5-Azacytidine depletes HSCs and synergizes with an anti-CD117 antibody to augment donor engraftment in immunocompetent mice

Abstract

Depletion of hematopoietic stem cells (HSCs) is used therapeutically in many malignant and nonmalignant blood disorders in the setting of a hematopoietic cell transplantation (HCT) to eradicate diseased HSCs, thus allowing donor HSCs to engraft. Current treatments to eliminate HSCs rely on modalities that cause DNA strand breakage (ie, alkylators, radiation) resulting in multiple short-term and long-term toxicities and sometimes even death. These risks have severely limited the use of HCT to patients with few to no comorbidities and excluded many others with diseases that could be cured with an HCT. 5-Azacytidine (AZA) is a widely used hypomethylating agent that is thought to preferentially target leukemic cells in myeloid malignancies. Here, we reveal a previously unknown effect of AZA on HSCs. We show that AZA induces early HSC proliferation in vivo and exerts a direct cytotoxic effect on proliferating HSCs in vitro. When used to pretreat recipient mice for transplantation, AZA permitted low-level donor HSC engraftment. Moreover, by combining AZA with a monoclonal antibody (mAb) targeting CD117 (c-Kit) (a molecule expressed on HSCs), more robust HSC depletion and substantially higher levels of multilineage donor cell engraftment were achieved in immunocompetent mice. The enhanced effectiveness of this combined regimen correlated with increased apoptotic cell death in hematopoietic stem and progenitor cells. Together, these findings highlight a previously unknown therapeutic mechanism for AZA which may broaden its use in clinical practice. Moreover, the synergy we show between AZA and anti-CD117 mAb is a novel strategy to eradicate abnormal HSCs that can be rapidly tested in the clinical setting.

Graphical abstract

Figure 1.

AZA depletes HSPCs in vivo. (A) Treatment schema. C57BL/6 mice were injected intraperitoneally with AZA at 5 mg/kg per day for 5 consecutive days. PB, spleens, and BM from both legs and spine were analyzed on days 6, 10, 15, and 20 after the start of AZA treatment. (B) Total live cells in the BM of untreated mice and AZA-treated mice on days 6, 10, 15, and 20 after the start of AZA treatment. (C) Hematoxylin and eosin staining of a BM section from 1 mouse femur on day 6 after treatment with AZA 5 mg/kg per day for 5 days compared with an untreated control. (D) PB cell counts for untreated mice compared with AZA-treated mice on days 6, 10, and 20 after the start of AZA treatment. (E) Representative flow cytometry contour plots of the HSPC compartment in the BM of untreated and AZA-treated mice on days 6, 10, 15, and 20. The figure shows our gating strategy beginning with Lin^– live cells, LSK cells, MPPs, LT-HSCs, and ST-HSCs. (F) Absolute cell counts from the different HSPC compartments on days 6, 10, 15, and 20 after AZA treatment compared with untreated control mice. LSK: Lin^–Sca1⁺c-Kit⁺; LT-HSC: LSKCD150⁺CD48^–; ST-HSC: LSKCD150^–CD48^–; MPP2: LSKCD150⁺CD48⁺; MPP3: LSKCD150^–CD48⁺Flt3^–; MPP4: LSKCD150^–CD48⁺Flt3⁺. Data are expressed as mean ± standard deviation (SD); n = 4 mice per group per time point. *P < .05. Hb, hemoglobin; WBC, white blood cell.

Figure 2.

AZA induces proliferation of HSCs in vivo and reduces growth and viability of proliferating HSCs in vitro. (A) Treatment schema. C57BL/6 mice were injected intraperitoneally with AZA at 2.5 or 5 mg/kg per day for 3 days. After the last AZA dose, BrdU was administrated intraperitoneally at 1 mg per mouse once every 6 hours for a total of 24 hours. Mice were euthanized on day 4 after the start of experiment, and BM from both legs was analyzed by flow cytometry for incorporation of BrdU. (B) Representative histogram of BrdU expression in LT-HSCs (gated from live Lin^–Sca1⁺c-Kit⁺CD150⁺CD48^– cells) from 1 untreated and 1 AZA-treated mouse BM sample. (C) BrdU expression in percent for LT-HSCs (n = 4 mice per group). (D) Schematic of in vitro culture protocol. FACS-sorted mouse Lin^–Sca1⁺c-Kit⁺CD48^– or human Lin^–CD34⁺CD38^– cells were plated in 96-well flat-bottom plates coated with fibronectin that contained 100 µL of HSC media per well, and were supplemented with 10 ng/mL SCF and 100 ng/mL thrombopoietin (TPO). Different concentrations of AZA (0.1, 0.5, and 1 µg/mL) were added on 2 consecutive days (at baseline and at 24 hours after cell culturing). Cell counting and cell imaging was performed once every 6 hours for a total of 48 hours on an ImagExpress Pico automated cell counting system. After 48 hours, the percentage of live and dead cells was assessed by using EarlyTox Live/Dead Assay Kit. (E) Proliferation curves of FACS-sorted mouse Lin^–Sca1⁺c-Kit⁺CD48^– cells in the presence of indicated concentrations of AZA (left). Cell viability was assessed by the percentage of calcein AM⁺ cells after 48 hours of cell culture (right) (data are from 3 independent experiments). (F) Proliferation curves of FACS-sorted human Lin^–CD34⁺CD38^– cells in the presence of indicated concentrations of AZA (left). Cell viability was assessed by the percentage of calcein AM⁺ cells after 48 hours of cell culture (right) (data are from 2 independent experiments). Data are expressed as mean ± SD. *P < .05; **P < .01; ***P < .001.

