WO2019200343A1 - S6k2 blockade perturbs redox balance and fatty acid metabolism, leading to oxidative cell death in mapk inhibitor resistant cancers - Google Patents

Abstract

NRAS Mutant melanoma cells involve certain pathways for induction of ferroptosis. Accordingly, our studies indicate that cancers with MAPK inhibitors will also be susceptible to treatment with the ferroptosis inducing agents. In a particular example, RAS mutant cancers, such as pancreas, lung, colon and thyroid, as non-limiting examples, would be susceptible to treatment with a ferroptosis/lipid peroxidation inducing agent as provided for in the examples, data, and methods as described herein.

Description

S6K2 BLOCKADE PERTURBS REDOX BALANCE AND FATTY ACID

METABOLISM, LEADING TO OXIDATIVE CELL DEATH

IN MAPK INHIBITOR RESISTANT CANCERS

GOVERNMENT SUPPORT

[0001] This invention was made with government support under the Wistar Cancer

Center Support Grant No. P30 CA010815 and NIH Grant No. K01 CA175269. The government has certain rights in the invention.

CROSS REFERENCE TO RELATED APPLICATIONS

[0002] This application claims the benefit of ET.S. Provisional Application Serial No.

62/657,541, filed April 13, 2018, the disclosure content of which is hereby incorporated by reference in its entirety.

FIELD OF INVENTION

[0003] This application is generally related to therapeutic formulations and methods of treatment of cancers, specifically the treatment of cancers patients having tumors resistant to MAPK inhibitors (MAPKi; e.g. RAF, MEK, ERK, RSK inhibitors) such as melanoma.

BACKGROUND OF INVENTION

[0004] Programmed cell death is crucial for various physiological processes in multicellular organisms. Deregulation of programmed cell death contributes to the development of multiple human diseases, such as cancer and neurodegeneration. While apoptosis is the best- studied form of programmed cell death, there are also programmed cell death processes that are not apoptosis. For example, necroptosis and ferroptosis are two distinct regulated necrosis pathways that are under precise genetic control and may function under diverse physiological and pathological contexts.

[0005] Ferroptosis is a programmed necrosis whose execution requires the accumulation of cellular reactive oxygen species (ROS) in an iron-dependent manner. Ferroptosis can be triggered by inhibiting the activity of cystine-glutamate antiporter, leading to the depletion of glutathione (GSH), the major cellular antioxidant whose synthesis requires cysteine. As such, ferroptosis is caused by the loss of cellular redox homeostasis. Further, it appears that lipid ROS/peroxides instead of mitochondrial ROS play more crucial roles in ferroptosis, and inactivation of glutathione peroxidase 4 (GPX4), an enzyme required for the clearance of lipid ROS, can induce ferroptosis even when cellular cysteine and GSH contents are normal. It is further understood that the intracellular metabolic pathway glutaminolysis also plays crucial roles in ferroptosis by promoting cellular ROS generation.

[0006] Indeed, the failure of cell death leads to proliferation of cells and growth in an uncontrolled, cancerous manner. Cancer is a leading cause of death in the developed world, with over one million people diagnosed and more than 500,000 deaths per year in the United States alone. Overall it is estimated that at least one in three people will develop some form of cancer during their lifetime. There are more than 200 different histopathological types of cancer, four of which (breast, lung, colorectal, and prostate) account for over half of all new cases in the U.S. However, melanoma cancers are also especially prevalent, as the population ages, and unhealthy habits and a culture of tanning for social purposes, leading to an explosion of melanoma cases, even in the under age 50 population. [0007] Many of these tumors arise from mutations that activate Ras proteins, which control critically important cellular signaling pathways that regulate growth and other processes associated with tumorigenesis. Indeed, RAS-driven tumors account for almost 40% of cases of melanoma, (due to mutations or alterations in RAS, NF1, receptor tyrosine kinases (RTKs) or G- coupled receptors). The name“Ras” is an abbreviation of“Rat sarcoma” reflecting the way the first members of the Ras protein family were discovered. The name“ras” also is used to refer to the family of genes encoding these proteins.

[0008] Ras-driven cancers have remained the most intractable diseases to any available treatment. New therapeutic and preventative strategies are urgently needed for such cancers. Drug discovery programs worldwide have sought Ras-selective drugs for many years, but heretofore no avail.

[0009] Ras proteins are key regulators of several aspects of normal cell growth and malignant transformation, including cellular proliferation, survival and invasiveness, tumor angiogenesis and metastasis. Ras proteins are abnormally active in most human tumors due to mutations in the ras genes themselves, or in upstream or downstream Ras pathway components, or other alterations in Ras signaling.

[00010] Genetic mutations in ras genes were first identified in human cancer over three decades ago. Such mutations result in the activation of one or more of three major Ras protein isoforms, including H-Ras, N-Ras, or K-Ras, that turn on signaling pathways, such as MAPK and PI3K/mTOR, leading to uncontrolled cell growth and tumor development. Activating Ras gene mutations occur de novo in approximately one third of all human cancers and are especially prevalent in pancreatic, colorectal, and lung tumors. Ras mutations also develop in tumors that become resistant to chemotherapy and/or radiation, as well as to targeted therapies, such as tyrosine kinase inhibitors. While ras mutations are relatively infrequent in other tumor types, for example, breast cancer, Ras can be pathologically activated by certain growth factor receptors (e.g. HER2/Neu) that signal through Ras.

[00011] Indeed, it is the formation and generation of resistance, for example in other pathways such as MAPK, wherein MAPKi resistance includes many therapeutic options, including MEK, ERK, and BRAF inhibitors.

[00012] There are no currently available cancer therapeutics approved by the ET.S. Food and Drug Administration that are known to selectively suppress the growth of tumors driven by activated Ras. In fact, Ras is often consider undruggable because of the relative abundance in cells and high affinity for its substrate, GTP.

[00013] Several approaches to treat diseases that arise from activating ras mutations have been undertaken. Because full maturation of the Ras protein requires lipid modification, attempts have been made to target this enzymatic process with inhibitors of famesyl transferase and geranylgeranyltransferase, but with limited success and significant toxicity. Targeting of downstream components of Ras signaling with inhibitors of Raf/Mek/Erk kinase components of the cascading pathway has been an extremely active area of pharmaceutical research, but also fraught with difficulties and paradoxes arising from complex feedback systems within the pathways.

[00014] Inhibitors targeting components within the PBK/Akt pathway also have not been successful as agents. Similarly, several other molecular targets have been identified from RNAi screening, which might provide new opportunities to inhibit the growth of Ras-driven tumors. [00015] Despite significant progress in treating melanoma and the recent approval of several drugs for metastatic disease, several challenges remain. For example, clinical responses are generally short-lived as tumors quickly become resistant to therapy and patients ultimately relapse. Moreover, tumors can acquire multiple mechanisms of resistance, making the development of second-line therapies extremely daunting. Therefore, it is crucial to identify therapeutic targets that are common for most resistant tumors and to generate methods of treatment for these tumors and cancer cells that heretofore had limited therapeutic treatment options.

[00016] Indeed, the prior art, e.g. Multi-stage differentiation defines melanoma Subtypes with Differential Vulnerability to Drug-Induced Iron-Dependent Oxidative Stress, Tsoi et al; 2018 Cancer Cell 33, 1-15; describes a joint therapeutic approach towards melanoma cancer treatment. Tsoi specifically identifies a combination therapy with a ferroptosis inducing agent a vemurafenib. However, Tsoi importantly argues and addresses that RAS mutation, with no change in differentiation status was insensitive to ferroptosis induction, thus confirming the need for and desire for a multi-tiered strategy.

SUMMARY OF INVENTION

[00017] We found that a protein called S6K is activated in melanoma, particularly in tumors that are resistant to inhibitors of a critical signaling pathway known as MAPK, which includes RAF, MEK and ERK. Selective Inhibition of the S6K2 isoform restrains tumor cell growth and survival. Notably, S6K is a common node for many resistance pathways. This provided a rationale for evaluating S6K as a novel target for melanoma therapy. We investigated the role of S6K in NR.AS mutant melanoma and determined the potential therapeutic value of targeting this protein. We have established the consequences of blocking S6K in melanoma and identified biological features that are regulated by S6K. We are now leveraging our findings to delineate combinatorial approaches that mimic the effects of S6K blockade and can lead to long term inhibition of melanoma, including tumor cells resistant to MAPK inhibitors.

[00018] In a preferred embodiment, a method of treating MAPKi resistant cancerous cells, comprising administering an effective amount of a ferroptosis inducer or an agent that induces lipid peroxidation. In particular embodiments, the ferroptosis inducer is erastin or RSL3, ML162, and ML210. In particular embodiments, the cancerous cells are MAPK inhibitor resistant melanoma cells.

[00019] In a preferred embodiment, a method of treating MEK inhibitor resistant cancerous cells, comprising administering to a patient an effective amount of a ferropoptosis inducer/lipidperoxidation agent. In particular embodiments, the ferropoptosis inducer is erastin, RSL3, ML162, and ML210. In particular embodiments, the cancerous cells are MEK resistant melanoma cells.

[00020] In a preferred embodiment, a method of treating ERK inhibitor resistant cancerous cells, comprising administering to a patient an effective amount of a ferropoptosis inducer. In particular embodiments, the ferropoptosis inducer is erastin or RSL3, ML162, and ML210. In particular embodiments, the cancerous cells are ERK inhibitor resistant melanoma cells.

[00021] In a preferred embodiment, a method of treating BRAF resistant cancerous cells, comprising administering to a patient an effective amount of a ferropoptosis inducer. In particular embodiments, the ferropoptosis inducer is erastin or RSL3, ML162, and ML210. In particular embodiments, the cancerous cells are BRAF inhibitor resistant melanoma cells. [00022] In a preferred embodiment, a method of treating MAPKi resistant melanoma cells comprising contacting said MAPKi resistant melanoma cells with a compound suitable to selective inhibition of S6K2. In a particular embodiment, the compound is suitable for selective inhibition of S6K2 without concomitant inhibition of S6K1.

[00023] In a preferred embodiment, a method of treating a tumor in a patient, comprising: obtaining a tumor sample from said patient and determining whether said tumor is kinase inhibitor resistant; administering to said patient a ferroptosis or lipid peroxidation inducer where said tumor is kinase inhibitor resistant.

[00024] In a preferred embodiment, a method of treating a MAPKi kinase inhibitor resistant tumor cell comprising: contacting said cell with a ferroptosis/lipid peroxidation inducer.

[00025] In a preferred embodiment, a method of treating a RAS mutant cancer, wherein said RAS mutant cancer comprises cells which are MAPK kinase inhibitor resistant and contacting said cancer cells with a ferroptosis/lipid peroxidation inducer.

[00026] In a preferred embodiment, a method of treating melanoma tumor comprising: taking a sample form said melanoma tumor and determining whether said tumor is NRAS mutant; administering to said patient a sufficient amount of a ferroptosis/lipid peroxidation inducer.

[00027] In a preferred embodiment, a method of treating a NRAS mutant melanoma tumor comprising contacting said mutant melanoma tumor with a ferroptosis/lipid peroxidation inducer. [00028] In a preferred embodiment, a method of treating a melanoma tumor by selective inhibition of SK2, comprising administering to said patient an effective amount of a

ferroptosis/lipid peroxidation inducing active agent.

[00029] In a preferred embodiment, any one of the methods described herein, wherein said ferroptosis inducing agent is selected from erastin, RSL3, ML162, and ML210.

[00030] In a preferred embodiment, a method of treating a cancer comprising

administering a first effective amount of a MEK inhibitor and a second effective amount of a ferroptosis /lipid peroxidation inducing agent.

[00031] In a preferred embodiment, a method of treating a patient having a MAPKi resistant tumor comprising, administering to said patient an effective amount of a

ferroptosis/lipid peroxidation inducing agent. In a preferred embodiment, wherein the method comprises further administering an agent for selective blockage of S6K2. In a preferred embodiment, wherein the selective blockage of S6K1 does not impact the levels of S6K1. In a preferred embodiment, wherein selective blockage of S6K2 is sufficient to generate acute depletion of S6K2 enhanced ROS production, lipid synthesis and accumulation of intracellular unsaturated fatty acids; wherein ROS susceptible (poly)-unsaturated fatty acids sensitized cells to ROS, resulting in lipid peroxidation and oxidative cell death.

[00032] In a preferred embodiment, a method of treatment of a patient having a NRAS mutant melanoma comprising administering to said patient an effective amount of a therapeutic sufficient to induce lipid peroxidation. In a preferred embodiment, wherein said therapeutic is a ferroptosis/lipid peroxidation inducing agent. In a preferred embodiment, wherein the step of inducing lipid peroxidation is generated by selective depletion of S6K2. In a preferred embodiment, further comprising administering to said patient a sufficient amount of erastin, RSL3, ML 162, ML210, FIN56 or an S6K2 inhibitor.

[00033] In a preferred embodiment, a method of attenuating the growth rate of a MAPKi sensitive tumor cells comprising: regulating the PI3K/mTOR and MAPK pathways by contacting said tumor cells with a ferroptosis/lipid peroxidation inducing agent.

[00034] In a preferred embodiment, a method for inducing NRAS melanoma cancer cell death by inducing ROS comprising contacting said melanoma cells with a sufficient amount of a polyunsaturated fatty acid. In a preferred embodiment, the method further comprising

administering to said patient a sufficient amount of a ROS inducer. In a preferred embodiment, the method further comprising wherein the ROS inducer is sorafenib or BSO. In a further embodiment, the method further comprising administering a sufficient amount of a ferroptosis inducing agent, such as erastin. In a preferred embodiment, a preferred method further comprises administering to said patient an effective amount of lipid ROS scavenger Fer-l.

