COMBINATIONS OF LSD1 INHIBITORS FOR USE IN THE TREATMENT OF SOLID TUMORS

Abstract

The instant invention relates to therapeutic combinations of LSD1 inhibitors and one or more other active pharmaceutical ingredient(s) or pharmaceutically acceptable salts thereof. The combinations are particularly useful for treating neoplastic diseases, such as cancer, particularly small cell lung cancer (SCLC).

Claims

1-30. (canceled)

31. A method for the treatment of a neoplastic disease, which method comprises administering an effective amount of an LSD1 inhibitor which is (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine or a pharmaceutically acceptable salt thereof, and administering an effective amount of one or more active pharmaceutical ingredients selected from ABT-199, ABT-263, ABT-737, Belinostat, Bendamustine, BGJ398, Carboplatin, Cisplatin, CPI-203, Docetaxel, Doxorubicin, EPZ-5676, EPZ-6438, Etoposide, Gemcitabine, GSK126, GSK1324726A, GSK1210151A, Irinotecan, (+)-JQ1, Lapatinib, LY2603618, MLN8237, OTX015, Paclitaxel, Panobinostat, SGC 0946, Temozolomide, Topotecan, Vincristine and pharmaceutically acceptable salts thereof, to a human being or animal.

32.-35. (canceled)

36. The method of claim 31, wherein the neoplastic disease is a cancer.

37. The method of claim 31, wherein the neoplastic disease is selected from breast cancer, prostate cancer, cervical cancer, ovarian cancer, gastric cancer, colorectal cancer, pancreatic cancer, liver cancer, brain cancer, neuroendocrine cancer, lung cancer, kidney cancer, hematological malignancies, melanoma and sarcomas.

38. The method of claim 31, wherein the neoplastic disease is a solid tumor.

39. The method of claim 31, wherein the neoplastic disease is small cell lung cancer (SCLC).

40. (canceled)

41. The method of claim 31, wherein the neoplastic disease is lung cancer.

42. The method of claim 31, wherein the neoplastic disease is neuroendocrine cancer.

43. The method of claim 31, wherein the neoplastic disease is blood cancer.

44. The method of claim 31, wherein the neoplastic disease is acute myeloid leukemia (AML).

45. The method of claim 31, wherein said method comprises administering an effective amount of one BCL2 inhibitor selected from ABT-199, ABT-263, ABT-737, and a pharmaceutically acceptable salt thereof.

46. The method of claim 31, wherein said method comprises administering an effective amount of ABT-199 or a pharmaceutically acceptable salt thereof.

47. The method of claim 31, wherein said method comprises administering an effective amount of one BET inhibitor selected from CPI-203, GSK1324726A, GSK1210151A, (+)-JQ1, OTX015, and a pharmaceutically acceptable salt thereof.

48. The method of claim 31, wherein said method comprises administering an effective amount of one DOT1 L inhibitor selected from EPZ-5676, SGC 0946, and a pharmaceutically acceptable salt thereof.

49. The method of claim 31, wherein said method comprises administering an effective amount of one DNA alkylating agent selected from Bendamustine, Carboplatin, Cisplatin, Temozolomide and a pharmaceutically acceptable salt thereof.

50. The method of claim 31, wherein said method comprises administering an effective amount of one HDAC inhibitor selected from Belinostat, Panobinostat, and a pharmaceutically acceptable salt thereof.

51. The method of claim 31, wherein said method comprises administering an effective amount of one topoisomerase inhibitor selected from Etoposide, Irinotecan, Topotecan, and a pharmaceutically acceptable salt thereof.

52. The method of claim 31, wherein said method comprises administering an effective amount of one anti-mitotic agent selected from Docetaxel, Paclitaxel, Vincristine, and a pharmaceutically acceptable salt thereof.

53. The method of claim 31, wherein said method comprises administering an effective amount of Docetaxel or a pharmaceutically acceptable salt thereof.

54. The method of claim 31, wherein said method comprises administering an effective amount of Paclitaxel or a pharmaceutically acceptable salt thereof.

55. The method of claim 31, wherein the LSD1 inhibitor is (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine bis-hydrochloride.

Description

DESCRIPTION OF THE DRAWINGS

[0309] FIG. 1A and FIG. 1B: In vitro differential activity of LSD1 inhibitors (i.e. (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine) in a panel of SCLC cell lines treated for 7 days. “Classic” neuroendocrine cell lines such as NCI-H1876 (FIG. 1A) and NCI-H510 (FIG. 1B) maintained a high level of sensitivity.

[0310] FIG. 2A and FIG. 2B: LSD1 inhibitors (i.e. (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine)improve potency and duration of SCLC standards of care (SOC) etoposide and carboplatin in vivo (FIG. 2A) as well as in vivo in mice (FIG. 2B).

[0311] FIG. 3: Principal component analysis score plot for principal component 1 (t[1], x-axis) and principal component 2 (t[2], y-axis) separates classic cell lines (C, black) from variant cell lines (V, gray) according to Example 5.

[0312] FIG. 4: Heat Map showing mRNA expression (as z-scores) for the gene panel of Example 6 comprising the genes of Table 8, Table 6 and MYC. These genes best predict response to an LSD1 inhibition therapy in the 19 cell lines of Table 6. Higher z-scores correlate with better sensitivity.

[0313] FIG. 5: Heat Map showing mRNA expression (as z-scores) for the neuroendocrine genes of Example 7 in the 19 cell lines of Table 6. Sensitive cell-lines display a stronger expression (higher z-score) of such neuroendocrine markers.

[0314] FIG. 6: Signature scores obtained by PLS analysis using the second principal component according to Example 8. Cell lines with score_1>0.5 are more likely to be sensitive to an LSD1 inhibition therapy.

[0315] FIG. 7: Signature scores obtained by PLS analysis using the first principal component according to Example 8. Cell lines with score_2>0.5 are more likely to be sensitive to an LSD1 inhibition therapy.

[0316] FIG. 8: Signature scores obtained by PLS analysis using the first principal component according to Example 8. Cell lines with score_3>0.45 are more likely to be sensitive to an LSD1 inhibition therapy.

[0317] FIG. 9: in vivo tumor growth inhibition of (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine in classic (C) cell line H-510A.

[0318] FIG. 10: Heat Map showing mRNA expression (as z-scores) patterns in SCLC patient samples.

EXAMPLES

[0319] The following examples 1 to 9 are provided for illustration of the invention. They should not be considered as limiting the scope of the invention, but merely as being representative thereof.

[0320] Methods

[0321] Expression Data

[0322] Expression data were obtained from whole transcriptomic RNA sequencing (RNA-seq) by Illumina, Inc. (San Diego, Calif.). The Illumina HiSeq machine generates raw base calls in reads of 50 or 100 bp length, which are subjected to several data analysis steps. The RNA-seq is conducted at 40 to 50 million reads per sample. This number provides relatively high sensitivity to detect low-expressed genes while allowing for cost-effective multiplexing of samples. RNA is prepared by standard kits and RNA libraries by polyA TruSeq Illumina kits. 100 ng of mRNA per cell line is used for each RNA-seq reaction. A number of quality control procedures are applied to the RNA-seq data for each sample. The Illumina HiSeq software reports the total number of clusters (DNA fragments) loaded in each lane, percent passing sequencing quality filters (which identifies errors due to overloading and sequencing chemistry), a phred quality score for each base of each sequence read, overall average phred scores for each sequencing cycle, and overall percent error (based on alignment to the reference genome). For each RNA-seq sample, the percentage of reads that contain mitochondrial and ribosomal RNA is calculated. The FASTQC package is used to provide additional QC metrics (base distribution, sequence duplication, overrepresented sequences, and enriched kmers) and a graphical summary. Raw reads were aligned against the human genome (hg19) using GSNAP and recommended options for RNASeq data. In addition to the genome sequence, GSNAP is given a database of human splice junctions and transcripts based on Ensembl v73. Resulting SAM files are then converted to sorted BAM files using Samtools. Gene expression values are calculated both as RPKM values following (Mortazavi et al..sup.107) and as read counts. Normalized read counts were obtained using the R package DESeq2.

