Engraftment and Long-Term Survival at Low Burden of Leukemic Blasts from Primary MRD+ Human Bone Marrow in a Xenotransplant Setting

Minimal residual disease (MRD) in bone marrow (BM) at the end of induction or re-induction chemotherapy is a critical prognostic factor for long-term survival of children with acute lymphoblastic leukemia (ALL). The presence of submicroscopic disease is a significant predictor of leukemia relapse, such that MRD positivity (0.01% to 5%) informs risk-adapted treatment intensification following induction. As most cases of relapsed ALL involve a clone that existed as a minor subclone at diagnosis, strategies that enable characterization of the leukemic blasts surviving at the end of induction could further influence treatment selection; however, the low numbers of leukemic blasts in an MRD sample makes evaluation of their sensitivity to chemotherapy or other treatment modalities challenging.

Immune-deficient NOD-scid/IL2Rγc null(NSG) mice are highly receptive to the engraftment and expansion of ALL blasts from diagnostic and relapse samples. However, little is known about the reproducibility of this approach for the small numbers of leukemic cells in primary MRD+ BM samples. To investigate the fate of MRD blasts following xenografting, MRD+ mononuclear cells from BM samples from 6 children with B-ALL in first (n=3) or subsequent remissions (n=3), with a median MRD level of 0.65% (range: 0.08-4.3%), were adoptively transferred into NSG mice by tail vein injection. Engraftment and expansion were monitored by regular flow cytometric evaluation of peripheral blood for human CD45 expressing cells, and mice were euthanized at onset of leukemia (human cell count > 15,000/ul) or after up to 1 year on study. BM, spleen and peripheral blood were harvested at time of death, and cells immunophenotyped with the identical 10-colour flow panel following a COG-validated, clinical MRD diagnostic algorithm.

Human cell engraftment was confirmed in all recipient mice. Recipients of two of the MRD samples progressed to frank leukemia: one mouse injected with an MRD burden of 4.3% presented with an elevated WBC after 6 months and a blast count of 67% in spleen, while another mouse euthanized one year after injection of a 0.16% MRD+ sample had 75% blasts in BM that comprised both leukemic subclones present in the original MRD sample. Of the two MRD samples, only the former was from a patient that subsequently relapsed. Although all other recipient mice continued to show fluctuating, but sustained low levels of peripheral human cells (<5% of human CD45+ cells among mouse CD45+ peripheral blood cells), none progressed to overt leukemia. At sacrifice (range: 179-390 days, median 260 days), all remaining recipient mice were confirmed to contain leukemic blasts at low burden in BM, ranging from 0.001% to 0.675% of immunophenotypically verified human leukemic blasts in 1-1.5 million of total cells. An in-depth comparison of primary MRD samples and the mouse engrafted blasts revealed that similar immunophenotypic blast populations identified in the patient MRD sample were sustained for as long as 1 year in NSG mice; in several mice there was evidence of maintenance of two to three distinct immunophenotypic subclones that were present at day 0 and/or day 29 in the patient. Furthermore, in all recipient mice the vast majority of human cells were phenotypically confirmed to be blasts and no significant amount of normal human hematopoiesis following injection of the MRD samples was detected.

This study demonstrates that heterogeneous leukemic cell populations from primary MRD+ BM samples can be successfully engrafted and sustained in NSG mice and that, in at least a subset of recipients, these blasts can progress to overt leukemia. Notably, one MRD-recipient mouse presented with leukemia prior to any progression in the patient, raising the possibility that this approach could be used to characterize potential relapse clones ahead of their clinical appearance. Furthermore, the sustained presence of leukemic blasts in all non-progressing recipients for as long as 1 year demonstrates the reproducibility of this approach for generating a stable, low level leukemia burden that resembles the MRD state. Such a setting may enable the evaluation of therapeutic interventions for efficacy against the residual blasts that will give rise to relapse in the patient.