Inclusion and ethics statement

This study has been a collaborative project between the Institute of Cancer Research, Barts Cancer Institute and the University of Oxford, with other significant contributions made across the UK and Europe. All experiments were conducted following extensive training and, where appropriate, following approval by ethical bodies.

Mice

All experiments on animals were performed under UK Home Office authorization under the project license PP4153210 at Barts Cancer Institute following approval by Queen Mary University of London AWERB. Animals were subject to an optimum dark–light cycle, with ambient temperature and humidity. All mice were on the C57BL/6 genetic background. Phd2fl/fl, shPhd2 and iMLL-AF9 mice were described previously28,29,30. Vav-iCre and NBSGW mice were purchased from The Jackson Laboratory63,64. All transgenic and knockout mice were CD45.2+. Congenic recipient mice were CD45.1+/CD45.2+. Mice used for support BM cells during transplantation experiments were CD45.1+. Unless stated otherwise, all in vivo experiments had mixed-sex animals aged between 8–12 weeks when experiments began. All animals were monitored twice daily and humanely killed at the experimental end point, which was not exceeded during this study. For animals analyzed in steady-state hematopoiesis, HSC or BM transplantation, or drug toxicity studies, the end point is as described in the experimental design stated in the figure legends for each experiment. For animals employed in AML studies, the humane end point was determined either by the experimental design or in the case of survival studies, by disease progression. Disease progression was measured against a clinical scoring system detailed in the UK Home Office project license. The parameters studied include animal weight, appearance, body condition, clinical signs and natural and provoked behavior.

Human tissue and ethical approvals

All use of human tissue was in compliance with the ethical and legal framework of the UK’s Human Tissue Act, 2004. Primary human AML samples were from Barts Cancer Institute Biobank (with approval of the Research Ethics Committee). Their use was authorized following ethical review by the Tissue Biobank’s scientific sub-committee and with the informed consent of the donors. For all samples used in this study, Barts Cancer Institute Biobank obtained informed consent from all participants.

Flow cytometry

All BM, fetal liver (FL) and splenic cells were prepared and analyzed as described6,47,58,65,66,67,68. BM cells were isolated by crushing tibias and femurs using a pestle and mortar. FL cells were prepared by mashing the tissue and passing through a 70-µm strainer. Single cell suspensions from BM, FL or PB were incubated with Fc block and then stained with antibodies. For HSC and progenitor cell analyses, following incubation with Fc block, unfractionated BM cells were stained with lineage markers containing biotin-conjugated anti-CD4, anti-CD5, anti-CD8a, anti-CD11b, anti-B220, anti-Gr-1 and anti-Ter119 antibodies together with BV711-conjugated anti-c-Kit, APC-Cy7-conjugated anti-Sca-1, PE-conjugated anti-CD48 and PE-Cy7-conjugated anti-CD150 antibodies. Biotin-conjugated antibodies were then stained with PB-conjugated streptavidin. For analyses of differentiated cells, following incubation with Fc block, spleen or BM cell suspensions were stained with PerCP-conjugated anti-B220 and APC-Cy7-conjugated anti-CD19 antibodies for B cells; APC-conjugated anti-CD11b and PE-Cy7-conjugated anti-Gr-1 for myeloid cells; PE-conjugated anti-CD4 and anti-CD8 antibodies for T cells.

To distinguish CD45.2+-donor-derived cells in PB or BM of transplanted mice, BV711-conjugated anti-CD45.1 and Pacific Blue-conjugated anti-CD45.2 antibodies were used. For HSC and progenitor staining in transplanted mice, APC-conjugated anti-c-Kit and PerCP-conjugated streptavidin was used; the remainder of the staining was as described above. For analyses of differentiated cells in PB and BM of transplanted mice, myeloid cells and lymphoid cells were stained as described above. TO-PRO-3 or 4,6-diamidino-2-phenylindole (DAPI) were used for dead cell exclusion.

To assess human AML burden, cells were stained with anti-human FITC-conjugated anti-CD45, APC-conjugated anti-CD33 and anti-human PE-conjugated CD14.

Flow cytometry analyses were performed using a LSRFortessa (BD). Cell sorting was performed on a FACSAria Fusion (BD). Representative flow cytometry gating strategies are available in Supplementary Figs. 10–17.

Leukemic transformation

Leukemic Meis1/Hoxa9, FLT3-ITD and PML-RARα cells were prepared as described68,69. Transduced cells were subjected to three rounds of colony-forming cell (CFC) assays in MethoCult M3231 (STEMCELL Technologies) supplemented with 20 ng ml−1 SCF, 10 ng ml−1 IL-3, 10 ng ml−1 IL-6 and 10 ng ml−1 granulocyte–macrophage colony-stimulating factor. Colonies were counted 5–7 days after plating and 2,500 cells were re-plated.

Syngeneic transplantation assays

CD45.1+/CD45.2+ recipient mice were lethally irradiated using a split dose of either; a 11 Gy (two doses of 5.5 Gy administered at least 4 h apart) at an average rate of 0.58 Gy min−1 using a Cesium 137 GammaCell 40 irradiator or 8 Gy (two doses of 4 Gy administered at least 4 h apart) at an average rate of 1.086 Gy min−1 using a RADSOURCE X-ray irradiator.

