Calcineurin A–Binding Protein, a Novel Modulator of the Calcineurin-Nuclear Factor of Activated T-Cell Signaling Pathway, Is Overexpressed in Wilms' Tumors and Promotes Cell Migration

Wilms' tumors (WT) represent ∼85% of childhood kidney cancer. It is believed to arise from nephrogenic rests resulting from a failure of mesenchymal progenitor cells to differentiate into nephrons. This concept is supported by both histologic and molecular data as the tumors typically contain all three cell types derived from mesenchymal progenitor cells (blastemal, epithelial, and stromal; ref. 1) and present a molecular signature consistent with an early kidney differentiation program (2). In this respect, WTs are a classic example of cancer resulting from developmental anomalies.

WTs can be treated with relatively high efficiency. However, 15% to 20% of WT patients experience relapse associated with resistance to treatments and poor prognosis (3, 4). An increased probability of relapse and mortality correlates with anaplasia and metastasis (5, 6) as well as increased chromosomal anomalies (7, 8). At the molecular level, relapse has been linked to consistent deregulation of a limited number of genes, including components of the calcineurin signaling pathway (9).

Evidence for both loss-of-function and gain-of-function mutations has been associated with WT development. Wilms' Tumor gene 1 (WT1) was the first gene identified with genomic alterations or point mutations (10-13) and accounts for 10% to 15% of WTs (14, 15). Consistently with the developmental origin of WTs, WT1 is expressed in the mesenchymal progenitor cells (metanephric mesenchyme) and is required for its normal differentiation into nephrons (16-18).

The transcription factor PAX2 is ectopically expressed in most WT samples. Initially described for a few tumors (19, 20), the consistent expression of PAX2 and its close family member PAX8 was subsequently observed in several other primary tumor samples (2, 21). A large-scale gene expression analysis further identified PAX2 as one of the signature genes for WTs (2).

Among the data supporting a functional link between PAX2 and WT development is the close connection between the murine Pax genes and Wt1, both at the transcriptional level and in terms of cellular functions such as mesenchymal-epithelial transition and cell survival. During the development of the definitive kidney (metanephros), Pax2 is up-regulated concomitantly with Wt1 expression in the mesenchymal progenitor cells before the epithelialization of the forming nephron (22). In addition, both genes seem to regulate each other at the transcriptional level (22, 23). As Wt1, Pax2 plays a crucial role in the initiation of metanephric kidney induction and nephron formation in the embryo (24). Later during development, Pax2 is involved in nephron maturation (25) and is shut down in terminally differentiated nephrons.

Expression of PAX genes in cancer correlates with reduced apoptosis as well as increased proliferative and invasive phenotypes (26-29). Inhibition of PAX2 in renal cell lines reduces proliferation as well as resistance to cisplatin treatment (26, 28). Analyses of the Pax2^+/− kidney phenotype also support a role for Pax2 in cell survival (30-33). Expression of PAX2 correlates with increased migration of Kaposi sarcoma cells and endothelial cells isolated from renal tumors (29, 34). Despite the association of PAX2 and other PAX transcription factors to cancer, little is known about the downstream pathways they regulate.

The consistent misexpression of PAX2 in WTs suggests a role in tumor initiation and/or progression (35). In a screen designed to isolate Pax2 target genes involved in normal and abnormal kidney development, we identified a gene coexpressed with Pax2 in the mesenchymal progenitor cells at the onset of metanephric kidney development. This gene encodes a membrane-anchored protein, which we named calcineurin A–binding protein (CnABP). Like PAX2, CnABP is overexpressed in most WT samples analyzed. Complementary gain-of-function and loss-of-function experiments in cell culture support a role for CnABP in promoting cell proliferation and migration. We finally show that CnABP binds and inhibits calcineurin Aβ, a central component of the calcium signaling pathway identified as a signature pathway of recurrent WTs.

