The zebrafish kidney mutant zeppelin reveals that brca2/fancd1 is essential for pronephros development

Abstract

The zebrafish kidney is conserved with other vertebrates, making it an excellent genetic model to study renal development. The kidney collects metabolic waste using a blood filter with specialized epithelial cells known as podocytes. Podocyte formation is poorly understood but relevant to many kidney diseases, as podocyte injury leads to progressive scarring and organ failure. zeppelin (zep) was isolated in a forward screen for kidney mutants and identified as a homozygous recessive lethal allele that causes reduced podocyte numbers, deficient filtration, and fluid imbalance. Interestingly, zep mutants had a larger interrenal gland, the telostean counterpart of the mammalian adrenal gland, which suggested a fate switch with the related podocyte lineage since cell proliferation and cell death were unchanged within the shared progenitor field from which these two identities arise. Cloning of zep by whole genome sequencing (WGS) identified a splicing mutation in breast cancer 2, early onset (brca2)/fancd1, which was confirmed by sequencing of individual fish. Several independent brca2 morpholinos (MOs) phenocopied zep, causing edema, reduced podocyte number, and increased interrenal cell number. Complementation analysis between zep and brca2^ZM_00057434-/- zebrafish, which have an insertional mutation, revealed that the interrenal lineage was expanded. Importantly, overexpression of brca2 rescued podocyte formation in zep mutants, providing critical evidence that the brca2 lesion encoded by zep specifically disrupts the balance of nephrogenesis. Taken together, these data suggest for the first time that brca2/fancd1 is essential for vertebrate kidney ontogeny. Thus, our findings impart novel insights into the genetic components that impact renal development, and because BRCA2/FANCD1 mutations in humans cause Fanconi anemia and several common cancers, this work has identified a new model to further study brca2/fancd1 in disease.

1. Introduction

The renal system is crucial for homeostasis because the kidneys filter the blood and reabsorb necessary metabolites and solutes while collecting the remaining waste products for their ultimate excretion. Vertebrate species form up to three kidneys of increasing complexity during their development, which can be vestigial or operational as these forms progressively degenerate and disappear when the subsequent organ form emerges (Dressler, 2006). These renal iterations arise from the intermediate mesoderm (IM), and are known as the pronephros, mesonephros and metanephros, respectively (Saxen, 1987). Each of these kidney forms contains individual excretory units known as nephrons, which typically have three parts termed the glomerular blood filter, tubule and duct, though exceptions such as aglomerular fish are existent (Reimschuessel, 2001).

The embryonic zebrafish kidney is a functional pronephros that contains two nephrons that join anteriorly at a single fused glomerulus and posteriorly at the cloaca (Drummond et al., 1998). Although this form of the kidney is much less complex than the metanephric kidneys found in amniotes, the segmental organization of the pronephros is relatively conserved with other vertebrate species, availing the opportunity to use zebrafish genetics for nephrogenesis research and renal disease modeling as well (Poureetezadi and Wingert, 2016). For example, zebrafish pronephric tubule segments are delineated by the expression of a specific cadre of solute transporter proteins that likewise typify the equivalent segment region in mice and humans (Wingert and Davidson, 2008). In addition, the zebrafish pronephric blood filter has similar components, including capillaries with a fenestrated endothelium, epithelial cells known as podocytes, and a specialized glomerular basement membrane (GBM) (Kroeger and Wingert, 2014), which show a conserved ultrastructure with mammals (Zhu et al., 2016). These parts are anatomically situated within the glomerulus such that the podocytes surround the capillaries and extend intricate cellular extensions known as foot processes, which interdigitate to form unique cell junctions known as the slit diaphragm (Wiggins, 2007). These junctions will express Nephrin and Podocin in mature podocytes and are crucial for the filtration function of the nephron. The combination of podocytes and the GBM forms a fine molecular sieve, allowing small compounds and solutes to pass into the tubule, while retaining cells, platelets, and large molecules, such as albumin, in circulation.

Clinical studies have shown that approximately 80% of end stage renal disease (ESRD) originates with podocyte dysfunction, as disruption of podocytes compromises nephron function and initiates a domino effect of cellular and molecular changes that trigger nephron damage (Wiggins, 2007). The dysfunction or loss of podocytes has such a drastic effect on kidney health, and hence health of the individual, because differentiated podocytes do not proliferate under normal conditions (Nagata et al., 1998). Indeed, mammals form all their podocytes during development, and these cells cannot be replenished with age (Kriz, 2012; Nagata, 2016). Extensive research has revealed that when podocytes are lost, the neighboring podocytes undergo hypertrophy to occupy the place of the missing cells (Kriz, 2012; Nagata, 2016). Thus, studying early development of podocytes may lend insights for future therapeutics that stimulate podocyte renewal.

In the zebrafish embryo, podocytes arise as two bilateral clusters of cells adjacent to the paraxial mesoderm that forms the third pair of trunk somites, in a region where the expression domains of the Wilms tumor 1 (WT1) paralogs, wt1a and wt1b, overlaps (Serluca and Fishman, 2001; Bollig et al., 2006). During early development these groups of cells coalesce and migrate toward the midline, recruiting vasculature and forming a fused glomerulus (Drummond et al., 1998). A number of transcription factors and signaling molecules are known to be expressed in podocytes, like lhx1a and mafba, and several have been identified to play important roles in podocyte development, such as wt1a, wt1b, foxc1a, lmx1b.1, osr1 and the Notch pathway effector encoded by rbpj (Bollig et al., 2006; Bollig et al., 2009; Gerlach and Wingert, 2013; Krauss et al., 1991; He et al., 2014; O'Brien et al., 2011; Picker et al., 2002; Swanhart et al., 2010; Tomar et al., 2014; Toyama and Dawid, 1997). Further, these crucial genetic factors are associated with various renal defects and disease states. For example, mutations in the human WT1 cause Wilms Tumor disease, a pediatric cancer of the kidney (Call et al., 1990).

Biochemical studies have suggested that several of these integral factors interact, forming a large multimeric protein complex that regulates podocyte differentiation (O'Brien et al., 2011). As the podocytes mature, expression of the slit diaphragm proteins, such as Nephrin, Podocin, CD2AP, and α-actinin4, is essential for proper glomerular assembly (De Zoysa and Topham, 2005; Huang et al., 2013; Kramer-Zucker et al., 2005; Zhou and Hildebrandt, 2012). The proteins described above are not the only factors implicated in vertebrate podocyte development: the morphogen retinoic acid (RA) is also essential for proper podocyte genesis, acting during proximo-distal pattern formation of the zebrafish IM to specify the podocyte lineage (Wingert et al., 2007; Wingert and Davidson, 2011). RA binds retinoic acid and retinoid X receptors (RARs, RXRs), to interact with RA response elements (RAREs) within DNA to regulate transcription of target genes (Maden, 2002). RA synthesis and degradation enzymes establish discrete sources and sinks of RA to form signaling gradients (Duester, 2008). During zebrafish pronephros formation, treatment with exogenous all-trans RA proximalizes the nephron, while the inhibition of RA biosynthesis distalizes the nephron, including the loss or reduction in podocyte number depending on severity of RA deficiency (Wingert et al., 2007; Wingert and Davidson, 2011). Interestingly, subsequent research uncovered a functional RARE in both the zebrafish wt1a and human WT1 enhancers (Bollig et al., 2009).

