| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1 Max Delbrück Center for Molecular Medicine, Berlin, Germany
2 Inserm, U698, Hôpital Bichat, and Université Paris 7, Paris, France
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
LOU rat; quantitative trait loci; congenic line; renal pelvic epithelial proliferation
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Clinically, because it is now possible to detect antenatal HN by ultrasound screening (4, 10) the question arises as to whether or not intervention is required to preserve renal function (4). Some cases of neonatal HN resolve spontaneously or have no adverse effects on renal function, whereas in others, usually involving obstruction, renal deterioration will occur if no intervention is made. Until the various mechanisms of HN formation are elucidated, it will be difficult to distinguish between the asymptomatic congenital form and the form that evolves rapidly toward destruction of the renal parenchyma, a distinction that is crucial for optimal postnatal clinical management.
One way of investigating the origin of HN is to perform genetic linkage studies in an animal model that develops the condition spontaneously. HN exists sporadically in the rat, and some inbred strains have been shown to develop it with a high frequency, suggesting that it is genetically determined. The Imamichi strain, obtained from a Wistar colony by selective breeding (39), presents a high incidence of severe unilateral HN, rapidly evolving toward renal damage (35, 46). On the other hand, the Brown Norway (BN) strain is affected by a milder form of congenital HN (32, 37), which does not evolve rapidly because renal function is reported to be preserved until quite late in life (33). Another strain, also derived from Wistar, appears to present intermediate characteristics (11, 43). Although several groups have investigated various aspects of this spontaneous HN in rats, in which there is undoubtedly a significant genetic component, no genetic linkage studies have yet been performed for this phenotype. We have thus carried out such a study for HN in a large cohort of rats originating from a cross between BN and LOU. Previous observations in our laboratory have shown that the LOU reference strain presents a much lower incidence of HN compared with the BN strain, where it is very frequent.
We chose to use backcross (BC) rats [(BN x LOU) x BN] for this linkage study because the same rats were used to study genetic linkage of several arterial phenotypes exhibited by the BN (20). A backcross was more suited to the study of two of these phenotypes, rupture of the abdominal aortic internal elastic lamina (IEL) and persistent ductus arteriosus (PDA), than an intercross (F2). Moreover, preliminary studies on F1 rats (BN x LOU) showed that HN was also a recessive phenotype for which the use of a backcross was well adapted.
In addition to identifying quantitative trait loci (QTLs) for HN, one of our aims was to determine whether this phenotype is linked to the various arterial phenotypes exhibited by the BN rat (IEL rupture, PDA, aortic elastin deficit), thus comprising a syndrome due to some common underlying mutation, or whether they are phenotypes that are independent.
Since we have observed during the course of this study a significant incidence of renal signs of hematuria at necropsy in the BC population, and subsequent observations showed that the rats of our BN colony also often present with intermittent macroscopic hematuria, we have included a description of this phenomenon in our study and analyzed its relation with HN.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Death and Phenotyping
Fourteen F1 rats (9 males and 5 females) and 239 BC rats (121 males and 118 females) were phenotyped for HN as described below. At 18 wk of age, under pentobarbital anesthesia (50 mg/kg ip), body weight and nose-rump length were recorded and a liver sample was snap-frozen in liquid N2 for DNA extraction. The kidneys were then perfusion-fixed with buffered formalin by placing a catheter in the abdominal aorta above the renal arteries as previously described for the abdominal aorta (20). Formalin-perfused kidneys were used for determining the HN score (see below).
Hydronephrosis scoring.
Formalin-fixed kidneys were weighed, cut into two halves sagittally, and examined under a dissecting microscope. The presence and severity of HN, as evaluated by the degree of dilation of the renal pelvis, was assessed for each kidney and graded on a scale ranging from 0 to 10 (see Fig. 1). Cases ranged from complete absence of HN, as in the LOU rat (score 0), through varying degrees of pelvic dilation ranging from a suspicion (score 1) to severe HN where, in extreme cases, very little renal parenchyma remained (scores 8–10). By adding the scores allotted to right and left kidneys, a global bilateral HN score was also calculated for each rat.
|
Histological studies.
