Cyst formation following disruption of intracellular calcium signaling

Significance

Autosomal dominant polycystic kidney disease is the most common cause of fluid-filled cysts within the kidney. However, how cyst formation occurs is not well understood. It is thought that proteins disrupted by this disease, such as polycystin 2, change calcium signaling, leading to the formation of cysts. In this study, we grow LLC-PK1 cells in a protein gel environment to enable the study of cysts in culture, which cannot be observed in traditional cell culture techniques. We demonstrate that loss of intracellular calcium release channels result in cyst growth and are correlated with a loss of a functional cellular component known as the primary cilia. These results demonstrate that calcium signaling is an important component in cyst development.

Abstract

Mutations in polycystin 1 and 2 (PC1 and PC2) cause the common genetic kidney disorder autosomal dominant polycystic kidney disease (ADPKD). It is unknown how these mutations result in renal cysts, but dysregulation of calcium (Ca²⁺) signaling is a known consequence of PC2 mutations. PC2 functions as a Ca²⁺-activated Ca²⁺ channel of the endoplasmic reticulum. We hypothesize that Ca²⁺ signaling through PC2, or other intracellular Ca²⁺ channels such as the inositol 1,4,5-trisphosphate receptor (InsP3R), is necessary to maintain renal epithelial cell function and that disruption of the Ca²⁺ signaling leads to renal cyst development. The cell line LLC-PK1 has traditionally been used for studying PKD-causing mutations and Ca²⁺ signaling in 2D culture systems. We demonstrate that this cell line can be used in long-term (8 wk) 3D tissue culture systems. In 2D systems, knockdown of InsP3R results in decreased Ca²⁺ transient signals that are rescued by overexpression of PC2. In 3D systems, knockdown of either PC2 or InsP3R leads to cyst formation, but knockdown of InsP3R type 1 (InsP3R1) generated the largest cysts. InsP3R1 and InsP3R3 are differentially localized in both mouse and human kidney, suggesting that regional disruption of Ca²⁺ signaling contributes to cystogenesis. All cysts had intact cilia 2 wk after starting 3D culture, but the cells with InsP3R1 knockdown lost cilia as the cysts grew. Studies combining 2D and 3D cell culture systems will assist in understanding how mutations in PC2 that confer altered Ca²⁺ signaling lead to ADPKD cysts.

The commonly occurring genetic kidney disorder, autosomal dominant polycystic kidney disease (ADPKD), is the result of mutations in polycystin 1 or 2 (PC1 or PC2). The progressive cyst formation within all segments of the nephron that defines the disorder leads to renal failure requiring treatment by dialysis and/or organ transplantation (1–3). Altered Ca²⁺ signaling is one of several pathways that have been implicated in the disease (4, 5). A major limitation toward elucidating the role of Ca²⁺ signaling in cyst formation has been the lack of easily manipulated, physiologically relevant experimental methodologies.

In the past, ADPKD research has relied largely upon data from mouse models and cells maintained in 2D cell culture. Mouse models have played a significant role in understanding the biology of cyst formation but are unable to fully recapitulate the physiology of disease progression in humans due to the inherent differences between the species including life span, genetics, and environment. Two-dimensional cell culture has the ability to provide information on signaling pathways and response to therapies in a fast, high-throughput manner, but is incapable of replicating the inherent 3D nature of cyst formation. Advances in 3D tissue culture over the past 2 decades have improved the ability to model cyst development in vitro. However, previously published 3D tissue models of ADPKD have relied upon short-term culture of Madin-Darby canine kidney (MDCK) cells (6–12) or cells from patients (13–18) or PC1-null mice (19, 20; for review, see ref. 21). Recently, 3D tissues have been developed that incorporate mouse cells containing a shRNA-mediated knockdown of PC1 (9, 19). The benefits of this system include the use of a cell line, thus eliminating the need to isolate primary cells, and the use of cells with a stable genetic background.

