WO/2010/018731 | A MAMMALIAN RNA DEPENDENT RNA POLYMERASE |
WO/2004/048540 | MRACS AS MODIFIERS OF THE RAC PATHWAY AND METHODS OF USE |
JP2009523452 | Methods for measuring kinase activity |
GARCÍA CAÑAVERAS JUAN CARLOS (ES)
JUAN VIDAL OSCAR (ES)
HERVÁS DAVID (ES)
CARRETERO ASUNCIÓN JULIÁN (ES)
PULIDO ENDRINO INÉS (ES)
MENA MOLLA SALVADOR (ES)
INSA MOLLÁ AMELIA (ES)
MARTÍN MARTORELL PALOMA (ES)
UNIV VALENCIA (ES)
INSTITUTO DE INVESTIG SANITARIA INCLIVA (ES)
PULIDO I ET AL: "Metabolic vulnerabilities of mesenchymal-like EGFR-mutant NSCLC cells with acquired resistance to tyrosine kinase inhibitors", vol. 78, no. 13, Supplement 1, 1 July 2018 (2018-07-01), pages LB - 099, XP009542911, ISSN: 1538-7445, Retrieved from the Internet
LI XIAO-YU ET AL: "Nicotinamide N-Methyltransferase: A Promising Biomarker and Target for Human Cancer Therapy", FRONTIERS IN ONCOLOGY, vol. 12, 9 June 2022 (2022-06-09), CH, XP093029003, ISSN: 2234-943X, DOI: 10.3389/fonc.2022.894744
MOTOAKI KOMATSUTAKESHI KANDAHIDENORI URAIARATA KUROKOCHIRINA KITAHAMASHUHEI SHIGAKITAKASHI ONOHIDEO YUKIOKAKAZUHIRO HASEGAWAHIROBU: "NNMT activation can contribute to the development of fatty liver disease by modulating the NAD + metabolism", SCI REP, vol. 8, no. 1, 5 June 2018 (2018-06-05), pages 8637
HARSHINI NEELAKANTAN ET AL.: "Small molecule nicotinamide N-methyltransferase inhibitor activates senescent muscle stem cells and improves regenerative capacity of aged skeletal muscle", BIOCHEM PHARMACOL., vol. 163, May 2019 (2019-05-01), pages 481 - 492, XP085659298, DOI: 10.1016/j.bcp.2019.02.008
DANG LWANG Y: "Prognostic value of nicotinamide N-methyltransferase in human cancers: Evidence from a meta-analysis and database validation", OPEN MED (WARS), vol. 17, no. 1, 14 February 2022 (2022-02-14), pages 292 - 303
WANG WYANG CWANG TDENG H: "Complex roles of nicotinamide N-methyltransferase in cancer progression", CELL DEATH DIS., vol. 13, no. 3, 25 March 2022 (2022-03-25), pages 267
OGAWA, M.TANAKA, A.NAMBA, K. ET AL.: "Tumor stromal nicotinamide N-methyltransferase overexpression as a prognostic biomarker for poor clinical outcome in early-stage colorectal cancer", SCI REP, vol. 12, 2022, pages 2767, Retrieved from the Internet
KOCINAJ ACHAUDHURY TUDDIN MS ET AL.: "High Expression of Nicotinamide N-Methyltransferase in Patients with Sporadic Alzheimer's Disease", MOL NEUROBIOL., vol. 58, no. 4, 2021, pages 1769 - 1781, XP037389039, DOI: 10.1007/s12035-020-02259-9
SAMPSON, C.M.DIMET, A.L.NEELAKANTAN, H. ET AL.: "Combined nicotinamide N-methyltransferase inhibition and reduced-calorie diet normalizes body composition and enhances metabolic benefits in obese mice", SCI REP, vol. 11, 2021, pages 5637, Retrieved from the Internet
BROZOVIC, A.IZRAELY, S. ET AL.: "Cancer drug resistance induced by FMT.- novel therapeutic strategies", ARCH TOXICOL, vol. 95, 2021, pages 2279 - 2297, XP037494085, Retrieved from the Internet
PADHYE, APARNA ET AL.: "Targeting CDK4 overcomes EMT mediated tumor heterogeneity and therapeutic resistance in KRAS-mutant lung cancer", JCI INSIGHT, vol. 6, no. 17, 8 September 2021 (2021-09-08), pages 148392
ESKIOCAK, B.MCMILLAN, E. A.MENDIRATTA, S.KOLLIPARA, R. K.ZHANG, H.HUMPHRIES, C. G.WANG, C.GARCIA-RODRIGUEZ, J.DING, M.ZAMAN, A.: "Biomarker Accessible and Chemically Addressable Mechanistic Subtypes of BRAF Melanoma", CANCER DISCOVERY, vol. 7, no. 8, 2017, pages 832 - 851, XP055611713, Retrieved from the Internet
WANG YZENG JWU W ET AL.: "Nicotinamide N-methyltransferase enhances chemoresistance in breast cancer through SIRT1 protein stabilization", BREAST CANCER RES., vol. 21, no. 1, 17 May 2019 (2019-05-17), pages 64
WANG, Y.ZENG, J.WU, W. ET AL.: "Nicotinamide N-methyltransferase enhances chemoresistance in breast cancer through SIRT1 protein stabilization", BREAST CANCER RES, vol. 21, 2019, pages 64, Retrieved from the Internet
SHLOMI TRABINOWITZ JD: "Metabolism: Cancer mistunes methylation", NAT CHEM BIOL., vol. 9, no. 5, 2013, pages 293 - 294
YONGZHI GAONATHANIEL I MARTINMATTHIJS J VAN HAREN: "Nicotinamide N-methyl transferase (NNMT): An emerging therapeutic target", DRUG DISCOV TODAY, vol. 26, no. 2, November 2021 (2021-11-01), pages 2699 - 2706
GAO YMARTIN NIVAN HAREN MJ: "Nicotinamide N-methyl transferase (NNMT): An emerging therapeutic target", DRUG DISCOV TODAY, vol. 26, 21 May 2021 (2021-05-21), pages 2699 - 2706
KIM DSHAM WSJANG WSCHO KSCHOI YDKANG SKIM BKIM KJLIM EJRHA SY: "Scale-Up Evaluation of a Composite Tumor Marker Assay for the Early Detection of Renal Cell Carcinoma", DIAGNOSTICS, vol. 10, no. 10, 25 September 2020 (2020-09-25), pages 750
CAMPAGNA RPOZZI VSPINELLI G ET AL.: "The Utility of Nicotinamide N-Methyltransferase as a Potential Biomarker to Predict the Oncological Outcomes for Urological Cancers: An Update", BIOMOLECULES, vol. 11, no. 8, 16 August 2021 (2021-08-16), pages 1214
KANNT, A.RAJAGOPAL, S.ADNUR, S. V. ET AL.: "A small molecule inhibitor of Nicotinamide N-methyltransferase for the treatment of metabolic disorders", SCI REP, vol. 8, 2018, pages 3660, Retrieved from the Internet
MUHAMMAD, N.LEE, H.M.KIM, J.: "Oncology Therapeutics Targeting the Metabolism of Amino Acids", CELLS, vol. 9, 2020, pages 1904
LYNCH, T.J. ET AL.: "Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib", N ENGL J MED, vol. 350, no. 21, 2004, pages 2129 - 39, XP002447439, DOI: 10.1056/NEJMoa040938
SHEVCHENKO AJENSEN ONPODTELEJNIKOV AVSAGLIOCCO FWILM MVORM OMORTENSEN PBOUCHERIE HMANN M: "Linking genome and proteome by mass spectrometry: large-scale identification of yeast proteins from two dimensional gels", PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA, vol. 93, 1996, pages 14440 - 14445, XP000960780, DOI: 10.1073/pnas.93.25.14440
CLAIMS 1. In vitro method for assessing the activity of nicotinamide N-methyltransferase (NNMT) in a biological sample which comprises: a. Determining the concentration level of at least the three metabolites: 1-MNA, 5-MTA and SAM; b. Wherein the identification of a concentration level of 1-MNA and SAM higher than a pre-established threshold level and a concentration of 5-MTA lower than a pre-established threshold level is an indication of high NNMT activity; or wherein the identification of a concentration level of 1-MNA and SAM lower than a pre-established threshold level and a concentration of 5-MTA higher than a pre-established threshold level is an indication of low NNMT activity. 2. In vitro method for selecting patients who may benefit from a treatment with NNMT inhibitors which comprises assessing the NNMT activity following the method of claim 1, wherein the identification of high NNMT activity is an indication that the patients may benefit from a treatment with NNMT inhibitors. 3. In vitro method for the prognosis and/or predicting the response of patients to NNMT inhibitors which comprises assessing the NNMT activity following the method of claim 1, wherein the identification of high NNMT activity is an indication of poor prognosis and/or that the patients may response to a treatment with NNMT inhibitors. 4. In vitro method for predicting whether a patient is resistant to a treatment with tyrosine kinase inhibitors which comprises the identification of high NMMT activity following the method of claim 1. 5. In vitro method for screening NNMT inhibitor candidates which comprises: a. Determining the concentration level of at least the three metabolites: 1-MNA, 5-MTA and SAM; b. Wherein the identification of a concentration level of 1-MNA and SAM lower than a pre-established threshold level and a concentration level of 5-MTA higher than a pre-established threshold level, is an indication that the candidate is a NNMT inhibitor. 6. In vitro method, according to any of the previous claims, wherein the patient is suffering from lung cancer, preferably from Non-small cell lung cancer, and most preferably from Non-small cell lung cancer with EGFR mutation. 7. In vitro method, according to any of the previous claims, wherein the biological sample obtained from the patients is serum, blood, plasma, sputum, biopsy, urine or cell cultures or cell supernatants. 8. In vitro method, according to any of the previous claims, wherein the method is a mass spectrometry-based method. 9. In vitro use of at least the metabolites: 1-MNA, 5-MTA and SAM, for assessing NNMT activity, for selecting patients which may benefit from a treatment with NNMT inhibitors, for the prognosis and/or predicting the response of patients to NNMT inhibitors, for predicting whether a patient is resistant to a treatment with tyrosine kinase inhibitors and/or for screening NNMT inhibitor candidates. 10. NNMT inhibitors for use in the treatment of patients, wherein the biological sample obtained from the patients is characterized by a concentration level of: a. 1-MNA and SAM higher than a pre-established threshold level; and b. 5-MTA lower than a pre-established threshold level. |
