To understand the role of the colonic extracellular calcium-sensing receptor (CaSR) in calcium chemoprotection against colon cancer, we activated the CaSR with 5 mM Ca2+ on HT-29 cells, an adenocarcinoma cell line. High Ca2+ stimulated the upregulation (as assessed by RT-PCR) and the secretion of Wnt5a (assessed by Western blot), a noncanonical Wnt family member. Inhibiting CaSR activity with a short interfering RNA (siRNA) duplex against the CaSR reduced CaSR protein and prevented the secretion of Wnt5a. Dominant negative CaSR (R185Q) or siRNA blocked the high Ca2+-mediated inhibition of the β-catenin reporter TOPflash. The CaSR/Wnt5a inhibition of β-catenin reporter was prevented by dominant negative ubiquitin ligase seven in absentia homolog 2 (Siah2). In low-calcium medium, overexpressing Wnt5a increased Siah2 amplicons and protein. Inducing the expression of full-length adenomatous polyposis coli (APC) prevented CaSRmediated increases of Siah2 and Wnt5a. Overexpressing the receptor tyrosine kinase-like orphan receptor 2 (Ror2) increased Wnt5a and CaSR-mediated inhibition of TOPflash. Conditioned medium from Wnt5a-transfected cells added to HT-29 cells in low-Ca2+ medium inhibited the β-catenin reporter. This inhibition was blocked dose responsively by Frizzled-8/Fc chimeric antibody. Overexpression of Ror2 in HT-29 cells in low-Ca2+ medium increased the inhibition of β-catenin reporter caused by recombinant Wnt5a protein compared with addition of Wnt5a protein alone. Our findings demonstrate that APC status plays a key role as a determinant of Wnt5a secretion and suggest that CaSR-mediated secretion of Wnt5a will inhibit defective Wnt signaling in APC-truncated cells in an autocrine manner.
calcium is a chemoprotective agent for colon cancer (3, 21, 29). Epidemiological studies have determined that increasing dietary calcium reduces the recurrence of colonic adenoma and the onset of colon cancer (47). The human extracellular calciumsensing receptor (CaSR), originally cloned from bovine parathyroid, may be activated by small changes (<5%) in millimolar levels of calcium as well as by spermine or aromatic amino acids such as phenylalanine (7). The CaSR is expressed on the apical and basolateral membranes of colonic crypt cells (11, 15, 17). Clearly, some understanding of how calcium works as a chemoprotective agent against colon cancer may be found in the nature of the signals transduced when the CaSR is activated on colonic epithelial cells.
Colon cancer is a disease of defective Wnt signaling (33, 41). Wnt proteins are cysteine-rich glycoproteins that interact with coreceptors low-density lipoprotein receptor-related protein (LRP) 5/6 and Frizzled, a seven-span transmembrane protein. Some Wnts (Wnt1, Wnt3, Wnt3a) stimulate β-catenin signaling by triggering the release of β-catenin from its destruction complex. Released β-catenin associates with DNA-binding factors of the T-cell factor family to activate target genes required for proliferation. Such signaling is canonical. A major protein of the destruction complex is adenomatous polyposis coli (APC). Mutations of APC are a determinant factor leading to intestinal adenomas with constitutive activation of proliferative genes (45). Other Wnts such as Wnt4, Wnt5a, and Wnt11 do not liberate β-catenin but signal noncanonically. Wnt5a may activate PKC (53), calcium-calmodulin kinase 11 (24), or JNK (54) or may induce phosphorylation of Dishevelled (19, 44). Wnt5a interacts with Wnt receptors Frizzled(s)-2 (48), -4 (13), -5 (46), and -7 (52), as well as the receptor tyrosine kinase-like orphan receptor 2 (Ror2) (37). Notably, Wnt5a can also inhibit Wnt3a stimulation of β-catenin signaling by increasing the rate of β-catenin degradation (51).
Increased expression of Wnt5a transcripts have been shown in primary breast cancer, malignant melanoma, thyroid carcinoma, gastric cancer, and colon cancer (26–28, 49, 53), whereas low expression of Wnt5a is associated with high-risk neuroblastoma (4, 5). Elegant studies in breast cancer have demonstrated that increased Wnt5a transcripts were not associated with increases in Wnt5a protein (30). This suggests that increases in Wnt5a protein expression rather than transcript changes may be more indicative of Wnt5a's biological effects. Consistent with this interpretation, recent evidence has shown that Wnt5a protein expression in primary Dukes B colon cancer was associated with a good prognosis (16). Although there are reports that some signaling cascades will upregulate Wnt5a RNA (6, 25), there is no understanding of determinants that lead to changes in Wnt5a protein secretion.
