Bhlhe40 Regulates Proliferation and Angiogenesis in Mouse Embryoid Bodies under Hypoxia

Abstract

Knowledge of the molecular mechanisms that underlie the regulation of major adaptive responses to an unbalanced oxygen tension is central to understanding tissue homeostasis and disease. Hypoxia-inducible transcription factors (HIFs) coordinate changes in the transcriptome that control these adaptive responses. Here, we focused on the functional role of the transcriptional repressor basic-helix-loop-helix family member e40 (Bhlhe40), which we previously identified in a meta-analysis as one of the most consistently upregulated genes in response to hypoxia across various cell types. We investigated the role of Bhlhe40 in controlling proliferation and angiogenesis using a gene editing strategy in mouse embryonic stem cells (mESCs) that we differentiated in embryoid bodies (EBs). We observed that hypoxia-induced Bhlhe40 expression was compatible with the rapid proliferation of pluripotent mESCs under low oxygen tension. However, in EBs, hypoxia triggered a Bhlhe40-dependent cell cycle arrest in most progenitor cells and endothelial cells within vascular structures. Furthermore, Bhlhe40 knockout increased the basal vascularization of the EBs in normoxia and exacerbated the hypoxia-induced vascularization, supporting a novel role for Bhlhe40 as a negative regulator of blood vessel formation. Our findings implicate Bhlhe40 in mediating key functional adaptive responses to hypoxia, such as proliferation arrest and angiogenesis.

1. Introduction

Oxygen homeostasis is essential to sustain complex organisms. An imbalance between oxygen consumption and demand (also referred to as hypoxia) triggers multiple adaptive responses at the cellular and organismal levels [1]. Knowledge of the cellular and molecular adaptive responses to hypoxia is central to understanding tissue homeostasis and the progression of highly prevalent pathologies such as cardiovascular diseases and cancer. Hypoxia-inducible transcription factors (HIFs) are a family of heterodimeric transcription factors composed of alpha (HIFα) and beta (HIFβ) subunits. The stability and transcriptional activity of HIFα is regulated by oxygen levels, while HIFβ remains unchanged [1]. HIFs coordinate changes in a large fraction of the transcriptome that vary between cell types and underpin the adaptive cellular responses to acute or intermittent hypoxic conditions [2]. Accordingly, HIF loss of function approaches abolish both activation and repression of gene expression by hypoxia. Although HIF directly controls gene upregulation, gene downregulation in hypoxia is controlled by indirect mechanisms that remain incompletely understood [3,4,5]. By integrating data from 43 individual RNA-sequencing (RNAseq) studies conducted across 34 distinct cell types, we have recently discovered a signature comprising 291 ubiquitously expressed genes that are consistently and robustly (FDR < 0.01 and |log₂FC| > 0.7) regulated by hypoxia [6]. Notably, the transcriptional repressor basic-helix-loop-helix family member e40 (Bhlhe40) emerged as one of the most consistently upregulated genes within this signature. In this study, we aim to analyze the role of the transcriptional repressor Bhlhe40 in limiting two key adaptive responses to hypoxia: cell proliferation and angiogenesis.

Cell growth and division are energy-intensive processes. It is, therefore, not surprising that hypoxia restricts the proliferation of many cell types. HIFs downregulate key positive modulators of the G1-S transition, while negative regulators such as Cyclin Dependent Kinase (CDK) inhibitors are upregulated [7,8]. Several genes involved in the control of DNA replication initiation are also repressed by HIFs [9]. Thus, an important part of the transcriptional response coordinated by HIFs is directed toward limiting cell cycle entry and DNA replication in the context of an unbalanced oxygen tension. Despite the generality and relevance of this response, the molecular knowledge of the mechanisms involved is still incomplete [10].

Angiogenesis, the primary mechanism that enables vascular expansion in the adult, is a fundamental adaptive response to hypoxia that aims to restore oxygen levels compatible with cell viability and functionality [11,12]. Vasculogenesis, or de novo vessel formation by differentiation of endothelial progenitor cells, also contributes to neovascularization in physiology and pathology and is activated by hypoxia. Emerging evidence indicates that endothelial cell (EC) proliferation, migration, and differentiation must be properly integrated and controlled to establish a functional vascular network. Both deficient and excessive EC proliferation result in a dysfunctional vasculature [13,14,15]. Furthermore, EC specialization into tip and stalk phenotypes and maturation must be strictly controlled for optimal functionality of the expanded vascular network. We have recently shown that hypoxia-induced endothelial cell cycle arrest is compensated by the induction of progenitor cell differentiation during angiogenesis [9]. However, the mechanisms that integrate proliferation arrest and expansion of the vascular network in hypoxia remain to be elucidated. Given the prominent role of Bhlhe40 in the hypoxic transcriptional profile and its implication in controlling proliferation and differentiation in different cellular contexts [16,17,18,19], we hypothesized that Bhlhe40 controls proliferation and prevents excessive angiogenesis under hypoxic conditions. To test this hypothesis, we used a loss-of-function approach by CRISPR gene editing to generate Bhlhe40 knockout mouse embryonic stem cell (mESC) lines that we differentiated in embryoid bodies (EBs) as a model of vascular development [20,21,22].

Our results support that Bhlhe40 restricts cell proliferation in hypoxia in a cell fate-dependent manner. Although hypoxia-induced Bhlhe40 expression is compatible with the rapid proliferation of pluripotent mESCs, it efficiently restricts the proliferation of most progenitor cells and ECs in EBs under hypoxic conditions. Furthermore, the knockout of Bhlhe40 in EBs increased basal angiogenesis and exacerbated hypoxia-induced angiogenesis. Taken together, our results support the role of hypoxia-induced Bhlhe40 in inhibiting cell proliferation and preventing excessive angiogenesis in mouse EBs under oxygen-limiting conditions.

2. Results

2.1. The Transcriptional Repressor Bhlhe40 Is Robustly Induced by Hypoxia in a Meta-Analysis of Transcriptomic Studies

In our recent meta-analysis of transcriptomic studies, which included 430 RNA-seq samples from 43 individual studies across 34 different cell types, we assessed the impact of hypoxia on the transcription of 20,918 genes identified across the datasets [6]. This analysis yielded a hypoxic transcriptomic profile, assigning the estimated effect size of hypoxia (log₂-Fold Change (log₂FC)) and statistical significance for the change in expression to each of the 20,918 genes. From these data, we generated a hypoxia signature comprising 291 ubiquitously expressed genes whose expression underwent a significant (False Discovery Rate (FDR) < 0.01) and robust (|log₂FC| > 0.7) regulation in response to hypoxia [6]. This hypoxia signature includes Bhlhe40, which is consistently the most hypoxia-induced transcriptional regulator across studies and cell types (Figure 1A). Its function as a transcriptional repressor in hypoxia was confirmed by the enrichment of Bhlhe40 binding in the regulatory regions of hypoxia-repressed genes [23].

Next, we interrogated whether genes implicated in cell cycle activation were enriched among genes repressed by hypoxia in the meta-analysis. Overlaying a recently published proliferative signature comprising 157 genes [24] onto the transcriptomic profile ranked by the log₂FC values revealed enrichment of this proliferative signature among hypoxia-repressed genes (Figure 1B). Furthermore, there was an overlap of 20 genes between the hypoxia (291 genes) and proliferative (157 genes) signatures, whose expression is repressed according to the meta-analysis (Figure 1C).

Functionally, Bhlhe40 has been implicated in regulating cell proliferation and differentiation in various cell types [16,17,18,19]. Therefore, based on the conclusions from our meta-analysis and the published functional role of Bhlhe40, we hypothesize that hypoxia-induced Bhlhe40 mediates the arrest of cell proliferation under hypoxic conditions. In this study, we have tested this hypothesis and its relationship to hypoxia-induced angiogenesis using stem cell models of vascular development.

