Channel Function of Polycystin-2 in the Endoplasmic Reticulum Protects against Autosomal Dominant Polycystic Kidney Disease

Abstract

Background: Mutations of PKD2, which encodes polycystin-2, cause autosomal dominant polycystic kidney disease (ADPKD). The prevailing view is that defects in polycystin-2-mediated calcium ion influx in the primary cilia play a central role in the pathogenesis of cyst growth. However, polycystin-2 is predominantly expressed in the endoplasmic reticulum (ER) and more permeable to potassium ions than to calcium ions.

Methods: The trimeric intracellular cation (TRIC) channel TRIC-B is an ER-resident potassium channel that mediates potassium-calcium counterion exchange for inositol trisphosphate-mediated calcium ion release. Using TRIC-B as a tool, we examined the function of ER-localized polycystin-2 and its role in ADPKD pathogenesis in cultured cells, zebrafish, and mouse models.

Results: Agonist-induced ER calcium ion release was defective in cells lacking polycystin-2 and reversed by exogenous expression of TRIC-B. Vice versa, exogenous polycystin-2 reversed an ER calcium-release defect in cells lacking TRIC-B. In a zebrafish model, expression of wild-type but not nonfunctional TRIC-B suppressed polycystin-2-deficient phenotypes. Similarly, these phenotypes were suppressed by targeting the ROMK potassium channel (normally expressed on the cell surface) to the ER. In cultured cells and polycystin-2-deficient zebrafish phenotypes, polycystin-2 remained capable of reversing the ER calcium release defect even when it was not present in the cilia. Transgenic expression of Tric-b ameliorated cystogenesis in the kidneys of conditional Pkd2-inactivated mice, whereas Tric-b deletion enhanced cystogenesis in Pkd2-heterozygous kidneys.

Conclusions: Polycystin-2 in the ER appears to be critical for anticystogenesis and likely functions as a potassium ion channel to facilitate potassium-calcium counterion exchange for inositol trisphosphate-mediated calcium release. The results advance the understanding of ADPKD pathogenesis and provides proof of principle for pharmacotherapy by TRIC-B activators.

Keywords: ADPKD; endoplasmic reticulum; ion channel; polycystin.

Graphical abstract

Figure 1.

Intracellular Ca²⁺ release defect in Pkd2-null cells and its rescue by TRIC-B. (A) Representative baseline normalized ATP-induced [Ca²⁺]_i transients (above the baseline level) shown as increases in the ratio of emission fluorescence at excitation wavelength 340 nm over 380 nm (Δ [F340/F380]). Each trace is the average of multiple cells on one individual coverslip. Control, Pkd2^−/− cells, and Pkd2^−/− cells stably expressing PC2 were studied. Please note that ATP-induced peak intracellular Ca²⁺ response is not significantly different with or without 1.2 mM [Ca²⁺] in the bath (Supplemental Figure 4), supporting that the peak Ca²⁺ response reflects ER Ca²⁺ release. (B) Resting (baseline) [Ca²⁺]_i shown as F340/F380. (C) ATP-induced [Ca²⁺]_i (i.e., Δ before and after ATP stimulation) is presented as Δ(F340/F380). (C) Mean±SEM of results of peak Ca²⁺ response (peak Δ[F340/F380] values) from three experiments each with three coverslips containing multiple cells. (D) Representative baseline normalized ATP-induced [Ca²⁺]_i transients for control, Pkd2^-/-, and Pkd2^−/− cells stably expressing TRIC-B. (E) Resting (baseline) [Ca²⁺]_i for experiments in (F) shown as F340/F380. (F) ATP-induced peak [Ca²⁺]_i release presented as Δ(F340/F380) for Pkd2^−/− cells with or without TRIC-B overexpression. Mean±SEM of results from three experiments each with three coverslips containing multiple cells. (G) Western blot of TRIC-B protein in control and Pkd2^−/− renal epithelial cells shows no difference. Inset shows increased abundance of TRIC-B in Pkd2^−/− cells stably expressing recombinant TRIC-B used in (D).

Figure 2.

