A 4.1-Mb Congenic Region of Rf-4 Contributes to Glomerular Permeability

Abstract

The combined transfer of two renal function quantitative trait loci (QTLs), Rf-1 (rat chromosome 1) and Rf-4 (rat chromosome 14), from the Fawn-hooded hypertensive rat onto the August Copenhagen Irish genetic background significantly increases proteinuria and demonstrates an interaction between these QTLs. Because the original Rf-4 congenic region is 61.9 Mbp, it is necessary to reduce this interval to feasibly search for variants responsible for renal susceptibility in this region. Here, we generated a minimal congenic line (Rf-1a+4_a) to identify a 4.1-Mb region of the Rf-4 QTL that significantly contributes to the severity of proteinuria in the Fawn-hooded hypertensive rat. Rf-1a+4_a animals have an increased glomerular permeability to albumin without significant changes in BP, indicating that at least one genetic element in this refined region directly affects renal function. Sequence analysis revealed no variants predicted to damage protein function, implying that regulatory elements are responsible for the Rf-4 phenotype. Multiple human studies, including recent genome-wide association studies, link the homologous human region with susceptibility to renal disease, suggesting that this congenic line is an important model for studying pathways that contribute to the progression of kidney disease.

The majority of ESRD cases are associated with diabetes, hypertension, or both; however, numerous studies in both humans and animal models suggest that renal disease susceptibility genes exist that are independent of the initiating factor.^1–6 The dissection of quantitative traits for CKD in humans remains challenging due to heterogeneity and environmental variability.^7,8 Consequently, there is a need to pursue other strategies for identifying genes and their associated pathways that are driving CKD. One solution is to use congenic rat models to investigate regions of the rat genome that are responsible for renal impairment.

One of the first examples of genetic dissection of renal impairment in rats was demonstrated in F2 crosses between the renal disease–susceptible Fawn-hooded hypertensive (FHH) rat and the renal disease–resistant August Copenhagen Irish (ACI) rat.^9,10 These studies led to the identification of five quantitative trait loci (QTLs) linked to the severity of proteinuria called Renal Failure-1 through -5 (Rf-1 through -5).^9,10 Gene–gene interactions were found between the various Rf QTLs, because the presence of homozygous FHH alleles in multiple Rf QTLs resulted in a synergistic increase in proteinuria severity.¹⁰ An interaction was specifically identified between Rf-1 and Rf-4, located on rat chromosome 1 and 14, respectively. Van Dijk et al. generated single (Rf-1a and Rf-4) and double (Rf-1a+4) congenic animals, and found that the Rf-1a and Rf-4 single congenic animals did not show increased proteinuria compared with the ACI control strain. Only the transfer of both Rf-1a and Rf-4, as in the Rf-1a+4 double congenic strain, conferred a significant increase in proteinuria.¹¹

The original Rf-4 region consisted of 61.9 Mb containing 499 known and predicted genes. To narrow the region of interest to begin a meaningful search for the causal variant, it is necessary to physically reduce the candidate region in turn reducing the number of candidate genes. In this study, we use congenic mapping to generate a minimal congenic line called Rf-1a+4_a, which carries only 4.1 Mbp of FHH genome in the Rf-4 region that we show significantly contributes to proteinuria. Van Dijk et al. previously demonstrated that genes in the Rf-1 region increase glomerular capillary pressure (P_GC) by impairing renal blood flow autoregulation.¹¹ To explain the interaction between Rf-4 and Rf-1, they hypothesized that the Rf-4 region affects integrity of the glomerular filtration barrier that is manifested when exposed to an increase in P_GC (i.e., Rf-1).¹² Here, we assessed glomerular permeability to albumin (P_alb) in the Rf-1a+4_a congenics to address this hypothesis, and found that P_alb was significantly higher in Rf-1a+4_a compared with control strains.

