Abstract 1432 Background:The origin of relapse in AML is believed to be related to persistence of resistant “leukemia stem cells.” in the bone marrow microenvironment where adhesion confers drug resistance. Engraftment of human AML in immunodeficient mice is dependent on CXCR4 (Tavor et al 2004). CXCR4 inhibitors, such as AMD3100 (plexiglass,hereafter P), overcome adhesion mediated chemotherapy resistance (Zeng et al 2006) and mobilize human leukemia engrafted in immunodeficient mice (Zeng et al 2009). P also mobilized leukemia in an APL murine model and in combination with chemotherapy reduced tumor burden (Nervi et al 2009). Methods:We studied the combination of P 5mg/kg daily sc X 3, cytarabine (araC=A) 300mg/kg IP X 3 and clofarabine (C), 20mg/kg IP X 3 in the NODscid IL2R γc−/− mouse engrafted with primary patient AML CD34+ cells after 350 cGy total body irradiation. We could first detect circulating human CD45+ or human CD34+ cells, denoting engraftment, by flow cytometry as early as 5–13 weeks. We then injected plerixafor to assess mobilization capability at 8–16 weeks, followed by the combination of plerixafor and chemotherapy. Animals were sacrificed by 14–38 days after chemotherapy, and assessed for AML in blood, marrow, and spleen. Results:A single 5mg/kg dose of P, produced a 2.26 ± 0.94 (SD) fold increase in peak mobilization (at 2 hours) compared to saline control, p=0.026. P-induced mobilization was directly related to expression of CXCR4, with a patient exhibiting 10.3% CXCR4 showing 0.86× baseline, as compared to a patient with 24.7% CXCR4 exhibiting a 2.2-fold increase, and 84.9% CXCR4, a 3-fold increase. Chemotherapy,described above, was given 2 hours after plerixafor. For animals that received P/A vs. P/A/C, there was no statistically significant difference in leukemic burden (in millions of human CD34+ AML cells ± SD) of the animals sacrificed 14 days after initiation of treatment: bone marrow six bones 116.3 ± 33.7 vs. 111.7 ± 29.2 (p=0.86), spleen 50.8 ± 10 vs. 43.7 ± 19.1 (p=0.59), blood 10.9 ± 9.6 vs. 3.1 ± 1.4 (p=0.16), or estimated total body burden 178.0 ± 45.3 vs. 158.5 ± 30 (p=0.52). A comparison of 4 groups of animals, P/A/C vs. P/A vs. A/C vs. A demonstrated a statistically significant difference between certain groups, at certain time points. For example, on day 10, P/A/C treated animals had a lower human PB CD34+ count than P/A, 0.07 vs. 0.24 × 109/L (p=0.034). On day 14, P/A/C had a lower CD34+ count than A/C, 0.08 vs. 0.16 (p=0.047). But at day 38, the leukemia had already recurred, and there was no statistically significant difference in the organ or total body involvement by the leukemia. There was a very low white blood count days 5–24 post chemotherapy, with only minimal residual disease detectable by flow cytometry (analogous to a period of remission in humans), but by day 38, the leukemia had recurred (Figure 1), and there was no statistically significant difference in the organ or total body involvement by the leukemia amongst the groups (Figure 2). Conclusion:This model demonstrates the efficacy of chemotherapy in reducing circulating leukemia, but resistant cells appear to remain sheltered and give rise to relapse. Despite the beneficial effects of CXCR4 inhibitors in vitro and in vivo in a murine APL model, the combination of plerixafor with chemotherapy did not prevent or postpone leukemic relapse. These results could be attributed to either inefficient scheduling of plerixafor (for example, continuous infusion may have worked better), or non-cycling cells may be preferentially mobilized (Bonig H et al., 2009), and thus less susceptible to cytarabine treatment. These concepts are currently being explored. Alternatively, concomitant inhibition of other adhesion receptors may be necessary to prevent leukemia from homing back to the marrow. Precise timing and degree of mobilization in combination with chemotherapy may be required to optimize this approach in the clinic. [Display omitted] [Display omitted] Disclosures:Becker:Sanofi-Oncology (Genzyme): Research Funding.