mtor Haploinsufficiency Ameliorates Renal Cysts and Cilia Abnormality in Adult Zebrafish tmem67 Mutants

Abstract

Background 

Although zebrafish embryos have been used to study ciliogenesis and model polycystic kidney disease (PKD), adult zebrafish remain unexplored.

Methods 

Transcription activator-like effector nucleases (TALEN) technology was used to generate mutant for tmem67, the homolog of the mammalian causative gene for Meckel syndrome type 3 (MKS3). Classic 2D and optical-clearing 3D imaging of an isolated adult zebrafish kidney were used to examine cystic and ciliary phenotypes. A hypomorphic mtor strain or rapamycin was used to inhibit mTOR activity.

Results 

Adult tmem67 zebrafish developed progressive mesonephric cysts that share conserved features of mammalian cystogenesis, including a switch of cyst origin with age and an increase in proliferation of cyst-lining epithelial cells. The mutants had shorter and fewer distal single cilia and greater numbers of multiciliated cells (MCCs). Absence of a single cilium preceded cystogenesis, and expansion of MCCs occurred after pronephric cyst formation and was inversely correlated with the severity of renal cysts in young adult zebrafish, suggesting a primary defect and an adaptive action, respectively. Finally, the mutants exhibited hyperactive mTOR signaling. mTOR inhibition ameliorated renal cysts in both the embryonic and adult zebrafish models; however, it only rescued ciliary abnormalities in the adult mutants.

Conclusions 

Adult zebrafish tmem67 mutants offer a new vertebrate model for renal cystic diseases, in which cilia morphology can be analyzed at a single-nephron resolution and mTOR inhibition proves to be a candidate therapeutic strategy.

TMEM67 is a major causative gene for Meckel syndrome (MKS; the gene was formerly named MKS3), which accounts for approximately 15% of MKS cases; it is also a causative gene for Joubert syndrome.^1–4 MKS is an autosomal recessive and perinatal lethal disorder that is characterized by a spectrum of abnormalities, including renal cysts in >95% of patients, central-nervous-system defects such as encephalocele, liver fibrosis, and sometimes polydactyly.^5–7 Spontaneous animal models have been used to study MKS3, including the wpk rat carrying a naturally occurring, single-point mutation in the Tmem67 gene; the bpck mouse with the causative Tmem67 gene encompassed in a large 245-kb deletion; and even a sheep model.^8–10 The murine models capture some characteristics of patients with MKS3, such as polycystic kidneys and hydrocephalus, but not encephalocele, biliary abnormalities, or polydactyly; whereas the sheep model has the characteristic hepatorenal disease.^8–12 A targeted knockout mouse (Tmem67^tm1[Dgen/H]) has also been generated that exhibits pulmonary hypoplasia, cardiac malformation, kidney cysts, and some encephalocele, and that dies by postnatal day 1 due to pulmonary and/or cardiac defects.^13–15 Of note, an effective therapy for MKS has not yet been reported.

TMEM67 encodes a transmembrane protein (meckelin) that is localized in the ciliary transition zone of the renal epithelium.^1,13,16 However, a range of ciliary phenotypes has been noted among different MKS3 models. For instance, bpck mice, wpk rats, and human MKS3 fetuses have elongated renal cilia; murine Tmem67 mutants exhibit shorter and fewer cilia; and ovine TMEM67 mutants display both very long and very short cilia in their kidneys.^9–11,13 Moreover, compared with those from wild-type mice, mouse embryonic fibroblasts derived from Tmem67-knockout mice present with longer, normal, or no cilia; similar findings have been obtained with cells expressing Tmem67 short hairpin RNA.^11,13,14,17 Therefore, additional studies with advanced imaging methods are needed to define ciliary defects and clarify the relationship between ciliary defects and cyst development.

Embryonic zebrafish have long been used to study cystogenesis because of their transparency and the efficiency of their genetic manipulation.^18–21 Embryonic zebrafish have also proven to be excellent models for in vivo analyses of ciliogenesis and cilium maintenance.²² The cilia in the zebrafish pronephric kidney include single motile cilia, arising from the majority of epithelial cells, and cilia arranged in multicilia bundles in the proximal straight tubule and distal early segment.^23–26 Single-ciliated cells (SCCs) and multiciliated cells (MCCs) form an intercalated “salt-and-pepper” pattern in the proximal straight tubule–distal early region, which is controlled by Notch signaling.^25,26 Because the pathogenesis of polycystic kidney disease (PKD) cannot be fully recapitulated during 8 days of embryogenesis, we turned our attention to adult zebrafish. The adult zebrafish mesonephric kidney undergoes similar branching morphogenesis and segment organization (proximal tubule [PT], distal tubule [DT], and collecting duct [CD]) as the mammalian kidney, although it contains much fewer nephrons (approximately 200 versus 1 million in humans) and lacks the loop of Henle. The loop of Henle functions to preserve water in mammals, and is, therefore, an unnecessary segment for freshwater zebrafish.^27–30 Whether or not adult zebrafish can be used to model PKD and develop therapies has not been explored.

Here, we report the establishment of tmem67 mutant as the first adult zebrafish model of kidney cysts. On the basis of the recent advent of tissue-clearing technology,^31–33 we developed a whole-mount imaging protocol that enables the characterization of an isolated zebrafish kidney at a single-nephron resolution. We defined ciliary abnormalities during embryogenesis and adulthood and correlated these defects with cyst development. Finally, we revealed the therapeutic benefits of mammalian target of rapamycin (mTOR) inhibition can be extended to tmem67-based cystic disease.