Closely intermingled with developing podocytes in the zebrafish embryo is a group of cells that gives rise to the interrenal gland that forms adjacent to the pronephric glomerulus (Hsu et al., 2003). The interrenal gland secretes steroid hormones, and is akin to the mammalian adrenal gland (Hsu et al., 2003). Similar to podocytes, the interrenal lineage is held to be derived from the wt1a expressing field, however it develops from a subset of this field that is triggered to express the steroid hormone receptor encoded by nuclear receptor subfamily 5, group A, member 1a (nr5a1a, formerly ff1b) (Hsu et al., 2003; Bollig et al., 2006). Prior studies have revealed a delicate balance between the specification of these lineages that needs to be maintained: a paucity of wt1a expression or reduction of Notch signaling leads to concomitant reductions in podocyte numbers and more interrenal cells (O'Brien et al., 2011; Miceli et al., 2014). These data suggest that the podocyte and interrenal lineages have a common precursor, or that one fate may transdifferentiate between the other fate depending on the combination of internal factors and external cues.

Here, we describe the novel zebrafish podocyte mutant zep, isolated in a F3 forward genetic screen using the chemical mutagen N-ethyl-N-nitrosurea (ENU), in which families that displayed congenital edema were further analyzed for kidney defects (Wingert and Davidson, 2011). From expression and functional analyses, we discovered that zep mutants have a dysfunctional glomerulus due to the loss of podocytes. Interestingly, while zep and wild-type (WT) embryos showed no difference in proliferation or cell death in the IM or subsequent pronephros, zep embryos had an increased interrenal gland size. Using WGS, we found that zep have a lesion in the gene breast cancer 2, early onset (brca2)/fancd1, which is broadly expressed during embryogenesis and maternally inherited as well. Through knockdown studies, we demonstrate that brca2 deficient embryos phenocopy zep in the acquisition of edema, reduction in the number of podocytes, and increased number of interrenal cells. Complementation studies between zep and brca2^ZM_00057434-/- fish, which possess an insertional mutant allele (Rodriguez-Mari et al., 2011), revealed that the combination of these genetic defects led to increased interrenal gland size, though podocyte development was not visibly altered. Through rescue studies, we found that brca2 overexpression was sufficient to partially restore podocyte formation in zep. Taken together, these data support the conclusion that deficiency of brca2 is sufficient to cause the zep phenotype. This study identifies for the first time that disruption of brca2 can lead to alterations in podocyte development, implicating this gene to be an essential renal factor critical for further study.

2. Materials and methods

2.1 Zebrafish maintenance and ethics statement

Zebrafish were cared for in the Center for Zebrafish Research at Freimann Life Sciences Center at the University of Notre Dame, based on Institutional Animal Care and Use Committee approved procedures outlined in protocols 13-021 and 16-025.

2.2 Live imaging

Embryos were kept in E3 media and incubated at 28°C until the desired time point. The embryos were then placed on a depression slide in 2% methylcellulose/E3 and 0.02% tricaine for imaging.

2.3 Dextran conjugated FITC injections

WT and zep mutant embryos were treated with 0.003% phenylthiourea (PTU; P7629) in E3 beginning at 24 hpf to reduce pigmentation. For microinjections, embryos were anesthetized in 0.02% tricaine and 40 kDa dextran-FITC molecules (Invitrogen D-1845) at a concentration of 5 mg/ml were injected into the circulation of 3 days post fertilization (dpf) zebrafish embryos. Embryos were transferred to fresh 0.003% PTU/E3 and kept in the dark at 28°C for 24 hours post injection, then imaged.

2.4 Whole mount and fluorescent in situ hybridization (WISH, FISH)

WISH was performed on WT and zep mutant embryos as previously described (Cheng et al., 2014). Antisense probes were prepared with digoxygenin or fluorescein labeled wt1a, wt1b, foxc1a, mafba, nphs1, and nr5a1a using template plasmids as previously described (Wingert et al., 2007; O'Brien et al., 2011). FISH was performed as previously described (Brend and Holley, 2009). Volumetric measurements were performed on interrenal glands analyzed by FISH using the Nikon C2 upright confocal microscope and Nikon NIS-elements software.

2.5 Chemical treatments

WT and zep mutant embryos were incubated in E3 media with control dimethyl sulfoxide (DMSO) vehicle, exogenous all-trans RA (Sigma R2625), or exogenous DAPT (EMD Millipore) dissolved in DMSO. RA was diluted to concentrations of 1×10^-6 M, 1×10-⁷ M, 1×10-⁸ M, and 1×10^-9 M in E3 (Wingert et al., 2007; Lengerke et al., 2011). DAPT was diluted to 100 μM in E3. Embryos were treated at 60% epiboly (RA) or 75% epiboly (DAPT) and placed at 28°C in the dark until the following morning, when the drug was replaced with standard E3 embryo media. The embryos were then fixed in 4% paraformaldehyde (PFA) at the developmental time point of 24 hpf and WISH was performed.

2.6 Cell proliferation, cell death and DNA damage assays

For whole mount immunofluorescence on WT and zep mutants, embryos were fixed overnight and washed in phosphate buffered saline (PBS) and then molecular grade water. Embryos were dehydrated in 100% methanol and placed at -20°C for at least 1 hour. The embryos were then stepwise rehydrated and blocked for 1 hour at room temperature (0.1 g BSA, 100 μl Triton x-100, 100 μl Tween-20, 100 μl DMSO, 500 μl fetal calf serum (FCS) filled to 10 ml with 1x PBS). Directly following blocking, primary antibodies were incubated at 4°C overnight: phospho-Histone H3 antibody diluted 1:200 (Millipore 06-570); anti-Caspase-3 diluted 1:100 (BD Biosciences 559565); anti-Histone H2A.X diluted 1:200 (GeneTex 127342). The following day anti-mouse or anti-rabbit secondary antibody was diluted 1:500 (Alexa Fluor, Invitrogen) and incubated at 4°C overnight. The embryos were then mounted for imaging. For acridine orange (AO; Sigma A6014) staining on WT and zep mutants, a 50x stock (250 μg/ml) of AO was diluted in 0.003% PTU. At specific time points, the 0.003% PTU was removed and replaced with the AO staining solution at room temperature. The embryos were washed three times with 0.003% PTU and imaged in 0.02% tricaine/2% methylcellulose/E3.

2.7 Chromogenic staining of 3β-hyrdoxysteroid dehydrogenase (3β -HSD)

Embryos were treated with 0.003% PTU at 24 hpf and allowed to develop an additional 24 hpf before fixation in 4% PFA for 1 hour at room temperature or 4°C overnight. Chromogenic staining was performed as described (Chai et al., 2003). In brief, the embryos were washed and the staining solution was prepared by combining 40 μl DMSO and 10 μl 0.1 mg/ml dehydroepiandrosterone (DHEA; Sigma D4000), then mixing in 6 μl 0.2 mg/ml NAD (Sigma N1511) diluted in 0.1 M phosphate buffer, pH 7.2 (Sigma p3288). The remainder of the staining solution was made in a separate tube and was composed of 941 μl PBS, 1 μl 0.1 mg/ml nicotinamide (B3; Sigma 72340) and 2 μl 0.1 mg/ml nitroblue tetrazolium chloride (NBT; Sigma N6876). The stain was left for 3-8 hours, quenched by two PBST washes and samples were fixed in 4% PFA. Area was obtained using imageJ and a Student t-test was performed to analyze results.

2.8 WGS and genotyping

WGS was performed as previously described (Leshchiner et al., 2012). Sequencing was performed on pools of 40 zep mutants and 40 phenotypically WT zep siblings as identified by WISH analysis for expression of the podocyte-specific gene, wt1b. DNA was isolated using the DNAeasy blood and tissue kit following the enclosed protocol (Qiagen 69504). Data was interpreted using SNPtrack (Leshchiner et al., 2012; Ryan et al., 2013). Genomic DNA of individual zep zebrafish was isolated and PCR amplifications of the brca2 locus were performed using the following primers: forward 5′ ATTTCAGCCAACAGCACAAGACGAGCTCGGTGGGACACGA- 3′, and reverse 5′ CTCAACAGCAGCATCTTGCTCCATTGCCTCATACAACTCA - 3′. Products were purified with the Qiagen MinElute PCR Purification Kit as per manufacturer instructions, and sequenced with the forward primer (Genomics Core, University of Notre Dame). brca2^ZM_00057434-/- were genotyped for the insertional allele using PCR with the primers forward 5′-CCACCTTCGACCTTGAGCCTAAAA-3′ and reverse 5′-GAAGCGAGAAGCGAACTGATTGGT-3′, and for the WT allele using PCR with the following primers, forward 5′- GCAGGTTGTGATGAAGCCACC-3′ and reverse GTGGTGTGAGGCCAGAGGTT-3′. Genotyping of the transgenics hsp70∷Gal4 and UAS∷NICD was performed as described (Drummond et al., 2017).

