Knockdown of Bicaudal C in Zebrafish (Danio rerio) Causes Cystic Kidneys: A Nonmammalian Model of Polycystic Kidney Disease

Abstract

Polycystic kidney disease (PKD) is one of the leading causes of end-stage renal disease in humans and is characterized by progressive cyst formation, renal enlargement, and abnormal tubular development. Currently, there is no cure for PKD. Although a number of PKD genes have been identified, their precise role in cystogenesis remains unclear. In the jcpk mouse model of PKD, mutations in the bicaudal C gene (Bicc1) are responsible for the cystic phenotype; however, the function of Bicc1 is unknown. In this study, we establish an alternative, nonmammalian zebrafish model to study the role of Bicc1 in PKD pathogenesis. Antisense morpholinos were used to evaluate loss of Bicc1 function in zebrafish. The resulting morphants were examined histologically for kidney cysts and structural abnormalities. Immunostaining and fluorescent dye injection were used to evaluate pronephric cilia and kidney morphogenesis. Knockdown of zebrafish Bicc1 expression resulted in the formation of kidney cysts; however, defects in kidney structure or pronephric cilia were not observed. Importantly, expression of mouse Bicc1 rescues the cystic phenotype of the morphants. These results demonstrate that the function of Bicc1 in the kidney is evolutionarily conserved, thus supporting the use of zebrafish as an alternative in vivo model to study the role of mammalian Bicc1 in renal cyst formation.

Polycystic kidney disease (PKD) is one of the leading causes of end-stage renal disease, with autosomal dominant PKD affecting 1 in every 500 to 1000 persons.²⁴ Autosomal dominant PKD is characterized by fluid-filled cysts in the kidneys and extrarenal abnormalities including hepatic and pancreatic cysts, hypertension, intracranial aneurysms, aortic dissection, cardiac valve abnormalities, and abdominal wall hernias.⁶² The rarer recessive form of the disease, autosomal recessive PKD, occurs in approximately 1 in every 20,000 live births and is characterized by cysts in the collecting ducts, biliary dysgenesis, and hepatic fibrosis.⁶² Feline PKD, which closely mimics human autosomal dominant PKD, is a prominent inherited disorder, with nearly 40% of Persian and Persian-related cats worldwide testing positive for PKD.^3,4,14 Despite advances in identifying some of the genes involved in PKD, there is currently no cure for the disorder, and the molecular pathways and mechanisms involved in cyst initiation, formation, and progression are not completely understood.

Animal models, particularly rodents, have been instrumental in the identification of genes that play a role in the development of cysts. One example is the jcpk mouse model in which homozygous affected animals have a phenotype that closely resembles human autosomal dominant PKD, with bilateral cysts in all parts of the nephron.²³ Mutations in Bicc1, the mammalian ortholog of the Drosophila Bicaudal C gene (BicC), are responsible for disease in jcpk model.¹³ In Drosophila, BicC is required for various aspects of oogenesis and specification of anterior–posterior patterning functions.⁴⁰ Female flies heterozygous for BicC mutant alleles produce embryos with a range of anterior–posterior patterning defects, including bicaudal embryos. Recent reports show that BicC is a RNA-binding protein that recruits a deadenylase complex to its target RNAs and functions to regulate oogenesis, cytoskeletal rearrangements, and its own translation.^11,37,52 Interestingly, although mutations in BicC cause developmental defects in Drosophila, mutations in the mouse ortholog, Bicc1, do not cause developmental defects but instead lead to PKD in affected animals.

Rodent models have played an important role in the identification of genes involved in cystogenesis; however, determining the in vivo function of these genes is cumbersome due to the lack of efficient means to manipulate gene expression in living animals. Zebrafish offer an attractive alternative to rodent PKD models due to the ease of visualizing organs and tissues in the transparent embryos and juveniles, the conservation of genetic pathways regulating organogenesis, and the ability to rapidly assay loss- and gain-of-function phenotypes for any gene.

