Aberrant centrosome biogenesis disrupts nephron and collecting duct progenitor growth and fate resulting in fibrocystic kidney disease

Mammalian kidney development is a highly coordinated process that requires reciprocal signaling between the cap mesenchyme (CM) and ureteric bud (UB) progenitor cells, which is crucial for their self-renewal, proliferation, and subsequent differentiation (Kopan et al., 2014; Little, 2015; Little and Lawlor, 2020; McMahon, 2016). The UB progenitors undergo extensive growth and several generations of branching morphogenesis to establish the collecting duct system. At each UB tip branching event, the CM cells similarly proliferate and undergo self-renewal. The progenitors then begin to differentiate, form a tight cluster called the pretubular aggregate located beneath the UB branch tip, then epithelialize to form a renal vesicle (RV). The RVs subsequently undergo a series of morphological changes and patterning to form the curved pretubular structures termed the comma- and S-shaped bodies. These ultimately give rise to the mature nephrons, including cells that make up the glomerulus, proximal tubules and distal tubules (Little and Lawlor, 2020; McMahon, 2016). Mutations that disrupt the reciprocal signaling between these two progenitor niches negatively impact their growth and differentiation, leading to defective branching morphogenesis, abnormal nephron development and endowment, congenital renal dysplasia and early-onset fibrocystic kidney diseases (Kopan et al., 2014; Little, 2015; Little and Lawlor, 2020; McMahon, 2016).

A significant number of genes that are mutated in individuals with cystic and fibrotic renal pathologies encode cilia- and centrosome-associated proteins (Barroso-Gil et al., 2021; Mansour et al., 2021; Reiter and Leroux, 2017). The centrosome is the major microtubule-organizing center in mammalian cells, with important roles in regulating the microtubule cytoskeleton in interphase and mitosis (Mascanzoni et al., 2022; Petry and Vale, 2015). Cilia are hair-like organelles that are templated by the centrosome and protrude from the apical surface of most cells, including the majority of cell types in the kidney (Hoshi et al., 2015; Marra et al., 2016). The centrosome-cilium complex acts as a signaling hub to regulate cell-cell communication, cell proliferation, differentiation, fate determination and ultimately tissue formation. Mutations in several centrosome-ciliary genes cause kidney diseases including Autosomal Dominant Polycystic Kidney Disease (ADPKD), Autosomal Recessive Polycystic Kidney Disease (ARPKD), and Nephronophthisis (NPH) (Harris and Torres, 2009; Lu et al., 2017; Ma et al., 2017). In addition, mutation in these organelles can cause multi-organ disease syndromes called ciliopathies, which include Joubert syndrome (JS) and Jeune Asphyxiating Thoracic Dystrophy (JATD) (Braun and Hildebrandt, 2017). Of note, these ciliopathies cause kidney phenotypes that can vary dramatically between individuals. For example, ADPKD is characterized by slow formation of cysts that leads to progressive enlargement of the kidneys, damaging of the renal parenchyma causing fibrosis and a gradual decline in kidney function leading to end-stage kidney disease (ESKD) by age 50-60 (Agborbesong et al., 2022; Budhram et al., 2018). In contrast, NPH is a rare autosomal recessive renal ciliopathy which typically causes ESKD in children and young adults. Unlike ADPKD, renal cysts are not a hallmark of NPH, which is characterized by smaller, hyperechogenic kidneys with cortico-medullary cysts and poor cortico-medullary differentiation, tubular atrophy, disintegration and irregular thickening of the tubular basement membrane, leading to fibrotic kidneys with shrunken appearance (McConnachie et al., 2021). Individuals with JS and JATD commonly display dysplastic kidneys and fibrocystic disease, and reach ESKD within adolescence (Meyer et al., 2022; Spahiu et al., 2022). So far, ciliopathies have been linked to mutations in ∼200 genes (Barroso-Gil et al., 2021; Reiter and Leroux, 2017), a significant number of which encode centrosomal proteins. Many of these mutations result in congenital kidney developmental defects and early-onset fibrocystic disease, pointing to a crucial role for centrosomes in the development of embryonic kidneys.

The centrosome comprises of a pair of centrioles surrounded by a mesh of proteins called the pericentriolar material. The older of the two centrioles templates the assembly of the cilium. Centrosomes are duplicated once every cell cycle and segregated to each daughter cell following mitosis. This process is tightly controlled such that each cell inherits a single centrosome and forms a single cilium (Fu et al., 2015; Werner et al., 2017). Mutations in genes that disrupt the centrosome duplication cycle can result in daughter cells that contain aberrant centrosome numbers, which in turn lead to pathological phenotypes (Nigg and Raff, 2009). Specifically, mutations in centriole duplication factors block centrosome biogenesis and result in cells lacking centrosomes after several rounds of cell division, a phenomenon termed Centrosome Loss (CL) (Fong et al., 2016; Meitinger et al., 2016). Several cellular and molecular changes occur following CL. As centrosomes facilitate the assembly of the mitotic spindle, cell cycle progression and mitotic timing are often impaired in cells lacking centrosomes (Fong et al., 2016; Meitinger et al., 2016). CL in wild-type (WT) cells leads to prolonged mitosis, activation of the p53BP1-USP28-TP53-dependent mitotic surveillance mechanism, cell cycle arrest and caspase-mediated apoptosis (Fong et al., 2016; Meitinger et al., 2016). Inducing CL globally in mice causes prometaphase delay and p53-dependent apoptosis, which prevents embryo development upon midgestation and causes lethality by embryonic day (E) 9 (Bazzi and Anderson, 2014). Conditional induction of CL in the developing mouse brain leads to neural progenitor cell death, as well as premature differentiation of the progeny into neurons, ultimately leading to microcephaly phenotypes (Phan et al., 2021). Inducing CL in the developing lung endoderm similarly causes p53-mediated apoptosis, but only in proximal airway cells with low extracellular signal-regulated kinase (ERK) activity (Xie et al., 2021). In contrast, high ERK activity in lung progenitors at the distal bud tips confers protection against CL-induced apoptosis, and these cells are able to survive and proliferate (Xie et al., 2021). Furthermore, centrosomes of the intestinal endoderm appear to be dispensable for progenitor cell growth and differentiation during development. Thus, the consequences of CL on progenitor cell growth and fate differs depending on the specific type of cell, tissue, organ and developmental context.

Centrosomal protein 120 (Cep120) is a daughter centriole-enriched protein that plays essential roles in centriole duplication, elongation and maturation (Lin et al., 2013; Mahjoub et al., 2010; Wu et al., 2014). Previous studies have shown that depletion of Cep120 in proliferating cells results in defective centriole biogenesis and loss of centrosomes after two cell divisions (Lin et al., 2013; Mahjoub et al., 2010; Wu et al., 2014). In addition, Cep120 depletion in non-dividing quiescent cells disrupts centrosome homeostasis, leading to aberrant ciliary assembly and signaling (Betleja et al., 2018). Genetic ablation of Cep120 globally in mice causes embryonic lethality (Wu et al., 2014), whereas siRNA-mediated depletion or conditional deletion of the gene in cerebellar granule neuron progenitors (CGNPs) causes hydrocephalus and cerebellar hypoplasia (Wu et al., 2014; Xie et al., 2007). Importantly, mutations in CEP120 were recently identified in JATD and JS, with affected individuals showing complex ciliopathy phenotypes including severe congenital kidney developmental defects and early-onset fibrocystic pathologies (Joseph et al., 2018; Roosing et al., 2016; Shaheen et al., 2015). Yet, how mutations in CEP120 impact kidney development and cause the fibrocystic kidney disease phenotypes remains unknown.

