|Home | About | Journals | Submit | Contact Us | Français|
To analyze the prevalence and clinical implications of Wilms' tumor 1 (WT1) single nucleotide polymorphism (SNP) rs16754 in the context of other prognostic markers in pediatric acute myeloid leukemia (AML).
Available diagnostic marrow specimens (n = 790) from 1,328 patients enrolled in three consecutive Children's Cancer Group/Children's Oncology Group trials were analyzed for the presence of SNP rs16754. SNP status was correlated with disease characteristics, WT1 expression level, and clinical outcome.
SNP rs16754 was present in 229 (29%) of 790 patients. The SNP was significantly more common in Asian and Hispanic patients and less common in white patients (P < .001). SNP rs16754 was also less common in patients with inv(16) (P = .043) and more common in patients with −5/del(5q) (P = .047). WT1 expression levels were significantly higher in patients with rs16754 or with WT1 mutations compared with WT1 wild-type patients (P = .021). Five-year overall survival (OS) for patients with and without the SNP was 60% and 50%, respectively (P = .031). Prognostic assessment by risk group demonstrated that in patients with low-risk disease, OS for those with and without SNP rs16754 was 90% versus 64% (P < .001) with a corresponding disease-free survival of 72% versus 53% (P = .041).
The presence of SNP rs16754 was an independent predictor of improved OS; outcome differences were most pronounced in the low-risk subgroup. The high prevalence of WT1 SNP rs16754, and its correlation with improved outcome, identifies WT1 SNP rs16754 as a potentially important molecular marker of prognosis in pediatric AML.
Single nucleotide polymorphisms (SNPs) account for much of the phenotypic diversity among individuals. Half of the known coding region SNPs in the human genome lead to a change in the resulting amino acid, whereas the other half do not (synonymous SNPs).1 Because synonymous SNPs encode a change in the DNA sequence without altering the resultant protein sequence, such silent SNPs were long assumed to be inconsequential. However, synonymous SNPs may represent genetic markers for functional molecular alterations with which they are in linkage disequilibrium; further, recent work has shown that synonymous SNPs may directly alter gene function and phenotype by various mechanisms, such as altering miRNA binding or protein folding, or by affecting mRNA splicing, stability, or expression.2 To date, silent SNPs have been reported in association with more than 40 diseases that have a genetic basis.3 Genomic studies in both pediatric and adult AML in the past decade have identified function-altering mutations in a host of genes, including FLT3, NPM1, NRAS, MLL, Wilms' tumor 1 (WT1), and CEBPA.4,5 The molecular characterization of AML is being continually redefined as novel alterations are discovered. We previously identified mutations in the zinc-finger domains of the WT1 gene in 8% of pediatric patients with AML.6 Although these mutations are predicted to lead to loss of function of WT1, we found that the presence of WT1 mutations had no independent significance in predicting outcome in pediatric AML. A recent study in adult AML also did not find WT1 mutations to be of independent prognostic significance; however, this study reported that presence of WT1 SNP rs16754 correlated significantly with improved survival outcomes.7
WT1 encodes a zinc-finger transcription factor protein that exists in multiple isoforms and is expressed primarily in tissues of the developing genitourinary and hematopoietic systems.8 WT1 is overexpressed in blasts cells of the majority of acute leukemia patients.9 The WT1 protein consists of a transcriptional regulatory domain (exons 1 to 6) as well as 4 C-terminal zinc-finger domains (exons 7 to 10) that facilitate DNA binding.8 Nearly all leukemia-associated WT1 mutations occur within the zinc-finger domains; most of these mutations occur within a hotspot in exon 7, also the location of SNP rs16754. In a study of 249 adult patients with normal-karyotype AML, Damm et al7 reported that SNP rs16754 independently predicted improved overall survival (OS) and relapse-free survival (RFS) in adult AML. In our study, we determined the prevalence of SNP rs16754 in a large cohort of pediatric AML patients enrolled on three consecutive CCG/COG trials. We then analyzed biologic or clinical differences among SNP-positive and SNP-negative patients and examined the prognostic significance of harboring at least one copy of the minor SNP allele in the context of other previously validated risk factors in pediatric AML.
