| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
From the Institute for Cancer Genetics,* the Department of Pathology, and the Department of Urology, Division of Pediatric Urology, Columbia University College of Physicians and Surgeons, New York, New York; the Department of Medicine, Mount Sinai School of Medicine, New York, New York; the Department of Internal Medicine,|| Division of Medical Genetics and the Cancer Center, University of Michigan Medical School, Ann Arbor, Michigan; and the Department of Pediatrics, ¶ University of Alberta and Cross Cancer Center, Edmonton, Alberta, Canada
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
10% of Wilms’ tumors, and this subset is enriched in tumors that occur in two heritable syndromes, Denys-Drash and Wilms’ tumor aniridia genitourinary anomaly retardation syndrome (WAGR).1,2
Both syndromes involve abnormalities of urogenital development, and consistent with a key role for WT1 in urogenital development, mice lacking this gene fail to form kidneys.3
WT1 encodes a transcription factor, and several candidate WT1 target genes are expressed in metanephric structures.4-6
But although WT1 is a prototypical tumor suppressor,7
it is not yet known how its loss predisposes to Wilms’ tumorigenesis.8
ß-Catenin, encoded by CTNNB1, is a nascent transcription factor that transmits signals from Wnt ligands to the cell nucleus. ß-Catenin mutations, occurring selectively at amino acid residues that are normal sites of phosphorylation, create a nondegradable form of this protein.9
Such mutations have been documented in Wilms’ tumors,10,11
and a previous study suggested that they might occur almost exclusively in cases with WT1 mutations.12
But 50% of WT1-mutant Wilms’ tumors lacked such mutations, leaving open the question of an essential link between loss of WT1 and activation of the ß-catenin/TCF pathway. We have now answered this question positively, using mutational analysis, gene expression profiling, and functional assays. In addition, our data uncover a novel ß-catenin/TCF target gene, GAD1, and also show relative overexpression of specific myogenic transcription factors, signaling molecules, and signaling inhibitors in WT1-mutant compared to WT1 wild-type Wilms’ tumors.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Sporadic Wilms’ tumors were newly diagnosed and archival cases resected at Babies Hospital of Columbia University and at Hackensack University Hospital. The newly diagnosed cases were studied with informed consent, whereas the archival cases were studied blind to patient identifiers, under an exempt protocol approved by the Columbia University Institutional Review Board. Cryopreserved biopsies of syndromic Wilms’ tumors were obtained anonymously from the National Wilms’ Tumor Study Group.
Oligonucleotide Microarray Analysis
HG-U95A GeneChips (Affymetrix), which query 10,000 genes (12,627 probe sets), were used for analyzing Wilms’ tumors. The cRNA probes were synthesized as described.13 After scanning, mRNA expression values were determined using Affymetrix GeneChip software version 5.0. Expression values for each probe set were first filtered for a minimal level of variation (1.5-fold variation from the experiment mean in at least six tumors) and for reliability (Affymetrix presence calls in at least five tumors). Genes that passed these criteria were subjected to nonsupervised clustering using the five-cluster K-means algorithm. For supervised analysis we labeled the cases as WT1 wild type or WT1-mutant. Starting with the probe sets that passed the initial criteria of variability and reliability, we asked for the genes that differed in their expression most strongly between these two groups of tumors, using the nonparametric Mann-Whitney U-test. Identical lists of the genes with the strongest differential expression were obtained when the genes were ranked using parametric tests. We used GeneSpring (Silicon Genetics, Redwood City, CA) for data analysis and for creating dendrograms.
Northern Blotting
RNA from Wilms’ tumor and fetal kidney tissues, pulverized under liquid nitrogen, and RNA from cell lines, was prepared using Trizol (Invitrogen, Carlsbad, CA) and were resolved on formaldehyde-containing agarose gels and transferred to Nytran membranes (Schleicher and Schull, Keene, NH). Probes were partial cDNAs prepared by reverse transcriptase-polymerase chain reaction (PCR) using gene-specific primers (sequences available on request). Hybridization was in ULTRAhyb (Ambion, Austin, TX) at 42°C overnight; washing was at 64°C in 0.1% sodium dodecyl sulfate/0.1x standard saline citrate. Band intensities on the autoradiograms were measured with a Storm Phosphorimager (model 840; Molecular Dynamics/Amersham Biosciences Corp., Piscataway, NJ).
