The endothelium is essential in maintaining vascular homeostasis (Cines et al., 1998; Vanhoutte et al., 2017). While the endothelium represents a key target of vascular estrogen action (Bolego et al., 2006a; Bolego et al., 2006b; Pinna et al., 2008; Arnal et al., 2010), endothelial biology is also governed by sex-specific factors (Mudrovcic et al., 2017; Stanhewicz et al., 2018). Moreover, sex hormones exert direct effects and interact with the effects of sex chromosomes during lifetime (Arnold et al., 2017; Ventura-Clapier et al., 2017), suggesting that sex hormones in the local microenvironment as well as intrinsic differences at the cellular level account for gender differences in endothelial function. Thus, understanding the pattern of sex chromosome as distinguished from sex hormone-driven phenotypic changes in the endothelium would be instrumental for the identification of clinically relevant preventive and/or therapeutic strategies.

Due to easy access and high purity, human umbilical vein endothelial cells (HUVECs) are a valuable model for the study of physiopathological processes. Because HUVECs have the fetus’ sex (Annibalini et al., 2014), sex-specific (XX and XY) chromosome-based differences may be explored. For instance, XX and XY cells have different susceptibility to cell death and survival (Malorni et al., 2007). Cultured HUVECs display sexual dimorphism in, e.g., the nitric oxide pathway (Addis et al., 2014; Cattaneo et al., 2017) as well as in the transcriptional profile and intracellular proteome under basal and stress conditions (Lorenz et al., 2015; Lorenz et al., 2019; Witt et al., 2019). Although an association between sex-specific gene variants and coronary endothelial dysfunction has been reported in humans (Yoshino et al., 2016), the relative contribution of sex and estrogenic agents to new vessel formation in vitro under physiological conditions remains to be fully characterized.

Endogenous signaling molecules including estrogen can stimulate the repair of damaged blood vessels as well as the formation of new blood vessels (Arnal et al., 2010). This process, known as angiogenesis, is important in the female reproductive tract physiology as well as in pathophysiological settings such as cardio/cerebrovascular disease, diabetes, endometriosis and cancer, supporting a role for sex hormones in the context of neovascularization process (Losordo and Isner, 2001; Fadini et al., 2008). Estrogen action in the endothelium is mediated by its interaction with estrogen receptor (ER)α and ERβ, which operate as hormone-regulated transcription factors. Membrane ERα and the recently described G protein-coupled estrogen receptor (GPER) also mediate rapid, non-genomic effects (Prossnitz and Barton, 2014). Notably, the ER expression pattern in ECs differs according to the location and vessel type (Franconi et al., 2017; Huxley et al., 2018). Estrogenic agents have been reported to enhance new vessel formation by interacting with ERα (Arnal et al., 2010; Sanchez et al., 2011). We recently showed that estrogen treatment enhances angiogenesis in HUVECs by interacting with GPER resulting in the rapid downstream activation of 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 3 (PFKFB3), a bifunctional protein linking glucose metabolism to endothelial cell proliferation (Trenti et al., 2017a). However, in the above study endothelial cells were not stratified by donor's sex, nor were sex-specific responses to estrogenic agents assessed.

The major steps of the angiogenetic process including EC migration, i.e. chemotaxis, and the ability to form tube-like structures when seeded onto extracellular matrix components involve the modification and spatial organization of actin cytoskeleton (Madri and Pratt, 1986). Estrogen treatment in HUVECs induces tyrosine phosphorylation and activation of focal adhesion kinase (FAK), a signaling protein located at sites of integrin clustering that controls cell migration (Sanchez et al., 2011). It has been reported that PFK15, a small molecular PFKFB3 inhibitor, causes a decrease in phosphorylated FAK in gastric cancer cells, thereby limiting cell invasiveness in vitro (Zhu et al., 2016). We found that digitoxin at clinically relevant concentrations elicits a potent anti-angiogenic effect in vitro and in vivo by inhibiting FAK activation (Trenti et al., 2017b). However, whether FAK activation and the responses to pharmacological agents affecting angiogenesis are sexually dimorphic is largely unknown.

Translation of research findings into clinical recommendations requires the identification of possible targets and/or mechanisms accounting for, e.g., gender-specific response to pharmacological treatment. Therefore, the aim of this study was to further explore the relative functional role of sex and estrogenic agents in the healthy endothelium using in vitro approaches that mimic the main steps of the angiogenic process. Additional specific aims were to unravel potential sex differences in key proteins involved in the migration and tubularization of HUVECs, and to test sex-specific responses to pharmacological treatment in vitro.