BRIEF DESCRIPTION OF THE DRAWINGS

[00035] FIG. 1A - 1G - MAPK and PI3K pathways differentially regulate S6K in NRAS- mutant melanoma cells resistant to MAPK inhibitors. (B-F) Cells were transduced with shRNA against S6K1 or S6K2. (B) S6K1/2 and cleaved-caspase 3 levels were analyzed by

immunoblotting. (C) Cell death was analyzed by PSVue/PI staining. (D) Bar graph comparing cell death in MAPKi-sensitive vs. MAPKi-resistant cells. (E) Reverse phase protein array identified proteins differentially affected by depletion of S6K2 vs. S6K1. (F) ROS levels were assessed using H2DCFDA and analyzed by flow cytometry. (G) S6K1- or S6K2-depleted cells were treated with the ROS scavenger NAC (lmM; cell death was assessed by PSVue/PI staining. [00036] FIGS. 2A - F NRAS mutant melanoma cells were transduced with S6K1 or S6K2 shRNA. (A) Cells were analyzed by reverse phase protein array. (B) Schematic of lipid synthesis and enzymes (red) involved in this process. (C) Unsaturated fatty acids were measured by the sulfo-phospho-vanillin method. (D) Transduced cells were stained with the lipid ROS sensor Bodipy Cl 1 and analyzed by flow cytometry. (E) Transduced cells were treated with the lipophilic antioxidant trolox (100 mM). Cell viability was determined by PSVue/PI and lipid peroxidation was measured by Bodipy Cl 1. (F) Transduced cells were treated with the SCD1 inhibitor, A939572 (5 mM), stained with Bodipy Cl 1 and analyzed by flow cytometry.

[00037] FIGS. 3 A-D (A) Ferroptosis is a form of regulated cell death by lipid

peroxidation. (B) S6K1 or S6K2 depleted cells were subjected to proteomic analysis. (C) S6K2- depleted cells were treated with inhibitors of necroptosis (Nec-l, 20mM), ferroptosis (Fer-l, 2mM), apoptosis (zVAD, 20mM), or combination of Fer-l and zVAD and analyzed by

PSVue/PI staining. (D) GPx activity was determined in transduced cells.

[00038] FIGS. 4A-E (A) Schematics showing strategies to phenocopy S6K2 depletion. (B-E) Cells were treated with GSK458 (PI3K/mTORi, 0.05 mM; B-C) or LY2584702 (panS6Ki, 10 mM; D-E). Cell death (B,D) and lipid peroxidation (C, E) were examined. (F-H) Cells were transduced with S6K2 shRNA alone or both S6K2 and S6K1 shRNA. Expression of SCD1 (F), lipid peroxidation (G) and cell death (H) were determined. (I-J) Cells were transduced with the constitutively active S6Kl(S389) and treated with LY2584702 (10 mM) for 2 days. Expression of S6Kl(S389) and cell death was examined.

[00039] FIGS. 5A-E (A,B) Cells were treated with BSO (ROS inducer), DHA

(polyunsaturated fatty acid) alone or in combination. Lipid peroxidation was assessed by Bodipy Cl 1 staining (A) and cell death by PSVue/PI staining (B). (C-D) cells were treated with erastin +/- the lipid ROS scavenger Fer-l. Lipid peroxidation (C) and cell death (D) were determined. (E) Mice bearing NRAS-mutant tumors were treated with erastin (30mg/kg) twice daily, every other day.

[00040] FIGS. 6A-E depicts that MAPK and PBK/mTOR differentially regulate S6K in NRAS-mutant melanoma cells sensitive v. resistant to MAPK inhibitors.

[00041] FIGS. 7A-F depicts that S6K2 silencing disrupts redox balance and induces cell death in NRAS-mutant melanoma cells.

[00042] FIGS. 8A-H depicts that depletion of S6K2 induces fatty acid synthesis and lipid peroxidation.

[00043] FIGS. 9A-I depicts that selective suppression of S6K2 is necessary to induce maximal lipid peroxidation and cell death.

[00044] FIGS. 10A-I depicts that induction of lipid peroxidation has tumor-suppressive activity in NRAS-mutant melanoma.

[00045] FIGS. 11 A-I depicts that S6K is mainly regulated by the PBK/mTOR but not the MAPK pathway in NRAS-mutant melanoma cells resistant to MAPKi.

[00046] FIGS. 12A-E depicts that S6K2 depletion selective induces cell death in MAPKi- resistant melanoma cells.

[00047] FIGS. 13A-K depicts that depletion of S6K2 leads to oxidative damage and triggers multiple forms of cell death.

[00048] FIGS. 14A-G depicts that concomitant suppression of S6K1 diminishes the effects of S6K2 depletion. [00049] FIGS. 15A-E depicts that Induction of lipid peroxidation restrains NRAS-mutant melanoma.

[00050] FIGS. 16A-G show that S6K2 is a vulnerability in NRAS-mutant melanoma resistant to MAPK inhibitors. FIG. 16(A) depicts NRAS-mutant melanoma cells were treated with the MEKi trametinib (M, 10 nM) or ERKi SCH772984 (E, 100 nM) for 48h and subjected to reverse phase protein array (RPPA) analysis. Heatmap depicts changes in protein expression or phosphorylation relative to the vehicle control (DMSO). FIG. 16(B) depicts cells that were transduced with doxycycline-inducible wild-type S6K2 (WT) or constitutively active S6K2E388 constructs, treated with doxycycline for 24h, and then treated with the ERKi SCH772984 (SCH984, 1000 nM) or the PBKi GSK2126458 (GSK458, 100 nM) for additional 24h.

Percentage of pS6 (S240/244)+ cells was determined by flow cytometry. Data are mean ± SD (n=2) from a representative experiment; p values were determined by Student’s t-test. Only p<0.0l are shown. FIG. 16(C) depicts a diagram illustrating that S6K is mainly regulated by the PBK/mTOR pathway in NRAS-mutant melanoma cells resistant to MAPKi, whereas it can be regulated by both the MAPK and the PI3K pathways in cells sensitive to MAPKi. FIGS. l6(D- G) depict cells that were transduced with lentiviruses encoding S6K1 or S6K2 shRNA. Cell death was analyzed by PSVue/PI staining (6-7 days post infection; dpi) in MAPKi-resistant (D) or MAPKi-sensitive (F) cells. FIG. 16(E) depicts levels of S6K1/2 and cleaved-caspase 3 that were analyzed by immunoblotting. Numbers indicate protein quantification relative to actin. Data shown in (D) are mean ± SD of two independent experiments, and in (F) are mean ± SD (n=2) of a representative experiment. FIG. 16(G) depicts percent cell death induced by S6K1 or S6K2 depletion in three MAPKi-R (D) and three MAPKi-S (F) cell lines was compared by unpaired, two-tailed Student’ s t-test. Data from vector control, two S6K1 or two S6K2 short RNA hairpins were pooled; n.s. denotes not significant. MAPKi-R cell lines, highlighted in orange; MAPKi-S cell lines, highlighted in blue.

[00051] FIGS. 17A-P relate to MAPK and PI3K/mTOR differentially regulate S6K in NRAS-mutant melanoma cells sensitive vs. resistant to MAPK inhibitors. FIG. 17(A) depicts NRAS-mutant (Q61K/R/L) melanoma cells that were treated with the MEKi trametinib (Tram;

0.1-1000 nM) or the ERKi SCH772984 (SCH984; 1-10 mM) for 72 h and cell viability was assessed by MTT assays. Area under the dose-concentration curve (AETC) was calculated as a measure of drug sensitivity. FIGS. 17 (B-C) depict cells that were treated with trametinib (100 nM) or SCH772984 (1000 nM) for 72 h. FIG. 17(B) depicts relative proliferation that was determined by BrdET incorporation relative to the vehicle (DMSO). Data represent mean ± SD (n=3). FIG. 17(C) depicts cell death that was assessed by PSVue/PI staining an analyzed by flow cytometry. Data from two independent experiments are shown as mean ± SD. FIG 17(D) depicts NRAS-mutant melanoma cells that were treated with SCH772984 (1000 nM) for 24h and analyzed by immunoblotting. FIG. 17(E) depicts cells that were treated with trametinib (100 nM) or SCH772984 (1000 nM) for the indicated times and analyzed by immunoblotting. FIG. 17(F) depicts cells that were treated with increasing doses of the MEKi MEK-162 or the ERKi

BVD523 for 72h and relative cell viability was determined by CellTiter-Glo. FIG. 17(G) depicts basal levels of pERK were determined by immunoblotting. Blots were quantified and pERK levels were normalized to actin (loading control); pERK levels relative to UACC1273 are shown. pERK basal levels in MAPKi-R and MAPKi-S cells were compared by unpaired t-test (right panel); ** p<0.0l. FIG. 17(H) is a scatterplot depicting correlation between sensitivity (defined by AUC) of NRAS-mutant melanoma cells to the MEKi trametinib or the ERKi SCH772984 and basal levels of pERK. FIG. 17(1) depicts cells that were treated with the PI3K/mTORi GSK458 (0.1 mM) for 24h. pS6(S240/244) levels and assessed by immunoblotting. FIG. l7(J) depicts cells that were treated with the PI3K inhibitor LY294002 (10 mM), the PBK/mTOR inhibitor GSK458 (0.1 mM), or the mTOR inhibitor INK128 (0.1 mM). Phosphorylation of S6 was determined by immunoblotting. All samples were run on the same gel (gaps indicate where lanes were removed for simplicity). FIG. 17(K) depicts cells transduced with doxycycline-inducible constitutively active S6K1E389 constructs and treated with doxycycline for 24h and then treated with SCH984 (1000 nM) or GSK458 (100 nM) for additional 24h. Percentage of

pS6(S240/244)+ cells was assessed by FACS. Data represent mean ± SD (n=2) from a representative experiment. * p<0.0l, ** p<0.00l when comparing vector- vs. S6K1E389- transduced cells. FIGS. l7(L-M) depicts cells that were transduced with lentiviruses encoding S6K1 or S6K2 shRNA for 6 days. FIG. l7(L) depicts transduced cells that were labeled with BrdET (4h), stained with PI and analyzed by flow cytometry. Percentage of BrdU+ cells is depicted in the upper right comer. FIG. 17(M) depicts S6K1 or S6K2 knockdown efficiency that was assessed by qRT-PCR. Data represent mean ± SD (n=3). FIG. 17(N) WM1361 A cells were transduced with lentiviruses encoding two additional S6K2 shRNA for 6-7 days. Knockdown efficiency of S6K2 was assessed by immunoblotting (upper panel). Cell death was determined by PSVue/PI staining (FACS; bottom panel). Data represent mean ± SD (n=2). FIG. 17(0) depicts MAPKi-S cells that were transduced with lentiviruses encoding S6K1 or S6K2 shRNA for 6-7 days. Knockdown efficiency of S6K1 or S6K2 was assessed by qRT-PCR. Data represent mean ± SD (n=3). FIG. 17(R) depicts human fibroblasts that were transduced with lentiviruses encoding shRNA against S6K1 or S6K2 for 6 days. Knockdown efficiency for S6K1 or S6K2 was assessed by immunoblotting (upper panel). Cell viability was determined by MTT assays (bottom panel). Data from a representative experiment are shown as mean ± SD (n=7). In FIGS. 17(M-R), statistical significance was calculated by Student’s t-test; ** p<O.Ol, *** p<0.00l, n.s. not significant. MAPKi-R cell lines: orange lines or highlighted in orange; MAPKi-S cell lines: blue lines or highlighted in blue.

[00052] FIGS. 18 A-J relate to depletion of S6K2 triggers lipid peroxidation facilitating oxidative cell death. NRAS-mutant melanoma cells were transduced with lentiviruses encoding S6K1 or S6K2 shRNA. FIG. 18(A) depicts NRAS-mutant melanoma cells that were analyzed by reverse phase protein array (4-6 dpi). Heatmap depicts fold change (log2) of sh/vector. FIG. 18(B) depicts transduced cells (6 dpi) that were stained with the ROS sensing dye H2DCFDA and analyzed by flow cytometry. FIG. 18(C) depicts transduced M93-047 cells that were subject to proteomic analysis (4 dpi). Heatmap shows selected proteins associated with lipid

peroxidation. FIG. 18(D) relates to relative levels of unsaturated fatty acids were determined in transduced cells (4 dpi) using the sulfo-phospho-vanillin method. FIG. 18(E) depicts transduced cells (6 dpi) that were stained with the lipid ROS/lipid peroxidation sensor Bodipy-Cl 1 and analyzed by flow cytometry. FIG. 18(F) depicts S6K2-depleted cells (1 dpi) that were treated with the SCD1 inhibitor, A939572 (5 mM, 3 days), stained with Bodipy-Cl 1 and analyzed by flow cytometry. Horizontal bars indicate gates for Bodipy-Cl 1+ cells. FIG. 18(G) depicts transduced cells (2 dpi) that were treated with the ROS scavenger NAC (1 mM) for additional 3 days and cell death was assessed by PSVue/PI staining (flow cytometry). FIG. 18(H) depicts S6K2-depleted cells (1 dpi) that were treated with the lipophilic ROS scavenger trolox (100 pM) for 3 days, stained with Bodipy-Cl 1 (left) or PSVue/PI (right), and analyzed by flow cytometry. FIG. 18(1) depicts S6K2-depleted cells (2 dpi) that were treated with the iron chelator DFO (5 pM) for additional 4 days, and analyzed by Bodipy-Cl 1 (left) or PSVue/PI staining (right). FIG.

