Effects of LMO2 on Human T-Cell Development Are Modulated by Notch Signaling

Abstract 2470

LMO2 is overexpressed in a significant percentage of human T cell acute lymphoblastic leukemia (T-ALL) and its locus has been the target of insertional mutagenesis in gene therapy trials. In the past years, 4 X-linked severe combined immunodeficiency (X-linked SCID) and one Wiskott-Aldrich syndrome (WAS) patients who were treated by retrovirus-mediated gene therapy developed T-ALL as a result of retroviral integration in the LMO2 locus. In these patients, leukemia developed 2 to 3 years after gene therapy without prior significant haematological abnormalities. However, both the latency of disease and the finding of additional somatic mutations and/or translocations in these leukemias suggest that the overexpression of LMO2 alone is insufficient to generate leukemia, a notion that has been supported by studies in mouse. Though LMO2 is typically recognized as a T-cell oncogene, reports have shown that it is also aberrantly expressed in acute myeloid leukemias (AML), chronic myeloid leukemia (CML), B-ALL and some non-hodgkin B cell lymphomas.

In order to study the impact of LMO2 overexpression on human hematopoietic stem/progenitor cells, a lentiviral vector was used to express this oncogene together with EGFP in lineage-depleted umbilical cord blood. In myeloid-promoting cultures, LMO2 had no effect on either differentiation or proliferation. Moreover, the expression of LMO2 did not modify the frequency or lineage distribution of colony forming progenitors compared to controls. However, significant differences were noted when transduced cells were assayed on OP9-Delta-Like 1 (DL1) stroma, an in vitro system that promotes T cell proliferation and differentiation. Cells overexpressing LMO2 were blocked at the double negative stage (CD4-CD8-) of differentiation and proliferated 50 to 100 times more than control cells. However, these cells were not immortalized as they proliferated for a median of 75 days, versus 50 days for controls.

Immunodeficient mice transplanted with primitive human hematopoietic cells expressing LMO2 (hereafter referred as LMO2 mice) had bone marrow engraftment levels comparable to controls at 20–24 weeks post-transplant. Neither B-lymphoid nor myeloid development were affected by LMO2 overexpression. Strikingly, in the thymus, the percentage of EGFP+ cells was significantly increased in LMO2 mice compared to controls (mean of 47.7% versus 8.8%, p=0.0001), clearly indicating that expression of this oncogene enhances thymic T-cell engraftment. We next analyzed the phenotype of LMO2-expressing T cells in the thymus and peripheral blood of these mice. Surprisingly, unlike our in vitro studies, there was no evidence of a block at the DN-stage of differentiation. Instead, there were significantly less EGFP+ DN cells in the thymi of LMO2 mice compared to controls (mean of 7.5% vs 14.5%, p=0.035). These results clearly demonstrate that unlike what was observed in OP9-DL1 co-cultures, LMO2 overexpression does not induce a block in T-cell differentiation in our in vivo system. One possible explanation for this difference is the constitutive NOTCH signaling provided via DL1 on stroma compared to the in vivo setting where LMO2-expressing cells would encounter different levels and forms of NOTCH signaling throughout development. To test this hypothesis, LMO2 cells were cultured on OP9-DL1 stroma for 50 days then switched onto OP9 stroma lacking NOTCH ligand. Upon transfer, the DN cells promptly stopped proliferating and differentiated into DP (CD4+CD8+) cells expressing CD3 and TCRαβ.

Thus, our results suggest that in the in vivo setting, as cells migrate through the thymus and face a decrease in NOTCH signaling, LMO2 overexpression alone can promote proliferation, but is not sufficient to maintain a differentiation block. However, constitutive NOTCH signaling can cooperate with LMO2 overexpression to block T cell differentiation at a proliferative DN stage. Thus, one can postulate that LMO2 exerts a proliferative effect on developing T-cells in thymic regions with high levels of NOTCH signaling, potentially providing a setting for the development of secondary leukemogenic events. NOTCH mutations are common in human T-ALL and can therefore allow for LMO2 overexpressing cells to become independent of the stromal niche. Taken together, our results suggest cooperation between LMO2 overexpression and NOTCH signaling in human T-cell leukemogenesis.