Autosomal dominant polycystic kidney disease (ADPKD) is characterized by large fluid-filled cysts and progressive deterioration of renal function necessitating renal replacement therapy. Previously, we generated a tamoxifen-inducible, kidney epithelium-specific Pkd1-deletion mouse model and showed that inactivation of the Pkd1 gene induces rapid cyst formation in developing kidneys and a slow onset of disease in adult mice. Therefore, we hypothesized that injury-induced tubular epithelial cell proliferation may accelerate cyst formation in the kidneys of adult Pkd1-deletion mice. Mice were treated with the nephrotoxicant 1,2-dichlorovinyl-cysteine (DCVC) after Pkd1-gene inactivation, which indeed accelerated cyst formation significantly. After the increased proliferation during tissue regeneration, proliferation decreased to basal levels in Pkd1-deletion mice just as in DCVC-treated controls. However, in severe cystic kidneys, 10–14 weeks after injury, proliferation increased again. This biphasic response suggests that unrestricted cell proliferation after injury is not the underlying mechanism for cyst formation. Aberrant planar cell polarity (PCP) signaling and increased canonical Wnt signaling are suggested to be involved in cyst formation. Indeed, we show here that in Pkd1 conditional deletion mice expression of the PCP component Four-jointed (Fjx1) is decreased while its expression is required during tissue regeneration. In addition, we show that altered centrosome position and the activation of canonical Wnt signaling are early effects of Pkd1-gene disruption. This suggests that additional stimuli or events are required to trigger the process of cyst formation. We propose that during tissue repair, the integrity of the newly formed Pkd1-deficient cells is modified rendering them susceptible to subsequent cyst formation.
Autosomal dominant polycystic kidney disease (ADPKD) is characterized by progressive deterioration of kidney function owing to the formation of fluid-filled cysts. Around the second decade, in general, only a few cysts can be detected by ultrasonography, whereas at middle age, renal function declines and thousands of cysts are usually being found. ADPKD has an incidence of 1 in 400 to 1 in 1000 in Caucasians and is in fact a systemic disease with cyst formation as the major hallmark. Extra-renal manifestations are hypertension, formation of aneurysms and formation of cysts in liver and pancreas (1). In most patients, ADPKD is caused by either a mutation in the PKD1 gene, encoding the protein polycystin-1 (PC1; 85% of clinical cases), or a mutation in the PKD2 gene, encoding polycystin-2 (PC2; 15% of clinical cases) (2–4).
The exact roles of PC1 and PC2 in cyst formation or other disease manifestations are not clear yet. It is known that, in kidneys, the primary defect occurs in the epithelium and that a balanced expression level of these proteins is critical to maintain renal epithelial architecture. In renal epithelial cells, deregulated PKD1/2 gene expression, either by dose reduction or by complete gene inactivation through somatic mutations, initiates cystogenesis (5–7). Both PC1 and PC2 are expressed in several cellular compartments. The proteins form different multimeric protein complexes in different subcellular locations and they regulate several signaling pathways that together control essential cellular functions such as proliferation, apoptosis, adhesion and differentiation (8,9). As all these activities are highly co-coordinated, disruption of any of these processes can lead to cyst formation as shown by a variety of mouse models (e.g. bcl2 knock-out, c-myc overexpression, β-catenin mutants, Fat4 knock-out) (10–13).
We recently generated an inducible Pkd1-deletion mouse model in which we can induce Pkd1 deletion specifically in renal epithelial cells by the administration of tamoxifen. These mice carry a tamoxifen-inducible KspCad–Cre as well as Pkd1lox2-11,del2-11 or Pkd1lox2-11,lox2-11 alleles (14,15). Upon the administration of tamoxifen, the Pkd1 gene is deleted and renal cysts are being formed. Interestingly, adult mice (3–5 months) in which the deletion was induced showed a relatively mild cystic phenotype after 3–5 months, whereas gene disruption in newborn mice (post-natal day 4) resulted in massive cyst formation after 1 month. Similar data were reported for other models (16–18). At neonatal stages, kidney development is not yet completed. The kidneys might be more susceptible for Pkd1-gene inactivation, given the developmental switch in gene expression between days PN12 and 14 (16). In addition, cell proliferation takes place to elongate the nephron, and therefore, we hypothesize that proliferation may accelerate cyst formation in (adult) ADPKD (15).