Methods

Zebrafish Strains

Zebrafish (WIK) were maintained under standard laboratory conditions, and experiments were carried out in accordance with the policies of the Mayo Clinic Institutional Animal Care and Use Committee. A hypomorphic mtor strain was identified from our insertional mutagenesis screen.³⁴

Zebrafish tmem67 mutants were generated using the Golden Gate transcription activator–like effector nucleases (TALEN) assembly protocol and library.^21,35–37 We designed TALEN pairs targeting exon 3 of tmem67, because pathogenic mutations are mostly found in the exons encoding the amino-terminal domains of meckelin.^3,38,39 TALEN mRNAs were injected into embryos at the one-cell stage. F1 adults carrying germ-line mutations were identified by PCR (forward primer, 5′- TGTATAGGACTGGCATGTGAG-3′; reverse primer, 5′-AGGGATTGCCATTCCCATC-3′) and XmnI restriction-enzyme digestion. Frameshift mutations were identified by sequencing, and the zebrafish were further outcrossed to reduce potential off-target effects. All experimental fish used in this study were F3 or F4 animals.

In Situ Hybridization

Whole-mount in situ hybridization was performed, as previously described, using riboprobes that were generated from T7 promoter sequence–tagged PCR products.⁴⁰

Histologic Analysis

Zebrafish embryos were analyzed by staining with hematoxylin and eosin, as we have described previously.²¹ Adult zebrafish kidneys were also collected as described in another study.²⁸ Briefly, each fish was euthanized with 0.2% tricaine methanesulfonate. All of its internal organs were removed except for the kidney, which was attached to the dorsal wall of the abdominal cavity. The fish body with the kidney was then fixed in 4% paraformaldehyde (PFA) overnight at 4°C. The next day, the kidney was carefully detached from the abdominal wall and subjected to paraffin embedding and staining with hematoxylin and eosin. If a tubule was dilated to an area >0.03% of the total kidney area, it was considered a cyst. The areas of renal cysts as percentages of the total tissue area were calculated using ImageJ software.

Immunofluorescence Labeling

Cilia in zebrafish embryos were visualized by whole-mount immunofluorescence staining, using antibodies against acetylated α-tubulin (Sigma-Aldrich), as previously described.^41,42 Antibodies against α6F (Developmental Studies Hybridoma Bank) were included for staining of the pronephros.

Immunofluorescence analysis of adult zebrafish kidneys was performed on cryo-sections as described previously.^21,28,29 The kidneys were dissected, fixed in 4% PFA/0.1% DMSO overnight at 4°C, and then permeabilized with 5% sucrose for 30 minute followed by 30% sucrose overnight. The next day, the kidneys were embedded in tissue-freezing medium (Electron Microscopy Sciences) and cryo-cut at 10 µm (Leica CM3050S). Immunofluorescence analysis of adult zebrafish kidneys was also conducted on paraffin sections, as described previously.⁴³ Renal tubular segments were labeled with alkaline phosphatase (AP; Invitrogen), rhodamine Dolichos biflorus agglutinin (DBA), and Lotus tetragonolobus lectin (LTL) (Vector Laboratories). Nuclei were labeled with SYTOX or, 4′,6-diamidino-2-phenylindole (DAPI, Vector Laboratories). Proliferating cells were labeled by anti–proliferating cell nuclear antigen (PCNA) antibody (Sigma-Aldrich). Images were acquired using a Zeiss Axioplan II microscope equipped with ApoTome and AxioVision software (Carl Zeiss Microscopy).

Optical Clearing of Adult Zebrafish Kidney and Whole-Mount Immunostaining

After fixation, permeabilization, and removal of melanocyte pigmentation, the kidney was cleared with X-CLARITY (Logos Biosystems) according to the manufacture’s instructions. Briefly, the kidney was incubated in hydrogel solution with 0.25% polymerization initiator overnight at 4°C, transferred to argon atmosphere (oxygen free) to induce hydrogel polymerization at 37°C for 3 hours, and was placed on a 37°C shaker for 3 days for tissue clearing. The samples were then washed in PBS overnight at room temperature and sequentially incubated with the following antibodies: polyclonal anti-ARL13B antibody for cilia labeling (gift from Zhaoxia Sun’s laboratory, Yale University School of Medicine), monoclonal anti-PKC antibody for renal tubular labeling (Santa Cruz Biotechnology, Inc.), Alexa 488 anti-rabbit IgG, and Alexa 568 anti-mouse IgG (Invitrogen). Each antibody was incubated with samples for 3 days on a 37°C shaker at a dilution of 1:200 in 6% BSA in PBS, 0.2% Triton X-100, and 0.01% sodium azide. Images were acquired using a Zeiss LSM 700 microscope and Zen software (Carl Zeiss Microscopy).

Ultrastructural Analysis

Adult zebrafish kidneys were fixed in Trump fixative (4% PFA, 1% glutaraldehyde). The remaining procedures for transmission electron microscopy and scanning electron microscopy (SEM) were performed according to standard methods at the Electron Microscopy Core Facility of the Mayo Clinic in Rochester, Minnesota.

