Inactivation of the INK4A/ARF (CDKN2) locus Cooperates with HMGA1 in T-Cell Leukemogenesis.

Abstract 3969

Poster Board III-905

Because T-cell acute lymphoblastic leukemia (T-ALL) is an aggressive lymphoid leukemia with a high rate of relapse in both children and adults, research is urgently needed to discover molecular pathways that could be targeted in therapy. Recent studies have identified NOTCH1 activating mutations, genomic classifiers, and other molecular pathways that are disrupted in T-ALL, although these findings have not yet translated into better therapy or improved long-term survival. Thus, we are investigating molecular pathways that play a causative role in T-ALL and could be modulated by therapy. We previously showed that HMGA1a is overexpressed in human T-ALL. Further, HMGA1a transgenic mice develop aggressive T-ALL with complete penetrance, indicating that HMGA1a induces leukemogenic transformation in vivo. Disruption of the CDKN2A (INK4A/ARF) tumor suppressor locus has been reported in up to 90% of T-ALL, suggesting that this contributes to leukemic transformation in T-cells. To determine if loss of function of the INK4A/ARF tumor suppressor locus could cooperate with HMGA1a in T-ALL, we crossed our HMGA1a transgenics with mice that are null for the INK4A/ARF locus. We found that T-ALL was markedly accelerated in the HMGA1a transgenics that are also null at this tumor suppressor locus. These mice died significantly earlier than the HMGA1 transgenics that are wildtype at the INK4A/ARF locus (4.8 months+/- 1.0 month versus 10.5 +/- 2.5 months, p < 0.0001 by the student's t-test). The HMGA1a transgenic, INK4A/ARF null mice recapitulate salient clinical and pathologic features of human T-ALL. All mice appeared ill and exhibit marked organomegaly with extensive leukemic infiltration. Immunophenotyping by fluorescence activated cell sorting showed T-ALL in all cases with identical T-cell markers (CD3, CD8, and αβ-TCR) to those previously reported for the HMGA1a transgenics at 8-10 months. Interestingly, the HMGA1a-INK4A/ARF+/- mice also developed early lymphoid malignancy in 6/17 cases. The accelerated lymphoid malignancy in these mice suggests that there is loss of heterozygosity at the INK4A/ARF locus. Alternatively, haploinsufficiency of INK4A and/or ARF may be sufficient to promote lymphoid tumorigenesis in the setting of HMGA1a overexpression. To determine if HMGA1 overexpression and loss of INK4A/ARF expression occur in human T-ALL, we analyzed gene expression profile analysis in leukemic blasts from children and young adults with T-ALL. We found high levels of HMGA1 expression and low levels of INK4A/ARF expression in most patients. Taken together, our results indicate that overexpression of HMGA1 and loss of function at the INK4A/ARF (CDKN2A) tumor suppressor locus cooperate in T-cell leukemogenesis and these pathways could represent rational therapeutic targets. Further, our mice provide a novel preclinical model for T-ALL.