Newly diagnosed pediatric patients with de novo AML enrolled in three consecutive pediatric AML protocols, CCG-2941, CCG-2961, and COG-AAML03P1, were eligible for this study. Details of these studies have been previously described.10,11 Diagnostic specimens from 790 of 1,328 patients were available for analysis. Patients with available specimens had similar clinical outcome compared with those without available specimens.6
Institutional review board approval was obtained before mutation analysis, and this study was approved by the COG Myeloid Disease Biology Committee. Informed consent was obtained in accordance with the Declaration of Helsinki.
Molecular genotyping of diagnostic specimens was performed as previously described.6 Reverse transcription was performed on 1 μg total RNA as per standard protocol (Invitrogen Corporation, Carlsbad, CA). Expression analysis by reverse-transcriptase polymerase chain reaction (PCR) was performed on cDNA transcripts, on a StepOne Plus real-time PCR instrument, using the TaqMan system (Applied Biosystems, Foster City, CA) per manufacturer's instructions. Patient samples were run in duplicate, with the beta glucuronidase (GUSB) housekeeping gene as an internal control. The TaqMan WT1 primer/probe set was designed to hybridize within WT1 exon 2. The ΔΔCT method was used to determine the relative expression levels of WT1. Normal bone marrow extracted from the same donor was used as a control on each run.
The Kaplan-Meier method was used to estimate OS and disease-free survival (DFS). OS was defined as time from study entry to death. DFS was defined as the time from end of course 1 for patients in complete remission (CR) until relapse or death. Estimates of relapse risk (RR) were obtained by the method of cumulative incidence that accounts for competing events. RR was defined as the time from end of course 1 for patients in CR to relapse or death due to progressive disease, where deaths from nonrelapse etiologies were considered competing events. The significance of predictor variables was tested with the log-rank statistic for OS and DFS, and with Gray's statistic for RR. Children who also received a stem-cell transplant while on study were censored at the time of transplant for all analyses. The significance of observed differences in proportions was tested by the χ2 test and Fisher's exact test when data were sparse. The Mann-Whitney test was used to determine the significance between differences in medians. Cox proportional hazard models were used to estimate hazard ratios (HR) for univariate and multivariate analyses for OS and RFS. RFS was defined as the time from end of course 1 for patients in CR to relapse or death due to progressive disease, where deaths from nonrelapse etiologies were censored.
All available specimens (n = 790) were subjected to direct sequencing of WT1 exon 7 to assess WT1 SNP status. SNP rs16754 represents an A>G substitution at nucleotide position 1293, the third position of codon 352 encoding arginine (CGA>CGG). At least 1 copy of the minor allele was detected in 229 patients (29.0%), 38 (16.6%) of whom were homozygous for the SNP (191 were heterozygous).
Demographic, laboratory, and clinical characteristics of patients with and without WT1 SNP rs16754 were compared (Table 1). There were no significant differences in sex, age, median diagnostic blast percentage, median diagnostic WBC count, or French-American-British class between SNP-positive and SNP-negative patients. There were significant differences in the racial distribution of the minor SNP allele. SNP rs16754 occurred at the highest frequency in patients of Asian (66%) and Hispanic (42%) descent and was less frequent in white (24%) and African American (22%) patients. Patients with the SNP had a lower prevalence of inversion 16 (6.3% v 12.6%; P = .043), and higher prevalence of −5/del(5q), (3.1% v 0.8% P = .047) compared with their SNP-negative counterparts. No other association with cytogenetic groups were identified.
The association between SNP rs16754 and other molecular alterations was investigated. FLT3/ITD, NPM1, and CEBPA mutations occurred at comparable frequencies in patients with and without the WT1 SNP (Table 1). Compared with the SNP-negative patients, who had a WT1 mutation prevalence of 10.4%, WT1 mutations were identified in only 3.2% of those with the SNP (P = .002). Further evaluation demonstrated that none of the 38 patients homozygous for SNP rs16754 had a concomitant WT1 mutation (Fig 1).
The proportion of SNP-positive and SNP-negative patients was similar when they were stratified into high-risk (−7, −5/del(5q), FLT3/ITD with high allelic ratio), low-risk (t,(8,21) inv(16)/t,(16,16) NPM1, or CEBPA mutations) or standard-risk groups (all other patients). Minimal residual disease data from the end of course 1, assessed by multidimensional flow cytometry, were available from 184 patients enrolled on AAML-03P1. Minimal residual disease higher than 0.01% was detected in 13 (24%) of 55 SNP-positive patients, compared with 40 (31%) of 129 SNP-negative patients (P = .405).