2.9 MO knockdown, reverse-transcription polymerase chain reaction (RT-PCR) and rescue studies

MOs were purchased from Gene Tools, LLC (Philomath, OR) and diluted in molecular grade water forstorage at -20°C. brca2 was knocked down using a translation blocking MO (5′- TTTCAAACATGCTGCCATGACTGTG - 3′), as well as 2 independent splice-inhibiting MOs designed between exons 20 and 21 (splice acceptor 5′- GCTTTTTCTTTTTACCTTATATGAG - 3′ and splice donor 5′-TAATGCAATCCCTACCCTACCCTCCATTTC -3′). Transcript analysis of brca2 splicing in WT, zep and brca2 splice morphants was performed using RT-PCR (Galloway et al., 2008). In brief, RNA was isolated from pools of 20-30 embryos, cDNA synthesized using random hexamers (Superscript IV, Invitrogen), and PCR performed with the zep genotyping primers (specified above). For rescue studies, full-length brca2 capped RNA (cRNA) was synthesized in vitro (Ambion) using an expression plasmid (Genscript) and applied in dosages ranging from 50-1500 picograms (pg). For all microinjections, embryos were manipulated at the 1-cell stage with MO or cRNA, then raised at 28°C until the desired stage for phenotypic analysis.

2.10 Histological analysis and imaging

Embryos were embedded for JB4 sectioning and sections were counterstained as described (Gerlach et al., 2014). Images for all studies were obtained on a Nikon Eclipse 80i or Eclipse Ni microscope with a Nikon DS-Fi1 or DS-Fi2 camera. Confocal images were obtained using a Nikon C2. Images were then processed with Adobe Photoshop CS5.

3. Results

3.1 zep encodes a recessive lethal mutation that is associated with the appearance of progressive edemaand a glomerular filtration defect

To broadly survey for genetic components necessary for renal development, an F3 forward genetic screen was performed in zebrafish to isolate families with recessive mutations associated with embryonic edema, as this morphological defect can be symptomatic of kidney failure through compromised development of the glomerular blood filter. A previously published edema mutant from the screen, lightbulb (lib), was shown to be deficient in RA biosynthesis due to a defect in the gene aldehyde dehydrogenase 1a2 (aldh1a2), suggesting the validity of the screening procedure (Wingert and Davidson, 2011). lib presented with edema at 48 hours post fertilization (hpf) along with reduced podocyte numbers (Wingert and Davidson, 2011). However, in comparison to lib and other screen families that presented severe edema at 48 hpf (data not shown), zep was unique in the late-presentation of overt edema at 96 hpf, and was selected for further phenotypic characterization.

During subsequent live time course studies, WT developing zebrafish displayed no edema throughout their lifespan depicted from 48 through 120 hpf (Figure 1A). Similarly, zep embryos were robustly healthy at 48 hpf with no edema (Figure 1A). However, the initial acquisition of slight to moderate edema in some zep embryos was observed by 72 hpf within the pericardial cavity (Figure 1A). By the 96 hpf stage, edema in zep embryos was observed consistently, and subsequently progressed in severity in the pericardial region as well as expanded to other tissues such as the trunk through 120 hpf (Figure 1A). This progressive edema was correlated with lethality between 6-10 dpf (data not shown).

Figure 1. zep mutants acquire edema and display altered fluid clearance.

(A) Compared to live WT siblings, zep mutant embryos have acquired an edematic phenotype that progresses in severity throughout the lifespan of the animal. In zep, the location of the pericardial cavity (black arrowhead) showed no edema at 48 hpf, while subtle edema was detected at 72 hpf which progressed in severity at 96 and 120 hpf. (B) Following microinjection of 40 kDa dextran-FITC into the vasculature at 72 hpf, WT siblings exhibited endocytosis by the proximal tubule (white box and inset) at 96 hpf, while zep embryos showed accumulation of fluorescence in edemic areas and the yolk (asterisk) and failed to exhibit proximal tubule uptake. Embryos are shown in lateral views.

As edema is a correlative symptom of kidney dysfunction, we performed a dextran-FITC injection assay to analyze renal clearance in zep along with the functionality of the glomerular filtration barrier. In this experiment, 40 kDa dextran-FITC was injected into the circulation of 3 dpf zebrafish. In WT embryos observed after 24 hours, the pronephric glomerulus was able to clear the dextran moiety from the circulation where it was partly reabsorbed in the proximal tubule and also excreted (Figure 1B), phenotypes that have been well-established by prior studies (Hentschel et al., 2007; McCampbell and Wingert, 2014). Conversely, when zep mutants were injected with the 40 kDa dextran-FITC molecules and observed after 24 hours, there were copious amounts fluorescence within the edemic regions and the embryos were deficient in proximal tubule reabsorption (Figure 1B). These data revealed that normal fluid flow, including glomerular filtration, was compromised in zep mutants.

3.2 zep mutants show a reduction or complete absence of podocyte-related gene expression

Given the renal filtration defect observed in zep zebrafish embryos, we next examined development of the pronephric glomerulus beginning with the pattern formation of podocytes. The emergence of podocyte precursors from the IM constitutes the initial stage of glomerular formation (Figure 2A) (Serluca and Fishman, 2001). The subsequent migration of these podocyte precursors to the midline and their progressive differentiation, including their expression and secretion of Vascular endothelial growth factor (Vegf), is required for recruitment of blood vessels to form the glomerulus by 48 hpf (Figure 2A) (Majumdar and Drummond, 1999). The blood filter structure is also elaborated by the establishment of the GBM, due to the composite activities of podocytes and endothelial cells in secreting this extracellular matrix.

Figure 2. zep mutants fail to normally develop the podocyte lineage during nephrogenesis.

(A) Location of podocyte and interrenal precursors and the result of their morphogenesis during early zebrafish embryonic development. (Left) The podocyte lineage (P, red) emerges rostral to the tubule lineages (T, purple) of the intermediate mesoderm at the 15 ss, while precursors of the interrenal gland (IR, green) are interspersed in the local vicinity of the podocyte precursors. (Right) By 48 hpf, morphogenesis events result in a single IR that is situated posterior to the fused glomeruli of the renal corpuscle (RC). Embryo drawings show lateral and dorsal views, with enlarged regions showing dorsal views. (B) Compared to WT siblings, zep mutant embryos have extremely reduced or absent expression of a suite of podocyte markers at the 24 and 48 hpf stages. Embryos are shown in dorsal views with each indicated gene expression stained in purple, where black boxes demarcate the cervical region where the podocytes develop.

To examine podocyte specification and differentiation in zep embryos, clutches obtained from pairwise matings of heterozygous zep adult zebrafish were collected and processed for whole mount in situ hybridization (WISH) to study the expression of a panel of previously established markers in developing WT and mutant podocytes (O'Brien et al., 2011). Between the 15 somite stage (ss) and 24 ss, zep embryos had strongly reduced or completely abrogated expression of wt1b compared to WT siblings (data not shown). By 24 hpf, WT embryos have two circular bilateral clusters of podocyte cells located laterally to somite 3 that can be identified by expression of the markers wt1a, wt1b, foxc1a, mafba, and nphs1 (Figure 2B).