18(J) depicts S6K2-depleted cells (1 dpi) that were treated with inhibitors of necroptosis (necrostatin-l/Nec-l, 20 mM), ferroptosis (ferrostatin-l/Fer-l, 2 mM), apoptosis (zVAD, 20 mM), or combination of Fer-l and zVAD for additional 3 days, and analyzed by PSVue/PI staining. In FIGS. l8(D, G-J), data from a representative experiment were plotted as mean ± SD (n=2).

Statistical significance was assessed by Student’s t-test. * p<0.05, ** p<0.0l, *** p<0.00l.

[00053] FIGS. 19 A-M show the depletion of S6K2 leads to widespread oxidative damage and cell death. FIG. 19(A) depicts a schematic of enzymes (red) involved in lipid synthesis. FIGS. 19(B-M) depict melanoma cells that were transduced with lentiviruses encoding S6K1 or S6K2 shRNA. FIG. 19(B) depicts expression of a panel of genes involved in fatty acid synthesis that was assessed by qRT-PCR in transduced cells (4 dpi). Data represent mean ± SD (n=3). FIG. l9(C-D) depict lipidomic analysis that was performed in transduced M93-047 cells (4 dpi). FIG. 19(C) depicts triacyl glycerol contents (upper panel) and percentage of TG subclasses in each treatment (bottom panel). Data represent mean ± SD (n=3). FIG. 19(D) is a heatmap depicting two-chain fatty acid phospholipids where most species (top3) were differentially affected by S6K2sh vs. S6Klsh (Student’s t-test, p<0.05). FIG. 19(E) depicts cells transduced with two additional S6K2 hairpins that were stained with Bodipy-Cl 1 and analyzed by flow cytometry. FIG. l9(F-G) depict S6K1- or S6K2-depleted M93-047 cells (6 dpi) that were stained for 4-HNE adducts and analyzed by flow cytometry (F) or immunofluorescence (G). Scale bar = 100 pm. FIGS. 19(H-I) depict transduced cells (6 dpi) that were stained with the DNA/RNA damage marker, 8-OHdG, and analyzed by flow cyto etry (H) or immunofluorescence (I). Scale bar = 100 pm. FIGS. l9(J-K) depicts DNA double strand breaks that were assessed by r-gH2AC (6 dpi, flow cytometry) (J) or 53BP1 staining (8 dpi, immunofluorescence) (K). In FIG. l9(J), data represent mean ± SD (n=2). In K, quantification of cells > 7 foci in six fields is depicted (mean ± SD). FIG. 19 (L) depicts S6K2-depleted cells (WM1361 A, 2 dpi) that were treated with DFO (2 mM) for additional 4 days, and analyzed by Bodipy-Cl 1. FIG. 19(M) depicts S6K2-depleted cells (1 dpi) that were treated with ferrostatin-l (Fer-l, 2 mM) or liproxstatin-l (Lipl, 1 mM) for 8 days, and analyzed by PSVue/PI staining. Data represent mean ± SD (n=2). Statistical significance was assessed by Student’ s t-test. * p<0.05, ** p<0.0l, *** p<0.00l.

[00054] FIGS. 20 A-l show that the depletion of S6K2 leads to S6K1 upregulation prompting lipid synthesis and cell death. FIGS. 20(A-C) depict M93-047 NRAS-mutant melanoma cells that were transduced with S6K1 or S6K2 shRNA (3 dpi) and analyzed by RNA- sequencing. FIG. 20(A) relates to genes overrepresented in S6K2-depleted cells predicted activation of SREBP (SREBF). Heatmap showing z-scores of the prediction. FIG. 20(B) depicts a heatmap showing expression of SREBP target genes. FIG. 20(C) depicts M93-047 cells that were transduced with an empty vector control (Vec) or S6K2 shRNA +/- SREBP siRNA. Cell death was assessed by Pi/PSVue staining. Results from one representative experiment are shown as mean ± SD (n=2). FIGS. 20(D-F) depict cells that were transduced with lentiviruses encoding shRNA against S6K2, S6K1 or S6K2 and S6K1 together. FIG. 20(D) depicts S6K1 and S6K2 mRNA levels that were determined by RNA-sequencing. FIG. 20(E) depicts expression of SREBP 1 and SCD1 (3 dpi) that was assessed by qRT-PCR. Results from one representative experiment are shown as mean ± SD (n=3). FIG. 20(F) depicts lipid peroxidation that was assessed in transduced cells (6 dpi) by Bodipy-Cl 1 staining. FIG. 20(G) depicts cell death that was determined by PSVue/PI staining (4 dpi; F). FIGS. 20(H-I) depict M93-047 cells that were transduced with empty vector control (Vec) or constitutively active FLAG-tagged S6K1S389. Transduced cells were then treated with DMSO (-) or the pan S6Ki LY4702 (10 mM) for 48h. Transduced cells were analyzed by immunoblotting with the indicated antibodies (H). Cell death was determined by PSVue/PI staining and flow cytometry (I). Results from one representative experiment are shown as mean ± SD (n=2). Statistical significance was assessed by Student’s t- test; *p<0.05, **p<0.0l, ***p<0.00l, n.s., not significant.

[00055] FIGS. 21 A-L relate to concomitant suppression of S6K1 and S6K2 diminishes the effects triggered by S6K2 depletion. FIGS. 21 (A-F) depict M93-047 NRAS-mutant melanoma cells that were transduced with S6K1 and/or S6K2 shRNA. FIG. 21(A) depicts levels of cleaved SREBP1 (nSREBPl) that were assessed by immunoblotting in S6K2-depleted cells (3dpi). nSREBPl levels were quantified and normalized to actin; numbers below each lane indicate normalized nSREBP levels relative to vector control. FIG. 21(B) depicts cells that transduced with vector control of S6K2 shRNA (2 dpi) were transfected with SREBPlsiRNA. S6K2 and SREBP1 mRNA levels were determined by qRT-PCR. FIG. 21(C) depicts S6K1 and S6K2 protein levels that were assessed by immunoblotting in transduced cells (4 dpi). S6K1/2 protein expression levels were normalized to actin. Numbers below each lane indicate normalized protein levels relative to vector control. FIG. 21(D) depicts PTGS2 mRNA levels (4 dpi) that were assessed by qRT-PCR. FIG. 21(E) depicts S6K1 and S6K2 mRNA levels for data shown in FIGS. 20(D-F) that were determined by qRT-PCR (4 dpi). Only changes > 50% are considered significant. Data from a representative experiment are shown in bar graphs as mean ± SD (n=3). *** = p<0.00l; n.s.= not significant when comparing mRNA levels to those in vector- transduced cells. FIG. 21(F) depicts representative images of cells transduced with the indicated shRNAs (10 dpi). Scale bar = 300 pm. FIGS. 2l(G-I) depict M93-047 cells that were treated with the S6Kli PF4708671 (10 pM) or pan-S6Ki LY2584702 (10 pM) for 48h or as indicated. FIG. 21(G) depicts expression of genes involved in fatty acid synthesis that was determined by qRT-PCR. FIG. 21(H) depicts lipid peroxidation that was determined by Bodipy-Cl 1. FIG. 21(1) depicts cell death that was assessed by PSVue/PI staining. Suppression of S6K activity was determined by immunoblotting of pS6 (inset). All proteins were run on the same gel (gaps denote lanes that were removed for simplicity). FIGS. 2l(J-L) depict NRAS-mutant cells that were treated with the PBK/mTOR inhibitor GSK458 (0.05 mM) (J) Expression of lipid synthesis genes was determined by qRT-PCR. FIG. 2l(K) depicts lipid peroxidation that was determined by Bodipy-Cl 1 staining after 24h treatment. FIG. 2l(L) depicts cell death that was determined by PSVue/PI staining 72h post treatment. Inhibition of S6K was assessed by immunoblotting of pS6 24h post treatment (inset). Data shown from one representative experiment as mean ± SD (n=3) for FIGS. 21(G) and 2l(J) and mean ± SD (n=2) for FIG. 21(1). In FIG. 2l(L), data represent mean ± SD of two independent experiments. Statistical significance was assessed by Student’s t-test. * p<0.05, ** p<0.0l, *** p<0.00l, n.s., not significant.

[00056] FIGS. 22 A-D relate to NRAS-mutant melanoma resistant to MAPKi is sensitive to induction of lipid peroxidation. FIG. 22(A) depicts cells that were treated with increasing doses of the ferroptosis inducers erastin or RSL3, for 24 h and lipid peroxidation was assessed by Bodipy-Cl 1 staining. FIG. 22(B) depicts cells that were treated with indicated concentrations of erastin or RSL3 ± the ferroptosis inhibitor ferrostatin-l (Fer-l, 2 pM) for 48 h. Cell death was determined by PSVue/PI staining. Data from a representative experiment are shown as mean ± SD (n=2). FIGS. 2l(C-D) depict mice bearing established NRAS-mutant tumors that were treated with erastin (30mg/kg, twice daily, every other day). FIG. 22(C) depicts tumor volume at day 12 (M93-047) or day 24 (WM1361 A). FIG. 22(D) depicts tumors that were analyzed by immunohistochemistry for 4-HNE and 8-OHdG. Scale bars = 200 pm.

[00057] FIGS. 23 A-I show how the suppression of S6K2 relieves negative regulation on PPARa facilitating lipid peroxidation. In FIGS. 23(A-D, G-H), M93-047 cells were transduced with S6K1 or S6K2 shRNA. In FIGS. 23(A-B), transduced cells (3 dpi) were analyzed by RNA sequencing. FIG. 23(A) shows a heatmap depicting fold change (log2 sh/vec control) for selected PPAR target genes. FIG. 23(B) shows a heatmap depicting Z-scores of predicted regulators in the PPAR family. Regulators with targets overrepresented were identified by IPA. FIG. 23(C) depicts PPAR mRNA expression that was determined by qRT-PCR (4 dpi). Data from a representative experiment are shown as mean ± SD (n=3). FIG. 23(D) depicts H2DCFDA (upper), Bodipy-Cl 1 (middle) and PSVue/PI (bottom) staining in transduced cells (6-7 dpi). Data from a representative experiment are shown as mean ± SD (n=2). FIG. 23(E) depicts S6K2 immunoprecipitation and immunoblotting with the indicated antibodies in M93-047. FIG. 23(F) depicts interaction of S6K2 with PPARa or NCoRl that was determined by proximity ligation assay (PLA). Left: Representative images. Scale bar = 30 pm. Right: Bar graphs show

quantification of images. Data indicate mean ± SD (n denotes numbers of cells). FIG. 23(G) depicts cell lysates from transduced cells (4 dpi) that were fractionated into cytosol (CE), nuclear (NE) and chromatin-bound (ChB) fractions and analyzed by immunoblotting. WE: whole cell extract. Protein levels relative to vector control are shown below each band. FIG. 23(H) depicts PLA that was performed in transduced cells (4 dpi). Left: Representative images. Scale bar = 30 pm. Right: Bar graphs show quantification of images. Data represent mean ± SD (n denotes fields quantified). Statistical significance was assessed by Student’s t-test. * p<0.05, ** p<0.0l, *** p<0.00l. FIG. 23(1) depicts a model illustrating regulation of PPARa by S6K2

[00058] FIGS. 24 A-F shows how PPARa facilitates cell death induced by S6K2 silencing. FIG. 24(A) depicts PPAR mRNA levels that were analyzed by qRT-PCR in S6K1- depleted cells (4 dpi). FIG. 24(B) depicts cells that were transduced with S6K2 and/or PPARa shRNAs. S6K2 and PPARa levels (3 dpi) of cells depicted in FIG. 22D were determined by immunoblotting and qPCR respectively. FIGS. 24(C-D) depict cells that were treated with the indicated concentrations of PPARa agonists (fenofibrate, in FIG. 24(C) or GW7647, in FIG. 24(D)) for 3 days and analyzed by H2DCFDA, Bodipy-Cl 1 or PSVue/PI staining. Data from a representative experiment are shown in bar graphs as mean ± SD (n=2). FIG. 24(E) depicts representative PLA images of single antibody control for FIG. 23F. Scale bar = 30 pm. FIG. 24(F) depicts S6K1 and S6K2 mRNA levels for FIG. 23H that were assessed by qRT-PCR. Data shown are mean ± SD (n=3). Statistical significance was assessed by Student’s t-test; * p<0.05, ** p<0.0l, *** p<0.00l, when comparing treated or transduced cells with untreated or vector transduced controls.

[00059] FIGS. 25 A-H shows how mimicking the effects of S6K2 loss restrains the growth of NRAS-mutant melanoma resistant to MAPKi. FIG. 25(A) depicts relative expression of PPARa in MAPKi -R and MAPKi-S cells determined by immunoblotting (see also FIG.