After being scored for HN, various kidneys of BC rats affected with different degrees of pelvic dilation and with hematuria were embedded in paraffin and processed for routine light microscopy. Sections were stained with Masson's trichrome for general architecture and picrosirius red to demonstrate collagen. Subsequently, kidneys of several male BN rats with hematuria, detected by observation of blood either in the urine or in the renal pelvis at death, were fixed by perfusion and examined histologically.
Genetic Studies
Microsatellite markers.
One hundred ninety-three microsatellite markers that exhibited allelic variation between the parental strains and that cover the 20 autosomal chromosomes of the rat genome were selected. The primer pair information of the microsatellite markers was obtained from RGD (http://www.rgd.mcw.edu/).
Genotyping.
DNA was extracted from the liver samples by the salting out method. Fluorescence-based semiautomated genotyping was carried out with fluorescence-labeled primers. The forward primer was always labeled with the fluorescent tag. After the individual PCR amplification, products were pooled and size-fractionated by electrophoresis on ABI Prism 3730 (Applied Biosystems) DNA capillary sequencers. The allele sizes were determined by using Genotyper V2 or Genemapper V3.0 (Applied Biosystems) software. The genetic map of the 193 microsatellite markers was established with MAPMAKER/EXP V3.0 (22), a computer package that performs full multipoint linkage analysis for the markers.
Linkage analyses.
Linkage analysis was conducted with Mapmaker software (23) using all 239 BC progeny and separately for the subsets of 121 females and 118 males. The HN data were checked for normal distribution in the population as described previously (20). Since the data were not normally distributed, nonparametric linkage analysis was performed with Mapmaker/QTL V1.9 (21). The thresholds to demonstrate significant and suggestive linkage by this method were Z scores >3.9 and 2.9, respectively. The QTL was determined by calculating the genetic distance based on the drop of 1.0 logarithm of odds (LOD) unit from the peak.
Candidate gene analysis by sequencing.
Mutation screening was performed on the candidate genes in the QTLs by sequencing genomic DNA. Direct sequencing was carried out with primers designed by OLIGO 6.0. PCR products were purified by shrimp alkaline phosphatase (Promega) and exonuclease I (Promega) treatment or with a gel purification kit (Qiagen) and were directly sequenced on an ABI 3730 Sequencer (Applied Biosystems).
Quantitative real-time PCR.
The expression of two candidate genes, Id2 and Agtr1b, in the kidneys and UPJs taken from 6-wk-old male BN and LOU rats was analyzed by quantitative real-time PCR as described previously (17).
Microsatellite-assisted production of a congenic rat line.
To prove the involvement in HN of one of the chromosome 6 loci identified, a line of congenic rats was produced by introgression of the chromosomal segment containing the part of the QTL with the highest Z score (D6Rat128/D6Rat115) from the BN donor strain onto the recipient LOU genome. This was achieved by performing eight backcrosses followed by one intercross to obtain homozygosity, as previously described (16). Twenty female congenic rats, homozygous BN/BN for the three markers—D6Rat128, D6Rat103, D6Rat115—spanning the segment, aged 
18 wk, were then scored for HN as described above.
The end points of the introgressed segment were determined by high-density single nucleotide polymorphism (SNP) mapping using molecular inversion probe (MIP) technology (13) on DNA extracted from liver of five female congenic rats showing the HN phenotype. The SNP panel used contained 5,180 SNPs that were discriminative between BN and LOU and distributed throughout the genome, with an average distance between SNPs of 516 kb.