Ca²⁺ signaling underpins many cellular processes ranging from cell proliferation to cell death. Intracellular Ca²⁺ levels can be modified by opening of the inositol 1, 4, 5-trisphosphate receptor (InsP3R) or other intracellular Ca²⁺ release channels, including PC2. Over 99% of PC2 resides on the endoplasmic reticulum (22), where it is known to act as a modulator of the InsP3R and the ryanodine receptor (RyR) (23), with the remainder on the primary cilia. PC2 itself can function as a Ca²⁺-activated Ca²⁺ release channel (22, 24).

Although it was demonstrated in 3D cultures that the knockdown of PC1 leads to cyst development (25), the effect of knocking down PC2 or other Ca²⁺-signaling proteins has not been shown. It has been hypothesized that the disruption of PC2, or the proteins that it interacts with, will result in cyst growth, as Ca²⁺ is a major signaling molecule (26, 27). Cells with decreased PC2 have been linked with decreased Ca²⁺ signaling (28), and overexpression of PC2 has been shown to act as an inhibitor of cell proliferation (29). Changing PC2 expression levels alters the uptake of Ca²⁺ into the endoplasmic reticulum, leading to liver cyst formation (30), but no direct link involving the release of Ca²⁺ from the endoplasmic reticulum has been implicated in renal cyst development. Similarly, changes in the expression of the InsP3R have been correlated with various disease conditions; for example, the InsP3R is upregulated in colorectal cancer (31), but downregulated in bile duct obstruction and cholestasis (32, 33).

Here, we demonstrate that cyst formation can be followed for several weeks using a 3D culture system and that the disruption of intracellular Ca²⁺ signaling, through the knockdown of either InsP3R or PC2, leads to cyst development.

Results

Combined Effects of InsP3R and PC2 on Ca²⁺ Signaling.

To determine if PC2 and InsP3R can work together to affect Ca²⁺ signaling, we first established that loss of InsP3R resulted in decreased Ca²⁺ signaling. Renal epithelial cells primarily express two of the three isoforms of InsP3R type 1 (InsP3R1) and type 3 (InsP3R3) (27). We transiently knocked down either InsP3R1 or InsP3R3 in LLC-PK1 kidney cells (Fig. S1A) and examined Ca²⁺ responses to 100 nM vasopressin (AVP). The siRNA had a Cy3 tag, enabling identification of cells that were transfected. In cells transfected with siRNA directed against either InsP3R1 or InsP3R3, Ca²⁺ transients were decreased compared with the scrambled siRNA control (Fig. 1 A and B). In cells with siRNA directed against InsP3R1, both the maximum amplitude and the time taken to return to 25% of the maximum amplitude (a measure of the duration of the transient) were significantly decreased compared with the scrambled siRNA control (Fig. 1 C and D). In cells with siRNA against InsP3R3, the maximum amplitude was decreased, but the time taken to return to 25% of the maximum amplitude was significantly increased compared with the scrambled siRNA control (Fig. 1 C and D). Previous studies with knockdown of PC2 demonstrated that loss of PC2 channel function decreases both the peak intracellular calcium signaling and the duration of the signal (22, 27, 34, 35). When the cells with a knockdown of either InsP3R subtype were also cotransfected with PC2 [enabling a three- to fivefold overexpression of the protein, as previously determined (35)], the amplitude of the Ca²⁺ transient and the total Ca²⁺ release was restored and there was no significant difference compared with the scrambled control (Fig. 1 C and D). The enhanced amplitude response elicited by PC2 expression is also significantly diminished with the knockdown of either InsP3R1 or InsPR3. These data demonstrate that the knockdown of InsP3R results in lower amplitude Ca²⁺ transients and that overexpression of PC2 can partially restore the Ca²⁺ transient to levels comparable to the scrambled control. However, the overexpression effects of PC2 are significantly diminished upon knockdown of the InsP3R, indicating that the release of calcium either is dependent on a direct interaction of PC2 with InsP3R or requires a certain threshold of calcium to be released by InsP3R before PC2 can release additional calcium. These results also show that there are InsP3R subtype differences in the duration of the Ca²⁺ transient in response to stimulation.

