Prognostic significance of Wilms tumor gene (WT1) mRNA expression in soft tissue sarcoma

Authors


Abstract

BACKGROUND

There have been several recent reports that Wilms tumor gene (WT1) mRNA is overexpressed in many types of neoplasms, and those results suggested that WT1 has oncogenic properties. The objective of the current study was to evaluate the prognostic significance of WT1 mRNA expression in patients with soft tissue sarcoma.

METHODS

Levels of WT1 mRNA expression were examined by quantitative, real-time reverse transcriptase-polymerase chain reaction analysis in frozen tissue samples from 52 patients with soft tissue sarcoma. Various clinicopathologic factors were analyzed along with the disease-specific survival rate for correlations with WT1 mRNA expression levels.

RESULTS

The levels of WT1 mRNA expression in a variety of soft tissue sarcomas were significantly greater compared with the levels in normal soft tissue samples (P = .0212). No significant correlation was observed between the level of WT1 mRNA expression and clinicopathologic factors, including gender, age, primary tumor site, tumor depth, tumor size, histologic grade, and distant metastasis at initial presentation. The disease-specific survival rate for patients with high WT1 mRNA expression levels was found significantly poorer compared with the rate for patients with low WT1 mRNA expression levels (P = .0182). Moreover, multivariate analysis indicated that a high WT1 mRNA expression level was an independent, adverse prognostic factor for disease-specific survival (hazards ratio, 2.6; P = .0488).

CONCLUSIONS

WT1 mRNA expression level can serve as a potent prognostic indicator in soft tissue sarcoma patients. Cancer 2006. © 2006 American Cancer Society.

The Wilms tumor gene (WT1) was identified first by positional cloning strategies from the tumors of children who developed Wilms tumor in the context of the Wilms tumor, aniridia, genitourinary abnormalities, and mental retardation (WAGR) syndrome.1, 2 The gene is inactivated in the germline of children with a genetic predisposition to Wilms tumor and in a subset of sporadic Wilms tumors. WT1 encodes a zinc-finger transcriptional factor that regulates target genes, some of which are related to cell differentiation, proliferation, and apoptosis, such as platelet-derived growth factor A, insulin-like growth factor-I receptor, c-myc and bcl-2.3 Under normal conditions, WT1 gene expression is observed in a limited set of tissues, including gonad, uterus, kidney, and mesothelium.4–6

The WT1 gene originally was categorized as a tumor suppressor gene; however, several recent studies have suggested that WT1 plays an important role as a tumor promoter for many types of neoplasms. First, WT1 was highly expressed in leukemic blast cells7, 8 and in breast cancer,9, 10 lung cancer,11 ovarian cancer,12 mesothelioma,13 renal cell carcinoma,14 and bone and soft tissue sarcomas.15 Second, high expression levels of WT1 mRNA clearly were correlated with a poor prognosis in patients with acute leukemia7 or breast cancer10 and with high tumor stage in patients with testicular germ cell tumors.16 Finally, treatment with WT1 antisense oligomers specifically caused growth inhibition in leukemic blast cells17, 18 and in cell lines derived from lung cancer, gastric cancer, colon cancer, and breast cancer.19, 20 Although the mechanism of aberrant expression and the biologic function of WT1 are not understood fully, those reports exposed an apparently dual nature of WT1: It is a tumor suppressor gene and has oncogenic properties in certain types of malignancies.

Soft tissue sarcomas are relatively rare mesenchymal malignant tumors that represent approximately 1% of adult malignancies and 15% of pediatric malignancies.21 Except for our previous preliminary report,15 the expression of WT1 mRNA in soft tissue sarcomas has not been examined. In the current study, we quantified WT1 mRNA expression levels in a variety of soft tissue sarcomas by real-time, quantitative, reverse transcriptase-polymerase chain reaction (PCR) analysis, and the results demonstrated frequent overexpression of WT1 mRNA in soft tissue sarcoma tissues and a significant correlation between high WT1 mRNA expression level and a poor prognosis for patients with soft tissue sarcoma.

