Hydroxyurea synergizes with valproic acid in wild-type p53 acute myeloid leukaemia

Abstract

Palliative care in acute myeloid leukaemia (AML) is inadequate. For elderly patients, unfit for intensive chemotherapy, median survival is 2-3 months. As such, there is urgent demand for low-toxic palliative alternatives. We have repositioned two commonly administered anti-leukaemia drugs, valproic acid (VPA) and hydroxyurea (HU), as a combination therapy in AML. The anti-leukemic effect of VPA and HU was assessed in multiple AML cell lines confirming the superior anti-leukemic effect of combination therapy. Mechanistic studies revealed that VPA amplified the ability of HU to slow S-phase progression and this correlated with significantly increased DNA damage. VPA was also shown to reduce expression of the DNA repair protein, Rad51. Interestingly, the tumour suppressor protein p53 was revealed to mitigate cell cycle recovery following combination induced arrest. The efficacy of combination therapy was validated in vivo. Combination treatment increased survival in OCI-AML3 and patient-derived xenograft mouse models of AML. Therapy response was confirmed by optical imaging with multiplexed near-infrared labelled antibodies. The combination of HU and VPA indicates significant potential in preclinical models of AML. Both compounds are widely available and well tolerated. We believe that repositioning this combination could significantly enhance the palliative care of patients unsuited to intensive chemotherapy.

Keywords: AML; DNA damage; hydroxyurea; p53; valproic acid.

Conflict of interest statement

CONFLICTS OF INTEREST

The authors declare no conflicts of interest.

Figures

Figure 1. Assessment of cell death induction and the enhanced potential of combining HU and VPA in AML cell lines

(A-D) Hoechst nuclear staining assay performed to generate dose response curves for cell death induction. HL-60, MOLM-13, OCI-AML3 and MV4-11 cells were treated with HU (25–200 μM) and VPA (0.25–2 mM) alone or in combination at a fixed ratio (1:10) for 72 hrs. N = 3. (E-G) Hoechst nuclear staining assay is performed to determine the % of viable cells in MOLM-13, OCI-AML3, MV4-11 (p53 wild-type), KG1-A, THP-1, K562 (p53 mutated) and HL-60 cell lines. All cell lines were treated with HU (60 μM), VPA (0.6 mM) and the combination. Results are pooled and presented as mutated/null p53 (Mut/null p53) vs. wild-type p53 cell lines (WT p53). A significant increase in sensitivity was observed in WT p53 cell lines under combination treatment (**P < 0.01) N = 3.

Figure 2. Investigating the role of p53 in HU and VPA combination therapy

(A) Lysate was produced from untreated MOLM-13 wt and shp53 cells. Immunoblotting was performed with antibodies targeting p53 and actin. N = 3. (B) + (C) MOLM-13 wt and shp53 cells were treated with HU (75 μM and 100 μM), VPA (0.75 mM and 1 mM) and the combination of both for 72 hrs and apoptosis was determined by Annexin-PI to generate dose response curves. Induction of apoptosis was compared between each cell line at particular treatment conditions (*P < 0.05, **P < 0.01). N = 3. (D–G) MOLM-13 wt and shp53 cells and (H-K) HL-60 and OCI-AML3 cells were treated with HU (100 μM) and VPA (1 mM), the combination or seeded without treatment. Following 72 hrs cells were washed 2 × in sterile saline and reseeded in wells. The cells were followed for a further 3 days and the number of viable cells was determined at 24 hrs intervals throughout the entire experimental course. Cell counts were performed using the Countess™ Automated Cell Counter (Invitrogen). N = 3.