Wilms’ Tumor Explants and Transforming Growth Factor (TGF)-ß2 Treatment
Wilms’ tumor explants were dispersed by mincing, and maintained and passaged in Dulbecco’s modified Eagle’s medium with 10% fetal bovine serum, 10 mmol/L penicillin-streptomycin, and 10 mmol/L L-glutamine. TGF-ß2 peptide (Sigma, St. Louis, MO) was added to an epithelial-appearing explant from a sporadic Wilms’ tumor (passage 4) at concentrations of from 0 to 50 ng/ml when the cells were 30% confluent. Plates were photographed after 4 days, cells were harvested, and total RNA was isolated for Northern blot analysis.
Sequencing of CTNNB1 and WT1
PCR primers for amplifying and sequencing exon 3 of CTNNB1 from genomic DNA, and for amplifying the entire CTNNB1 coding region from cDNAs, were based on sequences in GenBank. All CTNNB1 mutations found at the cDNA level were confirmed by exon-specific PCR of genomic DNA. WT1 was sequenced in each case both from genomic DNA (primers spanning individual exons 1 to 10) and from partial cDNAs, using primers located in exons 4 and 10. Primer sequences are available on request.
Tissue Microdissection
Paraffin-embedded tissue sections were stained with toluidine blue, and the epithelial and stromal components microdissected using a PixCell IIe LCM System (Arcturus, Mountain View, CA). For documentation, microscopic images were digitally captured before and after the microdissection procedure. Genomic DNA was purified from the tissue fragments using proteinase K/sodium dodecyl sulfate digestion followed by phenol/chloroform extractions. Pellet-paint co-precipitant (Novagen) was then added; the DNA was precipitated with 70% ethanol, dissolved in distilled water, and used for PCR followed by direct sequencing. Because relatively small amounts of DNA were obtained from the microdissected blocks, special precautions were taken to avoid PCR contamination; in addition to running standard minus-DNA controls, which were negative, the CTNNB1 PCR primers for the microdissection experiments were repositioned in more distal upstream and downstream intronic sequences flanking exon 3, outside of the sequences used in previous analyses of whole tissue DNA.
Promoter-Reporter Assays
Human 293 transformed kidney cells were maintained in Dulbecco’s modified Eagle’s medium with 10% fetal bovine serum and 10 mmol/L penicillin-streptomycin. The GAD1 (GAD67) promoter, nucleotides 73,250 to 72,781 of GenBank accession AC007504 (–400 to + 60 relative to the transcription initiation site), was cloned upstream of LUC in the pGL3 vector. A site-directed mutagenesis kit (QuikChange; Stratagene, La Jolla, CA) was used to mutate the TCF site (sequence CTTTGAT 
CTTTGGG). The dominant-negative TCF4 construct (pPGS-CITE-CMV-Tcf4dN, dnTCF4) and S37A ß-catenin construct, used to ascertain TCF-dependent promoter activity, have been previously described.14
To measure ß-catenin/TCF responses, cells in 35-mm plates were transfected in triplicate with 1.5 µg of plasmid DNA containing 0.4 µg of GAD1 promoter-reporter plasmids, with or without 0.6 µg of S37A ß-catenin plasmid and with or without 0.2 µg of dnTCF4 plasmid, and 0.1 µg of pCMV-ß-galactosidase (ß-gal). Carrier pCMV (pPGS-CITE-CMV-Neo) was added to make the total plasmid DNA concentration equal in each experimental condition. As positive and negative controls, pTopflash (containing four concatemerized TCF sites) and pFopflash (containing mutated versions of these sites) were transfected with 0.1 µg of the S37A ß-catenin plasmid. Luciferase and ß-galactosidase assays were performed according to the manufacturer’s protocol (Luciferase Assay System; Promega, Madison, WI).
Co-Transfection Assays for Activity of Mutant ß-Catenins
To assess the functions of the non-hotspot CTNNB1 mutations, the complete protein-coding regions of the mutant alleles were amplified by reverse-transcription (Superscript, Invitrogen) followed by PCR using CTNNB1 primers, and cloned into the pEGFP-N1 vector (Clontech, Palo Alto, CA), producing GFP-fusion proteins with expression driven from the CMV promoter. These plasmids, 0.5 µg, were co-transfected with the pTopflash or pFopflash luciferase reporter plasmids, 0.4 µg, together with 0.1 µg of the Renilla luciferase control plasmid, into 293 cells. Expression values were determined from the ratio of firefly luciferase to Renilla luciferase activity using the Dual Luciferase Assay System (Promega).