It has been proposed that tubular lengthening during renal development requires oriented cell division, which is the result of a correctly positioned orientation of the nuclear spindle axis. The planar cell polarity (PCP) pathway controls oriented cell division and is disturbed in non-orthologous models for PKD (19,20). This is one of the pathways mediated by Wnt signaling, since the intracellular response to Wnt depends on the signaling cascade activated in the responding cell. Cells can activate either the non-canonical pathway that controls PCP or the canonical pathway that modulates gene expression to control cellular differentiation and proliferation. A molecular switch between these pathways requires the involvement of the protein inversin (21).
Also, tissue regeneration after injury requires massive cell proliferation and proper differentiation. Therefore, we studied the effects of treatment with a nephrotoxic compound on cyst formation and proliferation in an adult inducible Pkd1-deletion model. Indeed, renal injury accelerated cyst formation but did not result in unrestricted cell proliferation. Furthermore, we show that, in Pkd1-deletion mice, tissue regeneration is accompanied by reduced expression of the PCP gene Four-jointed (Fjx1), together with increased expression of targets from the canonical Wnt signaling. In addition, we measured a deviation of the position of centrosomes in mutant compared with control mice. Overall, these data suggest that the integrity of the newly formed cells is altered.
Rate of cyst formation correlates with cell proliferation at the time of gene disruption
We previously showed that a substantial number of proliferating Ki-67-positive renal epithelial cells could be observed in newborn wild-type mice, whereas in adult mice, Ki-67-positive nuclei were hardly found (15). We hypothesized that the proliferation status of the tissue may contribute to the observed pathogenic differences between the adult and newborn tamoxifen-treated conditional deletion mice. In the mean time, we also generated mice with gene disruption at 5.5 weeks post-natal (PN40). These young-adult mice showed a more progressive phenotype compared with adult mice with gene disruption at 3 months, but less progressive compared with gene disruption at neonatal stages (not shown). Indeed, the proliferation index (PI) at 5.5 weeks is intermediate to newborn and adult stages (1.5 ± 0.57%; n = 6), whereas the mean PI in kidneys of young mice (post-natal week 1) is 6.8 ± 1.50% (n = 3), and in adult mice (>3 months), the mean PI is ∼0.2 ± 0.22% (n = 3) (Fig. 1).
Renal epithelial injury accelerates cyst formation in adult conditional Pkd1-deletion mice
Tissue regeneration upon injury is accompanied by increased cell proliferation. To determine whether injury-induced proliferation accelerates cyst formation, adult Cre;Pkd1del2-11,lox mice were treated with the nephrotoxic compound 1,2-dichlorovinyl-cysteine (DCVC) (cKO+DCVC) or vehicle (cKO-DCVC) (22,23). As control, Pkd1del2-11,wt and Pkd1del2-11,lox mice were treated with DCVC (control+DCVC). Blood urea (BU) concentrations were measured to monitor renal function, and mice were sacrificed at BU levels >20 mmol/l (Fig. 2A and Supplementary Material, Fig. S1). Cre-mediated recombination efficiency was 40–50% in all cKOs as determined by extension multiplex ligation-dependent probe amplification (eMLPA).
At 40 h after treatment, all DCVC-treated mice showed increased BU concentrations, indicative for renal injury. After 2 weeks, BU levels were normalized to baseline. After 10–14 weeks, BU levels increased again only in cKO+DCVC, and all groups were sacrificed. Kidneys of cKO+DCVC mice showed a severe cystic phenotype (Fig. 3C), whereas non-treated cKO mice showed only mildly dilated tubules (Fig. 3E) and occasionally a limited area with focal cysts after 10–15 weeks. The majority of cysts are from proximal tubular origin, the nephron segment that is most injured by DCVC treatment (Fig. 3F) (22,23). At 2 weeks after injury, only very mild tubular dilatation was observed in cKO+DCVC. This became more prominent at 5 weeks after injury, although dilatations were still mild (Fig. 3A and B, respectively).