The current experiments were designed to determine if activating the CaSR with increases in extracellular Ca2+ stimulated the secretion of Wnt5a from intestinal epithelial cells. We then determined some functional consequences of CaSR-induced Wnt5a secretion from colon cancer cells.
MATERIALS AND METHODS
HEK-293 and human colon adenocarcinoma HT-29 cell lines were purchased from the American Type Culture Collection (Rockville, MD). The CaR-HEK-293 cells were a generous gift from E. M. Brown (Harvard Medical School, Boston, Massachusetts). The HT-29-APC (APC inducible) and HT-29-βGal (β-galactosidase inducible) were kindly provided by B. Vogelstein (The John Hopkins School of Medicine, Baltimore, MD). HT-29 cells were cultured in DMEM supplemented with 10 or 20% FBS and with 100 μU/ml penicillin and were grown at 37°C in a humidified 5% CO2 atmosphere. The HT-29-APC and HT-29-βGal lines were cultured in McCoy's 5A Medium Modified supplemented with 10% FBS and 100 μU/ml penicillin. Cells were passaged weekly with 0.25% trypsin and were used for experimentation within the first six passages. HEK-293 and CaR-HEK-293 cell lines were cultured in DMEM supplemented with 10% FBS and 100 μU/ml penicillin and were passaged weekly with 0.05% trypsin for no more than six passages.
Transient transfections and luciferase reporter assays.
HT-29 cells were seeded at equal amounts into 24-well tissue culture plates and were grown in DMEM for 24 h. Cells were transfected with plasmid DNAs by using PolyFect transfection reagent (Qiagen, Mississauga, ON) according to manufacturer's recommendations. At least three different experiments were performed in triplicate. The following amounts of plasmids were used per well of a 24-well plate: 1.2 μg TOPflash (Upstate, Chicago, IL) 1.2 μg FOPflash (Upstate), 1.2 μg Wnt5a, 2 μg dominant negative seven in absentia homolog 2 (Siah2), 2 μg wild-type Siah2 (Y. Yang, National Institutes of Health, Bethesda, MD), 1.0 μg dominant negative mutant of the CaSR (R185Q; Dr. E. M. Brown) and 2 ng pRL-null (Promega, Madison, WI). The control plasmid phRL-null was cotransfected in all the experiments to normalize for transfection efficiencies within the experiments. On the basis of preliminary experiments using phRL-CMV, phRL-SV40, phRL-TK, and phRL-null as internal control, we decided to use phRL-null vector because the other vectors showed significant variation in the results. Eighteen hours after transfection, cells were incubated for 8 h in serum-free, Ca2+-free DMEM containing 4 mM l-glutamine, 0.2% BSA, 100 μU/ml penicillin, and 0.005 mM CaCl2. This medium was removed and was substituted with the same medium supplemented with 5 mM extracellular Ca2+ for another 18 h. Cells were harvested by lysing in 100 μl of reporter lysis buffer (Promega). HT-29-APC and HT-29-β-Gal were transfected by using essentially the same protocol and were cultured in the presence of 100 μM zinc for up to 18 h. Luciferase activity was measured using a Lumat LB9507 luminometer (Berthold Technologies, Bad Wildbad, Germany).
Preparation of Wnt5a-conditioned media.
HEK-293 cells were seeded into six-well tissue culture plates the day before transfection. Superfect transfection reagent (Qiagen) was used to transfect 1.2 μg of Wnt5a plasmid per well. The cells were grown in regular medium for 24 h after transfection, and this medium was substituted with DMEM containing 0.005 mM extracellular Ca2+ for 48 h. At this point, the medium was collected, filtered, sterilized, and stored at −70°C for later use.
