FIELD OF THE INVENTION

The invention relates to a combinatorial treatments with reactive oxygen inducing compounds to improve the effectiveness of cancer treatment.

BACKGROUND OF THE INVENTION

While targeted approaches are revolutionizing the treatment of cancer, the response to these treatments varies dramatically from patient to patient, and the management of both intrinsic and acquired therapy resistance remains a major limitation in several types of cancer. This is exemplified by the unprecedented, but transient, anti-tumour responses seen in patients with BRAF ^V600E mutant malignant melanoma exposed to agents that selectively inhibit oncogenic BRAF and downstream MEK [McArthur et at. (2014) Lancet Oncol. 15, 323-332] . Many of these patients show almost complete remission in response to such targeted agents, however, therapy resistance eventually develops in ~80% of all cases [Flaherty et a/. (2010) N. Engl. J. Med. 363, 809-819] .

One of the pathways that is emerging as a central player in multiple oncogenic processes and that functions downstream of a multitude of oncogenic signal transduction pathways is de novo lipogenesis. Accordingly, this pathway is specifically activated in many cancers [Menendez & Lupu (2007) Nat. Rev. Cancer 7, 763-777], in part through the induction of the transcription factor Sterol Regulatory Element Binding Protein (SREBP-1), a master regulator of lipogenesis [Griffiths et at. (2013) Cancer Metab. 1, 3] . Aberrant activation of the lipogenic pathway in cancer is required for the synthesis of phospholipids, which function as essential building blocks of membranes and that support cell growth and proliferation [Beckers et al. (2007) Cancer Res. 67, 8180-8187] . As this pathway mainly produces saturated and mono-unsaturated fatty acids, an increase in the proportion of these lipids in the cellular membrane composition of cancer cells is often observed [Currie et al. (2013) Cell. Metab. 18, 153-161] . Importantly, in contrast to poly-unsaturated fatty acid species which may be abundant in the diet, saturated and mono-unsaturated fatty acids are less prone to lipid peroxidation, thereby providing a survival advantage to cancer cells, particularly those exposed to severe forms of stress, including oxidative stress [Rysman et al. (2010) Cancer Res. 70, 8117-8126] .

Many genomic and non-genomic mechanisms have been described, all leading to re-activation of the MAPK- and/or PI3K- signalling pathways [Hugo et al. (2015) Cell 162, 1271-1285]. Moreover, different mutational events can be selected in distinct drug-resistant clones from the same patient and even co-occur within the same lesion.

Similar observations have been made in other tumour types and in response to other therapies. Specifically, prostate cancer is a leading cause of cancer-related male mortality. Although curative surgery is possible for localised disease, prostate cancer responds poorly to systemic treatment. Anti-androgen therapy remains the first line treatment for metastatic disease, however, transient therapy response remains a key challenge in the clinical management of prostate cancer.

These findings have highlighted the need to improve treatment protocols.

EP2018858 discloses compositions and methods for the treatment of primary and metastatic neoplastic diseases using reactive oxygen inducers such as arsenic trioxide.

SUMMARY OF THE INVENTION

In the present invention, we show that the lipogenic pathway is a key mediator of oncogenic signalling and functions as a modulator of therapy response by changing lipid polyunsaturation. We also show that targeted and other therapies induce lipid polyunsaturation by inhibiting lipogenesis and thereby sensitize cancer cell to inducers of reactive oxygen species (ROS). The present invention provides evidence for the use of targeted and other polyunsaturation-inducing treatments, PUFA activators, or lipogenesis inhibitors and inducers of reactive oxygen species in a novel combinatorial approach for therapy-responsive and therapy-resistant tumours.

The present invention reveals that a combination treatment of prostate cancer with an anti-androgen and arsenic trioxide has improved properties compared to the individual monotherapies.

Specific embodiments disclosed herein are:

1) An improved synergistic combination composition which comprises a) a Polyunsaturated fatty acid (PUFA) activator and b) a reactive oxygen species (ROS) activator for the treatment of a cancer. PUFA activator is defined as an inducer of the relative increase in cellular PUFA content.

2) The composition of statement 1, whereby the PUFA activator is a serine/threonine-protein kinase B-Raf inhibitor, tyrosine kinase inhibitor, MEK inhibitor, ERK inhibitor, fatty acid synthase inhibitor, acetyl-CoA carboxylase (ACACA) inhibitor, sterol regulatory element-binding protein (SREBP) inhibitor or combination thereof.

3) The composition of statement 1, whereby the PUFA activator is a serine/threonine-protein kinase B-Raf inhibitor of the group consisting of vemurafenib, dabrafenib and encorafenib.

4) The composition of statement 1, whereby the serine/threonine-protein kinase B-Raf inhibitor is dabrafenib.

5) The composition of statement 1, whereby the PUFA activator is a MEK inhibitor such as cobimetinib, trametinib or binimetinib.

6) The composition of statement 1, whereby the PUFA activator is combination of a serine/threonine-protein kinase B-Raf inhibitor and a MEK inhibitor.

7) The composition of statement 1, whereby the PUFA activator is an ERK inhibitor such as ulixertinib.

8) The composition of statement 1, whereby the PUFA activator is a fatty acid synthase inhibitor of the group consisting of TVB-2640, TVB3664, TVB3616, orlistat, C75 or cerulenin, (+)-Catechin hydrat, (-)-Epigallocatechin gallate, GSK837149A, Irgasan, Kaempferol, Luteolin, Osthole, Quercetin, (±)-Taxifolin hydrate, Triclosan, Cpd lOv, C93 (or FAS 93), EGCG, egcg, amentoflavone, - mangostin, -mangostin, cacalol, and diosgenin.

9) The composition of statement 1, whereby the PUFA activator is an acetyl-CoA carboxylase (ACACA) inhibitor such as ND646.

10) The composition of statement 1, whereby the PUFA activator is a sterol regulatory element-binding protein (SREBP) inhibitor.

11) The composition of statement 1, whereby the PUFA activator is a sterol regulatory element-binding protein (SREBP) inhibitor of the group consisting of betulin and fatostatin.

12) The composition of any one of the previous statements 1 to 11, whereby the reactive oxygen species (ROS) activator is an arsenic compound, preferably Arsenic Trioxide.

13) The composition of any one of the previous statements 1 to 11, whereby the reactive oxygen species (ROS) activator is a reactive oxygen species (ROS) scavenger inhibitor.

14) The composition of any one of the previous statements 1 to 11, whereby the reactive oxygen species (ROS) activator is a reactive oxygen species (ROS) scavenger inhibitor of the group consisting of Erastin, RSL3, beta-lapachone, buthionine sulfoximine, elesclomol, GSAO, imexon, Menadione, motexafin gadolinium, 5-ala and 2-methoxyestradiol and FK866. 15) The composition of any one of the previous statements 1 to 11, whereby the reactive oxygen species (ROS) activator is a ferroptosis inducer.

16) The composition of any one of the previous statements 1 to 11, whereby the reactive oxygen species (ROS) activator is a ferroptosis inducer of the group consisting of erastin, piperazine erastin, RSL3, altretamine, artesunate, sorafenib and regorafenib.