For primary transplantations of leukemic cells, 100,000 Meis1/Hoxa9-transduced c-Kit+ cells or 2,000 iMLL-AF9 LSK cells were transplanted into lethally irradiated CD45.1+/CD45.2+ recipient mice (together with 200,000 unfractionated support CD45.1+ wild-type BM cells). For secondary transplantations of leukemic cells, 50,000 cells collected from primary recipients were transplanted into lethally irradiated CD45.1+/CD45.2+ recipient mice (together with 200,000 unfractionated support CD45.1+ wild-type BM cells). Recipients were culled upon reaching their humane end point as recorded in survival curves.

For LDA analyses, increasing doses (10,000, 50,000 and 100,000) of CD45.2+ BM cells from primary transplantation were re-transplanted into lethally irradiated CD45.1+/CD45.2+ recipient mice (together with 200,000 unfractionated support CD45.1+ wild-type BM cells). LSC frequency was calculated using ELDA software70.

For primary transplantations of healthy total BM cells, 5,00,000 total BM cells were mixed with 5,00,000 support CD45.1+ BM cells. For primary transplantations of HSCs, 200 LSKCD48-CD150+ HSCs (per recipient) sorted from the BM of donor mice were mixed with 200,000 support CD45.1+ BM cells and transferred into lethally irradiated CD45.1+/CD45.2+ recipients. All recipient mice were killed and analyzed 16–20 weeks post-transplantation.

For animals treated with DOX, mice were provided ad libitum access to drinking water containing 2 mg ml−1 DOX with 30% sucrose.

In vivo treatment with PHD inhibitors

For AML experiments, 100,000 of THP-1, OCI-AML3 or MV411 cells were injected via tail vein into non-irradiated 10–12-week-old mixed-sex NBSGW mice and began drug treatment 14 days after transplantation. Mice were injected intraperitoneally (i.p.) twice daily for 21 days with 30 mg kg−1 Dap (an optimal in vivo concentration14,42), 30 mg kg−1 IOX5 or vehicle control. Mice in combination treatment experiments with venetoclax (ABT-199) (MCE) were dosed once daily with 100 mg kg−1 via oral gavage (o.g.).

For steady-state analyses, 8–10-week-old mixed-sex C57BL6 mice were injected i.p. twice daily for 14 days with 30 mg kg−1 IOX5 or vehicle control. Mice were bled before and after treatment and killed 12 h after the final dosing.

Cell proliferation and cell death analyses

Cells were cultured with 50 μM Dap (an optimal concentration in vitro71,72), 50 μM IOX5, 50 μM molidustat, 50 μM roxadustat (an optimal concentration in vitro73), 100 μM DM-NOFD, 0.01 μM venetoclax (MV411 and MOLM13), 0.1 μM venetoclax (THP-1 and OCI-AML3) or vehicle control. Viable cells were counted by Trypan blue exclusion at the indicated time points. To analyze cells undergoing apoptosis, cells were suspended in binding buffer containing PE-conjugated annexin V or FITC-conjugated annexin V and either PI or DAPI.

Primary human AML patient-derived samples

All patients involved in this study gave informed consent for storage and use of their tissue for research purposes. The study was approved by the Institutional Review Board of Barts Cancer Institute and all work was performed in accordance with the Declaration of Helsinki and the Local Research Ethics Committee requirements. Frozen AML samples were obtained from the Barts Cancer Institute Biobank and quickly thawed at 37 °C. Upon thawing, T cells were depleted from all samples using EasySep Human TCR Alpha/Beta Depletion kit (STEMCELL Technologies, 17847). Enriched samples were plated with concentrations 0.4–1.0 × 106 ml−1 in Myelocult H5100 medium (STEMCELL Technologies, 05150) supplemented with 20 ng ml−1 IL-3, granulocyte–macrophage colony-stimulating factor and TPO (BioLegend) in co-culture with irradiated MS-5 cells and treated with 50 μM Dap, 50 μM IOX5, 1 μM venetoclax or vehicle control for 7 days. After 3 days of treatment, cells were supplemented with fresh medium containing the corresponding agent in volume equivalent to 50% of initial volume. After 7 days, viable cell numbers were counted and viability was assessed using annexin V FITC/PI stain.

Synthesis of PHD inhibitors

Roxadustat (FG-4592) was from Cayman Chemical. Molidustat (BAY 85-3934) was from Selleckchem. Dap was synthesized following a reported procedure as described previously15.

1,3-Dicyclohexylpyrimidine-2,4,6(1H,3H,5H)trione (1a)

A suspension of N,N′-dicyclohexylcarbodiimide (3.5 g, 17.00 mmol) in THF (13 ml) was slowly added to a solution of malonic acid (884.5 mg, 8.50 mmol) in THF (13 ml) at 0 °C. After allowing the reaction mixture to warm up room temperature, it was stirred for 2 h. The resulting suspension was filtered and the solvent was removed in vacuo. Re-crystallization from ethanol gave a white fibrous solid (1.625 g, 5.56 mmol, 65%).

The melting point (m.p.) is 201.3–202.4 °C. 1H nuclear magnetic resonance (NMR) (600 MHz, DMSO-d6): δ (ppm) = 4.45 (tt, J = 12.2 Hz, 3.7 Hz, 2H,), 3.68 (s, 2H), 2.13 (qd, J = 12.6 Hz, 3.6 Hz, 4H), 1.76 (m, 4H), 1.58 (td, J = 16.2 Hz, 7.8 Hz, 6H), 1.25 (qt, J = 13.3 Hz, 3.6 Hz Hz, 4H), 1.08 (m, 2H).13C NMR (151 MHz, DMSO-d6): δ (ppm) = 166.0, 151.5, 53.7, 41.2, 28.7, 26.0 and 25.1. IR (ATR): \(\widetilde}}\) (cm−1) = 2,972, 2,930, 2,890, 1,695, 1,676, 1,412, 1,387, 1,330 and 1,204. ESI-HRMS (m/z): [M − H]− calculated for [C16H23N2O3]−: 291.1714, found: 291.1710. The analytical data are consistent with those previously reported15.