Controls+DCVC showed no cysts or signs of tubular dilatation after 10 weeks (Fig. 3D), or even after 6 months (not shown). Concomitantly, the renal cystic index was significantly increased in cKO+DCVC mice compared with both control groups, cKO-DCVC and controls+DCVC (Fig. 2B). The kidney-to-bodyweight (KW/BW) ratio, another marker for disease severity, is also markedly elevated in cKO+DCVC mice at 10–15 weeks (Fig. 2C).
From these data, we conclude that renal injury accelerates cysts formation in adult Pkd1-deletion mice.
Proliferation normalizes to baseline levels after tissue regeneration in Pkd1-deletion mice
To study the levels of epithelial cell proliferation at different stages upon renal injury, we determined PIs in tubular epithelial cells of the renal cortex, the cortico-medullary (CM) region and the medulla, using the proliferation marker Ki-67 (Fig. 4).
One week after injury, the PIs were increased in cKO+DCVC and controls+DCVC compared with untreated mutants, and the levels were higher in cortex and CM region in cKO+DCVC (Fig. 4A and B). After the initial repair phase, proliferation returned to the same level as in cKO–DCVC, reaching baseline levels at 2–5 weeks. This pattern is most clearly seen in the CM region and cortex, which are mostly injured by DCVC treatment (Fig. 4) (22,23). In severe cystic kidneys of cKO+DCVC at 10 weeks, proliferation was dramatically increased again. This biphasic response, in which proliferation is increased during repair but returns to basal levels before increasing again, was not observed in cKO–DCVC and controls+DCVC, which do not develop renal cysts at 10–14 weeks after injury.
These data suggest that cyst formation after injury is not the result of unrestricted cell proliferation induced by tissue repair.
PCP gene Fjx1 is differentially expressed upon renal injury in precystic Pkd1-deletion mice
Since proliferation reached baseline levels after tissue regeneration, but the DCVC-treated mutants form cysts much earlier than untreated mutants, we investigated whether the expression of genes involved in oriented cell division/PCP during tissue repair, recovery and at precystic and cystic stages was altered. Therefore, we studied mRNA expression of Four-jointed (Fjx1), Moesin (Msn) and Inscuteable (Insc) at different time points. These genes were selected since they showed differential expression upon microarray analysis of cKO mice in which the Pkd1 gene was disrupted at PN40. At 2 months upon gene disruption (PN40+2 months), which is at precystic stages with mild tubular dilatations, only a limited number of genes (298/34 000; manuscript in preparation) were differentially expressed, including Fjx1 (downregulated), Msn (upregulated) and Insc (upregulated) (Supplementary Material, Fig. S2A). For other genes involved in PCP, the expression was too low (Fat4, Prickle1), or did not show differential expression [Dachsous (Dchs1), Prickle2, Vangl and Celsr (Flamingo)].
Importantly, expression of Fjx1, a key regulator of PCP, was also reduced during the repair process at 1 and 2 weeks in cKO+DCVC, whereas the expression was increased in controls+DCVC (Fig. 5A). This difference persisted until 5 weeks after injury. At 10 weeks, in severe cystic kidneys, the expression of Fjx1 was significantly increased compared with cKO–DCVC and control+DCVC. The expression of Insc and Msn was not significantly different in the different groups, although the expression was slightly elevated at 5 weeks for Insc and significantly at 10 weeks for Insc and Msn.
