CSF1 Receptor Inhibition Augments Murine Immune Bone Marrow Failure: New Evidence Indicative of M1-Type Macrophage Polarization Driving Hematopoietic Failure

Background: Immune aplastic anemia (AA) is a disease of bone marrow failure (BMF) primarily driven by cytotoxic T-lymphocyte-mediated destruction of hematopoietic stem and progenitor cells, although other cellular pathways have been implicated in the disease. The depletion of macrophages in a murine BMF model ameliorated bone marrow (BM) destruction and alleviated peripheral pancytopenia, suggestive of the importance of macrophages and related cells in the pathogenesis of AA. In the current study, we utilized Pexidartinib (PLX, Selleck Chemicals) to further test the role of macrophages and other related cells in a well-established AA mouse model. PLX is a CSF1R inhibitor that depletes circulating non-classical monocytes and reduces tissue macrophage infiltration.

Methods: BMF was induced by injection of 5 × 10^6/mouse lymph node cells from C57BL/6 (B6) donors into heterozygous major histocompatibility complex mismatched CByB6F1 recipients pre-irradiated with 5 Gy total body irradiation. BMF mice were either left untreated, treated with PLX at 50 mg/kg three times per week for two weeks, or treated with JAK inhibitor baricitinib (BAR, LC Laboratories) at 25 mg/kg five times per week for two weeks. Both PLX and BAR were administered through oral gavage. Animals were either bled and euthanized at day 14 to measure cellular composition in the peripheral blood and BM, or monitored for survival.

Results: PLX treatment unexpectedly augmented BMF development in CByB6F1 mice, causing further declines in blood neutrophils (0.06 ± 0.02 vs 0.1 ± 0.02, 10⁹/L, P < .05), red blood cells (4.4 ± 0.4 vs 6.4 ± 0.4, 10¹²/L, P < .0001), hemoglobin (7.1 ± 0.7 vs 10.1 ± 0.7, g/dL, P < .0001), and platelets (32.4 ± 9.6 vs 73.1 ± 9.6, 10⁹/L, P < .001) when compared to untreated BMF mice which already exhibited significant peripheral blood pancytopenia. The acceleration of blood pancytopenia by PLX was accompanied by increased CD4+ (16.1 ± 1.0 vs 12.1 ± 1.0, %, P < .001) and a trend towards increased CD8+ (55.1 ± 2.3 vs 50.8 ± 2.3, %, P = .07) proportions, as well as increased Fas expression (61.6 ± 4.4 vs 45.4 ± 4.4, %, P < .001) and apoptosis (72.5 ± 4.1 vs 62.6 ± 4.1, %, P < .001) in non-T cell leukocytes. In the BM, PLX treatment caused further decline in total BM cells (87.7 ± 14.8 vs 122.3 ± 14.8, 10⁶, P < .05) relative to BMF mice without treatment, along with decreases in the proportions of CD11b myeloid cells (5.8 ± 4.5 vs 19.4 ± 4.5, %, P < .05) and granulocytic-myeloid derived suppressor cells (G-MDSC, 3.7 ± 4.3 vs 16.6 ± 4.3, %, P < .05). Survival was significantly worse in the PLX-treated mice (P < .05).

Additional experimentation was performed to better characterize the macrophage and MDSC populations. While there was no change in the overall number of macrophages, the PLX treatment group demonstrated a polarization towards the pro-inflammatory M1-type macrophage in comparison to the BM of untreated BMF mice (8.0 ± 1.0 vs 5.0 ± 1.0, %, P < .05). Because BAR attenuated BMF mice efficiently, we wanted to compare different effects on this cell component between BAR and PLX in order to explore the underlying mechanisms responsible for worsening the disease by PLX. Consistent with previously presented data, baricitinib treatment ameliorated BMF in CByB6F1 mice. In comparison to the BAR-treated population the overall macrophage proportion in the BM of PLX-treated mice was greater (13.1 ± 1.7 vs 4.0 ± 1.6, %, P < .0001), notably the M1-type macrophage (8.0 ± 1.0 vs 0.02 ± 0.9, %, P < .0001). PLX-treated mice exhibited decreased CD11b myeloid cells (2.0 ± 3.6 vs 31.9 ± 3.3, %, P < .0001), G-MDSC (0.02 ± 3.3 vs 24.2 ± 3.0, %, P < .0001), and monocytic-MDSC (0.003 ± 0.3 vs 2.0 ± 0.3, %, P < .0001).

Conclusions: Inhibition of CSF1R with PLX augmented murine BMF with increased T cell infiltration and enhanced leukocyte apoptosis, probably due to decreased G-MDSC and CD11b myeloid precursors. This finding is consistent with published results showing that G-MDSC are immunosuppressive and could be used as a potential therapeutic in treating murine BMF. The polarization towards the pro-inflammatory M1-type macrophage in PLX-treated mice supports the idea that this type of macrophage primes cytotoxic CD8+ T cells, whereas polarization towards the anti-inflammatory M2-type macrophage in BAR-treated mice is consistent with suppression of CD8+ T cell infiltration. More work is underway to better characterize the relationship between and function of macrophages and MDSCs in aplastic anemia.