|Home | About | Journals | Submit | Contact Us | Français|
Bicaudal-C (Bic-C) gene was originally discovered in Drosophila melanogaster. The gene product Bic-C is thought to serve as an RNA-binding molecule targeting diverse proteins at the post-transcriptional level. Recent research has shown this gene to be conserved in many species, from Caenorhabditis elegans to humans. Disruption of this protein can disturb the normal migration direction of the anterior follicle cell of Drosophila oocytes, while mutation of a mouse Bicc1 (a mouse homologue of Bic-C) results in phenotypes mimicking human hereditary polycystic kidney disease (PKD). However, the cellular function of Bicc1 gene products in mammalian system remains largely unknown. In this study, we established stable IMCD (mouse inner medullary collecting duct) cell lines, in which Bicc1 was silenced by short hairpin RNA inhibition (shRNA). We show that inhibition of Bicc1 disrupted normal tubulomorphogenesis and induced cystogenesis of IMCD cells grown in three dimensional cultures. To determine what factors contributed to the defect, we systematically examined biological changes of Bicc1-silenced IMCD cells. We found that the cells had significant defects in E-cadherin-based cell-cell adhesion, along with abnormalities in actin cytoskeleton organization, cell-extracellular matrix interactions, cell proliferation, and apoptosis. These findings suggest that lack of Bicc1 leads to disruption of normal cell-cell junctions, which in turn impedes establishment of epithelial polarity. These cellular defects may initiate abnormal tubulomorphogenesis and cystogenesis of IMCD cells grown in vitro. The observation of aberrant cellular behaviors in Bicc1-silenced IMCD cells reveal functions for Bicc1 in renal epithelial cells and provides insight into a potential pathogenic mechanism of polycystic kidney disease.
Inherited polycystic kidney disease (PKD) is a group of monogenic disorders that result in renal cyst development (Harris and Torres, 2008; Igarashi and Somlo, 2007). Based on different inherited traits, the two major types of PKD are classified as autosomal recessive polycystic kidney disease (ARPKD) or autosomal dominant polycystic kidney disease (ADPKD).
ARPKD is a common childhood renal disease with an incidence of 1:20,000. Only one causal gene PKHD1 has currently been identified. PKHD1 lies on chromosome 6p21.1−p12 and produces alternative spliced transcripts. The longest transcript encodes a large 4,074-amino-acid type I transmembrane protein, designated as fibrocystin or polyductin (FPC) (Onuchic et al., 2002; Ward et al., 2002; Xiong H, 2002).
ADPKD usually appears in adulthood with an incidence of 1:400-1000, and is the third most common single cause of end-stage renal failure worldwide. Approximately 85% of ADPKD patients have mutations in PKD1, which is located on chromosome 16p13 and encodes a large membrane-associated receptor-like protein, named polycystin-1 (PC1). PC1 is predicted to have 11 transmembrane domains, as well as a short intracellular cytoplasmic tail and a large extracellular domain (>3,000 aa) (Torres and Harris, 2006; Wilson P, 2004). The remaining 15% of ADPKD patients have mutations in PKD2, which maps to chromosome 4q21 and encodes a membrane-associated protein designated polycystin-2 (PC2) (Mochizuki et al., 1996). With six transmembrane domains and an intracellular COOH- and NH2-termini, PC2 is thought to be a receptor-operated, non-selective cation channel (Koulen P, 2002; Nauli SM, 2003). As PC2 has a modest degree of amino acids similarity to the transient receptor potential (trp), it is considered a member of the trp superfamily, and is often referred to TRPP2 (Montell C, 2002; Qamar et al., 2007). There are a few families that have been reported with ADPKD unlinked to either the PKD1 or PKD2 locus, suggesting there may be other ADPKD causal genes (Rossetti et al., 2007).
By studying various animal mutant models as well as human patient populations with PKD, more than 20 genes have been identified that can induce PKD phenotypes (Torres and Harris, 2006; Wilson PD, 2007). Therefore, PKD phenotypes may not only be caused by mutations in PKD1/2 and PKHD1, but may also result from dysfunction of the other PKD-casual genes. Further studies directed toward these genes may open new avenues in understanding the pathogenesis of human PKD. In the current study, we focus on the mouse PKD-casual gene Bicc1. Bicc1 is a mouse homologue of Drosophila Bicaudal-C (dBic-C) and its ortholog can be found in many other species (from C. elegans to humans) (Zhou L, 2008). Loss of dBic-C in Drosophila disrupts direction of anterior follicle cell migration and affects anterior-posterior patterning, producing abnormal embryos lacking head formation and having duplicate posterior segments (Mahone et al., 1995; Wessely et al., 2001). The Xenopus homologue of Bicaudal-C (xBic-C) is one of the few molecules that, when microinjected ectopically, results in endoderm formation in the absence of mesoderm induction (Wessely and De Robertis, 2000).