Compared to WT embryos, zep had a strong reduction or complete loss of these markers at 24 hpf, while expression of some of these markers in other tissues, such as the central nervous system, was unchanged (Figure 2B). In WT embryos, the expression of the podocyte markers was maintained through 48 hpf, as they migrated closer to the midline (Figure 2B). Notably, zep mutants did not gain expression of any these transcripts at 48 hpf through the 5 dpf stage (Figure 2B, data not shown). These finding suggest that the reduction and loss of these gene expression markers in zep is not due to developmental delay, but rather that the podocytes are absent in zep mutants.

To further examine the pronephros in zep, clutches obtained from pairwise matings of heterozygous zep adult zebrafish were collected to assay tubule development by WISH. During zebrafish embryogenesis, mesenchymal renal progenitors undergo an epithelial transition to form initially linear pronephric tubules by 24 hpf (Gerlach and Wingert, 2014; McKee et al., 2014). Thus, we used a riboprobe to label cells expressing cdh17, which is expressed throughout the tubule and duct of the pronephros at 24 hpf, along with a ribroprobe designed to detect wt1b transcripts, to label podocytes, and then measured absolute tubule length from the rostral to caudal extent (Poureetezadi et al., 2016). Compared to WT embryos, the zep-/- developed tubules that were statistically the same length, while they lacked bilateral clusters of wt1b+ expressing cells as we had observed previously (Figure 3A,B). Next, histological analysis of WT and zep-/- at the 5 dpf stage was performed following JB4 sectioning to assess glomerulus and lumen architecture (Figure 3C). While WT embryos formed a midline-situated glomerulus with flanking tubules, zep-/- lacked any discernible blood filter and the tubules displayed a distended appearance consistent with the morphological presence of edema (Figure 3C). Taken together, these studies suggest that tubule development initially proceeds normally in zep, though the pronephric tubules become disfigured at later stages concomitant with fluid accumulation.

Figure 3. zep mutants show normal tubule development and fail to form a normal glomerulus.

(A) WT and zep mutant embryos exhibit similar pairs of nephron tubules, which (B) are not statistically different in length. Embryos are shown in dorsal views. (C) Histological analysis of cross-sections from the cervical region of WT and zep embryos at 5 dpf. zep lack a midline glomerulus (G) ventral to the notochord (N), and the flanking tubules (T) are distended compared to WTs, consistent with the morphological appearance of edema.

3.3 Podocyte loss associates with an expansion in the interrenal lineage in zep rather than alterations in progenitor proliferation or survival

Next, to examine if the abrogation of podocytes in zep might be associated with changes in the expansion or survival of these cells, we conducted proliferation and apoptosis assays between the 15 ss and 24 hpf. Tissue proliferation was assessed by whole mount phospho-Histone H3 (anti-pH3) immunohistochemistry. There was no discernible qualitative difference between WT siblings and zep embryos in their pattern of pH3 staining, suggesting that a change in proliferation was not associated with the loss of podocytes in the mutants (Figure 4A, data not shown). Quantification of pH3+ cells in the cervical region of the embryo where podocytes are located confirmed that there was no statistically significant difference between WT and zep mutant embryos (Supplemental Figure 1A). To investigate cell death in zep, acridine orange (AO) staining was performed because this method robustly labels apoptotic cells. Using this assay, we observed no discernible qualitative difference between the WT and zep mutant embryos (Figure 4B, data not shown). Further, we quantified AO staining in other embryonic tissues, which similarly revealed that there was not a statistically significant difference between the WT and zep mutant embryos, and also performed whole-mount immunofluorescence to detect activated Caspase 3, which revealed no differences (Supplemental Figure 1B-E). The combination of these studies led us to conclude that neither elevated cell death nor reduced cell proliferation is associated with the developing pronephros or other nearby mesoderm populations in zep.

Figure 4. Reduction of podocytes in zep mutants occurs concomitantly with an expansion of the interrenal lineage.

WT siblings and zep mutants have no difference in (A) cell proliferation based on pH3 staining and showed no differences in the amount of cell death (B) based on AO staining in the area where podocytes and interrenal cells arise. In contrast, (C) WT siblings have bilateral clusters of podocytes (P), identified by wt1b transcript localization (purple), and a centralized interrenal (IR) gland at the midline, identified by localization of nr5a1a transcripts (red) at 36 hpf. zep mutants show abrogated wt1b expression and a distinctly increased region of cells that express nr5a1a transcripts. (D) The region of cells exhibiting 3β-HSD staining is increased in zep mutants compared to WT siblings, with the area quantified (E). (F,G) There is significant increase of nearly two-fold in the volume occupied by the hormone secreting interrenal gland cells in zep mutants compared to WT, as visualized by FISH and confocal microscopy. Asterisks (*) indicate p<0.001 using student T-test. Embryos are shown in dorsal views, where the white or black boxes demarcate the cervical region where podocytes and the interrenal gland develop. Insets show digital zoom view of 150%.

Due to the absence of detectable cellular changes in zep, we next examined the possibility of an alteration in progenitor fate. Previous research has demonstrated that genetic manipulations that reduce podocyte formation, including the knockdown of wt1a, can be accompanied by expansions in the interrenal lineage during zebrafish embryogenesis (O'Brien et al., 2011). Given the reduction of wt1a expression in zep embryos, we hypothesized that the interrenal fate was also altered. Thus, we assayed expression of wt1b and nr5a1a by double WISH in WT siblings and zep mutant embryos at 36 hpf, because transcripts of these genes are specific to the podocytes and interrenal lineages, respectively. The WT siblings had distinct bilateral podocytes, and a clearly defined region of the interrenal gland (Figure 4C). Conversely, in the zep mutants, there was an absence of wt1b expression accompanied by a vast increase in the region of cells expressing nr5a1a transcripts (Figure 4C). These data suggest that the formation of the interrenal cells in zep is related to their defect in podocyte formation, namely that interrenal expansion occurred at the expense of podocytes.

3.4 Chromogenic staining indicates an increase in differentiated, functional interrenal gland cells in zep

To further characterize the development of the interrenal gland in zep, 3β-hydroxysteroid dehydrogenase (3β-HSD) chromogenic staining was performed on WT siblings and zep mutant embryos to assess whether the expansion of the interrenal lineage was followed by normal differentiation of these cells. In this chromogenic assay, hormone-secreting cells of the mature interrenal gland are labeled (Chai et al., 2003). Compared to WT siblings, zep embryos displayed a larger area of 3β-HSD staining (Figure 4D) that was statistically significant (Figure 4E). These data demonstrate that the increased region of the nr5a1a expressing cells in zep progressed in development to become functional gland cells.

In addition, we performed fluorescent in situ hybridization (FISH) and used confocal microscopy to quantify the changes in interrenal gland volume in both WT siblings and zep mutant embryos. These analyses revealed that there was a statistically significant increase in the overall volume of the interrenal precursors, as designated by nr5a1a transcript expression (Figure 4F, 4G). Taken together, these data further support the conclusion that zep have a significant alteration of both interrenal precursors and functional gland cells. Given the close anatomical relationship of the interrenal and podocyte precursors, these findings are consistent with the notion that the expanded interrenal organ in zep can be explained by a switch in the fate of intermediate mesoderm progenitors from a podocyte lineage to an interrenal one.

