| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 The Jackson Laboratory
2 Howard Hughes Medical Institute, Bar Harbor, Maine 04609
3 Laboratory Animal Resource Research Center, University of Tsukuba, 1-1-1 Tennodai, Tsukuba, Ibaraki 305-8575, Japan
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
quantitative trait loci; hypertension
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Several of the rat and mouse hypertension QTL map to regions that contain homologous genes, suggesting that the genes determining hypertension are also homologous. Likewise, several of the limited number of human hypertension QTL also map to regions that contain genes homologous to those in rat and mouse hypertension QTL (28, 30). This concordance of hypertension QTL across species suggests that common allelic variations occur in only a subset of all the proteins that are involved in blood pressure regulation. These may be mutations in genes that code for rate-limiting or regulatory proteins since most proteins are present in excess, as shown by the large number of recessive diseases. The fact that QTL are concordant across species should help to guide the search for the genetic determinants of human hypertension, since many studies, such as narrowing a QTL region or carrying out expression studies, are more easily carried out in a rodent model.
In this study, we further exploited the mouse as a hypertension model. We conducted a QTL analysis of a population of (BALB/cJ x CBA/CaJ) F2 males to identify genomic regions and gene-gene interactions that regulate blood pressure, heart rate, and heart weight. We identified several significant QTL: two associated with blood pressure, three significantly associated with heart rate, and one associated with heart weight.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Phenotyping.
To measure blood pressure and heart rate, we used a blood pressure analysis system (model BP-2000; Visitech Systems, Apex, NC). The machine accommodated four mice and consisted of a warming plate, which kept the mice at 38°C, and four restraining units (dark metal tunnels open at one end). Computer-automated tail cuffs placed on the tails of each mouse rapidly inflated and deflated, and blood pressure was detected by a photoresistor cell below each tail and recorded by a computer (14). The machine also recorded the heart rate of each mouse.
We reduced the variation in blood pressure by acclimating mice to the room for at least 2 wk, measuring their blood pressures when they were 8 wk (±1 wk) old, and taking all measurements in the morning (9–12 AM). Additionally, we trained the mice to the machine and all the associated measurement procedures for 5 days, designated as days 1–5, without recording blood pressures, then measured blood pressures on days 8–12, each day taking 20 measurements for a total of 100 measurements per mouse (expressed in mmHg as means ± SE). One of us (F. Sugiyama) carried out all blood pressure measurements. Mice were killed, and body and heart weights were determined. The numbers of mice are listed in Tables 1–3 and Figs. 1–5.
|
|
|
|
|
|
|
|
Genotyping.
We extracted DNA from the tail of each mouse (31) and genotyped it with 91 simple sequence length polymorphic (SSLP) markers spaced at 
15-cM intervals throughout the mouse genome; 11 additional SSLP markers were added in specific locations after we detected QTL (Fig. 1). We reported map positions in centimorgans using the 2001 Mouse Genome Informatics database (http://www.informatics.jax.org). Because most genetic information for a quantitative trait is at the extreme ends of its distribution in a population, we selected for genotyping the mice in the upper and lower 20% of each phenotypic trait. Because we measured so many traits, selecting the upper and lower 20% for each trait resulted in 162 of the 195 (83%) F2 males being genotyped; all genotyped animals were used for the genome scan.
We purchased primers from Research Genetics (MapPairs, Huntsville, AL). The Allele Typing Service of the Jackson Laboratory genotyped the samples using fluorescent primers and the ABI model 3700 capillary electrophoresis apparatus. We reviewed the genotypes for the presence of double recombinants over short genetic distances, and questionable genotypings or missing data were repeated in our laboratory using standard procedures (17). The average spacing between markers was 15 cM. Four regions on chromosome (Chr) 10, 12, 13, and X had greater than 30 cM between markers. Since we found no polymorphic markers in these four regions (data not shown), we think these regions may be identical by descent because these two strains do share a common origin. Chromosome lengths were appropriate, and all markers fit a Hardy-Weinberg distribution.
Statistical analyses.
To compare blood pressures between parental and F1 mice, we used Student’s t-test. To detect QTL significantly associated with blood pressure, heart rate, or heart weight, we analyzed genome-wide scans in three steps (29). In the first step, we identified main QTL associated with each phenotype by computing a logarithm of the odds ratio (LOD) score at 2-cM steps over the entire genome and compared these to significance thresholds computed by permutation analysis (4). Confidence intervals were computed by the method of Sen and Churchill (24) by finding the region under the posterior density curve (10^LOD) that contains 95% of the total area. In the second step, we identified pairs of loci significantly associated with a phenotype by conducting a simultaneous search for pairs analysis (24). Genome-wide scans were implemented in MATLAB software (Mathworks, Natick, MA; source codes available at http://www.jax.org/research/churchill). In the third step, we integrated all the main and interacting QTL-phenotype associations detected in the first two steps into a multiple regression (using SYS/STAT Software, version 8; SAS Institute, Cary, NC). F-statistics based on adjusted (type III) sums of squares were used to determine the contribution of a QTL in combination with all other QTL. These statistical methods have been described previously (24).