Gender differences in the cardiovascular system have been largely ascribed to beneficial effects of estrogen on endothelial function. This view has been challenged not only by the mixed outcome of hormone intervention trials but also by emerging evidence of inherent sex differences in endothelial biology (Mudrovcic et al., 2017; Stanhewicz et al., 2018). We herein provide evidence that sex (ECs being XX or XY) and estrogenic agents affect HUVEC phenotypes by complementary modulation of key functional proteins. Notably, consistent with higher pro-angiogenic response in fHUVECs, we herein report sex-specific activation patterns of FAK, a signaling protein involved in cell migration and angiogenesis, as well as of PFKFB3, a master regulator of endothelial glucose metabolism in health and disease.

HUVECs, an established model to investigate endothelial pathophysiology, are suitable to determine the relative contribution of sex at the transcriptional and functional level because the umbilical cord vein is a foetal annex and maintains the foetus sex (Annibalini et al., 2014). In addition, sex differences in freshly isolated HUVECs are independent of maternal estrogen exposure (Hickey et al., 2014). Having assessed unequivocally the donor's sex at the genomic level, we used a combination of MTT assay and Trypan blue staining to measure viable cell numbers and cell proliferation rates, respectively. As a result, fHUVECs were more viable at different cell densities and serum concentrations and had a higher proliferation rate than mHUVECs. One study previously reported higher proliferation rate in fHUVECs than in mHUVECs but no difference in viability (Addis et al., 2014), whereas Cattaneo et al. (2017) reported no sex difference in the proliferation rate. These conflicting findings are likely due to discrepancies in experimental design and procedures compared with ours. In addition, those studies did not compare responses to estrogenic agents in mHUVECs vs fHUVECs.

Endothelial cell growth is known to be stimulated by estrogen treatment in vitro (Morales et al., 1995; Kim-Schulze et al., 1996; Geraldes et al., 2002; Oviedo et al., 2007), but sex differences in these responses have not been explored. Functional expression of ERα and ERβ in the endothelium is widely documented (Kim-Schulze et al., 1996; Rubanyi et al., 2002; Addis et al., 2014; Annibalini et al., 2014), with variable patterns according to vessel type (Simard et al., 2011) and donor's sex (Addis et al., 2014; Annibalini et al., 2014). In contrast to previous reports of nuclear GPER staining (Chakrabarti and Davidge, 2012), we herein first provide evidence for plasma membrane localization of GPER in mHUVECs and fHUVECs, with higher expression levels in the latter. Therefore, we tested the hypothesis that a GPER agonist and in parallel the nonselective ER agonist E2 would enhance HUVEC viability in fHUVECs more than in mHUVECs. Yet, treatment with increasing concentrations of E2 and G1 failed to enhance HUVEC viability, regardless of donors' sex. Previous work showed that the GPER agonist G1 triggers the growth rate of bovine vascular ECs (Xu et al., 2017), while others reported attenuated proliferation of the mouse brain microvascular EC line bEnd.3 (Holm et al., 2011) in response to G1, suggesting that EC origin affects endothelial function study outcomes (Huxley et al., 2018). Of note, FBS in our study did not undergo charcoal treatment as in other studies showing estrogen-induced EC proliferation (Morales et al., 1995; Kim-Schulze et al., 1996; Oviedo et al., 2007) because this procedure turns out to be associated with batch-to-batch variation and to remove nutrients essential for cell growth (Cao et al., 2009), thereby impairing HUVEC function including, e.g., sprouting ability (Vanetti et al., 2016). Hence, it is conceivable that increased proliferation rates detected in experiments using charcoal-treated serum (e.g. Morales et al., 1995) resulted at least in part from an anti-apoptotic effect of estrogen (Spyridopoulos et al., 1997; Su et al., 2017). Alternatively, serum supplementation in the culture medium may have provided enough estrogen level to yield a steady growth rate. It would be worthwhile testing the gender-specific effects of other steroid sex hormones on HUVEC growth and proliferation in future studies.

Among other direct effects on resident vascular cells, estrogen fosters endothelial healing also through angiogenesis. This process occurs physiologically in the female reproductive system (Reynolds et al., 1992) and favors neovascularization processes in the context of, e.g., tissue ischemia. Estrogen has been known to promote angiogenesis mainly by interacting with ERα (Losordo and Isner, 2001; Arnal et al., 2010; Sanchez et al., 2011). More recently, we reported that the angiogenic response to estradiol occurs, at least in part, through interaction with GPER, resulting in up-regulation of the glycolytic activator PFKFB3 (Trenti et al., 2017a). However, the relative role of sex and estrogen in angiogenesis has not been completely defined.