Table 3 Performance of 24-months overall survival prediction models of TKIs using 1-MNA, 5-MTA and other metabolites from the group NNMT-related metabolites. Area under the ROC curve (AUC), sensitivity and specificity for GBM model using combinations of 1-MNA, 5-MTA and a third metabolite from Nicotinamide, Methionine, Cystathionine, Cystine, Cysteine, Homocysteine, GluCys, SAH, SAM, CSSG, GSH, Spermine and/or Spermidine using UPLC-MS/MS quantification in blood of TKI-treated NSCLC patients against Overall Survival (OS) at 24 months. Consequently, the first embodiment of the present invention refers to an in vitro method for assessing the activity of nicotinamide N-methyltransferase (NNMT) in a biological sample (hereinafter “method of the invention”) which comprises: a) Determining the concentration level of at least a metabolite selected from the group comprising: 1-MNA, 5-MTA, CSSG, Cystathionine, Cysteine, Cystine, GluCys, GSH, Homocysteine, Methionine, Nicotinamide, SAH, SAM, Spermidine and/or Spermine, or any combination thereof; b) Wherein the identification of a concentration level of 1-MNA, Cystathionine, Cystine, Homocysteine, Nicotinamide, SAH, SAM, Spermidine and/or Spermine higher than a pre-established threshold level is an indication of high NNMT activity; or wherein the identification of a concentration level of 1-MNA, Cystathionine, Cystine, Homocysteine, Nicotinamide, SAH, SAM, Spermidine and/or Spermine lower than a pre-established threshold level is an indication of low NNMT activity; and c) Wherein the identification of a concentration level of 5-MTA, CSSG, Cysteine, GluCys, GSH and/or Methionine lower than a pre-established threshold level is an indication of high NNMT activity; or wherein the identification of a concentration level of 5-MTA, CSSG, Cysteine, GluCys, GSH and/or Methionine higher than a pre-established threshold level is an indication of low NNMT activity. Kindly note that the activity of a protein to be used as a biomarker does not always depend on, or is not always directly related to, expression levels or protein levels because, for example, there could be mutations or other interactions/collaborations with other activities/pathways that affect the activity, so that if we measure only the expression or protein level, we may not be correctly identifying the activity of the protein. In a preferred embodiment, the present invention refers to an in vitro method for selecting patients who may benefit from a treatment with NNMT inhibitors which comprises assessing the NNMT activity following the method of the invention, wherein the identification of high NNMT activity is an indication that the patients may benefit from a treatment with NNMT inhibitors. In a preferred embodiment, the present invention refers to an in vitro method for the prognosis and/or predicting the response of patients to NNMT inhibitors which comprises assessing the NNMT activity following the method of the invention, wherein the identification of high NNMT activity is an indication of poor prognosis and that the patients may response to a treatment with NNMT inhibitors. In a preferred embodiment, the present invention refers to an in vitro method for predicting whether a patient is resistant to a treatment with tyrosine kinase inhibitors which comprises the identification of high NMMT activity following the method of the invention. In a preferred embodiment, the present invention refers to an in vitro method for screening NNMT inhibitor candidates which comprises: a) Determining the concentration level of at least a metabolite selected from the group comprising: 1-MNA, 5-MTA, CSSG, Cystathionine, Cysteine, Cystine, GluCys, GSH, Homocysteine, Methionine, Nicotinamide, SAH, SAM, Spermidine, and/or Spermine, or any combination thereof, after administering the candidate NNMT inhibitor; a b) wherein the identification of a concentration level of 1-MNA, Cystathionine, Cystine, Homocysteine, Nicotinamide, SAH, SAM, Spermidine and/or Spermine lower than a pre-established threshold level; or wherein the identification of a concentration level of 5-MTA, CSSG, Cysteine, GluCys, GSH and/or Methionine higher than a pre-established threshold level, is an indication of the candidate is a NNMT inhibitor. In a preferred embodiment, the method of the invention comprises determining the concentration level of at two metabolites selected from the group: 1-MNA, 5-MTA, CSSG, Cystathionine, Cysteine, Cystine, GluCys, GSH, Homocysteine, Methionine, Nicotinamide, SAH, SAM, Spermidine and/or Spermine. In a preferred embodiment, the method of the invention comprises determining the concentration level of at least the following two metabolites: 1-MNA and 5-MTA, 1-MNA and SAM, and/or 5-MTA and SAM. In a preferred embodiment, the method of the invention comprises determining the concentration level of at two metabolites selected from the group: 1-MNA, 5-MTA, CSSG, Cystathionine, Cysteine, Cystine, GluCys, GSH, Homocysteine, Methionine, Nicotinamide, SAH, Spermidine and/or Spermine; in combination with SAM. In a preferred embodiment, the method of the invention comprises determining the concentration level of at least the following three metabolites: 1-MNA, 5-MTA and SAM. In a preferred embodiment, the method of the invention comprises the patient is suffering from lung cancer, preferably from non-small cell lung cancer, and most preferably from non-small cell lung cancer with EGFR mutation. In a preferred embodiment, the method of the invention comprises the biological sample obtained from the patients is serum, blood, plasma, sputum, biopsy, urine or cell cultures or cell supernatants. In a preferred embodiment, the method of the invention is carried by by mass spectrometry- based method. The second embodiment of the invention refers to the in vitro use of at least a metabolite selected from the group comprising: 1-MNA, 5-MTA, CSSG, Cystathionine, Cysteine, Cystine, GluCys, GSH, Homocysteine, Methionine, Nicotinamide, SAH, SAM, SAM, Spermidine and/or Spermine, or any combination thereof, for assessing NNMT activity, for selecting patients which may benefit from a treatment with NNMT inhibitors, for the prognosis and/or predicting the response of patients to NNMT inhibitors, for predicting whether a patient is resistant to a treatment with tyrosine kinase inhibitors and/or for screening NNMT inhibitor candidates. The third embodiment of the present invention refers to NNMT inhibitors for use in the treatment of patients, wherein the biological sample obtained from the patients is characterized by a concentration level of: a) 1-MNA, Cystathionine, Cystine, Homocysteine, Nicotinamide, SAH, SAM, Spermidine and/or Spermine higher than a pre-established threshold level; or b) 5-MTA, CSSG, Cysteine, GluCys, GSH and/or Methionine lower than a pre-established threshold level. In a preferred embodiment, the NNMT inhibitor is selected from the list comprising: SAH, Sinefungin, 1-MNA, 1-MQ, 5-amino-1-MQ, 8-Methyl-1-MQ, JBSNF-000088, JBSNF- 000265, AK-2, AK-4, MvH45, MS2756, MS2734, GYZ-78, NS1, LL320, AK-12, RS004, HS58A-C2, HS312, 4-Chloropyridine analogs, Yuanhuadine, EL-1, Cyclic peptide, 4‑hydroxy‑3‑methoxybenzaldehyde, other direct or indirect inhibitors and its derivatives. On the other hand, the present invention also refers to an in vitro method for predicting whether a patient is resistant to a treatment with tyrosine kinase inhibitors which comprises determining the level of any of the above metabolites or combinations thereof. Alternatively, the present invention also refers to a method for detecting the above cited metabolites or combinations thereof in a test sample from a human subject with the objective of detecting NNMT activity. In a preferred embodiment, the present invention is a computer-implemented invention, wherein a processing unit (hardware) and a software are configured to: ^ Receive the concentration level values of any of the above cited metabolites, ^ Process the concentration level values received for finding substantial variations or deviations, and ^ Provide an output through a terminal display of the variation or deviation of the concentration level. For the purpose of the present invention the following terms are defined: ^ The term "comprising" means including, but it is not limited to, whatever follows the word "comprising". Thus, use of the term "comprising" indicates that the listed elements are required or mandatory, but that other elements are optional and may or may not be present. ^ By "consisting of” means including, and it is limited to, whatever follows the phrase “consisting of”. Thus, the phrase "consisting of” indicates that the listed elements are required or mandatory, and that no other elements may be present. ^ By “therapeutically effective dose or amount” is intended an amount that, when administered as described herein, brings about a positive therapeutic response in a patient. The exact amount required will vary from subject to subject, depending on the age, and general condition of the subject, the severity of the condition being treated, mode of administration, and the like. An appropriate “effective” amount in any individual case may be determined by one of ordinary skill in the art using routine experimentation, based upon the information provided herein. ^ A reference control level is herein mentioned when referring to a “pre-established threshold level”. A reference value can be a “threshold value” or a “cut-off” value. Typically, a "threshold value" or "cut-off value" can be determined experimentally, empirically, or theoretically. According to the present invention, the “threshold” value refers the level of metabolite determined as mean, median, or the optimal cut-off value obtained using receiver operator characteristic analysis such as Youden’s index, Euclidean index or other recommended indexes. Description of the figures Figure 1. NNMT is correlated with a mesenchymal phenotype. (A) Correlation between NNMT expression and genes unique to epithelial or mesenchymal phenotypes in LAD patients (Battle Study). Pearson correlation coefficient and p-value were FDR adjusted. **** P<0.0001. (B) NNMT expression level positively correlates with gene expressions unique to a mesenchymal phenotype in lung adenoma carcinoma cell lines from CCLE dataset (Broad Institute). Figure 2. Increased NNMT expression correlates with poor prognosis in mutant EGFR NSCLC patients. NNMT expression analyzed by RT-PCR significantly correlates (p = 0.0454) with poor progression free survival in a cohort of patients with early stage surgically resected lung adenocarcinoma. Figure 3. Increased NNMT expression correlates with tissue