17) The composition of any one of the previous statements 1 to 11, whereby the PUFA activator is a combination of a lipogenesis inhibitor such as TVB3664, a serine/threonine-protein kinase B-Raf inhibitor and a MEK inhibitor and a ROS activator.

18) An improved synergistic combination composition which comprises a) Polyunsaturated fatty acid (PUFA) and b) a reactive oxygen species (ROS) activator for the treatment of a cancer.

19) The composition of statement 17, whereby the PUFA is a supplement from the group consisting of linoleic acid, linolenic acid, DHA or arachidonic acid or any combination thereof.

20) The composition of statement 17, whereby the reactive oxygen species (ROS) activator is arsenic trioxide.

21) The composition of statement 17, whereby the reactive oxygen species (ROS) activator is a reactive oxygen species (ROS) scavenger inhibitor.

22) The composition of statement 17, whereby the reactive oxygen species (ROS) activator is a ferroptosis inducer.

23) A pharmaceutical composition comprising a PUFA activator or PUFA and a ROS activator for use in treating a condition that benefits from increasing the tissue PUFA and reactive oxygen species (ROS).

24) The pharmaceutical composition of statement 23, whereby the PUFA activator is combination of a serine/threonine-protein kinase B-Raf inhibitor and a MEK inhibitor.

25) A pharmaceutical composition comprising a serine/threonine-protein kinase B-Raf inhibitor of the group consisting of vemurafenib, cobimetinib, dabrafenib and trametinib or a combination thereof and an arsenic compound, preferably arsenic trioxide for use in treating a condition that benefits from increasing the tissue PUFA and reactive oxygen species (ROS).

26) A pharmaceutical composition comprising dabrafenib and a reactive oxygen species (ROS) scavenger inhibitor of the group consisting of Erastin, RSL3, beta- lapachone, buthionine sulfoximine, elesclomol, GSAO, imexon, Menadione, motexafin gadolinium, 5-ala and 2-hydroxyestradiol and FK866 for use in treating a condition that benefits from increasing the tissue PUFA and reactive oxygen species (ROS).

27) A pharmaceutical composition comprising dabrafenib and an arsenic compound, preferably arsenic trioxide for use in treating a condition that benefits from increasing the tissue PUFA and reactive oxygen species (ROS).

28) The pharmaceutical composition of one of the previous statements 23 to 27, whereby the condition is a cancer that benefits from increasing the tissue PUFA and reactive oxygen species (ROS).

29) The pharmaceutical composition of statement 27, whereby the cancer is of the group consisting of melanoma, prostate cancer, glioma, breast cancer, lung cancer, pancreas cancer, liver cancer, or any other cancer type.

30) The pharmaceutical composition of any one of the previous statements 23 to 29, whereby the cancer is prostate cancer and whereby the treatment is combined with anti-androgen therapy, such as Enzalutamide, proxalutamide, darolutamide, apalutamide, seviteronel or abiraterone.

31) The pharmaceutical composition of any one of the previous statements 23 to 29, whereby the treatment is combined with carboplatin, paclitaxel and cabazitaxel.

32) The pharmaceutical composition of any one of the previous statements 23 to 29, whereby the treatment is combined with anti-EGFR therapy, such as RTK inhibitors.

33) The composition of any one of the previous statements further comprising a recipient or recipients.

34) The composition of any one of the previous statements whereby compound a) and compound b) are administered sequentially in separate single delivery forms.

35) An anti-androgen and arsenic trioxide for use in the treatment of prostate cancer.

36) The anti-androgen and arsenic trioxide, for use in the treatment of prostate cancer according to claim 35, wherein the anti-androgen is an androgen receptor binding compound.

37) The anti-androgen and arsenic trioxide, for use in the treatment of prostate cancer according to claim 35, wherein the anti-androgen is enzalutamide or apalutamide.

38) The anti-androgen and arsenic trioxide, for use in the treatment of prostate cancer according to claim 35, wherein the anti-androgen is abiraterone acetate. 39) An anti-androgen and arsenic trioxide for use in the treatment of prostate cancer in accordance to any one of claims 35 to 38, wherein the prostate cancer is castration resistant.

40) The anti-androgen and arsenic trioxide for use in the treatment of prostate cancer in according to any one claims 35 to 39, in a person that is responsive to anti-androgen monotherapy.

41) The anti-androgen and arsenic trioxide for use in the treatment of prostate cancer according to any one of claim 35 to 40, wherein the antiandrogen and the arsenic trioxide are administered simultaneously.

42) A pharmaceutical composition comprising an anti-androgen and arsenic trioxide, and a pharmaceutical acceptable carrier.

43) A method of treating prostate cancer in a subject, comprising the step of administering to said subject an effective amount of an anti-androgen and arsenic trioxide.

Detailed description

The present invention will become more fully understood from the detailed description given herein and the accompanying drawings which are given by way of illustration only, and thus are not limitative of the present invention :

Figure legends

Figure 1. Enzalutamide treatment decreases FASN expression in the enzalutamide responsive LnCap but not in therapy resistant 22RV1 cells (A). Membrane lipid poly-unsaturation incurred by enzalutamide treatment correlates with therapy response in prostate cancer cells (B).

Figure 2. Hydrogen peroxide sensitises anti-androgen therapy resistant 22RV1 cells to enzalutamide (A). The combination of arsenic trioxide and enzalutamide synergistically increase both cellular ROS (B) in mitochondrial ROS (C) in 22RV1 cells.

Figure 3. Arsenic trioxide further sensitises therapy responsive prostate cancer cells to enzalutamide and re-sensitises therapy resistant cells to enzalutamide.

Therapy of prostate cancer with arsenic trioxide is described in EP2018858, providing details on mode of administration and formulation.

"arsenic trioxide" refers to a pharmaceutically acceptable form of arsenic trioxide including salts, solutions, complexes, chelates and organic and inorganic compounds incorporating arsenic. Arsenic trioxide can be dissolved e.g. in an aqueous solution of sodium hydroxide, with the pH adjusted to a physiologically acceptable range, e.g. about pH 6-8.

The compound may be administered by parenteral administration such as intravenous, subcutaneous, intramuscular and intrathecal administration; oral, intranasal or rectal administration may also be used; directly into the tumour; transdermal patches; implant devices (particularly for slow release)

Arsenic trioxide can be formulated as sterile physiologically acceptable (aqueous or organic) solutions, colloidal suspensions, creams, ointments, pastes, capsules, caplets, tablets and cachets. The pharmaceutical compositions comprising arsenic trioxide can be contained in sealed sterile glass containers and/or ampoules. Further, the active ingredient may be micro-encapsulated, encapsulated in a liposome, noisome or lipofoam alone or in conjunction with targeting antibodies. Delayed slow or sustained release forms of administration are also considered.

"Anti-androgen" in the context of the present invention relates to compounds inhibiting androgen signalling. A recent review is e.g. Crawford et al. (2019) Prostate Cancer and Prostatic Diseases 22, 24-38.

Anti-androgens comprise LHRH agonist molecules such as leuprolide acetate, triptorelin pamoate, goserelin acetate and histrelin acetate.

Anti-androgens further comprises compounds blocking the androgen receptor such as enzalutamide and apalutamide.