Differential gene-expression analysis was performed by semiquantitative PCR by using a Mastercycler (Eppendorf, Hamburg, Germany). Total RNA from cells was isolated by the Trizol reagent by following manufacturer's instructions. The recovered RNA was quantitated by spectrophotometry, and aliquots of 1 μg total RNA from low-Ca2+ (0.005 mM) or high-Ca2+ (5 mM) samples were used for cDNA synthesis at 37°C for 1 h by using an Omniscript RT kit (Qiagen, Valencia, CA). The PCR reaction and cycle conditions have been described previously (40). Primers for CaSR were sense 5′-CGG GGT ACC TTA AGC ACC TAC GGC ATC TAA-3′ and antisense 5′-GCT CTA GAG TTA ACG CGA TCC CAA AGG GCT C-3′; the sequences used for Siah1 were sense 5′-CGG AAT TCA CCA TGC CCT GTA AAT ATG CGT CTT CTG-3′ and antisense 5′-CCG CTC GAG TCA ACA CAT GGA AAT AGT TAC ATT-3′; the sequences used for Siah2 were sense 5′-GCT AAT AAA CCC TGC AGC AA-3′ and antisense 5′-ACT TCT GGC GGC ATT GGT TA-3′; the sequences for Wnt5a were sense 5′-AAC CGG AAC CAT TTT TTT TC-3′ and antisense 5′-TCT TGT ATT ACC TTT TCA AAG ATC C-3′. The resulting bands were visualized on a 1.0% agarose gel stained with ethidium bromide and were compared with a 100-bp DNA ladder (Invitrogen) to confirm the predicted size. For positive amplification control we used the GAPDH gene, because this gene reveals no pseudogene organization, a problem that hampers the exact interpretation of PCR results at the cDNA level.
Western blot analysis.
Early-passage HT-29 cells were grown on six-well dishes. Cells were incubated for 18 h in serum-free, Ca2+-free DMEM containing 4 mM l-glutamine, 0.2% BSA, 100 μU/ml penicillin, and 0.005 mM CaCl2. This medium was removed and was substituted with the same medium supplemented with 5 mM Ca2+. At the end of the incubation period (18 h), conditioned medium was harvested and concentrated to 16× by Amicon Ultra filter devices (Millipore, Carrigtwohill, Ireland). Cells were washed once with ice-cold PBS containing 1 mM sodium vanadate and 25 mM NaF, and then 100 μl of ice-cold lysis buffer was added (20 mM Tris·HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 25 mM NaF, 1% Triton X-100, 10% glycerol, 1 mM dithiothreitol, 1 mM sodium vanadate, 50 mM glycerophosphate, and a cocktail of protease inhibitors). After being sonicated for 5 s, lysates were centrifuged at 10,000 g for 10 min at 4°C and were processed as described (30, 32). Equal amounts of protein from whole cellular lysates and conditioned medium were resolved on a gradient gel of 4–20% and were electrophoretically transferred to Immobilon membranes. Wnt5a was detected by using a 1:300 dilution of primary antibody in 1× PBS and 0.1% Triton X-100 with 1% BSA. Blots were washed for three 10-min periods at room temperature (1× PBS, 0.1% Triton X-100) and then were incubated for 1 h with a secondary antibody conjugated to horseradish peroxidase (1:2,000) in blocking solution. Blots were then washed a second time (3 × 10 min). For Siah2 determinations, HT-29-APC or HT-29-β-Gal cells were supplemented with 100 μM ZnCl2 for 18 h to induce the expression of full-length APC and β-Gal genes, respectively. Siah2 was detected by overnight incubation with a 1:500 dilution of mouse monoclonal antibodies against Siah2 (Sigma, St. Louis, MO). Whole cellular β-catenin levels were measured by using a rabbit polyclonal antibody (1:1,000; Santa Cruz Biotechnology, Santa Cruz, CA). Bands were visualized by chemiluminescence according to the manufacturer's instructions (SuperSignal; Pierce Chemical).
Data are presented as means ± SD of at least three separate experiments. Data were analyzed by Student's t-test or ANOVA when appropriate. P < 0.05 was considered a statistically significant difference.
CaSR activation increases Wnt5a secretion.
Earlier studies had reported that exposure to extracellular Ca2+ (1 mM) of some colonic cancer cell lines (CBS, Moser, SW480) resulted in decreases in β-catenin signaling together with increases in E-cadherin expression, suggesting that extracellular Ca2+ mediated these events (9, 10). For the current experiments, we used HT-29 adenocarcinoma cells, which have wild-type β-catenin and a truncated APC (20, 23). Rather than challenge HT-29 cells with nominally Ca2+-free conditions, we first screened various levels of extracellular Ca2+ to determine where the β-catenin reporter and proliferative rate were the highest (39).
