Therapeutic conversion of cancer cells

Abstract

The present invention relates to a method of treating RAS-dependent cancer in a subject by use of a farnesyltransferase inhibitor. Also described are compositions comprising a farnesyltransferase inhibitor; and an agent that increases the amount of geranylgeranyl pyrophosphate (GGPP). Lastly, the use of farnesyltransferase inhibitor for transdifferentiating a cell into a fibroblast-like stage is described herein.

Description

THERAPEUTIC CONVERSION OF CANCER CELLSCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of priority of Singapore provisional application no.10202403561X, filed 14 November 2024, the contents of it being hereby incorporated by reference in its entirety for all purposes.FIELD OF THE INVENTION

[0002] The present invention relates generally to the field of molecular biology. In particular, the present invention relates to the transdifferentiation of cellsBACKGROUND

[0003] HRAS, also known as GTPase HRAS, from “Harvey Rat sarcoma virus, is a commonly mutated oncogene and driver in a variety of human cancers. Therapeutically-intractable HRAS-driven cancers encompass 1% of all cancers (AACR Project GENIE consortium).

[0004] However, targeting HRAS has been challenging due to the absence of a druggable binding pocket in the HRAS protein.

[0005] Additionally, cell state transition contributes towards distinct step of malignant progression. Epithelial-to-mesenchymal (EMT) and the reverse process - mesenchymal-to-epithelia transition (MET) - confer phenotypic plasticity to cancer cells.

[0006] Thus, there is an unmet need for methods and compounds for transdifferentiation of cells.SUMMARY

[0007] In one aspect, disclosed herein is a method of treating RAS-dependent cancer in a subject. The method comprises administering a pharmaceutically effective amount of a famesyltransferase inhibitor to a subject in need thereof. The farnesyltransferase inhibitor can be, but is no limited to tipifarnib, lonafarnib, manumycin A and a-hydroxy farnesyl phosphonic acid Also, the RAS-dependent cancer can be a HRAS-dependent cancer. Furthermore, the method described herein can further comprise the additional use of geranylgeranyl pyrophosphate (GGPP). In one example, GGPP can be, but is not limited to exogenous GGPP, exogenous recombinant GGPP, lipid-based delivery system comprising GGPP, nanoparticles comprisingGGPP, and a combination thereof. In addition, the increase of GGPP can be compared to the level of GGPP in a subject not suffering from cancer.

[0008] In one example, the increase in GGPP is determined by the increase in the amount of geranylgeranylated proteins in the subject compared to a subject not suffering from cancer. It is also described that the farnesyltransferase inhibitor and the agent that increases the amount of GGPP can be administered either together in a single or separate compositions, separately, or sequentially. Also, the sequential administration described herein can involve the first administration of the farnesyltransferase inhibitor and a subsequent administration of the agent that increases the amount of GGPP; or the first administration of the agent that increases the amount of GGPP and a subsequent administration of the farnesyltransferase inhibitor. In the method described herein, cells of the RAS-dependent cancer are differentiated into fibroblast cells and the cells of the RAS-dependent cancer are in the mesenchymal state. The cancer is a KRAS-dependent cancer or a NRAS-dependent cancer. Examples for RAS-dependent cancers include but are not limited to breast cancer, bladder cancer, head and neck squamous cell cancer, lung cancer, thyroid cancer and melanoma.

[0009] Also, described is a composition comprising a farnesyltransferase inhibitor; and an agent that increases the amount of geranylgeranyl pyrophosphate (GGPP). Also, described herein is a method of transdifferentiating a cell into a fibroblast-like stage, the method comprising contacting the cell to a farnesyltransferase inhibitor wherein the method can further comprise contacting the cell to an agent that increases the amount of geranylgeranyl pyrophosphate (GGPP).BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The invention will be better understood with reference to the detailed description when considered in conjunction with the non-limiting examples and the accompanying drawings, in which:

[0011] FIG. 1 shows results indicating of a high-throughput small molecule inhibitor screen, identifying a farnesyltransferase inhibitor as a putative pro-mesenchymal cell state regulator. FIG. 1A shows a diagram of differential RAS protein prenylation and tipifarnib-mediated farnesyltransferase inhibition. Figure made with BioRender. FIG. IB shows bright-field images of 7-day control and tipifamib-treated (500 nM) NAMEC8 cells (N8) and NAMEC8-HRAS (N8H). Scale bar represents 300 pm. FIG. 1C shows results of Phalloidin staining of NAMEC8-HRAS control vs. 7-day tipifamib-treated (500nM) cells to quantify cell length (mm) usingImage! Scale bar represents 100 pm, n=100 cells. Statistical analysis was done using unpaired t-test, ****P<0.0001. FIG. ID shows results of protein expression analysis of cell state markers and EMT markers in tipifarnib-treated (500 nM) parental NAMEC8 and NAMEC8-HRAS cells over the course of 14 days. p-actin was used as loading control. FIG. IE shows global cell-type signatures of genes upregulated in tipifarnib-treated NAMEC8-HRAS cells. Enrichment scores for each cell-type signature were shown. FIG: IF shows results of single-cell UMAP visualization of control cells (global, grey) and tipifamib treated cells (global, black) as well as the four clusters identified represented in red. FIG. 1G shows UMAP visualization of fibroblast markers by gene weighted density: Vimentin, smooth muscle actin (SMA), fibroblast activation protein (FAP), fibroblast specific protein (FSP) andNidogen gene expression plots, respectively, in the global clustering. FIG. 1H shows results of protein expression of fibroblast proteins in tipifarnib-treated (500 nM, 7 days) mesenchymal NAMEC8, NAMEC8-HRAS cells and epithelial HMLE cells. 0- actin was used as loading control. Dotted line shows splicing of noncontiguous lanes. FIG. II shows bright-field images of tipifarnib-treated (500 nM, 7 days) HRAS-driven TNBC line, SUM159. Scale bar represents 300 pm FIG. 1J shows the results of protein expression analysis of fibroblast proteins in tipifarnib-treated (500 nM, 7 days) SUM159 cells 0-actin was used as loading control.

[0012] FIG. 2 shows the results indicating that pharmacological inhibition of famesyltransferase converts mesenchymal breast cancer cells into fibroblast. FIG. 2A shows a plot of tumour mass of in vitro tipifarnib pre-treated (500 nM, 7 days) NAMEC8-HRAS tumours. Data are represented as the mean mass + / - SD, n=4 for DMSO-treated NAMEC8-HRAS tumours, n=4 for tipifarnib-treated NAMEC8-HRAS tumours. Statistical analysis was done using unpaired t-test, ****P<0.0001. Representative images of tumours from mice sacrificed at week 6 post implantation (right). FIG. 2B shows images of immunofluorescence staining of epithelial lineage markers CK5, cytokeratin 5 and fibroblast markers FSP, fibroblastspecific protein; FAP, fibroblast activation protein; SMA, smooth muscle actin; and fibronectin in NAMEC8-HRAS tumour sections of in vitro pre-treated control tumours and tipifarnib tumours. Scale bar represents 100 pm. FIG. 2C shows graphs of tumour volume measurements of NAMEC8-HRAS tumours with in vivo tipifarnib treatment (200 mg / kg / day) for 7 days. Tumour tracking for 62 days is represented in the line plot (left). Data are represented as the mean volume + / - SD, n=3 for cyclodextrin control, n=3 tumours for tipifarnib-treated. Statistical analysis was done using unpaired t-test, **P<0.01. Tumour mass represented in bar graph (middle). Data are represented as the mean mass + / - SD, n=3 for cyclodextrin control, n=3tumours for tipifamib-treated. Statistical analysis was done using unpaired t-test, **P<0.01. Representative images of tumours from mice sacrificed at 62 days of implantation (right). FIG.2D shows images of immunofluorescence staining of epithelial lineage markers CK5, cytokeratin 5 and fibroblast markers FSP, fibroblast-specific protein; FAP, fibroblast activation protein; SMA, smooth muscle actin; and fibronectin in NAMEC8-HRAS tumour sections of cyclodextrin control tumours and in vivo tipifarnib treated tumours. Scale bar represents 100 gm. FIG. 2E shows images of hematoxylin and eosin (H&E) staining of in vivo cyclodextrin control vs. tipifarnib-treated NAMEC8-HRAS tumours. Scale bar represents 250 gm for 10X and 50 gm for 40X magnification. FIG. 2F shows the results of Masson’s tri chrome staining of in vivo control vs. tipifarnib-treated NAMEC8-HRAS tumours Blue staining indicates the presence of collagen. Scale bar represents 250 pm for 10X magnification.

[0013] FIG. 3 shows results indicating that extreme mesenchymal transition (xMT) diminishes the tumorigenicity of mesenchymal cancer cells. FIG. 3A shows bright-field images of NAMEC8-HRAS shControl, shControl tipifarnib-treated (500 nM, 7 days), FNTB-knockdown shl and sh2 cells. Scale bar represents 300 gm. FIG. 3B shows results of protein expression of fibroblast proteins in NAMEC8-HRAS FNTB-KD shl and sh2 cells. P-actin was used as loading control. Dotted line shows splicing of non-contiguous lanes. FIG. C shows a plot of tumour volume measurements of NAMEC8-HRAS FNTB-KD tumours. Tumour tracking for 59 days is represented in the line plot (top). Data are represented as the mean volume + / - SD, n=6 for shControl-KD. Statistical analysis was done using unpaired t-test, ****P<0.000l. Final tumour mass data are represented in the bar graph (bottom, left). Data are represented as the mean mass + / - SD, n=6 for shControl-KD, n=6 tumours for FNTB-KD cells. Statistical analysis was done using unpaired t-test, ****P<0.0001. Representative images of tumours from mice sacrificed at 59 days of implantation (bottom, right). FIG. 3D shows images of immunofluorescence staining of epithelial lineage markers CK5, cytokeratin 5 and fibroblast markers FSP, fibroblast-specific protein; FAP, fibroblast activation protein; SMA, smooth muscle actin; and fibronectin in control versus NAMEC8-HRAS FNTB-KD tumours. Scale bar represents 100 gm. FIG. 3E shows a graph of tumour volume measurements of NAMEC8-HRAS doxycycline-induced FNTB1216 KD tumours. Tumour tracking for 53 days is represented in the line plot (top). Data are represented as the mean volume + / - SD, n=9 for shControl-KD, n=9 tumours for FNTB-KD. Statistical analysis was done using unpaired t-test, ***P<0.001. Final tumour mass data are represented in the bar graph (bottom, left). Data are represented as the mean mass + / - SD, n=9 for shControl-KD, n=9 tumours for FNTB-KD. Statistical analysis wasdone using unpaired t-test, **P<0.01. Representative images of tumours from mice sacrificed at 59 days of implantation (bottom, right). FIG. 3F shows images of immunofluorescence staining of epithelial lineage markers CK5, cytokeratin 5 and fibroblast markers FSP, fibroblast-specific protein, FAP, fibroblast activation protein; SMA, smooth muscle actin; and fibronectin in NAMEC8- HRAS tumour sections of doxycycline-inducible control and FNTB-KD tumours. Scale bar represents 100 pm. FIG. 3G shows a column graph depicting the time to metastasis-induced death in immunocompromised N 1229 SG mice with intravenous injection of NAMEC8-HRAS and NAMEC8-HRAS tipifamib (500 nM, 7 days) in vitro pre-treated cells Data are represented as the mean + / - SD. Statistical analysis was done using unpaired t-test with Welch’s correction, n.s.=not significant.

[0014] FIG. 4 shows data indicating that the inhibition of the MAPK pathway in HRAS-dependent mesenchymal cells promotes xMT FIG. 4A shows a diagram of the canonical RAS activation pathway. FTase inhibition is achieved with tipifarnib and RAFI inhibition is achieved with either sorafenib treatment or gene knockdown. MEK inhibition is achieved with PD 98059 treatment. Figure made with BioRender. FIG. 4B shows bright-field images of NAMEC8-HRAS control, sorafenib-treated (0.5-1 pM) and MEK inhibitor-treated (1 pM) cells for 7 days. Scale bar represents 300 pm. FIG. 4C shows results of protein expression of fibroblast proteins and MAPK pathway proteins in NAMEC8-HRAS sorafenib-treated (0.5-1 pM) and MEK inhibitor-treated (1 pM) cells for 7 days. P-actin was used as loading control. Dotted line shows splicing of non-contiguous lanes. FIG. 4D shows bright-field images of NAMEC8-HRAS RAF1-KD shl and sh2 cells. Scale bar represents 300 pm. FIG. 4E shows the results of protein expression of fibroblast proteins and MAPK pathway proteins in NAMEC8-HRAS FNTB-KD and NAMEC8-HRAS RAF1-KD cells. P-actin was used as loading control. FIG. 4F shows the results of protein expression of HRAS-GTP and HRAS in NAMEC8-HRAS control vs. NAMEC8-HRAS RAF1-KD cells. P-actin (input) was used as loading control.