Anti-androgens further comprises compounds inhibiting nuclear translocation of the androgen receptor signalling.

Anti-androgens further comprises compounds inhibiting the enzyme CYP17, such as abiraterone acetate, typically used in combination with and prednisone.

Detailed information on anti-androgen monotherapy in prostate cancer is available in the art.

The antiandrogen and TAO can be provided in a single formulation. More typically they are each in a dedicated formulation allowing to adapt the frequency of administration and the amount of each of the compounds.

The methods of the present invention are suitable for treatment of a prostate cancer which responds to a monotherapy with anti-androgen. Herein the additional activity of TAO enhances the activity, and reduces the chances of the tumor becoming resistant to anti-androgens. Prostate cancer patients being treated with an anti-androgen monotherapy often relapse and develop resistance to the low androgen levels moving to the castrate resistant form of the disease (CRPC). The mechanisms of developing CRPC are not yet fully unravelled but ongoing studies evaluate different mechanisms of resistance which include:

i. AR splice variants. They normally lack the ligand-binding domain leading to continuous activity in the absence of ligands. AR-V7 is the well-characterized splice variant which confers resistance to antiandrogens

ii. Gene amplification and overexpression of AR. AR deregulation may lead to the restoration of the ADT effects.

iii. AR mutations. They provide broader ligand specificity or convert antagonists to agonists.

iv. Intraprostatic synthesis of testosterone and DHT. Upregulation of enzymes involved in de novo synthesis of androgens, starting from cholesterol; and enzymes involved in the conversion of adrenal androgens into DHT and testosterone restore the androgen level in the prostate.18

v. Cofactor perturbation and post-translational modifications. Enhanced AR activity can be achieved by upregulating the expression of coactivators or downregulating the expression of corepressors. In addition, post-translational modifications such as phosphorylation can modulate AR activity.

As demonstrated in the examples section, the combination therapy of the present invention allows to treat those patient which have become resistant ot an anti-androgen monotherapy.

Resistance to targeted therapy in cancer represents a major clinical challenge. This is partly a consequence of the fact that most therapeutic targets to date, including BRAF ^V600E , act in the proximal part of their signal transduction cascade. This offers multiple opportunities for cancer cells to bypass drug response through for instance the acquisition of mutation(s) that reactivate the pathway downstream (e.g., by MEK/ERK mutation). An attractive strategy to overcome therapy resistance is therefore the identification and exploitation of vulnerabilities, which are activated by and act downstream of such oncogenic pathways. Metabolic pathways are of particular interest in this context as they often rely on a few essential enzymes, are frequently rewired in cancer cells, provide essential survival/adaptive capabilities and can easily be pharmacologically targeted. Here, we identified the lipogenic transcription factor SREBP-1 as a key downstream target of oncogenic BRAF signalling. We have shown that sustained lipogenesis through the maintenance of active SREBP-1 is a key feature of therapy resistance to vemurafenib in BRAF-mutant melanoma, and that inhibition of SREBP-1 sensitizes melanoma to targeted therapy.

Critically, the addiction of therapy-resistant melanoma cells to SREBP-1 is independent of the described mechanisms exploited by the cancer cells to overcome drug response. We observed similar effects in cells that acquired resistance acquisition of through RTK upregulation (M229 R, M238 R), NRAS mutation (M249 R) or enhanced IGF1/PI3K signalling and RAF kinase flexibility (451lu R). In all sensitive BRAF-mutant models, vemurafenib caused a decrease in lipogenesis and attenuated the processing and thereby the activity of SREBP- 1. This was not seen in therapy-resistant models, which all showed high levels of lipogenesis even in the presence of the inhibitor. Together with our observation that pharmacological or genetic inactivation of SREBP-1 in resistant cells attenuates cell proliferation and sensitizes to vemurafenib (irrespective of the escape mechanism), these findings indicate that SREBP-l-mediated lipogenesis is a central pathway acting downstream of mutant BRAF.

Previous work has shown that SREBP-1 is activated through several mechanisms including, regulation by the PTEN/ PI3K/Akt/mTOR pathway, p53, modulation of MAPK signalling by KRAS46 and by direct SREBP phosphorylation by Erkl/2. Our work identifies mutant BRAF as another key modulator of SREBP-1 processing and function.

Consistent with our findings, SREBP and its downstream targets are highly expressed in many cancers. Importantly there is a growing body of evidence showing that SREBP-l-dependent activation of lipogenesis is required for tumour growth in multiple models, including in prostate cancer [Li et at. (2014) Mol. Cancer Ther. 13, 855-86] and EGFR-dependent glioma. SREBP-1 was shown to promote adhesion-independent growth and cell proliferation, including growth factor-independent proliferation.

Similarly, proteins involved in the post-translational processing of SREBPs have also been linked to oncogenic potential in multiple models. SCAP modification and inhibition inhibit tumour growth through SREPBs. Consistently, expression levels of SCAP inversely correlate with overall survival in multiple cancers in TCGA cohorts. Taken together, these data strongly support a pro-oncogenic role for SREBP processing in multiple cancers.

These data are in line with the well-established necessity of cancer cells to adapt their metabolism to their increased need of building blocks. Activation of SREBP- 1 and enhanced ability to generate lipids in a cell-autonomous manner is thought to be required to sustain rapid tumour cell proliferation. Earlier findings from our team have shown that lipogenesis also contributes to resistance to cell death by altering membrane lipid composition and susceptibility to lipid peroxidation [Rysman et at. (2010) Cancer Res. 70, 8117-8126]. Here we provide evidence that lipogenesis driven by oncogenic BRAF signalling promotes resistance to targeted therapy. In therapy-sensitive cells, inhibition of oncogenic BRAF decreases membrane lipid saturation. Inhibition of SREBP-1 mimics this effect in therapy-resistant cells, especially upon BRAF therapy. We further show that this effect is a consequence of increased cellular ROS and lipid peroxidation. In therapy-resistant cells, vemurafenib resulted in a substantial increase in mitochondrial ROS and lipid peroxidation. This is consistent with the well- established ability of vemurafenib to induce ROS production [Hall et at. (2013) Oncotarget 4, 584-599]. Similarly, although to a lesser extent, SREBP inhibition also increased mitochondrial ROS levels. Combined inhibition of SREBP and oncogenic BRAF further enhanced ROS and lipid peroxidation, which can either be rescued or enhanced by exogenous addition of oleate or PUFA, respectively. Furthermore, addition of the antioxidant NAC results in a rescue of cell proliferation under combined therapy. Importantly, SREBP inhibition enhanced the efficacy of vemurafenib in a pre-clinical PDX model of melanoma, emphasizing the clinical relevance of these findings. Our data support the growing interest in lipogenesis inhibition as a novel anti-neoplastic strategy and ongoing efforts aimed at identifying new classes of SREBP inhibitors, including those that interfere with the nuclear accumulation of mature SREBP. By showing that SREBP-1 has a key role in the resistance to mutant BRAF-targeted therapy our work identifies an important clinical setting in which such inhibitors may provide clear clinical benefit.

