| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Molecular Oncology, Markers, Clinical Correlates |
Institute of Human Genetics and Anthropology, University of Düsseldorf, 40225 Düsseldorf, Germany [V. S., S. Sc., S. So., B. R-P.]; Department of Pediatric Hematology and Oncology, Children’s Hospital, University of Heidelberg, 69120 Heidelberg, Germany [A. W.]; Institute of Pediatric Pathology, University of Kiel, 24105 Kiel, Germany [I. L., D. H.]; The Mount Sinai School of Medicine, New York, New York 10029 [J. L.]; and School of Biological Sciences, University of Manchester, Manchester M13 9PT, United Kingdom [S. R.]
 |
ABSTRACT |
|---|
 Top ABSTRACT IntroductionPatients and MethodsResultsDiscussionREFERENCES |
|---|
Experimental Design: Twenty-two new WTs were analyzed for WT1 mutations by PCR single-strand conformational polymorphism. Five tumors with WT1 mutations and six tumors without WT1 mutations were studied for the presence of WT1 transcripts and protein as well as for the expression of differentiation markers.
Results: Two new WT1 mutations were identified in stromal-predominant tumors, and none were identified in the other histological subtypes. Tumors with WT1 mutations expressed mutant messages, cytoplasmic truncated WT1 proteins, and muscle markers. In contrast, blastemal-predominant tumors without mutations showed nuclear WT1 protein staining. Both tumor types were positive for markers of early-induced mesenchyme and one marker of uninduced mesenchyme, but blastemal-predominant tumors also expressed cytokeratin, suggesting that these are further along the epithelial differentiation pathway.
Conclusions: Our data show that the two-hit inactivation of WT1 is operative in stromal-predominant WTs. Cells without functional nuclear WT1 protein start a faulty differentiation program. In contrast, blastemal-predominant tumors express wild-type WT1 and show early signs of epithelialization. The extensive rhabdomyomatous differentiation and the presence of WT1 mutations may be used as a diagnostic tool to identify a tumor subtype that seems to respond poorly to chemotherapy. These studies provide a foundation for improvement in tumor classification and ultimately for the development of more individualized tumor treatments.
|
Introduction |
|---|
|
TopABSTRACTIntroductionPatients and MethodsResultsDiscussionREFERENCES |
|---|
Nephrogenic rests, considered to be precursor lesions of WTs, are found in 30–44% of tumors (2) . ILNRs, located deep within the parenchyma and indicating an earlier developmental disturbance, are associated with intralobar WTs containing heterologous elements. ILNR-like WTs are observed in patients with sporadic WT often associated with GUs, anirdia, or Denys-Drash syndrome. PLNRs occur at the periphery of the renal lobe later in renal developmental and are associated with perilobar WTs lacking heterologous elements. PLNR-like WTs are often found in patients with hemihypertrophy and Beckwith-Wiedemann syndrome. Recently, Ravenel et al. (5) identified LOI of IGF2 as a valuable molecular tool for the classification of these two pathological subtypes. LOI and altered expression of IGF2 were found mainly in PLNR-like tumors.
The WT1 gene, a ZF transcription factor, was found to be mutated in 10–15% of sporadic WTs (for a review, see Ref. 6 ). Recently, we identified a high percentage of WT1 germline mutations in stromal-predominant/ILNR-like tumors (7) . Loss of the wild-type allele (LOH) was found in 7 of 13 analyzed cases, confirming Knudson’s two-hit hypothesis for this subclass of WTs. All but one were nonsense mutations predicted to result in protein truncation. Only one missense mutation was found in a blastemal-predominant tumor, suggesting that mutations in WT1 may play a minor role in the development of this subtype (7) .