MATERIALS AND METHODS

Tissue Samples and Patient Characteristics

Between September 1988 and March 2002, 118 soft tissue sarcoma samples were obtained at the Departments of Orthopedic Surgery, Osaka University Medical Hospital and Osaka Medical Center for Cancer and Cardiovascular Diseases, Japan. Fresh tumor tissue samples that were obtained from the patients were snap frozen in liquid nitrogen immediately after biopsy or surgical resection and stored at −80°C until use. Among these 118 samples, we chose 52 primary soft tissue sarcoma samples for investigation from which we could extract adequate amounts and quality of RNA for real-time PCR analysis. Of these 52 samples, 40 samples were obtained from biopsy specimens before any treatment, and the remaining 12 samples were from surgical specimens after preoperative chemotherapy. The histologic diagnoses of these samples are summarized in Table 1. Histologic grade was determined by using 3 histopathologic factors—evaluation of proliferative activity of the tumors, cellularity, and tumor necrosis—according to the criteria of the Osaka University grading system for soft tissue sarcomas, as described previously.22 Proliferative activity of the tumors was estimated by using silver staining for nuclear organizer regions (AgNOR) and/or Ki-67 (MIB-1) immunohistochemistry. Briefly, 1 point each was given for high proliferative activity (mean AgNOR count >7.00 per nucleus and/or Ki-67 labeling index >20%), for high cellularity, and for a >15% area of necrosis. The points were added up, and the tumors were divided into 3 histologic grades: Grade 1 (point 0), Grade 2 (point 1), and Grade 3 (points 2 and 3). Histologic Grades 1 and 2 were defined as low grade, and Grade 3 was defined as high grade in subsequent survival analyses. Normal tissue samples that were used as controls in this study were obtained from patients who had undergone surgical treatment for osteoarthritis or for benign bone or soft tissue tumors. Normal tissue samples included 5 fat tissues, 5 muscle tissues, and 3 synovial tissues. Written informed consent was obtained from each patient.

Table 1. Histologic Diagnosis
DiagnosisNo. of Patients
  • *

    Including 7 patients with myxoid liposarcoma, 1 patient with pleomorphic liposarcoma, and 2 patients with well differentiated liposarcoma.

Malignant fibrous histiocytoma14
Synovial sarcoma10
Liposarcoma10*
Malignant peripheral nerve sheath tumor5
Leiomyosarcoma5
Rhabdomyosarcoma2
Angiosarcoma2
Clear cell sarcoma1
Extraskeletal osteosarcoma1
Primitive neuroectodermal tumor1
Dermatofibrosarcoma protuberans1

The patients ranged in age from 12 years to 89 years (median, 58 years) and comprised 26 males and 26 females. The primary tumors were located in the upper extremities in 8 patients, the lower extremities in 37 patients, and the trunk in 7 patients. All but 1 of 52 patients underwent various kinds of definitive surgical treatment. Among these, 42 patients achieved adequate wide local excisions, and 9 patients underwent inadequate marginal or intralesional excisions. Adjuvant therapy in these 52 patients included chemotherapy in 30 patients, radiotherapy in 5 patients, and combined chemotherapy and radiotherapy in 2 patients.

Follow-up was calculated from the date of initial definitive surgery or preoperative chemotherapy to the date of either last follow-up or death. The median follow-up for survivors was 45 months (range, 9-205 months). During follow-up, 6 patients developed local recurrences, and 15 distant metastases developed in patients who did not have distant metastasis at presentation. Nineteen of 52 patients died because of their sarcomas, and 2 patients died from other causes.

RNA Extraction and Reverse Transcription

Total RNA was isolated from the sample tissues using Isogen (Nippon Gene, Toyama, Japan) according to the manufacturer's instructions. Six micrograms of total RNA were mixed with oligo (dT)12-18 primer and incubated at 70°C for 10 minutes. Then, complementary DNA (cDNA) was synthesized by using Superscript II (Life Technologies, Inc., Rockville, MD) at 42°C for 40 minutes followed by heating at 70°C for 15 minutes.