Ethyl (1,3-dicyclohexyl-6-hydroxy-2,4-dioxo-1,2,3,4-tetrahydropyrimidine-5-carbonyl)glycinate (1b)

1a (1.625 g, 5.56 mmol) and N,N-diisopropylethylamine (DIPEA; 1.94 ml, 11.12 mmol) were dissolved in CH2Cl2 (30 ml). Ethyl isocyanatoacetate (0.62 ml, 5.56 mmol) was added dropwise; the solution was then stirred at room temperature for 22 h. The reaction mixture was diluted with CH2Cl2 (15 ml), extracted with 6 M HCl (5 ml), then dried over sodium sulfate. After removing the solvent in vacuo, the solid was re-crystallized from cyclohexane, giving a white powder (2.303 g, 5.46 mmol, 98.3%).

The m.p. is 159.0–159.7 °C. 1H NMR, (600 MHz, CDCl3): δ (ppm) = 10.35 (s, 1H), 4.71 (m, 2H), 4.26 (q, J = 7.1 Hz, 2H), 4.15 (d, J = 5.7 Hz, 2H), 2.34 (tq, J = 15.5 Hz, 6.2 Hz, 4H), 1.83 (t, J = 11.4 Hz, 4H), 1.64 (m, 6H), 1.30 (m, 9H).13C NMR (151 MHz, CDCl3): δ (ppm) = 171.9, 169.3, 168.5, 163.5, 149.8, 81.0, 62.0, 41.6, 29.4, 29.1, 26.6, 26.6, 25.5, 25.4 and 14.3. IR (ATR): \(\widetilde\) (cm−1) = 2,981, 2,887, 1,751, 1,384, 1,252, 1,148 and 1,074. ESI-HRMS (m/z): [M + H]+ calculated for [C21H31N3O6]+: 422.2286, found: 422.2283. The analytical data are consistent with those previously reported in Yeh et al.15.

(1,3-Dicyclohexyl-6-hydroxy-2,4-dioxo-1,2,3,4-tetrahydropyrimidine-5-carbonyl)glycine (Dap)

1b (2.3 g, 5.46 mmol) was suspended in ethanol (40 ml), then 4 M sodium hydroxide solution (5 ml) was added slowly. After 2 h stirring at room temperature, 2 M HCl (12 ml) was added. The resulting precipitate was filtered and was stirred in water (50 ml) at 35 °C for 1 h and again filtered to give Dap (1.504 g, 3.82 mmol, 70.0%) as a white solid.

The m.p. is 221.6–223.0 °C. 1H NMR (600 MHz, DMSO-d6): δ (ppm) = 10.15 (t, J = 5.7 Hz, 1H), 4.63 (tt, J = 12.3 Hz, 3.7 Hz, 2H), 4.03 (d, J = 5.4 Hz, 2H), 2.28 (qd, J = 12.5 Hz, 3.6 Hz, 4H), 1.77 (dt, J = 13.2, 3.4 Hz, 5H), 1.64–1.50 (m, 6H), 1.26 (qt, J = 13.2 Hz, 3.6 Hz, 4H), 1.11 (qt, J = 13.0 Hz, 3.3 Hz, 2H). 13C NMR (151 MHz, DMSO-d6): δ (ppm) = 170.2, 170.1, 149.4 53.0, 41.5, 28.7, 26.1, 25.1. IR (ATR): \(\widetilde\) (cm−1) = 2,981, 2,933, 2,855, 2,665, 1,719, 1,663, 1,589, 1,489, 1,456 and 1,244. ESI-HRMS (m/z): [M + H]+ calculated for [C19H28N3O6]+: 394.1973, found: 394.1971. The analytical data are consistent with those previously reported in Yeh et al.15.

1H and 13C NMR of Dap are given in Supplementary Figs. 3 and 4, respectively.

Synthesis of IOX56-Chloro-4-methoxy-N-nicotinamide (2a)

6-Chloro-4-methoxy-nicotinic acid (250 mg, 1.33 mmol), C-(4-trifluoromethyl)-cyclohexylmethylamine (Fluorochem) (255 mg, 1.59 mmol), propylphosphonic anhydride (T3P, 827 mg, 2.6 mmol) were dissolved in dimethylformamide (DMF; 10 ml); DIPEA (803 µl, 4.67 mmol) was then added. The resultant mixture was stirred overnight at room temperature. EtOAc (20 ml) and H2O (100 ml) were added to the reaction mixture. The organic and aqueous fractions were then separated, and the aqueous layer was extracted with EtOAc (30 ml) twice. The combined organic fractions were washed with brine, then dried with anhydrous Na2SO4. The crude mixture purified using flash-column chromatography using cyclohexane (100 to 50%) and EtOAc (0 to 50%) over ten column volumes to give 2a (402 mg, 1.14 mmol, 86%) as a white solid.