2.2. Generation of Inducible CRISPR-Bhlhe40 Knockout Mouse Embryonic Stem Cell Lines

To address experimentally whether Bhlhe40 is involved in regulating cell cycle arrest and angiogenesis under hypoxic conditions, we utilized a loss-of-function approach in mESC that we differentiated in EBs. Stem cell-based models provide a unique opportunity to investigate the impact of external signals, such as hypoxia, on the regulatory mechanisms that control the transition from pluripotency to differentiation [25]. In addition, we used a shRNA interference strategy in human umbilical vein endothelial cells (HUVEC). The experimental approaches, models, and steps are summarized in Figure 2A.

First, we analyzed the kinetics of hypoxia-induced Bhlhe40 expression in the R3.8^Cas9 mESC line [26], which was not included in our previous meta-analysis. We observed that the expression of Bhlhe40 is rapidly induced when mESCs are exposed to hypoxia, showing a six-fold increase within 2 h (Figure 2B). The expression of Bhlhe40 peaked at approximately 16 times the basal expression between 4 and 8 h of hypoxia and remained elevated 16 h after exposure to hypoxic conditions. The kinetics of Bhlhe40 induction by hypoxia in mESCs was similar to that observed for Egln3, a prototypical gene induced by hypoxia in many cell types (Figure 2C).

We utilized the R3.8^Cas9 mESC line [26] to generate the doxycycline-inducible iBhlhe40^ΔSg R3.8^Cas9 mESC line in which M2-rtTA transactivator is constitutively expressed from the Rosa26 promoter, Cas9 expression is induced by doxycycline (DOX) treatment through activation of a Tet-O sequence inserted into the Col1a1 locus, and Bhlhe40-specific single guide RNAs (sgRNAs) are expressed from the U6-P promoter from the lentivirus used to transduced the R3.8^Cas9 mESCs with the sgRNAs (Figure 2D).

We used the Breaking-Cas9 bioinformatics tool [27] to predict the best sgRNAs for the CRISPR editing of the Bhlhe40 gene. We chose five sgRNAs that target exons 1–3 based on their high scores (Supplementary Figure S1A). We then quantified the effectiveness of these sgRNAs to knock down Bhlhe40 expression using qRT-PCR (Supplementary Figure S1B). The most effective sgRNA (sgRNA-4) was used in six editing experiments to confirm its efficacy in abrogating the induction of Bhlhe40 expression by hypoxia using qRT-PCR (Figure 2E). The efficiency of the sgRNA-4 in suppressing the induction of Bhlhe40 expression by hypoxia varied between 90.0% and 98.8%. This result was confirmed by western blot analysis (Figure 2F). The more efficient inducible cell line (iBhlhe40^ΔSg4 R3.8^Cas9) was used to generate control (Bhlhe40-Ctrl mESC (−DOX)) and Bhlhe40-knockout (Bhlhe40-KO mESC (+DOX)) mESC lines for further functional experiments.

2.3. Hypoxia-Induced Bhlhe40 Expression in Mouse Pluripotent Stem Cells Is Compatible with Their Rapid Proliferation

Given the induction of Bhlhe40 expression by hypoxia in mESCs (Figure 2A), we investigated the impact of hypoxia treatment on the proliferation of the R3.8^Cas9 mESC line. Figure 3A,B shows that the rapid proliferation rate of pluripotent R3.8^Cas9 mESCs was not significantly affected by hypoxia (Hx, 1% O₂), with a trend toward faster proliferation in hypoxic conditions.

The doubling time of R3.8^Cas9 mESCs was very similar in normoxic (N) and hypoxic (Hx) conditions (14.75 ± 0.45 h in N vs. 12.63 ± 0.73 h in Hx), with a trend toward lower doubling time values in hypoxia (Figure 3B). These results were verified in the R1 mESC line (Figure 3C,D). The doubling time of R1 mESCs was very similar in normoxic and hypoxic conditions (15.72 ± 1.92 h in N vs. 14.42 ± 1.00 h in Hx), with a trend toward lower doubling time values in hypoxia.

We confirmed the previous results by analyzing EdU incorporation by flow cytometry (FACS). Hypoxic conditions did not alter the percentage of S-phase cells quantified by FACS in the R3.8^Cas9 (Figure 3E) and R1 (Figure 3F) mESC lines.

Consistent with these findings, hypoxia did not significantly change the mRNA expression levels of cyclin D1 (Ccnd1), a critical positive regulator of the G1/S transition (Figure 3G). Additionally, we observed that the knockout of Bhlhe40 did not significantly change Ccnd1 expression levels in normoxic or hypoxic conditions (Figure 3G).

In conclusion, these results show that hypoxia-induced Bhlhe40 expression in mouse pluripotent cells is compatible with their rapid proliferation.

2.4. The Knockout of Bhlhe40 Prevents Hypoxia-Induced Cell Cycle Arrest in Most Progenitor Cells and Mature Endothelial Cells in Mouse Embryoid Bodies

To investigate the role of Bhlhe40 in mediating hypoxia-induced cell cycle arrest and angiogenesis, we differentiated pluripotent mESCs in EBs. This allows the development of progenitor cells and mature differentiated ECs organized in a vascular network [21,22]. To evaluate the impact of hypoxia treatment on EB growth and differentiation, we used a serum-containing medium without any added exogenous differentiation-inducing factors.

We made use of the doxycycline-inducible mESC line iBhlhe40^ΔSg4 R3.8^Cas9 to generate Bhlhe40-control (Bhlhe40-Ctrl mESC (−DOX)) and Bhlhe40-knockout (Bhlhe40-KO mESC (+DOX)) mESC lines (Supplementary Figure S1A,B). These cell lines were then differentiated without pluripotency factors (-LIF and -2i) in EBs generated and maintained in hanging drops from day 0 to day 4 (Figure 4A). From day 4 to day 10, the differentiation was continued under adhesion conditions. By day 10, progenitors coexisted with differentiated ECs organized in vascular structures in the EBs. In our experimental conditions for EB differentiation, the longest duration of 1% O₂ hypoxia treatment without compromising cell viability was 48 h. Considering all these facts, our experimental design included a 48-h hypoxia treatment of the EBs from differentiation day 8 to day 10, compared to normoxic conditions (Figure 4A).

We initially analyzed the effect of hypoxic treatment on Bhlhe40 expression in the EBs by qRT-PCR. Figure 4B shows that 16 h of hypoxia treatment in EBs differentiated for 8 days induced a 7-fold increase in Bhlhe40 mRNA. Prolonged hypoxia treatments of the EBs for 48 h also increased Bhlhe40 mRNA, although the fold change was smaller (Figure 4B). Additionally, we verified that the Bhlhe40 CRISPR knockout induced by doxycycline treatment in the iBhlhe40^ΔSg4 R3.8^Cas9 mESC line persisted in the EBs after 10 days of differentiation (Figure 4C).

To determine the percentage of cells in the S-phase in the EBs, we performed FACS analysis of EdU incorporation on the total cells disaggregated from the EBs (Figure 4D,E). The results showed that 48 h of hypoxia treatment from differentiation day 8 to day 10 decreased the percentage of S-phase cells compared to normoxia (fold % EdU N Bhlhe40-Ctrl (−DOX) over % EdU Hx Bhlhe40-Ctrl (−DOX) was 4.1) (Figure 4D). Remarkably, the Bhlhe40 knockout attenuated the decrease in S-phase cells induced by hypoxic conditions (fold % EdU N Bhlhe40-KO (+DOX) over % EdU Hx Bhlhe40-KO (+DOX) was to 1.7) (Figure 4D).

We also observed that the Bhlhe40 knockout increased the percentage of S-phase cells in hypoxia (fold % EdU Hx Bhlhe40-KO (+DOX) over % EdU Hx Bhlhe40-Ctrl (−DOX) was 4.3) and to a lesser extent in normoxia (fold % EdU N Bhlhe40-KO (+DOX) over % EdU N Bhlhe40-Ctrl (−DOX) was 1.8) (Figure 4D). These results were confirmed in an independent editing experiment (Supplementary Figure S2). Consistent with the above-described results, the Bhlhe40 knockout decreased the percentage of G0/G1 cells in EBs under hypoxic and normoxic conditions (Figure 4D).

