Heterocellular Contact Can Dictate Arterial Function

Abstract

Rationale:

Resistance arteries and conduit arteries rely upon different relative contributions of endothelial derived hyperpolarization (EDH) versus nitric oxide (NO) to achieve dilatory heterocellular signaling. Anatomically, resistance arteries use myoendothelial junctions (MEJs), endothelial cell (EC) projections that make contact with smooth muscle cells (SMCs). Conduit arteries have very few to no MEJs.

Objective:

Determine if the presence of MEJs in conduit arteries can alter heterocellular signaling.

Methods and Results:

We previously demonstrated that plasminogen activator inhibitor-1 (PAI-1) can regulate formation of MEJs. Thus, we applied pluronic gel containing PAI-1 directly to conduit arteries (carotid arteries, CAs) to determine if this could induce formation of MEJs. We found a significant increase in EC projections resembling MEJs that correlated with increased biocytin dye transfer from ECs to SMCs. Next, we used pressure myography to investigate whether these structural changes were accompanied by a functional change in vasodilatory signaling. Interestingly, PAI-1-treated CAs underwent a switch from a conduit to resistance artery vasodilatory profile via diminished NO signaling, and increased EDH signaling in response to the endothelium-dependent agonists Ach and NS309. Following PAI-1 application, we also found a significant increase in carotid expression of endothelial alpha globin, a protein predominantly expressed in resistance arteries. Carotids from mice with PAI-1, but lacking alpha globin (Hba1^−/−), demonstrated that L-NAME, an inhibitor of NO signaling, was able to prevent arterial relaxation.

Conclusions:

The presence or absence of MEJs is an important determinant for influencing heterocellular communication in the arterial wall. In particular, alpha globin expression, induced within newly formed EC projections, may influence the balance between EDH and NO-mediated vasodilation.

INTRODUCTION

Endothelial cell (EC) mediated vasodilation of arteries can generally be achieved either through production of nitric oxide (NO) and/or via endothelial derived hyperpolarization (EDH). NO is produced by endothelial nitric oxide synthase (eNOS) and diffuses to smooth muscle cells (SMCs) to promote vasodilation by binding to its cytosolic receptor soluble guanylyl cyclase, which produces cyclic guanosine monophosphate.¹ EDH refers to a signaling pathway that typically begins with the opening of small- and intermediate-conductance calcium-activated potassium channels (SK_Ca and IK_Ca) on ECs, leading to the efflux of potassium ions from smooth muscle cells (SMCs).^2–5 Hyperpolarization of smooth muscle due to reduced cytosolic positive charge leads to dilation. Both mechanisms of dilatory signals described above must be tightly regulated throughout the vascular tree to maintain blood pressure homeostasis. It has been established that the relative contribution of these endothelial-mediated vasodilatory mechanisms differ based on the size of the vessel; large conduit arteries such as the carotid and aorta rely on NO signaling to dilate, whereas in smaller, resistance arteries like the mesenteric arteries, EDH is an additional and important mechanism of dilation.^{1, 2, 6}

An anatomical difference exists in the structure of the conduit and resistance arteries that may account for the difference in functional dilation. ECs in resistance arteries have unique signaling microdomains named myoendothelial junctions (MEJs) that penetrate the fibrous internal elastic lamina (IEL) to make heterocellular contact with SMCs via gap junctions.^7–11 This direct heterocellular contact allows for electrochemical communication between ECs and SMCs to facilitate vasodilation predominantly through the EDH pathway (reviewed extensively^{4, 12–15}). In contrast, conduit arteries have a significantly reduced number of MEJs compared to resistance arteries, as well as a significantly thicker IEL (~0.5 μm in resistance arteries compared to up to 5 μm in aorta) that allows them to handle high transmural pressures in proximity to the heart.^16–19 Conduit arteries preferentially dilate via the NO pathway, presumably because NO is a highly diffusible signaling molecule that can cross the multiple thick IEL layers.^{15, 20–23} In contrast, the expression of alpha globin in endothelial cells of resistance arteries may provide a constraint on the ability of NO to diffuse from endothelium to smooth muscle, limiting the role of NO in this setting.^23–25

