BIOMARKERS AND THEIR USE

Related applications

[001] This application claims the benefit of U.S. Provisional Application No. 62/726,837, filed on September 4, 2018, the entire contents of which are incorporated herein by reference in its entirety.

Field of invention

[002] The present invention relates to the fields of molecular biology and medicine. The invention also relates to biomarkers for oral cancers. More specifically the invention relates to early cancer detection and prognostic methods .

[003] The present invention relates to non-invasive methods for the diagnosis and prognosis of cancer, and biosensor device, by determining the salivary levels of biomarkers related to oral cancers, including head-and-neck cancers and a kit.

Background of the invention

[004] Oral cancer is a public health problem, showing a high economic and social burden. It is the 6 ^th most common malignant tumor type worldwide. Until 2030, the incidence of oral cancer in general is predicted to exceed 1 million, in which 73% of cases will occur in less developed countries. For instance, this disease is highly frequent in Southern Asia (e.g. India and Sri Lanka) and in the Pacific Islands. Papua New Guinea shows the highest incidence rate worldwide in both sexes, and it is also the leading cause of cancer death among men in India and Sri Lanka. Tobacco smoking, alcohol consumption, and human papillomavirus infection are well-known risk factors for oral cancer. One of the major challenges of oral cancer is the late diagnosis. Oral cancer is associated with significant morbidity with a 5-year survival rate of 40-50% for the advanced cases. The time of diagnosis and tumor staging have a decisive influence on the survival rate, and in most cases, oral tumors are diagnosed in advanced stages (III and IV), which affects the therapeutic treatment with negative influence on prognosis.

[005] Oral squamous cell carcinoma (OSCC) is the most common type of oral cancer and is ranked in the eighth leading cause of cancer worldwide. OSCC exhibits high prevalence and morbidity with 300,000 new cases and 145,000 deaths per year worldwide. Standard multimodal management of OSCC is based on the Tumor-Node-Metastasis (TNM) classification which is a multimodal system of classification majorly used for prognosis and clinical treatment .

[006] TNM is based on the size of the primary tumor (T) , metastasis in lymph nodes (N) and occurrence of distant metastasis (M) . Still, this system presents some flaws, such as, patients diagnosed with the same TNM stage present different clinical behaviors, variations in treatment response, and variability of clinical results. Therefore, there is an urgent need to develop accessible diagnosis tools for prognosis of oral cancer to diminish the rates of mortality and contribute to the therapeutic treatment and patient monitoring. [007] Biomarkers are measurable markers or indicators of a certain biological state. Biomarkers associated to cancer are extremely useful to study the disease, identify patients in different clinical stages, and personalize therapies. The presence, absence, or changes in the levels of specific biomarkers may indicate an initial stage or progression in the development of cancer. The biomarkers may be composed by specific cells, genes, enzymes, proteins or hormones. An ideal cancer biomarker must present high sensibility and clinical specificity for efficient diagnosis and prognosis.

[008] Patent application US20050074793 discloses defined sets of genes that are used for identification and diagnosis of metastatic cancer and other conditions in a biological sample. The defined sets of genes can also be used for prognosis evaluation of a patient based on the gene expression pattern of a biological sample. The patent application US20050074793 also discloses: (a) a method of diagnosing the health status of a biological sample; (b) a method for prognostic evaluation of the metastatic potential of colorectal cancer; (c) a method for evaluating the progress of a treatment regimen for metastatic colorectal cancer; and (d) a method for evaluating the efficacy of drug candidates for use in the treatment of metastatic colorectal cancer involving the measuring of a gene expression pattern with reference values. The patent application US20050074793 also discloses a kit for diagnosing the health status of a biological sample said kit comprising nucleic acid probes that specifically bind to nucleotide sequences from reference sets and means of labeling nucleic acids. [009] Patent application US20090227533 discloses compositions and methods by identifying genes that are direct targets for miR-34 regulation or that are indirect or downstream targets of regulation following the miR-34- mediated modification of another gene(s) expression. In certain aspects of the invention disclosed in document US2090227533, compositions of the invention are administered to a subject having, suspected of having, or at risk of developing a disease, such as a cancerous disease or condition .