Previously, we demonstrated that hypoxia-induced cell cycle arrest is HIF-dependent [9]. Therefore, to serve as a control for bypassing hypoxia-induced cell cycle arrest, we generated a HIFβ (Arnt) knockout inducible cell line (iArnt40^ΔSg3 R3.8^Cas9). This cell line allowed us to generate by doxycycline treatment the Arnt-knockout cell line (Arnt-KO mESC (+DOX)), and without doxycycline, the Arnt-control cell line (Arnt-Ctrl mESC (−DOX)) (Supplementary Figure S1C–E). As expected, the Arnt knockout attenuated the cell cycle arrest in hypoxic conditions (fold % EdU N Arnt-Ctrl (−DOX) over % EdU Hx Arnt-Ctrl (−DOX) was 5.4; whereas fold % EdU N Arnt-KO (+DOX) over % EdU Hx Arnt-KO (+DOX) was 2.1) (Figure 4E). Therefore, the effect of the Arnt knockout on hypoxia-induced cell cycle arrest in the total cells disaggregated from the EBs was similar to that of the Bhlhe40 knockout.

The experimental conditions we used for EB differentiation primarily promoted mesodermal development, resulting in EBs composed mainly of mesodermal progenitors and a smaller number of mature ECs, along with other types of progenitors and differentiated cells [21,22]. Therefore, we analyzed the effect of hypoxia and Bhlhe40 knockout on the percentage of S-phase cells in mesodermal progenitors and ECs in vascular structures in the EBs.

We focused our analysis on a cluster of mesodermal progenitors we recently identified using single-cell RNA sequencing (scRNA-seq) [28]. This cluster is interesting in this study because it could be a source of progenitors that differentiate into ECs in EBs under hypoxic conditions [28]. We quantified the percentage of S-phase cells in microscopy fields of the EBs containing HOXD9⁺ mesodermal progenitors and ECs in vascular structures (Supplementary Figure S3).

In contrast with the expected results, we observed that hypoxia did not change the percentage of S-phase HOXD9⁺ mesodermal progenitors (fold % EdU⁺ HOXD9⁺ cells N Bhlhe40-Ctrl (−DOX) over Hx Bhlhe40-Ctrl (−DOX) was 1.0). The knockout of Bhlhe40 also did not affect the percentage of S-phase HOXD9⁺ mesodermal progenitors in normoxia (fold % EdU⁺ HOXD9⁺ cells N Bhlhe40-KO (+DOX) over N Bhlhe40-Ctrl (−DOX) was 1.0) or hypoxia (fold % EdU⁺ HOXD9⁺ cells Hx Bhlhe40-KO (+DOX) over Hx Bhlhe40-Ctrl (−DOX) was 1.1) (Figure 5A,B).

The HOXD9⁺ population represents approximately 13% of the total cells in the EBs under normoxic conditions [28] and may help to explain the partial effect observed in the Bhlhe40 knockout (Figure 4D). However, the magnitude of the impact of the Bhlhe40 knockout on the FACS analysis of EdU incorporation in the total cells of the EBs (Figure 4), together with the fact that HOXD9⁻ progenitors make up the majority of cells in the EBs [28], suggests that other types of progenitors in the EBs undergo a Bhlhe40-dependent cell cycle arrest under hypoxic conditions.

We confirmed the results of the FACS analysis of EdU incorporation in the total cells disaggregated from the EBs (Figure 4D) through immunofluorescence and three-dimensional Imaris analysis (Figure 5A–C). Hypoxia decreased the percentage of S-phase cells (fold % EdU⁺ total cells/field N Bhlhe40-Ctrl (−DOX) over Hx Bhlhe40-Ctrl (−DOX) was 1.6). The knockout of Bhlhe40 significantly increased the percentage of EdU⁺ total cells under hypoxic conditions (fold % EdU⁺ total cells/field Hx Bhlhe40-KO (+DOX) over Hx Bhlhe40-Ctrl (−DOX) was 2.4) and to a lesser extent under normoxic conditions (fold EdU⁺ total cells/field N Bhlhe40-KO (+DOX) over N Bhlhe40-Ctrl (−DOX) was 1.4).

We have previously reported that hypoxic treatment induced a cell cycle arrest of ECs in the vascular structures of the EBs [9]. Here, we investigated the impact of Bhlhe40 knockout on the percentage of S-phase ECs in vascular structures of the EBs differentiated in normoxia for 10 days or treated in hypoxia for the last 48 h. Figure 6A shows representative images of the vascular structures identified by double immunofluorescence using a membrane marker (CD31, red) and a nuclear marker (ERG, green) in all experimental conditions.

The quantification of EdU incorporation by image analysis of CD31⁺ ERG⁺ ECs in vascular structures in the EBs revealed that Bhlhe40 knockout partially prevented the hypoxia-induced inhibition of S-phase entry (fold % EdU⁺ ECs N Bhlhe40-Ctrl (−DOX) over Hx Bhlhe40-Ctrl (−DOX) was 3.1, whereas fold % EdU⁺ ECs N Bhlhe40-KO (+DOX) over Hx Bhlhe40-KO (+DOX) was 1.6) (Figure 6B). Furthermore, Bhlhe40 knockout increased the percentage of S-phase cells under hypoxic conditions (fold % EdU⁺ ECs Hx Bhlhe40-KO (+DOX) over Hx Bhlhe40-Ctrl (−DOX) was 2.1).

These results support that the cell cycle arrest of ECs in vascular structures in the EBs under hypoxic conditions is mediated by Bhlhe40.

We previously reported a hypoxia-induced cell cycle arrest in human umbilical vein endothelial cells (HUVEC) [9]. Here, we investigated the BHLHE40 dependency of hypoxia-induced cell cycle arrest in HUVEC using an interference strategy based on short harping RNAs (shRNAs) to silence the expression of BHLHE40 (Supplementary Figure S4A). We tested two different shRNA sequences specific for BHLHE40 (shBH-1 and shBH-5) independently or in combination, compared to a scramble sequence (shScr). Quantification by qRT-PCR showed that interference by shBH-5 was more efficient than shBH-1, and the combination of shBH-1 and shBH-5 was not more efficient than shBH-5 alone (Supplementary Figure S4B). These results were confirmed by western blot analysis of BHLHE40 protein levels (Supplementary Figure S4C). We next examined the effect of BHLHE40 interference using shBH-5 on EdU incorporation quantified by immunofluorescence in HUVEC exposed to normoxic or hypoxic conditions for 48 h (Supplementary Figure S4D,E). We observed that BHLHE40 interference did not affect EdU incorporation in normoxic or hypoxic conditions in HUVEC.

The difference in Bhlhe40 dependence between HUVEC and mouse ECs may be due to differences in the level of differentiation achieved in the two-dimensional culture of HUVEC compared to the three-dimensional growth of mouse ECs in vascular structures in EBs. Less differentiated ECs in the EBs are likely more dependent on Bhlhe40, similar to progenitor cells in the EBs.

2.5. Bhlhe40 Is a Negative Regulator of Blood Vessel Formation in Embryoid Bodies

In addition to regulating cell proliferation, Bhlhe40 also controls the differentiation of various cell types [16,17,18,19]. Therefore, we investigated the effect of Bhlhe40 knockout on the formation of hypoxia-induced vascular structures in the EBs. Blood vessel formation in the EBs involves the proliferation of ECs within pre-existing vessels to form new vessels (angiogenesis) and the differentiation of endothelial progenitor cells that integrate into the expanding vasculature (vasculogenesis). This model, therefore, provides an opportunity to test the effect of the Bhlhe40 knockout on the integrated regulation of proliferation and differentiation during blood vessel formation under hypoxic conditions.

We have quantified angiogenesis in control EBs (Bhlhe40-Ctrl (−DOX)) and Bhlhe40 knockout EBs (Bhlhe40-KO (+DOX)) that were differentiated for 10 days and exposed to hypoxia (1% O₂) or normoxia (21% O₂) for the last 48 h (Figure 4A). Figure 7A shows the vessel network visualized by CD31 immunostaining in representative EBs from all experimental conditions.

