Other genes and chromosomal alterations

1q. Gain of chromosome 1q is associated with an inferior outcome and is the single most powerful predictor of outcome.[53,54] Gain of chromosome 1q is one of the most common cytogenetic abnormalities in Wilms tumor and is observed in approximately 30% of tumors. In an analysis of FH Wilms tumor from 1,114 patients from NWTS-5 (COG-Q9401/NCT00002611), 28% of the tumors displayed 1q gain.[53] One study included a cohort of FH Wilms tumor that was enriched for patients who relapsed. The study found that the prevalence of 1q gain was higher in the relapsed Wilms tumor specimens (75%) than in the matched primary samples (47%).[55] The increased prevalence of 1q gain at relapse supports its association with poor prognosis and disease progression.

1q. Gain of chromosome 1q is associated with an inferior outcome and is the single most powerful predictor of outcome.[53,54] Gain of chromosome 1q is one of the most common cytogenetic abnormalities in Wilms tumor and is observed in approximately 30% of tumors. In an analysis of FH Wilms tumor from 1,114 patients from NWTS-5 (COG-Q9401/NCT00002611), 28% of the tumors displayed 1q gain.[53] One study included a cohort of FH Wilms tumor that was enriched for patients who relapsed. The study found that the prevalence of 1q gain was higher in the relapsed Wilms tumor specimens (75%) than in the matched primary samples (47%).[55] The increased prevalence of 1q gain at relapse supports its association with poor prognosis and disease progression.

The 8-year event-free survival (EFS) rate was 77% for patients with 1q gain and 90% for those lacking 1q gain (P < .001). Within each disease stage, 1q gain was associated with inferior EFS.

1q. Gain of chromosome 1q is associated with an inferior outcome and is the single most powerful predictor of outcome.[53,54] Gain of chromosome 1q is one of the most common cytogenetic abnormalities in Wilms tumor and is observed in approximately 30% of tumors. In an analysis of FH Wilms tumor from 1,114 patients from NWTS-5 (COG-Q9401/NCT00002611), 28% of the tumors displayed 1q gain.[53] One study included a cohort of FH Wilms tumor that was enriched for patients who relapsed. The study found that the prevalence of 1q gain was higher in the relapsed Wilms tumor specimens (75%) than in the matched primary samples (47%).[55] The increased prevalence of 1q gain at relapse supports its association with poor prognosis and disease progression.

The 8-year overall survival (OS) rate was 88% for those with 1q gain and 96% for those lacking 1q gain (P < .001). OS was significantly inferior in cases with stage I disease (P < .0015) and stage IV disease (P = .011).

1q. Gain of chromosome 1q is associated with an inferior outcome and is the single most powerful predictor of outcome.[53,54] Gain of chromosome 1q is one of the most common cytogenetic abnormalities in Wilms tumor and is observed in approximately 30% of tumors. In an analysis of FH Wilms tumor from 1,114 patients from NWTS-5 (COG-Q9401/NCT00002611), 28% of the tumors displayed 1q gain.[53] One study included a cohort of FH Wilms tumor that was enriched for patients who relapsed. The study found that the prevalence of 1q gain was higher in the relapsed Wilms tumor specimens (75%) than in the matched primary samples (47%).[55] The increased prevalence of 1q gain at relapse supports its association with poor prognosis and disease progression.

Similar results were reported in the International Society of Paediatric Oncology (SIOP) WT 2001 study of 586 children with Wilms tumor.[54]

16q and 1p. Additional tumor-suppressor or tumor-progression genes may lie on chromosomes 16q and 1p, as evidenced by LOH for these regions in 17% and 11% of Wilms tumor cases, respectively.[56]These conflicting results may arise from the greater prognostic significance of 1q gain described above. LOH of 16q and 1p loses significance as independent prognostic markers in the presence of 1q gain. However, in the absence of 1q gain, LOH of 16q and 1p retains their adverse prognostic impact.[53] The LOH of 16q and 1p appears to arise from complex chromosomal events that result in 1q LOH or 1q gain. The change in 1q appears to be the critical tumorigenic genetic event.[59]

