Introduction: Although risk-adapted therapy improved the prognosis of pediatric T-ALL in the last decades, patients with T-ALL still have a worse outcome compared to B Cell Precursor (BCP)-ALL, and therefore they would benefit from the identification of new prognostic markers for a better treatment stratification and novel strategies to complement current chemotherapy regimens.

Among newly reported genomic abnormalities, a subset of BCP-ALL patients is characterized by the over-expression of the Cytokine Receptor-like Factor 2 (CRLF2) gene, due to either an intra-chromosomal deletion causing the P2RY8-CRLF2 fusion or the IGH@-CRLF2 translocation. These two CRLF2 rearrangements were shown to correlate with poor outcome in BCP-ALL patients. In T-ALL, alterations of CRLF2 were not reported yet, but recently, mutations in its partner IL7Rα have been identified in about 10% of T-ALL patients.

Aim: To estimate the incidence of CRLF2 aberrations and their prognostic value in pediatric T-ALL.

Methods: We analyzed CRLF2 gene expression in 120 T-ALL patients, consecutively enrolled in the AIEOP-BFM ALL2000 study in Italian centers (AIEOP) from September 2000 to July 2005, and, as a validation cohort, in 92 consecutive patients treated with the same protocol in German centers (BFM-G), from July 1999 to December 2004. CRLF2 transcript levels were analyzed by RQ-PCR. Relative gene expression of CRLF2 was quantified by the 2-DDCt method. The DDCt for AIEOP and BFM-G samples was referred to the median DCt of their respective cohort.

Results: An heterogeneous expression of CRLF2 was observed among AIEOP T-ALL patients (range: 0.06 to 82 fold change). Seventeen patients (14.2%) presented an expression 5 times higher than the median (‘CRLF2-high’). Interestingly, none of the CRLF2-high cases resulted to be positive for P2RY8-CRLF2 fusion, 1/5 was positive for the IGH@ translocation and 1/7 showed a supernumerary X chromosome. JAK2 and CRLF2 mutations were absent in all 120 cases, while IL7R mutations were detected in 5/107 patients (4.7%), but unexpectedly they were not associated to CRLF2 over-expression.

CRLF2-high patients had a significantly inferior 5-y EFS (41.2%±11.9 vs. 68.9%±4.6, p=0.006) and an increased, although not statistically significant, cumulative incidence of relapse (CIR) compared to CRLF2-low patients (41.2%±11.9 vs. 26.3%±4.3, p=0.17).

The prognostic value of CRLF2 over-expression was confirmed in the BFM-G cohort (5-y EFS in 12 CRLF2-high patients was 50.0%±14.4 vs 83.8%±4.1, p-=0.008; CIR: 33.3%±13.6 vs. 11.3%±3.5, p= 0.06).

Interestingly, CRLF2-high patients were more frequently allocated to the high risk (HR) T-ALL subgroup (20.9% in HR vs.8.3% in no-HR in the two cohorts analyzed together). In this subgroup, CRLF2 over-expression was significantly associated to a poor prognosis (5-y EFS: 31.6%±10.7 vs. 62.5%±5.7, p-value=0.008; CIR: 47.4%±11.5 vs. 29.2%±5.4, p=0.14).

When analyzed according to prednisone (PDN) response, 17/28 (61%) were prednisone poor responders (PPR), and in the ‘PPR-only’ subgroup (non-HR by other features) 4/9 (44%) CRLF2-high patients relapsed versus 4/36 (11%) CRLF2-low.

Cox model analysis adjusted by risk group showed that CRLF2-high expression had a relevant prognostic impact with a 2-fold increased risk of relapse (Hazard ratio 2.12; 95% CI 1.07-4.21; p=0.03).

The mechanisms responsible for CRLF2 over-expression in T-ALL and its contribution to the pathogenesis of the disease is still being investigated. Notably, gene set enrichment analysis (GSEA) showed an inverse correlation between the expression of CRLF2 and positive cell cycle regulators, thus suggesting that CRLF2-high blasts have a low proliferating activity and may therefore be less sensitive to conventional chemotherapy.

Conclusions: CRLF2 over-expression is a poor prognostic marker not only in BCP-ALL patients, but also in T-ALL, identifying a subset of HR T-ALL patients with extremely severe outcome. Specifically, this marker would allow identifying T-ALL patients that could benefit from alternative therapy, potentially targetting the CRLF2 pathway.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution