Introduction

Autosomal dominant polycystic kidney disease (ADPKD) is the most frequent life-threatening genetic disease and one of the most common Mendelian human disorders with an estimated prevalence of 1/400-1000.^{1, 2} This equates to around 12.5 million affected individuals worldwide. About 5-10% of all patients requiring renal replacement therapy are affected by ADPKD. The majority of ADPKD patients carry a pathogenic germline variant in the PKD1 or PKD2 gene and present with the disease in adulthood.^2–4 However, occasionally, ADPKD can manifest in infancy or early childhood [< 2 years, very-early onset ADPKD (VEO-ADPKD)], and in late childhood or early teenage years [2-16 years, early-onset ADPKD (EO-ADPKD)].^{5, 6} VEO patients and fetuses often present with a Potter sequence and significant peri- or neonatal demise, which can be clinically indistinguishable from a typical Autosomal Recessive Polycystic Kidney Disease (ARPKD) presentation caused by PKHD1 mutations.^{7, 8} However, in contrast to VEO/EO-ADPKD, ARPKD kidneys invariably manifest as fusiform dilations of renal collecting ducts and distal tubules that usually remain in contact with the urinary system.⁴ Co-inheritance of an inactivating PKD1 or PKD2 mutation with an incompletely penetrant minor PKD allele in trans provides a likely explanation for VEO-ADPKD.⁹ In fact, we recently reported that the majority (70%) of VEO-ADPKD cases in an international diagnostic cohort had biallelic PKD1 variants (i.e., a pathogenic variant in trans with a hypomorphic, low penetrance variant), while cases of biallelic PKD2 and digenic PKD1/PKD2 were rather rare.¹⁰ In line with the dosage theory for PKD,¹¹ several other genes (GANAB, DNAJB11, ALG8, ALG9, IFT140) have been associated with a dominant, but late-onset atypical adult presentation and sometimes incomplete penetrance.^{4, 12–15} Yet, not all VEO/EO-ADPKD patients can be explained by monogenic inheritance suggesting digenic or oligogenic inheritance causes.

Previous data from mouse, Xenopus and zebrafish suggest a crucial role for the RNA-binding protein Bicc1 in the pathogenesis of PKD, although BICC1 mutations in human PKD have not been previously reported.^16–22 BICC1 encodes an evolutionarily conserved protein that is characterized by 3 K-homology (KH) and 2 KH-like (KHL) RNA-binding domains at the N-terminus and a SAM domain at the C-terminus, which are separated by a by a disordered intervening sequence (IVS).^23–28 The protein localizes to cytoplasmic foci involved in RNA metabolism and has been shown to regulate the expression of several genes such as Pkd2, Adcyd6 and Pkia in the kidney.^{22, 29} We now present data providing a mechanistic model linking BICC1 with the three major cystic proteins. We show that BICC1 physically interacts with the PKD1 (PC1) and the PKD2 (PC2) protein in human kidney cells. We also demonstrate that Pkd1 and Pkd2 modifies the cystic phenotype in Bicc1 mice in a dose-dependent manner and that Bicc1 functionally interacts with Pkd1, Pkd2 and Pkhd1 in the pronephros of Xenopus embryos. Finally, this interaction is supported by human patient data where Bicc1 alone or in conjunction with PKD1 or PKD2 is involved in VEO-PKD.

Methods

Cell Culture and Biochemical Studies

The characterization of the interaction between BICC1, PC1 and PC2 as well as the analysis of the human BICC1 variants were performed using standard approaches detailed in the Supplementary Methods.

Animal Studies

Mouse and Xenopus laevis studies were approved by the Institutional Animal Care and Use Committee at the Cleveland Clinic Foundation and LSU Health Sciences Center (present and former employer of Dr. Wessely) and adhered to the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Experimental design and data interpretation followed the ARRIVE1 reporting guidelines.³⁰

International Diagnostic Clinical Cohort

Research was performed following written informed consent and according to the declaration of Helsinki and oversight was provided by the Medizinische Genetik Mainz. DNA extraction was performed according to standard procedures (see Supplementary Methods for details).

