The Prognostic Effect of High WT1 Gene Expression in Pediatric AML Depends on WT1 SNP rs16754 Status: A Report From the Children's Oncology Group (COG)

WT1 is aberrantly over-expressed in the blast cells of most AML patients. Several adult AML groups have reported poor outcome in association with high WT1 expression. In contrast, in previously the largest pediatric study, COG investigators found no association between diagnostic WT1 expression and survival in 155 patients treated on the legacy POG-9421 trial. We recently demonstrated that the WT1 synonymous SNP rs16754 correlates with favorable outcome in childhood AML; further, SNP+ patients had higher median WT1 expression. In the present study we examine the clinical correlates of diagnostic WT1 expression within a contemporary COG trial, accounting for SNP rs16754 genotypes.

WT1 mRNA expression was measured via qRT-PCR in diagnostic specimens obtained from 225 patients enrolled on the pediatric trial COG-AAML03P1. Expression levels were normalized against pooled normal bone marrow (NBM) controls. Direct sequencing of WT1 exon 7 was performed to determine SNP rs16754 genotype. WT1 expression was correlated with disease characteristics, SNP status, and clinical outcome.

WT1 expression varied over 4 log folds across the study cohort (range: 0.00 to 3148.27 fold NBM, median: 75.91). In 29 (13%) patients, WT1 expression was undetectable. These patients had distinct biologic characteristics compared to the remainder of the cohort, as they were predominantly FAB class M5 (56%, p<0.001) and the majority (52%, p<0.001) had 11q23 rearrangements. This group was also younger (median age 5.1 years), and rarely harbored favorable cytogenetics (Inv(16): 1 patient; t(8;21): 1 patient) or molecular abnormalities (FLT3/ ITD: 1 patient; NPM1- mutated: 1 patient; no patients with mutated CEBPA or WT1). Survival outcomes for this group did not differ significantly from the study cohort at large. WT1 expression higher than NBM was detected in 183/196 of the remaining patients. We grouped patients into WT1 expression quartiles (Q1-Q4) for comparison of clinical characteristics and outcome. FLT3/ ITD (23% vs. 9%, p=0.017) and WT1 mutations (22% vs. 4%, p<0.001) occurred more frequently in those within the highest quartile (Q4; WT1 expression >302.97 fold NBM) compared to the lower 3 quartiles. Infant patients were also less common in Q4 (p=0.010). WT1 SNP rs16754 was equally distributed among quartile groups. Survival outcomes did not differ significantly between the quartiles, confirming that diagnostic WT1 expression has no prognostic impact in contemporarily-treated pediatric AML patients, when all patients were considered. However, given our recent findings regarding the clinical relevance of the WT1 SNP, we examined the prognostic impact of WT1 expression in SNP+ and SNP- subgroups. Median WT1 expression for SNP+ and SNP- patients was 149.76 (range: 0 – 2104.44) vs. 74.40 (range: 0 – 3148.27) respectively (p=0.299). When restricting analysis to SNP rs16754 negative patients (n=150), those in the highest quartile had significantly worse 3-year overall survival (OS) from study entry (51% vs. 72%, p=0.034) and event-free survival (EFS, 35% vs. 54%, p=0.033) than those in lower quartiles. When comparing SNP- patients with higher vs. lower than median WT1 expression (Q3 + Q4 vs. Q1 + Q2), patients with higher expression again had inferior EFS (39% vs 59%, p=0.008) with a corresponding relapse rate from CR of 45% vs. 25% (p=0.011). Notably, similar comparisons in WT1 SNP+ patients with high vs. low WT1 expression did not demonstrate significant outcome differences. Surprisingly, SNP+ patients within the lowest expression quartile had the poorest outcome, while those in the highest quartile had no relapses from CR.

WT1 expression levels vary across cyto-molecular subgroups of AML. WT1 expression was not prognostic when all AML patients were considered; however, stratifying patients by SNP rs16754 genotype revealed that high WT1 expression was significantly associated with poor clinical outcome in SNP- patients. Conversely, although the small sample size of SNP+ patients precluded significant comparisons, increasing WT1 expression trended towards improved outcome in the SNP+ group. The clinical and biologic significance of WT1 over-expression appears to differ depending on the presence or absence of SNP rs16754. WT1 expression must be considered in the context of SNP status, to determine which patients would most benefit from novel WT1 -targeted therapies.

No relevant conflicts of interest to declare.