WO2019140257A1

WO2019140257A1 - Compositions and methods for characterizing and treating prostate cancer

Info

Publication number: WO2019140257A1
Application number: PCT/US2019/013295
Authority: WO
Inventors: Pier Paolo Pandolfi; Ming Chen
Original assignee: Beth Israel Deaconess Medical Center, Inc.
Priority date: 2018-01-11
Filing date: 2019-01-11
Publication date: 2019-07-18

Abstract

As described below, the present invention features compositions and methods for characterizing prostate cancer for progression to metastasis, and for treating and/or preventing such progression.

Description

COMPOSITIONS AND METHODS FOR CHARACTERIZING AND TREATING

PROSTATE CANCER

CROSS-REFERENCE TO RELATED APPLICATIONS

This application claims the benefit of and priority to U.S. Provisional Patent

Application serial number 62/616,134, filed January 11, 2018, which is incorporated herein by reference in its entirety.

STATEMENT OF RIGHTS TO INVENTIONS MADE UNDER FEDERALLY

SPONSORED RESEARCH

This invention was made with government support under Grant No. R35 CA197529 awarded by the National Institutes of Health and under Grant No. W81XWH- 12- 1-0040 awarded by the Department of Defense. The government has certain rights in the invention.

BACKGROUND OF THE INVENTION

Although localized prostate cancer (CaP) is highly curable, metastatic CaP remains invariably fatal. While it has been postulated that a Western diet can promote CaP

progression, direct evidence supporting a strong association between dietary lipids and CaP is still lacking. Indeed, the rates of cancer mortality associated with metastatic disease are much higher in Western countries for many cancer types, including CaP, correlating with lifestyle factors such as diet. Moreover, the progression to metastasis represents a pivotal event influencing patient outcomes and the therapeutic options available to patients. There is an urgent need to understand the molecular events that underlie progression to metastasis, at both genetic and environmental levels. Such understanding is expected to significantly improve therapeutic options for patients and facilitate preventative interventions.

SUMMARY OF THE INVENTION

In one aspect, the present invention provides methods for treating prostate cancer in a subject, the methods comprising administering to the subject an agent that inhibits

AKT/mTOR or MAPK signaling. In another aspect, methods are disclosed for treating prostate cancer in a selected subject, the methods comprising administering to the subject an agent that inhibits

AKT/mTOR or MAPK signaling, wherein the subject is selected by detecting co-deletion of PTEN and PML or amplification of PPP1CA.

In another aspect, the invention provides methods for treating prostate cancer in a selected subject, the methods comprising administering to the subject an agent that inhibits AKT/mTOR or MAPK signaling, wherein the subject is selected by characterizing a biological sample of the subject for co-deletion of PTEN and PML·, amplification of

PPP1CA, and/or activation of SREBP. In some embodiments, activation of SREBP is characterized by assaying the lipidomic profile of a biological sample of the patient. In some embodiments of this aspect, the lipidomic profile is assayed by detecting an increase in fatty acyl chains, membrane phospholipids, or other lipids. In some embodiments of this aspect, the lipidomic profile comprises alterations in lysodimethylphosphatidylethanolamine, monoglyceride, phosphatidyl glycerol, and lysophosphatidyl glycerol.

In various embodiments of any of the above aspects, the agent that inhibits

AKT/mTOR is selected from the group consisting of rapamycin, Temsirolimus, Everolimus, and Ridaforolimus; and the agent that inhibits MAPK signaling is selected from the group consisting of SB203580, SB202190, and BIRB-796. In some embodiments of any of the above aspects, the methods further comprise administering an inhibitor of fatty acid synthesis, inhibition of cholesterol synthesis, or inhibition of SREBP. In some embodiments, the inhibitor of fatty acid synthesis is TOFA, the inhibitor of cholesterol synthesis is simvastatin, and the inhibitor of SREBP is Fatostatin. In some embodiments of any of the above aspects, the methods comprise administering to the subject a MEK inhibitor. In some embodiments, the MEK inhibitor is EG0126.

In some embodiments of any of the above aspects, the methods inhibit tumor growth and/or metastasis. In some embodiments of any of the above aspects, the methods further comprise prescribing a low-fat diet for the subject. In various embodiments of any of the above aspects, the characterization of a biological sample of a subject indicates that the prostate cancer has a propensity to metastasize.

Definitions

ETnless defined otherwise, all technical and scientific terms used herein have the meaning commonly understood by a person skilled in the art to which this invention belongs. The following references provide one of skill with a general definition of many of the terms used in this invention: Singleton et al., Dictionary of Microbiology and Molecular Biology (2nd ed. 1994); The Cambridge Dictionary of Science and Technology (Walker ed., 1988); The Glossary of Genetics, 5th Ed., R. Rieger et al. (eds.), Springer Verlag (1991); and Hale & Marham, The Harper Collins Dictionary of Biology (1991). As used herein, the following terms have the meanings ascribed to them below, unless specified otherwise.

By "agent" is meant any small molecule chemical compound, antibody, nucleic acid molecule, or polypeptide, or fragments thereof.

By“ameliorate” is meant decrease, suppress, attenuate, diminish, arrest, or stabilize the development or progression of a disease.

By "alteration" is meant a change (increase or decrease) in the expression levels or activity of a gene or polypeptide as detected by standard art known methods such as those described herein. As used herein, an alteration includes a 10% change in expression levels, preferably a 25% change, more preferably a 40% change, and most preferably a 50% or greater change in expression levels. "

By "analog" is meant a molecule that is not identical, but has analogous functional or structural features. For example, a polypeptide analog retains the biological activity of a corresponding naturally-occurring polypeptide, while having certain biochemical

modifications that enhance the analog's function relative to a naturally occurring polypeptide. Such biochemical modifications could increase the analog's protease resistance, membrane permeability, or half-life, without altering, for example, ligand binding. An analog may include an unnatural amino acid.

By“biologic sample” is meant any tissue, cell, fluid, or other material derived from an organism.

By“clinical aggressiveness” is meant the severity of the prostate cancer. Aggressive prostate cancers are more likely to metastasize than less aggressive prostate cancers. While conservative methods of treatment are appropriate for less aggressive prostate cancers, more aggressive prostate cancers require more aggressive therapeutic regimens.

By“control” is meant a standard of comparison. For example, the lipids present in a prostate cancer compared to the lipids present in a corresponding normal tissue.

By“diagnostic” is meant any method that identifies the presence of a pathologic condition or characterizes the nature of a pathologic condition (e.g., prostate cancer, metastatic prostate cancer). Diagnostic methods differ in their sensitivity and specificity. While a particular diagnostic method may not provide a definitive diagnosis of a condition, it suffices if the method provides a positive indication that aids in diagnosis.

In this disclosure, "comprises," "comprising," "containing" and "having" and the like can have the meaning ascribed to them in U.S. Patent law and can mean " includes,"

"including," and the like; "consisting essentially of' or "consists essentially" likewise has the meaning ascribed in U.S. Patent law and the term is open-ended, allowing for the presence of more than that which is recited so long as basic or novel characteristics of that which is recited is not changed by the presence of more than that which is recited, but excludes prior art embodiments.

“Detect” refers to identifying the presence, absence or amount of the analyte to be detected.

By "detectable label" is meant a composition that when linked to a molecule of interest renders the latter detectable, via spectroscopic, photochemical, biochemical, immunochemical, or chemical means. For example, useful labels include radioactive isotopes, magnetic beads, metallic beads, colloidal particles, fluorescent dyes, electron-dense reagents, enzymes (for example, as commonly used in an ELISA), biotin, digoxigenin, or haptens.

By“disease” is meant any condition or disorder that damages or interferes with the normal function of a cell, tissue, or organ. Examples of diseases include prostate cancer, particularly prostate cancer with a propensity to progress to metastasis.

By "effective amount" is meant the amount of a required to ameliorate the symptoms of a disease relative to an untreated patient. The effective amount of active compound(s) used to practice the present invention for therapeutic treatment of a disease varies depending upon the manner of administration, the age, body weight, and general health of the subject. Ultimately, the attending physician or veterinarian will decide the appropriate amount and dosage regimen. Such amount is referred to as an "effective" amount.

The invention provides a number of targets that are useful for the development of highly specific drugs to treat or a disorder characterized by the methods delineated herein. In addition, the methods of the invention provide a facile means to identify therapies that are safe for use in subjects. In addition, the methods of the invention provide a route for analyzing virtually any number of compounds for effects on a disease described herein with high-volume throughput, high sensitivity, and low complexity. By "fragment" is meant a portion of a polypeptide or nucleic acid molecule. This portion contains, preferably, at least 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, or 90% of the entire length of the reference nucleic acid molecule or polypeptide. A fragment may contain 10, 20, 30, 40, 50, 60, 70, 80, 90, or 100, 200, 300, 400, 500, 600, 700, 800, 900, or 1000 nucleotides or amino acids.

"Hybridization" means hydrogen bonding, which may be Watson-Crick, Hoogsteen or reversed Hoogsteen hydrogen bonding, between complementary nucleobases. For example, adenine and thymine are complementary nucleobases that pair through the formation of hydrogen bonds.

The terms "isolated," "purified," or "biologically pure" refer to material that is free to varying degrees from components which normally accompany it as found in its native state. "Isolate" denotes a degree of separation from original source or surroundings. "Purify" denotes a degree of separation that is higher than isolation. A "purified" or "biologically pure" protein is sufficiently free of other materials such that any impurities do not materially affect the biological properties of the protein or cause other adverse consequences. That is, a nucleic acid or peptide of this invention is purified if it is substantially free of cellular material, viral material, or culture medium when produced by recombinant DNA techniques, or chemical precursors or other chemicals when chemically synthesized. Purity and homogeneity are typically determined using analytical chemistry techniques, for example, polyacrylamide gel electrophoresis or high-performance liquid chromatography. The term "purified" can denote that a nucleic acid or protein gives rise to essentially one band in an electrophoretic gel. For a protein that can be subjected to modifications, for example, phosphorylation or glycosylation, different modifications may give rise to different isolated proteins, which can be separately purified.

By "isolated polynucleotide" is meant a nucleic acid (e.g., a DNA) that is free of the genes which, in the naturally-occurring genome of the organism from which the nucleic acid molecule of the invention is derived, flank the gene. The term therefore includes, for example, a recombinant DNA that is incorporated into a vector; into an autonomously replicating plasmid or virus; or into the genomic DNA of a prokaryote or eukaryote; or that exists as a separate molecule (for example, a cDNA or a genomic or cDNA fragment produced by PCR or restriction endonuclease digestion) independent of other sequences. In addition, the term includes an RNA molecule that is transcribed from a DNA molecule, as well as a recombinant DNA that is part of a hybrid gene encoding additional polypeptide sequence.

By an "isolated polypeptide" is meant a polypeptide of the invention that has been separated from components that naturally accompany it. Typically, the polypeptide is isolated when it is at least 60%, by weight, free from the proteins and naturally-occurring organic molecules with which it is naturally associated. Preferably, the preparation is at least 75%, more preferably at least 90%, and most preferably at least 99%, by weight, a

polypeptide of the invention. An isolated polypeptide of the invention may be obtained, for example, by extraction from a natural source, by expression of a recombinant nucleic acid encoding such a polypeptide; or by chemically synthesizing the protein. Purity can be measured by any appropriate method, for example, column chromatography, polyacrylamide gel electrophoresis, or by HPLC analysis.

By“marker” is meant any analyte (e.g., protein, polynucleotide, lipid) having an alteration in level or activity that is associated with a disease or disorder. In one embodiment, a marker is an alteration in a lipid described herein or the loss of PTEN or PML.

As used herein,“obtaining” as in“obtaining an agent” includes synthesizing, purchasing, or otherwise acquiring the agent.

By“reduces” is meant a negative alteration of at least 10%, 25%, 50%, 75%, or

100%.

By“reference” is meant a standard or control condition.

A "reference sequence" is a defined sequence used as a basis for sequence

comparison. A reference sequence may be a subset of or the entirety of a specified sequence; for example, a segment of a full-length cDNA or gene sequence, or the complete cDNA or gene sequence. For polypeptides, the length of the reference polypeptide sequence will generally be at least about 16 amino acids, preferably at least about 20 amino acids, more preferably at least about 25 amino acids, and even more preferably about 35 amino acids, about 50 amino acids, or about 100 amino acids. For nucleic acids, the length of the reference nucleic acid sequence will generally be at least about 50 nucleotides, preferably at least about 60 nucleotides, more preferably at least about 75 nucleotides, and even more preferably about 100 nucleotides or about 300 nucleotides or any integer thereabout or therebetween.

By“severity of prostate cancer” is meant the degree of pathology. The severity of a prostate cancer increases, for example, as the stage or grade of the prostate cancer increases. By "siRNA" is meant a double stranded RNA. Optimally, an siRNA is 18, 19, 20, 21, 22, 23 or 24 nucleotides in length and has a 2-base overhang at its 3' end. These dsRNAs can be introduced to an individual cell or to a whole animal; for example, they may be introduced systemically via the bloodstream. Such siRNAs are used to downregulate mRNA levels or promoter activity. In one embodiment, an agent for the treatment of prostate cancer is an siRNA.

By "specifically binds" is meant a compound or antibody that recognizes and binds a polypeptide of the invention, but which does not substantially recognize and bind other molecules in a sample, for example, a biological sample, which naturally includes a polypeptide of the invention.

Nucleic acid molecules useful in the methods of the invention include any nucleic acid molecule that encodes a polypeptide of the invention or a fragment thereof. Such nucleic acid molecules need not be 100% identical with an endogenous nucleic acid sequence, but will typically exhibit substantial identity. Polynucleotides having“substantial identity” to an endogenous sequence are typically capable of hybridizing with at least one strand of a double-stranded nucleic acid molecule. Nucleic acid molecules useful in the methods of the invention include any nucleic acid molecule that encodes a polypeptide of the invention or a fragment thereof. Such nucleic acid molecules need not be 100% identical with an endogenous nucleic acid sequence, but will typically exhibit substantial identity. Polynucleotides having“substantial identity” to an endogenous sequence are typically capable of hybridizing with at least one strand of a double-stranded nucleic acid molecule.

By "hybridize" is meant pair to form a double-stranded molecule between complementary polynucleotide sequences (e.g., a gene described herein), or portions thereof, under various conditions of stringency. (See, e.g., Wahl, G. M. and S. L. Berger (1987) Methods Enzymol. 152:399; Kimmel, A. R. (1987) Methods Enzymol. 152:507).

For example, stringent salt concentration will ordinarily be less than about 750 mM NaCl and 75 mM trisodium citrate, preferably less than about 500 mM NaCl and 50 mM trisodium citrate, and more preferably less than about 250 mM NaCl and 25 mM trisodium citrate. Low stringency hybridization can be obtained in the absence of organic solvent, e.g., formamide, while high stringency hybridization can be obtained in the presence of at least about 35% formamide, and more preferably at least about 50% formamide. Stringent temperature conditions will ordinarily include temperatures of at least about 30° C, more preferably of at least about 37° C, and most preferably of at least about 42° C. Varying additional parameters, such as hybridization time, the concentration of detergent, e.g., sodium dodecyl sulfate (SDS), and the inclusion or exclusion of carrier DNA, are well known to those skilled in the art. Various levels of stringency are accomplished by combining these various conditions as needed. In a preferred: embodiment, hybridization will occur at 30° C in 750 mM NaCl, 75 mM trisodium citrate, and 1% SDS. In a more preferred embodiment, hybridization will occur at 37° C in 500 mM NaCl, 50 mM trisodium citrate, 1% SDS, 35% formamide, and 100 pg/ml denatured salmon sperm DNA (ssDNA). In a most preferred embodiment, hybridization will occur at 42° C in 250 mM NaCl, 25 mM trisodium citrate,

1% SDS, 50% formamide, and 200 pg/ml ssDNA. Useful variations on these conditions will be readily apparent to those skilled in the art.

For most applications, washing steps that follow hybridization will also vary in stringency. Wash stringency conditions can be defined by salt concentration and by temperature. As above, wash stringency can be increased by decreasing salt concentration or by increasing temperature. For example, stringent salt concentration for the wash steps will preferably be less than about 30 mM NaCl and 3 mM trisodium citrate, and most preferably less than about 15 mM NaCl and 1.5 mM trisodium citrate. Stringent temperature conditions for the wash steps will ordinarily include a temperature of at least about 25° C, more preferably of at least about 42° C, and even more preferably of at least about 68° C. In a preferred embodiment, wash steps will occur at 25° C in 30 mM NaCl, 3 mM trisodium citrate, and 0.1% SDS. In a more preferred embodiment, wash steps will occur at 42 C in 15 mM NaCl, 1.5 mM trisodium citrate, and 0.1% SDS. In a more preferred embodiment, wash steps will occur at 68° C in 15 mM NaCl, 1.5 mM trisodium citrate, and 0.1% SDS.

Additional variations on these conditions will be readily apparent to those skilled in the art. Hybridization techniques are well known to those skilled in the art and are described, for example, in Benton and Davis (Science 196: 180, 1977); Grunstein and Hogness (Proc. Natl. Acad. Sci., USA 72:3961, 1975); Ausubel et al. (Current Protocols in Molecular Biology, Wiley Interscience, New York, 2001); Berger and Kimmel (Guide to Molecular Cloning Techniques, 1987, Academic Press, New York); and Sambrook et al., Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, New York.

By "substantially identical" is meant a polypeptide or nucleic acid molecule exhibiting at least 50% identity to a reference amino acid sequence (for example, any one of the amino acid sequences described herein) or nucleic acid sequence (for example, any one of the nucleic acid sequences described herein). Preferably, such a sequence is at least 60%, more preferably 80% or 85%, and more preferably 90%, 95% or even 99% identical at the amino acid level or nucleic acid to the sequence used for comparison.

Sequence identity is typically measured using sequence analysis software (for example, Sequence Analysis Software Package of the Genetics Computer Group, University of Wisconsin Biotechnology Center, 1710 University Avenue, Madison, Wis. 53705,

BLAST, BESTFIT, GAP, or PILEUP/PRETTYBOX programs). Such software matches identical or similar sequences by assigning degrees of homology to various substitutions, deletions, and/or other modifications. Conservative substitutions typically include substitutions within the following groups: glycine, alanine; valine, isoleucine, leucine;

aspartic acid, glutamic acid, asparagine, glutamine; serine, threonine; lysine, arginine; and phenylalanine, tyrosine. In an exemplary approach to determining the degree of identity, a BLAST program may be used, with a probability score between e ³ and e ¹⁰⁰ indicating a closely related sequence.

By "subject" is meant a mammal, including, but not limited to, a human or non human mammal, such as a bovine, equine, canine, ovine, or feline.

Ranges provided herein are understood to be shorthand for all of the values within the range. For example, a range of 1 to 50 is understood to include any number, combination of numbers, or sub-range from the group consisting 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15,

16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40,

41, 42, 43, 44, 45, 46, 47, 48, 49, or 50.

As used herein, the terms“treat,” treating,”“treatment,” and the like refer to reducing or ameliorating a disorder and/or symptoms associated therewith. It will be appreciated that, although not precluded, treating a disorder or condition does not require that the disorder, condition or symptoms associated therewith be completely eliminated.

Unless specifically stated or obvious from context, as used herein, the term "or" is understood to be inclusive. Unless specifically stated or obvious from context, as used herein, the terms "a", "an", and "the" are understood to be singular or plural.

Unless specifically stated or obvious from context, as used herein, the term“about” is understood as within a range of normal tolerance in the art, for example within 2 standard deviations of the mean. About can be understood as within 10%, 9%, 8%, 7%, 6%, 5%, 4%, 3%, 2%, 1%, 0.5%, 0.1%, 0.05%, or 0.01% of the stated value. Unless otherwise clear from context, all numerical values provided herein are modified by the term about. The recitation of a listing of chemical groups in any definition of a variable herein includes definitions of that variable as any single group or combination of listed groups. The recitation of an embodiment for a variable or aspect herein includes that embodiment as any single embodiment or in combination with any other embodiments or portions thereof.

Any compositions or methods provided herein can be combined with one or more of any of the other compositions and methods provided herein.