Figure 3. Complimentary regulation of cell cycle status following combination therapy in OCI-AML3 cells

(A) OCI- AML3 cells were treated with HU (100 μM), VPA (1 mM) and the combination of both for 24 hrs before cells were stained with PI and cell cycle status analysed by flow cytometry. Analyses was performed in 3 independent experiments with Figure (A) providing representative plots. (B) Comparison of OCI-AML3 cell populations defined as S-phase or Sub G1 (necrotic) at 24 hrs, 48 hrs and 72 hrs combination treatment with HU (100 μM) and VPA (1 mM). N = 3. (C) OCI-AML3 cells were incubated for 24 hrs with doses of HU (100 μM and 150 μM) and VPA (1 mM and 1.5 mM). Immunoblotting was with antibodies towards p21 and actin. (D) OCI-AML3 cells were treated with HU (100 μM), VPA (1 mM) and the combination of both for 72 hrs. Combination experiments were performed where monotherapies where added for the first 24 hrs with the secondary treatment added for the final 48 hrs of the experiment (Pretreatment with HU/VPA). Apoptosis was determined by Annexin-PI. N = 3. The two alternate sequence studies proved to significantly affect the compounds capacity to induce apoptosis (**P < 0.01).

Figure 4. Mechanistic studies demonstrate the combinations capacity to regulate DNA damage repair proteins

(A) OCI-AML3 cells were treated with HU (100 μM), VPA (1 mM) and the combination of both for 24 and 48 hrs. Immunoblotting was performed using antibodies towards γH2AX, p53, Chk1, Rad51 and actin. N = 3. (B) OCI-AML3 cells were treated with HU (100 μM), VPA (1 mM) and the combination of both for 24 hrs. Assesment of H2AX and γH2AX expression was performed using flow cytometry with the Muse™ H2A.X Activation Dual Detection Kit. N = 3. (C + D) OCI-AML3 cells were treated for 48 hrs with HU (100 μM), VPA (1 mM) or in combination, before cytospun, fixed and immunostained for Rad51. Representative image of three independent experiments is shown.

Figure 5. Annexin-PI studies demonstrate synergy in OCI-AML3, primary AML blasts and the non-toxic effect of the combination in PBMCs

(A) OCI-AML3 cells were treated with increasing doses of HU (0–200 μM) and VPA (0–2 mM) alone or in combination at a fixed ratio (1:10) for 72 hrs and apoptosis was determined by Annexin-PI to generate dose response curves. N = 3. (B) The data generated in figure (A) enabled CI vales to be plotted at fa (0–1.0). (C) Differences in means of viability between HU (75 μM), VPA (0.75 mM) and combination of both for 24 hrs for the pooled patient data analyzed by Annexin-PI. Results are given as means ± s.e. of mean (**P < 0.01, n = 10). (D) Bliss Independence analysis of expected and actual response for the combinational therapy of HU (75 μM) and VPA (0.75 mM) for each of the individual AML patient samples analyzed by Annexin-PI. (E) Peripheral blood mononucleocytes (PBMCs) obtained from four healthy donors were treated with HU (100 μM), VPA (1 mM) and the combination of both for 72 hrs and apoptosis was determined by Annexin-PI. Reduction of viability by the combinational treatment was compared with untreated control cells to reveal a non-significant (NS).

Figure 6. Combinational therapy of HU and VPA represses AML in vivo

(A) Survival data presented in Kaplan-Meyer curve illustrating the efficacy of HU and VPA and increased survival of combination therapy (log-rank P = 0.0003 vs controls, P = 0.0014 vs HU, P = 0.0001 vs VPA) in the OCI-AML3 orthotopic model of AML. Arrows indicate days on which animals were dosed with both compounds. Control (n = 8), HU (n = 8), VPA (n = 8) and combination (n = 7). (B) Primary AML cells from the same generation used to generate the PDX model were thawed and treated with HU (75 μM), VPA (0,75 mM) and the combination of both for 24 h and apoptosis determined by Annexin-PI staining. (C + D) Imaging with multiplexed mAbs was performed before initiation of therapy (day 21) and 1 week later (day 28). Control (n = 8), HU (n = 8), VPA (n = 8) and combination (n = 6). Representative images of mice after treatment reveal a significant reduction in total fluorescence (photon count per second) of combination treated mice compared with vehicle controls and monotherapies. Combination compared with control (p < 0.0001), VPA (p = 0.0002) and HU (p = 0.0005). (E) Schematic overview of the combination mechanism of action and key molecular events determining cell death or survival in AML cells exposed to HU and VPA. HU, Hydroxyurea; VPA, Valproic Acid; DSB, Double strand breaks; HR, homologous recombination.