3.5 WGS reveals the zep mutant involves a splicing site nucleotide change in brca2

To identify the gene responsible for the zep phenotype, we performed WGS (Leshchiner et al., 2012). Samples were analyzed for differences in single nucleotide polymorphisms (SNPs) and this indicated that the gene responsible for the zep mutant was located on chromosome 15 (Figure 5A). Through analyzing the SNP track record, the candidate brca2/fancd1, was noted as scoring highly using previously described metrics to restrict results (Ryan et al., 2013). Prior characterization of zebrafish brca2 has revealed that it contains 26 exons, with a total transcript length of 9167 basepairs that encodes a protein with 2874 amino acids (Rodriquez-Mari et al., 2011). The zep candidate lesion in brca2 revealed by WGS analysis was a T to C change located in the splice acceptor site between exons 20 and 21, thus predicted to interfere with mRNA processing (Figure 5B, 5C). Gene expression analysis of brca2/fancd1 in zebrafish has shown that transcripts are maternally deposited and zygotically produced throughout the embryo until the 24 hpf stage, with slightly stronger expression in the IM at 16 hpf, and subsequently expressed in the heart at the 48 hpf stage and then the brain at the 72 hpf stage (Titus et al., 2009; Shive et al., 2010). Analysis of brca2 expression in WT embryos confirmed the broad expression throughout tissues, with somewhat elevated expression in the location of the IM at the 12 ss, though we were unable to discern higher expression in putative podocyte or interrenal precursors specifically (Figure 5D).

Figure 5. WGS identifies that the zep mutation is located in the brca2 gene.

(A) SNP analysis localized the zep lesion to chromosome 15. Upon further analysis of the region of interest, (B) zep mutants were found to have a T to C mutation at the splice acceptor site between exons 20 and 21 of the brca2 gene. (C) Schematic of the brca2 gene with the zep and brca2^ZM_00057434 mutant alleles indicated. (D) brca2 expression is ubiquitous between the 12 ss and the 28 ss, though slightly elevated transcript levels are observed in the region of IM at 12 ss where the proximal pronephros and interrenal gland emerge (boxed area). Embryos shown in lateral views.

Next, individual WT and zep embryos were sequenced, and we confirmed that the mutation was correlated with the edemic phenotype (data not shown). Using RT-PCR to perform transcript analysis, we detected brca2 intronic sequence corresponding to the region between exons 20 and 21 from total RNA isolated from zep embryos, indicating that splicing was aberrant (Supplemental Figure 2A). Inclusion of this intronic sequence introduces a premature stop codon that is predicted to encode a truncated Brca2 peptide in which 451 residues at the C-terminus would be eliminated, thereby causing the loss of the predicted DNA binding domain based on prior annotation (Rodriquez-Mari et al., 2011).

When we performed WISH to assess brca2 transcript levels in clutches obtained from incrosses of heterozygous zep adults, we detected no spatiotemporal differences in brca2 expression in zep-/-compared to WT siblings between the 15 ss and the 72 hpf stage (data not shown), which suggested that nonsense mediated decay of brca2 mRNA was not significant in zep mutant embryos. Taken together, we concluded that zep encodes a mutation capable of altering brca2 splicing, though further molecular and biochemical work is needed to specifically define the properties of the altered protein product during development.

Previous work has established that brca2 is required for genome stability in zebrafish (Rodriguez-Mari, et al., 2011), and BRCA2 has been well established to function in homology-directed recombination to help repair DNA breaks (Titus, et al., 2009). Therefore, we used a zebrafish-specific anti-phospho-H2AX antibody (P-h2A.X) (Sidi, et al., 2008) to assess whether double stranded DNA repair is faulty in zep-/- mutants. Compared to WT controls, zep-/- mutants displayed elevated numbers of P-h2A.X+ cells throughout embryonic tissues at the 24 hpf stage (Figure 6A). The distribution of P-h2A.X+ cells in zep was not visibly enriched in the pronephros or interenal regions at this time point (Figure 6A). Statistical analysis determined that the difference in numbers of P-h2A.X+ cells was significant (Figure 6B). These data suggest that zep embryos have reduced DNA repair, and are consistent with the notion that Brca2 function is compromised as a result of the zep lesion.

Figure 6. zep mutants exhibit elevated DNA damage.

(A) WT and zep embryos were double labeled to detect podocytes by WISH for wt1b transcripts followed by whole mount immunohistochemistry to detect phosphorylated h2A.X (P-h2A.X), a marker of double strand breaks. Top panels show a dorsal view of the cervical region where podocytes develop, indicated by the white boxed region. Bottom panels show a lateral view of the tail, which was the area utilized for quantification. (B) Quantification of P-h2A.X+ cells in the tail region revealed a statistically significant increase in zep compared to WT. Asterisk (*) indicates p<0.0001 using student T-test.

3.6 brca2 knockdown recapitulates zep characteristics

To independently assess the result of Brca2 deficiency on pronephros development in the zebrafish, we performed knockdown studies and compared brca2 morphants to WT and zep mutant embryos. As previously described, WT zebrafish did not have any edema present (Figure 1A), however zep mutants acquired edema by 72 hpf (Figure 1A). Similarly, injection of a brca2 MO targeting the ATG start site caused edema by 72 hpf (Supplemental Figure 2B). Additionally, two splice blocking MOs were designed to target the splice acceptor and donor sites located at exons 20 and 21, respectively, and these similarly induced edema at 72 hpf (Supplemental Figure 2B). Microinjection of the exon 21 splice acceptor site blocking MO alone (brca2 MO2, which targets the site of the genetic lesion in zep) was sufficient to induce edema by 72 hpf (Supplemental Figure 2B); further, transcript analysis of these morphants using RT-PCR revealed the inclusion of the 92 basepair intron 20-21 (Supplemental Figure 2C), which encodes the aforementioned truncated Brca2 peptide due to an in-frame stop sequence located within the sequence of intron 20-21. Taken together, these data suggested that interference with splicing of brca2 at this location is sufficient to cause fluid dysregulation as seen in the zep mutant phenotype.

To more specifically characterize the effect of brca2 deficiency on podocyte genesis in these knockdown models, we next performed WISH studies using the marker wt1b. WT embryos injected with the brca2 start MO displayed a loss of podocytes compared to uninjected controls at 24, 48 and 72 hpf (Figure 7A). Additionally, knockdown using either the exon 21 splice acceptor site blocking MO alone or the combination of exon 20 and 21 splice donor and acceptor MOs either severely reduced or completely abrogated wt1b staining in embryos at 24, 48 and 72 hours of development (Figure 7A; data not shown).

Figure 7. Knockdown of brca2 causes a reduction in expression of the podocyte marker wt1b and an increase in the interrenal gland lineage.

(A-C) At 24, 48 and 72 hpf, WT embryos display normal wt1b expression in podocytes, while microinjection of an MO targeting the brca2 start (MO1) or the splice acceptor of exon 21 (MO2) recapitulated the zep mutant phenotype, as observed by a reduction in the expression of wt1b transcripts as assessed by WISH. Further, brca2 morphants have elevated (B) nr5a1a transcripts as assessed by WISH and (C) increased 3β-HSD chromogenic staining, which are quantified in (D,E), respectively. Asterisks (**) indicate p<0.001 using student T-test. Embryos are shown in dorsal views, where black boxes demarcate the cervical region where podocytes and the interrenal gland develop.

As the knockdown of brca2 caused a decrease in podocyte cells, we next examined if this change occurred concomitantly with a gain of interrenal gland cells, as in zep. We performed expression studies on brca2 morphants to determine if there was an increase in interrenal gland size at 36 hpf, which revealed an expanded area of nr5a1a staining (Figure 7B). Further, 3β-HSD chromogenic staining of brca2 morphants showed an expanded area of staining as well (Figure 7C). Quantification of these areas in morphants compared to WT control embryos showed that the increased areas of nr5a1a staining and 3β-HSD chromogenic staining were statistically significant (Figure 7D, 7E). Through double WISH, we found that brca2 morphants simultaneously had an increased interrenal gland, visualized again by nr5a1a transcripts, in conjunction with an abrogation of wt1b staining (data not shown). In sum, our brca2 knockdown studies corroborate the conclusion that defective expression of brca2 is sufficient to elicit the zep mutant phenotype, which includes the combined loss of podocytes and increase in interrenal gland size.