Subconfluent HT-29 adenocarcinoma cells were then challenged with high Ca2+ (5 mM) and were compared with low Ca2+ (0.005 mM) for 18 h. RT-PCR was performed by using primers for human Wnt3, Wnt3a, Wnt5a, and Wnt7. Only transcripts for Wnt5a were consistently and reproducibly altered with the high-Ca2+ challenge (Fig. 1A). We therefore focused our experiments on whether CaSR activation stimulated secretion of Wnt5a protein from this cell type. CaSR activation (5 mM Ca2+) caused secretion of Wnt5a into the medium of HT-29 cells (Fig. 1B). Although Wnt5a protein was detected in lysates of these cells in low Ca2+ (0.005 mM), no secreted Wnt5a protein was detected in the medium. In contrast, CaSR activation substantially increased the immunoreactive Wnt5a in the lysates and in the medium. We then used a short interfering RNA (siRNA) duplex against the CaSR and observed substantially reduced CaSR transcript and protein (Fig. 1C). Cells treated with siRNA duplex against the CaSR showed substantially reduced Wnt5a in the conditioned medium (Fig. 1B). As a positive control, we transfected cells with Wnt5a, then collected and fractionated the medium and probed for Wnt5a. There were both increases in Wnt5a in the lysate and a sharp immunoreactive band in the conditioned medium. We then assessed whether activation of other G protein-coupled receptors in these cells would stimulate Wnt5a secretion. Challenge by endothelin (1 nM), carbachol (1 μM), ADP (1 μM), or angiotensin (1 μM) for 18 or 24 h did not stimulate Wnt5a secretion (data not shown). In summary, CaSR activation increased Wnt5a protein in cell lysates that was secreted into the medium.
Functional correlates of CaSR-mediated Wnt5a secretion.
To understand the role of CaSR-mediated Wnt5a secretion in this adenocarcinoma cell line, we measured the Tcf/LEF reporter TOP/FOPflash activities (Fig. 2). We observed that after 18 h of exposure to 5 mM Ca2+, the TOPflash activity was reduced 78 ± 5% (P < 0.05; n = 6) compared with cells challenged with 0.005 mM Ca2+ (Fig. 2A). There was no difference in the amount of FOPflash activity between these two groups (12 ± 2 vs. 9 ± 4% of 0.005 mM Ca2+ TOPflash activity). Transfecting the HT-29 cells with Wnt5a caused a reduction in reporter activity of 70 ± 4% (P < 0.05; n = 8). There was no difference in the reduction of TOPflash caused by extracellular Ca2+ compared with Wnt5a transfection. Transfection of a dominant negative CaSR (R185Q) prevented the inhibition of TOPflash caused by 5 mM Ca2+ (96 ± 7 vs. 21 ± 6% of 0.005 mM Ca2+ activity; P < 0.05; n = 6; Fig. 2A). We then compared the effect of transient transfection of a siRNA duplex against the CaSR with cells treated with a scrambled siRNA duplex on TOPflash activity (Fig. 2B). Treatment with the siRNA designed against the CaSR prevented the reduction of TOPflash activity observed with 5 mM Ca2+ of scrambled siRNA duplex-transfected cells (93 ± 11 vs. 34 ± 6% of 0.005 mM Ca2+ activity; P < 0.05; n = 5; Fig. 2B). Because CaSR activity [prevented with either the dominant negative construct (Fig. 2A) or with siRNA against the CaSR (Fig. 2B)] reduced the β-catenin reporter comparably with the amount seen after Wnt5a transfection, we measured β-catenin levels in lysates of cells treated with either CaSR activation by 5 mM Ca2+ or following Wnt5a transfection in low-Ca2+ medium (Fig. 2C). Consistent with the reduction in β-catenin reporter activity, Western blot analysis demonstrated reduced total cellular β-catenin after 18-h challenge with 5 mM Ca2+ or after Wnt5a transfection in low-Ca2+ medium (Fig. 2C). Together, these results suggested to us that silencing the CaSR substantially reduced the amount of secreted Wnt5a and prevented the calcium-stimulated reduction of the β-catenin reporter activity.