Real-Time PCR

To determine relative WT1 mRNA expression levels in the sarcoma samples, 1.5 μL cDNA were added to PCR buffer (100 mM Tris-HCl [pH 8.3], 500 mM KCl; and 3 mM MgCl2) that contained 250 μM of each dinucleotide triphosphate, 1.25 U AmpliTaq Gold (PE Applied Biosystems, Foster City, CA), 0.5 μM forward and reverse primers, and 200 nM TaqMan probe in a total volume of 25 μL. The sequences of primers and probes used were as follows: WT1 forward primer, 5′-GATAACCACACAACGCCCATC-3′; WT1 reverse primer (R1), 5′-CACACGTCGCACATCCTGAAT-3′; WT1 probe, 5′-FAM-ACACCGTGCGT GTGTATTCTGTATTGG-TAMRA-3′; and β-actin forward primer, 5′-CCCAGCACAATGA AGATCAAGATCAT-3′; β-actin reverse primer, 5′-ATCTGCTGGAAGGTGGACAGCGA-3′; β-actin probe, 5′-FAM-TGAGCGCAAGTACTCCGTGTGGATCGGCG-TAMRA-3′. After activation of AmpliTaq Gold polymerase at 95°C for 10 minutes, PCR was performed for 40 cycles (at 95°C for 30 seconds and at 63°C for 60 seconds). WT1 reverse primer and β-actin forward primer contained sequences that spanned from exon 6 to exon 7 and from exon 4 to exon 5 of their respective genes to avoid amplification of the corresponding genome sequences. Standard curves for the quantification of WT1 and β-actin were constructed from the results of simultaneous amplification of serial dilutions of the cDNA from constantly WT1-expressing K562 leukemic cells, in which the WT1 mRNA expression level was defined as 1.0, as described previously.7WT1 mRNA expression levels in various types of sarcomas were determined according to the standard curves. Real-time PCR and subsequent calculations were performed on an ABI Prism 7700 Sequence Detector System (PE Applied Biosystems). To normalize the difference in RNA degradation and in RNA loading for reverse transcriptase-PCR in individual samples, the levels of WT1 gene expression divided by the levels of β-actin gene expression were defined as relative WT1 mRNA expression levels in the samples.

Immunoblotting

The frozen specimens were homogenized on ice and lysed in buffer that contained 4% sodium dodecyl sulfate (SDS); 125 mM Tris-HCl (pH 6.8), 10% glycerol; and 10% β-mercaptoethanol. After boiling for 5 minutes, debris was removed by centrifugation at 14000 × g for 15 minutes. The protein samples then were subjected to SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose membrane (Bio-Rad Laboratories, Hercules, CA). The membranes were blocked with 3% bovine serum albumin (BSA) in phosphate-buffered saline-0.1% Tween20 (PBS-T) for 1 hour at room temperature and incubated with anti-WT1 (DakoCytomation, Glostrup, Denmark) or anti-actin (Chemicon, Temecula, CA) mouse monoclonal antibody diluted in PBS-T that contained 1% BSA overnight at 4°C. The membranes were washed with PBS-T and incubated with alkaline phosphatase-conjugated goat antimouse immunoglobulin G (Promega, Madison, WI) in PBS-T for 1 hour at room temperature. After washing with PBS-T, blots were developed using mixture of nitro blue tetrazolium (Promega) and 5-bromo-4-chloro-3-indolyl-phosphate (Promega) in alkaline phosphatase buffer (100 mM Tris-HCl, pH 9.0; 150 mM NaCl; and 1 mM MgCl2).

Statistical Analysis

The correlation between WT1 mRNA expression levels in soft tissue sarcomas and in normal tissues was estimated using the Mann–Whitney U test. The correlation between the WT1 mRNA expression level and various clinicopathologic variables that are known prognostic factors was assessed by using the Fisher exact test. Cumulative disease-specific survival curves were calculated by using the Kaplan–Meier method, and the differences were estimated with the log-rank test in univariate survival analysis. A Cox proportional hazards model was used for multivariate survival analysis to assess the independent prognostic significance of clinicopathologic variables, including WT1 mRNA expression level. P values <0.05 were considered statistically significant. We performed all statistical analyses using the JMP software package (version 5.1.1; SAS Institute Inc., Cary, NC).