Canonical Wnt targets are differentially expressed upon renal injury in Pkd1-deletion mice
Since it is expected that altered PCP is associated with increased canonical Wnt signaling, we analyzed mRNA expression of several canonical Wnt targets and regulators at different time points upon renal injury. The genes selected showed differential expression on microarrays at precystic stages of cKO mice with Pkd1-gene disruption at post-natal day 40. The expression of these genes is increased, except for Notum and Secreted Frizzled related protein 1 (Sfrp1), which show strongly reduced mRNA levels (Supplementary Material, Fig. S2B). Both genes are canonical Wnt targets as well as negative regulators of Wnt signaling, suggesting that they are subject to complex regulation (24,25). For eight other canonical Wnt targets, sufficient signal was detected on the arrays but the expression was not significantly altered (not shown).
In the DCVC-treated animals, we analyzed the expression of Cyclin D1 (Ccnd1), Survivin (Birc5), Prominin-1 (Prom-1) and Notum (Notum), which are direct canonical Wnt targets with differential expression on microarrays at precystic stages. In addition, we analyzed the expression of Sfrp1, an indirect target.
In general, the expression of all these genes, except the negative regulators, peaked at 1 week upon DCVC. However, at precystic stages, mRNA levels were higher in cKO+DCVC compared with controls+DCVC as shown for Cyclin D1, Survivin and Prominin-1 (Fig. 5B–D). At 1 and 2 weeks, the relative expression was not always significantly increased, but at 5 weeks it was. Also in cystic kidneys, at 10 weeks, mRNA levels were significantly elevated.
For most genes, the expression was also (significantly) elevated in cKO-DCVC mice compared with controls+DCVC, which is in agreement with increased expression seen on the microarrays of PN40+2 month mice. Our data indicate that Pkd1-gene disruption induces canonical Wnt signaling already at very early, precystic stages.
The expression pattern for Cyclin D1 mRNA was confirmed by immunohistochemical staining for the Cyclin D1 protein. At 1 week after DCVC treatment, strong nuclear accumulation of Cyclin D1 could be observed in many nuclei in mutants and controls (Fig. 3G–I). After the initial increase of Cyclin D1 staining, the intensity decreased, and at 5 weeks, an increase in number and intensity of positive nuclei could be observed (Fig. 3 H and I). This was always higher in cKO+DCVC compared with controls+DCVC. Also in cKO–DCVC, Cyclin D1 staining increased over time.
The expression of the two negative regulators of Wnt signaling, Sfrp1 (Fig. 5E) and Notum (not shown), was significantly lower in cKO+DCVC compared with controls with DCVC. Notum can cleave Glypican-3, a canonical Wnt activator, and converts this proteoglycan from being a Wnt activator to a Wnt inhibitor. The expression of the inhibitor Notum was reduced in cKO+DCVC, and Glypican-3 (Gpc3; Fig. 5F) mRNA levels were substantially elevated.
From these data, we conclude that upon Pkd1-gene disruption, PCP signaling is altered and canonical Wnt signaling is increased.
Pkd1 deletion affects centrosome position
Mitotic spindle orientation has been used as a read-out system for aberrant PCP in kidney development (19,20). In the injury model, aberrant PCP signaling has to be determined after the repair of the injured tissue but in precystic stages. However, in these adult mice, cell proliferation is very low and only a few nuclear spindles can be observed. Since components of the primary cilia/basal body/centrosome complex also contribute to the formation of the nuclear spindles, we studied centrosome position in interphase cells as described by Jonassen et al. (26).
Centrosome position was measured in proximal tubules in the CM region at 2 weeks after DCVC treatment. In the absence of Pkd1, the normal bias towards a neutral centrosome position, close to the center of the apical surface of the cell, was lost (Fig. 6). This resulted in a significantly different distribution of angles defining centrosome position in cKO+DCVC (P < 0.0001) and cKO-DCVC (P < 0.05) compared with control+DCVC. As can be appreciated from Figure 6, this effect is stronger in cKO+DCVC, although this did not reach statistical significance.
These results indicate that after Pkd1 inactivation, centrosome position is altered.