In an effort to find new regulators of kidney development and tumorigenesis, we did a series of microarray experiments comparing (a) wild-type with Pax2 mutant renal cells, (b) early with later stages of mouse mesonephros (embryonic kidney) development, and (c) Pax2-expressing renal cells with surrounding Pax2-negative tissues (36, 37). Because Pax genes were found as signature markers for WTs and were previously shown to regulate the tumor suppressor gene Wt1, we examined our microarray results for genes potentially involved in WT. This was done by mapping differentially regulated expressed sequence tags (EST) to their chromosomal position in the mouse and comparing their syntenic position in human with known putative WT loci (e.g., 11p15, 1p35, 17q, and 16q; ref. 38). As expected, we identified Wt1, thereby validating the microarray approach used. Interestingly, we also identified an EST (AK076656) that showed 2.1-fold down-regulation in the absence of Pax2, 3.1-fold higher expression in the kidney than in surrounding tissues, and a 3.6-fold up-regulation in time between early and late mesonephros development (Table 1). Using the human and mouse UniGene databases, we found that EST AK076656 corresponds to a novel gene, which we name CnABP (see below), located on chromosome 16q in human.

To validate our microarray data, we mapped various potential PAX2-binding sites (PBS) in the CnABP locus based on promoter proximity and sequence conservation. Three putative sites were cloned upstream of a luciferase reporter construct and tested for induction by PAX2 (Fig. 1A). Of the three sites tested, only PBS2 exhibited a significant (P < 0.001) induction by PAX2 (Fig. 1B). We next introduced point mutations in this site to show the activity of PAX2 on this putative binding site. As expected, PBS2 mutation significantly (P < 0.01) reduced the activity of this enhancer element, although residual activity remained. Next, we wanted to validate whether the expression pattern of CnABP was compatible with Pax2 expression at different stages of mouse kidney development. In situ hybridization analysis at 12 days of embryonic development (E12.5) detected CnABP expression in the ureter as well as in the nephrogenic mesenchyme surrounding the ureter tip (Fig. 1D). This expression pattern was maintained until the end of embryonic development (E18.5) and was extended to differentiating nephrons (Fig. 1E). Importantly, the expression pattern of CnABP in the developing kidney overlaps with that of Pax2 (Fig. 1F and G). Hence, CnABP is regulated by Pax2 and expressed in the developing kidney, notably in the nephrogenic mesenchyme, believed to give rise to WTs.

According to the human UniGene database, CnABP is expressed as two major isoforms. Isoform-a is composed of three exons and represent ∼80% of CnABP EST sequences in human. Isoform-b extends and terminates exon 2 within intron 2 of isoform-a (Fig. 2A). This isoform, which corresponds to the mouse Riken clone 4732415M23Rik, was present in <20% of human UniGene ESTs. PCR on reverse-transcribed cDNA of WT and normal kidney samples revealed the expression of isoform-a, whereas isoform-b could not be detected (Fig. 2B). As it is the most prominent isoform, only isoform-a was studied further (Fig. 2B). CnABP isoform-a transcript is 2.2 kb long and encodes for a predicted 15.5-kDa protein. To identify paralog sequences for CnABP in mouse or human, we did blast analysis in National Center for Biotechnology Information databases. This search failed to reveal any paralog to CnABP in these species. In contrast, Blast analysis against vertebrate amino acid sequences revealed highly conserved domains between CnABP orthologs in mammals (Fig. 2C). The predicted CnABP protein could not be found in fishes or lower vertebrates. Although none of the conserved domains has been previously annotated, an NH₂-terminal myristoylation consensus site was found. Transfection experiments with green fluorescent protein (GFP)–tagged CnABP confirmed that CnABP association with the cell membrane depends on the integrity of this consensus sequence (Supplementary Fig. S1C and D). In addition, a consensus calcineurin docking site (PxIxIT) was identified in the most conserved region of CnABP (Fig. 2C).

The overlapping expression of CnABP and Pax2 in the developing kidney led us to hypothesize that, like Pax2, CnABP misexpression may be associated with WT. To test this, we assessed CnABP and PAX2 expression levels in 39 primary WTs relative to normal kidney tissue using quantitative reverse transcription-PCR. In line with previous reports, PAX2 was found overexpressed in 74% (29 of 39) of these samples (Fig. 3A). Strikingly, CnABP expression increased >2-fold the normal levels in 74% of WT samples analyzed (Fig. 3A). As much as 62% of the CnABP overexpressing tumors in fact showed >5-fold the normal expression levels. Of the tumors that overexpressed PAX2, 83% (24 of 29) also overexpressed CnABP. These results establish an association between elevated CnABP and Pax2 expression in WTs. We next assessed expression levels of PAX2 and CnABP in tumors with specific loss of heterozygosity (LOH). No association was observed in PAX2 expression level with respect to LOH status (Fig. 3B). However, a high proportion (100%, n = 10) of tumors with LOH at 11p exhibit elevated levels of CnABP compared with other tumors (Fig. 3C). Hence, CnABP is highly expressed in association with PAX2 in a majority of WTs.