Our lab has previously demonstrated that plasminogen activator inhibitor-1 (PAI-1) is enriched at the MEJ and that its depletion results in decreased MEJ formation.^{26, 27} However, it remains to be demonstrated whether PAI-1 is sufficient to induce MEJ formation. In this study, we investigate the formation of IEL holes in the carotid (a conduit artery) following the direct application of PAI-1 in live mice. We observed transient formation of EC projections (resembling MEJs) after 7 days that retreat by day 21. Interestingly, an increased proportion of non-NO, EDH-based arterial relaxation accompanies this anatomical phenotype. Additionally, we have demonstrated alpha globin expression in the carotid arteries upon induction of EC projections can dictate NO or EDH-based vasodilation. Thus, we have established a model of MEJ formation through the direct application of PAI-1 to a conduit artery and have shown that functional heterocellular contact is correlated with increased reliance on EDH-mediated pathways.

METHODS

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Please see Online Methods for extensive details on the experiments used in this manuscript.

Mice.

All mice and procedures were approved by the University of Virginia Animal Care and Use Committee. Male C57Bl/6 (Taconic Biosciences) and Hba1^−/− (Jackson Labs) mice were used between 10–12 weeks of age. Male mice were used in this study for consistency in comparison of results to our previous work, as well as to minimize possible variability given N-values.

Plasminogen activator inhibitor-1 (PAI-1).

Lyophilized PAI-1 (500 IU, Technoclone, Vienna) powder was reconstituted by addition of 200 μl distilled water followed by gentle agitation for 5 minutes at room temperature (2.5 × 10⁶ IU/L stock solution) that yields enzyme in its active form for ~12 hours or being bound to vitronectin. 40 μl/tube aliquots were stored at −70°C. At 4°C immediately prior to use, 40 μl stock solution of PAI-1 was mixed with 460 μl of sterile saline solution containing 35% F127 Pluronic Gel (Molecular Probes).

Mouse surgery.

Mice were anesthetized with an intraperitoneal (i.p.) injection of a freshly made ketamine/xylazine mixture (Fort Dodge; 80 mg/kg and Vedco; 6 mg/kg, respectively). The pluronic gel/PAI-1 mixture (preparation described above) and pipettes were kept at 4°C and placed on ice just before use. While holding the carotid artery (CA) away from surrounding tissues, the pluronic gel/PAI-1 mixture (100 IU PAI-1 in 500 μl or 2 × 10⁵ IU/L solution) or pluronic gel (500 μl of 35% pluronic gel) was topically applied to the outside of the left CA and solidified within one minute of application. Pluronic gel/PAI-1 mixture or pluronic gel only was kept on CAs for 1, 2, 3, 4, 5, 7, 14 or 21 days. At the end of the pluronic gel/PAI-1 or gel only incubation, mice were sacrificed using CO₂ asphyxia and CAs carefully removed (see Online Methods). Once free of surrounding tissues and cleaned, the CA was used for a variety of experiments (see Online Figure I for visual representation). Carotids for pressure myography were approximately 500 μm in diameter and 5 mm in length.

Visualizing holes and MEJs.

Three methods were used, including: (1) fluorescent microscopic measurement of holes in the IEL, which used Alexa Fluor 633 hydrazide (Molecular Probes) to visualize the IEL, (2) endothelial denuded CA viewed with scanning electron microscopy (SEM), and (3) transmission electron microscopy (TEM) on intact CAs to visualize endothelial extensions.

Measurements of carotid artery vasoreactivity.

