Ppargc1a Controls Ciliated Cell Development by Regulating Prostaglandin Biosynthesis.

Cilia are microtubule-based organelles that function in a multitude of physiological contexts to perform chemosensing, mechanosensing, and fluid propulsion. The process of ciliogenesis is highly regulated, and disruptions result in disease states termed ciliopathies. Here, we report that peroxisome proliferator-activated receptor gamma, coactivator 1 alpha (ppargc1a) is essential for ciliogenesis in nodal, mono-, and multiciliated cells (MCCs) and for discernment of renal tubule ciliated cell fate during embryogenesis. ppargc1a performs these functions by affecting prostaglandin signaling, whereby cilia formation and renal MCC fate are restored with prostaglandin E2 (PGE2) treatment in ppargc1a-deficient animals. Genetic disruption of ppargc1a specifically reduces expression of the prostanoid biosynthesis gene prostaglandin-endoperoxide synthase 1 (ptgs1), and suboptimal knockdown of both genes shows this synergistic effect. Furthermore, ptgs1 overexpression rescues ciliogenesis and renal MCCs in ppargc1a-deficient embryos. These findings position Ppargc1a as a key genetic regulator of prostaglandin signaling during ciliated cell ontogeny.

INTRODUCTION

Cilia are hair-like structures that project from the cell surface, where they perform extremely important tasks including chemosensing, mechanosensing, and fluid propulsion. Although most vertebrate cells possess a single immotile cilium, others are specialized to be motile multiciliated cells (MCCs) (Marra et al., 2016). Nodal cilia are a unique third class of ciliated cells that are transient and include a single motile cilium necessary for establishing left-right symmetry across vertebrates (Essner et al., 2002). Cilia structures consist of an axoneme, basal body, transition zone, ciliary membrane, and ciliary tip (Fliegauf et al., 2007). Disruptions to these ciliary structures are linked to heterogeneous human diseases termed ciliopathies, which are characterized by multisystem defects in organs such as the brain, heart, lungs, and kidneys (Mitchison and Valente, 2017; Reiter and Leroux, 2017). Despite recent advances in identifying steps of ciliogenesis, continued elucidation of the genetic mechanisms that regulate cilia structure and function, as well as the discernment of monociliated versus MCC fate, remain areas of central importance.

Several studies have highlighted key roles for prostaglandin signaling in ciliated cell ontogeny. During ciliogenesis, binding of bioactive prostanoid prostaglandin E2 (PGE2) to its receptor EP4 triggers a cyclic AMP-mediated signaling cascade, which promotes anterograde intraflagellar transport (IFT) of protein cargoes that extend cilia (Jin et al., 2014). Disruption of prostaglandin synthesis also leads to probasal body accumulation during multiciliogenesis, suggesting a role in regulating their trafficking or docking to the cell surface (Marra et al., 2019a). Additionally, prostaglandin signaling is necessary for renal progenitors in the zebrafish embryonic kidney, or pronephros, to adopt an MCC fate during nephron development, where a monociliated cell fate is chosen preferentially in the absence of normal prostaglandin synthesis (Marra et al., 2019a).

Interestingly, levels of PGE2 in the adult mammalian kidney are directly related to levels of peroxisome proliferator-activated receptor (PPAR) gamma, coactivator 1 alpha (PGC-1α; also known as Ppargc1a), which provide renoprotective effects following injury (Tran et al., 2016). PGC-1α is best known as a coactivator for the PPAR pathway, though it is now appreciated to serve a number of diverse tissue-specific functions (Puigserver and Spiegelman, 2003; Chambers and Wingert, 2020). Among these, ppargc1a was recently discovered to mediate cell fate in the developing zebrafish embryo kidney (or pronephros) by controlling regionalized expression of several transcription factors to establish segment domains within nascent nephrons (Chambers et al., 2018). However, despite the connection between PGC-1α and PGE2 in mature nephrons, there has been no previous evidence suggesting a relationship between these effectors during kidney or ciliated cell development.

Here, we report that ppargc1a is necessary for both proper ciliogenesis and MCC specification. ppargc1a zebrafish mutants and morphants exhibited cilium loss and/or shortened cilia in several tissues, including the ear, Kupffer’s vesicle (KV), and pronephros. ppargc1a-deficient embryos also displayed reduced numbers of renal MCCs, while monociliated cell numbers increased, thereby recapitulating the effects of attenuated prostaglandin signaling. Thus, we tested the hypothesis that these pathways were connected and found that PGE2 metabolites were decreased in ppargc1a-deficient zebrafish and that MCC fate choice and cilia length were restored in ppargc1a-deficient embryos treated with dmPGE2. Genetic interaction studies revealed a synergistic effect between ppargc1a and the prostaglandin biosynthesis enzyme encoded by prostaglandin-endo-peroxide synthase 1 (ptgs1; also known as cyclooxygenase; [cox1]), where combined suboptimal knockdowns resulted in ciliated cell defects. Consistent with these findings, ppargc1a-deficient embryos exhibited decreased ptgs1 transcript levels, where provision of ptgs1 mRNA was sufficient to rescue MCC number and cilia length. These results advance our fundamental understanding of ciliated cell patterning and differentiation by illuminating pivotal roles for ppargc1a in regulating prostaglandin production.

RESULTS

ppargc1a-Deficient Zebrafish Exhibit Ciliopathic Phenotypes

Hallmarks of underlying ciliary defects in the zebrafish embryo commonly present as pleiotropic phenotypes, including pronephric cysts, curved body axes, and edema (Drummond et al., 1998; Sun et al., 2004; Kramer-Zucker et al., 2005; Sullivan-Brown et al., 2008). Subsequent to detailing the role ppargc1a plays in nephron segmentation (Chambers et al., 2018), we unexpectedly detected a suite of morphological phenotypes suggestive of ciliary dysfunction in ppargc1a mutants. ppargc1a zebrafish (ZIRC) have a T-to-A mutation resulting in a premature stop codon in exon 7 (Chambers et al., 2018). At about 48 h post-fertilization (hpf), ppargc1a zebrafish developed curly tails, pronephric cysts, and pericardial edema compared with wild-type (WT) siblings (Figures 1A and 1B). These surprising phenotypes presented in a manner following the expected Mendelian ratio for a recessive allele bred through heterozygous incrosses (Figure 1A). Additionally, ppargc1a mutant embryos had atypical otic vesicle formation at 3 days post-fertilization (dpf), including the number of otoliths formed (Figure 1C) and diminished cilia formation on the cristae structures at 4 dpf (Figure 1D) (Wu et al., 2011; Yu et al., 2011; Stooke-Vaughan et al., 2012).

Figure 1.

ppargc1a-Deficient Zebrafish Display Phenotypes Commonly Associated with Cilia Defects

(A) WT sibling (top) and a ppargc1a mutant (bottom) zebrafish with a curly tail, pronephric cyst (blue arrow, inset), and pericardial edema at 3 dpf. Heterozygous crosses resulted in the expected Mendelian ratio shown at the bottom right of the respective panels (WT group includes homozygous WT and ppargc1a combined). Scale bar, 90 μm.