Immunohistochemistry
Procedures were essentially as previously described.13 Antigen retrieval for detection of ß-catenin was done by boiling the deparaffinized sections on slides for 10 minutes in 1 mmol/L ethylenediaminetetraacetic acid, pH 8.0, in a microwave oven. The anti-ß-catenin (ß-catenin; BD Transduction Laboratories, Lexington, KY) was used at a dilution of 1:400, and the anti-SIM2 antibody (C-17; Santa Cruz Biotechnology, Santa Cruz, CA) was used at a dilution of 1:500. For detecting ß-catenin, the secondary antibody (biotinylated horse anti-mouse; Vector Laboratories, Burlingame, CA) was used at a dilution of 1:400. For detecting SIM2, the secondary antibody (biotinylated rabbit anti-goat, Vector Laboratories) was used at a dilution of 1:400.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
We first sequenced WT1 in 19 sporadic and 17 syndromic Wilms’ tumors. Only a single sporadic tumor (WT564) showed a WT1 mutation, whereas each of six Denys-Drash-associated tumors and five of nine WAGR-associated tumors sequenced as positive controls showed the expected point mutations in the zinc finger region (Table 1)
. We next sequenced the mutational hotspot region of CTNNB1. This region, in exon 3, encodes phosphorylation sites (serine and threonine residues) that regulate ß-catenin stability. Among 17 WT1-mutant syndromic cases, 9 carried missense mutations or deletions affecting this exon of CTNNB1 (Table 1)
. In contrast, only 1 of the 19 sporadic Wilms’ tumors carried an exon 3 mutation and, strikingly, this was the single tumor with a WT1 mutation. Combining our analysis with the previous study,12
it is clear that CTNNB1 hotspot mutations occur in 50% of cases WT1-mutant Wilms’ tumors, but are rare or absent in their more common WT1 wild-type counterparts.
|
and data not shown). This analysis does not exclude that there could be a minor subset of Wilms’ tumors with nuclear ß-catenin even in the absence of CTNNB1 mutations, but it does indicate a good correlation between the presence of such mutations and the nuclear accumulation of ß-catenin.
|
. When we performed tissue microdissection and isolated DNA from the epithelial and stromal components of one of these cases, the heterozygous (gain-of-function) CTNNB1 mutation was clearly identified in both components (Figure 2)
. As expected, the WT1 homozygous mutation was also detected in both components (data not shown). Thus, mutation of the CTNNB1 gene occurred early enough in tumorigenesis that the mutation was present in all cells of the surgically resected tumor. These combined immunohistochemical and genetic data suggest that, despite its constitutive stabilization, a cytoplasmic or membrane component anchors the mutant ß-catenin in the cytoplasm in the epithelial tumor cells, while this mutant protein is free to translocate to the nucleus in the malignant stromal cells.
|
For convenience in the following discussion we refer to two Wilms’ tumor classes: class 1 tumors are WT1 wild-type whereas class 2 tumors are WT1-mutant. Our hypothesis in the next series of experiments was that constitutive activation of the Wnt/ß-catenin pathway, either by CTNNB1 mutations or by lesions in other components of this pathway, might occur in all WT1-mutant (class 2) Wilms’ tumors. Thus, sets of genes differentially expressed between class 1 and class 2 Wilms’ tumors might be enriched in ß-catenin/TCF targets. We performed expression profiling of 17 sporadic Wilms’ tumors and a comparison group of 11 syndromic Wilms’ tumors (seven WAGR syndrome and four Denys-Drash syndrome), with cases selected only on the basis of RNA availability. Both supervised and nonsupervised, hierarchical clustering of the microarray data readily distinguished the (primarily syndromic) WT1-mutant cases from the sporadic WT1 wild-type cases (Figure 3)
. Significantly, the single sporadic tumor with a WT1 deletion (WT564) showed a pattern of gene expression that closely matched the consensus of the syndromic WT1-mutant cases (Figure 3)
.
|
, arrowheads). Both of these outlier cases were classical triphasic Wilms’ tumors, but analysis of additional tumor samples will be necessary to determine whether a specific histology, perhaps stromal abundance analogous to that seen in many of the syndromic tumors, and/or an as yet unidentified genetic lesion, correlates with this intermediate pattern of gene expression.