[0323] Copy Number Variations (CNV)

[0324] To obtain copy number variation data genomic DNA were extracted and array CGH analysis were performed by Roche NimbleGen (Madison, Wis.) using their standard protocols. Normalized signal intensities and copy number changes were obtained using the segMNT algorithm. CGH microarrays contain isothermal, 45- to 85-mer oligonucleotide probes that are synthesized directly on a silica surface using light-directed photochemistry (Selzer et al..sup.108). The genomic DNA samples are randomly fragmented into lower molecular weight species and differentially labeled with fluorescent dyes.

[0325] Principal Component Analysis

[0326] Principal component analysis was carried out with Simca v 14 (Umetrics AB, Ume∪, Sweden).

[0327] Differential Gene Expression Analysis

[0328] Differential gene expression analysis used to generate data in Table 9 was carried out with the R package DESeq2 starting from raw read counts for 19 cell lines.

[0329] Heat Maps of Cell Lines

[0330] Heat maps of cell lines (as in FIG. 4 and FIG. 5) were generated using GenePattern v 3.9.4 (Reich M. et al..sup.109) to visualize color-coded gene expression levels. GenePattern takes in input the logarithm of normalized read counts (as reported in Table 10) plus one and applies a row-based normalization which consists of calculating z-scores for all expression levels of a given gene across the cell lines tested. A z-score of 0 corresponds to the mean of a distribution, and positive or negative value represent normalized gene expression levels above or below the mean, respectively. The color mapping capped the z-score range from −1.5 to +1.5, that is, z-scores above+1.5 are displayed in black and z-scores below −1.5 are in white. Intermediate values are displayed in different shades of gray. Gene Pattern performs hierarchical clustering to group and sort cell lines based on their gene expression profile.

Example 1—Differential Activity of LSD1 Inhibitors in SCLC Cell Lines

[0331] The differential activity of (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine in SCLC Cell Lines is presented in FIGS. 1A-1B. The activity of (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine was assessed in vitro in a panel of SCLC cell lines treated for 7 days. Cell lines characterized as “classic” neuroendocrine lineages, such as NCI-H1876 (FIG. 1A) and NCI-H510 (FIG. 1B), maintained a high level of sensitivity to (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine.

[0332] The compound potency determination was performed by culturing small cell lung cancer cell lines for 7 days at 37 degrees C. at 5% CO.sub.2 in humidified incubators in the presence of 15-serially dilutions (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine at the indicated concentration. Each of the cell lines was propagated and tested in distinct optimized media as recommended by ATCC or cell line source.

[0333] Cells were thawed from a liquid nitrogen preserved state. Once cells have been expanded and divided at their expected doubling times, screening was started. Cells were seeded in growth media in black 384-well tissue culture treated plates at 500 cells per well (except where noted in Analyzer). Cells were equilibrated in assay plates via centrifugation and placed in incubators attached to the Dosing Modules at 37° C. for twenty-four hours before treatment. At the time of treatment, a set of assay plates (which did not receive treatment) were collected and ATP levels were measured by adding ATPLite (Perkin Elmer). These Tzero (T0) plates were read using ultra-sensitive luminescence on Envision Plate Readers. Treated assay plates were incubated with (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine for one-hundred and sixty-eight hours. After one-hundred and sixty-eight hours, plates were developed for endpoint analysis using ATPLite. All data points were collected via automated processes; quality controlled; and analyzed using Horizon CombinatoRx proprietary software. Assay plates were accepted if they passed the following quality control standards: relative luciferase values were consistent throughout the entire experiment, Z-factor scores were greater than 0.6, untreated/vehicle controls behaved consistently on the plate.

[0334] Horizon Discovery utilizes Growth Inhibition (GI) as a measure of cell viability. The cell viability of vehicle was measured at the time of dosing (T0) and after one hundred and sixty-eight hours (T168). A GI reading of 0% represents no growth inhibition—cells treated with (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine and T168 vehicle signals were matched. A GI 100% represents complete growth inhibition—cells treated by compound and T0 vehicle signals were matched. Cell numbers have not increased during the treatment period in wells with GI 100% and may suggest a cytostatic effect for compounds reaching a plateau at this effect level. A GI 200% represents complete death of all cells in the culture well. Compounds reaching an activity plateau of GI 200% were considered cytotoxic. Horizon CombinatoRx calculates GI by applying the following test and equation:

[00004]$\mathrm{If}<V_0:100*\left(1-\frac{T-V_0}{V_0}\right)$$\mathrm{If}\ge V_0:100*\left(1-\frac{T-V_0}{V-V_0}\right)$

[0335] where T is the signal measure for a test article, V is the vehicle-treated control measure, and Vo is the vehicle control measure at time zero. This formula was derived from the Growth Inhibition calculation used in the National Cancer Institute's NCI-60 high throughput screen.

Example 2—Syneristic Effects of LSD1 Inhibitors Combined with Other Active Pharmaceutical Ingredients

[0336] Table 4 provides a heat map of synergy scores, the values indicating the strength of the synergistic effects. Synergy scores >6.4 were considered significant and warranted further validations. Cell lines that exhibited particular single agent response to (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine (NCI-H187, NCI-H1417, NCI-H1876, NCI-H510) were sensitized to the effects of a broad range of drug classes including HDAC and BET inhibitors, DNA alkylating agents, topoisomerase inhibitors, anti-mitotic agents, Aurora kinase inhibitors, BCL2 family inhibitors and Chk inhibitors. Similar leves of synergy were not uniformly observed in cell lines that were insensitive to (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine (NSCLC cell lines and SCLC cell line NCI-H1048, NCI-H446 and SBC-5). These data suggest the LSD1 inhibition may broadly sensitize SCLC cell lines to intervention by chemotherapeutics and targeted therapies.

[0337] Cells were thawed from a liquid nitrogen preserved state and expanded until they reached their expected doubling times. Each of the cell lines was propagated and tested in distinct optimized media as recommended by ATCC or cell line source.

[0338] Cells were seeded in 384-well assay plates at assigned densities (determined in the optimization phase). Cells were then equilibrated via centrifugation in incubators attached to the Dosing Modules for 24 hours before (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine pre-treatment. Assay plates were then treated with the assigned concentrations of (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine (determined in the optimization phase).

[0339] At the time of (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine treatment, a set of assay plates (which do not compound treatment) were collected and ATP levels measured by adding ATPLite (Perkin Elmer). These Tzero (T0) plates were read on Envision Plate Readers to measure luminescence. Treated assay plates were incubated with (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine for 96 hours before treatment with the second compound. After this time, assay plates were then treated with 8 point serial dilutions of enhancer compound in a 9×9 extended matrix and harvested after another 72 hours incubation. After a total of 168 hours from the initial (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine treatment time, plates were developed for endpoint analysis using ATPLite. All data points were collected via automated processes; quality controlled; and analyzed using Horizon CombinatoRx proprietary software. Assay plates were accepted if they passed the following quality control standards: relative luciferase values were consistent throughout the entire experiment, Z-factor scores were greater than 0.6, untreated/vehicle controls behaved consistently on the plate.