Bic-C, the Bicaudal-C gene product, contains several conserved N-terminal KH domains and a conserved C-terminal sterile-alpha motif (SAM) domain (Chen et al., 1997; Ponting CP, 1995). The KH domains have been shown to bind RNA (Bouvrette et al., 2008), while the SAM domain is predicted to have a high content of α-helices and play a role in protein-protein or protein-RNA interactions (Aviv T, 2003; Schultz et al., 1997; Stagner et al., 2009). There is a very conserved tyrosine residue at the 19th position of the SAM domian which is thought to have phosphorylation activity and may contribute to the protein-protein or protein-RNA interactions (Schultz J, 1997). A recent study demonstrated that Bicc1, the mouse Bic-C gene product, interacts with SamCystin, a protein whose mutation can cause PKD in Han:SPRD-cy rats. Stagner et al (Stagner et al., 2009) showed that Bicc1/SamCystin can form a complex under the mediation of specific mRNAs and suggested the complex may play an important role in regulation of normal epithelial behaviors. Thus, the Bicc1 is considered to be an RNA-binding molecule and believed to function in regualtion of mRNA post-transcription (Tran et al., 2007; Wessely and De Robertis, 2000).
The mouse Bic-C gene (Bicc1) was localized to the C1 region of Chromosome 10, and found to have two alternatively spliced transcripts: A and B (Cogswell et al., 2003). Transcript A encodes a 977 amino acid protein composed of Bicc1 exons 1–22, but without exon 21; while transcript B is composed of exons 1–21. There is a stop codon in exon 21, and the result is a shorter, 951 amino acids isoform. Whole-mount in situ hybridization indicates that the expression of Bicc1 mRNA can be first detected at the rostral tip of the primitive streak and Hensen's node at the late streak stage. At embryonic day (E) 6-7, Bicc1 expression specifically demarcates the layer of the node from which definitive endoderm and midline mesoderm arises. At day E8, Bicc1 is observed in the definitive endoderm, and strong expression can be detected in the caudal intestinal portal. At day E9, Bicc1 is still present in the hindgut, but transient expression can also be seen in tissues of neural and mesodermal origins (Maisonneuve et al., 2009; Wessely et al., 2001).
Two naturally occuring Bicc1-mutant mouse models (jcpk and bpk) have been reported (Cogswell et al., 2003). Although the underlying mutations of Bicc1 in these models are different, both models exhibit cystic phenotypes in the kidney which are similar to human polycystic kidney disease. The juvenile congenital polycystic kidney (jcpk) mutation was initiated with chlorambucil (CHL) mutagenesis, resulting in a point mutation (AG to AA) at the acceptor site of Bicc1 exon 3. This mutation causes abnormal splicing such that exon 3 is passed over for the acceptor site of exon 4, thus eliminating exon 3 (Cogswell et al., 2003). Mice homozygous for jcpk express cysts in all segments of the nephron. Death usually occurs before 10 days of age. Extrarenal involvement was also noted, with enlarged bile ducts, pancreatic ducts, and gall bladder often accompanying the PKD phenotype. In addition, approximately 25% of aged jcpk heterozygotes show evidence of glomerulocystic disease (Flaherty et al., 1995).
In BALB/c polycystic kidneys (bpk) mice, the mutation was spontaneously induced by BALB/c inbreeding. The bpk mutation is induced by a GC insertion at exon 22 of Bicc1, which can be fully penetrated with a recessive trait. Since bpk mutation only alters Bicc1 Transcript B, the cystic phenotype is milder than that observed with the jcpk mice. Homozygotes of bpk develop both dilatations of the renal collecting ducts as well as biliary dysgenesis. Death ensues within 4 weeks of birth, presumably due to renal insufficiency (Nauta et al., 1995).
Although the mutant mouse models provide some knowledge on the functional role of Bicc1 in mammalian development, the underlying cellular functions still remain largely unknown. To this end, we established stable Bicc1-silenced IMCD cell lines and characterized how loss of Bicc1 altered cellular functions. Our results indicate that the expression of normal Bicc1 protein is required to sustain renal tubulogenesis in 3-D cultures of IMCD cells. Lack of this protein resulted in obvious abnormalities in E-cadherin-based cell-cell contact, which is essential for epithelial polarization and regulation of tubulomorphagenesis. Our in vitro findings illustrate the critical importance for Bicc1 in renal epithelial differentiation and renal tubule development and may ultimately help clarify Bicc1 functional roles in PKD in vivo.
