Specific MLL Partner Genes in Infant Acute Lymphoblastic Leukemia (ALL) Associated with Outcome Are Linked to Age and White Blood Cell Count (WBC) at Diagnosis: A Report On the Children's Oncology Group (COG) P9407 Trial.

MLL translocations are the most pervasive molecular abnormalities in infant ALL and are poor prognostic factors; however, interrelationships between specific MLL partner genes, event-free survival (EFS) and clinical covariates have been elusive due to several factors. Infant ALL is a rare disease with limited cases within MLL translocation subsets, specialized methods are needed to identify the spectrum of MLL partner genes, and most prior clinical trials defined MLL status by karyotype and analyses of only common MLL partner genes. P9407, a COG pilot study testing the utility of intensified induction therapy, developed a detailed molecular cytogenetic classifier of MLL status and partner genes to investigate relationships with EFS and correlations with clinical covariates.

MLL status was defined by karyotype, FISH, Southern blot analysis, RT-PCR analysis of MLL translocations with AF4, ENL or AF9, and cDNA or genomic panhandle PCR. Kaplan-Meier plots were used to estimate EFS as a function of MLL status (rearranged, R; germline, G), MLL partner gene and, within MLL status and partner gene subsets, as a function of age at diagnosis. Fisher's exact test was used to assess correlations between MLL partner genes and age, WBC (K/ml³) or CNS status at diagnosis.

MLL status was determined in 210 of 221 total cases [95%; 158 R, 75%; 52 G, 25%]. Partner genes were definitively determined in 133 of the 158 MLL-R cases: AF4/78/49.4%; ENL/33/20.1%; AF9/15/9.5%; EPS15/4/2.5%; AF10/1/0.6%; ASAH3/1/0.6%; ACTN4/1/0.6%. The ACTN4 partner gene was identified in a new complex 3-way MLL/ACTN4/RYR1 rearrangement. In the other 25 MLL-R cases, RT-PCR with gene specific primers excluded AF4 (n=5) or AF4, AF9 and ENL (n=13) or MLL status was not defined beyond MLL-R or MLL-G (n=7). Analyses of 202 eligible subjects within the cohort of 210 in whom MLL status was assigned, indicated a 5-year EFS rate of 45.9±4.6%, which was worse in MLL-R than MLL-G cases (38.8±5.1% v. 66.2±9.3%, P=0.0003). By category, 5-year EFS rates were: ENL (29.0±10.0%), AF4 (34.2±7.4%), other partner genes (45.0±14.9%), AF9 (67.7±17.2%), MLL-G (66.2±9.3%) (P=0.0003). Within MLL-R, AF4 was associated with worse EFS than other partner genes including AF9 but excluding ENL (P=0.043), and ENL with worse EFS than other partner genes including AF9 but excluding AF4 (P=0.024). Also, EFS was worse when AF4 and ENL subsets were combined, compared to any other partner gene (P=0.025). Conversely, the higher EFS with AF9 than with any other partner gene became less significant when AF4 and/or ENL were excluded from the other MLL-R category. EFS in the AF9 subset was not different from that for MLL-G cases. Five-year EFS was worse in patients aged <90d than ≥90d at diagnosis within MLL-R cases overall (6.2±4.2% v. 47.4±6.5%, P<0.0001) and within the AF4 (4.6%±4.4% v. 46.6±9.5%, P<0.0001), or ENL subsets (0.0% v. 40.9±12.8%, P=0.0011), whereas age <90d did not adversely impact survival in the MLL-G cases or those with AF9 as partner gene. Compared to cases with different partner genes, cases with AF4, ENL or AF9 were not differently distributed in the <90d v. ≥90d age categories, but there were more other MLL-R cases in patients aged ≥90d at diagnosis (18% v. 3%, P=0.027). WBC was higher (≥50K) in cases with AF4 (P=0.0006) and lower (<50K) in cases with AF9 (P=0.0007). CNS status was not differently distributed by partner gene. In similar analyses of cohort 3 alone (n=136) in which treatment modifications were made for earlier toxicities, the same relationships were upheld.

MLL translocations, young age and high WBC have correlated with poor outcome in infant ALL before, but this MLL classifier in the largest cohort of infant ALL with detailed molecular characterization reported to date, adds new elements to the impact of MLL partner genes on outcome and their connection to prognostic factors. AF4 and ENL not only negatively impact EFS, but this effect is larger in the younger infants. Regardless of age, AF9 is a favorable prognostic factor, but the other partner genes associated with better outcome occur more often in older infants. Relationships between high WBC and AF4, and low WBC and AF9 were also discovered. The newfound associations of specific MLL partner genes with age and WBC indicate how disease biology and the classical prognostic factors in this disease are integrally connected.

No relevant conflicts of interest to declare.