To assess whether CnABP has an effect on cell proliferation, we overexpressed CnABP in the human embryonic kidney cell line HEK293 and in the WT cell line SKNEP-1 (Supplementary Fig. S2A). HEK293 cells transfected with a CnABP expression vector showed higher proliferation rates than cells transfected with vector control (Fig. 4B). After 96 hours in culture, CnABP-transfected HEK293 cells had a 50% increase in cell number compared with vector-transfected cells (P < 0.01; Fig. 4B). An equivalent 50% increase was also observed in CnABP-transfected SKNEP-1 cells compared with control cells at 96 hours (P < 0.01; Fig. 4C). To validate the effect of CnABP on cell growth, we knocked down CnABP in the renal cell carcinoma cell line ACHN, which expresses both PAX2 and CnABP endogenously (data not shown). Quantitative PCR showed that short hairpin RNA (shRNA) 7202 and 8392 reduced CnABP expression levels by 35% and 75%, respectively (Supplementary Fig. S2C). As expected, CnABP knockdown in ACHN cells resulted in significant reduction in cell growth. shRNA 7202 caused 70% reduction (P < 0.001) in cell number after 96 hours, whereas shRNA 8392 resulted in a virtually complete arrest of cell growth (P < 0.001; Fig. 4C). Together, these complementary approaches point to an important role of CnABP in regulating cell proliferation.

To test whether CnABP additionally play a role in cell migration, we transiently transfected HEK293 cells with CnABP or vector control and sorted cells according to expression levels of a cotransfected pEGFP reporter construct. Western blot analysis confirmed the correlation between CnABP and EGFP expression levels (Supplementary Fig. S2B). Migration assays in Boyden chambers revealed an increase in cell migration that followed CnABP expression levels (Fig. 5A). Cells expressing low CnABP indeed showed a 3-fold increase in cell migration in comparison with empty vector (P < 0.01), whereas high CnABP-expressing cells increased cell migration by 7-fold (P < 0.001; Fig. 5B). Conversely, knockdown of CnABP in ACHN cells reduced cell migration (Fig. 5A). At 35% knockdown, we observed an 8% decrease in transfilter migration, whereas 75% knockdown reduced migration by 75% (P < 0.01) of the control levels (Fig. 5C). These data indicate that CnABP promotes cell migration in a dose-dependent manner.

To gain more insight into the cellular function of CnABP, we sought to identify putative interaction partners. For this, we submitted the full-length CnABP protein to a yeast two-hybrid approach. Two independent screens, done with the LexA and Gal4 systems, led to the identification of the serine/threonine phosphatase calcineurin Aβ as one of CnABP interacting partners. To validate this interaction, we did a pull-down assay on extracts from HEK293 cells stably transfected with His/Myc-tagged CnABP using nickel beads. Western blot analysis with an anti–calcineurin Aβ antibody revealed a strong interaction between endogenous calcineurin Aβ and CnABP in cells expressing CnABP-His/Myc (Fig. 6A). The interaction between both proteins was additionally validated by immunoprecipitation with an anti-Myc antibody directed against CnABP-His/Myc (Fig. 6B). Using pull-down and immunoprecipitation, we could not confirm interaction of CnABP with other potential partners identified in the yeast two-hybrid screens.

As CnABP binding to calcineurin Aβ suggests a role for CnABP in calcineurin signaling, we next measured the effect of CnABP on calcineurin phosphatase activity based on phosphate release from a calcineurin-specific synthetic peptide (RII). The 10-fold difference in CnABP expression between cells transfected with 1.2 and 6 μg of cDNA was used to establish low (0.1) and high (1.0) relative expression levels. The comparison of CnABP and vector-transfected HEK293 cells revealed a slight reduction in calcineurin phosphatase activity in low (0.1) CnABP-expressing cells. However, high (1.0) expressing cells showed a 40% reduction in phosphatase activity (P < 0.01; Fig. 6C). From these data, we conclude that CnABP interacts with and negatively regulates calcineurin Aβ. This interaction gives rise to the name CnABP.

