The present invention refers to a method and kit for determining the response of prostate cancer patients to androgen receptor (AR) antagonists by determining the expression levels of selected genes or the protein binding and histone modification profiles at super-enhancers (SEs). Especially, the invention is related to kits to assess how a patient with prostate cancer responds to treatment with an AR antagonist, especially darolutamide. In a further aspect, the invention is related to the use of an AR antagonist, especially darolutamide, for the treatment of prostate cancer in a patient by analyzing a sample of body fluid or tumor tissue in vitro and determining how the patient responds to treatment.

AR stands for androgen receptor, NR3C4, AIS8, DHTR, HUMARA, HYSP1, KD, SBMA, SMAX1 and TFM.

In humans, it is encoded by the AR gene (Gene ID 367) and has the NCBI reference sequence identifier NM 000044.6. The corresponding AR protein has the identifier NP 000035.2. Human AR cDNA and protein sequences are shown in J. Trapman et ak, Biochem. Biophys. Res. Commun., 1988, 153:241-248, C.S. Chang et ak, Science, 1988, 40:324-326 and D.B. Lubahn et ak, Science, 1988, 240:327-330.

AR is a nuclear receptor that is bound and activated by androgens, mainly dihydrotestosterone and testosterone. Upon ligand binding it translocates into the nucleus where it binds to specific DNA sequences and locally interacts with different cofactors, ultimately leading to stabilization and activation of the transcription complex and expression of downstream genes (B. Haendler, Biomed. Pharmacother., 2002, 56:78-83; F. Claessens et ak, Cell. Mol. Life Sck, 2017, 74:2217-2228). The AR plays an essential role in early and late-stage prostate cancer (P.E. Lonergan and D.J. Tindall, J. Carcinog., 2011, 10:20; Q. Feng and B. He, Front Oneok, 2019, 9:858) and is therefore a target of choice.

MED1 stands for Mediator of RNA polymerase II transcription subunit 1. It is also named CRSP1, CRSP200, DRIP205, DRIP230, PBP, PPARBP, PPARGBP, RB18A, TRAP220 and TRIP2.

In humans it is encoded by the MED1 gene (Gene ID 5469) and has the sequence identifier NM 004774.4. The corresponding MED1 protein has the identifier NP 004765.2. Human MED1 cDNA and protein sequences are shown in P. Drane et ak, Oncogene, 1997, 15:3013-3024. MED1 belongs to the mediator complex, an essential coactivator for gene transcription by RNA polymerase II (M.T. Knuesel and D.J. Taatjes, Transcription, 2011, 2:28-31). MED1 interacts with the zinc finger transcription factor SP1 which binds to GC-rich regions often found in gene promoters (C. Rachez et al., Nature, 1999, 398:824-828). It binds to many other transcription factors, including the AR (J.W. Russo, Cancer Discov., 2019, 9:1490-1492).

FOXA1 stands for forkhead box protein Al. It is also named HNF3A, TCF3A.

In humans it is encoded by the FOXA1 gene (Gene ID 3169) and has the sequence identifier NM 004496.5. The corresponding FOXA1 protein has the identifier NP 004487.2. Human FOXA1 cDNA and protein sequences are shown in E. Lai et al., Genes Dev., 1991, 5:416-427 and C.D. Bingle and S. Gowan, Biochim. Biophys. Acta, 1996, 1307:17-20 .

FOXA1 belongs to the forkhead family of DNA-binding proteins and acts as pioneer factor that opens up the chromatin structure and facilitates gene transcription (K.M. Jozwik and J.S. Carroll, Nat. Rev. Cancer, 2012, 12:381-385). It is an essential antagonist of the epithelial-to-mesenchymal cell transition (M. Katoh et al., Cancer Lett., 2013, 328: 198-206). It binds to the AR to modulate its function (Y. Zhao et al., Int. J. Biol. Sci., 2014, 10:614-619). FOXA1 mutations that define different prostate cancer subgroups have been identified (E.J. Adams et al., Nature, 2019, 57:408-412).

H3 stands for histone 3 family members. They are encoded by the following genes (Gene ID and NM identifier are given) which code for the proteins indicated under NP identifier :H3-3A (Gene ID: 3020; NM_002107.6; NP_002098.1), H3-3B (Gene ID: 3021; NM_005324.5; NP_005315.1), H3-4 (Gene ID: 8290; NM_003493.2; NP_003484.1), H3-5 (Gene ID: 440093; NM_001013699.3; NP_001013721.2), H3C1 (Gene ID 8350; NM_003529.2; NP_003529), H3C2 (Gene ID: 8358; NM_003537.3; NP_003528.1), H3C3 (Gene ID: 8352; NM_003531.2; NP_003522.1), H3C4 (Gene ID: 8351; NM_003530.4; NP_003521.2), H3C6 (Gene ID: 8353; NM_003532.2; NP_003523.1), H3C7 (Gene ID: 8968; NM_021018.2; NP_066298.1), H3C8 (Gene ID: 8355; NM_003534.2; NP_003525.1), H3C10 (Gene ID: 8357; NM_003536.2; NP_003527.1), H3C11 (Gene ID: 8354; NM_003533.2; NP_003524.1), H3C12 (Gene ID: 8356; NM_003535.2; NP_003526.1), H3C14 (Gene ID: 126961; NM_021059.2; NP_066403.2),.H3C15 (Gene ID 333932; NM_001005464.2; NP_001005464.1), H3Y1 (Gen ID: 391769; NM_001355258.2; NP_001342187.1), H3Y2 (Gene ID 340096; NM_001371919.1; NP_001358848.1).

