The juxtaglomerular apparatus (JGA) is a specialized structure, formed by the afferent and efferent arteriole, thick limb of the ascending loop of Henle, and mesangial cells. The JGA embodies two major renal functions: tubuloglomerular feedback (TGF), which is part of the mechanism underlying renal autoregulation, and renin secretion. The JGA consists of different cell types with specialized functions: the macula densa (MD) cells, juxtaglomerular cells, mesangial cells, vascular smooth muscle cells (VSMCs), and endothelial cells (ECs). The MD cell of the thick limb of the ascending loop of Henle functions as a NaCl sensor. An increase in NaCl concentration in the tubular fluid eventually causes the VSMC in the afferent arteriole to contract to reduce glomerular filtration rate (GFR) and the juxtaglomerular cells of the afferent arteriole to reduce renin secretion. Mesangial cells are located between the MD cells (the sensor) and the afferent arteriole (the effector). They are believed to play an important role in the signal transduction from the MD to the afferent arteriole (Goligorsky et al., 1997).

The TGF signal originates in the MD cells and elicits local vasoconstriction in the juxtaglomerular part of the afferent arteriole. The response, however, encompasses the entire afferent arteriole, the distal part of the interlobular artery, and travels into the afferent arterioles of neighboring nephrons (Peti-Peterdi, 2006; Marsh et al., 2009). This indicates that the vasoconstrictor signal is conducted within the vascular wall most likely through gap junctions (GJs) coupling ECs to ECs, VSMCs to VSMCs, and ECs to VSMCs.

GJs are fluid-filled pores that allow the movement of small molecules and current between neighboring cells. GJs are made up of two hemichannels, connexons, built from connexins (Cxs) (Nielsen et al., 2012). Six Cxs of the same or different isoforms can assemble to form a connexon, and the two connexons from neighboring cells can dock to form a functional GJ channel (Moreno, 2004). Several Cx isoforms are expressed in the JGA (Hanner et al., 2010). Similarly, in preglomerular arterioles, expression of Cx37, Cx40, Cx43, and Cx45 (Wagner, 2008; Just et al., 2009; Schweda et al., 2009) is found, enabling the cells of the vessel wall, both smooth muscle and endothelial, to act as a syncytium (Goligorsky et al., 1997; Dora et al., 2003; Haefliger et al., 2004; Peti-Peterdi, 2006; Yao et al., 2009).

In the vasculature, Cx40 is the most widely expressed isoform and is predominantly found in the ECs (de Wit et al., 2003; Schweda et al., 2009), as well as in renin-producing cells (Haefliger et al., 2001). In mice, deletion of Cx40 leads to the development of renin-dependent hypertension (de Wit et al., 2000, 2003; Figueroa et al., 2003; Wagner et al., 2007; Just et al., 2009; Sorensen et al., 2012) and reduced conduction of vasodilation (de Wit et al., 2000). Cx45 is expressed in the VSMCs (Wagner, 2008) and mesangial cells (Kurtz et al., 2009; Hanner et al., 2010), and lack of Cx45 expression causes lethal malformations in the vasculature (Krüger et al., 2000; Hanner et al., 2008; Wagner, 2008). Therefore, only a conditional deletion of Cx45 is possible in mice. Previous studies in these mice show that Cx45 plays a role in the propagation of VSMC calcium waves and in TGF (Hanner et al., 2008; Møller et al., 2020). This calcium signaling may be one mechanism by which Cx45 affects vascular conducted responses.

Once the signal from MD has reached the afferent arteriole, the vascular conduction can spread dilation or constriction along the vessel wall, depending on the blood flow needs (Wagner, 2008). Conducted vasoconstriction occurs when depolarizing current propagates upstream via GJs (Gustafsson and Holstein-Rathlou, 1999), from the afferent arterioles to the larger arteries, and changes the vascular resistance, thus reducing renal blood flow (Wagner et al., 1997).

In this study, we aim to test if a lack of two specific Cxs affects vascular conduction in the afferent arteriole. Specifically, we explore the roles of Cx40 and Cx45 in the conduction of vasoconstrictor responses to electrical pulse stimulation.

The present study investigated the importance of intercellular communication via GJs, through Cx40 and Cx45, in the conduction of afferent arteriolar vasoconstriction in response to local electrical stimulation. We have previously shown that lack of Cx40 or Cx45 significantly reduces TGF (Sorensen et al., 2012; Møller et al., 2020). Here, we examine if this reduction is caused by a decrease in the vascular conduction of vasoconstriction. The local and conducted vasoconstrictor responses were measured during stimulation and compared to the diameter measured before stimulation. Decay of constriction was computed for the stimulation period for data pooled from each mouse strain.