3.7 zep fail to complement brca2^ZM_00057434 and can be rescued by overexpression of brca2

Previous researchers isolated and characterized brca2^ZM_00057434, which encodes a homozygous viable recessive allele that has an insertion that disrupts exon 11 allele and causes the production of two aberrant transcripts, one lacking the BRC repeats and another lacking the DBD (Rodriguez-Mari et al., 2011). Additionally, a brca2^Q658X allele has been characterized which encodes a nonsense mutation in exon 11, similar in location to BRCA2 mutations in humans with hereditary breast and ovarian cancer, which is likewise homozygous viable (Shive et al., 2010). Both the brca2^ZM_00057434 and brca2^Q658X -/- mutants undergo normal early development but later fail to form ovaries and become infertile males (Shive et al., 2010; Rodriguez-Mari et al., 2011). However, unlike zep, they were not reported to develop edema during embryonic stages or show outward signs of renal malfunction.

Thus, we next explored renal and interrenal development in the brca2^ZM_00057434 deficient strain and performed complementation tests with zep. Expression of brca2 in homozygous brca2^ZM_00057434mutants was normal at the 24 hpf stage (data not shown), as we had observed in zep-/- mutants between 24 and 120 hpf. To assess pronephros development in brca2^ZM_00057434-/- mutants, we used WISH to examine expression of wt1b transcripts between the 24 and 72 hpf stages in clutches obtained from matings of heterozygous brca2^ZM_00057434 adult zebrafish. This analysis did not reveal a detectable difference in podocyte formation (data not shown). Next, we analyzed expression of nr5a1a to assess the development of the interrenal lineage. Interestingly, the area of the interrenal gland at 36 hpf showed a slight increase in size in brca2^ZM_00057434 -/- mutant embryos compared to WT siblings (Figure 8A) that was statistically significant (Figure 8B). When pairwise matings of zep and brca2^ZM_00057434 heterozygous adults were performed, the embryos developed normally through the 7 dpf time point and did not display any alterations in fluid flow (data not shown), unlike clutches obtained from zep heterozygous adults in which 25% of embryos consistently developed edema by 96 hpf as described earlier in the present report. Consistent with this finding, clutches obtained from pairwise matings of zep and brca2^ZM_00057434 heterozygous adults showed no overt disruption of wt1b expression in podocyte precursors or decrease in podocyte numbers at the 24 hpf stage (data not shown). However, consistent with the interrenal phenotype in brca2^ZM_00057434 -/- mutant embryos, we found that the brca2^ZM_00057434 +/-; zep+/- compound heterozygous embryos had an increased area of nr5a1a expression at the 36 hpf time point (Figure 8A) that was equivalent to that observed in brca2^ZM_00057434-/- mutant embryos and was similarly statistically significant compared to WT siblings (Figure 8B). These findings revealed that the brca2^ZM_00057434 allele failed to complement zep with regard to formation of the interrenal lineage, though we were unable to detect a perceptible alteration in the formation of the podocyte lineage during embryogenesis.

Figure 8. zep and brca2^ZM_00057434 fail to complement one another.

(A) Gene expression analysis of the interrenal marker nr5a1a in WT, zep-/-, brca2^ZM_00057434 -/- and compound heterozygous brca2^ZM_00057434+/-; zep+/- mutants revealed that brca2^ZM_00057434 -/- have a modest but (B) statistically significant elevation in the area of nr5a1a expression that was similar to compound heterozygous embryos. Embryos are shown in dorsal views with insets showing digital zoom of 150%, where black boxes demarcate the cervical region where the interrenal gland develops. Asterisks (*) indicate p<0.05, (**) p<0.01 using student T-test.

Next, we performed rescue studies in clutches obtained from pairwise matings of zep heterozygous adults to assess whether brca2 cRNA was sufficient to rescue the alteration in podocyte development. Clutches were microinjected with brca2 cRNA at the 1-cell stage and then fixed at 24 hpf to perform WISH analysis for the expression of podocyte marker transcripts (Figure 9). brca2 overexpression was sufficient to partially rescue the zep podocyte phenotype based on partial restoration of wt1b expression (35/64, 55%) (Figure 9A). Interestingly, in a small percentage of WT siblings, brca2 overexpression induced ectopic expression of wt1b (24/177, 13.5%) (Figure 9A). To further validate these observations and also assess whether podocyte differentiation proceeded in zep-/- with brca2 overexpression, we examined expression of nphs1. We found that zep had a partial restoration of nphs1 expression (10/27; 37%), and also observed ectopic nphs1 expression in WT siblings (2/56, 3.5%) (Figure 9B). Taken together, these rescue studies provide strong evidence that an alteration in brca2 causes the defect in podocyte development in zep.

Figure 9. brca2 overexpression is sufficient to partially rescue zep and also induces ectopic podocytes in wild-type embryos.

Overexpression of full-length brca2 cRNA partially restored expression of (A) wt1b and (B) nphs1 in zep-/- and was associated with the induction of ectopic podocytes in a small percentage of WT siblings (see text for incidence). Embryos are shown in dorsal views with insets showing digital zoom of 150%, where black boxes demarcate the cervical region where the podocytes develop.

3.8 Assessment of RA and Notch signaling with respect to brca2 during pronephros development

To date, a collection of genetic, chemical, and biochemical studies have demonstrated that RA is required for normal podocyte formation in the embryonic zebrafish pronephros (Wingert et al., 2007; Wingert and Davidson, 2011), and further that RA signaling can regulate the wt1α promoter (Bollig et al., 2009). Therefore, we next examined whether a deficiency of RA might underlie the abrogation of podocyte development in zep. Previous studies have shown that the nephron patterning defects caused by RA deficiency can be rescued by the addition of exogenous all-trans RA to the embryo media beginning at gastrulation stages (Wingert and Davidson, 2011; Li et al., 2014; Cheng and Wingert, 2015; Marra and Wingert, 2016; Marra et al., 2016). To test whether zep fail to form podocytes due to insufficient RA levels, heterozygous zep zebrafish were mated, and their clutches were treated with varying levels of exogenous all-trans RA to interrogate whether supplementation of this metabolite was sufficient to rescue podocyte numbers in zep mutant embryos. From the 60% epiboly stage to approximately the 28 ss, clutches were incubated with either DMSO vehicle alone or various concentrations of RA/DMSO, which were selected because they have been shown previously to rescue podocyte ontogeny in RA deficient zebrafish embryos, then subsequently assayed by WISH for expression of podocyte markers (Figure 10). Consistent with prior observations, the DMSO control WT embryos had normal wt1b transcript levels, while these were absent in zep (Figure 10). zep mutants treated with one of several dosages of RA failed to develop podocytes (Figure 10). This indicated that addition of RA failed to rescue podocyte development. Addition of exogenous RA to control embryos had no effect on podocyte development, shown by normal wt1b staining (Figure 10). Next, to examine a later stage in podocyte formation, transcripts for the slit diaphragm gene, nphs1, were analyzed after RA treatment. While WT embryos displayed nphs1 expression, indicating a normal progression of podocyte differentiation, zep mutants lacked detectable nphs1 transcripts, again indicating the failure of podocyte genesis (Figure 10). Taken together, these data indicate that zep acts downstream of RA or in a separate pathway.

Figure 10. Addition of exogenous RA during nephrogenesis fails to rescue establishment of the podocyte lineage in zep mutant embryos.

(A) Schematic of exogenous RA treatment methodology in zebrafish embryos. (B) While WT siblings exhibit normal podocyte development after RA treatment, zep mutants fail to develop podocytes as assayed by expression of wt1b or nphs1. Embryos are shown in dorsal views, where black boxes demarcate the cervical region where the podocytes develop.

