Molidustat Induces Cell Death in Colonic Cell Lines independently of PHD-inhibition

A. Percentage of Cleaved Caspase 3/7 positive HT29 (top), and RKO (bottom) cells. Cells were treated with Molidustat for 48 hours at indicated concentrations, 10uM Staurosporine was used as a positive control (100% cell death). Mean + SEM is assessed by unpaired two tailed Student’s t-test, **p<0.01, (ns) non-significant. B. Representative images of Cleaved Caspase-3/7 signal in DMSO, Molidustat (90 μM), and Staurosporine treated cells. Scale bar: 300 μm. C. Representative Western Blot of PHD2 levels in HT29 cells. D. Percentage confluency of HT29 cells post-transfection with the indicated guide RNAs. E. Cleaved Caspase-3/7 signal in HT29 cells post-transfection with the indicated crRNAs. Mean + SEM is assessed by two-way ANOVA, *p<0.05. N = 3 biologically independent experiments.

Thermal Proteome Profiling Reveals Multiple Molidustat Targets.

A. Intact cells are treated with compound or vehicle control (1). Cells are collected (2), split into fractions (3) with each fraction heated to a different temperature (4). After heat treatment, cells are lysed and proteins are digested to peptides (5). The peptides in each fraction are labelled with a unique TMT tag, which can analysed using mass spectrometer (6), generating melting curves and Tm50 values for the whole proteome. This panel was created using BioRender.com. B. Volcano plot of hits identified using the Tm50 method. Tm50 shift and -log10 p-value are shown. Proteins exhibiting a significant shift (< 0.05) greater than 1 °C were considered potential targets. C. Volcano plot of all hits identified using the ΔY method. ΔY shift and -log10 p-value are shown. Proteins with a significant increase (< 0.05) in ΔY greater than 0.1 were taken forward as potential hits. D. Venn diagram showing the overlap of targets identified by both methods. Twelve proteins were identified by both methods are shown on the right.

Molidustat Inhibits GSTP1 Activity.

A. ω-aminohexyl sepharose beads were coupled with IOX4. Cell lysate is washed over empty beads (negative controls), IOX4 coupled beads (condition A), or IOX4 coupled beads after incubation with Molidustat (condition B). Samples are analysed by Western Blot or Mass Spectrometry. This panel was created using BioRender.com. B. Western blot of PHD2 pull-down with IOX4-conjugated beads across a concentration range (0.5-8 μM). C. Competition assay with Molidustat. Lysates were pre-incubated with Molidustat prior to incubation with IOX4-conjugated beads, PHD2 pulldown was assessed by Western Blot. D. Western Blot of GSTP1 binding to IOX4 beads with or without prior incubation with Molidustat. E. Molecular docking of Molidustat (left) or Eziostat (right) to GSTP1 protein. Residues in the active site are highlighted on the right. F. Recombinant (left) or cell-based (right) GSTP1 activity assay with Molidustat (1, 10, 100 μM) or DMSO; time-course of CDNB-GSH conjugation monitored at 340 nm and plotted as relative absorbance (arbitrary units, AU). Mean + SEM is assessed by two-way ANOVA, *p<0.05 N = 3 biologically independent experiments.

Combined PHD2 And GSTP1 Loss Drives Synergistic Proteomic Changes Associated with Apoptosis.

A. Heatmap of top 500 variable proteins in the tested data set. The p-values were calculated using one-way ANOVA. Hierarchical clustering of columns (samples) and rows (proteins detected) was performed using Euclidean distance. B. Protein enrichment analysis of bottom cluster using MSigDB.Hallmark database as a reference. C. MS analyses of GPX4 levels in the cell populations. GPX4 intensity is shown as a label-free quantitation (LFQ) values. Mean + SEM is assessed by unpaired two-tailed Student’s t test, **p<0.01. D. Heatmap of GO term “Intrinsic Apoptotic Signalling Pathway in Response to Oxidative Stress” proteins found in the tested data set. Hierarchical clustering of columns (samples) and rows (proteins detected) was performed using Euclidean distance. E. Synergy profiles of down-regulated (top) and up-regulated (bottom) proteins, expressed as per-protein z-scores. A total of 361 down-regulated (N = 361) and 314 up-regulated (N = 314) proteins exhibited synergistic behaviour.

Metabolic Imbalance and Apoptotic Vulnerability Following Combined PHD2 And GSTP1 Loss.

A. Heatmap of top 50 variable metabolites in the tested data set. The p-values were calculated using one-way ANOVA. Hierarchical clustering of columns (samples) and rows (metabolites detected) was performed using Euclidean distance. B. Metabolomics analyses of AMP/ATP, GSSH/GSH, NAD+/NADH and NADP/NADH ratios, in the indicated cell lines. Mean + SEM is assessed by unpaired two-tailed Student’s t test, **p<0.01. C. TCA cycle metabolite levels, normalised to sgControl. Mean + SEM is assessed by unpaired two-tailed Student’s t test, **p<0.01, ****p<0.0001. D. Left: Confluence curve, normalised to timepoint 0h in the indicated cell lines. Right: Representative Cleaved Caspase-3/7 signal curve in the indicated cell lines. Mean + SEM is assessed by two-way ANOVA, **p<0.01. N = 3 biologically independent experiments.