RESULTS

Relative Expression Levels of WT1 mRNA in Various Types of Soft Tissue Sarcomas and Normal Soft Tissue

The relative WT1 mRNA expression levels in a variety of soft tissue sarcoma samples were quantified by using a real-time PCR assay and are represented as a comparison with the levels in the human leukemia cell line K562 (defined as 1.0) (Fig. 1). Whereas WT1 mRNA expression levels in soft tissue sarcoma tissues varied widely (from <1 × 10−5 to ≈3 × 10−1), all levels in normal tissue samples were <3 × 10−3. WT1 mRNA expression levels in soft tissue sarcoma samples were significantly greater than the levels in normal soft tissue samples (median, 2.2 × 10−3 vs. 8.7 × 10−5; P = .0212; Mann–Whitney U test). No clear correlation was observed between various histologic types and WT1 mRNA expression levels in the sarcoma samples (Fig. 2).

Figure 1.

Wilms tumor gene (WT1) mRNA expression levels are illustrated in soft tissue sarcomas and in normal soft tissues. The upper and lower limits of the boxes and the line across the boxes indicate the 75th percentile, the 25th percentile, and the median, respectively. The WT1 level in leukemia cell line K562 was defined as 1.0.

Figure 2.

Wilms tumor gene (WT1) mRNA expression levels are illustrated in various histologic types of soft tissue sarcomas. MFH: malignant fibrous histiocytoma; SyS: synovial sarcoma; LS: liposarcoma; MRNST: malignant peripheral nerve sheath tumor; LMS: leiomyosarcoma; RMS: rhabdomyosarcoma; AGS: angiosarcoma.

Expression of WT1 Protein in Soft Tissue Sarcoma Tissues

To verify the correlation between WT1 mRNA and its protein expression level, we chose 2 samples each with low and high WT1 mRNA expression levels and subjected them to immunoblotting using anti-WT1 antibody (Fig. 3). Only trace levels of WT1 protein were detected in the pair of samples that had low WT1 mRNA expression (Fig. 3, Lanes 1 and 2); conversely, relatively high expression of WT1 protein was observed in the pair of samples that had high WT1 mRNA expression (Fig. 3, Lanes 3 and 4), suggesting a positive correlation between WT1 mRNA expression and WT1 protein expression.

Figure 3.

Wilms tumor (WT1) protein expression is illustrated in soft tissue sarcomas. Tissue lysates were analyzed by immunoblotting using anti-WT1 antibody. WT1 mRNA expression levels quantitated by real-time reverse transcriptase-polymerase chain reaction are shown at the bottom. Histologic diagnosis; Lane 1, malignant fibrous histiocytoma (MFH); Lane 2, well differentiated liposarcoma; Lane 3, MFH; Lane 4, angiosarcoma.

Correlation between WT1 mRNA Expression Levels and Various Clinicopathologic Factors

Because the median WT1 mRNA expression level in soft tissue sarcoma samples was 2.2 × 10−3, we chose a cut-off value of 0.01 to divide the patients into those with high and low expression levels of WT1. According to this cut-off value, 17 patients had a WT1 level ≥0.01,and 35 patients had a WT1 level <0.01. The correlations between WT1 mRNA expression levels and various clinicopathologic factors are shown in Table 2. No significant correlation was observed between the WT1 mRNA expression level and any clinicopathologic variable, including gender, age, site of the primary tumor, tumor depth, tumor size, histologic grade, distant metastasis at presentation, surgical treatment, radiation therapy, and adjuvant chemotherapy (Fisher exact test).

Table 2. Comparison of Clinicopathologic Factors with Wilms Tumor Gene mRNA Expression Level
FactorWT1 Level <0.01 (n = 35)WT1 Level ≥0.01 (n = 17)P Value
  • WT1: Wilms tumor gene.

  • *

    Tumor depth was defined relative to the superficial investing muscular fascia.

  • Tumor size was measured as the maximum tumor dimension (diameter).

  • Wide local excision.

  • §

    Marginal or intralesional excision or no surgery.

  • One patient received preoperative radiotherapy, and 4 patients received postoperative radiotherapy.