We next investigated whether the inhibition of calcineurin Aβ by CnABP would affect calcineurin-mediated signaling. The nuclear factor of activated T-cell (NFAT) family of transcription factors is known as one of the main signaling components downstream of calcineurin. On calcium release, calcineurin dephosphorylates NFAT, which leads to its nuclear translocation where it acts as a transcriptional regulator. To test whether CnABP affects this process, we first investigated the effect of CnABP on endogenous NFAT nuclear translocation in HEK293 cells. In the absence of stimulus, NFATc3 is mostly cytosolic (Fig. 7A). On activating calcium release by ionomycin, NFATc3 rapidly translocates into the nucleus (Fig. 7B). In striking contrast, ionomycin treatment on CnABP-expressing cells completely failed to translocate NFATc3 to the nucleus (Fig. 7D). This inhibition is similar to the response obtained with cyclosporin A (CsA), a potent inhibitor of calcineurin (Fig. 7C).

We next assessed whether inhibition of calcineurin Aβ by CnABP would result in a reduction of NFAT transcriptional response. The effect of CnABP was assessed by measuring the expression of cyclooxygenase-2 and c-Myc, which are two established nuclear targets of NFAT. As expected, ionomycin-dependent up-regulation of both genes could be efficiently inhibited by CnABP expression (Fig. 7E and F). To confirm that this effect is mediated through NFAT-specific transcriptional activity, we assessed the effect of CnABP on NFAT-responsive promoters driving reporter gene expression. We used two NFAT luciferase reporter constructs, the c-Myc (Del1) promoter (39) and the minimal IL2 promoter containing three upstream NFAT binding sites (40). On induction by ionomycin, we observed significant increases in luciferase expression. However, treatment with CsA or expression of CnABP resulted in strong reductions in luciferase expression. Together, these data show that calcineurin A inhibition by CnABP affects downstream signaling and nuclear response through the NFAT pathway.

The overexpression of PAX2 in renal tumors suggests that these cells adopt a developmental program taking part in the oncogenic process. However, the molecular pathways regulated by PAX2 transcriptional activity remain poorly understood. In this study, we identified a putative target gene of Pax2 in the kidney. We showed that this gene, CnABP, contains a PAX2-responsive enhancer and is expressed together with Pax2 in the embryonic kidney at different stages. Importantly, both CnABP and PAX2 are overexpressed in the majority of WTs tested, consistent with a role in tumorigenesis. In support of this, the expression of CnABP was shown to promote cell migration and proliferation. At the molecular level, we identified CnABP as an inhibitor of calcineurin Aβ, the catalytic subunit of a phosphatase known to act as an intermediate of calcium signaling involved in proliferation and migration (41-45). Together, these results suggest a role for CnABP in the oncogenic program leading to WTs.

We find CnABP and Pax2 coexpressed in the metanephric mesenchyme and other kidney compartments. Metanephric mesenchymal cells are thought to give rise to WTs when they fail to be successfully induced toward the nephron differentiation program. Strikingly, PAX2 and CnABP were also coexpressed in a large proportion of the 39 WT samples we analyzed. For PAX2, this observation is in perfect concordance with previous studies showing overexpression in 60% to 75% of tumors (46, 47). The presence of PAX2 and CnABP likely reflects the developmental origin of WTs, but the fact that they are present in such a large proportion of tumors strongly supports the idea that persistence of the developmental program participates in the oncogenic process.

Although we cannot determine exactly when CnABP exerts its action, some evidence points to a role in tumor progression rather than initiation. CnABP overexpression was indeed highest in WT samples with a LOH at chromosomal arms 16q and 11p, with overexpression in 70% and 100%, respectively. As LOH at 16q has been associated with later stages of tumor progression and poor prognosis (7, 8) and typically follows loss of Wt1 at 11p15 (48), the overexpression of CnABP likely comes as a later event in tumor progression. The 2- to 18-fold overexpression of CnABP in 16q LOH tumors is especially remarkable in that the CnABP gene is located on this chromosomal segment (16q24). CnABP is thus highly overexpressed in those tumors in spite of the presence of a single allele instead of two. In further support of a role in tumor progression is the link between CnABP and cell migration. This is especially relevant in the context of WT recurrence observed in 15% of cases. These tumors are generally resistant to treatment regardless of histologic features and acquire the capacity to migrate and invade surrounding tissues (5, 6). This link between WT recurrence and increased cell dissemination (6) is entirely compatible with the cellular functions we identified for CnABP. Expression of CnABP in embryonic kidney and WT cell lines indeed results in a modest increase in cell proliferation but most strikingly results in a 7-fold increase in cell migration. Our results therefore suggest a potential association between CnABP overexpression and adverse prognosis that could eventually be explored in the context of therapeutic interventions.