The pronephros is the first kidney to develop in vertebrates and is functional in larval zebrafish.²⁸ The pronephros consists of 2 fused glomeruli at the midline and paired tubules and collecting ducts, which empty into the cloaca. Although the zebrafish pronephros is more primitive than the metanephric mammalian kidney, it still contains the same specialized renal cell types in the glomerulus and tubules that are seen in the mouse. Unlike other alternative vertebrate model systems, such as Xenopus, zebrafish have a closed nephron system, which is more similar to the mammalian nephron.³² The zebrafish pronephros develops within 2 d after fertilization, and morphogenetic events model those in mammalian kidney development.²⁰ Notably, several studies have demonstrated that mutations in zebrafish orthologs of human and mouse PKD genes result in cystic phenotypes in fish.^39,45,54

In this study, we use the unique molecular tools and features of zebrafish to demonstrate that elimination of bicc1 function in the zebrafish causes cyst formation. We provide evidence that the cystic phenotype can be rescued with the addition of mouse Bicc1. These results demonstrate that the role of the Bicc1 protein in the kidney is evolutionarily conserved and suggest that the zebrafish is an excellent in vivo model system to investigate the role of Bicc1 in kidney development and PKD pathogenesis.

Materials and Methods

Animals.

Wild-type zebrafish were bred and raised inhouse by using breeding pairs that have been maintained as a closed colony for several generations. Fish were reared under the guidelines described in the Guide for the Care and Use of Laboratory Animals³³ and the Zebrafish Book.⁶¹ The protocol for the experimental use of fish was approved by the University of Missouri Institutional Animal Care and Use Committee. Adult fish were kept in a filtered, UV-treated recirculating aquatic system at 28.5 °C on a 14:10-h light:dark cycle. The adult fish were fed brine shrimp (Artemia nauplii) twice daily. Fish were spawned and embryos collected by using a mesh-bottom tank system. Embryos were collected in 90-mm culture dishes in E3 embryo medium⁶¹ and kept in an incubator at 28.5 °C for 4 d postfertilization (dpf). At 4 dpf, embryos were fed live paramecia (Paramecia multimicronucleatum) twice daily. At 9 dpf, embryos were moved to baby-rearing containers in the nursery and fed both paramecia and brine shrimp twice daily. Containers were cleaned carefully and replaced with clean system water twice daily. The proportion of brine shrimp was increased gradually while phasing out the paramecia until 21 dpf, at which point progeny were moved to adult tanks and fed live brine shrimp twice daily. Throughout the text, the developmental age of the embryos corresponds to the hours elapsed since fertilization (hours postfertilization [hpf] at 28.5 °C).

cDNA and sequence analysis.

A construct containing the full-length wild-type mouse Bicc1 (GenBank accession no. NM_031397) cDNA in pCS2⁺ was kindly provided by Oliver Wessely (Louisiana State University Health Sciences Center, Baton Rouge, LA). A zebrafish ortholog of mouse Bicc1 was identified by using Basic Local Alignment Search Tool (BLAST)² and designated bicc1a. Multiple sequence alignments were performed by using Clustal W2,¹⁰ and the images were viewed by using Jalview.^10,12 The following nucleotide (and protein) accession numbers were used: zebrafish bicc1a, NM_203420 (NP_981965); mouse Bicc1, NM_031397 (NP_113574); and human BICC1, ENST000000373886 (ENSP000000362993). Conserved protein domains were identified by using multiple programs available at ExPASy,²⁵ including MotifScan, ProSite, and Pfam.^{17,22,31,46,66}

RNA isolation and RT-PCR.

Total RNA was isolated from 20 to 30 dechorinated embryos at multiple stages of development using TRIZOL reagent (Invitrogen, Carlsbad, CA). Embryos less than 30 hpf were lysed by vortexing vigorously for 2 to 3 min; embryos at least 30 hpf were sonicated on ice in 5-s bursts for 1 min. RNA pellets were washed once with 75% ethanol and resuspended in RNA storage solution (Ambion, Austin, TX). One-step RT-PCR was performed (Superscript III, Invitrogen) according to the manufacturer's instructions. Gene-specific primers for bicc1a (5′ CAT GAC TGC AAA CAC TCC TTT GT 3′ and 5′ AAG GTC ACG CTT CTC TGC AT 3′) were designed by using Primer3.⁴⁸ The primers used for the housekeeping gene ef1α were 5′ TCA CCC TGG GAG TGA AAC AGC 3′ and 5′ ACT TGC AGG CGA TGT GAG CAG 3′.³⁰ A 25-μL reaction containing 100 ng template RNA and 0.2 μM gene-specific primers was amplified under the following conditions: 45 °C for 30 min; 94 °C for 2 min; 40 cycles of 94 °C for 30 s, 58.5 °C (bicc1a) or 66 °C (ef1α) for 30 s, 68 °C for 1 min, with a final extension of 7 min at 68 °C. After PCR, 15 μL of the amplified products were separated on a 3% agarose (in Tris-borate–EDTA) gel. Expected products of 179 bp for bicc1a and 648 bp for ef1α were visible after separation by gel electrophoresis.

