assay. Despite a large difference in the median values between the groups, the number of burst forming units-erythroid (BFU-Es) showed no significant difference (Figure 1D).

#### *2.2. Preeclampsia A*ff*ects the Gene Expression in UCB HSPCs*

Since there was no difference in the migration/homing or erythroid differentiation capacity of the UCB HSPCs (CD34+ CD45+ cells) that could explain the higher UCB erythroblast count documented in PE [18–21], cDNA subtractive hybridization was carried out to elucidate possible gene expression differences that might affect the maturation of erythroblasts. To perform cDNA subtractive hybridization, CD34+ CD45+ cells from normotensive and PE UCB samples were used as driver and tester groups, respectively. Sequencing of the differential fragments resulted in 26 protein-coding genes (Figure 2). Predictions by String suggested that the eukaryotic translation elongation factor 1 alpha 1 (EEF1A1) interacted with glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as well as several ribosomal proteins (RPs). The pathways of significance based on GSEA are presented in Supplementary Table S1. The RPs were also associated with significant hematological phenotypes, such as increased mean corpuscular volume, macrocytic anemia, persistence of HbF, and reticulocytopenia, as determined by ToppFun (false discovery rate (FDR) < 0.01).

**Figure 2.** Gene expression analysis in the UCB HSPCs using cDNA subtractive hybridization. The procedure is demonstrated from separating mononuclear cells (MNCs) from the umbilical cord blood (UCB) and isolating hematopoietic stem/progenitor cells (HSPCs) to the final list of genes that were found to be different in PE. The HSPCs (CD34+ CD45+) were sorted during SAM analysis and were used in this experiment (PE, *n* = 5 and NO, *n* = 10).

#### *2.3. Preeclampsia Is Associated with Changes in Metabolic and Protein Synthesis Pathways of In Vitro Di*ff*erentiated Erythroid Cells*

To investigate whether changes in ribosomal and metabolic pathways in the HSPCs affected late erythroid maturation steps in fetuses from PE pregnancies, proteomics analysis was performed using TMT-mass spectrometry on in vitro differentiated erythroid cells. After mapping the peptide sequences to proteins, 6222 proteins were detected (FDR ≤0.01) (Supplementary Table S2). At a threshold of fold change ≥20% and *p* value ≤0.05, a total of 90 proteins were increased and 14 proteins were decreased in PE vs. normotensive in vitro differentiated erythroid cells (Supplementary Table S2). The heat map of the differentially expressed proteins and the enriched pathways predicted by GSEA are presented in Figure 3. The protein–protein interaction network and the connection between the enriched pathways are presented in Supplementary Figure S2. The affected pathways were mainly related to ATP production (oxidative phosphorylation and the TCA cycle), as well as protein synthesis, transport, and metabolism (Figure 3).


**Figure 3.** The proteomics analysis heat map and the enriched pathways in the in vitro differentiated erythroid colonies. (**A**) The heat map for the significantly differentially expressed proteins in PE (*n* = 5) vs. normotensive (*n* = 5) in vitro differentiated erythroid cells. (**B**) The enrichment analysis in the gene set analysis in CPDB human network was performed based on the protein average fold ratio in PE and normotensive samples.

#### *2.4. Preeclampsia Does Not Alter the UCB Profile of Terminally Di*ff*erentiating Erythroblasts*

Considering that the in vitro analyses indicated no changes in molecular pathways rather than erythroid cell production, the frequency of various stages of terminally differentiating erythroid cells was investigated in the UCB erythroblasts between male and female fetuses from PE and normotensive pregnancies. The viable single cells were gated based on GPA and CD45 expression. The CD45<sup>−</sup>, GPA+ erythroid population was analyzed for surface expression of CD49d and Band 3 to evaluate the terminal erythroid differentiation stages of the UCB erythroblasts (Figure 4A). The erythroid precursors present in the samples were predominately basophilic erythroblasts II to orthochromatic erythroblasts. Comparing the erythroid profile of the samples, no significant differences were observed between the venous or arterial UCB from PE or normotensive pregnancies in male nor female fetuses (Figure 4B).

**Figure 4.** Comparison of the erythroid progenitor profile in the umbilical cord blood (UCB) from PE (*n* = 6) and normotensive (*n* = 7) pregnancies. (**A**) The flow cytometry profile of CD45− GPA+ cells expressing CD49d (Integrin a4) and Band 3 protein during their maturation from proerythrocytes (CD49d+ Band 3−) to mature erythrocytes (Cd49d− Band 3<sup>+</sup>). (**B**) No significant differences were observed between arterial or venous UCB from male or female in PE vs. normotensive pregnancies.

#### *2.5. Gene Expression Di*ff*erences between Male vs. Female Samples Are Irrespective of Pregnancy Outcome*

The absence of a significant increase in the frequency of immature erythroid cells in PE UCB samples was in line with the results from the in vitro differentiation cultures. Therefore, further RNA-sequencing analysis was performed to explore possible changes in the molecular pathways that may explain the higher erythroblast count. Gene expression of arterial vs. venous erythroid cells did not differ significantly (Supplementary Figure S3). Sample clustering by principal component analysis (PCA) indicated a major difference in gene expression in the male vs. female erythroid cells irrespective of pregnancy outcome (Figure 5A). A total of 35 genes were determined by EdgeR and Deseq2 to be differentially expressed (DE) in male vs. female fetuses, affecting pathways, such as RNA transcription (Supplementary Table S3) (Figure 5B). A more distinct clustering of PE vs. normotensive samples was observed in the samples from male fetuses (Figure 5A). Based on the DE genes confirmed by Deseq2 and EdgeR, 40 genes that affected metabolism and protein processing in endoplasmic reticulum (ER)/vesicle trafficking were downregulated in PE (Figure 5C). In addition, 21 genes that were involved in pathways, such as heat shock response and protein kinase activation by RHO GTPases, were upregulated (Figure 5C).

**Figure 5.** Primary gene expression analysis of the UCB erythroid progenitors. (**A**) A distinct clustering was observed between male and female samples in the principal component analysis (PCA) plot. Also, among the PE (*n* = 6) and normotensive (*n* = 7) samples, a clearer clustering was observed in the male compared to female fetuses. The heat maps of the differentially expressed genes and the related pathways are presented when comparing (**B**) male vs. female and (**C**) PE vs. normotensive samples. Shades of blue and red indicates down- or up-regulated genes/pathways, respectively.

#### *2.6. E*ff*ects of PE on Gene Expression in UCB Erythroid Cells Are Sex-Specific*

Taking into account the sex-specific clustering of the samples from the RNA-sequencing analysis, the effect of PE on male and female samples was analyzed separately. In the males, a total of 40 DE genes were determined in PE vs. normotensive groups by Deseq2 and EdgeR (Supplementary Table S3). The affected pathways included endocytosis, protein ubiquitination, regulation of cell cycle, and convergen<sup>t</sup> extension, i.e., the process of cell lengthening and narrowing along one axis (Figure 6A). Among the females, a total of 21 DE genes were confirmed in the PE group using Deseq2 and EdgeR (Supplementary Table S3). The altered pathways included metabolism, mTOR signaling, and cellular response to stress (via heat shock proteins) (Figure 6B).

**Figure 6.** Sex-specific transcriptome comparison in the UCB erythroid progenitors from PE vs. normotensive pregnancies. (**A**,**B**) The heat maps of the differentially expressed genes and the related pathways are presented when comparing PE vs. normotensive sample from (**A**) male and (**B**) female fetuses. Shades of blue and red indicate down- or up-regulated genes/pathways, respectively.