Previous work has also demonstrated that Notch signaling is essential for podocyte formation. In zebrafish pronephros development, the downstream Notch target gene hey1 is expressed in podocyte precursors (O'Brien et al., 2011). Further, genetic knockdown of the Notch transcriptional mediator rbpj alone, or in combination with knockdown of wt1α or foxc1α, partially reduced and abrogated podocyte numbers, respectively (O'Brien et al., 2011). Of these various knockdowns, only the combinatorial knockdown of wt1α and rbpj was associated with an increase in interrenal cells. Based on this knowledge, we hypothesized that reduced or abrogated Notch signaling might underlie the zep phenotype. To test this, we performed gain-of-function studies utilizing a transgenic model, Tg(hsp70∷GAL4, UAS∷NICD), that enables the ectopic activation of Notch signaling through heat-shock, and loss-of-function studies utilizing the gamma secretase inhibitor N-[N-(3,5-Difluorophenacetyl)-L-alanyl]-S-phenylglycine t-butyl ester (DAPT) (Figure 11). Podocyte formation was assessed by WISH for wt1b, and interrenal development was assessed by WISH for nr5a1α. NICD heat-shock at the 75% epiboly stage did not alter podocyte number in WT embryos or brca2 morphants (Figure 11). As expected, blocking Notch signaling by treatment with DAPT reduced podocyte formation in WTs, which was statistically significant, and DAPT treatment was associated with complete abrogation of podocytes in zep mutants (Figure 11). Upon analysis of interrenal gland development in these conditions, we found that there was not a statistically significant difference between untreated WTs, ectopic Notch activation in WTs, or DAPT treatment of WTs (Figure 11). Interestingly, brca2 morphants had a slightly reduced interrenal gland following ectopic NICD expression that was statistically significant compared to untreated knockdown controls (Figure 11). This finding warrants further investigation, as it may suggest that NICD was sufficient to alter the balance of podocyte versus interrenal lineages in brca2 deficient embryos, despite our inability to detect podocytes based on the assessment of wt1b expression.

Figure 11. Alterations in Notch signaling mediate podocyte and interrenal genesis.

(A,B) Gain and loss of function Notch experiments with Tg(hsp70∷GAL4, UAS∷NICD) and DAPT chemical treatments were performed in WT and brca2 morphants to assess podocyte development by WISH for wt1b and interrenal development by WISH for nr5a1a. Embryos are shown in dorsal views, where black boxes demarcate the cervical region where the podocytes and interrenal gland develop. (C,D) The standard deviation of triplicate groups, consisting of at least 5 embryos per group, was statistically analyzed by ANOVA with Tukey test. WT embryos had statistically significant reduced podocytes with DAPT. brca2 morphants failed to develop podocytes after NICD or DAPT. brca2 morphants had a reduced interrenal field following NICD that was statistically significant compared to non-heat shocked controls.

4. Discussion

The genetic mechanisms that specify nephron fates remain incompletely understood. Among nephron cell types, identification of the pathways that impact podocyte specification and differentiation has exceptionally high clinical importance because podocyte defects initiate a vast majority of chronic kidney disease states (Nagata, 2016). The central importance of podocytes, as well as their conservation across vertebrates, has led to their increasing study in a number of animal models, including the zebrafish (Morales and Wingert, 2017).

Here, we utilized the genetic tractability of zebrafish to identify essential renal factors, and discovered a critical requirement for brca2/fancd1 during development of the embryonic kidney through the cloning of the zep mutation. Through a WGS approach, zep was identified as a base pair alteration in a splice site of the brca2 gene, near its C-terminal end. This datum was confirmed by several approaches, including sequencing individual fish. By creating brca2 deficient zebrafish through multiple independent knockdown studies, we demonstrated that disruption of brca2 expression phenocopied the zep mutant. Notably, MOs designed to interfere with the same brca2 splice site that is mutated in zep recapitulated the zep podocyte and interrenal phenotypes, lending further credence to the conclusion that the brca2 locus contains the causative lesion in zep. Taken together, these data illustrate for the first time that brca2/fancd1 expression is required for normal podocyte ontogeny during zebrafish embryogenesis. Based on the consistent expansion of the interrenal lineage in zep, we hypothesize that brca2/fancd1 deficiency leads to an alteration in the balance of these fates during IM patterning.

4.1. Brca2 activity and the balance between podocyte genesis and interrenal cell fate

It is intriguing that the alteration in Brca2 caused by zep is associated with concomitant abrogation of podocytes and increased formation of interrenal gland cells. It is likely that this phenotype is the result of a change in the early cell fate decision that occurs specifically in the wt1α-expressing field of IM cells. In development there is a delicate balance between podocytes and the interrenal gland, in which there are essential cues, one of which we demonstrated here as brca2, that keep this balance. When this balance is disrupted, by knockdown of brca2 for example, one cell type can arise at the expense of the other, as identified in zep mutants and brca2 morphants. It is also possible that the loss of podocytes locally influences the expansion of the interrenal progenitors. However, at the time points we performed anti-phospho-histone H3 and AO staining, there was no appreciable difference between WT and zep mutants in cellular proliferation or cell death in the region where podocytes and interrenal cells arise, rendering these scenarios unlikely. 3β-HSD staining indicated the interrenal gland produced more hormones in zep mutants compared to WT embryos, possibly suggesting an alternative mechanism of an endocrine feedback loop leading to the interrenal expansion. However, as we have failed to detect an elevation in proliferation in this anatomical region, we favor the hypothesis of an alteration in cell fate decision during early podocyte/interrenal development.

At a molecular level, it is possible that the role of brca2 in podocyte/interrenal specification may involve downstream transcriptional promotion of a podocyte-specific gene program and/or inhibition of interrenal program. The expansion of the interrenal lineage might also represent an indirect consequence of the inability of podocyte precursors to be specified or progress in development to a renal fate. The observation that P-h2A.X+ cells are more abundant in zep as compared to WT embryos provides compelling evidence that DNA repair is compromised as a result of the zep mutation. How this relates to the alteration in podocyte/interrenal populations is unresolved in the present work, as we found increased P-h2A.X+ cells distributed throughout zep embryonic tissues at 24 hpf and did not detect an accumulation in the cervical region where the proximal pronephros and interenal gland develop. In sum, the observation of the podocyte/interrenal imbalance in zep leaves many questions to be answered about the interrelationship between these two cell types and the mechanistic consequences(s) of this genetic alteration in brca2/fancd1.

4.2. Epistatic relationship of brca2/fancd1 to RA and Notch signaling during pronephric development

Previous research from our lab and others has indicated the significant role of RA and Notch in patterning proximal cell fates within the kidney, including the podocyte lineage (Wingert et al., 2007; Bollig et al., 2009; Wingert and Davidson, 2011; Li et al., 2014; Cheng and Wingert, 2015; Marra and Wingert, 2016). Thus, one possible explanation for the dearth of podocytes in zep was a local reduction in RA levels or Notch signaling. The RA hypothesis was tested using chemical genetics, by addition of exogenous RA to zep mutant embryos to analyze any possible effect(s) on the podocyte lineage. However, the addition of RA did not rescue the loss of podocyte phenotype, implying that brca2 is required downstream or in a separate pathway than RA signaling. Future studies such as genetic modifier screens with zep will provide a valuable means to understand the regulatory networks surrounding brca2 renal functionality. Interestingly, we observed that ectopic expression of NICD led to a reduction in the interrenal gland size in brca2 morphants, though NICD was not sufficient to rescue development of podocytes based on the assessment of wt1b expression. These findings may suggest that Notch signaling acts downstream of Brca2 in the interrenal lineage. Further discernment of the effects of Notch in brca2 deficient embryos, however, will necessitate the identification of other IM markers of the podocyte lineage.

