The natriuretic peptide clearance receptor locally modulates the physiological effects of the natriuretic peptide system

Generation of Mutants.

We used homologous recombination in embryonic stem cells to generate animals in which the Npr3 gene was disrupted and partially deleted (Fig. 1 a and b). The resulting heterozygotes (+/−) survived and reproduced normally. The homozygous (−/−) mutants survived to birth, but about half died before weaning; about one-third of the survivors reproduced at least once. Ligand-binding analysis on lung membrane extracts demonstrated the complete absence of functional NPRC in Npr3 −/− animals but the continuing presence of NPRA (Fig. 1c). The amounts of the two guanylyl cyclase receptors (NPRA and NPRB) assessed by semiquantitative binding studies with isolated lung and kidney membrane preparations from Npr3 +/+, +/−, and −/− mice were indistinguishable in the three genotypes (data not shown).

Clearance of ANP.

The effect of absence of NPRC on clearance of natriuretic peptides from the circulation was assessed by measuring the disappearance of injected [¹²⁵I]ANP. The results (Fig. 2a and Table 1) show that ANP half-life in Npr3 −/− mice is two-thirds longer than in wild-type mice (P < 0.0001), demonstrating that NPRC plays a significant role in its clearance from the circulation. This conclusion agrees with earlier studies in rats in which clearance of [¹²⁵I]ANP was reduced when NPRC was blockaded with a truncated ANP (12). When ANP clearance decreases, its steady state plasma level should increase unless ANP production changes. However, in our Npr3 +/− and −/− mice, the steady state levels of ANP and BNP were not higher than in the wild type (Table 1); in fact, they tended to be lower than normal, although the difference did not reach significance. Two important inferences can be drawn from these measurements: first, that some factor(s) in the +/− and −/− mice have probably induced a homeostatic decrease in the cardiac secretion of the two peptides; and second, that any changes in the Npr3 +/− and −/− animals that we observe are not caused by increased levels of ANP or BNP in the circulation.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Clearance of ANP and water handling. (a) Decline in trichloroacetic acid precipitable radioactivity after injection of [¹²⁵I] rat ANP (1–28) into wild-type (n = 5, filled circles) and homozygous mutant mice (n = 4, open circles). The plotted points (log cpm) are means ± SE of logarithms of trichloroacetic acid-precipitable counts normalized as described in Materials and Methods. Where not shown, error bars are smaller than the symbols. ∗, P versus wild type < 0.0001. (b) Ability to dilute and concentrate urine assessed by loading and depriving water; wild-type (n = 7, filled circles), heterozygous (n = 6, open triangles), and homozygous mutants (n = 4, open circles). Urinary osmolalities (Uosm) at time 0, and 1, and 2 hours after 4% body weight water loading by gavage, and after 12 hours of water deprivation, are shown as means ± SE in mOsm/kgH₂O. ^†, P versus wild type < 0.05; ^‡, P versus wild type < 0.02.

Circulatory and Renal Effects.

Blood pressures in the homozygous mutants were significantly lower by 8 mmHg (1 mmHg = 133 Pa) than in the wild type; blood pressures in the heterozygotes were not significantly lower (Table 1). No significant effects of genotype were observed on plasma Na⁺, K⁺, Ca²⁺, Cl⁻, creatinine, and urea nitrogen concentrations or on daily urinary excretion of Na⁺, K⁺, Cl⁻, creatinine, or protein (data not shown). However, a substantial and progressive increase in the +/− and −/− mice relative to the wild type was seen in their total daily urine output, and their water intakes tended to increase (Table 1), suggesting alterations in renal function. Visual inspection of the Npr3 −/− mice shortly after birth suggested that they were dehydrated. Hematocrits, hemoglobin levels, and red blood cell counts of adults (Table 1) also increased progressively and highly significantly in the +/− and −/− mice relative to the wild type, indicating that a reduced amount of NPRC in the heterozygotes or its absence in the homozygotes decreases intravascular volume. As indicated above, this would be expected to induce a decrease in the cardiac secretion of ANP and BNP in the Npr3 +/− and −/− mice.