CAs were isolated after treatment as described above, and cannulated in a pressure myograph chamber (Danish MyoTechnology, DMT).²⁸

In order to evaluate the effect of PAI-1 on the vasoreactivity of CA, where indicated, experiments were conducted in the presence or absence of selective blockers of endothelium-mediated vasodilator pathways. In order to block NO generation alone, we used the nitric oxide synthase (NOS) inhibitor L-nitro-arginine methyl ester (L-NAME, 300 μmol/L; Sigma). In order to block both NO generation and the EDH pathway, we used (1) the cyclooxygenase inhibitor indomethacin (10 μmol/L; Sigma), (2) the intermediate-conductance calcium-activated potassium channel (IK_Ca) blocker 1-[(2-chlorophenyl) diphenylmethyl]-1H-pyrazole (TRAM-34, 10 μmol/L; Sigma), (3) the small-conductance calcium-activated potassium (SK_Ca) channel blocker apamin (300 nmol/L), and (4) L-NAME. After a 30-minute stabilization period in the circulating bath, the arteries were pre-constricted with 50 μmol/L phenylephrine (PE) followed by cumulative concentrations of Ach (10⁻⁹ to 10⁻⁴ mol/L). To activate IK_Ca and SK_Ca channels, NS309 (3 μmol/L), was added to pre-constricted CAs. All vessels were then washed with a Ca²⁺-free Krebs-HEPES solution to obtain maximal passive diameter of the vessels. Internal diameter was measured throughout the experiments using the DMT MyoVIEW software. Vasoconstriction to PE was calculated as percent of initial diameter: %constriction = D_PE / D_initial * 100, and relaxation to Ach or NS309 was calculated as a % relaxation: % relaxation = [(D_Ach − D_PE) * 100]/(D_max−D_PE), where D_PE was the diameter of the CA 10 minutes after application of 50 μmol/L PE; D_initial was the diameter prior to the addition of PE; D_Ach was the diameter of the CA after application of a given dose of Ach; and D_max was the maximal diameter of the CA measured at the end of experiment.

Biocytin dye transfer.

CAs were removed per above after experimental conditions, cannulated, and flushed with Ca²⁺-free Krebs-HEPES buffer, followed by perfusion through with new buffer supplemented with 0.01% Tween20 + 2.68 mmol/L (1 mg/ml) biocytin for 15 minutes at room temperature. After washing the lumen and bath with Ca²⁺-free Krebs-HEPES buffer, CAs were fixed in 4% PFA for an hour at room temperature, washed in PBS, and cut open longitudinally with the endothelium en face. Next, they were permeabilized in PBS containing 0.01% TritonX-100 for 30 minutes followed by an hour incubation in 1:500 streptavidin conjugated to Alexa Fluor 568 (also in PBS with 0.01% TritonX-100). CAs were imaged with a Olympus Fluoview 1000 confocal microscope.

Western blot.

Blotting was performed as previously described²⁹ and is briefly described in the Online Methods.

Statistics.

For in vivo experiments, animals were randomly assigned groups by an independent third party. An independent researcher performed vasoreactivity or morphology and scores were re-assigned to treatment groups afterward. Power analysis was performed (α=0.05; > 20% change) to determine group sizes and study power (>0.80) using G*Power^{30, 31}. Thus, some experiments only required an N=3 while others required N ≥ 4. One-way ANOVA followed by Tukey’s multiple comparisons test or two-way ANOVA followed by Sidak’s multiple comparisons test were used as necessary. For non-gaussian distributed data (Shaprio-Wilk normality test), non-parametric statistical tests were used and indicated in the figure legend (Kruskal-Wallis test followed by Dunn’s multiple comparisons test). Analysis was performed using GraphPad Prism version 7.0 for Mac OS X (GraphPad Software, La Jolla, CA) with p<0.05 considered significantly different. Data are expressed as mean +/− SEM.

RESULTS

Functional MEJs form in the IEL of carotids after application of PAI-1.