[0015] FIG. 5 shows results indicating that a switch from farnesylation to geranylgeranyl ati on drives cytoskeletal protein changes resulting in transdifferentiation into fibroblast. FIG. 5A is a schematic of prenylation and the effects of exogenous supplementation with GGPP. Figure made with BioRender. FIG. 5B shows the results of CLICK-IT immunofluorescence staining of farnesylated (Azide-FPP, left) and geranylgeranylated (Azide-GGPP, right) proteins in control, 2-day tipifarnib-treated and 7-day tipifarnib-treated (500 nM) NAMEC8-HRAS cells. Scale bar represents 100 pm. (Right) Quantification of Azide-FPP and Azide-GGPP in control, 2-day and 7-day tipifarnib-treated 1263 NAMEC8-HRAS cells. Errorbars indicate mean + / - SD. Statistical analysis was done using unpaired t-test, ****P<0.0001. FIG. 5C shows bright-field images of NAMEC8-HRAS shControl, shControl tipifarnib-treated, and GGPS1-KD cells treated with tipifarnib (500 nM) Scale bar represents 300 pm. FIG. 5D shows results of protein expression of fibroblast proteins in NAMEC8-HRAS GGPS1-KD cells treated with tipifarnib (500 nM for 7 days) followed by supplementation of GGPP (11 pM) at timepoints indicated. P-actin was used as loading control. FIG. 5E shows a schematic of NAMEC8-HRAS GGPS1-KD intratumoural xenograft GGPP supplementation (11 pM) after tipifarnib treatment (200 mg / kg / day, 7 days). Figure made with BioRender. FIG. 5F shows the results of immunofluorescence staining of epithelial lineage marker CK5, cytokeratin 5 and fibroblast markers FSP, fibroblast-specific protein; FAP, fibroblast activation protein; SMA, smooth muscle actin; and fibronectin in tipifarnib-treated NAMEC8-HRAS GGPS1-KD, vehicle- and GGPP-supplemented tumours. Scale bar represents 200 pm. FIG. 5G shows results of protein expression of fibroblast markers upon GGTI (1 pM) and tipifarnib treatment (500 nM, 7 days) in NAMEC8-HRAS cells. Schematic of drug treatment durations, made with BioRender. P-actin was used as loading control FIG. 5H shows results of protein expression of GGPS1 and Rho GTPase proteins in NAMEC8-HRAS control, tipifarnib-treated (500 nM, 7 days) and FNTB-KD cells P-actin was used as loading control FIG 51 shows the results of protein expression of Rho GTPase family effector proteins in NAMEC8-HRAS control, tipifarnib-treated (500 nM, 7 days) and FNTB-KD cells, p-actin was used as loading control. FIG. 5J shows results of protein expression of total and GTP -bound RHO A, RAC1 and CDC42 in NAMEC8-HRAS control vs. NAMEC8-HRAS tipifarnib-treated (500 nM, 7 days) cells. P-actin (input) was used as loading control. FIG. 5K shows results of protein expression of fibroblast proteins after RAC1 inhibition (NSC 23766, 500 nM) in tipifarnib pre-treated (500 nM, 7 days) NAMEC8-HRAS cells. Schematic of drug treatment durations, made with BioRender. P-actin was used as loading control. FIG. 5L shows the results of protein expression of fibroblast proteins after 1296 CDC42 inhibition (ML 141, 500 nM) in tipifarnib pre-treated (500 nM, 7 days) NAMEC8-HRAS cells. Schematic of drug treatment durations, made with BioRender. P-actin was used as loading control.

[0016] FIG. 6 shows data indicating that prenylation of RAC1 and CDC42 result in the downstream activation of AP-1 for transcriptional regulation of xMT. FIG. 6A shows results of an NCl-Nature 2016 pathway enrichment analyses of upregulated genes in tipifarnib-treated (500 nM, 7 days) NAMEC8-HRAS cells from RNA-seq. FIG. 6B shows results of an NCl-Nature 2016 pathway enrichment analyses of upregulated genes in NAMEC8-HRAS FNTB-KDcells from RNA-seq FIG. 60 shows the results of protein expression of fibroblast proteins in NAMEC8-HRAS control, tipifarnib -treated (500 nM), amlodipine-treated (500 nM) 2 days and 7 days cells. Schematic of drug treatment durations, made with BioRender. P-actin was used as loading control. FIG. 6D Protein expression of fibroblast proteins, FNK markers, AP-1 markers and their phosphorylated counterparts in NAMEC8-HRAS control, tipifarnib-treated (500 nM), and combination treatment of tipifarnib (500 nM) and JNK inhibitor (SP600125, 500 nM) for 2 days and 7 days. Schematic of drug treatment durations, made with BioRender, p-actin was used as loading control. FIG. 6E shows the results of protein expression of fibroblast proteins, FOS / JUN subunits and their phosphorylated counterparts in NAMEC8-HRAS tipifarnib-treated (500 nM) in combination with an AP-1 inhibitor (SRI 1302, 1 pM) for 2 days and 7 days. P-actin was used as loading control. FIG. 6F shows bright-field images of NAMEC8-HRAS control, tipifarnib-treated (500 nM), and in combination with AP-1 inhibitor (SRI 1302, 500 nM). Schematic of drug treatment durations, made with BioRender. Scale bar represents 300 pm. FIG.6G shows a schematic of the mechanistic regulation of xMT. Figure made with BioRender. FIG.6H shows a schematic of a revised epigenetic landscape of epithelial-mesenchymal plasticity that now includes transdifferentiation towards a fibroblast lineage.

[0017] FIG. 7 shows data indicating that drug screens identified a farnesyltransferase inhibitor as a putative promesenchymal cell state regulator. FIG. 7A shows results of dual-glo luciferase assays of NAMEC8-SERP1NE1 prom-Luc cells treated with serial dilutions of tipifarnib (1 nM, 25 nM, 50 nM, 75 nM, 100 nM, and 500 nM). Data are represented as the mean FireFly / Renilla luminescence + / - SD, n=5. Statistical analysis was done using unpaired t-test, n.s.=not significant, **P<0.01, ***P<0.001, ****P<0.0001. FIG. 7B shows results of cell viability assays (Cell Titre Gio) of parental mesenchymal NAMEC8 (top, left), NAMEC8-HRAS (top, right), NAMEC8-KRAS (bottom, left) and epithelial HMLE (bottom, right) control and tipifarnib-treated (500 nM) cells. Data are represented as the mean + / - SD. FIG. 7C shows results of dual-glo luciferase assays of NAMEC8-SERPINE1 prom-Luc cells treated with serial dilutions of lonafamib. Data are represented as the mean FireFly / Renilla luminescence + / - SD, n=3. Statistical analysis was done using unpaired t-test, n.s.=not significant, *P<0.05, **P<0.01, ***P<0.001. FIG 7D shows a list of global cell-type signatures of genes upregulated in tipifarnib-treated NAMEC8- HRAS cells and corresponding P-value scores (plotted -loglO). FIG. 7E shows the results of protein expression of fibroblast proteins in patient-derived tumour cells and matched patient fibroblasts (patients BC257 and BC258). P-actin was used as loading control. FIG. 7F a heatmap of cell-type clustering for the four distinct populations, highlighted inred: untreated (cluster 1), mixed (cluster 2) and tipifarnib-treated (500 nM, 7 days) (cluster 3a and 3b). FIG. 7G shows bright-field images of tipifarnib-treated (500 nM, 7 days) RL95-2 cell line (left). Scale bar represents 300 pm. Protein expression of fibroblast proteins in RL95-2 control and tipifarnib-treated cells (right). 0-actin was used as loading control. FIG. 7H shows bright-field images of tipifarnib-treated (500 nM, 7 days) KYSE30 cell line (left). Scale bar represents 300 pm. Protein expression of fibroblast proteins in KYSE30 control and tipifarnib-treated cells (right). P-actin was used as loading control. FIG. 71 shows bright-field images of tipifarnib-treated (500 nM, 7 days) MCF7-HRAS cell line (left). Scale bar represents 300 pm. Protein expression of fibroblast proteins in MCF7-HRAS control and tipifarnib-treated cells 1361 (right). P-actin was used as loading control. FIG. 7J shows (Left) bright-field images of NAMEC8-HRAS control and NAMEC8-HRAS tipifarnib-treated cells (500 nM, 7 days), at various timepoints (hours) upon wounding. Yellow lines demarcate area of wound. Scale bar represents 300 pm for 4X magnification. (Right) Quantitation of wound closure at various timepoints upon wounding. Statistical analysis was done using unpaired t-test with Welch’s correction, *P<0.05. FIG 7K (Top) shows representative bright-field images of trans-well assay of NAMEC8-HRAS vehicle control and NAMEC8-HRAS tipifarnib-treated cells (500 nM, 7 days), 24 hours after cell seeding. (Bottom) Quantitation of cells estimated by 1 / Integrated density + / - SD. Statistical analysis was done using unpaired t-test with Welch’s correction, ns=not significant.

[0018] FIG. 8 shows results indicating that pharmacological inhibition of farnesyltransferase converts mesenchymal breast cancer cells into fibroblast. FIG. 8A shows an image of a tumour mass of SUMI 59 control and tipifarnib in vitro pre-treated (500 nM, 7 days) tumours. Data are represented as the mean mass + / - SD, n=4 for DMSO control, n=4 tumours for tipifarnib-treated. Statistical analysis was done using unpaired t-test, **P<0.01. Representative images of tumours (right). FIG. 8B shoes results of hematoxylin and eosin (H&E) staining of SUMI 59 control and tipifarnib in vitro pre-treated (500 nM, 7 days) tumours. Arrows show extracellular matrix. Scale bar represents 100 pm for 20X and 50 pm for 40X magnification. FIG. 8C shows a heatmap of hierarchical clustering of mouse squamous cell carcinoma EMT subpopulations (32) (red) with untreated and tipifarnib-treated (500 nM, 7 days) NAMEC8-HRAS cells (blue). FIG. 8D shows images of UMAPs of single cell transcriptomics in different cancer types. Left: cell types, Middle: Epithelial markers (EPCAM, CK5, CK8, CK14, CK18, CK19) expression, Right: FSP expression. Black outlines in the right-most column depict putative fibroblast-like cells expressing FSP within the epithelium cell population. (Wu SZ, Al-Eryani G, Roden DL,Junankar S, Harvey K, Andersson A, et al. A single-cell and spatially resolved atlas of human breast cancers. Nat Genet. 2021,53(9): 1334-47; Nowicki-Osuch K, Zhuang L, Cheung TS, Black EL, Masque-Soler N, Devonshire G, et al. Single-Cell RNA Sequencing Unifies Developmental Programs of Esophageal and Gastric Intestinal Metaplasia. Cancer Discov. 2023; 13(6): 1346-63; Vazquez-Garcia 1, Uhlitz F, Ceglia N, Lim JLP, Wu M, Mohibullah N, et al. Ovarian cancer mutational processes drive site-specific immune evasion. Nature. 2022;612(7941):778-86; Glasner A, Rose SA, Sharma R, Gudjonson H, Chu T, Green JA, et al. Conserved transcriptional connectivity of regulatory T cells in the tumor microenvironment informs new combination cancer therapy strategies. Nat Immunol. 2023 ;24(6): 1020-35; Zhang Y, Narayanan SP, Mannan R, Raskind G, Wang X, Vats P, et al Single-cell analyses of renal cell cancers reveal insights into tumor microenvironment, cell of origin, and therapy response. Proc Natl Acad Set USA.2021; 118(24)).

[0019] FIG. 9 shows data indicating that extreme mesenchymal transition diminishes the tumorigenicity of mesenchymal cancer cells. FIG. 9A shows a graph of Farnesyltransferase B subunit (FNTB) gene expression 1395 in tipifamib-treated and FNTB-KD cells relative to DMSO-treated NAMEC8-HRAS control cells. Data are represented as the mean + / - SD, n=3. Statistical analysis was done using unpaired t-test, *P<0.05, ****P<0.0001. FIG. 9B shows results of protein expression of pan-farnesylation and epithelial lineage markers (pan-CK and CK5) in NAMEC8-HRAS shControl vs. FNTB-KD shl and sh2 cells, p-actin was used as loading control. Dotted line shows splicing of non-contiguous lanes. FIG. 9C shows the results of cell viability assay (Cell Titre Gio) of NAMEC8-HRAS shControl vs. FNTB-KD shl and sh2 cells Data are represented as the mean + / - SD, n=5. FIG 9D shows the results of colony formation assay of NAMEC8-HRAS shControl vs. FNTB-KD shl and sh2 cells, n=3 wells. Data are represented as the mean + / - SD. Statistical analysis was done using unpaired t-test, **P<0.01. FIG. 9E (left) shows bright-field images of NAMEC8-HRAS DMSO and NAMEC8-HRAS FNTB-KD cells at various timepoints (hours) upon wounding. Yellow lines demarcate area of wound. Scale bar represents 300 pm for 4X magnification. (Right) Quantitation of wound closure at various timepoints upon wounding. Statistical analysis was done using unpaired t-test with Welch’s correction, **P<0.01, ns=not significant. FIG. 9F shows the results of hematoxylin and eosin (H&E) staining of NAMEC8-HRAS shControl knockdown vs. NAMEC8-HRAS FNTB-KD tumours. Scale bar represents 100 pm for 20X and 50 pm for 40X magnification. FIG. 9G shows results of Masson’s trichrome staining in NAMEC8-HRAS shControl and NAMEC8-HRAS FNTB-KD xenograft tumour sections. Blue staining indicates the presence ofcollagen. Scale bar represents 250 pm. FIG 9H shows results of protein expression of pan-famesylation in NAMEC8-HRAS control, tipifarnib treated (500 nM, 7 days), uninduced FNTB-KD and doxycycline-induced FNTBKD cells. p-actin was used as loading control FIG. 91 shows results of hematoxylin and eosin (H&E) staining of NAMEC8-HRAS inducible shControl- KD vs. inducible FNTB-KD tumours. Scale bar represents 100 pm for 20X and 50 pm for 40X magnification.