[0010] Patent US 7,588,895 discloses molecular diagnostic and prognostic markers for head and neck squamous cell carcinoma metastasis and extracapsular spread (CTTN, MMP9, EGFR, BMP2, GTSE1) that are independent of histopathological evaluation. The patent US 7,588,895 also discloses a method of diagnosing, or providing prognosis for, oral tongue squamous cell carcinoma metastasis or extracapsular spread of an oral tongue squamous cell carcinoma in a patient wherein the levels of expression of such markers are measured and compared with the levels of expression in a population of human subjects having a carcinoma which did not metastasize. The overexpression of the RNA representing each of the markers disclosed in patent US 7,588,895 indicates that the metastasis or extracapsular spread is present or likely to develop.

[0011] Patent application W02011130435 revels biomarkers which constitute a metastasis-associated fibroblast ("MAF") signature, and the use thereof, in diagnosing and staging a variety of cancers. The biomarkers disclosed in this document (such as a COL11A1, COL10A1, COL5A1, COL5A2, COL1A1, COL1A2, THBS2 , INHBA, VCAN, FAP, MMP11, POSTN, ADAM12 , LOX, FN1 and SNAI2 gene products), are based on the discovery by the inventors that identifying differential expression of certain genes indicates a diagnosis and/or stage of a variety of cancers with a high degree of specificity. The invention disclosed in document W02011130435 provides for methods of diagnosis, diagnostic kits as well as methods of treatment that include an assessment of biomarker status in a subject.

[0012] Patent US 8,211,659 discloses a method of detecting metastasis of cancers in the body fluids from a mammal by measurement of the CAS protein level or CAS polypeptide level to screen or diagnose the metastatic cancers. The patent US 8,211,659 also disclose a kit for assaying CAS protein or CAS polypeptide in the body fluids from a mammal to detecting metastasis of cancers, which comprises CAS protein specific antibodies.

[0013] Patent application US20170037480 relates to Heat Shock Protein-1 (HSF1) gene and HSF1 gene products in tumor stroma. In some aspects, the invention disclosed in patent application US20170037480 provides methods of tumor prognosis, treatment-specific prediction, or treatment selection, the methods comprising measuring the level of HSF1 expression or HSF1 activation in a sample obtained from the tumor that comprises tumor-associated stromal cells. In some aspects, the invention relates to the discovery that increased HSF1 expression and increased HSF1 activation in tumor-associated stromal cells correlate with poor outcome in cancer. In some embodiments of the patent application US20170037480, the methods comprise measuring HSF-1 expression or activation specifically in tumor-associated stromal cells. In some embodiments of patent application US20170037480, the methods comprise measuring HSF1 expression or activation specifically in tumor-associated stromal cells and specifically in cancer cells. In some embodiments of patent application US20170037480, HSF1 expression or activation is measured using an antibody that specifically binds to HSF1. In some embodiments of patent application US20170037480 HSF1 expression or activation is measured by measuring expression of genes that are regulated by HSF 1 in tumor-associated stromal cells. In some aspects, the invention relates to inhibiting HSF1 in tumor-associated stromal cells as an approach to cancer therapy.

[0014] The invention disclosed in the patent application US20160078 is directed to methods compositions, tangible, computer-readable medium, and apparatuses related to assessing, prognosing, and/or treating cancer patients, particularly breast cancer patients. The invention of patent application US20160078 determines the prognosis of a breast cancer patient by evaluating a specified set of genes (such as RKIP, MMP1 , OPN, HMGA2 , CXCR4 , let-7, BACH I, among others) . Specifically, methods may comprise calculating a prognosis score based on a particular algorithm. Also disclosed are compositions, kits and methods for treating cancer in a subject in need thereof involving one or more upstream activators and/or downstream effectors of TETI.