BRIEF DESCRIPTION OF THE DRAWINGS FOR EXAMPLE 1 FIGs. 1A to IF show co-loss of PTEN and PML expression in advanced and metastatic human CaP. (a-c) Bar graph showing the percentage of deletion of PTEN { a), PML (b), or PTEN and PML (c) in LPC and mCRPC samples from the Grasso et al. dataset¹⁴. Co- loss events include homozygous, heterozygous or mixed co-loss event. (d,e) Bar graph showing the correlation of disease progression with low level of PTEN (d) or PML (e) protein in primary CaP. The color bar represents the intensity of staining, which was ranked into one of three groups: normal=2, low=l, and negative=0. (f) Bar graph showing the percentage of co-loss of PTEN and PML proteins in low-grade and high-grade primary CaP. (g) Univariate and multivariable Cox -proportional regression analysis of PTEN and PML loss, Gleason score, and pathologic stage (h) Kaplan-Meier plot of overall survival data for CaP patients after radical prostatectomy based on co-loss of PTEN and PML protein. In c and f, Fisher’s exact test (two-tailed) was used to determine significance. In d and e, Pearson’s chi-squared test was used to determine significance.

FIGs. 1-1 A to 1-1 J show co-loss of PTEN and PML expression in advanced and metastatic human CaP. (a) Bar graph showing the percentage of co-deletion of PTEN with 58 high-confidence TSGs32 in the Grasso et al. dataset of the mCRPC samplesl4, respectively (4 out of 62 TSGs from the Walker et al gene list, data not available). The genes co-deleted with PTEN only in metastatic disease among the top 25 TSGs are highlighted in Red. (b,c) Bar graph showing the percentage of deletion of PTEN or PML (b), or PTEN and PML (c) in mCRPC samples from the Robinson et al. dataset25. (d) Representative homozygous or hemizygous focal PML deletion from the Robinson et al. dataset (38%

(17/45) of PML deletion was focal)25. Copy number plots with x-axis representing chromosomal 15q and y-axis referring to copy number level. Red open circle indicates genomic position of PML. (e) Representative IHC staining of PTEN or PML showing examples of low, medium and high staining. Scale bar, 50 pm. (f) Table showing the significant correlation of co-loss of PTEN and PML protein expression during the disease progression. The number of cases in each expression category was listed together with Gleason score (g-j) Overall survival curves for CaP patients after radical prostatectomy based on the expression of PML protein (g), the expression of PTEN protein (h), Gleason score (i), or pathologic stage (j). In f, Pearson’s chi-squared test was used to determine significance.

FIGs. 1-1K to 1-1M depict the generation of Pmlflox/flox mice (a) Schematic map of the WT Pml locus (top), targeting vector (upper middle) and predicted targeted allele (lower middle) and floxed allele (bottom). The Pml genomic sequence was cloned and inserted into the pEZ-LOX-FRTDT vector. Black triangles mark the location of loxP sites that were utilized to excise the exon 2. Blue triangles mark the location of FRT sites that were utilized to excise the neomycin resistant cassette. The probes for Southern blot analysis are indicated (5' and 3' probes). BamHI digestion of genomic DNA from targeted ES cells was use to distinguish WT and targeted allele. BamHI (B), Seal (S), Notl (N). (b) Southern blot analysis of recombined ES cell clones after digestion with BamHI and hybridization with the 5’ probe (upper panel) and the neomycin (lower panel) probe. ES cell clones with corrected homologous recombination are highlighted in red. (c) Southern blot analysis of tail DNA from F2 mice after digestion with BamHI and hybridization with the 5’ probe (top), 3’probe and neomycine probe (bottom). The mice with deletion of neomycin resistant cassette are highlighted in red.

FIGs. 2A to 2F demonstrate that Pml loss renders localized Pten- null tumors lethal and metastatic to lymph node (a) IHC staining for Pml in the DLP and anterior prostate (AP) tissues from WT, Pten^pc and Pten^pc PmP^c mice at 12 weeks of age. Scale bar, 20 pm.

(b) Bar graph showing prostate tumor progression in Pten^pc and PtetP^PmP⁰^- mice at different age cohorts (h=10). The AP was excluded from the analysis due to cystic lesions after 12-week time point. HGPIN, high-grade PIN; LGPIN, low-grade PIN. (c) H&E-stained mouse prostate tissues and gross anatomy of representative WT, Pten^pc and

PtetP^-PmP⁰^- urogenital tract at 13 months of age. Scale bar, upper panel: 50 pm; lower panel: 5 mm. (d) Bar graph showing the sphere-forming ability of WT, Pten^pc and

PtetP^-PmP⁰^- prostatic epithelia following serial passage. For comparative purposes, the results were normalized as a percentage relative to the sphere number of WT (set at 100%) at the passages examined. Representative images of spheres at Pl shown at the lower panel. Scale bar, 20 pm. (e) Cumulative survival analysis of WT, PmP^c , Pten^pc and PterP^{z l} PmP^{z l} mice (f) Incidence of lymph node metastasis in cohorts of PterP^z and PtetP^-PmP⁰^- mice (g) H&E and IHC staining of lumbar lymph node metastases from two representative PterP^{z h}PmP^z mice. Arrows indicate metastases. Scale bar, 50 pm. In d, the results of one representative experiment are shown (n=5). Data are from three independent cultures and are mean ± s.e.m. Student’s t test (two-tailed) was used to determine

significance. In f, Fisher’s exact test (two-tailed) was used to determine significance.

FIGs. 2-1A to 2-1F show that Pml loss drives MAPK reactivation and metastatic progression in Pten-null CaP. (a) IHC staining for Pml in the VP tissues from WT, Ptenpc-/- and Ptenpc-/-Pmlpc-/- mice at 12 weeks of age. (b) H&E and IHC staining of the DLP tissues from Ptenpc-/- and Ptenpc-/-Pmlpc-/- mice at 20 weeks of age. Note that tumors in Ptenpc-/-Pmlpc-/- mice acquired invasive feature. Invasiveness was confirmed by the absence of SM-a-actin staining along with high level of Ki67 staining in the cancer cells. Arrows indicate invasive cancer (c) Higher magnification of Ptenpc-/-Pmlpc-/- tumors at 13 months of age showing predominate adenocarcinoma (arrows) in the presence of focal features of sarcomatoid carcinoma with high-grade pleomorphic spindle cells (arrowheads) (d) IB analysis of tissue lysates for Pml from a WT mouse (e) H&E-stained low-grade PIN in the VP and DLP tissues from a Pmlpc-/- mouse at 12 months of age. Insets show crowding cells with large nuclei (f) H&E and IHC staining of lumbar lymph node metastases from three Ptenpc-/-Pmlpc-/- mice. Arrows indicate metastases.

FIGs. 3A to 3F show that PML loss triggers MAPK reactivation in PTEN-mi\\ cells. (a,b) Immunoblot (IB) (a) and IHC (b) analyses of the DLP tissues from WT, PmP^z , PterP^z and PterP^z PmP^{z h} mice at 12 weeks of age. Scale bar, 50 pm. (c-f) IB of lysates from LNCaP (c) or PC3 (d) cells transfected with control or PML siRNA for 72 hours (hours), lysates from LNCaP (e) or PC3 (f) cells pretreated with vehicle or 1 pM arsenic trioxide (AS₂0₃) for 12 hours, followed by serum-starvation for 4 hours and stimulation with 10 ng/ml EGF for 5 min. In a, numbers indicate the relative ratios to PterP^z mice for phosphoprotein/total protein. In c-f, numbers indicate the relative ratios to controls for phosphoprotein/total protein.

FIGs. 3-1A and 3-1B show (a) IHC staining for phosphor-ERK in the DLP tissues from three pairs of Ptenpc-/- and Ptenpc-/-Pmlpc-/- mice at 12 weeks of age. (b) IB analysis of the DLP tissue lysates from WT, Ptenpc-/- and Ptenpc-/-Pmlpc-/- mice at 12 weeks of age. (c) IHC staining for Pml and phosphor-ERK in the DLP tissues from a

Ptenpc-/-Pmlpc+/- mouse at 12 weeks of age. Arrows indicate areas with lower level of Pml, but higher level of p- ERK. Arrowheads indicate areas with higher level of Pml, but lower level of p-ERK. Scale bars in all panels, 50 pm.

FIGs. 4A to 4G demonstrate that an SREBP-dependent lipogenic program is hyperactivated in prostate tumors from PtetP^PmP⁰^- mice (a) Top enriched‘biological process’ categories from GO enrichment analysis among WT, PterP^z and Pter ^PmP⁰^ prostates (b) GSEA enrichment plot for one of the top-scoring gene sets, SREBP targets, among WT, PterP^z and PtetP^PmP⁰^- prostates. (c,d) The validation of the expression changes in SREBP targets by the qPCR (c) and IB (d) analyses of the DLP tissues from WT, PterP^z and PterP^z PmP^z mice at 12 weeks of age. P and N shown in (d) denote the precursor and cleaved nuclear forms of SREBP-l or -2. (e) A scatter plot shows the distribution of all identifiable 1,743 lipid ions in prostate tissues from the three genotypes of mice. Each dot represents one lipid ion and each color represents a class of lipids. The representative lipid classes (6 out of 35) were shown. The insert displays the same chart with 3-D visual effect with Z-axis plotting the loglO mean intensity of each lipid ion across the three genotypes. (f,g) A statistically significant increase in abundance of various lipid classes (f) or fatty acyl chains (g) in PterP^PmP⁰^ tumors compared to PterP^z tumors. The bar graph represents fold change of the relative intensity. LdMePE:

lysodimethylphosphatidylethanolamine, MG: monoglyceride, PG: phosphatidylglycerol, and LPG: lysophosphatidyl glycerol. In c, the results of one representative experiment are shown (n=3). Data are from three independent mice per genotype and are mean ± s.e.m. In f and g, data shown are mean ± s.e.m. Student’s t test (two-tailed) was used to determine significance.

FIGs. 4-1A and 4-1B show the transcriptome and lipidomics profiling among WT, Ptenpc-/- and Ptenpc-/-Pmlpc-/- prostates (a) Representative H&E staining of DLP from WT, Ptenpc-/- and Ptenpc-/-Pmlpc-/- mice at 12 weeks of age. Scale bar, 50 pm. (b) Heat map of the SREBP targets in the microarray analysis of prostate tissues from the three genotypes of mice (c-f) GSEA enrichment plot for the targets of LXR, ChREBP, PPARy, and USF. The up- to down-regulated genes from the ranked gene list were analysed with the GSEA algorithm for enrichment of all gene sets in MSigDB among WT, Ptenpc-/- and Ptenpc-/-Pmlpc-/- prostates. (g,h) The relative intensity of all the identifiable 35 lipid classes (g) or the 30 most abundant fatty acyl chains (h) in prostate tissues from the three genotypes of mice (i) Heat map of the top 70 most regulated lipid ions in prostate tissues from the three genotypes of mice. Data shown in g and h are mean ± s.e.m FIGs. 5A to 5D show that SREBP is the downstream target of PML- loss induced MAPK activation. (a,b) qPCR of SREBP targets in LNCaP (a) or PC3 (b) cells. CaP cells were transfected with control or L□ siRNA for 48 hours, followed by l2-hr treatment with DMSO or 20 mM U0126 in media with 10% lipoprotein deficient serum. (c,d) IB analysis of total lysates or nuclear extracts from LNCaP cells. LNCaP cells were transfected with control or PML siRNA (c) or empty vector or HA-MEKl^S218D/S222D (d) for 24 hours, followed by 24- hr treatment with DMSO or 20 mM U0126 in media with 10% lipoprotein deficient serum. In a and b, the results of one representative experiment are shown (n=3). Data are from three independent cultures and are mean ± s.e.m.

FIGs. 6A to 6G show that an SREBP-dependent lipogenesis is critical for PML-loss induced CaP growth and metastasis. (a,b) Representative images and quantitation of migrated or invaded LNCaP cells in the migration and invasion assay. LNCaP cells were transfected with control or indicated siRNA against PML or/and SREBP-l (a), or pretreated with DMSO, 10 pg/ml TOFA or 10 pM Simvastatin (b), for 48 hours, then subjected to migration (24 hours) or invasion assay (48 hours). Scale bar, 50 pm. (c-g) Gross anatomy of representative PterP^{c h}PmP^c tumors (c), quantitation of tumor weight (d) and the incidence of metastasis (e), and IB (f) and IHC (g) analyses of tumors from PterP^{c h}PmP^c mice after the treatment of vehicle or l5mg/kg fatostatin for two months. Scale bar, 5mm (c) and 50 pm (g). Arrowheads in (g) indicate apoptotic cells. In a and b, the results of one representative experiment are shown (n=5). Data are from three independent cultures (4 fields per insert). Data shown in a, b, and d are mean ± s.e.m. Student’s t-test (two-tailed) was used to determine significance.

FIGs. 6-1A to 6-1C show that SREBP-dependent lipogenesis is important for PML- loss induced CaP growth and metastasis. (a,b) Representative images and quantitation of migrated and invaded PC3 cells transfected with siRNA against PML or/and SREBP-l (a), or LNCaP cells transfected siRNA against SREBP-2 (b), in the migration and invasion assays. CaP cells were transfected with control or indicated siRNA for 48 hours. PC3 cells were then subjected to 24-hr migration and invasion assay, while LNCaP cells were subjected to 24-hr migration and 48-hr invasion assay (c) H&E and IHC staining of metastases in the lumbar lymph node of two vehicle-treated Ptenpc-/-Pmlpc-/- mice. Arrows indicate metastases. In a and b, the results of one representative experiment are shown (n=3). Data are from three independent cultures (4 fields per insert). Data shown are mean ± s.e.m. Student’s t-test (twotailed) was used to determine significance. Scale bars in all panels, 50 pm. FIGs. 7A to 7K show that a high fat diet (HFD) drives metastatic progression in mouse models of CaP and increases lipid abundance in prostate tumors (a) HFD-fed mice gain body weight. Mice at 12 months of age were fed chow or HFD for 3 months (n=8).

Body weights of mice were measured per month (b-e) H&E and IHC staining of metastases in the lumbar lymph node and lung of a representative HFD-fed PtetP^PmP⁰^- mouse (b) or PterP^c mouse (d) and the comparison of the incidence of metastasis between chow- and HFD-fed PterP^c PmP^c mice (c) or PterP^c mice (e). (f) The ORO staining of chow- or HFD-fed PterP^c and PterP^PmP^ tumors (g-i) Heat map of the top 70 most regulated lipid ions (g), statistically significant increased lipid classes (h) or fatty acyl chains (i) in HFD-fed PterP^z and Pter ^^PmP^0-1- tumors compared to chow-fed counterparts (j) The ORO staining of vehicle or dietary lipids treated LNCaP cells (k) Representative images and quantitation of migrated or invaded LNCaP cells in the migration and invasion assay. LNCaP cells were pretreated with BSA vehicle control, 2% lipid mixture, 30 pm BSA-conjugate palmitic acid or oleic acid for 7 days, then subjected to migration (24 hours) or invasion assay (48 hours). Arrows in (b,d) indicate metastases. In k, the results of one representative experiment are shown (n=5). Data are from three independent cultures (4 fields per insert). In a, h, i and k, data shown are mean ± s.e.m. Student’s t-test (two-tailed) was used to determine significance. In c and e, Fisher’s exact test (two-tailed) was used to determine significance. Scale bars in all panels, 50 pm.

FIGs. 7-1A to 7-1G show that HFD drives metastatic progression in mouse models of CaP and increases lipid abundance in prostate tumors. (a,b) H&E and IHC staining of metastases in the lung of a representative Ptenpc-/-Pmlpc-/- mouse (a) or a Ptenpc-/- mouse (b). Arrows indicate metastases. (c) The survival analysis of Ptenpc-/- and

Ptenpc-/-Pmlpc-/- mice upon 3-month HFD feeding beginning at 12 months of age. (d,e) The relative intensity of all the identifiable 36 lipid classes (d) or the 30 most abundant fatty acyl chains (e) in prostate tissues from chow- or HFD- fed Ptenpc-/- and

Ptenpc-/-Pmlpc-/- mice (f) The ORO staining of vehicle or dietary lipids treated PC3 cells (g) Representative images and quantitation of migrated or invaded PC3 cells in the migration and invasion assay. PC3 cells were pretreated with BSA, 2% lipid mixture, BSA-conjugate palmitic acid or oleic acid for 7 days, then subjected to 24-hr migration and invasion assay. Data are from three independent cultures (4 fields per insert). Data shown in d, e, g are mean ± s.e.m. Student’s t-test (two-tailed) was used to determine significance. Scale bars in all panels, 50 pm. In g, the results of one representative experiment are shown (n=5).

FIG. 8 demonstrate that an SREBP signature is highly enriched in metastatic human CaP. (a,b) GSEA enrichment plot for gene set derived from SREBP-l signature. The up- to down-regulated genes from ranked gene list for metastasis versus primary samples in the Grasso et al.¹⁴ (a) or Taylor et al.¹⁵ (b) dataset were analyzed with the GSEA algorithm for enrichment of SREBP-l signature. SREBP-l signature was inferred by information-theory based CLR algorithm using the RNA-seq data from TCGA normal prostate samples. SREBP- 1 signature was split into two subsets.“POS” for positively correlated, and“NEG” for negatively correlated between SREBP-l and its targets (c) Mechanisms of aberrant lipid metabolism, regulated by PML, and the increased lipid influx by HFD as a potential risk factor for metastatic CaP.

FIG. 8-1A show the serum testosterone levels in chow- or HFD- fed Ptenpc-/- and Ptenpc-/-Pmlpc-/- mice at 14-15 months of age.

FIG. 8-1B show the validation of the expression changes of hypoxia-induced target genes by the qPCR among WT, Ptenpc-/- and Ptenpc-/-Pmlpc-/- prostates. Data shown in b are mean ± s.e.m.

FIGs. 9A to 9D demonstrate that amplification of PPP1CA occurs frequently in metastatic human CaP and is often mutually exclusive with co-deletion of PTEN and PML. (a,b) Bar graph showing the percentage of amplification of PPP1CA in LPC and mCRPC samples from the Grasso et al. dataset⁸ (a) and in mCRPC samples from the Robinson et al. dataset⁹ (b). (c) Bar graph showing the percentage of amplification of cyclin Dl alone or together with PPP1CA in mCRPC samples from the Robinson et al. dataset⁹ (d) Association between genomic amplification of PPP1CA and co-deletion of PTEN and PML from the Robinson et al. dataset⁹. Data were analyzed by Fisher’s exact test, P < 0.05 was consider significant.

FIGs. 9-1 A to 9-1B demonstrate that PTEN loss leads to feedback inhibition of ERK- MAPK signaling. (a,b) IB analysis of dorsal-later prostate tissues from wild type and prostate epithelium-specific Pten inactivation (Ptenpc-/-) mice at 12 weeks of age (a), lysates from primary Pteniox/ioxMEFs transduced with control or Cre retrovirus at 72 hours after selection (b).

FIGs. 10A to 10H demonstrate that PPla mediates PML- loss induced MAPK activation. (a,b) IB analysis of lysates from 293T cells transfected with empty vector (EV) or Flag-PPla for the indicated times (a), lysates from LNCaP or PC3 cells transfected with control or PPla siRNA for 48 hours (b), (c,d) Immunofluorescence (c) and fractionation (d) of PPla protein in WI-38 cells transfected with control, PTICN or RΊΈN plus PML siRNA for 48 hours. Quantification of the percentage of PPla protein in the cytosolic and nuclear fraction was carried out with Image J software after it was normalized against cytosolic marker HSP90 and nuclear marker Lamin Bl, respectively. Data shown are mean ± s.e.m. of three independent experiments. Scale bar, 5 pm. (e) Immunofluorescence and IB analysis of 293T cells transfected with EV, WT or NLS-tagged Flag-PPla. Scale bar, 5 pm. (f)

Endogenous reciprocal co-immunoprecipitation of PML and PPla in PC3 cells. Input is 10% of total cell extracts used for immunoprecipitation. indicates a nonspecific band (g) IB analysis of lysates from control or PML stable knocked-down PC3 cells transfected with control or PPla siRNA for 48 hours, followed by serum-starvation for 4 hours and stimulation with lOng/ml EGF for 5 mins. Quantification of p-ERK/ERK levels was carried out with Image J software. Numbers indicate the relative ratios to controls for

phosphoprotein/total protein. Data shown are mean ± s.e.m. of three independent

experiments (h) IB analysis of lysates from 293T cells transfected with EV, WT or phosphatase-inactive (H248K) Flag-PPla for 24 hours.

