Fatal gastrointestinal obstruction and hypertension in mice lacking nitric oxide-sensitive guanylyl cyclase

Generation of KO Mice.

For the generation of KO mice, we used the Cre/loxP system (16). Exon 10 of the β₁ subunit was floxed, and after ES transfer, it was deleted by using the EIIa/Cre mouse [ref. 17; Fig. 1A and supporting information (SI)]. The null mutation was confirmed by PCR and Southern blot analysis (Fig. 1B and SI). By using Western blot analysis, neither the β₁ subunit nor, surprisingly, the α subunits could be detected on the protein level (Fig. 1C). In line with this finding, NO-induced cGMP synthesis was neither detected in lung nor in brain (Fig. 1D) nor in other GC-rich tissues/cells (aorta and platelets; data not shown). We conclude that deletion of the β₁ subunit leads to a complete loss of NO-GC.

Mortality of GC-KO Mice.

Whereas mice heterozygous for the β₁ subunit of NO-GC were phenotypically undistinguishable from WT, homozygous GC-KO mice died prematurely and revealed two phases of mortality (Fig. 2A): 60% of born KO mice died within the first 2 days (indicated by I in Fig. 2A) whereas most (90%) of the remaining mice died between day 18 and 31 (indicated by III). To our surprise, in the 2 weeks between these two phases, practically all mice survived (indicated by II). The first period of lethality is indicative of maladaptation to the neonatal situation (18). As weaning and thus a change in diet takes place around day 18, we suspected gastrointestinal complications to underlie the second phase of mortality (see below).

When performing Caesarean section between prenatal days 17.5 and 18.5, mice fetuses were found to be phenotypically indistinguishable from WT or heterozygote siblings. Transmission of the mutation at the Mendelian ratio was found in prenatal offspring (n = 94 pups from eight litters) but was <25% in mice 1 day after birth (see numbers in Fig. 2). We assume that KO mice are born but die soon after birth because of reasons yet to be identified. Thus, the numbers in Fig. 2A underestimate the initial death rate of the KO mice by ≈50%.

Gastrointestinal Phenotype of GC-KO Mice.

The overall appearance of a 3-week-old GC-KO mouse and its WT sibling is shown in Fig. 2B; homozygous GC-KO mice exhibit considerable growth retardation, shown by a 40% reduction in body weight. Because this growth retardation may be explained by malnutrition, we therefore focused on the gastrointestinal tract: the vast majority of the GC-KO mice had a grossly enlarged and dilated caecum which, in most cases, was displaced from its normal position (Fig. 3A). To individually different degrees, these mice showed additional morphological changes such as an enlarged esophagus (Fig. 3B), a dilated stomach, a bloated gall bladder, and an enlarged colon (data not shown). To assess the functional relevance, we measured whole-gut transit time as a marker of gastrointestinal motility. After application of the red dye carmine into the stomach, the time until excretion was monitored (19). The whole-gut transit time in GC-KO mice was increased approximately three times compared with WT and heterozygote siblings (median of >475 min versus 160 and 151 min, respectively; Fig. 3D). This reduced gut transport may be the basis for further attendant symptoms in KO mice like ileus, peritonitis, intestinal necrosis (data not shown), and, in some cases, the perforation of the intestinal wall (Fig. 3C). Apparently, GC is indispensable for the physiological coordination of contraction/relaxation that is required for the segmental and propulsive movements of the gut, and the dysfunctional peristalsis due to lack of GC may well be the cause of death.

Rescue of GC-KO Mice.

To circumvent the second phase of mortality starting around day 18, we fed mice with fiber-free diet after weaning. As shown in Fig. 2A, KO mice could be rescued and survival improved immensely. Accordingly, the whole-gut transit time increased although it still did not equal that of WT or heterozygous mice (KO mice on fiber-free diet: median of 285 min vs. 170 and 160 min, respectively; see Fig. 3D). Despite the special diet, mice lacking NO-GC still appear to be susceptible to gastrointestinal complications because they still died prematurely. Taken together, the rescue of GC-KO mice indicates that mortality is indeed caused by severe gastrointestinal obstruction in animals on normal diet.

Elevated Blood Pressure and Reduced Heart Rate in GC-KO Mice.

The physiological role of NO as EDRF has been established in a series of elegant experiments leading to the award of the Nobel prize (20). To estimate the impact of the lack of the most important NO receptor, we measured blood pressure using a noninvasive tail-cuff system. Fig. 4A shows that GC-KO mice have a pronouncedly elevated systolic blood pressure that surmounts that of WT siblings by 26 mmHg [130 ± 2 mmHg (KO) vs. 104 ± 1 mmHg (WT)].

Lack of NO-Induced Relaxation in Aortae from GC-KO Mice.