(B) JB4 transverse section of a 4 dpf WT sibling (left) and ppargc1a mutant (right) zebrafish (n, notochord; tubule outlined in green, cysts outlined in red). Scale bar, 90 μm.

(C) Three days post-fertilization WT sibling (left) and ppargc1a mutant (right) otoliths. Arrowheads indicate abnormal otoliths present in the mutant. Scale bar, 90 μm.

(D) Four days post-fertilization otic cilia (green, acetylated α-tubulin) of the cristae in WT (left) and ppargc1a mutant (right) zebrafish. Scale bar, 10 μm.

(E) Dorsal view of 3 dpf WT sibling (top) and ppargc1a mutant (bottom) approximately 24 h after being injected with 40 kDa dextran-FITC to assay for pronephric function. White arrowheads indicate PCT fluorescence in WT. White box highlights the approximate area of the PCT with no fluorescence. Scale bar, 90 μm.

(F) Three days post-fertilization WT sibling (top) and ppargc1a mutant (bottom) stained via WISH for slc20a1a to mark PCT. Scale bar, 90 μm.

(G) Fifty-five hours post-fertilization WT and ppargc1a MO injected zebrafish stained via WISH for the heart marker myl7. Penetrance graph of atypical heart looping occurring in WT and ppargc1a morphants. Chi-square analysis indicates significance. n = 128 (WT), n = 103 (MO), p < 0.001. Scale bar, 90 μm.

(H) Whole-mount immunofluorescence of the Kupffer’s vesicle (KV) at the 10 ss stained for acetylated α-tubulin (cilia, green), anti-PKC (membrane boundary, red). Scale bar, 8 μm.

(I) Cilia length in micrometers of KV cilia.

(J) The number of cilia present in the KV.

(K) Twenty-four hours post-fertilization WT zebrafish stained via WISH to illustrate ppargc1a mRNA expression. Box is of approximate area of inset showing Ppargc1a (red) expression and the cilia marker acetylated α-tubulin (green). Scale bars, 65 μm for the whole embryo and 10 μm for the inset.

Data are represented as mean ± SD; **p < 0.01 and ***p < 0.001 (t test).

To further investigate the function of the pronephros, we microinjected 40 kDa dextran-FITC (fluorescein isothiocyanate) into 48 hpf WT and ppargc1a mutant embryos, then imaged the animals at the 72 hpf stage. Whereas WT embryos exhibited strong fluorescence signal in the proximal convoluted tubule (PCT), indicating normal fluid flow, ppargc1a mutants failed to take up the fluorescent conjugate (Figure 1E). Lack of fluid flow from ciliary defects disrupts collective cell migration and PCT morphogenesis in the pronephros (Vasilyev et al., 2009). Thus, to determine if the PCT was forming correctly, we performed whole-mount in situ hybridization (WISH) for slc20a1a and found a lack of coiling morphogenesis in ppargc1a compared with WT siblings (Figure 1F). These data suggest that cilia in the pronephros are not functioning properly.

Several studies of the KV have illustrated that the establishment of left-right symmetry is affected by loss of cilia function (Essner et al., 2005; Wang et al., 2011, 2012; Compagnon et al., 2014; Dasgupta and Amack, 2016). To assess left-right patterning in ppargc1a-deficient embryos, we examined its first overt morphological feature in vertebrates, cardiac looping. Using a previously validated ppargc1a morpholino oligonucleotide (MO) (Chambers et al., 2018), we performed WISH for the heart marker myl7 at 55 hpf and found that a randomization of heart looping occurred in ppargc1a morphants compared with WT controls (Figure 1G) (Bisgrove et al., 2005; Tian et al., 2009; Huang et al., 2011; Manning et al., 2013). To determine if the randomization of cardiac looping correlated with changes in KV cilia formation, we performed whole-mount immunofluorescence (IF) at the 10 somite stage (ss). Because KV phenotypes largely depend on age, we ensured that each sample was 10 ss before fixation. Subsequent analysis revealed that both the number and length of cilia in the KV were decreased in ppargc1a morphants compared with WTs (Figures 1H–1J).

Given these observations, we examined if the expression of ppargc1a and/or Ppargc1a corresponded with ciliated cell populations in the embryo. Previous reports have shown that ppargc1a transcripts are expressed in the renal progenitor field (Bertrand et al., 2007; Chambers et al., 2018). WISH studies in 24 hpf WT embryos confirmed that ppargc1a transcripts were spatially localized to the pronephros (Figure 1K) as well as another ciliated area, the nasal placodes (Figure 1K, inset). Interestingly, IF studies revealed that Ppargc1a protein is localized to the apical surface of the distal pronephros tubule cells (Figure 1K; Figure S1A). Taken together, these data suggest that ppargc1a is necessary for cilia formation, cilia function, or both, as it is expressed in ciliated areas, and zebrafish deficient in ppargc1a display multiple hallmarks of ciliopathies.

ppargc1a-Deficient Nephrons Have a Depleted MCC Lineage and Atypical Cilia Formation

On the basis of the renal phenotypes in ppargc1a mutants, we hypothesized that ciliated cell fate choice and/or ciliogenesis were compromised. The zebrafish pronephros has two types of ciliated cells that occupy the tubule: MCCs and monociliated transporter cells, where the latter includes several unique differentiated types that are organized into a series of discrete functional segments (Figure 2A) (Wingert et al., 2007; Liu et al., 2007; Ma and Jiang, 2007). Within the proximal tubule segments, MCCs are distributed in a “salt and pepper” fashion such that they are interspersed with the monociliated transporter populace. In contrast, the last distal tubule segment and pronephric duct regions are composed solely of monociliated transporter cells (Figure 2A) (Marra and Wingert, 2016). Both cell types possess motile cilia that function in fluid propulsion (Kramer-Zucker et al., 2005). Interestingly, ppargc1a-deficient animals have an increased proximal straight tubule (PST) segment, marked specifically by transcripts encoding trpm7, suggesting an increased number of monociliated transporter cells (Chambers et al., 2018). Furthermore, there is precedence that elevated monociliated cell number can occur at the expense of MCC fate choice (Marra et al., 2019a, 2019b).

Figure 2.

Loss of ppargc1a Results in Ciliogenesis Defects

(A) Schematic indicating location of MCCs in the proximal portion and monociliated cells in the distal portion of the zebrafish pronephros at 24 hpf.

(B) WT (top) and ppargc1a mutant (bottom) zebrafish at 24 hpf stained via WISH for odf3b to mark MCCs. Scale bar, 90 μm.

(C) Graph representing the number of MCCs present in WT and ppargc1a zebrafish at 24 hpf.

(D) Fluorescent in situ hybridization for MCC marker (odf3b, green) and transporter cell marker (trpm7, red) in WT and ppargc1a morphant zebrafish at 24 hpf. Scale bar, 10 μm.

(E and F) Graphs representing the percentage of odf3b+ (E) or trpm7+ (F) cells in WT (black) and ppargc1a morphants (white) pronephros between somites 9 and 11.

(G) Twenty-eight hours post-fertilization WT (top), ppargc1a MO-injected (middle), and ppargc1a (bottom) zebrafish stained via whole-mount immunofluorescence for acetylated α-tubulin (cilia, green), γ-tubulin (basal bodies, red), and DAPI in the proximal pronephros. Scale bar, 10 μm.