Like other conduit arteries, CAs have a thick IEL separating the endothelium and smooth muscle, and contain very few MEJs.^{32, 33} We have previously demonstrated that PAI-1 can regulate the formation of MEJs.²⁶ Using a pluronic gel, we applied PAI-1 to CAs of C57Bl/6 mice in order to determine whether these conduit arteries could form MEJs, and if so, whether they were functional (Online Figure I). PAI-1 was found throughout the carotid wall after application (Online Figure II). We initially screened for extent of hole formation in the IEL after PAI-1 application via autofluorescence of carotids laid en face (Online Figure III). We found that the maximum number of holes could be identified 7 days after PAI-1 application to CAs. After 7 days, the number of holes steadily regressed more than halfway to baseline by 21 days post PAI-1 application. Based on these results, we used time points at 0, 1, 7, and 21 days following PAI-1 application for further experiments. Through visualizing the IEL with autofluorescence, we quantified holes in the IEL, which we found to be significantly increased 7 days post-application of PAI-1 (Figure 1A–G). We confirmed the formation of holes in the CA IEL by imaging carotids stripped of endothelium with SEM (Figure 1H–N). It has previously been shown that not all IEL holes contain MEJ protein markers.¹² Therefore, we sectioned CAs in transverse at each of the time points and identified endothelial cell extensions through the IEL (Figure 1O–U). We observed a significant increase in the number of EC projections at day 7 following PAI-1 treatment, but these EC projections were transient, as their number reverted almost to baseline by day 21 with PAI-1. Although resembling MEJs, the EC projections could not be observed to physically touch SMC (Online Figure IV), likely due to width of the cellular projection when heterocellular contact was made. In addition, the sizes of the IEL holes in conduit arteries after PAI-1 was similar to those found in mesenteric or skeletal muscle beds (Online Figure V). These changes that occurred were independent of the thickness of the IEL (Online Figure VI) and normal EC junctions (e.g., VE-cadherin) were preserved (Online Figure VII). Overall, exogenous PAI-1 application to the carotid artery was sufficient to induce the formation of endothelial projections in the CA IEL, which is an anatomical hallmark normally restricted to small-diameter arterioles and resistance arteries.

Figure 1: Formation of holes and cellular extensions into the internal elastic lamina of carotids after PAI-1 application.

A–F, Representative confocal images of the IEL from CA en face preparations at 0, 7, and 21 days following PAI-1 treatment and from sham control. The number of holes per 1000 μm² is quantified in G. Scale bar in A is 20 μm and representative for A–F. In H-M, SEM images of the IEL from CA en face preparations at 0, 7, and 21 days following PAI-1 treatment and from sham control, with quantification of holes per 1000 μm² in N. Scale bar in H is 2 μm and representative for H-M. In O-T, TEM images on transverse sections at 0, 7, and 21 days following PAI-1 treatment and the sham controls, with number of projections per 100 linear μm quantified in U. In all cases, each dot on graphs in G, N, and U indicates the average of 3 random fields of view from 1 mouse. Scale bar in O is 1 μm and representative for O-T. In each graph, N=4 mice per experimental time point; *=p<0.05.

Although we could identify increased endothelial projections at day 7 of PAI-1 treatment, it was unknown if they made functional heterocellular contact. One important test of functional heterocellular contact at MEJs is the coupling of ECs to SMCs through gap junctions. Thus, we examined dye transfer from ECs to SMCs as an indicator of functional gap junction channels. ECs of a cannulated CA were loaded with biocytin using a pinocytotic method that loads >90% of ECs (Molecular Probes). A small, uncharged molecule, biocytin, can traverse gap junctions, and due to its biotin moiety, it can be detected with fluorescently-labeled streptavidin. Because of the lack of MEJs and heterocellular gap junctional communication in CAs at day 0, we were not able to observe biocytin in the SMC layer (Figure 2A–B). However, at day 7 post-PAI-1 application, we were able to measure biocytin transfer from ECs to SMCs via fluorescent signal present in the SMC layer. Although there are no truly specific pharmacological means to block gap junctions (e.g.,^{33, 34}), the addition of α-glycyrrhetinic acid (α-GA), a purported gap junction inhibitor, prevented the biocytin dye transfer (Figure 2A–B). Thus, the EC projections that we observed can couple ECs to SMCs, presumably allowing dye to pass through heterocellular gap junction channels. Next, we wanted to test for functional heterocellular contact using a vasoreactivity approach to assess changes in relaxation to NS309 (an activator of SK_Ca and IK_Ca channels; 3 μmol/L). Application of NS309 tests whether EDH signaling is active between endothelium and smooth muscle by inducing endothelial hyperpolarization to achieve subsequent smooth muscle vasodilation.³⁵ Sham- or gel-treated CAs did not relax more to NS309 treatment compared to vessels from day 0 PAI-1 treated animals (Figure 2C). However, at day 7 post-PAI-1 application, NS309 induced a significant increase in relaxation, which was lost after denuding endothelium or adding α-GA. Day 21 carotids also did not have a significant NS309 relaxation. In addition, Cx40 was observed in the holes of the IEL after 7 days of PAI-1 (Online Figure VII). Thus, we can correlate the formation of IEL holes and corresponding EC projections (both seen maximally at day 7 post-PAI-1 application) with functional EC and SMC coupling.

