1 Department of Physiology
2 Human and Molecular Genetics Center
3 Bioinformatics Research Center
4 Center for Biotechnology and Biomedical Engineering, Medical College of Wisconsin, Milwaukee, Wisconsin 53226-0509
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
genetics; salt-sensitive hypertension; renal damage; blood pressure
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Genetic linkage studies have traditionally combined results from both males and females for determination of quantitative trait loci (QTL), assuming that traits mapped to autosomes would be identical and independent of sex; whereas, sex differences would map to the X chromosome or Y chromosome or both. In the course of the initial characterization of the phenotypic differences between male and female BN/SsNHsdMcwi (Brown Norway; herein referred to as BN) and SS/JrHsdMcwi (Dahl S; herein referred to as SS) rats that were to be intercrossed to generate an F2 population, many phenotypic sex differences were observed in our laboratory. For this reason, and since females have not been phenotyped extensively in previous genetic studies of hypertension, we performed a genetic linkage analysis in female F2 rats derived from BN and SS parental rats.
Identification of the genes contributing to the development of hypertension and the associated end organ damage in human populations has been very difficult because of the polygenic nature of the disease (54), the small effect of individual genes (64), heterogeneity in onset of the disease (58), and the strong influence of the environment on the multiple pathways and genes controlling BP (20). Most efforts to identify genes of hypertension in human populations have focused on testing candidate genes, associated with hypertension in animal models (65), for linkage in population studies. Although several single gene mutations that influence the renal handling of sodium and BP have been successfully identified (21, 39, 40, 71), efforts to identify common mutations in genes that contribute to hypertension and end organ damage in the general population using whole genome scans have not been as successful (31, 52).
The failure to identify genes that contribute to variations in BP in human populations has led to the view that there is a need to identify potential "intermediate phenotypes" that might be better predictors of the onset and pathogenesis of hypertension than the level of pressure alone (18, 30, 35, 44, 45, 67). In principle, the intermediate phenotype mechanistically contributes to the BP and is therefore closer to the causal gene(s). Although the notion of an intermediate phenotype is easy to invoke and a large number of potential "intermediate phenotypes" have been proposed, to our knowledge none has met the strict criteria of mapping to the same location as BP and still correlating with BP in crosses, families, or affected sibpairs. In the present study, a genome-wide linkage analysis was performed using 236 blood pressure, neuroendocrine, metabolic, renal, and vascular phenotypes, measured in an F2 population of rats derived from a cross of hypertensive SS and normotensive BN rats. The SS rat was chosen as the model system for this study since it exhibits many of the features of salt-sensitive hypertension in humans (2). SS rats are salt sensitive (54), hyperlipidemic (47), and insulin resistant (63) and exhibit a low-renin form of hypertension (64). They also exhibit the rapid development of severe hypertension-induced renal, cardiac, and vascular end organ damage (23). The data were analyzed in two ways. First, we sought to map the likely determinants of BP and the end-organ damage associated with hypertension. Second, we sought to identify intermediate and secondary phenotypes related to the hypertension.
![]() |
METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Chronic phenotyping protocol.
Rats were housed in stainless steel metabolic cages, and food and water were available ad libitum. The light cycle was altered and set for 12 h lights on (2:00 AM to 2:00 PM) and 12 h lights off (2 PM-2 AM) for the duration of the study to allow for measurement of BP during both the active (lights off) and inactive (lights on) periods for rats during the course of the normal work day. Since the shift in the day/night cycle was only 4 h, 57 days during the recovery period was considered adequate for adaptation to the new light cycle prior to BP measurements. BP was directly measured while the rats were conscious, unrestrained, and housed in their home cages from a catheter chronically implanted in the femoral artery as described previously (13). Systolic (SAP), diastolic (DAP), and mean arterial blood pressures (MAP) and heart rate (HR) were recorded at 100 Hz on three consecutive days for 3-h periods from 9:00 AM to 12:00 PM (inactive phase) and for a 4-h period on the middle day from 2:00 PM to 6:00 PM (active phase). The data were averaged to 1-s intervals for time series analysis and also converted to 1-min averages and reduced to a single mean value for the entire recording session. The arterial pressure time series data was further analyzed using a recently developed mathematical model of arterial pressure control (34) to derive parameters related to the set points and gains of the baroreflex control system. A urine sample was collected for 24 h during the second day of the BP recording for measurement of urine volume and sodium, potassium, total protein, and creatinine excretions. After the third day of the BP measurement, a blood sample (500 µl) was collected from the femoral artery for measurement of plasma renin activity, hematocrit, and plasma protein and creatinine concentrations.