EXAMPLES

Example 1: Methods

Cell culture. A375 was obtained from ATCC. FLCM was generated from melanoma derived from Braf CA, Tyr: :CreER and Ptenlox4-5 mice. M202, M207, and M233 were gifted by professor A. Ribas. 451 and 451lu R, M229, M229 R, M238, M238 R, M249, and M249 R were gifted by R. Lo. A101D BMR and D10 BMR were kindly gifted by Professor Daniel Peeper. NHEM was obtained from melanocytes derived from the foreskin of a pool of three healthy neonatal donors. The procedure was approved by the ethical committee of the University of Leuven and executed according to Helsinki guidelines. All melanoma cell lines were propagated in DMEM High Glucose (Sigma), supplemented with 10% FBS (Gibco Lot 41F4234K) and 4 mM glutamine (ThermoFisher). A101D BMR and D10 BMR growth media was additionally supplemented with dabrafenib and trametinib. NHEM were cultured in Medium 254 (ThermoFisher) supplemented with HMGS (ThermoFisher) and Antibiotic-Antimycotic to l x (ThermoFisher). 451lu R SREBF-1 KO clones were grown in DMEM High Glucose supplemented with 30% FBS and 4 mM glutamine. All cell cultures were periodically tested for mycoplasma contamination. All experiments were performed in DMEM High Glucose, supplemented with 2% FBS (Gibco Lot 41F4234K) and 4 mM glutamine, except for ¹³ C-glucose metabolite tracer studies, where 4.5 g L-l 13C-glucose (Cambridge isotope laboratories) was supplemented to DMEM no glucose (ThermoFisher). The following compounds were used at the stated concentrations. All prostate cancer cell lines were propagated in RPMI 1640 (ThermoFisher), supplemented with 10% FBS (Gibco Lot 41F4234K) and 2 mM glutamine (ThermoFisher). NAC (120 mM) from Sigma, alpha-tocopherol (100 mM) from Sigma, ferrostatin (1.25 pM) from Sigma, vemurafenib (5 pM) from ApexBio, dabrafenib (2.5 pM) from Selleckchem, trametinib (0.5 pM) from Selleckchem, betulin (2 or 3 pM) from Sigma, fatostatin (0.5 or 1.5 pM) from Tocris Bioscience, oleic acid (20 pM) from Sigma, linoleic acid (10 pM) from Sigma, linolenic acid (10 pM) from Sigma, Hydrogen peroxide from Sigma, Piperazine Erastin from MedChemExpress enzalutamide (10 pM) was obtained from MedChem Express, Hydrogen peroxide (150 pM) from Sigma and arsenic trioxide from Sigma.

RNA-seq. RNA concentration and purity were determined spectrophotometrically using a Nanodrop ND-1000 (Nanodrop Technologies) and RNA integrity was assessed using a Bioanalyser 2100 (Agilent). Samples were analyzed on an HiSeq2000 (Illumina).

Plasmid transfections. M202 and M207 were transfected by electroporation (Neon Transfection System, ThermoFisher) with either pBabe-puro or pBabe- puro-BRAF ^V600E . A375, 451lu, and M249 were transfected by electroporation (Neon Transfection System, ThermoFisher) with pbabe-puro-HA-SREBF-1 (Y335R)-myc.

Construction of knockout cell lines. 451lu R cells were transfected (Neon Transfection System, ThermoFisher) with CRISPR-Cas9 plasmid constructs with a guide-RNA targeting human SREBF-1 exon 1 (VectorU6gRNA-Cas9-2A-GFP, target ID: HS0000039707 and HS0000039709) (Sigma). 72 h post-transfection, the top 10% of GFP expressers were sorted by FACS (BD bioscience, ARIA III) into single wells for colony formation. Targeted regions of individual clones were sequenced (Sanger sequencing, LGC) and indels in allelic sequences we genotyped using CRISP-ID.

¹⁴ C-acetate incorporation into lipids and ¹³ C-glucose tracing. Cells were grown in 6-well plates up to 80% confluence and were treated with 0.1 pCi acetate-2-14C (55 mCi/mmol; Amersham) for 4 h. After three washes with PBS (Sigma), cells were trypsinized and resuspended in 1 ml. PBS, followed by sonication. 0.125 ml. of lysate was set aside for DNA measurement and 0.7 mL of lysate was mixed with 0.9 mL MeOH solution (MeOH : HCI = 8: 1) (Sigma) and 0.8 mL CHCb (Sigma). After vortexing the mixture for 30 seconds followed by centrifugation for 10 min at 2000 x g, the organic fraction was counted for radioactivity (Tri-Carb 2810 TR scintillation counter, PerkinElmer). Mass spectrometry analysis of ¹³ C-glucose incorporation into palmitate and subsequent isotopomer spectral analysis was performed as described in Lorendeau et a/. (2017) Metab. Eng. 43, 187-197.

Untargeted metabolomics. Cell extracts were separated on an Acquity HSS T3 UPLC column (Waters Corp, 2.1 mm x 150 mm, 1.8 pm particle size) using an Ultimate 3000 HPLC (Dionex ThermoFisher Scientific Inc). Elution of metabolites was performed using a quaternary solvent system. The data was collected over a mass range of 50-1050 m/z. Data analysis was done using ThermoFisher Scientific Quan software (Xcalibur version 4.0) and manually verified. Further analysis was done using in-house software tools.

Cholesterol amount and uptake quantification. Cholesterol was quantified using Amplex Red cholesterol assay kit (ThermoFisher). Cholesterol uptake was quantified using NBD cholesterol (ThermoFisher)

Analysis of intact phospholipid species by ESI-MS/MS. 0.7 mL of homogenized tissue or cells were mixed with 0.9 mL MeOH : HCI(l N) (8: 1), 0.8 mL CHCI3 and 200 pg mL-1 of the antioxidant 2,6-di-tert-butyl-4-methylphenol (Sigma). The organic fractions were evaporated under vacuum, reconstituted in Me0H/CHCI3/NH40H (90: 10: 1.25) and lipid standards were added (Avanti Polar Lipids). Phospholipids were analyzed by electrospray ionization tandem mass spectrometry (ESI-MS/MS) on a hybrid quadrupole linear ion trap mass spectrometer (4000 QTRAP system, AB SCIEX) equipped w3ith a TriVersa NanoMate robotic nanosource (Advion Biosciences) as described in Rysman et a/, cited above.

Immunoblotting analysis. Following three ice-cold PBS washes, cells were collected in sample buffer (ThermoFisher) supplemented with DTT (Sigma), sonicated and boiled for 5 min. Equal amounts of protein were loaded onto precast gels (NuPAGE, ThermoFisher), transferred to nitrocellulose membranes, and incubated with antibodies against SREBP-1 (1/1000 dilution) (Active Motif, #39939), phospho-MEKl/2 ser217/221 (1/1000 dilution) (Cell Signaling, #9154); GAPDH (1/ 20000 dilution) (Cell Signaling, #5174), myc-Tag (1/2000 dilution) (Cell Signaling, #2276), HA-Tag (1/5000 dilution) (Cell Signaling, #3724), phosphor-AKT ser473 (1/1000 dilution) (ThermoFischer, #98H9L8), pS6 ser235 (1/1000 dilution) (Cell Signaling, #2211), and pERK 1/2 (1/1000 dilution) (Cell Signaling, #9101).