The refined interval is only 6.6% of the original Rf-4 congenic region containing just 67 known and predicted genes. To initiate the search for causative variants, we analyzed the genomic sequence of the entire congenic region. Within the coding sequence, we found only one benign nonsynonymous amino acid variant between ACI and FHH in the entire Rf-4_a region, suggesting that an intergenic, intronic, or untranslated variant(s) is likely responsible for the Rf-4_a renal phenotype. It has been demonstrated that conserved sequences suggest functionality,^13–15 so we prioritized noncoding variants based on evolutionary conservation. Using a congenic model and comparative genomic approach, we have reduced the candidates for the Rf-4 QTL to a handful of sequence variants. The results of this study are of particular interest because previous studies have indicated that the homologous region on human chromosome 4 is associated with various forms of CKD as indicated by both linkage¹⁶ and genome-wide association studies (GWASs).^17,18

Results

In Vivo Phenotyping of the Rf-1a+4_a Congenic Strain

To narrow the Rf-4 region, a series of subcongenic lines were created by a backcross and intercross approach (Supplemental Figure 1A). Subcongenic lines were screened for the development of albumin excretion (UAV) and we found that animals carrying FHH alleles from markers D14Rat78 to D14Hmgc18 demonstrated higher levels of UAV compared with other lines (Supplemental Figure 1B). To further investigate this candidate region, we generated a refined congenic line called Rf-1a+4_a (Figure 1). Rf-1a+4_a congenic animals are genetically identical to the Rf-1a congenics, with the exception of the Rf-4 region in which 4.1 Mb of FHH genome has been integrated onto rat chromosome 14 from ss262968744 (15.16 Mb) to D14Hmgc18 (19.3 Mb), a region containing 67 known and predicted genes (Supplemental Table 1).

Figure 1.

Schematic representation of the Rf-1a, Rf-1a+4, and Rf-1a+4_a genotypes on rat chromosomes 1 (RNO 1) and 14 (RNO 14) and the homologous region in the human. Black indicates FHH genotype and white indicates ACI genotype. The flanking short sequence length polymorphism and SNP markers for the Rf-1a, Rf-4 and Rf-4_a congenic regions are shown to the left of the rat chromosome, and the position in Mbp of each marker is shown to the right. The flanking markers for the Rf-4_a region are shown in bold. The homologous region in the human is located on chromosome 4 from approximately 74.3 to 78 Mb, and the orientation of this region in human is inverted compared with that of the rat. (¥) indicates human QTL for GFR,¹⁶ and (∆) and (§) indicate GWAS loci for GFR¹⁷ and serum magnesium,¹⁸ respectively.

As we have studied the congenics on a resistant genome background, we used a series of physiologic tools to help drive disease progression. Unilateral nephrectomy was performed on male rats at 5 weeks of age, and animals were placed on a purified AIN-76A rodent diet containing 0.4% NaCl and L-NAME–containing water after surgery. Van Dijk et al. observed the greatest differences between Rf-1a and Rf-1a+4 animals using this protocol.¹¹ Furthermore, Rf-4 was initially identified using an L-NAME and unilateral nephrectomy protocol to accelerate the renal disease from 9 months to <4 months.¹⁰ Animals receiving this treatment were phenotyped for UAV at 13 weeks. The Rf-1a+4_a minimal congenic animals excreted significantly higher levels of UAV (68.66±9.44 mg/d), compared with Rf-1a (30.95±6.07 mg/d; P=0.002) and compared with ACI animals (24.60±6.67 mg/d; P=0.001) at 13 weeks of age (Figure 2A).

Figure 2.

Rf-1a+4_a animals excrete higher levels of albumin compared with ACI and Rf-1a, but BP is not different between strains. (A) The susceptibility loci are on a resistant genome background (ACI), and therefore animals phenotyped for albumin excretion were unilaterally nephrectomized at 5 weeks of age and hypertension was induced by L-NAME diluted in the drinking water. Animals were placed into metabolic cages 8 weeks after unilateral nephrectomy, and 24-hour samples were analyzed for albumin excretion. Albuminuria is significantly elevated in Rf-1a+4_a compared with Rf-1 and ACI. n=7, n=10, and n=10 animals for ACI, Rf-1, and Rf-1a+4_a, respectively. **P≤0.01. ***P≤0.001. (B) MAP is not different in Rf-1a+4_a animals compared with ACI and Rf-1a. BP was recorded in 14-week-old awake animals by radiotelemetry after the metabolic cage experiment. n=9, n=12, and n=10 animals for ACI, Rf-1a, and Rf-1a+4_a, respectively.

