Clinical relevance of mutations in the Wilms tumor suppressor 1 gene WT1 and the cadherin-associated protein β1 gene CTNNB1 for patients with Wilms tumors

Results of long-term surveillance of 71 patients from International Society of Pediatric Oncology Study 9/Society for Pediatric Oncology

Abstract

BACKGROUND.

Mutations in the Wilms tumor (WT) suppressor 1 gene (WT1) and the cadherin-associated protein β1 gene (CTNNB1) are found predominantly in stromal type WT, defining a genetic subgroup. The clinical relevance of these mutations remains to be determined.

METHODS.

A long-term follow-up study was performed for 71 patients (International Society of Pediatric Oncology Study 9/Society for Pediatric Oncology; n = 77 tumors) with known molecular genetic status. Eight patients had bilateral disease, including 2 patients with a WT in both kidneys and 5 patients with a WT in 1 kidney and nephrogenic rests (NRs) in the other kidney. The response to preoperative chemotherapy, relapses, metastases, metachronous tumor development, and deaths were evaluated with a median follow-up of 12 years and 4 months.

RESULTS.

Nineteen patients (n = 24 tumors) had WT1 mutations, and 16 were constitutional mutations. Three patients with germline mutations had second tumor events: Two patients developed a WT in the kidney with NRs 3 years and 11 years after the first tumor; and 1 patient developed second tumors after 2 years, 1 in the kidney with a previous WT and 1 in the kidney with a previous NR. Eighteen of the WT1 mutant tumors were analyzed for CTNNB1 mutations, and all had mutations. A poor volumetric response (progression and <50% reduction) was observed in all patients who had tumors with a WT1 mutation and in 23 of 52 nonmutant tumors.

CONCLUSIONS.

Patients with WT1 germline mutations had an increased risk for bilateral disease and second tumor events. Therefore, the authors concluded that tumor surveillance until adulthood should be considered. Although tumors with both WT1 and CTNNB1 mutations had a poor volumetric response, there was no significant difference in overall survival in this cohort of patients with and without WT1 mutations. Cancer 2008. © 2008 American Cancer Society.

Wilms tumor (WT) is the most common malignant neoplasm of the kidney in childhood, and it is believed that WT arises from embryonic undifferentiated renal mesenchyme. WT provides one of the most impressive examples of success in the treatment of childhood cancers. In Europe, patients are treated according to the International Society of Pediatric Oncology (SIOP) nephroblastoma therapeutic protocols, with preoperative chemotherapy followed by nephrectomy and a risk-adapted postoperative chemotherapy for patients aged >6 months.1 For treatment decisions, the revised SIOP working classification is used taking into consideration the histologic features remaining after preoperative chemotherapy.2 Additional radiotherapy is performed only in patients who have adverse prognostic features. The 5-year overall survival rate for patients registered in the German SIOP-9/Society for Pediatric Oncology (GPOH) study is 89.5%.3 Long-term complications resulting from chemotherapy and radiotherapy include cardiotoxicity, second malignancies, reduced functioning of the remaining kidney or renal failure, vertebral damage because of radiation, and neurotoxicity.4 A long-term study of adult survivors of childhood cancer has revealed an increased incidence of chronic health conditions caused by the previous chemotherapy and radiotherapy.5 Therefore, the identification of low-risk patients who would benefit from a reduced therapy, resulting in a better quality of life. is an important goal for the future.

In the SIOP-9/GPOH study, Weirich et al6 have observed a poor volumetric response to preoperative chemotherapy in differentiated tumors, such as stromal or epithelial type tumors. Resistance or poor response to chemotherapy also was described for the rare histologic variant of WT, the fetal rhabdomyomatous nephroblastoma.7-9 In addition, a poor response was observed in patients with bilateral WTs and rhabdomyomatous histology.10 We and others have described that mutations in the WT1 gene are observed predominantly in stromal type tumors with rhabdomyomatous differentiation, suggesting that tumors with WT1 mutations may belong to the category of nonresponders.11-16

Although a poor volumetric response is observed, chemotherapy may have an effect on the tumor cells. After chemotherapy, some tumors consist mostly of differentiated elements, indicating that these are unaffected. In contrast, immature elements like blastemal cells are observed less frequently in tumors after treatment, pointing to their sensitivity.6, 17 Chemotherapy either may eradicate immature (dividing) cells, leaving mature (nondividing) cells intact, or induce the embryonic cells to differentiate. Currently, it is unknown whether differentiated tumor cells may be able to revert to a proliferative state.