To test whether Bicc1 plays a role in normal tubulomorphogenesis, we established a set of stable mouse Bicc1-silenced IMCD (inner medullary collecting duct) cell lines. We then used these cell lines to perform 3-D cultures, a technique that has been well-validated for analyzing tubulomorphogenesis in vitro (Chen D, 2004; Mai et al., 2005). At the beginning, we utilized Invitrogen's siRNA program to select a siRNA duplex based on the mouse Bicc1 cDNA sequence, which can inhibit Bicc1 mRNA in both mice and humans. The duplex was then inserted into a pRS-shRNA vector, and the construct designated Bicc1shRNA-C (Fig. 1A). A commercial pRS-shGFP vector TR30003, which contains a GFP-siRNA duplex, was used as a negative control to discern any potential interferon response. We transfected IMCD cells with the Bicc1shRNA-C construct as well as the control shRNA vector (TR30003) to produce stable cell pools designated IMCDshC and IMCDshGFP. Twelve single puromycin-resistant clones from each IMCDshC and IMCDshGFP cell pool were isolated respectively. Quantitative PCR and western blot analysis was performed to select IMCDshC cell clones in which Bicc1 was reduced from the Bicc1shRNA-C transfected cell pool. Compared to the control cell line, IMCDshGFP1, and the wildtype IMCD cell line, IMCDWT, the IMCDshC1C and IMCDshC4C cell lines (derived from IMCDshC transfected pool) had a 4-fold decrease in Bicc1 mRNA levels by quantitative PCR (*P<0.05) (Fig. 1B). Given that the control IMCDshGFP1 cell line did not demonstrate any significant inhibition on Bicc1 mRNA levels, while both Bicc1-silenced constructs (IMCDshC1C and IMCDshC4C) resulted in strong inhibition, suggests that the Bicc1shRNA-C has a specific Bicc1 silencing activity.
To further confirm this finding, cell lines from the IMCDshC transfected pool were used for western analysis with anti-Bicc1 polyclonal antibody (mBic-Np). This antibody was previously generated by our group and shown to be specific for Bicc1 (Supplemental Fig. 1) (Dai B, 2008). Western blots showed that the stable Bicc1-silenced cell lines IMCDshC1C and IMCDshC4C had the greatest downregulation of Bicc1 expression (*P<0.05) (Fig. 1C-D). The western results further confirm our quantitative-PCR data and provide unequivocal evidence that Bicc1 is downregulated via the shRNA approach at both RNA and protein levels in the IMCDshC1C and IMCDshC4C cell lines.
To enhance our understanding of phenotypic changes in the Bicc1-silenced IMCD cells, we used a similar approach to generate another Bicc1-silenced construct, Bicc1shRNA-N, in which the duplex target differs from Bicc1shRNA-C (Supplemental Fig 2A). Two Bicc1-silenced cell lines, IMCDshN3G and IMCDshN8E, were selected from the IMCDshN cell pool using the same protocol as that described above for the Bicc1shRNA-C construct. Quantitative-PCR and western analysis indicate that both IMCDshN3G and IMCDshN8E cell lines exhibit significantly downregulated Bicc1 mRNA and protein levels (Supplemental Fig 2B-C).
To determine if lack of Bicc1 alters normal tubulomorphogenesis, we used a previously established 3-D culture assay (Mai et al., 2005). In wildtype (IMCDWT) and control (IMCDshGFP1) cell lines, over 50% of the cells form 3 and more tubular branches (Fig. 2A, 2E) and approximately 40% demonstrated 1-3 tubular branches (Fig. 2B, 2E). These results are sharply in contrast to what is found with IMCDshC1C and IMCDshC4C cell lines in which no cells were found with 3 and more tubular branches (Fig. 2E). In addition, both IMCDshC1C and IMCDshC4C cells exhibited cyst formation (Fig. 2C) and cell aggregation (Fig. 2D). No tubulogenesis was found in 95% of IMCDshC1C and IMCDshC4C in 3-D cultures, while this phenomenon occurred in fewer then 5% of wildtype (IMCDWT) or control (IMCDshGFP1) cells (Fig. 2E). Similar results can be observed in the other two Bicc1-silenced cell lines, IMCDshN3G and IMCDshN8E (Supplemental Fig. 2D). These findings demonstrate that reduction of Bicc1 gene product Bicc1 prevents tubule formation and arrests normal branching morphogenesis in 3-D cultures.
The establishment of intercellular junctions and normal cytoskeletal assembly is essential for epithelial polarity and tubule formation (Higashiyama S, 1995; Matter K, 2003; Zegers MM, 2003). We therefore assessed the effects of Bicc1-silencing on these cell biological processes. To this end, the subcellular localization and distribution of E-cadherin and ZO-1 were compared between wildtype, control, and Bicc1-silenced IMCD cells utilizing immunofluorescence staining. In the wildtype and control cell lines, E-cadherin was predominantly seen at the cell-cell junctions (Fig. 3Aa-b); while in the Bicc1-silenced cells junctional staining of E-cadherin was nearly indistinguishable from cytosolic (Fig. 3Ac-d). The data from this panel of cell lines implies that silencing of Bicc1 alters the distribution of E-cadherin and thus, may impair the formation of adherent junctions (further evidence of impaired adherent junctions is provided in Supplement Fig. 3).