Fig. 1.

Effect of InsP3R knockdown and PC2 overexpression on Ca²⁺ signaling in LLC-PK1 cells. (A) Transient knockdown of InsP3R1 results in diminished Ca²⁺ signaling in response to 100 nM AVP (arrow), compared with scrambled control. Overexpression of PC2 results in a larger transient compared with control, and overexpression of PC2 restores the Ca²⁺ signal after InsP3R1 is knocked down. (B) Same as in A, but showing transient knockdown of InsP3R3 and overexpression of PC2 after InsP3R3 knockdown. The same PC2 overexpression data are shown for comparison in A and B. (C) Quantification of the maximum amplitude of the response. (D) Quantification of the time taken to return to 25% of the maximum response. “n” is representative of cells analyzed (n = 40–103) and is the average from at least four different experiments. Quantified data represent mean ± SEM. *P < 0.05, ***P < 0.001, ****P < 0.0001.

To determine if knockdown of the various subtypes of the InsP3R or PC2 induced 3D cyst growth, we generated shRNA constructs based on the same siRNA sequences used in the experiments described above to generate stable cell lines. Knockdown of each of the three proteins was greater than 60% and comparable to that achieved with transient knockdown (Fig. S1).

Effect of Hormone-Containing Media on 2D Cells.

The LLC-PK1 cell line has not generally been used in 3D tissue models due to its inability to reliably form large numbers of cysts over long periods of time. In one case, isolation of a specific clone was required to facilitate cyst formation in collagen gels (36). Therefore, to use the LLC-PK1 cell line to consistently induce cyst growth in 3D, we used a hormonal media composition previously shown to produce structural growth of immortalized human renal proximal tubule cells and primary human cortical cells in 3D tissues (37, 38). In addition, hormone-supplemented media has historically been used to support LLC-PK1 growth (39). We tested the effect of hormone-containing media on the expression of InsP3R and PC2, proliferation, and morphology of the scrambled control cells at 48 and 96 h after plating. Overall, there was no change in the expression of the InsP3R isoforms or PC2 after 48 h and no obvious effect on cell morphology or ciliation (Fig. S2 A–D). At 96 h, there was a decrease in InsP3R1 expression, but InsP3R3 and PC2 were unchanged (Fig. S2B). However, there was an increase in cell proliferation, as detected by Ki-67 immunoreactivity in cells grown in hormone-containing media (Fig. S2E). We also compared the effect of this hormone-containing media on cell proliferation in the knockdown cell lines in 2D with a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay (Fig. S2F). Similar to the Ki-67 staining, the scrambled cell line showed a significant increase in cell proliferation in the presence of the hormone-containing media compared with the basal media.

We examined the relative resting and endoplasmic reticulum stores of Ca²⁺ in the stable knockdown cell lines, as chronic reduction in Ca²⁺ channels has previously been associated with altered Ca²⁺ stores (40). The resting Ca²⁺ levels were similar in all cell lines, except the InsP3R1 knockdown cells where levels were statistically, but not dramatically, lower than the scrambled control (Fig. S3A). Similarly, the Ca²⁺ load in the endoplasmic reticulum was largely unchanged, as estimated by the addition of ionomycin (Fig. S3B).

Effect of PC2 and InsP3R on Cyst Growth and Formation.

For 3D tissues, cells were embedded in a previously described mixture of Matrigel and type I collagen (37). Cystic structures were monitored for size by both H&E staining and whole-tissue staining of tissues after 2, 4, 6, and 8 wk of culture (Fig. S4). Whole-tissue staining with carmine showed an increase in cyst size in all three knockdown cell lines compared with the scrambled control (Fig. 2).

Fig. 2.

Effect of shRNA knockdown of InsP3R and PC2 on cyst development. Carmine whole-mount staining of cyst development over an 8-wk period. (Scale bar, 200 μm in all images.)