[0015] Patent application US 20160139130 discloses methods for detecting cancer or monitoring cancer progression in a subject. The method comprises detecting the level of expression of one or more cancer markers in a biological sample obtained from the subject; and comparing the level of expression of the one or more cancer markers in the biological sample to a normal level of expression of the one or more cancer markers . The one or more cancer markers disclosed in the document US 20160139130 comprises CXCL16 or CXCR6 or both CXCL16 and CXCR6. The patent application US 20160139130 also discloses a kit for detecting cancer or monitoring cancer progression.

[0016] Patent application IN 201741007724 discloses a panel of 93 preferred salivary biomarkers for the detection/diagnosis of head and neck squamous cell carcinoma (HNSCC) wherein the preferred biomarkers are S100A7, CD44, COL5A1 , S100P, COL1A1, CD44, S100A11, alAT . Those biomarkers are useful for non-invasive early detection/prognosis and were identified in a biological sample such a salivary sample from an individual at different stages of oral cancer progression thorough proteomic profiling, where selected proteins that are differently regulated were detected and identified as candidate markers. The patent application IN 201741007724 also discloses methods of diagnosing and for providing a prognosis for oral cancer and periodontal disease using those biomarkers and a kit

[0017] Patent application US20180238896 refers to uncomplexed neutrophil gelatinase associated lipocalin (NGAL) that is present at increased levels in individuals with atypical ductal hyperplasia (ADH) , a major risk factor for future breast cancer development; in individuals that have ovarian cancer; and in individuals that have breast cancer, both invasive and noninvasive. Accordingly, the invention disclosed in patent application US20180238896 is directed to measuring uncomplexed NGAL levels in urine as a primary screen to determine if an individual is either at risk of developing, or has developed cancer, e.g., cancer of epithelial origin including breast and ovarian cancer.

[0018] Patent US 9,909,185 discloses genetic markers whose expression is correlated with breast cancer. Specifically, the invention disclosed in patent US 9,909,185 discloses sets of markers whose expression patterns can be used to differentiate clinical conditions associated with breast cancer, such as the presence or absence of the estrogen receptor ESR1, and BRCA1 and sporadic tumors, and to provide information on the likelihood of tumor distant metastases within five years of initial diagnosis. Moreover, the patent US 9,909,185 refers to methods of using these markers to distinguish these conditions, kits containing ready-to-use microarrays and computer software for data analysis using the statistical methods described in this document .

[0019] The present invention proposes i) a validated method for identification and quantitation of peptides; ii) antibodies against the peptides and/or proteins for specific identification/quantification of such biomarkers, iii) In addition, it is an object of the present invention to propose a biosensor comprising these antibodies to recognize the peptides/proteins described herewith.

[0020] The use of multiple biomarkers or biological samples disclosed in the present invention can provide a significant advance in the diagnosis and prognosis of oral cancer .

Brief summary of the invention [0021] It is an object of the invention to provide biomarkers for diagnosis/prognosis of metastasis of oral cancer cells wherein the biomarkers are predictive markers of clinical aggressiveness and are the proteins identified by proteomic profiling oral cancer, consisting of the proteins CSTB, LTA4H, PGKI, NDRG1 , COL6A1, and ITGAV that are under expressed in samples from patients with metastatic oral cancer.

[0022] In one embodiment of the invention, the biomarkers of the invention are biomarkers of oral cancer such as oral squamous cell carcinoma.

[0023] In one embodiment, the invention provides the use of biomarkers present in salivary samples of a patient with oral cancer for prognosis of oral cancer metastasis wherein the biomarkers are proteins CSTB, LTA4H, PGKI, NDRG1 , COL6A1, HEMO, NGAL, KLK1 , PLEC, and ITGAV.

[0024] In one embodiment, the invention provides with a method for detecting metastasis of cancer in a human subject, the method comprising the steps of measuring proteins levels of CSTB, LTA4H, PGKI, NDRG1 , COL6A1, HEMO, NGAL, KLK1 , PLEC, and ITGAV in a human body sample using a binding assay in the body fluid, comparing the protein level found in the body fluid with the protein level in a body fluid from a subject that does not have metastasis, and establish the prognosis of the individual having cancer metastasis. In one embodiment, the metastatic cancer is indicated when the levels of the proteins CSTB, LTA4H, PGKI, NDRG1 , COL6A1, HEMO, NGAL, KLK1 , PLEC, and ITGAV in the body fluid are decreased in relation to protein levels of a normal subject being it without cancer or with non-metastatic cancer. [0025] In one embodiment the body fluid is a biological sample, such as blood, serum, lacrimal fluid, urine, pleural effusion or saliva. In one preferred embodiment the sample is saliva sample.