[0340] Horizon Discovery utilizes Growth Inhibition (GI) as a measure of cell viability. The cell viability of vehicle is measured at the time of dosing (T0) and after one hundred and sixty-eight hours (T168). A GI reading of 0% represents no growth inhibition—cells treated with compound and T168 vehicle signals are matched. A GI 100% represents complete growth inhibition—cells treated by compound and T0 vehicle signals are matched. Cell numbers have not increased during the treatment period in wells with GI 100% and may suggest a cytostatic effect for compounds reaching a plateau at this effect level. A GI 200% represents complete death of all cells in the culture well. Compounds reaching an activity plateau of GI 200% are considered cytotoxic. Horizon CombinatoRx calculates GI by applying the following test and equation:

[00005]$\mathrm{If}<V_0:100*\left(1-\frac{T-V_0}{V_0}\right)$$\mathrm{If}\ge V_0:100*\left(1-\frac{T-V_0}{V-V_0}\right)$

[0341] where T is the signal measure for a test article, V is the vehicle-treated control measure, and Vo is the vehicle control measure at time zero. This formula is derived from the Growth Inhibition calculation used in the National Cancer Institute's NCI-60 high throughput screen.

[0342] Loewe additivity model is dose-based and applies only to the activity levels achieved by the single agents. Loewe Volume is used to assess the overall magnitude of the combination interaction in excess of the Loewe additivity model. Loewe Volume is particularly useful when distinguishing synergistic increases in a phenotypic activity (positive Loewe Volume) versus synergistic antagonisms (negative Loewe Volume). When antagonisms are observed, as in the current dataset, the Loewe Volume should be assessed to examine if there is any correlation between antagonism and a particular drug target-activity or cellular genotype. This model defines additivity as a non-synergistic combination interaction where the combination dose matrix surface should be indistinguishable from either drug crossed with itself. The calculation for additivity is:

I.sub.Loewe that satisfies (X/X.sub.1)+(Y/Y.sub.1)=1

[0343] where X.sub.1 and Y.sub.1 are the single agent effective concentrations for the observed combination effect I. For example, if 50% inhibition is achieved separately by 1 mM of drug A or 1 mM of drug B, a combination of 0.5 mM of A and 0.5 mM of B should also inhibit by 50%.

[0344] To measure combination effects in excess of Loewe additivity, the Horizon Discovery platform was utilized. This method devised a scalar measure to characterize the strength of synergistic interaction termed the Synergy Score. The Synergy Score is calculated as:

Synergy Score=log f.sub.X log f.sub.Y Σ max(0,I.sub.data)(I.sub.data−I.sub.Loewe)

[0345] The fractional inhibition for each component agent and combination point in the matrix is calculated relative to the median of all vehicle-treated control wells. The Synergy Score equation integrates the experimentally-observed activity volume at each point in the matrix in excess of a model surface numerically derived from the activity of the component agents using the Loewe model for additivity. Additional terms in the Synergy Score equation (above) are used to normalize for various dilution factors used for individual agents and to allow for comparison of synergy scores across an entire experiment.

[0346] Activity over Loewe additivity is most easily calculated using a simple volume score, where V.sub.Loewe=log f.sub.X log f.sub.Y Σ(I.sub.data−I.sub.Loewe), summed over all non-single agent concentration pairs and where log f.sub.X,Y are the natural logarithm of the dilution factors used for each single agent. This effectively calculates a volume between the measured and Loewe additive response surfaces, corrected for varying dilution factors. This volume score emphasizes the overall synergistic or antagonistic effect of the combination, thus minimizing the effects of outlying data spikes and identifying combinations with a robust synergy across a wide range of concentrations and at high effect levels. V.sub.Loewe is positive for mostly synergistic combinations and negative for antagonism. The uncertainty σ.sub.V can be calculated based on the measured errors σ.sub.1 and standard error propagation.

[0347] “Synergy Score” S=f.sub.cov ln f.sub.X ln f.sub.Y Σ max(0,I.sub.data) max(0,I.sub.data−I.sub.Loewe), which is a positive-gated, inhibition-weighted volume over Loewe additivity. This provides an additional prioritization favoring combinations whose synergy occurs at high effect levels, ignoring antagonistic portions of the response surface. Here f.sub.X,Y are the dilution factors used for each single agent and the coverage factor f.sub.cov accounts for missing data, scaling the score up by the ratio of total/tested combination dose matrix points. S is always positive, and its uncertainty as can be calculated based on the measured errors σ.sub.1 and standard error propagation. An alternative to the synergy score is the “Hit Score” H=f.sub.COV log f.sub.X log f.sub.Y Σ max(0,I.sub.data) max(0,I.sub.data−I.sub.HSA), which refers to the HSA model. The key distinctions between S and H lie in the different underlying models and also in how the single agents are used in the model calculations. In the Chalice Analyzer, the HSA model is calculated directly from the single agent responses at corresponding concentrations, while the Loewe additive model is derived from the sigmoidal fits to the single agent response curves.

[0348] To prioritize hits, distributions of a score (S or H) and its error can be used to define an appropriate selection cutoff. For example, combinations with S>3σ.sub.s are “individually significant” at ˜99% confidence, assuming normal errors. To estimate systematic experimental errors that are not tested by replicate plates, the distribution of synergy scores for any drug-with-itself combinations acquired during the experiment can be used to determine a plausible range for non-detections. Alternatively, the score distribution for the whole experiment can be used to identify outliers at a chosen confidence level.

Example 3—In Vitro Synergistic Effects for SCLC of LSD1 Inhibitors Combined with Other Active Pharmaceutical Ingredients

[0349] Table 5 provides a heat map of synergy scores, the values indicating the strength of the synergistic effects. Synergy scores >6.4 were considered significant and warranted further validations. A select panel of drug classes were prioritized for further evaluation in an expanded panel of SCLC cell lines based upon the level of synergy observed and potential clinical use of the compounds in a therapeutic regimen in SCLC. Targeted therapies and chemical probes that inhibited the epigenetic regulator BET (particularly Brd4) ((+)-JQ1, CPI-203, MS 436, GSK1324726A, GSK1210151A and OTX015) and anti-apopotic regulator BCL2 (Obatoclax, ABT-199, ABT-737, and TW-37) were highly synergistic with LSD1 inhibition. Synergy was also observed with other epigenetic regulators, EZH2 (e.g. CPI-169, EPZ005687, EPZ-6438, GSK126, GSK343) and DOT1L (e.g. EPZ-5676, SGC 0946), albeit at a low level compared to BET (particularly BRD4) and BCL2 inhibitors.

[0350] Inhibitors of the Notch, Hedgehog or Smoothened pathway were not synergistic with (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine treatment, indicating that LSD1 inhibition sensitizes cell lines to select compound and drug classes that include HDAC and BET inhibitors, DNA alkylating agents, topoisomerase inhibitors, anti-mitotic agents, Aurora kinase inhibitors, BCL2 family inhibitors, EZH2, DOT 1L and Chk inhibitors.

[0351] Cells were treated and data analyzed as described in Example 2 above.