Gender   
 Male1511 
 Female206.2367
Age, y   
 <60216 
 ≥601411.1401
Primary tumor site   
 Extremity3114 
 Trunk43.6700
Tumor depth*   
 Superficial82 
 Deep2715.4668
Tumor size, cm   
 ≤541 
 >531161.0000
Histologic grade   
 Low146 
 High21111.0000
Distant metastasis at presentation   
 Negative2913 
 Positive64.7110
Surgical treatment   
 Adequate2913 
 Inadequate§64.7110
Radiation therapy   
 Yes32 
 No3214.6434
Adjuvant chemotherapy   
 Yes237 
 No1210.1359

Univariate and Multivariate Analyses of Disease-Specific Survival According to WT1 mRNA Expression Levels and Clinicopathologic Variables

To investigate the prognostic implication of various clinicopathologic factors and WT1 mRNA expression levels in patients with soft tissue sarcoma, we calculated cumulative disease-specific survival rates by using the Kaplan–Meier method. Figure 4 shows that patients who had low WT1 mRNA expression levels had a significantly better disease-specific survival compared with patients who had high levels (P = .0182; log-rank test). The results of univariate analysis for disease-specific survival according to various clinicopathologic factors, including WT1 mRNA expression levels, are summarized in Table 3. Among other clinicopathologic factors, age <60 years (P = .0295), low histologic grade (P = .0357), and no distant metastasis at presentation (P<.0001) also were significantly favorable prognostic factors. Gender (male vs. female), primary tumor site (extremity vs. trunk), tumor depth (superficial vs. deep), and maximal tumor size (≤5 cm vs. >5 cm) did not correlate with disease-specific survival in the current series. We also performed a multivariate analysis for disease-specific survival by using a Cox proportional hazards model, which demonstrated that the WT1 mRNA expression level was a significant prognostic factor (hazards ratio, 2.6; P = .0488) (Table 4), independent of other prognostic factors, including age, histologic grade, and distant metastasis at presentation.

Figure 4.

This graph shows disease-specific survival curves according to Wilms tumor gene (WT1) mRNA expression levels.

Table 3. Disease-Specific Survival Rate According to Clinicopathologic Factors
Factor5-Year Survival Rate (%)P value
  1. NS: not significant; WT1: Wilms tumor gene.

Gender  
 Male45 
 Female60.0922 (NS)
Age, y  
 <6061 
 ≥6047.0295
Primary tumor site  
 Extremity58 
 Trunk0.3159 (NS)
Tumor depth  
 Superficial75 
 Deep46.1240 (NS)
Tumor size, cm  
 ≤560 
 >551.9077 (NS)
Histologic grade  
 Low76 
 High38.0357
Distant metastasis at presentation  
 Negative62 
 Positive0<.0001
WT1 mRNA expression level  
 <0.0167 
 ≥0.010.0182
Table 4. Multivariate Analysis of Prognostic Factors for Disease-Specific Survival Rate
FactorHR (95% CI)P value
  1. HR: hazards ratio; 95% CI: 95% confidence interval; NS: not significant; WT1: Wilms tumor gene.

Age (<60 y vs. ≥60 y)3.3 (1.2-9.3).0176
Histologic grade (low vs. high)2.5 (0.8-9.3).1062 (NS)
Distant metastasis at presentation (negative vs. positive)2.7 (1.6-4.8).0005
WT1 mRNA expression level (<0.01 vs. ≥0.01)2.6 (1.0-6.7).0488

DISCUSSION

In the current study, we observed that WT1 mRNA expression levels in soft tissue sarcomas were significantly greater than the levels in normal soft tissues. In our previous preliminary study, we demonstrated that all 6 different types of soft tissue sarcomas examined had no mutations in any of the 10 exons of the WT1 genomic DNA,15 suggesting that the proportion of WT1 gene mutation in soft tissue sarcomas may be low. Therefore, the wild-type WT1 gene may act as a tumor promoter for soft tissue sarcomas, as proposed previously for other malignancies.7, 8, 11, 17, 20

The expression levels of WT1 mRNA in soft tissue sarcoma samples varied widely. This may be contributed to in part by the heterogeneity of the origins of the soft tissue sarcomas. For example, WT1 mRNA was expressed highly in both angiosarcomas (2.5 × 10−1 and 5.2 × 10−2). A recent article reported that WT1 expression was increased in vascular smooth muscle cells and vascular endothelial cells after myocardial infarction.23 In addition, in rhabdomyosarcoma, an effect of WT1 on the tumor phenotype was deemed highly likely. Miyagawa et al.24 reported that high levels of myogenic gene expression were observed in Wilms tumor samples in which homozygous WT1 mutation was documented, and overexpression of the WT1 gene in C2 myoblasts resulted in regression of muscle differentiation. Thus, the low WT1 expression levels in the current series in 2 rhabdomyosarcomas (both <10−5) may reflect the myogenic lineage of differentiation. These observations suggest that various mesenchymal cells may be affected by the WT1 gene in various stages of differentiation, leading to the possibility that the cellular origins of soft tissue sarcomas are correlated with WT1 mRNA expression levels. To clarify the trend of WT1 mRNA expression levels in individual histologic types of soft tissue sarcomas, a larger cohort study is warranted.