Sodium and volume depletion.
On day 3, after obtaining the blood sample, the rats received an intraperitoneal injection of furosemide (10 mg/kg) and were switched to a low-salt (0.4%) diet to sodium deplete and volume deplete (VD) the animals.
The next morning, BP was recorded during a 30-min control period and again after delivery of two mild alerting stimuli (2 mA for 0.3 s) via a wire electrode chronically implanted in the back of the neck. The control MAP, the change in MAP, the time to peak, and the time to 90% recovery were recorded before, during, and after delivering the alerting stimulus. The stimulus was repeated 5 min later and the same parameters were measured. Thirty-six hours after VD with furosemide, BP was again recorded for 3 h during the lights on inactive cycle, and a 0.7-ml sample of blood was collected for measurement of plasma renin activity, plasma triglyceride, cholesterol, and creatinine concentrations, hematocrit, and white blood cell count. A 24 h urine sample was collected for measurement of total protein, creatinine, and electrolyte excretions.
Acute phenotyping protocol.
After completion of the chronic phenotyping protocol, the rats were maintained in the sodium- and volume-deplete state by feeding a low-salt diet. The following week, they were anesthetized with ketamine (30 mg/kg ip) and thiobutabarbital (Inactin, 50 mg/kg ip). New catheters were acutely implanted in the femoral artery and vein, and an electromagnetic flow probe (Carolina Instruments) was placed on the left renal artery via a midline incision. The rats received an intravenous infusion (50 µl/min) of 0.9% NaCl solution containing 1% bovine serum albumin to replace surgical fluid losses. After a 45-min equilibration period, MAP, and renal blood flow (RBF) were measured during a 15-min control period. Renal and peripheral vascular responses to 5 min intravenous infusions of angiotensin II (ANG II: 20, 100, and 200 ng·kg-1·min-1) and norepinephrine (NE: 0.5, 1, 3 µg·kg-1·min-1) were determined sequentially. Following return of MAP to control, renal vascular and systemic arterial responses to two 5-min intravenous infusions of acetylcholine (ACh; 0.1 and 0.2 µg·kg-1·min-1) were measured. Then, the renal and vascular response to blockade of NO synthesis with an intravenous bolus of L-NAME (dose 5 mg/kg) was determined. Finally, the rats again received intravenous infusions of ACh (same two sequential doses) to test for the degree of attenuation of the renal and vascular responses following blockade of NO synthesis.
Histology.
After completion of the acute phenotyping protocol, the heart and kidneys of the rats were collected and weighed to assess the degree of cardiac and renal hypertrophy. The right kidney was fixed in 10% buffered formalin solution and embedded in paraffin; then 3-µm-thick sections were prepared and stained with periodic acid-Schiff stain (PAS) for measurement of glomerular diameter, and the degree of glomerulosclerosis was rated on a 04 scale using the scoring system as previously described (12, 53).
Genotyping.
A portion of the liver of each animal was collected, frozen in liquid N2, and stored at -80°C for isolation of DNA.
The liver was subsequently thawed and minced, then incubated overnight in a lysis buffer containing 100 µg/ml of proteinase K. DNA was precipitated with isopropanol and pelleted by centrifugation at 12,000 g for 10 min. The pellet was washed with 75% ethanol, air dried, then resuspended in TE buffer, pH 7.5. The concentration of DNA in the samples was determined by measurement of 260-to-280 nm absorption ratios (Beckman model DU640).
The rats were genotyped using radioactively labeled primers, as previously described (26). In brief, genomic DNA (25 ng) was amplified by PCR in a 20-µl reaction containing 100 nmol/l of each primer, 200 nmol/l dNTP, 1.5 mmol/l MgCl2, and 0.05 U of Taq DNA polymerase. Both primers were labeled with [32P]ATP (Dupont, New England Nuclear) using T4 polynucleotide kinase (Boehringer). The reactions were initially denatured at 92°C for 3 min, followed by 34 cycles of denaturation at 92°C for 1 min, annealing for 2 min at 55°C, and extension at 72°C for 3 min, followed by final extension at 72°C for 7 min. The samples were denatured by adding a loading buffer, heating the samples to 94°C for 5 min, and then snap-cooling the reactions on ice. Aliquots of the denatured samples were loaded onto a 7% polyacrylamide gel containing 37.5% formamide, 2 mmol/l EDTA, and 8 mol/l urea in a 90 mmol/l Tris-borate buffer, and the reaction products separated at a constant power of 85 W for 34 h. The radiolabeled products were resolved by exposing the gels to X-ray film for 3 days. The films were read manually, and genotypes were determined.