TABLE-US-00004 TABLE 4 Synergistic effects of combinations of (trans)-N1-((1R,2S)-2- phenylcyclopropyl)cyclohexane-1,4-diamine with a panel of suitable active pharmaceutical ingredients (API). Synergy Scores in NSCLC Synergy Scores in SCLC CAL- NCI- NCI- NCI- NCI- NCI- NCI- NCI- SBC- 2nd API: 12T A549 H441 H187 H1417 H1876 H510 H1048 H446 5 (+)-JQ1 0.51 0.07 0.12 11.29 39.59 18.65 11.88 0.68 1.74 0.15 ABT-263 0.01 0.89 0.75 14.31 19.10 23.63 18.16 14.92 0.29 3.35 ABT-888 0.22 0.02 0.04 2.26 3.26 8.62 4.10 0.40 0.05 0.27 Belinostat 0.42 0.52 0.66 14.35 19.12 9.05 15.72 1.61 3.21 1.84 Benda- 0.10 0.02 0.01 8.19 5.20 12.38 1.12 2.01 0.10 0.37 mustine BGJ398 0.49 0.16 0.17 9.26 9.66 19.85 4.42 9.55 17.33 0.69 Carbo- 0.06 0.03 1.27 7.49 10.45 10.90 9.96 1.47 1.58 0.22 platin CGK 733 0.39 3.07 0.27 3.35 0.91 4.59 3.03 1.21 5.16 0.78 Cisplatin 0.85 0.04 1.80 4.91 15.96 5.34 9.89 3.72 1.56 0.48 Docetaxel 1.42 1.21 1.82 13.67 11.72 19.92 31.60 0.50 3.33 0.88 Doxo- 0.76 0.45 6.86 6.85 33.43 3.74 15.96 3.13 4.93 1.20 rubicin Erlotinib 0.43 0.18 0.70 3.17 2.87 5.21 3.29 0.50 1.99 0.32 Etoposide 0.15 0.00 5.83 8.28 13.22 11.17 11.55 1.04 0.12 0.54 Fluoro- 0.07 0.00 0.68 1.97 1.63 4.91 0.62 0.80 0.09 0.42 uracil Gem- 0.58 0.93 6.78 19.43 15.18 2.74 23.37 1.11 1.01 0.33 citabine GSK-J1 0.01 0.07 0.09 2.14 2.26 10.21 3.80 0.23 0.03 0.85 Irinotecan 0.14 0.02 3.45 13.74 21.62 8.75 17.17 2.00 1.16 0.48 Lapatinib 1.23 0.00 0.36 1.23 7.00 5.69 20.64 0.75 20.44 0.03 LY- 1.46 0.20 2.22 6.47 1.05 11.98 14.09 1.47 0.37 0.87 2603618 Menadione 0.01 0.06 0.02 3.89 4.18 8.81 3.63 0.45 0.06 0.52 Metho- 0.25 1.95 0.29 1.45 1.82 4.20 1.75 0.62 0.74 0.34 trexate MLN8237 0.76 0.08 0.81 7.60 9.15 23.31 28.91 1.02 1.16 0.51 Nutlin-3A 0.11 0.04 0.36 1.56 2.56 7.24 1.63 0.29 0.36 0.48 Paclitaxel 0.83 1.11 1.33 12.68 14.10 20.99 28.03 2.26 2.64 1.83 Pano- 1.11 1.34 0.45 10.41 17.19 2.60 12.47 0.58 15.30 1.13 binostat Peme- 2.12 0.01 2.76 4.10 2.54 4.27 0.86 0.20 0.18 0.11 trexed PF- 0.15 0.07 0.16 0.99 0.81 6.68 1.11 0.56 0.01 0.49 04217903 Temozo- 0.04 0.00 4.12 2.52 21.79 8.49 5.43 0.81 0.37 0.10 lomide Topotecan 0.08 1.92 1.98 11.49 2.24 6.23 23.12 3.73 6.87 0.20 Vincristine 0.49 1.23 0.04 18.07 27.52 7.80 10.01 0.08 2.31 0.15

TABLE-US-00005 TABLE 5 in vitro Synergistic Effects in SCLC cell lines of combinations of (trans)-N1- ((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine with a panel of suitable active pharmaceutical ingredients (API). Synergy Scores of SCLC cell lines NCI- NCI- NCI- NCI- NCI- NCI- SHP- NCI- DMS- NCI- 2nd API: H1876 H510 H1417 H187 H2171 H69 77 H526 114 H446 (+)-JQ1 9.15 24.21 7.87 2.83 8.69 6.44 0.45 8.06 1.55 4.03 ABT-199 4.63 13.46 11.01 1.95 0.04 5.07 0.33 3.47 0.03 0.07 ABT-263 21.80 10.70 12.21 2.51 6.19 8.22 3.59 4.18 0.26 1.41 ABT-737 19.84 11.80 25.26 4.17 3.15 4.01 1.88 11.44 0.42 1.73 BMS-906024 0.22 0.00 0.33 0.45 0.19 8.46 0.03 5.80 0.02 0.12 CPI-169 8.50 5.76 0.38 0.09 0.00 0.03 0.03 2.46 0.06 0.31 CPI-203 15.25 25.41 13.47 1.24 6.83 8.68 0.37 11.09 1.58 0.58 EPZ005687 4.57 10.77 0.46 0.10 0.48 0.80 0.23 3.66 0.14 0.35 EPZ-5676 6.68 19.72 0.87 0.11 0.05 0.70 0.12 1.20 0.00 0.24 EPZ-6438 8.18 9.06 0.31 0.03 0.01 0.68 0.02 1.82 0.04 0.00 FLI 06 6.95 0.83 0.28 1.71 1.76 4.01 6.32 2.18 5.48 3.29 GDC-0449 5.64 3.05 0.25 0.16 0.10 0.90 0.24 1.76 0.01 0.16 GSK1210151A 11.42 15.35 15.13 1.77 10.11 10.38 0.66 12.65 0.67 0.32 GSK126 10.80 12.69 0.41 0.35 0.80 1.63 0.02 3.20 0.51 0.41 GSK1324726A 15.98 21.32 21.84 2.66 9.12 10.17 0.72 12.18 1.22 1.25 GSK343 0.81 4.25 0.24 0.15 0.36 0.81 0.13 2.48 0.05 0.01 LDE225 0.57 5.81 1.01 0.73 0.34 0.78 0.53 2.17 0.08 0.02 LY-3039478 0.19 0.00 0.21 0.47 0.30 4.94 0.12 7.90 0.00 0.28 MK-0752 0.69 0.71 0.15 0.32 0.17 3.88 0.03 6.03 0.04 0.43 MS 436 5.74 5.52 12.37 0.06 1.37 3.82 0.23 15.84 0.18 0.42 Obatoclax 6.11 6.31 2.36 0.81 0.61 6.01 4.76 8.19 1.87 2.57 OTX015 14.88 11.52 14.26 0.88 5.30 8.65 0.78 16.72 0.10 0.44 PF-3084014 0.25 0.03 0.02 0.10 1.02 1.53 0.19 8.64 0.02 1.31 SGC 0946 6.81 24.63 0.28 0.12 0.07 4.50 0.06 1.80 0.00 0.06 Taladegib 1.64 2.50 0.19 0.09 0.11 2.35 0.19 3.66 0.10 0.01 TW-37 0.86 6.95 8.08 0.57 0.72 2.23 0.45 3.60 0.59 0.28

Example 4. LSD1 Inhibitors Improve Potency and Duration of SCLC-SOC In Vivo

[0352] In vivo, the effects of (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine treatment synergized with both etoposide and carboplatin to induce a cytotoxic response as can be seen from FIG. 2A. In the clinic the one standard of care (SOC) for SCLC is to combine etoposide and carboplatin.

[0353] In vivo, the combination of etoposide and carboplatin promotes rapid tumor regression during the dosing period in the NCI-1H526 model as can be seen from FIG. 2B. The addition of (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine to SOC improved the duration of tumor regression and significantly delayed tumor regrowth by 30 days after the suspension of dosing. Together the data suggest that LSD1 inhibition can sensitize cells to select chemotherapeutics and targeted agents in vitro and in vivo.

[0354] NCI-H526 Models:

[0355] 8-12-week old nu/nu mice were injected with 1×107 H526 cells or 5×106 SHP-77 resuspended in 100 μL of 1:1 mixture of Matrigel® and PBS. Tumors were staged at 100-150 mm.sup.3 animals and distributed into dosing groups. (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine was administered at a dose of 40 μg per kg (upk) five days on/two days off (5/2) for three weeks. Etoposide was administered i.p at a dose of 5 mg per kg (mpk) daily for five days (qdx5). Carboplatin was administered i.p. at a dose of 100 mpk weekly for three weeks (qwkx3). In combination with etoposide and carboplatin, (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine was administered at a dose of 20 upk five days on/two days off for three weeks. Tumor volume was measure biweekly using a digital caliber. The endpoint of the experiment was a tumor volume of 1000 mm.sup.3 or 90 days, whichever came first. Statistical analysis was performed using unpaired t-test and Gehan-Breslow-Wilcoxon test.