Because both NPRA and NPRC are strongly expressed in renal glomeruli (8), if NPRC is acting in the kidney as a clearance receptor, reduced amounts or absence of glomerular NPRC should expose the glomerular NPRA to progressively greater local concentrations of natriuretic peptides, thereby enhancing its guanylyl cyclase activity and increasing cGMP in the glomerular filtrate and in urine. To test this expectation, we determined the total daily excretion of cGMP in the urine of Npr3 +/+, +/−, and −/− mice. We found that a markedly increased cGMP excretion does indeed occur in the Npr3 +/− and −/− mice (Table 1), being respectively >200 and >300% normal. Because the levels of ANP and BNP in the circulation of the +/− and −/− mice were not greater than in wild-type mice, this result clearly demonstrates that decreasing NPRC substantially increases the renal effects of systemically delivered and/or locally synthesized natriuretic peptides. This in turn would be expected to alter the ability of the mice to concentrate their urine. We tested this expectation by comparing the urine osmolalities of the +/+, +/−, and −/− mice (Fig. 2b) and found that all genotypes can excrete dilute urine after water loading but that the +/− and −/− mice were progressively and significantly less able than wild-type mice to concentrate urine.

Interpretation of Renal and Hypotensive Effects.

In the kidney, ANP and BNP are derived from the circulation whereas urodilatin is derived from cells of distal tubules. NPRA is expressed strongly in the renal vasculature and glomerular podocytes and in medullary structures of the kidney. NPRC is expressed in endothelial and vascular smooth muscle cells throughout the circulation and is strongly expressed in glomerular podocytes, to a lesser degree in glomerular mesangial cells, and minimally in medullary interstitial cells (2, 21–23). In interpreting the effects of a decreased level or absence of NPRC, it is important to recollect that the levels of the natriuretic peptides in the systemic circulation and of the guanylyl cyclase receptors in the kidney are essentially normal in the Npr3 +/− and −/− mice. We therefore infer that the observed renal effects of decreasing NPRC are not caused by changes in the systemic levels of the natriuretic peptides. Rather, they are caused by local increases in the concentration of the peptides and possibly also by the effects of increased cGMP in the glomerular filtrate (24). We presume, but at present cannot prove, that the concentration of natriuretic peptides in the circulation in and downstream of the glomeruli will be less decreased than normally (i.e., increased relative to normal). As a result, the known effects of ANP on the structures exposed to glomerular and post-glomerular blood flow are likely to occur: namely, an increase in filtered volume and a decrease in water reabsorption (2, 25). In contrast, the effects of urodilatin produced in the distal tubules are unlikely to be affected because NPRC is not normally present in significant amounts in the distal nephron.

Our finding that the Npr3 +/− and −/− mice maintain their ability to dilute urine indicates that their defective ability to concentrate urine is unlikely to be caused by impairment of salt transport in the thick ascending limb. Anatomical inspection of the Npr3 +/− and −/− mice also indicates that the deficiency in urine concentration is not caused by a shortening of the renal papilla in these genotypes. Accordingly, the simplest interpretation of our findings is that they are the effect of higher than normal post-glomerular concentrations of natriuretic peptides in the blood and of cGMP in the ultrafiltrate.

We interpret the hypotensive effect of the absence of NPRC as being caused by a decreased blood volume resulting from the observed diuresis or from a locally increased concentration of the natriuretic peptides in endothelial and vascular smooth muscle cells in the systemic vasculature. [The latter would likely lead to vasorelaxation and/or to a shift of fluid from the intravascular to the interstitial compartment, with a resulting decrease in plasma volume (2).]