The m.p. is 67–70 °C. 1H NMR (600 MHz, CDCl3): δ 8.99 (s, 1H, H2), 7.41 (bs, 1H, H10), 6.92 (s, 1H, H5), 4.04 (s, 3H, H20), 3.33 (t, J = 6.1 Hz, 2H, H12), 2.01–1.88 (m, 5H, H14, H15, H16, H17, H18), 1.62–1.55 (m, 1H, H13), 1.31 (qd, J = 13.0, 3.2 Hz, 2H, H15′, H17′), 1.07–0.99 (m, 2H, H14′, H18′).13C NMR (151 MHz, CDCl3): δ 164.38 (C9), 163.06 (C6), 155.51 (C4), 153.85 (C2), 130.53–125.00 (q, J = 278.4 Hz, C19), 116.99 (C1), 106.99 (C5), 56.91 (C20), 45.60 (C12), 41.85 (q, J = 26.5 Hz, C16), 37.45 (C13), 29.33 (C14, C18), 24.8 (C15, C17). FT-IR Vmax (film): 3,411 and 1,652 cm−1. HRMS (ESI-TOF) calculated for C15H19O2N235ClF3 [M + H] +: 351.1081, found: 351.1078.

4-Methoxy-6-(1H-pyrazol-1-yl)-N- nicotinamide (2b)

2a (100 mg, 0.277 mmol), pyrazole (23 mg, 0.33 mmol), cesium carbonate (Cs2CO3,180 mg, 0.554 mmol), PdtBuXPhos G3 ((2-di-tert-butylphosphino-2′,4′,6′-triisopropyl-1,1′-biphenyl)-2-(2′-amino-1-1′-biphenyl)) palladium (II) methanesulfonate) (19 mg, 0.027 mmol) were placed under N2, then anhydrous tert-butanol (tBuOH, 2 ml) was added. The resultant mixture was heated for 16 h at 60 °C. The reaction mixture was allowed to cool to room temperature, then filtered through a Celite pad. EtOAc (20 ml) and H2O (100 ml) were added to the mixture and the organic and aqueous fractions were separated. The aqueous layer was extracted twice with EtOAc (30 ml). The organic fractions were combined, washed with brine and dried with anhydrous Na2SO4. The crude mixture was purified using flash-column chromatography using (0% to 100% EtOAc, in cyclohexane) over 20 column volumes to give 2b (49 mg, 0.127 mmol, 46 %).

1H NMR (400 MHz, CDCl3): δ 9.07 (s, 1H, H2), 8.62 (d, J = 2.6 Hz, 1H, H20), 7.75 (d, J = 1.6 Hz, 1H, H22), 7.60 (s, 1H, H5), 7.54 (bs, 1H, H10), 6.49 (dd, J = 2.7, 1.7 Hz, 1H, H21), 4.13 (s, 3H, H23), 3.35 (t, J = 6.4 Hz, 2H, H12), 2.06–1.90 (m, 5H, H14, H15, H16, H17), 1.69–1.57 (m, 1H, H13), 1.43–1.20 (m, 2H, H15′, H17′), 1.05 (qd, J = 12.8, 2.6 Hz, 2H, H14′, H18′). 13C NMR (101 MHz, CDCl3): δ 164.93 (C9), 163.40 (C6), 154.60 (C4), 153.02 (C2), 142.57 (C22), 130.71–125.16 (q, J = 278.2 Hz, C24), 127.76 (C20), 115.25 (C1), 108.25 (C21), 94.38 (C5), 56.47 (C23), 45.22 (C12), 41.68 (q, J = 26.4 Hz, C16), 37.22 (C13), 29.08 (C14, C18), 24.34 (C15, C17). FT-IR Vmax (film): 3,132, 1640, 1,550 cm−1. HRMS (ESI-TOF) calculated for C18H22O2N4F3 [M + H]+: 383.1689, found: 383.1689.

4-Oxo-6-(1H-pyrazol-1-yl)-N--1,4-dihydropyridine-3-carboxamide (IOX5)

2b (30 mg, 0.078 mmol) and lithium chloride (33 mg, 0.78 mmol) were dissolved in N,N-dimethylacetamide (DMAc, 2 ml). The resultant mixture was heated with microwave irradiation at 120 °C for 8 h. The resultant mixture was diluted with H2O (100 ml) and extracted with EtOAc (3 × 20 ml). The combined organic phases were washed with water, brine and dried with anhydrous Na2SO4. The volatiles were then evaporated in vacuo and the crude mixture was purified by flash-column chromatography using (0% to 10% methanol in CH2Cl2) over 15 column volumes to give IOX5 (8.5 mg, 0.023 mmol, 30 %) as a white solid.

The m.p. is 176–181 °C. 1H NMR (600 MHz, DMSO): δ 9.11 (app. bs, 1H, H10), 8.71 (s, 1H, H2), 8.61 (d, J = 2.5 Hz, 1H, H21), 8.13 (s, 1H, H3), 7.84 (d, J = 1.5 Hz, 1H, H23), 7.28 (s, 1H, H5), 6.58 (dd, J = 2.6, 1.6 Hz, 1H, H22), 3.19 (t, J = 6.4 Hz, 2H, H12), 2.22–2.13 (m, 1H, H16), 1.90–1.79 (m, 4H, H14, H15, H17, H18), 1.61–1.49 (m, 2H, H13), 1.21 (qd, J = 12.9, 3.4 Hz, 2H, H15′, H17′), 1.03 (qd, J = 13.0, 3.4 Hz, 2H, H14′, H18′).13C NMR (151 MHz, DMSO): δ 165.98 (C9), 163.17 (C6), 152.12 (C4), 148.55 (C2), 142.86 (C23), 130.96–125.42 (q, J = 278.1 Hz, C19), 127.97 (C21), 113.64 (C1), 108.83 (C22), 99.76 (C5), 44.61 (C12), 40.85–40.35 (q, J = 25.4 Hz, C16), 36.80 (C13), 28.55 (C14, C18), 24.20 (C15, C17).19F NMR (565 MHz, DMSO): δ −70.96, −72.84 (d, J = 9.0 Hz). Low levels of unassigned peaks in the NMR spectra likely reflect conformational isomers or pyrimidine/pyridine/amide conformational isomers as precedented with related compounds (as evidenced by variations in the intensities of some of them in variable temperature NMR studies74. FT-IR Vmax (film): 3,649, 3,442, 1,641, 1,574 cm−1. HRMS (ESI-TOF) calculated for C17H20O2N4F3 [M + H]+: 369.1532, found: 369.1533.