RNA extraction and RT-qPCR. RNA extraction and RT-qPCR were performed as described in Daniels et a/. (2014) PloS ONE 9, el06913.

Confocal microscopy. Cells were transfected (Neon Transfection System, ThermoFisher) with a plasmid coding an HA and myc-tagged transcriptionally inactive SREBP-1 (Y335R) (GenScript). The SREBP cDNA was obtained from pTK- HSV- BPla65. Cells were treated as indicated, fixed and incubated with antibodies against: myc-Tag (1/2000 dilution) (Cell Signaling, #2276) and HA- Tag (1/5000 dilution) (Cell Signaling, #3724), GM130 (1/500 dilution) (BD Biosciences, #610822), and PDI (1/1000 dilution) (ThermoFisher, #MA3-018). Samples were imaged using an Olympus FluoView FV1000 confocal microscope. Proliferation assays. Proliferation curves were generated using an IncuCyte ZOOM system (Sartorius) on cells seeded of microplates (TPP) based on phase contrast images taken at 2 h intervals for the duration of the experiments.

Prostate cancer proliferation assays were generated using an IncuCyte ZOOM system (Essen BioScience) on cells transduced with the IncuCyte® NucLight Green Lentivirus Reagent (Sartorius) based on green fluorescent images.

Colony formation in soft agar was performed as described by Franken et at. (2006) Nat. Protoc. 1, 2315-2319. Except, colonies were stained with Vybrant DyeCycle Green nuclear stain (ThermoFisher), imaged with a Typhoon FLA 9500 laser scanner (GE Healthcare) and quantified in ImageJ.

Cellular and Mitochondrial ROS measurement

Cellular and Mitochondrial ROS was measured using CellROX (ThermoFisher) and MitoSOX red mitochondrial superoxide indicator (ThermoFisher) according to the manufacturer's instructions. Cells were assayed using a FACS Verse flow cytometer (BD Biosciences).

Lipid peroxidation assay. Lipid peroxidation was quantified using the MDA assay kit (Sigma) according to manufacturer's instructions with some exceptions. Briefly, cells or tissue were collected in BHT supplemented PBS. TBA-acetic acid solution was buffered to pH 3.5. Plates were read using an EnSpire Multimode Plate Reader (PerkinElmer).

Animal experiments. The Mel006 PDX model was derived from a metastatic melanoma lesion carrying the BRAF ^V600E mutation (KU Leuven, TRACE). Tumours were transplanted into nude mice (NMRI-Foxlnu, Taconic) and when tumour size reached 200 mm ³ , mice were randomly assigned to a cohort and drugs or vehicles were blindly administered daily by oral gavage. Vemurafenib and fatostatin were both administered daily at 20 mg/kg. Fatostatin or vehicle (30% PEG (Sigma) in water (Baxter)) was administered an hour after vemurafenib or vehicle (2% hydroxypropylcellulose (Sigma) in water (Baxter)). Tumour size was measured blindly with digital calipers (Fowler Sylvac) every 3 days. Mice were sacrificed at 28 days following start of treatment, or when tumours reached a volume of 1500 mm ³ . The investigators were blinded for the evaluation of the results.

Statistical analysis. The results were analysed in GraphPad Prism 6.0 h using a t-test. In case of multiple comparisons a correction was applied using the Holm- Sidak method, p-values of <0.05 were considered to be statistically significant. All data presented represent means ± s.e.m.

Study approval. Written informed consent was obtained from the patient (Mel006) and all procedures involving human samples were approved by the UZ Leuven/KU Leuven Medical Ethical Committee (#ML8713/S54185). All procedures involving animals were carried out in accordance with the guidelines of the Animal Care and Use Ethical Committee (KU Leuven, P038/2015).

Example 2: De novo lipogenesis is inhibited by BRAF ^V600E -targeted therapy.

As in many cancers, there is evidence that de novo lipogenesis is activated in melanoma. We reasoned that ectopic MAPK- activation may be one key triggering event of such activation. To test this possibility, we assessed the impact of BRAF inhibition on lipid metabolism. We exposed BRAF-mutant, therapy-sensitive, melanoma cell lines (M249 and A375) to vemurafenib and profiled their transcriptome by RNA-seq. Ingenuity pathway analysis (IPA) identified fatty acid metabolism as one of the most affected pathways by the treatment. Consistently, expression of key lipogenic enzymes such as ATP citrate lyase (ACLY), acetyl-CoA carboxylase- 1 (ACACA), and fatty acid synthase (FASN) were consistently decreased. Alterations in the expression of these enzymes by mutant BRAF inhibition was confirmed by RT-qPCR on an extended panel of therapy-sensitive BRAF ^V600E parental and isogenic cell lines that have acquired resistance to vemurafenib through diverse mechanisms. These include Raf- kinase flexibility in MAPK signalling and in increased IGF-1R/PI3K signalling (451lu R), enhanced RTK signalling (M229 R and M238 R) and secondary acquisition of oncogenic NRASQ61K (M249 R). Whereas vemurafenib decreased the expression of lipogenic enzymes in all sensitive BRAF-mutant cell lines, this was not seen in normal neonatal human epidermal melanocytes (NHEM) and in the therapy-resistant lines. If anything, the opposite effect was observed in the vemurafenib-resistant cells. Direct measurement of the overall rate of lipogenesis by assessing ¹⁴ C-acetate incorporation into lipids confirmed an overall increase in lipogenesis in melanoma cell lines compared to NHEM. A marked decrease in de novo lipogenesis was observed in all BRAF ^V600E therapy- sensitive, but not resistant, cell lines upon vemurafenib exposure. These findings were further corroborated by isotopomer spectral analysis, a method that measures fatty acid biosynthesis rates by measuring the fraction of de novo synthesized palmitate. In general, there was a marked decrease in the fraction of de novo synthesized palmitate in therapy-sensitive lines. In contrast, vemurafenib did not cause any decrease in palmitate synthesis in some therapy- resistant cells or induced only a moderate reduction in others. We conclude that lipogenesis is sustained in therapy-resistant cells when compared to therapy- sensitive cells upon vemurafenib treatment. Notably, lipid uptake, cholesterol synthesis rate or cholesterol uptake were not affected in any of the conditions and cell lines, indicating that vemurafenib predominantly affects de novo fatty acid biosynthesis.

De novo lipogenesis mainly produces mono-unsaturated and saturated fatty acids with phospholipids as major end product. Consistently, mass spectrometry- based phospholipidome analysis revealed that inhibition of oncogenic BRAF in the therapy-sensitive lines caused an increase in the proportion of poly-unsaturated membrane phospholipid species at the expense of saturated and mono- unsaturated phospholipids. These are typical changes observed upon lipogenesis inhibition. Such a shift was either absent or less pronounced in the therapy- resistant lines and in NHEM. Taken together, these findings indicate that inhibition of oncogenic BRAF inhibits de novo lipogenesis and thereby enhances membrane poly-unsaturation. Example 3: BRAF ^V600E -induced lipid metabolism is mediated by SREBP-1.