To determine whether there were also abnormalities at the tight junctions, Bicc1-silenced cells were stained with a tight-tight junction marker, the ZO-1 antibody. Although ZO-1 staining was observed on the cell-cell junctions for wildtype, control, and Bicc1-silenced IMCD cells, a more diffuse and discontinuous pattern of junctional staining was seen in the Bicc1-silenced cells (Fig. 3Ba-b vs c-d). Because the Bicc1-silenced IMCD cells showed abnormalities in both adherent and tight junctions, we conclude that a lack of Bicc1 protein disrupts normal cell-cell contact.
Given the IF staining showed different distribution patterns for E-cadherin and ZO-1 in wildtype and control cells versus the Bicc1-silenced IMCD cells, we performed western blot analysis to determine whether there were variations in E-cadherin or ZO-1 expression levels. We found there was no detectable immunoreactive change in either protein among the cell lines (Fig. 3C), suggesting that lack of Bicc1 alters the cellular distribution of these proteins rather than their expression level.
To further evaluate alterations in cell-cell interactions, we measured cell transepithelial resistance (TER) to determine if cell-cell contact integrity is impaired. Our results indicate that in contrast to the wildtype and control cell lines, Bicc1-silenced cells demonstrated a low TER value after 3 days of transwell culture (*P<0.05) and this persisted out to day 6 (Fig. 3D). These results support that Bicc1-silenced cells lose their ability to establish normal cell-cell contacts.
Since the cell-cell interactions in Bicc1-silenced IMCD cells appeared to be abnormal (Fig. 3), we hypothesized that downregulation of Bicc1 might also alter the cytoskeleton. We performed rhodamine-phalloidin staining to label cellular cytoplasmic actin. We found that wildtype and control cells exhibit nominal cortical actin distribution and epithelial shape with fine and even stress fibers in sub-confluent cultures (Fig. 4Aa-b). Silenced-Bicc1 cells lose their normal cortical actin distribution and exhibit an irregular shape with thick and enriched stress fibers (Fig. 4Ac-d). In dense cultures, F-actin distribution of wildtype and control cells shows peripheral bands at cell-cell borders (Fig. 4Ba-b). In contrast, silenced-Bicc1 cells often display disorganized prominent stress fibers and lose the cell-cell borders, appearing to exhibit epithelial to mesenchymal transformation (EMT) (Fig. 4Bc-d). These findings indicate that lack of Bicc1 may disrupt normal cytoskeletal organization, and the changes may ultimately induce abnormal epithelial morphogenesis.
Cell-extracellular matrix (ECM) interactions play a critical role in branching morphogenesis, and perturbations of these interactions in IMCD cells result in decreased branching morphogenesis (Balda MS, 2003; Chen D, 2004). For this reason we investigated integrin-dependent adhesion to collagen I (CI). Cells in which Bicc1 is silenced showed significantly less adhesion (*P<0.05) to wells coated with 2 μg/ml collagen I then wildtype or control cell lines. We found IMCDshC1C and IMCDshC4C cell lines exhibited 30-50% cell adhesion compared to over 60% for IMCDWT and IMCDshGFP1 cell lines (Fig. 5A). To investigate the importance of Bicc1 in cell migration, we performed transwell migration assays. As shown in figure 5B, significantly fewer Bicc1-silenced cells migrated than was observed for the control groups (*P<0.05). These results indicat that inhibition of Bicc1 expression impedes normal cell migration and disrupts integrin-mediated cell-extracellular matrix interactions.
It has been reported that cyst formation closely associates with aberrant proliferation of renal epithelial cells (Wilson PD, 2007). We therefore performed proliferation assays for the cell lines we generated. Surprisingly, we found that Bicc1-silencing resulted in a significant decrease in tritiated thymidine uptake (*P<0.05), suggesting that downregulation of Bicc1 inhibits cell proliferation (Fig. 6A). To further validate this finding, we stained the tested cell lines with Phospho-Histone H3 antibody by which an index for cell mitotic and meiotic activity can be indicated. A similar result to tritiated thymidine uptake assay was obtained, and compared to wildtype and control cells, there is a significant reduction in the percentage of Phospho-Histone H3 postitive cells in Bicc1-silenced cells (Fig. 6B) (*P<0.05). Taken together these assays indicate that a downregulation in Bicc1 leads to a suppression in cell proliferation.
Given that programmed cell death is associated with tubulomorphogenesis (Zegers MM, 2003) and polycystic kidney disease (Wegierski T, 2009; Woo, 1995), we assessed the number of cells undergoing apoptosis using TUNEL assays. Only 5-10% of the wild-type and control cells were found to be apoptotic, while more than 25% of the Bicc1-silenced IMCD cells underwent programmed cell death (*P<0.05) (Fig. 6C). To confirm the results of the TUNEL assay, we employed an assay for active caspase-3 as an indictor to detect programmed cell death in our panel of cell lines (Fig. 6D). In agreement with the results from the TUNEL assay, the percentage of apoptotic cells seen with the Bicc1-silenced IMCD cell lines was significantly higher than with the wildtype or control cell lines (*P<0.05). Our results suggest that inhibition of Bicc1, either directly or indirectly, promotes programmed cell death during tubulomorphogenesis in vitro.