Genetic mapping.
A genome-wide scan was performed using 201 polymorphic simple sequence length polymorphism (SSLP) markers with an average spacing of 8.7 centimorgans (cM). Prior to performing the genetic linkage analysis, the distribution of the phenotypes in the F2 population was tested for normality using the Kolmogorov-Smirnov test. Phenotypes failing to meet the requirements of normality were transformed using either a logarithmic or square root transformation and retested for normality. Traits that were normally distributed or became normally distributed following transformation were analyzed with a parametric linkage analysis. Traits that were not normally distributed even following transformation were mapped with a nonparametric linkage analysis using nontransformed data. This analytical approach minimized the risk of the high rate of false-positive results associated with performing linkage analysis on traits that are not normally distributed (36).
Linkage analysis and the identification of QTL intervals were performed with MapMaker/EXP and MapMaker/QTL, respectively (38, 41, 42, 50). Linkage maps were constructed using the Kosambi mapping function for genetic distance calculations. For the parametric linkage analyses, suggestive and significant LOD thresholds of 2.8 and 4.3 were adopted (37). For the nonparametric analyses, a threshold of significance was determined to be a Z score 3.5. In a previous permutation study in our laboratory using a similar size F2 population and equal number of measured phenotypes, we found that these thresholds are quite conservative (64).
Definitions of "intermediate," "secondary," and "blood pressure-independent" phenotypes and the rationale for these definitions.
"Intermediate phenotypes" are defined from the genetic perspective in this analysis as those that mapped to the same region of the genome as one of the BP phenotypes and were correlated with this phenotype. "Potential intermediate phenotypes" were defined, from the genetic perspective, as traits that mapped to the same region of the genome as one of the BP phenotypes (resting SAP, DAP, and MAP, or the reduction of these pressures with salt and volume depletion) but were not significantly correlated with those BP phenotypes.
"Secondary phenotypes" were considered to be a consequence of high blood pressure and therefore defined as those that mapped to a region of the genome that did not contain a BP QTL but were significantly correlated with BP. For example, the severity of end organ damage would correlate with levels of BP in an F2 population and therefore was likely to be a consequence of the hypertension.
Finally, a large number of traits were identified that mapped to regions of the genome that did not contain a BP QTL and were not correlated with BP in either the F2 or Dahl S populations. These were classified as "blood pressure-independent phenotypes."
Statistics.
In addition to the genetic mapping techniques described above, mean values ± SE are presented, and significance of differences in corresponding values between SS, BN, and F2 rats were determined by analysis of variance followed by a Tukey or Fisher post hoc test (SigmaStat version 2.03; SPSS Inc., Chicago, IL). Correlation coefficients between the traits were determined using Pearson product moment correlation or Spearman rank order correlation (SigmaStat version 2.03). Variance of phenotype was tested using Levenes dispersion-variable test (SAS; SAS Institute, Cary, NC).
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
Plasma renin activity was lower in BN rats fed a high-salt diet than in SS rats, and it increased to a greater extent following VD. Plasma cholesterol and triglyceride concentrations were significantly higher in SS rats than in BN rats. Indices of renal end organ damage, such as plasma creatinine concentration, total protein excretion, kidney weight, glomerular size, and glomerular injury scores were all significantly higher in SS than in BN rats. Similarly, indices of cardiac hypertrophy such as heart weight and left ventricular weight were significantly higher in SS than BN rats.
The renal and peripheral vascular responses to vasoconstrictors and dilators also differed between SS and BN rats. SS rats exhibited elevated renal vasoconstrictor and BP response to NE and an elevated renal vasoconstrictor response to ANG II. The renal vascular responses and BP responses to ACh, and blockade of NO synthesis with L-NAME, were similar in the two strains.