ADPKD is an adult onset disease characterized by the formation of many fluid-filled cysts in both kidneys with advancing age. When the level of Pkd1 or Pkd2 drops below a critical threshold in renal epithelial cells, either by stochastic variation or because of somatic mutations, the differentiation status of the cells alters, resulting in increased cell proliferation, extracellular matrix synthesis and altered fluid transport. Previously, we showed that when we knock out the Pkd1 gene in newborn mice, rapid formation of cysts occurs, whereas in adult mice, there is a lag time of a few weeks between the moment of gene disruption and the onset of increased epithelial cell proliferation and cyst formation (15). Therefore, we hypothesized that the proliferative status of the renal tissue at the time of gene disruption is important for the rate at which cysts are formed after Pkd1 conditional deletion. In addition, a different gene expression profile in neonatal and adult kidneys may contribute to these phenomena as well (16).
Our data show that Pkd1-gene disruption at PN40 leads to a disease progression rate intermediate to that of neonatal and adult mice. At this stage, mice show levels of proliferation between neonatal and adult mice and are beyond the developmental switch in renal gene expression (PN12–14). We also show that DCVC-induced renal epithelial injury significantly accelerates the progression of renal cyst formation. The majority of the cysts were derived from the proximal tubules at the inner cortex, which is the nephron segment that is the main target of injury by this nephrotoxicant (22,23). A similar result has recently been shown for a conditional knockout model of the ciliary transport component Kif3a, which develops more progressive renal cystic disease upon ischemic injury (20).
Renal epithelial injury is followed by a repair phase that includes proliferation to restore the tubular epithelial architecture. Importantly, in DCVC-induced Pkd1-deletion mice, proliferation decreases to basal levels after repair at 2 and 5 weeks, similar to DCVC-induced controls. From this, we can conclude that, also in the context of the renal tissue, absence of Pkd1 expression does not lead to unrestrained epithelial cell proliferation. These results are corroborating previous reports showing that cystic epithelial cells of Pkd1 null mice do not grow in nude mice (27). In Pkd1-deletion mice, cell proliferation upon renal injury showed a biphasic response, e.g. after the initial peak in the repair phase, renal epithelial cell proliferation increased again during cyst formation at 10–14 weeks.
Although cell proliferation during tissue regeneration is controlled in DCVC-treated mutants, these mice form cysts much earlier than untreated mutants. Similar to embryogenesis and renal development, tissue repair reactions require properly regulated cell proliferation to restore the epithelial architecture. Defects in oriented cell division were observed during tubular elongation in renal development in non-orthologous animal models for PKD (19,20). In addition, loss of the PCP gene Fat4 leads to renal cysts formation (13). Altogether, these data suggest a potential role for aberrant PCP or oriented cell division, two closely linked pathways to which function only a few mammalian genes have been assigned as yet, in cystogenesis. Here, we investigated whether defects in PCP during tissue regeneration could be involved in accelerated cyst formation as observed in adult Pkd1-deletion mice. In adult precystic kidneys, the number of proliferating nuclei with a detectable nuclear spindle is too low to measure (20). Therefore, we studied centrosome position as described by Jonassen et al. (26), who showed that the intraflagellar transport protein IFT20 plays a role in controlling Wnt signaling. Furthermore, it is required for proper positioning of the centrosomes in non-dividing cells and for correct orientation of the mitotic spindles in dividing cells. We analyzed the position of centrosomes after repair of injury, but at a precystic stage. As expected, in the majority of cells of control mice (+DCVC), centrosomes are positioned close to the center of the apical membrane. In mutants with and without DCVC, however, a significantly different pattern in the distribution of centrosome position was observed. From these data, we conclude that improper positioning of centrosomes is an early event upon Pkd1-gene disruption. Jonassen et al. (26) suggested that in renal epithelial cells, the cilium may provide an external cue to position the mother centrosome, thereby coordinating the plane of cell division of an individual cell with neighboring tubular cells. Whether the effect of Pkd1 gene disruption on centrosome localization is a result of polycystin-1 requirement in normal ciliary structure and function, or through the interaction with Wnt signaling components, or both, requires further research.