Statistical analysis

Data are presented as mean values ± SEM. Paired and unpaired two-sided Student’s t test or ANOVA wereused for statistical analyses with a minimum of P < 0.05 indicating statistical significance. Measurements were taken from distinct biological samples. Analyses were carried out using Prism 10 (Graphpad).

Results

Interaction of BICC1 with PC1 and PC2

Loss of Pkd1 has been associated with lower expression of Bicc1.³¹ Furthermore, Bicc1 has been shown to regulate Pkd2 expression in cellular and animal models.^{22, 32, 33} However, whether this is due to direct protein-protein interactions between BICC1, PKD1 (PC1) and PKD2 protein (PC2) have not been investigated. In pilot experiments, BICC1 was detected by mass spectrometry in a pulldown assay from cells stably expressing a Polycystin-1 PLAT domain (Polycystin-1, Lipoxygenase, Alpha-Toxin)-YFP fusion.³⁴ The direct binding between the PC1-PLAT domain and Bicc1 was confirmed using in vitro binding assays, but also detected binding to the PC1 C-terminus (CT1) (Supplementary Fig. S1a,c).

Utilizing recombinant GST domains of PC1 and PC2, we demonstrated that BICC1 binds to both proteins in GST-pulldown assays (Fig. 1a, b). In the case of PC1, myc-mBICC1 strongly interacted with its C-terminus (GST-CT1) but its interaction was abolished by a PC1-R4227X truncation mutation (GST-CT1-R4227X) (Fig. 1b,c). In the case of PC2, myc-mBicc1 associated with both recombinant GST N-terminal (GST-NT2) and C-terminal (GST-CT2) fusions. To investigate whether binding was direct or indirect, we performed in vitro binding assays using in vitro translated myc-Bicc1 and recombinant PC1 and PC2 domains. GST-pulldowns confirmed a direct interaction between myc-Bicc1 and GST-CT1 but not GST-CT1-R4227X (Fig. 1d, e). Similarly, myc-Bicc1 interacted directly with GST-NT2. While binding was stronger with the distal sequence (NT2 aa101-223) both N-terminal fragments contributed to the overall binding to Bicc1 (Fig. 1d, e). Interestingly, no direct interaction between Bicc1 and GST-CT2 was detected (Supplementary Fig. S1b) suggesting that the observed in vivo interaction with Bicc1 is indirect. Finally, immunoprecipitation using lysates from human kidney epithelial cells (UCL93) to assay endogenous, non-overexpressed proteins showed that PC1, PC2 and BICC1 form protein complexes in vivo (Fig. 1f, g).

Figure 1.

Different Interaction Motifs for the Binding of Bicc1 to the Polycystins

To define the PC1/PC2 interaction domain(s) in Bicc1, we generated deletion constructs lacking the SAM domain (myc-mBicc1-ΔSAM) or the KH/KHL domains (myc-mBicc1-ΔKH) (Fig. 2a) and studied them by co-IP. Full-length PC1 co-immunoprecipitated with full-length myc-mBicc1 (Fig. 2b,c). Deleting the SAM domain reduced the association to PC1 by ∼55% compared to full-length myc-mBicc1. Surprisingly, an 8-fold stronger interaction was observed between full-length PC1 and myc-mBicc1-ΔKH compared to myc-mBicc1 or myc-mBicc1-ΔSAM. These results suggested that the interaction between PC1 and Bicc1 involves the SAM but not the KH/KHL domains (or the first 132 amino acids of Bicc1). It also suggests that the N-terminus could have an inhibitory effect on PC1-BICC1 association.

Figure 2.

Similar experiments were performed to define the Bicc1 interacting domains for PC2 (Fig. 2d,e). Full-length PC2 (PC2-HA) interacted with full-length myc-mBICC1. Unlike PC1, PC2 interacted with myc-mBICC1-ΔSAM, but not myc-mBICC1-ΔKH suggesting that PC2 binding is dependent on the N-terminal domains but not the SAM domain. Co-expression of BICC1, deletion constructs lacking the SAM domain (myc-mBicc1-ΔSAM) or the KH domains (myc-mBicc1-ΔKH) however had no effect on the interaction of PC1 with PC2 in co-immunoprecipitation assays (Fig 2f,g) suggesting that these interactions are not mutually exclusive.