fibrosis in Idiopathic pulmonary fibrosis (IPF) patients. Immunohistochemical analysis of Hematoxylin/eosin (HE), Masson's Trichrome (TRI) and α-SMA and NNMT stains on lung sections from IPF and control patients (C). Figure 4. NNMT is overexpressed in EGFR mutant cells resistant to TKI with a mesenchymal phenotype. (A) Transcriptomic and proteomic data integration. Protein and mRNA from 1484 single protein are represented as the median ratio between EGFR TKI resistant/TKI sensitive (log2 scale) from 3 different cell line models (HCC4006/Ge-R. HCC827/ERC4. H1975/CLR). X axis shows mRNA levels, Y axis, protein levels. Vertical and horizontal lines represent the threshold used (log2ratio >1 or. -1; >2 or <0.5 in a natural scale. (B) 72-hours cell viability assay with EGFR inhibitors. Data points are average of duplicate wells from three independent assays. Ge-R: gefitinib resistant, O-R: Osimertinib resistant, TGFB1: cells exposed to TGF-b1 to acquire a mesenchymal phenotype, ^NT: stably express shRNA against non-target (NT), and ^CDH1: stably express shRNA against E-cadherin (CDH1). Error bars, SD. (C) Increased expression of vimentin (VIM) and decreased expression of E-cadherin (CDH1) in EGFR mutant NSCLC cells grown resistant to EGFR TKIs correlates with NNMT expression. A representative immunoblots from 3 independent experiments. CDH1: E-cadherin, CDH2: N-cadherin, VIM: vimentin, NNMT, GAPDH. (D) NNMT expression starts to increase 85 days after exposure to the increasing concentrations of gefitinib during the generation of HCC4006 cells with acquired resistance to gefitinib. A representative immunoblot from 3 independent experiments (E) Induction of a mesenchymal phenotype by TGFß-1 exposure (left) or depletion of E-cadherin ( ^CDH1) (center) or by acquired osimertinib resistance (right) promote NNMT expression with increased expression of mesenchymal markers. A representative immunoblots from 3 independent experiments. CDH1: E-cadherin. VIM: vimentin. ACTB: ^-actin. Figure 5. NNMT is increased in EGFR mutant tumors resistant to TKI in NSCLC patients. (A) Clinicopathologic features of the 24 EGFRmut NSCLC patients analyzed by NNMT IHC. (B) Representative immunohistochemical analysis of tumors from NSCLC patients harboring EGFR secondary mutation (T790M+) or no T790M (T790M-) using anti- NNMT antibody before EGFR TKI therapies (PRE-TKI) and at the time of clinical progression on EGFR TKI therapies (POST-TKI). Figure 6. NNMT is necessary but not sufficient for the maintenance of a mesenchymal phenotype and an EGFR TKI resistance. (A) Immunofluorescence using antibodies against NNMT (green). E-cadherin (green, CDH1) and vimentin (blue, VIM) in EGFR TKI-resistant EGFR mutant cells stably expressing shRNA targeting non-target (shNT) or shRNA against NNMT (shNNMT) (B) Immunofluorescence using antibodies against NNMT (red). E-cadherin (green. CDH1). vimentin (blue. VIM). and nucleus (gray) in EGFR TKI-sensitive EGFR mutant cells stably expressing V5-tagged control (Flag) or NNMT cDNA. (C) Immunoblots demonstrating that NNMT depletion in EGFR TKI resistant cells with a mesenchymal phenotype (PC9O-R. HCC4006Ge-R. H1975O-R) promotes the expression of an epithelial marker E-cadherin (CDH1) while downregulates the expression of mesenchymal markers N- cadherin (CDH2), and vimentin (VIM). Representative immunoblots from 3 independent experiments are shown. Vinculin (VCL) as a loading control. (D) Immunoblots showing the status of canonical epithelial marker (CDH1: E-cadherin) and mesenchymal markers (CDH2: N-cadherin and VIM: vimentin) in EGFR TKI-sensitive EGFR mutant HCC4006 or H1975 cells stably expressing V5-tagged control (Flag) or NNMT cDNA and PC9 stably expressing mutant (MUT) or wild-type (WT) NNMT cDNA. A representative immunoblots from 3 independent experiments. (E) HCC4006Ge-R and PC9O-R cells stably expressing shRNA targeting NT or NNMT cells were treated with 1 µM of gefitinib and osimertinib. respectively for 72 hours and annexin V assay was performed. Percentages of annexin V positive cells are shown. Average of three independent experiments. Error bars: S.D; One-way ANOVA. N.S. Not significant. ** P<0.01, *** P<0.001, N.S. Not significant. (F) HCC4006 cells stably expressing V5-tagged control (Flag) or NNMT cDNA and PC9 stably expressing mutant (MUT) or wild-type (WT) NNMT cDNA were treated with 1 µM of TKI for 72 hours and annexin V assay was performed. Percentage of annexin V positive cells is represented. Average of three independent experiments. Error bars: S.D; One-way ANOVA. N.S. Not significant. * P<0.05, ** P<0.01, *** P<0.001. Figure 7. NNMT expression decreases the inhibition of cell viabilities by TKI treatment. (A) 72-hours cell viability assay with EGFR inhibitor in EGFR TKI-R HCC4006Ge-R and PC9O-R cells stably expressing non-target (shNT) or shRNA against NNMT (shNNMT). Data points are average of duplicate wells from three independent assays. Error bars. SD. (B) HCC4006Ge-R and PC9O-R cells stably expressing targeting non-target (shNT) or shRNA against NNMT (shNNMT)were treated with the indicated concentration of gefitinib and osimertinib respectively for 10 days and surviving cells were stained with crystal violet. Image is representative of three independent assays. (C) HCC4006Ge-R. H1975O-R and PC9O-R stably expressing targeting non-target (shNT) or shRNA against NNMT (shNNMT) shRNA targeting NT or NNMT cells were treated with DMSO or 500nM of the indicated TKI for 1 hour. Lysates were subjected to immunoblots with antibodies indicated. Images are representative of three independent experiments. (D) HCC4006 stably expressing control (Flag-V5) or NNMT-V5 expression vector and PC9 cells stably expressing WT or MUT NNMT cDNA were treated with DMSO or 500 nM of the indicated TKI for 1 hour. Lysates were subjected to immunoblot with antibodies indicated. Image is representative of three independent experiments. (E) The indicated cell lines overexpressing NNMT were treated with 5 or 15 nM of TKI for 10 days and surviving cells were stained with crystal violet. Image is representative of three independent assays. (F) 72-hours cell viability assay with osimertinib in HCC4006 stably expressing V5-tagged control (Flag) or NNMT cDNA treated with vehicle or 100 nM osimertinib for 30 days. Data points are average of duplicate wells from three independent assays. Error bars. SD. Figure 8. NAD+ metabolism is altered in EGFR mutant cells with a mesenchymal phenotype. (A) Graphic representation depicting the critical roles of NNMT in the methionine cycle and associated metabolic pathways. UPLC-MS/MS quantification of (B) 1-MNA. (C) NAD+. (D) nicotinamide. and (E) SAM/SAH ratio in the supernatants from EGFR mutant NSCLC cells (HCC4006 and H1975) and cells grown resistant to gefitinib (HCC4006Ge-R) or osimertinib (H1975O-R). Average of three independent experiments. Error bars: Min to max; T-test. * P<0.05, ** P<0.01, *** P<0.001, **** P<0.0001. Figure 9. Methionine. polyamine. trans-sulphuration metabolism is altered in EGFR mutant cells with a mesenchymal phenotype. (A) Graphic representation depicting the critical roles of NNMT in the methionine cycle and associated metabolic pathways. UPLC- MS/MS quantification of (A) methionine, (B) homocysteine, (C) 5-MTA, (D) cystathionine, (E) spermidine, (F) cysteine, (G) spermine and (H) GSH in in the supernatants from EGFR mutant NSCLC cells (HCC4006 and H1975) and cells grown resistant to gefitinib (HCC4006Ge-R) or osimertinib (H1975O-R). Average of three independent experiments. Error bars: Min to max; T-test. N.S. Not significant. * P<0.05, ** P<0.01, *** P<0.001. Figure 10. Metabolite biomarkers of NNMT activity. UPLC-MS/MS quantification of 1- MNA. NAD+. nicotinamide and SAM/SAH ratio in supernatants of (A) HCC4006Ge-R cells stably expressing non-target (shNT) or shRNA against NNMT (shNNMT), and (B) HCC4006 stably expressing V5-tagged control (Flag) or NNMT cDNA. Average of three independent experiments. Error bars: Min to max; T-test. N.S. Not significant, * P<0.05, ** P<0.01. Figure 11. EGFR TKI resistance by NNMT overexpression can be reverted by Daporinad via NAD+ depletion. (A) Graphical representation of NAMPT-mediated NAD+ synthesis pathway and inhibition by daporinad (DAP). (B) HCC4006/HCC4006Ge-R. H1975/H1975O- R and PC9/PC9O-R cells were treated with the indicated concentration of DAP for 10 days and surviving cells were stained with crystal violet. The displayed image is representative of three independent assays. (C) UPLC-MS/MS quantification of NAD+ in HCC4006O-R cells after 24-hours DAP (1 µM) and DAP with nicotinamide mononucleotide (NMN.100 µM) treatment combination. Error bars: Min to max; One-way ANOVA. N.S. Not significant. ** P<0.01. *** P<0.001. (D) HCC4006Ge-R. and PC9O-R cells were treated with the indicated concentration of DAP and DAP with NMN (100 µM) for 10 days and surviving cells were stained with crystal violet. Image is representative of three independent assays. (E) NNMT depletion diminishes DAP sensitivity in HCC4006Ge-R and PC9O-R cells. Following the 10-day DAP treatment. surviving cells were stained with crystal violet. Image is representative of three independent assays. (F) Percentages of annexin V positive cells after 72-hour DAP treatment in TKI- resistant HCC4006Ge-R cells stably expressing shRNA targeting non-target (shNT) or shRNA against NNMT (shNNMT). Average of three independent experiments. Error bars: S.D; One- way ANOVA. ** P<0.01. *** P<0.001. **** P<0.0001. (G) PC9O-R shNT cells were subject to colorimetric in vitro proliferation assay for 72 hours with DAP, olaparib, selisistat or its combination and compared to PC9O-R shNNMT cells treated with DAP. Data points are average of duplicate wells from three independent assays. Error bars, SD. Figure 