FIGs. 10-1 A to 10-11 demonstrate that PPla mediates PML-loss induced MAPK activation and promotes CaP cell invasiveness through activation of MAPK signaling (a) Immunoblot (IB) analysis of lysates from LNCaP cells transfected with EV or Flag-PPla for the indicated time periods. Quantification of p-ERK/ERK levels was carried out with Image J software. Numbers indicate or PP2A-C for the indicated times (c), lysates from LNCaP (d) or PC3 cells (e) transfected with control or PP2A-C siRNA for 48 hours (f) Genetic alterations of PPP2CA in the Robinson et al. datasets The gene alteration percentages are shown. It should be noted that neither genomic amplification nor deletion was observed for PPP2CA in this dataset of 150 samples from mCRPC patients (g) Endogenous coimmunoprecipitation of Rb with PPla and PML in PC3 cells. Input is 10% of total cell extracts used for

immunoprecipitation. (h) Cytochemical staining and quantification of senescence-associated b-galactosidase (SA-P-gal) activity in WI-38 cells transfected with control, PTEN siRNA or PTEN plus PML siRNA at 5 days post-transfection. Data shown are mean ± s.d. of three independent experiments. ***P<0.00l by unpaired two-tailed t-test. Scale bar, 100 pm. (i) IB analysis of lysates from serum-starved PC3 cells pretreated with tautomycin at the indicated concentration for 3 hours, followed by stimulation with lOng/ml EGF for 5 min. FIGs. 11A to 111 demonstrate that S6K1 phosphorylates PPla, induces the binding of PPla with 14-3-3g and triggers its cytoplasmic accumulation. (a,b) IB analysis of total lysates and immunoprecipitates from 293T cells transfected with Flag-PPla plus EV or the indicated HA-tagged constitutively active (CA) AGC family kinases for 48 hours (a), from 293T cells transfected with Flag-PPla, HA-S6K1-CA plus EV or the indicated HA-tagged 14-3-3 isoforms for 48 hours (b). (c) Fractionation of 293T cells transfected with Flag-PPla plus EV or HA-S6K1-CA for 48 hours (d) Immunofluorescence and quantitation of 293T cells transfected with Flag-PPla and HA-S6K1-CA for 48 hours. Arrowhead, HA-S6Kl^low cells; Asterisk, HA-S6K 1 ^high cells; Data shown are mean ± s.e.m. of three independent experiments. **P<0.0l, ***/^J<0 001 by unpaired two-tailed /-test. Scale bar, 10 pm. (e) IB analysis of lysates from 293T cells transfected with control or S6K1 siRNA in the absence and in the presence of Flag-PPla for 24 hours (f) A schematic showing the two highly conserved putative S6K sites, S224/T226 and T320, in PPla protein (g) In vitro S6K- mediated PPla kinase assays. Bacterial expressed WT or mutant GST-PPla was purified and incubated with S6K1 in the kinase buffer with [g-³²R] ATP. Reaction was stopped by sample buffer and resolved by SDS-PAGE. (h,i) IB analysis of total lysates and immunoprecipitates from 293T cells transfected with the indicated WT or mutant Flag-PPla constructs, HA- S6K1-CA plus EV or HA-14-3-3g for 48 hours (h), lysates from 293T cells transfected with EV or the indicated WT or mutant Flag-PPla constructs for 24 hours (i).

FIGs. 12A to 12F demonstrate that PPla acts as a B-Raf activating phosphatase and promotes CaP cell invasiveness through activation of MAPK signaling (a) Endogenous co- immunoprecipitation of PPla with A-Raf, B-Raf, C-Raf, MEK and ERK (the left panel) or B-Raf with PPla (the right panel) in PC3 cells. Input is 10% of total cell extracts used for immunoprecipitation. indicates a nonspecific band (b) IB analysis of lysates from 293T cells transfected with the indicated WT or mutant Flag-B-Raf constructs plus EV or Flag- PPla for 24 hours. Quantification of p-ERK/ERK levels was carried out with Image J software. Numbers indicate the relative ratios to controls for phosphoprotein/total protein. Data shown are mean ± s.e.m. of three independent experiments (c) In vitro ERK-mediated B-Raf kinase assays. Bacterial expressed WT or mutant GST-B-Raf was purified and incubated with ERK2 in the kinase buffer with [g-³²R] ATP. Reaction was stopped by sample buffer and resolved by SDS-page. (d) In vitro ERK-mediated B-Raf kinase and PPla phosphatase assay. Bacterial expressed WT or mutant GST-B-Raf was purified and incubated with ERK2 in the kinase buffer with [g-³²R] ATP in the absence or presence of recombinant PPla. Reaction was stopped by sample buffer and resolved by SDS-PAGE. (e) Representative images and quantitation of migrated and invaded LNCaP cells in the migration and invasion assay. LNCaP stable cells were subjected to migration (24 hours) or invasion assay (48 hours) in the absence or presence of 20 mM U0126. Western blotting confirmed the expression of phosphor-ERK and PPla. Data shown are mean ± s.e.m. of three independent experiments. **P<0.0l, ***/^J<0 001 by unpaired two-tailed /-test. Scale bar,

100 pm. (f) A model of the regulation of MAPK activation in PTEN- A\ cells by S6K, PPla, B-Raf and PML.

FIGs. 12-1A to 12-1F shows (a-c) IB analysis of lysates from 293T cells transfected with the indicated WT or mutant Flag-B-Raf constructs plus EV or Flag-PPla for 24 hours (d) Representative images and quantitation of migrated and invaded PC3 cells in the migration and invasion assay. PC3 stable cells were subjected to migration or invasion assay (24 hours) (n=3 per group, 4 fields per insert) in the absence or presence of 20 pM U0126. Western blotting confirmed the expression of phosphor-ERK and PPla. Data shown are mean ± s.e.m. of three independent experiments. *P<0.05, **R<0.01, ***P<0.00l by unpaired two-tailed t-test. Scale bar, 100 pm. (e) In vitro phosphatase assays showing that PML did not affect PPla phosphatase activity towards dephosphorylating Flag-CREB. (f) In vitro kinase and phosphatase assays showing that 14-3-3g did not affect PPla phosphatase activity towards dephosphorylating GST-B-Raf.

DETAILED DESCRIPTION OF THE INVENTION

The invention is based, at least in part, on the discovery of that PML is frequently co- del eted with PTEN in metastatic human prostate cancer (CaP). We demonstrate that conditional inactivation of Pml in the mouse prostate morphs indolent Pten- null tumors into lethal metastatic disease. We identify MAPK reactivation, the subsequent hyperactivation of an aberrant SREBP pro-metastatic lipogenic program, and a distinctive lipidomic profile, as key characteristic features of metastatic PmllPten double null CaP. Furthermore, targeting SREBP in vivo by fatostatin blocks both tumor growth and distant metastasis. Importantly, a high-fat diet (HFD) induces lipid accumulation in prostate tumors and is sufficient to drive metastasis in non-metastatic Pten- null mouse model of CaP, and an SREBP signature is highly enriched in metastatic human CaP. Thus, our findings uncover a pro-metastatic lipogenic program, and lend direct genetic and experimental support to the notion that a Western HFD could promote metastasis.

In other embodiments, the invention is based, at least in part, on the discovery that that genomic amplification of the PPP1CA gene is highly enriched in metastatic human CaP. We further identify an S6K/PPla/B-Raf signaling pathway leading to activation of MAPK signaling that is antagonized by the PML tumor suppressor. Mechanistically, PPla was found to act as a B-Raf activating phosphatase and was found to PML suppress MAPK activation by sequestering PPla into PML nuclear bodies, hence repressing S6K-dependent PPla phosphorylation, 14-3-3 binding and cytoplasmic accumulation. Our findings therefore reveal a PPla-PML molecular network that is genetically altered in human cancer towards aberrant MAPK activation, with important therapeutic implications.

Here, an SREBP pro-metastatic lipogenic program was identified as a key downstream effector of MAPK activation, opposed by the PML tumor suppressor, a potent failsafe mechanism that is genetically evaded in human cancer, and overcome by HFD in vivo , with ensuing implications in the prevention and treatment of metastatic CaP with stringent dietary regimens in combination with targeting of lipogenic enzymes.

PTEN

PTEN is among the most frequently lost or mutated tumor suppressor genes in human cancer^7,8. Partial loss of PTEN occurs early and is present in up to 70% of localized CaP (LPC)^{9 11}, while complete loss of PTEN is linked to metastatic castration-resistant CaP (mCRPC)^{12 15}. PTEN inactivation facilitates aberrant activation of the phosphoinositide-3- kinase (PI3K)/AKT pathway¹⁶. Intriguingly, studies from Pten knockout mouse models show that complete inactivation of Pten alone in mouse prostate leads to indolent tumors with minimally invasive features after a long latency^{17 19}. We and others have previously shown that tumor suppressor genes (TSGs), induced upon Pten loss, serve as failsafe mechanisms to restrict cancer progression and that inactivation of such barriers, such as Trp53 and Smad4 , promotes indolent Pten- null tumorigenesis^{20 21}. Therefore, identifying an additional event cooperating with PTEN loss to drive metastatic progression is important to model and study the lethal stage of CaP.

The Ras/Raf/MEK/ERK mitogen-activated protein kinase (MAPK) signaling cascade is another pathway that is often aberrantly activated in advanced metastatic CaP^15,22,23. However, little is known about the underlying molecular mechanisms leading to MAPK activation, since genetic alterations in MAPK signaling components are rare in human CaP^{14,24 29}. Despite the low frequency of Ras/Raf mutations in CaP, recent studies have shown that oncogenic K-ras^G12D or B-raf^/600E mutation, as a means to activate MAPK, cooperates with Pten loss to drive metastatic progression of CaP in GEMMs^30,31. These findings underscore the critical role of MAPK in cancer progression, but do not address how MAPK signaling is activated and what the key cellular events are in the majority of metastatic CaP.

SREBP Pro-Metastatic Lipogenic Signature

In one embodiment, the invention defines an up-regulated gene set activated by the SREBP family of transcription factors, the master regulator of fatty acid (FA) and cholesterol biosynthetic gene transcription. In particular embodiments, this lipogenic signature is present in Ptenpc-i-Pmlpc-i- prostates displayed higher level of precursor and nuclear SREBP- 1 and/or nuclear SREBP-2. In one embodiment, the lipogenic signature comprises increased lipid classes and fatty acyl chains, including but not limited to phosphatidylcholine (PC), sphingomyelin (SM), phosphatidylethanolamine (PE), triglyceride (TG),

lysophosphatidylcholine (LPC) and oleate (08: 1), palmitate (06:0), linoleate (18:2), arachidonate (20:4) and stearate (08:0), and/or increases in

lysodimethylphosphatidylethanolamine (LdMePE, 3 lipid ions), monoglyceride (MG, 7 lipid ions), phosphatidylglycerol (PG, 90 lipid ions) and lysophosphatidylglycerol (LPG). In other embodiments, this lipogenic signature is characterized in combination with dietary

assessment for a high fat diet. Methods for characterizing lipids, and fatty acyl changes are known in the art and described herein.

Types of biological samples

The invention provides compositions and methods for characterizing biological samples of a subject, including a subject having prostate cancer. In one embodiment, the biologic sample is a tissue sample that includes cells of a tissue or organ (e.g., prostatic tissue cells). Prostatic tissue is obtained, for example, from a biopsy of the prostate. In another embodiment, the biologic sample is a biologic fluid sample. Biological fluid samples include blood, blood serum, plasma, urine, seminal fluids, and ejaculate, or any other biological fluid useful in the methods of the invention. In another embodiment, the biological sample is biofluids, biopsy samples, and extracellular vesicles.

In one embodiment, a biological sample (e.g., prostate cancer biopsy) of a subject is characterized for co-deletion or mutation of PML and PTEN. In another embodiment, a lipogenic profile is characterized in a biological sample (e.g., prostate cancer biopsy). In yet another embodiment, activated SREBP pathway is characterized.

The lipogenic profile identified herein (e.g., increased phosphatidylcholine (PC), sphingomyelin (SM), phosphatidylethanolamine (PE), triglyceride (TG),

lysophosphatidylcholine (LPC) and oleate (08: 1), palmitate (06:0), linoleate (18:2), arachidonate (20:4) and stearate (08:0), increases in the abundance of

lysodimethylphosphatidylethanolamine (LdMePE, 3 lipid ions), monoglyceride (MG, 7 lipid ions), phosphatidylglycerol (PG, 90 lipid ions) and lysophosphatidylglycerol (LPG) and the genetic profile identified herein (e.g., Pml loss or mutation, TEN loss or mutation, activated SREBP pathway) can be measured in different types of biologic samples.

Diagnostic assays

The present invention provides a number of diagnostic assays that are useful for the characterization of prostate cancer and its propensity to metastasize. In particular, a lipid profile identified herein (e.g., increased phosphatidylcholine (PC), sphingomyelin (SM), phosphatidylethanolamine (PE), triglyceride (TG), lysophosphatidylcholine (LPC) and oleate (08: 1), palmitate (06:0), linoleate (18:2), arachidonate (20:4) and stearate (08:0), increases in the abundance of lysodimethylphosphatidylethanolamine (LdMePE, 3 lipid ions), monoglyceride (MG, 7 lipid ions), phosphatidylglycerol (PG, 90 lipid ions) and

lysophosphatidylglycerol (LPG) is characterized in a biological sample of a patient. In one embodiment, lipidomics detects less abundant lipid species at pmol level. With high-resolution untargeted liquid chromatography-tandem mass spectrometry (LC -MS/MS), the detecting sensitivity can reach to attomol level. Additionally, the cost to lipidomics is dramatically less than other biochemical assays, such as ELISA for PSA-based screening. In another embodiment, a genetic profile identified herein (e.g., Pml loss or mutation, PTEN loss or mutation, activated SREBP pathway) is characterized in a biological sample (e.g., prostate cancer sample). The characterization of these biomarkers may be used alone or in combination with other diagnostic assays. The Gleason scale is the most common scale used for grading prostate cancer. A pathologist will look at the two most poorly differentiated parts of the tumor and grade them. The Gleason score is the sum of the two grades, and so can range from two to 10. The higher the score is, the poorer the prognosis. Scores usually range between 4 and 7. The scores can be broken down into three general categories: (i) low-grade neoplasias (score < 4) are typically slow-growing and contain cells that are most similar to normal prostate cells;

intermediate grade neoplasias (4 < score <_7) are the most common and typically contain some cells that are similar to normal prostate cells as well as some more abnormal cells; high- grade neoplasias (8 < score < 10) contain cells that are most dissimilar to normal prostate cells. High-grade neoplasias are the most deadly because they are most aggressive and fast growing. High-grade neoplasias typically move rapidly into surrounding tissues, such as lymph nodes and bones.

Stage refers to the extent of a cancer. In prostate cancer, for example, one staging method divides the cancer into four categories, A, B, C, and D. Stage A describes a cancer that is only found by elevated PSA and biopsy, or at surgery for obstruction. It is not palpable on digital rectal exam (DRE). This stage is localized to the prostate. This type of cancer is usually curable, especially if it has a relatively low Gleason grade. Stage B refers to a cancer that can be felt on rectal examination and is limited to the prostate. Bone scans or CT/MRI scans are often used to determine this stage, particularly if prostate specific antigen (PSA) levels are significantly elevated or if the Gleason grade is 7 or greater. Many Stage B prostate cancers are curable. Stage C cancers have spread beyond the capsule of the prostate into local organs or tissues, but have not yet metastasized to other sites. This stage is determined by DRE, or CT/ MRI scans, and/or sonography. In Stage C a bone scan or a PROSTASCINT scan is negative. Some Stage C cancers are curable. Stage D cancer has metastasized to distant lymph nodes, bones or other sites. This is usually determined by bone scan, PROSTASCINT scan, or other studies. Stage D cancer is usually incurable, but may be treatable.

Selection of a treatment method

After a subject is diagnosed as having prostate cancer, a method of treatment is selected. Profiles that correlate with poor clinical outcomes, such as metastasis or death, are identified as aggressive prostate cancers. The genetic, lipid, or dietary profile of a subject identified as having prostate cancer is used in selecting a treatment method. In one embodiment, a subject having a pro-metastatic lipogenic signature is identified as having an aggressive prostate cancer. In one embodiment, less aggressive prostate cancers do not show loss of PML and do not have increased lipid levels or a lipogenic signature. In another embodiment, loss of PML, PTEN, and activation of SREBP regulated genes is used to select a treatment regimen, including treatment with an inhibitor of fatty acid synthesis (e.g.,

TOFA), inhibition of cholesterol synthesis (e.g., simvastatin), or inhibition of SREBP (e.g., Fatostatin). In another embodiment, subjects having a lipogenic signature are treated with a combination of agents that target AKT/mTOR and MAPK. Such therapies may be used alone or in combination with an aggressive treatment method. Aggressive therapeutic regimens typically include one or more of the following therapies: radical prostatectomy, radiation therapy (e.g., external beam and brachytherapy), hormone therapy, and

chemotherapy.

Patient monitoring

The diagnostic methods of the invention are also useful for monitoring the course of a prostate cancer in a patient or for assessing the efficacy of a therapeutic regimen. In one embodiment, the diagnostic methods of the invention are used periodically to monitor a subject’s pro-metastatic lipogenic signature. In one example, the prostate cancer is characterized using a diagnostic assay of the invention prior to administering therapy. This assay provides a baseline that describes the pro-metastatic lipogenic signature of the prostate cancer prior to treatment. Additional diagnostic assays are administered during the course of therapy to monitor the efficacy of a selected therapeutic regimen. A therapy is identified as efficacious when a diagnostic assay of the invention detects a decrease or normalization in lipid levels relative to baseline level of lipid.

Combination Therapies

MAPK activity can be suppressed/inhibited by small pharmacological inhibitors. The results provided herein indicate that patients with co-deletion of PTEN and PML or amplification of PPP1CA may benefit significantly from combinatorial therapy targeting both AKT/mTOR and MAPK signaling. In one embodiment an agent that inhibits AKT/mTOR is combined with an agent that inhibits MAPK signaling. Agents that inhibit AKT/mTOR include rapamycin, Temsirolimus, Everolimus, and Ridaforolimus. Agents that inhibit MAPK signaling include SB203580, SB202190, and BIRB-796. In one embodiment, an agent that inhibits MAPK and/or AKT/mTOR is used alone or in combination with an inhibitor of fatty acid synthesis (e.g., TOFA), inhibition of cholesterol synthesis (e.g., simvastatin), or inhibition of SREBP (e.g., Fatostatin). In another

embodiment, subjects having a lipogenic signature are treated with a combination of agents that target AKT/mTOR and MAPK and that inhibit fatty acid or cholesterol synthesis or that inhibit SREBP. If desired, such agents and combinations of agents of the invention are administered with any conventional anti -neoplastic therapy, including but not limited to, surgery, radiation therapy, or chemotherapy. Conventional chemotherapeutic agents include, but are not limited to, alemtuzumab, altretamine, aminoglutethimide, amsacrine, anastrozole, azacitidine, bleomycin, bicalutamide, busulfan, capecitabine, carboplatin, carmustine, celecoxib, chlorambucil, 2-chlorodeoxyadenosine, cisplatin, colchicine, cyclophosphamide, cytarabine, cytoxan, dacarbazine, dactinomycin, daunorubicin, docetaxel, doxorubicin, epirubicin, estramustine phosphate, etodolac, etoposide, exemestane, floxuridine, fludarabine, 5-fluorouracil, flutamide, formestane, gemcitabine, gentuzumab, goserelin,

hexamethylmelamine, hydroxyurea, hypericin, ifosfamide, imatinib, interferon, irinotecan, letrozole, leuporelin, lomustine, mechlorethamine, melphalen, mercaptopurine, 6- mercaptopurine, methotrexate, mitomycin, mitotane, mitoxantrone, nilutamide, nocodazole, paclitaxel, pentostatin, procarbazine, raltitrexed, rituximab, rofecoxib, streptozocin, tamoxifen, temozolomide, teniposide, 6-thioguanine, topotecan, toremofme, trastuzumab, vinblastine, vincristine, vindesine, and vinorelbine. In one preferred embodiment, an agent such as MEK inhibitor, EG0126, is administered in combination with other agents described herein (e.g., TOFA, simvastatin, Fatostatin).

Kits or Pharmaceutical Systems

The present compositions may be assembled into kits or pharmaceutical systems for use in ameliorating a prostate cancer. Kits or pharmaceutical systems according to this aspect of the invention comprise a carrier means, such as a box, carton, tube or the like, having in close confinement therein one or more container means, such as vials, tubes, ampoules, bottles and the like. The kits or pharmaceutical systems of the invention may also comprise associated instructions for using the agents of the invention. Kits of the invention include at least one or more agents that inhibit AKT/mTOR and MAPK signaling, agents that inhibit fatty acid synthesis (e.g., TOFA) or cholesterol synthesis (e.g., simvastatin), or an agent that inhibits SREBP (e.g., Fatostatin). In another embodiment, subjects having a lipogenic signature are treated with a combination of agents that target AKT/mTOR and MAPK.

Optionally, the kit includes instructions for administering the agents.