Cooperativity of BICC1 with other PKD genes

Since our biochemical analysis indicated a direct interaction between BICC1, PC1 and PC2, we wondered whether this is biologically relevant. If this were the case, BICC1 should cooperate with other PKD genes and reducing BICC1 activity in conjunction with reducing either PKD1 or PKD2 activity should still cause a cystic phenotype. We first addressed this question in the Xenopus system (Fig. 3), which is an easily manipulatable model to study PKD. The PKD phenotype in frog is characterized by dilated kidney tubules, the loss of the expression of the sodium bicarbonate cotransporter 1 (Nbc1) in the in the distal tubule and the emergence of body-wide edema as a sign of a malfunctioning kidney.^{21, 22, 35, 36} Knockdown of Bicc1, Pkd1, Pkd2 or the ARPKD gene Pkhd1 caused a PKD phenotype (Fig. 3e-i” and Supplementary Fig. S2a). The latter, Pkhd1 was included to assay not only ADPKD, but also ARPKD, which is generally thought to disturb the same cellular mechanisms. To test whether Bicc1 cooperated with the PKD genes we then performed combined knockdowns. We titrated each of the four MOs to a concentration that on its own resulted in little phenotypic changes upon injection into Xenopus embryos (Fig. 3j,k and Supplementary Fig. S2b). However, combining Bicc1-MO1+2 with Pkd1-sMO, Pkd2-MO or Pkhd1-sMO at suboptimal concentrations resulted in the re-emergence of a strong PKD phenotype.

Figure 3.

While injections with individual MOs developed edema in about 10% of the embryos, co-injections caused edema formation in almost 100% of the embryos (Fig. 3j, last 3 columns). A similar result was seen for the expression of Nbc1 in the late distal tubule, where individual MO injections showed some changes in gene expression, but double MO injections had a highly synergistic effect resulting in a near complete loss of Nbc1 (Fig. 3k).

We next investigated whether a similar cooperation between Bicc1 and Pkd1/Pkd2 can be observed in a genetic model. We initially focused on Bicc1 and Pkd2. Both Bicc1 and Pkd2 knockout mice develop cystic kidneys as early as E15.5.^{22, 37} As this is the earliest time point cystic kidneys can be observed, crossing those strains did not allow us to assess cooperativity (data not shown). Moreover, like in the case of compound Pkd1/Pkd2 mutants,³⁸ kidneys from Bicc1^+/-:Pkd2^+/- did not exhibit cysts (data not shown). Thus, we instead used mice carrying the Bicc1 hypomorphic allele Bpk, which develop a cystic kidney phenotype postnatally.^{18, 39} To assess cooperativity, we removed one copy of Pkd2 in the Bpk mice. Comparing the kidneys of Bicc1^Bpk/Bpk:Pkd2^+/-to those of Bicc1^Bpk/Bpk:Pkd2^+/+at postnatal day P14 revealed that the compound mutant kidneys were larger and more translucent (Fig. 4a) and the kidney/body weight ratios (KW/BW) were significantly increased (Fig. 4b). Moreover, analyzing survival the compound mutants showed a trend towards an earlier demise (Supplementary Table S1). We did not detect sex differences in the phenotype (Supplementary Figure S3c). Yet, the reduction in Pkd2 gene dose affected the progression of the disease, but not its onset. Performing the same analysis at postnatal day P4 did not show any differences (Fig. 4c).