We conclude that NPRC in the kidney and other tissues normally decreases in a “dose-dependent” manner the local concentration of systemically delivered or locally synthesized natriuretic peptides, thereby modulating and helping to compartmentalize their physiological effects. We suggest that human genetic variations (if they exist) that modestly decrease but do not abolish NPRC formation could have beneficial cardiovascular effects under some circumstances whereas genetic variations that increase NPRC formation could be detrimental. For example, hypertensive tendencies might be ameliorated by decreases in intravascular volume that accompany decreases in NPRC or might be exaggerated by increases in NPRC expression. Likewise, a genetic decrease (or increase) in NPRC might be protective (or detrimental) in situations that lead to cardiac hypertrophy because absence of NPRA in mice causes exaggerated and pathological cardiac hypertrophy (26). A search in humans for genetic variations affecting the expression of the gene coding for NPRC therefore appears to be worthwhile.

Skeletal Effects.

Unexpectedly, the −/− mice completely lacking NPRC exhibit striking skeletal abnormalities, recognizable 1 week after birth, including hunched backs, dome-shaped skulls, and elongated tails that, in some mice, were initially wavy (Fig. 3a). They had elongated femurs, tibias, metatarsal, and digital bones, longer vertebral bodies, increased body length, and decreased weight (Fig. 3 b–f). The thoracic cages of the −/− mice were smaller than the wild type and were constricted. Development of secondary ossification centers in long bones was delayed in the −/− mice (Fig. 4 a and b). In the cartilage growth plates of 10-day-old −/− mice, cellular expansion was apparent in the zone of hypertrophic chondrocytes although not in the zones with resting or proliferating chondrocytes (Fig. 4 c and d). This difference was no longer apparent in 3-month-old mice. The amino acid compositions of the organic bone matrix collagen in the wild-type and −/− mice were virtually indistinguishable (data not shown). The bony trabeculae in young −/− mice were thicker and longer than in the wild type (Fig. 4 e and f). In situ hybridization with an exon 1 probe showed strong expression of NPRC mRNA in the osteoblastic cells lining the bony trabeculae of developing wild-type bone (Fig. 4g). No signal was seen in the same cells of the −/− mice (Fig. 4h). The number and size of osteoclasts in cultures of bone marrow cells from 2-month-old −/− mice were nearly twice that of the wild type (Fig. 4i). The number of osteoblastic precursors from the −/− mice was ≈3× that of the wild type, and they proliferated almost twice as rapidly (Fig. 4j). These findings indicate a more active bone metabolism in the mutants.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Skeletal phenotypes of Npr3 +/+ and −/− mice. (a) Wild-type (left) and homozygous mutant mice (right) at 2 months of age. (b–e) Soft x-ray analysis of phalanges [wild type (b) and homozygous mutant (c)] and lower bodies [wild type (d) and homozygous mutant (e)] of mice at 3 months of age. (f) Body and bone dimensions of homozygous mutants at 3 months of age as percent difference from the wild type. ∗, P versus wild type < 0.001.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

(a and b) Histochemical localization of osteoclasts by tartrate-resistant acid phosphatase staining (dark brown) in developing femurs of 10-day-old Npr3 +/+ (a) and −/− mice (b); original magnification 40×. The arrow indicates a secondary ossification center. (c and d) Cartilagenous growth plates of femurs from 10-day-old +/+ (c) and −/− (d) mice; hematoxylin and eosin stain; original magnification 200×. (e and f) Picro-Sirus Red staining of bony trabeculae in femurs from 10-day-old Npr3 +/+ (e) and −/− mice (f); original magnification 400×. The amounts of unresorbed mineralized cartilage (MC) and newly deposited bone matrix (arrows) are both increased in the mutant. (g and h) In situ hybridization for mouse NPRC mRNA in trabecular bones of 10-day-old wild-type (g) and homozygous mutant mice (h); original magnification 400×. The signal is specifically localized to the osteoblastic cells lining the bony trabeculae (BT). (i) In vitro osteoclast formation assessed by osteoclast number and by the average number of nuclei per osteoclast. Open bars, wild type; filled bars, homozygous mutant. (j) In vitro osteoblast formation assessed by number of ALP-positive colonies and the average cell number per colony in bone marrow stromal cell cultures from wild-type (open bars) and homozygous mutant mice (filled bars). ∗, P versus +/+ < 0.05; ^†, P versus +/+ < 0.01; ^‡, P versus +/+ < 0.001.