RT-PCR Analysis

RNA was isolated using TRIzol, and cDNA synthesis was performed with the SuperScript III RT Kit (Invitrogen). The primers used for quantitative PCR were as follows: tmem67 forward, AGA​ATC​TGT​GTG​GCC​AAA​GG; tmem67 reverse, CAA​TGA​CAG​AGC​CCA​GGA​AT; mks1 forward, TGA​GGG​CTA​TGG​CTA​TTT​GG; mks1 reverse, CAG​TGG​TCT​GAG​TGC​GAA​AG; cep290 forward, ACC​ACT​GAC​GAA​GGT​CTT​GC; cep290 reverse, CAG​CTC​TTG​GGT​CAG​TTT​CC; cc2d2a forward, AAC​CTC​TGG​AAG​CCG​TTT​TT; and cc2d2a reverse, GGC​AGA​AAC​TGG​CGT​AAT​GT.

Western Blotting

Zebrafish kidneys were homogenized in RIPA lysis buffer (Sigma-Aldrich) as described previously.²¹ Antibodies against the following proteins were used: Tmem67 (LifeSpan BioSciences), β-actin (Sigma-Aldrich), mTOR, phospho-mTOR (S2448), phospho-S6K (T389), S6 ribosomal protein, and phospho-S6 protein (Ser240/244) (Cell Signaling Technology).

Rapamycin Treatment of Adult Zebrafish

Rapamycin (LC Laboratories) was administered at the specified dose to adult zebrafish via oral gavage, as described previously.⁴⁴ Treatment was conducted daily and lasted for 1 month.

Statistical Analyses

The data are presented as the mean±SD. Comparisons between two groups were performed with two-tailed t tests, and a P value of <0.05 was considered to indicate significance.

Results

tmem67^e3/e3 Embryos Exhibit Pronephric Cysts with Partial Penetrance

tmem67, the only zebrafish homolog of mammalian TMEM67, is located on chromosome 16 (Figure 1A).¹² Its transcript was detected in the kidney, neural tube, otic vesicle, brain, and retina during embryogenesis (Figure 1B), and in all major organs of adult fish (Supplemental Figure 1). We generated two tmem67 TALEN mutant alleles that resulted in a premature stop codon, including the M1 allele, containing a 1-bp insertion, and the M2 allele, containing a 5-bp deletion (Figure 1, A and C). Because both alleles exhibited the same phenotypes, we have presented data only from the M1 allele and have renamed this allele tmem67^e3.

Figure 1.: Zebrafish tmem67 ^e3/e3 embryos develop partially penetrant pronephric cysts. (A) Schematic diagram of the exon-intron structure and two mutant alleles. The TALEN recognition sequences are underlined. Mutations in both the M1 and M2 alleles cause a coding frameshift and premature stop codon (*). Additions are indicated in red, and deletions are indicated with dashes. The restriction-enzyme recognition site is highlighted in blue. The corresponding amino-acid sequence is shown below the DNA sequence. (B) In situ hybridization showing tmem67 expression in an 18-somite embryo. The pronephric kidney is indicated with an arrow, the neural tube with a dashed arrow, the otic vesicle with asterisks, the brain with a closed arrowhead, and the lens with an open arrowhead. (C) Genotyping of tmem67 ^e3/e3 (e3/e3) embryos. PCR products amplified from wild-type alleles could be digested by XmnI to produce 257-bp and 237-bp fragments, whereas PCR products amplified from mutated alleles could not be cleaved. (D) Quantitative-PCR analysis of tmem67 transcripts. Embryos at 5 days postfertilization (dpf) were subjected to RNA extraction, cDNA synthesis, and quantitative-PCR analysis. The results were normalized to those for glyceraldehyde-3-phosphate dehydrogenase (gapdh). The data are presented as the mean±SD from three independent experiments. Eight to ten embryos per genotype were examined. **P<0.01. (E) Gross morphology of tmem67 ^e3/e3 embryos. Mutants developed ventral curvature of the body (dashed arrow) and pronephric cysts (dashed circle). Shown are embryos at 4 dpf. (F) Enlarged view of the pronephric cyst in (E). (G) The glomerular neck region of the pronephros was dilated in some tmem67 ^e/e3 embryos (asterisks). Hematoxylin and eosin staining of JB-4 sections in day-3 embryos is shown. nc, notochord; wt, wild type.

In tmem67^e3/e3 embryos, tmem67 transcript levels were reduced by nearly 40%, likely due to nonsense-mediated mRNA decay (Figure 1D). Approximately 40% of the tmem67^e3/e3 embryos exhibited ventral body curvature, and 20% developed pronephric cysts (Figure 1, E–G). Unlike rodent knockout animals and zebrafish morphants, the tmem67^e3/e3 mutants did not exhibit apparent defects in convergence extension or neural-tube development and rarely developed hydrocephalus (Supplemental Figure 2 and data not shown).^8,9,12,14