[0020] FIG. 10 shows results indicating that fibroblast transdifferentiation is the result of unutilized famesyl pyrophosphate being atypically shunted into geranylgeranylation FIG. 10A shows bright-field images of NAMEC8-KRAS control and sorafenib-treated (1 pM, 7 days), and NAMEC8 control and sorafenib-treated (1 pM, 7 days) cells. Scale bar represents 300 pm. FIG.10B shows the results of protein expression of fibroblast proteins in sorafenib -treated (1 pM, 7 days) NAMEC8-KRAS and NAMEC8 cells. P-actin was used as loading control. FIG. 10C shows Geranylgeranyl pyrophosphate synthase (GGPS1) gene expression in GGPS1-KD NAMEC8-HRAS cells relative to shControl cells. Data are represented as the mean + / - SD, n=3. Statistical analysis was done using unpaired t-test, ****P<0.0001. FIG 10D shows results of hematoxylin and eosin (H&E) staining of NAMEC8-HRAS GGPS1-KD vs. GGPP supplemented GGPS1-KD tumours Scale bar represents 100 pm for 20X and 50 pm for 40X magnification. FIG. 10E shows protein expression of RHOA, RAC1 and CDC42 in NAMEC8-HRAS GGPS1-KD cells treated with tipifarnib (500 nM, 7 days) and exogenously supplemented with GGPP (11 pM). P-actin was used as loading control. FIG. 10F shows bright-field images of NAMEC8-KRAS control (top), tipifamib-treated (middle), and NAMEC8-KRAS FNTB-KD (bottom). Scale bar represents 300 pm. FIG. 10G shows results of protein expression of fibroblast proteins in NAMEC8-KRAS tipifarnib-treated and FNTB-KD cells. P-actin was used as loading control. Dotted line shows splicing of non-contiguous lanes. FIG 10H shows protein expression of total and GTP -bound HRAS and KRAS in NAMEC8-KRAS control and tipifamib-treated (500 nM, 7 days) cells. P-actin (input) was used as loading control. Numbers indicate relative band intensities normalized to P-actin. FIG. 101 shows graphs of tumour mass of NAMEC8-KRAS tumours represented in the bar graph (left). Data are represented as the mean mass + / - SD, n=3 for control in vitro pre-treated, n=3 tumours for tipifarnib in vitro pre-treated (500 nM, 7 days) cells. Statistical analysis was done using unpaired t-test, ns=not significant. Representative images of tumours from mice sacrificed at week 6 post implantation (right). FIG.10J shows results of hematoxylin and eosin (H&E) staining of NAMEC8-1459 KRAS control vs. tipifarnib in vitro pre-treated (500 nM, 7 days) tumours. Scale bar represents 100 pm for 20Xand 50 m for 40X magnification. FIG. 10K shows bright-field images of NAMEC8-HRAS+KRAS control cells (top), and tipifarnib-treated (500 nM, 7 days) (bottom). Scale bar represents 300 pm. FIG 10L shows results of protein expression of fibroblast proteins in NAMEC8-HRAS tipifarnib-treated (500 nM) cells (2 days and 7 days) vs NAMEC8-HRAS+KRAS tipifarnib-treated (500 nM) cells (2 days and 7 days), p-actin was used as loading control. FIG. 10M shows protein expression of total and GTP -bound HRAS and KRAS in NAMEC8-HRAS vs. NAMEC8-HRAS+KRAS cells control and tipifarnib-treated (500 nM, days) cells. P-actin (input) was used as loading control. Numbers indicate relative band intensities normalized to P-actin.

[0021] FIG. 11 shows data indicating that prenylation of RAC 1 and CDC42 result in the downstream activation of AP-1 for transcriptional regulation of xMT. FIG. 11A shows results of a cell viability assay (Cell Titre Gio) of NAMEC8-HRAS control vs. calcium channel inhibition (amlodipine at 500 nM, for 8 days). Data are represented as the mean + / - SD, n=5. FIG. 11B shows results of gene expression of AP-1 subunits in NAMEC8-HRAS control vs tipifarnib-treated cells. Data are represented as the mean + / - SD, n=4. Statistical analysis was done using unpaired t-test, ***P<0.001, ****P<0.0001, n.s.= not significant. FIG. 11C shows results of protein expression of AP-1 subunit proteins in NAMEC8-HRAS control and tipifarnib-treated (500 nM, 7 days) cells. P-actin was used as loading control. FIG. 1 ID shows results of a cell viability assay (Cell Titre Gio) of NAMEC8-HRAS control vs. JNK-inhibited (SP600125 at 500 nM, for 8 days) treatment. Data are represented as the mean + / - SD, n=5. FIG. 1 IE shows results of a cell viability assay (Cell Titre Gio) of NAMEC8-HRAS control vs. AP-1 -inhibited (SRI 1302 at 1 pM, for 8 days) treated cells. Data are represented as the mean + / - SD, n=5. FIG.1 IF shows results of protein expression of fibroblast proteins, INK markers, AP-1 markers and their phosphorylated counterparts in NAMEC8-HRAS control, tipifarnib-treated (500 nM, 7 days), and combination treatment of tipifarnib and RAC inhibitor (NSC 23766, at 500 nM) for 2 days and 7 days. Schematic of drug treatment durations, made with BioRender, p-actin was used as loading control. FIG. 11G shows results of protein expression of fibroblast proteins, JNK markers, 1493 AP-1 markers and their phosphorylated counterparts in NAMEC8-HRAS control, tipifarnib-treated (500 nM, 7 days), and combination treatment of tipifarnib and CDC42 inhibitor (ML 141, at 500 nM) for 2 days and 7 days. Schematic of drug treatment durations, made with BioRender. P-actin was used as loading control.

[0022] FIG. 12 shows results indicating that partial fibroblast transdifferentiation in N8Kras and H1299 (NRas-driven) tipifarnib-treated cells with GGPP supplementation in vitro. In moredetail, Figure 12 shows protein expression of fibroblast proteins and Rho-GTPases in NAMEC8-KRAS (left) and H1299 (right) treated with tipifarnib (500 nM for 7 days) followed by supplementation of GGPP (11 pM) at timepoints indicated.

[0023] FIG. 13 shows results indicating that tipifarnib, a farnesyltransferase inhibitor targeting HRAS promotes an extreme mesenchymal cell state in mesenchymal cancer cells and transdifferentiates cancer cells into fibroblast archetypes in vitro and in in vivo pre-clinical mouse models. FIG. 13A shows results of dual-glo luciferase assay of NAMEC8-SERPINE1 prom-Luc cells treated with serial dilutions of tipifarnib (1 nM, 25 nM, 50 nM, 75 nM, 100 nM, and 500 nM).Data are represented as the mean FireFly / Renilla luminescence + / - SD, n=5. Statistical analysis was done using unpaired t-test, n.s.=not significant, **P<0.01, ***P<0.001, ****P<0.0001. FIG. 13B shows results of protein expression of fibroblast proteins in tipifarnib -treated (500 nM, 7 days) mesenchymal NAMEC8, NAMEC8-HRAS cells and epithelial HMLE cells. 0-actin was used as loading control. Dotted line shows splicing of noncontiguous lanes. FIG. 13C shows results of immunofluorescence staining of epithelial lineage markers CK5, cytokeratin 5 and fibroblast markers FSP, fibroblast-specific protein; FAP, fibroblast activation protein; SMA, smooth muscle actin; and fibronectin in NAMEC8-HRAS tumour sections of cyclodextrin control tumours and in vivo tipifarnib -treated tumours. Scale bar represents 100 pm.

[0024] FIG. 14 shows results indicating that pharmacological inhibition of farnesyltransferase or genetic knockdown of farnesyltransferase subunits (FNTB) is sufficient to inhibit farnesylation and decrease tumorigenic capacity. FIG. 14A shows results of tumour volume measurements of NAMEC8-HRAS tumours with in vivo tipifarnib treatment (200 mg / kg / day) for 7 days. Tumour tracking for 62 days is represented in the line plot (left). Data are represented as the mean volume + / - SD, n=3 for cyclodextrin control, n=3 tumours for tipifarnib-treated. Statistical analysis was done using unpaired t-test, **P<0.01. Tumour mass represented in bar graph (middle). Data are represented as the mean mass + / - SD, n=3 for cyclodextrin control, n=3 tumours for tipifamib-treated. Statistical analysis was done using unpaired t-test, **P<0.01. Representative images of tumours from mice sacrificed at 62 days of implantation (right). FIG 14B shows results of tumour volume measurements of NAMEC8-HRAS FNTB-KD tumours. Tumour tracking for 59 days is represented in the line plot (left). Data are represented as the mean volume + / - SD, n=6 for shControl-KD. Statistical analysis was done using unpaired t-test, ****P<0.0001. Final tumour mass data are represented in the bar graph (middle). Data are represented as the mean mass + / - SD, n=6 for shControl-KD, n=6 tumours for FNTB-KD cells.Statistical analysis was done using unpaired t-test, ****P<0.0001. Representative images of tumours from mice sacrificed at 59 days of implantation (right).

[0025] FIG. 15 shows results indicating that activation of geranylgeranylation in HRAS-driven cancer cells promotes xMT transdifferentiation. FIG. 15A shows results of CLICK-IT immunofluorescence staining of farnesylated (Azide-FPP, left) and geranylgeranylated (Azide-GGPP, right) proteins in control, 2-day tipifarnib-treated and 7-day tipifarnib-treated (500 nM) NAMEC8-HRAS cells. Scale bar represents 100 pm. (Right) Quantification of Azide-FPP and Azide GGPP in control, 2-day and 7-day tipifarnib-treated 1263 NAMEC8-HRAS cells Error bars indicate mean + / - SD. Statistical analysis was done using unpaired t-test,****P<0.0001. FIG. 15B shows results of protein expression of fibroblast markers upon GGTT (1 pM) and tipifamib treatment (500 nM, 7 days) in NAMEC8-HRAS cells. Schematic of drug treatment durations, made with BioRender. [3-actin was used as loading control.

[0026] FIG. 16 shows results indicating that identification of Rho GTPase family members, RAC1 and CDC42 as key geranylated targets that subsequently undergoes activation and mediate transdifferentiation. FIG 16A shows results of protein expression of GGPS1 and Rho GTPase proteins in NAMEC8-HRAS control, tipifarnib-treated (500 nM, 7 days) and FNTB-KD cells [3-actin was used as loading control. FIG. 16B shows results of protein expression of total and GTP -bound RAC1 and CDC42 in NAMEC8-HRAS control vs. NAMEC8-HRAS tipifarnib-treated (500 nM, 7 days) cells. [3-actin (input) was used as loading control.

[0027] FIG. 17 shows results indicating that activation of INK signalling and AP-1 complex by RAC1 and CDC42 support a durable xMT transdifferentiation phenotype. FIG. 17A shows results of protein expression of fibroblast proteins, INK markers, AP-1 markers and their phosphorylated counterparts in NAMEC8-HRAS control, tipifarnib-treated (500 nM), and combination treatment of tipifamib (500 nM) and JNK inhibitor (SP600125, 500 nM) for 2 days and 7 days. Schematic of drug treatment durations, made with BioRender. [3-actin was used as loading control. FIG. 17B shows results of protein expression of fibroblast proteins, JNK markers, AP-1 markers and their phosphorylated counterparts in NAMEC8-HRAS control, tipifarnib-treated (500 nM, 7 days), and combination treatment of tipifamib and RAC inhibitor (NSC 23766, at 500 nM) for 2 days and 7 days. Schematic of drug treatment durations, made with BioRender. [3-actin was used as loading control. FIG. 17C shows results of protein expression of fibroblast proteins, JNK markers, 1493 AP-1 markers and their phosphorylated counterparts in NAMEC8-HRAS control, tipifarnib-treated (500 nM, 7 days), and combination treatment of tipifamib and CDC42 inhibitor (ML141, at 500 nM) for 2 days and 7 days.Schematic of drug treatment durations, made with BioRender. [J-actin was used as loading control.

[0028] FIG. 18 is an image of a gel, showing that the AP-1 complex and JNK are activated upon farnesyltransferase inhibitor and GGPP combinatorial treatment. In more detail, Figure 18 shows protein expression of fibroblast proteins, Rho-GTPases, JNK markers, AP-1 markers and their phosphorylated counterparts in NAMEC8-HRAS treated with tipifarnib (500 nM for 7 days) followed by supplementation of GGPP (11 pM) at timepoints indicated.DEFINITIONS

[0029] As used herein, the terms “RAS”, “HRAS”, “AP-1”, “N-Cadherin”, “Zebl”, “Snail”, “Slug”, “Twist” and “SMA” can refer to both the genes that express, and the proteins, RAS, HRAS, AP-1, N-Cadherin, Zebl, Snail, Slug, Twist and SMA, respectively. The terms “RAS", “HRAS”, “AP-1”, “N-Cadherin” , “Zebl”, “SnaiF, “Slu ', “Twist” and “SMA” (with italics) and “RAS”, “HRAS”, “AP-1”, “N-Cadherin”, “Zebl”, “Snail”, “Slug”, “Twist” and “SMA” (without italics) can be used interchangeably, and which is to be used is understood depending on the context of the present disclosure. This definition also applies to all other genes and their corresponding proteins mentioned in the present disclosure.

[0030] As used herein, the terms “RAS”, “HRAS”, “N-Cadherin”, “Zebl”, “Snail”, “Slug”, “Twist” and “SMA” have been identified with their human sequence identification numbers: HRAS (gene: NG 007666.1; protein: AAB02605.1), KRAS (gene: NG 007524.2; protein: NP 004976.2), NRAS (gene: NG 007572.1; protein: NP 002515.1), N-Cadherin (gene: EF444966.1; protein: NP_001783.2), Zebl (gene: NG_017048.2; protein:NP_001121600.1), Snail (gene: AJ245657.1; protein: CAB52414.1), Slug (gene: NG_012130.1; protein:NP_003059.1), Twist (gene: NG_008114.2; protein: NP_000465.1), and SMA (gene: NG 011541.1; protein: CAG38756.1).

[0031] As used herein, the term “RAS-dependent cancer” refers to a cancer in which RAS is shown to comprising one or more mutations, rendering it oncogenic and contributing to the development of the cancer. While RAS does not have to be the sole driver of the cancer, it plays a predominant role in the oncogenesis or carcinogenesis of the cancer.

[0032] As used herein, the term “HRAS -dependent cancer” refers to a RAS-dependent cancer that is driven by HRAS, and not KRAS or NRAS.

[0033] As used herein the terms “treating’ or “treatment” refer to any and all uses which remedy a disease state or symptoms, prevent the establishment of disease, or otherwise prevent,hinder, retard, or reverse the progression of disease or other undesirable symptoms in any possible way.