Using in vitro experimental approaches that mimic the major steps of the angiogenic process we found that, along with increased growth rate, fHUVECs displayed higher migratory properties than mHUVECs in response to 15% and 5% FBS as chemoattractant using the Boyden assay, consistent with previous findings generated using different time points, assays or chemotactic agents (Addis et al., 2014; Cattaneo et al., 2017). However, we here show for the first time that capillary-like tube formation was enhanced in fHUVECs with respect to mHUVECs under physiological conditions. Thus, while comparable observations were made previously in HUVECs exposed to either hyperoxia (Zhang and Lingappan, 2017) or serum starvation (Lorenz et al., 2015), sexual dimorphism in the formation of capillary-like structures occurs and may be relevant under conditions such as development (neovasculogenesis) or physical exercise. Overall, these data suggest a sex-specific chromosome-based control of ECs under basal conditions and in response to environmental cues triggering proliferation, migration and tubulogenesis that may be independent from the contribution of female hormones, at least in the present in vitro setting. It is conceivable that sex-specific chromosomes and hormones cooperate at different levels in modulating the multiple steps of the process. Sexual dimorphism also appears to regulate angiogenesis in in vivo models. For instance, gender does not affect angiogenesis and tissue reperfusion in a rabbit ischemia model, but estrogen deficiency negatively regulates it (Kyriakides et al., 2003). In a rat model of ischemia/reperfusion injury, the number of chimeric vessels and the total capillary density were higher with transplanted female than male endothelial progenitor cells (Fadini et al., 2008). Female gender or postmenopausal status is a negative predictor of collateral vessel formation in patients with severe coronary artery stenosis (Yetkin et al., 2015). Sex differences have also been recently demonstrated in adipose tissue angiogenesis, with important implications for the development of obesity-related cardiometabolic disease: an increased angiogenic response to high-fat diet in female mice may explain their increased resistance to diet-induced obesity and preserved metabolic homeostasis compared to males (Rudnicki et al., 2018). Overall, evidence of genomic contribution to sex differences in sexually immature animals is lacking. Thus, understanding the pathophysiological relevance of sexually dimorphic angiogenic responses and the interplay with female hormones in different settings deserves additional studies.

To further investigate whether angiogenesis-related sexually dimorphic traits occur through sex hormone-dependent or independent pathways, we focused on the regulation of key signaling molecules involved in the angiogenic process. Estrogenic agents upregulate the functional activity of PFKFB3 (Trenti et al., 2017a), a master regulator of both endothelial cell glucose metabolism and angiogenesis (Schoors et al., 2014; Xu et al., 2014; Peng et al., 2018). This may be in line with established gender differences in glucose homeostasis (Mauvais-Jarvis, 2018). Of note, PFKFB3 is a target of estrogen action in breast cancer cells (Imbert-Fernandez et al., 2014), which have high glycolytic rates similar to endothelial cells. In line with the knowledge that sex differences are mainly due to differential expression of genes present in both sexes (Gershoni and Pietrokovski, 2017), we evaluated PFKFB3 expression in HUVECs from male and female newborns and its role in FBS-induced migration. While PFKFB3 inhibition impaired migration in both mHUVEC and fHUVEC, no sex differences were observed in PFKFB3 baseline levels. In contrast, E2 treatment increased PFKFB3 with higher potency in fHUVECs than in mHUVECs (Figure 6B). Whether gender differences exist in the glycolytic metabolism of endothelial cells remains to be determined.

