The aspartyl-protease renin is the key regulator of the renin-angiotensin-aldosterone system, which is critically involved in salt, volume, and blood pressure homeostasis of the body. Renin is mainly produced and released into circulation by the so-called juxtaglomerular epithelioid cells, located in the walls of renal afferent arterioles at the entrance of the glomerular capillary network. It has been known for a long time that renin synthesis and secretion are stimulated by the sympathetic nerves and the prostaglandins and are inhibited in negative feedback loops by angiotensin II, high blood pressure, salt, and volume overload. In contrast, the events controlling the function of renin-secreting cells at the organ and cellular level are markedly less clear and remain mysterious in certain aspects. The unravelling of these mysteries has led to new and interesting insights into the process of renin release.

Development and Structure of Renin Secretory Vesicles

Renin is produced as pre-prorenin protein, which is transferred into the cisterns of the endoplasmic reticulum. The signal peptide is cleaved off during transfer, and prorenin is directed to the cis-Golgi cisterns. By default, cells export newly synthesized material from the Golgi apparatus to small clear vesicles and to immediate secretion by the constitutive pathway, unless the material is tagged by retention or processing signals to be directed elsewhere (FIGURE 1). The sorting of proteins to dense core secretory granules for regulated exocytosis involves a number of different mechanisms, including interaction with membrane-associated alpha helices, such as prohormone convertases, and creation of a local environment that favors aggregation of granule material (7). Correct sorting of prorenin to the regulated secretory pathway depends on the presence in the prorenin molecule of a paired basic amino acid site, which serves as a protease processing site for granule-attached prohormone convertases (4). In addition, lowering the pH in the early renin granules may help in the aggregation of the granule core. Consistent with this, protogranules often contain paracrystalline cores (89). In mice, which contain two renin genes (Ren-1d and Ren-2), deletion of the Ren-1d gene is associated with complete absence of dense core secretory granules in JG cells (14). The granulation of JG cells was rescued by transgenic over-expression of Ren-1d but not the Ren-2 gene (55). Based on their results, the authors concluded that the difference in the ability to form dense core secretory vesicles depended on structural differences of the renin protein. Renin originating from the Ren-1d gene has three N-linked glycosylation sites that Ren-2-derived renin does not possess, and it was suggested that renin glycosylation was necessary for renin trafficking to dense core secretory vesicles (55). Apart from potentially playing a role in granule formation, the glycosylation state of renin may also be important for mannose-6-phosphate receptor-mediated uptake of renin or prorenin in other tissues, such as the heart (62, 93).


After translation of renin, the preprosegment enters the endoplasmic reticulum
The pre-part is the signal peptide, and it is cleaved off when the ER cisterns are entered. Then it moves to the Golgi apparatus,where glycosylatation and tagging for regulated pathway takes place. The trans-Golgi release clear vesicles containing prorenin, which are secreted consitutitvely. Prorenin tagged for the regulated pathway is contained in protogranules that coalesce to form mature renin granules, in which the prosegment is cleaved off, and renin is activated. The mature granules are then stored and released by regulated exocytosis.

Protogranules that are destined for the regulated secretory pathway pinch off from the trans-Golgi network. They contain prorenin and proteases, such as prohormone convertases (1, 5) and cathepsin B (53, 58), that have the ability to cleave off the prosegment. Other proteases, such as kallikrein and plasmin, are also able to cleave the prosegment, but there is no evidence that they play a physiological role in the process. In dense core secretory granules, an acidic intracellular pH is created by vacuolar H+-ATPases. The low pH is optimal for the protease activity that activates prorenin to renin by splitting off the 43-amino acid NH2-terminal propeptide. Activation of the proteases from their own pro-forms by autoactivation is also dependent on a low pH of 4–6. The low pH of the renin secretory vesicles has been used experimentally to stain them with substances like the fluorescent dye quinacrin (1, 9, 63). When using quinacrine for dynamic visualization of live cells, it is possible to observe disappearance of fluorescence from individual secretory granules, representing exocytosis, a sudden increase in granule pH, or both (63).

The early secretory granules have the ability to take up extracellular markers, such as horseradish peroxidise, which is targeted to early granules without transit through the Golgi apparatus (88). Such retrograde transport linking early/recycling endosomes to the trans-Golgi network is a well known cellular phenomenon (e.g., Ref. 2) involving SNARE complexes and clathrin (74). The direct uptake route into the secretory granules could potentially be used to target the activation of renin.