[0034] In the context of the present disclosure, the terms “a pharmaceutically effective amount” or “an effective amount” includes within its meaning a non-toxic but sufficient amount of an agent to provide the desired pharmaceutical or therapeutic effect. As can be appreciated by a person skilled in the art, the exact amount required will vary from subject to subject depending on factors such as, but not limit to, the species being treated, age, gender, general medical condition of the subject, the particular agent being administered and the mode of administration and so forth. An appropriate “effective amount” can be determined by one of skill in the art using means known in the art for determining the same Single or multiple administrations of the agents according to the present disclosure can be carried out. One skilled in the art would be able, by appropriate means, to determine effective, non-toxic dosage levels of the agents of the present disclosure and an administration pattern which would be suitable for treating the disorders, diseases and / or infections to which the agents are applicable.

[0035] As used herein, the term “recombinant” refers to a recombinant nucleic acid or a protein that has a sequence that is not naturally occurring or was made artificially. Artificial combination is often accomplished by chemical synthesis or, more commonly, by the artificial manipulation of isolated segments of nucleic acids, e g., by genetic engineering techniques. Similarly, a recombinant protein is one encoded by a recombinant nucleic acid molecule or sequence. The recombinant nucleic acid or a protein can have been modified to contain desired traits or functions such as better bioavailability, increased stability in cells, etc.

[0036] The methods of the present disclosure include measuring RAS and mesenchymal biomarker expression levels in a biological sample. In the context of the present disclosure, gene expression level refers to the amount of gene product (protein or RNA) produced by a gene in a given biological sample, indicating gene activity under specific conditions. A person skilled in the art would be familiar with common methods for measuring gene expression at the protein level including methods such as, but not limited to, Western blotting, enzyme-linked immunosorbent assay (ELISA), and reporter gene assays. A person skilled in the art would also be familiar with common methods for measuring gene expression at the RNA level including methods such as, but not limited to, quantitative PCR (qPCR), reverse transcription PCR (RT-PCR), RNA sequencing (RNA-Seq), microarray assay, Northern blotting, and in situ hybridization (ISH). A person skilled in the art would understand that the methods ofmeasurement would depend on the biological sample type, experimental conditions, and the target RNA / protein to be measured.

[0037] In the context of the present disclosure, an individual can be said to be “[suitable] for cell differentiation therapy” or a specified treatment if the individual has identifiable factors or conditions that increase their likelihood of responding to the cell differentiation therapy. The identifiable factors or conditions can be based on a baseline that can be determined by one of skill in the art using appropriate means. Such factors can be, but are not limited to, presence or absence of a disease or symptom, biological markers (biomarkers) specific to a disease, as well as any other characterising factors considered to be pertinent by a person skilled in the art.

[0038] As used herein, the term, “mutation” in the context of an amino acid sequence or a protein refers to a change in the sequence of amino acids within a protein molecule as a result of e.g., genetic alteration, such as substitution, deletion, or insertion of one or more nucleotides in the coding sequence of a gene. Modifications in the amino acid sequence can lead to alterations in the structure, function, or properties of the protein, potentially affecting its biological activity, stability, interactions with other proteins, or binding specificity to other proteins. Mutation(s) (such as insertion or deletion) in an amino acid sequence can also result in a reading frame shift, leading to truncated or non-functional versions of the protein, catalytically inactive versions of the protein, or reduced protein levels. In the present disclosure, the term “mutated RAS”, for example, refers to a version of the RAS protein that comprises at least one mutation.

[0039] In the context of the present disclosure, the biological samples obtained from the subject can be, but are not limited to, a liquid sample, a tissue sample, a cell sample, and a nucleic acid sample. In one example, the liquid sample is a bodily fluid. In one example, the bodily fluid is selected from the group consisting of blood, bone marrow, cerebral spinal fluid, peritoneal fluid, pleural fluid, lymph fluid, ascites, serous fluid, sputum, lacrimal fluid, stool, urine, saliva, ductal fluid from breast, gastric juice / acid, and pancreatic juice. In one example, the bodily fluid is blood or blood plasma. In one example, the biological sample can be, but is not limited to, a tumour tissue, limited quantities of tumour cells (such as those present in blood), and cell free DNA present in blood. In one example, blood and tissue adjacent to the tumour obtained from the subject are used as “normal reference baseline” for identification of mutations and features unique to the tumour cells.DETAILED DESCRIPTION

[0040] HRAS, also known as GTPase HRAS, from “Harvey Rat sarcoma virus”, is a commonly mutated oncogene and driver in a variety of human cancers. However, targeting HRAS has been challenging due to the absence of a druggable binding pocket in the HRAS protein.

[0041] Lineage transdifferentiation offers an approach for treating highly malignant mesenchymal-like carcinomas, which are often able to survive standard of care therapy. Described herein is the extent of cellular plasticity by demonstrating extreme mesenchymal transition (xMT) and how it is used to regress carcinomas into more benign forms.

[0042] The term “transdifferentiation”, as used herein, refers to a process by which a mature somatic cell is converted into another mature somatic cell without undergoing an intermediate pluripotent state or progenitor cell type. In other words, the intermediate stem cell stage is skipped when a cell is transdifferentiated. This process can occur naturally or can be induced experimentally through the overexpression of specific transcription factors or exposure to chemical treatments.

[0043] Cellular plasticity is a hallmark of development, whereby cells undergo phenotypic changes to produce diverse functional cell types Most notably, embryonic stem cells and tissue stem cells undergo stepwise differentiation to enable organismal development, organ function, as well as tissue homeostasis and repair. Although less common, this hierarchical order is shown to be reversible, as exemplified by spontaneous de-differentiation of committed cell types towards the stem cell state. For instance, differentiated airway epithelial cells can revert into stable and functional stem cells in vivo, while heart regeneration can occur via cardiomyocyte dedifferentiation. The epithelial-to-mesenchymal transition (EMT) represents an example of cellular plasticity that elicits the phenotypic shifts of epitheli al -like cells, characterized by cellcell adhesion and apical-basal polarity, into mesenchymal -like cells that are spindle-shaped and do not form epithelial sheets. The acquisition of mesenchymal traits, in specific contexts, also confers cells with more stem-like and migratory characteristics, which are crucial for organ development. Moreover, this process has been shown to be exploited by cancer cells to facilitate disease progression, especially in overcoming barriers of the invasion metastasis cascade and to increase cancer stem-like properties.

[0044] Epithelial-to-mesenchymal transition (EMT) and its reverse - mesenchymal-to-epithelial transition (MET) - are dynamic, reversible processes. Without being bound by theory, while EMT facilitates invasion and dissemination of cancer cells from the primary tumour, thesubsequent activation of the MET is thought to be crucial for colonization of distant metastatic sites as cells in the epithelial state tend to be highly proliferative.

[0045] The epithelial and mesenchymal features were initially ascribed as binary cell states, whereby distinct phenotypic and molecular traits have specific roles in development and disease pathogenesis. More recently, epithelial-to-mesenchymal transition has been thought to encompass a continuous spectrum of amalgamated epithelial and mesenchymal characteristics, including a hybrid epithelial-to-mesenchymal transition state. Carcinoma cells possessing an intermediate phenotypic state along the epithelial-to-mesenchymal transition spectrum have been shown to be more malignant and successful at metastatic dissemination. Such hybrid states, marked by the co-expression of both epithelial and mesenchymal markers, have also been previously observed in circulating tumour cells of breast cancer patients.

[0046] Breast tumours are infiltrated by a variety of stromal components including cancer-associated fibroblasts (CAFs) and immune lymphocytes; these cells tend to be recruited from the surrounding tissue parenchyma or derived from the bone-marrow. However, it had been suggested that tumours can modify the niche by creating their own source of fibroblasts through a presumptive epithelial-to-mesenchymal transition-like process. For instance, a breast carcinoma-derived cell line exhibited robust fibroblastic traits but still contained residual keratin expression and formed epithelial foci. However, the experimental elucidation and mechanistic bases for putative transdifferentiation have remained unclarified. The incidence of transdifferentiation events in human tumours and the clinical impact of transdifferentiation on cancer progression need to be ascertained.

[0047] It has been shown that a more mesenchymal phenotype tends to reduce the tumorigenic properties of cancer cells. This points to the manipulation of cell states towards a more mesenchymal nature as a strategy of controlling cancer progression.

[0048] Discloses herein are therefore methods of exploiting or using the plasticity of cancer cells, for example, but not limited to breast cancer cells, for use in therapy or as a treatment for cancer. It had previously been demonstrated that epithelial-to-mesenchymal transition and mesenchymal-to-epithelial transition confer functional consequences in determining disease outcomes, as well as allowing the development of therapies and means to modulate the tumorigenic property of cancer cells.

[0049] Using small molecule drug library screens for cell state perturbations, the ability of mesenchymal cancer cells to transdifferentiate into fibroblast cells or fibroblast-like cells Fibroblast cells are defined by the expression of fibronectin, vimentin, SMA (a-smooth muscleaction), FAP (Fibroblast activation protein-a) and FSP (Fibroblast-specific protein (FSP-I )) Fibroblast-like cells are defined by the expression of a subset of these markers was observed, which has been termed “extreme mesenchymal transition” or “xMT” in the context of the present disclosure.

[0050] Generally speaking, the inhibition of farnesyltransferase (FTase; an enzyme of the prenyltransferase group) redirects farnesyl pyrophosphate (FPP) towards the accumulation of geranylgeranyl pyrophosphate (GGPP). Activation of geranylgeranyl ati on results in the prenylation of key cytoskeletal and fibroblast associated proteins that include the Rho GTPases RAC1 and CDC42, which activate c-JUN N-terminal Kinase (JNK). This, in turn, phosphorylates components of the AP-1 transcription complex that reprogram mesenchymal breast cancer cells into fibroblasts.

[0051] Shown herein is data indicating that farnesyltransferase inhibitors (FTls) mediate transdifferentiation of cancer cells into fibroblast-type cells and restrict tumorigenicity. Thus, in one example, transdifferentiation therapy is disclosed herein as a strategy to perturb or effect tumour growth.

[0052] It is further shown herein that cancer to fibroblast transdifferentiation involves the reprogramming of a mesenchymal cancer cell to an extreme mesenchymal state. Without being bound by theory, farnesyltransferase inhibitors are thought to lead to the geranylgeranylation of Rho GTPase family members, RAC1 and CDC42, which in turn activates JNK signalling and API complex. Consequently, this is thought to lead to the upregulation of fibroblast genes, resulting in transdifferentiation.

[0053] Thus, in one example, the method disclosed herein results in extreme mesenchymal transition (xMT), whereby mesenchymal-like cancer cells are induced to transdifferentiate into fibroblasts.

[0054] As disclosed herein, treatment of HRAS-dependent cancer cells with farnesyltransferase inhibitors (FTIs) is shown to convert these cancer cells into fibroblasts, and in doing so, control tumour growth.

[0055] As shown herein, farnesyltransferase inhibitors (FTIs) are identified, which are shown to regulate cell state to drive fibroblast transdifferentiation in mesenchymal breast cancer cells.

[0056] The data shown herein provides a mechanistic basis for the extreme mesenchymal transition (xMT) as described herein, showing that cancer cells, under specific stimuli, undergo extreme mesenchymal transition. This disclosure therefore represents another route towards modulating cell fates to control malignant progression.Inhibition of fa.rnesyltrans erase induces transdifferentiation of mesenchymal breast cancer cells into a fibroblastic phenotype

[0057] To understand the control of cell state transition and the degree of cellular plasticity that cancer cells can possess, a high-throughput chemical library screen was performed to identify molecules that were capable of converting invasive NAMEC8 (naturally arising mesenchymal epithelial cells, cell line 8) breast cells residing in the mesenchymal cell state to become more epithelial via mesenchymal-to-epithelial transition (MET). In general, cell state is defined by the cellular phenotype arising from the expression of specific molecular markers and acquisition of specific morphological features. Cells in the epithelial cell state express markers like E-cadherin and cytokeratins and show cell-cell adhesion and cobblestone-like organisation

[0058] Cytokeratins are the largest group of intermediate filaments. They are filamentous proteins which, together with other filaments, form the eukaryotic cytoskeleton. They have numerous functions, including, but not limited to, maintenance of epithelial structure, protection from injury, and communication with other cytoplasmic components. They are expressed in pairs, with expression patterns varying according to location, and classified numerically from 1 to 20 according to their molecular weight and isoelectric point. There are 2 major groups of cytokeratins: simple epithelial cytokeratins (CK7, CK8, CK18, CK19, and CK20) and more complex epithelial cytokeratins, such as those found in the skin (CK5 / 6, CK10, CK14, and CK15).

[0059] Cells in the mesenchymal cell state express markers like N-cadherin, Zebl, Snail, Slug, Twist and show spindle-like morphology and cell scattering (lacking cell-cell contacts).

[0060] While many compounds did not alter cell states, compounds, including retinoids, were shown to able to induce mesenchymal-to-epithelial transition to varying degrees based on the SERPINE1 -promoter luciferase reporter activity. Through this reporter screen, twelve (12) compounds that redirect the mesenchymal cell state towards a more epithelial phenotype, were identified. Of the 12 identified compounds, one molecule - tipifamib (FIG. 7A) - appeared to significantly upregulate SERPINE1 -promoter luciferase reporter activity. Tipifamib is shown herein to be able to consistently and significantly increase SERPINE1 -promoter luciferase activity in a dose-dependent manner (FIG. 7A) without affecting cell proliferation in NAMEC8, NAMEC8-KRAS, and human mammary epithelial (HMLE) cells. In NAMEC8-HRAS cells, tipifamib treatment reduced proliferation (FIG. 7B).

[0061] SERPINE1 is a mesenchymal marker known to be upregulated during epithelial -to-mesenchymal transition and is a marker of cytoskeletal changes associated with mesenchymalproperties Given that the NAMEC8 cells employed in the screen were already highly mesenchymal in nature, further upregulation of the SERPINE1 -promoter luciferase activity was unexpected.