For validation of the GeneChip data, we hybridized Northern blots with probes for 10 of the differentially expressed genes, selected based on their interesting potential biological functions. For each gene the pattern of expression closely matched that predicted from the microarray data (examples in Figure 4A
). Measuring the band intensities on these blots by Phosphorimaging confirmed that the GeneChips were giving a valid readout of differential mRNA expression (Figure 4B)
.
|
. Among five class 1 Wilms’ tumors examined, four showed no detectable SIM2 protein, while one case showed detectable SIM2 protein focally in the epithelial component, but not in the stromal component.
|
The genes with the greatest differential expression between the two Wilms’ tumor classes are described and ranked by significance in Tables 2 and 3
. Although the differentially expressed genes clearly distinguished the class 1 from class 2 tumors, additional analysis showed these two classes otherwise share a common expression profile. Specifically, of the 28 genes identified in our previous analysis as overexpressed in sporadic Wilms’ tumors compared to normal fetal kidney and several heterologous tumors and tissues,13
only 3 (GPR39, PAX2, PRAME) were differentially expressed in the comparison of class 1 versus class 2 Wilms’ tumors. This finding would be predicted based on general similarities in histogenesis. Although there are histological differences, notably a tendency for more mesenchymal and myogenic differentiation in the syndromic tumors,15,16
the pathological criteria for diagnosis are identical in these tumors and their sporadic counterparts.
|
). A gene encoding a third extracellular Wnt inhibitor, DKK1, was also differentially expressed, although at a less stringent P value (fold difference, 3.6; P value, 2.2 x 10–4) that was below the cutoff for the list in Table 3
. Consistent with the known tendency of syndromic Wilms’ tumors to undergo myogenic differentiation, two genes encoding myogenic transcription factors, MEOX2 (also known as MOX2) and LBX1, appeared in the class 2 signature set (Table 3
; see also Figure 4
for Northern blot confirmation of MEOX2 overexpression).
|
. In fact, when we exposed a class 1 Wilms’ tumor explant to exogenous TGF-ß2, the cells underwent a morphological epithelial-mesenchymal transition, and showed changes in the expression of several signature genes (upregulation of CSPG2 and down-regulation of IFI16 and PAX8) suggesting a partial transition to a class 2 phenotype (Figure 6)
. As discussed below, the versican gene (CSPG2) is likely to be a convergent target of activation by TGF-ß2 signaling and by the nuclear activity of ß-catenin/TCF in class 2 Wilms’ tumors. Additional signaling molecules in TGF-ß-related pathways, the agonist BMP2A and the antagonist follistatin (FST gene), were also found among the class 2 signature genes (Table 3)
.
|
Known and Novel ß-Catenin/TCF Targets Are Overexpressed in WT1-Mutant Wilms’ Tumors with and without CTNNB1 Hotspot Mutations
Three known ß-catenin target genes, encoding c-Myc, cyclin-D, and Fra-117-19
did not appear among the class 2 signature genes. Nevertheless, examination of the primary data showed that CMYC was in fact modestly but consistently overexpressed in the class 2 Wilms’ tumors, with a mean difference of 2.3-fold in class 2 compared to class 1 tumors. In addition, two recently reported ß-catenin target genes,20
encoding follistatin (FST) and versican (CSPG2), showed strongly differential expression and did appear in the class 2 signature set (Table 3)
.
The GAD1 gene, encoding glutamic acid decarboxylase, is necessary for production of 
-aminobutyric acid in neurons. But cleft palate observed in mice lacking Gad1 suggests that this gene also plays an essential role in epithelial-mesenchymal interactions.21,22
GAD1 has not been reported as a ß-catenin/TCF target, but it is consistently overexpressed in class 2 Wilms’ tumors (Figures 3 and 4)
. Analysis of sequences in GenBank (accession AC007504) showed that the proximal promoter of this gene contains a consensus TCF-binding site, located at 298 to 292 bp upstream of the transcriptional start site, and reporter assays in conjunction with site-directed mutagenesis of this site revealed that the GAD1 promoter is ß-catenin/TCF responsive (Figure 7)
. Thus, GAD1 is a novel direct target of ß-catenin/TCF uncovered by our screen.