Here, we investigated the consequences of CL on renal progenitor cell growth and differentiation, nephron formation, kidney development and function. We demonstrate that conditional knockout (KO) of Cep120 in the CM or UB cells during embryonic kidney development blocks centrosome biogenesis and causes reduced progenitor cell abundance. Cep120 loss caused premature differentiation of both progenitor niches, early formation of pre-tubular structures, decreased branching morphogenesis and small dysplastic kidneys at birth. Postnatally, the centrosome-less kidneys rapidly developed fibrosis and formed small cysts, which together resulted in a decline in renal function. Finally, transcriptional profiling of embryonic and postnatal kidneys from Cep120-KO mice identified several unique developmental, inflammatory, fibrogenic and cystogenic signaling pathways involved in the development of the disease phenotypes.

To disrupt centrosome biogenesis and cause CL in a spatiotemporally controlled manner in nephron and collecting duct progenitor cells, we used a recently developed mouse model harboring a floxed allele of Cep120 (Wu et al., 2014; Fig. 1A). To induce CL in the CM, Cep120^F/F mice were crossed with the Six2-Cre strain (Kobayashi et al., 2008) that expresses Cre-recombinase in the CM lineage (hereafter called Six2-Cep120; Fig. 1A,B). To cause CL in the UB cells, Cep120^F/F mice were mated to a Hoxb7-Cre strain (Zhao et al., 2004; hereafter named Hoxb7-Cep120; Fig. 1A,B). To compare the differences between Cep120 loss and a ciliary signaling defect in nephron progenitors and UB during embryonic kidney development, we crossed Pkd1^F/F mice (Starremans et al., 2008) to the same two Cre-recombinase expressing strains (hereafter named Six2-Pkd1 and Hoxb7-Pkd1; Fig. 1A,E). We examined the extent and specificity of Cep120 KO by staining kidney sections from E13.5 mice with antibodies against Cep120. Compared with WT control littermates, Cep120 expression was lost in the Six2⁺ mesenchyme but still evident in the UB cells of Six2-Cep120 mice (Fig. S1A). Similarly, Cep120 signal was missing in E-cadherin⁺ UB of Hoxb7-Cep120 KO mice, but present in the CM (Fig. S1A). Approximately 90% of CM cells had no Cep120 expression, whereas 85% of UB cells had lost Cep120 (Fig. S1B,C).

Next, we determined the effect of Cep120 loss on centrosome biogenesis in the two progenitor populations. In Six2-Cep120 kidneys there was a 71% decrease in centrosome numbers in the CM, whereas centrosome formation appeared to be normal in neighboring UB cells (Fig. 1B,C; Fig. S1E-H). Similarly, there was a 77% decrease in centrosome numbers in the UB cells of Hoxb7-Cep120 kidneys, whereas centrosome formation was unaffected in the CM (Fig. 1B,D; Fig. S1E-H). Importantly, there were no changes in Cep120 expression nor centrosome loss in either Hoxb7-Pkd1 or Six2-Pkd1 mice (Fig. S1D). Quantification of cilia number showed a large decrease in the CM of Six2-Cep120 and the UB cells of Hoxb7-Cep120 (Fig. S1I,J). Thus, loss of Cep120 results in a cell type-specific block in centrosome duplication and ciliogenesis during embryonic kidney development in our mouse models.

To determine the consequences of CL on overall kidney morphology, we performed hematoxylin and eosin (H&E) staining of kidney sections isolated at E13.5, E17.5 and postnatal day (P) 0. Compared with WT controls, kidneys of Six2-Cep120 and Hoxb7-Cep120 mice were smaller in size, which was evident as early as E13.5 and persisted until birth (Figs 1E and 6C; Fig. S2). In contrast, both Pkd1 KO mice displayed normal sized kidneys at embryonic stages, which rapidly developed cysts and were significantly enlarged compared with WT at P0 (Figs 1E and 6C; Fig. S2). These data indicate that impaired centrosome biogenesis results in unique embryonic kidney developmental defects, which are not evident upon loss of Pkd1 and ciliary signaling function.

To determine how centrosome biogenesis defects cause the small kidney phenotype, we investigated the growth of nephron progenitor cells and their differentiation into pretubular structures. First, we quantified the density of Six2⁺ CM at UB tip structures in the nephrogenic zone. Both Six2-Cep120 and Hoxb7-Cep120 KO kidneys had similar numbers of Six2⁺ CM cells compared with controls at E13.5 (Fig. 2A,C), indicating that Cep120 loss had not yet impacted the abundance of Six2⁺ cells at this stage. However, the amount of Six2⁺ mesenchyme was significantly reduced in both KO models at E17.5 (Fig. 2A,C; Fig. S4A). Notably, both Six2-Pkd1 and Hoxb7-Pkd1 KO kidneys had similar amounts of Six2⁺ cells compared with WT at E13.5 and E17.5 (Fig. 2A,C). These results suggest that Cep120 loss and ciliary signaling defects (due to Pkd1 ablation) have distinct effects on mesenchymal progenitor cell growth and renewal.

There are at least two mechanisms by which the growth of progenitor cells can be disrupted upon defective centrosome biogenesis: (1) CL can result in prolonged mitosis, activation of the p53-dependent mitotic surveillance pathway and caspase-mediated apoptosis, resulting in reduced stem cell growth (Fong et al., 2016; Lambrus et al., 2016; Meitinger et al., 2016; Phan et al., 2021); (2) CL can also disrupt cell fate and cause premature differentiation, depleting the pool of progenitor cells (Phan et al., 2021; Xie et al., 2021). To determine which of these processes is impacted in renal mesenchymal cells lacking centrosomes, we first asked whether cells with CL experience prolonged mitosis. This was determined by quantifying the fraction of mitotic cells (phosphorylated Histone H3 positive; pHH3⁺) among the total pool of Ki67⁺ cycling progenitors (Phan et al., 2021). Both Six2-Cep120 and Hoxb7-Cep120 kidneys showed an increased mitotic index when compared with control kidneys (Fig. S3A,B), suggesting that cells are delayed passaging through mitosis. We next evaluated whether this mitotic delay leads to p53-dependent apoptosis, by staining kidney samples for p53 and cleaved Caspase 3. There was a significant increase in the percentage of cells with nuclear-enriched p53 in both Six2-Cep120 and Hoxb7-Cep120 kidneys (Fig. S3C-E). The percentage of apoptotic cells (cleaved Caspase 3⁺) per unit area was also increased in kidneys of both Six2-Cep120 and Hoxb7-Cep120 at E13.5 (Fig. S3F-H).