26(A)) was compared by Student’s T-test. Data represent average levels from two independent experiments. FIG. 25(B) depicts cells that were treated with increasing doses of fenofibrate for 72h; cell viability determined by CellTiter-Glo. Data represent average levels from two independent experiments. MAPKi-R cell lines, orange; MAPKi-S cell lines, blue. FIG. 25(C) depicts AUC for MAPKi-R and MAPKi-S cells shown in FIG. 25(B), were compared by Student’s T-test. FIG. 25(D) shows M93-047 cells were treated with fenofibrate, DHA, as single agents or in combination for 72h. Lipid peroxidation and cell death were assessed by Bodipy- Cl 1 and PSVue/PI staining respectively. FIG. 25(E) depicts M93-047 cells that were treated with fenofibrate, erastin as single agents or in combination for 72h. Lipid peroxidation and cell death were measured by Bodipy-Cl 1 and PSVue/PI staining respectively. Data from a representative experiment are shown in bar graphs as mean ± SD (n=2). FIG. 25(F) depicts correlation of PPARa expression and MAPK signature in TCGA skin cutaneous melanoma dataset (NRAS-mutant melanoma patients, n=76). FIG. 25(G) depicts mice bearing established NRAS-mutant tumors that were treated with fenofibrate (200 mg/kg), DHA (300 mg/kg), or erastin (30 mg/kg) as single agents or in combination. Tumor volume after 24 days of treatment is shown. Statistical significance was assessed by Student’s T-test; *p<0.05, **p<0.0l,

***p<0.00l, n.s., not significant when comparing treated tumors vs. vehicle. FIG. 25(H) depicts a model illustrating that uncoupling S6K1 and S6K2 triggers oxidative cell death in NRAS- mutant melanoma resistant to MAPKi.

[00060] FIGS. 26 A-I show how PPARa agonists cooperate with lipid peroxidation enhancers to induce antitumor activity in NRAS-mutant melanoma. FIG. 26(A) depicts expression levels of PPARa and PPARy that were determined by immunoblotting. PPARa levels were normalized to actin (loading control); numbers indicate PPARa levels relative to

UACC1273. FIG. 26(B) depicts relative levels of PPARy in MAPKi-R and MAPKi-S cells was compared by Student’s t-test. FIG. 26(C), depicts mRNA expression that was determined by qRT-PCR in cells treated with the indicated doses of fenofibrate for 24h. Data from a

representative experiment are shown as mean ± SD (n=3). FIG. 27(D) depicts M93-047 cells that were treated with the SREBP1 activator U18666A (2.5 mM) and fenofibrate (25 or 50 pM) as single agents or in combination for 72h. Cell death was determined by PSVue/PI staining.

FIG. 27(E) depicts M93-047 cells that were treated with fenofibrate +/- DHA for 72h in the absence or presence of Lip-l (1 pM). Cell death (PSVue/PI) was determined by flow cytometry. FIG. 27(F) depicts WM1366 cells that were treated with fenofibrate or DHA as a single agent, or in combination for 72h. Lipid peroxidation and cell death were assessed by Bodipy-Cl 1 and PSVue/PI staining. FIG. 27(G) depicts WM1366 cells that were treated with fenofibrate (50 pM) and increasing doses (15-60 pM) of arachidonic acid as single agents or in combination for 72h. Lipid peroxidation (top panel) and cell death (bottom panel) were measured by Bodipy-Cl 1 and PSVue/PI staining respectively. FIG. 27(H) depicts WM1366 cells that were treated with fenofibrate +/- erastin for 72h. Lipid peroxidation and cell death were determined by Bodipy- Cl 1 (top) and PSVue/PI staining (bottom). FIGS. 27(D-H) show data from a representative experiment are shown as mean ± SD (n=2). FIG. 27(1) depicts weight of animals enrolled in study depicted in FIG. 25(G) after 24 days of treatment. Mice were treated with fenofibrate (200 mg/kg), DHA (300 mg/kg) or erastin (30 mg/kg) as a single agent, or in combination. Statistical significance was assessed by Student’ s t-test; *p<0.05, **p<0.0l, ***p<0.00l, n.s., not significant.

DETAILED DESCRIPTION OF THE EMBODIMENTS

[00061] In melanoma, RAS signaling is often deregulated; while activating mutations in NRAS are present in >25% of tumors, RAS-driven tumors account for almost 40% of cases. However, there are no effective therapies for this type of tumors. Although oncogenic NRAS activates the MAPK pathway, inhibition of this pathway alone has limited anti-tumor efficacy and most NRAS-mutant tumors are resistant to MAPK inhibitor (MAPKi) monotherapy. In this study, we investigated potential mechanisms that could be associated with resistance to MAPKi in NRAS mutant melanoma.

[00062] We demonstrated that NRAS-mutant cells display heterogeneous responses to MAPK inhibition. We also found that sensitivity to MAPK inhibition (MAPKi) is coupled to inhibition of S6 kinase (S6K) activity. We discovered that NRAS-mutant melanomas resistant to MAPKi are highly dependent on the ribosomal protein S6 kinase 2 (S6K2); selective depletion of S6K2 triggers multiple types of cell death in NRAS-mutant melanoma. Our data support a model whereby S6K2 depletion enhanced ROS production, lipid synthesis and accumulation of intracellular unsaturated fatty acids, leading to lipid peroxidation and cell death. Based on these findings, we evaluated strategies mimicking depletion of S6K2 and demonstrated that induction of lipid peroxidation suppresses melanoma tumor grow.

[00063] Accordingly, we posited to test the hypothesis that melanomas are dependent on S6K for proliferation and survival.

[00064] We discovered that NRAS-mutant melanomas resistant to MAPK inhibitors (MAPKi) are highly dependent on the ribosomal protein S6 kinase 2 (S6K2) for survival.

Specifically, we demonstrated that:

[00065] NRAS-mutant melanomas that are resistant to MAPKi have low levels of pERK and the regulation of S6K mainly relies on the PI3K pathway;

[00066] While the two S6K isoforms (S6K1 and S6K2) are essential for cell proliferation, S6K2 is required for survival of NRAS-mutant melanoma cells; only depletion of S6K2 (but not S6K1) leads to cell death.

[00067] We also identified that while depletion of S6K2 selectively kills NRAS mutant melanoma cells resistant to MAPKi, it has negligible effects on MAPKi-sensitive NRAS mutant melanoma or non-transformed human fibroblasts.

[00068] Mechanistically, depletion of S6K2 induces ROS levels, fatty acid synthesis and accumulation of unsaturated fatty acids, which together induce lipid peroxidation and oxidative cell death.

[00069] We further determined that S6K2 depletion triggers multiple forms of

programmed cell death including apoptotic and ferroptotic cell death. [00070] These results establish the critical role of S6K in melanoma cells resistant to MAPKi and could have important therapeutic implications, as the type of cell death triggered by a particular treatment may affect the tumor microenvironment and impact immune responses.

[00071] As a second strategy, we determined and evaluated the therapeutic value of blocking S6K in melanoma.

[00072] We discovered that selective inhibition of S6K2 is necessary to trigger melanoma cell death and that S6K2 blockade could be a strategy for melanoma therapy. This strategy is remarkably distinct from using PBK/mTOR inhibitors, which concurrently suppress S6K1 and S6K2, and are often cytostatic.

[00073] We found that:

[00074] Concomitant suppression of S6K1 and S6K2 by PBK/mTOR inhibitors or pan- S6K inhibitors fails to phenocopy depletion of S6K2.

[00075] Inhibition of S6K1 blunts the effects of S6K2 depletion.

[00076] Mimicking the effects of S6K2 loss by combining a ROS inducing agent with polyunsaturated fatty acids, or treating tumor cells with lipid peroxidation inducers restrains NRAS-mutant melanoma growth in vitro and in vivo.

[00077] NRAS mutant melanoma are highly sensitive to lipid peroxidation-mediated cell death.

[00078] Due to the challenges of developing selective S6K2 inhibitors, we posit that the selectivity could be achieved by identifying and targeting specific downstream effectors of S6K2. To this end, we performed RNA-seq, proteomics and lipidomics, and have identified candidate pathways that mediate S6K2 loss-triggered melanoma cell death. Our findings better elucidate the role of S6K in melanoma and to delineate effective therapeutic strategies to target NRAS driven tumors and melanoma resistant to MAPKi.

[00079] Accordingly, we identify that loss of S6K2 selectively leads to targeting of NRAS tumors and melanoma resistant MAPKi tumors for treatment. Therefore, NRAS mutated and RAS based cancers can be drugged by selectively targeting S6K2. We mimic loss or blockade of S6K2 by targeting selective upstream regulators of S6K2 or downstream effectors of S6K2.

[00080] In certain applications, we can then combine therapies to both include a ferroptosis/lipid peroxidation inducing agent and a secondary inhibitor to increase efficacy. While we would not use a selective BRAF inhibitor with an NRAS mutant tumor, certain of these tumors are susceptible to a multi-tiered approach using both the ferroptosis inducing agent and RAF, MEK, ERK or RSK inhibitors within the MAPK pathway, to allow for and generate an increased response when compared to MEK inhibitor alone. Such combination leads to an increase in ferroptosis and thus selective cell death of the particular tumor cells. Likewise, combinations of ferroptosis or lipid peroxidation agents could be combined with other inhibitors of cancer/melanoma-relevant pathways or processes to increase efficacy, duration of response or diminish toxicity by using sequential or alternate schedules or doses

[00081] Suitable ferroptosis-inducing agents include: erastin, PE, IKE, other erastin analogs; sulfasalazine; glutamate; BAY 43-9006; Sorafenib; L-Buthionine-(S,R)-Sulfoximine (BSO); N-Acetyl-4-benzoquinone imine (NAPQI); scetaminophen; (l S,3R)-RSL3; ML162,

DPI compounds 7,10, 12, 13, 17, 18, 19; GPx4 Antibody; Cystine/cysteine deprivation; BSO; DPI2; cisplatin; FIN56; FIN02; (-)-FIN02; statins (e.g., cerivastatin, simvastatin); cysteinase; silica-based nanoparticles; CCl₄; ferric ammonium citrate; trigonelline; brusatol; artemisinin; artesunate; and combinations thereof. [00082] Lipoxygenases: 5 -Lipoxygenase (potato); 5 -Lipoxygenase (human recombinant); 12-Lipoxygenase (platelet-type, mouse recombinant); Lipoxygenase from Glycine max

(soybean) - Purified; 15 -Lipoxygenase-2 (human recombinant)

[00083]

RESULTS

[00084] S6K is differentially regulated in NRAS-mutant melanoma cells sensitive vs. resistant to MAPK inhibitors

[00085] To identify factors responsible for the differential sensitivity to MAPK inhibitors (MAPKi) and potential vulnerabilities in cells resistant to MAPKi, we treated NRAS-mutant melanoma cells (Figure 11 A) with the MAPKi trametinib (MEK inhibitor) or SCH772984 (ERK inhibitor). MAPKi elicited variable effects on viability, proliferation, and cell death, despite effective and persistent inhibition of the MAPK pathway (Figures 11 A-E). Based on the biological response to MAPKi, we classified cells as MAPKi-resistant or -sensitive. Response to MAPKi was assessed by treating NRAS mutant melanoma cells with increasing doses of MEKi or ERKi. MAPKi-resistant cells were characterized by area under the curve/AUC>2.4

(trametinib) or AUC>2.3 (SCH772984), modest suppression of proliferation, and negligible induction of cell death (Figures 11 A-C). Sensitivity to MAPK inhibition was associated with basal levels of pERK (Pearson’s correlation |R|<-0.7 P<0.0l) (Figures 11F and 1A): whereas melanoma cells with high pERK levels were sensitive to MAPK inhibition, melanoma cells with low basal pERK levels were resistant to MAPKi. Trametinib or SCH772984 effectively suppressed the MAPK pathway in both sensitive and resistant cells (Figures 1B, 11D and 11E). Additionally, the PBK/Akt pathway was similarly activated in all MAPKi-treated cell lines (Figures 6B and 11D). In contrast, MAPKi suppressed phosphorylation of S6 kinase (S6K) and its substrate S6 in MAPKi-sensitive but not in MAPKi-resistant melanoma cells (Figures 6B and 11D), raising the possibility that S6K activity could be linked to response to MAPKi in NRAS- mutant melanoma cells. As S6K is a bonafide downstream effector of the PBK/mTOR pathway, PBK/mTORi suppressed phosphorylation of S6 (p-S6) (Figures 6C and 11G) and reduced cell viability (Figure 11H), indicating that the PBK/mTOR pathway can regulate S6K in NRAS- mutant melanoma cells in both MAPKi resistant and sensitive cells. To further examine the impact of MAPK or PBK/mTOR signaling on S6K (assessed by p-S6^S240/244, a site exclusively phosphorylated by S6K) (Roux, 2007 #421 } (Corcoran, 2013 #403 } (Meyuhas, 2008 #473 }, we ectopically expressed constitutively active S6K (S6K1E389 or S6K2E388) constructs in

MAPKi-sensitive or -resistant cells (Figures 1D and 1 II). Treatment of transduced cells with MAPKi or PBK/mTORi decreased phosphorylation of S6 in MAPKi-sensitive cells.

Additionally, constitutively active S6K (S6K1E389 or S6K2E388) restored phosphorylation of S6, indicating that S6K mediates both MAPK and PBK/mTOR signaling in MAPKi-sensitive cells (Figures 6D and 111). In contrast, while MAPKi had no effect on phosphorylation of S6 in MAPKi-resistant cells, PBK/mTORi completely suppressed p-S6. Furthermore, expression of constitutively active S6K constructs (S6K1E389 or S6K2E388) restored phosphorylation of S6 in cells treated with PBK/mTORi, indicating that in MAPKi-resistant cells, S6K activity is mainly controlled by the PBK/mTOR pathway (Figures 6D and 111). Together, these data indicate that S6K1/2 can be controlled by the PBK/mTOR and MAPK pathways in MAPKi- sensitive cells, but is primarily regulated by the PBK/mTOR pathway in MAPKi-resistant NRAS-mutant melanoma cells. (Figure 6E). Therefore, in MAPKi sensitive cells, MAPK inhibitors block S6K activity; in MAPKi-resistant cells, treatment with MAPKi does not affect S6K activity.