In an effort to better understand the molecular function of CnABP in the cell, we looked for protein partners and identified the signaling molecule calcineurin Aβ. We showed that CnABP interacts with calcineurin Aβ and interferes with its phosphatase activity as well as with its downstream NFAT-dependent nuclear response. At the cellular level, calcineurin-NFAT signaling has been associated with different outcomes such as increased proliferation and migration (42, 44, 45). This pathway was shown to play an important role in kidney development (49) and homeostasis (50). Interestingly, a number of calcineurin pathway components, including calcineurin A, were recently identified as specifically up-regulated in relapsing WTs (9). The calcineurin pathway itself may thus be involved in WT progression. Our results suggest that CnABP overexpression in WTs might impinge on this pathway.

In our experiments, the interaction between CnABP and calcineurin A leads to an inhibition of calcineurin phosphatase activity. This is somewhat surprising because both CnABP and calcineurin pathway activation are linked to promigratory and proproliferative activities. One possibility is that the nuclear response to NFAT signaling may not be as straightforward as previously thought. In this respect, it is interesting to note that NFATc3, which we found regulated by CnABP, may act as a tumor suppressor in certain systems and thus have antiproliferative activity (51, 52). Similarly, calcineurin regulation of NFATc2 inhibits cyclin-dependent kinase 4 expression, a cyclin-dependent kinase promoting reentry into the cell cycle (53). Hence, calcineurin signaling seems to have both proproliferative and antiproliferative activities depending on the cellular context. It is also possible that CnABP is involved in calcineurin-independent promigratory/proproliferative pathways yet to be identified. Alternatively, it is possible that CnABP is a substrate for calcineurin with its own NFAT-independent cellular activity, as is the case for the K⁺ channel TRESK (54). In this respect, it is remarkable that the best-conserved domain of CnABP contains a perfect PxIxIT consensus sequence (where x represents any amino acid), which is a calcineurin docking site found in NFATs and other calcineurin substrates (54, 55), as well as in calcineurin inhibitor proteins such as Cain/Cabin1 and A238L (56-58). This sequence was shown to be sufficient for binding to calcineurin (59). In fact, based on a mapped consensus binding strength, the PQIIIT sequence of CnABP is predicted as a strong binding site among calcineurin binding partners or substrates (59). Mutagenesis experiments will be important to determine the exact role of this site in CnABP function. The details of the intriguing relationship between CnABP and calcineurin signaling remain to be clarified and will be the focus of further studies.

The identification of Pax2 target genes in the mesonephros was described elsewhere (36). Briefly, mesonephric cells from C3H/He embryos were sorted by flow cytometry using a bacterial artificial chromosome transgene expressing GFP under the control of the Pax2 locus. Total RNA from sorted cells was submitted to linear amplification and hybridized to cDNA microarrays containing 25,000 ESTs. Sample comparisons were wild-type to Pax2 mutant, GFP-positive mesonephros to surrounding GFP-negative tissue, as well as samples from different stages of mesonephros development.

A region spanning 50 kb upstream to 50 kb downstream of the human CnABP gene was used to map potential PBSs. Sites were chosen based on sequence conservation between human and mouse or rat. Three PBSs (−7267, +2012, and +5698) were named PBS1, PBS2, and PBS3, respectively. A region of 200 bp surrounding each site was cloned into pGL3 vector upstream of a TATA box and firefly luciferase cDNA. Mutations to three critical base pairs were engineered into PBS2 by site-directed mutagenesis (Fig. 1A).

Cells in 12-well plates were transiently transfected with 0.25 μg of reporter plasmids and 0.25 μg of expression constructs with 0.01 μg of pCMV-βGAL as transfection control using Fugene (Roche). To test for PAX2 activation of putative PBSs, pGL3-PBS1, pGL3-PBS2, pGL3-PBS3, and pGL3-PBS2-Mut were transfected with pCDNA3.1 or pCDNA3.1-PAX2 (a gift from Dr. Paul Goodyer, McGill University, Montreal, Canada). To evaluate the effect of CnABP on NFAT-responsive promoters, pGL3-NFAT-Luc (40) and pBV-cMyc (Del1; ref. 39) were transfected with pCDNA4 or pCDNA4-CnABP. Treatment of 2 μmol/L ionomycin and/or 2 μmol/L CsA was administered 24 h after transfection. Forty-eight hours after transfection, cell lysates were collected. Luciferase activity was determined by luciferase assay (Promega) using a luminometer (Luminoskan, Thermo). Relative light units were normalized against β-galactosidase staining and vector control.