Figure 2: Endothelium and smooth muscle cells of the carotid are functionally coupled after PAI-1 application.

A, representative images of biocytin highlighted with streptavidin conjugated Alexa 568 fluorophore (red); nuclei in blue (DAPI). The XY-plane of endothelium is shown in the top row and smooth muscle in the one below. Note the addition of α-GA blocked biocytin transfer to smooth muscle in carotids treated with PAI-1 for 7 days. Scale bar is 20 μm. A Kruskal-Wallis test followed by Dunn’s multiple comparisons test was used to establish significance. B, Percent streptavidin-positive smooth muscle cells per 10,000 μm² after loading of endothelium with biocytin. In C, carotids were stimulated with 3 μmol/L NS309 and relaxation was measured. Each dot on graphs in B and C indicates the average of 3 random fields of view from 1 mouse, minimum N=3 mice per experimental condition. Note the addition of α-GA, and denuding endothelium, prevented relaxation of carotids treated with PAI-1 for 7 days. In each graph, *=p<0.05.

MEJs are necessary, but not sufficient, to alter endothelial-induced dilation.

In arteries with many MEJs (e.g., resistance arteries), endothelial derived vasodilation using cholinergic stimulation (via acetylcholine, Ach) is mainly through EDH^{4, 13–15}; whereas in arteries with few MEJs, NO signaling predominates.^{1, 3} We reasoned that the new formation of endothelial projections (i.e., MEJs) in a conduit artery could alter the endothelial-dependent mechanism of vasodilation. Dose-responses with Ach on control CAs (e.g., sham, pluronic gel only, or carotids with few MEJs) demonstrated that eNOS inhibition alone (with 300 μmol/L L-NAME, a non-specific NOS inhibitor) largely blunted relaxation (Figure 3A–B; Online Figure VIII); this is a widely described finding in CAs and other conduit arteries that demonstrate endothelial-derived vasodilation dominated by NO signaling.^{21, 36} However, at day 7 following PAI-1 application, when holes in the carotid IEL and EC projections were maximally present and heterocellular dye transfer and relaxation to NS309 both were significantly increased, L-NAME alone showed a reduced inhibition of relaxation. It was only with the addition of apamin, TRAM-34 and indomethacin (SK_CA, IK_CA, cyclooxygenase inhibitors, respectively) that the Ach response was significantly inhibited (Figure 3A–B; Online Figure VIII). In experiments on carotids treated with PAI-1 for 7 days, an inhibitory response was observed with the combination of indomethacin, TRAM-34, and apamin, as well as with only TRAM-34, apamin, or L-NAME (Online Figure IX). This predominance of non-NO based relaxation is usually only observed in resistance arteries where MEJs are prevalent.

Figure 3: When myoendothelial junctions are present in carotids, acetylcholine-induced relaxation is no longer reliant on nitric oxide.