[00086] S6K2 is required to maintain redox balance in NRAS-mutant melanoma cells

[00087] To further dissect the role of S6K in NRAS-mutant melanoma cells, we silenced

S6K1 or S6K2. Depletion of either S6K1 or S6K2 impaired cell proliferation. (Figures 7A and 12 A). Intriguingly, depletion of S6K2 but not S6K1, induced cell death as indicated by cleaved caspase 3 and PSVue/PI staining (Figures 7B, 7C and 12B). Notably, this effect appears to be selective to NRAS-mutant melanoma cells that are less reliant on the MAPK pathway, as depletion of S6K2 did not trigger cell death in MAPKi-sensitive NRAS-mutant melanoma cells (Figures 12C and 12D) or non-transformed human fibroblasts (Figure 12E).

[00088] To investigate the mechanism(s) whereby S6K2 depletion triggers cell death, we performed reverse phase protein array analysis, looking for proteins that were differentially affected by depletion of S6K2 versus S6K1. Depletion of S6K2 led to increased levels of proteins involved in oxidative stress response (ROS sensing/detoxifying enzymes), including VDAC1, PARK7 and SOD2 (Figure 7D). Consistent with induction of oxidative stress, depletion of S6K2 or S6K1 (to a lesser extent) led to increased ROS levels (Figure 7E). Treatment of transduced cells with the ROS scavenger NAC attenuated cell death in S6K2-depleted cells, indicating that ROS contributes to the cell death (Figure 7F).

[00089] Depletion of S6K2 induces fatty acid synthesis and lipid peroxidation, facilitating cell death [00090] We noted that depletion of S6K2 upregulated proteins involved in fatty acid synthesis, whereas depletion of S6K1 had opposite effects (Figure 8A). Accordingly, S6K2 depletion led to increased expression of a panel of genes involved in lipid synthesis (Figures 8B, 8C and 13 A); in particular, SCD1 (the rate-limiting enzyme for synthesis of unsaturated fatty acids) was upregulated in all cell lines tested (Figure 8C). Consistent with upregulation of SCD1, depletion of S6K2 was coupled to increased levels of unsaturated fatty acids (Figure 8D). As unsaturated fatty acids, particularly polyunsaturated fatty acids, are highly susceptible to ROS and can facilitate lipid peroxidation and cell death (Ayala, 2014 #349}, we examined if lipid peroxidation could contribute to S6K2 depletion-induced cell death. Indeed, depletion of S6K2, but not S6K1 elicited lipid peroxidation and subsequent generation of 4-HNE adducts (Figures 8E, 8F, 13B and 13C). Consistent with the reactive property of 4-HNE, S6K2-depleted cells exhibited oxidative damage of nucleic acids (8-OHdG) and DNA double-strand breaks (p- yH2AX and 53BP1) (Figures 13D-G). Furthermore, treatment of NRAS-mutant melanoma cells transduced with S6K2 shRNA with trolox (Barclay, 1995 #590}, a lipid ROS scavenger attenuated cell death, indicating that lipid peroxidation indeed contributes to cell death (Figure 8G). To assess if increased levels of unsaturated fatty acids prompt lipid peroxidation, we treated S6K2-depleted cells with the SCD1 inhibitor A939572. SCD1 blockade was coupled to decreased oxidation of the lipid peroxidation probe Bodipy-Cl 1, indicating that the lipid desaturase SCD1 and increased unsaturated fatty acids contribute to the detrimental

lipidperoxidation in S6K2-depleted cells (Figure 8H). Taken together, our results indicate that depletion of S6K2 triggers increased levels of ROS and unsaturated fatty acids which together prompt oxidative cell death. [00091] Depletion of S6K2 triggers multiple forms of cell death

[00092] Lipid peroxidation has been linked to apoptosis, necroptosis and ferroptosis (Ayala, 2014 #349}{Canli, 2016 #75}{Dixon, 2012 #88}. We observed increased levels of prostaglandin-endoperoxide synthase/cyclooxygenase 2 (PTGS2), a marker of ferroptosis (Yang, 2014 #180}, in S6K2-depleted cells, suggesting that these cells may undergo ferroptosis or a related form of cell death (Figure 13H).

[00093] Consistent with this, proteomic analysis of cells expressing S6K1 or S62 shRNA, revealed that S6K2 depletion perturbed pathways related to ferroptosis. Nevertheless, caspase 3 cleavage was also detected in S6K2-depleted cells (Figure 7B). We therefore asked if depletion of S6K2 could be triggering multiple forms of cell death. To this end, we treated S6K2-depleted cells with inhibitors of apoptosis (zVAD), ferroptosis (ferrostatin-l/Ferl) or necroptosis

(necrostatin-l/Necl) (Figure 13J). While necrostatin-l failed to prevent cell death in S6K2- depleted cells, zVAD and ferrostatin-l partially restored viability of S6K2-depleted cells (Figure 13K). Furthermore, combining ferrostatin-l and zVAD better prevented cell death, suggesting that S6K2 depletion and lipid peroxidation triggers both apoptosis and ferroptosis or a related form of cell death (Figure 13K).

[00094] Suppression of S6K1 offsets the effects of S6K2 depletion

[00095] Since depletion of S6K2 triggered lipid peroxidation and cell death in NR.AS- mutant melanoma cells resistant to MAPK inhibition, we next sought to examine if this mechanism could be exploited to kill melanoma cells. Currently, there are no selective S6K2 inhibitors available, and indirect suppression of S6K using PI3K/mTORi mainly causes cytostatic effects (Figures 11H and 9A). Accordingly, treatment of NRAS-mutant melanoma cells with PI3K/mTOR inhibitors did not induce ROS, fatty acid synthesis, or lipid peroxidation (Figures 14A-B, 9B). Suppressing S6K with a S6K1 inhibitor PF4708674 or a pan-S6K inhibitor LY2584702 also failed to induce synthesis of fatty acids, lipid peroxidation and cell death (Figures 9C-D and 14C-D). Consistent with depletion of S6K1, inhibiting S6K1 with PF4708674 induced moderate increase of ROS levels; however, ROS alone is not sufficient to induce lipid peroxidation (Figures 7E, 14C and 9D). Since S6K1 depletion suppresses fatty acid synthesis in contrast to S6K2 inhibition (Figure 8A), we hypothesized that a putative mechanism underlying these results could be that PI3K/mTOR or pan-S6K inhibitors concurrently suppress S6K1 and S6K2, and blunt induction of fatty acid synthesis by S6K2 inhibition. To test this possibility, we silenced S6K2 alone or in combination with S6K1. Depletion of S6K1 attenuated cell death and restored long-term survival induced by depletion of S6K2 (Figures 9E and 14E). Consistently, concurrent depletion of S6K1 and S6K2 attenuated levels of SCD1 and the transcription factor SREBP1, as well as lipid peroxidation in S6K2 depleted cells (Figures 9F, 14F and 14G). S6K1 depletion also suppressed the induction of PTGS2 triggered by S6K2 depletion (Figure 14G). Additionally, treatment of cells expressing a constitutively active S6K1 with the pan-S6K inhibitor LY2584702 (crafting a situation whereby S6K2 is“off’ and S6K1 is“on”) promoted cell death, further suggesting that selectively inactivation of S6K2 can trigger toxic effects (Figures 4H-I). Collectively, these results suggest that concurrent inhibition of S6K1 and S6K2 would not be optimal for inducing lipid peroxidation and cell death in NRAS-mutant melanoma, and that rather selective suppression of S6K2 is required to induce oxidative cell death.

[00096] Induction of lipid peroxidation triggers tumor cell death in NRAS-mutant melanoma [00097] We next examined whether mimicking the effects of S6K2 depletion would trigger cell death in NRAS-mutant melanoma. To this end, we used ROS inducers (BSO or sorafenib) in combination with PUFA (Docosahexaenoic acid/DHA) to induce lipid peroxidation and trigger cell death. Indeed, induction of ROS facilitated DHA-induced lipid peroxidation (Figure 10A) and the combination synergistically triggered cell death (Figures 10B-D and 15 A), phenocopying S6K2 depletion. We further leveraged compounds that induce lipid peroxidation and oxidative cell death such as erastin and RSL3 (Dolma, 2003 #353} (Yang, 2008

#121 } (Dixon, 2012 #88} to phenocopy S6K2 depletion. Both erastin and RSL3 triggered lipid peroxidation, leading to cell death (Figures 10E and F). Notably, ferrostatin-l and liproxstatin-l, which prevent accumulation of lipid ROS (Skouta, 2014 #87} (Friedmann Angeli, 2014 #164}, reversed cell death (Figures 10F and 15C), further supporting the notion that induction of lipid ROS can readily kill NRAS- mutant melanoma cells. To determine if induction of lipid peroxidation could be a potential strategy to suppress growth of NRAS-mutant melanoma in vivo, mice bearing NRAS mutant subcutaneous tumors were treated with DHA (300 mg/kg, oral gavage, 5 days/week), BSO (300 mg/kg, i.p., everyday) or combination of DHA and BSO for 2 weeks. Combining DHA and BSO delayed tumor growth.

[00098] We also examined the effect of lipid peroxidation inducers on tumor growth.

Mice bearing established subcutaneous NRAS mutant tumors were treated with erastin (30 mg/kg, twice daily, every other day). Erastin suppressed tumor growth (Figures 10F and 15F) coupled to lipid peroxidation and oxidative damage (Figures 10H). Altogether, our data indicate that depletion of S6K2 induces upregulation of ROS levels and unsaturated fatty acids. These two factors together trigger lipid peroxidation and oxidative cell death (Figure 10H).

Importantly, we demonstrate that induction of lipid peroxidation can be exploited as an anti- tumor strategy in NRAS-mutant melanoma, particularly in melanoma resistant to MAPKi, for which there are currently no effective therapies available.

[00099] Therefore, in preferred embodiments, it is sufficient to initiate ferroptosis through induction of lipid ROS, and that such induction is sufficient to kill NRAS-mutant melanoma cells. In a preferred embodiment, a method of treating NRAS-mutant melanoma cells is directed towards providing a preparation of a fatty acid and BSO sufficient to induce lipid ROS. In a preferred embodiment, the fatty acid is a PUFA, for example DHA.

[000100] In preferred embodiments, a method for treating a patient having NRAS- mutant melanoma comprises administering to said patient an effective amount of a PUFA and BSO. Said applications can provide for an oral application of the PUFA, with the BSO provided in a concomitant administration.

[000101] Accordingly, oncogenes often upregulate nutrient sensing pathways, leading to altered metabolism in tumor cells. RAS-driven tumors typify cancers undergoing marked metabolic reprogramming. In melanoma, RAS signaling is often deregulated; while activating mutations in NRAS are present in >25% of tumors, RAS-driven tumors account for almost 50% of cases. However, there are no effective therapies for this type of tumors. Although oncogenic NRAS activates the MAPK pathway, inhibition of this pathway alone has limited anti-tumor efficacy and most NRAS-mutant tumors are resistant to MAPK inhibitor (MAPKi) monotherapy.

[000102] As discussed above, we discovered that NRAS-mutant melanomas resistant to MAPKi are highly dependent on the ribosomal protein S6 kinase 2 (S6K2). We demonstrate that loss of S6K2, a signaling effector of the mTORCl nutrient-sensing pathway, triggers cell death selectively in NRAS-mutant melanoma cells resistant to MAPKi. Acute depletion of S6K2 enhanced ROS production, lipid synthesis and accumulation of intracellular unsaturated fatty acids. ROS susceptible (poly)-unsaturated fatty acids sensitized cells to ROS, resulting in lipid peroxidation and oxidative cell death. We further determined that S6K2 depletion was coupled to increased expression of markers of apoptosis and ferroptosis, suggesting that S6K2 blockade could trigger multiple forms of cell death, including apoptotic and ferroptotic-like cell death. Notably, co-inhibition of S6K1 (an isoform that is co-regulated by mTORCl) together with S6K2, diminished the effects of S6K2 blockade, suggesting that selective inhibition of S6K2 is required to induce cell death. While silencing of S6K2 triggered lipid peroxidation and oxidative cell death, and S6K1 silencing had negligible effects, concomitant depletion of S6K1 and S6K2 attenuated lipid peroxidation and cell death. Mimicking the effects of S6K2 loss by combining a ROS inducing agent with unsaturated fatty acids or treating tumor cells with lipid peroxidation inducers restrained NRAS-mutant melanoma growth in vitro and in vivo. Taken together, our studies have identified a critical vulnerability of NRAS-mutant melanoma by uncoupling S6K1 and S6K2 to trigger metabolic dysfunction and tumor cell death. This strategy is remarkably distinct from using PI3K/mTOR inhibitors, which concurrently suppress S6K1 and S6K2 and are often cytostatic. Such a therapeutic is ineffective and our studies confirm the same. Inhibition of both S6K1 and S6K2 counteracts the benefit of the S6K2 reduction. Accordingly, it is selective inhibition of S6K2 that provides the efficacy towards MEPKi resistant melanoma cells. [00090] In a particular embodiment, a method of treating cancer comprises administering to a patient having said MAPKi-resistant tumors an effective amount of a ferroptosis/lipid peroxidation inducing agent. In certain embodiments, said cancer is a solid tumor which is a NRAS mutant tumor. NRAS mutant tumors would be contraindicated for treatment with a BRAF inhibitor. Indeed, in certain embodiment it is necessary to administer only a first agent, wherein said agent is a ferroptosis inducing agent, e.g. erastin, (1S, 3R) RSL3, ML162, ML210 or FIN56. [000103] In certain embodiments, it is advantageous to co-administer a MEK inhibitor with a ferroptosis inducer. Together, this therapeutic approach can selectively inhibit S6K2, to increase lipid peroxidation and induce ferroptosis with the cell.