[0026] In one embodiment, it is measured the level of at least three target proteins selected from the group consisting of CSTB, LTA4H, PGKI, NDRG1 , COL6A1, HEMO, NGAL, KLK1 , PLEC, and ITGAV.

[0027] In one embodiment the cancer is an oral squamous cell carcinoma.

[0028] In one embodiment it is provided a method for detecting cancer metastasis in a human subject comprising the steps of measuring target peptide levels from the group consisting of Pep8_LTA4H, Pepl2_CSTB, Pep8_LTA4H, Pep9_COL6Al, Pepl2_CSTB, Pep8_LTA4H, Pep9_COL6Al, LTA4H in a human body fluid sample, comparing the peptide levels found in the body fluid with the peptide level in a body fluid from a subject that does not have metastasis, and establish the prognosis of the individual having cancer metastasis.

[0029] In one embodiment the metastatic cancer is indicated when peptide levels from the group consisting of Pep8_LTA4H, Pepl2_CSTB, Pep8_LTA4H, Pep9_COL6Al, Pepl2_CSTB, Pep8_LTA4H, Pep9_COL6Al, LTA4H, in the body fluid are decreased in relation to protein levels of a normal subject being it without cancer or with non-metastatic cancer.

[0030] In one embodiment a method for detecting cancer metastasis in a human subject comprises the measurement of at least one target peptide of the group consisting of Pep8_LTA4H, Pepl2_CSTB, Pep8_LTA4H, Pep9_COL6Al, Pepl2_CSTB, Pep8_LTA4H, Pep9 COL6A1, LTA4H. In one preferred embodiment the target peptides are Pep8_LTA4H, Pep9_COL6Al, Pepl2_CSTB.

[0031] In one embodiment the measurement technique is mass spectrometry (MS) .

[0032] In one embodiment it is provided a biosensor device for prognosis of oral cancer metastasis from a patient's saliva samples using the biomarkers described herewith wherein the biosensor is an electrochemical biosensor of nanomaterial as metal-organic structures (MOFs) where antibodies are immobilized.

[0033] In one embodiment, metal-organic structures (MOFs) is a porous zeolitic imidazolate-8 (ZIF-8) or 2- methylimidazole coated in interdigitated electrodes.

[0034] In one embodiment, the immobilized antibodies bind proteins selected from the group consisting of CSTB, LTA4H, PGKI , NDRG1 , COL6A1, HEMO, NGAL, KLK1 , PLEC, and ITGAV. In one preferred embodiment, the target proteins are at least three from the group consisting of CSTB, LTA4H, PGKI, NDRG1, COL6A1 , HEMO, NGAL, KLK1 , PLEC, and ITGAV.

[0035] In one embodiment, is provided a prognosis kit for point of care use comprising a method and biosensor to detect and measure the levels of proteins as defined herewith, the means to collect a biological sample, such as a salivary sample, reagents and instructions to use.

[0036] In one embodiment, it is provided biomarkers for diagnosis/prognosis of metastasis of oral squamous cell carcinoma wherein the biomarkers are predictive markers of clinical aggressiveness and are the proteins identified by proteomic profiling oral cancer, consisting of the proteins CSTB, LTA4H, PGKI, NDRG1 , COL6A1, HEMO, NGAL, KLK1 , PLEC, and ITGAV and its peptides that are under expressed in samples from patients with metastatic oral cancer.

[0037] In one embodiment, it is disclosed the use of biomarkers present in saliva samples of a patient with oral squamous cell carcinoma for prognosis of oral cancer metastasis wherein the biomarkers are proteins from the group consisting of CSTB, LTA4H, PGKI, NDRG1 , COL6A1, HEMO, NGAL, KLK1, PLEC, and ITGAV, and its peptides.