We measured BP in all strains to ensure that L-NAME treatment increased BP to a similar level in all three strains. We found no significant differences in BP between strains (Figure 2B).

In Vitro Glomerular Permeability to Albumin

At 90 seconds after bath exchange from 6% to 4% BSA, the fluorescent signal of the glomeruli fell to a greater percentage in Rf-1a and ACI animals relative to that seen in Rf-1a+4_a glomeruli. The distribution of Rf-1a+4_a glomerular fluorescence percentage baseline was significantly shifted to the right compared with both ACI (P<0.001) and Rf-1a glomeruli (P<0.001) indicating increased permeability to albumin in the Rf-1a+4_a glomeruli (Figure 3A). The albumin reflection coefficient (σalb) was reduced due to an increase in glomerular permeability to albumin (P_alb) as calculated by 1−σalb (σalb = actual change/expected [33%]). Average P_alb per animal was significantly higher in Rf-1a+4_a (0.467±0.0258) compared with ACI (0.349±0.0265; P=0.009) and compared with Rf-1a (0.373±0.0271; P=0.022) (Figure 3B).

Figure 3.

Rf-1a+4_a animals have increased glomerular permeability to albumin, and a higher percentage of glomerular sclerosis compared with ACI and Rf-1a. (A) Animals were infused with high molecular mass (250 kD) FITC-labeled dextran 3 minutes before sacrifice to allow perfusion of glomerular capillaries. Glomeruli were isolated in 6% BSA and total fluorescence was measured. Media were switched to 4% BSA and the change in fluorescence over time was measured. Distribution of glomerular swelling, as determined by decrease in fluorescence at 90 seconds after bath exchange from 6% to 4% BSA compared with baseline fluorescence of Rf-1a+4_a (green line) glomeruli is shifted to the right compared with Rf-1 (red line) and ACI (black line), indicating a significant increase in glomerular permeability. # and § indicate P<0.001 versus ACI and Rf-1a, respectively. n=117, n=144, and n=158 glomeruli measured for each group for ACI, Rf-1a, and Rf-1a+4_a respectively. (B) Glomerular permeability per animal was higher on average for Rf-1a+4_a compared with ACI and Rf-1a. P_alb (1 − σalb, where σalb = actual change/expected [33%]) was calculated for each glomeruli, and the mean of means for each group is represented. n=6, n=8, and n=6 animals for ACI, Rf-1a, and Rf-1a+4_a, respectively. (C) Rf-1a+4_a kidneys have increased presence of glomerular sclerosis compared with ACI and Rf-1a at 15 weeks of age. Thirty glomeruli from four kidneys for each strain were scored for percentage of glomerular sclerosis using a scoring scale from 0 (no sclerosis) to 4 (complete sclerosis). (D) ACI and Rf-1a kidneys present little to no glomerular sclerosis and interstitial fibrosis (panels 1 and 2), whereas Rf-1a+4_a kidneys showed an increased abundance of glomerular sclerosis (+) as well as interstitial fibrosis (→) (panel 3). Bar indicates 50 µM. *P≤0.05. **P≤0.01.

Histologic Analyses

The degree of glomerular sclerosis and basement membrane thickening was determined from histologic evaluation of kidney sections stained by Gomori’s one-step trichrome staining. Rf-1a+4_a animals demonstrated a significant increase in glomerular sclerosis (0.94±0.10) compared with both the Rf-1a (0.63±0.06; P=0.010) and ACI (0.63±0.02; P=0.011) animals as assessed by glomerular sclerosis score (Figure 3C). Histologic analysis revealed increased glomerular, as well as interstitial, fibrosis in Rf-1a+4_a kidneys compared with Rf-1a and ACI kidneys (Figure 3D).

Human Syntenic Region Analyses

The Rf-4_a congenic region in the rat is homologous to an approximately 3.7-Mb region on human chromosome 4. The orientation of this region is inversed in the rat compared with the human, indicating a chromosomal rearrangement between the human and rodent.