Human histone H3s are encoded by several genes. The main groups are H3.1, which includes HIST1H3A, HIST1H3B, HIST1H3C, HIST1H3D, HIST1H3E, HIST1H3F, HIST1H3G, HIST1H3H, HIST1H3I and HIST1H3J, H3.2, which includes HIST2H3A, HIST2H3C, HIST2H3D, and H3.3, which includes H3F3A and H3F3B. An overview of histone H3 cDNA and protein sequences is given in D. Wells and C. McBride, Nucleic Acids Res., 1989, 17 Suppl., r311-r346. Additional details are given in S.J. Clark et al., Nucleic Acids Res., 1981, 9:1583-1590; R. Zhong et al., Nucleic Acids Res., 1983, 11, 7409-7425; D. Wells and L. Kedes, Proc. Natl. Acad. Sci., 1985, 82:2834-2838; F. Marashi et al., Biochem. Cell. Biol, 1986, 64:277-289; D. Wells et al., Nucleic Acids Res., 1987, 15:2871-2889; W. Albig et al., Genomics, 1991, 10:940-948; E. Kardalinou et al., J. Cell. Biochem., 1993, 52:375-383; W. Albig et al., Genomics, 1995, 30:264-272; W. Albig and D. Doenecke, Hum. Genet., 1997, 101:284-294; W.F. Marzluff et al., Genomics, 2002, 80:487-498.

Histone H3 proteins are part of globular nucleosomes and play an essential role in controlling access to DNA. They exist in several variants with highly conserved sequences. They undergo many post- translational modifications, especially acetylation, which is associated with active genes and accessible DNA regions. Lysine 27 (K27) is a main site of acetylation, but additional lysine residues such as K9, K14, K18 and K23 can also be acetylated in histone H3 (S. K. Kurdistani et al., Cell, 117:721-733).

SEs are dense cluster of genomic regions regulating gene transcription and which are dysregulated in various cancer types (Y. He et al., Frontiers Pharmacol., 2019, 10:361). These regions are highly bound by cell type-specific transcription factors and transcription-associated factors such as MED1, and are also highly enriched in histone acetylation, especially histone H3 acetylation (J. Loven et al., Cell, 2013, 153:320-334; W. A. Whyte et al., Cell, 2013, 153:307-319). SEs may represent specific vulnerabilities in different tumor types, including prostate cancer (V. Zuber et al., BMC Genomics, 2017, 18:270).

Targeted cancer drugs have a direct or indirect effect on one or more relevant biochemical pathways. On the other hand, it is well known that when treating patients suffering from cancer, not all of them will respond to the treatment similarly and there may be varying degrees of responses. The serum levels of kallikrein-3/prostate-specific antigen are currently being used for prostate cancer screening and also for post-treatment monitoring and to provide guidance for subsequent therapies (B. Danielson et al., Can.

Urol. Assoc. J., 2019, doi: 10.5489/cuaj.5600). However, questions have been raised about over-diagnosis and over-treatment due to false positive or negative results (M.C. Bemal-Soriano et al., 2019, 98:el7451). Thus, it would be very useful to identify improved markers for the response of patients to treatment, so that those with the best chance of benefiting from the drug are properly treated and monitored.

Identification of responders in the sense of the invention also means the identification of a patient or a group of patients with shared biological characteristics by using molecular, biochemical and diagnostic testing to select the optimal treatment for the patients and achieve the best possible outcome. However, nothing is presently disclosed that describes the impact of darolutamide on the expression of the genes listed in Tables 1-3, or of the levels of MED1, AR or FOXA1 binding, or of histone H3 acetylation levels, more specifically H3 K27 acetylation, at the SEs mentioned in Table 4.

The term AR is used in the present invention for the AR gene (Gene ID 367, http://www.ncbi.nlm.nih.gov/gene/367) and AR protein (P10275.3).

The term MED1 is used in the present invention for the MED1 gene (Gene ID 5469, http://www.ncbi.nlm.nih.gov/gene/5469) and MED1 protein (AAH60758.1)

The term FOXA1 is used in the present invention for the FOXA1 gene (Gene ID 3169, http://www.ncbi.nlm.nih.gov/gene/3169) and FOXA1 protein (AAH33890.1).

The term SEs is used in the present invention for all genomic regions bound by the AR and identified using the ROSE algorithm and MED1 signals (Fig. 1). The identified SE list is given in Table 4.

It is thus an object of the present invention to find a method for evaluating the response of prostate tumor patients to treatment with an AR antagonist.