The tumor volume reduction during preoperative chemotherapy is a first indication of responsiveness, but this does not allow conclusions about the clinical outcome of these patients. Because we have analyzed a large number of patients/tumors from the SIOP-9/GPOH trial for WT1 and CTNNB1 mutations, we evaluated the clinical relevance of the presence or absence of mutations in WT1 and CTNNB1 by conducting a new follow-up study in 2006.

MATERIAL AND METHODS

Patients and Treatment

The SIOP-9/GPOH study enrolled 440 patients from January 1989 until March 1994. Frozen tumor samples were obtained from 93 patients who received preoperative chemotherapy. In 2 patients with bilateral WT, tumors from both sides were obtained. The accompanying molecular studies were approved by the local ethics committees (August 31, 1989; Heidelberg, Germany), and written consent was obtained from all parents.

WT1 and CTNNB1 Mutation Analysis

Overall, 67 frozen tumor samples from 65 patients and blood samples from 6 additional syndromic patients were analyzed for WT1 mutations, and those 71 patients were included in this study (details are given in Fig 1, top). For CTNNB1, there were no blood samples, but all frozen tumor samples plus paraffin sections were available for 3 of the 6 additional patients; ie, 70 tumors were analyzed because mutations in CTNNB1 are tumor-specific (for details, see Fig 1, bottom).

Figure 1. (Top) Seventy-one patients were analyzed for the Wilms tumor (WT) suppressor 1 gene (WT1). WT1 was analyzed in DNA isolated from 65 frozen tumors and from 6 blood samples in patients without tumor material. The results of the mutation analysis and details of the patients regarding sex and syndromes are indicated. An asterisk indicates a girl with a double kidney. BWS indicates Beckwith-Wiedeman syndrome; GU, genitourinary abnormalities; DDS, Denys-Drash syndrome; WAGR, Wilms tumor, aniridia, genitourinary tract abnormalities syndrome. (Bottom) Seventy tumor samples were analyzed for mutations of the cadherin-associated protein β1 gene (CTNNB1). Three of 61 unilateral tumors that were analyzed were from paraffin samples obtained specifically for CTNNB1 mutation analyses. In these patients, the corresponding WT1 analysis was performed on blood DNA samples.

Genomic DNA samples that were isolated from tumors or blood samples from 15 new patients were studied for all 10 exons of WT1 by single-strand conformation polymorphism, as described previously.11 In eight tumors, all 10 WT1 exons and flanking introns were sequenced from tumor DNA (Patients WTHD4, WTHD10, WTHD22, WTHD23, WTHD24, WTHD28, WTHD45, and WTHD52). The WT1 mutation analysis for 48 pretreated tumors was published previously.11, 15, 18 Exon 3 of the CTNNB1 gene was analyzed from the same tumor DNA that was used for WT1 mutation analysis—except that, in 3 patients, DNA was isolated from paraffin sections, as described previously.19

Tumor Histology

Representative tumor material was sent for central review to the Kiel Pediatric Tumor Registry (I.L.). For histologic subtyping, the revised SIOP working classification was used taking into account the treatment-induced changes.2 This classification distinguishes 3 tumor risk groups: completely necrotic (low risk), blastemal and diffuse anaplasia (high risk), and others (intermediate risk).

Tumor Volume

Diagnostic imaging was performed as described by Weirich et al.6 Tumor volumes were measured sonographically in 3 dimensions using the ellipsoid formula (length × thickness × depth × 0.523). For large tumors that could not be measured by ultrasound, other methods, such as computed tomography (CT) or magnetic resonance imaging (MRI), were used. The same method was used for the first and follow-up tumor volume determination. To minimize interobserver variability, tumor volume measurements of the participating centers were reviewed by the central diagnostic reference group by using the original images. The World Health Organization guidelines were followed for tumor volume measurements. The group of clinically poor responders included patients with tumor enlargement (progression), no reduction, and tumor shrinkage by <50%. A good response was defined as a volume reduction >50%.