Wnt5a mediates increases in Siah2, which contributes to Tcf-reporter decreases.
Increases in β-catenin degradation may be mediated by the E-type ubiquitin ligases Siah1 and Siah2 (51, 32, 35). Wnt5a has been shown to prevent Wnt3a-stimulated increases in TOPflash reporter activities by increasing Siah2 in heterologous transfected cells (51). We therefore assessed the effect of a dominant negative Siah2 on the CaSR-mediated inhibition of TOPflash. We found that transient transfection with dominant negative Siah2 inhibited the Ca2+-stimulated inhibition of reporter in HT-29 cells (Fig. 3A), whereas wild-type Siah2 did not increase the amount of inhibition caused by Wnt5a transfection. Because others (51) have reported that Wnt5a required intact APC to increase Siah2, we first used an engineered HT-29-APC cell line that had an integrated full-length APC transgene under the control of a zinc-inducible metallothionein promoter (38). As our control for these experiments, we used HT-29-β-Gal cells, which contained a β-galactosidase transgene expressed on the same zinc-inducible metallothionein promoter (38). Both cell types expressed the CaSR equivalently (data not shown). After Zn2+ treatment, we challenged cells with either low or high Ca2+ (0.005 vs. 5 mM) for 18 h, then performed RT-PCR (Fig. 3B). High calcium increased transcripts for Siah2 in both full-length APC and truncated APC cells and decreased transcripts for Siah1 in both cell types. Western blot analysis of Siah2 revealed that the high-Ca2+ (5 mM) challenge increased Siah2 protein only in the truncated APC cells (Fig. 3C). We also observed that Wnt5a transcripts increased with high-Ca2+ treatment in the truncated APC cells but not in cells overexpressing full-length APC (Fig. 3C). We screened conditioned medium from the Zn2+-induced full-length APC-expressing cells as well as two other cell lines with wild-type APC (SW48 and HCT116) and found no Wnt5a secreted after CaSR activation. When the Zn2+-induced vector-control HT-29 cells were transiently transfected with Wnt5a, RT-PCR revealed increases of Siah2 amplicons in low-Ca2+ medium (Fig. 3D). These experiments suggest that CaSR activation of HT-29 cells with wild-type APC overexpression do not show upregulation and secretion of Wnt5a. Increases in extracellular Ca2+ increased the Siah2 transcript in these cells, but there were no increases in Siah2 protein. Because Wnt5a transfection alone in low Ca2+ increased Siah2 transcripts, our current results suggest that it is the secreted Wnt5a that works in an autocrine manner on intestinal cells to inhibit β-catenin signaling by increasing Siah2. However, our data strongly suggest that APC must be truncated for Wnt5a to be secreted from these cells.
Secreted Wnt5a works in an autocrine manner to inhibit β-catenin.
We have shown that Wnt5a-transfected cells secrete Wnt5a (Fig. 1B); we then determined if this Wnt5a was biologically active by adding conditioned medium from Wnt5a-transfected (HEK-293) cells to the HT-29 in low-Ca2+ medium (0.005 mM) and chelating the Wnt5a with Frizzled-8/Fc chimeric antibody (Fig. 4). As before, Wnt5a-transfected HT-29 cells had a 72 ± 5% reduction in reporter activity compared with cells in 0.005 mM Ca2+. Conditioned medium added to nontransfected cells for 18 h showed a 59 ± 5% reduction in reporter activity that a 1:50 dilution of Frizzled-8/Fc increased to 89 ± 6% of the activity in 0.005 mM Ca2+ (P < 0.05; n = 5). A 1:100 dilution of Frizzled-8/Fc had a proportionately less effect on preventing the inhibition of the β-catenin reporter (53 ± 9% of activity in 0.005 mM Ca2+). Addition of a comparable amount of IgG (0.5 μg/ml) had no effect on the conditioned medium-stimulated reduction of reporter activity.
Effect of overexpression of wild-type Ror2 and dominant negative Ror2 on Wnt5a and CaSR inhibition of TOPflash.