Intermediate phenotypes have commonly been selected based on correlation analysis in hypertensive models and patients, and on mechanism-based physiology. Also, it was interesting to note that analysis of parental strains in the present study found a number of traits significantly correlated with BP in SS and BN parental rats (see Table 1).
Identification of QTL for blood pressure and related traits.
A comprehensive linkage analysis for all the phenotypes that mapped in the female F2 population is summarized in Fig. 1, coded for groupings of the QTL as given on Table 2. Schematic representations of these QTL, as well as and more detailed descriptions of the phenotypes, can be viewed on our web site (24). The female genome was calculated to be 1,941 cM in size and the marker coverage in this study averaged every 8.7 cM. The calculated size of the genome in the present study is consistent with previous estimates obtained in linkage analysis studies using male F2 populations (Bihoreau et al. 1,998 cM; Brown et al. 1,749 cM; Watanabe et al. 1,831 cM) (3, 5, 68).
|
|
Of the 20 traits reflecting various measurements of BP in the F2 population, 5 QTL were identified (Table 2). These included QTL for DAP on a high-salt diet mapped to chromosomes 14 and 15; the change in DAP and MAP following VD mapped to two different locations on chromosome 1; the change in SAP following VD that mapped to chromosome 12; and the change in BP following the electrical alerting stimulation that mapped to chromosome 11.
QTL for many other traits were identified and subdivided into functional categories that are indicated on Fig. 1 and Table 2. In addition to BP, these include BP variability, HR and HR variability, time series analysis of BP variability, BP and renal vascular resistance and RBF responses to vasoactive drugs, renal function, hormone levels, plasma lipid concentrations, morphometric measurements, and a miscellaneous category. When the QTL were sorted by these classifications, patterns became apparent. For example, a large number of QTL for BP and HR variability were identified. As can be seen in Table 2 and Fig. 1, 19 QTL reflecting the standard deviation of BP and HR were identified on 8 different chromosomes. Similarly, many of the parameters of the time series analysis reflecting BP variability also mapped to these same regions of the genome. Aggregates of these QTL were found on chromosomes 3, 6, 7, and 11. A QTL aggregate is defined as six or more QTL overlapping within their 95% confidence interval.
One of these large aggregates was identified on chromosome 6 and contained 10 QTL related to systemic vascular responses to ANG II, NE, and ACh (Fig. 1). The peaks of these QTL all fall within a narrow region flanked by markers D6Rat62 and D6Rat35, suggesting it is likely that at least one gene in this region is affecting these traits. The aggregate on chromosome 11 (Fig. 1) contained traits largely related to renal function and injury, including one QTL for BP, two QTL for BP variability, four QTL for renal function, two QTL for renal end organ damage, one QTL for the BP response to ACh, and one QTL for plasma triglycerides concentration. The genomic area harboring the peaks of these 11 QTL extended from D11rat27 to D11rat38, spanning a genetic distance of 38 cM, suggesting that there are a number of genes within the broad interval that contribute to the observed phenotypes. The suggestion of multiple genes contributing to the variability of the phenotypes in this region is supported by correlation analysis between phenotypes that mapped to this region. This analysis revealed that most of these traits were not correlated with any of the other traits that mapped to same region of chromosome 11. Several traits were correlated with each other; for example, plasma creatinine concentration positively correlated with triglyceride concentration in the F2 population (r = 0.56, P < 0.01), while both of these were inversely correlated to creatinine clearance. The BP response to ACh was also correlated with the BP response to the alerting stimulus. Total protein excretion was not correlated with glomerular diameter, despite the fact that both phenotypes mapped to a QTL in this region. This indicates that the genes responsible for the development of proteinuria and glomerular hypertrophy are likely to be different, but since they reside in the same region of chromosome 11, it is also likely that they are transmitted together.
As anticipated for multifactorial traits, multiple QTL located on different chromosomes appear to control the same phenotypes. For example, the RBF response to NE mapped to three different QTL located on chromosomes 3, 4, and 6. Baseline BP mapped to different QTL on chromosomes 14 and 15.
Some of the identified QTL appeared to interact with additional loci, as observed when the genetic variance of the first QTL was fixed (a command in MapMaker). For example, RBF mapped to marker D6rat62 on chromosome 6 with a LOD score of 4. This locus contributed 24.2% to the total variance of the trait. When the variance of this locus was fixed, a second QTL near marker D14rat90 on chromosome 14 appeared that explained another 13.2% of the total variance of this trait.