Statistics
Differences in HN and hematuria between groups were tested with the Mann-Whitney U-test or 
2-tests, as appropriate, and correlations between the various phenotypes were calculated with the Spearman test, using Statview 5 software.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
Histological studies of BC rat kidneys with different degrees of severity showed no evidence of altered renal architecture, except in rare, severe cases where the renal parenchyma was reduced to a thin shell (Fig. 1f), and there was no evidence of interstitial fibrosis, as shown by picrosirius red staining. This indicated that renal function was generally not impaired, except in the very severe cases. In accordance with this, there was no evidence for increased weight of kidneys contralateral to those with the higher unilateral HN scores (with 1 exception for a kidney scoring 10), indicating that, in general, contralateral compensatory hypertrophy had not occurred.
Linkage Analysis Results
Whole genome scans for left and right kidney HN and the bilateral score are shown in Fig. 2 and details of significant and suggestive linkage in Table 2 and Fig. 3. We have identified a region on chromosome 6 linked to bilateral HN containing two closely located, overlapping QTLs. One QTL, linked to the HN score in the whole population, is located between the markers D6Rat41 and D6Mit2, with a peak Z score of 4.4. The second QTL spans the interval between the markers D6Mit2 and D6Rat76 with a peak Z score of 4.8, in females only. The BN allele is associated with the higher HN scores at the peak markers D6Rat171 and D6Rat102, respectively (Fig. 3). Suggestive linkage was detected between both right kidney and bilateral HN scores and markers on chromosomes 2 and 4 (Table 2, Figs. 2 and 3).
|
|
|
Quantitative RT-PCR performed on kidneys and UPJ samples did not reveal any difference in the expression of either Id2 or Agtr1b between BN and LOU. Emilin1 (predicted), elastin microfibril interfacer 1, another possible candidate gene, was also present in one of the loci on chromosome 6, but sequencing of exons did not reveal any sequence variations.
Congenic Line LOU.BN.D6Rat128/D6Rat115
Details of the congenic line, with the introgressed BN segment of chromosome 6, carrying the locus linked to HN in females, are shown in Fig. 4. The position of the introgressed segment with respect to the microsatellite markers is shown on the linkage map in Fig. 4, top. Mean bilateral HN scores for female congenic rats were slightly greater than those of females of the LOU parental strain (P = 0.036, Mann-Whitney U-test, Fig. 4, middle), and the 2-test showed a greater incidence of moderate bilateral HN scores (>2–4) and a lower incidence of low bilateral scores (0–2) in the congenic females compared with LOU. This proves that one or more genes affecting the development of this form of HN are present in the introgressed chromosomal segment but suggests that they control only a small part of the phenotype.
|
Correlations Between Phenotypes
No correlation was found in the BC population between HN scores and any of the arterial phenotypes, namely, abdominal aortic IEL rupture, PDA, and the aortic elastin deficit described previously (20).
Hematuria
Examination of the kidneys of BC rats showed that a nonnegligible number (9.6%, 23/239) presented with blood in the renal pelvis (Fig. 5 a), suggesting hematuria at the time of death. Twenty-one of these rats were males (Table 3), and in all cases only the right kidney was affected. All cases of pelvic bleeding were in kidneys with HN scores 
2, and the incidence was almost doubled in BC rats with right kidney HN scores >3 (Table 3), suggesting a possible link between the two phenotypes. Although this phenotype could not readily be given a score, because it was by nature either present or absent, we nevertheless tested for linkage but found no significant locus.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
The first major observation in this study is that in the BC population HN is not linked to any of the other principal arterial phenotypes exhibited by the BN strain (IEL rupture, PDA, aortic elastin deficit) (20) and maps principally to a locus on chromosome 6, which is distinct from the QTLs identified for these arterial phenotypes. This finding suggests that each phenotype is under independent genetic control and we are not dealing with a "syndrome." This is important because in many mouse models and human cases HN exists as part of a syndrome, suggesting that mutations in some developmental gene may be involved.
The question arises as to whether all BN rats show all these phenotypes or whether their association is specific to our colony. Other groups have reported a high incidence of HN in the BN rat (32, 37), and we have observed it in BN rats of an origin different from that of those used here (M. Osborne-Pellegrin, unpublished observations). This is also the case for the above-mentioned arterial phenotypes (Refs. 3 and 6 and unpublished data). However, because we have not looked exhaustively at all available colonies of BN rats, we cannot be sure that these phenotypes represent universal hallmarks of this strain, which genetically is an outlier compared with most other laboratory rat strains (9, 38). Indeed, a recent study using SNP technology has shown that some genetic variation (<1%) does exist between various inbred BN strains (38), which could be responsible for variations in phenotype.