Next, we examined the impact of CL on the differentiation capacity of mesenchymal progenitors into nephron tubules. Nephron formation commences when a few Six2⁺ CM cells at a new branch tip undergo mesenchymal-to-epithelial transition (MET) and sequential morphological alterations to form a pretubular structure called the renal vesicle, which then develops into the comma-shaped body (CSB), S-shaped body (SSB), and ultimately forms the mature nephron (Kopan et al., 2014) (Fig. 1A). Quantification of RV and SSB density in Six2-Cep120 and Hoxb7-Cep120 kidneys showed no difference compared with control mice at E13.5 (Fig. 2B,D,F). However, there was an increase in the abundance of CSB in Hoxb7-Cep120 kidneys evident as early as E13.5 (Fig. 2B,E), suggestive of potential acceleration in nephron tubular epithelia specification. Indeed, both KO kidneys contained significantly higher levels of RV and SSB structures compared to WT at E17.5 (Fig. 2B,D,F). This suggests that Six2⁺ CM lacking centrosomes undergo premature differentiation into nephron tubular structures. In contrast, the densities of RV, CSB and SSB in Six2-Pkd1 and Hoxb7-Pkd1 KO kidneys were unchanged at all embryonic stages (Fig. 2B-F). Collectively, these data indicate that Cep120 loss causes both cellular apoptosis and premature differentiation of mesenchymal progenitors, which together result in reduced abundance of this stem cell population.

Next, we determined the effects of CL on the growth of UB cells and branching morphogenesis. Whole kidneys from Six2-Cep120 and Hoxb7-Cep120 KO mice at E15.5 were isolated and optically cleared using the CUBIC protocol, then stained for the UB epithelium marker E-cadherin. These samples were transparent and permissive for imaging by light sheet microscopy (Hasegawa et al., 2019) (Fig. 3A). Quantification of branch tips and nodes using the Imaris software Filament Analysis tool showed that Six2-Cep120 and Hoxb7-Cep120 KO kidneys contain significantly lower number of branch tips compared with WT kidneys at the same developmental stage (Fig. 3B,C). In addition, the total number of branching nodes and the density of branches were lower compared with WT (Fig. 3D,E). These data point to potential defects in UB cell growth and renewal, which are essential for continuous branching morphogenesis (Short et al., 2018). Finally, there was an increase in nephron tubule diameter in Six2-Cep120 and Hoxb7-Cep120 KO kidneys at E15.5 (Fig. 3F), suggesting that dilations in these tubules occur early during development. Finally, quantification of the number of glomeruli, proximal tubules, distal tubules and collecting ducts showed a significant decrease in both Cep120-KO models at P0 (Fig. S4B-E), indicative of reduced nephron endowment. In sum, these data indicate that Cep120 loss in the UB niche is deleterious for UB cell growth, which is required for branching morphogenesis during early embryonic kidney development.

To identify pathways that are disrupted upon CL that result in the dysplastic kidney phenotype, we performed transcriptional profiling of centrosome-less embryonic kidneys at the onset of the observed developmental defects. As the Hoxb7-Cre transgene co-expresses GFP in the UB lineage (Zhao et al., 2004), we isolated UB cells from Cep120 KO kidneys and WT control at E13.5 by fluorescence-activated cell sorting (FACS), followed by bulk RNA-seq analysis (Fig. 4A). Differential gene expression analysis was performed using |log2 fold-change|≥2 and P-value <0.05. There were a total of 518 differentially expressed genes (DEGs) (upregulated, 397; downregulated, 121) in the Cep120 KO UB cells compared with WT (Fig. 4B; Table S3). Hallmark pathway and CompBio (Comprehensive Multi-omics Platform for Biological Interpretation; Higgins et al., 2022; Zou et al., 2020) analyses identified several key biological and developmental signaling pathways impacted upon Cep120 loss (Fig. 4C,D).

There was upregulation of mitotic spindle regulators including NUMA1, Cdk5rap2 and PCNT, which play crucial roles in organizing the mitotic spindle poles and controlling spindle orientation in the absence of centrosomes (Chou et al., 2016; Lee et al., 2014; Luo and Pelletier, 2014; So et al., 2022; Watanabe et al., 2020). Moreover, there was upregulation of components of the TNFα-NF-κB pathway, UV stress response, hypoxia and TGFβ signaling (Fig. 4C). Immunoblotting of lysates prepared from E13.5 kidneys confirmed changes in downstream effectors of TNFα-NF-κB, including increased phosphorylation of c-Jun and p65 (Fig. 4E,F), which signifies activation of these inflammatory pathways (Hayden and Ghosh, 2014; Poulton et al., 2019). In conjunction, the RNA-seq analysis indicated downregulation of oxidative phosphorylation and fatty acid metabolism (Fig. 4C). Consistent with this, there was increased levels of carnitine palmitoyl transferase 1A (CPT1A), which controls the entry of fatty acids into the mitochondria for oxidation (Jang et al., 2020) and a corresponding decrease in translocase of the outer mitochondrial membrane member 20 (TOMM20), a member of the outer membrane translocator complex, which together verify mitochondria dysfunction (Johnson et al., 2015) (Fig. 4E,F). Altogether, these changes suggest that Cep120 loss likely causes oxidative stress, activation of pro-inflammatory signaling and mitochondrial defects, which together can result in tubular injury and inflammation leading to fibrosis (discussed below) (Console et al., 2020; Jang et al., 2020; Poulton et al., 2019; Simon and Hertig, 2015).

Other pathways impacted upon Cep120 loss include components of G₂/M progression and p53 signaling (Fig. 4C), which are known to result in activation of the mitotic surveillance pathway and cell apoptosis (Fong et al., 2016; Lambrus et al., 2016; Meitinger et al., 2016; Phan et al., 2021; Xie et al., 2021). We verified the activation of this p53-dependent cell death by immunofluorescence staining of kidney sections (Fig. S3C-H) and immunoblotting of kidney lysates (Fig. 4E,F). In addition, there were changes in expression of regulators of the cell apical junction/surface (Fig. 4C), indicative of potential defects in apicobasal and/or basolateral polarity during tubule formation (Pieczynski and Margolis, 2011; Schluter and Margolis, 2012; Wilson, 2011). In support of this, we observed abnormal orientation of mitotic spindles during cell division (Fig. 5A,B). Mitotic spindles in normal developing kidney tubules have a stereotypical orientation, with the axis of the spindle parallel to the long axis of the tubule. This orientation becomes defective in cystic kidney diseases and upon defects in centrosome function (Fischer et al., 2006; Jonassen et al., 2008; Saburi et al., 2008). In the absence of Cep120, the normal bias toward spindle orientation being parallel to the long axis of the tubule was lost, resulting in different mitotic spindle orientations between control and Cep120 KO mice (Fig. 5A,B).