It has now been found that a safe response monitoring is possible with the described inventive in vitro method for monitoring the response of prostate cancer patients to treatment with an AR antagonist.

The invention is an in vitro analysis method for determining how a patient suffering from prostate cancer responds to treatment with an AR antagonist by: i) determining the expression levels of the genes described in Tables 1-3 by measurement of the respective mRNA or derived cDNA expression levels in a sample of body fluid or tumor cells or tumor tissue of a treated patient, and comparing them with those measured before treatment, or with healthy prostate tissue samples and/or ii) determining the binding levels of MED1, AR or FOXA1 or the histone acetylation levels, more specifically histone H3, more specifically histone H3 K27 acetylation, at the SEs listed in Table 4 in a sample of body fluid or tumor cells or tumor tissue of a treated patient, and comparing them with those measured before treatment, or with healthy prostate tissue samples, and wherein the presence in said in vitro sample of an altered mRNA or derived cDNA level and/or reduced MED1, AR or FOXA1 binding, or histone H3 acetylation levels, more specifically H3 K27 acetylation, to SEs following treatment with an AR antagonist in comparison with a sample of the patient before treatment is suggestive of a better response to the treatment in said patient.

The invention is an in vitro analysis method for determining how a patient suffering from prostate cancer responds to treatment with darolutamide by: i) determining the expression level of the response markers described in Tables 1-3 by measurement of the respective mRNA or derived cDNA expression levels in a sample of body fluid or tumor cells or tumor tissue of treated patient, and comparing them with those measured before treatment, and/or ii) determining the binding levels of MED1, AR or FOXA1 or the histone H3 acetylation levels, more specifically H3 K27 acetylation, at the SEs listed in Table 4 in a sample of body fluid or tumor cells or tumor tissue of a treated patient, and comparing them with those measured before treatment, and wherein the presence in said in vitro sample of a modified mRNA or derived cDNA level and/or reduced MED1, AR or FOX1 binding, or histone H3 acetylation levels, more specifically H3 K27 acetylation, at SEs following treatment with darolutamide in comparison with a sample of the patient before treatment is suggestive of a better response to the treatment in said patient.

The determination of the expression level of the mRNA or derived cDNA from genes listed in Tables 1-3 and the determination of the bound levels of MED1, AR or FOXA1, or of the histone H3 acetylation levels, more specifically H3 K27 acetylation, at SEs listed in Table 4 can either be done combined, or separately. All combinations are possible to get a valuable result for pharmacodynamic response.

For example the following measurements are possible:

The respective mRNA or derived cDNA levels of the genes listed in Tables 1-3 combined with the MED1 protein binding level at the SEs listed in Table 4.

The respective mRNA or derived cDNA levels of the genes listed in Tables 1-3 combined with the AR protein binding level at the SEs listed in Table 4. The respective mRNA or derived cDNA levels of the genes listed in Tables 1-3 combined with the FOXA1 protein binding level at the SEs listed in Table 4

The respective mRNA or derived cDNA levels of the genes listed in Tables 1-3 combined with the histone H3 acetylation levels, more specifically H3 K27 acetylation, at the SEs listed in Table 4

The present invention concerns an analysis kit for monitoring the impact of an AR antagonist on treated prostate cancer patients.

In this regard the features are defined as follows:

Body fluid in the present invention means for example blood, plasma, serum, lymph, saliva, sweat, teardrops, urine or feces of a patient.

Tumor tissue in the present invention means for example primary tumor, metastases or circulating tumor cells.

An altered RNA level in a sample is suggestive of a better response to the treatment of prostate cancer in the patient, if the mRNA, cDNA or protein expression level is at least 4-fold different than before treatment.

More preferred is an expression level that is of at least 5 -fold different. It is also possible that an expression level is more than 6-fold different.

A reduced level of bound MED1 protein in a sample is suggestive of a better response to the treatment of prostate cancer in the patient, if the bound protein level is at least reduced 1.5 -fold than before treatment.

More preferred is a bound protein level that is at least 2-fold lower. It is also possible that a bound protein level is more than 2-fold lower.

A reduced level of bound AR protein in a sample is suggestive of a better response to the treatment of prostate cancer in the patient, if the bound protein level is at least reduced 1.5 -fold than before treatment.

More preferred is a bound protein level that is at least 2-fold lower. It is also possible that a bound protein level is more than 2-fold lower.

A reduced level of bound FOXA1 protein in a sample is suggestive of a better response to the treatment of prostate cancer in the patient, if the bound protein level is at least reduced 1.5 -fold than before treatment. More preferred is a bound protein level that is at least 2-fold lower. It is also possible that a bound protein level is more than 2-fold lower.

A reduced level of histone H3 acetylation levels, more specifically H3 K27 acetylation, in a sample is suggestive of a better response to the treatment of prostate cancer in the patient, if the bound protein level is at least reduced 1.5 -fold than before treatment.

More preferred is a histone H3 acetylation levels, more specifically H3 K27 acetylation, that is of at least 2 -fold lower. It is also possible that the histone H acetylation levels, more specifically H3 K27 acetylation, is more than 2-fold lower

A further aspect of the invention is the use of the method for in vitro analysis of prostate cancer in a patient. The patient is a mammal, especially a human.