The sorting of renin to the regulated secretory pathway is not highly efficient in juxtaglomerular cells; only 25% of the synthesized renin is directed to the dense core secretory granules, whereas 75% is constitutively secreted as prorenin (66). When added to prorenin originating from extrarenal sources, this helps to explain that the proform accounts for 80–90% of total renin in the circulation. Interaction of circulating prorenin with tissue renin receptors may lead to proteolytic or nonproteolytic activation and to local generation of angiotensin, as well as activation of second messenger pathways (reviewed in Ref. 15).

When folded, the renin molecule has a bilobar form, and its active site with two catalytic aspartic residues is located at the bottom of the cleft between the lobes. The prosegment forms a plug at the entrance to the cleft and hinders access of renin substrate to the active site. Reversible non-enzymatic changes in the folding of the prorenin molecule may lead to exposure of the active site (15), but the quantitative significance of this mechanism remains to be established.

The ability to segregate renin into dense core secretory granules for regulated exocytosis seems to be specific to native juxtaglomerular cells in situ (FIGURE 2). After isolation, the cells rapidly lose the ability to direct renin into this pathway, and the various renin-producing cell lines (e.g., As4.1, Calu-6) do not store active renin in secretory granules. In Connexin (Cx) 40−/ − mice, JG cells are not coupled by gap junctions, and the cells relocate to the extraglomerular mesangial area (51). However, the JG cells retain their ability to process active renin through the regulated secretory pathway. Therefore, neither communication through gap junctions nor localization in the afferent arteriole is a prerequisite for maintaining the regulated secretory pathway.


Mouse afferent arteriole displaying juxtaglomerular cells with many dense core secretory granules
Bar = 2.5 μm (adapted from Ref. 33).

Release of prorenin via the constitutive pathway depends on the activity of the renin synthetic pathway, i.e., the level of gene activation and transcription efficiency in the individual cell and the total number of renin-producing cells. Therefore, acute stimulation of renin release leads to an increase in release of mature renin secretory granules that contain only active renin, whereas chronic stimulation of the renin system increases both circulating prorenin and renin levels (90).

The number of renin granules per afferent arteriole varies inversely with the salt status and with concomitant pharmacological interventions. Thus, in rats on a 0.2% NaCl diet, the average number of granules per arteriole is 445, and the volume of JG cells per arteriole is 2,015 μm3. After 2 wk of low-sodium diet and enalapril treatment, new juxtaglomerular cells are recruited, the total JG cell volume increases to 33,000 μm3, and the number of granules per arteriole increases by a factor of 20–9,000 (67). The average volume of JG cell granules and the number of granules per JG cell volume did not differ significantly in the two situations (67). Thus the gross morphology of renin secretory granules in newly recruited juxtaglomerular cells is similar to that of the granules in the control situation. These findings are consistent with a view that long-term regulation of the renin-angiotensin system takes place by modification of the number of renin-producing cells in the afferent arteriole and not by modification of the processes involved in the control of renin secretion in the individual cell.

Release Mechanisms of Renin Vesicle Content

Renin release from single rat afferent arterioles is discontinous with a renin content per discharge, which corresponds to the calculated renin content of one secretory granule (81). The spontaneous discharge rate was rather slow, with a release episode occurring every 5 min, on average. Although slow, such a release rate would still be more than sufficient to keep up with the physiological demands that were calculated by Taugner et al. (87) to be in the order of 20–500 released granules/min (in mouse). The relative rarity of exocytosis events explains why it has been difficult to produce solid ultrastructural evidence for exocytosis (87). Consistent with the view that exocytosis is the predominant release pathway for renin, it was observed that acute reduction of renal perfusion pressure led to a reduction of the number of renin granules in the afferent arteriole from 9,000 to 5,000 granules per arteriole in rats whose renin system had been stimulated for 2 wk with low sodium intake and treatment with the ACE inhibitor enalapril (67). The average size of the granules did not change. Stimulation of renin release is associated with formation of deep membrane invaginations that are in contact with the surface membrane. This is likely to be the result of addition of membrane material from many secretory granules to the cell surface and to already fused secretory granules (compound exocytosis) (67, 72). Further secretion of renin into such deep invaginations may explain how renin granules, as observed by two-photon confocal microscopy, could disappear from deep within the JG cells without having to move to the cell perimeter (63).