12. Combination of NNMT and NAMPT inhibition improve EGFR TKI sensitivity. (A) HCC4006 stably expressing V5-tagged control (Flag) or NNMT cDNA and HCC4006Ge-R cells stably expressing shRNA targeting NT or NNMT were transfected with siRNA targeting NAMPT or Control. NAMPT. VCL. A representative immunoblots from three independent experiments. (B) 72 hours post-transfection with siRNA targeting NAMPT or Control. % Surviving cells were assessed using HCC4006Ge-R cells stably expressing shRNA targeting NT or NNMT (left) and HCC4006 cells stably expressing V5-tagged control (Flag) or NNMT cDNA (right). Error bars: S.D; One-way ANOVA. N.S. Not significant. ** P<0.01. (C) The quantification of NAMPT mRNA levels in HCC4006 and HCC4006Ge-R cells using Taqman RT-PCR. Error bars: S.D; T-test. ** P<0.01. (D) HCC4006Ge-R cells stably expressing NAMPT maintain NNMT protein levels. NNMT. NAMPT. ACTB: ^-actin. A representative immunoblots from three independent experiments. (E) HCC4006Ge-R cells ectopically overexpressing NAMPT cDNA were treated with the indicated concentration of DAP for 10 days and surviving cells were stained with crystal violet. Image is representative of three independent assays. (F) Annexin V assay was performed after 72-hour DAP treatment in HCC4006 cells stably expressing control (Flag-V5) or NNMT-V5 expression vector. Percentage of annexin V positive cells is represented. Average of three independent experiments. Error bars: S.D; One-way ANOVA. * P<0.05. ** P<0.01. *** P<0.001. (G) The indicated cell lines were treated with DAP for 10 days. Surviving cells were stained with crystal violet. Image is representative of three independent assays. (H) HCC4006 and HCC4006Ge-R cells were treated with DAP 1nM for 72h without changes in their epithelial or mesenchymal phenotype. E-cadherin (CDH1). N-cadherin (CDH2). CD44. vimentin (VIM). NNMT. NAMPT. GADPH. A representative immunoblots from 3 independent experiments. Figure 13. The inhibition of NNMT activity promotes mesenchymal to epithelial transition to ameliorate EGFR TKI sensitivity. (A) UPLC-MS/MS quantification of 1-MNA in H1975O-R and PC9O-R cells treated with 1-MQ (100 µM) for 30 days. Error bars: Min to max; T-test. **** P<0.0001. (B) Annexin V assay was performed in HCC4006Ge-R shNT cells after 72-hour treatment with DAP (5 nM) and DAP and 1-MQ (100 µM) combination. Percentage of annexin V positive cells is represented. Average of three independent experiments. Error bars: S.D; One-way ANOVA. N.S. Not significant. **** P<0.0001. (C) % NAD+ levels in HCC4006Ge-R cells treated for 72-hours with increasing concentration of DAP with and without 1-MQ (100 µM) compared to vehicle control. Average of three independent experiments. Error bars: S.D; One-way ANOVA. N.S. Not significant. * P<0.05. ** P<0.01. *** P<0.001. **** P<0.0001. (D) Immunofluorescence assays of NNMT. E-cadherin and vimentin in EGFR TKI-resistant EGFR mutant cells treated with 1-MQ (100 µM) for 30 days. A representative image from three independent experiments. Red, NNMT; green, E-cadherin; blue, vimentin; grey, nucleus. (E) The indicated cell lines were treated with indicated concentrations of osimertinib for 10 days. Surviving cells were stained with crystal violet. Image is representative of three independent assays. (F) H1975O-R and PC9O-R control and 1-month 1-MQ (100 µM) treated cells were subject to colorimetric in vitro proliferation assay for 72 hours with osimertinib. Data points are average of duplicate wells from three independent assays. Error bars. SD. (G) PC9 stably expressing WT or MUT NNMT protein were seeded at low density. treated with vehicle. osimertinib (100 nM) or osimertinib with 1- MQ (100 µM) and grown until confluency. Percent of confluent wells was assessed every week. Data points are average of three independent assays. Figure 14. Serum 1-MNA association with prognostic. Correlation between serum level of NNMT metabolite.1-MNA. and progression free survival (PFS) or Overall survival (OS) in a cohort of 69 EGFR mutant NSCLC patients. (A) PFS from all the patients (left), from patients with detectable T790M (middle), and from the patient without detectable T790M mutation upon EGFR TKI treatment (right). (B) OS from all the patients (left), from patients with detectable T790M (middle), and from the patient without detectable T790M mutation upon EGFR TKI treatment (right).1-MNAhigh/low stratified according above or below the median value respectively. Censored data are indicated by tick marks. Data from patients who had not died at the time of the analysis were censored on the basis of the last recorded date on which the patient was known to be alive. (C) Cox regression analysis of PFS. (D) Cox regression analysis of OS. Figure 15. Clinicopathologic features of the 69 EGFR mutant NSCLC patients in the study. A) B) C) Clinicopathologic features of the 69 EGFR mutant NSCLC patients in the study. Figure 16. Serum metabolites levels associated with overall survival. (A) UPLC-MS/MS quantification of NNMT related metabolites in blood of 69 EGFR mutant NSCLC patients. (B) Mean of each subgroup and p-value for Log-Rank test for correlation between serum level of NNMT metabolites and overall survival (OS) in a cohort of 69 EGFR mutant NSCLC patients from all the patients (left). from patients with detectable T790M (middle), and from the patient without detectable T790M mutation upon EGFR TKI treatment (right). High and low groups were stratified according above or below the mean value respectively for each metabolite. Censored data are indicated by tick marks. Data from patients who had not died at the time of the analysis were censored on the basis of the last recorded date on which the patient was known to be alive. Figure 17. Serum metabolites associated with prognostic. Correlation between serum level of (A) 1-MNA, (B) SAH, (C) SAM, (D) Spermine and (E) Spermidine metabolite and overall survival (OS) in a cohort of 69 EGFR mutant NSCLC patients from all the patients (left), from patients with detectable T790M (middle) and from the patient without detectable T790M mutation upon EGFR TKI treatment (right). High and low groups were stratified according above or below the mean value respectively for each metabolite. Censored data are indicated by tick marks. Data from patients who had not died at the time of the analysis were censored on the basis of the last recorded date on which the patient was known to be alive. Figure 18. Interrelation between metabolites related to NNMT activity. (A) Spearman correlation to check pairs correlation between metabolites. (B) EGFR status detailed. (B) Receiver-operating characteristics (ROC) curve and area under the ROC curve (AUC) for 1- MNAHigh/Low classification using UPLC-MS/MS quantification of Nicotinamide, Methionine, Cystathionine, Cystine, Cysteine, Homocysteine, 5-MTA, GluCys, SAH, SAM, CSSG, GSH, Spermine and Spermidine in blood. Figure 19. Performance of GBM prediction models by NNMT-related metabolites and 24-months overall survival. (A) Receiver-operating characteristics (ROC) curve and area under the ROC curve (AUC) for GBM model using UPLC-MS/MS quantification of Nicotinamide, Methionine, Cystathionine, Cystine, Cysteine, Homocysteine, 5-MTA, GluCys, SAH, SAM, CSSG, GSH, Spermine and Spermidine in blood and OS at 24 months. (B) Kaplan-Meier survival curves based for GBM model prediction. Good and Poor responders were stratified according GBM model for OS at 24 months. (C) Importance score of variables for the model. Figure 20. Performance of GBM prediction models by NNMT-related metabolites and 24-months overall survival. A) B) C) Receiver-operating characteristics (ROC) curve and area under the ROC curve (AUC) for GBM model using combinations from 2 to 15 UPLC- MS/MS quantification for Nicotinamide, Methionine, Cystathionine, Cystine, Cysteine, Homocysteine, 5-MTA, GluCys, SAH, SAM, CSSG, GSH, Spermine and Spermidine in blood and OS at 24 months. Figure 21. Performance of GBM prediction models by NNMT-related metabolites, age (years), sex, stage and smoking status for 24-months overall survival. (A) Receiver- operating characteristics (ROC) curve and area under the ROC curve (AUC) for GBM model using UPLC-MS/MS quantification of Nicotinamide, Methionine, Cystathionine, Cystine, Cysteine, Homocysteine, 5-MTA, GluCys, SAH, SAM, CSSG, GSH, Spermine and Spermidine in blood, clinical features and OS at 24 months. (B) Kaplan-Meier survival curves based for GBM model prediction. Good and Poor responders were stratified according GBM model for OS at 24 months. Detailed description of the invention The present invention is illustrated by means of the examples below without the intention of limiting its scope of protection. Example 1. Material and methods Example 1.1. NSCLC resistant cell lines HCC4006, HCC827, PC9 AND H1975 EGFR mutant cell lines were exposed to increasing concentrations of EGFR TKIs over 6 months in a manner similar to previously reported [Lynch, T.J., et al., Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib. N Engl J Med, 2004. 350(21): p. 2129-39]. For EGFR TKI-resistant HCC827 cells, clones were used as described previously [Lynch, T.J., et al., Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib. N Engl J Med, 2004. 