The regulation of blood pressure during pregnancy involves several biological pathways. Candidate genes implicated in hypertensive diseases during pregnancy include those of the renin-angiotensin system and nitric oxide synthase (NOS). We evaluated blood pressure and metabolic characteristics during pregnancy in mutant mice. These included mice with a null mutation in the endothelial NOS (eNOS) gene (Nos3−/− ), four copies of the angiotensinogen gene (Agt2/2 ), and mutations in both genes [four copies of Agt and heterozygous deficient for eNOS (Agt2/2Nos3+/− ), four copies of Agt and homozygous deficient for eNOS (Agt2/2Nos3−/− )]. Blood pressure measurements of nulliparous females from mutant strains were compared with two common laboratory strains C57Bl6/J and SV129 throughout their first pregnancy. Serum and urine analysis for the evaluation of renal and liver physiology were measured in the prepregnant state and during the third trimester of pregnancy. Throughout pregnancy blood pressures in all mutant strains were higher compared with controls.Agt2/2Nos3−/− showed the highest blood pressures and C57Bl6/J the lowest. Control mice, but not mutant mice, showed a second trimester decline in blood pressure. No immediate differences were noted regarding behavioral characteristics, renal or liver function parameters. Mice deficient for eNOS, mice with overexpression of Agt, and mice with mutations in both genes demonstrated higher blood pressure throughout pregnancy. There was no evidence of renal dysfunction, liver dysfunction, or hemolysis among any of the strains studied. We conclude that Nos3 andAgt are important genes in the regulation of blood pressure during pregnancy.
- nitric oxide synthase gene Nos3
nitric oxide (no) is a free radical derived from the conversion ofl-arginine to l-citrulline by the enzyme NO synthase (NOS) (17). Several isoforms of this enzyme exist. The endothelium-derived isoform (eNOS) leads to the generation of NO that functions to relax surrounding vascular smooth muscle (28). Targeted disruption of the eNOS gene (Nos3) has been successfully accomplished in mice (7,10, 22). The phenotype includes slower heart rates (22), increased renin plasma levels (22), developmental abnormalities of the limb (7), smaller adult size (22), and aberrant fertility (26). Nongravid mice deficient of NO derived from eNOS have been shown to have higher blood pressure compared with control mice (10,22).
The renin-angiotensin system through vasoconstriction and aldosterone secretion has an important role in blood pressure control, electrolyte, and fluid homeostasis (18, 19). Two mouse models with overexpression of the angiotensinogen gene (Agt) have been described. One is a transgenic Agt overexpressing mouse model in which expression of the Agt gene is driven by unknown promoter sequences (25). Another model was produced by creating a tandem duplication of Agt by gap repair targeting. In this case, Agt expression was driven by the wild-type promoter (23) and mice with two, three, and four copies of Agt were generated. A linear relationship between Agt copy number and blood pressure andAgt copy number and plasma angiotensinogen levels was established (12). Interestingly, although each model accomplished successful overexpression of Agt, blood pressure measurements of nonpregnant mice conflicted (12,25). In the case of the transgenic model, blood pressure was normal in the nonpregnant state, and for the gap repair targeting model, it was increased.
In humans, association studies have been used to help understand the contribution of single genes toward the development of pregnancy-induced hypertension. Regarding Agt (2, 11,29) and Nos3 (1, 2, 16), various polymorphisms have been found and shown to be associated with hypertension in the nonpregnant and pregnant states.
Recently, mouse models with multiple mutations have been developed to explore complex genetic traits (9, 13) and may provide novel approaches to understanding pregnancy-specific diseases such as those complicated by hypertension. Previously, female mice overexpressing the human Agt were found to have new onset increased blood pressure and proteinuria during pregnancy only after mating with a male overexpressing the human renin gene (25). Those observations led us to evaluate blood pressure characteristics in mice with mutations in two genes, each of which functions in a different biological pathway. Through cross-breeding, our laboratory successfully established a mouse model with simultaneous mutations in the genes Agt and Nos3. Our aim was to report blood pressure and metabolic characteristics during pregnancy in Nos3-deficient mice, in mice with a tandem duplication of the Agt gene, and in mice with mutations in both genes.