|
Some WT1-Mutant Wilms’ Tumors Carry Non-Hotspot CTNNB1 Mutations
As predicted from the hypothesis that ß-catenin stabilization is an obligate second event in WT1-mutant tumor precursor cells, a statistical analysis showed that bona fide ß-catenin/TCF target genes were relatively overexpressed both in the class 2 tumors with CTNNB1 hotspot mutations and in the remaining class 2 cases that lacked such mutations, in comparison with the class 1 Wilms’ tumors (Figure 8)
. To explain this general overexpression of ß-catenin/TCF target genes in class 2 Wilms’ tumors, we searched for mutations that might stabilize ß-catenin in those cases that lacked CTNNB1 hotspot mutations. Complete sequencing of CTNNB1 showed unusual heterozygous truncating and missense mutations, outside of exon 3, in several tumors in this category, with some of these mutations occurring in the APC/Axin-binding region (Table 1
and Figure 9A
). Other genes in the Wnt/ß-catenin pathway need to be examined, but based on these data hotspot plus non-hotspot CTNNB1 mutations already account for constitutive ß-catenin stabilization in 75% of the WT1-mutant tumors in our series.
|
|
S45, S37A), and with wild-type ß-catenin, in co-transfections with pTopflash and pFopflash reporters in 293 cells. As shown in Figure 9B
, both of the non-hotspot mutations tested proved to be gain-of-function alleles, reproducibly activating pTopflash transcription at levels twofold higher than wild-type ß-catenin. As might be expected from their less frequent biological selection during tumor formation, these rare alleles were somewhat less potent than the hotspot S45 and S37A mutant ß-catenins, which produced fourfold to sixfold increases in pTopflash transcription in this assay (Figure 9B)
. |
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
The strong expression of genes encoding extracellular inhibitors of Wnt signaling (WIF1, SFRP4) that we have shown is characteristic of the class 2 tumors is entirely consistent with the genetic data, because cells with constitutive activation of the intracellular arm of the Wnt pathway (via somatic mutations of CTNNB1) are expected to be at least partly resistant to growth inhibition by these extracellular molecules. One of these inhibitors, Dickkopf-1 (DKK1 gene) was previously reported as overexpressed in some Wilms’ tumors, although WT1 and CTNNB1 status was not determined.26
In principle, WT1 transcriptional targets might also be expected among the sets of differentially expressed genes reported here. Several targets of WT1 identified in cell line systems and in normal development were not differentially expressed between class 1 and class 2 Wilms’ tumors in our study, whereas other candidate WT1 target genes were found differentially expressed in the predicted direction, but did not pass our statistical criteria. Specifically, p21, podocalyxin, and amphiregulin do not appear among the class 1 signature genes. For podocalyxin, this is trivially accounted for by its lack of its representation on the GeneChip; for amphiregulin, the mRNA levels were not statistically different in the two classes of Wilms’ tumors, being low but detectable in both types of tumors, while p21 mRNA showed substantial case-to-case variability. Another candidate WT1 target, the cyclin E gene, was also expressed robustly in all of the Wilms’ tumors, regardless of WT1 status. The E-cadherin gene, which is WT1-responsive in some cell types, was variably expressed, although at levels much reduced from normal fetal kidney controls, without a significant difference between WT1-mutant and WT1-replete tumors. The connective tissue growth factor gene (CTGF/IGFBP8), previously identified as induced (3.1-fold) by a dominant-negative mutant form of WT1 in a cell line from an aggressive Wilms’ tumor27 did not pass our statistical and absolute expression value cutoffs, but examination of our primary expression data showed that it was in fact modestly (2.3-fold) overexpressed in class 2 (WT1-mutant) compared to class 1 (WT1-replete) tumors, consistent with its candidacy as a target gene for repression by wild-type WT1. Similarly, TGFB2 mRNA was identified in that previous screen (induced 2.9-fold by the dominant-negative WT1 construct), and this gene also appears in our class 2 signature set, consistent with its candidacy as a target for repression by wild-type WT1.
In agreement with our data, CMYC was recently independently reported as differentially expressed in WT1-mutant Wilms’ tumors,28 and from this evidence, as well as a previous study using cell transfections,29 this gene was proposed as a candidate WT1 target. Based on the uniform overexpression of ß-catenin/TCF targets in class 2 Wilms’ tumors, and the fact that CMYC is a ß-catenin/TCF target in other cell types, it seems likely that overexpression of CMYC in these tumors is at least in part because of hyperactivation of the ß-catenin/TCF pathway. Loss of WT1 may play a synergistic role in activating c-Myc in class 2 Wilms’ tumors.