[000104] NRAS Mutant melanoma cells involve certain pathways for induction of ferroptosis. Accordingly, our studies indicate that cancers with MAPK inhibitors will also be susceptible to treatment with the ferroptosis inducing agents. In a particular example, RAS mutant cancers, such as pancreas, lung, colon and thyroid, as non-limiting examples, would be susceptible to treatment with a ferroptosis/lipid peroxidation inducing agent as provided for in the examples, data, and methods as described herein.

FURTHER EMBODIMENTS

[000105] In another embodiment of the invention, uncoupling S6K1 and S6K2 perturbs lipid homeostasis triggering apoptosis and ferroptosis in MAPKi-resistant melanoma, and selective disruption of S6K2 perturbs lipid homeostasis triggering apoptosis and ferroptosis in MAPKi-resistant melanoma were analyzed.

[000106] Melanoma is often driven by NRAS mutations. Oncogenic NRAS activates MAPK signaling, yet inhibition of this pathway has limited efficacy in NRAS-mutant tumors. Hence, it is critical to understand the mechanisms underlying resistance to MAPKi and to develop strategies for tumors refractory to MAPK inhibitors. We discovered that depletion of S6K2, a mTORCl effector, triggers accumulation of unsaturated fatty acids, lipid peroxidation and cell death selectively in NRAS-mutant melanoma cells that are resistant to MAPK inhibition. S6K2 depletion induced upregulation of S6K1 and PPARa, triggering apoptotic and ferroptotic- like cell death. Conversely, depletion of S6K1 diminished lipid synthesis and cell proliferation, without eliciting cell death. Collectively, our data indicate that perturbing the balance between S6K1 and S6K2 activity and metabolic homeostasis triggers cell death of NRAS-mutant melanoma. Furthermore, our study indicates that selective inhibition of S6K2 prompts multiple forms of cell death and establishes S6K2 as a novel target for MAPKi-resistant NRAS-mutant melanoma.

[000107] RAS-mutant tumors are extremely aggressive and highly refractory to currently available therapies. Direct RAS targeting has been extremely challenging {Ryan, 2018 #780}, hence alternative approaches to kill RAS-mutant tumors are sorely needed. Potential approaches include identifying and targeting tumor dependencies or RAS effectors essential for RAS-mutant tumors. In melanoma, mutations in NRAS account for almost 30% of all tumors {Hayward, 2017 #796}, leading to activation of the RAF/MEK/ERK MAPK cascade. However, inhibitors of the MAPK pathway elicit only limited anti-tumor activity as single agents in NRAS-mutant melanoma cell lines and patients {Dummer, 2017 #788}{Atefi, 2015 #342}{Ascierto, 2013 #341 }. Suppression of MAPK often leads to feedback or compensatory activation of PI3K/AKT, another RAS effector pathway governing cell growth and survival in melanoma {Vu, 2016 #386} {Fattore, 2013 #4l7}{Gopal, 2010 #4l2}{Atefi, 2011 #416}. Although concomitant inhibition of the MAPK and PI3K pathways is effective in pre-clinical models, it is poorly tolerated in patients {Juric, 2014 #538}{Tolcher, 2015 #539}{Posch, 2016 #340}. Dual inhibition of MEK and PI3K/mTORCl/2 revealed that metabolic pathways were the most affected in NRAS-mutant melanoma cells {Posch, 2013 #406}. As Ras-mediated transformation often induces metabolic reprogramming to support macromolecular biosynthesis, cell proliferation, and tumor growth and progression {Pavlova, 2016 #531 } {Kimmelman, 2015 #534}, we reasoned that a metabolic dependency could constitute a vulnerability in RAS-mutant tumors.

[000108] A signaling pathway that plays a key role in regulating metabolism, including lipid biosynthesis, is the mechanistic target of rapamycin protein complex 1, mTORCl. (Duvel, 2010 #332}. mTORCl is a signaling hub that integrates upstream signals including the MAPK and PI3K pathways (Corcoran, 2013 #403 } (Magnuson, 2012 #295}, linking extracellular stimuli and signal transducers with cell growth and metabolism. mTORCl induces activation of the sterol responsive element binding protein SREBP, which regulates the expression of enzymes involved in fatty acid, triacylglycerol and phospholipid synthesis (Laplante, 2010 #791 } (Li,

2010 #790}{Eberle, 2004 #511 } (Dobrosotskaya, 2002 #786}. mTORCl can activate SREBP thru its downstream effector S6K1 (Duvel, 2010 #332}(Owen, 2012 #440} (Wang, 2011 #792}. The serine threonine kinase S6K1 and its homolog S6K2 belong to the AGC kinases super family. These kinases are highly homologous that have overlapping, but also distinct biological functions (Pardo, 2013 #294} (Magnuson, 2012 #295}(Karlsson, 2015 #179} (Pavan, 2016 #393} (PMID: 21444676). While both S6K1 and S6K2 have been implicated in regulating metabolism, particularly lipid homeostasis (Duvel, 2010 #332}(Kim, 20l2#363} (20493810?), their divergent activities and their contribution to melanomagenesis and response to therapy remains unclear.

[000109] Disruption of metabolic homeostasis can trigger lipid peroxidation, whereby reactive oxygen species (ROS) cause oxidative lipid degradation and cellular damage, and can lead to cell death. ETnsaturated fatty acids, particularly polyunsaturated fatty acids (PETFA), are highly susceptible to such ROS attack, resulting in self-propagation of radical chain reactions (Ayala, 2014 #349}. When production of lipid peroxides overwhelms the cellular detoxification capacity, excess lipid peroxides and their reactive end products such as 4-hydroxynonenal (4- HNE) cause widespread damage {Aitken, 2012 #469} (Ayala, 2014 #349}. Such excess lipid peroxidation has been linked to different types of cell death including necrosis, apoptosis, necroptosis (Ayala, 2014 #349}(Seiler, 2008 #85}(Aitken, 2012 #469}(Canli, 2016 #75} and ferroptosis (Dixon, 2012 #88}(Yang, 2014 #180}. Ferroptosis is a form of regulated cell death involving metabolic dysfunction and lipid peroxidation (Dixon, 2012 #88}. Recent studies on ferroptosis have identified key factors modulating oxidative cell death (Magtanong, 2016 #108}. For instance, enzymes that enhance the synthesis or membrane enrichment of PUFA promote cell death providing substrates for lipid peroxidation, and enzymes that detoxify 4-HNE suppress cell death by preventing further damage by the reactive molecules (Rees, 2016

#463}(Viswanathan, 2017 #40l }(Doll, 2017 #496}(Dixon, 2015 #147}. At present, the upstream events that trigger the imbalance of lipid redox homeostasis and lipid peroxidation remain elusive, but may yield clues about novel strategies to therapeutically exploit these processes.

[000110] Here, we provide evidence that acute depletion of S6K2 enhances ROS production, expression of enzymes involved in lipid synthesis, and accumulation of unsaturated fatty acids. This triggers lipid peroxidation and cell death selectively in NRAS-mutant melanoma resistant to MAPK inhibition. Mechanistically, we show that S6K2 depletion triggers cell death by inducing PPARa and S6K1 -dependent SREBP1 activation. Together, this work reveals that selective inhibition of S6K2 creates a molecular vulnerability in NRAS-mutant melanoma by triggering a lipid metabolic imbalance, leading to both apoptosis and ferroptosis. Further, we provide proof-of-concept that this vulnerability can be exploited to induce tumor cell death and thus inform the design of future strategies to combat NRAS-mutant melanoma. RESULTS

[000111] S6K2 is a vulnerability in NRAS-mutant melanoma resistant to MAPK inhibitors

[000112] Because MAPK inhibition elicits variable effects in NRAS-mutant melanoma, we first sought to identify factors responsible for differential sensitivity to MAPK inhibitors (MAPKi) and potential vulnerabilities in tumor cells resistant to MAPKi. We compared the effect of MAPK inhibition in NRAS-mutant melanoma cells (Table 1, below) treated with MEK or ERK inhibitors. MAPKi induced variable effects on viability, proliferation and cell death, despite effective and persistent inhibition of the MAPK pathway (Figures 17A-E). We classified cells as MAPKi-resistant or MAPKi-sensitive based on their response to MAPKi, defined as area under the curve/ AETC (trametinib: AETC>2.4, SCH779284: AETC>2.3 for resistant cells), as well as suppression of cell proliferation and induction of cell death (Figures 17A-C). MAPKi elicited modest suppression of proliferation and marginal cell death in the MAPKi-resistant cells (Figures 17B-C). This heterogeneous response to MAPKi was further validated in a larger panel of NRAS-mutant melanoma cells treated with two additional, clinically relevant MAPKi (MEKi: MEK-162 and ERKi: BVD-523; Figure 17F). These data confirm that pharmacological agents which target different MAPK effectors elicit heterogenous effects when used as single agents.

[000113] Sensitivity to MAPK inhibition was associated with higher basal levels of pERK (Pearson’s correlation |R|>0.5, P<0.05) (Figures 17G and 17H), suggesting that MAPKi-resistant cells have relatively low input to ERK. Biochemical analysis of cells treated with the MEK inhibitor trametinib or ERK inhibitor SCH772984 indicated that MAPKi suppressed the MAPK pathway similarly in both sensitive and resistant cells (Figures 16A and 17D). Likewise, the PI3K/AKT pathway was comparably activated in all MAPKi -treated cell lines (Figures 16A and 17D). In contrast, MAPKi treatment blocked phosphorylation of S6 kinase (S6K) and its substrate S6 in MAPKi-sensitive, but not MAPKi-resistant melanoma cells (Figures 16A and 17D). These data raise the possibility that S6K activity is linked to response to MAPKi in NRAS-mutant melanoma cells. As S6K is a bona fide downstream effector of the PBK/mTOR pathway, inhibition of this pathway suppressed phosphorylation of S6 (pS6) (Figures 171 and 17J), indicating that the PBK/mTOR pathway regulates S6K in both MAPKi-resistant and - sensitive NRAS-mutant melanoma cells. To further examine the impact of MAPK or

PBK/mTOR signaling on S6K (assessed by pS6S240/244, a site exclusively phosphorylated by S6K) (Roux, 2007 #42l;Corcoran, 2013 #403;Meyuhas, 2008 #473}, we ectopically expressed constitutively active S6K (S6K1E389 or S6K2E388) constructs in MAPKi-sensitive or -resistant cells (Figures 16B and 17K). Treatment of cells with MAPKi (SCH984) or PBK/mTORi (GSK458) decreased phosphorylation of S6 in MAPKi-sensitive cells. Moreover, expression of constitutively active S6K (S6K1E389 or S6K2E388) restored pS6 levels, indicating that S6K mediates both MAPK and PBK/mTOR signaling in MAPKi-sensitive cells (Figures 16B and 17K). In contrast, MAPKi had no effect on phosphorylation of S6 in MAPKi-resistant cells. In these cells, PBK/mTORi effectively suppressed pS6 (Figures 16B and 17K). Expression of S6K1E389 or S6K2E388 restored pS6 levels in cells treated with PBK/mTORi, indicating that S6K activity is mainly controlled by the PBK/mTOR pathway in MAPKi-resistant cells (Figures 16B and 17K). Together, these data indicate that S6K1/2 is controlled by the PBK/mTOR and MAPK pathways in MAPKi-sensitive cells, but is primarily regulated by the PBK/mTOR pathway in MAPKi-resistant melanoma cells (Figure 16C).

[000114] To dissect the role of S6K in NRAS-mutant melanoma cells, we silenced S6K1 or S6K2. Depletion of either S6K1 or S6K2 impaired cell proliferation (Figures 17L and 17M). Interestingly, only depletion of S6K2, but not S6K1, induced cell death (Figures 16D, 16E and 17N). This effect appears to be restricted to NRAS-mutant melanoma cells resistant to MAPKi, as depletion of S6K2 did not trigger cell death in MAPKi-sensitive NRAS-mutant melanoma cells (Figures 16F and 16G, and 170) or non-transformed human fibroblasts (Figure 17P). These data point to S6K2 as a novel vulnerability in NRAS-mutant melanoma resistant to MAPKi.

[000115] Depletion of S6K2 triggers lipid peroxidation, facilitating cell death

[000116] To investigate the mechanism(s) whereby S6K2 depletion triggers cell death in MAPKi-resistant cells, we surveyed the proteome to identify proteins that were differentially affected by depletion of S6K2 versus S6K1 in MAPKi-resistant cells. Depletion of S6K1 or S6K2 led to increased levels of proteins involved in oxidative stress response, including ROS sensing/detoxifying enzymes such as SOD2, PARK7 and VDAC1 (Figure 18 A), and increased ROS levels (Figure 18B). Depletion of S6K2, but not S6K1, was coupled to enhanced expression of proteins involved in lipid synthesis, fatty acid uptake and activation, and phospholipid remodeling (Figures 18 A, 18C, and 18 A). Consistently, S6K2 depletion upregulated the transcription of several lipogenic enzymes (Figures 19A and 19B). Interestingly, whereas depletion of S6K2 upregulated key regulators of fatty acid synthesis, depletion of S6K1 appeared to have the opposite effects (Figure 18 A). Lipidomic analysis indicated that S6K2 depletion was associated with accumulation of triacyl glycerols (Figure 19C). Notably, tri acyl glycerols as well as phosphatidylcholine, a major component of biological membranes, phosphatidyl ethanolamine and phosphatidyl glycerol showed increased degrees of unsaturation (Figures 19C and 19D), suggesting global fatty acid desaturation in S6K2-depleted cells. Indeed, S6K2 depletion increased relative levels of unsaturated fatty acids (Figure 18D). Together, these results suggest that S6K2 blockade is coupled to increased lipid metabolism, including increased fatty acid synthesis, desaturation and accumulation of triglycerides.