The impact of GC deficiency was further investigated in aortic smooth muscle. To date, it is still unclear whether NO induces relaxation of vascular smooth muscle solely via GC and/or via alternative mechanisms. To investigate this issue, we performed organ bath experiments (Fig. 4 B–D). The NO donor S-nitrosoglutathione (GSNO) led to relaxation of phenylephrine-precontracted WT aortic rings (EC₅₀ ≈0.04 μM; Fig. 4B). No relaxation was seen in aortic rings from GC-KO mice even at GSNO concentrations three orders of magnitude above those fully relaxing the WT aorta. Similarly, another NO donor, diethylamine-NO, failed to relax aortae from KO animals (Fig. 4D). In line with these data, carbachol did not induce a relaxing response in the KO animals (data not shown). Yet, in the GC-KO mice, the signaling cascade downstream of GC functioned normally because relaxation was induced by stimulation of cGMP-dependent protein kinase (PKG) using the membrane-permeable cGMP analogue 8-pCPT-cGMP (see Fig. 4A). In addition to GSNO and diethylamine-NO, we tested other NO donors [Proli-NO, Angeli salt, and SNP (all up to 100 μM)] that liberate either NO radical, NO⁻, or NO⁺ species, and they all failed to relax GC-deficient aortic rings (data not shown). Thus, in our aortic model, the relaxing effect of NO is solely conveyed by NO-sensitive GC.

Lack of Glycerol Trinitrate (GTN)-Induced Decrease in Systolic Blood Pressure from GC-KO Mice in Vivo.

The impact of GC deficiency was further investigated in vivo by testing the blood pressure-lowering effect of GTN (Fig. 4A). Anesthetized WT and GC-KO animals were injected with 5 mg/kg GTN, and the subsequent decrease in systolic blood pressure was recorded by using tail-cuff plethysmography. In WT animals, GTN induced a decrease in systolic blood pressure of ≈30 mmHg. In GC-KO animals, GTN did not decrease the systolic blood pressure.

Lack of NO-Induced Inhibition in Platelets from GC-KO Mice.

Similar to its role in smooth muscle, NO has been postulated to mediate platelet inhibition by cGMP-dependent as well as cGMP-independent mechanisms (9, 21–23). The Western blot in Fig. 5A shows that both α₁ and β₁ subunits are absent in GC-KO platelets and that the α₂ subunit is neither expressed in WT nor in KO platelets. Next, we studied the physiological impact of NO on platelet aggregation. Collagen (4 μg/ml) induced aggregation of both WT and GC-KO platelets (Fig. 5 C–F). Collagen-induced aggregation was inhibited by NO in WT platelets whereas NO had no inhibitory effect in KO platelets (Fig. 4D). Even with millimolar concentrations of Proli-NO (and also a variety of other NO donors including GSNO, diethylamine-NO, and SNP; data not shown), we were unable to induce inhibition of aggregation. In addition, aggregation of KO platelets by other agonists (thrombin or U46619, a thromboxane A2 mimetic) was also not inhibited by NO (data not shown).

The effect of NO-induced cGMP synthesis in platelets can be monitored by the activity of PKG, e.g., by the phosphorylation of substrate proteins such as vasodilator-associated phosphoprotein and phosphodiesterase type 5 (PDE5) (24, 25). The Western blot in Fig. 4B shows that, in contrast to WT, NO does not induce activation of PKG in KO platelets, as shown by the lack of vasodilator-associated phosphoprotein or PDE5 phosphorylation. However, similar to aortic tissue, the signaling cascade beyond NO-GC is still intact because stimulation of PKG using 8-pCPT-cGMP results in phosphorylation of vasodilator-associated phosphoprotein as well as in the inhibition of aggregation of both collagen-stimulated WT and KO platelets (Fig. 5 B, E, and F). Thus, we conclude that NO induces the inhibition of platelet aggregation solely via cGMP.

Generation of β₁-KO Mice.

Murine genomic DNA of the β₁ subunit was derived from BAC clones (Incyte Genomics, St. Louis, MO) containing genomic DNA of ES cells of 129/SvJ mice. The targeting vector (Fig. 1A) was constructed such that exon 10 of the β₁ subunit was flanked by a loxP-Tk/Neo-loxP cassette and a single loxP site using a triple loxP vector set (16). Chimaeras were bred with C57BL/6 mice, and the resulting heterozygotes (β₁^+/flox) were crossed with the general deleter mouse Ella-Cre (17), resulting in the deletion of the floxed exon 10 (β₁^+/−). To remove the Cre recombinase gene, the heterozygous KO mice were crossed with WT mice (C57BL/6). Finally, the generated heterozygotes (β₁^+/−) were intercrossed to yield homozygous mice (β₁^−/− = β₁-KO). Further information on the cloning strategy as well as on the genotyping can be found in the SI and can also be obtained from the corresponding author.