The canonical sequence of events that leads to secretion of dense core secretory vesicles is as follows: delivery of the granules to the membrane, reversible docking of the granules at the membrane, and the tethering of the granule and initiation of the fusion process. During exocytosis, a coil is formed between the three SNARE proteins syntaxin, synaptobrevin, and SNAP-25, bridging the vesicle and plasma membrane. Several molecular partners, including Sec1/munc18–1, are involved in the process (7, 65, 91). Since these processes occur in all living cells, the same is inferred to be the case for renin secretion from juxtaglomerular cells, but very little is yet known. In this context, it is interesting that the JG cells have been reported to express calcium-dependent activator protein for secretion (CAPS). This family of proteins is specifically involved in the secretion of large dense core secretory vesicles but not small clear vesicles. (e.g., Ref. 84).

Using the whole cell patch-clamp technique, it is possible to estimate cell membrane capacitance by measurement of capacitative currents induced by trains of small changes in membrane potential. Because biological membranes of secretory granules and the cell surface have similar specific capacitance (about 1 microFarad/cm2), addition of membrane area to the cell surface during exocytosis can be measured as an increase in membrane capacitance (57). This method has also been established for renin secretion from juxtaglomerular cells (21, 22, 24). Cell membrane capacitance as a measure of exocytosis is most valid in single cell measurements. In more intact preparations, the JG cells are coupled via gap junctions to each other, to endothelial cells, and to smooth muscle cells. Differences and time-dependent variability in coupling strength in such preparations will affect the ability to generate the capacitative currents that are used to estimate capacitance, making interpretation of the results complicated. Using the single cell patch-clamp approach, membrane capacitance (Cm) can be followed in JG cells for 10–20 min (FIGURE 3). With control solutions, the capacitance is constant or slightly increasing over time. The average capacitance of 106 mouse cells was 3.13 ± 0.13 pF (mean ± SE). With a specific membrane capacitance of about 1 μF/cm2, the average JG cell surface area was 313 μm2. Assuming perfectly spherical cells, this translates into a diameter of 10 μm and a volume of about 521 μm3 per cell. This size corresponds well to the observed size of isolated mouse JG cells (12). The capacitance of isolated rat JG cells is smaller (2.4–2.8 pF; Refs. 2123), corresponding to a surface area of about 260 μm2 and a volume of 394 μm3 per cell. A renin granule of a rat JG cell has a volume of 0.35 μm3 (67). If the granules are spheric, this corresponds to a surface area of 2.4 μm2. Mouse renin granules have a volume of 0.63 μm3 (33), corresponding to a surface area of 3.54 μm2.


Time course and relative changes of Cm
Top: time course of cell membrane capacitance (Cm) in single rat JG cell dialyzed with 1 μmol/l cAMP for 10 min. The increase in capacitance is a measure of exocytosis. Bottom: Relative changes of Cm in control cells and demonstration of enhanced Cm (exocytosis) in response to cAMP, to stimulation of beta adrenergic receptors with isoproterenol, to stimulation of protein kinase A (PKA) with SpcAMPs, and to inhibition of the cAMP response with the PKA-inhibitor RPcAMPs. Data are from Ref. 22.

An increase in cell membrane capacitance of 13.5–15.3% is seen after stimulation of renin release from rat JG cells by including 1 μM cAMP in the patch pipette (22, 23). Based on the figures above, this corresponds to secretion of 15–17 secretory granules, yielding a total released volume of 6.0 μm3. In mouse, inclusion of 1 μM cAMP in the patch pipette only leads to an increase of 7.1% in cell membrane capacitance (21). This corresponds to release of about 10.5 granules and a released volume of 6.6 μm3. Thus, although Cm may increase more in rat than in mouse in response to cAMP, the released volume of granule material per cell is about the same in the two species.

Increases in Cm that are higher than 15% in the rat are not observed in patch-clamp studies, suggesting that this is a maximal stimulation. Rats on a standard diet contain 445 granules in a volume of 2,015 μm3 of JG cells per afferent arteriole (67). Using the number of 394 μm3 volume per cell estimated from the patch-clamp data (22), this gives 5.1 JG cells per arteriole and therefore about 87 granules per cell. The released number of granules after maximal stimulation therefore constitutes about 20% of the total number and is likely to constitute the immediately releasable pool of renin granules. The existence of a releasable pool of renin granules is consistent with data from studies on isolated glomeruli (82) and isolated perfused kidney (50).