VSMCs and ECs work together to coordinate and propagate constriction and dilation along the vessel wall (Li et al., 2015). The conducted vasomotor response is important in the regulation of blood flow and resistance in vascular networks (Gustafsson et al., 2001) and depends on GJs (Segal et al., 1989; Hakim et al., 2008; Braunstein et al., 2009). In afferent arterioles, ECs are primarily coupled through Cx40 (Arensbak et al., 2001) and VSMCs through Cx45 (Møller et al., 2020). GJs also play a role in the signaling between ECs and VSMCs (Emerson and Segal, 2000) through myoendothelial GJs (MEGJs; Haefliger et al., 2004). It is unknown which Cxs are part of MEGJs in the renal vasculature (Wagner, 2008) but Cx45 or Cx40 are likely candidates. In Cx40 KO mice, renal endothelium derived hyperpolarization was significantly reduced compared to WT, suggesting that Cx40 is part of the renal MEGJ (Brasen et al., 2018). Also, Cx40 has been found in junctions between EC and renin secreting cells (transformed VSMC; Haefliger et al., 2001) supporting that Cx40 could be part of the MEGJ.

Both mouse strains used in the present study have been characterized previously. The Cx40 KO was hypertensive (de Wit et al., 2000; Wagner, 2008; Just et al., 2009), but RBF was similar in Cx40 WT and KO (Just et al., 2009). Segmental arterial and afferent arteriolar diameters were not significantly different (Sorensen et al., 2012), but heart weight and afferent wall trans-sectional area were increased in Cx40 KO mice (Sorensen et al., 2012; Jacobsen and Sorensen, 2015), possibly due to the hypertension. The Cx45 KO was reported to be slightly hypertensive (Hanner et al., 2008) or normotensive (Schmidt et al., 2012). Segmental arterial and afferent arteriolar diameters were not significantly different between Cx45 WT and KO, but wall trans-sectional area was decreased in afferent arterioles from Cx45 KO (Møller et al., 2020). In both strains, TGF was found to be reduced in the KO (Sorensen et al., 2012; Møller et al., 2020), whereas the myogenic response was unaffected by the reduction in intercellular communication (Jacobsen and Sorensen, 2015; Møller et al., 2020).

Electrical pulse stimulation induced a local and a conducted vasoconstriction response in afferent arterioles from Cx40 and Cx45 WT and KO mice. The local response measured at the stimulation site was significantly smaller in Cx40 KO compared to Cx40 WT. However, in cremaster arterioles from Cx40 KO mice, local vasoconstriction elicited with electrical stimulation or KCl was similar to that observed in WT mice (Figueroa et al., 2003; Wölfle et al., 2007). As shown before, the afferent arteriolar diameter measured in the juxtaglomerular segment in Cx40 KO was smaller than in WT and smaller than the diameter measured 100 μm upstream (Sorensen et al., 2012). This could suggest an increased baseline afferent arteriolar constriction at the site closest to the glomerulus when Cx40 is not present. Eliminating TGF did not normalize afferent arteriolar diameter, suggesting that TGF is not causing the constriction. Importantly, afferent arteriolar diameter measured upstream from the juxtaglomerular site was not significantly smaller in Cx40 KO (Sorensen et al., 2012). No significant difference in the local vasoconstriction was found between Cx45 WT and KO even though it appeared smaller in Cx45 KO. However, this could be due to the relatively small effects being measured resulting in low statistical power. In mice lacking Cx45 in VSMC, KCl-induced local vasoconstriction was similar to the constriction elicited in WT mice (Schmidt et al., 2012). In isolated segmental arteries from both strains, no difference was found between WT and KO when assessing the vasoconstrictor response to NE (Brasen et al., 2018; Møller et al., 2020). Collectively, these data suggest that the renal vascular reactivity in Cx40 and Cx45 KO is not reduced. Possibly, the decreased juxtaglomerular afferent arteriolar diameter in Cx40 KO affects the local vasoconstriction elicited at this site.

Although the decay rate of vasoconstriction seemed slightly increased in Cx40 KO, no significant difference was found between Cx40 WT and KO. This is in accordance with results obtained in skeletal muscle arterioles, where KCl-induced conducted vasoconstriction was unaffected in Cx40 KO mice (de Wit et al., 2000). It is generally believed that electrical signals travel predominantly in the endothelial layer (Kapela et al., 2018), which is coupled mainly by Cx40. In isolated preglomerular arteries from Cx40 KO, electrical stimulation did not elicit an increase in VSMC intracellular Ca²⁺ measured 500 μm from the stimulation site (Sorensen et al., 2012). However, no measurements were made at shorter distances. In addition, an electrical signal may travel further than what can be measured when using increases in intracellular Ca²⁺ or vasoconstriction as parameters. Comparing the constriction at a given distance against 0 revealed that in Cx40 KO the constriction was only significant at 50 μm. This could be due to the smaller local constriction.

In contract, in Cx45 KO, the decay rate was significantly increased compared to WT. Possibly, Cx45 is part of the MEGJ in the afferent arteriolar wall, and its removal by conditional knockout could reduce communication between the VSMC layer and the endothelium, hence reducing the mechanical response from the VSMC layer. Myoendothelial junctions are unevenly distributed along the wall. The spread of depolarization from VSMCs that are coupled to the endothelium, to VSMCs that are not, could be central for a distant mechanical response to electrical stimulation. In addition, although the endothelium is normally considered the main path for a longitudinal spread of current in the microcirculation, the presence of VSMC-VSMC coupling by Cx45-containing GJs may also contribute. A possible proposed mechanism is through Ca²⁺ signaling as Cx45 is known to be involved in the propagation of Ca²⁺ waves in VSMC isolated from afferent arterioles (Hanner et al., 2008). Comparing the constriction against 0 showed that in Cx45 KO the constriction was only significant up to 250 μm compared to 300 μm in the WT.