Bicc1 is a mouse homologue of Drosophila Bicaudal-C (dBic-C). Orthologs of dBic-C have been identified in many species, from C. elegans to human (Zhou L, 2008). Studies on dBic-C demostrated that the gene product may play cellular functions to monitor direction of cell migration and affects developmental patterning. Disruption of dBic-C resulted in abnormal embryos lacking head formation and having duplicate posterior segments (Mahone et al., 1995; Wessely et al., 2001). A recent study reported that dBic-C interacts with Trailar Hitch/Me31B to form a complex that mediates secretion of the TGF-α homolog Gurken (Grk) (Kugler et al., 2009). As Grk is localized within actin-coated structures during mid-oogenesis, dBic-C is thought to regulate oogenesis, cytoskeletal organization and its own expression by recruting CCR4-NOT deadenylase to target mRNAs (Chicoine et al., 2007; Snee and Macdonald, 2009). Thus, the evidence suggests that Bic-C may exert an important role in the spatial-temporal regulation of gene expression in embryonic development.
In mice, the Bicc1-mutant mouse models, jcpk and bpk (Cogswell et al., 2003; Zhou L, 2008), both exhibit cystic phenotypes in the kidney that are very similar to human polycystic kidney disease. Yet, the models seem to retain Bicc1 mRNA even in homozygote mutants (Cogswell et al., 2003). A recent study reports that targeted inactivation of Bicc1 results in a randomized left-right (LR) asymmetry. This phenotype may be induced by disrupting the planar alignment of motile cilia which can drive cilia-driven fluid flow. The study linked Bicc1 to the orientation of cilia with PCP, possibly by regulating RNA silencing in P-bodies (Maisonneuve et al., 2009). To better understand the role of Bicc1, we established stable Bicc1-silenced IMCD cell lines. With these cell lines, we demonstrate that the expression of Bicc1 is required to sustain renal tubulogenesis in vitro. Lack of the protein resulted in abnormalities in cell-cell contact, actin cytoskeleton organization, cell-extracellular matrix interactions, cell proliferation and apoptosis.
Notably, we found that E-cadherin-mediated cell-cell contacts were disrupted in Bicc1-silenced IMCD cells. E-cadherin-mediated cell-cell adhesion is an initial signal that subsequently triggers assembly of intercellular junctions, and thus is essential for epithelial polarity and tubule formation (Higashiyama S, 1995; Matter K, 2003; Zegers MM, 2003). Our finding that lack of Bicc1 disrupts normal tubulomorphogenesis in vitro is likely due to the failure of Bicc1-silenced cells to establish normal E-cadherin-mediated cell-cell adhesion. Once the assembly of adhesion and tight junctions is disrupted, the signals transmitted from the apical junctional complex to the cell interior modulate gene expression and regulate cell proliferation and differentiation (Matter K, 2003). The findings of this study indicate that dysfunction of Bicc1 disrupts the spatial-temporal formation of E-cadherin-mediated cell-cell adhesion, this impedes epithelial polarization leading to abnormal gene expression that results in diverse biological alterations such as were observed in our Bicc1-silenced IMCD cells.
Many studies have reported that loss of either ADPKD causal gene products (polycystin-1 and -2, PC1 and PC2) or ARPKD causal gene product (fibrocystin, FPC) can disrupt normal tubulobranching and tubulogenesis in a 3-D culture system (Boletta et al., 2000; Grimm et al., 2006; Kim I, 2009; Mai et al., 2005). For example, work by Boletta et al indicated that overexpression of PC1 promoted MDCK cells to spontaneously form branching tubules, while control cells formed cyst-like structures (Boletta et al., 2000). Recently, we reported that lack of either PC2 (whose mutations result in ADPKD) or FPC (whose mutations result in ARPKD) impairs normal tubulogenesis in a 3-D culture system (Kim I, 2009; Mai et al., 2005). That this same tubulomorphogenic defect was seen in Bicc1-silenced cells lends validity to the hypothesis that a loss of Bicc1 induces cyst formation via a similar mechanism.
A hemostasis between proliferation and apoptosis is essential for organogenesis and maintenance of organ size (Zhao et al., 2008). Abnormal cell proliferation and apoptosis have been reported to closely associate with pathogenesis of PKD (Wegierski T, 2009; Wilson PD, 2007; Woo, 1995). We recently reported that lack of human cystoproteins FPC (Mai et al., 2005) and PC2 (Kim I, 2009) increased apoptosis in renal epithelial cells. These findings coincide with our observation of enhanced apoptosis in the Bicc1-silenced cells, and suggest that Bicc1 may play a similar role in apoptotic function as polycystins and fibrocystin.