16q and 1p. Additional tumor-suppressor or tumor-progression genes may lie on chromosomes 16q and 1p, as evidenced by LOH for these regions in 17% and 11% of Wilms tumor cases, respectively.[56]These conflicting results may arise from the greater prognostic significance of 1q gain described above. LOH of 16q and 1p loses significance as independent prognostic markers in the presence of 1q gain. However, in the absence of 1q gain, LOH of 16q and 1p retains their adverse prognostic impact.[53] The LOH of 16q and 1p appears to arise from complex chromosomal events that result in 1q LOH or 1q gain. The change in 1q appears to be the critical tumorigenic genetic event.[59]

In large NWTS studies, patients with tumor-specific loss of these loci had significantly worse relapse-free survival and OS rates. Combined loss of 1p and 16q are criteria used to select FH Wilms tumor patients for more aggressive therapy in the current Children's Oncology Group (COG) study. However, a U.K. study of more than 400 patients found no significant association between 1p deletion and poor prognosis, but a poor prognosis was associated with 16q LOH.[57]

16q and 1p. Additional tumor-suppressor or tumor-progression genes may lie on chromosomes 16q and 1p, as evidenced by LOH for these regions in 17% and 11% of Wilms tumor cases, respectively.[56]These conflicting results may arise from the greater prognostic significance of 1q gain described above. LOH of 16q and 1p loses significance as independent prognostic markers in the presence of 1q gain. However, in the absence of 1q gain, LOH of 16q and 1p retains their adverse prognostic impact.[53] The LOH of 16q and 1p appears to arise from complex chromosomal events that result in 1q LOH or 1q gain. The change in 1q appears to be the critical tumorigenic genetic event.[59]

An Italian study of 125 patients, using treatment quite similar to that in the COG study, found significantly worse prognosis in those with 1p deletions but not 16q deletions.[58]

miRNAPG. Variants in selected miRNAPG are observed in approximately 20% of Wilms tumor cases and appear to perpetuate the progenitor state.[2,5-8] The products of these genes direct the maturation of miRNAs from the initial pre-miRNA transcripts to functional cytoplasmic miRNAs (see Figure 10).[60] The most commonly altered miRNAPG is DROSHA, with a recurrent variant (E1147K) affecting a metal-binding residue of the RNase IIIb domain, representing about 80% of DROSHA-altered tumors. Other miRNAPG that are altered in Wilms tumor include DGCR8, DICER1, TARBP2, DIS3L2, and XPO5. These variants are generally mutually exclusive, and they appear to be deleterious and result in impaired expression of tumor-suppressing miRNAs. A striking sex bias was noted for patients with variants in DGCR8 (located on chromosome 22q11), with 38 of 43 cases (88%) arising in girls.[5,6]Germline pathogenic variants in miRNAPG are observed for DICER1 and DIS3L2, with variants in the former causing DICER1 syndrome and variants in the latter causing Perlman syndrome.

miRNAPG. Variants in selected miRNAPG are observed in approximately 20% of Wilms tumor cases and appear to perpetuate the progenitor state.[2,5-8] The products of these genes direct the maturation of miRNAs from the initial pre-miRNA transcripts to functional cytoplasmic miRNAs (see Figure 10).[60] The most commonly altered miRNAPG is DROSHA, with a recurrent variant (E1147K) affecting a metal-binding residue of the RNase IIIb domain, representing about 80% of DROSHA-altered tumors. Other miRNAPG that are altered in Wilms tumor include DGCR8, DICER1, TARBP2, DIS3L2, and XPO5. These variants are generally mutually exclusive, and they appear to be deleterious and result in impaired expression of tumor-suppressing miRNAs. A striking sex bias was noted for patients with variants in DGCR8 (located on chromosome 22q11), with 38 of 43 cases (88%) arising in girls.[5,6]Germline pathogenic variants in miRNAPG are observed for DICER1 and DIS3L2, with variants in the former causing DICER1 syndrome and variants in the latter causing Perlman syndrome.