A Framingham study by Fox et al. mapped QTLs for both GFR and creatinine clearance to human chromosome 4, within the human homologous region to the Rf-4 QTL.¹⁶ The peak of this QTL was nearest to marker D4S2367, and the estimated span overlaps the human genome homologous to the Rf-4 minimal congenic region. Furthermore, a 2009 GWAS identified a locus significantly associated with GFR estimated by serum creatinine (eGFRcreat) on human chromosome 4.¹⁷ The most significant single nucleotide polymorphism (SNP) in this region, rs17319721, is located within an intron of the SHROOM3 gene, which is 1 of the 67 candidate genes in the Rf-4 minimal congenic region. SHROOM3 was also identified in a separate GWAS as being linked to serum magnesium concentration and kidney function as assessed by eGFR,¹⁸ making this an appealing renal function candidate gene. These human studies suggest that allelic variants in the region syntenic to the Rf-4 minimal congenic region could be causing kidney disease in both the human and rat. Overlap with these human loci is an important validation in the use of the rat to identify genetic variants that may have relevance to renal disease in humans.

Rf-4_a Sequence Annotation

After alignment of reads to the reference sequence of the rat genome, 89.1% and 86.1% of nongap reference bases were covered by ≥3 reads for the FHH and ACI, respectively. We identified 3257 sequence variants between ACI and FHH. Of these SNPs, four were found in exons, one of which resulted in an amino acid change in the albumin gene (I431V), which was considered to be a benign amino acid substitution according to the PolyPhen prediction algorithm (http://genetics.bwh.harvard.edu/pph/), reducing the likelihood that this variant is functionally detrimental to albumin function. None of the exon variants seemed to be potentially damaging, and we found no variant that might cause mis-splicing of the exon. To prioritize noncoding sequence variants, we assessed species conservation to help predict functionality of variants within intronic or intergenic regions. We identified 15 variants with a high conservation score (>0.75) based on the phasCons algorithm (Table 1).¹⁴

Table 1.

Variants between ACI and FHH in the Rf-4_a region that are highly conserved between species (conservation score >0.75)

Position	Start	Stop	ACI	FHH	Conservation Score
Intergenic	15,168,618	15,168,618	C	G	0.96
Intergenic	15,168,719	15,168,719	C	G	0.99
Intergenic	15,168,744	15,168,744	C	A	0.99
Intergenic	15,228,978	15,228,978	A	C	1.00
Intergenic	15,240,810	15,240,810	G	C	0.89
Intergenic	15,247,380	15,247,380	TA	T	0.85
Intergenic	16,411,415	16,411,415	AT	A	0.97
Intergenic	17,765,698	17,765,698	A	G	0.91
Intergenic	17,816,838	17,816,838	T	A	0.92
Intergenic	17,821,228	17,821,228	C	T	1.00
Intergenic	17,835,462	17,835,462	G	T	0.96
Intergenic	17,883,002	17,883,002	A	T	0.85
Intergenic	18,734,287	18,734,287	A	G	1.00
Exon (Afp: D282D)	19,101,441	19,101,441	A	G	0.99
Intergenic	19,247,351	19,247,351	C	T	1.00

Start and stop are the base positions of the variant on rat chromosome 14. The position column indicates if the variant is intergenic, intronic, or exonic (followed by the gene symbol and amino acid position if applicable).

Only one highly conserved intergenic variant found in the FHH strain, at position 17,821,228 of chromosome 14, was completely conserved between eight of the nine annotated species (there was no alignment for cow at this position), including ACI and Brown Norway. This suggests that this conserved allele is under evolutionary selection and, variation may be important to physiologic function. This highly conserved variant was predicted by TFSearch (http://www.cbrc.jp/research/db/TFSEARCH.html) to cause a loss of Nrf2 transcription factor binding site in FHH. To test the validity of this prediction, we performed electrophoretic mobility shift assay (EMSA) with nuclear proteins isolated from Brown Norway rat kidney and respective oligonucleotide probes from ACI and FHH. We found that nuclear protein binding to the ACI sequence is significantly higher compared with the FHH sequence in vitro (P<0.001). Furthermore, supershift assays using Nrf2 antibody supported that renal Nrf2 can bind to the ACI sequence, but the FHH variant nucleotide adversely affects binding of Nrf2 to this region (Supplemental Figure 2). These data demonstrate that intergenic variants can affect transcription factor binding in Rf-1a+4_a animals; therefore, transcriptional regulation of gene expression could contribute to the Rf-4_a renal phenotype.