Finally, we performed transcriptional profiling of Pkd1 KO embryonic kidneys at the same embryonic stage, to provide a comparison between CL and ciliary signaling dysfunction. UB cells were isolated from Pkd1 KO and WT control at E13.5 using FACS (Fig. 4A) and analyzed using bulk RNA-seq. Surprisingly, there were no DEGs in Pkd1 KO embryonic kidneys compared with WT control (Fig. S5A). This suggests that Pkd1 loss has no observable transcriptional changes at this early stage, unlike ablation of centrosomes.

Several of the pathways/processes identified in the Hallmark and CompBio analyses are downstream effectors of Wnt signaling (Fig. S5B). The DEGs are involved in both canonical Wnt-β-catenin and non-canonical Wnt/PCP pathways. The downstream effectors of canonical Wnt-β-catenin signaling were regulators of adhesion junctions, TGFβ, MAPK pathways and cell cycle progression (Fig. S5B). This indicates that defective canonical Wnt-β-catenin signaling may underlie the observed defects in growth and renewal of progenitor cells upon Cep120 loss. In addition, there were changes in downstream effectors of non-canonical Wnt/PCP pathways (Fig. S5B), which are essential for cellular motility (Sedgwick and D'Souza-Schorey, 2016), adhesion (Astudillo and Larrain, 2014) and rearrangements of the cytoskeleton (Torban and Sokol, 2021).

We found that the expression of several Wnt genes was altered in the Cep120-loss group (Fig. S5C), some of which (e.g. Wnt5/7/11) function through the non-canonical Wnt/PCP pathways (Torban and Sokol, 2021). The most downregulated gene was Wnt11, which is involved in UB branching morphogenesis and convergent extension during tubulogenesis; its absence leads to kidney hyperplasia due to defective UB branching (Majumdar et al., 2003). Wnt11 expression is highly restricted to UB tip progenitor cells, where it plays an important role in promoting their self-renewal, as well as communication with the CM cells (O'Brien et al., 2018). Immunofluorescence staining of WT kidneys at E13.5 showed that Wnt11 expression was restricted to the UB tip cells, but not in differentiated stalk cells, as expected (Fig. 5C,E). However, the expression of Wnt11 was reduced in the UB tip cells in both Six2-Cep120 and Hoxb7-Cep120 KO kidneys, consistent with the RNA-seq results (Fig. 5C,E; Fig. S5C).

The most upregulated Wnt gene in our dataset was Wnt7b (Fig. S5C). Wnt7b expression is typically restricted to the differentiated ureteric stalk epithelium where it activates Wnt/PCP signaling in the surrounding medullary interstitium (Yu et al., 2009). Wnt7b is required for elongation of the collecting duct and loops of Henle, by controlling the tubular luminal diameter and length. Wnt7b KO kidneys lack normal tubular elongation in the medullary region, suggesting that Wnt7b may act locally in the UB trunk or non-autonomously on the cells adjacent to developing tubules to stabilize polarity of proliferating UB cells (Torban and Sokol, 2021). Immunofluorescence staining of WT kidneys at E13.5 showed that Wnt7 expression was exclusively in the stalk, but not UB tip, cells (Fig. 5D,F). In contrast, Wnt7 levels were significantly increased in the UB tip cells in both Six2-Cep120 and Hoxb7-Cep120 KO kidneys (Fig. 5D,F). Moreover, immunoblot analysis of E13.5 kidney lysates showed reduced levels of phosphorylated GSK3β and a corresponding increase in phosphorylated β-catenin (Fig. 4E,F), which together indicate that canonical Wnt signaling is decreased (MacDonald et al., 2009) upon Cep120 loss. In contrast, there was elevated expression of Rock2 and a corresponding decrease in Rac2 (Fig. 4E,F), which indicate upregulation of the Wnt/PCP arm of the pathway (Qin et al., 2024). Altogether, these data suggest that Cep120 loss leads to abnormal expression and activation of Wnt signaling, likely resulting in the pre-mature differentiation of UB cells and abnormal branching morphogenesis.

We next determined the consequences of Cep120 loss in postnatal mice. Both Six2-Cep120 and Hoxb7-Cep120 mice were born in the expected Mendelian ratios (Table S1). Analysis of kidneys isolated from Six2-Cep120 and Hoxb7-Cep120 KO animals at P15 showed a reduction in size compared with WT, which was already evident at P0 and became more prominent by P15 (Fig. 6A,C). Notably, the Cep120 KO kidneys were significantly smaller than Hoxb7-Pkd1 or Six2-Pkd1 KO (Fig. 6A,C). Importantly, Cep120 loss resulted in early onset cyst formation in both KO models (Fig. 6A). Quantification of the cyst index indicated that cystogenesis in Six2-Cep120 and Hoxb7-Cep120 KO kidneys was milder than that of Six2-Pkd1 and Hoxb7-Pkd1 mutants (Fig. 6A,D). To identify the origin of cysts in the KO kidneys, we stained P15 sections for Lotus tetragonolobus lectin (LTL; a marker of proximal tubules) and Dolichos biflorus agglutinin (DBA; a marker of collecting ducts). In the kidneys of the Six2-Cep120 KO mice, ∼66% of the cysts were LTL⁺ and thus proximal tubule-derived, whereas in Hoxb7-Cep120 KO kidneys ∼75% of cysts were DBA⁺ (Fig. S6A). These results indicate that the majority of cysts are specifically derived from progenitor cells ablated of centrosomes. Finally, quantification of Cep120, centrosome and cilia number showed a significant reduction in cystic epithelial cells of Cep120 KO mice at P15 (Fig. S6B-E).

Next, we investigated the extent of fibrosis in the KO animals. Immunostaining of kidneys isolated at P15 with the fibrotic marker α-SMA showed increased levels in Six2-Cep120 and Hoxb7-Cep120 KO kidneys compared with WT (Fig. 6B,E). The extent of fibrosis was significantly higher than that observed in Hoxb7-Pkd1 or Six2-Pkd1 KO animals (Fig. 6B,E), suggesting that Cep120 loss causes more fibrosis compared with Pkd1 loss. Notably, immunostaining against α-SMA of kidneys isolated at P0 showed fibrosis occurs early during kidney development upon Cep120 loss (Fig. S6F). Finally, analysis of blood urea nitrogen (BUN) levels showed an increase in Six2-Cep120 and Hoxb7-Cep120 KO mice, but lower than those of Six2-Pkd1 and Hoxb7-Pkd1 KO (Fig. 6F). In sum, loss of centrosomes results in small and fibrocystic kidneys after birth that are distinct from those observed in early-onset ADPKD mice.