Figure 4:

Next, we performed a similar mouse study for Pkd1 by reducing the gene dose of Pkd1 postnatally in the collecting ducts using a Pkhd1-Cre as previously described⁴⁰ (in the following referred to as Pkd1^CD-). Similar to the Bicc1/Pkd2 scenario, Bicc1^Bpk/Bpk:Pkd1^+/CD-appeared larger when comparing kidneys from littermates (Fig. 4d) and littermates exhibited statistically significant differences in KW/BW ratio (Fig. 4e). Yet, the phenotype was rather subtle and aggregating all the data did not show differences in KW/BW ratios between Bicc1^Bpk/Bpk:Pkd1^+/+and Bicc1^Bpk/Bpk:Pkd1^+/CD-mice (Supplementary Fig. S3d). Thus, to further corroborate the genetic interaction, we determined the cystic index for proximal tubules and collecting ducts using LTA and DBA staining, respectively. This showed an increase in collecting duct cysts upon removal of one copy of Pkd1 (Fig. 4g). Like in the case of Pkd2, the phenotype seems to be correlated with cyst expansion and not the onset, as there was no difference at postnatal day P7 (Fig. 4f) and we did not detect increased mortality in the compound mutants (Supplementary Table S2). It is noteworthy that neither the Bicc1/Pkd2 nor the Bicc1/Pkd1 compound mutants showed an aggravated kidney function based on Blood Urea Nitrogen (BUN) levels (Supplementary Fig. S3a,b,e) likely due to the aggressive nature of the Bicc1^Bpk/Bpkphenotype. Together these data supported our biochemical interaction data and demonstrated that BICC1 cooperated with PKD1 and PKD2.

Finally, to better understand how Bicc1 would exert such a phenotype, we analyzed the expression of the PKD genes in the Bicc1^Bpk/Bpk mice. We have previously demonstrated that Pkd2 levels are reduced in a complete Bicc1 null mice,²² performing qRT-PCR of P4 kidneys (i.e. before the onset of a strong cystic phenotype), revealed that Bicc1, Pkd1 and Pkd2 were statistically significantly down-regulated (Fig. 4h-j). The effect on Pkd2 was confirmed by protein analysis for PC2 (Fig. 4k and Supplementary Fig. S3f).

BICC1 Variants in Patients with early and severe Polycystic Kidney Disease

To evaluate whether these interactions are relevant for human PKD, we analyzed an international cohort of 2,914 PKD patients by massive parallel sequencing (MPS) ^{41, 42} focusing on VEO-ADPKD patients with the hypothesis that BICC1 variants may lead to a more severe and earlier PKD phenotype. While variants in BICC1 are very rare, we could identify two patients with BICC1 variants harboring an additional PKD2 or PKD1 variant in trans, respectively. None of these BICC1 variants were detected in the control cohort or in non-VEO-ADPKD patients. Moreover, besides the variants reported below, the patients had no other variants in any of other PKD genes or genes which phenocopy PKD including PKD1, PKD2, PKHD1, HNF1ß, GANAB, IFT140, DZIP1L, CYS1, DNAJB11, ALG5, ALG8, ALG9, LRP5, NEK8, OFD1, or PMM2. The first patient was severely and prenatally affected demonstrating a Potter sequence with huge echogenic kidneys and oligo-/anhydramnios. Autopsy confirmed VEO-ADPKD with absence of ductal plate malformation invariably seen in ARPKD. The fetus carried the BICC1 variant (c.2462G>A, p.Gly821Glu) inherited from his father, who presented with two small renal cysts in one of his kidneys, and a PKD2 variant (c.1894T>C, p.Cys632Arg) that arose de novo (Fig. 5a). Individual in silico predictions (SIFT, Polyphen2, CADD, Eigen-PC, FATHMM, GERP++ RS and EVE), meta scores (REVEL, MetaSVM, and MetaLR) and other protein function predictions (PrimateAI, Alphamissense, ESM1b and ProtVar) indicate that this PKD2 missense variant is likely pathogenic (Supplementary Table S3). Moreover, structural analysis suggests that the hydrophilic substitution may interfere with the Helix S5 pore domain of PKD2 and change its ion channel function (Fig. 5b,c). Finally, PKD2 p.Cys632Arg has been previously reported as part of a PKD2 pedigree and implicated as a critical determinant for Polycystin-2 function.^{43, 44} On the other hand, the BICC1 p.Gly821Glu variant is located in an intrinsically disordered domain of BICC1 between the KH and the SAM domains (Fig. 6f). To address whether the variant is hypomorphic, we used CRISPR-Cas9-mediated gene editing to generate HEK293T cells lacking BICC1 or harboring the G821E mutation. These cells were analyzed for their impact on the translation of PKD2, a well-established target of Bicc1.²² As shown in Fig. 5d,e PC2 protein levels were strongly reduced in two independent HEK293T BICC1 (G821E) cells when compared to unedited HEK293T cells. Most notably, the PC2 levels were comparable to the levels found in HEK293T carrying a BICC1 null allele (HEK293T BICC1 KO) (Supplementary Figure S2c,d). Based on these data we hypothesize that the major disease effect results from the pathogenic PKD2 variant but is aggravated by the BICC1 variant.