Gene expression levels are assessed by determining the amount of RNA, for example mRNA or derived cDNA that is transcribed from a gene or gene sequence and coding for a peptide or protein. Today, the gene expression analysis can be done according to well-established and known processes. Methods for gene expression analysis include, but are not limited to, reverse transcription quantitative PCR, differential display PCR, hybridization-based microarrays and next-generation sequencing, including RNA-Seq (F. Ozsolak and P. M. Milos, Nat. Rev. Genet. 2011, 12:87-98).

For the measurement of gene expression, it is an advantage to amplify RNA, respectively cDNA. Today, well established processes are available for the generation of cDNA from an RNA template, using a reverse transcriptase (S. Hahn et ak, Cell. Mol. Life Sci., 2000, 57:96-105).

Gene expression profiles indicative of AR antagonist responders are preferably those which show at least a 4-fold difference following AR antagonist treatment with regard to the expression of the respective mRNA or derived cDNA of the genes listed in Tables 1-3.

An expression difference of 4-fold is clearly predictive of the influence of the AR antagonist on the prostate cancer patient. More preferred is a difference of 5 -fold and much more preferred is a difference of 6-fold, which more clearly indicates that the AR antagonist will inhibit the progression of the tumor.

Protein extracts can be prepared by methods including, but not limited to, ion exchange column, size exclusion chromatography, SDS polyacrylamide gel electrophoresis, high performance liquid chromatography or reversed-phase chromatography (N. E. Labrou, Methods Mol. Biol., 2014, 1129:3-10).

Protein levels and protein occupancy at chromatin can be measured by methods including, but not limited to, protein immuno staining and microscopy, immunoprecipitation, Immunoelectrophoresis, Western blot, spectrophotometry, mass spectrometry, radioimmunoassay and enzyme-linked immunosorbent assay, immuno-PCR, stable isotope labeling by amino acids, tissue microarrays, protein biochips, chromatin immunoprecipitation (ChIP), ChIP with subsequent deep DNA sequencing (ChIP-seq), proteomics and nanoproteomics (K. K. Jain, J. BUON, 2007, Suppl. 1 : S31 -S38: A. Brewis and P. Brennan, Adv. Protein Chem. Struct. Biol., 2010, 80:1-44; T. C. Collier and D. C. Muddiman, Amino Acids, 2012, 43:1109- 1117; S. E. Ong, Anal. Bioanal. Chem, 2012, 404:967-976; E. Rodriguez- Suarez and A. D. Whetton,

Mass Spectrom. Rev., 2013, 32:1-26).

Histone acetylation levels can be determined by liquid chromatography/mass spectrometry (Y. Zheng et al., Curr. Opin. Chem. Biol., 2016, 33:142-150), colorimetric and fluorometric assays, ELISA, Western blot analysis, immunocytochemistry, immunohistochemistry, ChIP, ChIP-seq (R.S. Jayani et al., Methods Cell Biol., 2010, 98:35-56), interaction assays based on fluorescence resonance energy transfer (K. Sasaki and M. Yoshida, Drug Discov. Today Technok, 2016, 19:51-56) or bioluminescence resonance energy transfer (M. Moustakim et al., Angew. Chem. Int. Ed. Engl., 2017, 56:827-831).

Protein occupancy and histone acetylation levels at SEs were measured by ChIP-seq in VCaP cells.

A bound MED1 protein level reduced 1.5-fold in patients clearly indicates that the patient responded to AR antagonist treatment. More preferred is a difference of 2-fold and much more preferred is a difference over 2-fold, which more clearly indicates that the AR antagonist will inhibit AR signaling.

A bound AR protein level reduced 1.5 -fold in patients clearly indicates that the patient responded to AR antagonist treatment. More preferred is a difference of 2-fold and much more preferred is a difference over 2 -fold, which more clearly indicates that the AR antagonist will inhibit AR signaling.

A bound FOXA1 protein level reduced 1.5-fold in patients clearly indicates that the patient responded to AR antagonist treatment. More preferred is a difference of 2-fold and much more preferred is a difference over 2-fold, which more clearly indicates that the AR antagonist will inhibit AR signaling.

A histone H3 acetylation level, more specifically H3 K27 acetylation, reduction 1.5-fold in patients clearly indicates that the patient responded to AR antagonist treatment. More preferred is a reduction of 2-fold and much more preferred is a reduction over 2-fold, which more clearly indicates that the AR antagonist will inhibit AR signaling.

Within the scope of the present invention, prostate cancer is understood as a disease of mammals, especially as a disease of the human and non-human mammal body, more specifically of the human body.

Prostate cancer in this regard means prostate adenocarcinoma, prostate glandular carcinoma, prostatic adenocarcinoma, prostate glandular cancer. AR antagonists that are known are for example those compounds that are disclosed in E.D. Crawford and al (J. Urol., 2018, 200:956-966) and A.E. Dellis and A.G. Papatsoris (Expert Opin. Pharmacother., 2019, 20:163-172). They include competitive AR antagonists such as cyproterone acetate, flutamide, bicalutamide, nilutamide, enzalutamide, apalutamide, darolutamide, proxalutamide. Other examples are compounds that address the AR N-terminal domain or DNA-binding domain, and AR degraders, including proteolysis-taegeting chimeras (C. Ferroni and G. Varchi, Current Med. Chem., 2019, 26:6053-6073).

