Hijacking the Niche: The Role of Stromal Osteopontin in the Induction of Leukemic Dormancy within the Bone Marrow Microenvironment

Acute lymphoblastic leukemia (ALL) is a treatable malignancy where initial induction chemotherapy achieves clinical remission in the majority of patients. Relapsed disease, however, occurs in many patients and is significantly more difficult to treat. The majority of relapsed cases are a result of minimal residual disease (MRD) that persists within the bone marrow (BM) after initial chemotherapy. Our evolving knowledge of the importance of the host microenvironment in tumor progression suggests that the stromal microenvironment can protect leukemic cells from chemotherapeutic assault, and that inhibiting the supportive relationship between leukemic blasts and the bone marrow microenvironment (BMM) will provide novel therapeutic opportunities. We therefore aimed to identify and characterize novel stromal signaling mechanisms that retain and support blasts within the malignant BMM. Our preliminary data suggest that osteopontin (OPN), normally secreted by osteoblasts within the marrow, is one such signaling chemokine that is highly upregulated in the leukemic niche. OPN has well-defined roles in both solid tumor metastasis and normal hematopoietic stem cell function within the BMM. Specifically, OPN expression at the endosteal bone surface functions to recruit hematopoietic progenitors to bone where they are induced to become quiescent and maintain long term repopulating potential. We hypothesized that a similar relationship exists between leukemia and OPN resulting in a quiescent population of chemoresistant leukemic blasts at the BM endosteum. Here, we demonstrate that stromal OPN negatively regulates leukemia cell proliferation in the BMM.

A GFP expressing clone of the pre-B ALL cell line Nalm-6 (10 × 10⁶ cells) was engrafted into SCID hosts (6-8 weeks old) via tail vein injection. In vivo imaging was accomplished in live anesthetized mice by reflecting the scalp and imaging the calvarial marrow compartment using real-time multi-photon confocal microscopy. OPN expression in the malignant marrow was imaged by injecting engrafted mice with fluorescently conjugated anti-OPN antibodies 18hrs prior to imaging. For OPN neutralizing experiments, engrafted mice were injected with a cocktail of anti-mouse and anti-human OPN antibodies at a dose of 3 mg/kg.

Using PCR, Western blot and ELISA assays, we show that the ALL cell line Nalm-6 expresses OPN and secretes large quantities of OPN into conditioned media. Flow cytometric analysis demonstrates that Nalm-6 also express the cell surface OPN receptors VLA-4 and VLA-5. Furthermore, Nalm-6 cells specifically adhere to OPN in vitro via specific engagement of these integrin receptors. In vivo imaging demonstrates that OPN expression in the BM increases as leukemia progresses and that OPN is highly expressed adjacent to areas of high tumor burden. Specifically, a significant amount of OPN is detected in bony tunnels surrounding the vasculature at the invading tumor front. Using Q-PCR and western analysis, we demonstrate that both host-derived and leukemia-derived OPN are upregulated in malignant BM. In vivo inhibition of the OPN signaling axis in the Nalm-6 xenograft model using neutralizing antibodies directed at both human and murine OPN increased overall tumor burden two-fold as measured by flow cytometry and in vivo imaging (p=0.02, N=7) while simultaneously increasing the Ki-67 positive proliferative tumor population (p=0.029, N=4). Furthermore, IHC analysis of a panel of diagnostic bone marrow biopsies from a diverse cohort of ALL patients demonstrated high OPN expression in the marrow of these patients.

Leukemic blasts that have hijacked normal stromal interactions to become quiescent may represent a major source of MRD and patient relapse in ALL. Our data demonstrate that the interaction of leukemic blasts with OPN in the stromal microenvironment reduces the number of cycling blasts and constrains tumor proliferation within the marrow. Current investigations are aimed at combining OPN neutralization with an in vivo model of MRD to determine whether OPN neutralization induces cycling of quiescent blasts, ultimately rendering them sensitive to chemotherapy. The ultimate goal of this work is the development of clinically relevant therapies designed to render leukemic cells more susceptible to chemotherapy by disengaging them from protective interactions with the BM microenvironment.

No relevant conflicts of interest to declare.