DICER1 syndrome is typically caused by inherited truncating variants in DICER1, with tumor formation following acquisition of a missense variant in a domain of the remaining allele of DICER1 (the RNase IIIb domain) responsible for processing miRNAs derived from the 5p arms of pre-miRNAs.[61] Tumors associated with DICER1 syndrome include pleuropulmonary blastoma, cystic nephroma, ovarian sex cord–stromal tumors, multinodular goiter, and embryonal rhabdomyosarcoma.[61] Wilms tumor is an uncommon presentation of the DICER1 syndrome. In one study, three families with DICER1 syndrome included children with Wilms tumor, with two of the Wilms tumor cases showing the typical second DICER1 variant in the RNase IIIb domain.[62] Another study identified DICER1 variants in 2 of 48 familial Wilms tumor families.[63] Large sequencing studies of Wilms tumor cohorts have also observed occasional cases with DICER1 variants.[6,7]

miRNAPG. Variants in selected miRNAPG are observed in approximately 20% of Wilms tumor cases and appear to perpetuate the progenitor state.[2,5-8] The products of these genes direct the maturation of miRNAs from the initial pre-miRNA transcripts to functional cytoplasmic miRNAs (see Figure 10).[60] The most commonly altered miRNAPG is DROSHA, with a recurrent variant (E1147K) affecting a metal-binding residue of the RNase IIIb domain, representing about 80% of DROSHA-altered tumors. Other miRNAPG that are altered in Wilms tumor include DGCR8, DICER1, TARBP2, DIS3L2, and XPO5. These variants are generally mutually exclusive, and they appear to be deleterious and result in impaired expression of tumor-suppressing miRNAs. A striking sex bias was noted for patients with variants in DGCR8 (located on chromosome 22q11), with 38 of 43 cases (88%) arising in girls.[5,6]Germline pathogenic variants in miRNAPG are observed for DICER1 and DIS3L2, with variants in the former causing DICER1 syndrome and variants in the latter causing Perlman syndrome.

Perlman syndrome is a rare autosomal recessive overgrowth disorder caused by variants in DIS3L2, which encodes a ribonuclease that is responsible for degrading pre-let-7 miRNA.[64,65] Heterozygous germline DIS3L2 pathogenic inactivations are also associated with Wilms tumor development.[66] Patients with Perlman syndrome have a poor prognosis, with a high neonatal mortality rate. In a survey of published cases of Perlman syndrome (N = 28), in infants who survived beyond the neonatal period, approximately two-thirds developed Wilms tumor, and all patients showed developmental delay. Fetal macrosomia, ascites, and polyhydramnios are frequent manifestations.[67]

SIX1 and SIX2.SIX1 and SIX2 are highly homologous transcription factors that play key roles in early renal development and are expressed in the metanephric mesenchyme, where they maintain the mesenchymal progenitor population. In patients with Wilms tumors, the frequency of SIX1 variants is 3% to 4%, and the frequency of SIX2 variants is 1% to 3%.[5,6]MLLT1. Approximately 4% of Wilms tumor cases have variants in the highly conserved YEATS domain of MLLT1 (ENL), a gene known to be involved in transcriptional elongation by RNA polymerase II during early development.[19] The altered MLLT1 protein shows altered binding to acetylated histone tails. Patients with MLLT1-altered tumors present at a younger age and have a high prevalence of precursor intralobar nephrogenic rests, supporting a model whereby activating MLLT1 variants early in renal development result in the development of Wilms tumor.

SIX1 and SIX2.SIX1 and SIX2 are highly homologous transcription factors that play key roles in early renal development and are expressed in the metanephric mesenchyme, where they maintain the mesenchymal progenitor population. In patients with Wilms tumors, the frequency of SIX1 variants is 3% to 4%, and the frequency of SIX2 variants is 1% to 3%.[5,6]MLLT1. Approximately 4% of Wilms tumor cases have variants in the highly conserved YEATS domain of MLLT1 (ENL), a gene known to be involved in transcriptional elongation by RNA polymerase II during early development.[19] The altered MLLT1 protein shows altered binding to acetylated histone tails. Patients with MLLT1-altered tumors present at a younger age and have a high prevalence of precursor intralobar nephrogenic rests, supporting a model whereby activating MLLT1 variants early in renal development result in the development of Wilms tumor.