The synthetic route for IOX5 is shown Fig. 4f. NMR spectra are described in Supplementary Figs. 5–9, specifically 1H spectra for 2a (Supplementary Fig. 5), 1H for 2b (Supplementary Fig. 6), 1H for IOX5 (Supplementary Fig. 7), 13C for IOX5 (Supplementary Fig. 8) and 19F for IOX5 (Supplementary Fig. 9).

Visualization of Dap and IOX5-binding mode models

The structures of Dap or IOX5 were drawn in Chemdraw v.19.1 and transferred to Chemdraw 3D. Minimized energy (MM2) calculations were applied to investigate the stable conformations of Dap or IOX5. The file was saved as a mol file and opened in the respective pdb model in PyMOL. Using the pair-fitting feature, the minimized structures of Dap or IOX5 were overlaid with the identified structures to give predicted binding modes of Dap or IOX5.

Protein production

Recombinant truncated forms of PHD2 (ref. 37), FIH75, JMJD6 (ref. 76), KDM4A77, KDM5B15,78 and KDM6B79 were produced and purified from Escherichia coli as described. Full-length FIH (M1–N349), full-length OGFOD1 (M1–G542), full-length JMJD6 (M1–R423), PHD2 (181–426), KDM4A (M1-L359) and KDM6B (D1141–E1590) were produced and purified from E. coli as described. Summaries of the general procedures used are given here.

DNA encoding for human FIH (M1–N349) and KDM4A (M1–L359), with an N-terminal His6-tag were cloned into the pNIC28-Bsa4 vector. DNA encoding for human JMJD6 (M1–R423), OGFOD1 (M1–G542) and PHD2 (181–426) containing an N-terminal His6-tag were cloned into the pET-28b vector and that for KDM6B (D1141–E1590) into pNH-TrxT vector. The constructs were transformed into E. coli strain BL21(DE3).

A 10-ml overnight culture of each target was used to inoculate each of 12 l Terrific Broth medium containing 100 μg ml−1 kanamycin. Cultures were grown at 37 °C until the OD600 reached 1.0, cooled to 18 °C and induced for 18 h with 0.5 mM IPTG. Cells were collected, resuspended in lysis buffer, lysed with three passages through a high-pressure cell breaker and clarified by centrifugation at 21,000 rpm for 30 min. The lysis buffer contained 50 mM HEPES, pH 7.4, 500 mM NaCl, 20 mM imidazole, 0.5 mM Tris (2-carboxyethyl)-phosphine (TCEP) and 5% glycerol. JMJD6 (M1–R423) and OGFOD1 (M1–G542) contained a Tris-based lysis buffer (50 mM Tris-HCl (pH 8), 200 mM NaCl, 20 mM imidazole, 0.5 mM TCEP and 5% glycerol). Protease inhibitor mix 1:2,000 was added for lysis (Complete, EDTA-free Protease Inhibitor Cocktail, Roche Diagnostics).

The clarified supernatant was passed through an Ni NTA gravity column. After ten column volume washes with lysis buffer, the His6-tagged proteins were eluted with 300 mM imidazole. For JMJD6 (M1–R423), OGFOD1 (M1–G542) and PHD2 (181–426) the eluted fraction was incubated with 100 mM EDTA for 1 h at 4 °C. The eluted fractions containing appropriately purified proteins were further purified with a S200 Gel Filtration column equilibrated with 50 mM HEPES (pH 7.4), 150 mM NaCl, 0.5 mM TCEP and 5% glycerol. For JMJD6 (M1–R423) and OGFOD1 (M1–G542), a Tris-based buffer was used (Tris-HCl, pH 8, 200 mM NaCl, 0.5 mM TCEP and 5% glycerol). The monodispersed peak containing the respective protein was collected. As reported, the purities and molecular masses of the desired products were validated by SDS–PAGE and intact electrospray ionization mass spectrometry (PHD2 (ref. 37), FIH75, JMJD6 (ref. 76), KDM4A77, KDM5B15,78 and KDM6B79).

KDM5B (M1–R822)

cDNA encoding for KDM5B (residues M1–R822) was cloned into the pFB-LIC vector, encoding for a protein with a TEV-protease cleavable N-terminal 6x-histidine tag via ligation independent cloning. Human KDM5B was then expressed in Sf9 (Spodoptera frugiperda) insect cells.

Exponentially growing Sf9 cells (2 × 106 cells per ml) were infected with high titer baculovirus stock and incubated in shaker flasks. The cells were shaken at 90 rpm at 27 °C and collected 72 h after infection by centrifugation (15 min, 800g, 4 °C). The cell pellets were washed and resuspended in PBS. The cells were centrifuged again and the cell pellets were stored at −80 °C.