Clinical outcome data were examined for the patients with known WT1 SNP rs16754 status (Fig 2). Patients with or without SNP rs16754 had similar CR rates after one course of induction (81.2% v 80.2%, P = .837). Overall survival at 5 years from study entry for SNP-positive patients was 60% (SE [±] 7%) versus 50% ± 5% for those without the SNP (HR, 0.76; P = .031). The corresponding DFS from CR was 51% ± 8% and 47% ± 5% for those with and without SNP rs16754 (HR, 0.86; P = .415). For patients who achieved an initial remission (n = 615), those with and without the SNP had similar relapse rates (40% ± 8% v 43% ± 5%; HR, 0.92; P = .763) and similar treatment-related mortality (8% ± 4% v 11% ± 3%; HR, 0.79; P = .442).
We performed univariate and multivariate Cox regression analysis to evaluate whether presence of the WT1 SNP as well as other known prognostic factors (ie, cytogenetics, mutations, diagnostic WBC, and race) were predictors of OS and RFS in the entire study cohort (Table 2). Patients with SNP rs16754 had an improved survival with an HR of 0.76 for death from enrollment (P = .031), and an HR of 0.92 for relapse from achieving an initial remission (P = .6). Patients with high-risk features had an HR of 1.76 for death from diagnosis (P < .001) and an HR of 1.91 for increased risk of relapse (P < .001), whereas low-risk group assignment was associated with improved survival (HR, 0.48; P < .001) and lower risk of relapse (HR, 0.49; P < .001). Diagnostic WBC higher than 50 ×109/L was a predictor of decreased OS (HR, 1.30; P = .005) and higher risk of relapse (HR, 1.26; P = .047). Non-white patients had decreased OS (HR, 1.34; P = .001) but not significantly reduced RFS (HR, 1.17; P = .152).
In the multivariate analysis, WT1 SNP rs16754 was an independent prognostic marker for improved OS with an HR for death of 0.64 compared with SNP-negative patients (P = .004). In this model, HR for relapse from remission for patients with WT1 SNP was 0.79 compared with their SNP negative counterparts (P = .197).
The prognostic impact of SNP rs16754 was evaluated in specific clinical risk groups (Fig 3). In patients with standard-risk disease (no high or low risk features, n = 294), WT1 SNP was identified in 79 patients (26%). Those with and without SNP rs16754 had identical OS at 5 years from diagnosis of 47% and a similar RR (48% ± 14% v 45% ± 9%; P = .77) and DFS (45% ± 14% v 47% ± 9%; P = .88) from remission. WT1 SNP was identified in 58 patients (26%) with low-risk AML (CBF translocations, CEBPA or NPM1 mutations, n = 221). In contrast to standard-risk patients, low-risk patients with SNP rs16754 had an actuarial OS of 90% ± 8%, versus 64% ± 9% for low-risk SNP-negative patients (HR, 0.27; P = .001). In low-risk patients who achieved an initial remission, DFS at 5 years from remission for SNP-positive patients was 72% ± 13% versus 53% ± 10% for SNP-negative patients (P = .041), with a corresponding RR of 21% ± 12% versus 41% ± 9% (HR, 0.49; P = .179) compared with their SNP-negative low-risk counterparts.
In patients with high-risk disease (−7, −5/del5q or high risk FLT3/ITD, n = 94), the WT1 SNP was identified in 33 patients (35%). Actuarial OS at 5 years from diagnosis was 41% ± 22% for SNP-positive patients versus 21% ± 12% for SNP-negative patients (P = .083). In high-risk patients who achieved an initial remission, DFS at 5 years from remission for SNP-positive patients was 30% ± 22% versus 10% ± 11% for SNP-negative patients (P = .273), with a corresponding RR of 60% ± 23% versus 78% ± 15% (P = .179) compared with their SNP-negative counterparts.
We further evaluated the prognostic significance of SNP rs16754 in patients with normal karyotype. Of the 125 patients without cytogenetic abnormalities, 38 had SNP rs16754 (30%). Overall survival at 5 years from study entry for patients with and without WT1 SNP was 45% ± 18% versus 39% ± 12% (P = .522). In patients who achieved an initial remission, DFS at 5 years from remission for SNP-positive patients was 51% ± 20% versus 41% ± 14% for SNP-negative patients (P = .273), with a corresponding RR of 46% ± 20% versus 47% ± 14% (P = .882) compared with their SNP-negative counterparts.