A, dose response curves to Ach; black circles are untreated control carotids, blue triangles are carotids treated with L-NAME, and red squares are carotids treated with L-NAME plus the inhibitors TRAM-34, indomethacin, and apamin. N values indicate the number of carotids. In all experiments, a minimum of N=3 mice were used per time point per experimental condition. B, representative traces of Ach dose-response curves in the presence of L-NAME; green arrow is initial phenylephrine (PE) constriction (in %), with each Ach dose indicated by a blue arrow. Red arrows indicate perfusion with Ca²⁺-free buffer. *=p<0.05 indicates control compared to L-NAME, TRAM-34, indomethacin, and apamin; #=p<0.05 indicates control compared to L-NAME.

Lastly, we tested whether the formation of MEJs alone was sufficient to alter the mechanism for endothelial-derived vasodilation. We have previously demonstrated that alpha globin is a potent NO-scavenger and is present in resistance arteries, with little to no expression in conduit arteries.^{23, 25} At day 7 post PAI-1 application, when MEJs and Ach-induced non-NO vasodilation are maximal, we found a significant increase in expression of alpha globin (Figure 4A–B). We also observed alpha globin in the holes of the IEL at day 7 (Online Figure VII). In mice with a disrupted Hba1 gene (Hba1^−/−), we found no alpha globin expression at day 7 post-PAI-1 application, whereas in littermate controls (Hba1^+/+) the protein was produced at normal levels similar to that seen in C57Bl/6 (Figure 4A–B; Online Figure X). Importantly, the absence of alpha globin expression at day 7 post-PAI-1 application in Hba1^−/− mice had no effect on (1) the number of holes in the IEL that were present (Figure 4C–D), (2) endothelial cell projections into the IEL (Figure 4E), (3) dye transfer to SMC (Figure 4F), or (4) relaxation to NS309 (Figure 4G). However, CAs of mice lacking alpha globin expression were unable to switch their dominant vasodilatory mechanism from NO to non-NO. As expected, CAs from Hba^+/+ littermate controls 7 days after PAI-1 application appeared similar to those seen in C57Bl/6 (Figure 4H), with non-NO dominating the vasodilatory response. In Hba1^−/− mice at the same treatment time point, L-NAME alone was able to significantly blunt relaxation, indicating that NO was still the dominant dilatory pathway (Figure 4I). Thus, MEJs are an important component of enabling EC-driven SMC vasodilation. Through limiting the role of NO signaling, the expression of alpha globin at MEJs may facilitate non-NO based signaling (e.g., EDH) to predominate in resistance arteries.

Figure 4: Induction of alpha globin expression within heterocellular contacts regulates nitric oxide diffusion, but not myoendothelial junction formation.

A, western blot for alpha globin from carotids, with VE-cadherin used as loading control for endothelium. In the right panel, expression of alpha globin at 7 days post-PAI-1 application is shown in samples from Hba1^+/+ and Hba1^−/− mice. Arrows indicate 10 kDa for alpha globin, and 100 kDa for VE-cadherin. In B, note the significant increase in alpha globin 7 days after PAI-1 treatment that decreases by day 21, matching hole and endothelial projection prevalence. In C–D, the number of holes in the IEL were assessed via fluorescent microscopy (C) and SEM (D). In E, endothelial projections down into the IEL were assessed the same way as in Figure 1. Coupling of endothelium and smooth muscle was experimentally determined using biocytin dye transfer (F) and relaxation to NS309 (G). In B-G, colors indicate identical time points; black is day 0 after PAI-1 on a C57Bl/6 mouse, grey is day 7 after PAI-1 on a C57Bl/6 mouse, brown is day 7 after PAI-1 on a Hba^+/+ mouse, and pink is day 7 after PAI-1 on a Hba^−/− mouse. In H-I, CAs were collected from Hba^+/+ (H) or HBA1^−/− (I) mice 7 days after PAI-1 application, cannulated, pressurized, and tested in the presence of L-NAME or L-NAME with the inhibitors, indomethacin, TRAM-34, and apamin. Loss of alpha globin had no effect on the number of holes in the IEL, cellular extensions, dye transfer, or NS309 induced relaxation (IK_Ca/SK_Ca), indicating MEJs were still capable of forming (C–G). However, data in H-I demonstrate the mechanism of endothelial vasodilation was altered without alpha globin present. In all cases, each dot on graphs in C–E indicates the average of 3 random fields of view from 1 mouse. In F–G, each dot represents 1 mouse. In F, a Kruskal-Wallis test followed by Dunn’s multiple comparisons test was used to establish significance In each graph, * and # = p<0.05. In B–G, * indicates control compared to respective carotid treatment. In H and I, * indicates control compared to inhibitors L-NAME, TRAM-34, indomethacin, and apamin; # indicates control compared to L-NAME.