Brief description of the drawings

[0038] Figure 1 discloses an experimental design. A) The ITF was delimited as a 1 mm depth from the edge of the tumor slice, and the inner tumor was defined as up to 1 mm from the epithelial tumor tissue origin. In more detail, the ITF and inner tumor with small and large neoplastic islands (arrows) are surrounded by the tumor stroma (dashed lines) (scale bars, 200 pm) . B) Laser microdissection of the six regions of interest. C) and D) Protein extraction from microdissected tissues and trypsin digestion. The peptide mixture was desalted in stage tips and analyzed by liquid chromatography coupled with tandem mass spectrometry (LC- MS/MS) . E) To select protein targets, MaxQuant and the Perseus package were used to identify and quantify proteins, and R software was used for statistical analysis of the clinic pathological parameters and for analysis of the proteins with positive staining for OSCC tissues in The Human Protein Atlas. F) Targeted proteins were evaluated following two different strategies by immunohistochemistry in OSCC tissues and by SRM of saliva samples from OSCC patients with (N+) or without (NO) lymph node metastasis.

[0039] Figure 2 discloses quantitative proteome analysis that indicates spatially distinct protein signatures. A) and B) Venn diagram of common and "exclusive" proteins identified for a neoplastic islands or b tumor stroma from the ITF and the inner tumor. C) and D) Clustering analysis of proteins identified in the ITF and the inner tumor of c neoplastic islands (n =20 samples) and d tumor stroma (n =17 samples) . Values for each protein (rows) and for each microdissected sample (columns) are colored based on the protein abundance, in which high (red) and low (blue) values (Z-scored log2 LFQ intensity values) are indicated based in the color scale bar shown in the top left of the figure. The colored bars shown on the top of the figure indicate samples from the ITF (blue) or from inner tumor (pink) . Hierarchical clustering was performed in the R environment using the Euclidean distance with complete ligation for neoplastic island data and the Euclidean distance with average ligation for stromal data. E) and F) Heat map of Pearson correlation coefficients derived from pairwise comparison of the 20-patient samples for e neoplastic island samples and for the f tumor stroma samples analyzed by discovery proteomics. Log2 LFQ intensity values of the protein dataset after filtering reverse and "only by site" entries were used to calculate the correlation coefficient using Perseus software, and the heat map was constructed using R language with the function heatmap.3. The dendrogram was built using Euclidean distance with complete ligation. Samples with low correlation values and low number of quantified proteins from tumor stroma dataset were removed. G) The five most enriched GO terms that distinguished neoplastic islands from tumor stroma are represented. GO terms for cellular metabolic processes are overrepresented for neoplastic island proteins, whereas cellular adhesion and protein cleavage processes are overrepresented for tumor stromal proteins. The statistically significant proteins between neoplastic islands compared with tumor stroma (two-sided Student's t test, P value < 0.05) also indicate, among other overrepresented processes, metabolic processes for proteins upregulated in neoplastic islands.

[0040] Figure 3 discloses abundance profile of proteins selected for verification steps. A) Differentially regulated proteins between the ITF and inner tumor for neoplastic islands, as determined by plotting Student's t test (P value < 0.05, two-sided) P values versus the log2 ratio of the LFQ intensity (ITF/Inner tumor), are represented on volcano plots. Significant proteins are indicated by red dots. B) Line plots depicting the abundance profiles (log2 of the LFQ intensity) of the proteins selected for the verification step. Shown are the abundances across all MS runs from discovery proteomics analysis of neoplastic island samples. Samples from the ITF are denoted by "F", and samples from inner tumor are denoted by "I". C) Differentially regulated proteins between the ITF and inner tumor for tumor stroma (see panel a for plot details) . D) Protein abundance profiles for tumor stroma samples (see panel b for plot details) .