Intracellular Signals Controlling Renin Secretion

At present, renin secretion is considered to be controlled by an interplay of three intracellular second messengers, namely cAMP, cGMP, and free cytosolic [Ca2+]i. Among these, the cAMP signaling cascade appears to be the central and stimulatory pathway for the exocytosis of renin (FIGURE 4). This concept was recently supported by an elegant in vivo study in mice generated with a specific deletion of the stimulatory G protein Gsα, which is responsible for the receptor-mediated activation of adenylate cyclases, in renin-producing cells of the kidney (11). This conditional knockout of Gsα resulted in a marked reduction of baseline levels of plasma renin concentration and in a virtual absence of the stimulation of plasma renin concentration in a variety of conditions that usually lead to a potent activation of renin secretion (11). These observations thus complete a series of previous studies demonstrating that all maneuvers that increase cellular cAMP levels, such as activation of adenylate cyclase, inhibition of cAMP-phosphodiesterases, and addition of membrane-permeable cAMP analogs, stimulate renin secretion. Thus activators of β-adrenoreceptors (27, 40) and other hormones stimulating adenylate cyclase activity, such as prostaglandins E2 and I2 (24, 32), adrenomedullin (34), dopamine (43), and the neurohormones CGRP (44) and PACAP (30), are known to enhance renin release, as does direct activation of AC activity by forskolin. Similarly, nonselective inhibition of cAMP phosphodiesterases by IBMX (isobutyl-methyl-xanthine) is a well established way to stimulate renin secretion (13, 16, 48, 97). The family of cAMP phosphodiesterases comprises four prominent members, termed PDE-1–4, two of which, namely PDE-3 and PDE-4, appear to play a major role in renin-secreting cells (22, 48, 68). PDE-4 is allosterically activated by cAMP itself and likely produces a counterregulatory modulatory role for the cAMP-signaling pathway (59). PDE-3 activity is inhibited by cGMP and thus provides a cross-link between cAMP- and cGMP-signaling pathways such that the cGMP pathway can act via the cAMP pathway through the inhibition of cAMP degradation (59). Nitric oxide (NO) likely acts by this mechanism in renin-secreting cells (see below).


Sketch summarizing the intracellular signaling pathways relevant for the control of renin exocytosis
ANP, atrial natriuretic peptide; ATP, adenosine-trisphosphate; cAMP, cyclic adenosine-monophosphate; AC5/6, adenylate cyclases 5/6; cGMP, cyclic guanosine-monophosphate; cGK, protein kinase G; GC-A, guanylate cyclase A (particulate guanylate cyclase); GP, GTP-binding protein; GTP, guanosine-trisphosphate; IP3, insitol-1,4,5 trisphosphate; NO, nitric oxide; PDE3a, cAMP-phosphodiesterase 3a; PKA, protein kinase A; PLC, phospholipase C; sGC, soluble guanylate cyclase.

Finally, membrane-permeable cAMP analogs, which act independently on adenylate cyclase activity and cAMP-phosphodiesterases, also stimulate renin secretion (26). It is reasonable to assume that cAMP-induced phosphorylation of downstream targets is involved in the induction of renin exocytosis, because the stimulatory effect of cAMP on renin release is abolished after pharmacological blockade of protein kinase A (21) (FIGURE 3). The downstream targets of protein kinase A relevant for triggering exocytosis of renin, however, are still unknown.

As already mentioned, and possibly confusing at first glance, NO, which is commonly thought to act via the cGMP pathway, also stimulates renin secretion along the cAMP-signaling cascade (48). NO is produced at two distinct sites in the immediate vicinity of renin-secreting cells, namely in endothelial cells of the afferent arteriole via endothelial NO synthase (38) and in the macula densa cells of the distal tubule via neuronal NO synthase (38). NO generates cGMP through soluble guanylate cyclase activity in renin-secreting cells, which in turn inhibits cAMP phosphodiesterase 3, leading to an increase of cAMP and, in consequence, a stimulation of renin secretion (48, 49). Surprisingly, activation of particular guanylate cyclase activity by natriuretic peptides preferentially inhibits renin secretion (42) by a process involving cGMP-dependent protein kinase type II (25, 94). As for the stimulatory effect of protein kinase A, the relevant downstream targets for the inhibitory effect of protein kinase G II on renin secretion also remain to be clarified. It will be of particular interest to understand how cGMP generated by soluble guanylate cyclase activity is primarily directed to inhibit cAMP phophodi-esterases, whereas that generated by plasma membrane-bound guanylate cyclase activity primarily activates protein kinase G.