We next sought to identify the pathways that cause the fibrotic and cystic disease phenotypes upon Cep120 loss. Kidneys were collected from Hoxb7-Cep120 KO and WT mice at P15, total RNA extracted and subjected to bulk RNA-seq analysis. There was a total of 94 DEGs, including 36 upregulated and 58 downregulated genes (Fig. 7A). The low number of DEGs at this time point compared with the embryonic kidneys (Fig. 4) is because centrosome-less cells could not be enriched using FACS (as the Hoxb7-CreGFP expression is only embryonic, and absent in adult kidneys). Hallmark pathway analysis identified gene sets that were mostly downregulated, with only the epithelial-to-mesenchymal transition (EMT) gene signatures upregulated (Fig. 7B). The downregulated pathways include xenobiotic metabolism and oxidative phosphorylation, which point to metabolic defects and dysfunctional mitochondria, both of which contribute to the development of cystic and fibrotic kidney diseases (Menezes and Germino, 2019). We further verified the involvement of mitochondrial defects by probing lysates of P15 kidneys for CPT1A and Tomm20 (Fig. 7D,E). In contrast, the upregulated gene set was mainly related to EMT-like and pro-inflammatory signaling (Fig. 7B). Previous studies have demonstrated that increased activity of the EMT-like pathway in tubular epithelial cells is associated with enhanced deposition of extracellular matrix (ECM), abnormal energy metabolism and aberrant proliferation of cyst epithelial cells (Fragiadaki et al., 2020; Wu et al., 2021). Similarly, CompBio analysis showed that the signaling networks most affected by the upregulated DEGs are ones involved in regulation of interstitial fibroblast differentiation, Wnt/PCP and cell differentiation, and ECM deposition, highlighting the processes that may be contributing to the fibrocystic phenotypes upon Cep120 loss (Fig. 7C).

We validated and confirmed the changes in pro-inflammatory signatures associated with TNFα-NF-κB signaling, including increased phosphorylation of c-Jun and p65 (Fig. 7D,E). We also verified the enhanced signatures associated with fibrosis and ECM deposition by immunoblotting for various components including matrix metallopeptidase 7 and 9 (Liu et al., 2020; Wang et al., 2010; Zhao et al., 2013), collagen I and α-SMA (Fig. 7D,E). Furthermore, we verified the increase in inflammatory/immune signatures by quantifying the extent of macrophage infiltration in the fibrocystic kidneys at P15 and found a ∼8- to 9-fold increase in the Cep120 null kidneys (Fig. S7A,B).

Next, we compared the signaling pathways that are disrupted in cystic kidneys following Cep120 loss versus ciliary signaling (Pkd1 KO) using bulk RNA-seq analysis of kidneys isolated from Hoxb7-Pkd1 KO mice at P15. There was a total of 1550 DEGs, including 786 upregulated and 764 downregulated genes (Fig. S8A). Hallmark analysis identified several signaling pathways known to be involved in cystogenesis and fibrosis in ADPKD (Fig. S8B). Subsequently, we compared the DEGs from the Cep120 KO and Pkd1 KO mice at P15 (Fig. S8C). There was a total of 1585 DEGs including 1032 upregulated and 553 downregulated genes in the Cep120 KO kidneys compared with Pkd1 (Fig. S8C). Hallmark analysis identified significant differences between the two cystic disease models with regards to the cellular processes and pathways impacted (Fig. S8D). The upregulated gene sets in Cep120 KO compared with Pkd1 KO mice include those involved in the G₂M checkpoint and E2F targets, which indicate that Cep120 loss causes more severe cell cycle defects compared with ADPKD models (Fig. S8D). Compared with Pkd1 loss, many signaling pathways involved in cystogenesis in ADPKD, such as TGFβ, JAK-STAT and KRAS signaling, are significantly lower in Cep120 KO kidneys. This suggests that these signaling pathways may be less active, and may explain the reduced severity of the cystic phenotype in Cep120 KO relative to Pkd1 KO kidneys.

In this study, we examined the consequences of disrupting centrosome biogenesis in nephron and collecting duct progenitor cells during embryonic kidney development. We observed that kidneys formed abnormally, were dysplastic and small in size, and became rapidly fibrotic and cystic. Ablation of Cep120 disrupted the balance in the growth and differentiation of progenitor cells, resulting in depletion of both the CM and UB pool of cells. There was premature cell differentiation and defective branching morphogenesis, ultimately leading to overall reduced nephron formation and abundance. The congenital developmental defects, as well as the early-onset fibrocystic phenotypes, in the Cep120 KO mice recapitulate the pathological kidney characteristics observed in individuals with NPH, JS and JATD ciliopathy. Importantly, the observed changes in progenitor cell growth and fate, nephron development, collecting duct branching and kidney morphology were different from the conditional ablation of Pkd1. Although the deletion of Pkd1 in mice causes a severe early-onset cystic disease phenotype that does not mimic the slow, degenerative human disease pathology (Rogers et al., 2016), we ablated the gene in the same progenitor cells to compare the consequences of defective centrosome biogenesis to ciliary signaling. Our results indicate that mutations which disrupt centrosome biogenesis and function cause defects in kidney progenitor cells that are more pronounced than defective ciliary signaling. Indeed, mutations in several other genes involved in centrosome biogenesis, such as POC1B and C2CD3, cause ciliopathies like JS and JATD, and these individuals develop the more severe early-onset fibrocystic kidney disease (Beck et al., 2014; Thauvin-Robinet et al., 2014).

We wondered why loss of Cep120 and centrosomes resulted in reduced abundance of nephron and UB progenitor cells, and development of the small, dysplastic kidney phenotype. Defective centrosome biogenesis can result in the loss of cell populations by a combination of cell cycle arrest, cell death or premature differentiation (Phan et al., 2021; Xie et al., 2021). We found that induction of CL in both CM and UB cells resulted in prolonged mitosis, activation of p53 and caspase-mediated apoptosis (Fig. S3A-F). This is consistent with previous studies in which depletion of centrosome proteins in neural progenitor cells in the brain causes prolonged mitosis and activation of the p53-mediated mitotic surveillance pathway, resulting in a reduced progenitor pool and differentiated neurons (Phan et al., 2021). In contrast, loss of centrosomes in Sox9⁺ lung progenitors, or in endodermal cells of the developing intestines, does not induce p53-mediated apoptosis. It has been proposed that high expression and activity of the ERK signaling pathway in the lung and intestinal progenitor cells confers protection against the p53-mediated apoptosis (Xie et al., 2021). However, this protective mechanism is unlikely to be in play in kidney progenitors, as there is consistent expression and activity of ERK in the CM and UB progenitor cells (Ihermann-Hella et al., 2018; Kurtzeborn et al., 2019), yet they do not survive following Cep120 loss. Moreover, it is unclear why, following the differentiation of the CM and UB into tubular epithelia, cells can survive and proliferate in the absence of centrosomes.

One possibility is that components of the mitotic surveillance pathway are differentially expressed in the nephron progenitor stage compared with differentiated tubular cells, such that the pathway is only activated in progenitors. Indeed, our RNA-seq analyses indicate that, following Cep120 loss, activators of the p53-mediated surveillance pathway (e.g. Pidd1; Burigotto and Fava, 2021; Weiler et al., 2022) were highly upregulated during the progenitor cell stage (E13.5), but not in fully differentiated tubular cells at P15 (Table S4). Thus, we believe that the observed apoptosis of the CM and UB populations may be at least partially a result of the p53-mediated surveillance mechanism. One way to test this would be to delete p53, which can enhance survival of cells lacking centrosomes in certain tissues (Bazzi and Anderson, 2014; Lambrus et al., 2016; Phan et al., 2021; Xie et al., 2021). However, conditional inactivation of p53 signaling in embryonic kidneys delays UB branching, disrupts proliferation of CM and causes renal hypoplasia (Saifudeen et al., 2009, 2012). Thus, we could not test the effects of blocking the p53-mediated surveillance pathway in centrosome-less kidneys.