Figure 5.

Figure 6:

The second patient presented perinatally with massively enlarged hyperechogenic kidneys, while the parents, both in their thirties, and the remaining family members were reported to be healthy (Fig. 5f-h). He carried a paternal canonic BICC1 pathogenic splicing variant (c.1179+1G>T), which is likely pathogenic as the protein is truncated after exon 10, and a novel heterozygous PKD1 variant (c.11942C>T, p.Ala3981Val) which has not been previously reported (Fig. 5f). While the Polycystin-1 variant appears minor in its amino acid change (i.e. Ala to Val), in silico analyses using individual predictions (SIFT, Polyphen2, CADD and EVE), Meta scores (REVEL) and other protein function predictions (PrimateAI, Alphamissense and ESM1b) indicate that missense variant is likely pathogenic (Supplementary Table S3). Structural analyses suggest that the Ala3981Val variant may not destabilize the Helix structure but may contact the TOP domain and interfere with domain flexibility and PC1 complex assembly.

A Sibling Pair of PKD Patients with a Homozygous BICC1 Mutation

The most insightful finding for a critical role for BICC1 in human PKD was the discovery of a homozygous BICC1 variant in a consanguinous Arab multiplex pedigree, two siblings, a boy and a girl, diagnosed with VEO-ADPKD (Fig. 6a-e). The affected female presented at a few months of age with kidney failure and enlarged polycystic kidneys that lacked corticomedullary differentiation. Histology after bilateral nephrectomy showed polycystic kidneys more suggestive of ADPKD than ARPKD without any dysplastic element (Fig. 6c). Her younger brother exhibited enlarged hyperechogenic polycystic kidneys antenatally by ultrasound (Fig. 6b). In addition, during early infancy, arterial hypertension and a Dandy-Walker malformation with a low-pressure communicating hydrocephalus were noted (Fig. 6d,e). By customized MPS, we identified the homozygous evolutionarily well-conserved missense variant (c.718T>C, p.Ser240Pro) in BICC1 (Fig. 6a). This variant was absent from gnomAD and perfectly segregated with the cystic phenotype present in this family. It results in a non-conservative change from the aliphatic, polar-hydrophilic serine to the cyclic, apolar-hydrophobic proline located in the second beta sheet of the first KHL1 domain and very likely disrupts the beta sheet and thus the RNA-binding activity of Bicc1 (Fig. 6f,g and Supplementary Table S4). In the more severely affected younger brother, we also detected the heterozygous PKD2 change (c.1445T>G, p.Phe482Cys), which results in a non-conservative change from phenylalanine to cysteine (Supplementary Table S3). It was previously reported that this Phe482Cys variant exhibited altered kinetic PC2 channel properties, increased expression in IMCD cells and a different subcellular distribution when compared to wild-type PC2,⁴⁵ These features suggested altered properties of this PC2 variant, yet its contribution to the case reported here remain untested.

