Prognostic Significance Of CRLF2 Over-Expression In Pediatric Acute Lymphoblastic Leukemia

Previous studies have reported that IKZF1 deletion was an adverse prognostic factor in pediatric B-cell-precursor ALL (BCP-ALL). However, the prognostic significance of CRLF2 over-expression (OE) is controversial. In order to assess the prognostic value of CRLF2 OE and genetic alterations involving CRLF2, we conducted genetic analysis of CRLF2 in pediatric BCP-ALL treated according to Japan Association of Childhood Leukemia Study (JACLS) ALL02 cohort.

We examined diagnostic bone marrow or peripheral blood samples of 167 pediatric BCP-ALL patients treated on the JACLS ALL02 protocol. The analyzed cohort included 68 patients classified in NCI-SR and 99 patients in NCI-HR. The deletion of IKZF1 and gain of CRLF2 were determined by MLPA. CRLF2 expression level was determined by quantitative RT-PCR and OE was defined as 10 times more than mean expression level of 167 cases. The P2RY8-CRLF2 fusion was detected by RT-PCR. To identify fusion transcript related to CRLF2 OE, we performed messenger RNA sequencing (mRNA-seq) on six of 17 patients with CRLF2 OE without CRLF2 gain and P2RY8-CRLF2 fusion.

CRLF2 OE was identified in 30 (18%) of 167 patients, which was similar to that reported previously. IKZF1 deletion was found in 25(15%) of 167 patients, which was found more in CRLF2 OE patients than non-CRLF2 OE patients (30% vs 11%, p<0.05). P2RY8-CRLF2 fusion was identified in only 3 (1.7%) of 167 patients, and all of them were classified in CRLF2 OE. In detail, only one of three P2RY8-CRLF2 positive patients had IKZF1 deletion. One of two P2RY8-CRLF2 positive patients without IKZF1 deletion harbored MLL-AF4 fusion. CRLF2 gain was identified in 18(11%) of 167 patients. Eleven of these 18 patients were classified in CRLF2 OE, suggesting CRLF2 gain was significantly related to CRLF2 OE (37% vs 5%, p<0.01). Interestingly, none of CRLF2 OE patients with CRLF2 gain did have IKZF1 deletion. We identified one patient with novel fusion transcript caused by the 32kb deletion from CRLF2 intron 1. This patient was classified in CRLF2 OE with IKZF1 deletion, suggesting that the novel fusion transcript was related to CRLF2 OE. However, we could not identify any fusion transcripts related to CRLF2, such as IgH@-CRLF2, in remaining five patients by mRNA-seq. Interestingly, we identified three EBF1-PDGFRb positive patients in CRLF2 OE patients with IKZF1 deletion who did not have genetic alterations including CRLF2. In survival analysis, significant difference on the 5-year event-free survival (EFS) and overall survival (OS) between patients with and without CRLF2 OE was not observed (71% vs 75%, log rank p=0.68, 96% vs 82%, log rank p=0.11). In addition, type of genetic alterations related to CRLF2, such as CRLF2 gain and P2RY8-CRLF2 fusion, did not show the impact on EFS and OS in CRLF2 OE patients. However, significant difference on 5-year EFS between CRLF2 OE patients with and without IKZF1deletion was observed (44% vs 83%, log rank p=0.02). In multivariate analysis of 167 patients, only IKZF1 deletion was significant predictors of inferior OS (Hazard ratio: 2.427, 95%CI:1.037-5.679, p=0.04).

The current analysis revealed that CRLF2 OE was caused by several different mechanisms, such as CRLF2 gain, P2RY8-CRLF2 fusion and rare fusion transcript related to CRLF2. However, regardless of type of genetic alterations of CRLF2, concomitant IKZF1 deletion holds the major impact on poor prognosis in CRLF2 OE patients. Especially, CRLF2 gain and IKZF1 deletion were mutually exclusive. Thus, CRLF2 OE caused by CRLF2 gain was not poor prognostic factor for high risk BCP-ALL. The prognostic significance of CRLF2 should be evaluated in consideration of IKZF1 status.

No relevant conflicts of interest to declare.