Morpholinos.

Translation-blocking or splice-blocking antisense morpholino oligonucleotides were designed by Gene Tools (Philomath, OR) and obtained from Open Biosystems (Huntsville, AL). The sequences for the morpholinos used are as follows: bicc1a translation-blocking (NM_203420), 5′ TCT CTG AGG CCG CCA TAG CAA GAC T3′; bicc1a splice-blocking, 5′ GGA AGC ATG ACT TTC CTC ACC TTT C 3′; and pkd2 translation-blocking (AY618926), 5′ AGG ACG AAC GCG ACT GGA GCT CAT C 3′.⁵⁴ A gli2 antisense mismatch morpholino (5′ CCT CTT ACC TCA GTT ACA ATT TAT A 3′) was used as a control.³⁵

RNA and morpholino injections.

The mBicc1–pCS2+ construct was linearized with NotI (New England Biolabs, Ipswich, MA) and full-length synthetic mouse Bicc1 mRNA was synthesized by using mMessage mMachine (Ambion) according to the manufacturer's instructions. Antisense morpholinos and mRNA were suspended in Danieau buffer⁴³ containing 0.1% phenol red, loaded into microcapillary needles (Stoelting, Wood Dale, IL), and injected into 1- to 4-cell embryos by using a microinjector (ASI, Eugene, OR). Dose–response experiments indicated that effective dose ranges were 1 to 20 ng for morpholino oligonucleotides and 150 to 750 pg for RNA rescue studies. At higher doses of morpholino oligonucleotides, the injected embryos exhibited nonspecific defects and died. Similarly, dose–response experiments defined the highest RNA dose at which nonspecific or toxic effects were absent.

Immunohistochemistry and in situ hybridization.

Whole-mount immunohistochemistry was performed as described previously⁶ with the following modifications. Embryos were fixed overnight in Dent¹⁸ (80% methanol, 20% DMSO), treated with 10% H₂O₂ in Dent for 2 h, and placed in 100% methanol. A monoclonal antibody against the chicken α1 subunit of Na⁺-K⁺ ATPase⁵⁵ was acquired from the Developmental Studies Hybridoma Bank (Iowa City, IA) and used at a 1:10 dilution. Synthesis of the digoxigenin-labeled probes and whole-mount in situ hybridization were carried out as described previously.⁵ The probes used for in situ hybridization were derived from EST clone AL923800 (clone ID 131-B03-2, OpenBiosystems) and linearized with XhoI (antisense; New England Biolabs) or XbaI (sense; New England Biolabs). Sense and antisense digoxigenin–dUTP-labeled probes were synthesized by using DIG RNA labeling mix (Roche Applied Science, Indianapolis, IN) according to the manufacturer's instructions, with 1 μg linearized plasmid DNA template and either T3 (antisense) or T7 (sense) RNA polymerase. Embryos were deyolked as necessary, mounted in 70% glycerol, and examined by fluorescence microscopy (BX60, Olympus, Center Valley, PA). In all comparisons, at least 8 embryos for each category were examined.

Histologic analysis.

Five-day-old larvae were fixed overnight in 2% glutaraldehyde, 2% paraformaldehyde in 0.1 M sodium cadodylate buffer, pH 7.35. Samples were washed in cadodylate buffer, taken through an ethanol dehydration series, embedded in JB4 (Polysciences, Warrington, PA), and serially sectioned at 4 μm by using a microtome (Ultracut UCT, Leica, Bannockburn, WI). Slides were stained with 1% alkaline toluidine blue in 1% sodium borate^8,42 and observed by light microscopy (Axiophot, Zeiss, Thornwood, NY). Images were acquired with an Olympus DP70 digital camera. In all comparisons, at least 6 embryos for each category were examined.

Rhodamine–dextran dye injections.