As expected, hypoxic conditions increased the vascularization of the EBs (number of EBs with high-medium angiogenesis (Figure 7B), total vessel length (Figure 7C), and branching points (Figure 7D) compared to normoxic conditions. Interestingly, the knockout of Bhlhe40 significantly increased the percentage of EBs exhibiting high and medium angiogenesis under hypoxic and normoxic conditions (Figure 7B). The increase in basal angiogenesis by Bhlhe40 knockout in normoxia could be explained by regions in the EBs where the three-dimensional growth generates hypoxic areas where Bhlhe40 could be induced. As hypoxia is a potent inducer of angiogenesis in the EB model, the effect of the Bhlhe40 knockout further increasing the vascularization of the EBs is more challenging to observe, and it did not reach statistical significance in the AngioTool parameters (Figure 7C,D). To overcome the limitations in the analysis by microscopy, we performed a FACS analysis of a more significant number of EBs per experimental condition using CD144 as an EC marker.

By FACS analysis, we verified that the Bhlhe40 knockout increased the percentage of CD144⁺ ECs in both normoxic and hypoxic EBs (fold % CD144⁺ N Bhlhe40-KO (+DOX) over N Bhlhe40-Ctrl (−DOX) was 2.2; and fold Hx Bhlhe40-KO (+DOX) over Hx Bhlhe40-Ctrl (−DOX) was 1.8) (Figure 7E). We confirmed these results in an independent Bhlhe40 editing experiment in EBs (Supplementary Figure S5).

Hypoxia-induced angiogenesis is HIF-dependent [30,31]. Therefore, as a control of our CRISPR gene editing strategy, we knock out the positive regulator Arnt using the inducible cell line iArnt40^ΔSg3R3.8^Cas9 (Supplementary Figure S6). As expected, the Arnt knockout significantly decreased the percentage of EBs with medium and high angiogenesis under normoxic and hypoxic conditions (Supplementary Figure S6B) to a similar extent but with an opposite trend to the Bhlhe40 knockout. Furthermore, under hypoxic conditions, the Arnt knockout decreased both the total vessel length and the branching index ((junctions/mm²)/EB) (fold total vessel length Hx Arnt-Ctrl (−DOX) over Hx Arnt-KO (+DOX) was 1.7; and fold branching index Hx Arnt-Ctrl (−DOX) over Hx Arnt-KO (+DOX) was 2.5). However, although there was a trend toward decreased angiogenesis in Arnt-KO (+DOX) EBs under normoxic conditions, it did not reach statistical significance (Supplementary Figure S6C,D). These data indicate that the observed effect on angiogenesis resulting from the knockout of Bhlhe40 is of a similar magnitude but with an opposite trend to that caused by the knockout of Arnt.

Taken together these findings support a novel role for Bhlhe40 as a negative regulator of blood vessel formation.

3. Discussion

Cell proliferation, migration, and differentiation are general cellular responses controlled by external stimuli critical for development and tissue homeostasis. To achieve optimal functionality, the control exerted by negative regulators is paramount in fine-tuning each of the integrated cellular responses to the correct level. In complex processes such as angiogenesis, where multiple cellular responses must be coordinated, negative regulators are critical in engineering the vascular network to achieve optimal functionality. The identification of the role of Notch as a key negative regulator of angiogenesis [30,31] has allowed a deeper understanding of the importance of fine-tuning each response, as hyperactivation of any of the integrated biological responses leads to dysfunction and even pathological outcomes [32,33,34]. Although hypoxia is an established inducer of vascular development [35,36] and a relevant stimulus for pathological angiogenesis [11,37,38], little is known about the hypoxia-induced negative regulators that modulate the cellular responses involved. Here, we identified a novel negative regulator of hypoxia-induced neovascularization.

In a recent meta-analysis of transcriptomic studies performed by our group [6], Bhlhe40 stood out as a transcriptional repressor that is consistently upregulated by hypoxia across multiple cell types, suggesting its potential role as a modulator of hypoxia-induced adaptive responses (Figure 1). Given the universal induction of Bhlhe40 by hypoxia and its general involvement in regulating cell proliferation and differentiation [16,17,18,19], we challenged the hypothesis that Bhlhe40 functions as a negative regulator of both cell proliferation and angiogenesis during adaptation to hypoxic conditions. We used a mouse stem cell-based model of vascular development in EBs to test this hypothesis.

The EB model integrates angiogenesis and vasculogenesis under the control of the same regulatory pathways governing vascular development in organisms. Blood vessel formation in the EBs involves the proliferation of ECs (angiogenesis) and the differentiation of endothelial progenitor cells, which are incorporated into the expanding vasculature of the EBs (vasculogenesis). Therefore, this model provides an opportunity to investigate the regulatory mechanisms that underpin the integrated regulation of proliferation and differentiation during blood vessel formation under hypoxic conditions.

Bhlhe40 is robustly induced by hypoxia in mESCs (Figure 2). Therefore, we first investigated the effect of hypoxia on the S-phase entry and the proliferation rate of mESCs growing under pluripotency conditions. Although the most common effect of hypoxia in mature terminally differentiated cells is cell growth arrest [7,8,9,38,39,40,41,42], S-phase entry and the fast proliferation rate of pluripotent cells were not restricted by hypoxic conditions in the two mESC lines tested (Figure 3). We also found neither hypoxia nor Bhlhe40 knockout altered cyclin D1 mRNA levels in mESCs (Figure 3G).

Thus, the induction of Bhlhe40 by hypoxia in pluripotent mESCs is compatible with their rapid proliferation. It is likely that in pluripotent mESCs, yet-to-be-identified bypass mechanisms compensate for the action of Bhlhe40 in suppressing cell proliferation under hypoxic conditions.

Next, we investigated the Bhlhe40 dependence of hypoxia-induced cell cycle arrest as a function of cell fate in the EB model, examining the impact of Bhlhe40 knockout on progenitors and differentiated cells (Figure 2A). We analyzed a late differentiation window from day 8 to day 10, during which mature differentiated cells of different types are present in optimal numbers and coexist with progenitor cells in the EBs. The expression of Bhlhe40 mRNA was induced after 16 h of hypoxia treatment on day 8 of differentiation, and its levels remained elevated above the basal normoxic levels for up to 48 h (Figure 4B,C). Hypoxia treatment reduced the percentage of S-phase cells in control EBs (Figure 4D,E), confirming our previous results [9]. Interestingly, the knockout of Bhlhe40 attenuated hypoxia-induced cell cycle arrest (Figure 4D). The magnitude of the impact of the Bhlhe40 knockout was similar to that of the Arnt knockout (Figure 4E).

We also observed that the knockout of Bhlhe40 or Arnt affected EdU incorporation in normoxic EBs. Applying a hypoxia classifier based on transcriptional changes previously developed by our group to scRNA-seq experiments in EBs, we found that in EBs under hypoxic conditions, the majority of the cells were classified as hypoxic and that in EBs under normoxic conditions, a small number of cells were also classified as hypoxic [43]. In normoxic conditions, the fraction of cells classified as hypoxic may correspond to those located in the core regions of EBs, where oxygen is compromised due to three-dimensional growth. Based on this, the effect of Arnt or Bhlhe40 knockdown in normoxia could originate from small hypoxic regions in the EBs differentiated under normoxic conditions. In any case, the magnitude of the change caused by Arnt or Bhlhe40 knockout was greater under hypoxic conditions than under normoxic conditions (Figure 4D,E).