PC2 rescues Ca²⁺ release defect in TRIC-B^−/− cells. (A) Gel picture showing genotyping of control, heterozygous, and homozygous TRIC-B^−/− cells. TRIC-B was deleted in HEK293 cells by a frameshift deletion of 91 bp within exon 2 by CRISPR-Cas9 genome editing. (B) Western blot shows the absence of TRIC-B protein in TRIC-B^−/− cells versus control. (C) Representative baseline normalized carbachol-induced [Ca²⁺]_i transients for control, TRIC-B^-/-, and TRIC-B^−/− cells stably expressing PC2. Resting (baseline) [Ca²⁺]_i levels among three groups are not different (not shown). Carbachol was used as the agonist because HEK cells express endogenous M3 muscarinic receptors, but no purinergic receptors. Inset shows increased abundance of PC2 in TRIC-B^−/− cells stably expressing recombinant PC2. (D) Carbachol-induced [Ca²⁺]_i release for control cells, TRIC-B^−/− cells, and TRIC-B^−/− cells stably expressing PC2. Mean±SEM of results of peak Ca²⁺ response from three experiments each with three coverslips containing multiple cells. (E) Localization of expressed TRIC-B in ER revealed by endogenous ER-resident protein calnexin. HEK293 cells were transiently cotransfected with HA-tagged TRIC-B. Immunofluorescence was obtained by staining with anti-HA and anticalnexin antibodies. Imaging was obtained using cofocal microscopy. Scale bars, 10 μm.

Figure 3.

Tail curvature in zebrafish pkd2-morphants is suppressed by Tric-b. (A) Representative pkd2-morphant (MO-based pkd2 knockdown) zebrafish embryos with dorsally curved body axis at approximately 3 dpf. Embryos were segregated as mild (<90°), moderate (90–180°), and severe (>180°) on the basis of the degree of dorsal curvature. (B) Percentage of embryos segregated on the basis of severity of tail curling in groups injected with pkd2-MO along with vehicle (None) or mRNA for mouse Pkd2, Tric-a or Tric-b. Inset shows no difference between TRIC-B protein in control and pkd2-morphant (pkd2-MO) embryos at approximately 3 dpf. Data presented are combined from at least three independent experiments. The total number (n) of embryos from combined multiple experiments are shown on top. For statistical significance, relative ratio of each category of curvature is compared between groups by chi-squared test (see Methods). Note that incomplete reversal of phenotypes in pkd2-MO is typical of these types of studies in zebrafish^, and is due to (uneven) dilution of injected RNA in rapidly dividing cells. (C) Total pronephric tubular cyst areas (as marked by asterisks in Supplemental Figures 10 and 11) are measured in zebrafish morphant embryos injected with pkd2-MO along with vehicle (None, n=12) or mRNA coding for PC2 (n=7) or Tric-b (n=5). Cystic area is presented as pixel square (px²). Data are presented as mean±SEM. (D) Both Pkd2 and Tric-b mRNA significantly rescue tail curling of pkd2-morphants, whereas Tric-b-K125Q does not. Tric-b-K125Q encodes loss-of-function Tric-b mutant.

Figure 4.

Suppression of pkd2-MO zebrafish and Pkd2^−/− cell phenotypes by ER-targeted ROMK K⁺ channels. (A) Percentage of embryos with severity of tail curling in groups injected with pkd2-MO along with vehicle (None), mRNA for Pkd2, Tric-b, WT-ROMK, or ROMK-KDEL. For statistical significance, relative ratio of each category of curvature is compared between groups by chi-squared test. (B) Resting [Ca²⁺]_i as F340/F380 and (C) ATP-induced [Ca²⁺]_i release as Δ(F340/F380), respectively, for control, Pkd2^−/−, and Pkd2^−/− cells stably expressing either PC2, TRIC-B, WT-ROMK, or ROMK-KDEL. Data are mean±SEM of the results from three coverslips. (D) Representative baseline normalized ATP-induced [Ca²⁺]_i transients (shown as Δ[F340/F380]) for control, Pkd2^−/−, and Pkd2^−/− cells stably expressing ROMK-KDEL.

Figure 5.