350(21): p. 2129-39]. All resistant cells are able to proliferate normally in the presence of 10 μmol/L EGFR TKIs. Upon confirming resistance, cells were cultured without drugs and their resistance to TKI was examined periodically. Example 1.2. Western blot analysis NSCLC cell lines were lysed for 1 hour in Cell Extraction Buffer (Biosource, Camarillo, CA) supplemented with protease and phosphatase inhibitors and precleared by centrifugation before protein quantification using the BCA Protein Assay kit (Pierce, Rockford, IL). Proteins (20 μg) resolved by SDS-PAGE were electrophoretically transferred to polyvinylidene difluoride membrane and Western blots were carried out using standard techniques. Cells were rinsed twice with cold PBS and proteins were extracted in ice-cold lysis buffer (Cell Signaling, Danvers, MA, USA) containing a protease and phosphatase inhibitors cocktail (Sigma-Aldrich, St. Louis, MO, USA), precleared by centrifugation before protein quantification using the BCA Protein Assay kit (Pierce, Rockford, IL). Proteins (20 μg) resolved by SDS-PAGE were electrophoretically transferred to polyvinylidene difluoride membrane using Bio-rad systems (Bio-rad, Richmond, CA, USA). Blots were incubated overnight at 4 °C with primary antibodies. Blots were incubated for 1 h at room temperature with the appropriate peroxidase-conjugated secondary antibody HRP-linked antibody (Cell Signaling, Danvers, MA, USA). Blots were visualized using a chemiluminescence detection kit ECL Western blotting substrate (Fisher Scientific, Madrid, Spain). Signals were captured by the ChemiDoc™ XRS + Imaging System (Bio-rad, Richmond, CA, USA). Example 1.3. Cellular viability assays Live cells were counted using Countess II (Thermo Fisher Scientific) and 5x103 cells per well were seeded in 96-well plates to compare cell growth kinetics using Sulforhodamine B assay as described by the NCI. The results were analyzed and graphed using Prism 8 (GraphPad Software). For the crystal violet cell viability assay, cells were seeded at low density (5x103 cells per well) in 6-well plates and left to adhere for 24 hours. Then media was replaced with vehicle, or the appropriate drug (DAP or TKI) and it was renewed every 3 days until approximately 70% confluency was reached. Supernatant was replaced with fixing solution and cells were stained with crystal violet 0.4% (v/v). Plates were then washed in running water, allowed to dry and scanned for imaging. Example 1.4. Cell death studies with annexin-V staining Adherent cells were trypsinized and pooled with the supernatant. Apoptosis was assessed using an Annexin V-FITC conjugate (ThermoFisher Scientific) and Propidium Iodide (ThermoFisher Scientific) staining according to the manufacturer's instructions. After incubation, samples were placed on ice and run using a BD LSRFortessa and data was analyzed with FlowJo software. Appropriate unstained and single-stained samples were used for controls and compensation. Example 1.5. Cell death analysis by fluorescence automatized microscopy INCell Analyzer 2000 (Ge Healthcare Life Sciences) was used to quantify cell death systematically. Briefly, 5x103 cells were seeded in black-wall 96-well plates and left 24 hours to attach. The appropriate siRNA was added to the wells and 72 hours after treatment, cell death was evaluated in situ with annexin V, propidium iodide and Hoechst 33342 staining. A 20x/0.45, Plan Fluor ELWD objective was used and INCell Investigator Image Analysis 1000 Workstation Software was used to analyze the acquired images. Example 1.6. siRNA transfection siRNA (pooled oligos) was transfected into the target cells using TransIT-siQUEST reagents (Mirus) according to manufacturer’s instructions. The catalog numbers for the siRNAs used are as follows: Example 1.7. cDNA/shRNA constructs and lentiviral infection pLX304 was a gift from David Root (Addgene plasmid # 25890; http://n2t.net/addgene:25890; RRID: Addgene_25890) (1). cDNA for NNMT (HsCD00442343) and NAMPT (HsCD00442891) were cloned to pLX304 vector and obtained from DNASU, Arizona State University Biodesign Institute. cDNA for WT and catalytically inactive NNMT (Y20A) mutant were cloned in pLVX-IRES-ZsGreen or pLVX-IRES-ZsTomato respectivelly. shRNA lentiviral vector construct designed by the RNAi Consortium was used as described previously [Lynch, T.J., et al., Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib. N Engl J Med, 2004. 350(21): p. 2129-39]. shRNA knockdown and viral transduction were performed as reported previously [Lynch, T.J., et al., Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib. N Engl J Med, 2004. 350(21): p.2129-39]. Using TransIT-LT1 transfection reagent (Mirus), the 293LTV cell line (Cell Biolabs) was transfected with pLKO.1 constructs and packaging plasmids, pCMV- ΔR8.2dvpr and pCMV-VSV-G. Lentiviral supernatants were collected, and quality of viral supernatant was routinely tested with Lenti-X Go Stix (Clontech). The supernatants were applied to target cells. Following puromycin selection, cell viability assays were performed 7 days post-infection. Example 1.8. RNA/DNA extraction RNA from cell lines and fresh frozen tumor or adjacent normal lung tissues was extracted and purified using trizol (TRI Reagent ^, Sigma) per manufacturer instructions. DNA was extracted with DNeasy blood & tissue kit (Qiagen). Samples were quantified using NanoDrop ND-200 Spectrophotometer. Example 1.9. Analysis of Microarray Gene Expression Data Gene expression profiles of indicated experimental conditions were generated using the Human Genome U133A 2.0 microarray (Affymetrix) (Genomics Core Facility, University of Valencia). Affymetrix GeneChipScanner was used for the export of summarized probe intensities. Next, the expression data was annotated and collapsed into Genepattern. Differential expression analysis using the t-test statistic and the false discovery rate (FDR) adjustment was carried out to find differentially expressed transcripts (http://www.broadinstitute.org/cancer/software/genepattern). GSEA (http://www.broadinstitute.org/gsea) was used to determine the gene set enrichment of experimental signatures related with EMT and E-cadherin knockdown obtained from the MsigDB (http://www.broadinstitute.org/gsea/msigdb). GSEA estimates whether the members of a given gene signature are found at the top or bottom of a list of genes ranked by signal-to- noise ratio, indicating they are associated with a specific phenotype, rather than being distributed uniformly or randomly across the gene list. An enrichment score (ES) is calculated to quantify the degree to which a gene signature is overrepresented at the top or bottom of the entire ranked list. GSEA normalizes the ES for each gene set to account for the variation in set sizes, yielding a normalized enrichment score (NES) and a false discovery rate (FDR). The FDR gives an estimate of the probability that a set with a given NES represents a false positive finding; it is computed by comparing the tails of the observed and empirical phenotype-based permutated null distributions for the NES. Example 1.10. Proteomic analysis SWATH LC-MS/MS analysis was used to study the proteomic content of individual samples. Briefly, samples were run in acrylamide gels and gel lanes slices and digested with trypsin as described previously [Shevchenko A, Jensen ON, Podtelejnikov AV, Sagliocco F, Wilm M, Vorm O, Mortensen P, Boucherie H, Mann M. Linking genome and proteome by mass spectrometry: large‐scale identification of yeast proteins from two dimensional gels. Proceedings of the National Academy of Sciences of the United States of America 1996; 93:14440‐14445].5 μl of each sample were loaded onto a trap column (NanoLC Column, 3μ C18‐CL, 75umx15cm; Eksigen) and desalted with 0.1% TFA at 3μl/min during 5 min. The peptides were loaded onto an analytical column (LC Column, 3 μ C18‐CL, 75µmx12cm, Nikkyo) equilibrated in 5% acetonitrile 0.1% FA. Peptide elution was carried out with a linear gradient of 5 to 35% B in 90 min (A: 0.1% FA; B: ACN, 0.1% FA) for at a flow rate of 300nl/min. Peptides were analyzed in a mass spectrometer nanoESI qQTOF (5600 TripleTOF, ABSCIEX). The tripleTOF was operated in information‐dependent acquisition mode, in which a 250‐ms TOF MS scan from 350–1250 m/z, was performed, followed by 150‐ms product ion scans from 350–1500 m/z on the 25 most intense 2‐5 charged ions. The rolling collision energies equations were set for all ions as for 2+ ions. ProteinPilot default parameters were used to generate peak list. The Paragon algorithm of ProteinPilot was used to search SwissProt protein database. Example 1.11. NAD+/NADH levels quantification NAD+/NADH-Glo Assay kit (Promega) was used per manufacturer advice as a faster method to quantify NAD+/NADH than UPLC-MS/MS. Briefly, 3.5x103 cells were seeded in 96-well plates and left 24 hours to adhere before DAP addition. After 24-hour of drug incubation NAD+/NADH detection reagent was added. One-hour post incubation supernatant was transferred to a white-well plate to evaluate luminescence in an Infinite M200pro (Tecan). Example 1.12. UPLC-MS targeted analysis UPLC separation was performed in an Acquity UPLC system (Waters, UK) equipped with an Acquity UPLC BEH C18 (1.7 μm, 2.1 × 100 mm; Waters) column. The temperatures of the column and the autosampler were set at 40°C and 4°C, respectively. The sample injection volume was 4 μl. Eluents consisted in 0.5% formic acid and 0.3% heptafluorobutyric acid in water (eluent A) and acetonitrile (eluent B). The flow rate was set at 0.3 ml/min. A 9-min elution gradient was performed as follows: initial eluent composition was set at 97% A and 3% B, which was linearly changed to 90% A and 10% B in 2 min; then the proportion of B was increased to 30% in the next 3 min, followed by a further increase to 100% B reached at min 6 and kept for 1.5 min. Finally, the initial conditions were recovered and maintained for 2 min for column conditioning. The MS analysis was performed using a Waters Xevo TQ-S mass spectrometer (Waters) equipped with an ESI source working in positive-ion mode multiple reaction monitoring (MRM) mode. A capillary voltage of 2 kV, a source temperature of 120°C and a desolvation temperature of 380°C were used. Desolvation and the cone gas flow were set as 800 liters/h and 