For protein purification, the cell pellet was thawed, resuspended in lysis buffer (50 mM HEPES, pH 7.4, 200 mM NaCl, 20 mM imidazole, 5% glycerol and 0.5 mM TCEP) in the presence of a protease inhibitor mix 1:2,000 (Complete, EDTA-free Protease Inhibitor Cocktail, Roche Diagnostics) and sonicated (2 min, amplitude 35%, on ice). The lysate was cleared by centrifugation (60 min, 36,000g, 4 °C). The protein was purified in the same methodology used for KDM4A (M1–L359) and KDM6B (D1141–E1590) production.

Inhibition assaysPHD2 solid phase extraction mass spectrometry assay

Inhibition of PHD2 (181–426) was measured using a solid phase extraction (SPE) mass spectrometry (MS) assay as reported previously35. In brief, compounds were dry dispensed into 384-well polypropylene plates using an ECHO 550 acoustic dispenser (Labcyte) approximating to a threefold dilution series across an 11-point concentration range (100 μM to 0.0017 μM). PHD2 (0.3 μM in 50 mM Tris-Cl, pH 7.5, 50 mM NaCl) was then dispensed across the plate using a Thermo Multidrop dispenser equipped with a small-volume dispensing cassette (25 μl per well). Compound dilutions were pre-incubated with PHD2 for 15 min; the reaction was initiated by dispensing 25 μl of a substrate solution (200 μM L-AA, 20 μM FAS, 20 μM 2OG and 10 μM CODD peptide) in 50 mM Tris-Cl, pH 7.5, 50 mM NaCl. The reaction was progressed for 15 min, then quenched by dispensing 5 μl 10% (v/v) aqueous formic acid. The assay plates were transferred to an Agilent RapidFire RF365 and processed as described in the section below (RapidFire SPE-MS Procedure).

FIH SPE-MS assays

Inhibition of FIH was measured using a SPE-MS assay and synthetic peptide substrate (HIF-1α788-822) and monitoring hydroxylation of the peptide product. In brief, compounds were dry dispensed into 384-well polypropylene plates using an ECHO 550 acoustic dispenser (Labcyte) approximating to a threefold dilution series across an 11-point concentration range (100 μM to 0.0017 μM). FIH (0.3 μM in 50 mM Tris-Cl, pH 7.5, 50 mM NaCl) was then dispensed (25 μl per well) across the assay plate using a Thermo Multidrop dispenser equipped with a small-volume dispensing cassette. Compound dilutions were pre-incubated with FIH for 15 min and then the reaction initiated by dispensing 25 μl of a substrate solution (200 μM L-AA, 20 μM FAS, 20 μM 2OG and 10 μM HIF-1α788-822) in 50 mM Tris-Cl, pH 7.5 and 50 mM NaCl. The enzyme reaction was allowed to progress for 15 min, then halted by dispensing 5 μl 10% (v/v) aqueous formic acid. Assay plates were transferred to an Agilent RapidFire RF365 and processed as described in the section below (RapidFire SPE-MS Procedure).

JMJD6 SPE-MS assay

The inhibitory effect of NOG, 2,4-pyridine dicarboxylic acid (2,4-PDCA), NOFD, Dap, roxadustat and molidustat on activity of full-length JMJD6 was measured using a 40-mer peptide substrate of bromodomain-containing protein 4 (BRD4511–550)80. Titrations of compounds were prepared using an ECHO 550 acoustic dispenser (Labcyte). An 11-point and threefold dilution for each compound (100 μM to 0.0017 μM) was prepared and dry dispensed into 384-well polypropylene plates. Full-length JMJD6 (1.0 μM in 50 mM Tris-Cl, pH 7.5) was then dispensed (25 μl per well) across the plate using a multidrop dispenser equipped with a low-volume dispensing cassette (Thermo). Compound dilutions were pre-incubated with JMJD6 for 15 min and the enzyme reaction was initiated by 25 μl dispense of the substrate mixture in 50 mM Tris-HCl, pH 7.5 (200 μM l-ascorbate, 20 μM ferrous ammonium sulfate, 20 μM 2-oxoglutarate and 10 μM JMJD6 substrate BRD4511–550). Reactions were progressed for 15 min at room temperature, then halted by dispensing 10% (v/v) aqueous formic acid (5 μl). The final concentration of DMSO was 0.5% (v/v). Assay plates were transferred to an Agilent RapidFire RF365 and processed as described in the section below (RapidFire SPE-MS procedure).

JMJD6 IOX5 IC50 determination

Inhibition of the catalytic activity of recombinant human JMJD6 was measured using an N-terminal peptide (RSKKRKKSKSRS) of RNA Binding Motif Protein 39 (RBM39 residues 31–42) and monitoring the appearance of the hydroxylated peptide product in 50 mM Tris-Cl, pH 7.5. Titrations of IOX5 for IC50 determinations (threefold and 11-point IC50 curves) were performed using an ECHO 550 acoustic dispenser (Labcyte) and dry dispensed into 384-well polypropylene assay plates. The final assay concentration of DMSO was kept constant at 0.5% (v/v). Full-length JMJD6 was prepared at a concentration of 1.0 μM in 50 mM Tris-Cl, pH 7.5 and 25 μl dispensed across the 384-well plates, JMJD6 was pre-incubated with compound dilutions for 15 min. The reaction was initiated by dispensing 25 μl of a substrate solution (20 μM FAS, 200 μM L-AA, 10 μM RBM3931–42 and 20 μM 2OG) across each 384-well assay plate. The reaction was allowed to progress for 30 min, then halted by dispensing 10% (v/v) aqueous formic acid (5 μl per well). Assay plates were analyzed by LC–MS as described in the section ‘Peptide detection by LC–MS’.