The functional link between FAK-related pathways and endothelial cell processes relevant to angiogenesis such as motility and proliferation is well established (Sulzmaier et al., 2014). Previous findings from our group showed a central role for FAK phosphorylation at Y576/577 (Trenti et al., 2017b; Trenti et al., 2018) in HUVEC migration and angiogenic potential. Notably, pharmacological inhibition of PFKFB3 reduced FAK Y576/577 phosphorylation and total FAK levels (Supplementary Figure S3), and impaired HUVEC migration, suggesting that the two enzymes are functionally linked and the functional activation of FAK is fostered by glycolysis. The precise mechanism underlying the above link remains to be determined. We herein report for the first time higher levels of FAK phosphorylation at Y576/577 in fHUVEC with respect to mHUVECs. This effect was unrelated to total FAK levels, which were even higher in mHUVECs, and was likely specific for FAK phosphorylation on Y576/577 by Src, which is required for maximal FAK activation (Zhao and Guan, 2011). While the mutual phosphorylation of FAK and Src family tyrosine kinases is required for cell migration (Hanks and Polte, 1997), we did not observe gender-related differences in Src phosphorylation. In addition, while the PI3K/Akt pathway is involved in FBS-induced migration, no gender differences were found in Akt and p-Akt levels (Figure 9B,C). Overall, gender differences in HUVEC migration are mainly related to differences in FAK activation, and such gender differences are abolished by LY294002 (PI3K inhibitor) and 3PO (PFKFB3 inhibitor). The scope for full mechanistic evaluation of these findings, however, goes beyond this current study. Whether other proteins that influence FAK activity and cytoskeleton remodeling, such as the Rho family of GTPases, are differentially expressed in fHUVEC vs. mHUVECs remains to be determined. While total FAK was more abundant and phospho-FAK less abundant in mHUVEC vs. fHUVEC, both FAK and phospho-FAK are more expressed in male vs. female vascular smooth muscle cells (Straface et al., 2009), in line with a key role of sex chromosomes in cytoskeleton rearrangement. In addition, consistent with the formation of multiprotein complexes of membrane ERs with Src and FAK, E2 treatment has been reported to induce FAK autophosphorylation at Y397 in HUVECs (Sanchez et al., 2011), although the donors’ sex was not indicated in that study. By contrast, E2 treatment did not change FAK levels in cancer stem cells (Zamani et al., 2020). This implies that, when looking at mechanisms and target proteins involved in functional responses, not only cell origin but also sexual dimorphism needs to be considered.

Identification of sexual dimorphism may be relevant for the outcome of pharmacological treatments targeting endothelial function. For instance, therapeutic agents targeting angiogenic factors such as bevacizumab, a monoclonal antibody that blocks VEGF, display sex differences in outcomes including longer overall survival in female with respect to male patients or sex- and dose-dependent progression-free survival (Brahmer et al., 2011; Reck et al., 2009). The reasons for these differential effects are unclear. Starting from our recent findings that the cardiac glycoside digitoxin at therapeutic concentrations impairs HUVEC migration, capillary tube-like formation and in vivo angiogenesis by inhibiting growth factor-induced FAK Y576/577 phosphorylation (Trenti et al., 2017b), we assessed whether such pharmacological responses were affected by HUVEC donors’ sex. In a proof-of-concept experiment, we showed that digitoxin treatment inhibited HUVEC migration in a sex-oriented fashion, with the gender gap being abolished at concentrations within the therapeutic range (Figure 10). The observation that digitoxin treatment does not prevents FAK autophosphorylation at Y397 (Trenti et al., 2017b) rules out overall blockade of cell migration in a FAK-dependent manner. Thus, the above-mentioned sex-specific FAK activation pattern not only contributes to the sexually dimorphic HUVEC angiogenic potential but may also affect the response to agents showing antiangiogenic potential such as digitoxin. These findings highlight the need for sex-oriented pharmacological targeting of endothelial function. Furthermore, since FAK promotes tumor progression through effects on multiple cells of the tumor microenvironment (Sulzmaier et al., 2014; Trenti et al., 2018), interfering with the FAK signaling pathway may be a promising strategy to target invasive cancer types that are prevalent in women (Sood et al., 2004; Lark et al., 2005).

This study has some limitations. First, experiments were performed solely in vitro. In addition, due to the heterogeneous pattern of phenotype and ER expression according to vessel type and location, results may not be generalizable to other endothelial cell models. Because of previously published findings from our group and others, responses to estrogenic agents were tested for selected study endpoints only. Despite these limitations, the present findings indicate that the hormonal microenvironment (i.e. estrogen) and donor's sex modulate HUVEC functional phenotypes and signaling molecules involved in angiogenesis by complementary mechanisms that warrant further investigation, and support mechanism-based approaches to gender-specific pharmacological targeting of endothelial function.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.

The studies involving human participants were reviewed and approved by Ethics Committee, Padua University Hospital. The patients/participants provided their written informed consent to participate in this study.

All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by AT, CaB, CBa, CS, LT, and CBo. The first draft of the manuscript was written by CBo and AC, and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

This study was funded by University of Padova PRID Grant 2018 (PI Chiara Bolego). Annalisa Trenti has been recipient of an award from the Centro Studi Nazionale su Salute e Medicina di Genere (Padova, Italy).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.