The selected lipogenic enzymes, the expression of which are downregulated upon oncogenic BRAF inhibition, are well-established transcriptional targets of SREBP-1. We therefore examined whether activity of SREBP-1 itself may be decreased by vemurafenib. Because SREBP-1 is synthesized as an inactive precursor, which is activated upon proteolytic cleavage, we used western blot analysis to assess the protein levels of both full-length and mature SREBP-1. Vemurafenib caused a decrease in the levels of the mature form of SREBP-1 in all BRAF ^V600E -therapy-sensitive, but not (or less so) in resistant cell lines. In contrast, the non-processed form was either unaffected or increased. To further substantiate this finding, we exploited the paradoxical activation of the MAPK pathway by vemurafenib in two NRAS mutant cell lines (M202 and M207) (32- 34). Vemurafenib resulted in both an expected increase in the levels of pMEK 1/2 proteins and of mature SREBP-1. In addition, over-expression of a BRAF ^V600E - encoding plasmid in these BRAF wild-type cells further support the ability of oncogenic BRAF to induce SREBP-1. RT-qPCR analysis of the transcripts encoding SREBP-la and SREBP-lc showed a decreased expression upon vemurafenib exposure in 451lu and A375, but not in the other cell lines. As SREBP expression is subject to autoregulation, these findings indicate that vemurafenib acts, at least in part, at the level of SREBP maturation. Note, however, that more direct transcriptional effects on the regulation of SREBF-1 transcription cannot be fully excluded. To further substantiate these findings, we looked at the effect of vemurafenib on the cellular distribution of SREBP-1. To this end, we generated a transcriptionally inactive recombinant full-length SREBP-1 construct with an N- terminal HA-tag and a C-terminal myc-tag, allowing visualization of the active and inactive forms of the protein. Western blotting established that the transgene is processed in a manner that is indistinguishable from endogenous SREBP-1. Similarly, to endogenous mSREBP-1, which localizes to the nucleus, the processed HA-tagged exogenous protein was detected in the nuclei of untreated melanoma cells. In contrast, both HA- and myc-tagged proteins co- localized in the ER-Golgi, indicating that proteolytic activation is halted in vemurafenib-treated cells. Taken together, oncogenic BRAF targeting inhibits the processing and activation of SREBP-1 in therapy-sensitive, but not therapy- resistant, melanoma cells and this effect is, by and large, mediated by a post- translational mechanism.

Interestingly, BRAF inhibition only induced a moderate decrease in mSREBP-1 levels and did not significantly affect lipogenesis in therapy-resistant cells. Since alternative activation of the ERK pathway is a common contributor to therapy resistance and a known regulator of SREBP we treated the therapy-resistant cell line 451lu R with the MEK inhibitor trametinib. As expected, these cells maintained high levels of pMEK upon vemurafenib treatment. Interestingly, MEK inhibition substantially decreased the levels of mSREBP-1 in these cells. Consistently, expression of well-established mSREBP-1 downstream targets, such as ACLY, ACACA, and FASN, was also reduced. These findings indicate that reactivation of the ERK pathway contributes to sustained SREBP-1 activity in therapy-resistant melanoma cells.

Example 4: Sustained SREBP-1 activity maintains lipogenesis in therapy- resistant cells.

To further assess the role of SREBP-1 in the changes in lipid metabolism evoked by vemurafenib, we inhibited SREBP-1 in 451lu R cells using pharmacological and genetic approaches. We used two small-molecule inhibitors, betulin and fatostatin, which inhibit the trafficking of SREBP to the Golgi, and thereby its proteolytic activation through similar but distinct mechanisms 37,38. Exposure of 451lu and 451lu R cells to these inhibitors induced the expected dose-dependent decrease in the levels of mature SREBP-1; an effect that was more pronounced in the therapy-sensitive cell line. Phospholipidomic analysis revealed that chemical inhibition of SREBP-1 dose-dependently depleted mono-unsaturated and fully saturated phospholipid species and increased membrane poly- unsaturation, partially recapitulating the effect of BRAF inhibition on the therapy- sensitive cell line. Furthermore, these effects were further enhanced with the addition of vemurafenib, whereby the phospholipidome of the resistant line under SREBP-1 inhibition and vemurafenib closely resembled that of the sensitive line in response to vemurafenib.

To corroborate these data using a genetic approach, we generated a heterozygous and two homozygous SREBF-1 knockout clones from the therapy- resistant cell line 451lu R using CRISPR-Cas9. Both Sanger sequencing and western blot analysis confirmed partial or full SREBF-1 deletion in the heterozygous and homozygous mutant cells, respectively. Similar changes in the lipid membrane composition to the ones observed upon pharmacological inhibition were observed in the heterozygous SREBF-1 KO clone, and to an even greater extent, in homozygous KO clones. Since inhibition of SREBP-1 activation resulted in membrane lipid changes that closely mimicked the effects of vemurafenib, we concluded that in this context, SREBP-1 is the major mediator of BRAF inhibition-dependent lipid metabolism rewiring.

Example 5: Inhibition of SREBP-1 re-sensitizes resistant cells to BRAF targeting therapy.

In order to investigate whether vemurafenib-induced processing of SREBP-1 contributes to its anti-tumour response, we assessed the ability of melanoma cells to grow in both 2D and 3D cultures in response to vemurafenib upon pharmacological or genetic inactivation of SREBP-1. Vemurafenib potently inhibited cell proliferation of the therapy-sensitive, but not the therapy-resistant line. Proliferation of the therapy-resistant cells was strongly inhibited upon exposure to vemurafenib and SREBP-1 inhibitors. Similarly, SREBP-1 inhibitors synergized with vemurafenib in inhibiting the ability of therapy-resistant cells to form colonies. Genetic ablation of SREBF-1 inhibited the rate of cell proliferation in 2D cultures of therapy-resistant cells when compared to the resistant parental cell line. In contrast, inactivation of only one SREBF-1 allele alone had no effect on cell proliferation. The rate of proliferation of these cells decreased significantly upon exposure to vemurafenib. Similarly, these heterozygous SREBF-1 knockout cells were able to form colonies in 3D cultures; an ability that was reduced in the presence of vemurafenib.

In order to assess whether sustained SREBP-1 processing also mediates therapy resistance to vemurafenib in BRAF ^V600E mutant cells that are endogenously resistant to BRAF targeting therapy, we treated M233 40 with betulin and fatostatin. Both betulin and fatostatin treatment further sensitized the cell line to vemurafenib.

In order to assess whether our results directly reflect the effects of BRAF inhibition and are not off-target effects of vemurafenib, in parallel we treated 451lu and 451lu R cells with dabrafenib. Dabrafenib invoked a decrease in both ¹⁴ C-acetate incorporation into lipids and mature SREBP-1 protein levels in 451lu cells, but not in 451lu R cells. Furthermore, combined betulin and fatostatin treatment with dabrafenib inhibited proliferation in 451lu R cells, mirroring the effects of vemurafenib.

Since combined BRAF and MEK inhibition treatment is a standard of care for BRAF-mutant melanoma patients, we further assessed the cell proliferative response to SREBP inhibition in A101D BMR and D10 BMR under dabrafenib and trametinib. These cell lines are resistant to both dabrafenib and trametinib (A101D BMR is partially resistant and retains some level of sensitivity). Here we show that SREBP inhibition by either betulin or fatostatin sensitized the cells to combined BRAF and MEK inhibition. Taken together, these data indicate that SREBP-1 contributes to the anti-tumour response induced by BRAF inhibition and that SREBP-1 inhibition sensitizes therapy-resistant melanoma cells to MAPK- targeting therapy.