Calcium appears to exert a rather unusual effect on secretion in renin-producing cells in such a way that increases of the cytosolic calcium concentration attenuate the release of renin. This assumption results from the observations that calcium-mobilizing hormones such as angiotensinII, endothelins, or vasopressin inhibit renin secretion (41, 70, 92). In line with this, induction of calcium store-operated calcium entry by inhibitors of the endoplasmatic reticulum calcium ATP-ase (SERCA) also suppresses renin release (79). Finally, renin secretion is virtually inversely related to the extracellular concentration of calcium (80).

Since calcium usually facilitates exocytosis, its inhibitory effect on renin secretion has been coined as the “calcium paradox of renin release” (27). Despite numerous efforts, the mystery of this paradox remained unresolved until recently, when two groups in parallel provided evidence that the cytosolic calcium concentration likely influences renin secretion indirectly, in such a way that it modulates the cytosolic cAMP concentration. It was found that [Ca2+]i and cAMP levels in juxtaglomerular cells are inversely related. Thus a reduction of [Ca2+]i increased cAMP levels (60), whereas an increase in [Ca2+]i suppressed intracellular cAMP levels (26). Moreover, clamping of the intracellular cAMP levels by the membrane-permeable cAMP analog cAMP-AM prevented the inhibition of renin release in response to calcium mobilizing hormones such as angiotensin II, endothelin, or the SERCA inhibitor thapsigargin, indicating that the calcium-induced reduction of cAMP mediates or is at least required for the inhibition of renin release by [Ca2+]i (26). The expression of calcium-inhibited adenylate cyclases AC5 and AC6 (26, 60) in juxtaglomerular cells provides the molecular basis for the calcium-dependent reduction of cAMP levels. Accordingly, the inhibition of cAMP generation either by pharmacological blockade of AC activity (61) or by siRNA-mediated downregulation of the respective adenylate cyclases (26) abolished the inhibitory effects of [Ca2+]i on renin release and prevented the stimulation of renin release by the β-adrenoceptor agonist isoproterenol.

In summary, the cAMP-signaling pathway has been corroborated as the central, direct, and stimulating trigger factor for the exocytosis of renin. This pathway thus explains the physiologically relevant stimulatory regulation of renin secretion by renal sympathetic nerves and circulating catecholamines that act via β1-adrenoreceptors (27). The cAMP pathway, moreover, provides an essential permissive basis for all other (patho)physiological modulators of renin secretion in vivo and which in their majority exert a negative control of renin secretion. These factors comprise circulating or tissue levels of angiotensin II (27), the intrarenal arterial pressure (27), the salt load of the body (27), and also the intrarenal control of renin secretion by tubular feedback (75). Negative interference with or inhibition of these factors by angiotensin II antagonists, a fall in renal blood pressure, or salt deficiency by treatment with loop diuretics, respectively, causes a stimulation of renin secretion in a dis-inhibitory manner. These stimulations are markedly attenuated or even abolished in mice with a deletion of Gsα in renin-producing cells (11), indicating that a permissive stimulation of renin secretion via the cAMP-signaling pathway, probably maintained by β-adrenoreceptor activity (36), is required for the modulatory action of all physiologically relevant factors controlling renin secretion in vivo.

Electrical Characteristics of Renin-Secreting Juxtaglomerular Cells

Ion channels play an important role in the secretory properties of the majority of endo- or exocrine glands. The knowledge available today suggests that the electrical properties of renin-producing juxtaglomerular cells are very similar to those of the neighboring vascular smooth muscle cells (6, 45, 52). Moreover, renin-producing cells are electrically well coupled with the smooth muscle cells (Refs. 6, 45, 52; see also Role of Intercellular Communication for the Control of Renin Secretion), suggesting an electrical syncytium in the juxtaglomerular region. Direct electrophysiological and indirect functional findings suggest that renin-producing cells express at least three types of regulated potassium channels, which importantly influence the membrane potential of renin-producing cells. The first is the anomalous inward rectifier Kir (45, 52), which can be inhibited by hormones such as angiotensin II via a G-protein-dependent process (45). The second is the outwardly rectifying potassium channel (45), which has been identified as the zero- splice variant of BKca (23). These channels are activated by an increase in the cyclic AMP concentration (23). Third, renin-producing cells appear to have KATP channels, the activation of which causes hyperpolarization of the cells (18, 35, 71). In addition, renin-producing cells likely possess a high number of calcium-activated chloride channels (45) as well as L-type calcium channels (23).