tmem67^e3/e3 Embryos Progressively Lose Single Cilia in the Pronephros

Next, we noted significantly shorter cilia in the distal pronephros in tmem67^e3/e3 embryos, compared with wild type, at 26 hours postfertilization (Figure 2, A and F), when developing pronephros do not yet have filtration functions and when multicilia clusters are not yet present.^18,25,26 When pronephric cysts became visually detectable at 2 days postfertilization,¹⁸ manifesting as dilations in the glomerulus-neck region, we grouped the mutants into those with pronephric cysts (designated as +EC) and without pronephric cysts (designated as −EC). Compared with wild-type organisms, both +EC and −EC mutants had significantly shorter and fewer single cilia at the segments caudal to the multiciliated region (Figure 2, B, E, G, and H), and had less, but only marginally shorter, single cilia at the segments rostral to the multiciliated region (Figure 2, B, C, and G). The overall reductions in both cilium length and number were comparable between +EC and −EC embryos (Figure 2, G–H). In contrast, multicilia bundles and MCCs, as indicated by odf3b expression, remained normal at this stage (Figure 2, B, D, and I–K). At 4 days postfertilization, distal single cilia were absent from all mutants, but the number of multicilia bundles continued to be normal; however, the numbers of cells with MCC fate were increased in the +EC embryos, suggesting the number of cilia bundles may increase at a later time (Supplemental Figure 3). In contrast to the pronephros, other tissues, such as the Kupffer vesicle and neural tube, had unaffected cilia (Supplemental Figure 4).

Figure 2.: tmem67 ^e3/e3 (e3/e3) embryos have shorter and fewer single cilia in the distal pronephros than wild-type embryos. (A) Pronephric cilia in 26 hours postfertilization (hpf) embryos were examined by whole-mount immunostaining using an α-acetylated tubulin antibody. (B) Pronephric cilia from a wild-type embryo and a mutant without visually detectable pronephric cysts at 2 dpf were illustrated by coimmunostaining using antibodies against α-acetylated tubulin (green) and sodium ion/potassium ion ATPase (α6F, red; labels the pronephros). The boxed areas are enlarged in (C–E). (C–E) Enlarged representative images of α-acetylated tubulin staining. (C) Proximal single cilia, (D) multicilia bundles, and (E) distal single cilia. Boxed areas in (D) show further enlarged cilia bundles. (F) Quantification of distal cilium length in 26 hpf embryos. (G) Quantification of renal single-cilium length in 2 dpf embryos. Single cilia in the tubular segment immediately rostral to the multiciliated region (PT) or caudal to the multiciliated region (DT) were measured. (H) The total numbers of single cilia in the DTs of both pronephros were counted at 2 dpf. (I) Quantification of multicilia bundles at 2 dpf. (J and K) MCCs were examined by in situ hybridization using an odf3b riboprobe. Shown are representative embryos (J) at 26 hpf and (K) 2 dpf. The data are presented as the mean±SD from (F) two or (G–I) four independent experiments. A total of 16–22 embryos per group per age were examined. Scale bars, 5 µm in A, 20 µm in C and D, 40 µm in E, and 50 µm in B. *P<0.05, **P<0.01. +EC, mutants with pronephric cysts; −EC, mutants without pronephric cysts; wt, wild type; NS, not statistically significant.

To assess the contribution of cilia motility to cyst formation, we measured pronephric fluid flow and found no significant difference between wild-type siblings and mutants and between the +EC and −EC embryos (Supplemental Figure 5),²¹ suggesting a largely normal kidney cilia motility. Of note, pronephric cysts in the mutant embryos mostly remained small and rarely led to cardiac/whole-body edemas during embryogenesis, which also argues against substantial cilia motility defects.

Adult tmem67^e3/e3 Fish Develop Progressive Mesonephric Cysts

Nearly all tmem67^e3/e3 embryos, including cystic ones, survived for at least 15 months, with approximately 40% of fish exhibiting wavy bodies (Figure 3A). Independent of body shape, the mutant kidneys were significantly enlarged (Figure 3, B and C), and the renal tubules were progressively dilated in male fish and, to a much less extent, female fish (Figure 3, D and E, and data not shown). Because the phenotype was milder than that in rodent models, we assessed the possibility that the mutations were hypomorphic. We did not note any skipping of the targeted exon via alternative splicing events (Supplemental Figure 6); however, we detected approximately 10% of Tmem67 proteins in the mutants (Figure 3F), suggesting a hypomorphic nature of the mutant. We also examined the expression of other MKS genes. Although tmem67 transcript levels were consistently reduced by approximately 50%, mks1, cep290, and cc2d2a were activated in the mutant kidney, suggesting a genetic compensation (Figure 3G).

Figure 3.: Adult tmem67 ^e3/e3 fish develop progressive renal cysts. (A) tmem67 ^e3/e3 (e3/e3) fish were either morphologically indistinguishable from their wild-type siblings or curved in shape. Shown are 12-month-old fish. (B and C) tmem67 ^e3/e3 fish had larger kidneys than wild-type fish. (B) Fish (12 months old) were fixed in 4% PFA, and then the kidneys were dissected for size measurement. (C) The kidney area was normalized by the body weight (BW). Eight male fish from each group were analyzed. (D and E) tmem67 ^e3/e3 fish exhibited renal cysts. Kidneys were collected at the indicated ages, and staining with hematoxylin and eosin was performed on paraffin sections. Representative images (D) are shown. (E) The cyst area/total kidney area (cystic index) was calculated. Four to 11 male fish from each group at each time point and three sections per kidney were analyzed. (F) Tmem67 protein levels in the adult fish kidney. Five wild-type and five mutant kidneys were harvested at 7 months, and the representative images are shown. (G) Quantitative PCR analysis of other MKS genes. Kidneys from three wild-type siblings and three tmem67 ^e3/e3 fish were collected at 9 months, and gene expression was normalized by gapdh. The data are presented as the mean±SD. Scale bars, 7 mm in A, 1 mm in B, and 100 µm in D. *P<0.05, **P<0.01. 6M, 6 months; wt, wild type.