MATERIAL AND METHODS
Mice deficient for eNOS (Nos3−/− ) were previously generated in our laboratory (7). Breeding pairs of mice with a tandem duplication of Agt on chromosome 8 and therefore driven by the wild-type promoter were provided (12,23). Through a breeding scheme (Fig.1 A) we succeeded in generating mice with four copies of Agt(Agt2/2 ). A second scheme (Fig. 1 B) was used to generate mice with four copies of Agt that were also either heterozygous deficient (Agt2/2Nos3+/− ) or homozygous deficient (Agt2/2Nos3−/− ) for eNOS. In addition to these mice two common laboratory strains (C57Bl6/J and SV129) were studied with respect to blood pressure and metabolic characteristics during pregnancy.
Animal husbandry practices followed guidelines established by the Animal Care Committee of Baylor College of Medicine. Mice were subjected to daily 12:12-h alternating periods of light-dark cycles within a humidity- and temperature-controlled environment. Food and water were provided ad libitum.
Genotyping for Nos3 was done as described previously (7). Briefly, exon 1 of the Nos3 gene was removed through targeted mutagenesis using a neomycin-resistance gene. A unique EcoR I restriction site was created that allowed genotyping by Southern blotting.
Because of the cross-breeding between F1Nos3−/− mice (created from C57Bl6/J and SV129 strain mice) and the Agt mutant mice, we found it necessary to create a strategy for establishing the Agt copy number after crossing with the eNOS-deficient mouse line. The Agtgenotyping approach originally described (23) took advantage of the fact that the tandem duplication for Agtwas engineered within chromosome 8 of the SV129 strain. For detecting the tandem duplication of Agt, the strain-specific D8Mit56 marker was used (23). Therefore, cross-breeding mice with mutations at the Agt locus with our eNOS-deficient mice made this genotyping strategy unhelpful.
We designed PCR strategies to amplify genomic DNA probes capable of detecting the deleted exon 2 (725-bp probe) or the tandem duplication of Agt (443-bp probe). Primer sequences for each probe were designed after consideration of the construct used to generate null and tandem duplication mice in addition to the published genomic sequence (3, 23). For the 725-bp probe, the sequence of the forward 22-mer primer sequence was 5′-GTA TAC ATC CAC CCC TTC CA-3′. This primer started at position 985 within exon 2. The reverse 22-mer primer sequence was 5′-GGA AGT GAA CGT AGG TGT TGA-3′ and started at position 1732 within exon 2. PCR conditions were as follows: denaturing 94°C for 5 min, 30 cycles of 94°C for 45 s, 58°C for 45 s, 72°C for 45 s, and extension of 72°C for 5 min. PCR products were resolved on a 1.5% low-melting agarose gel. The 725-bp fragment was recovered from the gel and purified.
For the 443-bp probe, the sequence of the forward 20-mer primer sequence was 5′-GCA GGA GAG GAG GAA CAG CC-3′. This primer started at position 2428 within exon 5. The reverse 22-mer primer sequence was 5′-GGG GGT GCT TTT TGG TGT CA-3′ and started at position 2870 (3′ of the coding sequence). PCR conditions were as follows: denaturing 94°C for 5 min, 30 cycles of 94°C for 45 s, 57°C for 45 s, 72°C for 45 s, and extension of 72°C for 5 min. PCR products were resolved on a 1.5% low-melting agarose gel. The 443-bp fragment was recovered from the gel and purified.