Follow-up

Follow-up status was determined by using specific forms that were sent every 6 months until at least 5 years after diagnosis to report recurrences and any tumor-related or nontumor-related events. One of the authors (A.W.) updated these data in 2006 by contacting the participating centers or individual patients.

RESULTS

WT1 and CTNNB1 Mutations

In our first consecutive series of WT1 analyses, we observed that stromal type tumors had a higher percentage of WT1 mutations.11 To confirm this observation, we analyzed all tumors of stromal histology first and then analyzed the tumors with other histologic subtypes. To date, 20 of 20 stromal tumors (100%), 9 of 10 blastemal tumors (90%), 14 of 21 mixed type tumors (67%), 21 of 38 tumors with regressive histology (55%), 2 of 3 epithelial tumors, and 1 of 2 anaplastic tumors were analyzed. The analysis of all tumor samples will be completed in the near future. In addition, blood DNA samples from 6 syndromic patients also were analyzed for WT1 mutations. Nineteen patients had 24 tumors with WT1 mutations, and 52 patients had 53 tumors without mutations. The top graph in Figure 1 summarizes the results of the WT1 analyses in all 71 patients. Eight patients had bilateral disease, and 1 of those patients had Beckwith-Wiedeman syndrome (BWS). In this study, we used the term ‘bilateral disease’ for patients who had bilateral tumors (synchronous bilateral WTs) and for patients with a tumor in 1 kidney and nephrogenic rests (NRs) in the other kidney. Seven patients with bilateral disease had germline WT1 mutations, and details on these patients are discussed below. The patient with BWS had no WT1 mutation.

All DNA samples that were isolated from frozen tumors also were studied for CTNNB1 mutations. In 2 patients with bilateral disease, tumor DNA from both sides was available and was studied for CTNNB1 mutations. In addition, paraffin sections from 3 tumors were analyzed. The bottom graph in Figure 1 describes the analysis of samples for CTNNB1 mutations. Seventy tumors were analyzed for CTNNB1 mutations, and 23 tumors (33%) had mutations identified.

Sixty-three patients presented with unilateral WT, and 12 of those patients had a WT1 mutation; in 3 patients, these mutations were tumor-specific. Three new constitutional WT1 mutations were identified: 1 in a boy who had a unilateral WT without genitourinary abnormalities; (Patient WTHD7; p172fs); 1 in a boy with bilateral WT without genitourinary abnormalities; (Patient WTHD10; c423+1G→A;), and 1 in a girl with unilateral WT (Patient WTHD4; p.S223N). The mutations in Patients WTHD4 and WTHD10 were detected only by sequencing.

Response to Preoperative Chemotherapy

The reduction in tumor volume after preoperative chemotherapy was correlated with the presence or absence of WT1 mutations. All 24 tumors/ILNRs from the 19 patients with WT1 mutations showed a poor response with either enlargement (n = 6), no reduction (n = 2), or a reduction between 3% and 48% (n = 16). Nineteen mutant WT1 tumors had stromal histology. Of the 53 tumors without WT1 mutations, 23 tumors had a poor response with progression observed in 7 cases.

When correlating tumor-specific CTNNB1 mutations with response, 20 of 23 tumors had a poor response, and 3 of 23 tumors had a good response. In summary, the presence of a WT1 mutation was correlated significantly with a poor volumetric response to preoperative chemotherapy. Table 1 summarizes all data described in this report, including the WT1 mutations and the presence (+) or absence (−) of a CTNNB1 mutation.