[000117] The effects of S6K2 depletion on redox balance and increased lipid metabolism raised the possibility that S6K2 depletion could be triggering cell death by inducing lipid peroxidation (Dixon, 2012 #88 J { Vi s wan at h an, 2017 #401 }. In support of this, S6K2 depletion enhanced lipid peroxidation, as indicated by the lipid ROS sensor Bodipy-Cl 1 and generation of 4-HNE adducts (Figures 18E, 19E-G). S6K2 depletion also led to nucleic acid oxidative damage (8-OHdG) and DNA double-strand breaks (r-gH2AC and 53BP1) (Figures 19H-K). We next assessed if enhanced lipid metabolism modulated the sensitivity of S6K2-depleted cells to oxidative stress. As SCD1 (the rate-limiting enzyme for fatty acid desaturation) was upregulated upon S6K2 depletion (Figures 18A and 19B), we examined the role of SCD1 in oxidative cell death. S6K2-depleted cells were treated with the SCD1 inhibitor A939572. SCD1 blockade led to decreased oxidation of the lipid peroxidation probe Bodipy-Cl 1, indicating that SCD1 contributes to lipid peroxidation in S6K2-depleted cells (Figure 18F). Unsaturated fatty acids are more susceptible to ROS {Ayala, 2014 #349}. Therefore, SCD1 likely facilitates lipid peroxidation and cell death by increasing the desaturation of fatty acids (Figures 18D and 19C). Treatment of NRAS-mutant melanoma cells transduced with S6K2 shRNA with the ROS scavenger NAC or the lipid ROS scavenger trolox {Barclay, 1995 #590} attenuated cell death (Figures 18G and 18H), indicating that ROS, particularly lipid ROS, contributes to cell death triggered by S6K2 depletion. Together, these results indicate that depletion of S6K2 leads to increased levels of ROS and accumulation of unsaturated fatty acids, which jointly trigger oxidative cell death.

[000118] Depletion of S6K2 triggers multiple forms of cell death

[000119] Availability of PUFA and accumulation of lipid peroxides is associated with induction of ferroptosis, an iron-dependent form of regulated cell death {Doll, 2017 #496} {Rees, 2016 #463}{Viswanathan, 2017 #40l }{Dixon, 2012 #88}{Yang, 2014 #180}. Proteomics analysis indicated that S6K2 depletion enhanced pathways associated with ferroptosis and the ferroptosis marker PTGS2 {Viswanathan, 2017 #40l;Shimada, 2016 #l07;Magtanong, 2016 #l08;Friedmann Angeli, 2014 #l64;Kagan, 2017 #580;Doll, 2017 #496;Yang, 2014 #180} (Figure 18C). We therefore assessed if S6K2 depletion could be triggering ferroptosis. Treatment of S6K2-depleted cells with the iron chelator deferoxamine (DFO) attenuated lipid peroxidation and cell death (Figures 181 and 19L), supporting the notion that S6K2 depletion was associated with iron-dependent accumulation of lipid ROS and oxidative cell death. Moreover, treatment of NRAS-mutant melanoma cells with the ferroptosis inhibitors ferrostatin-l (Fer-l) {Skouta, 2014 #87} or liproxstatin-l (Lip-l) {Friedmann Angeli, 2014 #!64}{Kagan, 2017 #580} attenuated cell death triggered by S6K2 depletion (Figures 18J and 19M), further suggesting that ferroptosis occurs in these conditions. Since caspase-3 cleavage was also detected in S6K2-depleted cells (Figure 16E), we asked if depletion of S6K2 could be triggering multiple forms of cell death. To address this possibility, we treated S6K2-depleted cells with inhibitors of apoptosis (zVAD), ferroptosis (Fer-l) or necroptosis (necrostatin-l; Nec-l). Both zVAD and Fer-l partially restored viability of S6K2-depleted cells, whereas Nec-l did not prevent cell death (Figure 18J).

Combining Fer-l and zVAD further prevented cell death, suggesting that S6K2 depletion and lipid peroxidation trigger both apoptosis and ferroptosis, or a related form of cell death (Figure 18J).

[000120] Depletion of S6K2 leads to S6K1 upregulation prompting lipid synthesis and cell death

[000121] We next sought to further dissect the mechanism whereby S6K2 modulates lipid metabolism and lipid peroxidation. Toward this goal, we assessed the transcriptional changes triggered by S6K2 depletion. In cells expressing S6K2 shRNA, target genes of the master regulator of lipid homeostasis, SREBP, were overrepresented, suggesting that suppression of S6K2 activated SREBP (Figure 20A and Table 2, below). We therefore asked if SREBP could be contributing to the upregulation of lipid synthesis and subsequent oxidative events (Figures 18 A, 18E and 19B). Indeed, S6K2 silencing prompted activation of SREBP as indicated by upregulation of SREBP target genes and an increase of nuclear SREBP1 (nSREBPl)(Figures 20B, 19B and 21 A). Additionally, SREBP1 silencing partially blocked cell death in S6K2- depleted cells, supporting a role of SREBP 1 in triggering oxidative cell death (Figures 20C and 21B). Since S6K2 knockdown often leads to upregulation of S6K1 (Figures 20D and 21C) {Pardo, 2013 #294}{Karlsson, 2015 #179}, and since S6K1 can stimulate SREBP1 processing and activation {Duvel, 2010 #332;Owen, 2012 #440}, we assessed the contribution of S6K1 to oxidative cell death in S6K2-depleted cells. We found that S6K1 depletion diminished the expression of SREBP1 targets involved in lipid synthesis (Figure 20B). Concurrent suppression of S6K1 and S6K2 blunted the effects of S6K2 depletion in inducing SREBP1, SCD1 and the ferroptosis marker PTGS2 (Figures 20E, 21D and 21E). Furthermore, concomitant depletion of S6K1 and S6K2 attenuated lipid peroxidation, cell viability and cell death (Figures 20F, 20G and 21F). Together, these data indicate that activation of S6K1 contributes to the effects triggered by S6K2 depletion. These results led us to postulate that selective inhibition of S6K2 in the context of active S6K1 would be required to induce lipid peroxidation in NRAS-mutant melanoma cells resistant to MAPKi. Consistent with this premise, concomitant inhibition of S6K1 and S6K2 with a pan-S6K inhibitor (LY2584702; LY4702) {Tolcher, 2014 #77l }{Hollebecque, 2014 #772} or inhibition of S6K1 with an isoform selective inhibitor (PF4708671) {Pearce, 2010 #770} suppressed (rather than induced) the expression of enzymes involved in lipid synthesis (Figure 21G). Concurrent inhibition of S6K1 and S6K2 with LY4702 or a PI3K/mTOR inhibitor (GSK458) did not trigger lipid synthesis, lipid peroxidation, or cell death (Figures 21G-L). In contrast, mimicking an“S6K1 on/S6K2 off’ scenario by ectopically expressing constitutively active S6K1S389 and treating cells with the pan-S6K inhibitor LY4702, led to enhanced cell death compared to LY4702 alone (Figures 20H and 201). Collectively, these results indicate that selective inhibition of S6K2, without significantly perturbing S6K1, is required to induce lipid peroxidation and subsequent death of MAPKi-resistant NRAS-mutant melanoma cells. Table 2. List of overrepresented gerses arid z-scores predicting activated SREBFs (SREBPs) in S6K2-dep!eted ceiis.

RNA sequencing data were analyzed by iPA.

Type: regulator type

Regulator: targets of this regulator are overrepresented in the list

P: nominal p-value of the overrepreseniation

Z: z-score of the prediction - positive for activated, negative for inhibited function

i: number of regulator's target genes in the list

pos: number of unregulated molecules

neg: numbets of downreguiated molecules

Molecules: genes (or complexes) from the gene list known to be regulated

[000122] NRAS-mutant melanoma resistant to MAPKi is sensitive to induction of lipid peroxidation

[000123] Since depletion of S6K2 triggered lipid peroxidation and cell death in MAPKi- resistant NRAS-mutant melanoma cells, we next sought to examine if this mechanism could be therapeutically exploited. As there are no selective S6K2 inhibitors, we evaluated alternative pharmacological approaches that could phenocopy S6K2 depletion. To this end, we leveraged compounds that induce lipid peroxidation and oxidative cell death to mimic the effects of S6K2 depletion. Both erastin and RSL3 efficiently triggered lipid peroxidation (Figure 22A) and cell death (Figure 22B), and cell death was reversed by the ferroptosis/lipid ROS inhibitor Fer-l (Skouta, 2014 #87} (Figure 22B). Moreover, erastin suppressed the growth of NRAS-mutant subcutaneous tumors (Figure 22C). This effect was coupled to lipid peroxidation (4-HNE) and oxidative damage (8-OHdG) (Figure 22D), indicating that pharmacological induction of lipid peroxidation could be a viable strategy to restrain NRAS-mutant melanoma. [000124] Suppression of S6K2 relieves negative regulation on PPARa, facilitating lipid peroxidation

[000125] We noted that several PPAR targets, including modulators of ferroptosis (e.g. ACSL4 and LPCAT3), were upregulated upon S6K2 depletion (Figures 18C, 23 A, and 23B, and Table 3, below) (Kersten, 2014 #50l;Rakhshandehroo, 2010 #505} . We therefore asked if PPAR was contributing to cell death by inducing lipid peroxidation. S6K2 depletion led to upregulation of PPARa, and, to a much lesser extent, PPAR mRNA levels (Figures 23C and 24A), suggesting that PPARa might be a key effector. Indeed, depletion of PPARa substantially decreased ROS levels, lipid peroxidation and cell death in S6K2-depleted cells (Figures 23D and 24B). Conversely, treatment of NRAS-mutant melanoma cells with PPARa agonists (fenofibrate or GW7647) induced ROS, lipid peroxidation and cell death (Figures 24C and 24D). Together, these data suggest that PPARa contributes to lipid peroxidation and cell death triggered by S6K2 depletion.

Table 3. List of overrepresented genes and z-scores predicting increased/inhibited function

of PPARs in S6K1 or S6K2 -depleted cells.

RNA sequencing data were analyzed by IPA.

Type: regulator type

Regulator: targets of this regulator are overrepresented in the list

P: nominal p-value of the overrepresentation

Z: z-score of the prediction - positive for activated, negative for inhibited function.

N: number of regulator’s target genes in the list

pos: number of upregulated molecules

neg: numbets of downregulated molecules

[000126] Since mTORCl suppresses the transcriptional activity of PPARa by regulating the nuclear receptor corepressor NCoRl (Sengupta, 2013 #368} and S6K2 has been implicated in this process in the liver (Kim, 2012 #363}, we wondered if S6K2 was negatively regulating PPARa in a similar fashion in NRAS-mutant melanoma cells. Immunoprecipitation and proximity ligation assays (PLA) indicated that S6K2 interacts with PPARa and its co-repressor NCoRl (Figures 23E, 23F and 24E). Because S6K2 shuttles between the cytosol and nucleus (Magnuson, 2012 #295}, we then examined if S6K2 affected the localization of NCoRl and PPARa. We found that S6K2 was predominantly localized to the nucleus and that depletion of S6K2 led to decreased nuclear and chromatin-bound NCoRl (Figure 23G). Conversely, S6K2 depletion increased PPARa chromatin binding (Figure 23G). S6K2 depletion also diminished the interaction of NCoRl with PPARa (Figures 23H and 24F). Together, these data suggest that S6K2 depletion relieves the negative regulation of NCoRl on PPARa, leading to activation of the PPARa axis and enhanced lipid peroxidation (Figure 231).

[000127] Based on our data that PPARa facilitates lipid peroxidation, we next asked if PPARa agonists could induce anti-tumor effects in NRAS-mutant melanoma. We noted that PPARa (but not PPARy) levels were higher in MAPKi-resistant than in MAPKi-sensitive NRAS-mutant melanoma cells (Figures 25A, 26A and 26B). These MAPKi-resistant/PPARa- high cells were particularly sensitive to the PPARa agonist fenofibrate (FNB) compared to MAPKi-sensitive cells (Figures 23B and 23C). We also noted that FNB induced lipid peroxidation and cell death especially at doses that upregulated FASN and SCD1 (Figures 24C and 26C). However, FNB inhibited the expression of SREBP1 (Figure 26C), a facilitator of oxidative cell death in our system. Therefore, we hypothesized that enhancing fatty acid synthesis/desaturation or potentiating SREBP activity could sensitize tumor cells to FNB treatment. Indeed, the SREBP activator U18666A (Schmitt, 2017 #787} cooperated with FNB in triggering lipid peroxidation and cell death (Figure 26D). Likewise, addition of PUFAs (docosahexaenoic acid/DHA or arachidonic acid) potentiated the effect of FNB in inducing lipid peroxidation and cell death (Figures 25D, 26E and 26F). Additionally, the lipid peroxidation inhibitor liproxstatin-l protected the cells, attenuating cell death (Figure 26G). Furthermore, blockade of the anti-oxidant defense system by treatment with erastin also potentiated the effects of FBN (Figures 25E and 26H). These results indicate that MAPKi-resistant melanoma cells are sensitive to pharmacological activation of PPARa in combination with enhanced fatty acid desaturation or lipid peroxidation. [000128] To further explore the relevance of our findings, we analyzed the skin cutaneous TCGA database, and found that PPARa levels in NRAS-mutant tumors (but not in BRAF- mutant or BRAF/NRAS wild-type) are inversely associated with a set of six genes included in a MAPK transcriptional signature (MPAS) (Wagle, 2018 #747} (Figure 25F). MPAS is derived from the transcript levels of 10 MAPK target genes which correlate with MAPK inhibitor sensitivity in multiple cell lines and tumor samples. This association between PPARa levels and MAPK target gene expression raised the possibility that PPARa agonists could potentially be used to target MAPKi -resistant NRAS-mutant melanoma. Indeed, combining FNB

(200mg/kg/day) with DHA (300mg/kg/day) or erastin (30mg/kg, twice daily, every other day) significantly suppressed the growth of MAPKi-resistant M93-047-derived tumors without overt toxicity in any treatment cohort (Figures 25G and 261). These results provide proof-of-principle that combining a PPARa agonist such as FNB with a lipid peroxidation enhancer can elicit anti tumor activity and restrain MAPKi-resistant NRAS-mutant melanoma.

