- Top of page
- MATERIALS AND METHODS
- LITERATURE CITED
The association of microglia with brain vasculature during development and the reduced brain vascular complexity in microglia-deficient mice suggest the role of microglia in cerebrovascular angiogenesis. However, the underlying molecular mechanism remains unclear. Here, using an in vitro angiogenesis model, we found the culture supernatant of BV2 microglial cells significantly enhanced capillary-like tube formation and migration of brain microvascular endothelial cells (BMECs). The expression of angiogenic factors, ephrin-A3 and ephrin-A4, were specifically upregulated in BMECs exposed to BV2-derived culture supernatant. Knockdown of ephrin-A3 and ephrin-A4 in BMECs by siRNA significantly attenuated the enhanced angiogenesis and migration of BMECs induced by BV2 supernatant. Our further results indicated that the ability of BV2 supernatant to promote endothelial angiogenesis was caused by the soluble tumor necrosis factor α (TNF-α) released from BV2 microglial cells. Moreover, the upregulations of ephrin-A3 and ephrin-A4 in BMECs in response to BV2 supernatant were effectively abolished by neutralization antibody against TNF-α and TNF receptor 1, respectively. The present study provides evidence that microglia upregulates endothelial ephrin-A3 and ephrin-A4 to facilitate in vitro angiogenesis of brain endothelial cells, which is mediated by microglia-released TNF-α. Anat Rec, 297:1908–1918, 2014. © 2014 Wiley Periodicals, Inc.
Angiogenesis is the process that generating new blood vessels from pre-existing vascular networks by capillary sprouting (Weis and Cheresh, 2011; Marcelo et al., 2013). Angiogenesis is not only critical for embryogenesis and adult ovulation under physiological conditions, but also essential in many pathological processes such as wound healing, diabetic retinopathy, and tumorigenesis (Weis and Cheresh, 2011; Marcelo et al., 2013). Evidences suggest that angiogenesis is a multistep process involving endothelial cell migration and formation of lumen-containing tubular structures, regulated by the interactions between endothelial cells and perivascular supporting cells (Carmeliet and Jain, 2011; Weis and Cheresh, 2011).
In central nervous system (CNS), active angiogenesis has been demonstrated during brain development (Lee et al., 2009). Increased angiogenesis of brain microvessels was also associated with a variety of brain diseases such as infection, stroke, and neoplasia. At the microenvironment where angiogenesis occurs, there are multiple cell types including endothelial cells, pericytes, neurons, and glial cells, which as a whole are now recognized as a multi-cellular complex called neurovascular units (NVU) (Zlokovic, 2011; Drewes, 2012). The coordination of NVU implies the contribution of other cell types in NVU to the angiogenic process in brain. It has been shown that pericytes, which are closely associated with brain microvascular endothelial cells, play important roles in endothelial cell stimulation and guidance, as well as in endothelial maturation and stabilization (Ribatti et al., 2011; Sa-Pereira et al., 2012). Also, astrocytes are involved in the angiogenesis in developing retina and ischemia-induced retinal neovascularization (Scott et al., 2010; Weidemann et al., 2010; Hirota et al., 2011; Stenzel et al., 2011).
Microglia are bone marrow-derived macrophages in CNS and its role in pathological immune responses are clearly documented, whereas their crucial roles under physiological conditions were revealed until recently (Pont-Lezica et al., 2011). It was identified that microglia invade mouse brain early (E10.5) during embryonic development (Kierdorf et al., 2013). In brain parenchyma, the microglia were closely associated with blood microvessels (Grossmann et al., 2002; Arnold and Betsholtz, 2013). Moreover, in mice lacking microglia, Fantin et al. found the blood vessel intersections were reduced in the hindbrain (Fantin et al., 2010). These findings indicated the role of microglia in brain vascular angiogenesis; however, the involved molecular mechanism remains elusive.