Example 5. Cell Response to LSD1 Inhibition

[0356] The compound potency determination was performed by culturing 19 small cell lung cancer cell lines (of various tumor origins) for 4 days at 37 degrees C. at 5% CO.sub.2 in humidified incubators in the presence of serially diluted (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine bis-hydrochloride.

[0357] As a positive control for cytotoxicity the Hsp90 inhibitor 17-N-allylamino-17-demethoxygeldanamycin (17-AAG, a geldanamycin analogue) was used as positive control in serial dilution. Each of the cell lines was propagated and tested in distinct optimized media as recommended by ATCC or cell line source.

[0358] Small cell lung cancer cell lines can be categorized as “classic” or “variant”, based on their enzymatic activities, cellular morphologies, and growth phenotypes (Desmond et al.sup.110. Shoemaker R.H..sup.111). Classic cells lines express elevated levels of L-dopa decarboxylase, bombesin-like immunoreactivity, neuron-specific enolase, and the brain isozyme of creatine kinase; variant cell lines continue to express neuron-specific enolase and the brain isozyme of creatine kinase, but have undetectable levels of L-dopa decarboxylase and bombesin-like immunoreactivity. Unlike classic cell lines, some variant cell lines are amplified for and have increased expression of the c-myc (MYC)oncogene.

[0359] Some cell lines exhibit features specific to both a classic and variant subtype. For example, SHP-77 has biochemical properties of classic SCLC (e.g. elevated levels of L-dopa decarboxylase and bombesin-like immunoreactivity) but the morphology of a variant. According to the literature, SHP-77 is considered classic based on its biochemical profile but variant based on its morphology and growth characteristic.

[0360] For NCI-H2029 and SBC-5 no subtype is reported in literature however their transcriptomic profile (mRNA expression levels of DDC/GRP) clearly shows their class membership which is provided in brackets in Table 6.

[0361] Depending on their responses to (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine bis-hydrochloride, cell lines are classified as either “sensitive” [S], defined as having EC50<0.05 μM, or “resistant”, defined as having EC50>=0.05 μM [R].

[0362] Cell-based response to (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine bis-hydrochloride was greater in classic SCLC cell lines compared to variant SCLC cell lines (p-value 0.0055 Table 6). Out of the 19 SCLC cell lines tested, 9 out of 11 classic cell lines [C] are sensitive [S], and 7 out of 8 variant cell lines [V] are resistant [R] (Table 7).

[0363] The variant and classic subtypes predict response to an LSD1 inhibitor therapy with a sensitivity of 82% and specificity of 88%.

[0364] Higher copy number variations (CNV) in the MYC gene (Ensemble Gene ID: ENSG00000136997) are associated with small cell lung cancer of variant subtype (V) (Am J Pathol. 1988 July; 132(1): 13-17). Indeed, among the 19 cell lines here described, high copy number variations of the MYC gene (CNV>>2) were found exclusively in cell lines with a variant subtype (NCI-H2171, NCI-H446, NCI-H82, see Table 6). Furthermore, all three cell lines with high copy number variations of MYC were resistant to LSD1 inhibition, indicating that the presence of MYC amplification can predict resistance (R) to an LSD1 inhibition therapy.

[0365] Principal component analysis carried out from RNA-seq data for the cell lines of Table 6 surprisingly revealed that classic and variant SCLC cell lines form distinct clusters. (FIG. 3).

TABLE-US-00006 TABLE 6 Cell-based response to (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cy- clohexane-1,4-diamine bis-hydrochloride in classic SCLC cell lines [C] as compared to variant SCLC cell lines [V]. SubType Max. EC50 Sensitivity to MYC Cell Line Lit. Response (%) (μM) LSD1 inh. CNV NCI-H1876 C 145 4.32 × 10.sup.−5 S 0.81 NCI-H69 C 44 5.85 × 10.sup.−4 S 1.14 NCI-H510A C 68 3.15 × 10.sup.−4 S 2.56 NCI-H146 C 48 1.00 × 10.sup.−4 S 1.24 NCI-H187 C 61 1.36 × 10.sup.−4 S 1.12 NCI-H2081 C 10 9.20 × 10.sup.−4 S 2.22 NCI-H345 C 7 2.75 × 10.sup.−5 S 1.26 NCI-H526 V 35 6.32 × 10.sup.−4 S 1.07 NCI-H748 C 13 3.00 × 10.sup.−4 S 1.05 NCI-H1417 C 77 3.02 × 10.sup.−4 S NA DMS-114 V 0  >5 × 10.sup.−2 R 1.21 NCI-H1048 V 27  >5 × 10.sup.−2 R 0.98 NCI-H2029 (C) 0  >5 × 10.sup.−2 R 1.23 NCI-H2171 V 0  >5 × 10.sup.−2 R 7.46 NCI-H2227 C 0  >5 × 10.sup.−2 R 0.81 NCI-H446 V 7  >5 × 10.sup.−2 R 6.72 NCI-H82 V 0  >5 × 10.sup.−2 R 9.44 SHP-77 V(C) 0  >5 × 10.sup.−2 R 1.36 SBC-5 (V) 23  >5 × 10.sup.−2 R 1.21

TABLE-US-00007 TABLE 7 Contingency matrix showing the number of classic and variant cell lines that are sensitive or resistant to an LSD1 inhibition therapy Classic [C] Variant [V] Sensitive [S] 9 1 Resistant [R] 2 7

Example 6. Gene panel to predict response to LSD1 inhibition

[0366] Differential gene expression analysis between two resistant cell lines that have features of a classic subtype (SHP-77 and NCI-2029) and classic and variant cell lines which are sensitive (NCI-H1876, NCI-H69, NCI-H510A, NCI-1H146, NCI-1H187, NCI-H2081, NCI-H345, NCI-H526, NCI-H748) interestingly revealed that lower mRNA expression levels of HOXA10 correlate with resistance to an LSD1 inhibition therapy (Table 8). This suggests that low levels of HOXA10 mRNA may predict resistance to an LSD1 inhibition therapy even in the presence of a classic phenotype.

[0367] A predictive mRNA expression signature of response to an LSD1 inhibition therapy was defined by selecting top differentially expressed genes between classic and variant cell lines (Table 9). Based on adjusted p-values, DDC (adjusted p-value 4.37E−23), which encodes the enzyme L-dopa decarboxylase, and GRP (adjusted p-value 5.19E−14), which encodes bombesin-like immunoreactivity peptides rank as second and sixth most differentially expressed genes. The most differentially expressed gene is ASCL1 (adjusted p-value 2.6E−23). ASCL1 is a transcription factor required for proper development of pulmonary neuroendocrine cells, and is essential for the survival of a majority of lung cancers (Augustyn et al..sup.112).

[0368] As discussed in Example 5 above, MYC amplification can predict resistance to LSD1 inhibition therapy.

[0369] Table 10 lists normalized read counts of DDC, GRP, and ASCL1 across the 19 cell lines of Table 6 described while Table 11 lists the corresponding z-scores.

[0370] The heat map of FIG. 4 visually shows that sensitive cell lines can be distinguished from resistant cell lines based on mRNA expression levels of genes listed in Table 9, and based on expression levels of HOXA10 and copy number variations of MYC.

