Mass Cytometry Analysis Dissects CRLF2-Driven Signaling Pathways in Childhood B-Cell Precursor Acute Lymphoblastic Leukemia (BCP-ALL)

BACKGROUND: Rearrangements of the CRLF2 gene, present in 7-15% of childhood BCP-ALL, are responsible of the overexpression of Thymic Stromal Lymphopoietin Receptor (TSLPR) and they are correlated with poor prognosis (Chen IM Blood 2012). TSLPR overexpression can be associated with JAK2 mutations, which leads to aberrant activation of JAK/STAT and PI3K/AKT pathways. Although the cross talk of the signaling pathways is still under investigation, there is a rationale for the use of targeted tyrosine kinase inhibitors (TKIs) to treat this subgroup of patients (Maude SL Blood 2012).

We focused on the dissection of CRLF2-driven signaling in primary CRLF2 rearranged(r) BCP-ALL samples by using single cell mass cytometry (CyTOF) analysis. We leveraged the high dimensional single cell capability of the CyTOF to understand, with previously unattainable resolution, the activation of these pathways simultaneously in single cells and their response to inhibition with TKIs and anti-TSLPR monoclonal antibodies (mAbs). This revealed heterogeneity in signaling response, identifying subpopulations which differentially activate intracellular signals through TSLPR and differentially respond to ex vivo treatment.

METHODS: Twelve BCP-ALL primary samples, 6 CRLF2r and 6 CRLF2 wild type (wt), were investigated and the expression of 24 phenotypic and 15 functional proteins were measured at single cell level using CyTOF as previous described (Bendall SC Science 2011). To assess the response to ex-vivo TSLP stimulation (10ng/mL) and TKIs/mAbs treatment, data were normalized to the basal levels of each phosphoprotein and significance was calculated using student`s t-test. One million cells per condition were treated with different TKIs, Dasatinib, Ruxolitinib and BEZ-235, and two different clones of anti-TSLPR mAbs (130A10 and 130H3) from MRC Technology.

RESULTS: As expected, we observed an aberrant TSLP-induced activation of pSTAT5 and prpS6 in CRLF2r patients as compared with CRLF2wt, used as control group (p=0.0055, p= 0.0007). Of note, we also observed a previously not described TSLP-dependent activation of pERK and pCREB (p=0.0313, p=0.0261) suggesting a cross-talk of the TSLPR-driven signaling also with the RAS/MEK pathway.

Treatment with TKIs revealed strong inhibitory activity of Dasatinib, which completely inhibited the TSLP-mediated phosphorylation of STAT5, rpS6, CREB and ERK in CRLF2r treated blasts compared to CRLF2r not treated cells (p= 0.0040, 0.0017, 0.0007, 0.0114 respectively). Ruxolitinib, JAK1/2 inhibitor, also reduced rpS6, CREB and ERK phosphorylation (p=0.0025, 0.037, 0.0132). Interestingly one of the two anti-TSLPR tested mAbs (130A10) was also able to significantly inhibit the TSLP-mediated activation of STAT5, rpS6, and ERK (p= 0.0071, 0.0006, 0.0323). Finally, the PI3K/TORk inhibitor, BEZ-235, did not show any statistically significant reductions.

Single cell analysis revealed a population of TSLPR overexpressing blasts (range 20-50%) in which the TSLP stimulation resulted in activation of prpS6 but not pSTAT5, present in all the CRLF2r patients. This rpS6 activation could be inhibited by anti-TSLPR mAb, Dasatinib, Ruxolitinib and BEZ-235, except for one patient in which the activation was blunted only by anti-TSLPR mAb and Dasatinib suggesting an activation of prpS6 through a non canonical pathway. This data reveals heterogeneous signaling populations present within this subtype of leukemia driven by TSLPR overexpression.

Finally in 3 additional CRLF2r primary samples, we investigated signaling profile of residual blasts (MRD) at Day8 and Day15 post induction initiation. TSLPR expression was consistently maintained in all patients at both time points. Furthermore, residual blasts were still able to respond to TSLP and the induced pSTAT5 could be effectively inhibited by 130A10 anti-TSLPR clone and Ruxolitinib.

CONCLUSION: In summary, these data suggest heterogeneity of TSLPR-related signaling with activation of the expected JAK/STAT and PI3K pathways but also RAS/MEK and CREB activation. Further, TSLPR+ blasts exhibit heterogeneous responses to both treatment with TSLP in combination with TKIs or mAb. Finally, the MRD detection by CyTOF allowed the study of the functional activity of the TSLPR positive resistant cells suggesting a role of CRLF2r in the persistence of the leukemic cells and its targeting to treat late and refractory stages of the disease.