Neurovascular coupling is a set of mechanisms regulating the exchanges between nerves and blood vessels, and more precisely at the cellular level, between endothelial cells and neural cells. Some of these mechanisms have been described [37, 38], but most of them remain unclear, which confirm the need for further study of the interplay between these cell types that can support the regulation of angiogenesis by nerve fibers. In more details, through this coupling, local neuronal activity plays an important role on angiogenesis and physiology of neighboring microvessels. In this context and based on our previous data obtained in vitro on the effect of sensory neurons on endothelial cells [3], our aim here was to discriminate the cell signaling involved in the neurovascular coupling and to focus on the role of the two neurotransmitters CGRP and SP produced by sensory neurons on the cell communication with endothelial cells.

Among the responses of endothelial cells to different chemical (e.g., neurotransmitters) and physical stimuli (e.g., shear stress pulsatile stretch), the increase in intracellular Ca2+ concentration can further trigger diverse signaling pathways, known to regulate the expression of angiogenic markers and EC functions [39].

In this work, we showed that the stimulation of SNs with capsaicin results an increase on intracellular calcium level in ECs (Fig. 1c-d). The effect of stimulation of SNs by capsaicin was observed on a group of endothelial cells whose resting [Ca2 + i] had been measured. This observation suggests that the stimulation of SNs by capsaicin and the subsequent release of vesicles [40], including CGRP and SP, can be responsible for the calcium influx observed in ECs. The role of CGRP and SP has been previously identified in our work as mediators stimulating the gene expression of angiogenic markers of endothelial cell functions [3]. Our results also suggest that [Ca2 + i] is increased in ECs in response to SNs stimulation with capsaicin and that ECs communicate quickly to present a harmonized response.

In particular, calcium signaling is known to occur via gap junctions and connexin (Cx) hemichannels [30]. Gap junctions are constituted by Cx that contain multiple phosphorylation sites and the phosphorylation state is well known to regulate the opening or closing of channels [41]. Among the Cx, Cx43 is one of the most widely expressed connexin in tissues and cell lines, especially in the vascularized bone tissue [41]. Phosphorylation of Cx43 is known to be involved in the regulation of hemichannels and gap junctional communication through trafficking, assembly/disassembly, degradation as well as gap junction channel gating [42]. More precisely, it has been described that Cx43 phosphorylation decreases intercellular communication [42,43,44,45], however, Ca2+ transfer is rarely affected. Indeed, even if Cx43 phosphorylation can inhibit other molecules transfer such as IP3, it allows Ca2+ movement [46]. Thus, the calcium signaling in ECs observed in presence of SNs in the Fig. 1 might still be mediated through Cx43 activity.

When looking at Cx43 in our co-culture model of ECs with SNs, an increase in Cx43 expression over time likely due to the increase in culture confluence (Fig. 2c), and a tendency in the phosphorylation rate were observed over time in the presence of SNs (Fig. 2d). Moreover, Cx43 protein expression tends to be upregulated in the presence of SNs relative to ECs cultured alone at day 4. An upregulation of Cx43 has been shown to positively modulate the angiogenic potential of endothelial cells by stimulating endothelial cell migration and network formation [47]. However, here, antagonists of CGRP and SP had no significant influence on the Cx43 protein expression and phosphorylation ratio. Antagonist of SP may have showed a tendency to decrease protein expression (p = 0.07), but it did not influence significantly the phosphorylation ratio. These results should be interpreted with caution and should be confirmed by further experiments before concluding on the role of SNs and these neuropeptides to control site-specific connexin phosphorylation, associated with a control of gap junctional communication through Cx43.

NO is also a potent mediator produced by the eNOS in ECs and is involved in endothelial functions by maintaining vascular homeostasis for example [48]. A Ca2+-dependent regulation of eNOS producing NO has already been described [49]. The eNOS is an important signal generator involved in the control of vascular tone and angiogenesis, regulated by several kinases acting at different sites to activate or inhibit NO production. Our results show that SNs increase NO production in ECs, probably involving a decrease in the eNOS T495 phosphorylation site without affecting eNOS protein expression, resulting thus in a stimulatory effect on eNOS activity. While SP seems not to be involved in the increase of NO production (Fig. 3e-f), the neuropeptide CGRP seems to be one of the mediators of this communication, since NO production is decreased in the presence of AntCGRP in the co-culture of ECs and SNs and increased when ECs are stimulated with synthetic CGRP. Our data are in accordance with the well-known vasodilator role of CGRP [50] and importantly, identify an important mechanism in which SNs can regulate ECs functions, in a NO-dependent manner.

Interestingly, a study conducted by Gaete and coauthors have suggested a contradictory effect of CGRP on NO production in mesenteric arterial beds [22]. The authors showed that CGRP signaling inhibits NO production through gap junction protein pannexin-1 channel activation in ECs. Besides we did not evaluate the role of pannexin-1, here we used ECs isolated from bone marrow microvasculature. The differences in the cells and tissues origin as well as the presence of different cell types in mesenteric arterial beds may explain the contradictory results observed in both situations. We developed an in vitro model to elucidate the cell communication and signaling between dissociated Dorsal Root Ganglia (DRG) cultures (including a mix of SNs and Schwann cells) and endothelial cells. Our approach does not reflect the complexity of the vascular tissue composition and response to SNs stimulation, however it allows us to describe the particular ECs response and to identify the main actors of their communication with SNs here and in our previous study [3].