HN in both humans and rodents can occur unilaterally or bilaterally, and some forms can predominate in males. It is not clear what determines the sex difference and whether HN occurs bilaterally or unilaterally and, in the latter case, what determines the side preference. In the mouse, in unilateral cases the right kidney is usually more affected, and the reason may be anatomic because the right kidney is more rostral (1). This is also the case in humans, where right-sided HN is also predominant. In the rat, according to published reports, although congenital HN may be bilateral it is more often unilateral. Unilateral HN in the two above-mentioned colonies selected from Wistar rats showed no predisposition for side or sex in one colony (39) and was right-sided and predominated in males in the other (11). Although not having been compared directly, it appears that the former is a more severe form (35, 46). Previous reports on HN in the BN rat describe it as unilateral (32), whereas we found it to be more generally bilateral despite a more rostrally situated right kidney, but such a distinction may depend on the degree of scrutiny applied. In fact, although the more affected kidney showed no side predominance when BN males and females were considered together, separate analysis showed that males were more affected on the right and females on the left. BN males had globally higher scores than females. In our parental LOU population, among the occasionally affected rats only males showed higher scores on the right side. In the BC cohort, both sexes were predominantly affected on the right side, but this trend was stronger in males.
In our linkage studies, linkage to chromosome 6 was found for both right and left kidney HN scores separately but was stronger for the left score, which thus determined a greater part of linkage of the bilateral score (Fig. 2). The finding of additional suggestive loci on chromosome 2 for right kidney HN suggests that an independent mechanism is at play on this side, and, indeed, genotype-phenotype analysis suggested that LOU alleles contribute to right-sided HN. This contrasts with left-sided and bilateral HN, which were largely determined by BN alleles (Fig. 3).
Although overall male BN and BC rats had higher bilateral HN scores than females, this was mainly due to the contribution of the right kidney. Left kidney HN was not different between sexes. Our linkage studies showed that one of the overlapping chromosome 6 loci was linked to bilateral HN only in females. Because this was the locus with the highest Z score, we chose this segment for producing a congenic line. Although these rats had significantly higher HN scores than parental LOU rats, the difference was small, because the congenics presented very mild HN. No obvious side preference was observed in this small population. Clearly, further studies will be required to unravel the causes of these side and sex preferences in the development of HN. The very mild phenotype in the congenic rats shows that the chromosome 6 QTL introgressed has only a small influence on HN or that the LOU background prevents it from being expressed. Unfortunately, this will not facilitate the pursuit of this study using genotype-phenotype analysis, and the production of other congenic sublines will be necessary.
We did not perform renal functional studies in the whole BC population used for linkage, although we did do urine analysis and attempt quantification of microalbuminuria in a small group (see below). This is evidently a limitation to our study. However, we confirmed by histological examination of hydronephrotic kidneys from the BC population that this moderate form did not cause greatly altered renal architecture, except in a few extreme cases, and, more importantly, no evidence of tubulointerstitial fibrosis was seen, as described in one colony of congenitally HN rats (46). This confirms previous studies in the BN showing that renal function was maintained in the hydronephrotic kidney until quite late in life. (33). We bring additional evidence in favor of preservation of renal function in the hydronephrotic kidneys of the BC population, because we found no evidence of compensatory hypertrophy (as judged by increased weight) in kidneys contralateral to those with moderate to high HN scores. This indicated that renal function was not transferred from the hydronephrotic kidney to the opposite kidney.