TABLE-US-00008 TABLE 8 Principal component analysis for HOXA10 carried out from RNA-seq data for selected cell lines (*http://www.ensembl.org/, Cunningham F. el al..sup.106) log2Fold Ensembi Gene ID* Gene baseMean Change pvalue ENSG00000253293 HOXA10 2717.58 8.21 7.45E−023

TABLE-US-00009 TABLE 9 Genes sorted according to pvalue obtained through principal component analysis carried out from RNA-seq data for selected cell lines (*http://www.ensembl.org/, Cunningham F. et al..sup.106). Ensembl Gene ID* Gene baseMean log2Fold Change pvalue ENSG00000139352 ASCL1 43665.33 6.82 2.62E−023 ENSG00000132437 DDC 15817.8 6.42 4.37E−023 ENSG00000086548 CEACAM6 210.89 6.34 1.23E−017 ENSG00000188306 LRRIQ4 90.81 5.1 4.61E−016 ENSG00000131910 NR0B2 600.58 6.35 5.15E−015 ENSG00000134443 GRP 6711.45 6.52 5.19E−014 ENSG00000105388 CEACAM5 1788.17 6.22 9.23E−014 ENSG00000125285 SOX21 523.59 5.88 2.29E−013 ENSG00000167332 OR51E2 3047.56 6.39 3.37E−013 ENSG00000166562 SEC11C 36139.18 3.33 5.01E−013 ENSG00000164929 BAALC 1833.4 4.33 1.66E−012 ENSG00000141519 CCDC40 2309.83 2.26 2.07E−012 ENSG00000169213 RAB3B 28247.78 3.64 2.80E−012 ENSG00000091844 RGS17 2783.99 3.2 3.72E−012 ENSG00000164163 ABCE1 13643.12 −1.08 4.99E−012 ENSG00000157557 ETS2 11829.42 3.06 5.19E−012 ENSG00000197599 CCDC154 1198.98 4.61 7.21E−012 ENSG00000077327 SPAG6 767.39 5.34 7.85E−012 ENSG00000005421 PON1 334.17 5.15 1.53E−011 ENSG00000002933 TMEM176A 3224.04 5.38 7.65E−011 ENSG00000175262 C1orf127 596.15 5.04 1.19E−010 ENSG00000073792 IGF2BP2 2414.53 −5.17 1.28E−010 ENSG00000115461 IGFBP5 86866.7 4.41 1.38E−010 ENSG00000162981 FAM84A 4954.8 3.93 1.45E−010 ENSG00000125798 FOXA2 4530.46 5.12 1.71E−010

TABLE-US-00010 TABLE 10 Normalized read counts from mRNA expression levels. Cell Line ASCL1 DDC GRP HOXA10 NCI-H1417 42666.4 16161.1 10935.2 3327.72 NCI-H1876 34116.3 986.718 43.7461 2779.5 NCI-H69 19902.1 25773.6 3256.24 4271.2 NCI-H510A 79879.7 19456.3 27861 2730.12 NCI-H2227 4515.83 2005.02 645.86 2.59381 NCI-H2029 127171 39070.6 1800.43 10.0396 NCI-H146 59238.2 45308.8 426.015 2126.39 NCI-H187 71323.6 4363.62 130.681 2448.85 NCI-H2081 69670.9 29683.5 2.97459 3423.76 NCI-H345 81805.8 16935.7 30601.3 263.11 SHP-77 115523 71808.9 39002.6 4.72759 NCI-H748 122007 27938.7 12773.8 3940.53 DMS-114 59.1696 16.3227 12.242 1462.92 NCI-H1048 38.9626 90.2292 0 1168.88 NCI-H2171 1115.78 368.976 0 1248.61 NCI-H446 13.1805 32.0098 11.2976 2818.75 NCI-H82 577.05 486.304 9.30725 221.047 SBC5 4.51028 13.5308 0 617.908 NCI-H526 11.9576 38.2644 4.78305 4091.9

TABLE-US-00011 TABLE 11 Z-scores generated by GenePattern from normalized mRNA read counts. Cell Line ASCL1 DDC GRP HOXA10 NCI-H1417 0.63 0.69 1.09 0.67 NCI-H1876 0.57 −0.24 −0.34 0.6 NCI-H69 0.42 0.85 0.78 0.78 NCI-H510A 0.8 0.76 1.34 0.59 NCI-H2227 0.02 0 0.35 −2.31 NCI-H2029 0.93 0.99 0.62 −1.82 NCI-H146 0.72 1.04 0.25 0.48 NCI-H187 0.77 0.26 −0.06 0.54 NCI-H2081 0.77 0.9 −0.98 0.69 NCI-H345 0.81 0.71 1.36 −0.43 SHP-77 0.9 1.19 1.43 −2.11 NCI-H748 0.92 0.88 1.13 0.75 DMS-114 −1.16 −1.59 −0.66 0.32 NCI-H1048 −1.27 −1.04 −1.34 0.22 NCI-H2171 −0.36 −0.57 −1.34 0.25 NCI-H446 −1.55 −1.38 −0.68 0.6 NCI-H82 −0.54 −0.48 −0.73 −0.51 SBC5 −1.81 −1.65 −1.34 −0.06 NCI-H526 −1.58 −1.32 −0.88 0.76

Example 7. Neuroendocrine Gene Panel to Predict Response to LSD1 Inhibition

[0371] mRNA expression levels for a second set of genes according to Table 12 (NCAM1, NCAM2, NEUROD1, KRT8, EN12, AVP, OXT, SYP, CHGA, CHGB, SOX21, BCL2) that includes genes representative of a neuroendocrine phenotype and that are used as immunohistochemical markers for diagnosing lung neuroendocrine tumors are strongly downregulated in resistant cell lines DMS 114, SBC5, and NCI-1H1048, as illustrated in FIG. 5. This is an agreement with our hypothesis that an LSD1 inhibition therapy stops cellular growth in tumors of neuroendocrine origin.

[0372] Tables 13A and 14B list normalized read counts of the genes of Table 12 across the 19 cell lines of Table 6 described.

TABLE-US-00012 TABLE 12 Genes of the second neuroendocrine gene panel (*http://www.ensembl.org/, Cunningham F. et al..sup.106). Ensembi Gene ID* Gene ENSG00000149294 NCAM1 ENSG00000154654 NCAM2 ENSG00000162992 NEUROD1 ENSG00000170421 KRT8 ENSG00000111674 ENO2 ENSG00000101200 AVP ENSG00000101405 OXT ENSG00000102003 SYP ENSG00000100604 CHGA ENSG00000089199 CHGB ENSG00000125285 SOX21 ENSG00000171791 BCL2

TABLE-US-00013 TABLE 13A Normalized read counts from mRNA expression levels. Cell Line NCAM1 NCAM2 NEUROD1 KRT8 ENO2 AVP NCI- 52961.1 230.0 257.7 32261.1 32287.3 5.8 H1417 NCI- 12131.4 111.0 143.4 36460.8 37021.4 33.2 H1876 NCI- 53702.4 16861.8 295.0 28560.6 28765.0 18.6 H69 NCI- 21010.6 197.4 255.2 67662.7 11901.4 1.7 H510A NCI- 42956.2 32469.4 1273.6 181.6 35558.6 2.6 H2227 NCI- 37343.8 70.3 244.3 76401.1 22753.0 0.0 H2029 NCI- 39176.8 1929.1 173.4 50190.4 32430.6 5.5 H146 NCI- 47022.6 8.5 31.3 61809.4 32195.9 2.8 H187 NCI- 37569.1 1279.1 2427.3 26842.7 32137.5 0.0 H2081 NCI- 62260.5 131.6 96.7 46256.4 32848.5 45.6 H345 SHP- 21787.1 990.4 0.0 35148.0 8851.6 0.0 77 NCI- 21844.8 892.7 12.1 1508.8 44468.6 0.9 H748 DMS- 95.9 512.1 18.4 377.5 3260.5 0.0 114 NCI- 14740.2 760.8 0.0 12726.4 38304.4 0.0 H1048 NCI- 16524.2 35.4 60402.8 26223.8 212034.0 0.0 H2171 NCI- 79657.4 3747.0 19164.5 45.2 36229.5 0.0 H446 NCI- 20878.5 437.4 34283.3 27.9 22702.7 0.0 H82 SBC- 130.8 19026.6 9.0 640.5 160.1 0.0 5 NCI- 44561.3 0.0 23.9 38233.3 24912.5 0.0 H526