(H) Cilia length in micrometers for the proximal pronephros.

(I) Graph of the percentage of ciliated basal bodies/total basal bodies per 100 μm in the proximal pronephros.

(J) Twenty-eight hours post-fertilization WT (top) and ppargc1a MO-injected (middle) and ppargc1a (bottom) zebrafish stained via whole-mount immunofluorescence for acetylated α-tubulin (cilia, green), γ-tubulin (basal bodies, red), and DAPI in the distal pronephros. Scale bar, 10 μm.

(K) Cilia length in micrometers for the distal (right) pronephros.

(L) Graph of the percentage of ciliated basal bodies/total basal bodies per 100 μm in the distal pronephros.

(M and N) Fluorescent intensity plots of WT (black) and ppargc1a MO-injected (red) zebrafish in proximal (M) and distal (N) pronephros.

Data are represented as mean ± SD. *p < 0.05, **p < 0.01, and ***p < 0.001 (one-way ANOVA or t test).

Therefore, we first sought to determine if MCCs were developing properly in ppargc1a-deficient embryos. ppargc1a mutant zebrafish were assessed with WISH for the MCC marker odf3b, which showed that they had significantly fewer MCCs per nephron compared with WT control embryos (Figures 2B and 2C). Using fluorescent in situ hybridization (FISH) to enable precise quantification of both MCCs and monociliated cells, we found that ppargc1a-deficient embryos had a decreased number of odf3b+ cells that was proportional to the increased number of trpm7+ cells (Figures 2D–2F). Notably, the number of nuclei marked by DAPI in WT and ppargc1a-deficient embryos was not significantly different within this segment domain, suggesting that altered cell number and/or morphology does not account for the changes in MCC/transporter cell ratio (Figure S1B). Given the reduction in maturing MCCs, we used WISH to examine expression of the MCC progenitor marker jag2b (Marra et al., 2019a). ppargc1a-deficient embryos had significantly reduced numbers of jag2b cells compared with WT controls (Figures S1C and S1D). These findings support the conclusion that fate choice between MCC and monociliated transport cell identity in the developing PST is altered as a consequence of ppargc1a deficiency.

Next we explored ciliary development in the pronephros using whole-mount IF to detect acetylated α-tubulin (cilia), γ-tubulin (basal bodies), and DAPI in WT and ppargc1a-deficient embryos at the 28 hpf stage. In the proximal pronephros, cilia length was significantly shorter in both ppargc1a and ppargc1a morphant embryos compared with WT (Figures 2G and 2H). Additionally, we noted that ppargc1a-deficient embryos possessed fewer ciliated basal bodies (Figure 2I), suggesting a defect in initiating cilium formation. Consistent with reduced cilium formation and shorter cilium length, the fluorescent intensity of the acetylated α-tubulin signal was reduced in ppargc1a-deficient embryos compared with WT controls (Figure 2M). As the proximal pronephros is composed of MCCs and transporter cells, these data suggest ciliogenesis defects in both populations.

To investigate this further, we examined the distal tubule, as it consists of a homogeneous population of monociliated cells. Similar to the proximal tubule, cilia length in the distal tubule domain was significantly decreased in ppargc1a-deficient embryos compared with WT controls (Figures 2J and 2K). Likewise, distal tubule cells in ppargc1a-deficient embryos possessed fewer ciliated basal bodies (Figure 2L). Furthermore, the fluorescent intensity of the acetylated α-tubulin signal was lower in ppargc1a-deficient embryos compared with WT controls, consistent with decreased cilia formation (Figure 2N). Taken together, these data documenting shortened or absent cilia in tubule cells of the proximal and distal pronephros indicate that cilia formation is significantly disrupted in two independent ppargc1a-deficient zebrafish models and support the conclusion that ppargc1a is requisite for proper ciliogenesis.

To assess if the MCC abrogation and defunct ciliogenesis phenotypes are related outcomes unique to ppargc1a deficiency, we examined MCC formation in ift88-deficient embryos. Loss of the IFT factor ift88 in zebrafish, which is an ortholog of the mouse Polaris gene, causes cystic kidney, reduced cilia length, abrogated fluid flow in the pronephros, and randomized heart looping (Kramer-Zucker et al., 2005; Vasilyev et al., 2009; Manning et al., 2013; Gerlach and Wingert, 2014), mimicking cystic kidney disease and laterality randomization in mice with hypomorphic mutations in Polaris (Moyer et al., 1994; Yoder et al., 1995). Interestingly, MCC number was not significantly different between WT control embryos and ift88 morphants (Figures S2A and S2B). This finding suggests that ciliogenesis and MCC fate choice are not intrinsically linked developmental pathways. Injection of a standard control MO also resulted in no significant changes to the number of MCCs, proximal and distal cilia length, or proximal and distal ciliated basal bodies (Figures S2C–S2J). Taken together, these data led us to conclude that ppargc1a deficiency specifically leads to two independent developmental defects in nephrons, namely, a depleted MCC population (that is replaced by an expanded monociliated transport cell populace) and atypical cilia formation.

Ciliogenesis and Renal MCC Fate in ppargc1a-Deficient Zebrafish Can Be Rescued by PGE2

Recent studies have established that prostaglandin signaling is needed for proper cilium elongation, basal body trafficking, and MCC versus transporter cell fate choice during pronephros development (Jin et al., 2014; Marra et al., 2019a). Strikingly, ppargc1a-deficient embryos possess both the ciliogenesis and renal cell fate defects as embryos lacking prostaglandin signaling. A link between ppargc1a and the prostanoid PGE2 was previously reported in the injured adult mammalian kidney, where decreased ppargc1a resulted in decreased PGE2 levels and increased ppargc1a correlated with increased PGE2 levels (Tran et al., 2016). Additionally, ciliary and renal abnormalities from defective prostaglandin synthesis or export are rescued by provision of PGE2 or its analog 16,16-dimethyl-PGE2 (dmPGE2), the latter of which has a longer half-life (Jin et al., 2014; Poureetezadi et al., 2016; Marra et al., 2019a). Therefore, we hypothesized that ppargc1a-deficient animals lack sufficient PGE2 and that supplementing with a form of PGE2 during development might restore ciliogenesis and MCC fate choice.

First, to determine if ppargc1a-deficient zebrafish had decreased PGE2, we performed an ELISA for PGE2 metabolites (Esain et al., 2015). ELISA results indicated a significant decrease in PGE2 metabolites in 28 hpf ppargc1a-deficient zebrafish (Figure 3A). As prostaglandin signaling is necessary for MCC cell fate and maturation (Marra et al., 2019a), and ppargc1a-deficient embryos similarly have decreased MCC progenitors (jag2b+ cells) and maturing MCCs (odf3b+ cells) (Figures 2B and 2C; Figures S1C and S1D), we examined the effect of dmPGE2 on MCC development in ppargc1a-deficient zebrafish. To do this, we completed WISH for odf3b in 24 hpf stage WT controls, ppargc1a morphants, and ppargc1a morphants supplemented with 100 μM dmPGE2 (Figure 3B). The ppargc1a morphant model was selected for these studies because we had not observed any discernible phenotype differences between mutant and morphant embryos (Figures 1 and 2). Importantly, the use of morphants is highly tractable, as it enabled us to reduce the number of animals needed for experiments and increased the efficiency of the studies because this approach does not necessitate genotyping. dmPGE2 rescued MCC number in ppargc1a-deficient embryos to WT levels (Figures 3B and 3C). Furthermore, changing the time of addition for dmPGE2 supplementation to 8, 10, or 12 hpf led to comparable rescue of MCC number in ppargc1a-deficient embryos (Figure S3A). We observed no gain-of-function phenotype in the number of MCCs (Figure S3B), suggesting that there is a threshold above which normal MCCs develop.