Altered PCP signaling was further studied by gene expression analysis. Microarray-based gene expression profiling in precystic kidneys of conditional Pkd1-deletion mice (PN40+2 months) revealed differential expression of the PCP-related genes Fjx1, Insc and Msn, and the expression of these genes was studied during tissue regeneration. Although the subcellular localization of the encoded proteins is probably also important, the expression pattern of Fjx1 is particularly interesting. Four-jointed is a Golgi-associated kinase that phosphorylates the extracellular domains of Fat and Dachsous (28). In the first week after DCVC treatment, Fjx1 mRNA levels are elevated in control mice, suggesting that Four-jointed is needed during tissue repair. In Pkd1-deletion mutants, however, the expression of Fjx1 goes down, suggesting defects in PCP signaling.
PCP is regulated by the non-canonical Wnt pathway. So, defects in PCP signaling could be expected to be accompanied by an increase in canonical Wnt signaling. Indeed, microarray gene expression analysis showed elevated mRNA levels of canonical Wnt targets in precystic stages of conditional Pkd1-deletion mice. Moreover, the expression of the negative Wnt regulatory genes Sfrp1 and Notum was decreased. After renal injury, most of the Wnt targets analyzed showed a similar biphasic response for gene expression as observed for proliferation, with higher levels in the precystic stage in DCVC-treated mutants than in DCVC-treated controls. Cyclin D1 expression, however, did not show the second increase in gene expression, e.g. mRNA levels were not higher in cystic kidneys than in untreated mutants. Regulation of this gene and some other Wnt targets is complex, implicating several parallel signaling pathways with positive and negative feedback loops.
Since also in untreated Pkd1-deletion mice centrosome position and the expression of canonical Wnt targets are altered, we conclude that these are early responses to Pkd1 gene disruption. We speculate that these early alterations create a permissive condition for accelerating cyst formation and that additional factor(s) are required. Abnormal integrity of the newly formed cells, resulting from the absence of Pkd1 expression during repair, might be such a factor. The polycystins form multimeric protein complexes at the cell membrane and function in cell–cell and cell–matrix interactions, mechanosensation and signal transduction (8,29–32). It is likely that polycystins are required during repair for the proper formation of these protein complexes. Without Pkd1, cellular structure, geometry and cohesion will be modified. In agreement with this hypothesis, polycystin-1/-2 levels are elevated after ischemic injury (33–35). Without renal injury, gradually, cell renewal will result in Pkd1-deficient cells with strongly disturbed cellular integrity that are more sensitive to triggers for subsequent cyst formation.
Abnormal Wnt signaling has been implicated in the pathogenesis of polycystic kidney disease; ciliary defects affect Wnt signaling, overexpression of Pkd1 is accompanied by nuclear localization of β-catenin, constitutively active β-catenin induces cyst formation (12,21,36). Recent data also suggest that the C-terminal tail of PC1 may have a direct role in modulating Wnt signaling by inhibiting the interaction between β-catenin and the T-cell factor (TCF) (37). This study also demonstrated the activation of the canonical Wnt signaling pathway in cysts of different sizes, isolated from ADPKD patient tissues. Our data indicate that, already at precystic stages, canonical Wnt signaling is activated. In ADPKD, activation of canonical Wnt signaling seems to be rather subtle since we did not observe nuclear localization of β-catenin in Pkd1-deletion mice at any stage (data not shown) and as we reported previously, we could not detect nuclear accumulation of non-phosphorylated β-catenin in cyst-lining epithelium in end-stage ADPKD kidneys (38). This is different from various forms of cancer where core components of canonical Wnt signaling are constitutively activated, leading to pathophysiological levels of active β-catenin, although PC1 only seems to modulate the canonical Wnt pathway (39).