[0062] Tipifarnib is a famesyltransferase (FTase) inhibitor that binds to and inhibits FTase, thereby preventing downstream protein farnesylation (FIG. 1A). To confirm the specificity of FTase inhibition, another famesyltransferase inhibitor, lonafarnib, which similarly upregulated the SERPINE1 -promoter luciferase activity (FIG. 7C) was also used. The resulting FTase inhibitor-treated cells exhibited a spindle-like and fibroblastic morphology (FIG. IB), which was quantified with phalloidin staining (FIG. 1C). Epithelial -to-mesenchymal transition (EMT) transcription factors, such as ZEB1 and SLUG, were upregulated transiently on day 1, suggesting a shift towards a more mesenchymal cell state, but decreased in expression over the course of 14 days of tipifarnib treatment (FIG. ID). Next, it was sought to deconvolute the identity of these tipifarnib-treated cells. RNA-seq gene expression analyses in NAMEC8-HRAS cells treated with tipifarnib for 7 days were performed, and the expression signatures were mapped to the ARCHS4 tissue databases. Signatures associated with cells of the mesenchymal tissue lineage, such as fibroblasts (Adj. p-value 3.62 x 10'35), vascular smooth muscle (Adj. p-value 2.25 x 10'39) and respiratory smooth muscle (Adj. p-value 7.10 x 10'33) emerged, indicating that the resultant cells could be fibroblastic in nature (FIG. IE and FIG. 7D).

[0063] Using patient-derived breast cancer cells and matched adjacent fibroblasts, a panel of biomarkers that could help to define the fibroblast marker signatures was selected and validated (FIG. 7E). Markers, such as fibroblast specific protein (FSP) and fibroblast activation protein (FAP), were shown to be unique to fibroblasts and not expressed within the epithelial tissue lineage. To differentiate between the epithelial lineage and fibroblast cells that might result from the transdifferentiation process, single-cell RNA-seq was performed on untreated and tipifarnib-treated NAMEC8-HRAS cells to compare their profiles against a cell reference panel. From the single-cell clustering analyses, it was shown that tipifarnib treatment induced a high degree of transcriptomic shift (FIG. IF). While untreated cells showed strong correlation with epithelial cell types, tipifarnib-treated cells possessed distinct signatures that correlated with tissues of mesodermal origin, such as smooth muscle, cardiomyocyte, and adipocyte (FIG. 7F). Tipifarnib treatment was shown to increase the acquisition of fibroblast archetype markers, such as FAP, FSP, smooth muscle actin (SMA) and nidogen 1 (NIDI), while retaining vimentin expression (FIG. 1G). Similarly, immunoblotting showed that tipifarnib-treated cells expressed vimentin, SMA, FAP, and FSP (FIG. 1H). Using another mesenchymal breast cancer cell line, SUM159,similar morphological changes were observed (FIG. II), as well as the acquisition of fibroblast markers with tipifamib treatment (FIG. 1J). This cell state transition is referred to herein as “extreme mesenchymal transition” (or “xMT”).

[0064] Thus, in one example, there is described a method of transdifferentiating a cell into a fibroblast-like stage, the method comprising contacting the cell to a famesyltransferase inhibitor (FTI). In another example, the method further comprises contacting the cell to an agent that increases the amount of geranylgeranyl pyrophosphate (GGPP).

[0065] To further elucidate the permissiveness of this cellular plasticity, HMLE cells, which reside in the epithelial cell state were exposed to tipifarnib. HMLE cells are parental cells from which the isogenic mesenchymal NAMEC8 cells are derived. In contrast, tipifarnib-treated HMLE cells did not exhibited any change in their morphology, nor did tipifarnib-treated HMLE cells express FSP and FAP (FIG. 1H). Three other cancer cell lines, RL95-2 (endometrial), KYSE30 (esophageal), and MCF7-HRAS (breast), all bearing an epithelial cell phenotype, were also exposed to tipifarnib. No indication of transdifferentiation into fibroblasts was observed (FIG. 7G to FIG. 71). This highlights the cellular plasticity of cells residing within the mesenchymal cell state, but not epithelial cell state, and the propensity of these cells to undergo xMT transdifferentiation.

[0066] The induction of extreme mesenchymal transition (xMT) and fibroblast transdifferentiation indicated a switch from a malignant cell type to a benign cell type, which was thought to decrease cellular motility and invasive properties. To assess this, a wound-healing assay with control and tipifarnib-treated NAMEC8-HRAS cells was performed. Tipifarnib-treated cells showed a delay in wound closure compared to control cells (FIG. 7J). Transwell invasion assay also showed that tipifarnib could, on average, inhibit cellular invasion and migration (FIG. 7K).

[0067] In another example, the method is a method of treating RAS-dependent cancer in a subject, wherein the method comprises administering a pharmaceutically effective amount of a famesyltransferase inhibitor (FTI) to a subject in need thereof. In one example, the method further comprises administering an agent that increases the amount of geranylgeranyl pyrophosphate (GGPP).

[0068] In another example, the famesyltransferase inhibitor and the agent that increases the amount of GGPP are administered either together in a single or separate compositions, separately, or sequentially. In a further example, sequential administration involves a. the first administration of the famesyltransferase inhibitor and a subsequent administration of the agentthat increases the amount of GGPP; or b. the first administration of the agent that increases the amount of GGPP and a subsequent administration of the famesyltransferase inhibitor.

[0069] In another example, there is described a composition comprising a famesyltransferase inhibitor (FTI); and an agent that increases the amount of geranylgeranyl pyrophosphate (GGPP). In a further example, the composition is a pharmaceutical composition.

[0070] Next, it was sought to understand the nature of extreme mesenchymal transition (xMT) transdifferentiation and, specifically, its impact on the tumorigenicity of mesenchymal-like cancer cells. An amount of 5xl05NAMEC8-HRAS cells that had been treated with tipifamib in vitro, or that had been vehicle-treated for 7 days, were subcutaneously transplanted into immunocompromised NOD-SCID-I12rg209 null (NSG) mice. Tipifamib-treated NAMEC8-HRAS cells produced smaller tumours than their respective controls (FIG. 2A). Immunofluorescence staining revealed upregulated expression of fibroblast markers: SMA, FSP and fibronectin, but not cytokeratin 5 (CK5), which is an epithelial tissue lineage marker (FIG.2B). This indicates that short-term exposure to tipifamib in culture is capable of programming mesenchymal cells to undergo xMT transdifferentiation in vivo. Likewise, cells of the HRAS-driven breast cancer cell line, SUMI 59, transiently exposed to tipifamib in vitro, also displayed similar reductions in tumour mass (FIG. 8A). The observation of an oncogenic activity inhibition that disrupts tumorigenicity through the induction of a benign cell fate or phenotypic switch had not been previously reported.

[0071] To further evaluate the effects of extreme mesenchymal transition (xMT) on the tumorigenic potential of mesenchymal-like cancer cells in vivo, untreated NAMEC8-HRAS cells were implanted into NSG mice. When tumour volumes reached 100 mm3, mice were dosed with 200 mg / kg / day tipifamib daily, for 7 days. Tumour growth arrests were observed in tipifarnib-treated mice compared to vehicle-treated mice (FIG. 2C). Tipifarnib-treated tumours were shown to have upregulated expression of human-specific SMA, FSP and FAP but downregulated CK5 expression (FIG. 2D). Since these human-specific antibodies were co-stained with exogenously introduced GFP+cancer cells, the in vivo results confirmed that tipifamib treatment induced direct transdifferentiation of cancer cells to become fibroblasts. This confirmed that the SMA+, FAP+or FSP-cells were not recruited mouse stroma. Histopathological analyses of tumours arising from in vivo tipifamib-treated NAMEC8-HRAS (FIG. 2E) and SUM159 cells (FIG. 2B) revealed the increased abundance of extracellular matrices and fibrocollagenous tissue. Fibrocollagenous tissue is characterized by the wide scattering of fibroblasts and the abundance of extracellular matrix components with collagen and elastic fibres being dominant. To confirmthe fibrocollagenous phenotype, Masson’s trichrome staining was performed on tumour sections where collagen was stained blue. In tipifamib -treated tumours, an increased in collagen was seen in areas with the elongated cell morphology characteristic of fibroblast (FIG. 2F). These observations show that the induction of cellular transdifferentiation was associated with the reduction in tumorigenicity.

[0072] Thus, in one example, the cells disclosed herein are differentiated or transdifferentiated into fibroblast cells. In another example, cells of RAS-dependent cancers are differentiated into fibroblast cells.

[0073] The above findings led to questions on the extent of cellular transdifferentiation and its impact on tumorigenicity in other functional studies, which also examined the aggressiveness of cancer cells within the epitheli al -to-mesenchymal transition (EMT) spectrum. Bulk transcriptomic profiles of untreated and tipifarnib -treated cells were compared with previously defined sub-populations of mouse squamous cell carcinoma tumour cells that were defined along the EMT spectrum. Hierarchical clustering first revealed the close similarity between untreated NAMEC8-HRAS cells and the aggressive epithelial -like Epcam+carcinoma subpopulation (FIG 8C). Tipifarnib-treated NAMEC8-HRAS cells were shown clustered closest to the most mesenchymal and least tumorigenic CD51+ / CD61+and CD51+ / CD61+ / CD106“ sub -populations (FIG. 8C). These analyses showed that tipifamib induced an extreme mesenchymal state that is associated with reduced tumorigenicity. To further examine if transdifferentiation of epithelial cells into fibroblasts could be observed in human cancers, it was thought that the co-occurrence of epithelial and fibroblast markers would indicate this possibility. Single-cell transcriptomic studies were interrogated to assess the presence of cancer epithelial cells that co-expressed fibroblast marker, FSP, in human tumours. Across several cancer types, rare populations of FSP+cells were found to reside within the epithelial compartment, marked by the expression of keratins and EPCAM (FIG. 8D). This showed that a small subset of epithelial cancer cells possessed the potential to undergo transdifferentiation toward a fibroblast-like phenotype in human tumours.

[0074] Famesyltransferase is a heterodimeric enzyme. The alpha subunit is conserved between all prenyltransferases, but its beta subunit (encoded by FNTB) is specific to the famesyltransferase enzyme. To confirm that the induction of xMT transdifferentiation was indeed due to the loss of FTase function, shRNA knockdown (KD) of FNTB in NAMEC8-HRAS cells was performed. The FNTB-KD led to a loss of FTase mRNA expression (FIG. 9A) that was accompanied by decreased pan-famesylated protein expression (FIG. 9B) and cellularproliferation (FIG. 9C) Compared to shControl cells, FNTB-KD cells recapitulated the spindlelike fibroblast morphology (FIG. 3A) and acquired fibroblast-specific markers previously seen in tipifamib-treated cells (FIG. 3B), as well as showing reduced soft agar colony-forming ability (FIG. 9D). Moreover, in agreement with tipifamib-treated NAMEC8-HRAS cells, FNTB-KD cells (shl) showed a delay in wound closure compared to control cells (FIG. 9E). In mice, FNTB-KD resulted in decreased xenograft tumour growth (FIG. 3C). Residual tumours displayed a fibrocollagenous tumour architecture, characterized by high collagen levels (FIG. 9F and 9G) that was accompanied by the gain in FSP, FAP, SMA and fibronectin expression, as well as the loss of epithelial CK5 marker (FIG. 3D).

[0075] To further support that xMT transdifferenti tion of mesenchymal cancer cells occurred in response to FTase inhibition and that xMT transdifferentiation was not due to compromised cell viability in vivo, doxycycline (Dox)-inducible shRNA knockdown was utilised. In cell cultures, Dox induction of FNTB-KD led to loss of pan-farnesylation (FIG. 9H), as expected. Next, uninduced FNTB-KD cells were xenografted into NSG mice. When tumours reached 100 mm5, mice were fed doxycycline drinking water (2 mg / ml in 30% fructose) to activate FTase shRNA expression. In vivo induction of FNTB-KD arrested tumour growth (FIG. 3E). Compared to Dox-induced shControl tumours, the residual FNTB-KD tumours were fibrocollagenous in nature (FIG. 91), increased the expression of fibronectin, FSP and SMA, and lost CK5 expression (FIG. 3F). To assess if transdifferentiation affected metastatic potential, NAMEC8-HRAS cells were pre-treated with tipifarnib, 500nM in vitro for 7 days and the cells injected intravenously into NSG mice to recapitulate lung metastasis. Compared to untreated cells, tipifarnib -treated cells delayed metastasis-associated death by approximately 0.3 months (FIG. G), suggesting that transdifferentiation alters metastatic potential.

[0076] In one example, the famesyltransferase inhibitor is, but is not limited to, the group consisting of tipifarnib, lonafamib, manumycin A, a-hydroxy famesyl phosphonic acid, and combinations thereof. In one example, the famesyltransferase inhibitor is tipifarnib. In another example, the famesyltransferase inhibitor is lonafamib (data not shown).

[0077] Thus, the data shown herein confirms that the on-target inhibition of FTase results in transdifferentiation of mesenchymal-like cancer cells to become fibroblasts in vitro and in vivo via the activation of xMT. The data presented herein shows that disruption of protein prenylation can program a previously unreported transdifferentiation toward an alternate tissue lineage. Inhibition ofMAPK pathway in HRAS-dependent mesenchymal cells promotes xMT

[0078] To determine the mechanism by which inhibition of farnesylation-mediated xMT, two possibilities were evaluated. First, FTase inhibition inactivates HRAS and its canonical downstream signalling pathways (FIG. 4A) Without being bound by theory, it was thought that exposure of NAMEC8-HRAS cells to tipifamib leads to the downregulation of the RAF-MEK-ERK mitogen-activated kinase (MAPK) pathway. To assess if direct disruption of the MAPK pathway led to xMT transdifferentiation, RAFI was inhibited with sorafenib. MEK was inhibited with a highly selective inhibitor, PD 98059. By 7 days of treatment, cells adopted an elongated morphology characteristic of fibroblasts (FIG. 4B) There was a gain in FAP and FSP in sorafenib-treated NAMEC8-HRAS cells and FSP expression in MEK inhibitor-treated cells (FIG. 4C), thus indicating the involvement of the MAPK pathway in transdifferentiation. To further confirm this, shRNA-mediated knockdown of RAFI was performed in NAMEC8-HRAS cells. In line with sorafenib treatment, RAFI knockdown resulted in cells with elongated morphology (FIG. 4D), dysregulated MAPK pathway, and upregulated expression of fibroblast markers (FIG. 4E). NAMEC8-HRAS RAF1-KD cells downregulated HRAS-GTP (FIG. 4F). In contrast, xMT was weakly induced in sorafenib -treated NAMEC8-KRAS cells, marked by weak morphological changes and a slight increase in expression of FSP and FAP (FIG. 10A and B], This showed that the induction of xMT was most effectively induced in cells with activated HRAS-dependent RAF-MEK-ERK pathway.Rechanneling of farnesy I pyrophosphate (FPP) towards geranylgeranylation promotes xMT

[0079] Without being bound by theory, it was thought that a block in HRAS prenylation could have other unexpected consequences. Besides its role in protein farnesylation, farnesyl pyrophosphate (FPP) can be converted to geranylgeranyl pyrophosphate (GGPP) through geranylgeranyl pyrophosphate synthase 1 (GGPS1) (FIG. 5 A). GGPP is a key substrate for geranylgeranylation of proteins - a post-translational modification catalysed by geranylgeranyltransferase (GGTase) to attach a 20-carbon lipid tail onto the CaaX motif of designated proteins. To detect changes in the levels of total farnesylated and geranylgeranylated proteins, CLICK chemistry, which relies on copper to act as a bioconjugate of an azide-tagged substrate and an alkyne fluorophore-detection molecule for visualization, was employed.