Alternative splicing of the WT1 gene affecting exons 5 and 9 results in the synthesis of four major isoforms of the protein with different nucleic acid binding and transcriptional properties (8, 9, 10) . The WT1 protein has many postulated functions both in regulating transcription and in posttranscriptional RNA processing (reviewed in Ref. 11 ). It is targeted to the nucleus by the presence of at least two nuclear localization signals, one lying within ZF I and one in ZF II and/or III (12) . Reporter construct experiments have identified many putative in vivo target genes for both transcriptional activation and repression by WT1 (13 , 14) . However, significant variation has been found in these experiments, depending on the cellular context and the promoter used to express WT1 (15) . At present, little is known about the in vivo situation. By analyzing WT1 missense mutants, English and Licht (16) have found that activation and not repression by WT1 is critical for growth control. WT1 is predominantly expressed in the developing kidney, where it may play a role in mesenchymal to epithelial transition (17 , 18) . This hypothesis is supported by the findings that in WTs missing functional WT1, the mesenchymal cells can no longer differentiate into epithelia but can take on different fates, either stromal or myogenic (7 , 19) . During early nephrogenesis, a low level of WT1 is first detected in the uninduced mesenchyme and then increases dramatically during induction (20) . Expression then becomes restricted to the posterior part of newly formed epithelium and is limited to the podocytes of the adult kidney. In tumors, WT1 is seen in the malignant counterparts of those elements that express the protein during normal development (that is, blastema and glomeruloid structures) but is absent from stromal cells (21) .
To confirm and extend our previous observations that most stromal-predominant/ILNR-like tumors are caused by WT1 mutations, we have studied more patients for mutations and examined these for the expression of a mutant WT1 transcript and protein. Furthermore, we have studied the expression of a set of specific kidney development/differentiation markers as a first approach to identify a set of useful classification markers for WT.
|
Patients and Methods |
|---|
|
TopABSTRACTIntroductionPatients and MethodsResultsDiscussionREFERENCES |
|---|
Mutation Analysis of the WT1 Gene.
Of the 86 collected tumor specimens, 64 were previously analyzed for WT1 mutations by PCR single-strand conformational polymorphism and direct sequencing (7)
. Here we have analyzed 22 additional patients with the same method.
Northern Blot.
Total RNA (10 µg) was loaded on a 1% agarose gel in 4-morpholinepropanesulfonic acid buffer containing ethidium bromide and 0.22 M formaldehyde. Separation was at 30 V overnight. After photography, the gel was rinsed in H2O for 5 min and transferred overnight in 10x SSC to gene screen membranes (New England Nuclear, Dupont). After transfer, the RNA was UV-cross-linked, and the filter was baked for 2 h at 80°C. Preparation of RNA, hybridization probes, and conditions were as described previously (24)
.
RT-PCR.
For the identification of mutant WT1 transcripts, different oligonucleotide primers were used: GW21 (5'-CCGCCTCACTCCTTCATC-3') and GW25 (5'-ACCGAGTACTGCTGCTCAC-3'); and Sus39 (5'-TCGCAATCAGGGTTACAG-3') and Sus40 (5'-GTGGGTCTTCAGGTGGTC-3'); P1 and P4 were as described previously (25)
. Five µl of cDNA were amplified in a 50-µl PCR reaction containing 25 pmol of each primer, 200 µM each deoxynucleotide triphosphate, 1x Taq DNA polymerase buffer, and 1 unit of recombinant Taq polymerase (Life Technologies, Inc.). PCR cycle conditions were as follows: (a) for P1/P4, 3 min x 94°C; 30 cycles of 94°C x 1 min, 60°C x 2 min, and 72°C x 2 min; followed by a final elongation step at 72°C for 10 min; and (b) for GW21/25 and Sus 39/40, 3 min x 94°C; 30 cycles of 94°C x 1 min, 53°C x 1 min, and 72°C x 1 min; followed by a final elongation step at 72°C for 10 min.
Restriction Enzyme Digestion.
Wild-type or mutant WT1 alleles and transcripts were identified by digesting 15 µl of PCR products with enzymes specific for the mutant site. For digestion with BsiEI (New England Biolabs), MgCl2 was adjusted to 10 mM final concentration. Products were visualized on a 3% NuSieve-agarose gel stained with ethidium bromide.