Vasoconstricting hormones, such as angiotensin II, lead to depolarization of renin-producing cells (6, 45, 47), probably by inhibition of Kir and by calcium mobilization activating chloride channels (45). Activation of BKca by cAMP would be expected to lower the membrane potential of renin-producing cells. Such an effect of the cAMP-signaling pathway on the membrane potential, however, has not been consistently observed (6, 20). It may become more apparent, however, at more depolarized membrane potentials. A similar explanation holds for the function of L-type calcium channels in renin-producing cells. Calcium measurements and functional studies have not attributed a major role to these L-type calcium channels in renin-producing cells (46, 78), probably because these channels become activated only upon very strong depolarizations (23). Although the precise role of all of the aforementioned channels in the control of exocytosis in renin-secreting cells is not yet understood in detail, there emerges a common pattern in the fact that maneuvers that depolarize renin-secreting cells, such as vasoconstricting hormones (6, 45) or increases in the extracellular concentration of potassium (46), hinder secretion. Maneuvers that are expected to lower membrane potential, such as activation of BKca (23) or KATP (18, 35, 71) channels or that block chloride entry (10), are known to favor renin secretion. How changes of the membrane potential precisely influence renin secretion remains to be elicudated, but at least a major mediating role for L-type calcium channels appears less likely in this context.

Role of Intercellular Communication for the Control of Renin Secretion

It has been long known that renin-producing juxtaglomerular cells form prominent gap junctions not only between each other (5) but also with their adjacent cells, in particular with the endothelium of the afferent arteriole (54, 86) and with the extraglomerular mesangium (85), a phenomenon that also remains notably stable in vitro (73). Gap junctions are made up of connexins (83), and in the last few years evidence has indicated that renin-producing cells predominantly express Cx40 and, to a lesser extent, Cx37 as main connexins (3, 28, 31, 98). It was reported previously that replacement of Cx43 by Cx32 lowers renin expression in the kidney (29), but neither of these connexins is expressed by the juxtaglomerular cells. Therefore, the functional meaning of gap junctions in renin-producing cells in general, and the function of specific connexins in particular, remained obscure until recently. Examination of mice lacking Cx40 revealed a strong renin phenotype of these mice with regard to both the structure of the cells and the regulation of renin synthesis and secretion (39, 51, 95).

Normally, renin-producing cells replace and continue the layer of vascular smooth muscle cells at the terminal part of the afferent arteriole and thus form the entrance port into the glomerular capillary network. Chronic stimulation of the renin system increases the number of renin-producing cells in the walls of afferent arterioles, likely by a retrograde metaplastic transformation of vascular smooth muscle cells into renin-producing cells containing prominent secretory granules (8, 15). Ectopic expression of renin is a rare event and occurs, if at all, in the extraglomerular mesangium (37).

In kidneys lacking Cx40, renin-producing cells are absent from the medial layer of afferent arterioles. Instead, they are found in the extraglomerular mesangium, in the glomerular tuft, as well as in the periglomerular interstitium (51). The cells contain numerous granules and display an irregular fibroblast-like shape. The number of renin-expressing cells is increased, but there is a striking heterogeneity in the number of renin-expressing cells around the individual glomeruli, ranging from clusters of up to 30 cells to episodic few cells per glomerulus (51). Altogether, it appears as though Cx40 is essential for the correct “homing” of renin-producing cells in the kidney.

Circulating renin is strongly elevated in Cx40-deficient mice, much more strongly than renal renin mRNA abundance (39, 95), suggesting an increased stability of renin in the circulation or, more likely, an increased translational efficiency of renin gene transcripts. Such a change in the translational efficacy of renin mRNA is not unusual and was first noticed during treatment with angiotensin II antagonists (56). In this context, it may be coincident that the inhibitory effect of angiotensin II on renin gene expression and renin secretion is markedly attenuated in Cx40-deficient kidneys (95). In parallel, the characteristic negative feedback of blood pressure on renin gene expression and renin secretion is also markedly attenuated in these kidneys (39, 95) (FIGURE 5).


