Preclinical studies reveal that LSD1 inhibition results in tumor growth arrest in lung adenocarcinoma independently of driver mutations

Abstract

Lung adenocarcinoma (LUAD) is the most prevalent subtype of non-small cell lung cancer. Despite the development of novel targeted and immune therapies, the 5-year survival rate is still only 21%, indicating the need for more efficient treatment regimens. Lysine-specific demethylase 1 (LSD1) is an epigenetic eraser that modifies histone 3 methylation status, and is highly overexpressed in LUAD. Using representative human cell culture systems and two autochthonous transgenic mouse models, we investigated inhibition of LSD1 as a novel therapeutic option for treating LUAD. The reversible LSD1 inhibitor HCI-2509 significantly reduced cell growth with an IC₅₀ of 0.3-5 μmin vitro, which was linked to an enhancement of histone 3 lysine methylation. Most importantly, growth arrest, as well as inhibition of the invasion capacities, was independent of the underlying driver mutations. Subsequent expression profiling revealed that the cell cycle and replication machinery were prominently affected after LSD1 inhibition. In addition, our data provide evidence that LSD1 blockade significantly interferes with EGFR downstream signaling. Finally, our in vitro results were confirmed by preclinical therapeutic approaches, including the use of two autochthonous transgenic LUAD mouse models driven by either EGFR or KRAS mutations. Importantly, LSD1 inhibition resulted in significantly lower tumor formation and a strong reduction in tumor progression, which were independent of the underlying mutational background of the mouse models. Hence, our findings provide substantial evidence indicating that tumor growth of LUAD can be markedly decreased by HCI-2509 treatment, suggesting its use as a single agent maintenance therapy or combined therapeutical application in novel concerted drug approaches.

Keywords: HCI-2509; KDM1A; LSD1; epigenetic alterations; histone methylation; lung adenocarcinoma.

Figure 1

HCI‐2509 reduces the viability and invasion capacities of LUAD cells. (A–D) MTT assay after treating the different cell lines for 5 days with a range of HCI‐2509 concentrations as indicated. The viability of untreated cells (Ctrl) was set to 100%. (A) EGFR‐mutated cell lines PC9 and H1975, (B) KRAS‐mutated cell lines A549 and H460 and (C) cell lines with an EML4/ALK translocation H2228 and H3321. SD and the significances of all values were calculated using analysis of variance (ANOVA) followed by Dunnett's post‐hoc test. (D) Calculated IC ₅₀ values of NSCLC cells treated with HCI‐2509. (E) Cell cycle analysis by flow cytometry using PI staining of the indicated cell lines (A549, H460, H1975, PC9, H3122 and H2228) treated with 0 or 2 μm HCI‐2509 for 72 h. The percentage of cells (left) was calculated using appropriate gating (Fig.  S2B). The percentage of each cell cycle phase of untreated cells per cell line was set to 0 and the percentage of treated cells was calculated accordingly (right). The significance for the relative amount was calculated by means of two‐way ANOVA (left) and the fold change (right) was calculated using Student′s t‐test. (F) Invasion assays performed with Boyden chambers. Invaded cells were stained with crystal violet after 48 h and photographed five times. The pixel value was measured for each photograph and the value of the untreated cells was set to 1. The HCI‐2509‐treated cells were calculated accordingly. Significances were calculated with Student's t‐test. Significances are indicated: *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001.

Figure 2

The cell cycle, as well as the EGFR signaling pathway, is dysregulated after HCI‐2509 treatment. (A) The top 20 most significant pathways revealed by reactome pathway analysis. Pathways related to cell cycle control are indicated in dark blue and light blue bars indicate other pathways. (B) Heatmap of expression profiles shown by microarray analysis (MA) were validated by qPCR using RNA of untreated (Ctrl) and HCI‐2509‐treated A549 and PC9 cells. Validation experiments were performed in biological triplicates. (C) Phosphorylation array results in PC9 and A549 cells showing the relative normalized integrated intensity. Values of untreated (Ctrl) cells were set to 0 and HCI‐2509‐treated cells were calculated, respectively. Significances were calculated using Student's t‐test. Significances are indicated: *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. (D) Western blot analysis on PC9 and A549 untreated (Ctrl) and HCI‐2509‐treated (48 h, 2 μm HCI‐2509) immunoblotted against LSD1, AKT, p‐AKT (S373), MEK, p‐MEK, p‐ERK, ERK and the normalization protein β‐actin.

Figure 3

Impeded tumor growth in an EGFR‐driven transgenic mouse model by HCI‐2509 treatment. (A) Exemplary μCT scans of C57BL/6N^{TG ( EGFR L858R) ×  TG ( CC 10‐rt TA )} either treated for 2 or 4 weeks with HCI‐2509 or not treated (Ctrl). (B) The relative tumor growths of the control and treated mice were normalized to total lung volume. Significances were calculated using Student's t‐test (*P ≤ 0.05). (C) Representative staining (H&E, LSD1 IHC and Ki67 IHC) of EGFR‐driven lung tumors of mice not treated (Ctrl) or with HCI‐2509 treatment over a period of 4 weeks for EGFR‐driven lung tumors. Scale bars = 200 μm. (D) Ki67 expression quantification by counting of total positive cells per view. (E) Immunosquare blot analysis of C57BL/6N^{TG ( EGFR L858R) ×  TG ( CC 10‐ RTTA )} mice treated for 4 weeks with control feed (n = 7) or HCI‐2509 diet (n = 8), using antibodies against LSD1, H3K4me2, H3K9me2 and β‐actin. The signals were measured using image lab 4.0.1 (Bio‐Rad) and the signal values of LSD1, H3K4 and H3K9 were normalized using the β‐actin signals. The box plot includes the mean value of each group and each target. Outliers were calculated using Tukey's test and significances were calculated using Student's t‐test. Significances are indicated by stars: *P ≤ 0.05, **P ≤ 0.01.

Figure 4

Reduction of tumor growth of KRAS G12V driven LUAD after treatment with HCI‐2509. (A) Tumor occurrence in the control (Ctrl) (n = 15) or HCI‐2509‐treated (n = 16) group divided by mice showing clear tumor nodules in the H&E‐stained lung half and mice that were tumor‐free. Significances were calculated using Fisher's exact test. (B) The tumor area was measured using cellp software (Olympus, Hamburg, Germany). For mice with tumors, the mean tumor area of all nodules is depicted. Tumor‐free mice have a tumor area of 0. Points indicate single mice, the bars indicate SEM and the significances (*P < 0.05) were calculated by Student's t‐test. (C) Example of stainings (H&E, LSD1 IHC and Ki67 IHC) of mouse lungs with tumors and without tumors for two KRAS‐driven control mice. Scale bars = 200 μm. (D) Quantification of Ki67‐positive cells. (E) LSD1, H3K4me2, H3K9me2 protein levels were studied using immunosquare blot analysis of lung tissue lysates of C57BL/6N^{( KRAS G12V)} mice treated with control diet (n = 15) or HCI‐2509 diet (n = 15). The signals were measured using image lab 4.0.1 (Bio‐Rad) and the signal values of LSD1, H3K4 and H3K9 were normalized using the β‐actin signals. The box plot includes the mean value of each group and each target. Outliers were calculated using Tukey's test and significances were calculated using Student's t‐test. Significances are indicated by stars: *P ≤ 0.05, **P ≤ 0.01. ***P ≤ 0.001.