Table 1. Clinical and Molecular Characteristics of 71 Study Patients

Case	Age at DX, mo	Stage*	Histology	% Muscle in Stroma	Regression	Tumor Volume, mL		Volumetric Reduction, %	Response (50% Reduction)	Follow-up		WT1 Mutation†	WT1 Type†	CTNNB Mutation
						Before CHT	After CHT			Status	Months
DX, indicates diagnosis; CHT, chemotherapy, WT1, Wilms tumor-suppressor gene 1; CTNNB, cadherin-associated protein β1 gene; WTHD, study patients; S, stromal; (L), left; (R), right; 1CR, first complete remission; +, positive; NA, not analyzed; ILNR, intralobal nephrogenic rests; S, stromal; M, mixed; 2CR, second complete remission; B, blastemal; Regr, regressive; Epit, epithelial predominant; Anapl, anaplastic. * Stage V is not given for bilateral disease, but local stage is listed. † In the list of mutations, (P) indicates polymorphism; (C), constitutional; (T), tumor-specific. ‡ Lost to follow-up after the time indicated. § Because of the high percentage of regressive changes, these tumors should be in the “good” category; however, if the tumor volume alone is used, then they belong to the group of poor responders. ‖ This patient had a P281S missense mutation that, in the interim, was identified as a rare polymorphism. ¶ Patients where only blood DNA was analyzed for WT1 mutations; no frozen tumor samples were available.
Patients with WT1 mutations
WTHD1a	7	Local I	S (L)	70	0	104	109	Enlarged	Poor	1CR	120	p.S313X	(C)	+
WTHD1b		Local I	S (R)	70	3	486	759	Enlarged	Poor					NA
WTHD2a	15	Local I	S (L)	60	5	280	300	Enlarged	Poor	1CR	181	p.S122fs	(C)	+
WTHD2b		Local I	ILNR (R)	75	60	2	2	0	Poor					+
WTHD3a¶	11	Local I	M (L)	80	5	635	564	11	Poor	1CR	139	p.R362X	(C)	NA
WTHD3b		Local III	ILNR (R)	70	30	22	45	Enlarged	Poor					NA
WTHD4¶	10	Local I	S (R)	90	45	402	1092	Enlarged	Poor	2CR (bilateral disease)	148	p.S223N	(C)	NA
WTHD5¶	24	II	S	90	40	243	243	0	Poor	1CR	168	p.R390X	(C)	+
WTHD6	9	Local I	S	90	40	165	167	Enlarged	Poor	Death (progress of bilateral disease)	40	c.849+1G→C	(C)	+
WTHD7	14	I	S	90	20	1384	1346	3	Poor	1CR	120	p.P172fs	(C)	+
WTHD8	29	III	S	80	35	352	341	3	Poor	1CR	148	p.G253A	(C)	+
WTHD9	29	I	S	65	30	393	377	4	Poor	1CR	186	p.E272fs	(T)	+
WTHD10a	14	Local I	S (L)	15	1	357	286	20	Poor	1CR	165	c.422+1G→A	(C)	+
WTHD10b		Local III	S (R)	10	3	220	212	4	Poor					NA
WTHD11a	3	Local III	ILNR (L)	0	25	10	9	10	Poor	2CR (local relapse)	164	p.A10fs	(C)	+
WTHD11b		Local I	S (R)	25	10	680	420	38	Poor					+
WTHD12	20	I	S	70	5	186	166	11	Poor	1CR	26‡	del 11p13	(C)	+
WTHD13	7	III	S	70	10	398	347	13	Poor	1CR	172	p.A307fs	(T)	+
WTHD14	12	I	S	60	0	893	713	20	Poor	1CR	184	p.S136fs	(C)	+
WTHD15¶	7	I	S	70	55	301	238	21	Poor	1CR	57‡	c.894+2T→G	(C)	NA
WTHD16	34	II	S	30	30	33	26	21	Poor	1CR	161	p.S313X, p.372fs380X	(C) (T)	+
WTHD17¶	36	I	S	90	30	595	354	41	Poor	1CR	140	p.C385R	(C)	+
WTHD18¶	19	I	Regr	NA	80	245	140	43	Poor	1CR	153	p.R394W	(C)	+