In a further preferred embodiment of the method for identification of responders according to the invention, body fluid or body tissue, preferably blood, alternatively whole blood, serum or available plasma, is taken from the patient to be examined, and the analysis is made in vitro, respectively ex vivo, which means outside the human body.

Within the scope of the invention AR is to be understood as a human protein or polypeptide encoded by the AR gene shown in Gene ID 367 and having the amino acid sequence shown in P10275.3 , or a fragment of the AR protein sequence of at least 15 amino acids.

Within the scope of the invention MED1 is to be understood as a human protein or polypeptide encoded by the MED1 gene shown in Gene ID 5469 and having the amino acid sequence shown in AAH60758.1, or a fragment of the MED1 protein sequence of at least 15 amino acids.

Within the scope of the invention FOXA1 is to be understood as a human protein or polypeptide encoded by the FOXA1 gene shown in Gene ID 3169 and having the amino acid sequence shown in AAH33890.1, or a fragment of the FOXA1 protein sequence of at least 15 amino acids.

It may be beneficial to combine AR antagonists for example with one or more compounds selected from:

1311-chTNT, abarelix, abiraterone, aclarubicin, ado-trastuzumab emtansine, afatinib, aflibercept, aldesleukin, alemtuzumab, Alendronic acid, alitretinoin, altretamine, amifostine, aminoglutethimide,

Hexyl aminolevulinate,amrubicin, amsacrine, anastrozole, ancestim, anethole dithiolethione, angiotensin II, antithrombin III, apalutamide, aprepitant, arcitumomab, arglabin, arsenic trioxide, asparaginase, atezolizumab, avelumab, axitinib, azacitidine, basiliximab, BAY 1895344, belotecan, bendamustine, belinostat, bevacizumab, bexarotene, bicalutamide, bisantrene, bleomycin, bortezomib, buserelin, bosutinib, brentuximab vedotin, busulfan, cabazitaxel, cabozantinib, calcium folinate, calcium levofolinate, camrelizumab, capecitabine, capromab, carboplatin, carfilzomib, carmofur, carmustine, catumaxomab, celecoxib, celmoleukin, cemiplimab, ceritinib, cetuximab, chlorambucil, chlormadinone, chlormethine, cidofovir, cinacalcet, cisplatin, cladribine, clodronic acid, clofarabine, copanlisib, crisantaspase, cyclophosphamide, cyproterone, cytarabine, dacarbazine, dactinomycin, darbepoetin alfa, dabrafenib, dasatinib, daunorubicin, decitabine, degarelix, denileukin diftitox, denosumab, depreotide, deslorelin, dexrazoxane, dibrospidium chloride, dianhydrogalactitol, diclofenac, docetaxel, dolasetron, doxifluridine, doxorubicin, doxorubicin + estrone, dronabinol, durvalumab, eculizumab, edrecolomab, elliptinium acetate, eltrombopag, endostatin, enocitabine, enzalutamide, epirubicin, epitiostanol, epoetin alfa, epoetin beta, epoetin zeta, eptaplatin, eribulin, erlotinib, esomeprazole, estradiol, estramustine, etoposide, everolimus, exemestane, fadrozole, fentanyl, filgrastim, fluoxymesterone, floxuridine, fludarabine, fluorouracil, flutamide, folinic acid, formestane, fosaprepitant, fotemustine, fulvestrant, gadobutrol, gadoteridol, gadoteric acid meglumine, gadoversetamide, gadoxetic acid, gallium nitrate, ganirelix, gefitinib, gemcitabine, gemtuzumab, Glucarpidase, glutoxim, GM-CSF, goserelin, granisetron, granulocyte colony stimulating factor, histamine dihydrochloride, histrelin, hydroxycarbamide, 1-125 seeds, lansoprazole, ibandronic acid, ibritumomab tiuxetan, ibrutinib, idarubicin, ifosfamide, imatinib, imiquimod, improsulfan, indisetron, incadronic acid, ingenol mebutate, interferon alfa, interferon beta, interferon gamma, iobitridol, iobenguane (1231), iomeprol, ipilimumab, irinotecan, Itraconazole, ixabepilone, lanreotide, lapatinib, Iasocholine, larotrectinib, lenalidomide, lenograstim, lentinan, letrozole, leuprorelin, levamisole, levonorgestrel, levothyroxine sodium, lisuride, lobaplatin, lomustine, lonidamine, masoprocol, medroxyprogesterone, megestrol, melarsoprol, melphalan, mepitiostane, mercaptopurine, mesna, methadone, methotrexate, methoxsalen, methylaminolevulinate, methylprednisolone, methyltestosterone, metirosine, mifamurtide, miltefosine, miriplatin, mitobronitol, mitoguazone, mitolactol, mitomycin, mitotane, mitoxantrone, mogamulizumab, molgramostim, mopidamol, morphine hydrochloride, morphine sulfate, nabilone, nabiximols, nafarelin, naloxone + pentazocine, naltrexone, nartograstim, nedaplatin, nelarabine, neridronic acid, nivolumabpentetreotide, nilotinib, nilutamide, nimorazole, nimotuzumab, nimustine, niraparib, nitracrine, nivolumab, obinutuzumab, octreotide, ofatumumab, olaparib, omacetaxine mepesuccinate, omeprazole, ondansetron, oprelvekin, orgotein, orilotimod, oxaliplatin, oxycodone, oxymetholone, ozogamicine, p53 gene therapy, paclitaxel, palifermin, palladium- 103 seed, palonosetron, pamidronic acid, pamiparib, panitumumab, pantoprazole, pazopanib, pegaspargase, PEG-epoetin beta (methoxy PEG-epoetin beta), pembrolizumab, pegfilgrastim, peginterferon alfa-2b, pembrolizumab, pemetrexed, pentazocine, pentostatin, peplomycin, Perflubutane, perfosfamide, Pertuzumab, picibanil, pilocarpine, pirarubicin, pixantrone, plerixafor, plicamycin, poliglusam, polyestradiol phosphate, polyvinylpyrrolidone + sodium hyaluronate, polysaccharide-K, pomalidomide, ponatinib, porfimer sodium, pralatrexate, prednimustine, prednisone, procarbazine, procodazole, propranolol, PSMA-TTC, quinagolide, rabeprazole, racotumomab, radium-223 chloride, radotinib, raloxifene, raltitrexed, ramosetron, ramucirumab, ranimustine, rasburicase, razoxane, refametinib, regorafenib, ribociclib, risedronic acid, rhenium- 186 etidronate, rituximab, romidepsin, romiplostim, romurtide, roniciclib, rucaparib, samarium (153Sm) lexidronam, sargramostim, satumomab, secretin, sipuleucel-T, sizofiran, sobuzoxane, sodium glycididazole, sorafenib, stanozolol, streptozocin, sunitinib, talaporfin, talazoparib, tamibarotene, tamoxifen, tapentadol, tasonermin, teceleukin, technetium (99mTc) nofetumomab merpentan, 99mTc-HYNIC-[Tyr3]-octreotide, tegafur, tegafur + gimeracil + oteracil, temoporfin, temozolomide, temsirolimus, teniposide, testosterone, tetrofosmin, thalidomide, thiotepa, thymalfasin, thyrotropin alfa, tioguanine, tocilizumab, topotecan, toremifene, tositumomab, trabectedin, tramadol, trastuzumab, trastuzumab emtansine, treosulfan, tretinoin, trifluridine + tipiracil, trilostane, triptorelin, trametinib, trofosfamide, thrombopoietin, tryptophan, ubenimex, valrubicin, vandetanib, vapreotide, veliparib, vemurafenib, vinblastine, vincristine, vindesine, vinflunine, vinorelbine, vismodegib, vorinostat, vorozole, yttrium-90 glass microspheres, zinostatin, zinostatin stimalamer, zoledronic acid, zorubicin.