There were 44 phenotypes that were not significantly different between parental female SS and BN rats that still mapped to the genome. These traits include plasma renin activity on a high-salt diet, creatinine clearance on a high-salt diet and after salt depletion, and the BP response to the alerting stimulation.
Evaluation of intermediate phenotypes.
The analysis found that left kidney weight divided by body weight was significantly correlated with resting DAP and mapped to the same QTL as the resting DAP (see Fig. 1). This phenotype therefore represents a true intermediate phenotype and one that could potentially be used to identify causal genes of hypertension.
Table 3 shows seven "potential intermediate phenotypes" that mapped to the same region of the genome as four arterial pressure phenotypes but were not significantly correlated with these BPs. Four of the potential intermediate phenotypes were linked to the same region as the resting DAP QTL on chromosomes 14 and 15. These included the renal vascular resistance responses to intravenous infusions of ANG II and ACh, arterial pressure responses to NE, and a time series analysis trait reflecting BP variability.
|
Evaluation of secondary phenotypes.
Secondary phenotypes are herein defined as traits that were highly correlated with BP in the F2 population but mapped to regions of the genome devoid of BP QTL (Table 4). Seven phenotypes were identified that fit this classification. These include the change in MAP following VD (salt sensitivity), HR, the ratio of heart weight to body weight, urine flow, the degree of glomerular sclerosis, and plasma triglyceride concentrations. Of these phenotypes, heart weight corrected by body weight, the degree of glomerular injury, and elevations in plasma triglyceride concentration secondary to proteinuria have all been considered as clinical phenotypes reflecting renal and cardiac end organ damage. However, other phenotypes that are clinical indices of hypertension-induced end organ damage such as total protein excretion or the renal and peripheral vascular responses to ACh (index of endothelial function) were not correlated with BP in the F2 population. Phenotypes such as plasma cholesterol concentration and BP response to L-NAME were correlated with BP in the F2 population but did not map in this cross.
|
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
In general, sex differences have not been extensively explored. Previous genetic studies in Dahl S rats have focused mostly on mapping genes related to arterial pressure, body weight, or heart weight in male F2 rats (32, 54). However, sex difference effects have been reported in these traits in F2, backcross, and congenic Dahl S studies (79, 55, 56). Although hypertension has been linked to 56 homologous regions of the genome in human and experimental animal studies, to date the positional cloning of QTL has not yet progressed to the point of final proof that an identified mutation in a gene contributes to the development of hypertension or end organ damage. Even narrow congenics (17) do not meet this proof of concept. It is well recognized that the regulation of arterial pressure is one of the most redundant physiological control systems and entails the integration of many complex and interacting pathways impacting on renal, vascular, and neuroendocrine function. Furthermore, the phenotypic consequences of sustained hypertension are difficult to distinguish from potential causes of this disease. Thus the strategy of the present study was to measure as many phenotypes as possible in each animal to provide a greater breadth of phenotypic data with which to analyze the genetic determinants of this complex disease.
To increase the power of QTL detection in this study, different types of stressors were utilized, such as salt loading, acute sodium and volume depletion, an electrical alerting stimulus, active vs. resting BPs, and stressors of vascular function such as ANG II, NE, ACh, and the NO synthase inhibitor L-NAME (24). The results identified four chromosomes, 3, 6, 7, and 11, that were especially notable in the regulation of arterial pressure and renal function in female rats of the F2 population. Aggregates of 811 QTL with overlapping 95% confidence limits, were mapped to these chromosomes. Most of the traits that mapped within these QTL aggregates were not correlated with each other. Overall, 126 QTL were identified for 96 of the 236 measured or derived traits.
The finding of these functional cassettes of genetic regions that are important for the control of BP and cardiovascular traits in female rats demonstrates the power of animal models in the dissection of multifactorial disorders, aw well as the utility of this approach for studying complex diseases. The genomic regions harboring genes linked to many of the features of hypertension, and related metabolic and organ damage phenotypes, delineates the locations in the genome for which it will be especially interesting to develop consomic and congenic animal models. Such model systems will enable one to address distinct questions as to the causal relationship of disease phenotypes and the mechanisms underlying the progression of the disease, which then can be directly translated to the homologous regions of the human genome (27, 66).