TABLE-US-00014 TABLE 13B Normalized read counts from mRNA expression levels. Cell Line OXT SYP CHGA CHGB SOX21 BCL2 NCI- NA 6220.2 44388.5 11152.1 20.4 6170.7 H1417 NCI- 4.2 13216.2 7061.0 3968.7 1201.4 4126.7 H1876 NCI- 9.5 10950.9 16527.4 52724.6 20.9 10853.4 H69 NCI- 1.8 9116.9 22660.3 20808.2 79.1 27378.7 H510A NCI- 0.0 19962.0 11537.3 14927.4 2.6 1136.1 H2227 NCI- 0.0 8905.1 16397.9 5776.1 786.4 8687.6 H2029 NCI- 16.9 14940.0 22829.6 9597.3 660.2 10340.5 H146 NCI- 0.0 5696.0 23923.2 6804.0 264.2 14934.6 H187 NCI- 0.0 14334.6 79374.1 10934.6 44.6 2778.3 H2081 NCI- 0.0 9686.8 22971.1 7702.7 4953.5 39332.3 H345 SHP- 0.0 7861.2 47453.1 61511.4 480.6 7364.0 77 NCI- 2.7 19958.6 46176.5 7932.8 1408.9 11595.8 H748 DMS- 0.0 4299.0 1897.5 6161.8 10.2 185.7 114 NCI- 0.0 260.4 16.4 6.2 4.1 8063.2 H1048 NCI- 0.0 12335.6 9407.4 23159.9 0.0 1065.6 H2171 NCI- 0.0 7403.7 10702.6 5688.3 1.9 2398.9 H446 NCI- 4.7 31714.5 19382.3 7303.9 4.7 148.9 H82 SBC- 0.0 3642.1 311.2 203.0 4.5 306.7 5 NCI- 0.0 12538.8 9920.1 9877.0 0.0 16511.1 H526

Example 8. Signature Scores to Predictive Response to LSD1 Inhibition

[0373] Normalized expression levels (Norm_read_count) of ASCL1, DDC, GRP, and HOXA10 and MYC copy number variations (Copy_number_variation) were used to generate a gene signature to predict response to an LSD1 inhibition therapy as follows: A score was generated from the following equation, obtained by partial least square (PLS) analysis using the second principal component:

[00006]$\mathrm{Signature}\mathrm{Score}1=0.0900693+\mathrm{Norm\_read}\mathrm{\_count}\left(\mathrm{ASCL}1\right)\times 0.00000211296+\mathrm{Norm\_read}\mathrm{\_count}\left(\mathrm{DDC}\right)\times 5.36658\times {10}^{-7}+\mathrm{Norm\_read}\mathrm{\_count}\left(\mathrm{GRP}\right)\times 0.00000297345+\mathrm{Norm\_read}\mathrm{\_count}\left(\mathrm{HOXA}10\right)\times 0.000234721-\mathrm{Copy\_number}\mathrm{\_variation}\left(\mathrm{MYC}\right)\times 0.0537056.$

[0374] A Signature Score 1>0.5 predicts response to an LSD1 inhibition therapy (Fisher's exact test two-tailed p 0.0001, sensitivity 90%, specificity 100%) as depicted in FIG. 6.

[0375] Alternatively, a score was generated from the following equation, obtained by partial least square analysis using the first principal component:

[00007]$\mathrm{Signature}\mathrm{Score}2=0.483918+\mathrm{Norm\_read}\mathrm{\_count}\left(\mathrm{ASCL}1\right)\times 0.00000188066+\mathrm{Norm\_read}\mathrm{\_count}\left(\mathrm{DDC}\right)\times 0.00000188066+\mathrm{Norm\_read}\mathrm{\_count}\left(\mathrm{GRP}\right)\times 0.00000352033-\mathrm{Copy\_number}\mathrm{\_variation}\left(\mathrm{MYC}\right)\times 0.0407898.$

[0376] A Signature Score 2>0.5 predicts response to an LSD1 inhibition therapy (Fisher's exact test two-tailed p 0.0055, sensitivity 90%, specificity 77.8%) as depicted in FIG. 7.

[0377] Further, a score was generated from the following equation, obtained by partial least square analysis using the first principal component:

[00008]$\mathrm{Signature}\mathrm{Score}3=0.393569+\mathrm{Norm\_read}\mathrm{\_count}\left(\mathrm{ASCL}1\right)\times 0.00000182731+\mathrm{Norm\_read}\mathrm{\_count}\left(\mathrm{DDC}\right)\times 0.00000189964+\mathrm{Norm\_read}\mathrm{\_count}\left(\mathrm{GRP}\right)\times 0.00000342046.$

[0378] A Signature Score 3>0.45 predicts response to an LSD1 inhibition therapy (Fisher's exact test two-tailed p 0.0055, sensitivity 90%, specificity 77.8%) as depicted in FIG. 8.

[0379] A signature score above the reference level indicates a high likelihood of response to treatment with an LSD1 inhibitor, whereas a signature score below said level indicates a lower likelihood to respond to such treatment. A higher score is associated with higher mRNA expression of ASCL1, DDC, GRP, HOXA10, and with lower copy number variations in MYC.

Example 9. In Vivo Tumor Growth Inhibition

[0380] NCI-H510A Models:

[0381] 7-8-week old athymic nude mice animals were injected with 5×10.sup.6 H510A cells resuspended in 100 μL of 1:1 mixture of Matrigel® matrix (Corning Inc., Tewksbury/MA, C. S. Hughes et al..sup.113) and PBS. Tumors were staged at 200-300 mm.sup.3 animals and distributed into dosing groups. (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine was administered at a dose of 40 μg per kg (upk) five days on/two days off until end of study. Tumor volume was measure biweekly using a digital caliber. The study was concluded when mean tumor volume within control group reached 2000 mm.sup.3 or 28 days post-staging. Statistical analysis was performed using unpaired t-test.

[0382] NCI-H526 and SHP-77 Models:

[0383] 8-12-week old nu/nu mice were injected with 1×107 H526 cells or 5×10.sup.6 SHP-77 resuspended in 100 μL of 1:1 mixture of Matrigel® and PBS. Tumors were staged at 100-150 mm.sup.3 animals and distributed into dosing groups. (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine was administered at a dose of 40 upk five days on/two days off until end of study. Tumor volume was measure biweekly using a digital caliber. The study was concluded when mean tumor volume within control group reached 2000 mm.sup.3 or 28 days post-staging. Statistical analysis was performed using unpaired t-test.

[0384] The in vitro activity of the LSD1 inhibitor (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine translated into in vivo growth inhibition in the H510A xenograft model as shown in FIG. 9. Treatment of (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine in the “responsive signature” positive cell line H510A model resulted in a modest but measurable tumor growth inhibition of 34% compared to untreated controls after 21 days of dosing. These results suggest that the 15 gene response signature as previously defined may predict in vivo sensitivity to (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine. The in vivo activity of (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine has also been assessed in the “response signature positive” SHP-77 and “response signature negative” H526 xenografts to validate the predictability of the gene signature from in vitro results.