Previous studies have documented a correlation between high WT1 mRNA expression levels in tumor samples and a poor prognosis in patients with leukemia and breast cancer.7, 10 However, there have been no studies analyzing the prognostic significance of WT1 mRNA expression levels in patients with soft tissue sarcoma. Moreover, although a number of significant clinicopathologic prognostic factors have been reported,25–27 documentation of molecular prognostic variables for soft tissue sarcomas remains limited because of their relative rarity and histologically heterogeneous diversity. The current study showed for the first time that the WT1 mRNA expression level is a strong prognostic factor for patients with soft tissue sarcoma and was independent of other various clinicopathologic prognostic factors in a Cox multivariate survival analysis. These results suggest that the WT1 mRNA expression level may be a new parameter with which to predict the prognosis of patients with soft tissue sarcoma. In the current study, we used immunoblotting in 4 samples to determine that the WT1 mRNA expression level was correlated with the WT1 protein level; and we previously reported 3 samples with WT1 mRNA overexpression that also showed positive staining for WT1 protein in immunohistochemistry using formalin fixed, paraffin embedded sections15. Therefore, a larger scale investigation of the prognostic implications of WT1 expression levels, combined with other known clinicopathologic prognostic factors for soft tissue sarcomas, should be performed using archival soft tissue sarcoma samples. To determine whether the WT1 expression level provides additional prognostic information within a single tumor type, in which the other prognostic factors are matched, larger sets of individual tumor types matched in terms of a multifactorial grading system, such as the French or National Institutes of Health system,28, 29 would be needed.

The WT1 gene is spiced alternatively mainly at 2 sites (17AA and KTS) and yields 4 spliced variant forms, each of which is suggested to have different functions.30–32 Previously, we examined the ratio of the 4 WT1 spliced forms in several samples of soft tissue sarcoma and found that the 17AA(+)KTS(+) WT1 spliced variant was expressed dominantly among the 4 WT1 spliced forms (data not shown). Constitutive expression of the 17AA(+)KTS(+) WT1 spliced variant restores the growth inhibition induced by treatment with WT1 antisense oligomer in K562 leukemia17 and AZ-521 gastric cancer cells.20 Therefore, WT1 overexpression, and particularly of the 17AA(+)KTS(+) spliced variant, may stimulate the proliferation and/or inhibit apoptosis of sarcoma cells. Further study to analyze the functions of the various forms of WT1, especially in soft tissue sarcoma, is needed.

Several recent reports have indicated that WT1 protein is a tumor-associated antigen, and cytotoxic T lymphocytes can be elicited by immunization with 9-mer WT1 peptides.33 These WT1 peptides can bind to major histocompatibility complex Class I molecules with obvious antitumor activity in both in vivo murine models33, 34 and in vitro human colon cancer cell lines.35 These reports suggest that WT1 protein can serve as a target antigen for tumor-specific immunotherapy. In fact, the Osaka University Cancer Immunotherapy Group recently conducted a Phase I clinical trial of cancer immunotherapy targeting the WT1 protein for several human cancers36 and already is proceeding to a Phase II clinical trial using WT1 peptides for patients with various types of cancers, including bone and soft tissue sarcomas.

The results of the current study demonstrated that WT1 mRNA is overexpressed frequently in various types of soft tissue sarcomas and that the WT1 mRNA expression level is a significant prognostic indicator in patients with soft tissue sarcoma. These results strongly suggest that WT1 may be a candidate for a potent molecular marker to predict patient prognosis and a promising target as a tumor antigen for tumor-specific immunotherapy against soft tissue sarcomas.

Acknowledgements

The authors thank Ms. M. Okamoto (Department of Orthopedics, Osaka University Graduate School of Medicine) for her technical assistance in tissue preparation.

Ancillary