Mouse genomic DNA for genotyping was obtained from proteinase K digested tails using phenol-chloroform extraction. Genomic DNA was digested using the restriction enzyme Sac I. Southern blots of genomic DNA were hybridized at high stringency with either the 725- or the 443-bp probes, which had been labeled using random primers ([α-32P]dCTP). The Sac I restriction fragment lengths of mice with zero, one, two, three, and four copies of the Agt gene is shown in Table1. The genotyping strategy shown in Table1 does not allow for a distinction between mice with three and four copies of Agt. For this reason, care was taken to use onlyAgt2/0 mice from litters in which mice with three or four copies of Agt were possible († in Fig.1 A). Figure 2 illustrates Southern blots probed with the 725- and the 443-bp probe, respectively. A 7.5-kb restriction fragment was indicative of the chromosome with no copy of Agt (Fig. 2 A) and a 3-kb fragment established a chromosome with two copies (Fig. 2 B). On the basis of these genotyping strategies and the breeding schemes shown mice with zero, one, two, three, and four copies of Agt and either homozygous deficient, heterozygous, or wild-type for eNOS could be distinguished accurately.
Blood Pressure Measurements
Blood pressure characteristics were evaluated forNos3−/− (n = 6),Agt2/2 (n = 6),Agt2/2Nos3+/− (n = 5), andAgt2/2Nos3−/− (n = 5) mice throughout pregnancy. In addition to these mice, two common laboratory strains [C57Bl6/J (n = 6) and SV129 (n = 3)] were studied.
Nulligravid 8- to 10-wk-old mice were first conditioned in the tail-cuff blood pressure apparatus (NarcoBiosystems, Austin, TX) for 5 days. Conditioning was performed daily for 15 min between 8 AM and noon. After this period, prepregnancy blood pressure measurements were obtained on 3 consecutive days. Subsequently, timed mating with a C57Bl6/J fertile male was performed between 9 and 11 AM daily. At the conclusion of the cohabitation period, the presence of a plug was confirmed visually and by gentle probing of the vaginal orifice using a blunt tapered glass rod. The identification of a plug defined day 0 of gestation. Conditioning to the blood pressure apparatus was continued until pregnancy was established.
For taking blood pressure, mice were placed onto a warming plate with the temperature preset at 37°C. A 17-mm tail cuff was applied to the base of the tail, and a pneumatic pulse transducer was then applied distal to the tail cuff. Output was via a two-channel physiograph that was calibrated before each run. The tail cuff was programmed to insufflate to a maximal pressure of 250 mmHg. A rest period of 15 s was allowed until the next insufflation. Blood pressure was recorded daily between 8 AM and 12 noon beginning on gestational day 1. We recorded blood pressures until five tracings without movement artifacts were obtained. Five successive readings were averaged to establish the blood pressure reading of any given day. The average time of blood pressure measurement was 15 min. We continued to measure blood pressure for an additional 5 days after pregnancies were completed.
Evaluation of Metabolic Characteristics
Timed mating was performed as previously described. Two separate cohorts of mice were studied. Mice were either nulligravid (8–10 wk, 5 mice of each strain) or third trimester primigravid (gestational day 15–17, 5 mice of each strain). Mice were placed into a metabolic cage (Tecniplast, Birchrunville, PA) that housed one mouse at a time for 24 h each. Urine production, food intake, and water consumption were recorded. Weight (g) on entry and removal from the cage was noted, change in weight during a 24-h exposure period (%) was calculated. Urine was frozen at −20°C, and specimens were analyzed in batch. After the 24-h period, mice were anesthetized (by inhalation with metophane) and a cardiac puncture was performed. Blood obtained was centrifuged (5,000 rpm for 5 min), and the serum was stored at −20°C. Specimens were analyzed in a batch. A chemistry analyzer (Cobas Mira; Hoffman-LaRoche, Nutley, NJ) was used to measure serum blood urea nitrogen (BUN), serum creatinine, and serum total protein to assess renal function, aspartate aminotransferase and alanine aminotransferase to assess liver function, and lactate dehydrogenase as a marker of hematological dysfunction. Total protein, total creatinine, and sodium concentration were determined on 24-h urine collections. Creatinine clearance was calculated by the formula: urine[creatinine] × urine[volume]/serum[creatinine]. Water and sodium balances are given as ratios of 24-h water (sodium) intake/24-h water (sodium) urine excretion.