The following examples are offered by way of illustration, not by way of limitation. While specific examples have been provided, the above description is illustrative and not restrictive. Any one or more of the features of the previously described embodiments can be combined in any manner with one or more features of any other embodiments in the present invention. Furthermore, many variations of the invention will become apparent to those skilled in the art upon review of the specification. The scope of the invention should, therefore, be determined not with reference to the above description, but instead should be determined with reference to the appended claims along with their full scope of equivalents.

It should be appreciated that the invention should not be construed to be limited to the examples that are now described; rather, the invention should be construed to include any and all applications provided herein and all equivalent variations within the skill of the ordinary artisan.

The practice of the present invention employs, unless otherwise indicated, conventional techniques of molecular biology (including recombinant techniques), microbiology, cell biology, biochemistry and immunology, which are well within the purview of the skilled artisan. Such techniques are explained fully in the literature, such as, “Molecular Cloning: A Laboratory Manual”, second edition (Sambrook, 1989);

“Oligonucleotide Synthesis” (Gait, 1984);“Animal Cell Culture” (Freshney, 1987);

“Methods in Enzymology”“Handbook of Experimental Immunology” (Weir, 1996);“Gene Transfer Vectors for Mammalian Cells” (Miller and Calos, 1987);“Current Protocols in Molecular Biology” (Ausubel, 1987);“PCR: The Polymerase Chain Reaction”, (Mullis, 1994);“Current Protocols in Immunology” (Coligan, 1991). These techniques are applicable to the production of the polynucleotides and polypeptides of the invention, and, as such, may be considered in making and practicing the invention. Particularly useful techniques for particular embodiments will be discussed in the sections that follow.

The following examples are put forth so as to provide those of ordinary skill in the art with a complete disclosure and description of how to make and use the assay, screening, and therapeutic methods of the invention, and are not intended to limit the scope of what the inventors regard as their invention. EXAMPLES

Although localized CaP is highly curable, metastatic CaP remains invariably fatal¹. While it has been postulated that a Western diet can promote CaP progression^2,3, direct evidence supporting a strong association between dietary lipids and CaP is still lacking⁴. Indeed, the rates of cancer mortality associated with metastatic disease are much higher in Western countries for many cancer types, including CaP, correlating with lifestyle factors such as diet^2,3,5. Moreover, the progression to metastasis represents a pivotal event

influencing patient outcomes and the therapeutic options available to patients. Thus, understanding the molecular events that underlie progression to metastasis, at both genetic and environmental levels, has the potential to significantly improve therapeutic options for patients and facilitate preventative interventions. However, to date, CaP metastasis has proven to be particularly challenging to model in vivo , and the progression to metastasis from either a primary indolent or advanced stage disease is rarely observed for the majority of the genetically engineered mouse models (GEMMs) of CaP⁶.

PTEN is among the most frequently lost or mutated tumor suppressor genes in human cancer^7,8. Partial loss of PTEN occurs early and is present in up to 70% of localized CaP (LPC)^{9 11}, while complete loss of PTEN is linked to metastatic castration-resistant CaP (mCRPC)^{12 15}. PTEN inactivation facilitates aberrant activation of the phosphoinositide-3- kinase (PI3K)/AKT pathway¹⁶. Intriguingly, studies from Pten knockout mouse models show that complete inactivation of Pten alone in mouse prostate leads to indolent tumors with minimally invasive features after a long latency^{17 19}. We and others have previously shown that tumor suppressor genes (TSGs), induced upon Pten loss, serve as failsafe mechanisms to restrict cancer progression and that inactivation of such barriers, such as Trp53 and Smad4 , promotes indolent Pten- x\\ tumorigenesis^20,21. Therefore, identifying an additional event cooperating with PTEN loss to drive metastatic progression is crucial to be able to model and study the lethal stage of CaP.

Example 1

Concomitant loss of PTEN and PML expression in metastatic human CaP

To identify cooperating genomic alterations that may associate with PTEN loss in metastatic human CaP, a recent array-based comparative genomic hybridization (aCGH) dataset¹⁴ of 59 LPC and 35 mCRPC samples was used to evaluate the frequencies of co- deletion of PTEN with 58 high-confidence TSGs³². In this dataset, RΊΈN was lost in 14% of LPC and 66% of mCRPC (FIG. 1A). The frequencies of co-deletion of the other TSG with PTEN in metastatic disease ranged from 0 to 50% (FIG. 1-1 A). Among the top 25 TSGs co- del eted with PTEN, PML was lost in 31% of mCRPC, while remaining intact in LPC (FIG. IB), and was among several TSGs that were co-deleted with PTEN solely in metastatic disease (FIG. 1-1A). Concomitant deletion of P TEN and PML occurred in 20% of mCRPC and was significantly associated with metastatic disease (FIG. 1C), suggesting that PML might suppress metastatic CaP in coordination with PTEN. To validate these findings, a recent whole-exome sequencing dataset²⁵ of 150 mCRPC samples was analyzed. P TEN and PML were lost in 38% and 30% of mCRPC, respectively and co-deletion of P TEN and PML occurred in 11% of mCRPC (FIGs. 1-1B and 1-lC). In contrast to frequent homozygous PTEN deletion in mCRPC (FIG. 1A and FIG. 1-1B), homozygous PML deletion was less common than hemizygous PML deletion, which was frequently observed and focal in many cases (FIG. IB and FIG. 1-ID)

We next evaluated PTEN and PML protein expression in human CaP and sought to identify if PTEN and PML loss could represent cooperative predictors of overall survival (OS) after prostatectomy. Tissue microarray (TMA) analysis was performed in prostatectomy specimens from 144 men with primary CaP (Table 1). Loss of PTEN or/and PML significantly correlated with disease progression (FIG. ID and IE and FIGs. 1-1E and 1- 1F). Complete loss of PTEN and PML at the protein level occurred in 15% of the high-grade, but not in the low-grade CaP (FIG. IF). We then applied Cox proportional hazards models to identify risk factors for predicting OS after prostatectomy. Univariate analysis showed that PTEN and PML loss, Gleason grade (>7 us <7), and pathologic stage (pT3-4 or Nl us pTl-2 NO) were associated with adverse prognosis (FIG. 1G). Importantly, on multivariable analysis, PTEN and PML loss was confirmed as a statistically significant independent prognostic factor for OS (FIG 1G), and PTEN and PML loss displayed the most statistically significant power to stratify patient survival time (FIG 1H and FIGs. 1-1H to 1-1 J). These results further support a model in which loss of PML and PTEN might cooperate to promote advanced CaP.

Table 1: siRNA Sequences

Pml loss promotes metastatic progression in Pten- null CaP

The tumor suppressive function of the PML gene alone or in cooperation with PTEN has, to date, been studied in the context of primary CaP initiation and progression³³.

However, it was postulated that generating conditional Pten and Pml compound inactivation would allow us to conduct a long-term follow-up and model a continuum of lesions beyond minimally invasive Pten- null cancer and to determine whether Pten/Pml co-deletion would favor metastatic progression. We generated conditional Pml flox/flox mice (FIGs. 1-1K to 1- 1M) and crossed them with mice carrying Pb-Cre4 and floxed Pten alleles to obtain prostate epithelium-specific Pten/Pml double null mice (PterP^{z l} PmP^{z l '}).

Immunohistochemistry (IHC) assays revealed that Pml protein levels were markedly increased in Pten^pc prostate intraepithelial neoplasia (PIN) compared to wild type (WT) prostate epithelium (FIG. 2A and FIG. 2-1A), suggesting that Pml might serve as a potential failsafe mechanism to restrict cancer progression in Pten- null indolent tumors, while Pml protein was almost undetectable in Pten^{pc h}PmP^c PIN, confirming prostate epithelium- specific inactivation of Pml (FIG. 2A and FIG. 2-1A).

We next evaluated the consequences of Pml loss for Pten-nuW prostate tumorigenesis in mice. Histopathological analyses of prostate tumors were performed in cohorts of Pten^pc and Pten^-PmP⁰-¹- mice at 12 to 40 weeks of age. At 12 weeks of age, Pten^pc and Pten^pc PmP^c mice displayed low to high-grade PIN with higher penetrance of high-grade PIN in Pten^pc PmP^c mice (FIG. 2B). Consistent with previous reports²⁰, Pten^pc mice developed invasive prostate adenocarcinoma after a 28 to 40-week latency and, in sharp contrast, Pten^pc PmP^c mice developed invasive prostate adenocarcinoma as early as 20 weeks with higher penetrance at three time points examined (FIG. 2B and FIG. 2-1B). Furthermore, Pten^pc PmP^c tumors progressed to aggressive, poorly differentiated adenocarcinoma with little or no glandular structure along with focal features of sarcomatoid carcinoma with high-grade pleomorphic spindle cells at 52-60 weeks, whereas age-matched Pten^pc tumors remained indolent (FIG. 2C and FIG. 2-1C). Culture and serial passage of prostate spheres derived from WT, Pten^pc and Pter ^PmP^ mice revealed that

Pten^-PmP⁰^- prostate epithelial cells had significantly enhanced stem/progenitor self renewal capacity and growth compared to both WT and Pten^pc spheres (FIG. 2D). These differences resulted in drastically impaired survival of the Pten^pc PmP^c mice compared to Pten^pc mice. All Pten^pz mice survived beyond 18 months, in contrast, Pter ^PmP^ mice succumbed to localized disease, probably due to bladder obstruction and renal failure, or were euthanized due to extensive tumor burden from 13 months of age. While, 21% (4 out of 19) of PmP^{z h} mice exhibited low-grade PIN in ventral (VP) and dorsolateral (DLP) prostates after 12 months of age, but none died of the disease (FIG. 2E and FIGs. 2-ID and 2-1E). These results support the notion that inactivation of PML can favor both CaP initiation and progression.

Critically, Pml loss can promote metastasis in Pten- null CaP. While no Pten^pz mouse developed distant metastasis in an 18-month follow-up²⁰, 30% of Pten^^PmP⁰^ mice developed lumbar lymph node metastasis at 13-15 months of age (FIG. 2F). Histological and molecular pathological analysis revealed that these metastases resembled primary prostate tumors and showed high levels of phosphorylated AKT as well as nuclear AR and CK8 staining (FIG. 2G and FIG. 2-1F).

PML loss reactivates MAPK signaling in TE/V-null cells

To determine how Pml inactivation might impact Pten- null tumorigenesis to favor a pro-metastatic switch, the expression of CaP relevant metastatic pathways^21,30,31’³⁴ was analysed. Consistent with previous findings³³, compared to Pten^pc prostates,

Pten^-PmP⁰^- prostates displayed up to a 1.5-fold increase in the level of Akt

phosphorylation, and in stark contrast, a 19- to 35-fold increase in the level of Erk

phosphorylation (FIG. 3a). In line with the notion that PTEN loss/ AKT-mTOR activation leads to a feedback inhibition on MAPK pathway^35,36, Pten^pc prostates displayed even lower level of Erk phosphorylation than WT (FIG. 3A), suggesting that Pml loss triggers relief of feedback inhibition and consequent reactivation of the MAPK signaling in Pten- null prostates. IHC assays confir ed that Pt t ^^PmP^^1- prostates showed similar widespread membrane staining for phosphor- Akt as Pten^pc prostates, along with a rare and scattered nuclear phosphor-Akt staining, while a markedly increased phosphor-Erk compared to Pten^pc prostates (FIG. 3B and FIG. 3-1A). In contrast, negligible changes were observed in other pathways that are involved in metastasis^21,34 (FIG. 3-1B). Consistent with the notion that PML is haploinsufficient for some of its tumor suppressive functions³⁷, Pten^pc^Pml^pc^ prostates also displayed elevated levels of Erk phosphorylation (FIG. 3-1C). Collectively, the magnitudes of change in MAPK signaling strongly implicate Pml- loss induced MAPK reactivation in metastatic progression of Pten^{pc h}PmP^c tumors.

We then investigated the effects of knockdown of PML on MAPK activation via small interfering RNA (siRNA) in PTEN- A\ CaP cell lines, LNCaP and PC3. Knockdown of PML consistently resulted in MAPK activation (FIGS. 3C and 3D). Moreover, the degradation of PML protein triggered by arsenic trioxide also led to upregulation of EGF- induced ERK phosphorylation (FIGs. 3E and 3F), lending further support to MAPK suppression by PML.

An SREBP-dependent lipogenic program is hyperactivated in Pten/Pml double null CaP

To gain insight into what effectors lie downstream of this aberrant signaling to drive metastatic progression in Pten^^PmP⁰^ tumors, a transcriptome analysis was performed using microarray from 12-week-old WT, Pten^pz and Pten^pz PmP^z prostates (n=3) (FIG. 4-1A). We identified 101 genes whose expression was significantly upregulated and 158 genes whose expression was significantly down-regulated at least 1.5-fold (P<0.0l) in PtetP^-PmP⁰^- prostates compared to PterP^z prostates. Gene ontology analysis revealed that the three most significantly enriched gene-categories in the PterP^{z h}PmP^z signature were“cell migration” (7 X0017),“cell motility” (7 X0014) and“lipid metabolic process” (/^J=0.007) (FIG. 4A). While enrichment of cell migration and cell motility genes is consistent with the increased metastatic potential of PterP^{z h}PmP^z tumors, the lipid metabolic process signature is surprising, as activation of lipogenesis is an early event in the development of CaP and the potential role of aberrant lipogenesis in metastasis in vivo remains largely elusive^38,39.

Additional unbiased gene-set enrichment analysis (GSEA) showed that

Pten^-PmP⁰^- prostates displayed a significantly up-regulated gene set activated by the SREBP family of transcription factors, the master regulator of fatty acid (FA) and cholesterol biosynthetic gene transcription⁴⁰ (FIG. 4B, the right panel; FIG. 4-1B), confirming gene ontology analysis (FIG. 4A). Compared to WT, lipid metabolic process and SREBP signature were also mildly upregulated in PterP^z prostates (FIG. 4A, grey column, <0.00l7 and FIG. 4B, the left panel), consistent with the finding that activation of

AKT/mTOR/S6K signaling promotes lipogenesis by activating SREBP^41,42, but were more extensively enhanced in Pter ^PmP^ prostates (FIG. 4A, orange column, <0.00005 and FIG. 4B, the middle panel). qPCR analysis confirmed SREBP target genes were further upregulated in PterP^z PmP^z prostates compared to PterP^z prostates (FIG. 4C) and western blot assays demonstrated the increased protein expression of Mel, Elovl6, Hmgcs and Idil, four key lipogenesis/cholesterol biosynthesis enzymes, in Pten^^PmP⁰^ prostate lysates (FIG. 4D). The SREBP family is comprised of two genes: SREBP-l and SREBP-2. Pten^-PmP⁰^- prostates displayed higher level of precursor and nuclear SREBP-l and, to a lesser extent, nuclear SREBP-2 (FIG. 4D). In contrast, no significant differences in the gene signatures of other lipogenic transcription factors were observed (FIGS. 4-1C to 4-1F). Therefore, the SREBP pathway is hyperactivated and appears to be the main transcriptional factor responsible for the changes in lipid metabolism in Pten!Pml double null CaP.

Lipidomic analysis confirms that de novo lipogenesis positively correlates with the aggressiveness and metastatic potential of CaP

To confirm the aberrant alterations of de novo lipogenesis in Pter ^PmP^ tumors, a global lipidomic analysis was performed in l2-week-old WT, Pten^pc and Pten^{pc l} PmP^{c l} prostates (n=3) using untargeted high-resolution liquid chromatography-tandem mass spectrometry (LC-MS/MS)^{43 45}. These analyses identified 1,743 lipid ions in mouse prostate, which belong to 35 classes of lipids. Within all the identified lipid ions, 127 distinct fatty acyl chains, varying in a chain length from 4 to 38 carbons and double bonds from 0 to 7, were found. The most abundant lipid classes and fatty acyl chains in mouse prostate were phosphatidylcholine (PC), sphingomyelin (SM), phosphatidylethanolamine (PE), triglyceride (TG), lysophosphatidylcholine (LPC) and oleate (Cl8: l), palmitate (Cl6:0), linoleate (18:2), arachidonate (20:4) and stearate (Cl8:0), respectively (FIG. 4-1G and 4-1H).

We next performed unsupervised hierarchical clustering to identify the lipid ions with significant changes in abundance among WT, Pten^pc and Pten^pc PmP^c prostates.

Among the top 70 most significantly differential lipid ions, consistent with the results of the transcriptome analysis in lipogenic signatures, compared to WT, the abundance of the majority of lipid ions was mildly increased in Pten^pc prostates. The abundance was more extensively enhanced in Pten^{pc l} PmP^{c l} prostates (FIG. 4-11). To further confirm the alterations in lipid abundance among the three genotypes of mice, a scatterplot was used aiming to illustrate the estimated number of down- or up-regulated lipid ions, comparing two of the three genotypes of mice at a time^44,45. By doing so, it allowed us to reveal the overall change in relative lipid abundance of all identified 1,743 lipid ions. The scatterplot plots the log2 (MS1 Peak Area ratio between Pten^{pc /} PmP^{c /} and WT) versus the log2 (MS1 Peak Area ratio between Pten^pc and WT) on the y and x axes, respectively. A diagonal line (y=x) was also added to facilitate the comparison between Pten^^PmP⁰^ and Pten^pc prostates (FIG. 4E). Overall, 1,131 out of 1,743 lipid ions had x-coordinate value greater than zero, while 1,261 out of 1,743 lipid ions had y-coordinate value greater than zero, indicating a prominent up-regulation of lipid abundance in both Pten^pc and PterP^{c l} PmP^{<: l} prostates compared to WT. Importantly, 1,154 lipid ions were above the diagonal line, revealing a further augmentation of lipid abundance in Pten^{pc l} PmP^{c l} prostates compared to Pten^pc prostates (FIG. 4E). These data are consistent with the transcriptome analysis showing that aberrantly increased lipogenesis correlates with the aggressiveness and metastatic potential of CaP. Distinctive alterations in lipid species and lipid saturation in Pten/Pml double null CaP

We subsequently examined the qualitative changes in lipid profiles. 25 out of 35 lipid classes identified had increased abundance in PterP^c PmP^{c h} prostates compared to

Pten^pc prostates. Among them, the increases in the abundance of

lysodimethylphosphatidylethanolamine (LdMePE), monoglyceride (MG),

phosphatidylglycerol (PG) and lysophosphatidylglycerol (LPG) were statistically significant (FIG. 4F and FIG. 1-4G). LdMePE, PG and LPG are membrane phospholipids. Among them, PG, present at a level of 1-2% in most animal tissues, and together with its hydrolytic product, LPG, can serve as the precursor of cardiolipin (CL) found in mitochondrial membranes. PG lipids have recently been found to be the most significantly upregulated lipid species in MFC-driven cancer^{46 48}, suggesting a possible link between PG lipids and tumorigenesis. Additionally, MG lipids, a class of glycerolipids, can be hydrolyzed by monoacylglycerol lipase (MAGL) and serve as the precursors to synthesize pro-metastatic lipid messengers, including lysophosphatidic acid (LPA) and prostaglandin (PGE₂)⁴⁹.

Increased levels of saturated fatty acyl chains have been reported to protect cancer cells against oxidative damage by reducing lipid peroxidation⁵⁰. We therefore compared the levels of the 30 most abundant fatty acyl chains to assess the alterations in lipid saturation among the three genotypes of mice (FIG. 4-1H). A total of 7 fatty acyl chains had significantly increased abundance in PtetP^PmP⁰^- prostates compared to Pten^pc prostates (FIG. 4G and FIG. 4-1H). Only one was polyunsaturated FA (22:6), and the other six were saturated and monounsaturated FAs (14:0, 16:0, 18:0, 16: 1, 18: 1, 24: 1), indicating that hyperactivation of lipid metabolism in Pten^^PmP⁰^ prostates leads to higher levels of saturated and monounsaturated fatty acyl chains. These findings are consistent with earlier studies showing that increased levels of saturated and monounsaturated FAs are associated with aggressive breast cancer⁵¹.