Kidney RNA from 6-week-old animals was analyzed for expression of genes up- or downstream of the Nrf2 binding site variant as well as known genes in the entire Rf-4_a region. Fold change of Rf-1a and Rf-1a+4_a gene expression was compared with ACI (Supplemental Table 2). We detected expression of all known genes except for Ereg and Epgn in the kidney, but found no genes within the congenic region to be significantly differentially expressed between strains.

Discussion

Of the five Rf- QTLs mapped in crosses between ACI and FHH, our group has previously refined three of them (Rf-1, Rf-2, and Rf-3)^19,20 using a congenic strategy, and we have successfully identified causative genes underlying Rf-1 (J. Lazar and H. J. Jacob, 2011, unpublished data) and Rf-2.²⁰ In this study, we utilized a similar approach to greatly refine the Rf-4 QTL as well. We reduced this region to a small enough interval in which we can initiate a feasible search for causative variants by sequence annotation.

Combining the Rf-1 and Rf-4 QTLs in the double congenic model was crucial for successfully refining the Rf-4 region. Previous studies showed that Rf-4 contributes to the development of proteinuria only in the presence of Rf-1a.¹¹ The Rf-1a+4 double congenic rat allowed us reduce the size of the critical interval of Rf-4 from 61.9 Mb to just 4.1 Mbp. This interval, called Rf-4_a, is just 6.6% of the original Rf-4 region, reducing the candidate gene list to just 67 known and predicted genes. Indeed, Rf-1a+4_a congenic animals excreted significantly higher levels of albumin than Rf-1a congenics, and this increased albuminuria was associated with an increase in glomerular permeability to albumin, indicating damage to the glomerular filtration barrier.

The long-term goals of this renal failure (“Rf”) project have been to identify the underlying causal variants and to study the nature of the Rf QTL interactions. It is known that albuminuria can occur as a result of changes in renal hemodynamics and resultant elevations in glomerular capillary pressure (P_GC), leading to increases in GFR and the filtered load of protein, changes in the glomerular filtration barrier to proteins, and/or altered protein trafficking and reuptake of filtered proteins in the proximal tubule.^21,22 Van Dijk et al. proposed that the Rf-4 region affects the integrity of the glomerular filtration barrier that is manifested when exposed to an increase in P_GC (Rf-1), particularly in the presence of stressors such as L-NAME and reduced renal mass.¹¹ To address this hypothesis, we measured P_alb in glomeruli isolated from 12- to 13-week-old rats, before significant development of proteinuria, and found that indeed the in vitro permeability to albumin was significantly higher in Rf-1a+4_a compared with Rf-1a and ACI, supporting the hypothesis that an insult at the glomerulus in combination with changes in P_GC are required for the presence of measureable proteinuria in this model. This is the first instance in which we have identified a mechanism of renal impairment in two separate Rf QTL, and demonstrated that the interaction between these two mechanisms is required to produce a proteinuria phenotype.

We found greater heterogeneity in the measurement of in vitro glomerular permeability in Rf-1a+4_a compared with Rf-1a and ACI. Increased heterogeneity would be expected given the focal nature of segmental glomerular sclerosis in FHH rats.^23–25 That is, the barrier function of all glomeruli is not equally affected; only isolated regions of a kidney display impaired (eventually sclerosed) glomeruli, whereas glomeruli in other regions of the kidney appear healthy. Because of this heterogeneity, it was necessary to assess a larger population of glomeruli from each strain to improve the sensitivity of the assay. By using change in fluorescence as a high throughput measurement of change in volume (ΔV), we were able to assess on average approximately 21 glomeruli per animal, about four times the number of glomeruli assessed using the traditional in vitro P_alb methods.²⁰ Mean arterial pressure (MAP) is not elevated in the Rf-1a+4_a congenics, further supporting the hypothesis that genes in this region are directly affecting the kidney, and the observed UAV is not secondary to hypertension.