Increased proliferation rates have been previously reported in many polycystin-deficient renal epithelial cells and tissues (Nishio et al., 2005; Wilson PD, 2007). However, we found the opposite effect, with a clear decrease in cell proliferation when expression of Bicc1 was inhibited. To confirm our findings, we used a different Bicc1 knockdown construct (Bicc1shRNA-N) to produced IMCDshN8E and IMCDshN3GBicc1-silenced cell lines. Similar proliferation rates can be observed when compared to IMCDshC1C and IMCDshC4CBicc1-silenced cell lines (Supplemental Fig. 2E-F). In addition, we also performed parallel proliferation assays with Pkd2-null renal collecting cells. The data showed that Pkd2-null renal collecting cells displayed a significant increase in proliferation rates (Kim I, 2009), while Bicc1-silenced IMCD cells consistently displayed a reduction in their proliferation rates. In prior experiments, we demonstrated that Pkhd1-silenced IMCD cells (Pkhd1 encodes FPC) also resulted in a decreased proliferation rate (Mai et al., 2005). That Pkhd1- and Bicc1-silenced cells yield a decreased proliferation rate while Pkd2-null cells display an increased proliferation rates hints that these cystoproteins may exert opposing effects to maintain hemostasis during organogenesis and thus modulate tubulomorphogenic and organ size. Disruption of this hemostasis may eventually result in cyst formation such as is found in PKD.
Through our characterization of stable Bicc1-silenced IMCD cells, we demonstrate that downregulation of Bicc1 disrupts normal tubulomorphogenesis and tubulobranching in vitro, which result from the failure of establishing normal E-cadherin-mediated cell-cell adhesion. This E-cadherin failure causes lose of signals essential for assembling normal cell-cell/cell-ECM interactions, establishing normal epithelial polarization, and modulating normal differentiation, proliferation, and apoptosis. The data obtained from the Bicc1-silenced IMCD cell lines reveals new functions for Bicc1 in mammalian epithelia, and the information gained may eventually give rise to a new cystogenic mechanism for human polycystic kidney disease.
The commercial reagents and antibodies used in this study were as follows. pRS-shRNA vector system (Origene), pCMV-Tag4 expression vector (Stratagen), Rat type I collagen (CI) (Becton Dickinson), gelatin (Sigma), paraformaldehyde (Sigma), Rhodamine-phalloidin (Vector Laboratories), anti-ZO-1 antibody (Zymed Lab, Inc.), anti-E-cadherin antibody (BD Transduction Laboratories), and DAPI (Invitrogen). Secondary antibodies included Cy3-conjugated rabbit anti-mouse IgG and Cy2-conjugated goat anti-rabbit IgG (Jackson Laboratories).
Mouse inner medullar collecting duct (IMCD) cells were cultured using the conditions described in the American Type Culture Collection manual. The 3-D tubulogenesis assay was as previously described (Mai et al., 2005). Briefly, the Matrigel (MG)/collagen I (CI) gels were a 1:1 mixture of CI and MG with a final concentration of 0.5 mg/ml for CI and 0.5 mg/ml for MG. Ten percent fetal calf serum (FCS) was used for the 3-D MG/CI gel cultures. Tubule formation was determined in five randomly picked high-power fields. For the cell lines used to establish cell polarity, cells were plated to confluence for at least 5 days on either 12-mm transwell plates or 6-mm culture plates (Costar). To establish stable Bicc1-silenced and control IMCD cell lines, pRS-Hush-RNAi-C and pRS-Hush-GFP (Fig. 1A) were transiently transfected into subconfluent IMCD cells. Twenty-four hours later, the media was changed to 7% FCS DMEM/F12 with Puromycin at a concentration of 1 μg/ml to select for Puromycin-resistant clones. After one week, the remaining cells were resuspended and seeded in 100 mm culture plates (Costar) at a cell density of 103 per plate. Once Puromycin-selected colonies formed, single colonies were picked using an inverted microscope and transferred into a new set of 24-well cultured plates (Costar).
A pGEX3 GST expression vector (Amersham) was used as a backbone for producing Bicc1 GST-fusion proteins. The amino acid sequence of the fusion portion is from 61E to 199A, and designated mBic-N. The mBic-N cDNA insert was fused in-frame to construct the pGEX3-mBic-N-GST expression vector. This vector was transformed to Rossetta host cells (Novagen) to produce GST-fusion antigens. The antigen was subcutaneously injected into New Zealand white rabbits at a dosage of 0.6 mg per week for 3 injections. The polyclonal antiserum was affinity purified and was tested for the Bicc1-specificity (Supplemental Fig. 1) (Dai B, 2008).
Total RNA was isolated from the IMCD-derived stable cell lines, using Trizol reagents (Invitrogen) according to the manufacturer's instructions. The cells were harvested after 3 days of confluence. Quantitative PCR was performed using the iCycler iQ Real Time PCR Detection System (Bio-Rad). The relative expression level of Bicc1 was normalized using the 2-ΔΔCT method (Schmittgen and Livak, 2008). A pair of primers was designed to bridge exons 2 and 3 of Bicc1 cDNA. The forward primer: 5′-TCT TGC ACA GCC CG GAG-3′, and reverse primer: 5′-CGA TCT TCA GTT TGG ACG GC-3′ were used for quantitative PCR with the iQ SYBR Green Supermix kit (Bio-Rad, Richmond, CA). PCR was performed for 35 cycles, each consisting of denaturation at 94°C for 30 seconds, annealing at 56°C for 30 seconds, and extension at 72°C for 1 minute. GAPDH (forward primer, 5′-GAC CAC AGT CCA TGC CAT CAC T-3′, reverse primer, 5′-TCC ACC ACC CTG TTG CTG TA-3′) was amplified to confirm an equal amount of total mRNA in samples and was used to normalize samples.