OGFOD1 SPE-MS assays

The inhibition of OGFOD1 activity was measured using a 30-mer peptide substrate of the ribosomal protein RPS23 (RPS2347–76). IC50 determinations were performed in 384-well-plate format using polypropylene plates (Greiner Bio One, cat. no. 781096). Compounds were prepared as 20 mM DMSO stock solutions and all compound dispenses were performed using an ECHO 550 acoustic dispenser (Labcyte). A positive control compound (2,4-PDCA, 100 mM) was dispensed into column 1 (250 nl) and DMSO was dispensed into column 13 (250 nl). All test compounds were serially diluted (an approximately threefold dilution series across an 11-point IC50, 100 μM to 0.0017 μM) and 250 nl of each dilution dispensed in duplicate into the polypropylene plate. OGFOD1 was diluted to 0.3 μM in assay buffer (50 mM Tris-Cl, pH 7.5) and was dispensed (25 μl) into the 384-well compound plates using a multidrop combi reagent dispenser (Thermo Scientific, 5840300) with a small-tube plastic-tip dispensing cassette (Thermo Scientific, 24073290). The compounds were pre-incubated with OGFOD1 for 15 min and the reaction was initiated by dispensing 25 μl of a substrate solution (200 μM L-AA, 20 μM FAS, 20 μM 2OG and 10 μM RPS23 (47–76) peptide) in assay buffer. The final concentration of DMSO in the assay was 0.5%. The reactions were allowed to progress for 20 min, then halted by addition of 10% (v/v) formic acid (5 μl). Assay plates were transferred to an Agilent RapidFire RF365 and processed as described in the section below (RapidFire SPE-MS procedure).

Lysine demethylase LC–MS assay

The inhibitory activity of IOX5 was measured by monitoring demethylation of peptide substrates for KDM4A, KDM5B and KDM6B. The peptide substrate for KDM4A was a 15-mer histone-H3 derivative (ARTAQTARK(me3)STGGIA) as reported previously81 and synthesized by GL Biochem. The peptide substrate for KDM5B was a 21-mer histone-H3 peptide (ARTK(me3)QTARKSTGGKAPRKQLA), synthesized by Peptide Protein Research. The KDM6B peptide substrate was a 17-mer histone-H3 peptide (LATKAARK(me3)SAPATGGVK), synthesized by GL Biochem.

KDM4A LC–MS assays

KDM4A reactions were performed under optimized buffer conditions (50 mM MES, pH 7.0). KDM4A (0.15 μM) was pre-incubated for 15 min in the presence of IOX5 (100 μM) and the enzyme reaction initiated by addition of a substrate solution (100 μM L-AA, 10 μM FAS, 10 μM 2OG and 10 μM peptide substrate). The reaction was progressed for 50 min and then stopped by the addition of formic acid to a final concentration of 1% (v/v). Control reactions in the presence of 0.5% (v/v) DMSO and a control in the presence of a known inhibitor of KDM4A (50 μM 2, 4-pyridine dicarboxylic acid82).

KDM5B LC–MS assays

KDM5B enzyme reactions were performed under optimized buffer conditions (50 mM MES, pH 7.0, 50 mM NaCl and 1 mM TCEP). KDM5B (0.15 μM) was pre-incubated for 15 min in the presence of IOX5 (100 μM) and the enzyme reaction was initiated by addition of substrate (100 μM L-AA, 10 μM FAS, 10 μM 2OG and 5 μM peptide). The enzyme reaction was progressed for 30 min, then halted by addition of formic acid to a final concentration of 1% (v/v). Control reactions included a 0.5% DMSO control and a reaction with a known inhibitor of KDM5B (10 μM KDOAM25 (ref. 83)).

KDM6B LC–MS assays

KDM6B reactions were performed under optimized buffer conditions (50 mM MES, pH 7.0). KDM6B (0.15 μM) was pre-incubated for 15 min in the presence of IOX5 (100 mM) and the enzyme reaction initiated by addition of substrate (100 μM l-ascorbate, 10 μM ferrous ammonium sulfate, 10 μM 2OG and 5 μM peptide). The enzyme reaction was progressed for 30 min, then halted by addition of formic acid to a final concentration of 1% (v/v). Control reactions included a 0.5% DMSO control and a reaction with a known inhibitor of KDM6B (10 μM GSKJ1 (ref. 84)).

Peptide detection by LC–MS

KDM4A, KDM5B and KDM6B reactions were transferred to a 96-well polypropylene plate and peptide analysis was performed by LC–MS using an Agilent 1290 infinity II LC system equipped with an Agilent 1290 multisampler and an Agilent 1290 high-speed pump and connected to an Agilent 6550 Accurate Mass iFunnel quadrupole-time of flight (QTOF) mass spectrometer. Then, 4 μl of the enzyme reaction was injected and loaded onto a ZORBAX RRHD Eclipse Plus C18 column (Agilent Technologies). Solvent A consisted of LC–MS grade water containing 0.1% (v/v) formic acid and solvent B consisted of acetonitrile containing 0.1% (v/v) formic acid. Peptides were separated using a step-wise gradient (0 min in 95% solvent A; 1.0 min in 80% solvent A; 3.0 min in 45% solvent A; 4.0 min in 45% solvent A; 5.0 min in 0% solvent A; 6.0 min in 0% solvent A; and 7.0 min in 95% solvent A). This was followed by a 3-min post run with 95% solvent A to re-equilibrate the column; all flow rates were 0.2 ml min−1. The mass spectrometer was operated in the positive ion mode with a drying gas temperature of 280 °C, drying gas flow rate of 13 l min−1, nebulizer pressure of 40 psig, sheath gas temperature of 350 °C, sheath gas flow rate of 12 l min−1, capillary voltage of 4,000 V and nozzle voltage of 1,000 V. All acquired data were analyzed using Agilent MassHunter Qualitative Analysis (v.B.07.00) software.