In situ hybridization was done as described previously (60). CnABP cRNA probe was generated from IMAGE clone ID 6771161 (Invitrogen), containing the full coding sequence. Pax2 cRNA probe was described previously (61, 62).

Thirty-nine WT and 2 normal kidney samples were obtained from the Children Oncology Group (Arcadia, CA). All but one tumor were removed before chemotherapeutic treatment and were previously characterized by the Children Oncology Group for LOH at positions 16q, 1p, and 11p. Tumors and normal tissues were obtained in respect of institutional and national bioethical guidelines.

RNA from frozen tumor samples or from cells in culture was extracted with Trizol (Invitrogen). cDNA was synthesized by reverse transcription of 1 μg RNA using SuperScript II (Invitrogen). Expression analysis of CnABP isoform-a and isoform-b was done by PCR on reverse-transcribed cDNA with isoform-specific primers. Genomic DNA was used as positive control for isoform-b. All quantitative PCRs were done in duplicates on a Light Cycler (Roche) using SYBR Green (Sigma) as detection method. Primers used to amplify human CnABP are as follows: isoform-a, 5′-GAGAAGTTGATGGAGAAGCATTCC-3′ (+83 to +105; forward) and 5′-TGCTGTACTGGGGTTCCTGTGC-3′ (+369 to +390; reverse); isoform-b, 5′-AAAGCAAAGAGAAACCCAGGAAAG-3′ (+53 to +76; forward) and 5′-CCTCCAAGCCACTGACCTCG-3′ (+277 to +296; reverse). Primers used to amplify human cyclooxygenase-2 are 5′-AACACCCTCTATCACTGGCATCC-3′ (forward) and 5′-CCTGCTCTGGTCAATGGAAGC-3′ (reverse). Primers used to amplify c-Myc (63), PAX2, and β2-microglobulin (B2M; ref. 27) were described previously. Relative expression of mRNA transcripts was calculated with respect to B2M transcript levels.

Cell lines used (HEK-293, SKNEP-1, and ACHN) were obtained from the American Type Culture Collection. The HEK-293 cell line was maintained in DMEM (Wisent) supplemented with 10% fetal bovine serum (Wisent), SKNEP-1 WT cell line was maintained in McCoy's 5A (Wisent) supplemented with 15% fetal bovine serum, and ACHN renal cell carcinoma cell line was maintained in Eagle's MEM (Wisent) supplemented with 0.1 mmol/L nonessential amino acids, 1 mmol/L sodium pyruvate, and 10% fetal bovine serum. Wild-type coding sequence of CnABP was PCR amplified from human CnABP cDNA clone obtained from the Image Clone Collection (ID 2688932; Invitrogen) and cloned into pCDNA4/Myc-His A (Invitrogen). Myc-His–tagged CnABP was generated by site-directed mutagenesis of CnABP stop codon in pCDNA4/Myc-His A. Cells were either transiently transfected or selected as pools with zeocine (Invitrogen) for 2 wk before doing functional assays. Knockdown of CnABP was achieved by lentiviral-mediated expression of shRNA targeting CnABP transcript. shRNA hairpins expressed in pLKO.1 were obtained from Open Biosystems (IMAGE IDs: TRCN0000137202 and TRCN0000138392). Viral particles were collected 48 h after transfection for up to 7 d. Cells were infected daily with filtered virus (0.45 μm filters, PALL) and 4 μg/mL polybrene. Infected cells were selected with 2 ng/mL puromycin (Calbiochem) for 2 wk before doing functional assays. Experiments done with lentivirus were carried out according to institutional guidelines.