Bone Metabolism.

We therefore compared the levels of several relevant blood and urine components in 2-month-old +/+, +/−, and −/− mice. The data (Table 2) strongly indicate that the homozygous −/− mice, but not the +/− mice, have substantially increased bone turnover. Their plasma ALP, an indicator of bone formation, is 150% of the wild-type level, and urinary pyridonoline and deoxypyridonoline, indicators of bone resorption, are all significantly increased (>200% that of the wild type).

Interpretation of Bone Effects.

Various components of the natriuretic peptide system are known to be present in cells or tissues involved in bone formation. For example, CNP and NPRB (the guanylyl cyclase receptor essentially specific for CNP) and the corresponding mRNAs have been demonstrated in cultured fetal mouse tibia (27), and mRNA for NPRC is readily detected in osteoblasts (ref. 28 and our present data). A simple hypothesis consistent with previous and our present observations is that NPRC in growing bone modulates the autocrine/paracrine effects of locally produced natriuretic peptides (mainly CNP). In favor of this hypothesis are observations that CNP, much more than ANP, increases cGMP production in the fetal mouse tibia organ cultures and causes an increase in longitudinal bone length, that this bone growth stimulation is mimicked by 8-bromo-cGMP, and that it is inhibited by HS-142-1, a nonpeptide GC-coupled natriuretic peptide receptor antagonist (27). The growth stimulation in the organ cultures was accompanied by an increase in the height of the proliferative and hypertrophic chondrocyte zones in the growth plate that also was seen in our Npr3 −/− mice (Fig. 4d). Additionally 1,25 dihydroxyvitamin D3, a key regulator of mineral metabolism, stabilizes NPRC mRNA (28), indicating a role for NPRC in this regulation.

Other Mutant Animals.

Spontaneous and N-ethyl-N-nitrosourea-induced mutations in mice causing skeletal abnormalities identical to those described here have recently been identified in the Npr3 gene (J. Jaubert, personal communication). Similar abnormalities also have been seen in transgenic mice grossly over-expressing BNP (29), which could be due partly to cross-stimulation of NPRB by the high levels (200 times normal) of circulating BNP and partly to chronic blockade of NPRC by this BNP.

Conclusions.

We do not expect the effects of changes in NPRC levels to be restricted to those that we have described here in the kidney and bone. Effects can be expected in any tissues or organs in which at least one of the natriuretic peptides, an active receptor, and the clearance receptor are all present in significant amounts. The interplay of these elements is complex and is likely affected by the precise anatomical location of the cells synthesizing each component of the system, which may differ in different organs and will need to be more carefully delineated. In several different situations, expression of NPRC is controlled differently from the expression of the two biologically active natriuretic peptide receptors and of the natriuretic peptides themselves. For example, chronic high salt intake leads to a decrease in expression of NPRC in kidney glomeruli and papilla without affecting the expression of the biologically active receptors (30). In the heart, progressive hypertrophy induced by an aortovenocaval fistula in the rat is accompanied by a striking gradual disappearance of NPRC transcripts whereas expression of ANP, BNP, and NPRA increases (31). In bone, 1,25 dihydroxyvitamin D3 affects the expression of CNP and NPRC but not that of NPRB (28). We conclude from these and our present results that NPRC provides a biologically effective variable allowing the physiological effects of the natriuretic peptide system to be tailored to local needs because the several elements of the system (ligands, active receptors, and clearance receptor) can be changed independently. The potential flexibility of the natriuretic peptide system and its possible compartmentalization are thereby substantially enhanced.