To assess the evidence for multiple linked QTL on the same chromosome, we calculated LOD scores for one, two, and three QTL models on each chromosome that showed a significant main effect. The difference in LOD score between the single and multiple QTL models was used to compute a likelihood ratio test (and thus a P value) for the evidence in favor of multiple linked QTL. These test statistics have nonstandard distributions (23), so we conducted simulations to estimate significance thresholds and P values. For the comparison of 2-QTL vs. 1-QTL models, the 95% significance threshold is 3.4 for additive QTL and 3.9 if we allow for interaction between the two QTL. The estimates are based on 1,000 simulated intercrosses with 200 mice and a single QTL with an effect size of 0.5 standard deviations. LOD scores for 3-QTL vs. 2-QTL models in this study were clearly nonsignificant, and no simulations were carried out.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The heart rates of BALB mice were significantly slower than those of CBA mice (484 ± 9 vs. 620 ± 7 beats/min, P < 0.0001, Table 1); those of the F1 progeny were similar to each other and intermediate to the heart rates of the parental strains with a tendency to be more like CBA (Table 1). The hearts of BALB mice weighed significantly more than did those of CBA mice (120 ± 2 vs. 108 ± 2 mg, P < 0.0001, Table 1). The genes underlying the relatively slower heart rates and larger heart weights of BALB mice were probably recessive, because these phenotypes in F1 and F2 progeny were more CBA-like than BALB-like (Table 1). The mean blood pressures, heart rates, and heart weights of the F2 progeny were distributed normally around the mean (Fig. 2).
Identification of QTL for blood pressure.
The genome-wide scan indicated significant QTL for blood pressure on Chrs 7 and 15 (Fig. 3A). The Chr 15 QTL, already named Bpq6 (for "blood pressure QTL 6") was previously found in two QTL crosses between strains A/J and C57BL/6J and between strains BPH and BPL [these latter strains are "blood pressure high" and "blood pressure low" strains developed by Schlager (22)]. However, in this cross the Chr 15 QTL is broad as shown by the interval map together with the 95% confidence interval (Fig. 4A). Confidence intervals were computed by the method of Sen and Churchill (24) by finding the region under the posterior density curve (10^LOD) that contains 95% of the total area. Equivalent results are obtained by the method of dropping down 1.5 units from the peak of the LOD curve. We calculated the LOD scores for the 1-QTL, 2-QTL, and 3-QTL models for Chr 15 without interaction effects. The difference in LOD scores for 2-QTL vs. 1-QTL models is 1.9 (P = 0.51) and is not statistically significant. We note, however, that this does not rule out the possibility of a second QTL on chromosome 15 (see DISCUSSION). The difference between the 2-QTL and the 3-QTL model is a LOD score of 0.5, which is not significant.
The Chr 7 QTL was also found previously in the A/J x C57BL/6 cross as an interaction (24); it is now named Bpq7. The interval map for Bpq7 and the 95% confidence interval, as determined by posterior probability density (24), is shown in Fig. 4B. The allele effects of Bpq6 and Bpq7 are shown in Fig. 5, A and B. The Bpq6 allele for high blood pressure came from strain CBA, and the Bpq7 allele for high blood pressure came from strain BALB.
The scan for interacting QTL did not reveal any for blood pressure in this cross. The multiple linear regression (Table 2) shows that the Chr 7 and 15 QTL accounted for 22.5% of the variance. Table 3 lists the chromosomal location, LOD scores, allele giving the highest value, and peak markers for each blood pressure QTL and for the heart rate and heart weight QTL discussed below.
Identification of QTL for heart rate and heart weight.
The genome-wide scan revealed significant QTL for heart rate on Chrs 2 and 15 (Fig. 3B). The interval maps for these QTL and the 95% confidence intervals are shown in Fig. 4, C and D. The allele effects of the Chr 2 QTL, named Hrq1 (for "heart rate QTL"), and the Chr 15 QTL, named Hrq2, are shown in Fig. 5, C and D. In both cases strain CBA contributed the allele for high heart rate; the CBA allele is additive for Hrq1 but dominant for Hrq2. This is consistent with the heart rates in F1 progeny (Table 1), which show that heart rate is between the two parents but tends to be more like CBA. Hrq2 on Chr 15 appears to be different than Bpq6 on Chr 15; their 95% confidence intervals overlap only slightly (0–25 cM for Bpq6; 20–35 cM for Hrq2).