Antibodies.
For immunohistochemistry, the following antibodies were used in this study: anti-collagen I (COL-1; 1:5000; Sigma); anti-collagen IV (CIV 22; 1:50; DAKO); anti-cytokeratin (KL1; 1:10; Immunotech); anti-desmin (D33; 1:250; DAKO); anti-fibronectin (1:2000; DAKO); anti-LamA (4C7; 1:300; DAKO); anti-MyoD1 (5.8A; 1:10; DAKO); anti-myogenin (F5D; 1:30; DAKO); anti-SA (Sr-1; 1:10; DAKO); anti-vimentin (Vim3B4; 1:10; Roche Diagnostics); and anti-WT1 [C19 (1:500); WT-180 (1:100); Santa Cruz Biotechnology).
Immunostaining.
Immunoperoxidase staining was performed using the Labeled StreptAvidin Biotin kit from DAKO according to the manufacturer’s recommendations. After fixation in acetone or a 1:1 mixture of acetone and methanol for 10 min at -20°C and blocking of unspecific sites, sections were stained with the first antibody for 30 min followed by incubation with the biotinylated secondary antibody and the streptavidin-peroxidase reagent. Negative controls consisted of consecutive tissue sections of each case in which the primary antibody was omitted.
Neutralization of WT1 Antibody.
To demonstrate the specificity of the WT180 antibody in immunostaining, experiments were repeated after preincubation of the antibody with an excess of a purified WT1 protein for 1 h at room temperature. A fusion plasmid of GST and WT1 (26)
was used for bacterial expression, and purification of the GST-tagged recombinant WT1 protein was performed using GSTrap columns from Amersham Pharmacia according to the manufacturer’s recommendations.
Western Blot.
Total protein extracts were separated on 10% SDS-PAGE and transferred to a nitrocellulose membrane (Schleicher & Schuell). The proteins were visualized using enhanced chemiluminescence (Amersham Pharmacia Biotech) as described by the manufacturer’s recommendations. The membrane was sequentially probed with three primary antibodies: (a) WT180 (1:500; Santa Cruz Biotechnology); (b) WTC19 (1:1000; Santa Cruz Biotechnology); and (c) 
-tubulin (1:500; Sigma).
Cell Culture and Transient Transfection.
NIH3T3 and Cos-7 cells were cultured in DMEM (Biochrom) containing 10% FCS (Biochrom). For transfection, the cells were seeded at a density of 1–3 x 105 cells/coverslip and transfected according to the manufacturer’s (Roche Applied Science) recommendations by using 3 µl of FuGENE 6 and 1 µg of plasmid DNA. The culture medium was changed after 16 h, and the cells were maintained for another 48 h to allow optimal expression of the fusion protein. The transfected cells were examined under a Zeiss Axiophot fluorescence microscope.
Plasmid Construction.
By using different pairs of primers, three fragments were PCR amplified from a WT1 cDNA clone named CMV-WIT2F (kindly provided by Dr. Paul Baird, Australia) and cloned into pCRII from the TA cloning kit (Invitrogen). The primers used to generate fragments were as follows: (a) for the WT450 fragment, VS1a (5'-CGAATTCTATGGGCTCCGACGTGCGGGA-3') and VS5a (5'-CAGTCGACTCAAAGCGCCAGCTGGAGTTT-3'); (b) for the WT380 fragment, VS1a and VS4a (5'-CAGTCGACTCAACCTGTATGTCTCCTTTG-3'); and (c) for WT220, VS1a and VS3a (5'-CAGTCGACTCACCTCAGCAGCAAAGCCTG-3'). The amplicons were sequenced with IRD800-labeled M13 primers on a LICOR automatic sequencer to verify that no mutations had been introduced. All primers contained a start or a stop codon as well as EcoRI or SalI restriction site that permitted subsequent directional in-frame cloning into pEGFP-C1 (Clontech), generating the following plasmids: (a) pEGFP-WT450, representing the wild-type WT1 isoform containing exon 5 and KTS; (b) pEGFP-WT380, representing a truncated WT1 form with a stop codon in exon 9; and (c) pEGFP-WT220, representing a truncated WT1 form with a stop codon in exon 3.