Relationship between renal artery pressure and renin secretion in isolated perfused kidneys from normal mice and mice lacking connexin 40
Figure is adapted from Ref. 95.

On the other hand, other physiological regulators of the renin system, such as catecholamines or salt balance, appear to function quite normally (39, 95), suggesting that renin synthesis and secretion do not generally escape from physiological control in the absence of Cx40. Considering the aberrant localization of renin-producing cells, it is easy to attribute the defective feedback inhibitions by angiotensin II and blood pressure mainly to the anomalous position of the renin-producing cells. However, nonselective inhibitors of gap junctions also abrogate the pressure control of renin secretion in normal kidneys (95), underlining the relevance of gap junctions per se in the control of renin secretion. It should be added in this context that the endothelial cells of the afferent arterioles also predominantly express Cx40 (3, 28, 31, 98), thus providing the possibility of Cx40 holochannels (gap junctions) between endothelial cells and renin-producing cells. One could speculate, therefore, that the endothelium might be critically involved in the signaling pathways by which blood pressure and angiotensin II inhibit renin synthesis and renin secretion. Since Cx40 is also expressed by the extraglomerular mesangium (3, 28, 31, 98), renin-producing juxtaglomerular cells could form Cx40 holochannels (gap junctions) with the lacis cells. In fact, in Cx40-deficient kidneys, no gap junctions can be encountered in the lacis cell field (51). In parallel, the inhibitory effect of the macula densa on renin secretion is clearly attenuated (51), corroborating the concept of a crucial role of gap junctions of the lacis cell field for the signaling of macula densa cells into the juxtaglomerular region (64, 69, 75).

Experiments with isolated perfused kidneys suggest that calcium is somehow enrolled in the control of renin secretion involving Cx40 (95). It is already well known that lowering of the extracellular concentration of calcium enhances renin secretion and at the same time abrogates the inhibitory effects of perfusion pressure and angiotensin II on renin secretion (76, 77). Obviously, lowering of the extracellular calcium concentration in normal kidneys strikingly mimics the behavior of Cx40-deficient kidneys, with regard to renin secretion (95). Conversely, lowering the extracellular calcium in Cx40-deficient kidneys has no apparent effect on renin secretion, quite in contrast to normal kidneys (95).

Although it is yet unknown how extracellular calcium and connexin functions merge in their effects on renin secretion, it should be noted in this context that, in the juxtaglomerular cell line As4.1, which has a connexin expression pattern similar to native juxtaglomerular cells (73), spreading of calcium waves from cell to cell is mediated by ATP release (96). ATP is known to be released by connexon hemichannels (17), and it can also pass through holochannels (gap junctions). One might speculate, therefore, that it is a main function of connexins, either of hemi- or holochannels, in the juxtaglomerular region to enable spreading of ATP and, in consequence, calcium waves, which in turn act to inhibit renin secretion (see Intracellular Signals Controlling Renin Secretion).

Altogether, Cx40, either as hemichannel or as a constituent of gap junctions, appears to be critically involved in the calcium-dependent control of renin secretion and probably also renin synthesis.


Although we have learned at lot about the function of renin-secreting cells during the last years, new important questions have emerged in this context. Two issues should be highlighted in this context. First, a protein kinase A-dependent process apparently is the core process for initiating exocytosis of renin. The molecular mechanisms of this critical step in the release of renin together with the identification of molecular components of the secretory machinery in renin-producing cells now await future elucidation. Second, we meanwhile also know that intercellular communication through connexons is very important not only for the final destination of renin-producing cells in the mature kidney but also for the physiological control of renin production and secretion by these cells. To identify the mode of this intercellular signaling as well as the chemical nature of the relevant signals remains a challenge for future experiments. In view of the relevance of the renin system both for normal function of the organism and also for the development and progression of diseases, we expect that solving these above-mentioned questions will also be of clinical relevance. Thus recent evidence already suggests that alterations in the expression of connexins relevant for renin secretion could be of relevance for the development of hypertension also in humans (19).


The authors’ work is supported by the Deutsche Forschungsgemeinschaft (SFB 699) (F. Schweda, C. Wagner, A. Kurtz) and by the Danish Medical Research Council, the Danish Heart Foundation, and the Novo Nordisk Foundation (U. Friis, O. Skott).


View Abstract