A potential explanation for why loss of centrosomes disrupted nephron and collecting duct development is due to abnormal differentiation of the progenitor cells. Previous studies have established an important role for centrosomes in regulating stem cell differentiation in various organs. For example, disrupting centrosome biogenesis in the developing mouse brain disrupts the balance of asymmetric neuronal stem cell divisions by causing premature differentiation, and thus depletion of progenitor cells, leading to the small brain phenotype (Phan et al., 2021). The premature differentiation is due to defective inheritance of centrosomes by the daughter cells (Wang et al., 2009; Yamashita et al., 2003), which is linked to abnormal segregation of fate determinants in those cells. Therefore, we hypothesize that Cep120 (and centrosome) loss may be leading to premature differentiation of CM and UB progenitor cells by abnormal expression and/or segregation of some essential fate determinants. Our RNA-seq and immunoblot analyses of Cep120 KO embryonic kidneys identified prominent changes in several components of Wnt signaling (Fig. 4), several of which localize to centrosomes and have been shown to regulate cell fate (Itoh et al., 2009; Zhang et al., 2007). For example, diversin (Ankrd6), a known antagonist of β-catenin stability, depends on its centrosomal localization to appropriately promote β-catenin degradation (Itoh et al., 2009). Cells without centrosomes undergo attenuated response to Wnt and accumulate a distinct, higher-molecular-weight species of phosphor-β-catenin, which results in cell fate changes (Itoh et al., 2009; Zhang et al., 2007). Finally, mouse models of Wnt11 and Wnt7b (the two most differentially impacted Wnt genes in our dataset) show overlapping features with the Cep120 KO animals. Loss of Wnt11 results in ureteric branching morphogenesis defects and consequent kidney hypoplasia in newborn mice (Majumdar et al., 2003), whereas loss of Wnt7b disrupts tubular elongation in the medullary region, similarly disrupting embryonic kidney patterning (Torban and Sokol, 2021). In sum, we predict that the reduced abundance of progenitor cells observed upon CL is likely a combination of cell cycle delay, cell death and Wnt-associated premature differentiation, which together cause the observed developmental patterning defects.

Unlike the progenitor cell stage, loss of Cep120 and centrosomes paradoxically led to rapid growth of interstitial myofibroblasts (that cause fibrosis) and tubular epithelial cells (that form cysts) in adult kidneys. This indicates that CL has differential effects on the progenitor versus differentiated populations of kidney cells; the differentiated kidney cells may be more resistant to CL, or may be ‘protected’ against the cell cycle arrest and apoptosis observed during the progenitor stages. One potential mechanism for this protection is via enhanced EMT and pro-inflammatory signaling, which have been shown to help cells against the deleterious effects of CL and permit their growth. Our RNA-seq analyses of postnatal Cep120 KO fibrocystic kidneys identified strong signatures of EMT and pro-inflammatory signaling, suggesting that these pathways may be key for cell survival and growth in the absence of centrosomes, and progression to the fibrocystic kidney phenotype. Our future studies will test whether modulation of these pathways can inhibit the abnormal proliferation of differentiated interstitial and epithelial cells, and improve kidney morphology and function upon centrosome loss.

Finally, we wondered how loss of Cep120 (and centrosome) function impacts the stromal progenitor populations of the developing kidney, and how that contributes to the overall renal dysplasia, fibrosis and cyst formation observed in patients. In an accompanying study (Langner et al., 2023 preprint) we induced conditional deletion of Cep120 in the stromal progenitor cells of the developing kidney. Cep120 deletion disrupted centrosome biogenesis in stromal progenitor-derived cell types including pericytes, interstitial fibroblasts, mesangial, and vascular smooth muscle cells, and resulted in reduced abundance of several stromal cell lineages (interstitial pericytes, interstitial fibroblasts and mesangial cells). This led to development of small, hypoplastic kidneys with visible signs of medullary atrophy and delayed nephron maturation by P15 (Langner et al., 2023 preprint). The reduced interstitial cell populations were due to a combination of defective cell cycle progression of stromal progenitors lacking centrosomes, p53-mediated apoptosis, and changes in signaling pathways essential for differentiation of stromal lineages. This indicates that aberrant centrosome biogenesis in all three kidney progenitor populations is contributing to the small dysplastic/hypoplastic kidney phenotype in patients. Yet, there was no spontaneous fibrosis or early-onset cystogenesis after birth upon Cep120 loss and CL in the stromal progenitors, unlike the those observed in the Six2-Cep120 and Hoxb7-Cep120 KO mice (Fig. 6). However, CL in the interstitium sensitized kidneys of adult mice, causing rapid fibrosis via enhanced TGFβ/Smad3-Gli2 signaling after renal injury (Langner et al., 2023 preprint). Thus, our data collectively indicate that loss-of-function mutations in Cep120 (and other centrosome biogenesis genes) in the nephron and collecting duct progenitors likely play a major role in spontaneous fibrosis and cystogenesis, whereas mutations in the stromal progenitors contribute to the developmental defects and enhanced fibrosis following renal injury. In sum, our study provides a detailed characterization of the underlying molecular and cellular defects in renal centrosomopathies, and identifies unique pathways that may be leveraged for therapy.

All animal studies were performed following protocols that are compliant with guidelines of the Institutional Animal Care and Use Committee at Washington University and the National Institutes of Health. Conditional Cep120 and Pkd1 KO strains were generated by crossing Cep120^F/F or Pkd1^F/F mice with Six2-Cre or Hoxb7-Cre-expressing strains, respectively. Litters were genotyped by PCR using the following primers: Cep120 Flox forward, 5′-CCTCTGCCTCCTTAGTGGATC-3′; WT forward, 5′-ATCACTGTGGAGCCTTGGGCA-3′; WT reverse, 5′-TGTTACTCAGCAGCTGGTACC-3′; Pkd1 forward, 5′-GCCCACAGCTATTGTTCCTAA-3′; Pkd1 reverse, 5′-GGATAAAGTGATCAAGCAGCA-3′; Cre forward, 5′-CCAATTTACTGACCGTACACC-3′; Cre reverse, 5′-CGTAACAGGGTGTTATAAGCAA-3′.

Both kidneys were isolated from embryos (E13.5 and E17.5) or postnatal mice (P0 and P15), fixed with 4% paraformaldehyde in PBS for 24 h at 4°C, then embedded in paraffin. Kidney blocks were sectioned at 5-7 µm thickness using a microtome (Leica, RM2125 RTS), placed onto microscope slides (Thermo Fisher Scientific) and stored at room temperature. H&E staining of paraffin-embedded sections was performed using standard protocols (Feldman and Wolfe, 2014).