|
Results |
|---|
|
TopABSTRACTIntroductionPatients and MethodsResultsDiscussionREFERENCES |
|---|
Taken together with our previous study (7) , we have now analyzed a total of 86 WT patients and found WT1 mutations/deletions in 21 cases (24%). The mutation frequency was significantly different among the various histological subtypes: mutations were seen in 63% of tumors of stromal-predominant histology; 14% of tumors of triphasic histology; and 6% of tumors of blastemal-predominant histology. Of the 21 mutations, 17 were present in the germline, and 3 of these were deletions in patients with Wilms’ tumor, aniridia, genital malformation, mental retardation syndrome. Six of 10 stromal tumors with WT1 mutations showed LOH of 11p13 markers, and 1 tumor had a second somatic mutation.
These results confirm that the presence of WT1 mutations correlates with the specific histological subtype of stromal-predominant or ILNR-like WTs.
Expression and Subcellular Localization of Mutant WT1 Transcripts and Proteins in WTs.
Using our previously described protocol (23)
, DNA, RNA, and protein were extracted simultaneously, allowing us to analyze the same material for WT1 mutations as well as for RNA and protein expression. To examine whether tumors with WT1 mutations express WT1, we first performed Northern blot analyses on WTs of different histology. As has been observed previously (21)
, tumors with blastemal-predominant histology expressed high levels of WT1, whereas stromal-predominant tumors expressed only low or undetectable levels of WT1. A Northern blot of different WTs probed for WT1 and actin is shown in Fig. 1A
. In the three stromal-predominant WTs 9274, 9177, and 9184, a WT1 transcript with the highest level can be seen in 9274, where a smaller amount of RNA was loaded when compared with the actin signal. It was estimated that a mRNA can be detected by Northern blots if it comprises 0.001% of the total mRNA, corresponding to 50 copies/cell (27)
. Therefore, these stromal-predominant tumors contain an amount of RNA within the normal range of a low abundant message. From the right tumor of patient 9200, which shows no signal on the Northern blot (Fig. 1A)
we have performed real-time RT-PCR to quantify WT1 expression level. The expression was 178-fold lower than that in blastemal-predominant tumor 9168 (data not shown). Because tumor 9200 contains a mixture of cells, some of which most likely do not express WT1, we can deduce that this tumor contains about 5 copies or less of WT1 RNA, which explains why, in this case, no transcript can be seen in Northern blots (Fig. 1A)
.
|
shows an example for patient 9318, where a constitutional mutation (C>A in exon 7) causes the loss of a BsiEI restriction site. The analysis was performed on tumor and blood DNA (DNA-PCR) and tumor RNA (RT-PCR) using two different pairs of primers. As seen in Fig. 1B
, right, the tumor DNA only shows the mutant allele resulting in a 235-bp undigested fragment of exon 7. Amplification of the cDNA and digestion with BsiEI leads to an undigested 771- and 720-bp fragment, containing or lacking exon 5, respectively (Fig. 1B
, left). This fragment represents the mutant transcript. A schematic drawing of the fragment sizes after digestion of the DNA-PCR and RT-PCR product is given below the figure.