The Chr 2 QTL is broad with a distinct cusp (Fig. 4C). We calculated the LOD scores for the single QTL and the two QTL model for Chr 2; the difference between the 1-QTL LOD and the 2-QTL LOD with interaction is 3.15 (P = 0.19). Although the test failed to achieve the standard 0.05 level of significance, we interpret this as suggestive evidence for multiple QTL. The maximal peak for a single QTL on Chr 2 occurs at 72 cM and we name this QTL Hrq1. If a second locus exists, it would localize somewhere in the region 40–80 cM, so it is tightly linked. The two loci appear to interact epistatically, but with tight linkage not all genotype combinations occur in the cross, and it is difficult to characterize the interaction.
The search for gene-gene interaction showed a significant interaction between D1Mit10 (57 cM) and D2Mit304 for heart rate. We named this interacting QTL on Chr 1 Hrq3. The magnitude of the interaction between Hrq3 and Hrq1 (Fig. 5F) shows that heart rate for mice with a D1Mit10 CBA allele in either a homozygous or heterozygous state are very affected by the alleles at D2Mit304 with the CBA allele contributing to higher heart rate. For mice that are homozygous for the BALB allele at D1Mit10, the Hrq3 locus has no effect. The multiple regression analysis for heart rate (Table 2) shows that these three QTL account for 32.2% of the variance in heart rate.
The genome scan for heart weight was carried out on the log of heart weight and also on the log of heart weight adjusted with body weight as a covariate. The two genome scans gave equivalent results; the scan for log heart weight is shown (Fig. 3C). One significant QTL was found at D14Mit67 with the peak at cM 38 and the 95% confidence interval at 20–43 cM (Fig. 4E); we name this QTL Hwq1 for "heart weight QTL 1." Two additional QTL that reach suggestive significance were also found, both on Chr 12. These were not named because they are only suggestive, but both were included in the multiple regression analysis since we used a high threshold for suggestive. The allele effects of Hwq1 are shown in Fig. 5E, and the regression analysis is shown in Table 2. The significant locus on Chr 14, and the two suggestive loci on Chr 12 account for 13.1% of the variance.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
In our current study, we conducted a QTL analysis of (CBA x BALB) F2 progeny. Ordinarily, the blood pressure difference of 8 mmHg between the two parental strains with standard deviations in the range of 5 would not be sufficient to justify a QTL analysis, since this represents a strain difference of less than 2 standard deviation units. However, we planned this cross for other purposes, and the wide distribution of blood pressure among F2 progeny encouraged us to search for QTL for this trait. We identified two QTL associated with blood pressure, Bpq6 and Bpq7. The fact that BALB and CBA each contributed one allele for high blood pressure may explain the small interstrain blood pressure difference between the two parents compared with the difference in the F2 population. We identified the effect of Bpq6 on blood pressure in a previous study (29) as an interacting QTL; Bpq6 had an effect on blood pressure through its interaction with Bpq5 on Chr 6. We have assumed that Bpq6 found in this cross and the previous studies identify the same gene, although further data may reveal that these are distinct genes. Bpq7 was also found in our previous cross as an interacting QTL between A/J and C57BL/6, although we did not report it in our original paper (29). It was found and reported later as our methods for finding interacting QTL improved (24). We also identified three QTL significantly associated with heart rate; Hrq1, Hrq2, and Hrq3 and one associated with heart weight, Hwq1.
Two of the QTL (Bpq6 and Hrq1) reported in this study appear to consist of multiple loci. We feel that it is important to report indications that a QTL region may harbor multiple loci, because it is our experience that upon further dissection (e.g., by analysis of congenics) such QTL may break apart into multiple independent effects. The lack of strong statistical support for multiple QTL is likely to be a consequence of low power to detect linked QTL. Although we can provide some indication of the complexity of these QTL using statistical indices, it is difficult to precisely localize multiple effects in an intercross population of this size.