In rodent Tmem67 models, renal cysts present as a mixture of PT dilation and CD enlargement in newborns and are mainly restricted to the CD at the later stage.^8,9 In the tmem67^e3/e3 fish, the proportion of PT cysts, marked by AP, decreased from 6 to 12 months; DT cysts, labeled by DBA in zebrafish, had no significant change; and AP (−)/DBA (−) cysts, presumptively a mixture of CDs and tubules with dedifferentiated epithelial cells, significantly increased (Figure 4, A and B).²⁸ We then costained samples with LTL, which uniquely labels both PTs and the major CDs in zebrafish,²⁹ and DBA, and found that the numbers of unstained cysts, presumably dedifferentiated, increased with time (Figure 4, C and D). Moreover, by subtracting AP (+) cysts (Figure 4B) from LTL (+) cysts (Figure 4D), more CD cysts were estimated at a later stage (approximately 38% at 12–14 months versus approximately 2% at 6–7 months).

Figure 4.: tmem67 ^e3/e3 (e3/e3) fish exhibit a switch in cystic tubule origin along disease progression and hyperproliferation of cyst-lining epithelial cells. (A and B) Dynamic changes in PT cysts. (A) The tubular origins of cysts were revealed by costaining of frozen sections with AP (blue; labels PTs), DBA (red; labels DTs), and SYTOX (green; labels nuclei). (B) The percentage of segment-specific cysts/total cysts was quantified. (C and D) Dynamic changes in other cysts. (C) The identities of cystic tubules were indicated by costaining of paraffin sections with LTL (green; labels PTs and major CDs), DBA (red), and DAPI (blue; labels nuclei). (D) The percentage of segment-specific cysts/total cysts was quantified. (A and C) The dashed circle indicates an unstained tubule. (E and F) Kidney epithelial cell proliferation. (E) Paraffin sections of 9-month-old fish kidneys were immunostained using a PCNA antibody (red), and then costained with LTL (green) and DAPI (blue). (F) The percentage of PCNA(+) LTL-labeled cells (arrow in E) were quantified in 6- and 9-month-old kidneys (F). (B and D) Four to six male fish per genotype per time point, three sections per kidney, were used for renal cyst quantification, and approximately 200 LTL-labeled cells per section were analyzed in (E). The data are presented as the mean±SD. Scale bar, 20 µm. *P<0.05, **P<0.01. 6M, 6 months; wt, wild type.

One of the characteristics of mammalian PKD is increased proliferation of cyst-lining epithelial cells.^45,46 Consistently, we detected more cells positive for proliferating cell nuclear antigen in the tmem67^e3/e3 mutants than in those of wild-type fish (Figure 4, E and F).

The Mesonephric Kidneys of tmem67^e3/e3 Fish Have Shorter and Fewer Single Cilia but More MCCs than Wild-type Fish

Because adult zebrafish kidney cilia have not been characterized, we conducted electron-microscopy analysis. In agreement with the light-microscopic observations,^27,29 we found some tubules were lined with thick brush borders, which are indicative of PT segments (Figure 5, A and A′), whereas others had scanty microvilli and could not be clearly distinguished as DTs or CDs (Figure 5, B and B′). Cilia bundles were observed in PTs, as noted in other teleost fish,⁴⁷ and also in DTs/CDs (Figure 5, A, A′, B, B′, and G). Approximately 20–30 cilia, measuring 15–21 µm in length, were typically clustered together with the distinctive “9+2” motile cilia structure (Figure 5, C and G). Single cilia were scarcely detected by SEM (Figure 5H), probably because longitudinal sections were rarely obtained from the thin kidney tissues. In the tmem67^e3/e3 kidneys, the epithelial cells in dilated PTs often had loose brush borders (Figure 5, D and D′). Clusters of cilia that were 15–21 µm long also resided in the PTs and DTs/CDs and retained the 9+2 organization (Figure 5, D–F and I). Notably, the mutants had shorter single cilia than the wild-type fish, although only a few were identified (Figure 5J), and contained more cilia bundles in the PT segment (Figure 5K). By immunostaining, single cilia were found in PTs, DTs, and unstained tubules, and so were cilia bundles, as specified by strong α-tubulin staining. Consistently, more cilia bundles were observed in the mutants (Supplemental Figure 7).

Figure 5.: Adult tmem67 ^e3/e3 (e3/e3) kidneys have more MCCs than wild-type kidneys. (A–C) Transmission electron microscopy (TEM) analysis of cilia in the mesonephric kidneys of wild-type fish. Multicilia bundles (dashed squares) were observed in (A) the PTs (with brush borders) and (B) DTs/CDs (without brush borders), and (A′ and B′) were enlarged. These cilia had typical 9+2 motile cilia structures (C). (D–F) TEM analysis of cilia in the mesonephric kidneys of tmem67 ^e3/e3 fish. Multicilia bundles were also observed in (D and D′) the PTs and (E and E′) DTs/CDs, and (F) had 9+2 structures. (G and I) Multicilia bundles and (H and J) single cilia in the PT segments of (G and H) wild-type fish and (I and J) tmem67 ^e3/e3 mutants were revealed by SEM analysis. (K) The percentage of MCCs in the PT segment was quantified via TEM analysis. Kidneys were collected from 9-month-old male fish, six wild-type siblings, and three mutants, and approximately 200 PT cells per kidney were scored. The data are presented as the mean±SD. Scale bars, 5 µm in A, B, D, and E; 1 µm in A′, B′, D′, and E′; 50 nm in C and F; 2 µm in G, H, and J; and 8 µm in I. *P<0.05. wt, wild type.