Next we analyzed whether a truncated protein is synthesized from these mutant messages. By immunohistochemistry, we studied frozen sections for the presence of a wild-type protein using the WTC19 antibody directed against the COOH terminus and the presence of a truncated as well as a wild-type protein using the WT180 antibody against the NH2 terminus. As described previously (17
, 20)
WT1 staining was restricted to the nucleus of podocytes and their precursors in normal embryonic kidney (Fig. 1C
, NEK), demonstrating that both antibodies were specific. Although all six blastemal-predominant tumors showed nuclear staining with both antibodies (tumor 9345 in Fig. 1C
and Table 1
), no immunoreactivity was seen with the COOH-terminal antibody in stromal-predominant WTs (tumor 9200 in Fig. 1C
and Table 1
), demonstrating that no wild-type protein was synthesized. Using the NH2-terminal antibody, all five cases with nonsense mutations had cytoplasmic staining in cells with a skeletal muscle morphology. In Fig. 1C
, tumor 9200 shows a positive reaction with the WT180 antibody, and it can be seen that the cells, which were cut longitudinally, show the elongated morphology of muscle cells with small nuclei and staining of the cytoplasm. No reactivity was seen in the surrounding stroma. Most of these tumors had lost the wild-type allele, confirming that only a truncated mutant WT1 protein is synthesized. The specificity of the NH2-terminal WT180 antibody was examined by the use of three different batches and by preabsorption with a purified GST-WT1 fusion protein containing aa 1–181 before incubation on tissue sections. This lead to a greatly reduced staining (data not shown).
|
. No wild-type or truncated WT1 protein could be detected with either antibody in the stromal-predominant tumors (Fig. 1D, s)
. This indicates that the amount of mutant WT1 protein is below the detection limit of the Western blot method, which is not suitable to detect a protein of low abundance (29)
, specifically if the cells used for protein extraction represent a mixture of WT1-expressing and non-WT1-expressing cells. In Fig. 1C
, it can be seen that only a few cells in the stromal-predominant tumors are positive for WT1, further explaining why the truncated protein may not be visible in Western blot analysis.
Analysis of Subcellular Localization of Truncated WT1 Proteins in Cell Culture.
To support our findings of a cytoplasmic truncated WT1 protein in tumors, we studied the subcellular localization of EGFP-WT1 fusion proteins of different length after transient transfection in NIH3T3 and Cos-7 cells. The full-length construct contained a cDNA encoding all 450 residues of the WT1 protein including exon 5 and KTS and an NH2-terminal fusion of EGFP (EGFP-WT450). Two WT1 proteins truncated at aa residues 220 and 380 were constructed and designated EGFP-WT220 and EGFP-WT380, respectively (Fig. 2A)
. These corresponded to the naturally occurring mutations in patients 9184 and 9394. EGFP-WT220 lacked the entire ZF domain, whereas EGFP-WT380 lacked parts of ZF III and the entire ZF IV. The nuclear localization signal in ZF I was still present here. The constructs were verified by transient transfection in NIH3T3 cells and subsequent Western blot analysis (data not shown). In contrast to previous studies on subcellular localization of WT1 proteins (12
, 30)
, we have directly visualized the EGFP-WT1 fusion proteins in native cells under a fluorescence microscope without fixation. Whereas the wild-type EGFP-WT450 fusion protein was restricted to the nucleus in both cell types, the deletion mutant EGFP-WT380 showed nuclear as well as cytoplasmic distribution (Fig. 2B)
. In contrast, the EGFP-WT220 fusion protein was mostly excluded from the nucleus in NIH3T3 cells and was present in the nucleus and the cytoplasm in Cos-7 cells. Fig. 2C
summarizes the results.
|
).
Analysis of Kidney Differentiation Markers.