Example 6: SREBP-1 protects vemurafenib-resistant cells from lipid peroxidation.

The findings described above predict that SREBP-l-mediated therapy resistance is a consequence of enhanced membrane lipid saturation and, consequently, of decreased lipid peroxidation. To test this hypothesis, we mimicked the effect of SREBP-1 inhibition on the lipid composition of therapy-resistant cells by treatment with the poly-unsaturated fatty acids linoleic and linolenic acid (PUFA). When combined with vemurafenib, PUFA addition slightly attenuated the growth of the cultures. Conversely, supplementing cells with the final product of lipogenesis, oleic acid, slightly enhanced proliferation of cells under vemurafenib treatment. Both of these effects were significantly enhanced upon pharmacological or genetic inactivation of SREBP-l/SREBF-1, indicating that membrane lipid saturation contributes to SREBP-l-mediated resistance to BRAF inhibition.

We have previously reported that membrane lipid saturation protects cancer cells from ROS-and chemotherapy-induced cell death. Here, we observed that both fatostatin and vemurafenib increase the levels of mitochondrial ROS independently, and that a combination of the two leads to an additional effect. In addition, mitochondrial ROS levels were either further enhanced or decreased by addition of PUFAs and oleic acid, respectively. Under combined vemurafenib and fatostatin treatment, addition of oleic acid reduced the levels of mitochondrial ROS to levels observed in cells treated with vemurafenib alone. ROS has been linked to membrane lipid peroxidation, which results in the accumulation of toxic by-products. Since poly-unsaturated lipids are more prone to lipid peroxidation than saturated lipids, we next measured lipid peroxidation by measuring levels of cellular malondialdehyde (MDA), which is a direct by-product of lipid hydro- peroxide degradation. Whereas generally low in melanoma cells, levels of MDA increased upon exposure to fatostatin, and to an even greater extent upon vemurafenib treatment. Combined fatostatin and vemurafenib treatment resulted in a further increase in MDA levels, which is directly in line with the levels of membrane lipid peroxidation incurred by the treatments. The levels of MDA under combined therapy increased further with addition of exogenous PUFA and decreased to levels seen under vemurafenib alone upon addition of exogenous oleic acid.

Interestingly, steady state metabolomics analysis of cellular AMP, ADP, and ATP indicated that cellular ATP levels or energy charge are not markedly altered by combined BRAF and SREBP-1 inhibition. Metabolomics analysis of NAD, NADH, NADP, NADPH, GSH, and GSSG ratios revealed that combined SREBP-1 and BRAF inhibition significantly lowers the cell antioxidant potential.

To corroborate our findings that SREBP-1 inhibition sensitizes cells to vemurafenib through lipid peroxidation, we supplemented cells with the antioxidants alpha-tocopherol, ferrostatin, and N-acetyl-cysteine (NAC). Under combined SREBP-1 and oncogenic BRAF inhibition, addition of antioxidants partially rescued cell proliferation. To show that lipid poly-unsaturation and lipid peroxidation also plays a role in vemurafenib response in drug sensitive cells, we treated therapy-sensitive cell lines M229 and 451lu with either betulin or fatostatin and found that these compounds further enhance the cytostatic effects of the BRAF-inhibitor. Furthermore, we show that treatment with alpha- tocopherol, in part, rescued the cytostatic effects of vemurafenib in the M229 cell line, and alpha-tocopherol, NAC and ferrostatin treatment rescued the proliferation of 451lu cells. We conclude that SREBP-1 inhibition sensitizes cells to vemurafenib, at least partly, though alterations of membrane poly- unsaturation and, thereby, lipid peroxidation.

Example 7: SREBP-1 inhibition sensitizes melanoma to vemurafenib in vivo.

To assess the therapeutic potential of these findings we investigated the impact of SREBP-1 inhibition in an in vivo pre-clinical BRAF ^V600E -mutant melanoma model. We chose PDX MEL006, which has been extensively characterized previously and was shown to poorly respond to BRAF inhibitors alone. Mouse cohorts were treated blindly with either vemurafenib alone, fatostatin alone or a combination of the two. Fatostatin treatment alone inhibited tumour growth more potently than vemurafenib. Importantly, combined vemurafenib/ fatostatin co-treatment had a greater anti-tumour effect than any of the monotherapy regimens. Phospholipidomic analysis of the various treated melanoma lesions revealed a correlation between the changes in the poly-unsaturation of phospholipids and anti-tumour growth response, whereby membrane poly- unsaturation was synergistically enhanced by the combination treatment. MDA analysis revealed that whereas fatostatin or vemurafenib treatment alone did not significantly increase lipid peroxidation, the combined vemurafenib/fatostatin treatment greatly enhanced lipid peroxidation. In the Mel006 tumour treated with a combination of dabrafenib and trametinib, an increase in MDA was found shortly after the start of treatment and, to a lesser extent, after establishment of resistance.

Taken together, combined fatostatin and vemurafenib therapy enhanced therapy response in vivo and increased membrane lipid poly-unsaturation and lipid peroxidation. These data support the concept of a novel combinatorial approach to overcome therapy resistance in BRAF ^V600E mutant melanoma models.

Example 8: Arsenic trioxide re-sensitizes melanoma cells to Map Kinase inhibition.

Since we have shown that Map Kinase inhibition dramatically increases membrane lipid poly-unsaturation through lipogenesis inhibition and somewhat in therapy resistant cells, we assessed whether further ROS induction can sensitize therapy resistant cells. To this end, we used the ROS inducers hydrogen peroxide and Piperazine Erastin in combination with BRAF and SREBP inhibitors, showing that BRAF inhibition sensitizes cells to ROS induction, especially when combined with SREBP inhibition. Next, the FDA approved ROS inducer Arsenic trioxide (ATO) was added to 451lu R cells in combination with both dabrafenib and Vemurafenib, showing that ROS induction sensitizes cells to BRAF inhibition. In the BRAF and Mek inhibition resistant cell line D10BMR, we show that ATO sensitizes cells to combinatorial dabrafenib and trametinib treatment. Importantly, in the preclinical PDTX mel006 model we show that BRAF inhibition by dabrafenib potently synergizes with ATO to cause dramatic tumour shrinkage.

Example 9: Membrane lipid poly-unsaturation by PUFA supplementation sensitizes cells to ATO.

We have shown that membrane lipid-poly-unsaturation through lipogenesis inhibition sensitizes cells to ROS inducers. Since multiple cancers reply on lipid uptake from the extracellular environment or surrounding stromal cells, we investigated the effect of membrane lipid saturation modulation by the exogenous supplementation of lipids. To this end, we supplemented DIO BMR cells with a mixture of cell media supplemented with 20 uM oleic acid or a 1 : 1 mixture of 10 uM linoleic and 10 uM linolenic acid. Using phospholipidomics, we show that these fatty acids are incorporated in membrane phospholipids. We next showed that under dabrafenib and trametinib treatment, membrane lipid poly-unsaturation by PUFA supplementation further sensitizes cells to Arsenic trioxide, and that saturated lipid supplementation is protective. Taken together, in addition to membrane lipid poly-unsaturation by lipogenesis inhibition, PUFA supplementation sensitizes cells to ROS induction.