Over the past decades, considerable progress has been obtained in elucidating the molecular mechanisms of vascular angiogenesis. VEGF family and angiopoietin/Tie2 family in angiogenesis have been studied extensively (Carmeliet and Jain, 2011). Recently, the importance of ephrin/Eph family in angiogenesis has been revealed in the literature. Ephrin/Eph family is the largest known subfamily of receptor tyrosine kinases, containing at least 8 ephrin ligands and 14 Eph receptors, both are anchored to the plasma membrane (Surawska et al., 2004). On the basis of sequence similarity and binding affinities, both ephrin ligands and Eph receptors are divided into two subclasses, A and B. The ephrin-A ligands are membrane-bound through glycosylphosphatidyl linkage, whereas ephrin-B ligands are anchored to the membrane via transmembrane domain. The stimulatory effect of ephrin-A ligands in vascular angiogenesis has been reported. Ephrin-A1 can promote in vitro vascular assembly of lung microvascular endothelial cells (Brantley-Sieders et al., 2004). Tube formation of RF/6A retina endothelial cells is enhanced by ephrin-A4 (Du et al., 2012).
In this study, we found that two angiogenic factors, ephrin-A3 and ephrin-A4, were simultaneously upregulated in brain endothelial cells treated with culture supernatant from microglial cells, which is required for the microglia-induced in vitro angiogenesis in brain endothelial cells.
- Top of page
- MATERIALS AND METHODS
- LITERATURE CITED
The important role of microglia in brain angiogenesis was recently identified by Fantin et al. utilizing microglia-deficient mice (Fantin et al., 2010). Nevertheless, in their efforts to explore the involved mechanism, the obtained evidence only argued against the involvement of VEGF in microglia-mediated vessel fusion during developmental brain angiogenesis, thus the mechanism underlying microglia-induced brain vascular angiogenesis remained to be determined. In this study, using an in vitro angiogenesis model, we found the culture supernatant derived from microglial cells promoted in vitro migration and angiogenesis of brain endothelial cells. Our results demonstrated that ephrin-A3 and ephrin-A4, but not VEGF, in brain endothelial cells were specifically upregulated by microglia, which is necessary for the microglia-induced angiogenesis.
We found that the microglia-induced brain endothelial angiogenesis was effectively prevented by knockdown of endogenous ephrin-A3 and ephrin-A4, respectively. Interestingly, there was no additive effect after simultaneous knockdown of ephrin-A3 and ephrin-A4 in brain endothelial cells (Fig. 3B–E), indicating that ephrin-A3 and ephrin-A4 act in the same pathway. It was thought that ephrin-A3 and ephrin-A4 ligands could bind with the same receptor, EphA2, respectively (Surawska et al., 2004). The role of EphA2 receptor in promoting endothelial cell migration and angiogenesis reported in our lab (Zhou et al., 2011) and in other research group (Brantley-Sieders et al., 2004; Sainz-Jaspeado et al., 2013), allowed us to infer that microglia could upregulate endothelial ephrin-A3 and ephrin-A4, which coordinately interact with EphA2 receptor on brain endothelial cells, resulting in augmented endothelial angiogenesis. It should be mentioned that ephrin-A1 in brain endothelial cells was decreased in response to microglial supernatant treatment (Fig. 2B). Given that the precise effect of ephrin-A1 on angiogenesis remained controversial (Surawska et al., 2004; Ojima et al., 2006), it will be interesting to clarify the exact roles of ephrin-A1 in microglia-induced angiogenesis.
In our study, the culture supernatant derived from microglia enhanced angiogenesis in brain endothelial cells. It has been reported that mouse EOC2 microglial cells stimulated vessel sprouting and branching in aortic ring cultures in the absence of a direct contact (Rymo et al., 2011), which is consistent with our findings. The facilitating effect of microglial supernatant on brain endothelial angiogenesis pointed out the existence of soluble angiogenic factors released from microglia. Our further results identified high level of TNF-α in microglial supernatant. The promoting effect of recombinant TNF-α on migration and angiogenesis of HBMECs was observed (data not shown), which is in line with the angiogenic potential of TNF-α (Fajardo et al., 1992). Moreover, the microglial supernatant-induced angiogenesis of HBMECs was effectively blocked by the neutralizing antibody against mouse TNF-α and human TNF receptor (TNFR1), respectively. These data indicated that TNF-α, released from murine BV2 microglial cells, bound with TNFR1 receptor on human brain endothelial cells to promote the endothelial angiogenesis. Our results, together with the inability of VEGF in microglia-induced angiogenesis in brain (Fantin et al., 2010), as well as the evidence against the involvement of VEGF-A in microglia-stimulated angiogenesis in an aortic ring model (Rymo et al., 2011), indicated that microglia-derived TNF-α, rather than VEGF, plays a critical role in microglia-induced brain endothelial angiogenesis.