WTHD19	12	II	S	80	25	680	352	48	Poor	1CR	147	p.R301X	(T)	+
Patients without WT1 mutations
WTHD20	49	IV	Regr	0	100§	293	837§	Enlarged	Poor§	1CR	136	–		–
WTHD21	58	IV	B	0	50	36	137	Enlarged	Poor	2CR (local relapse)	161	–		–
WTHD22	21	I	S	70	30	162	178	Enlarged	Poor	1CR	54‡	–		+
WTHD23	60	II	S	20	5	659	690	Enlarged	Poor	1CR	89	–		–
WTHD24	23	I	S	0	25	336	345	Enlarged	Poor	1CR	165	–		–
WTHD25	155	II	B	0	30	132	379	Enlarged	Poor	1CR	159	–		–
WTHD26	20	II	Regr	70	70	1387	1387§	0	Poor§	1CR	120	–		–
WTHD27	12	III	B	0	3	400	460	Enlarged	Poor	1CR	169	–		–
WTHD28	46	I	S	60	<5	184	160	13	Poor	1CR	91‡	–		–
WTHD29	27	III	Regr	0	75§	469	458	2	Poor§	1CR	120	–		–
WTHD30	153	IV	B	0	10	1224	1170	4	Poor	1CR	179	–		+
WTHD31	40	IV	Regr	0	95§	486	357	27	Poor§	2CR (lung metastases)	195	–		–
WTHD32	7	I	M	0	50	54	51	6	Poor	1CR	67§	–		–
WTHD33	61	I	Epit	0	<5	375	330	12	Poor	1CR	120	–		–
WTHD34	10	I	M	0	5	154	117	24	Poor	1CR	86‡	–		–
WTHD35	24	I	M	0	65	281	210	25	Poor	1CR	120	–		–
WTHD36	25	IV	S	0	<5	516	382	26	Poor	1CR	135	–		–
WTHD37	26	I	M	29	25	360	245	32	Poor	1CR	127	–		–
WTHD38	39	I	M	0	50	201	129	36	Poor	1CR	90‡	–		–
WTHD39	36	I	B	0	70	518	322	38	Poor	Death (lung+abdominal metastases)	21	–		–
WTHD40	47	II	B	0	60	621	376	39	Poor	2CR (lung metastasis)	155	–		–
WTHD41	114	II	Regr	0	93	685	378	45	Poor	1CR	201	–		–
WTHD42	59	II	Regr	0	80	164	88	46	Poor	1CR	112‡	–		–
WTHD43	24	III	Regr	0	95	378	174	54	Good	1CR	166	–		–
WTHD44	24	I	Regr	0	93	108	46	57	Good	1CR	190	–		–
WTHD45	45	I	S	80	25	92	39	58	Good	1CR	166	–		+
WTHD46	36	IV	M	30	60	889	352	60	Good	1CR	163	–		+
WTHD47	38	III	Regr	0	75	572	226	60	Good	1CR	174	–		–
WTHD48	30	III	B	0	25	424	158	63	Good	1CR	143	– (P)‖		–
WTHD49	24	IV	M	5	5	523	183	65	Good	1CR	165	–		–
WTHD50	49	I	M	NA	<5	714	244	66	Good	1CR	33‡	–		–
WTHD51	31	IV	Regr	0	70	1163	390	66	Good	1CR	86‡	–		–
WTHD52	12	I	S	0	10	170	53	69	Good	1CR	154	–		–
WTHD53	45	IV	M	30	50	644	190	70	Good	2CR (lung metastases)	162	–		–
WTHD54	57	II	Regr	80	90	630	176	72	Good	1CR	126	–		–
WTHD55	106	II	M	0	55	816	225	72	Good	1CR	146	–		–
WTHD56	99	IV	B	0	50	635	176	72	Good	1CR	136	–		–
WTHD57	31	IV	Regr	90	15	151	43	72	Good	1CR	171	–		–
WTHD58	48	I	Regr	0	70	164	43	74	Good	1CR	136	–		–
WTHD59	111	II	Regr	0	75	524	130	75	Good	1CR	147	–		–
WTHD60	89	III	M	0	50	16	4	75	Good	1CR	149	–		–
WTHD61	43	I	M	0	20	243	58	76	Good	1CR	155	–		–
WTHD62a	61	V	Epit (L)	0	60	200	46	77	Good	1CR	162	–		–
WTHD62b		V	M (R)	5	70	214	45	79	Good					NA
WTHD63	46	I	Regr	20	75	411	81	80	Good	1CR	195	–		–
WTHD64	24	III	Anapl	5	30	335	65	81	Good	1CR	179	–		–
WTHD65	12	I	B	0	40	563	92	84	Good	1CR	149	–		–
WTHD66	41	II	B	0	40	490	57	88	Good	1CR	145	–		–
WTHD67	52	IV	Regr	0	75	1036	99	90	Good	Death (lung metastases)	19	–		–
WTHD68	42	III	Regr	0	100	1707	169	90	Good	1CR	159	–		–
WTHD69	40	IV	M	0	50	453	29	94	Good	2CR (lung metastases)	189	–		–
WTHD70	23	II	B	0	65	753	43	94	Good	1CR	120	–		+
WTHD71	60	I	Regr	0	75	228	7	97	Good	1CR	128	–		–