Because the orphan tyrosine kinase receptor Ror2 has been shown to mediate Wnt5a-stimulated signaling, we performed further experiments to determine whether Ror2 was involved in the CaSR/Wnt5a inhibition of the β-catenin reporter. We first compared the effect of wild-type Ror2 overexpression on Wnt5a-mediated inhibition of TOPflash in HT-29 cells. As illustrated in Fig. 5A, HT-29 cells transfected with Wnt5a in these experiments had ∼75% reduction in TOPflash activity (30 ± 10% activity of 0.005 mM Ca2+; n = 5). Cells transfected with wild-type Ror2, together with Wnt5a, showed a further reduction in reporter activity (16 ± 4 vs. 30 ± 10% activity of 0.005 mM Ca2+; n = 5). These cells were transfected with a membrane-tethered variant of Ror2 (Ror2-TM) previously shown to work as a dominant negative (36). Transfection with Ror2-TM was sufficient to inhibit the ability of wild-type Ror2 to enhance the Wnt5a-mediated inhibition of TOPflash (37 ± 14 vs. 16 ± 4% activity of 0.005 mM Ca2+; n = 5; Fig. 5A). When the CaSR was activated, in the presence of overexpressed Ror2, there was a further increase in the amount of TOPflash inhibition compared with 5 mM Ca2+ alone (10 ± 3 vs. 29 ± 6% activity of 0.005 mM Ca2+; n = 5; Fig. 5A). We then confirmed that Ror2 was overexpressed by measuring Ror2 expression in the transfected cell lysates (Fig. 5B).
We then added recombinant human (rh) Wnt5a protein (200 ng/ml) to the HT-29 cells in low-Ca2+ medium with and without Ror2 overexpression (Fig. 5C). Addition of rhWnt5a protein reduced the activity of the β-catenin reporter (65 ± 8% of activity in 0.005 mM Ca2+; P < 0.05; n = 6). Overexpression of Ror2 increased the inhibition of β-catenin reporter caused by addition of rhWnt5a protein (36 ± 11% of activity in 0.005 mM Ca2+; P < 0.05; n = 6). Together, these results suggest that Wnt5a can interact with Ror2 to inhibit β-catenin signaling in colon cancer cells and that manipulating wild-type Ror2 expression increased the ability of Wnt5a and CaSR to inhibit the β-catenin reporter.
The current experiments demonstrate that activating the CaSR on an adenocarcinoma cell line stimulated the secretion of Wnt5a. Inhibiting CaSR activity with siRNA reduced CaSR protein and prevented the secretion of Wnt5a. Inhibiting CaSR activity with either siRNA or a dominant negative CaSR blocked the calcium-mediated inhibition of β-catenin reporter. The CaSR/Wnt5a inhibition of β-catenin reporter activity was prevented by dominant negative Siah2. The activation of the CaSR on cells with a truncated APC, but not full-length APC, increased Siah2 transcripts and protein. Overexpression of the orphan tyrosine kinase Ror2 increased the Wnt5a and CaSR-mediated inhibition of β-catenin signaling in these cells. The CaSR-mediated Wnt5a secretion also required truncated APC, because overexpressing wild-type APC in these cells prevented the CaSR-mediated synthesis and secretion of Wnt5a. Little is known about determinants of the expression and secretion of Wnt5a (29). The current experiments are the first demonstration that Wnt5a synthesis and secretion may be regulated by the activity of a G protein-coupled receptor.
Expression of Wnt5a protein has been shown to be associated with longer survival in primary Dukes B colon cancer (16). Expression of Wnt5a is a predictor of longer disease-free survival in human breast cancer (26, 30). Studies have shown that Wnt5a mRNA may be increased by phorbol esters in breast carcinoma (25) and that TNF-α will increase mRNA in macrophages (6), but little in known about the regulation of Wnt5a protein expression in the intestine. Wnt5a transcripts have been localized to the mesenchyme of normal small intestinal villi, with abundance at the villus tips and in the colon. Wnt5a transcripts are found in the mesenchyme beneath the surface epithelium (18). Wnt5a protein expression has been shown in normal human colon epithelia at the base of the crypts (16). In situ hybridization studies have shown strong expression of Wnt5a mRNA in normal mucosa and in colorectal cancer cell lines (49). It is not clear that this reflects the expression of Wnt5a protein in these cells. Recent studies of breast carcinomas lacking Wnt5a protein still had detectable and elevated mRNA levels, consistent with the idea that in breast epithelia, Wnt5a expression was regulated at a posttranscriptional level (30). Therefore, we focused our experiments on Wnt5a protein. Secreted Wnt5a has been shown to regulate dopaminergic neurogenesis (8), whereas tetanic stimulation leads to N-methyl-d-aspartate receptor-dependant synaptic Wnt3a release during hippocampal long-term potentiation (14). However, little is understood about the proximal signals determining Wnt secretion (37). The current experiments demonstrate that stimulating the CaSR increased the secretion of Wnt5a protein from HT-29 adenocarcinoma cells. Furthermore, our current studies suggest a novel relationship between APC status and regulated Wnt5a secretion from intestinal epithelia.