Rhodamine–dextran injections were performed as previously described³⁶ with the following modifications. Embryos (3 to 3.5 dpf) were anesthetized with 0.02% MS222 (Tricaine, Sigma–Aldrich, St Louis, MO) and mounted on slides in 3% methylcellulose (Sigma). A 5% solution of 70-kDa tetramethylrhodamine–dextran (Invitrogen) was injected into the common cardinal vein; the embryo was covered with a coverslip and examined by using fluorescence microscopy (BX60, Olympus). Ten wild-type embryos and 8 bicc1a morphants were examined.

Immunostaining of cilia.

Embryos (28 to 30 hpf) were fixed in Dent overnight, slowly rehydrated in PBS containing 1% DMSO and 0.1% Tween, and blocked 2 h in PBS containing 1% DMSO, 0.1% Tween, and 10% normal goat serum. Primary antibody (mouse antiacetylated α tubulin; dilution factor, 1:400; T6793, Sigma) was added to PBS containing 1% DMSO, 0.1% Tween, and 1% normal goat serum and incubated overnight. Samples were washed 5 times for 30 min in PBS containing 1% DMSO, 0.1% Tween, and 1% normal goat serum with agitation. Secondary antibody (dilution factor, 1:500; Alexa-Fluor-568–conjugated goat antimouse, Molecular Probes, Invitrogen) was added and incubated overnight. Nuclei were counterstained with 4’,6-diamidino-2-phenylindole (Roche Applied Science) at a final concentration of 13 μM. Samples were washed 5 times for 30 min each and mounted in 70% glycerol. Whole mounts were examined by using a 2-photon point-scanning confocal microscope (LSM 510 META NLO, Zeiss). Six wild-type embryos and 5 bicc1a morphants were examined.

Results

Characterization of a zebrafish bicc1 ortholog.

Using the Basic Local Alignment Search Tool (BLAST),² we compared the mouse Bicc1 (GenBank accession no., NM_031397) nucleotide sequence with sequences in the public database at the National Center for Biotechnology Information (NCBI) and identified a zebrafish bicc1 ortholog (NM_203420; designated here as bicc1a). This bicc1a ortholog maps to zebrafish chromosome 14 (Ensembl assembly, Zv8) and shares 24% overall nucleotide identity with mouse Bicc1. The bicc1a transcript is approximately 2950 bp in length and encodes a protein consisting of 849 amino acids (Figure 1 A).

Figure 1.

Bicaudal C (Bicc1) protein comparison. (A) Comparison of zebrafish, mouse, and human Bicc1 protein sequences. The number immediately after the peptide sequences represents the percentage identity of that sequence compared with the zebrafish sequence, and the number in parentheses represents the overall conservation of the peptide sequence compared with the zebrafish sequence. Light gray blocks of amino acids indicate residues that are conserved between at least 2 of the species, whereas dark gray blocks of amino acids represent residues that are conserved among all 3 species. (B) Schematic of the predicted Bicc1 protein and its functional domains. (C) Comparison of human, mouse, and zebrafish amino acid sequences for Bicc1 functional domains. The percentages of amino acid conservation for the 3 KH domains and the SAM domain are listed.

The zebrafish Bicc1a predicted peptide sequence contains 2 conserved functional domains: 3 K homology (KH) domains in tandem near the N terminus, and a C-terminal sterile α motif (SAM; Figure 1 B). KH domains bind RNA, whereas SAM domains have been implicated in protein–protein interactions during development.^1,50 Overall, the mouse, human, and zebrafish Bicc1 protein sequences share 44% amino-acid identity (Figure 1 A). The level of amino-acid conservation is markedly higher in the KH and SAM domains across all 3 species: 82% in KH1, 87% in KH2, 72% in KH3, and 82% in SAM (Figure 1 C). These data strongly suggest that the functional domains are vital for Bicc1 function.

Zebrafish bicc1a is expressed throughout early development.

To determine the temporal expression pattern of bicc1a in the developing zebrafish embryo, we performed RT-PCR by using total RNA from embryos ranging in age from 1 to 72 hpf. The constitutively expressed gene elongation factor α (ef1α) served as a control.³⁰ bicc1a transcripts were present in all stages tested (Figure 2 A). Expression at the 4- to 8-cell stage (1 hpf) is due to maternally contributed bicc1a transcripts, consistent with BicC expression in Drosophila and Xenopus.^40,59 These results indicate that zebrafish bicc1a is expressed continuously during the period that coincides with pronephros development.