For western blot analysis, cell lysates from 3-day confluent cultures were extracted in lysis buffer (0.5% Nonidet P-40/5% sodium deoxycholate/50 μM NaCl/10 μMtris/HCl (pH 7.5)/1% BSA) and centrifuged at 4°C for 15 mins. The supernatant was mixed with 2× loading buffer and boiled for 5 mins after which the samples were separated using 4%-10% gradient SDS/PAGE gels (Bio-Rad) and transferred to nitrocellulose membranes. The membranes were incubated with 5% fat-free milk at room temperature for 1 h, blotted with primary antibodies at room temperature for 4 hrs, washed, and then incubated with peroxidase-conjugate secondary Ab. Detection of immunoreactivity was with enhanced chemiluminescence (Amersham Pharmacia). Normalized quantitative analyses of western blot results derive from densitometry values of immunoreactive bands for Bicc1 and β-actin (loading control).
For immunofluorescence staining, cultured cells were washed with 1×PBS twice and fixed with 4% paraformaldehyde at 4°C for 30 mins. The cells were permeablized in 0.1% Triton-X100 on ice for 5 mins. After blocking in 2% BSA solution for one hour, the cells were incubated with primary Abs for 2 hrs and washed again with 1×PBS three times. Washed cells were treated with fluorescence-conjugated secondary Abs for 1 hr. For confocal microscopy, the images of antibody staining were collected as Z series sections by using a Zeiss LSM 510 confocal microscope system with a 40× or 63× oil objectives. Multiple sections (0.3 μm in thickness) were projected onto one plane for presentation.
Cells (40,000 per well) were placed in 24-well tissue culture plates. After 5 days in culture, cells were pulsed for 24 hrs with 3H-thymidine (1 μCi/well). The media was then removed and the plates washed with PBS 3 times to remove free 3H-thymidine. The cells were lysed in 0.2% NaOH after being fixed with 10% trichloroacetic acid and the lysates were measured with a β-counter. For the Phospho-Histone H3 cell proliferation assay, cells (20,000 per well) were placed in 24-well plates that had glass coverslips in each well and cultured for 3 days. The cells were then stained with phospho-histone-H3 antibody (Cell Signaling Technology Inc.). Positive stained cells were counted from three randomly picked high-power fields (40×).
For apoptosis studies, cells were cultured in 24-well plates and grown to sub-confluence in 7% FCS DMEM/F-12 (1:1) medium (Gibco) under 5% CO2 at 37°C. The cells were incubated with 0.5 μM Ionomycine (Alexis Corp.) under serum-free conditions. Six or 12 hrs later, TUNEL assay (DeadEnd™ florometric TUNEL system, Promega) or Caspase-3 Active Apoptosis Kit I (BD Biosciences) were used according to the manufactor's manual, respectively. The apoptotic cells were counted in three randomly picked high-power fields (40×).
The 96-well cell culture plates (Nunc) were coated with CI at the indicated concentrations in PBS for 12 hrs at 4°C. The negative controls were performed on plates coated with 1% BSA. Positive controls were cells plated onto tissue culture plates in the presence of 10% FCS. The plates were washed with PBS and incubated with PBS containing 1% BSA for 60 mins to block nonspecific binding. Aliquots (100 μl) of single-cell suspensions (106 cells/ml) in serum free DMEM/F12 containing 0.1% BSA were added in triplicate to 96-well plates with 0.03 - 2 μg/ml CI and incubated for 60 mins at 37°C. Non-adherent cells were removed by washing the wells with PBS. Cells were then fixed with 4% formaldehyde, stained with 1% crystal violet, solubilized in 20% acetic acid, and the OD of the cell lysates read at 570 nm. Cells bound to FCS were used as a positive control to indicate maximal cell adhesion, and the amount of cells bound to CI-uncoated wells (BSA only) was used as the background, and this OD was subtracted from that obtained with serum or ECM proteins. We evaluated the cell-matrix adhesion by the formula: (OD value of tested cells minus OD of background/OD of positive controls minus OD of background) × 100.
For the transwell filter migration assay, cells were placed onto polyvinylpyrrolidone free polycarbonate filters with 8 μm pores (Costar). Aliquots (100 μl) of cell suspension (5×105 cells/ml) in 7% FCS medium were added to the wells, and cells were allowed to migrate to the underside of the transwell for 24 hrs. Cells on the top of the filter were removed by wiping and the filter was then fixed in 4% formaldehyde in PBS. Migrating cells were stained with 1% crystal violet and five randomly chosen fields from triplicate wells were counted at 20× magnification.