Example 10. Expression Patterns in SCLC Patient Samples

[0385] Gene expression patterns in a set of SCLC patient samples were found to be similar to those observed in SCLC cell lines (Example 6, FIG. 4), suggesting that use of LSD1 inhibitor response gene signature, particularly the use of the (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine response gene signature, may increase the likelihood of identifying patients who will clinically benefit from LSD1 inhibitor based therapies, particularly from (trans)-N1-((1R,2S)-2-phenylcyclopropyl)cyclohexane-1,4-diamine based therapies.

[0386] FIG. 10 provides a Heat Map showing mRNA expression (as z-scores) patterns in SCLC patient samples comprising the genes of Table 8, Table 9 and MYC. Higher z-scores correlate with better sensitivity.

[0387] The dataset used corresponds to the dataset EGAD00001000223 obtained from European Genome-phenome Archive (EGA) of the The European Bioinformatics Institute, Part of the European Molecular Biology Laboratory (EMBL-EBI, Hinxton/UK, https://www.ebi.ac.uk/ega/datasets/EGAD00001000223). The patient ID (SAMxxxxxx) corresponds to the ID used in this dataset. [0388] .sup.1 Shi et al. (2004) Cell 119:941 [0389] .sup.2 WO 2011/131697 A1 [0390] .sup.3 WO 2012/135113 A2 [0391] .sup.4 WO 2013/057322 A1 [0392] .sup.5 WO 2010/143582 [0393] .sup.6 WO 2011/131576 [0394] .sup.7 WO 2013/022047 [0395] .sup.8 WO 2013/025805 [0396] .sup.9 WO 2014/058071 [0397] .sup.10 WO 2014/084298 [0398] .sup.11WO 2014/085613 [0399] .sup.12 WO 2014/086790 [0400] .sup.13 WO2014/164867 [0401] .sup.14 WO 2014/194280 [0402] .sup.15 WO 2014/205213 [0403] .sup.16 WO 2015/021128 [0404] .sup.17 WO 2015/031564 [0405] .sup.18 WO 2015/089192 [0406] .sup.19 WO 2015/120281 [0407] .sup.20 WO 2015/123465 [0408] .sup.21 WO 2015/123437 [0409] .sup.22 WO 2015/123424 [0410] .sup.23 WO 2015/123408 [0411] .sup.24 WO 2015/134973 [0412] .sup.25 WO 2015/156417 [0413] .sup.26 WO 2016/004105 [0414] .sup.27 WO 2016/007722 [0415] .sup.28 WO 2016/007727 [0416] .sup.29 WO 2016/007731 [0417] .sup.30 WO 2016/007736 [0418] .sup.31 WO 2016/034946 [0419] .sup.32 WO 2016/037005 [0420] .sup.33 CN 105541806 [0421] .sup.34 WO 2016/123387 [0422] .sup.35 WO 2016/130952 [0423] .sup.36 CN 105924362 [0424] .sup.37 CN 105985265 [0425] .sup.38 WO 2016/161282 [0426] .sup.39 CN 106045862 [0427] .sup.40 CN 106045881 [0428] .sup.41 WO 2016/172496 [0429] .sup.42 WO 2016/177656 [0430] .sup.43 WO 2017/004519 [0431] .sup.44 WO 2017/027678 [0432] .sup.4 WO 2010/138588 A2 [0433] .sup.46 US 2007/0027135 A1 [0434] .sup.47 WO 2005/049594 A1 [0435] .sup.48 US 2006/0229289 A1 [0436] .sup.49 WO 2011/091213 A2 [0437] .sup.50 WO 2009/040517 A2 [0438] .sup.51 DD 159877 A1 [0439] .sup.52 WO 2006/000420 A1 [0440] .sup.53 WO 2012/129353 A1 [0441] .sup.54 ES 544159 A1 [0442] WO 2007/015632 A1 [0443] .sup.56 Rosenberg B. et al. Nature (1969) 222(5191):385-6 [0444] .sup.57 WO 2013/120104 A2 [0445] .sup.58 WO 2014/134583 A2 [0446] .sup.59 EP 253738 A1 [0447] .sup.60 DE 2510866 A1 [0448] .sup.61 Daigle S.R. et al. Cancer Cell (2011) 20(1):53-65 [0449] .sup.62 WO 2012/118812 A2 [0450] .sup.63 WO 2012/075381 A2 [0451] .sup.64 WO 2012/142504 A1 [0452] .sup.65 CH 514578 A [0453] .sup.66 WO 2013/178821 A1 [0454] .sup.67 U.S. Pat. No. 2,802,005 [0455] .sup.68 WO 2006/028958 A2 [0456] .sup.69 GB 2136425 [0457] .sup.70 WO 2011/140324 A1 [0458] .sup.71 WO 2011/054843 A1 [0459] .sup.72 WO 2011/140325 A1 [0460] .sup.73 WO 2012/052390 A1 [0461] .sup.74 U.S. Pat. No. 6,121,451 A [0462] .sup.7 WO 2011/143651 A1 [0463] .sup.76 WO 99/35146 A1 [0464] .sup.77 WO 2010/033481 A1 [0465] .sup.78 WO 2006/105262 A1 [0466] .sup.79 WO 2013/016081 A1 [0467] .sup.80 U.S. Pat. No. 2,331,725 [0468] .sup.81 WO 2002/081435 A1 [0469] .sup.82 WO 2008/063525 A1 [0470] .sup.83 WO 2012/116170 A1 [0471] .sup.84 US 2005/0282803 [0472] .sup.85 WO 2004/106328 A1 [0473] .sup.86 U.S. Pat. No. 5,712,274 A [0474] .sup.87 EP 253739 A1 [0475] .sup.88 WO 2002/022577 A2 [0476] .sup.89 EP 432677 A1 [0477] .sup.90 US 2007/0265272 A1 [0478] .sup.91 US 2005/0215610 A1 [0479] .sup.92 WO 93/07148 A1 [0480] .sup.93 Yu et al., Nature Communications (2012) 3(1288):1-11 [0481] .sup.94 JP 10152462 A [0482] .sup.95 WO 2010/147917 A1 [0483] .sup.96 DE 3231255 A1 [0484] .sup.97 EP 321122 A2 .sup.98 WO 2006/023778 A2 [0485] .sup.99 BE 624076 [0486] .sup.100 S. P. Parker, Ed., McGraw-Hill Dictionary of Chemical Terms (1984) McGraw-Hill Book Company, New York [0487] .sup.101 Eliel, E. and Wilen, S., “Stereochemistry of Organic Compounds”, John Wiley & Sons, Inc., New York, 1994 [0488] .sup.102 Cahn et al. Angew. Chem. Inter. Edit. 1966, 5, 385; errata 511 [0489] .sup.103 Ansel's Pharmaceutical Dosage Forms and Drug Delivery Systems (2004) Lippincott, Williams & Wilkins, Philadelphia [0490] .sup.104 Rowe R. C, Handbook of Pharmaceutical Excipients (2005) Pharmaceutical Press, Chicago [0491] .sup.105 Remington: The Science and Practice of Pharmacy (2000) Lippincott, Williams & Wilkins, Philadelphia [0492] .sup.106 Cunningham F. et al., Nucl. Acids Res. (2015) 43(D1): D662-D669 [0493] .sup.107 Mortazavi et al. Nat Methods (2008) 5(7):621-8 [0494] .sup.108 Selzer et al., Genes Chromosomes Cancer (2005) 44(3):305-319 [0495] .sup.109 Reich M. et al., Nature Genetics (2006) 38(5): 500-501 [0496] .sup.110 Desmond et al., Cancer Res (1985) 45(6):2913-2923 [0497] .sup.111 Shoemaker R. H., Nature Reviews Cancer (2016) 6:813-823 [0498] .sup.112 Augustyn et al., Proc Natl Acad Sci USA (2014) 111(41):14788-93 [0499] .sup.113 Hughes et al., Proteomics (2010) 10(9):1886-90