BP phenotypes, per se, of the female F2 rats as determined by direct measurements mapped to chromosomes 1, 12, 14, and 15. As recently reviewed by J. Rapp (54), there is evidence from a number of previous studies demonstrating that chromosome 1 contains at least one blood pressure gene based upon male F2 intercross populations involving not only Dahl S, but also spontaneously hypertensive (SHR), SBH, and FHH rats (54). Many of these BP QTL, including those in our female population, fall within the region of chromosome 1 containing the SA gene and other interesting candidate genes including renal kallikrein, and the ß- and -subunits of the epithelial sodium channel (Scnn1b and Scnn1g). With regard to chromosome 12, there is one report showing linkage of BP to the same broad QTL region found in the present study in a male F2 population of Dahl S and Wistar-Kyoto (WKY) rats (32). The BP QTL identified on chromosome 14 in the present study is unique in that this region has not been previously linked to BP in any previous study involving Dahl S rats. However, the
-adducin gene located in the same region on chromosome 14 has been linked to BP in Milan hypertensive rats (MHS) (1). The BP QTL found in the present study on chromosome 15 is the first that has been reported in any F2 intercross study and could be unique to female populations.
The extent to which the ovulatory cycle influenced the QTL analysis in this study is an issue that cannot be overlooked. The estrus cycle was not synchronized in the present study, nor was the cycle pharmacologically eliminated. It was not practical in this study to carry out vaginal smears to select the rats at the same phase of their estrus cycle for all of these studies. For the linkage analysis, these cycles were therefore considered as environmental or background "noise." We found, however, that in only 5% of the traits was the statistical variance of the measured traits greater in the F2 females than was found in the F2 males of this intercross, indicating that the estrus "noise" was not greater than that created by other environmental factors and variations of measurements.
Identification of "intermediate phenotypes."
The present study also sought to identify intermediate phenotypes, since any such leads might be useful in segregating hypertensive patient populations into more homogeneous subgroups, thereby reducing variance and facilitating the identification of genes that regulate BP in humans. It has been difficult to distinguish between phenotypes that are causally linked to the function of genes contributing to the development of hypertension vs. phenotypes that are a consequence of hypertension. It has also been difficult to distinguish those traits that become genetically fixed in inbred animal models of hypertension by virtue of being carried on the same chromosome as the hypertension-causing gene. These phenotypes may or may not be causal of the elevated BP, but are part of the disease process and correlated with the pressure. The concept of "intermediate phenotypes" arose from the widely recognized need to find traits that are relatively easy to measure and regulated by less redundant pathways than BP and therefore are presumably "closer" in a functional sense to the causal gene(s). A number of intermediate phenotypes such as BP responses to air jet stress have been proposed in rodent models of hypertension (14), and some have been proposed in human hypertension such as urinary free cortisol (43) and kallikrein concentrations (29), indices of sympathetic nervous system function (45), and nonmodulation of plasma renin activity with changes in salt intake (70). However, none of these putative intermediate phenotypes have been yet found to map to the same region of the genome as BP, a prerequisite if the trait is to be used as a surrogate to map for (BP) genes of hypertension.
An intermediate phenotype is expected to exhibit several basic characteristics. First, the phenotype is expected to map to the same region of the genome as BP. Second, the phenotype must be correlated with BP in the population studied. Third, the phenotype is expected to be linked closely to the primary action of genes that directly contribute to hypertension (27). In the present study, one phenotype met all of these criteria: the left kidney weight divided by body weight. This trait was identified on chromosome 15 in the same region as the QTL for resting DAP. This is therefore the first identified intermediate phenotype that meets all of the expected criteria. This is relevant not only because at this time none other has been identified, but also because previous studies have indicated that the risk of hypertension and progressive renal disease is enhanced in subjects with reduced nephron numbers and greater kidney weight (4, 33). It has recently been reported that patients with hypertension had significantly fewer glomeruli per kidney than matched normotensive controls (702,379 vs. 1,429,200 glomeruli) (33). Furthermore, kidney weight divided by body weight of the hypertensive subjects was 11% greater than in normotensive subjects (33). Similar relationships have been described in SHR compared with normotensive WKY rats (62). In the present study, the average kidney weight of SS rats was 38% greater than that of the normotensive BN rats. Since kidney size is relatively easy to quantify noninvasively in humans using ultrasound imaging techniques, this trait might one day prove to be a useful "intermediate phenotype" of hypertension.