Recently, an interesting relation between NO, Cx43 and Ca2+ signaling has been suggested to regulate endothelial cell migration, proposed by Espinoza and Figueroa [51]. Using rat primary microvascular endothelial cells and a 2D migration scratch assay, the authors suggested the involvement of NO-mediated S-nitrosylation in Cx43 hemichannels’ functions [51]. The activation of Cx43 hemichannels by S-nitrosylation should play a critical role in the long-lasting increase in [Ca2 + i] that directs endothelial cell migration and network formation, suggesting that this mechanism may contribute to the NO-mediated effects in angiogenesis. The data previously published by our team showing that SNs and the neuropeptides SP and CGRP participate in the upregulation of angiogenesis markers in ECs such as Angpt1, Col4, Mmp2 and interestingly, VegfA [3]. As well known, VEGF play an important role in ECs by contributing to vascular integrity and endothelial homeostasis [52, 53] and but also it stimulates eNOS expression and activation [49, 54]. Here, we show that CGRP regulates NO production, likely indirectly participating in the activation of Cx43 channels through NO-mediated S-nitrosylation, permitting the Ca2+ entry and thus triggering multiple mechanisms stimulating ECs functions. In this sense, our findings presented here show the increase in [Ca2 + i] in ECs after SNs stimulation with capsaicin, a significant upregulation of NO production in ECs co-cultured with SNs, and only a tendency in Cx43 protein expression in ECs co-cultured with SNs at day 4 (p = 0,07). These data are in accordance with Espinoza’s and Figueroa’s observations and confirm that SNs play an important role in the control of the endothelial cell functions such as migration and network formation.

MAPK signaling proteins such as ERK1/2 and p38 are also known to be involved in processes such as angiogenesis by regulation of eNOS [55, 56] for example. Moreover, the ERK1/2 and p38 have been shown to be activated by the recognition of SP and CGRP, respectively, by their receptors in mesenchymal stem cells [57]. This activation leads to the migration of MSCs to bone formation sites and increased osteogenesis. Here, p38 protein expression and p38 phosphorylation ratio are not modulated by the presence of SNs nor by the presence of AntCGRP and AntSP at day 4 and day 7. Regarding Erk1/2, a downregulation of phosphorylation ratio was observed in ECs co-cultured with SNs at day 7 relative to day 4 and ECs cultured alone at day 7 (Fig. 5c), followed by an upregulation of Erk1/2 phosphorylation ratio in the presence of AntSP at day 7 (Fig. 5f). However, the effect of the neuropeptide SP on Erk1/2 phosphorylation was not confirmed when ECs alone were cultured in well plates in the presence of 1µM SP. It is important to note that when we identified the downregulation of Erk1/2 phosphorylation ratio at day 7 and the upregulation of Erk1/2 phosphorylation ratio in the presence of AntSP, the ECs were co-cultured with SNs, reflecting a more complex physiological communication between both cell types relative to the culture of ECs alone with the culture medium supplemented with CGRP and SP individually. Indeed, the neuropeptides availability and concentrations are not the same for both situations, which can explain the absence of effect of SP on Erk1/2 phosphorylation ratio when SP was added to the ECs culture medium (Fig. 5i). It is known that SNs secrete other molecules and neuropeptides, which could take over or act in synergy with SP, explaining why SP in ECs culture medium does not significantly affect Erk1/2 phosphorylation ratio. A similar scenario was observed in our previous study [3], in which we identified that SP when in culture medium can upregulate Mmp2 expression, however when ECs were co-cultured with SNs in the presence of AntSP, Mmp2 expression was only significantly reduced when AntSP and AntCGRP were used combined, suggesting that other molecules might be involved in the regulation of Mmp2 expression by SNs.

It is well described in the literature that in ECs, Erk1/2 phosphorylation increases as a quick response followed by an angiogenic signal, and then the phosphorylation level is decreased due to the activity of Erk1/2 phosphatases [58, 59]. Here, the time course used for the analysis of Erk1/2 expression and phosphorylation in the co-cultured ECs (4 and 7 days allowing the neurite outgrowth), that could impact angiogenesis, could not be appropriate with a signal of the MAPK activities. However, in our study, when we consider the Erk1/2 phosphorylation ratio in ECs co-cultured with SNs from day 4 to 7, we observe a decrease in the phosphorylation ratio. This downregulation could be the result of the Erk1/2 phosphatases activity, acting on a return of Erk1/2 phosphorylation ratio to a base level after an earlier stimulation. Moreover, EC culture medium is a growth factor-rich environment, including VEGF-A in the composition, which could explain the high levels of Erk1/2 phosphorylation ratio in the group of ECs cultured alone at both time points.