150 liters/h, respectively, and the collision gas was 0.25 ml/min. Transitions, cone voltages, and collision energies were automatically tuned for each metabolite using the Quanoptimizer software (Waters). The data station operating software used was MassLynx 4.1 (Waters). MS conditions (including quantification and qualification transitions, cone voltage, collision energy, retention time and internal standard used for normalization purposes) for each metabolite and internal standard are detailed in the following table: Calibration curves were generated by plotting the peak area ratio of the respective compound to the corresponding internal standard versus the nominal concentration. The line of best fit was determined by linear-weighted (1/×) least-squares regression. The linearity acceptance criterion for the correlation coefficient was 0.99 or better. Each back calculated standard concentration should be within ±15% deviation from the nominal value, except for the LLOQ, for which the maximum acceptable deviation was ±20%. Standards calibration curves, with concentrations in the 0.62–20,000 nM range, were prepared by serial half dilutions. IS concentrations were kept constant at all the calibration points at 3000nM for 3Aminobenzamide and CystineD6, 300 nM for MetD3 and 5MTA, 200 nM for MNAD3 and 100nM for theanine and PheD5. Example 1.13. Sample preparation Serum samples. First, 100 μl of serum sample was allowed to thaw on ice and then 25 μl of 40 mM N-ethylmaleimide (NEM) were added to derivatize free thiols. Protein precipitation was performed by adding 3 volumes of cold methanol. After vortexing samples were incubated at -20 °C for 30 min. To remove protein precipitate, samples were centrifuged twice at 10,000 g for 10 min at 4°C, and supernatants were transferred to clean tubes. A 400 μl aliquot was evaporated to dryness and resuspended in 80 μl of eluent A containing internal standards. Another aliquot was diluted 1/20 into eluent A containing internal standards. Cultured cell samples. For 6-well plates. Media was removed and the cells washed twice with cold PBS. Cells were scraped using 1000 μL of 8mM NEM in water. Cells were lysed by using 3 freeze-thaw cycles. Then, a 250 μl was extracted using 750 μl of cold methanol. After vortexing samples were incubated at -20 °C for 30 min. To remove protein precipitate, samples were centrifuged twice at 10,000 g for 10 min at 4°C, and supernatants were transferred to clean tubes. An 800 μl aliquot was evaporated to dryness and resuspended in 80 μl of eluent A containing internal standards. Another aliquot was diluted 1/20 into eluent A containing internal standards. Cell culture spent media. First, 20 μl of spent media was allowed to thaw on ice and then 5 μl of 40 mM NEM were added. Protein precipitation was performed by adding 3 volumes of cold methanol. After vortexing samples were incubated at -20 °C for 30 min. To remove protein precipitate, samples were centrifuged twice at 10,000 g for 10 min at 4°C, and supernatants were transferred to clean tubes. An aliquot was diluted 1/20 into eluent A containing internal standards. Example 1.14. Histology and immunohistochemistry All immunohistochemistry procedures were performed in the Department of Pathology at Hospital Quirónsalud (Valencia, Spain). In murine studies, tumors were fixed in 4% formalin overnight at room temperature. The specimen was then embedded in paraffin, from which 5- μm sections were generated. Tumors were sectioned on slides with 5 μm thickness. Then paraffin sections were deparaffinized in xylene and rehydrated in a graded alcohol series and boiled with citrate buffer pH 6. Sections were incubated with primary antibodies specific for NNMT or ki67 All biopsy specimens from patients with NSCLC were stained with hematoxylin and eosin and NNMT and reviewed by the board-certified pathologists in the Department of Pathology at Hospital Quirónsalud (Valencia, Spain). For antibody detection, Envision detection system peroxidase/DAB (Dako) was used. The procedure was performed on a Dako Autostainer Link 48. Image acquisition and analyses were performed with a Leica DMD108 microscope. The NNMT staining score were developed by the board-certified pathologists and designated as follows: -, negative expression; +, weak expression; ++, moderate expression; +++, strong expression. Lung tissue samples from idiopathic fibrotic patients were fixed in 10% formalin neutral buffer solution for 24 h, and then embedded in paraffin before being cut into 4-μm-thick sections. For histological examination, sections were stained with Mayer’s hematoxylin and then with 1% eosin alcohol solution (H&E staining). For staining of collagen (Masson’s trichrome staining), sections were treated sequentially with solution A (5% w/v potassium dichromate and 5% w/v TCA). And for immunohistochemical analysis of α-SMA, samples were incubated 12 h with an antibody against α-SMA from Abcam. Example 1.15. Biopsies from patients This study was carried put in accordance with the Declaration of Helsinki including all amendments and participating hospitals required Institutional Research Ethics Board approval. All patients provided their written informed consent prior to being included in the study and before study procedures, sampling, and analyses. Following routine clinical practice, patients’ data were retrospectively obtained from patients’ medical records. All 24 EGFR-mutant NSCLC patients who progressed on EGFR TKI therapy underwent second biopsy of their tumor at Hospital Clinico Universitario and Hospital Universitario La Fe in Valencia, Spain and were included in the study cohort. Samples used for analyses had to have both pre- and posttreatment tumor specimens available. The medical records were reviewed retrospectively by board-certified physicians to obtain clinical information under an Institutional Review Board–approved protocol. All relevant ethical regulations were followed. The mutational status of EGFR (both primary and secondary mutations) was determined by the institutions as part of their diagnostic routine. Example 1.16. Murine drug treatment studies All animal treatment studies were reviewed and approved by the Institutional Animal Care and Use Committee at University of Valencia. Approximately 2×106 total cells were injected subcutaneously in the flanks of NOD scid gamma (NSG) mice. The size of the engrafted tumors was measured with digital calipers and grown until approximately 100 mm3. The tumors were then randomized, and mice were treated (intraperitoneally) daily for 20 days either with vehicle or daporinad (10 mg/kg). Immunohistological analysis of mouse xenografts was performed similarly to samples from EGFR-mutant NSCLC patients. Example 1.17. Immunofluorescence Briefly, 0,1x106 cells were plated in 12 well plates and left 24 hours to adhere. After 24 hours cells were fixed in 4% PFA. After permeabilization with 0.2% Triton X-100 cells were blocked. Primary antibodies were incubated overnight in a wet-chamber and the conjugated secondary was incubated at room temperature for one hour. Nuclei were labelled with Hoechst 33342 (Molecular Probes, Life Technologies). Image acquisition and analyses were performed with the Olympus FV1000 confocal microscope with a 40X objective. Example 1.18. Predictive Modeling A total of 69 patients were stratified into good-responders with overall survival (OS) longer than 24 months and poor-responders with OS shorter than 24 months. At baseline, the serum samples were collected from patients for targeted metabolites profiling analysis. Categorical variables associated for each patient (sex, smoking status or TNM staging) were recategorized into binary components. We developed machine learning-based risk prediction models applying the Gradient Boosting Model (GBM) algorithm with k-fold cross-validation (k = 5) performed on Python 3.7.3 using all analyzed metabolites. After, we developed new models including demographic, clinical, genetic characteristics and metabolites. Importance score of each feature was automatically calculated. A higher importance score indicated that the variable had a higher predictive value for the model. The classification quality was evaluated using a comparison of real and the predicted class for each patient. The metrics were ROC curves to obtain Area Under ROC Curve (AUC), sensitivity and specificity and Kaplan-Meier curves. Example 2. Results Example 2.1 Increased NNMT expression and activity induce mesenchymal phenotype and correlate with poor prognosis and resistance to cancer therapy The development of irreversible EGFR TKIs has provided treatment options to the NSCLC patients with acquired T790M mutations following reversible EGFR TKI treatment; however, NSCLC patients with acquired EGFR TKI resistance with an EMT phenotype do not benefit from targeted therapies. To find the Achilles’ heel that can be targeted to destroy the subset of EGFR TKI resistant NSCLC with an EMT phenotype, we interrogated public databases of lung tumour samples for targets that correlates with mesenchymal-specific genes and the EMT signatures we derived from the EGFR TKI resistant NSCLC cells. Among the target candidates, NNMT (Nicotinamide N-methyl transferase) positively correlated with mesenchymal genes and negatively correlated with epithelial genes (Figure 1B). Additionally, NNMT significantly associates with EMT signatures unique to EGFR TKI resistant cells. NNMT is an enzyme that transfers methyl groups from the universal donor S- adenosylmethionine (SAM) to nicotinamide (NAM) producing 1-MNA. This use of SAM reduces histone methylation potential, promoting gene expression changes in cancer cells. To investigate if the NNMT expression determines clinical outcome, we measured NNMT expression in tumors and in the adjacent normal tissue from NSCLC patients with various genotypes (Figure 1A & 5B). In the cohort, tumors with elevated NNMT expression levels compared to the normal adjacent tissues had statistically worse overall survival (p=0.0454) than patients with lower NNMT expression in their tumors (Figure 2). We then sought to measure serum 1-MNA levels as a surrogate marker of NNMT activity. Using UPLC-MS, we measured 1-MNA level in serum samples from EGFR mutant NSCLC patients undergoing EGFR TKI therapies (Figure 14). The patients with higher serum 1-MNA level had worse OS than the patients with lower serum 1-MNA level (Figure 14). Based on the strong correlation between the EMT and CDH1KD signatures derived from EGFR mutant NSCLC cells grown resistant to EGFR TKI (Figure 1B), we sought to investigate consecutive formalin-fixed paraffin-embedded (FFPE) tissue samples from advanced stage EGFR-mutant NSCLC patients who radiographically progressed on EGFR TKI treatment for the presence of a T790M secondary mutation and the overexpression of NNMT (Figure 5B). 