Another argument in favor of a lack of functional impairment of the hydronephrotic kidney is the fact that there was no correlation in either male or female rats between IEL ruptures in the renal artery (20) and the contralateral HN score. We showed long ago (30) that renal artery IEL rupture is influenced by flow, and an artery irrigating a kidney having undergone compensatory hypertrophy presents higher numbers of IEL ruptures than a control renal artery. In our BC population, renal artery lesions were not increased contralateral to kidneys with higher HN scores, again indicating that this type of moderate HN does not impair renal function and thus does not lead to contralateral renal hypertrophy.
There was, however, one extreme case in a female BC rat with a HN score of 10 (see Fig. 1e), where renal parenchyma was reduced to a thin shell and renal function had been largely transferred to the opposite kidney. In this rat, we observed 10 ruptures of the IEL in the renal artery on the side of the functioning kidney. Since no other female in this population had more than four IEL ruptures in the two renal arteries taken together (mean value: 0.5) (20), this elegantly illustrates the fact that increased flow related to compensatory renal hypertrophy increases renal artery lesions, as already described (30).
Both the significant locus on chromosome 6 and the suggestive loci on chromosome 2 contain large numbers of genes, among which we searched for possible candidates. In a recent study, Aoki et al. (1) showed that adult inhibitor of DNA binding 2 knockout (Id2–/–) mice develop hydronephrosis with a high frequency and subsequently noted that Id2 haploinsufficiency also resulted in the same phenotype. Their results suggested that there was obstruction of the urinary flow at the UPJ in hydronephrotic kidneys of Id2 mutant mice. Id2 is a member of the helix-loop-helix family of transcription factors that promote cellular growth and inhibit differentiation (29, 44, 45). The function of Id proteins is to block the heterodimerization of the basic helix-loop-helix proteins (5). It was also shown that Angiotensin receptor II, type 1 (Agtr1b) is substantially reduced in the region around the UPJ of hydronephrotic kidneys of adult Id2 mutant mice. Many previous studies have shown that Agtr1–/– mice show morphological changes in the kidney leading to HN with urinary outflow obstruction (27, 41). In the rat, Id2 is localized on chromosome 6 and Agtr1b on chromosome 2. Since both these genes are situated within the QTLs we identified for HN in this study and so can be considered as candidates, we screened for sequence variations.
Complete sequencing of Id2 and Agtr1b genes did not reveal any mutations in the coding region, and the insertions we found in the BN strain appeared not to cause differential expression of these two genes in the UPJ between BN and LOU. We chose to do expression studies in 6-wk-old rats because one would expect a causative gene to be expressed more highly in young, growing rats than in adult rats of the age used for phenotyping. The 21-bp insertion we detected downstream to the 3' end of Id2 was found to lie in the intronic region of a predicted gene, Ac2-300, whose function is not known. The search for miRNA sequences or enhancer elements in this region did not reveal any results of interest. These insertions were found to be present only in the SBH rat, among 14 other strains of laboratory rat tested. To our knowledge, this rat strain has not been reported to have HN, and so the functional significance of these insertions in the context of this study is at present not clear. Emilin1 (predicted), another possible candidate gene present in one of the chromosome 6 HN loci, is a constituent of elastic fibers, but no mutations were found. Recently, a microarray analysis (36) was performed on rats derived from a colony with congenital HN (39). Differential expression highlighted several genes such as ion transporters, growth factors, some types of collagen, and genes involved in apoptosis, none of which is located in the genomic regions we have identified on chromosomes 6 or 2. However, although these microarray studies were performed on young rats, it is still possible that they reveal the consequences rather than the cause of the condition. Moreover, apart from Id2 and Agtr1b mentioned above, none of the other genes whose knockout or mutation in the mouse cause HN is known to be present in our identified QTLs.
In our study, because HN was not associated in the BC population with the IEL rupture phenotype or the aortic elastin deficit, we have no reason to assume that it is based on an underlying extracellular matrix anomaly. A gene involved specifically in renal development or in smooth muscle function would perhaps be a better candidate. It is noteworthy that, in the male BC population only, HN scores correlated positively with various weight-related phenotypes (body weight, aortic dry weight, and elastin content expressed as mg/cm). In addition, the suggestive loci for HN on chromosome 2 overlapped with the locus for aortic elastin content and elastin-to-collagen ratio previously described (20). The significance of such findings is not clear at the present time.