The molecular links between TNF-α and ephrin-As remain unclear because of limited number of studies. In our study, TNF-α of microglial cells promoted transcription of ephrin-A3/4 in brain endothelial cells, leading to increased levels of ephrin-A3/4 protein. A previous study indicated that TNF-α could induce expression of ephrin-A1 in endothelial cells, which is mediated by JNK and p38-MAPK signaling pathway (Cheng and Chen, 2001). These studies suggested exogenous TNF-α could regulate ephrin-As expression through intracellular signaling. In contrast, a recent study found that TNFR1, directly interacts with EphA7 receptor upon stimulation with ephrin-A5, the ligand of EphA7, forming a three-protein complex (Lee et al., 2013). Also, the induction of cell apoptosis by ephrin-A5 (Yue et al., 1999; Depaepe et al., 2005; Lee et al., 2013; Park et al., 2013) is in line with the pro-apoptotic effect of TNF-α (Aggarwal, 2003). Thus, the physical association of TNFR1 with EphA7 and the functional similarity between their ligands suggested that TNF-α may directly affect the signaling responses to ephrin-A5 via the membrane-anchored TNFR1-EphA7 complex. Do these observations imply that TNF-α could regulate ephrin-As activity through both intracellular signaling and membrane-bound receptors? This is an interesting question needs to be addressed in future studies.
Our analysis revealed that endothelial ephrin-A3/4 protein was significantly increased at 4 and 8 h after BV2 supernatant treatment (Fig. 5B,D). Also, this increase was significantly attenuated by neutralizing antibody against TNF-α and TNFR1, respectively (Fig. 5B,D). These data were consistent with the realtime RT-PCR results (Fig. 5A,C). In contrast, at 12 h post-treatment, the ephrin-A3/4 protein was only slightly increased, without statistical significance. Thus the decrease of ephrin-A3/4 protein at 12 h in response to the neutralizing antibody was not obvious compare to that at 4 and 8 h time point. This discrepancy between mRNA and protein changes at 12 h post-treatment, but not earlier, suggested a translational regulation of endothelial ephrin-A3/4 at the later stages of BV2-supernatant treatment. It is thus suggested that ephrin-A3/4 is primarily involved in the early steps of microglia-induced angiogenesis.
Apart from the stimulatory effect on endothelial migration and angiogenesis, we found microglial supernatant suppressed the growth of brain endothelial cells (Fig. 1E), which was not likely caused by the TNF-α in microglial supernatant because recombinant TNF-α had no obvious effect on endothelial proliferation (data not shown). We thought this inhibitory effect of microglial supernatant on the proliferation of brain endothelial cells was associated with the presence of proliferation inhibitory cytokines, rather than TNF-α, in the microglial supernatant. In consistent with our results, it had been reported that endothelial proliferation was reduce by the supernatant derived from resting microglia, which was partially due to TGF-β1, but not TNF-α (Welser et al., 2010).
In summary, our data provides a novel mechanism for microglia-induced angiogenesis, that is, microglial cells upregulate angiogenic factors ephrin-A3 and ephrin-A4 in brain endothelial cells, which is mediated by microglia-released TNF-α, to promote angiogenesis of brain endothelial cells. Further study is required to investigate the ephrin-A3/4 initiated intracellular signaling pathways in brain endothelial cells. In addition, the expression of ephrin-A3 and ephrin-A4 in mouse brain microvessels was recently identified in our lab (data not shown), thus it will be interesting to explore the in vivo significance of ephrin-A3/4 in microglia-induced brain vascular angiogenesis.