Within the SNP-positive group, homozygous (n = 38) and heterozygous (n = 191) patients had similar 5-year OS (61% ± 8% v 56% ± 18%, P = .960), DFS (46% ± 8% v 41% ± 17%, P = .720), and RR (38% ± 9% v 50% ± 20%, P = .312).
Because the presence of this synonymous SNP imparts prognostic significance, we sought to determine a potential mechanism by which a silent polymorphism may be translated into a functional alteration. Differential RNA editing is one mechanism by which a silent genomic mutation may create a functional alteration in the pre-mRNA transcript, leading to a change in the protein product. An example of RNA editing has been previously described in codon 280 of WT1.12 To determine whether differential RNA editing might occur as a result of the synonymous SNP rs16754, we sequenced WT1 exon 7 from cDNA transcripts obtained from 50 patients at random from COG AAML-03P1. Of the 50 patients, 13 (26%) had at least 1 minor allele of the WT1 SNP; 4/13 SNP-positive patients were homozygous for the SNP. Correlation of genotyping data performed on genomic DNA with these cDNA data showed identical sequences, demonstrating that RNA editing is not involved in the pathogenesis of AML in regard to SNP rs16754.
WT1 is highly expressed in most patients with AML.13 We questioned whether patients with SNP rs16754 may have an altered WT1 expression level. WT1 expression levels were evaluated by quantitative reverse-transcriptase PCR in 114 unselected patient samples from COG-AAML03P1 with known WT1 SNP rs16754 (SNP positive, n = 30) and WT1 mutation (mutation positive, n = 13) status (Fig 4). WT1 expression levels for each sample were normalized to the expression in normal marrow control. WT1 expression was highly variable, ranging from 0.00- to 2949.42-fold normal marrow expression (median 70.4-fold normal bone marrow). Eleven patients did not have detectable WT1 expression (10%) and 97 patients (85%) had WT1 expression higher than observed in normal marrow controls.
Median WT1 expression in WT1 wild-type patients (without SNP rs16754 or a WT1 mutation, n = 71) was 40.61 times that of normal marrow controls (range, 0.00 to 2949.42), whereas the median expression in patients with the WT1 SNP (n = 30) was 215.76 times that of normal marrow control (range, 0.00 to 1,399.23; P = .0214). We further compared WT1 expression levels in those with and without WT1 mutations. Of the 114 patient samples tested, 13 (11.4%) were positive for a WT1 mutations (12 exon 7 mutations, one exon 8 mutation) and only one patient harbored both SNP rs16754 and a WT1 mutation. Median WT1 expression level for this cohort was 327.14-fold normal marrow control (range, 54.17- to 902.3-fold; P = .0005) compared with the WT1 wild-type cohort.
In this study, we identified WT1 SNP rs16754 in 29% of childhood patients with AML, the presence of which was either directly (−5/del5q) or inversely (inv 16, WT1 mutation) associated with specific disease characteristics. Further, rs16754 was disproportionately distributed across ethnic groups, with higher prevalence in Asian and Hispanic patients. More importantly, despite being a silent polymorphism, the presence of this SNP highly correlated with improved disease outcome. The protective impact of this polymorphism appears to be most pronounced in patients with favorable risk features.
The elucidation of possible mechanisms by which a synonymous SNP may be biologically significant is an important pursuit. Such a SNP may be commonly inherited as part of a haplotype and exist in linkage disequilibrium with yet-unidentified, disease-associated molecular markers. Studies are ongoing to identify possible disease-associated alterations that may exist in linkage disequilibrium with SNP rs16754. Alternatively, a synonymous SNP may lead to alterations in a microRNA binding site, alternative splicing, protein folding, or mRNA expression. In the case of the MDR1 gene, Kimchi-Sarfaty et al14 showed that a synonymous SNP substituting a rare codon for a common codon encoding the same amino acid results in decreased function of the encoded P-glycoprotein. As the tRNA pool for each codon is proportional to how frequently that codon is utilized, replacing a commonly used codon with a rare codon (or vice versa) may directly impact translation kinetics of the protein in question. Altering translation kinetics is known to affect protein folding, resulting in altered protein function. In the case of WT1 SNP rs16754, the minor allele results in a CGA>CGG transversion, which represents a change from a rare (CGA, 6.2 per thousand) codon encoding arginine to a more frequently used codon (CGG, 11.4 per thousand; frequencies obtained from the Codon Usage Database15). Substitution of a rare codon for a more frequently used codon leads to increased translation kinetics, which can also have functional protein consequences, as has been demonstrated in vitro in E coli.16
It is notable that the incidence of WT1 zinc-finger mutations was lower in SNP-positive patients than in their SNP-negative counterparts. Further, such mutations occurred only in heterozygous patients; none of the homozygous SNP-positive patients also had a concomitant WT1 mutation. Although the mechanism is unclear, if further validated, it is intriguing to speculate that harboring the WT1 SNP rs16754 may render unnecessary the acquisition of a WT1 mutation during myeloid leukemogenesis.
