Abstract 3767

Poster Board III-703


We and others recently identified the type I-like cytokine receptor CRLF2 as an oncoprotein in adult and high-risk pediatric precursor B-cell acute lymphoblastic leukemia (B-ALL). Wild-type CRLF2 heterodimerizes with IL7R and signals in response to thymic stromal lymphopoietin (TSLP) through unclear mechanisms that include STAT5 activation. In cases of B-ALL that overexpress CRLF2, the CRLF2 locus is rearranged through either a translocation with IGH or an intrachromosomal deletion. CRLF2 rearrangements are confined to patients who lack translocations involving MLL, E2A, TEL, and BCR/ABL. A subset of cases with CRLF2 overexpression also harbors a CRLF2 Phe232Cys gain-of-function mutation, while a separate subset has mutations in Janus Kinases (JAK). Cells transformed by CRLF2 Phe232Cys or the combination of wild-type CRLF2 with JAK2 exon 16 mutants have constitutive activation of STAT5 and ERK, and upregulation of STAT target genes. However, only the cells that express JAK2 mutants have constitutive phosphorylation of JAK2, suggesting that CRLF2 Phe232Cys signals through an alternate kinase. We used transcriptional profiling to identify candidate pathways and then screened available kinase inhibitors for selective toxicity in cells transformed by CRLF2 and JAK2 mutant alleles.


We reviewed transcriptional profiles obtained from 22 adult patients with B-ALL enrolled on Gruppo Malattie Ematologiche dell'Adulto (GIMEMA) protocols LAL2000 and LAL2004. All samples were confirmed to have high (n=8) or low (n=14) CRLF2 expression by quantitative RT-PCR. Samples were assayed on the Affymetrix HG-U133A platform, which contains a single probe set (208303_s_at) that targets the CRLF2 transcript (both complete and partial coding sequence, as well as expression sequence tags). Supervised analysis was performed using a t-test: probe sets were 1.5, and expression≥0.005 and fold change ≤required to have a p-value >100 in at least one group. For the kinase inhibitor screen, BaF3 cells that stably express CRLF2, IL7R and/or JAK2 alleles were incubated with compound for 48 hours in the absence of IL3. Viability was measured by ATP content with luminescence using CellTiterGlo and read on an Envision plate reader.


Supervised analysis defined a 105 gene (130 probe set) signature associated with CRLF2 overexpression. This signature included upregulation of CD10, CD99, the STAT5 induced regulator SOCS2 and Protein Kinase C (PKC) iota. Downregulated genes included class I and class II major histocompatibility antigens. Based on the involvement of JAKs in CRLF2 signaling and the overexpression of PKC iota, we carried out a kinase inhibitor screen. Compared with cells that express wild-type CRLF2/IL7R grown in the presence of TSLP, cells transformed by CRLF2 Phe232Cys or wild-type CRLF2/mutant JAK2 had 10-100-fold lower EC50 for small-molecule inhibitors of JAK signaling and agents with activity against PKC family kinases (Go6976, UCN-01, k252a). Assays to confirm these findings in primary B-ALL specimens are underway.


CRLF2/JAK signaling confers a unique pattern of kinase inhibitor sensitivity that can be elucidated through transcriptional profiling and targeted with available agents.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.