Long-term Survival

Next, we examined whether the clinical outcomes were different between patients in our cohort who had tumors with or without WT1 mutations. The clinical follow-up was ≥10 years for 58 patients, 10 patients were lost to follow-up between 2 years and 9 years, 1 patient was followed for only 7.4 years, and 3 patients died of the disease. Two of the deaths occurred in patients without a WT1 mutation because of local relapse and progression of lung metastases in 1 patient (WTHD39) and distant lung metastasis in another patient (WTHD67). One death occurred in a patient (WTHD6) with a germline WT1 mutation and bilateral disease.

Standard log-rank tests based on Kaplan-Meier estimates were used to analyze survival data. In the 19 patients with WT1-associated tumors, the 10-year event-free survival was 0.89 (95% confidence interval, 0.76-1.00), and the 10-year overall survival was 0.94 (95% confidence interval, 0.84-1.00). In the 52 patients without WT1 mutations, the 10-year event-free survival was 0.87 (95% confidence interval, 0.78-0.96), and the 10-year overall survival was 0.96 (95% confidence interval, 0.91-1.00). These data demonstrate that overall and event-free survival did not differ statistically between patients with and without WT1 mutations. Figure 2 shows the Kaplan-Meier curves for event-free survival (Fig. 2, top) and overall survival (Fig. 2, bottom) for both groups.

Figure 2. Kaplan-Meier curves for patients who had Wilms tumors with or without Wilms tumor suppressor gene 1 (WT1) mutations illustrate (Top) event-free survival and (Bottom) overall survival. CI indicates confidence interval.

Clinical Follow-up of Patients With Bilateral Disease and WT1 Germline Mutations

The disease/tumor management and clinical course of the 7 patients with bilateral disease and WT1 mutations are shown in Figure 3. Four patients were in first complete remission (Patients WTHD1, WTHD2, WTHD3, and WTHD10): Two of those patients had a tumor in both kidneys at the time of diagnosis (Patients WTHD1 and WTHD10), and both patients received 8 weeks of preoperative chemotherapy with vincristine (VCR) and actinomycin D (ACTD) and with doxorubicin was added during the last 4 weeks. Two patients had a tumor in 1 kidney and ILNR in the contralateral kidney (Patients WTHD2 and WTHD3). Patient WTHD2 was treated preoperatively for 7 weeks with ACTD and VCR and, because, of veno-occlusive disease, was not treated with ADR. After nephrectomy and removal of the ILNR, the patient was given ACTD, VCR, and ADR and remained in first complete remission for 15 years. Patient WTHD3 received 3 months of preoperative chemotherapy with VDR, ACTD, and ADR without volume reduction. Postoperative treatment for this patient included ADR, etoposide, ifosfamide, and carboplatin, and the patient remained disease free for up to 12 years.

Figure 3. The clinical course of 7 patients with bilateral disease and germline Wilms tumor (WT) suppressor 1 gene (WT1) mutation. The events and the clinical management of the 7 patients are indicated. A distinction is made between complete kidney removal, including the tumor (nephrectomy), and kidney-sparing tumor surgery (partial nephrectomy). The time from initial diagnosis to the second event and the length of follow-up (FU) are given. GU, genitourinary abnormalities; L, left; R, right; ILNR, intralobal nephrogenic rests; CR, complete response; NR: nephrogenic rests; GU, genitourinary abnormalities.

Three patients (WTHD4, WTHD6, and WTHD11) who had a tumor on 1 side and NRs or ILNRs in the contralateral kidney had second tumor events. Patient WTHD4 was treated preoperatively with VCR, ACT, and carboplatin as well as ADR for 7 weeks; during that time, the tumor increased in size. A biopsy of the suspicious area in the left kidney revealed benign NR, which was not removed and remained silent for 11 years. The patient was in regular surveillance; and, in 2005, the NR increased in size, and a tumor was diagnosed. Fortunately, at that time, the patient underwent kidney transplantation, and the WT could be removed.

Patient WTHD6, who had a unilateral WT at first diagnosis, was treated preoperatively for 4 weeks with ACTD and VCR and, during that time, the tumor increased in size. MRI revealed a suspicious area in the contralateral kidney. The tumor was removed, and the suspicious area was biopsied; however, pathology could not confirm the diagnosis of NR. Therefore, the patient was classified and treated as a patient with unilateral WT. Because of the genital abnormalities, Denys-Drash syndrome (DDS) was suspected, but neither the histology of the kidneys nor the clinical picture confirmed that diagnosis. In addition, the patient had a splice site mutation in WT1 and not a typical DDS missense mutation. A metachronous tumor was identified after 3.2 years. At that time, the parents refused further treatment, and the patient died 2 months later.

Patient WTHD11 received 4 weeks of VCR and ACTD preoperatively and underwent a partial nephrectomy on both sides; postoperatively, he was treated again with VCR and ACTD and once with epirubicin, which was discontinued because of cardiotoxicity. After 2 years, he developed a local recurrence on the side with the WT and a WT in the contralateral kidney that previously had an ILNR. The patient was pretreated again with ACTD and VCR, and the tumor on the right was reduced, whereas the tumor on the left increased in size. The tumors were surgically removed again, and postoperative treatment included 5×etoposide and carboplatin. The patient underwent kidney transplantation in 2005 and has remained in second complete remission for 13.6 years.

DISCUSSION

The 71 study patients were treated for WT between 1989 and 1994, and a new survey of their clinical status was conducted in 2006. This study revealed that the presence of a WT1 mutation correlated significantly with a poor volumetric response to chemotherapy; however, when we compared patients with (n = 19) and without (n = 52) WT1 mutations, no statistically significant difference in overall or event-free survival was observed.

Although no difference in event-free survival was observed, patients with germline WT1 mutations have a higher risk of bilateral disease, including synchronous and metachronous WT and developing recurrent disease. Risk factors for the development of contralateral WT were reported previously and included young age at diagnosis and the presence of NR.20 Bilateral disease and young age at diagnosis are more frequent in patients with germline WT1 mutations.21 Three of the study patients with germline WT1 mutations and bilateral disease had second tumor events. Three tumors developed from previous ILNR or NR, in one patient after 11 years. This suggests that tumors later in life may develop only if persistent clusters of embryonal cells (NRs) are present in the kidneys after birth. NRs may be the precursor lesions from which WT can develop, although most NRs do not form WTs and spontaneously regress.22 Recently, the molecular analysis of 3 ILNR lesions from 2 WTs revealed that WT1 mutations were present, whereas mutations in CTNNB1 were acquired at a later stage and were observed only in the associated tumors.23 These results suggest that, in WT1 mutant tumors, loss of a functional WT1 may result in the formation of the precursor lesion; and other mutations, eg, in CTNNB1, are needed for tumor formation. The presence of only a mutation in CTNNB1 in some tumors suggests that either an undetectable alteration in WT1 was present or that in, these tumors, another genetic event cooperates with the mutant CTNNB1.