RapidFire SPE-MS procedure

Assay plates were then transferred to an Agilent RapidFire RF365 machine connected to an Agilent 6550 Accurate Mass iFunnel QTOF mass spectrometer. Samples were aspirated under vacuum, loaded onto a C4 SPE cartridge and then washed with 0.1% (v/v) aqueous formic acid to remove buffer salts. Peptides were eluted from the C4 SPE cartridge using 0.1% (v/v) formic acid in 85% (v/v) acetonitrile and 15% (v/v) LC–MS-grade water onto the QTOF mass spectrometer. The mass spectrometer was operated in the positive ion mode with a drying gas temperature of 280 °C, drying gas flow rate of 13 l min−1, nebulizer pressure of 40 psig, sheath gas temperature of 350 °C, sheath gas flow rate of 12 l min−1, capillary voltage of 4,000 V and nozzle voltage of 1,000 V. Ion chromatogram data for the most predominant charge state was extracted and peak area data for the substrate peptide and hydroxylated peptide integrated using RapidFire Integrator (Agilent, v.4.3.0.17235). The percentage conversion was calculated using Excel and IC50 curves generated using GraphPad Prism (v.5.04).

Biophysical methodsNMR spectroscopic studies

NMR spectra were measured using a Bruker AVIII 700 MHz NMR spectrometer equipped with a TCI inverse cryoprobe using 3-mm diameter high-throughput NMR tubes (Norell). Samples were recorded at 298 K. Data were processed with TopSpin v.3.6.2 software.

1H CPMG NMR experiments

For 1H Carr–Purcell–Meiboom–Gill (CPMG) NMR spectroscopy, the assay mixtures contained 50 µM apo-PHD2181-426, 200 µM Zn (II), 50 µM 2OG, increasing concentrations of ligand (up to 400 µM) in 50 mM Tris-D11, 150 mM NaCl (pH 7.5), in 90 % H2O and 10 % D2O (v/v). Typical experimental parameters for CPMG NMR spectroscopy were total echo time, 100 ms (τ = 1 ms, n = 50); number of points, 32,768; sweep width, 16 ppm; relaxation delay, 2 s; and number of transients, 64. The PROJECT-CPMG sequence (90°x-(τ -180°y-τ -90°y-τ -180°y-τ)n-acq) was employed. Pre-saturation was used to achieve water suppression15.

Western blot assays

For immunoblotting in HEK293T cells, cells were treated with increasing concentrations of IOX5, 20 μM of roxadustat or vehicle control (as indicated in Fig. 4d) using cell lysis buffer (1× RIPA buffer, Sigma, R0278) supplemented with protease inhibitors (Complete, Mini, EDTA-free Protease Inhibitor Cocktail, Roche). Protein extracts were subjected to SDS–PAGE separation (NuPAGE 4–12% Bis-Tris Plus Gel, Thermo Fisher Scientific, NP0323BOX), then transferred onto a polyacrylamide membrane using wet transfer. The gels were run in 1× Tris/glycine/SDS running buffer (20× NuPAGE MES SDS Running Buffer, Life Technologies) at 180 V for 45 min (Mini Gel Tank, Life Technologies). Membranes were blocked with 5% milk powder in 1× PBS-T for 30 min then incubated overnight at 4 °C with anti-HIF-1α (BD Biosciences, cat. no. 610959, 1:1,000, ON at 4 °C) and anti-GAPDH (Invitrogen, cat no. MA5-15738, 1:1,000, ON at 4 °C). After incubation with appropriate horseradish peroxidase-coupled secondary antibody (Cell Signaling Technology, rabbit anti-mouse IgG, (D3V2A), 1:5,000, 2 h at room temperature), proteins were detected with by GE Healthcare Amersham ECL Prime Western Blotting Detection Reagent (RPN2236) and acquired on the Bio-Rad Universal Hood III.

For immunoblotting in AML cells, MOLM13, OCI-AML3, MV411 and THP-1 cells treated with 50 μM Dap, 50 μM IOX5 or vehicle control using cell lysis buffer (Cell Signaling Technology, 9803) supplemented with protease and phosphatase inhibitors (Merck, 20-201, 524624). Total protein extracts (30 or 60 μg) were subjected to SDS–PAGE separation (NuPAGE 4–12% Bis-Tris Plus Gel, Thermo Fisher Scientific, NP0323BOX) and then transferred onto polyvinylidene difluoride membranes using wet transfer. Membranes were blocked in 5% milk-TBST (TBS with 0.1% Tween20) and probed with anti-HIF-1α (BD Biosciences, 610959, 1:1,000, ON at 4 °C), anti-HIF-2α (Cell Signaling Technology, 59973, 1:1,000, ON at 4 °C), anti-BNIP3 (Abcam, EPR4034, 1:2,000, ON at 4 °C), anti-β-actin (Cell Signaling Technology, 3700, 1:10,000, 30 min at room temperature) and anti-histone-H3 (Cell Signa