SREBP is the downstream target of L-loss induced MAPK activation

Since the level and transcriptional activity of SREBP is known to be regulated by growth factor induced phosphorylation^52,53, SREBP was tested as a target of MAPK activation, induced by PMLA oss. PTEN-mx\\ CaP cells were cultured in media supplemented with 10% lipoprotein deficient serum to limit the availability of exogenous lipids and increase endogenous synthesis. qPCR analysis showed that treatment of LNCaP cells with MEK inhibitor, U0126, resulted in decreased basal expression of all SREBP targets examined (FIG. 5A). Knockdown of PML led to upregulation of SREBP target genes. EG0126 could reverse PML -depletion induced upregulation of SREBP target genes (FIG. 5A). Similar results were also obtained in PC3 cells (FIG. 5B). Cell fractionation experiments showed that EG0126 strongly downregulated both nuclear SREBP-l and SREBP-2 protein levels along with a large reduction in SREBP target gene expression (FIGs. 5C and 5D). In contrast, MAPK activation through either knockdown of PML or a constitutively active MEK led to higher levels of nuclear SREBP proteins as well as SREBP target gene expression. These increases were largely suppressed by EG0126 (FIGs. 5C and 5D). We, and others, have previously shown that MAPK signaling inhibits the tumor suppressor function of the TSC complex to activate mTOR/S6K^54,55. Consistently, the levels of both p-ERK and p-S6K phosphorylation were found to mirror the changes in SREBP and their target genes. As such, SREBP cannot be excluded as an indirect target of MAPK signaling through mTOR/S6K, but it was concluded that P L-loss induced MAPK activation is sufficient to drive an SREBP- dependent lipogenic program.

An SREBP-dependent lipogenesis is critical for PML-loss induced CaP growth and metastasis

We then assessed whether PML-loss induced pro-metastatic phenotype is dependent on SREBP-mediated lipogenic program. In an in vitro cell migration and invasion assay, knockdown of either SREBP-l or SREBP-2 by siRNA significantly inhibited LNCaP and PC3 cell migration and invasion (FIG. 6A and FIGS. 6-1A and 6-1B). Inhibition of fatty acid synthesis by TOFA (acetyl-CoA-carboxylase inhibitor) or cholesterol synthesis by simvastatin (HMG-CoA reductase inhibitor) mimicked the effect of SREBP knockdown on cell migration and invasion (FIG. 6B). In contrast, CaP cells transfected with PML siRNA showed significantly increased migration and invasion. This effect was nearly abolished when cells were co-transfected with SREBP-l siRNA (FIG. 6A and FIG. 6-1A).

To further corroborate the crucial role of SREBP-dependent lipogenesis in PML-loss driven CaP growth and metastasis, in vivo preclinical studies of fatostatin targeting SREBP in P/e«^pc_/_Pw/^pc_/_ mice were carried out. Fatostatin is a recently discovered SREBP chemical inhibitor that directly binds SREBP cleavage activating protein and blocks the ER to Golgi transport of SREBP and its subsequent activation^{56 58}. Treatment of fatostatin for two months in P/e«^pc_/_PwP^c_/_ mice inhibited both prostate tumor growth (FIGs. 6C and 6D) and distant lymph node metastasis (FIG. 6E and FIG. 6-1C). This potent antitumor and anti -metastatic activity of fatostatin is presumably due to the suppression of SREBP pathway, since fatostatin-treated PterP^z PmP^z tumors displayed markedly lower expression of SREBP- regulated enzymes for fatty acid synthesis and for cholesterol synthesis (FIGs. 6F and 6G). Consistent with the vital role of lipid metabolism in various aspects of cancer development³⁹, fatostatin-treated PterP^z PmP^z tumors displayed a drastic decrease in the frequency of mitotic cells positive for Ki67 staining, along with a concomitant induction of apoptosis as indicated by higher cleaved Parp expression and cleaved Caspase 3 staining (FIGs. 6F and 6G). These functional data suggest that SREBP-mediated lipogenesis is a key downstream effector of PML- loss driven CaP growth and metastasis.

A high fat diet (HFD) drives metastatic progression and increases lipid abundance in prostate tumors

Given that the HFD feeding stimulates the expression of SREBP and subsequent lipogenic gene expression⁵⁹ and induces lipid accumulation in non-adipose tissues^40,60, 12- month-old PterP^c and PterP^{c l} PmP^{c l} mice were fed a HFD for 3 months to test whether a causal relationship exists between lipogenesis, lipid accumulation and metastasis. To mimic hyperactivated de novo lipogenesis (new lipids enriched in saturated and monounsaturated fatty acyl chains), lard-based HFD was chosen that was enriched in saturated and

monounsaturated FAs and capable of inducing the classic HFD effect in rodents⁶¹. As expected, HFD-fed mice gained considerably more weight (FIG. 7A). Pten^vz PmP^z mice, which displayed limited metastasis to lymph node when fed chow, now developed lymph node metastases in 6 out of 8 cases, and lung metastasis in 4 out of 8 cases upon HFD feeding (FIGs. 7B and 7C and FIG. 7-1A). PterP^{z h} mice, non-metastatic when fed chow²⁰, developed lymph node metastases in 3 out of 8 cases, and lung metastases in 3 out of 8 cases upon HFD feeding (FIGs. 7D and 7E and FIG. 7-1B). These metastases resembled primary prostate tumors and displayed high levels of phosphorylated ART as well as nuclear AR staining. Strong CK8 staining was also observed in lymph node metastases, but less so in lung metastases (FIGs. 7B and 7D and FIG. 7-1B). No other distant metastases or significant impact on the survival were observed in HFD-fed Pten^pz ox PterP^PmP^ mice (FIG. 7-1C).

To further establish the causal relationship between dietary lipids and metastasis, Oil Red O (ORO) staining was conducted to determine whether lipids are accumulated in prostate tumors following HFD feeding. Compared to tumors from chow-fed mice, tumors from HFD-fed PterP^z or PterP^PmP^ mice showed much stronger staining for ORO (FIG. 7F). To validate the results of the ORO staining, an additional global lipidomic analysis was performed in prostate tumors from chow- or HFD-fed Pten^pc and

Pter ^-PmP^- mice. Analyses of the top 70 most significantly differential lipid ions (FIG. 7G), significantly increased lipid classes (FIG. 7H and FIG. 7-1C), and significantly increased fatty acyl chains from the 30 most abundant chain types (FIG. 71 and FIG. 7-1E) revealed that the abundance of the majority of lipids was markedly higher in HFD-fed Pten^pc or Pter ^^PmP^0-1- tumors compared to respective chow-fed tumors. Among the four groups of tumors, while Pten^pc tumors had the lowest lipid abundance when fed chow, in response to HFD feeding, Pten^pc tumors displayed higher lipid abundance than chow-fed Pter ^-PmP^- tumors but a largely similar lipid profile as HFD-fed Pten^pc PmP^c tumors (FIGs. 7G to 71). These findings indicate that HFD feeding increases lipid abundance in prostate tumors to a much greater extent than de novo lipogenesis driven by Pml inactivation, and suggest that HFD affects metastatic progression, at least in part, through increased lipid accumulation in prostate tumors.

We then sought to determine whether dietary constituents of the HFD could affect CaP cell invasiveness recapitulating aspects of the HFD-induced metastatic phenotype. CaP cells were cultured in media supplemented with a lipid mixture, palmitic acid, or oleic acid, consisting of main components of lard-based HFD⁶¹, respectively. Consistent with these findings in HFD-fed prostate tumors, CaP cells treated with dietary lipids possessed markedly increased lipid droplet accumulation, compared to vehicle-treated cells, as shown by ORO staining (FIG. 7J and FIG. 7-1F). Importantly, they displayed significantly increased cell migration and invasion (FIG. 7K and FIG. 7-1 G), suggesting that dietary lipids can directly affect CaP cells and are sufficient to recapitulate aspects of the in vivo HFD-induced metastatic phenotypes.

A highly enriched SREBP signature in metastatic human CaP

To assess the relevance of these findings to human CaP, a GSEA analysis was performed of the metastasis-gene signatures in mCRPC us LPC from the Grasso el al.

dataset¹⁴. Our analysis revealed significant enrichment of SREBP- 1 target genes in mCRPC compared to LPC (FIG. 8A). Additionally, the Taylor et al. dataset¹⁵, which contains 181 primary and 37 metastatic CaP samples, was analyzed. Consistent with the above dataset, the SREBP-l signature is highly enriched in metastatic us primary CaP (FIG. 8B). Thus, the highly enriched SREBP/lipogenesis signature is frequently observed in metastatic human CaP, a finding consistent with our transcriptome data showing enrichment of the SREBP signature in Pten and Pml double null metastatic CaP (FIG. 4B) and our in vivo preclinical studies showing that lipid accumulation promotes metastatic progression (FIGs. 7B to 71).

Our data provide a strong genetic foundation to the mechanisms underlying metastatic progression, and demonstrate how environmental dietary factors can boost progression from primary to metastatic cancer, intertwining with the genetic makeup of cancer (FIG. 8C). We have demonstrated that SREBP-dependent lipogenesis, which can be hyperactivated by concomitant activation of the PI3K/AKT and MAPK pathways, or a HFD regimen, functions as an underlying rheostat towards metastatic cancer progression. Furthermore, PML was identified as a critical mediator of feedback inhibition of MAPK signaling and lipogenesis, driving metastatic progression in PTEN loss/PI3K-AKT driven cancers. Interestingly, Cph2, the yeast homologue of SREBP, has been shown to be involved in invasive/pseudohyphal growth in yeast, suggesting a conserved evolutionary function of SREBP in invasiveness⁶².

Consistent with earlier studies showing that lipid metabolism is required for metastasis in other cancer systems^63,64, our analysis reveals a highly enriched SREBP-l dependent lipogenic signature in metastatic human CaP. Concordantly, a HFD enriched in saturated and monounsaturated FAs triggers metastasis in a non-metastatic Pten^pc CaP model and further enhances metastasis in a metastatic Pter ^PmN^-1- CaP model. In addition to our genetic and experimental evidence in GEMMs of CaP, it has been reported that HFD favors metastatic progression in xenograft models from cell lines of various histological origin^63,65,66. Numerous mechanisms have been proposed to explain a possible association between dietary lipids and CaP⁶⁷, including paracrine mechanisms through secreted cytokines from adipose tissues, endocrine mechanisms through an alteration of androgen levels (FIG. 8-1 A), and an induction of basal -to-luminal cell differentiation caused by immune cell infiltration⁶⁸. However, specific genetic perturbations or HFD was determined to like be able to exert a direct effect on metastasis through increased lipid accumulation. The intracellular lipid changes in GEMMs of CaP was characterized and qualitative changes were detected in four different lipid classes as well as in the saturation of fatty acyl chains. Together, these results establish a strong mechanistic and causal link between aberrant lipogenesis, excess lipid accumulation and metastasis, providing a compelling rationale for integrating lifestyle data (e.g. diet) and tumor genetics into clinical practice to identify patients at high-risk of metastasis. Additionally, lipid metabolism itself is an attractive therapeutic target through inhibition of lipogenic enzymes^39,69. Notably, such inhibitors result in reduced CaP cell viability only in the absence of an exogenous lipid source such as lipoprotein⁷⁰, highlighting the importance of integrating pharmacologic approaches with stringent dietary regimens to prevent metastasis. Future studies are warranted to evaluate whether specific lipid subsets/signatures can serve as prognostic biomarkers to distinguish CaP with metastatic potential from indolent disease.

Lastly, given that PML is lost in human cancer of multiple histological origins⁷¹, our study suggests that PML loss may underlie MAPK activation in cancers lacking genetic alterations in MAPK signaling components^{26 28,72}. In an accompanying study, PPla was demonstrated to be a B-Raf activating phosphatase genetically amplified in CaP, and that PML can suppress PP la-dependent activation of MAPK signaling⁷³. Moreover,

PtetP^-PmP⁰^- tumors displayed a significantly up-regulated gene set induced by hypoxia, consistent with earlier studies showing that PML is a key player in orchestrating the cellular response to hypoxia through repression of mTOR⁷⁴ (FIG. 8-1B). This finding has equally important implications for turn ori genesis, because PML loss in the hypoxic core or tumoral lesions would not only activate mTOR, resulting in sustained HIF-l activation, but would also relieve the feedback inhibition of MAPK signaling triggered by mTOR activation, thus leading to simultaneous activation of both mTOR and MAPK signaling. Taken together, our study provides a roadmap for targeted therapies tailored to the individual patient for the prevention and treatment of metastatic cancer.

The following materials and methods were used to generate the data reported in Example 1.

Murine models. All animal experiments were approved by the Beth Israel Deaconess Medical Center IACUC Committee on Animal Research. Floxed Pml allele was created following the scheme shown in FIG. 1-1K. The Pb-Cre4 transgenic mice and Pten^nox/n°^x mice have been previously described²⁰. Pten^nox/n°^x mice were first crossed with Pb-Cre4 mice. The resulting compound mice or Pb-Cre4 transgenic mice were then crossed with Pm ⁰^ mice to generate conditional knockout of Pten and/or Pml in the prostate epithelium. The three genotypes of mice were maintained on a mixed C57BL/6 (80%) x l29Sl/SvImJ (20%) background. Ten mice per genotype were randomly chosen and used to examine the tumor grade at the indicated ages. The histological grade was determined blindly by the pathologist. HFD was achieved by feeding male mice a dietary chow consisting of 60% kcal fat (Teklad Diet TD.06414) beginning at the age of 12 months for 3 months. All other mice were fed standard chow consisting of 17% kcal fat (Lab Diet 5008). Male mice were housed one per cage during preclinical studies of both HFD and fatostatin.

Plasmids, reagents and antibodies. HA-MEKl^S218D/S222D was purchased from Upstate. Two independent siRNA duplexes targeted to PML, SREBP-l, SREBP-2, and control non-target siRNA were purchased from Dharmacon or Sigma. U0126 was from Selleck Chemicals. TOFA and Simvastatin were from Cayman Chemical. Fatostatin was from EMD Millipore. EGF, Lipofectamine 2000, Lipofectamine RNAiMAX, RPMI, DMEM, Opti-MEM reduced serum media and fetal bovine serum (FBS) were from Invitrogen. TransYT-Xl was from Mirus Bio LLC. Fatty acid free BSA, sodium palmitate, oleic acid, lipid mixture, Polybrene and puromycin were from Sigma. We used the following primary antibody: Anti-p-ERK, anti-ERK, anti-p-MEK, anti-MEK, anti-p-Akt, anti-Akt, anti-PTEN, anti-EZH2 and anti- GAPDH were from Cell Signaling Technology; Anti-Rb (IF8) and anti-HSP90 (H-l 14) were from Santa Cruz; Anti-IDIl (NBP1-57587) was from Novus Biologicals; Anti-SREBPl (2A4) was from Active Motif; Anti-P-actin (AC-74) was from Sigma; Anti-SREBP2 (A303- 125 A), anti -PML (A301-167A), Anti-FASN (A301-324A) and anti-HMGCSl (A304-590A) were from Bethyl Laboratories; Anti-PML (MAB3738) for detecting mouse Pml protein was from Millipore; Anti-AR (EPR1535) and anti-SMAD4 (EP618Y) were from Abeam; Anti- Ki67 (SP6) was from Thermo Fisher Scientific; Anti-CK8 (MMS-162P) was from Covance. Anti-a-SM-actin (1 A4) was from Dako; Anti-ELOVL6 (PA5-20520), Anti-HMGCR (PAS- 37367) and Anti-MEl (PA5-40600) were from Invitrogen.

Cell culture and transfection. All cell lines were obtained from ATCC and checked for mycoplasma by MycoAlert Mycoplasma Detection Kit (Lonza). Cells were maintained in DMEM or RPMI supplemented with 10% fetal bovine serum, 2 mM glutamine, 100 U/ml penicillin and streptomycin (Invitrogen). To study the effect of PML&MAPK signaling on SREBP and SREBP targets, LNCaP or PC3 cells were seeded and cultured in media with 10% lipoprotein deficient serum (Kalen Biomedical) throughout the course of the experiments. Transfections were performed using Lipofectamine 2000, Lipofectamine RNAiMAX reagent or TransTT-X2 according to manufacturer’s instruction. In brief, 50nM mixtures of two independent siRNA pairs targeting each gene or 1 pg of DNA plasmids were transfected into lxlO⁵ cells in a 6-well dish. Cells were recovered into completed media after l2-hr transfection and then harvested at the indicated times. For the treatment of dietary lipids, palmitic and oleic acid were conjugated to fatty acid-free BSA as described⁷⁵. CaP cells were cultured with BSA vehicle control, 2% lipid mixture, 30 pm BSA-conjugated palmitic acid or oleic acid for 7 days, then subjected to ORO staining, cell migration, and invasion assay.

Western blotting. For western blotting, prostate tissues or cells were lysed in RIPA buffer (Sigma) supplemented with protease (Roche) and phosphatase (Sigma) inhibitor. Proteins were separated on 4-12% Bis-Tris gradient gels (Invitrogen), transferred to polyvinylidine difluoride membranes (Immobilon P, Millipore) and the blots were probed with the indicated antibodies. Densitometry quantification was performed with ImageJ. Nuclear/cytoplasmic fractionation was performed as described⁷⁶.

Histology, IHC and ORO staining. Individual prostate lobes were dissected and fixed in 4% paraformaldehyde for histology and IHC analysis, or cryoembedded in OCT compound (Sakura) for ORO staining. For staining, the tissues were embedded in paraffin in according with standard procedures. 5 pm sections were cut and processed for histology or

immunostaining. The following primary antibodies were used for IHC: PML (MAB3738,

1 :300), p-ERK (20G11, 1 : 100), p-AKT (D9E, 1 : 100), CK8 (MMS-162P, 1 :200), AR

(EPR1535, 1 : 100), Ki67 (SP6, 1 :200), Cleaved Caspase-3 (9661, 1 :300), FASN (A301-324A, 1 : 100), HMGCR (PA5-37367, 1 : 100) and smooth muscle a actin (1 A4, 1 : 1000). For ORO staining, cells were prepared by drying drop on poly-l-lysine slides after 7-day treatment of dietary lipids. Frozen tissues or cells were stained by ORO working solution as described⁷⁷. The stained slides were visualized by a bright-field microscope.

Prostate sphere assay. The culture and passage of prostate spheres were carried out as described⁷⁸. Dissociated prostate epithelial cells were prepared from mice at 12 weeks of age. To initiate sphere formation, unsorted mouse prostate cells were prepared in PrEGM medium (Lonza) at a density of 2.5 x 10⁵ cells per ml. 40 pl of cell suspension was mixed with 60 pl cold Matrigel (BD Bioscience), and pipetted around the rim of a well of a 12-well plate and allowed to solidify at 37°C for 30 min. lml warm PrEGM was then added to each well. The spheres were cultured and monitored for 14 days with 50% medium change every 3 days. To passage spheres, Matrigel was digested by 1 mg/ml dispase solution (StemCell Technologic) for 30 minutes at 37°C. Digested cultures were collected, pelleted, resuspended and subjected to sequential digestion by 2mg/ml type I collagenase (Sigma) for 1 hr and 0.05%

Trypsin/EDTA (Invitrogen) for 5 min at 37°C, and then passed through a 27-gauge syringe 5-10 times, and filtered through a 40 pm cell strainers. Cells were counted by hemocytometer and replated at the density of lxlO⁴ cells per l2-well.

Microarray analysis. RNA was extracted from WT and knockout mice using QIAzol (Qiagen). Two hundred nanograms of total RNA was hybridized to Affymetrix Mouse Gene 2.1 ST arrays by the Beth Israel Deaconess Medical Center Genomics and Proteomics Core. The obtained raw intensity .cel files were normalized by robust multichip analysis

(Bioconductor release 3.1) and differential expression was determined using the limma Bioconductor package by fitting a linear model. Gene set enrichment analysis was conducted with the gene sets from the Molecular Signatures Database (MolSigDB v3. l). Gene Ontology analysis was conducted with the Panther Annotation System (version 9.0). qPCR. Total RNA was prepared using the TRIzol (Invitrogen). cDNA was obtained with the iScript cDNA Synthesis kit (Bio-Rad). Triplicate samples for qPCR were run in the

Lightcycler 480 (Roche) using the SYBR Green I Master (Roche). Each value was adjusted using the level of Sdha (for mouse genes) or RPLP0 (for human genes) as a reference.

Lipidomics by untargeted high-resolution liquid chromatography-tandem mass spectrometry (LC-MS/MS). The lipidomic analysis was performed as previously described⁴⁴. Briefly, non-polar lipids were extracted from 5 mg prostate tissues using MTBE. The upper phase containing the non-polar lipids was dried using a speedvac with no heat. Lipid samples were resuspended in 35 pl of 50% isopropanol (IPA)/50% MeOH. 10 mΐ of samples were injected for reversed-phase (Ci₈) LC-MS/MS using a hybrid QExactive Plus Orbitrap mass spectrometer (Thermo Fisher Scientific) coupled to an Agilent 1100 HPLC in DDA mode using positive/negative ion polarity switching (Top 8 in both modes). The lipidomics data were analyzed using LipidSearch 4.1.9 software. The software identifies intact lipid molecules based on their molecular weight and fragmentation pattern using an internal library of predicted fragment ions per lipid class and the spectra are then aligned based on retention time and MS1 peak areas are quantified across sample conditions. Excel 2010 was used to produce intensity and the R program (version 3.2.5) was used for data manipulation and statistical analyses, including unsupervised hierarchical clustering and heat map visualization.