To determine whether ILNR-like, stromal-predominant WTs with WT1 mutations can be distinguished from PLNR-like, blastemal-predominant WTs by their differentiation stage, we have studied a set of markers. Markers for uninduced and early-induced mesenchyme were vimentin, collagen I, and fibronectin, and markers for condensed mesenchyme were cytokeratin, laminin A, and collagen IV (31)
. Using this set, we could show that cells in both tumor subtypes correspond to early-induced mesenchyme but express one marker of uninduced mesenchyme. The organoid structure of cell clusters in blastemal tumors and the beginning of cytokeratin expression demonstrates that this subclass is already one step further toward epithelial differentiation. Fig. 3
shows a representative example of one stromal-predominant tumor (tumor 9200) and one blastemal-predominant tumor (tumor 9345). Cytokeratin staining is negative in tumor 9200 and is seen only within blastemal clusters of tumor 9354. Vimentin staining is seen in stromal cells of both tumors and in blastemal clusters of tumor 9345, whereas muscle cells in tumor 9200 are negative (cluster of blue-stained cells in the middle of the picture). Staining for the extracellular matrix proteins collagen IV and fibronectin, as well as the membrane protein laminin A, surrounds the entire blastemal clusters in blastemal-predominant tumors, except in one case, where the individual tumor cells were positive (data not shown). In contrast, in stromal-predominant tumor 9200, each individual cell was surrounded by positive staining. Collagen I staining is negative in both tumors.
|
summarizes the immunohistochemical results of five stromal-predominant tumors with WT1 mutations and six blastemal-predominant WTs without WT1 mutations as well as normal kidney specimens. Expression of markers was evaluated only in those cells that were positive for WT1 in both subclasses and compared with the expression of these markers as described in the kidney development database.4 |
Discussion |
|---|
|
TopABSTRACTIntroductionPatients and MethodsResultsDiscussionREFERENCES |
|---|
Our previous work and data presented in this report demonstrate that WT1 mutations occurred in a high precentage (63%) in one subset of WTs, the stromal-predominant or ILNR-like tumors. Importantly, 11 were present in the germline, and all but 1 were mutations resulting in protein truncation or loss of the entire gene (Ref. 7 ; data reported here). The germline mutations were found in three normal females, one normal male, and seven males with GUs of varying degrees. Interestingly, two males without GU had a tumor-specific mutation. Unilateral tumors developed in eight patients with germline mutations. An important clinical information from these data is that patients with germline mutations can present with unilateral tumors without associated abnormalities, and it should be kept in mind that these patients have a high risk for a contralateral tumor. Six of 10 (60%) stromal-predominant tumors had lost the wild-type DNA allele, and 1 tumor had a second somatic mutation, confirming Knudson’s two-hit inactivation model for WT1 in the development of this tumor subclass. In contrast, only one missense mutation in exon 2 was found in 16 blastemal-predominant WTs, suggesting that WT1 mutations play a minor role in the development of this tumor subtype.
Mutant WT1 mRNA was found in stromal-predominant WTs, and in cases where LOH occurred, only mutant transcripts were expressed. To our knowledge, only one other group found a mutant transcript in a tumor with a WT1 alteration (32) . In quantitative RT-PCR and Northern blot experiments, we have determined that the amount of the mutant transcript in different stromal-predominant tumors is between 5 and 50 copies/cell, which is in the normal range of low abundant messages (27) . This could correspond to the amount found in early-induced cells, at a time when WT1 expression starts. Blastemal-predominant WTs contain about 180-fold more WT1 mRNA, which may correspond to the normal range in condensed, epithelialized cells.
In stromal-predominant WTs with a mutation, the WT1 protein could be detected by immunohistochemistry in individual cells only with an antibody that recognizes the NH2 terminus, but not the COOH terminus, demonstrating that no intact wild-type protein was present. The mutant form was aberrantly localized in the cytoplasm, resulting in the absence of a functional WT1 protein in the nucleus. This was confirmed by analyzing the subcellular localization of different EGFP-WT1 mutants in unfixed, native cells. It is interesting that Englert et al. (30) found a WT1 protein truncated at aa 326 localized exclusively in the nucleus of Saos-2 and U2OS cells, two embryonal tumors not derived from kidney. Bruening et al. (12) , who used the same cells as we did (NIH3T3 and Cos-7), also noted a cytoplasmic localization with some constructs. They have deduced that a second nuclear localization signal is present between aa 291 and 350, which is absent in our almost exclusively cytoplasmic construct WT220. Because all truncated proteins still contain the dimerization motif, it is possible that at a time where LOH has not yet occurred, the wild-type protein might be excluded from the nucleus by dimerization with the cytoplasmic truncated protein. This could be the first step in an abnormal differentation of early-induced cells. In contrast, blastemal clusters in blastemal-predominant tumors expressed wild-type WT1 protein in the nucleus. Therefore, one molecular distinction between these two subtypes of WTs is the presence of a nuclear wild-type WT1 protein in blastemal-predominant tumors and its absence in stromal-predominant tumors.