Example 10: Androgen receptor inhibition therapy sensitizes cells to ATO

Anti-androgen therapy is a standard of care in prostate cancer, where either endogenous or acquired androgen insensitivity in prostate cancer limits treatment. Importantly, it is shown that anti-androgen therapy promotes membrane lipid poly-unsaturation in prostate cancer and that lipogenesis, namely through SREBP-1 promotes multiple oncogenic processes including metastasis. To this end, we treated multiple androgen responsive and androgen resistant prostate cancer cell lines with Enzalutamide in combination with lipogenesis and with arsenic trioxide. We show that in general, Enzalutamide synergizes with arsenic trioxide to either further sensitize therapy responsive prostate cancer cells, or to sensitize therapy resistant cells. In addition, we show that lipogenesis inhibition further enhances the magnitude of these effects.

Anti-androgen therapy inhibits lipogenesis by inhibiting SREBP activation and thereby leads to membrane lipid poly-unsaturation by reducing the membrane fraction of de novo synthesised saturated lipids. We therefore reasoned that 1 - lipogenesis would be inhibited more strongly in anti-androgen responsive cell lines than in resistant cell lines (Figure 1A) - upon enzalutamide treatment, the fraction of poly-unsaturated membrane lipids would reflect enzalutamide response (Figure IB).

Since it has been extensively described that poly-unsaturated membrane lipid species are more reactive to ROS, we reasoned that lipogenesis may be a driver of therapy resistance in prostate cancer and may pose a promising anti- neoplastic target. To this end, we show that enzalutamide sensitises therapy resistant 22RV1 cells to hydrogen peroxide (Figure 2A). Using a more clinically relevant ROS inducer in the form of arsenic trioxide, we show in 22RV1 cells that enzalutamide works synergistically with arsenic trioxide to elevate the levels of both cellular and mitochondrial ROS (Figure 2B and 2C). In order to show that translational potential of this finding, we treated a panel of therapy responsive and therapy resistant prostate cancer cells with a combination of enzalutamide and arsenic trioxide (figure 3). In all cases, the combination treatment synergistically decreases cell proliferation in both therapy responsive and therapy resistant cell lines, implying the potential of this therapy combination in further sensitising cancer cells to anti-androgen response and in overcoming therapy resistance.

Example 11: Carboplatin therapy sensitizes cells to ATO

Ovarian cancer remains poorly responsive to targeted approaches, where the cytotoxic drugs carboplatin and paclitaxel (in combination) remain the first line therapy in the clinic. Importantly, we have shown that carboplatin treatment is therapy responsive ovarian cells causes a dramatic increase in membrane lipid poly-unsaturation, which also occurs to an extent in therapy resistant cells. Furthermore, we have shown that cell death in therapy responsive cells is co- incident with a dramatic increase in mitochondrial ROS and membrane lipid poly- unsaturation, occurring to a lesser extent in therapy resistant cells. We next treated therapy resistant cells with a combination of carboplatin and arsenic trioxide, showing that arsenic trioxide treatment causes a potent cytotoxic effect in combination with carboplatin.

Example 12: Membrane lipid poly-unsaturation sensitizes cells to multiple ROS inducers

We have shown that as a general phenomenon, membrane lipid poly- unsaturation caused by either lipogenesis inhibition, lipid uptake or by other indirect chemical interventions sensitises cells to ROS induction, especially in combination with Arsenic Trioxide. In order to investigate if this phenomenon occurs with more general ROS inducers, we next investigated whether membrane lipid poly-unsaturation in 451lu R and D10BMR cells caused by lipogenesis inhibition as well as PUFA supplementation sensitized cells to a range of FDA approved, in clinical trials or commonly used small molecules that induce cellular ROS through multiple mechanisms. We tested a selection of compounds including 2- methoxyestradiol, beta-lapachone, elesclomol, menadione and 5- aminolevulinic acid. Example 13: Membrane lipid poly-unsaturation sensitises cells to glycolytic inhibitors

We have shown that an increase in lipid-poly-unsaturation sensitizes cells to oxidative stress. As glycolytic inhibitors have been shown to increase ROS levels, we reasoned that lipid poly-unsaturation sensitizes cells to glycolytic inhibition. We therefore investigated in multiple cancer cell line models (PANC-1, HepG2, GL261, 451lu, 451 lu R, A549, MDA-231, A2780) whether membrane lipid poly- unsaturation caused by lipogenesis inhibition or PUFA supplementation sensitized cells to 3-bromopyruvate.

Example 14: Gemcitabine sensitises pancreatic cancer cells to ATO

Pancreatic cancer responds poorly to chemotherapy with agents such as gemcitabine being first-line treatment. Gemcitabine is the most widely used agent in treating pancreatic cancer, although cancers often respond either partially or transiently to treatment. In spite of progress in the treatment of many types of cancer, pancreatic cancer retains a dismal prognosis. It has previously been shown that gemcitabine response correlates inversely with lipogenesis and FASN expression, and that FASN inhibition re-sensitizes pancreatic cancer models to gemcitabine treatment. Here we show that the response of PANC-1 cells to gemcitabine is 1 - enhanced by ATO treatment and that 2 - the combination of FASN inhibition and ATO treatment have a greater synergistic effect in inducing cell death.

Example 15: lipogenesis inhibition sensitises multiple non-solid tumour cells to ATO

ATO is approved in treating acute promyeloytic leukaemia and is thought to work by promoting the proteasomal degradation of the PML-PARA fusion onco-protein. However, surprisingly, it has been shown that ATO synergizes with proteasomal barrel inhibitors to promote cancer cell death, and that this is in part through enhancing cellular ROS. In order to investigate whether ATO kills non-solid tumour including APL models by enhancing ROS, we treated a panel of leukemia and lymphoma cell lines with a combination of ATO and FASN inhibitors/ PUFA supplementation and measured cell death and cellular and mitochondrial ROS production, showing that membrane lipid poly-unsaturation sensitizes cells to ATO treatment in causing cell death and is co-incident with an increase in ROS levels. Example 16: ATO sensitises cancers to anti-angiogenesis therapy

Although in general cancer respond poorly to anti-angiogenesis therapy, it remains the first-line treatment in RCC with some patients showing a very good and sustained response, some patients showing a transient response, and others showing no response. It has recently been observed that cancer cells can circumvent the nutrient deprivation effects of anti-angiogenesis treatment by enhancing the uptake of lipids from surrounding adipose tissue and by upregulating beta oxidation, thereby utilizing lipids for energy and to gain metabolic feedstock. Since cancer cells with enhanced beta oxidation are more reliant on functional mitochondria and thereby possibly more sensitive to alterations mitochondrial membrane potential, they may be sensitive to ATO treatment as ATO enhances mitochondrial ROS and uncouples mitochondrial respiration. To this end, we tested the effects of anti-angiogenesis treatments in mouse models of RCC combines with ATO treatment.