The medication may comprise an AR antagonist and a therapeutic composition selected from the group consisting of tyrosine kinase inhibitors, CDK inhibitors, MEK inhibitors, PI3K inhibitors, MAP kinase inhibitors, Aik inhibitors, mTOR inhibitors, P-TEFb inhibitors, ATR inhibitors, PARP inhibitors, apoptosis modulators, hedgehog inhibitors, proteasome inhibitors, HDAC inhibitors, methotrexate, dexamethasone, PSMA-based compounds and combinations thereof.

The invention further relates to a kit for in vitro analysis of the response of a prostate cancer patient to treatment with an AR antagonist containing the steps i) determining the expression level of the genes listed in Tables 1-3 by measurement of the respective mRNA or derived cDNA expression levels in a sample of body fluid or tumor tissue of a treated patient, and comparing the expression level with that before treatment, and/or ii) determining the bound MED1, AR or FOXA1, or histone H3 acetylation levels, more specifically H3 K27 acetylation, at the SEs listed in Table 4 in a sample of body fluid or tumor tissue of a treated patient, and comparing it with the levels before treatment, wherein the presence in said in vitro sample of a modified mRNA or derived cDNA and/or reduced bound MED1, AR or FOXA1, or reduced histone H3 acetylation levels, more specifically H3 K27 acetylation, following treatment with an AR antagonist in comparison with the untreated patient is suggestive of a better response of the patient.

Within the kit, the determination of the expression level of the mRNA or derived cDNA and the determination of the bound MED1, AR or FOXA1, or histone H3 acetylation levels, more specifically H3 K27 acetylation, can either be done combined, or separately. When using the kit, all combinations are possible to get a valuable result for pharmacodynamic response.