To more accurately monitor cilia in the adult kidney, we performed optical clearing, which removes lipids from tissues and normalizes the refractive index to enable whole-mount imaging of the kidney. We were able to capture intact individual nephrons with both single cilia and multicilia bundles (Supplemental Videos 1 and 2). Single cilia were noticed in the glomerulus, glomerulus-PT junction, between multicilia clusters, in the distal half of the nephron, and in the CD-like structure; brightly stained multicilia bundles were present in the proximal half of the nephron (Figure 6, A, B1–3, and B6). In addition, DTs joining to another DT or CD-like tubule were spotted (Figure 6, B4 and B5). In tmem67^e3/e3 zebrafish, we observed tubular dilation (Figure 6, D and G), densely populated cilia bundles (Figure 6, E–G), and fewer and stunted distal single cilia (Figure 6, H and I). Together, our data reveal the occurrence of consistent ciliary defects in tmem67^e3/e3 embryos and adult fish.

Figure 6.: Optical-clearing, whole-mount immunostaining of adult tmem67 ^e3/e3 (e3/e3) kidneys reveals loss of distal single cilia. (A) Three-dimensional reconstruction of a single intact nephron showing multicilia/single cilia localization in a wild-type kidney. Whole-mount immunostaining of an isolated, optically cleared kidney was performed using anti-Arl13B (green) and anti-PKC (red) antibodies. (B) Representative images showing single cilium and cilia bundles at the glomerulus-PT region (B1), single cilium interpolated between cilia bundles (B2), the transition area from cilia bundles to single cilium (B3), the merging of a single cilium-containing DT with another DT (B4), the merging of a DT to a CD-like structure (B5), and a CD-like tubule (B6). (C) Schematic showing proposed cilia distribution in a whole nephron. (D) PKC-based, whole-mount immunostaining showing tubular dilation in tmem67 ^e3/e3 kidneys. (E) Three-dimensional reconstruction of a single intact nephron showing multicilia/single cilia localization in a tmem67 ^e3/e3 kidney. (F) Representative images showing cilia bundles and single cilia in wild-type and tmem67 ^e3/e3 kidneys. (G) At higher magnification, compressed cilia bundles in the mutants were strongly stained, expanded, and emanated into cystic lumen; (H) distal single cilia were reduced in numbers and length. (I) Quantification of single-cilium length in the DTs and CD. Asterisk, open triangle, closed triangle, and arrow indicate glomerulus, single cilium, cilia bundles, and tubular dilation, respectively. Dashed lines outline tubules. Four fish per group, 100–160 distal single cilia or 50–60 CD single cilia were measured. The data are presented as the mean±SD. Scale bars, 20 µm in B1–B3, 10 µm in B4–B6, 100 µm in D and F, and 50 µm in G and H. **P<0.01. DE, distal early; DL, distal late; GL, glomerulus; PCT, proximal convoluted tubule; PST, proximal straight tubule; wt, wild type.

Cyst Development during Embryogenesis Does Not Predispose but Rather Delays tmem67 Fish to Mesonephric Cyst Formation

To assess whether pronephric tubule dilation affects mesonephric cyst formation, we raised +EC embryos separately from −EC embryos. The cystic index in +EC fish progressively increased from 4 to 12 months; however, surprisingly, it was lower at 4 months in +EC fish than in −EC fish (Figure 7A). Consistently, epithelial cell hyperproliferation was detected only in −EC kidneys (Figure 7B).

Figure 7.: Development of pronephric cysts is oppositely associated with mesonephric cyst formation at early adulthood. (A) Adult tmem67 ^e3/e3 fish that exhibited pronephric cysts (+EC) and those that did not develop pronephric cysts (−EC) were subjected to renal cyst analysis at the indicated ages. The cystic indices quantified from seven to 12 male fish per group per time point are shown. (B) Cell proliferation in the +EC and −EC fish at 4 months, as indicated by PCNA immunostaining. (C–E) Renal cilia in the +EC and −EC fish at 4 months. Whole-mount immunostaining of cleared kidneys was carried out using anti-Arl13B (green) and anti-PKC (red) antibodies. Shown are (C) representative images of distal single cilium, (D) quantification of the single-cilium lengths, and (E) representative images of cilia at the glomerulus-PT junction. Asterisk indicates glomerulus; dashed lines outline proximal convoluted tubule region. Three fish per group and 63–167 distal single cilia were measured. The data are presented as the mean±SD. Scale bars, 10 µm in C, 20 µm in E. *P<0.05, **P<0.01. 4M, 4 months; wt, wild type.