Our evidence that activating the CaSR mediated the upregulation and secretion of Wnt5a came from two sets of experiments. Downregulating CaSR expression by siRNA reduced CaSR protein, prevented the CaSR-mediated secretion of Wnt5a, and reversed the calcium-mediated inhibition of the β-catenin reporter. Interfering RNA against the CaSR has been shown to block calcium-stimulated secretion of bone morphogenic protein-2 from colonic myofibroblasts (40) and functioning of the CaSR in aortic endothelia (56). Transient transfection experiments using R185Q, a dominant negative CaSR, reversed the high Ca2+-mediated inhibition of the β-catenin reporter. Because R185Q will generate the formation of heterodimers of CaSR to produce a rightward shift in the EC50 for extracellular Ca2+ (2, 34, 40, 42), our current results strongly suggest that CaSR mediates the effect of the Ca2+ challenge. We believe we are reporting an autocrine effect of CaSR-mediated Wnt5a secretion, because when conditioned medium from Wnt5a-transfected HEK-293 cells was added to the HT-29 cells, a reduction in the β-catenin reporter ensued, which was comparable with the decrease observed if the cells were secreting Wnt5a or following the addition of recombinant Wnt5a protein. This decrease in reporter activity was prevented by anti-Frizzled-8 antibody, an established chelator of Wnt protein (22), consistent with the interpretation that the secreted Wnt5a was biologically active.
Insight into how the CaSR and Wnt5a inhibited β-catenin signaling in these cells was provided by our finding that a dominant negative E3 ubiquitin ligase, Siah2, reversed the CaSR-mediated inhibition of TOPflash. Others have shown that Wnt5a will inhibit canonical Wnt3a signaling by degrading β-catenin by upregulating Siah2 independent of nuclear factor of activated T cells and calmodulin kinase II (51). Our current results extend these findings by showing that only when APC was truncated did Wnt5a induce Siah2 transcript and protein. Because full-length APC prevented CaSR-activated cells from upregulating Wnt5a transcripts, whereas high-Ca2+ challenge of these cells increased Siah2 amplicons, our current results suggest that the secreted Wnt5a must act together with the CaSR to increase Siah2 protein for inhibition of β-catenin signaling. The mechanisms by which the CaSR or Wnt5a will increase Siah2 transcript and protein are not currently known.
CaSR activation of the truncated APC colon cancer cells stimulated the synthesis and secretion of Wnt5a. How does this relationship differ from other types of cancer where the CaSR activation stimulates rather than inhibits proliferation? CaSR activation of breast cancer cells (43) and Leydig cancer cells (50) generates parathyroid hormone-related protein secretion and proliferation. Activating the CaSR on prostate cancer cells also stimulates parathyroid hormone-related protein secretion (55) and cell proliferation (31). Indeed, Ca2+ activation of the CaSR on rat calvarial osteoblasts stimulates substantial upregulation of cyclin D genes and JNK-dependent proliferation (12). We have observed that CaSR activation of human mesenchymal stem cells as well as CaSR-expressing hypothalamic neurons causes the synthesis and secretion of the canonical Wnt3a (I. Pacheco and R. J. MacLeod, unpublished observations). These results allow us to speculate that when extracellular calcium is a mediator of prostate cancer skeletal metastasis, a critical determinant for the subsequent proliferative response may be the CaSR's production of a canonical Wnt family member that then works in an autocrine manner (1). In contrast, we speculate that in breast cancer the differences in metastatic potential of cells will be strongly related to whether the CaSR stimulates both Wnt5a secretion and Ror2 production, i.e., both increases in Wnt5a and Ror2 expression are required to diminish migration. Clearly, knowledge of the determinants of both the Wnt family member that is secreted as well as how the CaSR changes the “receptor context” of these cells will be important for therapeutic considerations.