Five patients in the current study with bilateral disease received prolonged preoperative chemotherapy with ACTD, VCR, and ADR, and 3 patients also received various combinations of ifosfamide, etoposide, and carboplatin. It is noteworthy that 2 of the patients who developed tumors from NRs obtained a ‘milder’ treatment with ACTD and VCR only. However, it should be noted that 7 patients with WT1 germline mutations and unilateral WT were treated with the same ‘milder’ chemotherapy regimen and remained free of disease for at least 10 years. It is unknown whether any of these patients had NRs at the time of diagnosis that were undetectable with the methods used and may have been eradicated by the chemotherapy. More intensive treatment options with combinations of drugs have been tested in patients with bilateral tumors but only rarely have resulted in a tumor response.10 Shamberger et al,24 who analyzed patients with bilateral WTs from the National Wilms Tumor Study, discussed the importance of identifying markers that can differentiate bilateral WTs, which are managed best with early resection and less intensive therapy after nephrectomy. It is possible that the presence of WT1 mutations identifies such tumors of low responsiveness and favorable prognosis, because tumors from patients with WT1 mutations seldom metastasize. However, not all patients with bilateral tumors will have an overall favorable outcome: Several patients were described in the literature who died of untreatable tumor progression; and some of these patients were treated preoperatively for an extended period.10, 24 Unfortunately, none of the cases described in the literature were studied for WT1 mutations.

One factor that may influence tumor recurrence is the type of mutation in the WT1 gene, because we have demonstrated that a higher proportion of bilateral tumors occur in patients with mutations in the N-terminal part of the WT1 gene.21 Furthermore, we recently described a patient who had a germline WT1 truncation mutation close to the N terminus and bilateral WT who was treated with kidney-sparing tumor surgery. The child did not receive postoperative chemotherapy and developed recurrences in both kidneys after 1 year. Molecular analyses demonstrated that the second tumors had different CTNNB1 mutations than the first tumors and, thus, were not recurrences.25 This case demonstrates that a germline WT1 mutation poses an increased risk for the development of independent tumors and that there is a high selection pressure for acquiring CTNNB1 mutations in cells with WT1 loss. An important conclusion from these observations is that, for patients with germline WT1 mutations, it is advisable initially to use a highly sensitive method that is able to detect NRs in the contralateral kidney, such as CT or MRI.26 A careful follow-up of these lesions is required to determine their fate of either regression or tumor development, and the patients should remain under close tumor surveillance for a long time. Future studies will have to investigate whether second events only occur when patients initially have residual rests.

A higher probability of a WT1 germline mutation is present in patients with bilateral disease, a young age at diagnosis, and syndromic features. Therefore, blood DNA from these patients should be studied for the presence of a WT1 mutation immediately at the time of diagnosis. If a WT1 germline mutation is detected and there is no volumetric response after 8 weeks of preoperative chemotherapy, then early tumor resection should be considered. In countries without preoperative treatment, this analysis should be initiated at the time of diagnosis, and the result will help in further treatment decisions.

In this study, we confirmed the correlation of stromal type histology and rhabdomyoblastic differentiation with WT1 mutations. In addition, all WT1 mutant tumors that were analyzed also had tumor-specific CTNNB1 mutations. The presence of rhabdomyoblasts in the tumor may be relevant for the identification of WT1-associated tumors, because stromal type tumors without any muscle differentiation may be derived from tumors of other histologic subtypes before chemotherapy. In 7 tumors with stromal type histology, no mutations in WT1 were observed by sequencing all 10 exons. It is possible that a mutation that was present in a small proportion of cells was missed. In addition, intragenic or large deletions, as well as promoter alterations or epigenetic changes, cannot be detected with these methods. Five of the stromal type tumors had a poor response, and 2 had a good response. In 3 of those tumors, no rhabdomyoblastic differentiation was observed.

Histologic and gene expression studies of WT1 mutant tumors have indicated that chemotherapy may induce a more mature muscle differentiation.27-29 A maturation accompanied by (terminal) differentiation and an inability to divide would explain why tumors with WT1 mutations may have less aggressive behavior and seldom metastasize. However, it is not known whether differentiated tumor cells may be able to revert to a state that permits cell division and growth when chemotherapy is stopped. In conclusion, although no volume reduction was observed, cells in the tumors may have responded by differentiation.

Acknowledgements

We thank Dr. Ludwig, coordinator of the International Society of Pediatric Oncology 9/Society for Pediatric Oncology trial for continuous support and encouragement.