The cysts arising from the InsP3R1 knockdown cells having the most rapid growth outgrew the scrambled control cell cysts by 2 wk and were consistently larger than the scrambled control, the InsP3R3 knockdown, and PC2 knockdown cysts (Fig. 3A). The PC2 knockdown cysts also outgrew the scrambled control by 2 wk, and the InsP3R3 knockdown cysts were significantly larger than the scrambled control by 4 wk. The cysts of all three knockdown lines remained larger than the scrambled control for the remainder of the time course. This relative increase in cyst size was due to both an actual change in cyst size for each knockdown cell line and to the fact that the cysts of the scrambled control did not change in size after 2 wk in 3D culture (Fig. 3B). The cysts arising from the InsP3R1 knockdown cell line cumulatively grew in size across each 2-wk time period (Fig. 3B) until week 6, and there was no significant growth between weeks 6 and 8. The PC2 knockdown cysts did not grow significantly until week 4 and then showed a significant increase in size for the remainder of the time course (Fig. 3B). Throughout the experiment, InsP3R1 knockdown had the most dramatic increase in both cyst size and rate of growth. Taken together, these data indicate that loss of PC2, InsP3R1, or InsP3R3 leads to increased cyst size over time and that a loss of InsP3R1 has the most significant impact upon cyst size.

Fig. 3.

Quantification of cyst size over 8 wk. (A) Comparison of the changes in cyst size among the four cell lines. At week 2, the InsP3R1 knockdown cysts are larger than scrambled. By week 4, the InsP3R1, InsP3R3, and PC2 knockdown cysts are larger than scrambled. By week 4, the InsP3R1 knockdown cysts are larger than the PC2 and InsP3R3 knockdown cysts. Individual symbols represent a single cyst. (B) Comparison of the biweekly size of the cysts. The change in size for the InsP3R1 knockdown cysts is significant between 2 and 4 weeks and significantly larger with every subsequent time point. The change in size for the InsP3R3 knockdown cysts is significant between the 2- to 4-wk and the 4- to 6-wk culture period. The change in size for the PC2 knockdown cysts is significant between the 4- to 6-wk and the 6- to 8-wk culture period. There is no change in size of cysts in the scrambled cell line. Analysis was based upon 95–203 individual cysts from four separately cultured tissues. ****P < 0.0001, ***P < 0.001, **P < 0.01.

This increase in cyst size was confirmed by H&E of the 3D tissues, where the larger cysts showed apparent dead cells and partial hollowing out of the center of the cysts (Fig. 4). Maintenance of the knockdown in 3D culture was confirmed with immunofluorescence microscopy (Fig. S5). InsP3R3 largely colocalized with PC2 in the cytoplasm in scrambled cells, and InsP3R1 was primarily confined to the basolateral surface (Fig. S6). In InsP3R3 knockdown tissues, InsP3R1 redistributed to the cytoplasm of the cell and colocalized with PC2 (Fig. S6). In PC2 knockdown tissues, there was no obvious difference in InsP3R1 or InsP3R3 distribution compared with control, with InsP3R1 distributed on the basolateral surface (Fig. S6). With InsP3R1 knockdown tissues, InsP3R3 and PC2 distribution was unchanged compared with scrambled control (Fig. S6). These data suggest that there is little redistribution under knockdown conditions of InsP3R or PC2 that could contribute to cyst growth.

Fig. 4.

H&E staining of cysts. Increase in cyst size was confirmed by H&E of the 3D tissues. (Scale bar, 50 μm in all images.)