13 of the 22 post-EGFR TKI specimens (59.1%) harboured a T790M mutation. Board-certified pathologist scored IHC samples for NNMT protein expression and 6 of the 22 (27.3%) patients showed increased NNMT expression at the time of disease progression; of these 6 patients, 5 patients (83.3%) did not have concurrent T790M secondary mutation and only 1 of 13 patients (7.7%) who harboured a T790M secondary mutation had increased NNMT level (Figure 5A & Figure 5B). To investigate whether NNMT expression correlates with the expression of mesenchymal and epithelial genes in NSCLC cell lines, we queried the gene expression data of the CCLE database. A significant positive correlation of NNMT expression with mesenchymal genes and a negative correlation with epithelial -specific genes was found (Figure 1B). The EMT and CDH1KD signatures derived from EGFR mutant NSCLC cells also strongly correlates with NNMT expression. To test if the induction of a mesenchymal phenotype in EGFR mutant NSCLC cells is sufficient to promote the overexpression of NNMT, we induced a mesenchymal phenotype in otherwise epithelial EGFR mutant NSCLC by chronically expose the cells to TKIs or TGF ^1 or depleting E-cadherin (1-3). As we previously have shown (3), the induction of a mesenchymal phenotype in EGFR mutant NSCLC promotes EGFR TKI resistance (Figure 4A & 4B). Considering that the induction of mesenchymal phenotype is epigenetically regulated, we compared global transcriptomic, proteomic and epigenetic changes in three parentals (i.e. HCC827, HCC4006, H1975) EGFR mutant NSCLC against their respective EGFR TKI resistant cell lines (i.e. ERC4, HCC4006Ge-R, H1975O-R) with a mesenchymal phenotype derived from the parental cell lines. To easy compare the data, all the changes are represented as median of log2ratio. We compared log2 rations of gene expression (X axis), protein expression (Y axis) and the promoter methylation score (color scale in the secondary Y axis). NNMT appears in the upper- right quadrant close to a set of genes with promoter hypomethylation resulting in the robust mRNA expression and protein expression (p<0,05, comparing all 3 pairs, Figure 4A). In our EGFR TKI resistant EGFR mutant NSCLC cells, NNMT is overexpressed in the cells presenting E-cadherin low and vimentin high expressions, a feature common among the mesenchymal cells (Figure 4C). In contrast, EGFR TKI resistance by MET amplification (ERC13, Figure 4B) or T790M in EGFR mutant NSCLC cells that maintain epithelial phenotype did not increase the NNMT expression. It is during the acquisition of EGFR TKI resistance in HCC4006 cells, that NNMT is upregulated concomitantly to the loss of E- cadherin expression and the increase of vimentin and N-cadherin expression (Figure 4D). Previously, we have shown that TGF ^1-induced EGFR TKI resistance and the mesenchymal phenotype are reversible in EGFR mutant NSCLC cells by culturing the cells in the absence of TGF ^1, and that the TGF ^1-induced NNMT expression was also reversible (Figure 4E). Furthermore, the induction of a mesenchymal phenotype by the depletion of CDH1 also resulted in an increase in NNMT expression (Figure 4E). In summary, we discovered that NNMT expression is a unique event in EGFR TKI resistant EGFR mutant NSCLC cell lines with a mesenchymal phenotype. Furthermore, this NNMT increasing has been also observed in paraffin-embedded human idiopathic pulmonary fibrosis patients, suggesting that NNMT is important in the pathogenesis of IPF, modulating EMT process (Figure 3). To investigate if NNMT is necessary to maintain the EMT-mediated EGFR TKI resistant phenotype, we repressed NNMT by shRNA in three EGFR TKI-resistant mesenchymal cell lines (HCC4006Ge-R, H1975O-R and PC9O-R) and their respective parental cells. NNMT repression in mesenchymal cell lines transformed the cells to a more epithelial-state as determined by CDH1 re-expression and down-regulation of other canonical mesenchymal markers (VIM, CDH2, ZEB1) (Figure 6A & 6B). To investigate if NNMT is enough for the induction of a mesenchymal phenotype, NNMT was ectopically overexpressed in EGFR TKI-sensitive cells (HCC4006, H1975 and PC9) and changes in EMT phenotype were investigated with confocal microscopy (Figure 6B). In NNMT overexpressing cells, vimentin intermediate filaments are detected in the cytoskeleton and CDH1 cell-cell junctions are reduced. Lysates from the NNMT overexpressing cells were resolved on western blot to show no apparent changes in canonical markers for EMT (Figure 6D). These results indicate that NNMT is necessary for the maintenance but not enough for the induction of the mesenchymal phenotype. Next, we assessed if the NNMT expression affects EGFR TKI sensitivity. Depleting NNMT in the mesenchymal EGFR TKI-resistant cells increases cell death significantly following EGFR TKI treatment (Figure 6E), decreases the IC50 for EGFR TKIs (Figure 7A) and reducing the number of surviving colonies after long-term treatment with EGFR TKI (Figure 7E). Although overexpressing NNMT did not affect phenotype, NNMT overexpressing cells increase EGFR TKI resistance, reducing significantly cell death after short term drug treatment (Figure 6F) and increasing the number of surviving colonies (Figure 7E). In HCC4006 cells, NNMT is necessary to promote the EGFR TKI resistant phenotype which is further enhanced with EGFR TKI treatment (100 nM, 30 days). (Figure 7F). This increased sensitivity to EGFR TKIs after NNMT repression could be explained to a regained EGFR signaling addiction that EGFR TKI treatment can effectively shut down by a decreased phosphorylation in EGFR/Akt/ERK/S6 signaling pathway (Figure 7C and 7D). In summary, we find that increased NNMT expression and activity is present in the induction of EMT process (in cancer and other pathologies) and significantly correlated with poor prognosis and therapy’s resistance. Example 2.2. Increased expression of NNMT alters NAD+ metabolism in EGFR mutant NSCLC cells, determining a new vulnerability based on NAD+ depletion NNMT transfers methyl groups from the universal donor SAM to NAM, producing 1-MNA regulating the availability of NAM for NAD+ synthesis (Figure 8A). Based on the discovery that NNMT expression is increased in EGFR TKI-resistant NSCLC cells, we hypothesized that NNMT metabolites from the methionine and NAD+ pathways could be altered in EGFR TKI resistant cells with a mesenchymal phenotype. The level of 1-MNA, which is a metabolite of NAM produced by NNMT, is significantly increased in EGFR mutant cells grown resistant to EGFR with a mesenchymal phenotype (Figure 8A). Similarly, the levels of NAM, which is a substrate of NNMT, are statistically lower in EGFR TKI-resistant cells (Figure 8B). The increased NNMT activity promoted an increased production of 1-MNA resulting in reduced NAD+ levels in EGFR TKI-resistant cells (Figure 8C). Finally, we measured the SAM/SAH ratio to evaluate SAM utilization, that are lower in cell lines with high NNMT than in cell lines with low NNMT (Figure 8D). The increased use of SAM in mesenchymal cells has to be replenished by methionine leads to reduced levels of this metabolite (Figure 9A). The polyamine synthesis pathway is essential for cell growth and proliferation and consumes SAM producing 5-MTA. We didn’t find significant changes in 5-MTA (Figure 9B) confirming that in our models, SAM/SAH lower ratio levels are not due to an increase in this pathway. Lastly, a direct consequence of lower SAM/SAH ratio in NNMThigh cells is a reduction in glutathione levels (Figure 9H). Since NAD+ metabolism is dysregulated in EGFR TKI-resistant cells with a mesenchymal phenotype, we hypothesized that the resistant cells adapt their metabolism depending on NNMT activity. Overexpressing NNMT in an EGFR TKI-sensitive cell line (HCC4006) or depleting NNMT in EGFR TKI resistant mesenchymal cells (HCC4006Ge-R) results in a deep metabolism rewiring. First, 1-MNA, NNMT sub-product, is detected 3,000 times more in NNMT overexpressing cells (HCC4006NNMT and HCC4006Ge-R shNT). Second, SAM/SAH ratio and NAM, surrogate markers of NNMT activity decreased when NNMT levels are higher in the cells. Finally, NAD+levels decreased in the cells with NNMT overexpression and depletion of NNMT restored NAD+ levels (Figure 10A& 10B). NAD+ is an essential co-factor in several metabolic processes therefore its levels are highly regulated. Among the three NAD+ salvage pathways, NAMPT catalyzes the main reaction using NAM as substrate to produce NAD+ (Figure 11A). FK866/APO866, daporinad (DAP) is a potent and specific NAMPT inhibitor, which is being used in different cancer clinical trials, that compromises the pool of cellular NAD+ (von Heideman, 2010; Goldinger 