Formally, we cannot exclude that the early changes in gene expression in our mouse model are without biological significance and that stronger Wnt signaling observed in advanced stages of PKD is the result of secondary events during cyst formation. However, Lal et al. (37) suggested that PC1 expression is necessary to inhibit basal Wnt signaling, probably by an interaction of the C-terminal 200 amino acids cleavage product of polycystin-1 with β-catenin.
In conclusion, nephrotoxic injury dramatically accelerates polycystic kidney disease in adult Pkd1-deletion mice. Proliferation after injury first decreases to basal levels, indicating that accelerated cyst formation is not the result of unrestricted cell proliferation. Altered centrosome position and the activation of canonical Wnt signaling are early effects of Pkd1 gene disruption but additional stimuli or events are required to trigger the process of cyst formation.
MATERIALS AND METHODS
Experimental animals and study design
The inducible Pkd1-deletion mice (tam-KspCad-CreERT2;Pkd1del2-1l/lox2-11) and tamoxifen treatment have been described previously (14,15). The Pkd1 gene was inactivated in adult mice (∼11 weeks) by oral administration of tamoxifen (5 mg/day, 3 consecutive days). Renal injury was induced 6 days later by i.p. administration of 15 mg/kg DCVC or vehicle (40). Cre;Pkd1del,wt, Pkd1del,wt and Pkd1del,lox mice all received DCVC.
Blood samples were obtained from the tail vein at indicated time points with lithium–heparin-coated capillary tubes (Roche Diagnostics). BU concentrations were determined using the Reflotron Plus (Roche Diagnostics). Mice were sacrificed at previously determined time points, or when BU concentrations exceeded 20 mmol/l (indicating severe renal failure). After sacrifice, KW/BW ratio was determined for each animal.
Recombination efficiency was determined by eMLPA analysis as recently described by Leonhard et al. (41).
Local animal experimental committee of the Leiden University Medical Center and the Commission Biotechnology in Animals of the Dutch Ministry of Agriculture approved the experiments performed.
mRNA expression analysis
For gene expression analysis, reverse transcriptase multiplex ligation-dependent probe amplification (RT-MLPA) reactions were performed as described previously (42). Briefly, kidneys of mice sacrificed at indicated time points were homogenized and total RNA was isolated. From an aliquot of 60–200 ng total RNA, cDNA was synthesized. On this cDNA, two separate hybridization reactions were performed.
To each sample, MLPA probe mix (two oligonucleotides per probe) and SALSA MLPA buffer were added (MRC-Holland), and samples were incubated at 95°C for 1 min followed by incubation at 60°C for 4 h. Ligation of annealed oligonucleotides was performed at 54°C for 10–15 min followed by ligase inactivation at 98°C for 5 min, and ligation products were amplified by PCR using SalsaTaq (MRC-Holland) and FAM or HEX-labeled primers. The amplified samples were mixed with Hi-Di formamide containing GeneScan-500 ROX size standard (Applied Biosystems), heated for 5 min at 95°C, and run on a 3730 DNA analyzer (Applied Biosystems). Data were analyzed using GeneScan analysis 3.5 and Microsoft Excel software. The expression of the housekeeping genes Ywhaz (tyrosine 3-monooxygenase/tryptophan 5-monooxygenase activation protein, zeta polypeptide) and Hprt (hypoxanthine phosphoribosyl transferase) served as reference for gene expression.
Peak ratios of target gene and housekeeping genes were calculated and results of the different groups were normalized to the median of the control+DCVC mice at t= 1 week, which has been set to 1. Sequences and details of probes, primers and competitors are available upon request.
Quantitative gene expression analysis for Msn was performed on a 7900HT Fast Real-Time PCR System (Applied Biosystems) using SYBR Green technology with 2× SybrGreen mastermix (Applied Biosystems) according to manufacturers’ protocol. Hprt was used as a housekeeping gene. Primer sequences are available upon request.