Bpq6, Bpq7, Hrq1, and Hrq3 each harbor promising candidate genes. We previously reported that Bpq6 contains two candidate genes; prostaglandin receptor gene Pterep2 and Npr3 (natriuretic peptide receptor 3; cM 6.7). However, further mapping studies showed that Pterep2 was really on Chr 14 and the prostaglandin receptor on Chr 15 was a different member of the same gene family, Ptgerep4 (prostaglandin E receptor EP4 subtype; cM 6.4). Like the Ptgerep2, Ptgerep4 codes for a G protein-coupled receptor, and a number of studies support its role in blood pressure regulation (1–3, 7, 20). The most promising candidate genes for Bpq7 are Adm (adrenomedullin; cM 50.5), which encodes a potent vasodilator peptide that exerts major effects on cardiovascular function (21); Pth (parathyroid hormone; cM 52.5), Calc (calcitonin, 
-calcitonin gene-related peptide; cM 54), which shares some structural homology with Adm and is a potent vasodilator (2, 20), and Scnn1b and 1g (sodium channel, non-voltage-gated 1ß and 1
; cM 56). Mice that lack calcitonin develop hypertension (8). Scnn1b and Scnn1g, the sodium channel, non-voltage-gated 1ß or 1, participate in the control of sodium flux in the kidney; mutations in these proteins have been associated with Liddle syndrome, a rare form of monogenic human hypertension (9, 26). Some interesting candidate genes for heart rate are located on Chrs 1 and 2. Acetylcholine is known to affect heart rate (3), and two classes of cholinergic receptors are known; the muscarinic receptors, of which five are known, and the nicotinic receptors, which exists as a pentamer composed of four different subunits. The muscarinic receptors are G-coupled proteins and are known to affect heart rate; the nicotinic receptor is a ligand-gated ion channel that affects sodium and potassium flux and may affect heart rate (3, 10, 11, 25, 32). Chrna1 (cholinergic receptor, nicotinic, polypeptide-) is located on Chr 2 at cM 43, and two other subunits of this receptor are located on Chr 1 at cM 52.3 (Chrnd, Chrng; cholinergic receptor, nicotinic, polypeptides 
and ). These subunits and one other form a pentamer that is the receptor. Two of the muscarinic receptors are also located on Chr 2 in the Hrq1 region, Chrm4 and Chrm5 (cholinergic receptor, muscarinic, 4 and 5; cM 49 and 58).
Evidence for the concordance of QTL across species is increasing. Previously, several studies demonstrated that a QTL in one experimental animal model can predict the location of a QTL in a homologous region in another species, especially humans (13, 28, 29). A recent review provides evidence for additional QTL that are concordant in mice, rats, and humans (12). Our current study offers more evidence for this type of concordance: both murine hypertension QTL we identified were concordant with blood pressure QTL in humans. Bpq6 is homologous to human 5p12–14 and 8q22–24; a human hypertension QTL has been found in both of these regions region (7, 19). Bpq7 is homologous to three human chromosomal regions: cM 35–44 is homologous to human 15q23–26; cM 44–54 are homologous to human 11p12–13 and 11q13–21; and cM 55–56 are homologous to human 16p12–13. A human hypertension QTL has been found for 15q25.1–26.1 (16, 34), and a rare form of monogenic human hypertension, Liddle syndrome, maps to 16p12–13 (9, 26). The rat homology to Bpq6 is not known except for a 2-cM region homologous to rat 2; this small region of rat Chr 2 is part of a hypertension QTL (32). Homology to rat Chr 7 begins at the distal edge of Bpq6 at cM 25; this region of rat Chr 7 does contain a hypertension QTL (5, 6), but the most likely candidate gene Cyp11b1 (7) maps to mouse cM 44, which is clearly out of the Bpq6 region. Since the Chr 15 QTL appears to be complex, there may be a second QTL on Chr 15 which is homologous to rat 7, but it is not the region we define as Bpq6. Bpq7 is homologous to rat Chr 1, which does contain a QTL for hypertension (32). A QTL for heart rate has been found in rats (15); its location on rat Chr 3 has a peak homologous to mouse Chr 2, cM 35. This is not the location of Hrq1, which is at cM 72, but we note that the Chr 2 locus appears to be complex and may have a second peak about cM 40; this second peak is homologous with the rat QTL.
These results establish the laboratory mouse as an excellent model for essential hypertension. Further work in additional inbred mouse strains will more clearly characterize quantitative blood pressure and other cardiovascular traits and enable us to elucidate the genes that regulate these traits in humans.
|
ACKNOWLEDGMENTS |
|---|
This work was supported by grants from the National Heart, Lung, and Blood Institute’s Program for Genomic Applications (HL-66611) and the SCOR program (HL-55001), by the American Heart Association Grant 50564N (G. A. Churchill and R. Li), by the American Health Assistance Foundation G1999023, and by the Japanese Ministry of Education, Science, Sport, Culture, and Technology (F. Sugiyama). S. W. M. John is an assistant investigator of the Howard Hughes Medical Institute.