[0041] Figure 4 discloses immunohistochemical staining of targeted proteins. Oral SCC tissue samples from a set of 125 cases were used to verify the abundance of a CSTB, LTA4H, NDRG1, and PGK1 neoplastic island proteins, and 96 cases were used to verify the abundance of b COL6A1, ITGAV, and MB tumor stroma proteins. Among the cases, it was possible to analyze 114 cases for CSTB, 118 cases for LTA4H, 119 cases for NDRG1, 118 cases for PGK1, 93 cases for COL6A1, 80 cases for ITGAV, and 86 cases for MB. The scores that represent the sum of the intensities and percentage of protein staining in the ITF or the inner tumor are shown as a heat map. Differential staining in ITF and inner tumor in both neoplastic islands and tumor stroma was identified, in agreement with the MS discovery analysis however, we also identified negative cases for each protein and cases with gradual staining according to the color key shown at the top right. For proteins selected in neoplastic island, increased CSTB and NDRG1 expression was identified in the inner tumor, according to the MS results, with staining only inside neoplastic cells of OSCC. However, LTA4H and PGK1 presented peripheral staining in neoplastic cells and were also detected in cells from tumor stroma, such as inflammatory cells. For proteins selected in tumor stroma, greater COL6A1 and MB expression was identified in the ITF area, which demonstrates the preferential localization in OSCC regions compared with that observed in the MS discovery analysis. COL6A1, ITGAV, and MB proteins were preferentially present in tumor stroma (Histological images obtained using a ^c 40 objective, scale bars, 200 pm) .

[0042] Figure 5 discloses Kaplan-Meier survival analysis of IHC and clinical outcomes. Overall survival (OS), disease- free survival (DFS) and specific survival (SS) were available in relation to a second primary tumor, local, locoregional or lymph node relapse. A) Patients with lower CSTB expression in the ITF had a higher risk of local relapse and worse survival (P value < 0.05, log-rank test) . In addition, a lower NDRG1 expression in the ITF was associated with a higher risk of the patient presenting a second primary tumor and a worse DFS (P value < 0.05, log-rank test) . Equal expression between the ITF and the inner tumor or higher CSTB and NDRG1 expression in the ITF did not influence the local relapse or second primary tumor. Patients with higher PGK1 expression in the ITF had a worse survival and early locoregional recurrence (P value < 0.05, log-rank test) . B) Patients with higher expression of ITGAV in the ITF have a greater risk to present lymph node metastasis relapse and poor survival (P value < 0.05, log-rank test) .

[0043] Figure 6 discloses targeted proteomics of saliva proteins. A), B) Volcano analyses show log2 ratio of N+/N0 of a proteins and b peptides according to the adjusted P value. Proteins that met the indicated statistical cut-off criteria (Mann-Whitney U test, with P values adjusted for multiple comparisons using the Benj amini-Hochberg FDR method, adjusted P value < 0.05) are colored in red. C) The graph demonstrates individually the L/H intensity ratio (not log transformed) of six differentially expressed proteins CSTB, COL6A1 , ITGAV, LTA4H, PGK1, and NDRG1 between N+ and NO saliva samples. *P value <0.05, Mann-Whitney U test. D) Peptide relative quantification (log2 L/H ratio) between N+ and NO saliva samples. For each protein, 2-3 proteotypicpeptides were monitored, with exception for NDRG1, only one proteotypic peptide was monitored. The light peptide (corresponding to the endogenous peptide present in saliva) and the heavy peptide (which corresponds to the synthetic peptide spiked-in saliva) were monitored, and the light/heavy ratio for each of the 14 peptides was obtained by Skyline. Box plots represent the median and interquartile range, whiskers represent the 1-99 percentile, and outliers are represented by empty circles. E) Bar plots represent the relationship between MS Discovery analysis of tissue and SRM-MS of saliva in the identification of potential prognostic signatures. The log2 N+/N0 ratio for saliva samples and log2 ITF/inner ratio for neoplastic islands and tumor stroma from microdissected tissues are represented in the graph. F) Representative figure illustrates the gradient dynamics of the protein abundance between tissue (ITF and inner tumor) and saliva (N+ and NO), which indicates that the abundance of proteins in saliva and its association with prognosis (N+ and NO) is not necessarily associated with the proximity of the altered oral epithelium. Other components, such as water, electrolytes, DNA, RNA, and microorganisms, were not included. Images in f were adapted from files provided by Servier Medical Art (https://smart.servier.com/, licensed under a Creative Commons Attribution 3.0 Unported License) .

