Splenic CD24^low Red Pulp Macrophages Provide an Alternate Niche for Chronic Myeloid Leukemia Stem Cells

Chronic myeloid leukemia (CML) is a typical stem-cell driven malignancy, driven by leukemia stem cells (LSCs). LSCs are resistant to conventional therapies. This resistance is mediated by cell-intrinsic mechanisms and interactions with the microenvironment. LSCs depend on signals from a specialized microenvironment, a so called niche, to maintain their stem cell characteristics. In CML the bone marrow (BM) as a niche is well-investigated and several therapeutic targets, which aim at LSCs by interrupting their interaction with the BM-niche are under investigation. However, even though splenomegaly is a hallmark of CML the contribution of the splenic microenvironment to CML development has not been studied so far. This project aims to investigate the role of the splenic microenvironment as an independent secondary LSC niche and its contribution to disease development.

To induce a CML-like disease in mice we retrovirally transduced FACS-sorted Lineage^- Sca-1⁺ cKit⁺ BM cells with pMSCV-p210BCR/ABL-IRES-GFP and injected the transduced cells into non-irradiated mice.

To find out if the spleen contributes to disease development we induced CML in splenectomized and sham operated mice. Splenectomized mice survived significantly longer compared to sham operated controls (median survival 31 vs. 22 days; p=0.0006) with 20% of the splenectomized mice surviving longer than 90 days. Moreover, the number of LSCs in the BM of splenectomized mice was reduced 3.7-fold (p=0.002).

Flowcytometric analysis of the spleen and BM compartments of CML bearing mice revealed that the majority of the leukemic stem and progenitor cells (LSPCs) were located in the spleen (19-fold more LSCs in the spleen; p =0.007). Moreover we found the leukemic compartment in the spleen to be enriched for LSPCs compared to the BM (20 % spleen vs. 10 % BM; % LSPCs of total leukemic cells; p=0.01). To confirm this phenotypic observation functionally we performed a limiting dilution transplantation of leukemic cells from spleen and BM. In line with the phenotypic observation we found a higher frequency of LSCs in the spleen compared to the BM (1/41'703 vs. 1/432'594; p=0.02). We next analyzed the gene expression of LSPCs from spleen and BM. We found that the gene expression profile of splenic LSPCs showed higher expression of stemness-related genes and reduced expression of myeloid differentiation genes compared to BM LSPCs, indicating that the spleen is more supportive of primitive LSPCs.

Knowing that the spleen contributes to disease development by providing an alternate niche for LSCs we next analyzed the spleens using confocal microscopy. We found that the LSCs resided exclusively in the red pulp. Previous studies have shown that HSCs reside in direct contact with red pulp macrophages (RPMs) during extramedullary hematopoiesis (Dutta et al., JEM, 2015). In addition we found that in spleens from human CML patients CD34+ leukemia cells localized together with macrophages (p=0.001). Furthermore we could show that RPMs are capable of producing both SCF and G-CSF. To test the role of RPMs as a potential niche component in vitro we co-incubated LSCs and RPMs overnight before plating the LSCs in a colony formation assay. We found that the co-incubation with RPMs improved the colony formation capacity of LSCs (CFUs 166 vs. 138; p=0.0356). To test the role of RPMs in vivo we depleted macrophages in CML mice using clodronate liposomes. This resulted in significantly reduced splenomegaly (867mg vs. 249mg; p<0.0001) and reduced numbers of splenic LSCs (23-fold; p=0.001). This was further confirmed in a genetically engineered mouse model lacking RPMs (Spic^-/-) (735mg vs. 450mg; p=0.0112 and 22-fold fewer LSCs; p<0.0001). Finally, while in naïve mice RPMs can be differentiated into a CD24^low and CD24^high population, the CD24^high population is lost in CML bearing mice.

In summary, we found that the spleen provides an alternate niche for LSCs, thereby contributing to CML development. Compared to the BM niche the splenic niche is more supportive of primitive LSPCs, as shown by the higher frequency of LSCSs found in the spleen and the higher expression of stemness related genes in splenic LSCs compared to BM LSCs. Moreover we identified CD24^low RPMs as a unique and central component of the splenic LSC niche. Even though we could show that RPMs are capable of producing SCF and G-CSF the exact mechanisms by which RPMs support LSCs remains to be investigated.