Interestingly, in stromal tumors, the truncated protein was restricted to cells with a rhabdomyomatous phenotype expressing desmin, MyoD, myogenin, and sarcomeric actin. It was described previously that tumors with WT1 mutations may aberrantly differentiate into skeletal muscle (19) . This suggests that the lack of nuclear WT1 protein in multipotential mesenchymal cells results in the failure of a proper epithelial differentiation and alternative differentiation programs; predominantly skeletal muscle as well as fat, cartilage, and bone are turned on. All these structures are not found during normal kidney differentiation and in blastemal-predominant tumors, suggesting that alternative programs are normally repressed. Whether WT1 is this repressor is not known. It should be noted, however, that all stromal-predominant tumors that we have studied were obtained after chemotherapy, and therefore rhabdomyomatous differentiation in response to chemotherapy cannot be excluded as has been proposed by Brisigotti et al. (33) . However, we have studied several untreated stromal-predominant WTs for the presence of muscle differentiation by the expression of desmin and found them positive, suggesting that at least some muscle differentiation is present before chemotherapy.5
During early induction of kidney differentiation, vimentin, collagen I, and fibronectin are expressed, and at the stage of condensation, cytokeratin, laminin A, and collagen IV are expressed (31) .4 A switch from fibronectin to cytokeratin occurs at the stage of mesenchymal epithelial conversion. In stromal-predominant tumors, the cells expressing mutant WT1 and showing a rhabdomyomatous phenotype are at the differentiation stage of condensed cells except for the expression of fibronectin, which is normally turned off at that stage. Cells expressing wild-type WT1 in blastemal-predominant tumors are positive for markers of early-induced and condensed mesenchyme and may be one step further in epithelial differentiation because they start to express cytokeratin. Laminin A and collagen IV staining was not seen around each individual cell as in the case of stromal-predominant tumors but was seen surrounding a cluster of blastemal cells, suggesting that these may be further along the development to a more organoid structure. The expression of laminin A increases when epithelial cell polarization begins (34) . It was shown that laminin A, collagen IV, and heparan sulfate proteoglycan first appear as spots on aggregating cells and then form a semicontinuous sheet on the surface of comma- and S-shaped bodies, suggesting a role in cell aggregation and later in epithelial differentiation (35) . Forced expression of WT1 in mesenchymal NIH3T3 cells leads to up-regulation of E-cadherin, cytokeratin, heparan sulfate proteoglycan, and collagen IV, suggesting that WT1 is a positive regulator of epithelial differentiation (36) . We have shown previously (37) that blastemal-predominant WTs expressing wild-type WT1 are positive for PAX-8, which is turned on when epithelial differentiation begins, further supporting the notion that tumor cells expressing wild-type WT1 are at an early stage of epithelialization.
Taken together, these data show that two distinct subclasses of WTs can be distinguished on the basis of the presence or absence of WT1 mutations and by their associated expression of specific differentiation markers. A similar subclassification was made on the basis of the presence or absence of LOI in PLNR-associated/blastemal-predominant WTs and ILNR-associated/stromal-predominant WTs by Ravenel et al. (5) . This will be the basis to distinguish these two subtypes in future studies, rather than simply placing all WTs into one category. By doing so, it is expected that many of the previously unresolved inconsistencies might be resolved.