Our data also have implications for the histopathology of syndromic Wilms’ tumors. These WT1-mutant tumors are known to have, in general, a more extensive stromal (mesenchymal) component than their sporadic, WT1-replete, counterparts.16 The TGF-ß2 protein, which is a class 2 Wilms’ tumor signature gene, is a potent mesenchyme-inducer in many cell systems, and we have shown here that it can have this effect on Wilms’ tumor cells. Relative overexpression of TGFB2 may therefore contribute to the abundance of the neoplastic stromal component in the class 2 tumors. Interestingly, we observed nuclear ß-catenin only in the malignant stromal component, suggesting that the stromal component may be important for the malignant behavior of these tumors. Similarly, the strong overexpression of the myogenic transcription factor genes MEOX2 and LBX1, also identified as class 2 signature genes in this study, may account for the focal myogenic (skeletal muscle) differentiation that is another consistent feature of WT1-mutant Wilms’ tumors.15,30
Overall, our findings, together with the previous observation that homozygous mutation of Wt1 is cell-lethal in a mouse knockout3 suggests that ß-catenin stabilization in WT1-mutant tumor progenitor cells may be necessary for survival and proliferation of these cells. Our data in no way exclude that WT1 transcriptional targets may play a direct role in tumor suppression by inhibiting cell proliferation, but our positive findings support an alternative model in which the tumor suppressor role of the WT1 gene is explained by a less direct biological mechanism. In this scenario, loss of function of WT1 in primitive metanephric cells predisposes these cells to apoptosis, thereby providing an ongoing selection, during kidney growth, for cell rescue and clonal expansion via activation of the Wnt/ß-catenin pathway. Alternatively, absence of WT1 and stabilization of ß-catenin may be synergistic and continuously necessary for growth of class 2 Wilms’ tumors. These hypotheses make different predictions, which may be testable in mouse genetic models.
Lastly, we have pointed out a minor subset of class 1 Wilms’ tumors that, despite lacking CTNNB1 mutations, show an expression profile intermediate between the class 1 and class 2 consensus patterns. In addition, 25% of class 2 Wilms’ tumors lack CTNNB1 mutations, but the tumor cells in these cases nonetheless overexpress ß-catenin/TCF target genes, adhering to the consensus class 2 signature. Whether these tumors contain alternative genetic lesions that activate the Wnt/ß-catenin signaling pathway will be an interesting topic for future investigation.
|
Acknowledgements |
|---|
S45 ß-catenin plasmid; and Alain Borczuk for help with tissue microdissection. |
Footnotes |
|---|
Supported by grants from the National Institutes of Health (to J.D.L. and B.T.) and the Stewart Trust (to B.T.).
Accepted for publication August 18, 2004.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  R. Fukuzawa, R. W Heathcott, H. E More, and A. E Reeve Sequential WT1 and CTNNB1 mutations and alterations of {beta}-catenin localisation in intralobar nephrogenic rests and associated Wilms tumours: two case studies J. Clin. Pathol., September 1, 2007; 60(9): 1013 - 1016. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. I. DeAlmeida, L. Miao, J. A. Ernst, H. Koeppen, P. Polakis, and B. Rubinfeld The Soluble Wnt Receptor Frizzled8CRD-hFc Inhibits the Growth of Teratocarcinomas In vivo Cancer Res., June 1, 2007; 67(11): 5371 - 5379. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Uschkereit, N. Perez, C. de Torres, M. Kuff, J. Mora, and B. Royer-Pokora Different CTNNB1 mutations as molecular genetic proof for the independent origin of four Wilms tumours in a patient with a novel germ line WT1 mutation J. Med. Genet., June 1, 2007; 44(6): 393 - 396. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Packer, S. Pavey, A. Parker, M. Stark, P. Johansson, B. Clarke, P. Pollock, M. Ringner, and N. Hayward Osteopontin is a downstream effector of the PI3-kinase pathway in melanomas that is inversely correlated with functional PTEN Carcinogenesis, September 1, 2006; 27(9): 1778 - 1786. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Yuan, C.-M. Li, D. J. Yamashiro, J. Kandel, H. Thaker, V. V. Murty, and B. Tycko Genomic Profiling Maps Loss of Heterozygosity and Defines the Timing and Stage Dependence of Epigenetic and Genetic Events in Wilms' Tumors Mol. Cancer Res., September 1, 2005; 3(9): 493 - 502. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