Figure 3.

Anti-CD117 mAb (ACK2) combined with AZA enhances HSC depletion and delays HSC recovery in vivo. (A) Treatment schema for panels B-F. C57BL/6 mice were injected intravenously with a single dose of ACK2 or 2B8 at 500 µg 5 days before the start of treatment with AZA. AZA was administered intraperitoneally at 5 mg/kg once per day for 5 days. Mice were euthanized at 6, 10, or 20 days after the first dose of AZA; spleens and BM from both legs and spine were harvested and analyzed by flow cytometry for HSC depletion and depletion of mature myeloid and lymphoid cells. (B) Hematoxylin and eosin staining of a BM section of 1 femur at day 6 from a mouse treated with ACK2 only, AZA only, or ACK2-AZA. (C) Absolute cell counts of the different HSPC compartments in the BM of untreated controls and ACK2-AZA–treated mice on days 6, 10, and 20 after the start of AZA treatment as measured by flow cytometry. (D) Comparison of the absolute cell counts of different HSPC compartments in the BM of mice treated with AZA only or ACK2-AZA at baseline and at 6, 10, and 20 days after start of AZA. (E) Comparison of the absolute cell counts of LT-HSCs in the BM of mice treated with 2B8-AZA vs ACK2-AZA at baseline and on days 10 and 20 after the start of AZA treatment. (F) Comparison of the absolute cell counts of ST-HSCs in the BM of mice treated with 2B8-AZA vs ACK2-AZA at baseline and on days 10 and 20 after the start of AZA treatment. For panels C-F, data for each experimental group were pooled from 2 independent experiments; n = 8-9 mice per group per time point. (G) Treatment schema for panels H-I. C57BL/6 mice were injected intravenously with a single dose of ACK2 500 µg 5 days before the start of treatment with AZA (administered intraperitoneally at a dose of 2.5 mg/kg per day for 3 days). Mice were euthanized 4 days after the first dose of AZA, and BM from both legs was harvested and analyzed by flow cytometry for annexin V and propidium iodide (PI) staining. (H) Representative flow cytometry plots of annexin V and PI staining (gated from Lin^–Sca1⁺c-Kit⁺) in untreated controls and mice treated with ACK2, AZA 2.5 mg/kg, or ACK2-AZA. (I) Top: annexin V⁺; bottom: annexin V⁺/PI⁺; stained cells are shown as a percentage of LSK cells in the different treatment groups compared with untreated controls (n = 3-5 mice). Data are expressed as mean ± SD. *P < .05; **P < .01; ***P < .001; ****P < .0001. ns, nonsignificant.

Figure 4.

ACK2 synergizes with AZA and permits engraftment of congenic HSCs in immunocompetent mice. (A) Schematic of congenic transplantations. C57BL/6 (H-2^b, Thy1.1, CD45.1) mice were donors for C57BL/6 (H-2^b, Thy1.1, CD45.1/CD45.2) recipients. ACK2 at 500 µg was injected intravenously 10 days before cell infusion, and AZA was administrated on days −5 through −1 at a dose of 2.5 or 5 mg/kg per day. Recipients received 20 × 10⁶ whole BM cells (WBM) or 5 × 10⁴ FACS-sorted LSK cells on day 0 via retro-orbital injection. PB chimerism was assessed by flow cytometry using the CD45 allelic marker to distinguish between donor and recipient live total cells, myeloid cells (Gr1⁺Mac1⁺), B cells (CD19⁺CD3^–), and T cells (CD19^–CD3⁺). (B) Higher levels of sustained multilineage donor engraftment were observed in the ACK2-AZA group compared with single-agent ACK2 or single-agent AZA groups after transplantation of 20 × 10⁶ WBM cells (n = 3-6 mice per group). (C) ACK2-AZA at 5 mg/kg per day enables sustained multilineage engraftment of 5 × 10⁴ FACS donor congenic LSK cells (n = 6-7 mice per group). Data were pooled from 2 independent experiments and represent mean ± SD.

Figure 5.

ACK2 synergizes with AZA and permits engraftment of allogeneic HSCs in immunocompetent mice. (A) Schematic of allogeneic transplantations. BALB.B mice (H-2^b, Thy1.2, CD45.2) were recipients for B6 donors (H-2^b, Thy1.1, CD45.1). ACK2 was injected intravenously 10 days before transplantation, AZA was administered intraperitoneally on days −5 through −1 at 5 mg/kg per day, and anti-CD4/anti-CD8 mAbs were injected intravenously at 100 µg for each mAb on days −2, −1, 0 (day of transplant). Recipients received 5 × 10⁴ LSK cells at day 0 via retro-orbital injection. PB chimerism analysis was assessed by flow cytometry using a CD45 marker to distinguish between donor and recipient live total, myeloid (Gr1⁺Mac1⁺), B cells (CD19⁺CD3^–), and T cells (CD19^–CD3⁺). (B) Multilineage donor-derived chimerism in PB at 4, 14, and 26 weeks after conditioning with single-agent ACK2, single-agent AZA 5 mg/kg per day, or ACK2-AZA and transplantation of 5 × 10⁴ LSK cells. Data represent mean ± SD (n = 3-5 mice per group).