In independent editing experiments, we observed that Bhlhe40 knockout partially prevented hypoxia-induced cell cycle arrest in the EBs. This partial effect of Bhlhe40 knockout may result from incomplete cell cycle arrest under hypoxia and the presence in the EBs of cell types whose proliferation is not arrested under hypoxia or is independent of Bhlhe40. Using scRNA-seq analysis, we have recently shown that all these scenarios coexist in the EBs differentiated under hypoxic conditions: (i) Despite hypoxia causing a pervasive cell cycle arrest, reducing S-phase cells in almost all clusters of progenitors and mature cells, a small percentage of S-phase cells remained in all clusters [28]; (ii) While Bhlhe40 was induced by hypoxia in the majority of clusters, it was not induced in all cells of the EBs [28].

Here, we also studied a population of mesodermal progenitor cells in the EBs that expresses HOXD9 and exhibits a faster proliferation than differentiated ECs. The proliferation of HOXD9⁺ progenitor cells was not affected by hypoxia or Bhlhe40 knockout (Figure 5). This population accounts for approximately 13% of the total cells in the EBs under normoxic conditions by scRNA-seq analysis [28] and could contribute to explaining the partial effect observed upon Bhlhe40 knockout. To support these results, we performed a computational prediction using the scRNA-seq data from reference [28]. Unfortunately, the limited number of Hoxd9 reads in this sequencing dataset hindered our ability to identify the Hoxd9⁺ cell population accurately. Consequently, the population of cells with at least one read of Hoxd9 did not exhibit the high proportion of cycling cells expected for these progenitor cells observed in our EdU pulse labeling experiments (Figure 5B). Also, because of this limitation, the predicted proportion of cells in the S-phase appeared to be reduced by hypoxia in both Hoxd9-expressing and non-expressing cells according to the scRNA-seq data, which contrasts with the experimental data (Figure 5B,C). Beyond the uncertainty regarding Hoxd9⁺ cells in the scRNA-seq dataset, the discrepancy between experimental observations and computational predictions may also be because the computational approach relies on gene expression to indirectly assign cells to specific cell cycle phases. In contrast, the experimental technique used in this study provides a direct assessment of the proportion of cells in the S-phase.

Given that the HOXD9⁺ progenitors were not arrested by hypoxia, they could serve as a homeostatic reservoir to compensate for the loss of progenitor cells resulting from the induction of differentiation under hypoxic conditions.

However, our results in the total cells of the EBs (mainly composed of progenitor cells [28]) (Figure 4 and Figure 5) are compatible with a hypoxia-induced and Bhlhe40-dependent cell cycle arrest in most progenitor cells of the EBs.

Recently, a scRNA-seq analysis identified a Hoxd9⁺ Col4A1⁺ ECs cluster with progenitor characteristics predominant in the actively expanding neovasculature of tumors compared to normal samples [44]. It would be interesting to analyze whether hypoxia has a differential effect on the proliferation of Hoxd9⁺ versus Hoxd9⁻ ECs in the vascular structures of the EBs.

We next examined the effect of Bhlhe40 on hypoxia-induced cell cycle arrest in mature ECs in organized vascular structures in the EBs. Since ECs are a minority in the EBs, we quantified EdU incorporation by immunofluorescence and image analysis in confocal microscopy fields (Figure 6). The Bhlhe40 knockout attenuated the hypoxia-induced decrease of EdU incorporation in CD31⁺ ERG⁺ ECs in vascular structures in the EBs. To extrapolate this result to human cells, we used an interference strategy in HUVEC. In contrast to our results in mouse ECs in EBs, we found that shRNA interference of BHLHE40 expression in HUVEC did not prevent the cell cycle arrest induced by hypoxia (Supplementary Figure S4). The difference in Bhlhe40 dependence between HUVEC and mouse ECs may be due to variations in the level of differentiation achieved in two-dimensional culture (HUVEC) versus vascular structures in EBs. In line with this, it has been previously reported that Bhlhe40 restricts the proliferation of the earliest germinal center B cells but not early memory B cells or plasmablasts [45].

Collectively, our results indicate that Bhlhe40 plays a role in hypoxia-induced cell cycle arrest in a cell fate-dependent manner. Pluripotent mESCs and HOXD9⁺ mesodermal progenitor cells in the EBs bypassed the proliferation restriction imposed by Bhlhe40 in hypoxia by mechanisms yet to be characterized. However, the proliferation of most progenitor cells and ECs in vascular structures within the EBs was restricted by Bhlhe40 under hypoxic conditions.

Determining the mechanism by which hypoxia-induced Bhlhe40 restricts cell proliferation of progenitors and mature ECs would be relevant, as proliferation is a central response integrated into the mechanism of angiogenesis. In other cellular contexts, it has been identified that Bhlhe40 inhibits proliferation by downregulating cyclin D1 [46,47], stabilizing cyclin E [48], and downregulating ID3 [49].

The role of the HIF/Bhlhe40 axis in vascular remodeling has not been previously investigated. A potential link between Bhlhe40 and angiogenesis is supported by the reported pro-angiogenic role of PPARγ [50,51] and Bhlhe40 repression of PPARγ in other contexts [52]. PPARγ is expressed in ECs and is upregulated by hypoxia [53,54]. Tie2CrePPARγflox/flox mice showed impaired angiogenesis and vasculogenesis in vivo [51]. Silencing of PPARγ attenuates the angiogenic response of mature ECs in tube formation assays on Matrigel, and their migration, survival, and proliferation were also impaired [51].

Based on this evidence, we hypothesized that Bhlhe40 may act as a negative regulator of hypoxia-induced angiogenesis, fine-tuning proliferation, and differentiation to ensure productive angiogenesis. Consistent with this hypothesis, we found that the Bhlhe40 knockout increased the vascularization of the EBs under hypoxic conditions (Figure 7). Our results support a novel role for Bhlhe40 in fine-tuning hypoxia-induced neovascularization that may be executed through the integration of (i) reduced proliferation of progenitor and mature endothelial cells, (ii) inhibition of hypoxia-induced differentiation of progenitor cells, and (iii) inhibition by Bhlhe40 of hypoxia-induced specialization of committed endothelial cells to stalk and tip phenotypes (summarized in the scheme of Figure 7F). In agreement with the proposed inhibitory effect of Bhlhe40 on the differentiation of progenitors, Bhlhe40 overexpression in pluripotent P19 cells inhibits mesodermal differentiation [55].

To consolidate the role of Bhlhe40 as a negative regulator of angiogenesis, it would be necessary to perform the following: (i) Alternative approaches, such as Bhlhe40 overexpression; (ii) Alternative vascular models, such as vascularization of the mouse retina and other organs [56]; (iii) Transcriptomic analysis by bulk-RNAseq and sc-RNAseq to identify the molecular mediators of the functional effects of Bhlhe40 on proliferation and differentiation under hypoxic conditions.

Although the EB model allowed us to address Bhlhe40’s role in regulating physiological vascularization, it would be interesting to study its contribution to pathological angiogenesis, such as, for example, in the context of cancer. Bhlhe40 mediates the promotion or suppression of tumors depending on the cancer type (reviewed in [57]). Likely, many of the direct effects of Bhlhe40 in cancer cells are mediated by its role as an effector of p53 and TGF-b [57]. However, its impact on the components of the tumor microenvironment is less characterized. Evidence supports the role of Bhlhe40 as a regulator of the immune response [58,59,60]. The implication of Bhlhe40 in regulating tumor angiogenesis has not been explored yet. Recently, Bhlhe40 was identified as a driver of EMT in colorectal cancer [61]. Although EMT is associated with tumor angiogenesis, this study did not directly explore whether Bhlhe40 regulates tumor angiogenesis. It may also be possible that the regulatory mechanisms that operate in the context of cancer differ from those that operate in physiological conditions, which is the context of our findings on the role of Bhlhe40 in hypoxia-induced vascularization in mouse embryoid bodies.

Overall, our data implicate Bhlhe40 in controlling critical functional adaptive responses to hypoxia, such as proliferation arrest and angiogenesis.