Ciliary localization is not essential for PC2 function. (A) R6G-PC2 is absent from the primary cilium. Pkd2-null renal epithelial cells stably expressing R6G-PC2 were stained for acetylated α-tubulin and for Myc-tagged R6G-PC2 and imaged by confocal microscopy. Imaging obtained by ×63 objective. Scale bars, 20 μm. (B) ATP-induced [Ca²⁺]_i in control, PC2-null cells, and PC2-nullc stably expressing WT-PC2 or R6G-PC2 mutant presented as Δ(F340/F380). Mean±SEM of results from three coverslips of a single experiment. Four separate experiments were performed with similar results. (C) Percentage of embryos with severity of tail curling in groups injected with pkd2-MO along with vehicle (None), mRNA for Pkd2, Tric-b, Pkd2-R6G, or Pkd2L1. Data presented are from at least three independent experiments. The total number (n) of embryos from combined multiple experiments are shown on top. For statistical significance, relative ratio of each category of curvature is compared between groups by chi-squared test. Results show mRNA coding for ciliary-targeting defective PC2 mutant (Pkd2-R6G) suppresses tail curling phenotype as effective as WT-Pkd2, indicating that ciliary targeting is not required.

Figure 6.

Transgenic Tric-b expression ameliorates PKD phenotype in Pkd2 conditional knockout mice. (A) Kidney to body weight ratio of Pkd2^f/f;Pax8-LC1 (i.e., Pkd2^f/f;Pax8-rtTA;TetO-Cre), Pkd2^f/f;Pax8-LC1;cTg-Tric-b (i.e., Pkd2^f/f;Pax8-rtTA;TetO-Cre;cTg-Tric-b), and control mice (Pkd2^f/f, Pkd2^f/f;Pax8-rtTA or Pkd2^f/f;TetO-Cre). All male mice were treated with doxycycline (2 mg/ml in drinking water) from weeks 5 to 7 of age for 2 weeks, analyzed 3 months after the start of treatment. (B) Hematoxylin and eosin–stained kidney sections of the mice described above in (A). Scale bars, 100 μm. (C) Cyst index of the kidney sections. (D) BUN of the mice described above. (E) c-Myc expression in kidneys of these mice was analyzed by qRT-PCR.

Figure 7.

Synergistic effect of conditional knockout of Pkd2 and Tric-b in kidney. (A) Generation of Tric-b-floxed mice. LoxP sites were inserted into intron 2 and 3 flanking exon 3 of Tric-b gene by gene targeting. Exon 3 deletion mediated by Cre recombinase leads to loss of functional TRIC-B protein. Inset shows PCR results of various Tric-b alleles using genotyping primers as shown. F/R2 PCR does not detect wild-type Tric-b allele with this PCR protocol. (B) Tric-b^f/f;Ksp-cre mice are grossly normal. They have normal body weight at 4 weeks of age. (They live past 6 months with no apparent phenotype and are fertile; data not shown.) (C) Breeding scheme for production of Pkd2^f/f;Ksp-Cre;Tric-b^f/f and Pkd2^f/f;Ksp-Cre offspring. Multiple crossings were performed first to produce mice with linked alleles of Pkd2^f/+ and Ksp-Cre (Pkd2^f/+,Ksp-Cre). Scheme illustrates crossing between parents with Pkd2^f/+;Ksp-Cre;Tric-b^f/+, and parents with Pkd2^f/f;Tric-b^f/+ will produce pups with Pkd2^f/f;Ksp-Cre;Tric-b^f/f and Pkd2^f/f;Ksp-Cre at mendelian ratio of 12.5% each. In the far right (with predicted mendelian ratio 50%), the unfilled green line box indicates undermined floxed-Tric-b allele status (i.e., 50% with Pkd2^f/+ with or without Tric-b–floxed allele). (D) Survival curve of Pkd2^f/f;Ksp-Cre mice versus Pkd2^f/f;Ksp-Cre;Tric-b^f/f. Nine of 11 mice with homozygous deletion of Pkd2 died between postnatal days 19 and 25. For comparison, 12 of 12 mice with double deletion of Pkd2 and Tric-b died by postnatal day 24. P=0.02 double deletion of Pkd2 and Tric-b versus single deletion of Pkd2. Statistical analysis was performed by log-rank (Mantel–Cox) test.

Figure 8.

Cyst formation in Pkd2^f/+;Tric-b^f/f;Ksp-Cre mouse kidneys. (A) Kidney to body weight ratio of control (Ctrl), Pkd2^f/+;Ksp-Cre, and Pkd2^f/+;Tric-b^f/f;Ksp-Cre mice at the age of 6 months. (B) Cyst index of the kidneys of these mice on the basis of analysis of their kidney sections. (C) Representative kidney sections of Pkd2^f/+;Ksp-Cre and Pkd2^f/+;Tric-b^f/f;Ksp-Cre mice. Scale bars, 100 or 500 μm as indicated.