Figure 2.

Gene expression of bicc1a in whole embryos. (A) Primers specific for zebrafish bicc1a or ef1α (control) were used in RT-PCR to amplify total RNA from whole embryos at time points from 1 h postfertilization (hpf) to 72 hpf. Amplicon sizes are 179 bp for bicc1a (upper panel) and 648 bp for ef1α (lower panel). L is a 1-kb ladder molecular size standard (Promega, Madison, WI), and NTC indicates no-template control (negative control). (B) Whole-mount in situ hybridization of 16-hpf embryo. In this lateral view, bicc1a expression is seen at the midbrain–hindbrain boundary (asterisk) and in the developing pronephric duct (arrowheads). (C) Dorsal view of the same embryo, with bicc1a expression seen bilaterally in both pronephric ducts (arrowheads).

To determine the spatial expression pattern of bicc1a, we performed whole-mount in situ hybridization on 12- to 48-hpf embryos. A digoxigenin-labeled probe derived from EST clone AL923800, which contains the nucleotide sequence near the 5′ end of bicc1a and includes part of the KH domains, was used. Signal corresponding to bicc1a transcripts was observed in 10- to 18-hpf embryos in the developing pronephric ducts as well as at the midbrain–hindbrain boundary (Figure 2 B, C).

Knockdown of bicc1a generates kidney cyst-associated morphologic defects.

To test whether bicc1a plays a role in kidney development and physiology, we examined the effects of loss of bicc1a expression by using antisense morpholino oligonucleotides.⁴³ Initially, dose–response experiments were performed to determine the concentration range in which specific, nontoxic effects could be obtained. As a positive control, we used a morpholino oligonucleotide against pkd2 that previously led to kidney cysts in zebrafish.⁵⁴ Injection of bicc1a morpholino oligonucleotides over a broad concentration range (1 to 20 ng/embryo) induced dorsal curvature in the trunk and tail, with the severity of phenotype directly proportional to the dose (Figure 3 D through F; Table 1). Importantly, the abnormal morphology observed in the bicc1a morphants is similar to the phenotype of pkd2 morpholino-injected embryos (Figure 3 C) and is a feature consistently observed in other cystic zebrafish mutants and morphants.^21,43,54 In contrast, uninjected and control oligonucleotide-injected embryos developed normally (Figure 3 A, B). At a dose of 2.5 ng or higher, the trunk phenotype was fully penetrant, with 100% of embryos exhibiting dorsal curvature (Table 1). Morpholino-injected embryos exhibited no morphologic defects up to 30 hpf. The trunk phenotype became visible by 40 hpf and was fully expressed by 48 to 60 hpf, suggesting that this phenotype is a specific effect of morpholino oligonucleotide injection. Many embryos also exhibited various degrees of pericardial edema, but the penetrance of this phenotype was variable between experiments, ranging from 0% to 50%. Therefore, only the trunk phenotype was used as a reliable indicator of bicc1a morphant activity in all subsequent experiments.

Figure 3.

Morphologic defects in bicc1a morphant embryos. Panels show bright-field images of 72-hpf embryos (magnification, ×250). (A) Uninjected fish (control). (B) Control morpholino (MO). (C) pkd2 MO at 1-ng dose. (D) bicc1a MO at 1-ng dose. (E) bicc1a MO at 2.5-ng dose. (F) bicc1a MO at 5-ng dose. No gross abnormalities were detected in uninjected animals or those injected with the control MO. In contrast, embryos injected with the pkd2 or bicc1a MO showed pronounced dorsal curvature of the trunk (asterisk).

Table 1.

Phenotypes observed in morpholino studies using a translation-blocking oligonucleotide toward bicc1a

	Dorsal curvature			Kidney cysts
	% affected	total no. embryos evaluated	no. of experiments	no. of embryos affected	total no. embryos evaluated
Uninjected	0	266	7	0	10
Control MO 1 ng	5.3	112	4	0	4
bicc1a MO 1 ng	96.5	29	1	5	5
bicc1a MO 2.5 ng	100	265	7	13	13
bicc1a MO 5 ng	100	84	3	8	8
bicc1a MO 10 ng	100	84	3	not available
bicc1a MO 20 ng	100	18	1	not available
pkd2 MO 1 ng	100	87	3	14	14
Bicc1 RNA 150 pg	21.6	37	1	0	4

MO, morpholino oligonucleotide

Renal cysts develop in bicc1a morphants.