The detection in the BC population of several cases of blood in the renal pelvis remaining after perfusion with formalin was an unexpected finding. We have shown that BN rats, but not LOU rats, frequently develop papilloma-type lesions of the renal pelvic epithelium that appear to represent the cause of this bleeding. One may suppose that the lesions causing bleeding in the BC population were the same as those in BN rats. However, unfortunately, we did not keep all the fixed kidneys from the BC population after scoring, and only a selection with various HN scores were retained for histological studies. This prevented systematic screening for papilloma lesions in the whole BC population. However, we observed such lesions in some BC rats with and without presence of pelvic blood at the time of death (see Fig. 5f), suggesting that the incidence of this pathology was higher in this population than evaluated by renal pelvic blood at necropsy. The fact that no significant linkage was found for pelvic bleeding is thus not surprising in view of the fact that it probably only represents part of the population developing the pelvic epithelial lesions we describe.
Our results suggest that there was a relation between HN and hematuria, because only kidneys of BC rats with HN presented with blood in the pelvis. The same was observed in male BN rats of the parental strain, but very few BN rats were devoid of HN. However, occurrence of pelvic bleeding was not restricted to the most severely hydronephrotic kidneys. This was even more evident in the parental BN strain (Table 3). It is conceivable that, in animals susceptible to developing papilloma-like lesions, the retention of urine within the pelvis due to HN, whether moderate or severe, may initiate or aggravate such lesions. It is intriguing that only right kidneys were affected in the BC population, but because the numbers of hematuric kidneys were small this may just be a reflection of the fact that right-sided HN dominated in this population.
Such pathology has already been observed in a colony of rats of a hybrid strain of BN x Lewis rats with intermittent hematuria (40), which also suggests that BN alleles are implicated in this phenotype. HN was also observed in these rats, but the relation between the two was not discussed. Much earlier studies in the 1970s reported that BN rats are susceptible to the development of epithelial tumors of the ureter and urinary bladder, but no mention of the renal pelvis was made (7, 8). We performed renal function studies (using metabolic cages for urine collection) of the last 36 of the BC rat series used here for linkage, some of their siblings, and parental BN and LOU rats, and these revealed that microscopic hematuria was quite common in BN and BC populations and that it was intermittent and related to the presence of the above-described epithelial lesions in at least one kidney. The extent to which this pathology is genetically determined and its exact relation with HN obviously require further studies. Nevertheless, in view of our observations, it appears that measurements of microalbuminuria as an index of renal functional degradation in rats originating from crosses involving BN rats should be avoided unless one is sure that the BN parental colony is unaffected with this pathology.
In conclusion, we have shown that the BN rats from the colony studied here present HN, along with several interesting arterial phenotypes. Although the HN, at least at the ages studied, is mild and appears nonsymptomatic as far as renal function (glomerular filtration and tubular reabsorption) is concerned, it is sufficiently detectable anatomically to enable scoring for severity. Such scoring in a large population of BC rats, originating from crossing BN and LOU, has permitted the identification of two significant overlapping QTLs on chromosome 6 for bilateral HN and two suggestive loci on chromosome 2 linked to right kidney HN. Although we have not identified any candidate genes in these loci, this study, in association with other complementary studies, may contribute to elucidation of the genetic basis of this poorly defined condition. Furthermore, our observations of cases of hematuria in the BC and BN populations raise the question of a possible relationship between renal pelvic epithelial lesions and urine retention in HN.
|
GRANTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
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.
1 The online version of this article contains supplemental material. 
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. H. Wiessner, M. R. Garrett, R. J. Roman, and N. S. Mandel Dissecting the genetic basis of kidney tubule response to hyperoxaluria using chromosome substitution strains Am J Physiol Renal Physiol, August 1, 2009; 297(2): F301 - F306. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