We also correlated cyst growth with an increase in terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL)-positive cell staining over time, confirming that the cells at the center of the cyst were dead (Fig. 5A). Notably, the InsP3R1, InsP3R3, and PC2 knockdown cell lines all demonstrated positive staining with caspase 3, whereas little to no positive staining was observed in the scrambled control (Fig. 5B). These data indicate that the hollowing of the cysts is due to apoptosis-mediated cell death, consistent with previous studies (3, 41). Tissues were also assessed for cell proliferation at 4 and 6 wk. At week 4, no proliferating cells were observed in scrambled control tissues; however, the three knockdown tissues still had proliferating cells, with InsP3R1 knockdown cell lines having the highest number of proliferating cells (Fig. 5C). By week 6, only InsP3R1 knockdown cell lines had proliferating cells (Fig. S7).

Fig. 5.

Cell death and cell proliferation of cysts. (A) The increase in cyst size was correlated with an increase in TUNEL-positive cell staining over time. TUNEL staining is in green; nuclei (DAPI) are in blue. (B) Caspase 3 staining (green) at 4 wk suggests that the cause of death for the cells inside the cysts is due to apoptosis. (C) Staining of proliferating cells with Ki-67 (green) at 4 wk. Note that InsP3R1 knockdown tissues have a higher number of Ki-67–positive cells (proliferating cells in InsP3R3 and PC2 knockdown tissues are denoted by yellow arrow). Nuclei (DAPI) are in blue and actin (phalloidin) is in red. Images are representative of three independent tissues.

Cilia in 3D Cysts.

Primary cilia have been proposed as a link between the development of kidney cysts and the polycystin proteins (42–45). Therefore, we looked at the distribution of cilia in the 3D tissues (Fig. 6). The cilia were detected by immunoreactivity to Arl13B, a protein localized to cilia. After 2 wk in 3D culture, cilia were observed in all cell lines at similar amounts, and they were generally found on the apical side of the cells, facing into the cyst lumen (Fig. 6A). At 6 wk, although cilia were observed in scrambled, InsP3R3, and PC2 knockdown cell lines, in InsP3R1 knockdown cells, a mixed population of cysts containing Arl13B staining and those without cilia were observed (Fig. 6B). In the scrambled, InsP3R3, and PC2 knockdown cell lines, cilia were present up to 8 wk (Fig. 6C). However, by 8 wk, all InsP3R1 knockdown cysts examined were immunonegative for Arl13B staining, indicating a lack of cilia (Fig. 6C). These data suggest that the cysts can continue to develop in the presence or absence of intact cilia and that loss of cilia in the InsP3R1 knockdown cell line is suggestive of a more proliferative phenotype.

Fig. 6.

Cilia are present in cysts with knockdown of InsP3R3 or PC2. (A) Sections from tissue grown for 2 wk in 3D culture were stained for cilia (immunoreactivity to ARL13B, green). Sections were counterstained for calnexin (endoplasmic reticulum, red) and DAPI (nuclei, blue). (B) Cilia were detected in some InP3R1 knockdown cysts, but not in others after 6 wk of 3D culture. (C) After 8 wk of growth in 3D culture, cilia were not observed in the InsP3R1 knockdown cysts, although cilia were present in the other three knockdown cell lines. Images are representative of four independent tissues at all time points.

InsP3R1 and InsP3R3 Are Differentially Localized in Kidney.

To ascertain why knockdown of the two different InsP3R isoforms resulted in different patterns of cyst development, we assessed the relative distribution of both InsP3R1 and InsP3R3 isoforms in mouse and human kidney tissue. In kidney tissue taken from C57/Bl6 mice, InsP3R1 and InsP3R3 were largely excluded from the glomerulus (Fig. S8A, region denoted by “G”). The proximal tubule was identified by cross-reactivity with the Phaseolus vulgaris (PHA-E) lectin (Fig. S8 A and B). InsP3R1 and InsP3R3 were present in the proximal tubule, with InsP3R1 and InsP3R3 expressed throughout the cells (Fig. S8 A and B). As cysts in ADPKD also arise from the collecting duct, these tubules were identified by immunoreactivity to Dolichos biflorus agglutinin (DBA). Similar to the proximal tubule, InsP3R1 and InsP3R3 were present in the collecting duct, although InsP3R1 appeared to be more abundant in the collecting duct than InsP3R3 (Fig. 7 A and B; note arrows in A). In littermate Pkd2^+/− mice (which do not develop kidney cysts), InsP3R1 was less expressed in DBA-positive tubules (note arrows in Fig. 7A). InsP3R3 was decreased throughout the kidney (Fig. 7B).