Genotyping of Mice.

Genotyping was performed by PCR analysis and Southern blotting of tail DNA (Fig. 1B). The WT (680 bp) and β₁-KO (830 bp) fragments were amplified in a multiplex PCR using the primer lox-β₁-U1 (AAG ATG CTG AAG GGA AGG ATG C) versus lox-β₁-L1 (CAG CCC AAA GAA ACA AGA AGA AAG) and del-β₁-L1 (GAT GTG GGA TTG TTT CTG AGG A). NheI digests of genomic DNA and a 536-bp β₁-genomic probe (corresponding to the DNA sequence from exon 12 to exon 13) were used for Southern blots. This probe identifies a 5.5-kb fragment from the endogenous allele and a 10.2-kb fragment from the β₁-KO allele. Studies were performed with ≈3-week-old mice of either sex with a hybrid background (129/SvJ × C57BL/6, F3–F9 generation). In each experiment, littermates of the same sex were used wherever feasible.

Analysis of Murine Tissues.

The preparation of tissues, Western blot analysis, determination of GC activity, and cGMP-RIA were performed as described (24).

Total Gut Transit Test.

Carmine (50 μl; 3 mg of carmine in 0.5% methylcellulose or, in the case of animals on fiber-free diet, in tap water) was orally administered to each mouse. Mice were then returned to individual cages and placed on a white sheet of paper. The time taken for excretion of the first colored feces was recorded. Gut transit time was measured in 5–10 animals of each genotype.

Measurement of Blood Pressure.

Systolic blood pressure was measured in conscious and anesthetized mice by tail-cuff plethysmography (BP-98A; Softron, Tokyo, Japan). Measurements of blood pressure in conscious mice were taken from a total of seven WT mice (five females and two males) and eight β₁-KO mice (six females and two males). For habituation, the mice were trained daily (3–7 days). After the training period, 7–10 measurements per mouse were recorded daily for 5 days. In a different set of experiments, mice (nine WT and six GC-KO of either sex) were anesthetized with ketamine/hydrazine. In these animals, the change in blood pressure subsequent to i.p. injection of 5 mg/kg GTN was recorded.

Isolation of Platelets from Mice and Aggregation Experiments.

WT and β₁-KO mice were anesthetized by ether inhalation. Blood (≈800 μl) was collected from the orbital sinus and was drawn into 200 μl of heparin solution (50 units/ml). After dilution (1 ml of PBS, Ca²⁺- and Mg²⁺-free) and centrifugation (15 min at 90 × g and 20°C), the upper phase [platelet-rich plasma (PRP)] was collected. The lower phase was washed with a volume of PBS similar to that of the removed plasma and then recentrifuged; the upper phase was mixed with the PRP. Optical aggregation of PRP (2–3 × 10⁷ platelets per 200 μl) was performed in a four-channel aggregometer (Chronolog, Havertown, PA) after the addition of collagen (4 μg/ml; Chronolog). Proli-NO (4 mM) was administered simultaneously with collagen, and other NO donors or 8-pCPT-cGMP (40 μM) were administered 3–20 min before addition of collagen.

Organ Bath Experiments.

Thoracic aortae were cut into rings and mounted on fixed segment support pins in two four-chamber myographs (Myograph 610; Danish Myo Technology, Aarhus, Denmark) containing 5 ml of Krebs–Henseleit solution (118 mM NaCl, 4.7 mM KCl, 2.5 mM CaCl₂, 1.2 mM KH₂PO₄, 1.2 mM MgSO₄, 25 mM NaHCO₃, 7.5 mM glucose, pH 7.4) and gassed with 95% O₂ and 5% CO₂. Diclofenac (3 μM) and N-nitro-l-arginine methyl ester (200 μM) were present. Resting tension was set to 2.5 mN. After equilibration (at least 60 min at 37°C), rings were precontracted with phenylephrine (1 μM; Sigma, Taufkirchen, Germany), and the vasorelaxing effects of Proli-NO (4 mM; Alexis, Lausen, Switzerland) and 8-pCPT-cGMP (40 μM; Biolog, Bremen, Germany) were assessed. Each experiment was performed in parallel with two to four aortic rings derived from KO and WT mice.

Statistics.

If not stated otherwise, data are expressed as mean ± SEM (n = number of animals). To compare gut transit time results from KO, heterozygote, and WT animals, all groups of the experiment were compared by Kruskal–Wallis test. If the resulting P was <0.05, two individual groups were compared by Mann–Whitney tests in a hierarchical order. Specifically, we first tested GC-KO animals versus WT animals, then heterozygote animals versus WT animals, and then heterozygote versus GC-KO animals. The comparison was rated significant if P < 0.05, and hierarchical testing of individual groups was stopped if P > 0.05 to correct for multiple testing.