[0080] In untreated NAMEC8-HRAS cells, there was an abundance of farnesylated proteins, which dramatically decreased upon tipifarnib treatment over the course of 7 days. More importantly, there was a concomitant increase in geranylgeranylated proteins (FIG. 5B). Owing to the accumulation of FPP, it was thought that a flux towards GGPP was created. This indicatesthat FTase inhibition induced a shift in protein prenylation from farnesylation to geranylgeranylation.

[0081] It was sought to understand how geranylgeranylation could cause xMT. Since GGPS1 converts FPP to GGPP, its functional disruption was thought to alter the levels of available GGPP and thereby abrogate FTase inhibitor-induced transdifferentiation. As a pharmacological inhibitor of GGPS1 was not available, GGPS1 was knocked down, which decreased its expression by up to 4-fold (FIG. 10C). NAMEC8-HRAS GGPS1-KD cells neither produced a fibroblast-like morphology nor gained expression of fibroblast markers upon tipifarnib treatment compared to their shControl counterparts (FIG. 5C and D). To directly and more conclusively demonstrate that GGPP was a component driving xMT, GGPS1-KD cells were supplemented with exogenous GGPP to bypass the loss of GGPS1 activity. It was shown that GGPP supplementation was sufficient to restore tipifarnib -induced fibroblast protein expression in GGPS1-KD cells in cell cultures (FIG. 5D). Thus, GGPP supplementation rescued the phenotype of GGPS1-KD cells and provided direct evidence that geranylgeranylation is involved in driving x\IT

[0082] In one example, administration of a famesyltransferase inhibitor (FTI) and an agent that increase the amount of GGPP increases the levels of GGPP in the subject or in a cell. In another example, administration of the famesyltransferase inhibitor and agent that increase the amount of GGPP increases the amount of geranylgeranylated proteins in the subject or in a cell.

[0083] NAMEC8-HRAS GGPS1-KD cells were transplanted into NSG mice to evaluate the contributions of GGPP in vivo. When tumour volumes reached 100 mm3, tipifarnib was delivered intraperitoneally for 7 days before intratumoral supplementation with GGPP (11 pM) was introduced to the right flank tumour every 2 days, while the contralateral left flank tumour was vehicle-treated (FIG 5E). After 3 rounds of dosing, the tumours were resected. Consistent with cell cultures, GGPP supplementation, on its own, increased the abundance of fibroblast-like cells within the right flank tumours only (FIG. 10D). These resected tumours expressed FAP and FSP, which colocalized with GFP expression, indicating transplanted cells had undergone transdifferentiation (FIG. 5F). Thus, this underscored the association between increased geranylgeranylation and xMT. To demonstrate the contribution of geranylgeranylation to the induction of xMT, GGTasel was inhibited using a small molecule inhibitor (GGTI 298) in NAMEC8-HRAS cells. Indeed, tipifarnib treatment in cells pre-treated with GGTase inhibitor were shown to resist xMT as shown by decreased SMA, FAP and FSP expressions (FIG. 5G).

[0084] It was sought to investigate how geranylgeranylation directed cellular transdifferentiation. The major targets of geranylgeranylation include RAC1, CDC42 and RHOA, which are members of the Rho family of GTPases belonging to the RAS superfamily; these signalling molecules need to be geranylgeranylated for lipid anchoring to the inner cell membrane. These proteins are known to be associated with cell motility and cytoskeletal remodelling, as well as being biomarkers of the fibroblast lineage. In tipifarnib -treated or FNTB-KD NAMEC8-HRAS cells, GGPS1 was upregulated, in parallel with RAC1, CDC42 and RHOA (FIG. 5H), as were Rho GTPase-family effectors, p21 -activating proteins, PAK1 and PAK2, and LIM domain kinases, LIMK1 and LIMK2, which are RAC1 and CDC42 effectors (FIG. 51). In contrast, ROCK1 and ROCK2, which are RHOA effectors, were weakly expressed in tipifamib-treated or FNTB-KD NAMEC8-HRAS cells (FIG. 51). Additionally, RHOA, RAC1 and CDC42 expressions were rescued with exogenous supplementation of GGPP in tipifarnib -treated GGPS1-KD cells (FIG. 10E). These findings prompted us to assess whether the increased Rho family of GTPases expressions would correlate with an increase in their activity. Quantification of GTP-bound proteins showed that during tipifarnib treatment, RAC1-GTP and CDC42-GTP levels were increased (FIG. 5 J). In contrast, RHOA did not show changes in activation state upon tipifarnib treatment (FIG. 5J). Small molecule inhibition of RAC1 (with compound NSC 23766) (FIG. 5K) and CDC42 (with compound ML 141) (FIG. 5L) were able to reduce the extent of xMT, as demonstrated by the decrease in gain of fibroblast markers FAP and FSP by day 7 of treatment, thereby indicating that RAC1 and CDC42 have active roles in facilitating xMT transdifferentiation.

[0085] In the context of cancer, tipifarnib blocks HRAS activity through the inhibition of famesylation. However, in contrast to HRAS which can only be activated by farnesylation, KRAS and NRAS can be activated by both geranylgeranylation and famesylation. Thus, KRAS-and NRAS-driven cancer cells might be able to bypass famesylation-specific inhibition through geranylgeranylation as an alternative protein prenylation mechanism. This led us to reason that the specific nature of RAS oncogene addiction might detennine the propensity of mesenchymal-like cancer cells to undergo xMT. To test this hypothesis, NAMEC8-KRAS cells that were exposed to tipifarnib alone or contained FNTB shRNA were analysed Unlike NAMEC8-HRAS cells, they were unable to undergo xMT in both contexts. They neither recapitulated the spindle like morphology (FIG. 10F) nor expressed fibroblast markers (FIG. 10G). Next, it was assessed whether KRAS was able to sufficiently divert the geranylgeranylation away from Rho GTPases, thereby preventing fibroblast transdifferentiation. An increase in KRAS-GTP upon tipifarnibtreatment in NAMEC8-KRAS cells was observed, indicating the geranylgeranylation and subsequent activation of KRAS (FIG. 10H). In line with the lack of transdifferentiation, tipifamib treatment alone did not affect the growth of NAMEC8-KRAS tumours (FIG. 101) or resulted in the formation of fibroblast-like cells of human origin (FIG. 10J). Lastly, NAMEC8 cells overexpressing both mutant-HRAS and -KRAS exposed to tipifamib were unable to adopt a clear fibroblast phenotype (FIG. 10K) and did not display the same drastic increase in fibroblast proteins expressions (FIG. 10L). The channelling of GGPP towards oncogenic KRAS function, leading to the observed upregulation of KRAS-GTP (FIG. 10M) indicates that a reduction in the geranylgeranylation of other targets that are essential for driving xMT. Taken together, the data shows that geranylgeranylation was responsible for cancer transdifferentiation under circumstances where GGPP can be effectively channelled towards the prenylation of Rho GTPases, such as in HRAS-driven cancers.

[0086] Thus, in one example, the cancer is a RAS-dependent cancer. In another example, the RAS-dependent cancer is a HRAS-, KRAS-, or NRAS-dependent cancer. In yet another example, the cancer is a HRAS-dependent cancer.

[0087] In another example, the RAS-dependent cancer is, but is not limited to, breast cancer, bladder cancer, head and neck squamous cell cancer, lung cancer, thyroid cancer, and melanoma.

[0088] In yet another example, the cells of the RAS-dependent cancer are in the mesenchymal state. The characterisation of the mesenchymal state in cells can be determined, for example, based on the presence of absence of any one or more of the following markers N-Cadherin, Zeb I, Snail, Slug, Twist, and SMA.

[0089] In another example, there is disclosed a method of identifying a subject’s suitability for cell differentiation therapy, wherein the method comprises: obtaining a biological sample from the subject; measuring the expression level of one or more RAS oncogenes; and measuring the expression level of at least one or more mesenchymal biomarker as disclosed herein, and wherein the presence or increase in a number of mutations in the RAS oncogene or in the expression level of the one or more RAS oncogenes and the number of mesenchymal biomarkers indicates suitability of the subject for the cell differentiation therapy. In one example, a number of mutations in the RAS oncogene, or the expression level of the RAS oncogene, and / or the number of mesenchymal biomarkers is compared to a subject not known to suffering from a disease, or not known to be suffering, or not suffering from cancer.

[0090] In one example, the expression levels referred to herein can be, but are not limited to, expression at the protein, RNA, and / or mRNA level.

[0091] In another example, the cell or subject disclosed herein is exposed to an agent that increases the amount of geranylgeranyl pyrophosphate (GGPP). Such agents can be, but are not limited to, exogenous GGPP, exogenous recombinant GGPP, lipid-based delivery system comprising GGPP, nanoparticles comprising GGPP, and a combination thereof.

[0092] In another example, the biological sample is selected from the group consisting of tumour tissue, tissue adjacent to the tumour tissue and blood. In one example, the biological sample is obtained from a subject suffering or thought to be suffering from cancer. In another example, the biological sample is a reference sample obtained from a subject not suffering from cancer.Prenylation of RAC1 and CDC42 results in the downstream activation of AP-1 for the transcriptional regulation of xMT

[0093] While geranylgeranyl ati on of mesenchymal cell-associated Rho GTPases appeared to be a mechanism necessary for xMT, their direct activation alone was unlikely to fully account for dramatic phenotypic shifts. Transcription factors are key drivers that exert a pleiotropic effect on cellular plasticity or drive cell state transitions. To dissect the transcriptional control of xMT, bulk RNA seq gene expression analyses of NAMEC8-HRAS tipifamib-treated and FNTB-KD cells was first performed In both instances, perturbation to the transcription regulatory network associated with Activator Protein 1 (AP-1) was among the most significantly enriched gene sets (FIG. 6A and 6B). The AP-1 complex is composed of heterodimerized JUN and FOS subunits, which are phosphorylated for AP-1 complex activation.

[0094] Thus, in one example, the increase in GGPP is determined by the increase in the amount of geranylgeranylated proteins in the subject compared to a subject not suffering from cancer.

[0095] To understand how AP-1 might be activated for the promotion of xMT, two possible routes were examined. First, protein kinase C (PKCa and PKCo) have been established to activate AP-1 in a calcium signalling-dependent manner. Calcium channels were blocked via amlodipine in NAMEC8-HRAS cells for two days before tipifarnib exposure. The perturbation of calcium signalling did not significantly disrupt the gain of fibroblast markers (FIG. 6C) or affect cell proliferation (FIG. 11 A). Second, AP-1 is a direct phosphorylation target of c-Jun N-terminal kinases (JNK1 and JNK2). Importantly, RAC1 and CDC42 are known activators of JNK. Having already shown that RAC1 and CDC42 inhibition ablated the gain in FAP and FSP (FIG. 5K and 5L), it was thought that the RAC 1 / CDC42-JNK- AP-1 axis could be responsible for the transcriptional regulation of xMT. It was shown that members of the AP-1 subunits (J UNB,JUND, C-JUN and C-FOS) were generally upregulated in their gene transcripts or protein levels, especially their phosphorylated counterparts (p-C-FOS and p-C-JUN), upon tipifamib treatment (FIG. 6D and 6E; FIG. 11B and 11C] As JNK is known to phosphorylate certain AP-1 subunits (C-JUN and C-FOS), NAMEC-HRAS cells pre-treated with tipifamib were subjected to a JNK inhibitor, SP600125. This reduced the gain of FAP and FSP (FIG. 6D), similar to what was observed upon RAC1 and CDC42 inhibition but did not affect cell proliferation (FIG. 11D). Concomitantly, JNK inhibition resulted in decreased expression of p-C-JUN and p-C-FOS (FIG.6D).

[0096] Finally, to confirm the role of AP-1 in directing xMT, the function of AP-1 was disrupted with a small molecule inhibitor, SR11302. While loss of AP-1 activity in NAMEC8-HRAS cells exposed to tipifamib did not affect cell proliferation (FIG. 1 IE), this led to reduced phosphorylated C-JUN and C-FOS levels, and the inability to gain FAP and FSP (FIG. 6E) and adopt fibroblast morphologies (FIG. 6F). Moreover, JNK signalling and AP-1 activation were perturbed by the inhibition of RAC 1 (FIG. 1 IF) and CDC42 (FIG. 11G), thereby demonstrating the link between RAC1 and CDC42 activation, and their downstream JNK / AP-1 signalling Mechanistically, the switch from famesylation to geranylgeranylation led to the activation of RAC1 and CDC42 that promoted 468 JNK-dependent AP-1 phosphorylation, which resulted in the transcriptional control of xMT trans differentiation (FIG. 6G). While FTase inhibitors such as tipifamib have been conventionally thought to mediate their effects through the blockade of growth-promoting RAS signalling, it is shown here that they also have roles in abrogating tumorigenicity through lineage transdifferentiation.