Fig. 7.

InsP3R1 and InsP3R3 are differentially expressed in Pkd2^+/− mouse kidney and in human kidney. (A) InsP3R1 staining (red) in WT and Pkd2^+/− mouse kidney tissue. DBA denotes the collecting duct. Yellow arrows point to staining in the collecting duct. Note that InsP3R1 staining is lost in the Pkd2^+/− mouse (Lower). (B) InsP3R3 staining (red) in WT (Upper) and Pkd2^+/− (Lower) mouse kidney tissue. “DBA” denotes the collecting duct. Mouse images are representative of three (Pkd2^+/−) and four (WT) mice. (Scale bar, 20 μm.) (C) InsP3R1 (red) and InsP3R3 (blue) staining in human kidney tissue. “DBA” denotes the collecting duct (green). Yellow arrows denote InsP3R3 staining at the basolateral membrane. Human images are representative of three different samples. (Scale bar, 25 μm.)

To confirm that the InsP3R subtypes had a similar distribution in humans, we examined tissue sections taken from three independent nondiseased human samples. In contrast to the mouse, InsP3R1 was distributed throughout the cell, whereas InsP3R3 was predominantly found at the basolateral membrane (Fig. 7C, arrows). The differential distribution of the InsP3R subtypes and the resulting Ca²⁺ signals are known to promote distinct cellular responses (33) and therefore could contribute to specific patterns of cyst growth in ADPKD.

Discussion

The association between cystogenesis and mutations to PC2, a Ca²⁺-dependent Ca²⁺ channel, has led many to suggest that disruption of Ca²⁺ signaling is a factor in cyst formation. Here we provide direct evidence linking the disruption of intracellular Ca²⁺ proteins, InsP3R and PC2, to cyst formation and that the degree of cyst growth depends on the subtype of InsP3R being altered. We demonstrate that LLC-PK1 cells can form cysts in 3D culture when either PC2 or InsP3R is disrupted. We show that intact cilia are found in the cyst structures, but lost when InsP3R1 is knocked down. Finally, we reveal a differential distribution of InsP3R in WT and Pkd2^+/− mice. These data suggest that dysregulation of Ca²⁺ signaling in the endoplasmic reticulum is sufficient to induce cyst development and that cilia are not required to maintain cyst growth.

A number of studies have suggested that disrupted Ca²⁺ signaling as a consequence of PC2 mutations is associated with cyst development (24, 27, 30); however, a clear link between Ca²⁺ channel disruption and cyst development has not been recapitulated in 3D tissues. The majority of PKD studies using 3D tissue culture have used MDCK cells as the cell line of choice (6–12), but PC2-dependent Ca²⁺ signaling has not been studied extensively in this cell line. Here, we demonstrate another cell line, LLC-PK1, in which much is known about PC2-dependent Ca²⁺ signaling, also has the capacity to support cysts when PC2 or InsP3R expression levels are decreased and when cultured in hormone-containing media. This result points to a common denominator of intracellular Ca²⁺ signaling as being key in the development of cysts. Overexpression of PC2 has been associated with decreased cellular proliferation (29), and this result is in line with our finding that loss of PC2 results in cyst formation.

Mechanisms linking the disruption of Ca⁺² signaling to cyst formation remain unclear. A link between ciliary Ca²⁺ signaling, the polycystin proteins, and cyst development has recently gained attention (42, 46–51). The fact that the cilia in the InsP3R3 and PC2 knockdown cysts were still present, but absent in InsP3R1 knockdown cysts, by week 8, suggests that cilia may not be essential to the latter stages of cystogenesis and is broadly consistent with recent findings (52). Our finding that the cilia were no longer observable, and presumably resorbed, when InsP3R1 was knocked down is compatible with the changes associated with the cells undergoing proliferation (see reviews in refs. 53–55). However, this latter finding is inconsistent with other studies demonstrating that loss of cilia can halt cyst development after PC1 or PC2 inhibition (52). The strikingly different rate of cyst growth after loss of InsP3R1 suggests that activation of additional pathways contributing to cystogenesis may not be occurring in the cysts arising from loss of PC2 or InsP3R3.