The distinction of Wnt protein signaling into canonical and noncanonical varieties has recently been clarified by the demonstration that Wnt5a protein will signal canonically or noncanonically depending on receptor context (37, 51). Wnt5a has previously been reported to stimulate a variety of second-messenger cascades (19, 24, 44, 53, 54). The recent demonstration that the orphan tyrosine kinase Ror2 was required for Wnt5a to inhibit β-catenin stabilization prompted us to determine if Ror2 was involved in the intestinal CaSR-mediated Wnt5a inhibition of β-catenin signaling. Our current results confirm that overexpression of wild-type Ror2 increased Wnt5a inhibition of β-catenin signaling. Consistent with other reports (51), we found that deletions in the intracellular domain of Ror2 acted in a dominant negative fashion against overexpressed Ror2 and Wnt5a. Our current results showing that the CaSR-mediated Wnt5a inhibition of β-catenin signaling was increased by overexpressing Ror2 strongly suggest that Ror2 is involved in the intestinal epithelial response to secreted Wnt5a.
The current study has not addressed whether CaSR activation of truncated APC colon cancer cells influences other secreted Wnt antagonists such as Dickkopf-1 or its receptor LRP5/6. However, in parallel studies, we have discovered that CaSR activation of the stromal, subepithelial colonic myofibroblast stimulates the synthesis and secretion of both Wnt5a and Dickkopf-1, whereas CaSR activation of intestinal epithelia, irrespective of the status of APC, upregulates transcript and protein of Ror2 and LRP6, the respective receptors of these Wnt antagonists (R. J. MacLeod and I. Pacheco, unpublished observations). Our studies favor the interpretation that CaSR activation changes the “receptor context” of intestinal epithelia (Frizzled-to-Ror2) when Wnt5a has been secreted by CaSR activation of subepithelial myofibroblasts. However, without the secretion of Wnt5a (because the epithelia express full-length APC), it is possible that CaSR activation of intestinal epithelia changes receptor context, favoring noncanonical Wnt5a signaling. Additional studies are required to test this hypothesis.
Our current findings are summarized in the schematic illustrated as Fig. 6. In colonic epithelial cancer cells with a truncated APC, activation of the CaSR by ionized Ca2+ resulted in the upregulation of Wnt5a transcripts and increases in Wnt5a protein. Activation of the CaSR by Ca2+ stimulated the secretion of Wnt5a. The secreted Wnt5a then activated in an autocrine fashion either the orphan tyrosine kinase Ror2 or an uncharacterized Frizzled receptor to increase transcripts and protein of the E-ubiquitin ligase, Siah2. Increases in β-catenin degradation mediated by Siah2 reduced β-catenin signaling. When full-length APC was overexpressed in these cells, Ca2+ stimulation of the CaSR did not upregulate Wnt5a or increase Siah2 protein. How full-length APC prevents the CaSR-mediated upregulation of Wnt5a is not yet understood.
In summary, our findings suggest that Wnt5a secretion stimulated by CaSR activation may inhibit defective Wnt signaling in an autocrine manner. Epithelial CaSR activation (by Ca2+ supplementation) after APC mutations may be an important determinant of calcium's chemoprotective effect against colon cancer.
This work was funded by operating grants from the Canadian Institutes of Health Research, Crohns and Colitis Foundation of Canada, and the Dairy Farmers of Canada to R. J. MacLeod. It was also supported by the Canadian Foundation for Innovation and the Canada Research Chairs Program. Support is also acknowledged from the Gastrointestinal Diseases Research Unit/Canadian Institutes of Health Research training grant and a fellowship from Canadian Institutes of Health Research/Ferring/Canadian Association of Gastroenterology to I. Pacheco.
We thank R. Nusse for Ror2-TM, Y. Minami for wild-type Ror2, Y. Yang for wild-type and dominant negative Siah2, and E. M. Brown for R185Q plasmids, respectively. We also acknowledge the technical assistance of Craig Spencer.
R. J. MacLeod is the Canada Research Chair (Tier 2) in Gastrointestinal Cell Physiology.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society