Colonic Organoids Display Genotype-Specific Sensitivity to Molidustat.

A. WT, APCfl/fl(CMS2), AKP (CMS3) and KPN (CMS4) organoids were treated with increasing concentrations of Molidustat for 48 h. Viability percentage was assessed using an MTT Assay. Mean + SEM is assessed by Two-way Anova with Dunnett’s Correction for Multiple Comparisons, *p<0.05, ***p<0.001, ****p<0.0001. B. 2-day old wild type (WT) or APCfl/fl organoids were treated for 48h hours with the indicated PHD inhibitors. Organoid viability percentage was assessed with an MTT assay. Mean + SEM is assessed by unpaired two tailed Student’s t-test, ***p<0.001, ****p<0.0001, (ns) non-significant. C. Top: representative images of Cleaved Caspase-3/7 signal in wild type and APCfl/fl organoids treated with DMSO or Molidustat at 24 h timepoint. Scale bar: 1000 μm. Bottom: quantification of Cleaved Caspase-3/7 signal area in organoids at the indicated timepoints. Mean + SEM is assessed by unpaired two tailed Student’s t-test, ***p<0.001, (ns) non-significant. N = 3 biologically independent experiments.

PHD2 as a hit from TPP Easy50 Script.

A. Raw values and sigmoidal curve fits for PHD2 in samples treated with Molidustat (blue) or DMSO (black). B. Raw values and sigmoidal curves for PLOD1 in samples treated with Molidustat (blue) or DMSO (black). N = 3 biologically independent experiments.

Molidustat Inhibits GSTP1 Activity.

A. Mass Spectrometry analyses of GSTP1 levels in the indicated conditions. Protein intensity is shown as a label-free quantitation (LFQ) values. B. HT29 cells were treated with increasing concentrations of Molidustat. ROS were measured using Peroxy Orange 1 (PO1) and expressed as arbitrary units (AU). At 50 μM Molidustat, a significant increase in PO1 fluorescence was observed. Co-treatment with 5 mM N-acetylcysteine (NAC), a ROS scavenger, abolished the Molidustat-induced signal. Mean + SEM is assessed by one-way ANOVA, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, (ns) non-significant. N = 3 biologically independent experiments. C. Recombinant (left) or cell-based (right) GSTP1 assay with ethacrynic acid (EA) or DMSO; time-course of CDNB-GSH conjugation monitored at 340 nm and plotted as relative absorbance (arbitrary units, AU). D. Cleaved Caspase-3/7 signal curve in the indicated treatments. Mean + SEM is assessed by two-way ANOVA, N = 1 biologically independent experiment.

Proteomics analyses of HT29 knockout cell models.

A. Mass Spectrometry analyses of PHD2 (left), or GSTP1 (right) levels in the cell populations. Protein intensity is shown as a label-free quantitation (LFQ) values. B Mass Spectrometry analyses of HIF1A levels in the cell populations. HIF1A intensity is shown as a label-free quantitation (LFQ) values. C. Protein enrichment analysis of top cluster from Figure 4A using MSigDB.Hallmark database as a reference. D. Protein enrichment analysis of synergistic, down-regulated proteins (as seen in Fig. 4E) using MSigDB.Hallmark database as a reference. D. Protein enrichment analysis of synergistic, up-regulated proteins (as seen in Fig. 4E) using MSigDB.Hallmark database as a reference.

Metabolomics analyses of HT29 knockout cell models.

A. Enrichment analysis of metabolomics and proteomics data, showing top 15 enriched metabolic processes in sgPHD2 and sgGSTP1 double knockout cell line comparing to sgControl. B. Representative merged images of Phase and Cleaved Caspase 3/7 signal in the indicated cell lines. Scale bar: 400 μm. N = 3 biologically independent experiments.

Evaluation of Molidustat in CRC in vivo model.

A. Epo mRNA levels after Molidustat treatment. Mean + SEM is assessed by Mann-Whitney U test, **p<0.01. B. Representative immunohistochemistry images of rolls of the upper section of the intestinal epithelium with a large number of macroscopic and microscopic tumours. Sections were stained with BrdU to stain DNA of actively dividing cells. C. Tumour burden of mice treated with DMSO or Molidustat. Mean + SEM is assessed by unpaired two tailed Student’s t-test, (ns) non-significant. D. Tumour number in mice treated with DMSO or Molidustat. Mean + SEM is assessed by unpaired two tailed Student’s t-test, (ns) non-significant. E. Tumour size of mice treated with DMSO or Molidustat. Mean + SEM is assessed by unpaired two tailed Student’s t-test, (ns) non-significant. F. Survival curve of mice treated with DMSO or Molidustat. Log-rank Mantel Cox test, p=0.2924. N = 10 DMSO, N = 10 Molidustat

The effects of different PHD inhibitors on murine colonic organoids viability.

Wild type (WT) or APC organoids viability after 48 hour treatment with the indicated PHD inhibitors. Organoid viability percentage assessed using MTT assay. Mean + SEM is assessed by two-way ANOVA, (ns) non-significant. N = 2 biologically independent experiments. B. Representative images of WT and APC organoids treated with 30 and 90μM Molidustat. Scale bar: 1000 μm. N = 3 biologically independent experiments.