Virtually all SIX1 and SIX2 variants are in exon 1 and result in a glutamine-to-arginine variant at position 177 (Q177R).

SIX1 and SIX2.SIX1 and SIX2 are highly homologous transcription factors that play key roles in early renal development and are expressed in the metanephric mesenchyme, where they maintain the mesenchymal progenitor population. In patients with Wilms tumors, the frequency of SIX1 variants is 3% to 4%, and the frequency of SIX2 variants is 1% to 3%.[5,6]MLLT1. Approximately 4% of Wilms tumor cases have variants in the highly conserved YEATS domain of MLLT1 (ENL), a gene known to be involved in transcriptional elongation by RNA polymerase II during early development.[19] The altered MLLT1 protein shows altered binding to acetylated histone tails. Patients with MLLT1-altered tumors present at a younger age and have a high prevalence of precursor intralobar nephrogenic rests, supporting a model whereby activating MLLT1 variants early in renal development result in the development of Wilms tumor.

Variants in WT1, AMER1, and CTNNB1 are infrequent in cases with SIX1, SIX2, or miRNAPG variants. Conversely, SIX1 or SIX2 variants and miRNAPG variants tend to occur together.

SIX1 and SIX2.SIX1 and SIX2 are highly homologous transcription factors that play key roles in early renal development and are expressed in the metanephric mesenchyme, where they maintain the mesenchymal progenitor population. In patients with Wilms tumors, the frequency of SIX1 variants is 3% to 4%, and the frequency of SIX2 variants is 1% to 3%.[5,6]MLLT1. Approximately 4% of Wilms tumor cases have variants in the highly conserved YEATS domain of MLLT1 (ENL), a gene known to be involved in transcriptional elongation by RNA polymerase II during early development.[19] The altered MLLT1 protein shows altered binding to acetylated histone tails. Patients with MLLT1-altered tumors present at a younger age and have a high prevalence of precursor intralobar nephrogenic rests, supporting a model whereby activating MLLT1 variants early in renal development result in the development of Wilms tumor.

In Wilms tumor, SIX1 and SIX2 variants are associated with the high-risk blastemal subtype and the presence of undifferentiated blastema in chemotherapy-naïve samples.

SIX1 and SIX2.SIX1 and SIX2 are highly homologous transcription factors that play key roles in early renal development and are expressed in the metanephric mesenchyme, where they maintain the mesenchymal progenitor population. In patients with Wilms tumors, the frequency of SIX1 variants is 3% to 4%, and the frequency of SIX2 variants is 1% to 3%.[5,6]MLLT1. Approximately 4% of Wilms tumor cases have variants in the highly conserved YEATS domain of MLLT1 (ENL), a gene known to be involved in transcriptional elongation by RNA polymerase II during early development.[19] The altered MLLT1 protein shows altered binding to acetylated histone tails. Patients with MLLT1-altered tumors present at a younger age and have a high prevalence of precursor intralobar nephrogenic rests, supporting a model whereby activating MLLT1 variants early in renal development result in the development of Wilms tumor.

In a study of 82 cases of FH Wilms tumor, SIX1 Q177R hotspot variants were identified at a higher rate in tumor specimens at relapse (11 cases; 13.4%) than in those at diagnosis (4%). For 45 cases that had both diagnostic and relapse specimens, there were 6 cases with SIX1 Q177R at relapse, 3 of which did not have SIX1 Q177R at diagnosis. This finding suggests that this variant is not required for tumor development in some individuals with Wilms tumor.[55]

TP53 (tumor suppressor gene). Most anaplastic Wilms tumor cases show variants in the TP53 tumor suppressor gene.[68-70] TP53 may be useful as an unfavorable prognostic marker.[68,69] In a study of 118 prospectively identified patients with diffuse anaplastic Wilms tumor registered on the NWTS-5 trial, 57 patients (48%) demonstrated TP53 variants, 13 patients (11%) demonstrated TP53 segmental copy number loss without variants, and 48 patients (41%) lacked both (wild-type TP53 [wtTP53]). All TP53 variants were detected by sequencing alone. Patients with stage III or stage IV disease with wtTP53 had a significantly lower relapse rate and mortality rate than did patients with TP53 abnormalities (P = .00006 and P = .00007, respectively). The TP53 status had no effect on patients with stage I or stage II tumors.[71]