Brief description of the abbreviations

AR Androgen receptor

ATCC American Tissue Culture Collection

ChIP Chromatin immunoprecipitation ChIP-seq Chromatin immunoprecipitation with deep sequencing chr Chromosome

DMEM Dulbecco's modified Eagle's medium DMSO Dimethyl sulfoxide FOXA1 Forkhead box protein A1 H3 Histone 3

H3K27ac Histone 3 lysine 27 acetylation HEPES 4-(2-hydroxyethyl)- 1 -piperazineethanesulfonic acid MED1 Mediator of RNA polymerase II transcription subunit 1 R1881 Methyltrieneolone RPMI Roswell Park Memorial Institute SE Super-enhancer

BRIEF DESCRIPTION OF THE FIGURES

Table 1 shows the genes with over 6-fold down or up-regulation following treatment of VCaP cells stimulated with 1 nM R1881 and the AR antagonist darolutamide (2 mM for 8 or 22 hours), in comparison to R1881 treatment only, as measured by sequencing on a hiSeq2500 device and mapping of FASTQ reads to the human genome GRCh38 and quantification with featureCounts from the Subread package. Differentially expressed genes were identified with DESeq2.

Table 2 shows the genes with 5- to 6-fold down or up-regulation following treatment of VCaP cells stimulated with 1 nM R1881 and the AR antagonist darolutamide (2 mM for 8 or 22 hours), in comparison to R1881 treatment only, as measured by sequencing on a hiSeq2500 device and mapping of FASTQ reads to the human genome GRCh38 and quantification with featureCounts from the Subread package. Differentially expressed genes were identified with DESeq2.

Table 3 shows the genes with 4- to 5-fold down or up-regulation following treatment of VCaP cells stimulated with 1 nM R1881 and the AR antagonist darolutamide (2 mM for 8 or 22 hours), in comparison to R1881 treatment only, as measured by sequencing on a hiSeq2500 device and mapping of FASTQ reads to the human genome GRCh38 and quantification with featureCounts from the Subread package. Differentially expressed genes were identified with DESeq2.

Table 4 shows the list of genomic coordinates (assembly GRCh37) of SEs changing in bound MED1,

AR, FOXA1 and histone H3 acetylation levels following treatment with 1 nM R1881 in comparison to DMSO treatment only.

Table 5 shows the ratio of medians of mean values of protein occupancy and H3K27 acetylation levels for the treatment conditions indicated, as depicted in Fig. 2.

Fig. 1 shows the detection of SEs at AR-binding genomic regions using the ROSE algorithm and MED1 signals as cut-off threshold.

Fig. 2 shows the averaged signals of transcription-associated factors and histone H3 K27 acetylation at scaled SE regions after treatment of VCaP cells with DMSO, R1881 (1 nM) or darolutamide (2 mM) + R1881 (1 nM). Average signals for MED1 binding at SEs are reduced 2-fold between the R1881 and the R1881 + darolutamide groups. Average signals for AR binding at SEs are reduced 1.5-fold between the R1881 and the R1881 + darolutamide groups. Average signals for FOXA1 binding are reduced 1.5-fold between the R1881 and the R1881 + darolutamide groups. Average signals for H3 K27 acetylation are reduced 1.5-fold between the R1881 and the R1881 + darolutamide groups.

BIOLOGICAL EXAMPLES The following examples describe the feasibility of the present invention. The VCaP cell line (ATCC CRL2876) was selected as model for androgen-dependent prostate cancer.

1. CELL CULTURE

VCaP cells were routinely cultured in an incubator at 37°C with 5% carbon dioxide in DMEM/glutamine medium. For the assays the cells were starved for 2 days in RPMI1640 without phenol red supplemented with 10% charcoal-stripped, heat-inactivated and filtered fetal bovine serum, before treatment with the synthetic androgen R1881 (methyltrienolone) at a concentration of 1 nM. For gene expression studies, darolutamide was added at a final concentration of 2 mM, and the cells were harvested after 8 or 22 hours post-treatment. For protein binding studies, darolutamide was added at a concentration of 2 pM and the cells harvested 22 hours post-treatment.

2. RNA EXTRACTION AND SEQUENCING

Cells were lyzed and RNA was isolated using RNeasy columns with on-column DNA digestion, as described by the manufacturer (Qiagen, Hilden, Germany). RNA library preparation was performed after mRNA purification using poly-T beads, as described by the manufacturer (TruSeq Stranded mRNA Kit; Illumina, San Diego, CA, USA). Five biological replicates per condition were sequenced on a hiSeq2500 device via single-end, 50 base-pair reads with an average depth of 21 million reads per sample (Illumina, HiSeq2500 HTv4, SR, dual-indexing, 50 cycles).

3. DETERMINATION OF EXPRESSION LEVELS

FASTQ reads were mapped via STAR aligner to the human genome GRCh38 and quantified with featureCounts from the Subread package (Y. Liao, Nucleic Acids Res., 2019, 47:e47). Differentially expressed genes were identified with DESeq2 (M.I. Love et al., Genome Biol., 2014, 15:550). Significantly regulated genes were defined as having adjusted p-values lower than 0.05 and absolute log2- fold values higher than one for at least one time point.