Figure 3.

ppargc1a Loss of Function Phenotypes Can Be Rescued by Supplementing with PGE2

(A) Relative concentration of PGE2 in WT and ppargc1a morphants.

(B) MCCs stained via WISH for odf3b at 24 hpf in WT (left), ppargc1a MO (middle), and ppargc1a morphants treated with dmPGE2 (right). Scale bar, 90 μm.

(C) Graph depicting the number of MCCs present in WT, ppargc1a MO, and ppargc1a MO treated with dmPGE2.

(D) Twenty-eight hours post-fertilization WT (top), ppargc1a MO-injected (middle), and ppargc1a MO-injected + dmPGE2 (bottom) zebrafish stained via whole-mount immunofluorescence for acetylated α-tubulin (cilia, green), γ-tubulin (basal bodies, red), and DAPI in the proximal pronephros. Scale bar, 10 μm.

(E) Cilia length in micrometers for the proximal pronephros of WT, ppargc1a morphants, and ppargc1a morphants treated with dmPGE2.

(F) Graph of the percentage of ciliated basal bodies/total basal bodies per 100 μm in the proximal pronephros.

(G) Twenty-eight hours post-fertilization WT (top), ppargc1a MO-injected (middle), and ppargc1a MO-injected + dmPGE2 (bottom) zebrafish stained via whole-mount immunofluorescence for acetylated α-tubulin (cilia, green), γ-tubulin (basal bodies, red), and DAPI in the distal pronephros. Scale bar, 10 μm.

(H) Cilia length in micrometers for the distal pronephros of WT, ppargc1a morphants, and ppargc1a morphants treated with dmPGE2.

(I) Graph of the percentage of ciliated basal bodies/total basal bodies per 100 μm in the distal pronephros.

(J and K) Scatterplots illustrating the difference in fluorescent intensity signals of WT, ppargc1a MO, and ppargc1a MO + dmPGE2 in proximal (J) and distal (K) cilia.

(L) Representative images of ppargc1a in crosses shown at 48 hpf WT zebrafish treated with DMSO (top), curly ppargc1a treated with DMSO (middle), and a WT-appearing, genotype-confirmed ppargc1a rescued with dmPGE2 (bottom). Scale bar, 90 μm.

(M) The percentage of zebrafish with WT (black) or curly (white) phenotypes at 48 hpf after treatment with DMSO or dmPGE2. Chi-square test indicates significance.

Data are represented as mean ± SD. *p < 0.05, **p < 0.01, and ***p < 0.001 (one-way ANOVA or t test); n.s., not significant.

Next, to test if dmPGE2 treatment could rescue the ciliogenesis phenotypes observed in ppargc1a deficiency, we treated ppargc1a-deficient embryos with 100 μM dmPGE2 beginning at 6 hpf, and pronephros ciliogenesis was assessed at 28 hpf using whole-mount IF. In the proximal tubule, dmPGE2 rescued the appearance of cilia in ppargc1a-deficient embryos compared with vehicle-treated controls (Figure 3D). Quantification of ciliary length showed a significant restoration in the dmPGE2-treated ppargc1a morphants (Figure 3E), which was consistent with measurements of cilia fluorescent intensity (Figure 3J). Additionally, the number of ciliated basal bodies in dmPGE2-treated ppargc1a morphants was restored to a WT level (Figure 3F). Evaluation of these features in WT embryos treated with dmPGE2 did not reveal a gain-of-function effect (Figures S3C and S3D). These findings suggested that cilia formation in monociliated cells and MCCs in ppargc1a-deficient embryos is rescued by provision of a bioactive PGE2 prostanoid. Similar analyses were performed in the distal pronephros to explicitly examine if ciliogenesis in monociliated cells of this region was similarly restored (Figure 3G). Again, we observed an overt rescue of ciliogenesis in dmPGE2-treated ppargc1a morphants compared with vehicle controls (Figure 3G), which was significant when ciliary length and ciliated basal bodies were quantified (Figures 3H and 3I), and consistent with measurements of cilia fluorescent intensity (Figure 3K). Together, these findings led us to conclude that ciliogenesis throughout the pronephros was restored in ppargc1a-deficient animals by addition of dmPGE2.

Furthermore, we investigated if dmPGE2 supplementation was sufficient to rescue the curved body axis phenotype observed at the 48 hpf stage. We obtained three independent clutches from matings of ppargc1a heterozygous adults, then split each clutch at 6 hpf to treat half of the embryos with media containing DMSO vehicle, while the other half were immersed in media with dmPGE2; subsequently, the media was refreshed at 24 and 33 hpf to offset degradation of the prostanoid. At the 48 hpf stage, DMSO-treated cohorts displayed the expected Mendelian ratio (25%) of mutant phenotype with 24.33% curly, and the corresponding dmPGE2-treated cohorts displayed an incidence of mutant phenotype with 13% curly (Figures 3L and 3M). Representative genotype analysis of random embryos from each cohort confirmed that some ppargc1a mutants displayed WT morphology following dmPGE2 administration (Figure 3L). We concluded that the regimen of dmPGE2 supplementation rescued ciliogenesis in some cases, and we speculate that rescue of all mutants was not obtained even with fresh additions during the experiment, likely because of fluctuations in the quantity of active dmPGE2 over this protracted developmental interval. In sum, these data show that provision of dmPGE2 is sufficient to rescue ciliogenesis and renal MCC fate choice in ppargc1a-deficient zebrafish.

ppargc1a Regulates the Prostaglandin Pathway Component ptgs1

Prostaglandin synthesis occurs when a cyclooxygenase enzyme (Cox1 or Cox2) converts arachidonic acid to an intermediate that is subsequently converted to an active prostanoid such as PGE2. Compromised prostanoid biosynthesis due to the loss of ptgs1, which encodes Cox1 in zebrafish, abrogates ciliogenesis and skews renal MCC fate choice (Marra et al., 2019a). Therefore, we designed genetic interaction studies to explore the relationship between ppargc1a and ptgs1. Suboptimal doses of MOs are commonly used to test for a synergistic effect that would suggest multiple genes act in a shared pathway (DiBella et al., 2009; Wagle et al., 2011; Choi et al., 2015; Kallakuri et al., 2015). We conducted suboptimal MO (SMO) microinjections with ppargc1a and ptgs1 on the basis of previously established dosages (North et al., 2007; Jin et al., 2014; Esain et al., 2015; Poureetezadi et al., 2016; Chambers et al., 2018; Marra et al., 2019a). Furthermore, we performed IF to evaluate Ppargc1a expression following ppargc1a SMO, which revealed diminished signal in the distal pronephros (Figure S1A), which was statistically significant when quantified on the basis of fluorescent intensity (Figure S4A). Assessment of ciliogenesis using IF in the proximal pronephros (Figures 4A–4C) and distal pronephros (Figures 4D–4F) revealed no statistical difference in cilia length or ciliated basal bodies in WT, ppargc1a SMO, or ptgs1 SMO, while there was a statistically significant decrease in ppargc1a + ptgs1 SMO. Additionally, the mean fluorescent intensity was lower for the combination knockdown compared with WT or single knockdowns (Figures S4B and S4C). Next, we tested the effect of incubating ppargc1a + ptgs1 SMO knockdowns with dmPGE2. Addition of dmPGE2 rescued cilia length, the percentage of ciliated basal bodies, and cilia fluorescent intensity in the proximal (Figures 4A–4C; Figure S4B) and distal pronephros (Figures 4D–4F; Figure S4C).