Formalin-fixed, paraffin-embedded kidney sections (4 µm) were analyzed using standard hematoxylin and eosin staining, periodic acid Schiff (PAS) staining or were used for immunohistochemistry. Sections were subjected to heat-mediated antigen retrieval procedure (10 mm citrate buffer, pH 6.0, for anti-Ki-67, anti-cyclinD1, 10/1 mm TRIS/EDTA, pH 9.0, for anti-γ-tubulin and anti-collagen IV). After blocking of endogenous peroxidase activity for 15 min in 0.1% H2O2 in water, sections were incubated with primary antibodies diluted in 1% BSA in PBS. Following incubation with secondary antibody, immune reactions were revealed using diaminobenzidine as a chromogen and counterstained with hematoxylin, dehydrated and mounted. Slides for immunofluorescence were mounted with vectashield containing DAPI after secondary antibody incubation.
Primary antibodies: rabbit polyclonal anti-megalin [1:500, Pathology LUMC, Leiden, The Netherlands (43)], goat polyclonal anti-uromodulin (1:500, Organon Teknika-Cappel, Turnhout, Belgium), rabbit polyclonal anti-aquaporin-2 (1:4000 Calbiochem, Amsterdam, The Netherlands), polyclonal anti-Ki-67 (1:3000, Nova Castra), rabbit monoclonal anti-cyclin D1 (1:200 clone, 92G2, Cell Signaling), rabbit polyclonal anti-collagen IV IgG1 (1:500, Abcam) and mouse monoclonal anti-γ-tubulin IgG1 (1:750, clone GTU-88, Sigma-Aldrich). Secondary antibodies: anti-rabbit envision HRP (DakoCytomation, Glostrup, Denmark), rabbit-anti-goat HRP (1:300), goat-anti-rabbit IgG Alexa546 and goat-anti-mouse IgG1 Alexa 488 (both 1:200, Invitrogen).
The PI was assessed as described previously (15). Briefly, the percentage of Ki-67-positive renal epithelial cells was determined in three segments of the kidney. Between 1600 and 2200 tubular epithelial nuclei per section were counted in cortex and CM junction and 350–600 per section in medulla. Only clearly definable nuclei with heavy or granular staining were identified as Ki-67 positive.
PAS-stained transverse kidney sections were used for the determination of cystic index. Cystic index was defined as the area percentage of lumen over the total image area. Three to five images were obtained from cortex and CM region. For both renal areas, mean cystic index was calculated, after which total cystic index was calculated as the mean of both areas. Areas were determined using Image J software (public domain software, NIH, USA).
For centrosome orientation, Z-stacks were acquired using a Leica DM500B microscope with a Leica DM DFC 350 FX camera using Leica Application Suite 1.8.0 build1346. Analyzed images are maximum projections of 25–30 Z-images taken 0.24 µm apart. Centrosome positions in transverse sectioned proximal tubules at 2 weeks after injury were determined essentially as described by Jonassen et al. (26) using Image J software. In brief, the position of the centrosome was determined by measuring the angle between a line through the center of the nucleus perpendicular to the basement membrane and a line through the centrosome towards this point at the basement membrane. Collagen IV staining was used to indicate the basement membrane (Supplementary Material, Fig. S3). In total, 950 angles were measured blindly in three cKO+DCVC, three cKO-DCVC and six controls+DCVC.
Statistical comparisons between groups concerning the development of cystic kidneys within 15 weeks were performed with the Mantel–Haenszel test in which severe cystic kidneys are defined by a combination of KW/BW ratio and renal cystic index at the indicated time point.
Differences between groups in cystic index, KW/BW and mRNA expression were assessed using one-way ANOVA.
The Kolomogorov–Smirnov Z-test was used to determine differences in angle distribution between groups. In addition, a Mann–Whitney test was performed to test whether differences between groups were significant.
A P-value <0.05 was considered statistically significant.
This research was supported by grants from the Dutch Kidney Foundation (NSN C05-2132) and the Polycystic Kidney Disease Research Foundation (175G08a).
We thank Peter-Bram 't Hoen for support in microarray data analysis, the Leiden Genome Technology Center for performing the microarray experiment, Ron Wolterbeek for statistical support.
Conflict of Interest statement. None declared.