4.3. Brca2 activity and other aspects of renal ontogeny and maintenance

Previous research on the developmental functions of brca2 in zebrafish identified essential roles in gonad development and spermatogenesis through analysis of brca2 alleles with disruptions in exon 11 of the gene (brca2^Q658X and brca2^ZM_00057434)(Shive et al., 2010; Rodriguez-Mari et al., 2011). Further, these brca2 exon 11 alleles were not homozygous lethal, and embryonic or adult renal phenotypes were not reported in these zebrafish models (Shive et al., 2010; Rodriguez-Mari et al., 2011). Here, we have isolated a brca2 disruption that affects the C-terminal portion of the gene, in the DNA binding domain. How this mutation alters the expression and functionality of Brca2 requires future work. It is intriguing that zep display renal failure and is embryonic lethal, features that are distinct from brca2^Q658X and brca2^ZM_00057434. Across vertebrate models to date, BRCA2 null alleles are embryonic lethal in mouse and humans while viable in rats (Moynahan, 2002; Cotroneo, et al., 2007). Identification of whether zep represents a genetic null, hypomorphic, or hypermorphic allele will require many additional studies, and these will be necessary to further elucidate the molecular consequences that lead to the zep phenotype. Given the complexity of Brca2, zep provides a context that can help uncover more insights into the activities of this protein using the zebrafish.

4.4. Mechanisms of BRCA2 activity and implications in renal cancer

Many years of research have shown that various mutations in human BRCA2 cause breast, ovarian and other cancers of the reproductive system, in addition to leukemia (Wagner et al., 2004; Alter, 1996; Alter et al., 2007; Alter, 2014; D'Andrea and Grompe, 2003; Rosenberg et al., 2003). In addition, recent research has linked BRCA2 to several other cancers, such as that of the pancreas, prostate, and kidney (Lucas et al., 2014; Schutte et al., 1995; Edwards et al., 1998; Castro et al., 2013; Reid et al., 2005; Breast Cancer Linkage Consortium, 1999). Interestingly, brca2 mutation has also been shown to cause tumorigenesis in zebrafish (Shive et al., 2014; Rodriguez-Mari et al., 2011). Intriguingly, humans and mice seem to have the type and severity of cancers caused by a BRCA2 mutation correlated to the site of the lesion within BRCA2 (Reid et al., 2005; McAllister et al., 2002).

In addition to these cancers, genetic diseases such as Fanconi anemia, are linked to BRCA2 mutations. The Fanconi anemia complex is highly conserved amongst vertebrates (Titus et al., 2006). There are at least 16 Fanconi anemia complementation groups, one of which, FANCD1, has been renamed BRCA2 (D'Andrea and Grompe, 2003). Humans that develop Fanconi anemia have a higher incidence of leukemia, as well as other solid tumors (Hirsch et al., 2004). In fact, several studies have directly linked human Fanconi anemia/BRCA2 mutations to horseshoe kidney and childhood familial Wilms tumor, a pediatric cancer of the kidney (de Chadarevvian, 1985; Reid et al., 2005; Rizk et al., 2013; Compostella et al., 2010; Auerbach, 2009; de Kerviler et al., 2000; Jurca et al., 2014). Research has shown that many children with Fanconi anemia develop tumors (Malric et al., 2015). Specifically, there was a study that showed there is a 97% probability a child with biallelic BRCA2 mutation will develop cancer by 5.2 years of age (Hirsch et al., 2004; Compostella et al., 2010). In another recent study, a clear cell sarcoma of the kidney was discovered in a child diagnosed with Fanconi anemia acquired by biallelic BRCA2 mutation (Trejo-Bittar et al., 2014). This evidence is highly suggestive that BRCA2 is a tumor suppressor gene, with its most studied function in DNA damage repair, complexing with other Fanconi anemia complementation groups, in addition to RAD51 and BRCA1 to perform DNA repair (D'Andrea and Grompe, 2003). As discussed earlier, further exploration of whether DNA repair intersects with podocyte/interrenal development is a point for future research. Based on our current work, we believe that there is a novel role for brca2 through our characterization of the zep mutant phenotype, where it appears to have critical functionality in early cell fate decisions within IM descendants.

In summary, we have demonstrated through a forward genetic screen that podocyte genesis in the zebrafish pronephros relies on the expression of the brca2 gene. We were able to induce the zep mutant phenotypes using several independent brca2 MOs, validating our WGS findings. Our data suggests that a splicing mutation in brca2 causes an alteration of early cell fate decisions, resulting in a loss of the podocyte lineage and a gain of the interrenal gland cell lineage. Because previous work has not identified specific functions for brca2 in IM tissues, this novel finding can help inform future research in the early development of podocytes. Ultimately, because the nephron is highly conserved amongst vertebrates and human patients with BRCA2 mutations have presented with Wilms tumor and other kidney ailments, such as Fanconi anemia, it is reasonable to suggest that this research is applicable to mammalian studies and can elucidate new pathways necessary for renal development and maintenance, furthering the understanding of the causes of podocyte diseases.

Supplementary Material

1. Supplementary Figure 1. Further assessment of cell proliferation and death in zep mutants.

(A) The number of pH3+ cells was quantified in the cervical region of WT and zep embryos at the 24 hpf stage. The quantity of pH3+ cells was not statistically different as assessed by student T-test. (B) Whole mount AO staining and (C) anti-Caspase-3 immunostaining of the tail region in WT and zep embryos revealed similar levels of cell death, which was assessed by student T-tests (D, E) respectively.

2. Supplementary Figure 2. brca2 transcript analysis in zep mutants and brca2 morphants, and live morphology of brca2 morphants.

(A) RT-PCR analysis of brca2 transcripts in pools of zep embryos compared to WT control siblings (which included WT and heterozygous embryos), where embryos were scored into these groups based on edema at embryonic day 6. WT/heterozygous pools had normally spliced transcripts as well as variants that corresponded to inclusion of the 92 basepair intron 20-21 (see band doublet). zep mutant pools contained only the longer variant with the intron (top band). (B) RT-PCR of brca2 transcripts in brca2 MO2 injected embryos compared to WT control siblings revealed inclusion of the 92 basepair intron 20-21 only in the morphant pool, and WT transcripts were not detected. (C) WT embryos were microinjected with a brca2 start morpholino (MO1) or a morpholino designed to the splice acceptor site (MO2) of exon 21, and live morphology was observed over the subsequent 6 days of development. Morphants recapitulated the zep mutant embryo phenotype such that they developed pericardial edema (black arrowhead), starting by 72 hpf, which progressed in severity through the 144 hpf stage. Embryos are shown in lateral views.

Highlights.

• brca2 is required for podocyte development in the zebrafish embryo kidney

• the balance of podocyte and interrenal fates is mediated by brca2

• overexpression of brca2 induces ectopic podocytes in zebrafish

• retinoic acid signaling acts upstream or independently of brca2 in podocyte formation

Abbreviations

AO

acridine orange

brca2

breast cancer 2, early onset, also known as fancd1

cRNA

capped RNA

DAPT

N-[N-(3,5-Difluorophenacetyl)-L-alanyl]-S-phenylglycine t-butyl ester

dpf

days post fertilization

ESRD

end stage renal disease

FISH

fluorescent in situ hybridization

GBM

glomerular basement membrane

hpf

hours post fertilization

IM

intermediate mesoderm

lib

lightbulb

MET

mesenchymal to epithelial transition

MO

morpholino oligonucleotide

P

podocytes

PTU

1-phenyl-2-thiourea

RA

retinoic acid

RARE

retinoic acid response element

RAR

retinoic acid receptor

RXR

retinoid X receptor

RT-PCR

reverse-transcription polymerase chain reaction

ss

somite stage

WGS

whole genome sequencing

WISH

whole mount in situ hybridization

WT

wild-type

WT1

Wilms tumor 1

zep

zeppelin