The Rf-4_a syntenic region has been associated with multiple renal function loci in humans. Two separate GWASs have linked the SHROOM3 gene to renal function,^17,18 indicating that certain human alleles in, or around, this gene confer susceptibility to renal function. Shroom3 encodes an F-actin binding protein that has been shown to play a role in epithelium-like cell motility.²⁶ Actin binding is known to be important in maintaining the glomerular filtration barrier integrity,^27,28 making Shroom3 an interesting candidate for glomerular permeability in the Rf-4_a region. However, we did not find amino acid changes between ACI and FHH, suggesting that this gene is not segregating with the phenotypic differences between ACI and FHH mapped to this QTL.

We found only four coding variants within the entire Rf-4_a region, and only one of these resulted in a nonsynonymous, benign amino acid change. Few of these coding variants were conserved among species, and none were predicted to be damaging to protein function. Therefore, the Rf-4_a causative variant(s) is likely not coding, but rather an intergenic or intronic regulatory element. Some GWASs have shown significant associations of intergenic regions with disease,^13,29,30 indicating that nongenic sequences can also influence a complex disease. Noncoding sequence accounts for a vast majority of sequence variants found in mammalian genomes; however, determining functionality of these sequences remains a challenge.¹⁴ Comparative sequence analysis has been one strategy to identify nongenic sequences that have regulatory function.³¹ On the basis of prioritizing intergenic variants by conservation, we preliminarily narrowed the list of likely causative variants to just 15 in the Rf-4_a region.

The Rf-4_a region carries a variant, which is otherwise most highly conserved, and is predicted to cause the loss of an Nrf2 transcription factor binding site. Activation of Nrf2, a master regulator of antioxidant molecules, plays a protective role against renal disease progression.^32,33 Specifically, Nrf2-deficient mice demonstrate increased glomerular sclerosis in response to hyperglycemia,³⁴ whereas elevated Nrf2 levels by using a pharmacological agent inducer improve renal function in mice.³⁵ Thus, decreased binding affinity of Nrf2 to the regulatory region of yet to be identified gene(s) in Rf-1a+4_a animals could adversely influence the antioxidant response and thereby contribute to increased glomerular damage. Although we found no gene expression differences in kidneys of young (6-week-old) animals, the role of Nrf2, either alone or in association with other regulatory elements, could be influencing gene expression only after the induction of oxidative stress by unilateral nephrectomy and L-NAME treatments, because these stressors are required for the manifestation of the Rf-4 phenotype. Further expression analysis at different ages and stages of proteinuria could demonstrate differences among our strains. Additional studies of transcription factor activity in our congenic strains are also required to resolve the precise mechanism of renal damage in the Rf-4_a interval.

Many GWAS SNPs are noncoding variants, and therefore few GWAS have pinpointed the causative element underlying loci. Rf-1a+4_a may serve as a useful model to provide insight into the characterization of noncoding sequence variants that affect a complex disease, and these sequences can be compared across species to help identify causative variation in human forms of renal disease. In this study, we characterized a mechanism of proteinuria in the Rf-4_a congenic region. This 4.1-Mb interval is homologous to human loci and QTL that have been implicated in renal function, supporting the need to further investigate this region of the genome. It is likely that a variant(s) in noncoding sequence is responsible for the Rf-4_a phenotype, making the Rf-4_a a useful model for studying the physiologic effect of nongenic sequences in renal disease. Additional studies will be required to prove the specific variant(s) responsible for increased glomerular permeability and the molecular mechanisms causing the renal impairment phenotype in the Rf-1a+4_a animals.

Concise Methods

Animal Care

The rats were housed in the Biomedical Resource Center of the Medical College of Wisconsin, an American Association for the Accreditation of Laboratory Animal Care–approved facility. The local Animal Care and Use Committee approved all protocols used in these studies.