The structure of the vascular wall may affect both the local vasoconstriction and conduction properties. The vascular wall in renal vessels with a diameter <40 μm is thicker compared to vessels from skeletal muscles (Hosoyamada et al., 2015), which would reduce conduction in afferent arterioles. In both KO strains, changes in wall trans-sectional area (WTA) and stress sensitivity in the afferent arteriole have been found (Jacobsen and Sorensen, 2015; Møller et al., 2020). However, WTA and stress sensitivity seemed to be increased in Cx40 KO, whereas it was decreased in Cx45 KO. Whether this can account for the increase in decay rate observed in Cx45 KO remains to be established. Another possibility is that deletion of Cx40 or Cx45 changes the expression of other GJs or ion channels responsible for eliciting the vasoconstriction. Deletion of Cx40 reduced expression of Cx37 in the vasculature (Simon and McWhorter, 2003; de Wit, 2010), which could affect the vascular response. Whether this could also be the case for vascular ion channels is currently not known.

Differences in myogenic activity at the chosen RPP of 95 mmHg may also influence the obtained results. Resting membrane potential of VSMCs in the afferent arteriole in isolated perfused rat kidneys was approximately −40 mV at 80 mmHg RPP (Loutzenhiser et al., 1997). We assume that a comparable resting membrane potential is present in mouse afferent arterioles at an RPP of 95 mmHg. Under these conditions, the afferent arteriole has substantial myogenic tone and addition of nifedipine, an inhibitor of L-type Ca²⁺ channels, significantly increased afferent arteriolar diameter (Sorensen et al., 2012). In Cx40 KO mice, afferent arteriolar diameter remained constant even though RPP was increased from 75 to 95 mmHg (Sorensen et al., 2012) and in Cx45 KO mice the same RPP increase induced a significant reduction in afferent arteriolar diameter (Møller et al., 2020), suggesting that both strains have myogenic tone in the present experimental setup. As no differences have been found in the myogenic response between KO and WT from both strains (Jacobsen and Sorensen, 2015; Møller et al., 2020) it does not seem likely that differences in myogenic tone are affecting the electrically induced vascular responses observed in the present experiments.

A number of limitations are present in our experiments. Due to the experimental setup, it is not possible to position the stimulation pipette next to the vascular wall. A varying layer of tubular tissue will separate the pipette tip and the vascular wall. This limitation has been mitigated by only including arterioles with a minimum local constriction of 5%. In addition, the preparation is not fixed and the arterioles have some vasomotion that could affect the measurements. Consequently, we used an average of the diameter measured over 30 s. Vasoconstriction is not the most sensitive measure to quantify vascular conduction. Changes in membrane potential can be measured using sharp electrodes or patch-clamp (Li et al., 2004). This method is not applicable in our preparation, where the cells of the vascular wall are not freely accessible. Another factor that should be taken into consideration is the age span of the mice (2–14 months). Laboratory mice reach sexual maturity at 8–12 weeks of age (Dutta and Sengupta, 2016); past the age of 18 months, these mice are considered old (Hagan, 2017). We know that age does not affect the myogenic response (Lott et al., 2004), unless subjects are sedentary (Hong and Lee, 2017). In WKY and SHR rats, both cortical and juxtamedullary autoregulatory activity were similar between ages 10 and 70 weeks (Iversen et al., 1998) even though MAP increased significantly in SHR. These findings indicate that vascular function is intact at 14 months of age, and we are not aware of any study that has directly addressed the effect of aging on vascular conduction in the mouse afferent arteriole. Lastly, we have used mixed sexes. Renal autoregulation in the normal physiological state is identical in males and females (Hilliard et al., 2011; Brown et al., 2012) as is renal vascular responses to angiotensin II and PE (Schneider et al., 2010). Thus, the use of mixed sexes would likely not affect the results.

In conclusion, we have shown that knockout of Cx40 reduces local vasoconstriction but does not significantly change the conduction of vasoconstriction. Vascular deletion of Cx45 significantly reduces the vascular conduction induced by electrical stimulation in afferent arterioles. Consequently, vasoconstriction in response to a comparable stimulus is conveyed shorter when intercellular communication between cells in the vascular wall is reduced in the renal microcirculation. This may in turn affect the ability to regulate and autoregulate renal perfusion.

All datasets generated and source code for this study are included in the article/Supplementary Material.

The animal study was reviewed and approved by the Danish National Animal Experiments Inspectorate, Ministry of Environment and Food of Denmark, and Danish Veterinary and Food Administration.

SM and CS designed the animal experiments. SM performed animal experiments and drafted manuscript. SM, CS, and JJ analyzed data and prepared figures and tables. SM, JJ, N-HH-R, and CS interpreted results, edited and revised manuscript and approved final version of manuscript.