Fig 2: LSD1 inhibition sensitizes AML cells to physiological concentrations of RA.(A) Heat map representing the results of cell proliferation assays performed on 21 AML cell lines treated with MC_2580 (2 µM) and/or RA as indicated. Values are normalized on dimethyl sulfoxide (DMSO) treatment. The AML French-American-British classifications of each cell line are in parentheses. (B) Growth curves of NB4 APL cells treated as indicated. (C) Morphological analysis by May-Grünwald-Giemsa staining of NB4 cells treated for 96 hours in liquid culture as indicated. (D) Colony-forming ability, scored after 7 days, of 1000 NB4 cells plated in methylcellulose medium and treated with MC_2580 (2 µM) and/or 0.01 µM RA and 1 µM RA. Means and SDs of three independent experiments are shown. C.F.U., colony-forming units. (E) Analysis of CD11b mRNA levels in NB4 cells treated for 96 hours in liquid culture as indicated. Values are normalized against GAPDH (glyceraldehyde phosphate dehydrogenase) and referred to DMSO. The graph represents the mean and SD of three independent experiments. FC, fold change. (F) Western blot showing depletion of LSD1 in NB4 cells knocked out for LSD1 (LSD1 KO). WT, wild type. (G) Proliferation assays of NB4 and NB4 LSD1 KO cells after 24, 48, and 72 hours of the indicated treatments (DMSO as control). (H) Expression of CD11b by fluorescence-activated cell sorting (FACS) in NB4 and NB4 LSD1 KO cells after 24, 48, and 72 hours of the indicated treatments (DMSO as control). (I) Kaplan-Meier survival plots for mice transplanted with murine APL cells and treated as indicated (n = 6 for each treatment group). Pink shaded area indicates the duration of RA treatment (21 days, pellet), while LSD1i (DDP_38003) was administrated twice a week orally (OS) for the entire duration of RA treatment (total, six times). P values were obtained using analysis of variance (ANOVA).

Fig 3: Pharmacological inhibition of LSD1 disrupts its interaction with GFI1.(A) Schematic representation of SILAC mass spectrometry approach to identify LSD1 interactors in NB4 cells. LC-MS/MS, liquid chromatography–tandem mass spectrometry. (B) Scatterplot showing log2 (heavy/light) ratio of forward reaction on the x axis (Rep1) and the log2 (light/heavy) ratio of reverse reaction on the y axis (Rep3). In the top right quadrant are represented LSD1 interactors. The blue dashed lines define the threshold used to define the LSD1 interactors from the background. Proteins belonging to the CoREST complex are shown in red dots. (C) Venn diagrams with numbers of individual and overlapping putative LSD1 interactors identified in the three different SILAC replicates. (D) Western blot analysis of LSD1 and some identified interactors in LSD1 IPs, with or without blocking peptide. Lamin B1 is used as loading control. (E) Schematic representation of SILAC mass spectrometry approach to identify recruited and evicted interactors of LSD1, upon LSD1 pharmacological inhibition with 2 µM MC_2580 for 24 hours. (F) Scatterplot showing log2 (heavy/light) ratio of forward reaction on the x axis and the log2 (light/heavy) ratio of reverse reaction on the y axis. Proteins recruited by LSD1 after inhibition are present in the top right quadrant, while proteins evicted from the interaction with LSD1 after drug treatment are found in the bottom left quadrant. Proteins previously identified as interactors of LSD1 in NB4 are shown as red dots. The blue dashed lines define the threshold used to determine recruited and evicted proteins. (G) Western blot analysis of LSD1 and GFI1 in LSD1 IPs in NB4 cells treated for 24 hours with DMSO, 2 µM MC_2580, or 2 µM DDP_38003. (H) Western blot analysis of LSD1 and GFI1 in GFI1 IPs in NB4 LSD1 KO cells transduced with empty vector, wild-type, or catalytic inactive K661A-LSD1, treated with 2 µM MC_2580 or DMSO.

Fig 4: The catalytic activity of LSD1 is dispensable.(A) Western blot analysis of LSD1 levels in NB4 LSD1 KO cells transduced with expression vectors encoding for wild-type LSD1 or the catalytically inactive mutant K661A-LSD1. Empty vector was used as control, and actin served as loading control. (B) Growth curves of NB4 LSD1 KO cells transduced with empty vector, LSD1, or K661A-LSD1 and treated with 0.01 µM RA (DMSO as control treatment). (C) Percentage of CD11b-expressing cells assessed by FACS in NB4 LSD1 KO cells transduced with expression vectors encoding for LSD1, K661A-LSD1, or empty vector, treated with 0.01 µM RA (or DMSO as control) for 24 hours. (D) H3K4me2 enrichment at the promoter of PI16 (left), ITGAM (center), and TGM2 (right) in NB4 and NB4 LSD1 KO cells transduced with expression vectors encoding for LSD1, K661A-LSD1, or empty vector, assessed by ChIP-qPCR. The results represent percentage of input chromatin. Error bars indicate SD from triplicate experiments, and P values were determined by a two-tailed unpaired Student’s t test (*P < 0.05, **P < 0.01, and ***P < 0.001).

Fig 5: The interaction of LSD1 with GFI1 is fundamental for LSD1 activity in APL cells.(A) Heat maps showing ChIP-seq signal of LSD1 and GFI1 ranked according to decrescent LSD1 signal. bp, base pairs. (B) Representative examples of overlapping LSD1 and GFI1 binding regions. University of California Santa Cruz Genome Browser profile of LSD1 and GFI1 ChIP-seq on the PI16 gene promoter and interferon regulatory factor 8 (IRF8) putative enhancer is shown. (C) HDAC1 recruitment (measured by ChIP-qPCR) at three genomic regions bound by both LSD1 and GFI1, before and after treatment with MC_2580. n.s., not significant. **P < 0.01 and ***P < 0.001. (D) Western blot analysis of LSD1 and GFI1 in GFI IP assays performed in NB4 LSD1 KO cells reconstituted with wild-type LSD1, LSD1-D553,555,556A mutant, or empty vector (as control). IgG, immunoglobulin G. (E) Growth curves of NB4 LSD1 KO cells transduced with wild-type LSD1, LSD1-D553,555,556A, or empty vector and treated for the indicated time with 0.01 µM RA (or DMSO as control).