[0097] Thus, in one example, the administration of the farnesyltransferase inhibitor alone increases the activity of Activator Protein 1 (AP-1). In another example, the administration of the farnesyltransferase inhibitor and an agent that increase the amount of GGPP increases the activity of Activator Protein 1 (AP-1).

[0098] In another example, the increase in GGPP is determined by the increase in the amount of geranyl geranylated proteins in the subject compared to a subject not suffering from cancer.

[0099] The Waddington’s epigenetic landscape depicts multiple unidirectional paths that a cell may take towards tissue lineage and cell fate commitment. The resulting metastable cell states preserve the cellular identity, and molecular barriers prevent or restrict dramatic shifts in cellular phenotypes. These phenotypic shifts within the epithelial tissue lineage pose low energy barriers, thereby permitting some cancer cells to manifest functional plasticity that enables disease progression. Overexpression of Yamanaka factors’ or exposure to combinations ofpathway-specific inhibitors robustly reprograms differentiated and lineage-committed cell types such as fibroblasts and peripheral blood cells into induced pluripotent stem cells (iPSCs). Likewise, overexpression of tissue lineage-specific factors such as Gata4 / Mef2c / Tbx5 and MyoDl induced the transdifferentiation of fibroblasts into differentiated cardiomyocytes and myoblasts, respectively, while T lymphocytes could be transdifferentiated into neuronal cells with Brn2 / Ascll / Mytll / Ngn2 transcription factor combination. The observations point to the plasticity inherent to certain cell types, which under the appropriate stimuli, can facilitate transdifferentiation into specified cell lineages.

[0100] In the context of cancer, otherwise lineage-restricted mammary gland cells can undergo lineage switches with oncogenic PIK3CA stimuli to give rise to different tumour subtypes. Exposure of murine and human mesenchymal breast cancer cells to a combination of MEK inhibitors and Rosiglitazone was found to transdifferentiate metastatic breast cancer into adipocytes as shown by lipid accumulation in the periphery of tumours. While transdifferentiation into adipocytes did not ablate the tumorigenic capacity of mesenchymal cancer cells, it was able to reduce the formation of metastatic lesions

[0101] The present disclosure shows that mesenchymal like cancer cells can undergo dynamic phenotypic switching within tumours beyond their otherwise restricted tissue lineage (FIG. 6H). Extreme mesenchymal transition (xMT) was shown to program mesenchymal breast cancer cells to acquire the cell identity of fibroblasts, resulting in the abrogation of tumorigenic properties.

[0102] In summary, at least two prerequisites are necessary for xMT. First, xMT through FTase inhibition requires the addiction of a cancer cell to HRAS activity and its downstream signalling pathways. The consequent accumulation of GGPP from FPP is rechannelled towards geranyl geranyl ati on of RAC1 and results in the activation of downstream JNK The phosphorylation of AP-1 complex proteins, in turn, directs the transcriptional reprogramming of mesenchymal carcinoma cells toward the fibroblast lineage, as seen in NAMEC8-HRAS cells and the HRAS-driven mesenchymal triple-negative breast cancer cell line, SUMI 59. This observation can be abolished through the concomitant overexpression of KRAS in the same cell, thought to be due to the sustained demand for GGPP utilization in KRAS activation. Second prerequisite is that the plasticity that resides within a mesenchymal cell state, wherein only the mesenchymal NAMEC8, but not epithelial HMLE cells, are capable of undergoing tran sdi fferenti ati on .

[0103] The invention illustratively described herein may suitably be practiced in the absence of any element or elements, limitation or limitations, not specifically disclosed herein. Thus, for example, the terms "comprising", "including", "containing", etc. shall be read expansively and without limitation. Additionally, the terms and expressions employed herein have been used as terms of description and not of limitation, and there is no intention in the use of such terms and expressions of excluding any equivalents of the features shown and described or portions thereof, but it is recognized that various modifications are possible within the scope of the invention claimed. Thus, it should be understood that although the present invention has been specifically disclosed by preferred embodiments and optional features, modification and variation of the inventions embodied therein herein disclosed may be resorted to by those skilled in the art, and that such modifications and variations are considered to be within the scope of this invention.

[0104] As used in this application, the singular form “a,” “an,” and “the” include plural references unless the context clearly dictates otherwise. For example, the term “a genetic marker” includes a plurality of genetic markers, including mixtures and combinations thereof.

[0105] As used herein, the term “about”, in the context of concentrations of components of the formulations, typically means + / - 5% of the stated value, more typically + / - 4% of the stated value, more typically + / - 3% of the stated value, more typically, + / - 2% of the stated value, even more typically + / - 1% of the stated value, and even more typically + / - 0.5% of the stated value.

[0106] Throughout this disclosure, certain embodiments may be disclosed in a range format. It should be understood that the description in range format is merely for convenience and brevity and should not be construed as an inflexible limitation on the scope of the disclosed ranges. Accordingly, the description of a range should be considered to have specifically disclosed all the possible sub-ranges as well as individual numerical values within that range For example, description of a range such as from 1 to 6 should be considered to have specifically disclosed sub-ranges such as from 1 to 3, from 1 to 4, from 1 to 5, from 2 to 4, from 2 to 6, from 3 to 6 etc., as well as individual numbers within that range, for example, 1, 2, 3, 4, 5, and 6. This applies regardless of the breadth of the range.

[0107] Certain embodiments may also be described broadly and generically herein Each of the narrower species and sub-generic groupings falling within the generic disclosure also form part of the disclosure. This includes the generic description of the embodiments with a proviso or negative limitation removing any subject matter from the genus, regardless of whether or not the excised material is specifically recited herein.

[0108] The invention has been described broadly and generically herein. Each of the narrower species and sub-generic groupings falling within the generic disclosure also form part of the invention. This includes the generic description of the invention with a proviso or negative limitation removing any subject matter from the genus, regardless of whether or not the excised material is specifically recited herein.

[0109] Other embodiments are within the following claims and non- limiting examples. In addition, where features or aspects of the invention are described in terms of Markush groups, those skilled in the art will recognize that the invention is also thereby described in terms of any individual member or subgroup of members of the Markush group.EXPERIMENTAL SECTIONCell lines and cell culture

[0110] Human mammary epithelial cell line (HMLE) was generated through immortalization using retroviral expression of human telomerase enzyme, telomerase reverse transcriptase (hTERT) and SV-40 large T antigen. Naturally arising mesenchymal cell lines (NAMEC) and NAMEC8-HRASG12V (NAMEC8-HRAS) were generated as described. To generate NAMEC cells, HMLE cells were grown to 50% confluency, followed by differential trypsinization for 1 min with 0.05% trypsin. Detached cells were carefully collected and re-plated at approximately 200 cells per well of a 24-well plate. Upon expansion, wells were screened for populations which were morphologically mesenchymal and can be stably propagated. NAMEC8-HRAS was established with pWZL-Blast-Ras. NAMEC8-KRASG12V (NAMEC8-KRAS) was established with pMN-KRASG12V. HMLE and NAMEC8 cells were maintained in MEGM growth medium composed of MEBM basal medium and MEGM SingleQuot Kit (Lonza). All cell cultures were maintained at 37°C with 5% CO2. SUMI 59 cells were maintained in Gibco Fl 2 supplemented with 5% fetal bovine serum (FBS). MCF7 and 293FT cells were maintained in Gibco DMEM supplemented with 10% FBS. KYSE30 was maintained in RPMI 1640 and Ham’s F12 (1:1) with 2 mM glutamine and 2% FBS. RL95-2 was maintained in DMEM-F12 supplemented with 10% FBS. At 80% confluency, cells were passaged at 1:10. All cell cultures were maintained at 37°C with 5% CO2.Primary human tumor and fibroblast cell lines establishment

[0111] National Cancer Centre Singapore (NCCS) provided primary human breast cancer samples with patients’ consent and under CIRB 2015 / 2976, A* STAR IRB approved protocols (2020-149 and 2024-047). The protocol permits the collection of fresh patient breast tumour andadjacent normal tissue for in vitro and in vivo experimentation. Cells were collected from dissociated tissue samples using 1 mg / ml collagenase IV (Gibco) and 1 mg / ml dispase II (Thermo Fisher Scientific) in DMEM. Cells were seeded onto irradiated 3T3 feeders and collected for downstream analysis. For fibroblast generation from patient tissues, cells collected from dissociated tissue samples were seeded in DMEM media with 20% FBS supplemented with basic Fibroblast Growth Factor (20 ng / ml).Dual-Gio luciferase assay

[0112] Luciferase activity was measured in 96-well plates according to manufacturer’s instructions. Media was aspirated from each well. 30 pl Dual -Gio Luciferase substrate (Promega) was added to the wells to read Firefly luciferase activity, and the plate was incubated for 10 minutes at room temperature. Then, 15 pl Dual-Glo Stop & Gio substrate was added to each well to read Renilla luciferase ac 580 tivity, and the plate was incubated for 10 minutes at room temperature. The TEC AN Ml 000 Pro plate reader was used to measure luminescence at each stage. Promoter activity was calculated by the ratio of FireFly over Renilla luminescence signal. Lentiviral. shRNA

[0113] Lentiviral shRNAs hairpins sequences (see sequence listing) were obtained from the RNAi Consortium (TRC) collection from the Broad Institute and synthesized by i-DNA Biotechnology (Singapore). The shRNA sequences were cloned into pLKO.l puromycin (Addgene plasmid #8453) or pLKO.l-Tet ON (Addgene plasmid #21915).Retrovirus generation

[0114] Retroviral particles for pMN-KRASG12V were produced in Plat A cells cultured in DMEM / F12 supplemented with 10% FBS. Briefly, 5-10 pg of the plasmid of interest, 1 pg of pVSVG (Addgene plasmid #8454), 1 pg of pdelta 8.2, were transfected in one 10 cm2dish with Lipofectamine 2000 according to manufacturer’s instructions. Media was refreshed after approximately 16 hours with DMEM / F12 with 10% FBS. Retrovirus-containing media was collected at 24 hours and 48 hours and strained with a 0.45 pm strainer (Sartorius).RNA isolation, reverse transcription, and real-time PCR analysis

[0115] Cells were collected and rinsed with PBS twice. Total RNA was extracted by TRIzol (Invitrogen) and purified using the RNeasy column (Qiagen). A total of 500 ng of total RNA was used for cDNA synthesis with a SuperScript III First-strand synthesis system (Thermo Fisher Scientific). cDNA was diluted 5-folds and 1 pl was used for real time PCR using a 2X mastermix (MiRXES). Gene expression quantification with gene-specific primers (Table 4) was carried outusing ABI Prism 7900HT Sequence Detection System 2.2 (Applied Biosystems). Expression of genes of interest was normalized to the expression of GAPDH, P-actin and 18S.Cell viability / proliferation assays

[0116] Cells were seeded into 96-well plates at a density 611 of 2000 cells per well in 100 pl of media for treated and untreated conditions. The cells were allowed to rest for 24 hours before the addition of drugs in 50 pl. The relative number of viable cells was determined by the quantification of ATP using the CellTiter-Glo Luminescent Cell Viability Assay (Promega G7571) system. The proliferation of these cells was analysed every two days for a period of 10 days. For NAMEC8-HRAS pLKO Tet-ON cells, 1 pg / ml doxycycline (Sigma) was used to induce shRNA expressionProtein extraction and immunoblotting

[0117] 1.5 x 106cells were lysed in RIPA buffer (Thermo Fisher 89900) with IX inhibitor (Thermo Fisher 78430) or Laemmli buffer on ice for 30 minutes. After centrifugation at 12000 rpm for 20 minutes, the supernatant was collected. Total protein concentration was measured by Bradford assay and 20-25 pg of protein was loaded for each sample. Cell lysates were subjected to electrophoresis using a gradient 4-12% SDS-PAGE (NP0322BOX) gel and transferred using Bio-Rad Transblot onto nitrocellulose or PVDF membranes. Protein membranes were blocked with 5% skimmed milk in Tris buffered saline (TBS, 1st BASE) with 0.1% Tween-20 (Santa Cruz). Blots were probed with primary antibodies (1:2000-1:5000 dilution) (Table 1) overnight at 4°C. Secondary HRP antibodies (Table 1) were incubated at room temperature for 1 hour at a 1:2000-1:5000 dilution and then visualized using chemiluminescence (WestemBright Sirius HRP substrate, K-12043-D20) with a Bio-Rad ChemiDoc imager.Wound healing assay

[0118] A 12-well plate was coated with 0.04 mg collagen (Corning) in 1 ml of 0.12% acetic acid for at least 1 day at 37 °C. 600-800K cells of each condition were seeded into a single 12-well respectively. After cells attachment, an equal width wound was made across the centre of the well using a 200 pl tip. Detached cells were removed by the replacement of 1 ml of media. The area of the wound was imaged immediately (defined as 0 hour) and at several timepoints up to 60 hours. Quantification of the wound area was done using Fiji with a Phase Wound plugin developed by AMP, A* STAR, Singapore.Transwell assay

[0119] Twenty-four-well (24-well) transwells (8 pm pore size) were coated with 6.6 pg Matrigel (Coming) in 100 pl MEGM media and incubated at 37°C for 1 hour. 100K cells in 150 pl MEGM media were seeded in each transwell The bottom compartment was filled with 1 ml of MEGM supplemented with 20% FBS as attractant. The plate was incubated at 37°C for 24 hours. Cells that are left on the top chamber were gently removed using a vacuum suction. Cells that have migrated to the lower surface of the membrane were fixed and stained with crystal violet dye. Transwells were imaged with the GelCount plate reader (Oxford Optronix). Quantitation of cells was done using Fiji using the integrated density function.Immunofluorescence (IF) staining of paraffin-embedded tumor sections

[0120] Deparaffinization of embedded tissue sections was done with 3 washes of 658 xylene (5 minutes each wash), 2 washes of 100% ethanol, 1 wash of 95% ethanol, 1 wash of 70% ethanol and 2 washes of deionized water (DI H2O). Once deparaffinized, slides were incubated with endogenous blocking buffer (65 ml 100% methanol, 3.5 ml 30% hydrogen peroxide, 31.5 ml DI H2O) for 30 minutes at room temperature. Antigen retrieval was performed in a pressure cooker for 30 minutes (10% citrate buffer). After cooling on ice, samples were incubated with blocking buffer (5% horse serum, 10% BSA, in PBS) and 0.2% Triton-X for an hour Blocking buffer was aspirated from the circled tissues, and PBS was washed 3 times for 5 minutes each time. The samples were treated with TrueBlack (GOLDBIO TB-250-1), a lipofuscin autofluorescence quencher. The samples were again washed 3 times in PBS for 5 minutes each wash. Primary antibodies (Table 1) were diluted 1:400 in blocking buffer and incubated overnight at 4°C. After which the samples were washed 3 times in PBS and incubated with secondary fluorescent-labelled antibodies (Table 1) for 1 hour at room temperature The samples were washed 3 times in PBS for 5 minutes and then incubated with DAPI (0.25 pg / ml) for 20 minutes. After DAPI, samples were washed once in PBS for 5 mins. Glycergel mounting medium (Dako) was added to the glass slides and affixed with a cover slip for imaging and storage at 4°C.Soft agar colony formation assay

[0121] Soft agar assay consists of 2 layers of agar, bottom (denser layer) and top (layer with cells). The bottom layer was a 1:1 mixture of 1.4% noble agar (Sigma, A5431-250G) and 2X DMEM-F12 supplemented with 20% FBS. The top layer was a 1:1 mixture of 0.7% noble agar and 2X DMEM-F12 supplemented with 20% FBS. 2 x 104 cells were seeded into each well. Each condition was done in triplicates. 2X DMEM-F12 10% FBS was added to each well every3 weeks to prevent the gel from drying up. Colonies were stained with crystal violet and counted after approximately 2 months.In vivo experiments using NSG Mice

[0122] All animal experiments were approved by the Agency for Science Technology and Research Singapore - Biological Resource Centre IACUC (Protocol number 241874).

