Clinical Evaluation of An in Vitro Assay for Rapid Assessment of Cytosine Arabinoside Response in Patients with Acute Leukaemia and Chronic Lymphocytic Leukaemia.

Abstract 1608

Poster Board I-634

Cytosine arabinoside (Ara-C) is the mainstay of treatment for acute non lymphoblastic leukaemia (ANLL). Currently there is no rapid, inexpensive test to assess patient sensitivity to Ara-C prior to treatment. We have previously reported a bioluminescent 8-hour assay which assesses Ara-CTP levels in KG-1a and THP-1 cell lines (Smith et al., Blood 2005; 106(11): p695a). Here we present results of this assay using six further ANLL cell lines, 45 patients with ANLL, 2 patients with Ph+ acute lymphoblastic leukaemia (ALL) and 12 with B-CLL. Over 30% of patients with ANLL fail to respond to Ara-C. Potential causes of resistance include lack of hENT-1 transporter activity, over expression of cytidine deaminase (cdd) and over expression of Y-family DNA polymerases. The 8-hour assay was initially tested on eight ANLL cell lines with known response to Ara-C (CCRF-CEM, K562, MV-4-11, HEL, KG-1a, HL-60, THP-1, Mo7e) and this correlated with a commercially available 3-day cytotoxicity assay (CellTiterGlo‘, Promega). ANLL patient samples were sourced at presentation from patients with blast burdens >80%. The distribution of FAB sub-types was M4 (26%), M2 (20%), secondary AML (20%), M0 (14%), M1 (6%), M3 (6%), JAK2+ (3%) and Ph+ ALL (5%). Patient ages ranged from 24 to 94 (median 67.5 years) and included 25 peripheral blood and 22 bone marrow samples. The Ara-C concentration used in vitro was 25 μM which is equivalent to 2g/m² in the clinical setting. The luminescence cut-off for sensitivity was an increase in light production of more than 10%. The 8-hour luminescence assay showed 100% correlation with the 3-day CellTiterGlo” test results, both indicating that 51% of patients were Ara-C sensitive and the remainder resistant. In the 32 patients for whom clinical outcome data was available, assay results correlated with clinical outcome in 30 patients (p<0.05). The remaining two patients showed results which conflicted with clinical outcome and are therefore under further investigation to examine Y-family DNA polymerase expression. Confirmation of the predictive capacity of the 8-hour assay was observed in two patients with M0 ANLL, initially resistant to Ara-C, now corroborated by the assay, who only achieved remission with FLAG-Ida. The bioluminescence assay was therefore modified to allow a 4-hour pre-incubation with fludarabine and showed marked improvement in blast cell sensitivity to combination treatment compared with Ara-C alone. A further 2 patients with Ph+ ALL and 12 patients with B-CLL were assessed using the 8-hour assay. For the Ph+ ALL patients, one entered complete remission and the second failed to respond to treatment. In both cases this was corroborated by the 8-hour assay and the 3-day CellTiterGlo” assay. The 12 B-CLL patients, all exhibited a highly sensitive profile with the 8-hour assay but in contrast showed resistant results with the 3-day CellTiterGlo” assay. This may indicate the ability of B-CLL cells to incorporate and metabolise Ara-C to Ara-CTP yet this may not translate into cytotoxic effect due to over expression of low fidelity DNA polymerases This new assay system may enable response prediction prior to patients receiving Ara-C based chemotherapy. This may be of importance in high risk patients with adverse cytogenetics/immunophenotype who may require fludarabine with Ara-C to achieve response. Similarly patients with a highly sensitive profile may benefit from a lower dose of Ara-C in vivo thus minimising unwanted side effects. In the event of relapse, repeat analysis may give insight into true clonal evolution and acquisition of chemo-insensitivity.