For immunofluorescence staining of paraffin-embedded sections, antigen unmasking was performed by boiling the slides in antigen-retrieval buffer (10 mM Tris Base, 1 mM EDTA, 0.05% Tween-20, pH 9.0) for 30 min. Samples were permeabilized with 0.05% Triton X-100 in PBS (PBS-T) for 10 min at room temperature, incubated in blocking buffer (3.0% bovine serum albumin and 0.1% Triton X-100 in PBS) for 1 h, followed by staining with primary antibodies overnight at 4°C (Table S2). After three washes with PBS-T, samples were incubated with secondary antibodies for 1 h at room temperature. All Alexa Fluor dye-conjugated secondary antibodies (Thermo Fisher Scientific) were used at a final dilution of 1:500. Nuclei were stained with DAPI and specimens mounted using Mowiol containing n-propyl gallate (Sigma-Aldrich). Images were captured using a Nikon Eclipse Ti-E inverted confocal microscope equipped with 10× Plan Fluor (0.30 NA), 20× Plan Apo air (0.75 NA), 40× Plan Fluor oil immersion (1.30 NA), 60× Plan Fluor oil immersion (1.4 NA) or 100× Plan Fluor oil immersion (1.45 NA) objectives (Nikon). A series of digital optical sections (z-stacks) were captured using a Hamamatsu ORCA-Fusion Digital CMOS camera at room temperature and three-dimensional image reconstructions were produced. Images were processed and analyzed using Elements AR 5.21 (Nikon), Adobe Illustrator and Photoshop software.

For quantification of Cep120 expression, centriole and centrosome number in the kidney tissue, kidney sections were immunostained with antibodies against Cep120, Cep135, PCNT and γ-tubulin (Table S2). A minimum of five tissue sections – a midsagittal section and sections generated in 5-15 µm increments in both directions from the midsagittal section – were analyzed to ensure equivalent representation of tissue regions (as previously described by Dionne et al., 2018). Sections from both kidneys were analyzed. A minimum of 1000 cells per mouse tissue was scored. Normal centrosome number was defined as cells containing one or two foci of each marker and loss of centrosomes was defined as cells containing zero foci.

Sagittal sections of P0 kidneys were immunostained for α-tubulin to mark spindles, E-cadherin to mark the tubular cells and pHH3 to mark DNA of cells in mitosis. Mitotic cells were imaged using 1 µm z-steps, 3D volume rendered and the angle between the long axis of the tubule and the spindle measured using Nikon Elements AR 5.21. The angles were then grouped into 30° bins and represented using Origins software.

To determine the density of Six2⁺ CM at different developmental stages, sections from WT, Six2-Cep120, Hoxb7-Cep120, Six2-Pkd1 and Hoxb7-Pkd1 kidneys at E13.5 and E17.5 were immunostained for Six2 (Table S2). For quantification of the density of the intermediate structures of the development of the nephron – RVs, CSBs and SSBs – samples were stained with antibodies against the structure markers Pax2 and WT1. The relative density of CM was determined by quantifying the number of Six2⁺ cells per unit area. Similarly, the relative abundance of RV, CSB and SSB structures was measured per unit area. The density of these structures was calculated by dividing the total number of each structure by the whole area of the whole kidney section area. Kidneys of five mice were analyzed per genotype.

Kidney weight (KW) and body weight (BW) were measured following isolation of kidneys at P0 and P15. KW/BW ratio was then calculated by dividing the average kidney weight by body weight (g).

The entire sagittal kidney section was used to calculate cyst burden. Sagittal kidney sections were stained with H&E and examined by light microscopy. All kidneys were imaged under the same magnification. ImageJ analysis software was used to calculate the cyst index, which refers to the cumulative area of cysts within the total area of the kidney.

For quantification of kidney fibrosis, kidney sections from WT, Six2-Cep120, Hoxb7-Cep120, Six2-Pkd1 and Hoxb7-Pkd1 mice at P0, P7 and P15 were immunostained with antibodies against α-SMA. Multiple regions of each kidney were imaged using a 20× objective then analyzed using ImageJ software. The data are expressed as the mean area of α-SMA+ foci per unit area (mm²) of each kidney section.

For quantification of macrophage infiltration, kidney sections from WT, Six2-Cep120 and Hoxb7-Cep120 mice at P15 were immunostained with antibodies against F4/80 to mark macrophages. Several regions of each kidney were imaged using a 20× objective then analyzed using ImageJ software. The data are expressed as the mean area of F4/80+ foci per unit area (mm²) of each kidney section.

For analysis of kidney function, the level of BUN was measured in mice at P15. Blood serum was separated (6000 g, 15 min at 4°C) from blood isolated by sub-mandibular bleeding and used for the BUN assay (BUN-Urea, BioAssay Systems) following the manufacturer's protocol.

Kidneys of WT, Six2-Cep120, Hoxb7-Cep120 at E15.5 were isolated, fixed in 4% paraformaldehyde in PBS for 24 h at 4°C, washed with PBS three times for 2 h each at 4°C. Whole kidneys were placed in CUBIC clearing buffer {25% urea, 25% Quadrol [N,N,N′,N′-Tetrakis (2-hydroxypropyl)ethylenediamine; EDTP], 15% Triton X-100 in ddH₂O} for 3 days with shaking at 37°C. The cleared, transparent kidneys were then washed three times with PBS-T for 2 h each at 4°C and stained with antibodies against E-cadherin (Table S2) for 48 h at 4°C. Subsequently, samples were washed with PBS-T three times for 2 h each at 4°C, then incubated with secondary antibody (goat anti-mouse IgG2a Alexa Fluor 594, 1:500; see Table S2) for 48 h at 4°C, washed with PBS-T three times for 2 h each at 4°C. Samples were then transferred into 20-30 ml of CUBIC imaging buffer [RI 1.46, Histodenz (Sigma-Aldrich, D2158) in 0.02 M PBS-0.1% Triton X-100, 0.01% sodium azide, pH 7.5] in a 50 ml Falcon tube, then imaged on a Zeiss light sheet 7 microscope.

To quantify UB branching events, the 3D data files from light sheet microscopy were imported into Imaris (Bitplane). A surface of the 3D image was first created to digitally render the UB branching and tips. Dendrites/branches were selected using the autopath algorithm (no loop) of Imaris filament module, based on the local contrast. The starting point (largest diameter) and seed point (thinnest diameter) of dendrites were set to 50 µm and 5 µm, respectively. The thresholds for the starting points and seed points were automatically determined. Next, supervised automatic tracing within the filament tracer module was performed to ensure that automatically generated paths were not duplicated or artifacts inaccurately rendered as UB branches. The total number of branch tips and nodes were estimated from the number of dendrite terminal parts in the entire z-stack. The density of UB branches was determined relative to the kidney area within selected 3D stacks (100 slices). One kidney per animal was used for quantification. All parameters of surface and filament objects were exported from Imaris for statistical analysis.