With regard to the specific pathways for cyst development, our observation that cysts develop in both InsP3R subtype knockdown cells and PC2 knockdown cells suggests that the Ca²⁺ signal from the endoplasmic reticulum is an essential component of cystogenesis, in addition to, or independent from, any Ca²⁺ signal originating from a polycystin complex that may reside in the cilia or plasma membrane. The 3D cell growth system that we have developed will provide a useful tool for investigating both of these scenarios as we have created a 3D tissue culture system that can be easily manipulated and controlled to examine 3D cyst development. Furthermore, these systems can be maintained in vitro for weeks and months, providing a view of slower and more realistic physiological conditions in many cases.

The results shown here suggest that several factors and pathways are likely to contribute to cyst development. For example, it is known that the different isoforms of the InsP3R respond to elevation in cAMP, and hence to PKA phosphorylation, in distinct patterns that potentially fulfill functionally alternate roles (56–58). Additionally, the subcellular distribution of InsP3R and PC2 can differ, providing another aspect that may alter Ca²⁺ signaling in distinct regions of a cell; InsP3R is excluded from the cilia. These differences in subcellular localization may explain, at least in part, the strikingly different cyst growth in InsP3R1 knockdown cells. Examples of specialized distribution, as demonstrated here in the renal tubules, are found in a range of epithelial cells. In MDCK cells, InsP3R3 is found near the tight junctions (59). In hepatocytes, the InsP3R1 is predominantly expressed in the perinuclear and cytosolic environment, whereas InsP3R2 is restricted to regions under the apical membrane (33). In cholangiocytes, InsP3R3 is highly expressed in the apical membrane (60). Finally, our Ca²⁺ imaging data support the idea that the two isoforms confer differing Ca²⁺ signals to the cell. Thus, each of these isoforms of InsP3R have subtle, but unique, patterns of Ca²⁺ transients that can have specific functions in cells (61).

That the disruption of InsP3R results in cyst formation is perhaps surprising, as cyst development has not been reported in the various InsP3R transgenic mouse models (62, 63). However, mice with InsP3R1 knocked out globally suffer from severe ataxia and are malnourished, leading to death around postnatal day 18–20. Moreover, although experiments were conducted on the kidney in InsP3R1 knockout mice, they were limited to the glomeruli, not the tubules, where cysts develop (63). A renal phenotype has not been reported in the global InsP3R3 knockout mouse (64), although explicit studies on the kidney were not conducted. Our study suggests that Pkd2 haploinsufficiency results in a redistribution and decreased expression of InsP3R. Thus, the functional contribution of InsP3R in cyst development cannot be discounted in future studies, where InsP3R and the Pkd genes are selectively knocked down in the kidney. Indeed, the one human gene expression study on ADPKD (with Pkd1 mutations only) reveals an altered expression of InsP3R between cystic and renal-cell carcinoma tissue (65).

In conclusion, we have established that knockdown of PC2, as well as InsP3R, leads to the formation of cysts in an LLC-PK1 3D tissue culture model. These results validate the use of the LLC-PK1 in 3D cell cultures and also demonstrate that Ca²⁺ signaling in the endoplasmic reticulum is a critical signaling pathway in cystogenesis.

Materials and Methods

See SI Materials and Methods for complete information.

LLC-PK1 pig epithelial cells were grown in 2D and 3D matricies as previously described (37). Three-dimensional cultures were stably transfected with shRNA directed against InsP3R1, InsP3R3, PC2, or scrambled control.