P321: tnf-mediated cell death: an actionable target for immunotherapy in t-cell acute lymphoblastic leukemia

Background: T-cell acute lymphoblastic leukemia (T-ALL) is an aggressive hematological malignancy with a dismal prognosis in the case of primary refractory disease or relapse. The phenotypic and oncogenic landscapes of T-ALL are complex, hindering the design of efficient targeted therapy. Previously published results reported on the induction of apoptosis secondary to a chronic T cell receptor engagement with anti-CD3ε antibody (OKT3) via a signaling program related to developmental negative selection in sCD3+ leukemia models. Aims: We aim to evaluate the effect of a polyclonal IgG cocktail, thymoglobulin (ATG), on human T-ALL. We hypothesized that this treatment contains antibodies against different targets expressed at the cell surface of T-ALL (CD3 among others) and could potentially be an efficacious therapy for T-ALL. Methods: We tested the cytotoxicity of ATG in T-ALL cell lines and a series of human T-ALL ex vivo and in vivo (in curative settings) representing the different subtypes of all T-ALL in terms of phenotype and molecular landscape. We RNAsequenced sensitive and resistant cells exposed to ATG and compared results to control to identify enriched pathways after ATG exposure in both groups. Results: We screened a large series of 45 Patient-Derived-Xenograft (PDX) for their response to ATG. 27 T-ALL PDXs (60%) were responsive as early as 24h of exposure. This elevated response rate was obtained in most sCD3+ T-ALL (n=14/17), and strikingly, in a significant proportion of sCD3- T-ALL (n=13/28). We then tested in vivo 9 distinct human T-ALL PDX models, 2 of which expressed a surface CD3 (sCD3) and 7 were absent for sCD3. For all predicted ATG-responsive T-ALL, ATG treatment showed strong anti-leukemic effects, as demonstrated by the delayed dissemination of leukemic blasts in the blood and improved mice survival (Fig1A). We RNA-sequenced 7 T-ALL PDXs that were responsive ex vivo. Differential gene expression analysis in ATG vs control condition was performed and GSEA identified the TNFα signaling via NF-κB pathway as the only enriched pathway in these PDXs. No significant enrichment of this pathway was identified in resistant T-ALL (Fig1B). In line with this, activated caspase 8 was induced after ATG in ALL SIL and JURKAT responsive cell lines compared to the resistant cell line RPMI-8402. Similar results were also observed in 4 ATG-responsive PDXs. Moreover, after a shorter exposure to ATG, ALL-SIL and JURKAT (ATG-responsive) induced expression of pMLKLS358 protein, which plays a major role in TNFRSF-induced cell death through necroptosis unlike RPMI-8402. These results suggest the role of cell death induced by the TNF receptor superfamily leading to apoptosis (caspase 8 activation) or necroptosis (RIPK3 and MLKL activation) after ATG exposure (Fig1C). In parallel, we identified through RNAsequencing a cluster of genes constantly increased through ATG exposure (H1, H4, H24) and belonging to TNFα signaling via NF-κB pathway. The BIRC3 gene encoding for the cIAP2 protein was strongly induced after ATG in sensitive and resistant T-ALL PDX (n=3). We subsequently tested the effect of the combination of birinapant and ATG ex vivo in 22 PDXs. All six ATG-responsive benefited from the combination. Among the 16 ATG-resistant cases, 13 had a decreased cell viability but 3 remained resistant. Critically, all the PDXs resistant to Birinapant and ATG alone (n=11) obtained an additive (n=5) or synergic effect (n=6) with a significant decrease in cell viability (Fig1D). Summary/Conclusion: These results provide a strong rationale for the combination of SMAC mimetic with ATG immunotherapy inducing TNFα signaling in T-ALL.Keywords: T cell acute lymphoblastic leukemia