DISCUSSION

We have previously demonstrated that PAI-1 is enriched in MEJs.²⁶ In this study, we further investigated the involvement of PAI-1 in MEJ formation by directly applying PAI-1 to the CAs of mice via pluronic F-127 gel, a useful medium that can deliver high concentrations of PAI-1 to carotid arteries while avoiding adverse systemic effects. The local increase in PAI-1 at the CA resulted in major phenotypic shifts in structure and function to make this conduit resemble a resistance artery. For example, CAs normally have very few holes in the IEL, nor EC projections (i.e., MEJs) through those holes; however, following 7 days of PAI-1 treatment, both of these anatomical phenomena developed. In addition, indications of direct heterocellular coupling between endothelium and smooth muscle were also apparent. This included Cx40 expression in the newly formed IEL holes, blocked biocytin transfer with gap junction inhibition via α-GA, and NS309 dilation. However, direct evidence of heterocellular contact via electrical coupling (e.g.,^{5, 37}) could not be demonstrated. Regardless, the significant upregulation of IEL holes, endothelial projections, and functional heterocellular coupling represents an important anatomical and functional shift in a conduit artery.

It isn’t clear why the IEL holes are temporary, but it could represent either an unresolved inflammatory response, binding to vitronectin, or degradation of the exogenously applied PAI-1; this will require elaboration in future studies. However, this temporal loss of the IEL holes formed following PAI-1 treatment speaks to the dynamic nature of IEL holes/MEJs and the potential necessity of a constant stimulus to maintain the ability to form MEJs. Here we report not only an increase in IEL holes but also the formation of endothelial projections. It has previously been shown that not all IEL holes contain markers for MEJ proteins.¹² Although this manuscript does not address the proportion of IEL holes that contain MEJ markers, the significant upregulation of endothelial projections is itself an important anatomic change. Since exogenous application of PAI-1 alone induces formation of IEL holes and MEJ-like projections within the carotid, this represents an important advancement in our understanding of the development of these critical vascular signaling microdomains. The spatial distribution of PAI-1 across the vascular wall may be one such avenue to explore; with PAI-1’s local translation and enrichment at MEJs³⁸, it is possible the MEJ signaling domain may be capable of regulating its own formation.

Functionally, the induction of endothelial projections (i.e., MEJs) in the carotid was able to introduce a larger contribution of EDH into the vasodilatory component of these conduit arteries. Pressure myography experiments revealed relaxation via the IK_Ca/SK_Ca-channel activator NS309 only in PAI-1-treated CAs at 7 days post-treatment. In addition, L-NAME, an inhibitor of NO production, was unable to prevent vasodilation via the Ach pathway in PAI-1-treated carotids. These pressure myography results indicate a switch from primarily NO-based signaling, normally characteristic of conduit arteries, to a predominantly EDH-based mechanism, characteristic of resistance arteries. The correlation of this switch in dilatory mechanism alongside increased MEJ formation mimics the anatomical and functional observations of resistance vessels; thereby implicating MEJs as a potentially critical endothelial signaling microdomain that determines the dominant dilatory pathway of arteries.