For analysis of GFP⁺ UB cells with and without centrosomes during embryogenesis, kidneys of WT and Hoxb7-Cep120 mice were isolated at E13.5 and dissociated into single cell suspension as previously described (Jain et al., 2014). Briefly, E13.5 embryonic kidneys were dissected and collected in 1.5 ml tube. Collected kidneys were suspended with Trypsin-EDTA solution (Sigma-Aldrich, T3924) with 200 µg/ml DNase I (Sigma-Aldrich) and incubated for 10 min at 37°C. After the incubation the specimens were triturated with a P-200 pipette and rinsed once with DMEM/F12 containing 10% fetal calf serum (FCS) to inactivate Trypsin. The cell preparation was then treated with collagenase dissolved in DMEM/F12 (1 mg/ml) and incubated at 37°C for 10 min. The specimens were again triturated and fully dissociated cells were rinsed twice in 500 µl of PBS/5% FCS and subjected to FACS (Sony Synergy-HAPS 1, 100 micron), performed using the GFP channel.

For analysis of gene expression changes in fibrocystic kidneys of postnatal mice, whole kidneys were isolated from WT and Hoxb7-Cep120 animals at P15. The kidneys were weighed and then ground using pestle and mortar in a liquid N₂ bath and then lysed in Trizol. Total RNA was isolated using Direct-zol^TM RNA MiniPrep Plus (Zymo Research).

For bulk RNA-seq analysis, samples were prepared according to the library kit manufacturer's protocol (SMARTer Ultra Low RNA kit for Illumina Sequencing; Takara-Clontech), indexed, pooled and sequenced on an Illumina HiSeq. Basecalls and demultiplexing were performed with Illumina's bcl2fastq software and a custom python demultiplexing program with a maximum of one mismatch in the indexing read. RNA-seq reads were then aligned to the Ensembl release 76 primary assembly with STAR version 2.5.1a. Gene counts were derived from the number of uniquely aligned unambiguous reads by Subread:featureCount version 1.4.6-p5. Isoform expression of known Ensembl transcripts were estimated with Salmon version 0.8.2. Sequencing performance was assessed for the total number of aligned reads, total number of uniquely aligned reads, and features detected. The ribosomal fraction, known junction saturation and read distribution over known gene models were quantified with RSeQC version 2.6.2.

All gene counts were then imported into the R/Bioconductor package EdgeR and TMM normalization size factors were calculated to adjust for samples for differences in library size. Ribosomal genes and genes not expressed in the smallest group size minus one sample greater than one count-per-million were excluded from further analysis. The TMM size factors and the matrix of counts were then imported into the R/Bioconductor package Limma. Weighted likelihoods based on the observed mean-variance relationship of every gene and sample were then calculated for all samples with the ‘voomWithQualityWeights’. The performance of all genes was assessed with plots of the residual standard deviation of every gene to their average log-count with a robustly fitted trend line of the residuals. Differential expression analysis was then performed to analyze for differences between conditions and the results were filtered for only those genes with Benjamini–Hochberg false-discovery rate adjusted P-values ≤0.05.

For each contrast extracted with Limma, global perturbations in known Gene Ontology (GO) terms, MSigDb and KEGG pathways were detected using the R/Bioconductor package GAGE to test for changes in expression of the reported log 2 fold-changes reported by Limma in each term versus the background log 2 fold-changes of all genes found outside the respective term. The R/Bioconductor package heatmap3 was used to display heatmaps across groups of samples for each GO or MSigDb term with a Benjamini–Hochberg false-discovery rate adjusted P-value ≤0.05. Perturbed KEGG pathways where the observed log 2 fold-changes of genes within the term were significantly perturbed in a single-direction versus background or in any direction compared with other genes within a given term with P-values ≤0.05 were rendered as annotated KEGG graphs with the R/Bioconductor package Pathview.

To find the most crucial genes, the raw counts were variance stabilized with the R/Bioconductor package DESeq2 and then analyzed via weighted gene correlation network analysis with the R/Bioconductor package WGCNA. Briefly, all genes were correlated across each other by Pearson correlations and clustered by expression similarity into unsigned modules using a power threshold empirically determined from the data. An eigengene was then created for each de novo cluster and its expression profile was then correlated across all coefficients of the model matrix. Because these clusters of genes were created by expression profile rather than known functional similarity, the clustered modules were given the names of random colors, where gray is the only module that has any pre-existing definition of containing genes that do not cluster well with others. These de-novo clustered genes were then tested for functional enrichment of known GO terms with hypergeometric tests available in the R/Bioconductor package clusterProfiler. Significant terms with Benjamini–Hochberg adjusted P-values <0.05 were then collapsed by similarity into clusterProfiler category network plots to display the most significant terms for each module of hub genes in order to interpolate the function of each significant module. The information for all clustered genes for each module was then combined with their respective statistical significance results from Limma to determine whether or not those features were also found to be significantly differentially expressed.

The DEGs identified by bulk RNA-seq at E13.5 and P15 were further analyzed using the CompBio software package. CompBio uses contextual language processing and a biological language dictionary that is not restricted to fixed pathway and ontology knowledge bases, such as KEGG or GO, to extract ‘knowledge’ from all PubMed abstracts that reference entities of interest. Conditional probability analysis calculated the statistical enrichment of biological concepts (processes/pathways) over those that occur by random sampling. Related concepts built from the list of differentially expressed entities were further clustered into higher-level themes (e.g. biological pathways/processes, cell types and structures, etc.). Themes generated by CompBio incorporate genes in each pathway or process, independent of whether they have a positive or negative function.

Kidneys of E13.5 and P15 were collected and homogenized in T-PER lysis buffer (Thermo Fisher Scientific) supplemented with 2× protease (Pierce) and 3× phosphatase inhibitors (Roche). The homogenized lysate was incubated at 4°C for 20 min, and cell debris cleared by centrifugation at 12,000 g for 10 min. For E13.5 samples, four kidneys from two embryos per genotype were pooled together and analyzed. For P15, one kidney per genotype was used for analysis. We analyzed 50 µg of protein lysate from each sample using SDS-PAGE. Membranes were incubated overnight at 4°C with the primary antibodies (Table S2); secondary antibodies include donkey anti-rabbit IgG-horseradish peroxidase (Sigma-Aldrich) and goat anti–mouse IgG-horseradish peroxidase (Jackson ImmunoResearch). Blots were imaged on a BioRad scanner.

Statistical analyses were performed using GraphPad PRISM 9.0. The box plots show median values (middle bars) and first to third interquartile ranges (boxes); whiskers indicate the maximum and minimum values for each dataset analyzed. Collected data were examined by one-way ANOVA or two-tailed unpaired t-test as specified in the figure legends. Statistical significance is shown as *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

We thank the Washington University Genome Technology Access Center (GTAC) for help with bioinformatic analyses. We also thank Dr Jeff Miner for sharing reagents used in this study. We thank members of the Mahjoub lab and the Washington University Ciliopathy Research Group for helpful feedback on this project, and critical reading of the manuscript. Finally, we thank members of the Washington University Center for Cellular Imaging (WUCCI) for assistance with light sheet microscopy and Imaris data analysis [supported in part by The Children's Discovery Institute of Washington University and St. Louis Children's Hospital (CDI-CORE-2015-505 and CDI-CORE-2019-813) and the Foundation for Barnes-Jewish Hospital (3770 and 4642).

The peer review history is available online at https://journals.biologists.com/dev/lookup/doi/10.1242/dev.201976.reviewer-comments.pdf.