Based on the present data, we propose a model for the development of these two distinct subclasses of WTs (Fig. 4)
. Tumor cells in blastemal-predominant/PLNR-like WTs correspond to an early stage of induction/conversion to an epithelial phenotype. Cells arrested at the proliferative stage of early-induced mesenchyme are blocked from further differentiation, possibly by continuous expression of WT1 and/or LOI of IGF2. In contrast, the first event leading to stromal-predominant/ILNR-like WTs is either a germline or early somatic truncation mutation in a multipotential kidney stem cell (Fig. 4
, nucleus marked in red). This is followed by a second hit in a cell, most likely after induction by the ureteric bud, resulting in complete absence of a functional nuclear WT1 protein. Mesenchymal stem cells lacking WT1 fail to epithelialize and stochastically differentiate into other mesenchymal lineages. This ultimately leads to a heterogeneous tumor containing multiple cell types at the time of diagnosis. Microdissection of different cell types in two stromal-predominant tumors revealed LOH in all, indicating that this genetic event occurred in a multipotential stem cell early in tumor development (data not shown). Therefore, cells with a normal stromal appearance are also tumor cells, as has been described previously (38)
; however, they are negative for the WT1 protein. In summary, the two-hit inactivation of WT1 is essential in stromal-predominant/ILNR-like WTs.
|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 Supported by the Elterninitiative "Kinderkrebsklinik e.V.," Düsseldorf. 
2 To whom requests for reprints should be addressed, at Institute of Human Genetics, University of Düsseldorf, Postfach 101007, D40001 Düsseldorf, Germany. E-mail: royer{at}uni-duesseldorf.de
3 The abbreviations used are: WT, Wilms’ tumor; ILNR, intralobar rest; GU, genitourinary malformation; PLNR, perilobar rest; LOI, loss of imprinting; LOH, loss of heterozygosity; GST, glutathione S-transferase; RT-PCR, reverse transcription-PCR; EGFP, enhanced green fluorescent protein; ZF, zinc finger; aa, amino acid(s).
4 http://www.ana.ed.ac.uk/anatomy/database.
5 V. Schumacher and B. Royer-Pokora, unpublished observations.
Received 11/ 4/02; accepted 2/10/03.
|
REFERENCES |
|---|
|
TopABSTRACTIntroductionPatients and MethodsResultsDiscussionREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  V. A. Schumacher, S. Jeruschke, F. Eitner, J. U. Becker, G. Pitschke, Y. Ince, J. H. Miner, I. Leuschner, R. Engers, A. S. Everding, et al. Impaired Glomerular Maturation and Lack of VEGF165b in Denys-Drash Syndrome J. Am. Soc. Nephrol., March 1, 2007; 18(3): 719 - 729. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Srivastava, B. Fuchs, K. Zhang, M. Ruan, C. Halder, E. Mahlum, K. Weber, M. E. Bolander, and G. Sarkar High WT1 Expression Is Associated with Very Poor Survival of Patients with Osteogenic Sarcoma Metastasis. Clin. Cancer Res., July 15, 2006; 12(14): 4237 - 4243. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C.-M. Li, C. E. Kim, A. A. Margolin, M. Guo, J. Zhu, J. M. Mason, T. W. Hensle, V. V.V.S. Murty, P. E. Grundy, E. R. Fearon, et al. CTNNB1 Mutations and Overexpression of Wnt/{beta}-Catenin Target Genes in WT1-Mutant Wilms' Tumors Am. J. Pathol., December 1, 2004; 165(6): 1943 - 1953. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Cell Growth & Differentiation |
undigested sample. C, immunostaining of WT1 in normal embryonic kidney (NEK), one blastemal-predominant tumor WT (Tumor 9345), and one stromal-predominant WT (Tumor 9200) using an antibody directed against the COOH terminus (WTC19) and one directed against the NH2 terminus (WT180). All final magnification, x400 . D, Western blot analysis of two blastemal-predominant tumors (NL and 9168) and four stromal-predominant tumors (9385, 9274, and 9318) using the two WT1 antibodies (WTC19 and WT180) as well as the internal control 