Unfortunately, both siblings passed away and besides DNA and the phenotypic analysis described above neither human tissue nor primary patient-derived cells could be collected. Thus, to validate the pathogenicity of this point mutation, we turned to the amphibian model of PKD.^{21, 22} In Xenopus, knockdown of Bicc1 using antisense morpholino oligomers (Bicc1-MO1+2) causes a PKD phenotype, which can be rescued by co-injection of synthetic mRNA encoding Bicc1.²¹ To test whether BICC1 p.(Ser240Pro) has lost its biological activity, we introduced the same mutation into the Xenopus gene (BicC1* S236P). Xenopus embryos were injected with Bicc1-MO1+2 at the 2-4 cell stage followed by a single injection of 2 ng wild type or Bicc1* S236P mRNAs at the 8-cell stage. At stage 39 (when kidney development has been completed) embryos were analyzed by whole mount in situ hybridization for the expression of Nbc1 in the late distal tubule of the pronephric kidney, one of the most reliable readouts for the amphibian PKD phenotype.²¹ As shown in Fig. 6h, wild type Bicc1 mRNA restored expression of Nbc1 on the injected side in 63% of the embryos. However, Bicc1* S236P did not have any effect, and the embryos were indistinguishable from those injected with the Bicc1-MO1+2 alone. This suggested that Bicc1* S236P was functionally impaired. To address this hypothesis, we first assessed the subcellular localization of Bicc1 to foci that are thought to be involved in mRNA processing.^{19, 22, 27, 46} Transfection of Flag-tagged Bicc1 into HEK293T cells reproduced this pattern (Fig. 6i). Surprisingly, Bicc1 S236P-Flag was no longer detected in these cytoplasmic foci but rather homogenously dispersed throughout the cytoplasm (Fig. 6j). Western blot analysis demonstrated that this was accompanied by a reduction in protein levels (Fig. 6k). In vitro transcription/translation detected no differences between the proteins suggesting that the wildtype and Bicc1 S236P-Flag are translated equivalently (data not shown). Yet, in an in vivo pulse-chase experiments the S240P variant was less stable than its wildtype counterpart (Fig. 6k). However, whether the reduced protein level was due to an inherent instability of the mutant protein or a consequence of its mislocalization remains to be resolved. Finally, as in the case of BICC1 G821E we engineered HEK293T cells to harbor the BICC1 S240P mutation. Western blot analysis demonstrated a reduction in PC2 levels in the BICC1 S240P cells when compared to unedited cells and that this reduction was comparable to PC2 levels in BICC1 KO cells (Figs. 5d,e and 6l).

Finally, to determine to what extent the BICC1 S240P variant differs from a BICC1 loss of function, we performed mRNA sequencing (mRNA-seq) of the genetically engineered HEK293T cells. Differential gene expression analysis identified several genes that were differentially up-or down-regulated in the BICC1 S240P and the BICC1 KO cells compared to their unedited counterpart (Supplementary Fig. 4a,e). Approximately 24% and 18% of the differentially expressed genes were shared between BICC1 S240P or the KO cells, respectively (Supplementary Fig. S4b,f). Yet, a substantial number of genes were specific to either cell line. The BICC1 S240P-enriched/depleted transcripts were generally also enriched/depleted in the BICC1 KO cells, but did not reach statistical significance (Supplementary Fig. 4c,g). Conversely, many of the BICC1 KO enriched transcripts were specifically enriched/depleted in the BICC1 KO cells and not in the BICC1 S240P cells (Fig. 6m and Supplementary Fig. 4d). This suggested that there are qualitative differences between a null phenotype and the BICC1 S240P variant, supporting our hypothesis that BICC1 S240P acts as a hypomorph. Indeed, Gene Set Enrichment Analysis (GSEA) using the hallmark gene sets and comparing BICC1 KO and S240P cells revealed a statistically significant enrichment for the Hallmark_Epithelial_Mesenchymal_Transition set (Fig. 6n and Supplementary Table S5), a pathway previously implicated in ADPKD.^{47, 48}