[000129] Together, our results indicate that depletion of S6K2 induces upregulation of ROS levels and lipid synthesis, which jointly trigger lipid peroxidation and oxidative cell death (Figure 25H). By uncoupling the activities of S6K1 and S6K2, we have identified a strategy whereby enhanced activation of PPARa and enhanced desaturation of fatty acids could be potentially exploited as an anti-tumor strategy in NRAS-mutant melanoma resistant to MAPKi.

Table 4. SREBF1 siRNA and sequences of primers

DISCUSSION

[000130] The MAPK pathway, a key downstream effector of oncogenic NRAS, plays a pivotal role in melanoma. While MAPK inhibitors improve the survival of BRAF-mutant melanoma patients (Chapman, 2011 #800}{Hauschild, 2012 #801 }, they often produce heterogeneous responses and disappointing clinical outcomes in patients with NRAS-mutant melanoma { Solit, 2006 #345}{Dummer, 2017 #788}. Hence, understanding the mechanisms underpinning response and resistance to MAPKi will be critical to maximize the likelihood of selecting patients who will benefit from these regimens, and to develop effective strategies for tumors refractory to MAPK inhibitors. We have identified S6K as a critical pathway associated with resistance to MAPKi in NRAS-mutant melanoma. Furthermore, we have uncoupled the distinct roles of S6K1 vs. S6K2 in promoting survival of NRAS-mutant melanoma resistant to MAPKi. Additionally, we have uncovered a potential strategy to kill NRAS-mutant melanoma cells resistant to MAPKi, whereby selective S6K2 blockade prompts oxidative cell death. We found that depletion of S6K2 induces ROS and perturbs lipid metabolism. Suppression of S6K2 was coupled to increased unsaturated fatty acids, leading to harmful accumulation of lipid peroxides and cell death. The excessive accumulation of lipid peroxides triggered both apoptotic and ferroptotic-like cell death, linking the S6K2 effector pathways to multiple forms of cell death.

[000131] Notably, we found that silencing or inactivating S6K1 blunts the detrimental effects triggered by S6K2 depletion, indicating that S6K1 is required to induce cell death of MAPKi-resistant NRAS-mutant melanoma following S6K2 depletion. This suggests that selective inhibition of S6K2 could restrain melanoma through a mechanism distinct from PI3K and mTOR inhibitors, which concomitantly inhibit S6K1 and S6K2, and are predominantly cytostatic. Accordingly, treatment of NRAS-mutant melanoma cells with PI3K/mTOR inhibitors attenuated ROS levels, expression of enzymes involved in lipid metabolism, and lipid peroxidation. These results might partially explain why PI3K/mTOR inhibitors fail to induce cytotoxicity in NRAS-mutant melanoma as single agents. [000132] Our data support a model whereby uncoupling S6K1 and S6K2 creates a lipid metabolic imbalance that can prompt oxidative cell death when S6K2 is selectively inhibited. Nevertheless, selective S6K2 inhibitors are not yet available. Although some potential strategies aiming at specifically targeting S6K2 have been proposed, these remain to be fulfilled (Pardo, 2013 #294}{Karlsson, 2015 #179}. Mechanistically S6K2 suppression triggered oxidative cell death, by inducing PPARa and SREBP-mediated lipogenesis. These results prompted us to test the combination of the PPARa agonist fenofibrate with the polyunsaturated fatty acid DHA (both in clinical use) to mimic the effects of S6K2 depletion. This combination efficiently restricted tumor growth in a xenograft model, inducing both apoptosis and ferroptosis.

[000133] While recent studies indicate that advanced stage tumors, including melanomas refractory to therapy, are sensitive to ferroptosis (Viswanathan, 2017 #40l }{Tsoi, 2018 #674}, drugging this pathway remains challenging due to the lack of suitable inhibitors. Hence, defining key regulators and/or effectors of this cell death pathway and identifying strategies to induce ferroptosis could have a major impact for the treatment of melanoma, particularly for tumors driven by oncogenic NRAS. Our studies, provide proof-of-concept for the development of therapeutically viable options. Additionally, it would also be important to define markers that could help select tumors with the highest likelihood of benefiting from these strategies.

Consistent with previous studies (Atefi, 2015 #342}{Wagle, 2018 #747}, we found that low MAPK activity and high PPARa levels are associated with resistance to MAPKi and enhanced sensitivity to PPARa agonists. While these findings deserve further investigation, basal pERK and PPARa levels could potentially be used to identify patients who may benefit from therapies inducing lipid peroxidation. [000134] Intriguingly, we found that S6K2 depletion selectively killed NRAS-mutant melanoma cells that are resistant to MAPKi. As PPARa is a master regulator of lipid

metabolism, MAPKi-resistant melanoma cells might have a distinct metabolic state.

Conceivably, this distinct metabolic state might make MAPKi-resistant cells susceptible to the metabolic perturbations triggered by S6K2 depletion. In addition to inducing cytotoxic lipid peroxides, S6K2 also regulates anti-apoptotic signals (Liwak, 2012 #400 J { Basu, 2017 #381}; hence, suppression of S6K2 might also make cells prone to apoptosis. In contrast, in MAPKi- sensitive cells, high activity MAPK could be activating an S6K-independent anti-apoptotic program (Luciano, 2003 #472}{Hata, 2015 #47l }{Atefi, 2015 #342}, thereby protecting MAPKi-sensitive cells from cell death induced by S6K2 depletion.

[000135] Upregulation or mutations in NBAS have been implicated in regulating the mTORCl/S6Kl axis (Teh, 2018 #695}. Additionally, enhanced and/or persistent

phosphorylation of S6 has been linked to resistance to MAPKi in BRAF-mutant (Villanueva, 2013 #746} (Corcoran, 2013 #403} and MEKi/Cdk4i in NRAS-mutant melanoma (Teh, 2018 #695} (Romano, 2018 #694}. Whereas inhibiting the mTORCl/S6Kl axis could restrain specific tumors, such as BRAF- or NRAS-mutant melanoma with acquired resistance to MAPKi and/or CDK4i (Villanueva, 2013 #746} (Corcoran, 2013 #403}(Teh, 2018 #695}(Romano,

2018 #694}, our data suggest that the balance between S6K1 and S6K2 activity could provide an opportunity for synthetic lethality in NRAS-mutant melanoma.

[000136] Based on our results and given that S6K is upregulated in several tumor types (Pardo, 2013 #294}(Perez-Tenorio, 2011 #474} (ref), selectively targeting S6K2 or its downstream effectors may provide specificity toward tumor cells. Furthermore, as S6K2 depletion did not induce cell death in non-transformed human primary fibroblasts, and mice lacking S6K2 are viable {Pende, 2004 #499}, it is likely that selective abrogation of S6K2 will not trigger overt collateral toxicity. Therefore, S6K2 represents a viable therapeutic target in NRAS-mutant tumors. Further identification of the mechanisms regulating S6K2 activity and its effectors, particularly those responsible for redox and metabolic homeostasis, would provide important information about the role of this under-explored kinase in melanoma. This knowledge could pave the way for the development of novel strategies phenocopying the effects of S6K2 abrogation.

[000137] Together, our results underscore the significance of oncogenic NRAS-induced metabolic dependency and the role of S6K2 in this context. Furthermore, our studies establish that harnessing this S6K-dependent addiction and lipid homeostasis, can elicit both apoptotic and ferroptotic-like cell death, and could thereby be exploited as a potential strategy for NRAS- mutant melanoma, a tumor type with limited treatment options..

Claims

We claim:

1. A method of treating a tumor in a patient, comprising: a. taking a tumor sample from said patient and determining whether said tumor is kinase-inhibitor resistant; and b. administering to said patient a ferroptosis inducer where said

tumor is kinase-inhibitor resistant.

2. The method of claim 1, wherein the kinase inhibitor is MAPKi.

3. A method of treating a MAPkinase inhibitor (MAPKi) resistant tumor cell

comprising: contacting said cell with a ferroptosis/lipid peroxidation inducer.

4. A method of treating a RAS mutant cancer, wherein said RAS mutant cancer comprises cells which are MAPKi resistant and contacting said cancer cells with a ferroptosis inducer.

5. A method of treating melanoma tumor comprising: taking a sample form said melanoma tumor and determining whether said tumor is NRAS mutant;

administering to said patient a sufficient amount of a ferroptosis inducer.

6. A method of treating a NRAS mutant melanoma tumor comprising contacting said mutant melanoma tumor with a ferroptosis/lipid peroxidation inducer.

7. A method of treating a melanoma tumor by selective inhibition of RSK2,

comprising administering to said patient an effective amount of a ferroptosis inducing active agent.

8. The method of any claim wherein said ferroptosis inducing agent is erastin.

9. The method of any claim, wherein said ferroptosis inducing agent is selected from erastin, RSL3, FIN56, ML162, and ML210.

10. A method of treating a cancer comprising administering a first effective amount of a MEK inhibitor and a second effective amount of a ferroptosis inducing agent.

11. A method of treating a patient having a MAPKi resistant tumor comprising,

administering to said patient an effective amount of a ferroptosis inducing agent.

12. The method of claim 11, wherein the method comprises further administering an agent for selective blockage of S6K2.

13. The method of claim 11, wherein the blockage of S6K1 does not impact the levels of S6K1.

14. The method of claim 12, wherein selective blockage of S6K2 is sufficient to

generate acute depletion or inhibition of S6K2, enhanced ROS production, lipid synthesis or accumulation of intracellular unsaturated fatty acids; wherein ROS susceptible (poly)- unsaturated fatty acids sensitized cells to ROS, resulting in lipid peroxidation and oxidative cell death.

15. A method of treatment of a patient having a NRAS mutant melanoma comprising administering to said patient an effective amount of a therapeutic sufficient to induce lipid peroxidation.

16. The method of claim 15, wherein said therapeutic is a ferroptosis/lipid peroxidation inducing agent.

17. The method of claim 15, wherein the step of inducing lipid peroxidation is

generated by selective depletion of S6K2.

18. The method of claim 15 comprising administering to said patient an effective amount of a ferroptosis inducing agent.

19. A method of attenuating the growth rate of a MAPKi sensitive tumor cells

comprising: regulating the PBK/mTOR and MAPK pathways by contacting said tumor cells with a ferroptosis inducing agent.

20. A method for inducing NRAS melanoma cancer cell death by inducing ROS comprising contacting said melanoma cells with a sufficient amount of a polyunsaturated fatty acid.

21. The method of claim 20, further comprising administering to said patient a

sufficient amount of a ROS inducer.

22. The method of claim 20, wherein the ROS inducer is sorafenib or BSO.

23. The method of claim 20, further comprising administering a sufficient amount of a ferroptosis inducing agent.

24. The method of claim 23, wherein said ferroptosis agent is erastin.

25. The method of claim 21 or 23 further comprising administering to said patient an effective amount of lipid ROS scavenger Fer-l.

26. A method for treating a patient having NRAS mutant melanoma tumors comprising administering to said patient an effective amount of a polyunsaturated fatty acid.

27. The method of claim 26, further comprising administering to said patient a

sufficient amount of a ROS inducer.

28. The method of claim 26, wherein the ROS inducer is sorafenib or BSO.

29. The method of claim 26, further comprising administering a sufficient amount of a ferroptosis inducing agent.

30. The method of claim 29, wherein said ferroptosis agent is erastin.

31. The method of claim 27 or 29 further comprising administering to said patient an effective amount of lipid ROS scavenger Fer-l.

32. The method of any claim wherein said ferroptosis inducing agent is selected from the group consisting of erastin; PE; IKE; sulfasalazine; glutamate; BAY 43-9006; Sorafenib; L-Buthionine-(S,R)-Sulfoximine (BSO); N-Acetyl-4-benzoquinone imine (NAPQI); scetaminophen; (lS,3R)-RSL3; ML162, DPI compounds 7,10, 12, 13, 17, 18, and 19; GPx4 Antibody; Cystine/cysteine deprivation; BSO; DPI2; cisplatin; FIN56; FIN02; (-)-FIN02; statins (e.g., cerivastatin, simvastatin); cysteinase; silica-based nanoparticles; CCl₄; ferric ammonium citrate;

trigonelline; brusatol; artemisinin; artesunate; and combinations thereof.