4. PREPARATION OF ChIP-SEQ SAMPLES

Three replicates of eight million cells were treated with 2 pM darolutamide plus 1 nM R1881 for 22 h, subsequently fixed with 1% formaldehyde and processed in the case of AR and H3K27 acetylation as described in S.J. Baumgart et al., Nucleic Acids Res., 2017, 45:7722-7735, and in the case of MED1 and FOXA1 as described in iDeal ChIP-seq kit for Transcription Factors, Diagenode, Seraing, Belgium. For each ChIP reaction, 3 million cells were used and probed with antibodies specific for MED1 (A300-793A, Bethyl Laboratories, Montgomery, TX, USA), for AR (ab74272, Abeam, Cambridge, UK), for FOXA1 (ab23738, Abeam) or for histone H3 K27 acetylation (C15410196, Diagenode, Seraing, Belgium). ChIP experiments were performed in biological triplicates and library preparation was done as described by the manufacturer (MicroPlex Library Preparation Kit v2; Diagenode SA, Seraing, Belgium). The libraries were sequenced on a HiSeq2500 Illumina machine with 50 base pair, single-end reads to an average depth of 25-30 million reads per sample. The February 2009 human reference sequence (GRCh37) was used for mapping of the SE regions.

5. ChIP-SEQ BIOINFORMATIC ANALYSIS

Sequencing reads were mapped to human genome GRCh37 using the BWA aligner with default settings (H. Li and R. Durbin, Bio informatics, 2010, 26:589-595). Narrow peaks were called by MACS2 (Y.

Zhang et al., Genome Biol., 2008, 9:R137) with default parameters and a q-value cutoff of 0.05. Human genome blacklisted regions (ENCODE consortium) were excluded from further analysis. Peaks present in at least two replicates were used for further analysis. BAM files were merged and converted to bigwig format via Deeptools2 bamCoverage with default parameters and reads per kilo base million normalization (RPKM) (A.S. Richter et al., Nucleic Acids Res., 2016, 44:W160-W165).

6. DETECTION OF SE REGIONS

The ROSE algorithm was used to define SEs. Default conditions were applied, and transcriptional start sites regions excluded (J. Loven et al., Cell, 2013, 153:320-334; W.A. Whyte et al., Cell, 2013, 153:307- 319). We identified SEs in the VCaP cell line based on AR binding sites present after treatment with 1 nM R1881 for 22 hours. The MED1 signal was used to select groups of AR-binding regions qualifying as SEs with a cut-off set by the ROSE algorithm at 2737.7 (Fig. 1).

Signals of MED 1, AR, FOXA1 and histone H3 K27 acetylation were averaged in VCaP cells treated with DMSO only, or with R1881 (1 nM), or with darolutamide (2 mM) plus R1881 (1 nM) for 22 hours (Fig. 2). Mean values of RPKM normalized signals were measured at SEs scaled to the length of the shortest SE. The mean values of all SEs or of the top 20 SEs were compared between treatments using boxplots (Table 5). The boxplots show the minimum, first quartile, median, third quartile, and maximum, and outliers are not shown. The ratios between the median values of the treatments were compared. Table 1

List of gene transcripts down- or up-regulated over 6-fold in the 2 mM darolutamide + 1 nM R1881 samples, compared to the 1 nM R1881 samples, following treatment for 8 or 22 hours.

Table 2

List of gene transcripts down- or up-regulated 5- to 6-fold in the 2 mM darolutamide + 1 nM R1881 samples, compared to the 1 nM R1881 samples, following treatment for 8 or 22 hours.

Table 3 List of gene transcripts down- or up-regulated 4- to 5-fold in the 2 mM darolutamide + 1 nM R1881 samples, compared to the 1 nM R1881 samples, following treatment for 8 or 22 hours.

Table 4

List of SEs where MED1, AR and FOXA1 binding, and histone H3 K27 acetylation is strongly reduced in the 2 mM darolutamide + 1 nM R1881 samples, compared to the 1 nM R1881 samples, following treatment for 22 hours. The February 2009 human reference sequence GRCh37 was used as reference for mapping the SE regions. Ranking was made according to descending MED1 occupancy strength when comparing the 1 nM R1881 -treated to the DMSO group at the 22 hours time point.

Table 5

Ratio of medians of mean values of protein occupancy and H3K27 acetylation levels for the treatment conditions indicated, as shown in Fig. 2. Figure 1

Detection of SEs with the ROSE algorithm by using MED1 signals for SE cut-off measurements and AR binding sites as genomic regions qualifying for SE. The cut-off for SE definition is shown by the dotted horizontal grey line. Each red dot indicates an AR binding cluster either above or below the SE threshold.

Figure 2

Averaged signals of bound factors and histone H3 lysine 27 acetylation at scaled SE regions after treatment of VCaP cells with DMSO, R1881 (1 nM) or darolutamide (2 mM) + R1881 (1 nM) for 22 hours. AR-bound SEs were detected with the ROSE algorithm by using MED1 signals as cut-off threshold (see Fig. 1). The results are shown for the top 20 SEs defined in Table 2 (top) and for all SEs identified (bottom).