Although the morphologic defects of bicc1a morphants resemble those previously described in zebrafish mutants and morphants for PKD-causing genes, kidney cysts were not visible grossly in live bicc1a morphant embryos.^20,54 Therefore, we examined cross-sections of 5-d-old zebrafish embedded in JB4 for structural defects in the larval pronephros. In uninjected embryos, the glomerulus and pronephric tubules were normal (Figure 4 A, B), with cellular morphologies resembling those described previously for wild-type embryos.^19,21 In contrast, bicc1a morphants had prominent cysts in the pronephric tubules as well as cystic dilations in the glomerulus and pronephric ducts (Figure 4 C through F). The cysts were lined by epithelial cells, which appeared normal. The cysts were noted in every bicc1a morphant embryo examined (n = 18). These results demonstrate that abrogation of bicc1a expression leads to the formation of cysts in the developing pronephros. Importantly, similar gross morphology and cystic defects were observed by using a second bicc1a morpholino designed to target the splice junction between exons 2 and 3 of bicc1a.(Table 2). This splice-blocking morpholino results in a transcript that mimics the defective transcript in the jcpk mouse. The observation that 2 independent morpholinos targeting the same gene induce a similar phenotype strongly suggests that the formation of pronephric cysts is specific to the loss of bicc1a function.

Figure 4.

Kidney cysts in bicc1a morphant embryos. Histologic analysis of cross-sections from 5-d-old embryos. (A) Control morphant; magnification, ×400. (B) Boxed area in panel A; magnification, ×1000. (C) bicc1a morphant at 1-ng dose of bicc1a MO; magnification, ×400. (D) Boxed area from panel C; magnification, ×1000. (E) bicc1a morphant at 2.5-ng dose of bicc1a MO; magnification, ×400. (F) Boxed area from panel E; magnification, ×1000. Nc, notochord; Pd, pronephric duct; T, pronephric tubule; G, glomerulus; asterisks, cysts. Sections are representative of embryos examined (control MO, n = 6; uninjected, n = 8; bicc1a morphants, n = 18). Scale bar, 50 μm (panels A, C, and E); 10 μm (panels B, D, and F).

Table 2.

Phenotypes observed in morpholino studies using a splice-blocking oligonucleotide toward bicc1a

	Dorsal curvature			Kidney cysts
	% affected	total no. embryos evaluated	no. of experiments	no. of embryos affected	total no. embryos evaluated
Uninjected	0	102	3	0	8
Control MO 1 ng	4	51	3	0	3
bicc1a MO 2.5 ng	52	69	2	not available
bicc1a MO 5 ng	59	85	3	not available
bicc1a MO 10 ng	80	99	3	5	5

MO, morpholino oligonucleotide

Kidney morphogenesis and function appear normal in bicc1a morphants.

To test whether the formation of cysts in bicc1a morphants was accompanied by changes in kidney development and physiology, we stained 48-hpf embryos with anti-Na⁺-K⁺ ATPase α subunit monoclonal antibody α6F, which labels the pronephric ducts and tubules.²¹ The morphology and arrangement of these structures was not affected in bicc1a morphants (Figure 5 A, B), indicating that cyst formation does not result from aberrant development of the pronephros.

Figure 5.

Kidney morphogenesis and function. The upper panels show whole-mount immunostaining of 72-hpf embryos with α6F antibody. Embryos were (A) injected with the control MO or (B) injected with bicc1a MO. Both panels A and B are dorsal views, with anterior to the left. The arrowhead indicates a duct, and the asterisk indicates the location of a paired tubule (out of focus). D, duct; T, tubule. The boxed region (magnification, ×2000) is magnified in the inset (magnification, ×600. A total of 8 embryos were examined in each group, and representative embryos are depicted. (C) Uninjected wild type control (n = 10) and (D) bicc1a morphant (n = 8) animals injected with fluorescently labeled dextran; magnification, ×400. Arrows indicate excretion of the labeled dextran from the cloaca by way of the pronephros. Confocal microscopy (magnification, ×400 with ×3 optical zoom) images of pronephric cilia in 28- to 30-hpf (E) wild-type (n = 5) and (F) bicc1a morphant (n = 6) embryos. Scale bar, 50 μm (panels C and D); 10 μm (panels E and F).