Figure 4.

ppargc1a and Prostaglandin Pathway Component ptgs1 Act Together to Properly Form Ciliated Cells

(A) Twenty-eight hours post-fertilization WT (top), ppargc1a suboptimal dose MO (SMO), ptgs1 SMO, ppargc1a + ptgs1 SMO, and ppargc1a + ptgs1 SMO + dmPGE2-treated zebrafish stained via whole-mount immunofluorescence for acetylated α-tubulin (cilia, green), γ-tubulin (basal bodies, red), and DAPI in the proximal pronephros. Scale bar, 7 μm.

(B) Cilia length in micrometers for the proximal pronephros.

(C) Graph of the percentage of ciliated basal bodies/total basal bodies per 100 μm in the proximal pronephros.

(D) Twenty-eight hours post-fertilization WT (top), ppargc1a SMO, ptgs1 SMO, ppargc1a + ptgs1 SMO, and ppargc1a + ptgs1 SMO + dmPGE2-treated zebrafish stained via whole-mount immunofluorescence for acetylated α-tubulin (cilia, green), γ-tubulin (basal bodies, red), and DAPI in the distal pronephros. Scale bar, 7 μm.

(E) Cilia length in micrometers for the distal pronephros.

(F) Graph of the percentage of ciliated basal bodies/total basal bodies per 100 μm in the distal pronephros.

(G) MCCs stained via WISH for odf3b at 24 hpf in WT, ppargc1a SMO, ptgs1 SMO, ppargc1a + ptgs1 SMO, and ppargc1a + ptgs1 SMO + dmPGE2-treated zebrafish. Scale bar, 50 μm.

(H) Graph depicting the number of MCCs present.

(I) Forty-eight hours post-fertilization zebrafish injected with ppargc1a SMO, ptgs1 SMO, ppargc1a + ptgs1 SMO, and ppargc1a + ptgs1 SMO + dmPGE2-treated displaying either a WT phenotype (left) or an affected phenotype (right). The respective penetrance is listed within each image area. Scale bar, 100 μm.

Data are represented as mean ± SD. *p < 0.05, **p < 0.01, and ***p < 0.001 (one-way ANOVA); n.s., not significant.

Because previous research identified an essential role for ptgs1 in MCC development (Marra et al., 2019a), we investigated if the ppargc1a/ptgs1 relationship was relevant to MCC fate choice. WISH was used to detect expression of the MCC marker odf3b, in WT, ppargc1a SMO, ptgs1 SMO, and the ppargc1a + ptgs1 SMO (Figure 4G). Significantly decreased MCC numbers were observed in ppargc1a SMO and ptgs1 SMO compared with WT controls (Figures 4G and 4H). MCC number in ppargc1a + ptgs1 SMO embryos was significantly less than the quantity seen in either single SMO injected group, excluding an additive consequence (Figures 4G and 4H). Furthermore, MCC number was rescued with the addition of dmPGE2 in ppargc1a + ptgs1 SMO embryos (Figures 4G and 4H). These data provide evidence of a synergistic effect on renal MCC fate choice in the context of ppargc1a and ptgs1 suboptimal knockdown.

As loss of ppargc1a or ptgs1 results in morphological defects, including a curved body, we also performed live imaging studies with our suboptimal knockdown model. The combined loss of ppargc1a and ptgs1 dramatically increased the incidence of embryos with affected phenotypes (curved body axis) compared with single SMO knockdowns (Figure 4I; Figure S4D). Additionally, the number of affected ppargc1a + ptgs1 SMO embryos was nearly eliminated following dmPGE2 treatment (Figure 4I; Figure S4D). Overall, this evidence strongly indicated a link between ppargc1a and ptgs1 in ciliogenesis and MCC development in the pronephros.

To further interrogate the relationship between ppargc1a and ptgs1, we examined the promoter region 1.5 kb upstream of the ptgs1 start site. Our analysis identified six putative Ppargc1a consensus sites within 1.5 kb of the ptgs1 open reading frame (Figure 5A) (Charos et al., 2012). Subsequently, we investigated the expression of each gene by WISH and quantitative PCR in reciprocal knockdown studies. The expression of ppargc1a was unchanged in ptgs1 morphants (Figures S5A and S5B). However, the expression of ptgs1 was decreased in ppargc1a-deficient zebrafish (Figures 5B and 5C). These results, in light of the previous rescue experiments, suggested that ppargc1a may directly enhance prostaglandin signaling output (Figures 3 and 4). To test this potential relationship, we performed rescue experiments with ptgs1 capped mRNA injected in ppargc1a morphants, then completed IF for cilia and basal bodies or WISH for odf3b. We observed a rescue of cilia length, percentage of ciliated basal bodies, and mean fluorescent intensity in both the proximal and distal portions of the pronephros (Figures 5D–5I; Figures S5C and S5D). Also, the number of MCCs present in ppargc1a morphants injected with ptgs1 cRNA was rescued compared with ppargc1a morphants (Figures 5J and 5K). Altogether, these data led us to conclude that ppargc1a promotes ptgs1 expression in the pronephros and is needed for proper MCC development and ciliogenesis (Figure 5L).

Figure 5.

ptgs1 Functions Downstream of ppargc1a during Ciliogenesis

(A) Schematic depicting several Ppargc1a consensus sites located within 1.5 kb of the ptgs1 open reading frame (ORF).

(B) Twenty-four hours post-fertilization WT (top) and ppargc1a (bottom) zebrafish stained via WISH for ptgs1 (n > 60). Scale bar, 100 μm.

(C) Relative ptgs1 mRNA expression in WT and ppargc1a morphants.

(D) Whole-mount immunofluorescence for acetylated α-tubulin (cilia, green), γ-tubulin (basal bodies, red), and DAPI in the proximal pronephros of WT (top), ppargc1a morphants (middle), and ppargc1a morphants injected with ptgs1 cRNA (bottom). Scale bar, 7 μm.

(E) Cilia length for the proximal pronephros treatment groups shown in (D).

(F) Graph of the percentage of ciliated basal bodies/total basal bodies per 100 μm in the proximal pronephros.