Cell migration and invasion assay. PC3 or LNCaP cells were transfected with indicated siRNA or pretreated with 10 pg/ml TOFA or 10 mM Simvastatin for 48 hours, then detached into single-cell suspension. LNCaP (lxlO⁵) or PC3 (5xl0³ and 5xl0⁴ for migration and invasion assay, respectively) cells were resupsended in 100 pl of RPMI medium containing 0.1% FBS and placed into the top chamber of 8 pm transwell inserts for migration assay or Matrigel-coated transwell inserts for invasion assay (BD Biosciences). The bottom wells contained 600 pl RPMI supplemented with 10% FBS. After 24 hours or 48 hours (for LNCaP invasion assay only), cells on the upper surface of the inserts were removed with a cotton swab. Migrated cells were fixed in 10% formalin, then stained with 0.2% crystal violet for 10 minutes. Cells were counted in four microscopic fields under 20x magnification. Results are representative of three independent experiments.

In vivo treatment. Pter ^PmP⁰^ mice at 12-13 months were treated with fatostatin (15 mg/kg) or com oil (vehicle control) by intraperitoneal (i.p.) injections, every other day for two months. Mice were then euthanized and mouse tissues, including prostate tumors, were dissected, weighed and processed for histopathology and molecular analyses.

Blood sampling and testosterone ELISA. After 90 days of chow- or HFD-feeding, blood samples were collected from the mice through cardiac puncture into BD Vacutainer SST™ Serum Separation Tubes (BD 367986) and immediately mixed by proper inversion. The SST™ blood specimens were allowed to clot for 30 minutes and then centrifuged at lOOOg for 10 minutes in a swing bucket centrifuge to recover the serum. Serum levels of testosterone were measured using a competitive ELISA kit according to manufacturer’s instructions (Abeam 108666).

Array CGH analysis. We downloaded the data from GEO database (Grasso: GSE35988; Taylor: GSE21032) or cBioportal (for Robinson et al. dataset²⁵) with focus on the aCGH datasets. We wrote R scripts to process the data and generate the heatmap based on the log2- transformed ratio. The cutoff threshold used is -0.35 to -0.8 as heterozygous deletions, and those lower than -0.8 as homozygous deletions.

TMA analysis. All the prostate specimens were obtained upon informed consent and with approval from the Memorial Sloan-Kettering Cancer Center (MSKCC) ethic committee. Clinic-histopathological and follow-up information is included in the Table 1. The study cohort was comprised of radical prostatectomy specimens from 144 patients with primary CaP. Tumor samples were collected at the time of surgical resection with written informed consent. The patients were treated and followed at Memorial Sloan-Kettering Cancer Center. PML (Santa Cruz) and PTEN (Cell Signaling Technology) staining were performed as previously described⁷¹. Cases that had more than 50% of the core composed of tumor cells were analyzed.

Statistical analysis. No statistics was applied to determine sample size. The studies involved mice were randomized. The investigators were not blinded to allocation during experiments and outcome assessment. For analysis of average data, datasets were compared using unpaired two-tailed Student’s t tests. For analysis of categorical data (for example, copy number alteration), 2X2 contingency tables were constructed, and datasets were compared using Fisher’s exact test, For the correlation of TMA staining with clinical parameters, datasets were compared using Pearson’s chi-squared test. Survival outcomes were evaluated using Kaplan-Meier survivor estimates, Log-rank (Mantel-Cox) test and univariate and multivariable Cox-proportional hazards models. All tests were two-sided and an a-error of 5% was considered as significant. Univariate exploratory analyses showed that grouping PTEN and PML loss (defined as lack of both marker or lack of one marker and low expression of the other), Gleason score (< 7 vs >7), and pathologic stage (pTl-2 NO vs pT3-4 or Nl) maximized the likelihood ratio chi-squared for overall survival; these groupings were used in the multivariable model. P values of <0.05 were considered to be statistically significant. Statistical tests were executed using the statistical software R (version 3.1.2) or GraphPad Prism software.

Example 1: References

1. Wu, J.N., Fish, K.M., Evans, C.P., Devere White, R.W. & Dall'Era, M. A. No

improvement noted in overall or cause-specific survival for men presenting with metastatic prostate cancer over a 20-year period. Cancer 120, 818-23 (2014). 2. Yang, M. et al. Dietary patterns after prostate cancer diagnosis in relation to disease- specific and total mortality. Cancer Prev Res (Phila) 8, 545-51 (2015).

3. Gronberg, H. Prostate cancer epidemiology. Lancet 361, 859-64 (2003).

4. Markozannes, G. et al. Diet, body size, physical activity and risk of prostate cancer:

An umbrella review of the evidence. Eur J Cancer 69, 61-69 (2016).

5. Zhu, Y. et al. Dietary patterns and colorectal cancer recurrence and survival: a cohort study. BMJ Open 3(2013).

6. Ittmann, M. et al. Animal models of human prostate cancer: the consensus report of the New York meeting of the Mouse Models of Human Cancers Consortium Prostate Pathology Committee. Cancer Res 73, 2718-36 (2013).

7. Steck, P.A. et al. Identification of a candidate tumour suppressor gene, MMAC1, at chromosome l0q23.3 that is mutated in multiple advanced cancers. Nat Genet 15, 356-62 (1997).

8. Li, J. et al. PTEN, a putative protein tyrosine phosphatase gene mutated in human brain, breast, and prostate cancer. Science 275, 1943-7 (1997).

9. Yoshimoto, M. et al. Interphase FISH analysis of PTEN in histologic sections shows genomic deletions in 68% of primary prostate cancer and 23% of high-grade prostatic intra-epithelial neoplasias. Cancer Genet Cytogenet 169, 128-37 (2006).

10. Sircar, K. et al. PTEN genomic deletion is associated with p-Akt and AR signalling in poorer outcome, hormone refractory prostate cancer. J Pathol 218, 505-13 (2009).

11. Han, B. et al. Fluorescence in situ hybridization study shows association of PTEN deletion with ERG rearrangement during prostate cancer progression. Mod Pathol 22, 1083-93 (2009).

12. Lunardi, A. et al. A co-clinical approach identifies mechanisms and potential

therapies for androgen deprivation resistance in prostate cancer. Nat Genet 45, 747-55 (2013).

13. Beltran, H. et al. Targeted next-generation sequencing of advanced prostate cancer identifies potential therapeutic targets and disease heterogeneity. Eur Urol 63, 920-6 (2013).

14. Grasso, C.S. et al. The mutational landscape of lethal castration-resistant prostate cancer. Nature 487, 239-43 (2012).

15. Taylor, B.S. et al. Integrative genomic profiling of human prostate cancer. Cancer Cell 18, 11-22 (2010).

16. Song, M.S., Salmena, L. & Pandolfi, P.P. The functions and regulation of the PTEN tumour suppressor. Nat Rev Mol Cell Biol 13, 283-96 (2012).

17. Ma, X. et al. Targeted biallelic inactivation of Pten in the mouse prostate leads to prostate cancer accompanied by increased epithelial cell proliferation but not by reduced apoptosis. Cancer Res 65, 5730-9 (2005).

18. Trotman, L.C. et al. Pten dose dictates cancer progression in the prostate. PLoS Biol 1, E59 (2003).

19. Svensson, R.LT. et al. Slow disease progression in a C57BL/6 pten-deficient mouse model of prostate cancer. Am J Pathol 179, 502-12 (2011).

20. Chen, Z. et al. Crucial role of p53-dependent cellular senescence in suppression of Pten-deficient tumorigenesis. Nature 436, 725-30 (2005).

21. Ding, Z. et al. SMAD4-dependent barrier constrains prostate cancer growth and

metastatic progression. Nature 470, 269-73 (2011).

22. Gao, H. et al. Combinatorial activities of Akt and B-Raf/Erk signaling in a mouse model of androgen-independent prostate cancer. Proc Natl Acad Sci USA 103,

14477-82 (2006). 23. Kinkade, C.W. et al. Targeting AKT/mTOR and ERK MAPK signaling inhibits hormone-refractory prostate cancer in a preclinical mouse model. J Clin Invest 118, 3051-64 (2008).

24. Wang, X.S. et al. Characterization of KRAS rearrangements in metastatic prostate cancer. Cancer Discov 1, 35-43 (2011).

25. Robinson, D. et al. Integrative clinical genomics of advanced prostate cancer. Cell

161, 1215-28 (2015).

26. Palanisamy, N. et al. Rearrangements of the RAF kinase pathway in prostate cancer, gastric cancer and melanoma. Nat Med 16, 793-8 (2010).

27. Moul, J.W., Friedrichs, P.A., Lance, R.S., Theune, S.M. & Chang, E.H. Infrequent RAS oncogene mutations in human prostate cancer. Prostate 20, 327-38 (1992).

28. Gumerlock, P.H., Poonamallee, ET.R., Meyers, F.J. & deVere White, R.W. Activated ras alleles in human carcinoma of the prostate are rare. Cancer Res 51, 1632-7 (1991).

29. Barbieri, C.E. et al. Exome sequencing identifies recurrent SPOP, FOXA1 and

MED12 mutations in prostate cancer. Nat Genet 44, 685-9 (2012).

30. Wang, J. et al. B-Raf activation cooperates with PTEN loss to drive c-Myc expression in advanced prostate cancer. Cancer Res 72, 4765-76 (2012).

31. Mulholland, D. J. et al. Pten loss and RAS/MAPK activation cooperate to promote EMT and metastasis initiated from prostate cancer stem/progenitor cells. Cancer Res 72, 1878-89 (2012).

32. Walker, E.J. et al. Monoallelic expression determines oncogenic progression and outcome in benign and malignant brain tumors. Cancer Res 72, 636-44 (2012).

33. Trotman, L.C. et al. Identification of a tumour suppressor network opposing nuclear Akt function. Nature 441, 523-7 (2006).

34. Varambally, S. et al. The polycomb group protein EZH2 is involved in progression of prostate cancer. Nature 419, 624-9 (2002).

35. Serra, V. et al. PI3K inhibition results in enhanced ITER signaling and acquired ERK dependency in ITER2-overexpressing breast cancer. Oncogene 30, 2547-57 (2011).

36. Carracedo, A. et al. Inhibition of mTORCl leads to MAPK pathway activation

through a PI3K-dependent feedback loop in human cancer. J Clin Invest 118, 3065-74 (2008).

37. Wang, Z.G. et al. Role of PML in cell growth and the retinoic acid pathway. Science

279, 1547-51 (1998).

38. Swinnen, J.V. et al. Overexpression of fatty acid synthase is an early and common event in the development of prostate cancer. Int J Cancer 98, 19-22 (2002).

39. Baenke, F., Peck, B., Miess, H. & Schulze, A. Hooked on fat: the role of lipid

synthesis in cancer metabolism and tumour development. Dis Model Mech 6, 1353-63 (2013).

40. Goldstein, J.L. & Brown, M.S. A century of cholesterol and coronaries: from plaques to genes to statins. Cell 161, 161-72 (2015).

41. Porstmann, T. et al. SREBP activity is regulated by mTORCl and contributes to Akt- dependent cell growth. CellMetab 8, 224-36 (2008).

42. Duvel, K. et al. Activation of a metabolic gene regulatory network downstream of mTOR complex 1. Mol Cell 39, 171-83 (2010).

43. Trevino, M.B. et al. Perilipin 5 regulates islet lipid metabolism and insulin secretion in a cAMP-dependent manner: implication of its role in the postprandial insulin secretion. Diabetes 64, 1299-310 (2015). 44. Breitkopf, S.B., Yuan, M., Helenius, K.P., Lyssiotis, C.A. & Asara, J.M. Triomics Analysis of Imatinib-Treated Myeloma Cells Connects Kinase Inhibition to RNA Processing and Decreased Lipid Biosynthesis. Anal Chem 87, 10995-1006 (2015).

45. Breitkopf, S B. et al. A relative quantitative positive/negative ion switching method for untargeted lipidomics via high resolution LC -MS/MS from any biological source. Metabolomics 13(2017).

46. Shroff, E.H. et al. MYC oncogene overexpression drives renal cell carcinoma in a mouse model through glutamine metabolism. Proc Natl Acad Sci USA 112, 6539-44 (2015).

47. Eberlin, L.S. et al. Alteration of the lipid profile in lymphomas induced by MYC

overexpression. Proc Natl Acad Sci USA 111, 10450-5 (2014).

48. Perry, R.H. et al. Characterization of MYC-induced tumorigenesis by in situ lipid profiling. Anal Chem 85, 4259-62 (2013).

49. Nomura, D.K. et al. Monoacylglycerol lipase regulates a fatty acid network that

promotes cancer pathogenesis. Cell 140, 49-61 (2010).

50. Rysman, E. et al. De novo lipogenesis protects cancer cells from free radicals and chemotherapeutics by promoting membrane lipid saturation. Cancer Res 70, 8117-26 (2010).

51. Hilvo, M. et al. Novel theranostic opportunities offered by characterization of altered membrane lipid metabolism in breast cancer progression. Cancer Res 71, 3236-45 (2011).

52. Kotzka, J. et al. Insulin-activated Erk-mitogen-activated protein kinases

phosphorylate sterol regulatory element-binding Protein-2 at serine residues 432 and 455 in vivo. J Biol Chem 279, 22404-11 (2004).

53. Abidi, P., Zhang, F., Li, C. & Liu, J. Blockage of the ERK signaling pathway

abrogates the SCAP ligand-induced transcriptional activation of the LDL receptor gene in HepG2 cells. Int J Mol Med 16, 779-85 (2005).

54. Ma, L., Chen, Z., Erdjument-Bromage, EL, Tempst, P. & Pandolfi, P.P.

Phosphorylation and functional inactivation of TSC2 by Erk implications for tuberous sclerosis and cancer pathogenesis. Cell 121, 179-93 (2005).

55. Roux, P.P., Ballif, B.A., Anjum, R., Gygi, S.P. & Blenis, J. Tumor-promoting phorbol esters and activated Ras inactivate the tuberous sclerosis tumor suppressor complex via p90 ribosomal S6 kinase. Proc Natl Acad Sci USA 101, 13489-94 (2004).

56. Williams, K. J. et al. An essential requirement for the SCAP/SREBP signaling axis to protect cancer cells from lipotoxicity. Cancer Res 73, 2850-62 (2013).

57. Li, X., Chen, Y.T., Hu, P. & Huang, W.C. Fatostatin displays high antitumor activity in prostate cancer by blocking SREBP -regulated metabolic pathways and androgen receptor signaling. Mol Cancer Ther 13, 855-66 (2014).

58. Kamisuki, S. et al. A small molecule that blocks fat synthesis by inhibiting the

activation of SREBP. Chem Biol 16, 882-92 (2009).

59. Lin, J. et al. Hyperlipidemic effects of dietary saturated fats mediated through PGC- lbeta coactivation of SREBP. Cell 120, 261-73 (2005).

60. Unger, R.H. Lipotoxic diseases. Annu Rev Med 53, 319-36 (2002).

61. Buettner, R. et al. Defining high-fat-diet rat models: metabolic and molecular effects of different fat types. JMol Endocrinol 36, 485-501 (2006).

62. Lane, S., Zhou, S., Pan, T., Dai, Q. & Liu, H. The basic helix-loop-helix transcription factor Cph2 regulates hyphal development in Candida albicans partly via TEC 1. Mol Cell Biol 21, 6418-28 (2001). 63. Pascual, G. et al. Targeting metastasis-initiating cells through the fatty acid receptor CD36. Nature 541, 41-45 (2017).

64. Nieman, K.M. et al. Adipocytes promote ovarian cancer metastasis and provide

energy for rapid tumor growth. Nat Med 17, 1498-503 (2011).

65. Park, H. et al. A high-fat diet increases angiogenesis, solid tumor growth, and lung metastasis of CT26 colon cancer cells in obesity-resistant BALB/c mice. Mol Carcinog 51, 869-80 (2012).

66. Kim, E. J. et al. Dietary fat increases solid tumor growth and metastasis of 4T1 murine mammary carcinoma cells and mortality in obesity-resistant BALB/c mice. Breast Cancer Res 13, R78 (2011).

67. Sonn, G.A., Aronson, W. & Litwin, M.S. Impact of diet on prostate cancer: a review.

Prostate Cancer Prostatic Dis 8, 304-10 (2005).

68. Kwon, O.J., Zhang, B., Zhang, L. & Xin, L. High fat diet promotes prostatic basal-to- luminal differentiation and accelerates initiation of prostate epithelial hyperplasia originated from basal cells. Stem Cell Res 16, 682-91 (2016).

69. Menendez, J.A. & Lupu, R. Fatty acid synthase and the lipogenic phenotype in cancer pathogenesis. Nat Rev Cancer 7, 763-77 (2007).

70. Ros, S. et al. Functional metabolic screen identifies 6-phosphofructo-2- kinase/fructose-2, 6-biphosphatase 4 as an important regulator of prostate cancer cell survival. Cancer Discov 2, 328-43 (2012).

71. Gurrieri, C. et al. Loss of the tumor suppressor PML in human cancers of multiple histologic origins. J Natl Cancer Inst 96, 269-79 (2004).

72. Tilch, E. et al. Mutations in EGFR, BRAF and RAS are rare in triple-negative and basal -like breast cancers from Caucasian women. Breast Cancer Res Treat 143, 385- 92 (2014).

73. Chen, M. et al. Deregulated PPla phosphatase activity towards MAPK activation is antagonized by a tumor suppressive failsafe mechanism. Nat Commun Accept(20l7).

74. Bemardi, R. et al. PML inhibits HIF-lalpha translation and neoangiogenesis through repression of mTOR. Nature 442, 779-85 (2006).

75. Das, S.K., Mondal, A.K. & Elbein, S.C. Distinct gene expression profiles characterize cellular responses to palmitate and oleate. J Lipid Res 51, 2121-31 (2010).

76. Trotman, L.C. et al. Ubiquitination regulates PTEN nuclear import and tumor

suppression. Cell 128, 141-56 (2007).

77. Mehlem, A., Hagberg, C.E., Muhl, L., Eriksson, Li. & Falkevall, A. Imaging of neutral lipids by oil red O for analyzing the metabolic status in health and disease. Nat Protoc 8, 1149-54 (2013).

78. Lukacs, R.U., Goldstein, A.S., Lawson, D.A., Cheng, D. & Witte, O.N. Isolation,

cultivation and characterization of adult murine prostate stem cells. Nat Protoc 5, 702-13 (2010).

The entire contents of the above publications are incorporated herein in their entirety.

Example 2

Activation of signaling pathways, such as the phosphoinositide-3 -kinase (PI3K)/AKT and mitogen-activated protein kinase (MAPK), is regulated by feedback inhibition in both normal and cancer cells^{1, 2}. Evasion of feedback inhibition or fail-safe mechanisms resulting from aberrant activation of major oncogenic pathways represents one of the critical mechanisms underlying tumor progression in tumors of diverse histological origin^{3, 4}. On the other hand, relief of negative feedback by anti-cancer drugs constitutes a major hurdle to limit the success of several targeted therapies⁵. Hence, identification of the key pathways that govern such regulation is of utmost importance for tumor-specific therapeutic targets.

Prostate cancer (CaP) is the most common malignancy found in men, and an estimated 1 in 7 men in the U.S. will be diagnosed with CaP during their lifetime⁶. In the past 25 years, CaP mortality has declined by nearly 40%, however, improvement in survival for patients with metastatic disease has not contributed substantially to the observed drop in CaP mortality⁷. More than 26,000 men in the U.S. die annually of metastatic CaP⁶. Recent whole- exome sequencing studies have revealed that copy number alterations, recurrent somatic mutations and genomic rearrangements are among the driving forces for metastatic castration-resistant prostate cancer (mCRPC) and have identified distinct molecular subtypes of mCRPC based on alterations in existing signaling pathways^{8, 9}.

Co-activation of the PI3K/AKT and MAPK pathways is frequently observed in advanced and metastatic CaP and is found to be associated with disease progression and poor prognosis¹⁰. One of the dominant mechanisms underlying PI3K/AKT activation is inactivation of PTEN (phosphatase and tensin homolog)¹¹. In contrast, the mechanisms underlying MAPK activation, to date, remain largely elusive since activating mutations or gene rearrangements among MAPK signaling components are extremely rare in human CaP^8, 12, 13, 14, is, 16, 17

PTEN loss/PBK-AKT activation occurs as an early event in the development of human CaP¹⁸, leading to feedback inhibition on Ras/Raf/MAPK signaling^19, ²⁰ (FIGs. 9-1A and 9-1B) How human CaP evades this feedback inhibition to frequently co activate the PI3K/AKT and MAPK signaling is also poorly understood. In view of these critical gaps in the field, the mechanistic basis of MAPK activation in metastatic human CaP was investigated.