4. Materials and Methods

4.1. Cell Lines, Culture Conditions, and Lentiviral Infection

R3.8^Cas9 mESC line was generously provided by Fernández-Capetillo [26], and R1 mESC line was generously provided by Claesson-Welsh [62]. mESCs lines were cultured on a monolayer of mitomycin C (Sigma, Burlington, MA, USA, M4287) inactivated Mouse Embryonic Fibroblasts (MEFs) or on gelatin-coated multiwell plates (Sigma, G1890) in mESC cell medium: Dulbecco’s Modified Eagle Medium (DMEM) (Gibco, Carlsbad, CA, USA, 41966052) supplemented with 50 U/mL penicillin (Gibco, 15140122), 50 μg/mL streptomycin (Gibco, Carlsbad, CA, USA, 15140122), 2 mM GlutaMAX (Sigma, Burlington, MA, USA G8541), non-essential amino acid solution (Gibco, Carlsbad, CA, USA, 11140050), 0.1 mM 2-mercaptoethanol (Gibco, Carlsbad, CA, USA, 21985023), and 15% (v/v) fetal bovine serum (Sigma, Burlington, MA, USA F7524). Cells were maintained in a pluripotent state using 1000 U/mL ESGRO Leukaemia Inhibitory Factor (LIF) (Chemicon, Burlington, MA, USA ESG1106), and two inhibitors (2i): 1 μM PD0325901 (MEK inhibitor, Selleckchem, Houston, TX, USA, S1036) and 3 μM CHIR-99021 (GSK-3 inhibitor, Selleckchem, Houston, TX, USA, S1263). Cells were grown at 37 °C and 5% CO₂ in a humidified incubator and tested regularly for mycoplasma contamination. Medium was renewed every other day, and cells were sub-cultured using TrypLE Express (Gibco, Carlsbad, CA, USA, 12604013) when mESC clones reached an optimal size. CRISPR-inducible mESC lines to generate control and knockout Bhlhe40 (iBhlhe40^ΔSg4 R3.8^Cas9) or knockout Arnt (iArnt^ΔSg3 R3.8^Cas9) were generated by lentiviral transduction of R3.8^Cas9 cells with the corresponding single guide RNAs (sgRNAs) cloned in vector pKLV2-U6gRNA5(BbsI)-PGKpuro2ABFP-W (Addgene, Watertown, MA, USA, #67974) as previously described [26]. Four sgRNAs targeting Arnt and five sgRNAs targeting Bhlhe40 were designed with the Breaking Cas tool (https://bioinfogp.cnb.csic.es/tools/breakingcas/ accessed on 24 February 2024) [27]. Transduced pools of cells were sorted by Blue Fluorescent Protein (BFP) expression in BD FACSAria II SORP cell sorter (BD, Franklin Lakes, NJ, USA). Gene edition was induced in mESC lines by 5-day treatment with 2 μg/mL Doxycycline (Sigma, Burlington, MA, USA, D9891). Cell lines expressing the sgRNA that yielded the best percentage of editing were used for further experiments. The sequence of the sgRNAs of the pool cell lines generated is indicated in Supplementary Figure S1. Primers used to check the efficiency of editing by qRT-PCR are indicated in

Table 1. Primer sequences and short hairpin RNA sequences. Sequences of the primers for qRT-PCR analysis and of the shRNAs used for BHLHE40 silencing.

Name	Sequence 5′-3′
Primer
qPCR_Bhlhe40-sgRNA1-F	CTGGAGCACGGAGACCTG
qPCR_Bhlhe40-sgRNA1-R	CTCCGTTTTATTCCCCGCCT
qPCR_Bhlhe40-sgRNA2-F	GCTGGAGCACGGAGACCT
qPCR_Bhlhe40-sgRNA2-R	CTCCGTTTTATTCCCCGCCT
qPCR_Bhlhe40-sgRNA3-F	CCACATGTACCAAGTGTACAAG
qPCR_Bhlhe40-sgRNA3-R	GGCAGTTTGTAAGTTTCCTTGCT
qPCR_Bhlhe40-sgRNA4-F	TGCCCACATGTACCAAGTG
qPCR_Bhlhe40-sgRNA4-R	GGCAGTTTGTAAGTTTCCTTGCT
qPCR_Bhlhe40-sgRNA5-F	GAGCGAAGACAGCAAGGAAAC
qPCR_Bhlhe40-sgRNA5-R	TGGGCAATGCACTCGTTA
qPCR_Arnt-sgRNA1-F	GGGACGGTGCCATCTCGA
qPCR_Arnt-sgRNA1-R	TCAAAATCCAGCCCTGACCG
qPCR_Arnt-sgRNA2-F	TGCCATCTCGACCATGGC
qPCR_Arnt-sgRNA2-R	TCAAAATCCAGCCCTGACCG
qPCR_Arnt-sgRNA3-F	CAGAAATGACATCAGATGTACCAT
qPCR_Arnt-sgRNA3-R	TCAAAATCCAGCCCTGACCG
qPCR_Arnt-sgRNA4-F	GAGGGCTATTAAGCGACGG
qPCR_Arnt-sgRNA4-R	CTCATCATCCGACCTGGCAA
qPCR_mBhlhe40-F	GGAGACCTGTCAGGGATGGA
qPCR_mBhlhe40-R	CTCCGTTTTATTCCCCGCCT
qPCR_mArnt-F	TATTAAGCGACGGTCAGGGC
qPCR_mArnt-R	CTCATCATCCGACCTGGCAA
qPCR_hBhlhe40-F	GCCCTGCAGAGTGGTTTACA
qPCR_hBhlhe40-R	CGAAGACTTCAGGTCCCGAG
qPCR_mβ-Actin_F	CACACCCGCCACCAGTTCGC
qPCR_mβ-Actin_R	TTGCACATGCCGGAGCCGTT
qPCR_hβ-Actin_F	ATTGGCAATGAGCGGTTCC
qPCR_hβ-Actin_R	GGGAAATGAGGGCAGGACTTAG
qPCR_mEgln3_F	AGGCAATGGTGGCTTGCTAT
qPCR_mEgln3_R	GACCCCTCCGTGTAACTTGG
qPCR_mCcnd1-F	CTTCAAGTGCGTGCAGAAGG
qPCR_mCcnd1-R	TCTGCTCCTCACAGACCTCC
shARNs
pLV-EGFP-U6-sh-1	AGAAAGGATCGGCGCAATTAA
pLV-EGFP-U6-sh-5	CATGTGAAAGCACTAACAAAC
pLV-EGFP-U6-Scramble	CCTAAGGTTAAGTCGCCCTCG

.

Human umbilical vein endothelial cells (HUVEC) from pools of donors were purchased from Lonza (Lonza, Basel, Switzerland, C2519A), grown in Endothelial Medium Bullet Kit (EGM-2, Lonza, CC-3162), and used from passage 5–7 for experiments. To silence BHLHE40 expression in HUVEC, we used two short hairpin RNAs (shRNAs) targeting different regions (Table 1) by lentiviral transduction of the expression vectors pLV-EGFP-U6-sh-1 (Vector Builder, Neu-Isenburg, Germany, VB230523-1740jrm) and pLV-EGFP-U6-sh-5 (VectorBuilder, Neu-Isenburg, Germany, VB230523-1784wkp). A shRNA scramble (pLV-EGFP-U6-Scramble; VectorBuilder, Neu-Isenburg, Germany, VB010000-0001mty), with a sequence that does not interfere with any human gene, was used as a control. For lentiviral transduction, HUVEC were seeded at a density of 104 cells/cm² in 24-well multiwell plates (MW24) the day before infection, and lentiviruses were used at a multiplicity of infection (MOI) of 1–2 for 8 h. After 24 h, cells were cultured in normoxia (21% O₂) or hypoxia (1% O₂) for 48 h and then processed for microscopy analysis. Efficiency of lentiviral transduction was quantified by ImageJ software version 2.0.0-rc-69/1.52p (https://ccrod.cancer.gov/confluence/display/ROB2/Home accessed on 24 February 2024) (4 December 2018) counting of Green Fluorescent Protein (GFP) expression in 5 × 10³ cells per experimental condition.