Generation of the Rf-1a+4_a Strain

The congenic breeding strategy is outlined in Supplemental Figure 1A. The development of the Rf-1a+4_a minimal congenic strain was initiated by crossing Rf-1a (ACI.FHH-[D1Rat74-D1Rat90]) single congenic males to Rf-1a+4 (ACI.FHH-[D1Mit18-D1Rat90]/[D14Mit11-D14Rat33/D14Rat65-D14Rat90]) double congenic females. Fine map genotyping in our hands revealed that the Rf-1a congenic region is identical in both Rf-1a and Rf-1a+4 strains, spanning from D1Mit18 to D1Rat90; therefore, F1 generation animals were heterozygous for the Rf-4 QTL but remained FHH homozygous in the Rf-1a region. F1 animals were intercrossed and F2 generation offspring were genotyped using a previously described fluorescent genotyping protocol to identify animals with a desirable recombination within the Rf-4 region.³⁶ Animals that had inherited a recombination of interest were then backcrossed to the Rf-1a congenic line, and offspring carrying the Rf-4 recombined congenic interval were intercrossed to fix the recombinant region to homozygosity. Subcongenic animals were phenotyped for UAV to map the genetic region contributing to the renal impairment phenotype, and we found that FHH genotype in the region between approximately D14Rat78 (17.6 Mbp) and D14Hmgc18 (19.3 Mbp) caused increased UAV (Supplemental Figure 1B). To investigate this small region of the genome, we generated a minimal congenic line called Rf-1a+4_a (ACI.FHH-[D1Mit18-D1Rat90]/[D14Rat98-D14Hmgc18]/Mcwi) (RGD ID 4145374). We utilized SNP genotyping to fine map the upper boundary of Rf-4_a, and found that this region extends to ss262968744, located at 15.16 Mbp.

Phenotyping for Albuminuria and BP

Male rats aged 5–6 weeks were anesthetized with a cocktail of ketamine (30 mg/kg), xylazine (2.5 mg/kg), and acepromazine (0.6 mg/kg). The right kidney was exposed by a retroperitoneal incision, the renal artery and vein were ligated, and the kidney was removed. After surgery, 150 mg/L of L-NAME (Sigma-Aldrich, St. Louis, MO) was added to the rats’ drinking water and their chow switched from Laboratory Rodent Diet 5001 (PMI Nutrition International Inc, Brentwood, MO) to a purified AIN-76A rodent diet containing 0.4% NaCl (Dyets Inc, Bethlehem, PA). This chow and L-NAME dose were maintained ad libitum throughout the remainder of the protocol. Eight weeks after unilateral nephrectomy, the rats were housed in metabolic cages (Nalgene, Rochester, NY) for urine collection. Rats were acclimated to metabolic cages for 2 days, and urine was then collected for two consecutive 24-hour periods and albumin concentration was determined using the Albumin Blue 580 assay (Molecular Probes, Eugene, Oregon).^37,38

BP was measured 9 weeks after unilateral nephrectomy, immediately after the final urine collection. MAP was measured in awake rats by radiotelemetry (Data Sciences Inc, St. Paul, MN) as described previously.³⁹ Telemetry transmitters (TA11PA-C40) were implanted subcutaneously (under isoflurane anesthesia), and the catheter was inserted into the abdominal aorta via the femoral artery. Animals were allowed 4 days for recovery after surgery, and BP was then recorded at 500 Hz in conscious, freely moving animals for 3 consecutive days. Ten-second intervals were continuously recorded every two minutes, and these data were averaged over a 3-hour period each day to estimate MAP.

In Vitro Fluorescent Glomerular Permeability Assay

Male animals aged 12–13 weeks were anesthetized with isofluorine and infused with a bolus of high molecular mass (250 kD) FITC-labeled dextran (Sigma-Aldrich) dissolved in 0.7 mL of saline at a dose of 75 mg/kg body wt via femoral vein catheterization. After 3 minutes of equilibration, animals were sacrificed, both kidneys were removed, and the glomeruli were isolated using a differential sieving technique as described previously.⁴⁰ The in vitro glomerular permeability assay was adapted from a previously described method.^20,40 Savin et al. determined ΔV by measuring glomerular diameter, whereas we directly assessed ΔV by dilution of a fluorescent volume marker (250 kD-FITC dextran). The total fluorescent intensity of each glomerulus was measured and recorded in real time (1 observation/sec) using the Incyte1 program. After baseline recording, the perfusion media were switched from 6% BSA to 4% BSA to expose glomeruli to a hypo-oncotic environment. The fluorescent level 90 seconds after bath exchange to the 4% BSA was used to calculate the albumin reflection coefficient (σalb) and P_alb. The measured change in volume/expected change (33%) determined the σalb, and P_alb was calculated by 1 − σalb.