(G) Whole-mount immunofluorescence for acetylated α-tubulin (cilia, green), γ-tubulin (basal bodies, red), and DAPI in the distal pronephros of WT (top), ppargc1a morphants (middle), and ppargc1a morphants injected with ptgs1 cRNA (bottom). Scale bar, 7 μm.

(H) Cilia length for distal pronephros treatment groups shown in (G).

(I) Graph of the percentage of ciliated basal bodies/total basal bodies per 100 μm in the distal pronephros.

(J) WISH for the MCC marker odf3b in WT (top), ppargc1a morphants, and ppargc1a morphants injected with ptgs1 cRNA (bottom). Scale bar, 90 μm.

(K) Number of MCCs per nephron.

(L) Proposed mechanism by which ppargc1a regulates MCC development and ciliogenesis via PGE2 levels by controlling ptgs1 expression.

Data are represented as mean ± SD. *p < 0.05, **p < 0.01, and ***p < 0.001 (one-way ANOVA or t test); n.s., not significant.

DISCUSSION

There is broad significance to understanding the genetic mechanisms that control ciliated cell development during embryogenesis. Ciliated cell development involves the distinct processes of fate choice, where for example a cell adopts a monociliated or MCC phenotype, and ciliogenesis, whereby a cell builds the correct number, length, and arrangement of cilia. Although there have been advances in the molecular understanding of ciliated cell development, much remains to be discovered. Recently, prostaglandin signaling was found to be necessary to promote ciliogenesis in numerous vertebrate cell types and during the specification of renal MCCs (Jin et al., 2014; Marra et al., 2019a). However, the genetic mechanisms by which prostaglandin signaling is regulated to perform these tasks were not determined. Here, we found that ppargc1a is necessary for ciliogenesis throughout the developing zebrafish and that it also dictates monociliated versus MCC fate choice in the pronephros, in both cases by modulating prostaglandin biosynthesis (Figure 5L). Through a number of experiments, we found that provision of the active prostanoid dmPGE2 rescued ppargc1a loss-of-function phenotypes. Furthermore, the results of genetic interaction studies and rescue experiments nicely complemented our evidence of decreased ptgs1 RNA and PGE2 metabolites and led us to determine that ppargc1a acts to mitigate ptgs1 expression that subsequently affects prostanoid levels. These findings unearth fascinating molecular functions of Ppargc1a as a regulator of prostaglandin signaling by exerting control over prostanoid production during ciliated cell development, where a necessary threshold of active prostanoid is essential to control cell fate and cilium growth (Figure 5L). Notably, our studies here also demonstrate that diminished MCC fate and defunct ciliogenesis are not inextricably linked phenotypes, as altered ciliogenesis in ift88-deficient embryos occurred simultaneously with normal MCC development in the pronephros. This leads us to conclude that the effects of Ppargc1a/Ptgs1/PGE2 activity on MCC fate and ciliogenesis are independent outcomes.

Long known for its roles in directing cellular mitochondrial biogenesis and metabolism, PGC-1α can function as a transcription factor but can also interact with chromatin remodeling machinery and RNA processing complexes (Knutti and Kralli, 2001). Interestingly, there has been mounting evidence that PGC-1α is vital to adult kidney health (Lynch et al., 2018). Additionally, although roles for PGC-1α during mammalian development have received comparatively little scrutiny, we found that embryonic kidney formation in the zebrafish is reliant on its ortholog ppargc1a (Chambers et al., 2018). In this context, ppargc1a is necessary for the proper segmentation of embryonic nephrons via reciprocal antagonism with the essential PST transcription factor Sim1a (Cheng and Wingert, 2015), a genetic interaction that negotiates the caudal boundary of the PST segment (Chambers et al., 2018). The consequence of ppargc1a deficiency is that embryos form an expanded PST (Chambers et al., 2018). In the present report we have now elucidated that skewing of tubule fate choice to favor the monociliated cell identity underlies this nephron composition change, where fewer MCC progenitors form in the absence of ppargc1a. As other organs contain MCCs, such as the respiratory and reproductive tracts, our findings lay the groundwork for future investigations of ppargc1a and prostanoid signaling in these tissue types, which could provide additional insights into MCC pattern formation.

In the context of the adult mammalian kidney, PGC-1α provides renoprotective effects in several injury states including ischemia, toxins or sepsis, where PGC-1α expression promotes the metabolic recovery of nephron tubular cells (Portilla et al., 2002; Tran et al., 2011, 2016; Ruiz-Andres et al., 2016; Lynch et al., 2018). Specifically, in the post-ischemic mouse kidney, it was found that PGC-1α regulates the biosynthesis of nicotinamide adenine dinucleotide (NAD), leading to augmentation of fat catabolism and increased prostaglandin production (Tran et al., 2016). Using the innate ability of the zebrafish to regenerate nephrons post-AKI and tracking the endogenous levels of ppargc1a during the regenerative process could help us link essential factors leading to nascent nephrons (Zhou et al., 2010; Diep et al., 2011; McCampbell et al., 2015; McKee and Wingert, 2015). There are also links between PGC-1α and chronic kidney disease, where decreased PGC-1α expression has been found in the human diabetic kidney (Sharma et al., 2013), as well as rat and murine models of diabetic nephropathy (Guo et al., 2015; Long et al., 2016). In these studies, cilium structure and function have not yet been described but would be of great importance to assess given our findings.

Interestingly, PGC-1α expression is also diminished in Pkd1-deficient rodents, a model of polycystic kidney disease (PKD), which has profound cilia defects (Ishimoto et al., 2017). Additionally, PGC-1α levels are directly correlated with HNF1B activity in renal cells, and HNF1B directly binds to the ppargc1a promoter in mouse kidney cells (Casemayou et al., 2017). Mutations in HNF1B are known to cause a spectrum of kidney diseases, including PKD, a common ciliopathy (Verhave et al., 2016). In contrast, global or nephron-specific knockout of PGC-1α alone has not yet been linked to changes in renal development or ciliary defects in mice (Lin et al., 2004; Tran et al., 2011; Svensson et al., 2016). However, in light of the present work, this may require further scrutiny. There have been conflicting reports about whether global knockout of PGC-1α results in progeny below the expected Mendelian ratio (Lin et al., 2004; The Jackson Laboratory, 2020), and PGC-1α-knockout pups are more prone to post-natal lethality, heart failure, and neurodegeneration, all facets that might signal the existence of an underlying ciliopathy (Lin et al., 2004; Arany et al., 2006). Alternatively, the functions of PGC-1α in higher vertebrates may not align with Ppargc1a in the zebrafish, despite the high sequence similarity of these orthologs that suggests functional conservation (Chambers et al., 2018).

An additional layer of complexity to fully discerning the roles of PGC-1α during embryogenesis is the relatively recent appreciation that mammals express numerous splice variants of PGC-1α, which exhibit tissue-specific expression patterns and exert different functions (Martínez-Redondo et al., 2015). The existence of these other PGC-1α isoforms could possibly underlie the absence of more severe phenotypes in murine knockout animals that survive the post-natal period, because the floxed region does not target some of the variants identified to date (Lin et al., 2004; Tran et al., 2011, 2016; Svensson et al., 2016; Zerbino et al., 2018). Furthermore, the nephron-specific PGC-1α conditional knockout mice exhibited significantly reduced, but not abrogated, expression of PGC-1α transcripts and therefore likely represent a hypomorphic loss-of-function model (Svensson et al., 2016). Future efforts to exhaustively isolate and characterize variants of PGC-1α are needed to assess whether there are functional isoforms that are essential to renal lineage ontogeny in mammals.