Discussion

BICC1 has been extensively studied in multiple animal models, which have suggested a critical role for BICC1 in several different developmental processes and in tissue homeostasis.²³ This study functionally implicates it to human disease in general and PKD in particular by identifying the homozygous BICC1 S240P variant, which was sufficient to cause a cystic phenotype in a sibling pair of human PKD patients. It is noteworthy that another study identified heterozygous BICC1 mutations in two patients with mildly cystic dysplastic kidneys.⁴⁹ Yet, the variants were both also present in one of the parent. While such a situation is extremely rare and does not significantly contribute to the mutational load in ADPKD or ARPKD, it demonstrated that loss of BICC1 is sufficient to cause PKD in humans. In addition, variants in BICC1 and PKD1 and PKD2 co-segregated in PKD patients from an International Clinical Diagnostic Cohort. While we have not yet shown the pathogenesis of the variants when introduced as a compound situation, we postulate that PKD alleles in trans and/or de novo exert an aggravating effect and contribute to polycystic kidney disease. A reduced dosage of PKD proteins would severely disturb the homeostasis and network integrity, and by this correlates with disease severity in PKD. ADPKD is quite heterogeneous and - even within the same family - shows quite some phenotypic variation.^{50, 51} It is thought that stochastic inputs, environmental factors and genetics influence PKD.⁵¹ The demonstrated interaction of BICC1, PC1 and PC2 now provides a molecular mechanism that can explain some of the phenotypic variability in these families. Of note, while our mouse studies support the cooperation between Bicc1, Pkd1 and Pkd2, the genetic proof for Bicc1 acting as a disease modifier, i.e. reduction of Bicc1 activity in a heterozygous Pkd1 or Pkd2 background is still outstanding.

The second important aspect of the study is that BICC1 emerges as a central in the regulation of PKD1/PKD2 activity. Functional studies reported here and previously^{22, 32, 33} demonstrate that BICC1 regulates the expression of PKD1 and PKD2. Moreover, we now show that BICC1 and PC1/PC2 physically interact and that lowering the expression levels of both proteins was sufficient to cause a PKD phenotype in frogs. Finally, in mice the reduction of the gene dose for Pkd1 or Pkd2 in a hypomorphic allele of Bicc1 results in a more severe cystic kidney phenotype. These results in the kidney are paralleled and augmented in studies of left/right patterning, where PC2 can activate Bicc1 and where Bicc1 triggers critical aspects in establishing laterality.^{19, 27, 52, 53} Thus, it is tempting to speculate that BICC1/PKD1/PKD2 are components of a critical regulatory network in maintaining epithelial homeostasis. BICC1 functions as a posttranscriptional regulator modifying gene expression modulating the effects of microRNAs (miRNAs), regulating mRNA polyadenylation and translational repression and activation.^{22, 23, 29, 54–57}. While PKD2 is the most appealing target in respect to ADPKD, there are undoubtable others (e.g. adenylate cyclase-6)²⁹ that are equally critical. In the future, it will be interesting to see how all these targets are integrated and together contribute towards protecting kidney epithelial cells from undergoing cystic malformations. Lastly, BICC1 has been implicated in the regulation of miRNAs such as the ones of the miR-17 family.²² A connection between miR-17 activity and PKD is well-established ^58–63 and is even considered as a PKD therapeutic.⁶⁴ Thus, the interaction between BICC1, PKD1 and PKD2 and miRNAs - even though not examined in this study - may be the ultimate regulatory complex that is responsible for several of the aspects of human ADPKD.

Data sharing statement

The datasets are presented in the figures and the supplementary information. The mRNA-seq data are deposited into the Gene Expression Omnibus (GEO) database (GSE262417) and are available online. Additional information is available from the corresponding author on reasonable request.

Acknowledgements

The authors would like to thank the patients and their families for their cooperation and interest in the study. This work was supported by grants from NIH/NIDDK (R01DK080745) and a philanthropic gift for PKD research at CCF to OW, Kidney Research UK and the PKD Charity UK (PKD_RP_005_20211124), the Sheffield Hospitals Charity and the Sheffield Kidney Research Foundation to AJS and ACMO, the Deutsche Forschungsgemeinschaft (DFG, BE 3910/8-2, BE 3910/9-1, Project-ID 431984000 - Collaborative Research Center SFB 1453), the Federal Ministry of Education and Research (BMBF, 01GM1903I and 01GM1903G) and the European Union’s Horizon Europe research and innovation programme (grant agreement 101080717, TheRaCil) to CB. DS was supported by a Faculty PhD Scholarship from the University of Sheffield. We thank Drs. S. Somlo, P. Igarashi, and K. Dell for mouse strains, S. Feng, and L. Chang for technical assistance and R. Allen Schweickart for bioinformatical support.

Additional files

Supplemental Material