[0044] Figure 7 discloses prognostic signature in saliva distinguishes OSCC patients. A) Workflow for machine learning approach to measure the predictive power of peptides and proteins. B) , C) The predictive relevance of individual proteins and peptides to distinguish NO from N+ patients is represented by a bar chart indicating their cross-validation ROC AUC (100 repetitions of stratified tenfold cross- validation) . The most relevant protein and peptide ordered by the AUC is LTA4H and Pep8_LTA4H, respectively. When only the AUCs of the individual signatures (size 1) are considered, the three highest areas at the protein level are LTA4H (73.9%), COL6A1 (62.1%), and ITGAV (60.5%) and at the peptide level are Pepl2_CSTB (73.5%), Pep8_LTA4H (72.8%), and Pep9_COL6Al (71.0%) . D) Cross-validation estimated ROC curves of the best protein and peptide signatures. E) Box plots representing the AUC of all possibilities of signatures for both imbalanced and balanced (SMOTE) cross-validation. At the peptide level, 1024 signatures were tested. At the protein level, 63 signatures were tested. Signatures formed by peptides from different proteins SI {Pep8, Pepl2} and S2 {Pep8, Pep9, Pepl2} have approximately 10.5% higher AUC than the peptide signature formed by LTA4H (S4) . S2 peptide signature outperformed both SI and S4 signatures. The candidate signatures are indicated by labels: SI, S2, S3, and S4. Peptide sequences: Pepl_MB: HGATVLTALGGILK; Pep2_MB: YLEFISECI IQVLQSK; Pep3_PGKl : VLNNMEIGTSLFDEEGAK; Pep4_PGKl : VLPGVDALSNI; Pep5_ITGAV: LQEVGQVSVSLQR; Pep6_ITGAV: STGLNAVPSQILEGQWAAR; Pep7_LTA4H: LTYTAEVSVPK; Pep8_LTA4H: DLSSHQLNEFLAQTLQR; Pep9_COL6Al: GLEQLLVGGSHLK; PeplO_COL6Al : TAEYDVAYGESHLFR; Pepll_NDRGl: EMQDVDLAEVKPLVEK; Pepl2_CSTB: HDELTYF; Pepl3_CSTB: SQWAGTNYFIK; and Pepl4_CSTB: VHVGDEDFVHLR . Four peptides were not included in the training model because they did not pass the filtering step (step 2 from Part 2 of Fig. 7a; P value < 0.1, Mann-Whitney U test) . Box plots represent the median and interquartile range, whiskers represent the 1-99 percentile, and outliers are represented by "+".

[0045] Figure 8 shows a scanning electron microscopy (SEM) images for ZIF-8 deposition on the interdigitated electrode (IDE) . The pristine IDE surface before (a) and after (b) ZIF-8 deposition. The characteristic tetrahedrally coordinated MOF is observed in (c) . XRD pattern for ZIF-8 deposited for 6 hours in methanol solution, with an insert of the ZIF-8 unit cell33 made by zinc ions coordinated by four imidazolate rings (d) .

[0046] Figure 9 shows impedance measurements for modified IDE with ZIF-8 in the presence of the proteins Trx-1, Cyto and Cyto-F in the concentrations of 0.5, 2.0 and 4.0 mM in PBS solution.

[0047] Figure 10 discloses the verification of specificity of antibodies used for IHC analysis. Forty micrograms of protein extracts derived from 7 cell lines (BJ-5ta ATCC CRL-4001, a fibroblast immortalized cell line; CAF, primary oral cancer associated fibroblasts; SCC-9 ATCC CRL-1629, a tongue cancer cell line; HSC3, a tongue cancer cell line; SK-MEL-28, a malignant skin-derived melanoma cell line; MCF7 ATCC HTB22, a breast cancer cell line; and A549 ATCC CCL-185, an epithelial lung cancer cell line) were analyzed on a 10% SDS-PAGE and subjected to western blot with the same antibodies used for IHC analysis.