We gave careful consideration to analysis of blood pressure data over time and concluded that data points for each mouse were not independent values (i.e., blood pressure on day 5 of pregnancy was likely dependent on blood pressure on day 4). To account for the lack of independence, the area under the curve (AUC) was used for blood pressure comparisons between and among groups. During pregnancy, the AUC for each group was calculated after smoothing by adjacent averaging of the nearest five data points (Microcal Origin 4.1; Microcal, Northampton, MA). Two and four nearest points were used to perform smoothing for prepregnancy, and postpartum blood pressure calculations, respectively. For statistical analysis, pregnancy was also broken down into trimesters every 6 days starting with day 1 of pregnancy. Although animals were time mated, not all animals delivered after the same duration of pregnancy (range 17–19 days). To calculate the AUC, each mouse contributed 6 days for each trimester. In cases where a missing value occurred (n = 12), a blood pressure value was substituted for the missing information by averaging all values obtained in the respective trimester for the same mouse. In cases in which a mouse delivered on gestational day 19 (n = 6) the value was not included in the analysis. Statistical analysis (SigmaStat version 2.0; Jandel Scientific, San Rafael, CA) employed one-way ANOVA with Tukey's test of multiple comparisons when significant differences were identified (P < 0.05). For evaluation of metabolic characteristics, one-way ANOVA with Tukey's test of multiple comparisons were used across each time point studied andt-tests with Bonferroni correction were used between each time point studied.
Blood Pressure Before, During, and After Pregnancy
We graphed blood pressure values 3 days before pregnancy, for each day of pregnancy, and for 5 days after pregnancy for all six groups of mice studied (Fig. 3). We observed several important features within this graph. First, mutant mice showed higher blood pressure values in the prepregnancy state. Second, the common laboratory strains C57Bl6/J and SV129 mice demonstrated lower blood pressure across all time points during pregnancy than any of the mutant mice. Third, each of the common laboratory strains demonstrated a second trimester decline in blood pressure that was not observed among any of the mutant mouse strains. Fourth, blood pressure values remained high for all mutant strains in the postpartum period. As the known biology would predict, the highest blood pressure at the beginning of pregnancy and throughout pregnancy was identified in theAgt2/2Nos3−/− cohort.
AUC for blood pressure in the prepregnancy period was compared for each strain (Fig. 4 A). TheAgt2/2Nos3−/− mutant mice had the greatest values (378.9 ± 22) and the C57Bl6/J strain had the lowest value (294.5 ± 39.9). AUC for blood pressure during the entire pregnancy was compared for each strain (Fig.4 B). TheAgt2/2Nos3−/− mutant mice had the greatest values (3,051.5 ± 80.6) and the C57Bl6/J strain had the lowest value (1,897.8 ± 280.1). Significant differences were found between the following strains:Agt2/2Nos3−/− and C57Bl6/J (P = 0.002), Agt2/2 and C57Bl6/J (P = 0.003), and Nos3−/− and C57Bl6/J (P = 0.01). AUC for blood pressure was also compared in the postpartum period for each strain (Fig. 4 C). The Agt2/2Nos3+/− mutant mice had the greatest values (3,051.5 ± 80.6) and the C57Bl6/J strain had the lowest value (724.4 ± 40.8). Significant differences were found between the following strains:Agt2/2Nos3+/− and C57Bl6/J (P = 0.005), Agt2/2 and C57Bl6/J (P = 0.004), andAgt2/2Nos3−/− and C57Bl6/J (P = 0.02), and Nos3−/− and C57Bl6/J (P = 0.01).