[0123] Single cells (5 x 10') were counted and mixed in a 1:1 mixture of serum-free MEGM and Matrigel (BD 356231) and injected subcutaneously into the flanks of 4- to 6-week-old NOD. Cg-PrkdcscldI12rgtml"j1SzJ (NSG mice, Jackson Laboratories bred #005557). Six weeks later, or when tumour sizes exceeded 2 cm in length, mice were sacrificed, and tumours were harvested for downstream analysis (tumour mass comparisons, H&E, immunohistochemistry, immunofluorescence). The tumours were measured every 3 days for tumour tracking. For doxycycline-induction experiments, Dox -water (2 mg / ml in 30% fructose) was refreshed every 3 days when the tumours were measured. For intratumoral GGPP supplementation experiments, tumours were allowed to form to 100 mm3in volume. Supplementation of GGPP (11 iiM in 70% methanol) was performed using a 30G insulin needle on days 1, 3, and 5. Tumours were harvested on day 5 after 2 hours of incubation. Resected tumours were incubated in 4% paraformaldehyde for 24 hours prior to dehydration with 70% ethanol and sample submission to A* STAR IMCB Histopathology Lab (AMPL) for paraffin embedding, sectioning and H&E staining.In vivo metastasis assay

[0124] NAMEC8-HRAS cells were treated with DMSO or tipifamib (500 nM) for 7 days. One (1) million cells were resuspended in 200 pl MEGM media and injected into the tail vein of NSG mice. Moribund mice or mice with significant weight loss (>20%) were sacrificed and lungs were harvested and inspected for the presence of metastases.Hematoxylin and Eosin Staining of Tumour Sections

[0125] Resected xenograft tumours were weighed, fixed in 4% paraformaldehyde for 3 hours before being transferred to 70% ethanol and stored at 4 °C (in case of overnight storage). Samples were sent to the Advanced Molecular Pathology Lab, IMCB, for paraffinization, sectioning, hematoxylin and eosin staining, and histopathological evaluation by an in-house pathologist.Masson ’s trichrome staining

[0126] Five (5) pm thick paraffin sections of tumours were cut and placed onto glass slides. The slides were dewaxed in xylene and hydrated with descending series of ethanol before beingstained for Masson’s trichrome. The slides were placed in Bonin’s fixative solution in the oven at 60°C and washed in tap water. Weigert’s hematoxylin stain and Biebrich scarlet-acid fuchsin solution were applied sequentially and washed after each stain. Slides were then differentiated in 5% phosphomolybdic - phosphotungstic acid and transferred to 2.5% aniline blue. The slides were washed in distilled water and then differentiated in 1% acetic acid. After which, the slides were dehydrated though ascending series of ethanol to xylene before being mounted. scRNA-seq library preparation and data processing

[0127] Fresh single-cell suspensions were loaded into the automated Chromium system (10X Genomics) targeting 6000 cells per sample. Barcoded sequencing libraries were generated using the Chromium Next GEM Single Cell 3' Reagent Kits v3.1. All libraries were sequenced on NovoSeq 6000 until sufficient saturation was reached. After QC, raw sequencing reads were aligned to the human reference genome, GRCh38, and processed using CellRanger v.7.0.1. Quality control and transformation of 737 scRNA-seq data

[0128] CellRanger v.7.0.1 was used to assign the sequenced reads to the barcoded droplets. The raw expression matrix contains the unique molecular identifier counts (UM1 counts; indicative of a number of unique RNA molecules detected) for each genein each droplet. First, droplets that were potentially empty or containing multiple cells were removed by keeping cells that satisfied the following conditions: number of detected genes (nGenes) between 300 to 12500, total UM1 count between 500 to 200000 and percentage of mitochondrial reads less than 15%. Next, DoubletFinder v.2.0.3 was used with 15 principal components (PC) and assuming a doublet rate of 1% per 1000 cells to identify and remove any residual multiplets in each sample. Finally, the UM1 counts were normalised so that each cell had a total of 10000 UMIs across all genes and log-transformed the normalized counts with a pseudocount of 1 using the dataLogNormalize function in the RCAv2 package.Reference-based cell clustering

[0129] The Pearson correlation was computed between each log-transformed single cell transcriptome and each of the 179 cell type references in GlobalPanel provided in the RCAv2 package. Graph-based Louvain clustering from Seurat v. 4.2.0 was used to cluster the cells using the correlation projection space. Clustering on the correlation projection space has been shown to be more robust to batch effects. The 13358 high-quality cells clustered into 4 main clusters. The expression of canonical markers (vimentin, SMA, FAP, FSP, nidogen) was visualised using the plot density function from Nebulosa package (FIG. 1G). The correlation between each cluster and cell type references are shown in a heatmap (FIG. 8F) provided at least one correlation valueexceed the threshold of 2.0. scRNA-seq data was deposited in NCBI GEO (GSE264484, private access token: qfyhyseylpafdgj).Differential gene expression

[0130] The expression profde of the 4 main clusters was compared using the FindMarkers function (logfc. threshold = 0.585, p_val_adj = 0.05).BulkRNA-seq

[0131] 100 ng of DNAsel-treated RNA was used to input into the Bioanalyzer to attain the RIN score (RNA Integrity Number). Samples were loaded onto an Agilent RNA 6000 nanochip (Agilent cat no. 5067-1511) and only samples that have RIN scores above 9 were used. Library construction was done using Illumina Stranded mRNA seq and sequencing was done with Illumina HiSeq 4000 by the A*STAR GIS Next Generation Sequencing Platform. Data was deposited in NCBI GEO (GSE264732, private access token: sxgxkcgavrubncn).ARCHS4 analysis

[0132] For tipifamib-treated cells, genes filtered by FDR<0.5, with a 1.5 log2FC and a p. value of 0.001 were selected. A total of 694 genes were identified and input into Enrichr software, and ARCHS4 RNA-seq database was used to establish cell type transformation. Graphs were plotted with cell types on the y-axis and enrichment score or adjusted p-value on the x-axis. To deconvolute the transcriptional activation pathways, upregulated RNAseq genes were matched against the list of human transcriptional factors. Transcription factors with a 1 logFC and a p-value of 0.05 were selected. From these genes, zinc fingers were identified and removed and the list of genes (totalling 45 genes) were input into Enrichr and NCI-Nature 2016 database for gene ontology analyses. Graphs were transcriptional pathways on the y-axis and combined score or adjusted p-value on the x-axis.C ICK-H ' assay Immunofluorescence

[0133] 5000 cells were counted and plated on day 5 of tipifamib treatment. Azide-FPP (Cayman chemicals #13269) or Azide-GGPP (Thermo Fisher C10249) was supplemented 24 hours before the CLICK-IT assay was performed. On day 7 of tipifarnib treatment, a PBS wash was performed twice, and the cells were fixed with 4% paraformaldehyde for 15 minutes at room temperature. After incubation, cells were permeabilized for 15 minutes with 0.25% Triton-X (diluted in PBS). After permeabilization, fixed cells were washed in 3% BSA (diluted in PBS), repeated twice. A CLICK-IT cell reaction buffer kit was utilized, and components and volumes were used according to manufacturer’s instructions. Images were acquired with a Zeissmicroscope Axio Observer DI. Fluorescence intensity was quantitated using the integrated density function in Fiji.Single cell transcriptomic data UMAP generation

[0134] UMAPs of single cell transcriptomic data were mined and extracted from the CZ CELLxGENE Discover database.Active Rael (Cell Signaling #8815), Rho (Cell Signaling #8820), Ras (Cell Signaling #8821) detection assay

[0135] 2 x 106cells were pelleted into a 15 ml Falcon tube and resuspended in 10 ml of ice-cold PBS. The cells were pelleted at 4°C at 3000 rpm for 5 minutes and washed with ice-cold PBS. The pellet was resuspended in 500 pl of lysis buffer (provided by kit) and PMSF (final concentration 1 mM). Tubes were incubated on ice for 5 minutes and centrifuged at 16000 g for 15 minutes. Supernatant was collected and lysate concentration was determined by BCA protein assay. Each lysate was diluted to 1 mg / ml with lysis buffer. Spin cups were placed into 2 ml collection tubes (kit provided), 100 pl of glutathione beads were added and spun down, and the flowthrough was discarded. Then, 400 pl of lysis buffer was added to wash the beads, spun down at 6000 g for 30 seconds and flow through discarded. 20 pg of GST-PAK1-PBD (for Rael and CDC42), 400 pg of GST-Rhotekin-PBD (for Rho), or 80 pg of GST-Rafl-PBD (for RAS) was added to the beads. 500 pl of lysates were added to their respective tubes, and samples were incubated at 4°C for 1 hour with gentle rocking. After incubation, the mixture was spun down and the resin was washed 3 times with 400 pl of lysis buffer. After the last wash, spin cups were placed into a new collection tube. 50 pl of reducing sample buffer (with DTT, final concentration 200 mM) was added to each spin cup and incubated at room temperature for 2 minutes The mixture was spun down at 6000 g for 2 minutes. Samples were labelled and heated at 95°C for 5 minutes, and immunoblotting was performed.Statistical analyses

[0136] Statistical analyses of group comparisons were performed using unpaired, two-tailed t-test (GraphPad Prism).TABLES

[0137] Table 1 - List of antibodies used herein.&&

[0138] Table 2 - List of chemicals / inhibitors used hereinSEQUENCE LISTING

Claims

CLAIMS1. A method of treating RAS-dependent cancer in a subject, wherein the method comprises administering a pharmaceutically effective amount of a farnesyltransferase inhibitor to a subject in need thereof.

2. The method of claim 1, wherein the farnesyltransferase inhibitor is selected from the group consisting of tipifamib, lonafarnib, manumycin A and a-hydroxy famesyl phosphonic acid.

3. The method of any one of claims 1 or 2, wherein the RAS-dependent cancer is a HRAS- dependent cancer.4 The method of any one of the preceding claims, wherein the method further comprises administering an agent that increases the amount of geranylgeranyl pyrophosphate (GGPP).

5. The method of claim 3, wherein the agent that increases the amount of GGPP is selected from the group consisting of exogenous GGPP, exogenous recombinant GGPP, lipid- based delivery system comprising GGPP, nanoparticles comprising GGPP, and a combination thereof6 The method of any one of claims 4 to 5, wherein the increase in GGPP is compared to the level of GGPP in a subject not suffering from cancer.

7. The method of claim 6, wherein the increase in GGPP is determined by the increase in the amount of geranylgeranylated proteins in the subject compared to a subject not suffering from cancer.

8. The method of any one of claims 4 to 7, wherein the farnesyltransferase inhibitor and the agent that increases the amount of GGPP are administered either together in a single or separate compositions, separately, or sequentially.

9. The method of claim 8, wherein the sequential administration involves:a. the first administration of the farnesyltransferase inhibitor and a subsequent administration of the agent that increases the amount of GGPP; orb. the first administration of the agent that increases the amount of GGPP and a subsequent administration of the farnesyltransferase inhibitor.10 The method of any one of claims 1 to 9, wherein cells of the RAS-dependent cancer are differentiated into fibroblast cells.

11. The method of claim 10, wherein the cells of the RAS-dependent cancer are in the mesenchymal state.

12. The method of any one of claims 4 to 11, wherein the RAS-dependent cancer is a KRAS- dependent cancer or a NRAS-dependent cancer.13 The method of any one of claims 1 to 12, wherein the RAS-dependent cancer is selected from the group consisting of breast cancer, bladder cancer, head and neck squamous cell cancer, lung cancer, thyroid cancer and melanoma.

14. A composition comprising a farnesyltransferase inhibitor; and an agent that increases the amount of geranylgeranyl pyrophosphate (GGPP).15 The composition of claim 14, wherein the composition is a pharmaceutical composition.

16. A method of transdifferentiating a cell into a fibroblast-like stage, the method comprising contacting the cell to a farnesyltransferase inhibitor.

17. The method of claim 16, further comprising contacting the cell to an agent that increases the amount of geranylgeranyl pyrophosphate (GGPP).

Images