Disruption or blockage of fluid flow can result in the formation of cysts in the pronephric kidney.³⁶ To determine whether the formation of cysts in bicc1a morphants was the result of blocked fluid flow through the pronephros, we examined the excretion of rhodamine–dextran from the cloaca immediately after injection into the common cardinal vein of 3- to 3.5-d-old larvae. In control animals (n = 10), boli of fluorescent material were excreted within 2 to 3 min of injection (Figure 5 C). Rhodamine–dextran also was excreted efficiently by bicc1a morphants (n = 8; Figure 5 D), although the process was slightly slower than in controls, with excretion between 5 to 8 min of injection. These results suggest strongly that the formation of cysts in bicc1a morphants does not interfere with pronephric fluid flow.

A number of PKD-causing genes encode proteins that localize to the primary cilia of renal epithelial cells, suggesting that cilia structure or function plays an important role in the formation of cysts.^{15,16,27,29,34,38,41,45,47,51,56,60,63–65,67} To determine whether the primary cilia are affected in bicc1a morphants, we performed immunostaining using a mouse antiacetylated-α-tubulin monoclonal antibody in 28- to 30-hpf wild-type and bicc1a morphant embryos. Confocal fluorescence microscopy revealed that the length, number, and direction of the cilia in bicc1a morphants (n = 5) closely resembled that of the wild type controls (n = 6; Figure 5 E, F). Taken together, these results indicate that the formation of cysts in bicc1a morphants is not due to overt defects in pronephros development, defects in cilia formation or structure, or inhibition of fluid flow through the pronephros.

Expression of mouse Bicc1 can prevent cyst formation in zebrafish bicc1a morphants.

To further demonstrate that pronephric cysts in bicc1a morphants are specifically due to the knockdown of bicc1a, we tested whether expression of the mouse Bicc1 ortholog could rescue the cystic phenotype in zebrafish. Synthetic full-length mRNA encoding mouse Bicc1 that lacks the bicc1a morpholino-binding site was coinjected with bicc1a morpholino oligonucleotide and embryos were examined histologically for morphologic defects and cyst formation (Figure 6, Table 1). Uninjected and mouse Bicc1 mRNA-injected embryos were indistinguishable up to 5 dpf (Figure 6 A, B, E, F), indicating that injection of mouse Bicc1 mRNA alone has no detrimental effects on larval development or physiology. In embryos coinjected with both the bicc1a morpholino oligonucleotide and mouse Bicc1 mRNA, the occurrence of dorsal trunk curvature was markedly reduced (Figure 6 D), occurring in approximately 60% fewer embryos than those injected with the bicc1a oligonucleotide alone (Figure 6 C). Histologic examination of coinjected embryos (n = 6) revealed reversal of the cystic phenotype, with the embryos more closely resembling the uninjected controls rather than the bicc1a morphants (Figure 6 E, G, H). The ability of mouse Bicc1 to rescue the cystic morphants strongly suggests that the formation of renal cysts in bicc1a morphants is a direct result of the loss of bicc1a and demonstrates that the functional role of Bicc1 is conserved in mouse and zebrafish.

Figure 6.

Mouse Bicc1 rescues the zebrafish bicc1a morphant phenotype. Bright-field images (magnification, ×250) of representative embryos that were (A) uninjected, (B) injected with 150 pg mouse Bicc1 RNA, (C) injected with 2.5 ng zebrafish bicc1a MO, or (D) coinjected with 150 pg mouse Bicc1 RNA and 2.5 ng zebrafish bicc1a MO. Embryos injected with the bicc1a MO showed pronounced dorsal curvature of the trunk (C). Panels E through H show the histologic analysis of cross-sections of representative fish from each treatment group; magnification, ×400. Nc, notochord; Pd, pronephric duct; T, pronephric tubule; G, glomerulus; asterisks, cysts. Scale bar, 50 μm (E through H). (I) Histogram of the number and percentage of animals observed in each group exhibiting normal outward morphology. In the morphants that were coinjected with mouse Bicc1 RNA and bicc1a MO, 56% did not exhibit outward morphologic defects and showed little or no dorsal curvature of the trunk. Statistical analysis using Fisher's Exact nonparametric test⁷ (Sigma Plot) showed a significant difference (*, P < 0.05) between groups.