4.2. Generation and Differentiation of Mouse Embryoid Bodies

EBs were generated by the hanging drop method and differentiated in two dimensions (2D) in tissue culture plates coated with 1% gelatin (Sigma, G1890) in Phosphate-buffered Saline (PBS). The mESCs were aggregated in 25 μL droplets containing 1.5 × 10³ cells/drop in mESCs medium in the absence of LIF and 2i. After differentiating for 4 days in hanging drops, EBs for microscopy analysis were seeded using a magnifying glass in 12 mm diameter coverslips in 24-well multiwell plates (1 EB/well) coated with 1% gelatin in PBS. For FACS analyses, EBs were seeded on p60 culture plates (30 EBs/plate) coated with 1% gelatin in PBS. The 2D differentiation of EBs continued until day 10 at 37 °C and 5% CO₂ in normoxia (21% O₂) in a cell incubator. Hypoxia treatment consisted of maintaining EBs at 37 °C in a gaseous mixture of 1% O₂, 5% CO_2, and 94% N₂ in a Whitley Hypoxystation H35 hypoxia chamber (Don Whitley Scientific, Victoria Works, Yorkshire, UK) for the last 48 h of differentiation.

4.3. Immunofluorescence Staining and Confocal Microscopy Analysis

EBs and HUVEC were fixed with 4% Paraformaldehyde (PFA) in PBS for 20 min at room temperature (RT). The remaining PFA was blocked with 0.1 M glycine in PBS for 20 min at RT, followed by washing with PBS. The fixed EBs and HUVEC were stored at −20 °C in 70% ethanol until processing. After washing with PBS for 10 min at RT, they were permeabilized with 0.5% Triton X-100 in PBS for 20 min at RT, followed by two washes with PBS and blocking with 3% Bovine serum albumin (BSA) in PBS for 20 min at RT. Antibodies for immunostaining were diluted in PBS with 3% BSA and 0.1% Tween-20 as follows: 1:800 rat anti-mouse CD31 (BD, Franklin Lakes, NJ, USA, 553370), 1:1000 rabbit anti-ERG (Abcam, Cambridge, UK, ab92513), 1:250 rabbit anti-GFP (Thermo Fisher Scientific, Waltham, MA, USA, A-6455) and 1:100 rabbit anti-mouse HOXD9 (Santa Cruz Biotechnology, Dallas, TX, USA, sc-8320); and 1:500 goat anti-rat AF564, 1:500 goat anti-rabbit AF488, and 1:300 goat anti-rabbit AF594. Primary antibodies were incubated overnight at 4 °C in a humified chamber and washed with 0.1% Tween-20 in PBS (5 washes of 20 min). Secondary antibodies were incubated at RT for 1.5 h in the humified chamber and washed with 0.1% Tween-20 in PBS (6 washes of 20 min). DNA was stained with 1.25 μg/mL DAPI (Molecular Probes, Eugene, OR, USA, D1306) in PBS for 10 min for nuclei visualization. Coverslips were mounted on slides with Prolong (Molecular Probes, Eugene, OR, USA, P36970) and allowed to dry overnight at RT, protected from light. The slides were then stored at 4 °C until visualization.

4.4. Quantification of EdU Incorporation by Flow Cytometry and Microscopy

4.4.1. EdU Quantification by Flow Cytometry in Mouse Embryoid Bodies

We added 10 μM EdU during the last hour of EB differentiation. 30–60 EBs per experimental condition were used. EBs were dissociated into single cells by treatment with 1 mg/mL Dispase II (Roche, Basel, Switzerland, 04942 078001) and 2.5 mg/mL collagenase I (Gibco, 17018029) in DPBS (Cytiva, Marlborough, USA, SH30264.01) for 20 min at 37 °C, followed by 2 min incubation with 0.5 mL non-enzymatic cell dissociation solution (Sigma, C5914) and mechanical dissociation by pipetting. Cells were centrifuged for 5 min at 400× g. For EdU detection, the pellet was resuspended in 50 μL of PBS, fixed with 1 mL of 70% ethanol, and incubated for 20 min at 4 °C before storing at −20 °C. For EdU detection we used a modified Click-iT reaction with biotin amplification as follows: cells were centrifuged at 500× g 7 min, washed with 1 mL 0.1% Tween-20 (Sigma, P9416) in PBS, resuspended in 100 μL of freshly prepared EdU detection reaction buffer (PBS, 2 mM CuSO4 (Sigma, C7631), 0.05 mM Biotin-Azide (Molecular Probes, B10184), and 5 mM ascorbic acid (Sigma, A-4544)), and incubated 30 min at RT protected from light. After washing with 0.1% Tween-20 in PBS, cells were resuspended in 100 μL Streptavidin-AF647 (Molecular Probes, B10184) diluted 1: 200 in PBS containing 0.1% Tween-20 and 3% BSA (NZYTech, Lisbon, Portugal, MB04602) and incubated 30 min at RT protected from light. After the Click-iT reaction, the cells were washed with 1 mL PBS and centrifuged at 500× g for 5 min. FACS was performed on a three-laser (405 nm, 488 nm, 633 nm) flow cytometer (FACSCantoII with BDFACSDIVA software 6.2, BD Biosciences). A minimum of 10⁴ events per experimental condition were acquired in slow rate mode to avoid doublets. Data analysis was performed with FlowJo 9/10 software (Ashland, Wilmington, Delaware) as previously described [9].

4.4.2. EdU Quantification by Flow Cytometry in Mouse Embryonic Stem Cells

We grew 2 × 10⁵ mESCs on 0.1% gelatin-coated p35 plates and cultured in normoxia or hypoxia for 48h, and 10 μM EdU was added during the last 30 min. Cells were centrifuged at 500× g for 7 min; the pellet was resuspended by vortexing, and cells were fixed in 0.5 mL of 70% ethanol and incubated for 20 min at 4 °C before storing at −20 °C until processing for EdU detection as indicated above.

4.4.3. EdU Quantification by Microscopy in Vascular Structures in Mouse Embryoid Bodies

EdU was added during the last 3 h of EB differentiation. For microscopy analysis of EdU incorporation in EBs, we used Click-iT, following the manufacturer’s instructions (Invitrogen, Waltham, MA, USA, C10340). Images of 4–10 fields of vascular structures per EB, out of a total of 6 EBs, were acquired with a 20× objective (Plan-APOCHROMAT 20×/0.75 multi-immersion DIC) using a Stellaris 8 Tau STED spectral confocal microscope (Leica Microsystems, Wetzlar, Germany) with laser lines 405, 499, 557, and 653 for each experimental condition. To quantify the percentage of S-phase cells in vascular structures, we used a semi-automatic macro that allowed the automatic counting of ERG⁺ nuclei in CD31⁺ endothelial cells (CD31⁺ ERG⁺ mask), followed by manual counting of EdU⁺ EC nuclei in the maximum intensity Z-projection, as described in Acosta-Iborra et al. [9]. We quantified 1.5 × 10³–3 × 10³ ECs per experimental condition.

4.4.4. EdU Quantification by Microscopy in HOXD9⁺ Progenitor Cells and Total Cells in the Core Regions of Mouse Embryoid Bodies

EdU was incorporated during the last 3 h of EB differentiation. For microscopy analysis of EdU incorporation in EBs, we used Click-iT, as indicated above. EdU was quantified in images of microscopy fields containing vascular structures and progenitors at approximately 800 mm from the outer edge of the EB, see Supplementary Figure S3. Images were acquired with a 20× objective (Plan-APOCHROMAT 20×/0.75 multi-immersion DIC) using a Stellaris 8 Tau STED spectral confocal microscope with laser lines 405, 491, 557, 590, and 653 for each experimental condition and analyzed using Imaris 10.0 software (Oxford Instruments, Abingdon, Oxfordshire, UK). We quantified images of 5–10 fields per EB out of a total of 6 EBs. To quantify the percentage of EdU incorporation in HOXD9⁺ progenitor cells and in the total cells, we used the Imaris 10.0 3D Surfaces classifier tool. Before quantification, apoptotic bodies were removed using size filters and intensity filters. Nuclei were visualized using DAPI to quantify the total cell number. We quantified 2 × 10⁴–6 × 10⁴ cells per experimental condition. The images shown in Figure 5 are sections of images taken with a 40× objective (Plan-APOCHROMAT 40×/1.3 Oil DIC) using the same confocal microscope.