For each trimester of pregnancy (6 days) we determined the AUC for each strain of mice studied (Fig. 5). In the first trimester Agt2/2 andAgt2/2Nos3−/− mice showed statistically significant higher blood pressure compared with C57Bl6/J mice (P = 0.003 and P < 0.001, respectively). In the second trimester all mutant mice demonstrated higher blood pressure than the C57Bl6/J strain (Agt2/2 , P < 0.001;Agt2/2Nos3+/− , P = 0.01; and Agt2/2Nos3−/− ,P < 0.001, Nos3−/− ,P = 0.02). During the third trimester,Nos3 −/− mice had significantly higher blood pressure than that observed for C57Bl6/J (P = 0.02).
Behavioral and Metabolic Characteristics
Behavioral and metabolic characteristics as well as renal parameters of all mouse strains studied in the prepregnant state and in the third trimester are given in Table 2. Change in weight during 24 h and food and water consumption were considered behavioral characteristics. Independent of genotype, weight loss occurred in each group. There was no apparent relationship between genotype and any of the behavioral characteristics evaluated. Sodium and water balances were not significantly different between or among any group studied. The BUN values (Table 2) appear to decrease when prepregnant state and third trimester of pregnancy are compared.Nos3−/− mice demonstrate the highest BUN at baseline and during the third trimester of pregnancy. There was no clear difference in bilirubin, aspartate aminotransferase, alanine aminotransferase, or lactate dehydrogenase for any of the mouse genotypes or strains studied (Table 2).
We evaluated mice for evidence of proteinuria by measuring urine protein concentration (g/dl) and calculating total protein excreted in 24 h. When measured and reported as a concentration, urine protein was not significantly different between or among groups studied. In Fig. 6, the total protein excretion for each group is shown. After Bonferroni correction, there were no significant differences despite the clear increase in protein excretion in the third trimester of pregnancy for all mouse groups except the C57Bl6/J strain.
This is the first study to report on blood pressure and metabolic characteristics during pregnancy in mice with mutations inNos3 or in mice with simultaneous mutations inNos3 and Agt. We found that mice deficient for eNOS, mice overexpressing Agt, and mice with mutations in both genes showed higher blood pressures throughout pregnancy compared with common laboratory strains. No clear differences were seen with respect to the metabolic characteristics studied.
Hypertension is a common complication of human pregnancies. One disorder that is part of this spectrum, preeclampsia, is thought to impact 6–8% of pregnancies in the United States (20). Preeclampsia manifests as new onset or worsening (in a chronic hypertensive patient) hypertension in the third trimester and new onset proteinuria and may manifest as multiorgan dysfunction leading to renal or hepatic insufficiency as well as hemolysis. Patients with chronic hypertension are known to be at increased risk to develop preeclampsia. Blood pressure values have been shown to decrease during the second and early third trimester in both normotensive and chronic hypertensive women (5). During the third trimester, however, blood pressure in hypertensive women returns to its former hypertensive level. Absence of the second trimester physiological decline in blood pressure is considered a heralding sign of preeclampsia (5).
In the prepregnancy state, all mutant mouse groups studied showed elevated blood pressure values as expected and thus our model may be viewed as one of chronic hypertension during pregnancy. Double mutant mice showed the greatest elevation in prepregnancy blood pressures. As pregnancy progressed, those differences became more apparent. Blood pressure among eNOS-deficient mice was observed to increase from the first to the third trimester of pregnancy. With respect to the influence of Nos3 on blood pressure in pregnancy, our data indicate a role for endothelial-derived NO in blood pressure regulation and support previously published data describing elevated blood pressure in rats treated with NO inhibitors (6, 30). Blood pressure was highest with a slight downward trend during the third trimester for mice with four copies of Agt independent ofNos3 genotype. A second trimester decline in blood pressure, as the normal physiology of human pregnancy would predict, was only observed in common laboratory strains C57Bl6/J and SV129 mice, while being absent in all mutant strains. Of note, blood pressure of the common strains did not approach that of any mutant strain during pregnancy. In the postpartum period blood pressure in all mutant strains remained high, whereas in the common laboratory strains a slight increase of blood pressure toward prepregnant values was observed.