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: B. Paigen, The Jackson Laboratory, 600 Main St., Bar Harbor, ME 04609 (E-mail: bjp{at}jax.org)
10.1152/physiolgenomics.00002.2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. Howden, E. Liu, L. Miller-DeGraff, H. L. Keener, C. Walker, J. A. Clark, P. H. Myers, D. C. Rouse, T. Wiltshire, and S. R. Kleeberger The genetic contribution to heart rate and heart rate variability in quiescent mice Am J Physiol Heart Circ Physiol, July 1, 2008; 295(1): H59 - H68. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Sheehan, S.-W. Tsaih, B. L. King, C. Stanton, G. A. Churchill, B. Paigen, and K. DiPetrillo Genetic analysis of albuminuria in a cross between C57BL/6J and DBA/2J mice Am J Physiol Renal Physiol, November 1, 2007; 293(5): F1649 - F1656. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. C. Solberg, A. E. Baum, N. Ahmadiyeh, K. Shimomura, R. Li, F. W. Turek, J. S. Takahashi, G. A. Churchill, and E. E. Redei Genetic analysis of the stress-responsive adrenocortical axis Physiol Genomics, November 21, 2006; 27(3): 362 - 369. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. B. Gould, F. C. Phalan, S. E. van Mil, J. P. Sundberg, K. Vahedi, P. Massin, M. G. Bousser, P. Heutink, J. H. Miner, E. Tournier-Lasserve, et al. Role of COL4A1 in Small-Vessel Disease and Hemorrhagic Stroke N. Engl. J. Med., April 6, 2006; 354(14): 1489 - 1496. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. E. Baum, L. C. Solberg, P. Kopp, N. Ahmadiyeh, G. Churchill, J. S. Takahashi, J. L. Jameson, and E. E. Redei Quantitative Trait Loci Associated with Elevated Thyroid-Stimulating Hormone in the Wistar-Kyoto Rat Endocrinology, February 1, 2005; 146(2): 870 - 878. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Varona, J. Casellas, J. Piedrafita, A. Sanchez, P. Garcia-Casado, M. Arque, and J. L. Noguera Bayes factor analysis for the genetic background of physiological and vitality variables of F2 Iberian x Meishan newborn piglets J Anim Sci, February 1, 2005; 83(2): 334 - 339. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. F. Deschepper, J. L. Olson, M. Otis, and N. Gallo-Payet Characterization of blood pressure and morphological traits in cardiovascular-related organs in 13 different inbred mouse strains J Appl Physiol, July 1, 2004; 97(1): 369 - 376. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. J. Martin, A. G. Comuzzie, G. E. Sonnenberg, J. Myklebust, R. James, J. Marks, J. Blangero, and A. H. Kissebah Major Quantitative Trait Locus for Resting Heart Rate Maps to a Region on Chromosome 4 Hypertension, May 1, 2004; 43(5): 1146 - 1151. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. DiPetrillo, S.-W. Tsaih, S. Sheehan, C. Johns, P. Kelmenson, H. Gavras, G. A. Churchill, and B. Paigen Genetic analysis of blood pressure in C3H/HeJ and SWR/J mice Physiol Genomics, April 13, 2004; 17(2): 215 - 220. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Schwartz, A. Duka, E. Triantafyllidi, C. Johns, I. Duka, J. Cui, and H. Gavras Serial analysis of gene expression in mouse kidney following angiotensin II administration Physiol Genomics, December 16, 2003; 16(1): 90 - 98. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Le Corvoisier, H.-Y. Park, K. M. Carlson, D. A. Marchuk, and H. A. Rockman Multiple quantitative trait loci modify the heart failure phenotype in murine cardiomyopathy Hum. Mol. Genet., December 1, 2003; 12(23): 3097 - 3107. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. D. L. Woo and I. Kurtz Mapping blood pressure loci in (A/J x B6)F2 mice Physiol Genomics, November 11, 2003; 15(3): 236 - 242. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Wong, N. R. Mahapatra, S. Chitbangonsyn, P. Mahboubi, M. Mahata, S. K. Mahata, and D. T. O'Connor The angiotensin II receptor (Agtr1a): functional regulatory polymorphisms in a locus genetically linked to blood pressure variation in the mouse Physiol Genomics, June 24, 2003; 14(1): 83 - 93. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. W. Cowley Jr. Genomics and homeostasis Am J Physiol Regulatory Integrative Comp Physiol, March 1, 2003; 284(3): R611 - R627. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