TP53 (tumor suppressor gene). Most anaplastic Wilms tumor cases show variants in the TP53 tumor suppressor gene.[68-70] TP53 may be useful as an unfavorable prognostic marker.[68,69] In a study of 118 prospectively identified patients with diffuse anaplastic Wilms tumor registered on the NWTS-5 trial, 57 patients (48%) demonstrated TP53 variants, 13 patients (11%) demonstrated TP53 segmental copy number loss without variants, and 48 patients (41%) lacked both (wild-type TP53 [wtTP53]). All TP53 variants were detected by sequencing alone. Patients with stage III or stage IV disease with wtTP53 had a significantly lower relapse rate and mortality rate than did patients with TP53 abnormalities (P = .00006 and P = .00007, respectively). The TP53 status had no effect on patients with stage I or stage II tumors.[71]

In-depth analysis of a subset of 39 patients with diffuse anaplastic Wilms tumor showed that 7 patients (18%) were wtTP53. These wtTP53 tumors demonstrated gene expression evidence of p53 pathway activation. Retrospective pathology review of wtTP53 tumors revealed no or very low volume of anaplasia in six of seven tumors. These data support the key role of TP53 loss in the development of anaplasia in Wilms tumor and support its significant clinical influence in patients who have residual anaplastic disease after surgery.[71]

FBXW7.FBXW7, a ubiquitin ligase component, is an established tumor suppressor gene that has been identified as recurrently altered at low rates in Wilms tumor and other malignancies. Variants of this gene have been associated with epithelial-type tumor histology.[72]; [73][Level of evidence C1]

TRIM28.TRIM28 encodes a multidomain protein involved in the regulation of many cellular processes and is an autosomal dominant Wilms tumor predisposition gene. TRIM28 accounts for about 8% of familial Wilms tumor and 2% of unselected Wilms tumor.[74-77]; [73][Level of evidence C1]

TRIM28.TRIM28 encodes a multidomain protein involved in the regulation of many cellular processes and is an autosomal dominant Wilms tumor predisposition gene. TRIM28 accounts for about 8% of familial Wilms tumor and 2% of unselected Wilms tumor.[74-77]; [73][Level of evidence C1]

A strong association between TRIM28 variants and epithelial Wilms tumor has been observed, and most individuals with a TRIM28 variant have a Wilms tumor of predominantly epithelial histology.[74-76]; [73][Level of evidence C1]

TRIM28.TRIM28 encodes a multidomain protein involved in the regulation of many cellular processes and is an autosomal dominant Wilms tumor predisposition gene. TRIM28 accounts for about 8% of familial Wilms tumor and 2% of unselected Wilms tumor.[74-77]; [73][Level of evidence C1]

In a cohort of 91 affected individuals from 49 families with Wilms tumor pedigrees, 33 individuals were identified as having constitutional cancer-predisposing variants, 21 of whom had a variant in TRIM28. There was a strong parent-of-origin effect, with all ten evaluable cases having inherited variants that were maternally transmitted.[73][Level of evidence C1]

TRIM28.TRIM28 encodes a multidomain protein involved in the regulation of many cellular processes and is an autosomal dominant Wilms tumor predisposition gene. TRIM28 accounts for about 8% of familial Wilms tumor and 2% of unselected Wilms tumor.[74-77]; [73][Level of evidence C1]

Most TRIM28-altered cases have either frameshift, nonsense, or splice-site variants in one allele combined with LOH in the second allele, leading to loss of TRIM28 protein expression in the tumor. Immunohistochemistry staining for loss of TRIM28 protein expression can be used to identify most patients whose tumors have TRIM28 variants.[77]