Here it is demonstrated that an S6K-PPla-B-Raf pathway activates MAPK signaling in PI3K/AKT-driven cancers and is opposed by the promyelocytic leukemia (PML) tumor suppressor. We further demonstrate its importance in regulating CaP cell migration and invasion and in metastatic human CaP. Amplification of PPP1CA in metastatic human CaP.

It is now well recognized that the MAPK cascade is negatively regulated through inhibitory phosphorylation of components of the pathway, in particular, Raf kinases, the major upstream activators of MAPK signaling^{21, 22, 23, 24, 25}. Raf kinases can not only be switched on by acquiring activating mutations, but also through phosphatase-mediated dephosphorylation at their inhibitory sites to relieve inhibition and to allow reactivation^{21, 22,}

^{26, 11}. Given that activating mutations in Raf kinases are rare in human CaP, it was postulated that aberrant phosphatase activity might promote Raf kinases activity and subsequent MAPK activation in metastatic human CaP.

We focused on PP2A and PP1, two major eukaryotic protein phosphatases that are reported to contribute to >90% of serine/threonine dephosphorylation and regulate a variety of cellular processes through the dephosphorylation of distinct substrates²⁸, and it was initially sought to determine if genetic alterations to either of these protein phosphatases could help establish a role in the context of metastatic cancer. Interestingly, the catalytic subunit of PPla, encoded by th e PPPICA gene in human, is located on chromosomal band 1 lql3, one of the regions frequently amplified in CGH analysis of human CaP^{29, 30}.

Moreover, enhanced cytoplasmic PPla immunostaining strongly correlates with high

Gleason Score³⁰, suggesting that PPla may be involved in prostate tumorigenesis. To confirm the relevance of PPP1CA to human CaP, the genomic status of PPP1CA in a recent array-based comparative genomic hybridization (aCGH) dataset⁸ of 59 localized (LPC) and 35 mCRPC CaP was evaluated. We found that PPP1CA was amplified in 7% of LPC and 17% of mCRPC, respectively (FIG. 9A). To independently validate the findings from the aCGH dataset, a recent large whole-exome sequencing dataset⁹ of 150 samples from mCRPC patients was analyzed. Consistent with the aCGH analysis, PPP1CA was amplified in 25% of mCRPC (FIG. 9B). Notably, PPP1CA, which is ~2Mb away from cyclin Dl a proto- oncogene also associated with metastatic CaP¹⁸, was more frequently amplified than cyclin Dl in mCRPC (25% us 5%, FIG. 9C). Also of interest, PPP1CA was co-amplified with cyclin Dl in 5 out of 7 cases where the latter was amplified (FIG. 9C). Thus, amplification of PPP1CA occurs frequently in metastatic human CaP.

PPla mediates ML-loss induced MAPK activation.

We previously reported that co-deletion of PTEN and PML leads to MAPK

reactivation in mouse prostate epithelial cells and human CaP cells and frequently occurs in metastatic human CaP³¹. To determine if PPP1CA genomic amplification could cooperate with co-deletion of PTEN and PML in metastatic human CaP, whether amplification of PPP1CA was correlated with co-deletion of PTEN and PML in the Robinson el al. dataset⁹ was determined. Strikingly, PPP1CA genomic amplification and co-deletion of PTEN and PML was found to be often mutually exclusive (FIG. 9D), supporting their proto-oncogenic functions in the same pathway. In keeping with this notion, overexpression of PPla induced a marked increase in ERK phosphorylation (FIG. 10A and FIG. 10-1A). Conversely, knockdown of PPla via small interfering RNA (siRNA) attenuated ERK phosphorylation in both LNCaP and PC3 cells (FIG. 10B), suggesting that PPla is the principle phosphatase that positively regulates the MAPK signaling pathway. In contrast, although PP2A has previously been identified as the phosphatase mediating the dephosphorylation and reactivation of Raf kinases (FIG. 10-1B)²², the overexpression or silencing of the PP2A catalytic subunit (PP2A-C) did not affect ERK phosphorylation in either 293T cells or CaP cell lines (FIG. 10-lC to 10-1E), presumably due to its complex effects on various components of the MAPK cascade³², and no consistent genomic alterations in the PPP2CA , the gene encoding PP2A-C protein, have been found in mCRPC samples from the Robinson et al. dataset⁹ (FIG. 10-1F).

Although co-deletion of PTEN and PML is often mutually exclusive with

amplification of PPP1CA , the possibility that co-loss of PTEN/PML might lead to aberrant PPla phosphatase activity was not ruled out, thereby contributing to PMLAoss induced MAPK activation. It is well established that PML is induced upon PTEN inactivation in a p53-dependent and -independent manner^{4, 33, 34} and that PML, through its nuclear body (NB)- dependent scaffold activity, can interact with a variety of proteins, including phosphatases, to regulate their functions³⁵. Additionally, PPla is a known Rb phosphatase, and, like PML, can bind Rb (FIG. 10-1G) and cooperate with Rb function in the induction of cellular

senescence³⁶. We therefore considered that upon PTEN loss, PML might recruit PPla into the PML-NBs to promote PTEN-loss induced cellular senescence, a potent failsafe mechanism that restricts tumorigenesis⁴, and simultaneously restrict PP la-induced MAPK activation. WI-38 human diploid fibroblasts, the well-accepted cell models used to study the senescence with intact PTEN and PML protein expression, were used to test this possibility. We found that acute knockdown of PTEN in WI-38 cells induced senescence, PML upregulation, strong co-localization of PML and PPla, and suppression of MAPK signaling (FIG. 10-1H and FIGs. 10C and D) On the other hand, it was reasoned that in PTEN/PML double-null cells, due to the lack of sequestration to NBs, PPla could contribute to MAPK activation through the dephosphorylation and reactivation of Raf kinases. In support of this hypothesis, the cytoplasmic accumulation of PPla was significantly increased upon simultaneous knockdown of PTEN and PML along with concomitant MAPK reactivation in WI-38 cells (FIGs. 10C and D). Moreover, compared to wild type (WT) PPla, a PPla mutant constitutively targeted to the nucleus by fusing the NLS sequence derived from c-Myc³⁷, displayed a drastically decreased capacity to activate ERK (FIG. 10E), further corroborating the role of subcellular compartmentalization in the regulation of PP la-induced ERK activation. Importantly, PML and PPla interact with each other in vivo in PTEN- xW CaP cells. Immunoprecipitation of PML led to the co-immunoprecipitation of PPla, and vice versa in PC3 cells (FIG. 10F).

To further investigate whether PPla might mediate PML- loss driven MAPK activation, siRNA targeting PPla was used in stable PML knockdown cells. We found that the induction of basal and EGF-stimulated ERK phosphorylation through stable knockdown of PML was suppressed by PPla downregulation (FIG. 10G). Therefore, PML- loss driven MAPK activation is mediated, at least in part, by PPla. Notably, the phosphatase activity of PPla is required for MAPK activation since expression of the phosphatase-inactive PPla mutant (H248K)³⁸ did not affect ERK phosphorylation (FIG. 10H). In line with this, PC3 cells treated with tautomycin, a more selective inhibitor for PPla^{39, 40, 4 K 42}, displayed a dose- dependent inhibition of EGF-induced ERK phosphorylation (FIG. 10-11). Thus, it was concluded that in PTEN and PML double-null cells, sequestration of PPla to NBs is impaired and in turn facilitates the aberrant cytosolic localization of PPla and its subsequent activation of MAPK signaling.

S6K induces PPla phosphorylation, 14-3-3 binding and cytoplasmic accumulation.

To gain further mechanistic insights into how PPla is delocalized into the cytoplasm, which AGC kinase can trigger cytoplasmic accumulation of PPla given that the AGC family kinases are commonly activated upon PTEN loss⁴³ was examined. We found that S6K1, but not other AGC kinases such as AKT and SGK, phosphorylated PPla in vivo (FIG. 11 A), induced the interaction of PPla with 14-3-3g (FIG. 11B), and led to increased cytoplasmic accumulation of PPla (FIGs. 11C and 11D). Notably, knockdown of S6K1 via siRNA largely suppressed PP la-induced ERK activation (FIG HE), suggesting that activation of ERK by PPla is dependent on S6K. In support of PPla being an S6K substrate, two highly conserved imperfect AGC family kinase-recognition motifs (RxRxxpS/T) located at

S224/T226 and T320 of PPla protein, respectively, were identified (FIG. 11F). We generated PPla mutants in which S224/T226 (S224A/T226A), T320 (T320A) or all three sites (3 A), were mutated to alanine. In vitro kinase assays confirmed that recombinant S6K1 could also phosphorylate PPla (FIG. 11G). Moreover, S224A/T226A and 3A PPla mutants, but not T320A, displayed drastically reduced S6K1 -dependent PPla phosphorylation (FIG. 11G). Consequently, S224A/T226A and 3 A PPla mutants had a lower capacity to interact with 14-3-3 and to activate MAPK signaling (FIGs. 11H and 111), suggesting that S6K mediated PPla phosphorylation on S224/T226 is critical for the binding of PPla with 14-3-3 and for the ability of PPla to activate MAPK.

PPla dephosphorylates B-Raf inhibitory phosphorylation sites.

Next, it was determined if PPla could interact with Raf family kinases.

Immunoprecipitation of PPla revealed a strong and specific interaction between PPla and B- Raf and to a lesser extent, A-Raf, while PPla did not interact with C-Raf or other

components of the MAPK cascade, such as MEK and ERK (FIG. 12A, the left panel). Given that A-Raf shows weak interaction with PPla and has low intrinsic activity towards MAPK activation⁴⁴, further studies investigating the effect of PPla on Raf family kinases were focused on B-Raf. The interaction between PPla and B-Raf was further confirmed by reciprocal immunoprecipitation in which PPla co-immunoprecipitated with anti -B-Raf precipitates (FIG. 12B, the right panel).

The observation that PPla and B-Raf interact and that the phosphatase activity of PPla is required for MAPK activation led us to investigate whether B-Raf might be a putative substrate of PPla. We therefore asked if B-Raf itself is dephosphorylated by PPla. Since ERK -mediated feedback inhibitory phosphorylation on S151, T401, S750, and T753 of B-Raf protein negatively regulates its kinase activity²², it was reasoned that PPla could dephosphorylate these inhibitory sites and, in turn, relieve feedback inhibition and reactivate MAPK. To address whether PPla activates B-Raf through these inhibitory sites, B-Raf protein mutants were used in which an individual inhibitory site, as well as all four sites (4A), were mutated to alanine. As expected, cells overexpressing either PPla or WT B-Raf had higher ERK phosphorylation than cells transfected with empty vector (FIG. 12B).

Additionally, cells overexpressing both PPla and WT B-Raf displayed even higher ERK phosphorylation (FIG. 12B), suggesting that PPla can enhance B-Raf activity. On the other hand, the 4A B-Raf mutant, but not single alanine B-Raf mutants, displayed an enhanced basal ability to activate MAPK compared to WT B-Raf (FIG. 12B), indicating that these four sites indeed negatively regulate B-Raf activity, while dephosphorylation of a single inhibitory site is not sufficient to increase B-Raf activity. Critically, the 4A B-Raf mutant was insensitive to PPla activation (FIG. 12B); In contrast, mutation of other known B-Raf inhibitory phosphorylation sites, including S365A/S429A/T440A²⁴, S465A/S467A²⁵, and S614A²³, to alanine failed to blunt PPla-mediated ERK activation (FIGs. 12-1A to 12-1C). Therefore, PPla appears to exert its effect on B-Raf primarily through the ERK -regulated inhibitory sites.

To determine which site could be dephosphorylated by PPla, GST-3 A-B-Raf proteins in which one inhibitory site was WT and the other three sites were mutated to alanine were purified. GST-WT-B-Raf and GST-4 A-B-Raf protein were included as the positive and negative control, respectively. We then phosphorylated GST-B-raf in vitro by incubating B- Raf with recombinant ERK2 protein, and next used the phosphorylated form of B-Raf as a substrate for recombinant PPla. We confirmed that GST-B-Raf was phosphorylated by ERK2 in vitro ²², mainly on S151 and T753 (FIG. 12C), and found that PPla

dephosphorylated B-Raf on both ERK phosphorylation sites (FIG. 12D). Thus, PPla appears to dephosphorylate these inhibitory sites of B-Raf, triggering relief of feedback inhibition and consequent activation of the MAPK pathway.

PPla promotes CaP cell invasiveness via MAPK signaling.

The identification of PPP1CA amplification in metastatic human CaP is consistent with a pro-metastatic role for PPP/CA in the prostate (FIGs. 9A and 9B). We therefore examined whether overexpression of PPla affects CaP cell migration and invasion. Indeed, it was found that LNCaP cells stably overexpressing PPla exhibited higher ERK activation along with significantly increased cell migration and invasion (FIG. 12E). Notably, treatment with the MEK inhibitor, EG0126, in LNCaP cells repressed not only basal but also PPla- induced cell migration and invasion (FIG. 12E). Similar results were also obtained in PC3 cells (FIG.12-1D). These functional data, together with the human genetic and mechanistic analyses, implicate PPP1CA as a pro-metastatic proto-oncogene in human CaP and MAPK signaling as one of the key downstream effectors of PP la-induced cell invasiveness.

Integrated genetic and molecular analyses allowed us to identify an S6K-PPla-B-Raf pathway towards the aberrant activation of MAPK signaling. We find that this pathway is suppressed by the PML tumor suppressor through sequestration of PPla into NBs in PTEN- null cells as a result of failsafe mechanisms evoked by P TEN loss (FIG. 12F). Although the critical role of PP1 and PP2A mediated dephosphorylation in the activation of Raf kinases has previously been reported^{26, 27, 45, 46}, it was shown here for the first time that 1) PPla is amplified in metastatic human CaP and is the principle phosphatase positively regulating MAPK activation and that 2) S6K -mediated PPla cytoplasmic accumulation is essential for the activation of MAPK by PPla. Importantly, given that PML is frequently lost in human cancer⁴⁷, our study suggests that aberrant cytoplasmic retention of PPla caused by PML loss, or the amplification of PPP1CA might represent a common mechanism underlying MAPK activation in cancers that lack activating mutations or gene rearrangements among MAPK signaling components, such as breast cancer and CaP^{13, 14, 15, 16, 17, 48, 49}. Notably, since nuclear PPla has been shown to activate Rb tumor suppressor through dephosphorylation⁵⁰, PML and PML-NBs do serve as rheostats to switch PPla activity from tumor suppressive to tumor promoting. It is also worth noting that, although both PML and 14-3-3g interact with PPla, neither of them affect PPla phosphatase activity towards dephosphorylating its respective nuclear and cytosol substrates (FIGs. 12-1E and 12-1F), suggesting that PML or 14-3-3g primarily functions as a scaffold/chaperone for PPla rather than as a direct regulator of PPla phosphatase activity.

Additionally, as S6K is a downstream target of the ERK pathway^{51, 52}, our study suggests that, in the context of PML loss or PPP1CA amplification, the S6K-PPla-B-Raf- ERK pathway represents a feed-forward loop supporting sustained ERK activation. However, it was previously also shown that AKT/mTOR/S6K activation triggers a negative feedback on MAPK signaling pathway, presumably as a result of the upstream IRS inactivation induced by S6K^{19, 53} (FIGs. 9-1A and B). Thus, depending on the genetic context, S6K is a double-edge sword in the regulation of ERK activation, since it can act as both a suppressor of ERK activation, in the context of intact PML function, and as a promoter of sustained ERK activation, in the context of PML loss or PPP1CA amplification.

Collectively, our findings have important implications for tumorigenesis at large because simultaneous activation of MAPK and PI3K/AKT signaling undoubtedly represents a devastating force in propelling cancers into more advanced and aggressive diseases. Since MAPK activity can be suppressed/inhibited by small pharmacological inhibitors, this study provides a compelling rationale for treating patients with co-deletion of PTEN and PML or amplification of PPP1CA with combinatorial therapy targeting both AKT/mTOR and MAPK signaling that is worthy of further investigations.

Example 2 Methods

Plasmids, reagents and antibodies. Human WT and c-terminal NLS^{c Myc}-tagged³⁷ PPla cDNA were cloned into pCMV-Tag2B vector to generate PPla expression plasmid.

pCDNA3. l-hygro-B-Raf was purchased from Addgene. B-Raf and PML-I cDNA were subcloned into pGEX-4T-l vector and used to express GST-B-Raf and GST-PML protein.

All mutant constructs of B-Raf and PPla were generated using a QuickChange Lightning Site-Direct Mutagenesis (Agilent Technologies) and all mutations were confirmed by sequencing. H A-Myr- AKT 1 / AKT2/ AKT3 , HA-SGKlA60, HA-S6K1-CA and HA-14-3-3 isoforms were previously described⁵⁴. The SMART pool or two independent siRNA duplexes targeted to PML, PTEN, PPla, PP2A-C, S6K1 and control non-target siRNA were purchased from Dharmacon or Sigma. The sequences for the siRNA are listed in Table 1. The target sequences in the pLKO-PML shRNA vector against human PML were 5'- GTGTACGCCTTCTCCATCAAA-3 ' and 5 '-CACCCGCAAGACCAACAACAT-3 '.

Tautomycin was from Enzo life sciences. EG0126 was from Selleck Chemicals. EGF,

Lipofectamine 2000, Lipofectamine RNAiMAX, RPMI, DMEM, Opti-MEM reduced serum media and fetal bovine serum (FBS) were from Invitrogen. Polybrene and puromycin were from Sigma. We used the following primary antibodies for immunoblotting: Anti-p-ERK (Cell Signaling Technology, 9101, 1 : 1000), anti-ERK (Cell Signaling Technology, 9102, 1 : 1000), anti-p-MEK (Cell Signaling Technology, 9154, 1 : 1000), anti-MEK (Cell Signaling Technology, 9126, 1 : 1000), anti-A-Raf (Cell Signaling Technology, 4432, 1 : 1000), anti-B- Raf (Santa Cruz Biotechnology, sc-5284, 1 : 1000), anti -B-Raf (Santa Cruz Biotechnology, sc- 9002, 1 : 1000), anti-C-Raf (Cell Signaling Technology, 9422, 1 : 1000), anti-p-AKT substrate (RXXS*/T*) (Cell Signaling Technology, 9614, 1 : 1000), anti-p-AKT (Cell Signaling Technology, 9271, 1 : 1000), anti-AKT (Cell Signaling Technology, 9272, 1 : 1000), anti-p- S6K (Cell Signaling Technology, 9234, 1 : 1000), anti-S6K (Cell Signaling Technology, 2708, 1 : 1000), anti-IRSl (Cell Signaling Technology, 3407, 1 : 1000), anti-IRS2 (Cell Signaling Technology, 4502, 1 : 1000), anti-PP2AC (Cell Signaling Technology, 2259, 1 : 1000), anti-p- CREB (Cell Signaling Technology, 9198, 1 : 1000), anti-GAPDH (Cell Signaling Technology, 2118, 1 :6000), anti-PML (Bethyl Laboratories, A301-167A, 1 :2000), anti-HA (Santa Cruz Biotechnology, sc-805, 1 : 1000), anti-HSP90 (BD Biosciences, 610419, 1 :4000, anti-PPla (Novus Biologicals, NB- 110-57428, 1 : 1000), anti -Flag (Sigma- Aldrich, F1804, 1 :4000), anti- b-actin (Sigma-Aldrich, A5316, 1 :4000), anti-Rb (Santa Cruz Biotechnology, sc-50, 1 : 1000), and anti-Lamin Bl (Abeam, abl6048, 1 : 1,000).

Cell culture, transfection and lentivirus production. All cell lines were purchased from the American Type Culture Collection (ATCC), cultured in RPMI or DMEM supplemented with 10% FBS and tested for mycoplasma contamination every month. Transfections were performed using Lipofectamine 2000 or Lipofectamine RNAiMAX reagent (Invitrogen) according to manufacturer’s instruction. In brief, lxlO⁵ cells were transfected with 1 pg of DNA plasmids or 50nM siRNA in a 6-well dish. Cells were recovered into completed media after l2-hr transfection and then harvested at the indicated times. For lentivirus production, human PPla cDNA was subcloned into the pWPI-Puro lentiviral vector to generate pWPI- Puro-PPla. pWPI-Puro Vector or pWPI-Puro-PPla (6 pg), pMD2.G (1.5 pg) and psPAX2 (4.5 pg) were co-transfected into 293 T cells using Lipofectamine 2000. The virus

supernatants were collected 48 hours after transfection and filtered through a 0.45 pm filter. Freshly made virus supernatants supplemented with 4 pg/ml polybrene were added to exponentially growing LNCaP and PC3 cells. After 5 hours, fresh medium was added. CaP cells were then selected with 2 pg/ml puromycin (Sigma) for 48 hours after two-day infection and then used for the cell migration and invasion assay.