It is interesting that of the many phenotypes determined in this study, only kidney weight divided by body weight met the strict genetic definition of an intermediate phenotype. This may be explained by a lack of statistical power to map a sufficient number of BP phenotypes. It might also be due to the complex nature of BP regulation, whereby it is difficult for any one phenotype to both map to the same region of the genome as BP and also correlate with BP. It may also be due to a phenotyping window effect, as modeled by Schork et al. (59), whereby a true intermediate phenotype exists but contributes to the BP only earlier in life and no longer correlates with BP at later times in life despite mapping to the same QTL confidence interval.
"Potential intermediate phenotypes."
A number of other phenotypes warrant consideration in that they exhibited QTL that mapped to the same region of the genome as BP or BP responses with salt and volume depletion (referred to as blood pressure "salt sensitivity") but were not correlated with arterial pressure. These included renal vascular resistance responses to ANG II that mapped both to the DAP QTL on chromosome 14 and to a QTL on chromosome 1 representing the extent to which BP decreased following VD. Notably, parental SS rats exhibited significantly greater reductions in RBF than BN rats in response to intravenous infusion of ANG II (Table 1). The renal vascular response to ANG II has also been proposed as an intermediate phenotype in humans and has been used to stratify patients with essential hypertension according to whether they exhibit a blunted increase of adrenal aldosterone secretion or a blunted reduction of RBF in response to ANG II, so called "nonmodulators" (70). Approximately 4050% of all hypertensive patients appear to be "nonmodulators."
The renal vascular responses to infusions of ACh mapped to the QTL for DAP on chromosome 15. Although this may be a difficult phenotype to routinely assess in humans, it is of interest because of the important role that NO plays in the regulation of vascular tone and in the long-term regulation of renal function and arterial pressure (11). Another of the interesting phenotypes that fell into the category of a potential intermediate phenotype were several derived traits designed to be related to the behavior of the arterial baroreceptor reflexes, as recently described by Kendziorski et al. (34). These traits were obtained utilizing time series analysis techniques applied to continuous MAP recordings. This analysis utilized BP measurements obtained from parental and F2 female rats to generate autocorrelations and partial autocorrelation functions for each F2 rat. As seen in Table 3, the second moment of the arterial pressure that represents the variability of pressure (Bptsm_tpm_alpha3) mapped to the same region of chromosome 15 as DAP. These results suggest that indices of BP variability may serve as a "potential intermediate phenotype" for hypertension. Continuous recording of arterial pressure has now become routine in clinical practice, and these noninvasive methods could be used to test whether various indices of BP variability might have predictive value for patients that later develop hypertension (49, 51).
"Intermediate phenotype": a commonly used or abused term.
Semantics can often cloud scientific understanding when moving between two different disciplines, in this case genomics and physiology. In the present study the term "intermediate phenotype" has been defined in the manner accepted generally by the genetics community. A different understanding of this term is often assumed by physiologists. For a physiologist, the idea that generally comes to mind when hearing the term "intermediate phenotype" for hypertension is that of a biological pathway that is an important determinant of arterial pressure. For this reason, we have previously referred to the many phenotypes measured in linkage analysis studies as "likely determinant phenotypes" (64).
The diagram shown in Fig. 2 is included to clarify the terms "intermediate," "potential intermediate," and "secondary" phenotypes of arterial BP. The genome is reflected on the left of this diagram with genes connecting through a variety of complex biological pathways that likely determine the intermediate phenotypes (or potential ones). These phenotypes may or may not be causal of the hypertension but are part of the disease syndrome and highly correlated with the BP. Some individual genes can affect more than a single phenotype (pleiotropy) such as the gene for vascular NO synthase (NOS III) that would be expected to affect the systemic and renal vascular responsiveness to administration of ACh, NE, and ANG II as illustrated in Fig. 2. Gene(s) exerting a measurable effect on a phenotype that was influencing arterial pressure, however, would be expected to segregate with both this trait and also the arterial pressure (e.g., map to the same region as the BP in the F2 generation offspring). If the phenotype was part of the syndrome of hypertension, and if it was to be a useful predictor of hypertension, then it would also be expected to correlate with BP. In the present study, the trait shown in bold type (kidney weight divided by body weight) fulfilled both of these criteria while the others failed to correlate with resting diastolic pressure and were therefore called "potential intermediate phenotypes."
|
Identification of "secondary phenotypes."