Prompted by the significant difference in cyst formation between +EC and −EC fish at 4 months of age, we compared their renal cilia. We found more PT-MCCs in the +EC animals than in the −EC animals, although the numbers of MCCs were also substantially increased in the −EC group. On the other hand, the single-cilium number was reduced comparably in the two groups (Supplemental Figure 8). Tissue-clearing, whole-mount imaging further revealed the distal single-cilium length was similarly shortened in +EC and −EC fish (Figure 7, C and D). Interestingly, cilia bundles at the glomerulus-PT junction of the −EC kidney were less condensed than in the +EC fish, or even replaced by single cilia (Figure 7E).

mTOR Inhibition Ameliorates Cyst Formation, Increased Proliferation of Cyst-Lining Cells, and Ciliary Abnormality in tmem67^e3/e3 Fish

We then asked whether signaling pathways that are dysregulated in PKD were also altered in the tmem67 mutant.^46,48 We found hyperphosphorylation of mTOR, S6K, and S6 protein in tmem67^e3/e3 kidneys (Figure 8, A and B). This upregulation seemed specific to mTORC1, because AKT was not activated (data not shown). We also found mTOR activation in bpck mice, suggesting a conserved signaling event (Supplemental Figure 9). Additionally, we noted overexpression of the Wnt/β-catenin target gene axin2 (a specific target) and cyclin D1 but normal MAPK activity (Supplemental Figure 10).¹⁴

Figure 8.: mTOR haploinsufficiency reverses cystogenesis, overproliferation, and ciliary abnormality in tmem67 ^e3/e3 fish. (A and B) mTOR signaling was activated in the kidneys of tmem67 ^e3/e3 (e3/e3) fish. (A) Phosphorylation of mTOR and S6 proteins was examined by immunoblotting. (B) Administration of 7.3 µg/g of rapamycin (rapa) to the adult fish via gavage effectively inhibited phosphorylation of S6K. Representative images of three independent experiments are shown. Samples were harvested 2 hours after rapamycin or vehicle administration. Total of 12–14 fish at the age of 7–9 months per genotype per treatment were used. (C) Rapamycin treatment suppressed cystogenesis in tmem67 ^e3/e3 fish. Rapamycin or vehicle (dimethylsulfoxide) was given to 7-month-old fish by daily gavage for 1 month. Then, renal cyst formation was examined by hematoxylin and eosin staining of paraffin sections. Representative images are shown, and the cystic indices were calculated. Four wild-type male fish and eight male mutants per group were analyzed. (D) tmem67 ^e3/e3 ;mtor ^+/− double mutants exhibited fewer renal cysts than tmem67 ^e3/e3 single mutants. Eight male fish per genotype at 9–12 months were examined. (E) tmem67 ^e3/e3 ;mtor ^+/− double mutants had fewer renal epithelial cells positive for PCNA than tmem67 ^e3/e3 single mutants. (F and G) Ciliary phenotypes in tmem67 ^e3/e3 mutants were rescued in tmem67 ^e3/e3 ;mtor ^+/− double mutants, as indicated by the percentages of (F) MCCs and (G) single-ciliated cells (SCCs). (E–G) Four male fish per genotype at 9 months were examined, and all analyses were performed on LTL tubules. The data are presented as the mean±SD. Scale bar, 100 µm. *P<0.05, **P<0.01. wt, wild type.

Because experimental mTOR inhibition results in significant regression of renal cysts in rodent models of PKD,^49–51 we went on to test whether mTOR inhibition also exerts beneficial effects in tmem67^e3/e3 fish. We found that rapamycin treatment abolished phosphorylation of S6K and S6 protein and diminished mesonephric cysts, without causing apparent defects in wild-type fish (Figure 8, B and C, and data not shown). Moreover, genetic inhibition of mTOR signaling with a hypomorphic mtor^+/− mutant also significantly inhibited cyst growth in the adult fish (Figure 8D).³⁴ At the cellular level, we noticed that hyperproliferation of renal epithelial cells was reversed; the numbers of MCCs and single-ciliated cells were normalized in tmem67^e3/e3;mtor^+/− double mutants (Figure 8, E–G). However, although mTOR signaling was activated and short-term rapamycin treatment alleviated pronephric cysts in +EC embryos, rapamycin was unable to mitigate ciliary defects in the embryos (Supplemental Figure 11), suggesting the therapeutic benefit of sustained mTOR inhibition likely does not occur via direct regulation of ciliogenesis.

Discussion

tmem67^e3/e3 Mutant Is the First Adult Zebrafish Model of Kidney Cysts

In this study, we have demonstrated, for the first time, that renal cystic disease can be studied in adult zebrafish. Characteristics of mammalian PKD were conserved in tmem67^e3/e3 fish mesonephric kidneys, including progressive cystogenesis, an age-dependent switch in tubular cyst origin, and increased cell proliferation. Compared with pronephros consisting of a single pair of nephrons, the adult fish mesonephros possesses numerous nephrons and enables the quantification of cyst size and number, so that disease progression can be monitored. Moreover, longitudinal studies can be conducted to identify compensatory versus decompensatory events, as indicated by our observation that embryonic cysts do not predispose, but rather delay, mesonephric cyst formation in young adults.