Other intriguing directions include elucidating the relationship between ppargc1a with other factors that control MCC fate and ciliogenesis. Previous research has found that ptgs2, which encodes the cyclooxygenase Cox2, is also important for MCC specification and ciliogenesis during MCC maturation (Marra et al., 2019a). This study did not reveal a significant difference in MCC development when comparing loss of ptgs1 with loss of ptgs2. However, further research could be completed to identify if ptgs2 expression is also affected by ppargc1a expression or if it acts in a compensatory manner when ppargc1a and ptgs1 are absent. Other options such as CaMK-II and histone deacetylases (HDACs), which were recently shown to have a role in zebrafish ciliogenesis, present possible avenues of exploration (Rothschild et al., 2018). Furthermore, future studies to explore the relationship of ppargc1a/prostanoid signaling with requisite MCC regulators such as Notch signaling, transcription factors (such as mecom, etv5a, irx2a, and e2f5), and ciliogenesis regulators (such as foxj1) have merit as well (Deblandre et al., 1999; Hellman et al., 2010; Li et al., 2014; Liu et al., 2007; Ma and Jiang, 2007; Marra et al., 2016, 2019b; Tsao et al., 2009; Xie et al., 2020).

In conclusion, in this work we have determined essential functions of ppargc1a in ciliated cell development in the zebrafish, identifying ppargc1a as a key regulator of prostaglandin synthesis through its control of ptgs1 expression. These results may be relevant for understanding the mechanisms of ciliated cell formation across tissues given the pleiotropic cilia defects that occur in ppargc1a zebrafish mutant embryos. Furthermore, considering the human disease burden related to ciliary defects, these findings may suggest PGC-1α as a target of interest for therapeutic applications.

STAR⋆METHODS

RESOURCE AVAILABILITY

Lead Contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Rebecca A. Wingert (rwingert@nd.edu).

Materials Availability

This study did not generate any unique reagents. The zebrafish ppargc1a mutant line is distributed by the Zebrafish International Resource Center.

Data and Code Availability

All used software is listed in the Key Resources Table. This study did not generate any unique datasets or new code.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Anti-tubulin acetylated	Sigma	T6793; RRID:AB_477585
Anti-γ-tubulin	Sigma	T5192; RRID:AB_261690
Anti-Ppargc1a	Abcam	AB54481; RRID:AB_881987
Anti-PKC	Santa Cruz	SC216; RRID:AB_2300359
Goat anti-Mouse IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor 647	Invitrogen	A21236; RRID:AB_2535805
Goat anti-Rabbit IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor 647	Invitrogen	A21245; RRID:AB_141775
Goat anti-Mouse IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor 568	Invitrogen	A11031; RRID:AB_144696
Goat anti-Rabbit IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor 594	Invitrogen	A11037; RRID:AB_2534095
Goat anti-Mouse IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor 488	Invitrogen	A11029; RRID:AB_138404
Goat anti-Rabbit IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor 488	Invitrogen	A11034; RRID:AB_2576217
Chemicals, Peptides, and Recombinant Proteins
16,16-Dimethyl-prostaglandin E2	Santa Cruz Biotechnology, Inc	sc-201240
40 kDa dextran-FITC	Invitrogen	D-1845
Critical Commercial Assays
Prostaglandin E2 Express ELISA Kit	Cayman Chemical	500141
mMESSAGE mMACHINE SP6 Transcription kit	Ambion	AM1340
PerfeCTa SYBR Green SuperMix with ROX	Quantabio	VWR 101414-160
qScript cDNA SuperMix	Quantabio	VWR 101414-106
TSA Plus Fluorescein	Akoya Biosciences	NEL741001KT
TSA Plus Cyanine	Akoya Biosciences	NEL744001KT
Experimental Models: Organisms/Strains
ppargc1a-sa13186	ZIRC	N/A
Oligonucleotides
ppargc1a qRT-PCR forward AATGCCAGTGATCAGAGCTGTCCTT	This paper	N/A
ppargc1a qRT-PCR reverse GTTCTGTGCCTTGCCACCTGGGTAT	This paper	N/A
18S qRT-PCR forward TCGGCTACCACATCCAAGGAAGGCAGC	This paper	N/A
18S qRT-PCR reverse TTGCTGGAATTACCGCGGCTGCTGGCA	This paper	N/A
ptgs1 qRT-PCR forward CATGCACAGGTCAAAATGAGTT	This paper	N/A
ptgs1 qRT-PCR reverse TGTGAGGATCGATGTGTTGAAT	This paper	N/A
ppargc1a-sa13186 genotyping forward primer GGGCCGGCATGTGGAATGTAAAGACTTAAACATGCCAACCTCCACTACTACGACATATCGTTGTCTTCCACCCCCCCTTCGTCTTCCTCACTGGCCAGG	Chambers et al., 2018	N/A
ppargc1a-sa13186 genotyping reverse primer TCCCACTACCCCGCTATAGAAGGCTTGCTGAGGCTTTCCAAAGTGCTTGTTGAGCTCGTCCCGGATCTCCTGGTCCCTAAGAAGTTTCCTGCCACCAGAA	Chambers et al., 2018	N/A
ppargc1a morpholino CCTGATTACACCTGTCCCACGCCAT	Gene-Tools	ZFIN: MO1-ppargc1a
ptgs1 morpholino TCAGCAAAAAGTTACACTCTCTCAT	Gene-Tools	N/A
standard control morpholino. CCTCTTACCTCAGTTACAATTTATA	Gene-Tools	N/A
ift88 morpholino AGCAGATGCAAAATGACTCACTGGG	Gene-Tools	N/A
Software and Algorithms
Prism v8	Graphpad	https://www.graphpad.com/scientific-software/prism/
ImageJ	Fiji	ImageJ (https://imagej.nih.gov/ij/)

EXPERIMENTAL MODEL AND SUBJECT DETAILS

The Center for Zebrafish Research at the University of Notre Dame maintained the zebrafish used in these studies and experiments were performed with approval of the University of Notre Dame Institutional Animal Care and Use Committee (IACUC), under protocol numbers 16–07-3245 and 19–06-5412.

Animal models

Tübingen strain zebrafish were used for all studies, with the ppargc1a mutation propagated on this background. Zebrafish were raised and staged as described (Kimmel et al., 1995). For all studies, embryos were incubated in E3 medium at 28°C until the desired developmental stage, anesthetized with 0.02% tricaine, and then fixed using 4% paraformaldehyde/1x PBS (PFA) (Westerfield, 1993). Embryos were analyzed before sex determination, so the influence of gender cannot be reported in the context of this study.