Our evaluation of mice with higher blood pressure antedating pregnancy failed to demonstrate any abnormalities in renal function, hepatic function, or evidence of hemolysis. We used a metabolic cage to collect specimens and to determine food and water consumption over 24 h. In performing studies of this nature, others have stated that food and water should be provided ad libitum (15). Although we cannot completely exclude the possibility of small particles of food contaminating urine specimens, the values we obtained are comparable to those reported (14). Although the creatinine clearance values we report are low, a wide range in values appears in the literature (21, 24). We believe these low values can be accounted for by low normal urine production over 24 h. The reason for the reduced urine production may be due to behavioral changes in that we did not condition animals to the metabolic cage before initiating sample collection. Reassuring to us was the observation of an increased creatinine clearance in the third trimester of pregnancy among all mouse groups except C57Bl6/J. Furthermore, strain variation in these and all parameters measured cannot be excluded when interpreting data of this nature. Urine protein excretion was increased in the third trimester of pregnancy compared with baseline for all groups except C57Bl6/J. Of note, mice overexpressing Agtindependent of Nos3 genotype showed the greatest increase (significant or near significant differences) before Bonferroni correction. These findings are in agreement with a previous report that mice overexpressing Agt develop proteinuria during pregnancy (25). However, we would add caution to this interpretation, because mice have proteinuria in the absence of pregnancy (14). As expected, an increase in renal perfusion, as occurs during pregnancy, would result in an increase in total protein excretion and thus no change in protein concentration as we observed. Overall, despite the observed elevated blood pressure values during pregnancy and the absence of a decline in blood pressure in the second trimester of pregnancy, we have to conclude that our mutant mice are not sufficient models for preeclampsia.
In this study only the maternal genotype was evaluated without regard to paternal genotype. Each female was mated to a C57Bl6/J male. Therefore, all offspring were heterozygous for wild-typeNos3 and Agt (from the father) and the engineered aberration of Nos3 and Agt (from the mother). This study made no attempt to evaluate the reproductive outcomes of fetal pups carrying these mutations.
In summary, we established a mouse model with simultaneous mutations in the Nos3 and Agt genes. Mice deficient for eNOS and mice overexpressing Agt, as well as mice with double mutations in both genes, showed elevated blood pressures at the beginning and throughout pregnancy. Mice with four copies ofAgt and deficient for eNOS showed the highest blood pressure. There was no evidence of renal dysfunction, liver dysfunction, or hemolysis among any of the strains studied.
These data contribute further evidence that physiological blood pressure regulation in pregnancy is complex and can be regulated by multiple genetic pathways simultaneously. Given the complex nature of the syndrome labeled preeclampsia, the evaluation of the mutant mouse models used in this study begins to unravel the relative contributions of specific genes to this phenotype. The current literature would suggest a polygenic and perhaps environmental component in the pathogenesis of this disorder. We would speculate that mouse models with mutations in multiple different biological pathways will give way to the most cogent models of this disease. Data are currently available that support a role for cell-adhesion molecules (27, 31), cytokines (4, 8), and blood pressure-regulating genes (2, 29); however, a critical combination has not yet been described. There seems no doubt that mouse models will be instrumental in guiding our probe into the critical combinations of genes needed to evolve the complex phenotype of preeclampsia.
This work was supported in part by the Erwin-Schroedinger-Auslandsstipendium J1,839-MED with funding by the Fonds zur Foerderung der Wissenschaftlichen Forschung to L. A. Hefler.
Address for reprint requests and other correspondence: A. R. Gregg, Dept. of Obstetrics and Gynecology, Baylor College of Medicine, 6550 Fannin St., Smith Tower Suite 901-A, Houston, TX 77030.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2001 the American Physiological Society