Glomerular Histology

After the metabolic cage experiment, 14-week-old unilaterally nephrectomized animals were sacrificed and the left kidney was removed and immediately placed in 10% buffered formalin (Sigma-Aldrich) for fixation. Fixed kidneys were sectioned and stained using Gomori’s one-step trichrome stain for histologic analysis. Individual glomeruli were scored on a scale of a 0 through 4, with 0 indicating no damage, 2 representing loss of 50% of glomerular capillary area, and 4 indicating complete loss of the glomerular filtration area.

Congenic Region Sequence Analyses

ACI and FHH genomic DNA was sequenced on an Illumina HiSEquation 2000 according to the manufacturer’s instructions (Illumina, San Diego, CA). We used Illumina's CASAVA software to align the paired-end reads to the reference genome and identify variants. Variants with a read depth >3 and occurring in >50% of the reads were annotated with ANNOVAR software.⁴¹ Sequence comparisons were made between the ACI and FHH within the minimal congenic region.

EMSA

Nuclear protein preparation and EMSA were performed as described previously.⁴² Briefly, kidney was isolated from male Brown Norway rat, homogenized in PBS, and protein extracts were prepared in the absence of EDTA and quantified. Expression of Nrf2 in the nuclear fraction was confirmed by Western blot analysis (data not shown). For DNA probe, high-performance liquid chromatography purified oligonucleotides with forward and reverse sequences of Rf-4_a variant region of FHH and the respective region of ACI were synthesized by Integrated DNA Technologies (IDT, Coralville, IA). The ACI oligonucleotide had the forward sequence of 5′-TGGGTGACTTTGTAGACTCTTCCGGTTTTCCGTGGTA-3′, whereas C at position 19 was replaced with a T in the FHH. Respective forward and reverse oligonucleotides were annealed and the double-stranded DNA was labeled with DIG-11-ddUTP using a recombinant Terminal Transferase (20 U/μl) (DIG Gel Shift Kit, Second Generation; Roche, Indianapolis, IN). Binding assays were performed with 5 μg of nuclear proteins and 0.08 pmol of each DIG-labeled probe. For the supershift assay, binding was performed with nuclear proteins incubated with 4 μg of Nrf2 (C-20, sc-722 X) antibody (Santa Cruz Biotechnology, Santa Cruz, CA). After gel electrophoresis and blotting onto membrane, chemiluminescence detection and x-ray film exposure were performed, and band intensities were measured with ImageJ software (National Institutes of Health).

RNA Extraction and Quantitative PCR

Six-week-old animals were sacrificed and whole kidneys were removed and placed into RNAlater (Ambion Life Technologies, Grand Island, NY). Total RNA was extracted from whole kidneys with Trizol reagent (Invitrogen) and cDNA was synthesized using Superscript III reverse transcription (Invitrogen) according to the manufacturer’s instructions. Quantitative PCR was performed by using the QuantiTect SYBR Green PCR kit (Qiagen, Valencia, CA) and the Rotorgene 3000. Cycling was performed at 95°C for 15 minutes, 94°C for 20 seconds, 60°C for 20 seconds, 70°C for 20 seconds for 40 cycles, and 70°C for 5 minutes. mRNA expression was normalized to β-actin as an endogenous control, and relative expression was calculated by 2^(-ΔΔCT) method versus ACI.

Statistical Analyses

Data are presented as means ± SEM. Albuminuria, glomerular index score, and average P_alb data were analyzed using ANOVA, followed by a Holm–Sidak multiple comparison test using Sigma Plot 11.0 software. The BP data set was not normally distributed and was analyzed by a Kruskal–Wallis one-way ANOVA on ranks. Difference in distribution of fluorescent dilution of FITC dextran between the three strains was analyzed using standard SAS statistical software package (version 9.2; SAS Institute, Cary, NC). Kolmogorov–Smirnov nonparametric tests were used for normality tests and two sample comparisons.

Disclosures

None.