Shroom3 contributes to the maintenance of the glomerular filtration barrier integrity

Abstract

Genome-wide association studies (GWAS) identify regions of the genome correlated with disease risk but are restricted in their ability to identify the underlying causative mechanism(s). Thus, GWAS are useful "roadmaps" that require functional analysis to establish the genetic and mechanistic structure of a particular locus. Unfortunately, direct functional testing in humans is limited, demonstrating the need for complementary approaches. Here we used an integrated approach combining zebrafish, rat, and human data to interrogate the function of an established GWAS locus (SHROOM3) lacking prior functional support for chronic kidney disease (CKD). Congenic mapping and sequence analysis in rats suggested Shroom3 was a strong positional candidate gene. Transferring a 6.1-Mb region containing the wild-type Shroom3 gene significantly improved the kidney glomerular function in FHH (fawn-hooded hypertensive) rat. The wild-type Shroom3 allele, but not the FHH Shroom3 allele, rescued glomerular defects induced by knockdown of endogenous shroom3 in zebrafish, suggesting that the FHH Shroom3 allele is defective and likely contributes to renal injury in the FHH rat. We also show for the first time that variants disrupting the actin-binding domain of SHROOM3 may cause podocyte effacement and impairment of the glomerular filtration barrier.

Figure 1.

Introgression of the BN Shroom3 gene onto FHH background improves glomerular and overall kidney function. (A) At 14 wk of age, both homozygous and heterozygous FHH.BN14a congenic animals showed a significantly lower degree of albuminuria compared to FHH (n = 4, 7, 8, and 3, respectively). (B) Both heterozygous and homozygous FHH.BN14a demonstrated significantly improved glomerular permeability (Palb) compared to FHH. (n = 4 animals/115 glomeruli, 3 animals/76 glomeruli, and 2 animals/70 glomeruli, respectively. Palb in BN was obtained from previously published data [Rangel-Filho et al. 2005].) (C) FHH.BN14a kidney showed a decreased presence of glomerular sclerosis compared to FHH at 14 wk of age. A minimum of 30 glomeruli from three kidneys for each strain were scored for a percentage of sclerosis using a scale from 0 (no sclerosis) to 4 (complete sclerosis). (D) Representative trichrome-stained images of glomeruli from FHH and FHH.BN14a are shown. Fibrotic tissues are indicated by blue stain. Scale bars = 50 µm. (E) Electron microscopic images of glomeruli showed podocyte foot process fusion (indicated by arrow) in FHH compared to FHH.BN14a animals at 18 wk of age. Scale bars = 500 µm. (CL) Capillary lumen, (*) P < 0.05 vs. FHH, (#) P < 0.05 vs. BN.

Figure 2.

FHH Shroom3 is defective and contributes to glomerular dysfunction. (A) Schematic representation of the rat Shroom3 protein is shown. Vertical lines represent the amino acid variants found in FHH Shroom3. Variants predicted to be damaging by PolyPhen-2 are shown in red. (B) Co-injection of shroom3 + tp53 morpholino (MO) with full-length BN, but not FHH, Shroom3 mRNA rescued the edema phenotype. ([***] P < 0.001 vs. MO and MO + FHHmRNA) and (C) cell death induced by shroom3 + tp53 MO ([*] P < 0.05 vs. uninjected control). (D) Representative fluorescence images of individual dorsal aorta at 1, 24, and 48 h following 70-kDa dextran injection are shown. (E) Co-injection with BN but not FHH Shroom3 mRNA rescued the dextran leakage induced by knockdown of endogenous shroom3 in zebrafish. (n = 35, 26, 23, and 22, respectively. [*] P < 0.05 vs. uninjected and MO + BNmRNA.)

Figure 3.

The G1073S variant disrupts the function of the FHH Shroom3 gene. (A) Schematic of the different recombinant Shroom3 cDNAs, where a specific region of the BN Shroom3 sequence was replaced by FHH. (B) Co-injection of shroom3 + tp53 MO with Shroom3∆641–3044 or Shroom3∆4117–5966 mRNA, but not Shroom3∆3044–4117 mRNA, rescued dextran leakage induced by the MO (n = 23, 9, 18, 28, and 15, respectively). (C) Schematic of Shroom3 single-amino acid mutants created by site-directed mutagenesis. (D) Co-injection of ∆Y1291C or ∆A1356V restored normal glomerular permeability, while ∆G1073S failed to exhibit functional rescue (n = 17, 12, 17, 23, and 21, respectively). (*) P < 0.05, (**) P < 0.001 vs. uninjected.

Figure 4.

The G1073S variant decreases the actin-binding affinity of SHROOM3 protein. (A) FLAG-tagged BN, FHH, or ∆G1073S SHROOM3 proteins were overexpressed in HEK293 cells, followed by immunoprecipitation against FLAG. Immunoprecipitated lysates were immunoblotted using antibodies against FLAG, ROCK1, and ACTB. A representative Western blot of immunoprecipitated lysate is provided. (B) Quantification of the Western blot showed that FHH SHROOM3 and ∆G1073S mutant had significantly reduced actin-binding affinity compared to BN SHROOM3. (n = 3 per group. [*]P < 0.05 vs. BN.) (C) ROCK1-binding affinity was not different among the three alleles (n = 3 per group).

Figure 5.

rs181194611 (p.P1244L) associated with nondiabetic ESKD impairs SHROOM3 function in vivo. (A) Proline at the amino acid position 1244 in SHROOM3 is evolutionarily conserved. (B) Representative images of dorsal aorta at 1, 24, and 48 h following 70-kDa dextran injection are shown. (C) Co-injection of nonmutated human SHROOM3 mRNA restored normal glomerular permeability, while ∆P1244L failed to show functional rescue. (*) P < 0.05 vs. uninjected.

Figure 6.

Podocyte-specific disruption of shroom3 causes increased glomerular permeability and podocyte effacement in zebrafish. (A) Specific transgenic (TG) lines were crossed to obtain embryos expressing podocin:Gal4 and podocin:Gal4;UAS:shrm3DN. The control and mutants were injected with 70-kDa FITC-labeled dextran at 4.5–5 d post-fertilization (dpf) and analyzed for dextran clearance at 24 and 48 h post-injection (hpi). (DA) Dorsal aorta. (B) Representative fluorescence images of individual dorsal aorta at 1, 24, and 48 hpi are shown. (C) podocin:Gal4;UAS:shrm3DN mutants had significantly decreased FITC signal at 24 and 48 hpi compared to control, indicating that the glomerular filtration barrier was disrupted. (n = 20 and 25, respectively. [***] P < 0.001.) (D) Electron micrograph revealed intact podocyte foot processes (indicated by arrowhead) in podocin:Gal4 animals. Foot process effacement (indicated by arrow) was observed in the podocin:Gal4;UAS:shrm3DN mutants. Scale bars = 500 µm. (CL) Capillary lumen. (E) podocin:Gal4;UAS:shrm3DN had a significantly reduced number of foot processes contacting glomerular basement membrane (GBM) and increased foot process diameter compared to the control. (n = 3 per group. [**] P = 0.01, [***] P = 0.0008.)