4.4.5. EdU Quantification by Microscopy in HUVEC

HUVEC were seeded at a density of 10⁴ cells/cm² in 12 mm diameter coverslips in MW24, and 10 μM EdU was added in the last 3 h. For EdU detection, we used Click-iT. Images were acquired using the mosaic tool of a Zeiss Axio Observer Z1 (Zeiss, Oberkochen, Germany) with a 10× objective (Plan-APOCHROMAT 10×/0.45) and analyzed with ImageJ software version 2.0.0-rc-69/1.52p (4 December 2018). Quantification of the percentage of cells in S-phase in HUVEC was performed on binary images by automatic counting using Fiji’s ‘analyze particles’ command, and the total number of cells quantified was 1.2 × 10⁴–2.0 × 10⁴ cells per experimental condition.

4.5. Quantification of Angiogenesis by FACS and Microscopy

Angiogenesis was quantified by FACS using anti-CD-144 antibodies to VE-cadherin. 30–60 EBs per experimental condition were used. Cells were resuspended in 100 μL of 3% BSA in PBS, and 1 μg of anti-VE-Cadherin-PE antibody (BD, 560410) or isotype control antibody was added and incubated 40 min at 4 °C protected from light. After washing with 3 mL of PBS, cells were centrifuged at 500× g for 5 min and resuspended in 200 μL of PBS containing 1 μg/mL of DAPI. FACS was performed in FACSCantoII BD Biosciences. For microscopy analysis, images of 11–18 EBs per experimental condition were acquired using the mosaic tool of a Zeiss Axio Observer Z1 microscope (Zeiss) with a 10× objective (Plan-APOCHROMAT 10×/0.45). Angiogenesis was quantified using AngioTool software version 0.6a (https://ccrod.cancer.gov/confluence/display/ROB2/Home accessed on 24 February 2024) (October 2014) [29]. Before AngioTool analysis, images were processed with Fiji software version 2.0.0-rc-69/1.52p (https://imagej.net/software/fiji/ accessed on 24 February 2024) (4 December 2018) to optimize the analysis by applying the Gaussian Blur filter (Sigma 2.00).

4.6. RNA Extraction and qRT-PCR

30–60 EBs per experimental condition were used. RNA extraction and purification were performed using NucleoSpin RNA Plus Kit (Marcherey Nagel, Düren, Nordrhein-Westfalen, Germany, 740955) following the manufacturer’s instructions. For RNA quantification, RNA from each sample was reversed-transcribed into complementary DNA (cDNA) using the NZY First Strand cDNA Synthesis Kit (NZYTech, MB12501). cDNA was diluted to 5 ng/μL and used as a template for amplification reactions, carried out with the Power SYBR Green PCR Master Mix (Applied Biosystems, Waltham, MA, USA, A25741). PCR amplifications were performed on the StepOne Realtime PCR System (Applied Biosystems, 4376357). Data were analyzed with StepOne software version v2.1, and expression levels were calculated using ΔΔCT and β-actin as reference. The mean expression value across all samples was used as the sample reference for the graphs. The sequences of the primers are included in Table 1.

4.7. Western Blot

HUVEC cultured in MW6 at a density of 10⁴ cells/cm² and mESCs cells cultured in 0.1% gelatin-coated p35 plates were lysed in RIPA buffer (50 mM Tris-HCl pH 8, 150 mM NaCl, 0.02% NaN3, 0.1% SDS, 1% NP-40) containing protease inhibitors (Complete ULTRA table, Roche, 06538304001). Samples were sonicated for 5 min and centrifuged for 10 min at 10,000× g at 4 °C. Protein concentration was quantified using Bio-Rad DC protein assay (Bio-Rad, Hercules, CA, USA, 5000112), and samples of 20 μg were resolved by SDS-polyacrylamide electrophoresis. Proteins were transferred to polyvinylidene difluoride (PVDF) membranes (Immobilon-P, Millipore, Burlington, MA, USA, IPVH100010). Membranes were blocked for at least 1 h with 5% non-fat dry milk and 3% BSA in TBS-T (50 mM Tris-HCl pH 7.6, 150 mM NaCl, 0.1% Tween-20) and incubated with the corresponding antibodies: 1:1000 rat anti-HIF1α (provided by CNIO Monoclonal Antibodies Unit), 1:1000 rabbit anti-BHLHE40 (Novus, St. Louis, MO, USA, NB100-1800), 1:1000 mouse anti-CAS9 (Santa Cruz Biotechnology, SC517386), 1:5000 mouse anti-Tubulin (Sigma, T6199), 1:5000 Goat Anti-Mouse-HPR (Santa Cruz Biotechnology, SC516102), 1:5000 Goat Anti-rat-HPR (Santa Cruz Biotechnology, SC2065), and 1:10,000 goat anti-rabbit-HPR (Jackson ImmunoResearch, West Grove, Pennsylvania, USA, 11-035-045). Primary antibodies were incubated overnight at 4 °C, while secondary antibodies were incubated for 1 h at room temperature. Washings after antibody incubations were performed in TBS-T. Western blots were developed by enhanced chemiluminescence (Clarity western ECL substrate Bio-Rad, 170-5060) and visualized with a digital luminescent image analyzer Fusion SOLO 7S Edge (Vilber, Marne-la-Vallée, France). Densitometry analysis was carried out using the Fiji software version 2.0.0-rc-69/1.52p (https://imagej.net/software/fiji/ accessed on 24 February 2024) (4 December 2018).

4.8. Proliferation Curves and Cell Viability

5 × 10³ mESCs cells were plated on 0.1% gelatin-coated in 6-well multiwell plates and cultivated in normoxia (21% oxygen) or hypoxia (1% oxygen) for the indicated time. mESCs cells were dissociated into single cells by treatment with TripLE Express, centrifuged at 500× g 7 min, resuspended in 170 μL of PBS containing 1 μg/mL DAPI (Molecular Probes, D1306), and counted by FACS using Perfect Counts microspheres (Cytognos, Santa Marta de Tormes, Salamanca, Spain, CYT-PCM-50) (30 μL per experimental condition). Sample measurements were performed with BD FACSDIVATM Software (Version 6.2, BD Biosciences). Data acquisition and analysis were performed following Perfect Counts microsphere instructions. Doubling time was calculated based on the coefficients obtained from the linear regression analysis using the following equation:

$${log}_2{N}_t={log}_2{N}_0+\frac{1}{T}t$$

t (time hours), N_t (cell number at time t), N₀ (initial cell number) and T (doubling time).

4.9. Gene Set Enrichment Analysis (GSEA)

We compared a proliferative gene signature from [24] against a gene list ranked according to the estimated effect of hypoxia on their expression, extracted from a meta-analysis of 43 individual RNA-seq studies across 34 distinct cell types [6]. This ranked list encompasses all genes detected across the cell lines in the 43 studies, totaling 22,182 genes. The analysis was conducted using GSEA v4.3.2 for Linux, utilizing the GSEA-Preranked tool with default parameters.

4.10. Statistical Analysis

Figure 1A,C was generated using R statistical software (version 4.3.3, 29 February 2024, “Angel Food Cake”) and the packages “pheatmap” (version 1.0.12) and “ggplot2” (version 3.4.2).

GraphPad Prism version 10.0.0 for Windows, 11 July 2023, GraphPad Software, Boston, MA, USA, www.graphpad.com, was used for statistical analysis. Normality and homogeneity of variance were determined for each dataset using Shapiro–Wilk’s and Brown–Forsythe tests, respectively. The statistical significance of differences in group means was determined using a parametric test (ANOVA and Student’s t-test) when normality and homogeneity were met; otherwise, non-parametric analyses (Mann-Whitney and Friedman tests) were performed.