9q22.3 microdeletion syndrome. Patients with 9q22.3 microdeletion syndrome have an increased risk of Wilms tumor.[78] The chromosomal region with germline pathogenic deletion includes PTCH1, the gene that is altered in Gorlin syndrome (nevoid basal cell carcinoma syndrome associated with osteosarcoma). 9q22.3 microdeletion syndrome is characterized by the clinical findings of Gorlin syndrome, as well as developmental delay and/or intellectual disability, metopic craniosynostosis, obstructive hydrocephalus, prenatal and postnatal macrosomia, and seizures. Five patients who presented with Wilms tumor in the context of a constitutional 9q22.3 microdeletion have been reported.[78-80]

MYCN. Genomic alterations involving the MYCN network (e.g., MYCN, MAX, MGA, NONO) have been reported to occur in 25% to 30% of Wilms tumor cases.[55] Specific genomic alterations associated with the MYCN network include the following:

MYCN. Genomic alterations involving the MYCN network (e.g., MYCN, MAX, MGA, NONO) have been reported to occur in 25% to 30% of Wilms tumor cases.[55] Specific genomic alterations associated with the MYCN network include the following:

MYCN copy number gain was observed in approximately 13% of Wilms tumor cases. MYCN gain was more common in anaplastic cases (7 of 23 cases, 30%) than in nonanaplastic cases (11.2%), and it was associated with poorer relapse-free survival (RFS) and overall survival, independent of histology.[81] MYCN tandem duplication was reported in 11 of 82 (13%) FH Wilms tumor specimens from relapse.[55]

MYCN. Genomic alterations involving the MYCN network (e.g., MYCN, MAX, MGA, NONO) have been reported to occur in 25% to 30% of Wilms tumor cases.[55] Specific genomic alterations associated with the MYCN network include the following:

Germline pathogenic copy number gain at MYCN has been reported in a bilateral Wilms tumor case,[81] and germline MYCN pathogenic duplication was also reported for a child with prenatal bilateral nephroblastomatosis and a family history of nephroblastoma.[82]

MYCN. Genomic alterations involving the MYCN network (e.g., MYCN, MAX, MGA, NONO) have been reported to occur in 25% to 30% of Wilms tumor cases.[55] Specific genomic alterations associated with the MYCN network include the following:

Variants at codon 44 (p.P44L) of MYCN are observed in approximately 3% to 4% of Wilms tumor cases at diagnosis [81,83] and in 8.5% of cases at relapse.[55] In a study of 810 Wilms tumor cases, 24 (3%) had MYCN P44L hotspot variants. RFS was significantly lower (68.6%) in patients with P44L variants than in patients with wild-type MYCN status (87.1%).[83]

MYCN. Genomic alterations involving the MYCN network (e.g., MYCN, MAX, MGA, NONO) have been reported to occur in 25% to 30% of Wilms tumor cases.[55] Specific genomic alterations associated with the MYCN network include the following:

The MYCN interacting protein MAX was altered at codon 60 (R60Q) in 7 of 782 Wilms tumor cases (0.9%).[83] RFS was significantly lower in patients with the MAX R60Q hotspot variant than in patients with wild-type MAX status.

CTR9. Inactivating CTR9 germline pathogenic variants were identified in 4 of 36 familial Wilms tumor pedigrees.[11,84] CTR9, which is located at chromosome 11p15.3, is a key component of the polymerase-associated factor 1 complex (PAF1c), which has multiple roles in RNA polymerase II regulation and is implicated in embryonic organogenesis and maintenance of embryonic stem cell pluripotency.

REST. Inactivating germline pathogenic variants in REST (encoding RE1-silencing transcription factor) were identified in four familial Wilms tumor pedigrees.[10] REST is a transcriptional repressor that functions in cellular differentiation and embryonic development. Most REST variants clustered within the portion of REST encoding the DNA-binding domain, and functional analyses showed that these variants compromise REST transcriptional repression. When screened for REST variants, 9 of 519 individuals with Wilms tumor who had no history of relatives with the disease tested positive for the variant; some had parents who also tested positive.[10] These observations indicate that REST is a Wilms tumor predisposition gene associated with approximately 2% of Wilms tumor.