Discussion

The utility of zebrafish as a model for PKD has been demonstrated in several studies in which PKD-causing genes in mammals, including Pkd2, Nek8, and Invs, have been shown to cause cyst formation in the zebrafish pronephros.^39,45,54 In addition, in several independent genetic screens, at least 20 zebrafish pronephric mutants have been recovered that include cyst formation as a component of their phenotype.^9,21,26 The relatively short developmental cycle of zebrafish and the abilities to readily visualize organs and tissues, test drugs by their addition to the water, and rapidly assay gain- or loss-of-function phenotypes make zebrafish an attractive animal system for studying kidney development and function. The developmental similarities between the zebrafish pronephros and the mammalian kidney are well-established, further strengthening the utility of zebrafish as a nonmammalian model for study of the kidney.

Our data validate the use of a bicc1a zebrafish model as a complementary alternative system in which to study the role of Bicc1 in cystogenesis. A primary goal of this study was to demonstrate conservation of Bicc1 function in the kidneys of zebrafish and mice. We show that the morpholino-induced loss of zebrafish bicc1a results in abnormal morphology and kidney cyst formation. The consequence of the mutation in the jcpk mouse PKD model is a severely truncated protein missing both the KH and SAM domains, and this protein is expected to be functionally null.¹³ Knockdown of zebrafish bicc1a presumably decreases the amount of Bicc1a protein, closely mimicking the situation seen in jcpk homozygous mice. Similar to the phenotype in jcpk mice, bicc1a morphant fish exhibit cysts in all parts of the nephron, including the glomerulus.

In the current study, we used 2 distinct morpholinos that specifically targeted bicc1a. Both yielded similar phenotypes, suggesting that the cystic phenotypes are a direct result of the loss of bicc1a function. Importantly, the cystic phenotype can be reversed with the addition of mouse Bicc1. These data provide further evidence that cyst formation in bicc1a morphants is due to knockdown of bicc1a and that the function of Bicc1 in the kidney is conserved.

Many genes involved in PKD code for proteins that localize to the primary cilia or basal body, suggesting that normal ciliogenesis and normal ciliary structure, and function are important for proper kidney morphogenesis and maintenance.^{15,16,27,29,34,38,41,45,47,51,56,60,63–65,67} Likewise, several cystic zebrafish mutants are due to defects in genes known to be involved in ciliary structure or function.^{36,44,47,54,60} Electron microscopic studies of jcpk homozygous mice reveal morphologically normal cilia present on the apical surface of the renal epithelial cells (data not shown), and the mouse Bicc1 protein does not localize to the primary cilia of mouse inner medullary collecting duct cells.⁵³ In addition, Xenopus embryos injected with morpholinos to knockdown expression of xBicC display a cystic phenotype but do not have any obvious differences in the number or length of cilia.⁵⁷ Likewise, assessment of the pronephric cilia in zebrafish bicc1a morphants revealed no apparent defects in cilia, and fluid flow through the pronephros was not compromised. Given all these independent data, we propose that Bicc1 is not involved directly in the structure or formation of cilia but instead acts in a downstream or parallel pathway that, when perturbed, leads to cystogenesis.

Drosophila BicC localizes RNA and regulates translation by interacting with components of the polyadenylation machinery and associating with specific mRNAs, including the 5′ untranslated region of its own mRNA.^11,37,49,52 Studies in Xenopus suggest that the KH RNA-binding domains are important for Bicc1 function in the pronephros.⁵⁷ Bicaudal C binds RNA through its KH domains,^11,37,49,52 and a recent report indicates that Xenopus Bicc1 functions to repress translation of the PKD-protein polycystin 2.⁵⁸ Our previous work has demonstrated that Bicc1 interacts with another PKD-related protein, SamCystin, in a RNA-dependent manner,⁵³ however, the precise molecular function of mammalian Bicc1 in the kidney remains unclear. The ability to exploit the zebrafish as an alternative model system in which to study Bicc1 specifically in the kidney will allow us to readily and rapidly determine the functional role of Bicc1 in cyst formation and PKD pathogenesis.