2016). EGFR TKI-resistant cells lead NAM to produce 1-MNA instead of NAD+ which prompted us to investigate a potential selective vulnerability of EGFR TKI-resistant cells to NAD+ depletion. Genetically repressing NAMPT reduces viability almost up to 50% in gefitinib resistant HCC4006Ge-R cells (Figure 12B). DAP treatment of NNMT overexpressing cells shows that mesenchymal cells are exceptionally sensitive to this inhibitor (Figure 11B). The anti-tumor effect seen in NNMThigh cells is, as expected due to a drop in NAD+ levels that can be rescued with nicotinamide mononuclueotide (NMN), NAMPT product (Figure 11C) reducing cell death induced by DAP (Figure 11D). To confirm that DAP targets selectively NNMThigh cells by an energetic collapse due to a drop in NAD+ levels, we measured NAMPT mRNA levels in HCC4006 and HCC4006Ge-R cells. Although NAMPT mRNA is lower in HCC4006Ge-R cells, overexpressing NAMPT (Figure 12C) does not affect NNMT expression (Figure 12D) or change DAP sensitivity (Figure 12E) indicating that EGFR TKI-resistant cells are specifically sensitive to DAP not due to NAD+ synthesis deficiencies but because the increased NNMT activity compromises NAD+ levels. We confirmed that NNMT overexpressing cells are remarkably sensitive to NAMPT inhibition and repressing NNMT in resistant HCC4006Ge-R cells reduces significantly DAP-induced cell death (Figure 11E & 11F). Moreover, overexpressing NNMT in epithelial HCC4006 cells is enough to sensitize cells to DAP treatment (Figure 12F). High sensitivity to DAP is also extrapolated to alternative forms of EMT activation with increased NNMT expression, like chronic TGFß1 exposure or stable CDH1 repression cell models (Figure 12G). Although DAP severely drops NAD+ levels but inhibiting the main NAD+consuming enzymes with olaparib (PARP) or selisistat (sirtuins) does not prevent DAP cytotoxic effect. Only pharmacological or genetic inhibition of NNMT can revert it (Figure 11G), suggesting that this cytotoxic effect of DAP on mesenchymal EGFR TKI-resistant cells is NNMT dependent. In summary, we find that NNMT activity correlates with metabolites from methionine cycle, trans-sulfuration pathway. cystine/cysteine cycle, polyamine pathways and NAD+ salvage cycle, specially, NNMT activity direct product 1-MNA. This phenomenon is also observed in TKI-resistance cells and NSCLC patients. NNMT contributes to NAD+ pathway homeostasis, so an NAD+ production inhibitor as NAMPT inhibitor, could generate a vulnerability specific to cells overexpressing NNMT, reducing EMT process and improving prognosis and reducing the acquisition of resistance against therapies. On the other hand, identification of patients with high levels of NNMT could serve to predict the response to NAMPT inhibitors. Example 2.3. Therapies directed against NAD + and methionine metabolism, such as NNMT or NAMPT inhibition, contribute to promote mesenchymal to epithelial transition and to ameliorate EGFR TKI sensitivity Since the accumulation of 1-MNA limits NNMT enzymatic activity, inhibitory activity of other small N-methylated molecules has been studied, including 1-methylquinolinium (1-MQ). Considering the increased NNMT activity in EGFR TKI resistant cells with a mesenchymal phenotype, we hypothesized that pharmacological inhibition of NNMT with 1-MQ could mimic the depletion of NNMT using shRNA. We sought that inhibition of NNMT activity with 1-MQ in mesenchymal resistant cells could prevent DAP cytotoxic effect. Treatment of mesenchymal EGFR TKI resistance cells with 1-MQ resulted in a significant decrease in 1- MNA level that indirectly suggests a decreased NNMT activity (Figure 13A).1-MQ treatment abolishes completely DAP cytotoxicity in mesenchymal cells (Figure 13B). This reduction in cell death can be explained because 1-MQ treatment prevented the abrupt drop of NAD+ levels (Figure 13C) following DAP treatment in EGFR TKI-resistant NNMThigh cells. Long-term treatment of EGFR TKI-resistant cells with 1-MQ promotes the mesenchymal-to-epithelial transition, increasing CDH1 cell-to-cell junctions and reducing vimentin levels (Figure 13D). The transition to the epithelial state by 1-MQ treatment associates with an increased EGFR TKI sensitivity (Figure 13E & 13F) similar to the depletion of NNMT by shRNA (Figure 7E & 7E). We have previously shown that NNMT expression is increased during EGFR TKI resistance (Figure 4D) and also that 1-MQ can inhibit NNMT favoring a more pronounced epithelial phenotype. Thus, we hypothesized that the small molecule 1-MQ could prevent or delay EGFR TKI-acquired resistance. As expected, stably expressing an active NNMT protein allows cells to acquire osimertinib resistance faster than those cells with mutated NNMT protein. In both cell lines, pharmacological inhibition of NNMT with 1-MQ prevents the appearance of drug resistant cells (Figure 13G). Motivated by the results that the EGFR TKI-resistant tumors with a mesenchymal phenotype is sensitive to single-agent DAP treatment in vitro (Figure 11B), we sought to test the impact of combination of NNMT and NAMPT. We observed that combination of both inhibitors reduced aggressiveness and drug resistance for TKI therapies (Figure 12B). In summary, we find that inhibition of NNMT activity combined with TKI presents improved results than the use of TKI alone. And given that NNMT increase correlates with NAD+ dependency, combination of NNMT inhibitors with inhibitors of the production of this metabolite, such as NAMPT inhibitors, considerably reduce the acquisition of resistance to TKIs. Example 2.4. Detection of NNMT related metabolites to predict survival of NSCLC’s patients treated with TKIs Since NNMT could be a useful therapeutic target against different diseases, we set out to identify biomarkers of its activity. We have observed that certain metabolites related to NNMT activity, such as S-Adenosyl methionine (SAM), 5′-methylthioadenosine (5-MTA), Methionine, Cystine, Nicotinamide, Spermidine, Spermine, Cystathionine, Cysteine, HomoCysteine, Glutamyl-L- cysteine (GluCys), S-adenosyl-L-homocysteine (SAH), Glutathione Oxidized (CSSG), Glutathione (GSH), Glutathione disulfide (GSSG), or NAD+ are altered in cell lines that have acquired resistance to therapy against TKI (Figure 8 & 9). We have also been able to observe that the levels of these metabolites are directly related to NNMT activity, performing cell models with down-expression of NNMT (Figure 10A) or with over-expression (Figure 10B). We have observed a good correlation between serum level of NNMT metabolite, 1-MNA, and progression free survival (PFS) or Overall survival (OS) in a cohort of 69 EGFR mutant NSCLC patients. Patients with high levels of 1MNA, representative of high NNMT activity, had worse survival than those patients with low levels of 1MNA (Figure 14A). 1- MNAhigh/low stratified according to above or below the median value respectively. Similar results were observed with other metabolites related to NNMT activity, such as SAH, Spermine or Spermidine, taking as cut-off point the mean instead of the median of each of these metabolites (Figure 17). In order to improve the predictive capacity for response to TKI therapy, we designed different machine learning models based on gradient boosting machine 24 months survival after TKI treatment in a cohort of 69 EGFR mutant NSCLC patients. We observed that with the combination of these 15 metabolites we improved the predictive capacity compared to the individual analysis of the metabolites (Figure 20). The best of the models obtained a predictive capacity of 0.83 for the AUC (Area Under the Curve) ROC (Receiver Operating Characteristics) curve with a 0.84 for Sensitivity and 0.8 for Specificity (Figure 19A). With this model, we classified patients as poor or good responders for TKI therapy and generated a Kaplan-Meier curve for OS, demonstrated that the model had high clinical utility. Our GBM model was able to stratify patients with a p-value <0.005 (Log- Rank test) (Figure 19B). But we were also able to verify that if we add demographic information such as age and gender to the information provided by these biomarkers, or if the patient is a smoker or not, we can further improve the ability to predict the response to TKI therapy. Our predictive model associated to 2-years overall survival obtained a p value <0.005 in Kaplan-Meier analysis (Figure 21B) with high accuracy (AUC: 0.94; Accuracy: 0.97; Sensitivity: 1.0; Specificity: 0.88) (Figure 21A). The results confirmed a high discriminatory power of blood NNMT related metabolites to distinguish between good and poor responders against TKI therapy in NSCLC. These data indicate that our method using mass spectrometry to measure blood NNMT related metabolites and machine learning models is a promising tool that could be used to support a better management of cancer’s patients. Despite the fact that these results have been obtained in a small cohort of patients, so it should be validated in a larger number of patients, it should be noted that the models obtained are based on machine learning strategies that allow the improvement of performance with adding new data. In summary, using mass-spectrometry quantification of NNMT related metabolites and machine learning models we can identify which patients present an increase in NNMT levels and therefore a weak response to TKI treatment, resulting in a low survival compared to those patients with low NNMT levels and a better response to TKIs. This method could be used in clinical practice for the rapid and accurate identification of patients with a high risk of progression who may benefit from more aggressive treatment and may improve therapeutic strategies for patients with NSCLC.
Next Patent: BIO BASED POLYAMIDES AND CORRESPONDING POLYMER COMPOSITIONS