Abstract 3251

Recently, an extremely poor risk subset of Philadelphia chromosome (9;22) negative ALL patients was identified that nonetheless have a gene expression profile similar to BCR-ABL positive ALL. The majority of cases in this ALL subset were found to have genetic alterations in IKZF1 (IKAROS). In addition, a significant proportion of these cases had CRLF2 rearrangements, and over half of the CRLF2 rearranged cases also had activating mutations in a JAK kinase. Identifying a biologically-targeted therapy for these patients is important as they are poor responders to conventional chemotherapeutics. Unfortunately, IKZF1 and CRLF2 are not currently targetable, and less than one half of these patients have JAK mutations.

mTOR signaling has been shown to be activated in patients with BCR-ABL+ leukemia, in part through an Akt-independent mechanism. Accordingly, mTOR inhibition is effective in BCR-ABL+ ALL, while Akt inhibition is not. As patients with IKZF1 mutations have a gene expression profile similar to BCR-ABL+ ALL, we hypothesized that mTOR may be activated and that targeting mTOR signaling would be effective in this poor risk ALL subtype, whereas targeting Akt would be ineffective.

To test this hypothesis, we established xenografts from 21 patients with this poor risk ALL subtype using NOG mice. 17 of 21 samples successfully engrafted with sufficient disease burden to generate second and third generation xenografts. We randomized second-generation xenografts from 4 patient samples to treatment with sirolimus (mTOR inhibitor), perifosine (Akt inhibitor), or vehicle control. Disease was evaluated at weekly intervals by FACS of peripheral blood for human CD19+/CD45+ ALL cells. To assess if there was a differential effect based on JAK mutations and/or CRLF2 rearrangements, we treated mice bearing patient ALL samples from each of the following sub-types: Jak+/CRLF2+, Jak+/CRLF2-, Jak-/CRLF2+, and Jak-/CRLF2-.

Mice treated with sirolimus had a profound response with markedly decreased disease in blood and in spleen at sacrifice in all 4 samples (Table 1). Perifosine had no anti-leukemic effect. Both drugs were well tolerated. We confirmed the drugs were hitting target pathways in treated animals by immunoblot for pS6 and pAKT.

SampleAvg (range) pb blasts control mice at sacrificeAvg (range) pb blasts sirolimus treated at sacrificeAvg (range) pb blasts perifosine treatedsirolimus vs control; perifosine vs control (t-test)
Jak-/CRLF2- 5,596 (4,203 – 7,922) 120 (61 – 248) 4,498 (3,522 – 5,474) p = 0.01; p = 0.11 
Jak-/CRLF2+ 30,009 (21,173 – 41,166) 12,764 (7,171 – 14,945) 24,587 (12,980 – 44,498) p = 0.01; p = 0.32 
Jak+/CRLF2- 50,6730 (24,415 – 90,144) 15,523 (13,814 – 17,915) 31,512 (22,621 – 43,669) p = 0.02; p = 0.10 
Jak+/CRLF2+ 6,799 (2,165 – 13,095) 1,676 (398 – 3,431) 5,042 (1,632 – 9,442) p = 0.05; p = 0.31 
SampleAvg (range) pb blasts control mice at sacrificeAvg (range) pb blasts sirolimus treated at sacrificeAvg (range) pb blasts perifosine treatedsirolimus vs control; perifosine vs control (t-test)
Jak-/CRLF2- 5,596 (4,203 – 7,922) 120 (61 – 248) 4,498 (3,522 – 5,474) p = 0.01; p = 0.11 
Jak-/CRLF2+ 30,009 (21,173 – 41,166) 12,764 (7,171 – 14,945) 24,587 (12,980 – 44,498) p = 0.01; p = 0.32 
Jak+/CRLF2- 50,6730 (24,415 – 90,144) 15,523 (13,814 – 17,915) 31,512 (22,621 – 43,669) p = 0.02; p = 0.10 
Jak+/CRLF2+ 6,799 (2,165 – 13,095) 1,676 (398 – 3,431) 5,042 (1,632 – 9,442) p = 0.05; p = 0.31 

In conclusion, targeting mTOR with sirolimus maybe an effective treatment for IKZF1 and JAK mutated ALL.

Disclosures:

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.