The tested cells were plated on transwell filters (12 well, 0.4 μm pore size, Corning Costar) with 5×104 cells/well and allowed to attach overnight, to form a confluent monolayer in normal culture medium. Transepithelial resistance (TER) was measured after 24 hrs of plating and every 24 hr thereafter using an EVOM/STX2 (World Precision Instruments) electrical resistance measurement system to measure the value for 6 days and the results were expressed in Ω/cm2.
All assays were repeated at least 3 times in duplicate or triplicate and the graphic data presented as the mean ± SD unless otherwise stated. Statistical analysis was performed where appropriate using the Student's t-test or one-way analysis of variance (ANOVA) followed by the Tukey's Multiple Comparison Test. Differences with P-values < 0.05 were considered statistically significant.
Specificity of the mBic-Np antiserum. (A) Western blot of lysates from Rosetta bacteria with pGEX3-mBic-N expression vector. Lane 1 shows no band was detected with pre-immune (pre-IM) serum from the same rabbit. Lane 2 shows pGEX3-mBic-N was detected by mBic-Np (1:10000 dilution) at the expected size (~40kD). Lane 3 shows the same size band was detected using an anti-GST-monoclonal antibody. (B) pCMV-Tag4-mBicc1 expression vector in which a Flag tag was fused in-frame with mouse Bicc1 full length cDNA and transiently transfected into HEK293 cells, a cell line which has very low endogenous levels of Bicc1. Duplicate samples show there is no band in non-transfected cells when probed with the mBic-Np antibody (lane 1) and no specific band in present in lysates from HEK293 transfected cells probed with pre-IM serum from the same rabbit (lane 2). Bands of ~105kD were detected in lysates from HEK293 transfected cell when probed with the affinity-purified mBic-Np antibody (lane 3), or an anti-Flag antibody (lane 4). β-actin is used to verify equivalent protein loading.
Establishment of stable Bicc1-silenced IMCD cell lines using Bicc1-shRNA-N construct. (A) Bicc1-shRNA-N was inserted into the pRS-shRNA vector. (B) Quantitative PCR was used to examine Bicc1 mRNA expression levels in wildtype (IMCDWT), control (IMCDshGFP1) and Bicc1-silenced (IMCDshN3G and IMCDshN8E) cell lines (*P<0.05). The bars and error lines represent the mean and standard error of triplicates. (C) The wildtype (IMCDWT), control (IMCDshGFP1) and Bicc1-silenced (IMCDshN3G and IMCDshN8E) cell lines were subjected to western blot analysis with mBic-Np. Lysates from HEK293 cells transfected with an expression vector, which contains full-length Bicc1 cDNA, were used as a positive control. Bicc1-silenced cell lines (IMCDshN3G and IMCDshN8E) showed a significant reduction in Bicc1 protein expression when compared with IMCDWT and IMCDshGFP1 cells. Equivalent protein loading was demonstrated using anti-β-actin antibody on the same western blot. (D) Wild-type (IMCDWT), control (IMCDshGFP1) and Bicc1-silenced (IMCDshN8E and IMCDshN3G) cells were cultured for 7 days in 3-D MG gels. A significantly higher rate of cyst formation and cell aggregation was observed for the Bicc1-silenced IMCD cell lines (IMCDshN8E and IMCDshN3G) when compared to either wild-type or controls cell lines (IMCDWT and IMCDshGFP1). White bars indicate no tubular branching; gray and black bars show 1-3 and >3 tubular branches respectively. (E) The same cell lines were incubated with 3H-thymidine after which the rate of 3H-thymidine incorporation was determined as described in Materials and Methods for at least three times. 3H-thymidine values were significantly different between cells with and without Bicc1-silencing (*P<0.05). (F) The wildtype (IMCDWT), control (IMCDshGFP1) and Bicc1-silenced (IMCDshN8E and IMCDshN3G) cells were used in Phospho-Histone H3 staining to evaluate cell proliferation. The percentage of cells found to be positive for Phospho-Histone H3 was significantly different between cells with and without Bicc1-silencing (*P<0.05).
Three-dimensional cultures of wildtype (IMCDWT) (a-c) and Bicc1-silenced cells (IMCDshC1C) (d-f) were placed onto glass-slides and stained with an anti-E-cadherin antibody (red). Co-staining with DAPI, a dye for nucleic acids (blue), is also shown (b and e). Merged microscopic images indicate a more diffuse E-cadherin distribution in Bicc1-silenced cells (c vs f). Bar: 5 μm in a-f.
This work was supported by National Natural Science Foundation of China (30672483 and 30870501), Changjiang Scholarship of China, State Key Laboratory of Molecular Oncology of China, National Institutes of Health of USA (DK062373 and DK71090) and the National Cancer Institute Specialized Programs of Research Excellence (5P50 CA095103), NIH, USA to G.W.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.