Seven phenotypes that were correlated with BP mapped to regions of the genome devoid of BP QTL, and were therefore classified as secondary phenotypes (Table 4) and likely to be a consequence of hypertension in the F2 population. It is interesting that HR was one of these phenotypes. It is well recognized that HR is elevated in essential hypertension (48) and indeed is elevated in SS rats compared with BN rats. An elevated HR has often been assumed to reflect excess sympathetic nervous system activity and has been taken as evidence of the relative role of the nervous system in hypertension (28, 44).
Plasma triglyceride concentration and the degree of glomerular injury were highly correlated with MAP in the F2 female rats, but these phenotypes failed to map within BP QTL regions. It is well recognized that hyperlipidemia, proteinuria, glomerular injury, and renal hypertrophy are associated with hypertension. The present analysis would indicate that these phenotypes were secondary to the hypertension in the F2 female rats. Consistent with these findings was the high correlation found between both the glomerular injury score and plasma lipids and the MAP.
There has been considerable debate about whether renal dysfunction in hypertension is a cause or consequence of essential hypertension. We recognize that the present definitions do not resolve this issue, but rather serve as a way to begin thinking more precisely about the complex relationship between cause and consequence. For example, if one accepts the proposed definition of a "secondary phenotype," then the data would indicate that both the renal injury and hyperlipidemia fall into this category. However, one would expect that the genetic background of each of the individual F2 rats in our population, as in the general human population, would influence the effects of arterial pressure upon end organ damage. If so, then individuals in the population with a permissive genetic background to develop hypertension-induced renal damage should exhibit a high correlation between glomerular injury and the level of BP. We would define the glomerular injury in these individuals as "secondary trait" and a consequence of the interaction with hypertension. Formally, until the entire genetic basis of hypertension is defined in this cross, we cannot exclude that these secondary traits are contributing to BP. However, several studies in the Dahl S (16, 17), FHH (61), GH (22), SHR (25), SHRsp (57), and the Munich-Wistar-Fromter rats (60) have shown that renal failure, left ventricular hypertrophy, and stroke are genetically independent from BP, but interact with it. Consequently, it is likely that the same is true for the SS females.
Blood pressure-independent phenotypes.
Most of the QTL for phenotypes measured in the present study were not correlated with BP traits in either the F2 or SS rats. All that can be concluded from these observations is that although these traits may differ in the SS and BN rats, these phenotypes appear to be BP independent. We conclude that these phenotypes may have been captured by chance or because they are close to hypertensive loci that were selected for during the initial breeding of the SS rat for hypertension (19).
Summary.
The present study represents a first attempt to develop a detailed systems biology map of cardiovascular function specific for female rats. Major regions of the genome harboring QTL for BP and related metabolic, vascular, and end organ disorders were identified. Four chromosomes (3, 6, 7, and 11) were especially important and harbored aggregates of eight or more QTL for various indices of BP and cardiovascular and renal function. The present data prioritize the regions of the genome of most interest to create consomic and congenic strains of SS rats to deconvolute the genes dictating the complex interactions among hypertension, renal, and cardiovascular disease. The analysis also identified left kidney weight (divided by body weight) as an "intermediate phenotype," which upon further study might prove useful in helping to unravel the genetic basis of hypertension in animal models of hypertension and in patients. With the recent release of the sequence of the rat genome and its incorporation into a variety of comparative genomics tools, it is also now possible to directly compare these results homologous regions of the mouse and human genomes. This will allow for translation of the information contained in this female-specific systems biology map of cardiovascular function in rats to identify target regions and genes for further testing in female patient populations.
![]() |
GRANTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
ACKNOWLEDGMENTS |
---|
Editor C. Sigmund served as the review editor for this manuscript submitted by Editors A. S. Greene, R. J. Roman, H. J. Jacob, and A. W. Cowley, Jr.
![]() |
FOOTNOTES |
---|
Address for reprint requests and other correspondence: A. W. Cowley, Jr., Dept. of Physiology, Medical College of Wisconsin, 8701 Watertown Plank Rd., Milwaukee, WI 53226 (E-mail: cowley{at}mcw.edu).
10.1152/physiolgenomics.00105.2003.
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|