Western blot was done according to standard procedure. Rabbit polyclonal anti-CnABP antibodies were generated by injection of a 21–amino acid synthetic peptide corresponding to the COOH terminus of CnABP. Antibody specificity for endogenous and transfected CnABP was tested by peptide competition at 10× or 100× the concentration of the antibodies used. Concentrations used for Western blot analysis were 1:1,000 for anti-CnABP and 1:1,000 for anti-tubulin (Abcam). Secondary antibodies conjugated to horseradish peroxidase (1:10,000; Amersham) and enhanced chemiluminescence Plus reagent (Amersham) were used for detection. Stripping of blots was done using Re-Blot Plus Strong Solution (Chemicon).

Experimental and control cells were plated at densities of 10,000 per well (HEK-293 and ACHN) or 50,000 per well (SKNEP-1) in 12-well plates. Cells were counted using Coulter counter over 4 d; each time point was measured in triplicates. Statistical values were obtained with Student's t test.

Transwell filter assay was done as previously described (64). Cells (HEK-293 and ACHN) were seeded on 8-μm polycarbon filters with serum-free medium. Complete medium is added below the filters. Cells were allowed to migrate for 24 to 48 h at 37°C, 5% CO₂. Filters were fixed in 10% formalin and stained with crystal violet solution (0.1% crystal violet, 20% methanol). Nonmigratory cells were manually removed with cotton tips. Each experiment was done in triplicates and four fields of 1 mm² were used to quantify the number of migrated cells.

Yeast two-hybrid assays were done in collaboration with Hybrigenics. For this, the full-length CnABP bait was expressed in frame with LexA or Gal4DBD and each construct was screened against an embryonic cDNA library. Calcineurin Aβ was identified in both LexA- and Gal4AD-based screens.

Cells were lysed in binding buffer containing 20 mmol/L Tris (pH 8), 25 mmol/L NaCl, 1% Triton X-100, 10% glycerol, and protease inhibitors (Roche). Histidine-tagged CnABP was pulled down with Ni Sepharose 6 Fast Flow beads (Amersham) according to the manufacturer's instructions. Immunoprecipitation was done overnight with c-Myc antibody (clone 9E10; Santa Cruz Biotechnology) and incubated for the last 4 h with protein A/G Plus-agarose beads (Santa Cruz Biotechnology). The beads were washed thrice with binding buffer and proteins were resolved by SDS-PAGE. Western blots were done with calcineurin Aβ (1:1,000; Upstate) and c-Myc (1:1,000; Upstate) antibodies.

The calcineurin activity assay was done using the Calcineurin Cellular Assay Kit Plus (Biomol International) according to the manufacturer's instructions. The assay was made in HEK-293 cells transfected with different amounts (1.2 and 6 μg) of CnABP or pCDNA4/Myc-His vector DNA using Fugene 6 (Roche) according to the manufacturer's instructions. Western blot using Alexa-conjugated secondary antibodies and the Typhoon detection system (Amersham) were used to quantify the level of CnABP expression. P values were obtained by Student's t test.

For immunofluorescence staining, cover slides were pretreated with 12 N HCl for 5 min, washed with water, and dried in an oven at 60°C overnight. Cells (250,000) were plated in six-well plates and allowed to rest for 24 h. Cells were then transfected with either CnABP or pCDNA4/Myc-His A control vector. Twenty-four hours after transfection, the cells were treated with 2 μmol/L ionomycin and/or 2 μmol/L CsA for 15 min. Cells were briefly rinsed with PBS and fixed in 4% paraformaldehyde for 15 min. Samples were blocked with 10% normal goat serum and 0.2% Triton X-100 in PBS for 1 h at room temperature. Slides were incubated with anti-NFATc3 (1:50; Santa Cruz Biotechnology) and/or anti-Myc (1:100; Upstate) in 1% normal goat serum and 0.2% Triton X-100 in PBS overnight at 4°C. Subsequently, samples were incubated with Alexa-labeled secondary antibodies (Invitrogen) and 1 μg/mL of 4′,6-diamidino-2-phenylindole for 1 h at room temperature. Cells were washed thrice with PBS for 10 min between each step. Samples were mounted on cover slides using Prolong Antifade mounting medium (Invitrogen) and cured overnight at room temperature. Images were obtained on a LSM 5 Pascal confocal microscope (Zeiss).

The authors declare no conflicts of interest.

We thank the Children Oncology Group and the Cooperative Human Tissue Network for providing WT samples, Hybrigenics for the yeast two-hybrid assay, and Jerry Pelletier and Arnim Pause and the members of the Bouchard lab for critical reading of the manuscript.