Compared with patients with MKS3 and corresponding rodent models, the zebrafish tmem67 mutants exhibited milder cystic phenotypes. This difference is unlikely an issue related to freshwater species per se, because naturally occurring pc/glis3 mutants of medaka, another freshwater teleost, develop massive cysts within 6 months.^47,52 In addition to a potential hypomorphic mutation, there are two plausible reasons for this. First, zebrafish have a strong regenerative capacity, not only repairing nephrons (as mammals do) but also forming new ones de novo.^53–57 Because impairments in renal tubule repair after injury have been suggested to trigger cystogenesis,^58,59 an elevated regeneration potential would certainly help to slow disease progression. Second, genetic compensation by other MKS genes was noted in the tmem67 TALEN mutants. Recent studies have shown that a TALEN-mediated indel mutation, if resulting in a premature termination codon, often activates nonsense-mediated mRNA decay of the mutant mRNA and induces the expression of genes with sequence homology.^60–62 Future endeavors toward identification of critical regeneration/compensation genes in this fish model might suggest new therapeutic targets for MKS3 treatment in mammals.

Relationship between Ciliary Abnormality and Cystogenesis in the Zebrafish tmem67 Models

We successfully applied optical-clearing, three-dimensional imaging to the adult zebrafish kidney for precisely defining cilia in normal and diseased states. Through monitoring cilia and renal cysts at a single-nephron resolution, and together with electron-microscopy and 2D analysis, we made the following discoveries. First, we described cilia distribution in the nephron. Single cilium exists in all segments and multicilia bundles in the proximal part of the nephron. We will determine the segment identity in the future using segment-specific reporter strains. Second, single-cilia defects were detected in both +EC and −EC embryos, a phenotype also reported in zebrafish mutants for the causative gene for Joubert syndrome, ahi1,⁶³ and all adult mutants, including +EC fish at the precyst stage of 4 months. On the other hand, MCC expansion was observed after pronephric cyst formation and was reversely correlated with cyst severity in the young adult zebrafish. These longitudinal studies suggest Tmem67 exerts a primary function on single-cilia biogenesis, probably via controlling cell type–specific entrance of ciliary proteins into cilia, and that MCC expansion is more likely a compensatory event.

MCCs are rarely detected in adult mammals. However, they have been observed in human fetuses, in patients with hypercalcemia or nephrotic syndrome, and in rodent models of autosomal recessive PKD.^{11,14,64–69} With regard to TMEM67 in particular, more MCCs have been noted in MKS3 fetuses than in normal fetuses.¹¹ In addition, multiple centrosomes and more than one cilium have been detected in wpk rat and Tmem67-deficient cells in vitro, although further investigation is required to determine whether these cells are MCC-like.^11,14 Nonetheless, the reappearance of MCCs in diseased kidneys indicates their importance and has been postulated as an adaptive event to relieve local fluid accumulation.⁷⁰ Our data showing opposite correlation between MCCs and cyst severity provide the first experimental evidence to support this hypothesis.

The zebrafish tmem67 model does not recapitulate the ciliary elongation phenotype present in some mammalian models.^9,11,13,15 Although renal cilia are motile in zebrafish but mostly immotile in mammals, the machinery for both motile and nonmotile cilium assembly and maintenance is conserved.^71–74 Of note, the ciliary elongation phenotype has also been proposed as an adaptive response.^75,76 Whether or not such an adaptive event predominantly presents as cilium lengthening in mammals but MCC expansion in zebrafish, due to species-specific expression of ciliogenic factors and/or different susceptibility to cell-fate transition, warrants further study.

The Adult Zebrafish MKS3 Model Facilitates the Identification of Candidate Therapeutic Strategies Such as mTOR Inhibition

Consistent with findings in patients with PKD and other types of animal models, mTOR signaling is hyperactivated in +EC (but not −EC, data not shown) embryos and all adult mutants, and mTOR inhibition shows beneficial effects. Through extending the mTOR therapy to MKS, a syndromic form of PKD, our results underscore the use of an adult zebrafish model for the identification of therapeutic strategies and support the hypothesis that abnormal mTOR signaling is a common pathologic event for cystogenesis of different etiologies and that mTOR inhibition is a broadly applicable therapeutic strategy.

mTOR signaling has been shown to regulate cilium biogenesis, but the conclusions of the related studies have not always been consistent, and the relationship between mTOR and cilia during cystogenesis remains elusive.^77–83 Here, we found that long-term mTOR inhibition normalized ciliary abnormalities in adult fish, whereas short-term inhibition in embryos did not, suggesting an indirect role of mTOR in ciliogenesis. These data lead to the hypothesis that sustained mTOR inhibition exerts therapeutic benefits via a cilia-independent mechanism. Future studies using the adult zebrafish model are anticipated to uncover this mechanism, facilitating our understanding of the interplay among mTOR signaling, ciliogenesis, and cystogenesis.

Disclosures

P.C. Harris reports having patents and inventions with Amgen, Bayer, Genzyme, GlaxoSmithKline, Millipore, Mitobridge, and Vertex; having consultancy agreements with Mitobridge, Otsuka, Regulus, and Vertex; and receiving research funding from Otsuka Pharmaceuticals. All remaining authors have nothing to disclose.

Funding

This work was supported by the Mayo Clinic Translational Polycystic Kidney Disease Center Pilot and Feasibility grant NIDDK DK90728 (to X. Lin), the Mayo Foundation for Medical Education and Research (to X. Xu), and the Mayo Clinic Center for Biomedical Discovery (to X. Xu and PC, Harris).