S128 BCP‐ AND T‐ALL CELLS HIDE IN DISTINCT NICHES OF THE BONE MARROW

Background: Persistence of miminal residual disease (MRD) is a strong indicator for drug resistance in acute lymphoblastic leukemia (ALL). The contribution of the bone marrow microenvironment to resistance remains incompletely understood. In‐vivo imaging in a murine model of T‐cell ALL (T‐ALL) infirm the existence of distinct predilection sites for leukemia cells in the bone marrow. A recent report based on experiments with patient‐derived‐xenografts (PDX) has suggested a role for a dormant subpopulation in B‐cell precursor ALL (BCP‐ALL), possibly in vicinity to the bone endothelium, to confer chemoresistence. Aims: We here set out to compare the topology of the bone marrow microenvironment in primary patient‐derived BCP‐ and T‐ALL under therapeutic pressure in PDX models. Methods: We established a PDX model of ALL induction chemotherapy with Dexamethasone, Vincristine and Doxorubicin in NOD/scid ILR2gamma null mice and analyzed the distribution of residual ALL cells in relation to bone and vascular structures using 3D confocal immunofluorescence microscopy of different bones at specific timepoints before and after treatment. We used cell‐type specific immunofluorescence staining combined with 3D confocal z‐stack tiling and applied distance analysis algorithms to define the cellular components of the leukemia niche. Results: Analyzing long bones from 5 BCP‐ALL and 3 T‐ALL PDX both at early engraftment (4 or 11 days after transplantation) and at MRD (after induction chemotherapy), we show that BCP‐ALL cells are located in close contact to the extravascular side of endothelial cells of bone marrow sinusoids. BCP‐ALL cells were enriched within a distance below 5 μm from endothelial cells and reproducibly more distant from arterioles and transition zones. In contrast, T‐ALL PDX at early engraftment and after induction chemotherapy were spread over the bone marrow with a marked tendency to cluster close to the endosteum after induction chemotherapy (∼30 % of all T‐ALL cells). We next attempted to identify subsets with slow cell cycle rates using Carboxyfluoresceinsuccinimidylester (CFSE) labelled ALL PDX. In our model of induction chemotherapy, proliferation of BCP‐ as well as T‐ALL cells was transiently decelerated. However, we could not detect any persistence of a reproducible population with reduced proliferative activity, challenging the notion that a dormant ALL subpopulation can be detected by this approach. Moreover, residual CFSE retaining cells after chemotherapy localized next to bone marrow sinusoids for BCP‐ALL and more widespread within the bone marrow for T‐ALL. Finally, we confirm an increase of CFSE retention in distal adipocytic rich yellow bone marrow (e.g. tail vertebra) compared to the red bone marrow of long bones, suggesting differences that may impact on the activity of therapeutic agents targeting rapidly cycling leukemia cells. Summary/Conclusion: Our data reconcile previous reports by revealing robust differences between BCP‐ and T‐ALL with respect to the topology of the leukemia niche in MRD. BCP‐ALL MRD was detected mostly in the perivascular space, while T‐ALL MRD was less specifically distributed with some clusters at the endosteum. This provides the basis to investigate molecular determinants of BCP‐ALL interactions with the vascular compartment, which may leverage new candidate targets for therapeutic intervention.