METHOD DETAILS

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

WISH was performed as previously described (Cheng et al., 2014; Galloway et al., 2008; Lengerke et al., 2011; Marra et al., 2019c) with antisense RNA probes either digoxigenin-labeled (ppargc1a, odf3b, slc20a1a, trpm7, jag2b, myl7, ptgs1) or fluorescein-labeled (odf3b) via in vitro transcription using IMAGE clone templates as previously described (Wingert et al., 2007; O’Brien et al., 2011; Gerlach and Wingert, 2014). FISH was performed as described (Brend and Holley 2009; Marra et al., 2017) using TSA Plus Fluorescein or Cyanine Kits (Akoya Biosciences). For all gene expression studies, every analysis was done in triplicate for each genetic model, and measurements were scored in a blinded fashion, with sample sizes of n > 20 per replicate.

Sectioning

Zebrafish were embedded for JB4 sectioning and counterstained as previously described (Gerlach and Wingert, 2014).

Dextran-FITC injections

ppargc1a heterozygous incrosses were incubated in 0.003% phenylthiourea (Sigma, P7629) in E3 from 24 hpf until 48 hpf. Zebrafish were anesthetized in 0.02% tricaine and microinjected was performed directly into the circulation as described (Kroeger et al., 2017) at the 48 hpf time point with 5 mg/ml 40 kDa dextran-FITC (Invitrogen, D1845). Embryos were again incubated in 0.003% PTU until 72 hpf when they were imaged.

Immunofluorescence (IF)

Whole mount IF experiments were completed as previously described (Gerlach and Wingert, 2014; Kroeger et al., 2017; Marra et al., 2017, 2019c). For cilia and basal bodies, anti-tubulin acetylated diluted 1:400 (Sigma T6793) and anti γ-tubulin diluted 1:400 (Sigma T5192) were used, respectively. The anti-Ppargc1a was diluted 1:150 (Abcam, AB54481) (Ran et al., 2017, Blechman et al., 2011), and anti-PKC was diluted 1:500 (Santa Cruz, SC216). Anti-rabbit and anti-mouse secondary antibodies were diluted 1:500 (Alexa Fluor, Invitrogen).

Rescue Experiments with dmPGE2

Chemical treatments were completed as previously described (Marra et al., 2019a; Poureetezadi et al., 2014; Poureetezadi et al., 2016). 16,16-Dimethyl-prostaglandin E2 (Santa Cruz Biotechnology, Inc, SC-201240) was dissolved in 100% dimethyl sulfoxide (DMSO) to make 1 M stocks then diluted to the 100 μM treatment dose. Treatments were completed in triplicate with n > 20 embryos per replicate.

PGE2 metabolite quantification

PGE2 metabolite quantifications were completed according to the manufacturer’s protocol (Cayman Chemical #500141). In brief, groups of 50 WT or ppargc1a MO injected zebrafish were pooled, anesthetized, and flash frozen in 100% ethanol. Lysates were homogenized and supernatant was isolated after centrifugation at 4 degrees (12,000 RPM for 10 minutes). Kit reagents were used to complete plate set up and manufacturer’s protocol was followed for assay completion using a plate reader (SpectraMax ABSPlus) at 420 nm.

Quantitative real-time PCR

Groups of 30 zebrafish (WT, ppargc1a morphants, or ptgs1 morphants) were pooled at 24 hpf. Trizol (Ambion) was used to extract RNA, qScript cDNA SuperMix (QuantaBio) was used to make cDNA. PerfeCTa SYBR Green SuperMix with ROX (QuantaBio) was used to complete qRT-PCR with 100 ng for ptgs1 and ppargc1a and 1 ng for 18S controls being optimal cDNA concentrations. The AB StepOnePlus qRT-PCR machine was used with the following program: 2 minute 50°C hold, 10 minute 95°C hold, then 35 cycles of 15 s at 95°C and 1 minute at 60°C for denaturing and primer annealing and product extension steps respectively. Each target and source were completed in biological replicates and technical replicates each with the median Ct value normalized to the control. Data analysis was completed by using delta delta Ct values comparing WT uninjected to the respective morphant groups with 18S as a reference. Primers used include: ppargc1a forward 5′–AATGCCAGTGATCAGAGCTGTCCTT–3′ and reverse 5′–GTTCTGTGCCTTGCCACCTGGGTAT–3′. To target 18S: forward 5′–TCGGCTACCACATCCAAGGAAGGCAGC–3′ and reverse 5′–TTGCTGGAATTACCGCGGCTGCTGGCA–3′. To target ptgs1: forward 5′- CATGCACAGGTCAAAATGAGTT- 3′ reverse: 5′-TGTGAGGATCGATGTGTTGAAT-3′.

Genetic models

The ppargc1a line (ppargc1asa13186/sa13186) was obtained from ZIRC (Eugene, OR) (Busch-Nentwich et al., 2013). Homozygous mutant and heterozygous zebrafish were identified as previously described (Chambers et al., 2018) using genotyping primers, forward 5′–GGGCCGGCATGTGGAATGTAAAGACTTAAACATGCCAACCTCCACTACTACGACATCATCGTTGTCTTCCACCCCCCC TTCGTCTTCCTCACTGGCCAGG–3′, and reverse 5′–TCCCACTACCCCGCTATAGAAGGCTTGCTGAGGCTTTCCAAAGTGCTTGTTGAGCTCGTCCCGGATCTCCTGGTCCCTAAGAAGTTTCCTGCCACCAGAA–3′. Antisense morpholino oligonucleotides (MOs) were obtained from Gene Tools, LLC (Philomath, OR). MOs were solubilized in DNase/RNase free water to generate 4 mM stock solutions which were stored at −20°C. Zebrafish embryos were injected at the 1-cell stage with 1–2 nL of diluted MO. ppargc1a was targeted with the following validated MO: 5′–CCTGATTACACCTGTCCCACGCCAT–3′ (400 μM optimal, 200 μM suboptimal) (Hanai et al., 2007; Bertrand et al., 2007, Chambers et al., 2018). ptgs1 was targeted with the following validated MO: 5′-TCAGCAAAAAGTTACACTCTCTCAT-3′ (400 μM optimal, 200 μM suboptimal) (Marra et al., 2019a; Poureetezadi et al., 2016, North et al., 2007). The ift88 MO was 5′-AGCAGATGCAAAATGACTCACTGGG-3′ (100uM) (Vasilyev et al., 2009; Gerlach and Wingert, 2014). Control MO was 5′-CCTCTTACCTCAGTTACAATTTATA-3′ (400uM) (Gene Tools, LLC).

Image acquisition and phenotype quantification

A Nikon Eclipse Ni with a DS-Fi2 camera was used to image WISH samples and live zebrafish. Live zebrafish were mounted in methyl cellulose with trace amounts of tricaine present. IF and FISH images were acquired using a Nikon C2 confocal microscope.

QUANTIFICATION AND STATISTICAL ANALYSIS

Cilia phenotypes were quantified using ImageJ/Fiji (https://imagej.nih.gov) software tools. All measurements were completed on representative samples imaged at 60X magnification. The multi-point tool was used for counting. The segmented line tool was used for length measurements. Fluorescent intensity plots were generated with the plot profile function. Each experiment was completed in triplicate. From these measurements an average and standard deviation (SD) were calculated, and unpaired t tests or one-way ANOVA tests were completed to compare control and experimental measurements using GraphPad Prism 8 software. Statistical details for each experiment are located in the figure legends.