Interestingly, increased alpha globin expression was also detected in CAs following PAI-1 treatment. Alpha globin expression is normally limited to resistance arteries, but not conduit arteries.³⁹ Alpha globin has been shown to complex with eNOS to regulate NO-signaling in the resistance vasculature.^{24, 25} Its expression at the MEJ may act to finely control vasodilation by limiting the diffusion of eNOS-derived NO to the SMC.³⁹ Contact between ECs and SMCs is required for the expression of alpha globin in ECs^{23, 40}, and in the newly formed EC projections that resulted from PAI-1 treatment, alpha globin expression is significantly increased which correlated with limited dilatory effect of NO (Figure 4H). However, in mice with a disrupted alpha globin gene product (due to the insertion of a neomycin cassette into exon 2 of the Hba1 gene), the dependence on NO signaling is not lost: relaxation is almost completely blunted with L-NAME alone, with little additive effect from EDH inhibitors (Figure 4I). Within the vasculature, cell-cell interactions induce protein expression to influence cellular function^{23, 40–43} and our results embody this motif: alpha globin expression is induced upon heterocellular contact (EC-SMC coupling) and alters cellular function (altered mechanism of arterial relaxation). This data also fits well with previous work demonstrating alpha globin’s regulation of eNOS^{24, 25} and suggests that alpha globin expression is important for EDH based signaling to predominate. Indeed, disruption of the alpha globin/eNOS interaction via a mimetic peptide leads to increased NO signaling in resistance arteries.^{25, 39} Based on this and the data we show here, alpha globin at newly formed endothelial projections in PAI-1-treated carotids likely participates in the switch to EDH-based relaxation through limiting the effects of NO-based vasodilation. It is also possible that the PAI-1 enzyme may be simultaneously acting on eNOS to reduce NO-availability, enhancing the inhibitory effect to ensure EDH predominates. The interaction between PAI-1 and eNOS may be an interesting avenue to explore in future studies.

Since the major anatomical differences between a conduit and resistance artery are the IEL and points of heterocellular contact via MEJs, our finding that the induction of EC projections (i.e., MEJs) can switch a conduit artery’s dilatory mechanism from NO-dominated to likely EDH-dominated demonstrates a match between form and function. Our data suggests that MEJs with alpha globin are an important signaling microdomain and could be the reason why EDH predominates in resistance arteries.

NOVELTY AND SIGNIFICANCE.

What Is Known?

• IEL holes and MEJs are hallmarks of resistance arteries, but both are sparse in conduit arteries.

• Endothelial mediated vasodilation occurs in resistance arteries largely via EDH and predominantly through NO in larger arteries.

• Alpha globin predominantes in endothelial cells of resistance arteries.

What New Information Does This Article Contribute?

• IEL holes and EC projections (i.e., MEJs) can form in conduit arteries following in vivo treatment with PAI-1.

• The mechanism of endothelial mediated vasodilation is dependent on cellular contact between endothelium and smooth muscle.

• Alpha globin may in part regulate the altered signaling mechanisms in resistance and conduit arteries.

Our data demonstrate that plasminogen activator inhibitor-1 (PAI-1) induces the carotid artery (CA), a conduit artery, to phenocopy the arterial wall structure and function of a resistance artery; this includes holes in the internal elastic lamina (IEL), EC projections (i.e., myoendothelial junctions [MEJs]), and switching to predominantly endothelial derived hyperpolarization (EDH) vasodilation. PAI-1-treated CAs further resembled resistance arteries with alpha globin expression. Using PAI-1-treatd CAs from a global alpha globin knockout mouse, arterial relaxation predominantly occurred through nitric oxide (NO) signaling, suggesting a role for alpha globin in facilitating the switch to predominately EDH-based relaxation.

Nonstandard Abbreviations and Acronyms:

PAI-1

plasminogen activator inhibitor-1

MEJ

myoendothelial junction

Ach

acetylcholine

IEL

internal elastic lamina

SEM

scanning electron microscope

TEM

transmission electron microscope

CA

carotid artery

TDA

thoracodorsal artery

EDH

endothelial derived hyperpolarization

PE

phenylephrine

NO

nitric oxide

NOS

nitric oxide synthase

L-NAME

L-nitro-arginine methyl ester

TRAM

1-[(2-chlorophenyl)-diphenylmethyl]-1H-pyrazole

EC

endothelial cell

SMC

smooth muscle cell

α-GA

α-glycyrrhetinic acid