In multivariate analysis, SNP rs16754 is an independent predictor of improved OS. This WT1 SNP may be in linkage disequilibrium with a haplotype of another gene which affects response to chemotherapy. Survival differences were most pronounced in patients assigned to the low-risk group, with SNP-positive patients having an excellent OS of 90%. However, the prognostic utility of SNP status in this risk group is uncertain, as SNP-negative low-risk patients have better survival outcomes than standard-risk patients. Likewise, although SNP-positive high-risk patients trended toward improved outcomes compared with SNP-negative high-risk patients, their outcomes were still inferior to those of standard-risk patients. Thus, on the basis of SNP status alone, we can make no recommendation for change in risk stratification of patients with AML. Of note, the presence of SNP rs16754 had no impact on survival outcomes in pediatric AML patients currently assigned to the standard-risk group.
We also found that median expression levels of WT1 in patients with SNP rs16754 were significantly higher than those in patients with wild type WT1. SNP rs16754 is a putative cis-acting regulatory SNP; such SNPs may regulate expression levels of their respective genes, contributing to allelic imbalance.17 Although its role in leukemogenesis is unclear, the frequency and degree of WT1 overexpression in AML suggests that this expression can be used both as target of immunotherapy, as well as a potential marker of minimal residual disease.13 Identifying a population with more favorable outcome within the subset of patients with high WT1 expression may help in refining the utility of this parameter in response evaluation. Thus, WT1 mutation and WT1 SNP status should be prospectively evaluated alongside WT1 expression in future pediatric AML trials.
We thank the patients and families who consented to the use of biologic specimens in these trials; the Children's Oncology Group AML Reference Laboratories for providing specimens; and V. Shanker, PhD, for scientific editing.
Supported by the American Society of Clinical Oncology Young Investigator Award, University of Washington Child Health Research Center New Investigator Award, and Mary Claire Satterly Foundation AML Research Grant (P.A.H.), Grants No. R01 CA114563 and R21 CA10262-01 from the National Institutes of Health (S.M.), Children's Oncology Group Statistics and Data Center Grant No. U10 CA98413 and Chair's Grant No. NIH U10 CA98543.
Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
Employment or Leadership Position: Janet L. Franklin, Amgen (C) Consultant or Advisory Role: None Stock Ownership: Janet L. Franklin, Amgen Honoraria: None Research Funding: None Expert Testimony: None Other Remuneration: None
Conception and design: Phoenix A. Ho, Soheil Meshinchi
Financial support: Soheil Meshinchi
Administrative support: Alan S. Gamis
Provision of study materials or patients: Susana C. Raimondi, Janet L. Franklin, Beverly Lange
Collection and assembly of data: Phoenix A. Ho, Julia Kuhn, Robert B. Gerbing, Rong Zeng, Kristen L. Miller, Susana C. Raimondi
Data analysis and interpretation: Phoenix A. Ho, Julia Kuhn, Robert B. Gerbing, Jessica A. Pollard, Rong Zeng, Nyla A. Heerema, Susana C. Raimondi, Betsy A. Hirsch, Janet L. Franklin, Beverly Lange, Alan S. Gamis, Todd A. Alonzo, Soheil Meshinchi
Manuscript writing: Phoenix A. Ho, Julia Kuhn, Robert B. Gerbing, Jessica A. Pollard, Rong Zeng, Kristen L. Miller, Nyla A. Heerema, Susana C. Raimondi, Betsy A. Hirsch, Janet L. Franklin, Beverly Lange, Alan S. Gamis, Todd A. Alonzo, Soheil Meshinchi
Final approval of manuscript: Phoenix A. Ho, Julia Kuhn, Robert B. Gerbing, Jessica A. Pollard, Rong Zeng, Kristen L. Miller, Nyla A. Heerema, Susana C. Raimondi, Betsy A. Hirsch, Janet L. Franklin, Beverly Lange, Alan S. Gamis, Todd A. Alonzo, Soheil Meshinchi