Immunofluorescence. Cells were grown on coverslips, fixed with 4% paraformaldehyde and permeabilized with ice-cold methanol. Cells were rinsed with PBS, blocked with 10% goat serum and then incubated with primary antibody overnight, followed by incubation with Alexa Fluor conjugated secondary antibodies (Life Technologies). Coverslips were mounted with ProLong Gold Antifade reagent with DAPI (Life Technologies). The following primary antibodies were used for immunofluorescence: anti -Flag (Sigma-Aldrich, F1804, 1 :400), anti- HA (Santa Cruz Biotechnology, sc-805, 1 :400), anti-PML (Santa Cruz Biotechnology, sc- 966, 1 :400), and anti-PPla (Bethyl Laboratories, A300-904A, 1 :400). The stained slides were visualized by a bright-field or confocal microscope. SA-P-gal activity in prostate tissue was measured with the senescence detection kit (Calbiochem) on 5 pm-thickness frozen section.

Western blotting and immunoprecipitation. Cell lysates or prostate tissues were prepared in RIPA buffer (Sigma) supplemented with protease (Roche) and phosphatase (Sigma) inhibitor. Proteins were separated on 4-12% Bis-Tris gradient gels (Invitrogen), transferred to polyvinylidine difluoride membranes (Immobilon P, Millipore) and the blots were probed with the indicated antibodies. Densitometry quantification was performed with ImageJ.

Nuclear/cytoplasmic fractionation was performed as described⁵⁵. For immunoprecipitation, cells were lysed in lysis buffer (50 mM Tris at pH 7.5, 10% glycerol, 5 mM MgCl2, 150 mM NaCl, 0.2% NP-40, protease (Roche) and phosphatase (Sigma) inhibitor) and the lysates were incubated with anti-PML (Santa Cruz Biotechnology, sc-966, 1 : 100) or anti -PP la

(Invitrogen, 43-8100, 1 :50) or anti-B-Raf (Santa Cruz Biotechnology, sc-5284, 1 : 100) or anti-Rb (Santa Cruz Biotechnology, sc- 102, 1 : 100) or anti -Flag (Sigma- Aldrich, F1804,

1 : 100) antibody overnight at 4°C. The protein G sepharose (GE Healthcare) was then added and incubated for another 2h. The immunoprecipitates were washed with wash buffer (50 mM Tris at pH 7.5, 5 mM MgCl2, 150 mM NaCl, 0.1% NP-40, protease (Roche) and phosphatase (Sigma) inhibitor) three times and eluted with 2XSDS sample buffer. Prostate tissues from WT and PterP^z mice as well as Primary Pten^lox/lox MEFs and their transduction with retrovirus expressing PEIRO-IRES-GFP or Cre-PETRO-IRES-GFP have been previously described⁴.

In vitro kinase and phosphatase Assays. In vitro kinase assays were performed as described⁵⁶. Briefly, bacterial expressed GST-PPla-H248 (the phosphatase-dead mutant to avoid the autodephosphorylation of PPla) or GST-B-Raf was purified using Glutathione Sepharose 4B (GE Healthcare) according to the manufacturer’s instructions. To determine the residue(s) on PPla protein phosphorylated by S6K1, recombinant S6K1 (R&D Systems) was incubated with 1 pg of GST-PPla-H248. For the in vitro phosphatase assays, recombinant ERK2 (R&D Systems) was incubated with 1 pg of GST-B-Raf in the absence or presence of PPla (Lifespan Bioscience) or PP2A-C (Abeam). Both reactions were incubated in kinase assay buffer (50 mM Tris-HCl pH 7.5, 2 mM MgCh, 0.1 mM EDTA, 2 mM DTT, 0.1 mM ATP) with 5 mCi [g-³²R] ATP. For PPla treatment, 1 mM MnCh was added to the kinase assay buffer to promote PPla activity. The reaction was initiated by the addition of GST- PPla-H248 or GST-B-Raf in a volume of 30 pl for 30 min at 30 °C followed by the addition of SDS-PAGE sample buffer to stop the reaction before resolved by SDS-PAGE. To determine B-Raf kinase activity towards phosphorylating MEK1, in vitro kinase assays were performed as described previously⁵⁷. Briefly, B-Raf kinase was immune-purified from 293T cells transfected with Flag-B-Raf constructs. GST-MEK1 was expressed in BL21 E.coli and purified using Glutathione Sepharose 4B media (GE Healthcare). B-Raf kinase was incubated with 0.2 pg of GST-MEK1 in the absence or presence of PPla or PP2A-C in kinase assay buffer (10 mM HEPES pH 8.0, 10 mM MgCh, 1 mM dithiothreitol, 0.1 mM ATP). Reaction was initiated by the addition of GST-MEK1 in a volume of 30 mΐ for 15 min at 30 °C followed by the addition of SDS-PAGE sample buffer to stop the reaction before resolved by SDS-PAGE. Nuclear PPla phosphatase activity was determined by in vitro phosphatase assays using CREB as the substrate as described previously⁵⁸. Phosphorylated CREB protein was immune-purified from 293T cells transfected with Flag-CREB. The phosphatase assay was carried out using Flag-CREB and PPla in IX NEBuffer for PMP supplemented with 1 mM MnCh (New England Biolabs). Bacterially expressed and purified GST-PML was added as indicated in the experiments.

Cell migration and invasion assay. CaP cells stably expressing pWPI-Vector or pWPI-PPla were detached into single-cell suspension. LNCaP (lxlO⁵) or PC3 (5xl0³ and 5xl0⁴ for migration and invasion assay, respectively) cells in 100 mΐ of 0.1% FBS -containing RPMI medium in the absence or presence of 20 mM EG0126 were placed into the top chamber of 8 pm transwell inserts for migration assay or Matrigel-coated transwell inserts for invasion assay (BD Biosciences). A volume of 600 pl of 10% FBS-containing RPMI in the absence or presence of 20 mM EG0126 were added to the bottom wells. After 24 hours or 48 hours (for LNCaP invasion assay only), cells on the upper surface of the inserts were removed with a cotton swab. Migrated cells were fixed in 10% formalin, then stained with 0.2% crystal violet for 10 min. Cells were counted in four microscopic fields under 20x magnification. Results are representative of three independent experiments.

Array CGH analysis. The data were downloaded from GEO database (Grasso: GSE35988) or cBioportal (for Robinson et al. dataset⁹) with focus on the aCGH datasets. The R scripts were used to process the data. The cutoff threshold used was -0.35 to -0.8 as heterozygous deletions, those lower than -0.8 as homozygous deletions, 0.6 to 0.8 as l-copy gain, those higher than 0.8 as >l-copy gain.

Statistical analysis. 2X2 contingency tables were constructed and were used to analyze categorical data (for example, copy number alteration). The datasets were compared using Fisher’s exact test. For analysis of average data, datasets were compared using unpaired two- tailed Student’s t tests. P values of <0.05 were considered to be statistically significant. Data availability. The authors declare that the data supporting the findings of this study are available within the article and its Supplementary Information files. All relevant data are available from the authors upon request.

Our data provide a strong genetic foundation to the mechanisms underlying metastatic progression, and demonstrate how environmental dietary factors can boost progression from primary to metastatic cancer, intertwining with the genetic makeup of cancer (FIG. 8C). We have demonstrated that SREBP-dependent lipogenesis, which can be hyperactivated by concomitant activation of the PI3K/AKT and MAPK pathways, or a HFD regimen, functions as an underlying rheostat towards metastatic cancer progression.

Furthermore, PML was identified as a critical mediator of feedback inhibition of MAPK signaling and lipogenesis, driving metastatic progression in PTEN loss/P 13 K-AKT driven cancers. Interestingly, Cph2, the yeast homologue of SREBP, has been shown to be involved in invasive/p seudohyphal growth in yeast, suggesting a conserved evolutionary function of SREBP in invasiveness⁶².

Consistent with earlier studies showing that lipid metabolism is required for metastasis in other cancer systems^63,64, our analysis reveals a highly enriched SREBP-l dependent lipogenic signature in metastatic human CaP. Concordantly, a HFD enriched in saturated and monounsaturated FAs triggers metastasis in a non-metastatic PterP^c CaP model and further enhances metastasis in a metastatic PtetP^PmP⁰^ CaP model. In addition to our genetic and experimental evidence in GEMMs of CaP, it has been reported that HFD favors metastatic progression in xenograft models from cell lines of various histological origin^63,65,66. Numerous mechanisms have been proposed to explain a possible association between dietary lipids and CaP⁶⁷, including paracrine mechanisms through secreted cytokines from adipose tissues, endocrine mechanisms through an alteration of androgen levels (FIG. 8-1 A), and an induction of basal -to-luminal cell differentiation caused by immune cell infiltration⁶⁸. However, it was shown here that specific genetic perturbations or HFD are likely able to exert a direct effect on metastasis through increased lipid accumulation. We also characterized the intracellular lipid changes in GEMMs of CaP and detected qualitative changes in four different lipid classes as well as in the saturation of fatty acyl chains. Together, these results establish a strong mechanistic and causal link between aberrant lipogenesis, excess lipid accumulation and metastasis, providing a compelling rationale for integrating lifestyle data (e.g. diet) and tumor genetics into clinical practice to identify patients at high-risk of metastasis. Additionally, lipid metabolism itself is an attractive therapeutic target through inhibition of lipogenic enzymes^39,69. Notably, such inhibitors result in reduced CaP cell viability only in the absence of an exogenous lipid source such as lipoprotein⁷⁰, highlighting the importance of integrating pharmacologic approaches with stringent dietary regimens to prevent metastasis. Future studies are warranted to evaluate whether specific lipid subsets/signatures can serve as prognostic biomarkers to distinguish CaP with metastatic potential from indolent disease.

Lastly, given that PML is lost in human cancer of multiple histological origins⁷¹, our study suggests that PML loss may underlie MAPK activation in cancers lacking genetic alterations in MAPK signaling components^{26 28,72}. In an accompanying study, it was demonstrated that PPla is a B-Raf activating phosphatase genetically amplified in CaP, and that PML can suppress PP la-dependent activation of MAPK signaling⁷³. Moreover, PtetP^-PmP⁰^- tumors displayed a significantly up-regulated gene set induced by hypoxia, consistent with earlier studies showing that PML is a key player in orchestrating the cellular response to hypoxia through repression of mTOR⁷⁴ (FIG. 8-1B). This finding has equally important implications for turn ori genesis, because PML loss in the hypoxic core or tumoral lesions would not only activate mTOR, resulting in sustained HIF-l activation, but would also relieve the feedback inhibition of MAPK signaling triggered by mTOR activation, thus leading to simultaneous activation of both mTOR and MAPK signaling. Taken together, our study provides a roadmap for targeted therapies tailored to the individual patient for the prevention and treatment of metastatic cancer.

Cell culture and transfection. All cell lines were obtained from ATCC and checked for mycoplasma by MycoAlert Mycoplasma Detection Kit (Lonza). Cells were maintained in DMEM or RPMI supplemented with 10% fetal bovine serum, 2 mM glutamine, 100 U/ml penicillin and streptomycin (Invitrogen). To study the effect of PML&MAPK signaling on SREBP and SREBP targets, LNCaP or PC3 cells were seeded and cultured in media with 10% lipoprotein deficient serum (Kalen Biomedical) throughout the course of the experiments. Transfections were performed using Lipofectamine 2000, Lipofectamine RNAiMAX reagent or TrcmsTT-X2 according to manufacturer’s instruction. In brief, 50nM mixtures of two independent siRNA pairs targeting each gene or 1 pg of DNA plasmids were transfected into lxlO⁵ cells in a 6-well dish. Cells were recovered into completed media after l2-hr transfection and then harvested at the indicated times. For the treatment of dietary lipids, palmitic and oleic acid were conjugated to fatty acid-free BSA as described⁷⁵. CaP cells were cultured with BSA vehicle control, 2% lipid mixture, 30 pm BSA-conjugated palmitic acid or oleic acid for 7 days, then subjected to ORO staining, cell migration, and invasion assay.

immunostaining. The following primary antibodies were used for IHC: PML (MAB3738, 1 :300), p-ERK (20G11, 1 : 100), p-AKT (D9E, 1 : 100), CK8 (MMS-162P, 1 :200), AR

Prostate sphere assay. The culture and passage of prostate spheres were carried out as described⁷⁸. Dissociated prostate epithelial cells were prepared from mice at 12 weeks of age. To initiate sphere formation, unsorted mouse prostate cells were prepared in PrEGM medium (Lonza) at a density of 2.5 x 10⁵ cells per ml. 40 pl of cell suspension was mixed with 60 pl cold Matrigel (BD Bioscience), and pipetted around the rim of a well of a 12-well plate and allowed to solidify at 37°C for 30 min. lml warm PrEGM was then added to each well. The spheres were cultured and monitored for 14 days with 50% medium change every 3 days. To passage spheres, Matrigel was digested by 1 mg/ml dispase solution (StemCell Technologie) for 30 minutes at 37°C. Digested cultures were collected, pelleted, resuspended and subjected to sequential digestion by 2mg/ml type I collagenase (Sigma) for 1 hr and 0.05%

Trypsin/EDTA (Invitrogen) for 5 min at 37°C, and then passed through a 27-gauge syringe 5-10 times, and filtered through a 40 mih cell strainers. Cells were counted by hemocytometer and replated at the density of lxlO⁴ cells per l2-well.

Lipidomics by untargeted high-resolution liquid chromatography-tandem mass spectrometry (LC-MS/MS). The lipidomic analysis was performed as previously

described⁴⁴. Briefly, non-polar lipids were extracted from 5 mg prostate tissues using MTBE. The upper phase containing the non-polar lipids was dried using a speedvac with no heat. Lipid samples were resuspended in 35 mΐ of 50% isopropanol (IPA)/50% MeOH. 10 mΐ of samples were injected for reversed-phase (Ci₈) LC-MS/MS using a hybrid QExactive Plus Orbitrap mass spectrometer (Thermo Fisher Scientific) coupled to an Agilent 1100 HPLC in DDA mode using positive/negative ion polarity switching (Top 8 in both modes). The lipidomics data were analyzed using LipidSearch 4.1.9 software. The software identifies intact lipid molecules based on their molecular weight and fragmentation pattern using an internal library of predicted fragment ions per lipid class and the spectra are then aligned based on retention time and MS1 peak areas are quantified across sample conditions. Excel 2010 was used to produce intensity and the R program (version 3.2.5) was used for data manipulation and statistical analyses, including unsupervised hierarchical clustering and heat map visualization. Cell migration and invasion assay. PC3 or LNCaP cells were transfected with indicated siRNA or pretreated with 10 pg/ml TOFA or 10 mIUI Simvastatin for 48 hours, then detached into single-cell suspension. LNCaP (lxlO⁵) or PC3 (5xl0³ and 5xl0⁴ for migration and invasion assay, respectively) cells were resupsended in 100 mΐ of RPMI medium containing 0.1% FBS and placed into the top chamber of 8 pm transwell inserts for migration assay or Matrigel-coated transwell inserts for invasion assay (BD Biosciences). The bottom wells contained 600 pl RPMI supplemented with 10% FBS. After 24 hours or 48 hours (for LNCaP invasion assay only), cells on the upper surface of the inserts were removed with a cotton swab. Migrated cells were fixed in 10% formalin, then stained with 0.2% crystal violet for 10 min. Cells were counted in four microscopic fields under 20x magnification. Results are representative of three independent experiments.

Blood sampling and testosterone ELISA. After 90 days of chow- or HFD-feeding, blood samples were collected from the mice through cardiac puncture into BD Vacutainer SST™ Serum Separation Tubes (BD 367986) and immediately mixed by proper inversion. The SST™ blood specimens were allowed to clot for 30 minutes and then centrifuged at lOOOg for 10 minutes in a swing bucket centrifuge to recover the serum. Serum levels of

testosterone were measured using a competitive ELISA kit according to manufacturer’s instructions (Abeam 108666).

Array CGH analysis. We downloaded the data from GEO database (Grasso: GSE35988; Taylor: GSE21032) or cBioportal (for Robinson et al. dataset²⁵) with focus on the aCGH datasets. We wrote R scripts to process the data and generate the heatmap based on the log2- transformed ratio. The cutoff threshold used was -0.35 to -0.8 as heterozygous deletions, and those lower than -0.8 as homozygous deletions.

Data availability. Microarray data in this paper have been deposited in NCBI's Gene Expression Omnibus and are accessible through GEO series accession number GSE98493. All other data supporting the findings of this study are available within the article and its Supplementary Information files and are available from the authors upon request.

URLs. Gene Ontology Analysis (GO), http://pantherdb.org; Gene Set Enrichment Analysis (GSEA), http://software.broadinstitute.org/gsea/msigdb/; NCBI Gene Expression Omnibus (GEO), http://www.ncbi.nlm.nih.gov/geo/. 1. Wu, J.N., Fish, K.M., Evans, C.P., Devere White, R.W. & Dall'Era, M. A. No improvement noted in overall or cause-specific survival for men presenting with metastatic prostate cancer over a 20-year period. Cancer 120, 818-23 (2014).

2. Yang, M. et al. Dietary patterns after prostate cancer diagnosis in relation to disease- specific and total mortality. Cancer Prev Res (Phila) 8, 545-51 (2015).

3. Gronberg, H. Prostate cancer epidemiology. Lancet 361, 859-64 (2003).

An umbrella review of the evidence. Eur J Cancer 69, 61-69 (2016).

Other Embodiments

From the foregoing description, it will be apparent that variations and modifications may be made to the invention described herein to adopt it to various usages and conditions. Such embodiments are also within the scope of the following claims.

The recitation of a listing of elements in any definition of a variable herein includes definitions of that variable as any single element or combination (or subcombination) of listed elements. The recitation of an embodiment herein includes that embodiment as any single embodiment or in combination with any other embodiments or portions thereof.

All patents and publications mentioned in this specification are herein incorporated by reference to the same extent as if each independent patent and publication was specifically and individually indicated to be incorporated by reference.

Claims

What is claimed is:

1. A method for treating prostate cancer in a subject, the method comprising administering to the subject an agent that inhibits AKT/mTOR or MAPK signaling.

2. A method for treating prostate cancer in a selected subject, the method comprising administering to the subject an agent that inhibits AKT/mTOR or MAPK signaling, wherein the subject is selected by detecting co-deletion of P TEN and PML or amplification of PPP1CA.

3. A method for treating prostate cancer in a selected subject, the method comprising administering to the subject an agent that inhibits AKT/mTOR or MAPK signaling, wherein the subject is selected by characterizing a biological sample of the subject for co-deletion of PTEN and PM ; amplification of PPP1CA ; and/or activation of SREBP.

4. The method of any of claims 1-3, wherein the agent that inhibits AKT/mTOR is selected from the group consisting of rapamycin, Temsirolimus, Everolimus, and Ridaforolimus; and the agent that inhibits MAPK signaling is selected from the group consisting of SB203580, SB202190, and BIRB-796.

5. The method of claim 3, wherein activation of SREBP is characterized by assaying the lipidomic profile of a biological sample of the patient.

6. The method of claim 5, wherein the lipidomic profile is assayed by detecting an increase in fatty acyl chains, membrane phospholipids, or other lipids.

7. The method of claim 6, wherein the lipidomic profile comprises alterations in

lysodimethylphosphatidylethanolamine, monoglyceride, phosphatidylglycerol, and lysophosphatidyl glycerol.

8. The method of any one of claims 1-7, wherein the method further comprises administering an inhibitor of fatty acid synthesis, inhibition of cholesterol synthesis, or inhibition of SREBP.

9. The method of claim 8, wherein the inhibitor of fatty acid synthesis is TOFA, the inhibitor of cholesterol synthesis is simvastatin, and the inhibitor of SREBP is Fatostatin.

10. The method of claim 4, wherein the method comprises administering to the subject a MEK inhibitor.

11. The method of claim 10, wherein the MEK inhibitor is EG0126.

12. The method of any one of claims 1-11, wherein the method inhibits tumor growth and/or metastasis.

13. The method of any one of claims 1-12, wherein the method further comprises prescribing a low fat diet for the subject.

14. The method of any one of claims 1-13, wherein the characterization indicates that the prostate cancer has a propensity to metastasize.