Detailed description of the invention

[0048] It is an object of the present invention the use of antibodies against the proteins identified above in two different methods for biomarker detection:

[0049] 'Absolute' quantitation method of OSCC markers using saliva and detected by mass spectrometry (MS) .

[0050] Point-of-care biosensor of OSCC markers using saliva and detected by an electrochemical response.

[0051] In the present invention, the inventors combine discovery and proteomics approaches to identify prognostic signatures for OSCC patients. In the initial discovery phase, the inventors integrate knowledge of the histopathology, discovery proteomics analysis of formalin-fixed paraffin- embedded (FFPE) OSCC tissues and clinical features of patients. Assessing the protein profiles of large and small neoplastic islands and their surrounding stroma by combining laser microdissection (LMD) and proteomics reveal several proteins including CSTB, NDRG1 , LTA4H, PGK1. COL6A1, ITGAV and MB with distinct expression patterns between ITF and inner tumor, suggesting a potential prognostic value by clinic-pathological association analysis. In the subsequent targeted phase, it was used two follow up approaches to verify these signatures in two patient cohorts. First, analysis of clinical significance and immunohistochemical staining are performed in 125-OSCC patient cohort, indicating CSTB, at low expression levels in the ITF, as an independent marker for local recurrence. Second, selected reaction monitoring mass spectrometry (SRM-MS) is applied to study the abundance of the above-mentioned seven proteins in saliva samples from an independent 40-OSCC patient cohort. Analyzing the SRM-MS results with machine-learning approaches demonstrates that a combination of LTA4H-, COL6A1-, and CSTB-specific peptides in saliva are able to distinguish patients with and without lymph node metastasis with good estimated prediction performance, outperforming predictors based on individual or grouped proteins.

[0052] The proteins identified above are used in the present invention as a prognostic signature that may assist in the clinical decision-making process leading to appropriate treatment, thus improving the prognosis and survival of patients with OSCC.

Definitions

[0053] The term "oral cancer" refers to a group of neoplasms affecting any region of the oral cavity, pharyngeal regions and salivary glands. This term is used interchangeably with the term "oral squamous cell carcinoma" or "OSCC" along this document and refers to malignant epithelial neoplasm affecting the oral cavity. It is estimated that more of 90% of all oral cancers are OSCC.

[0054] The term "metastasis" used herewith refers to the spreads of cancer from the primer tumor or origin to other tissues and parts of the body, such as lymph nodes.

[0055] The term "prognosis" used herewith refers to the prediction of the likelihood of metastasis, predictions of disease free and overall survival, the probable course and outcome of cancer therapy, or the likelihood of recovery from cancer, in a subject.

[0056] The term "diagnosis" used herewith refers to the identification of the disease state, such as a cancer or metastatic cancer, in a subject.

[0057] The terms "overexpressed", "overexpression" or overexpress" interchangeably refer to protein or nucleic acid (or both) that is transcribed or translated at a detectably greater level, usually in a cancer cell, in comparation to a normal cell.

[0058] The term "biological sample" used herewith refers to any type of biological specimen such as cells, internal organs, veins, or even fluids. Examples of different types of biological samples include, but are not limited to, blood, plasma, serum, tissue (from biopsy, etc.) _/ urine, saliva, oral cells, etc.

Results

[0059] Proteins that were spatially organized in distinct histological areas of tongue squamous cell carcinoma were identified. For this purpose, the inventors mapped the proteome of neoplastic islands and their surrounding tumor stroma from the ITF and inner tumor FFPE tissue samples from 20 patients (Fig. la, b) . Using histology-guided LMD, it was isolated six different areas of the tumor: (1) small neoplastic islands from the ITF; (2) large neoplastic islands from the ITF; (3) small neoplastic islands from the inner tumor; (4) large neoplastic islands from the inner tumor; (5) stroma from the ITF; and (6) stroma from the inner tumor.

[0060] The six proteomes were analyzed using quantitative mass spectrometry and label-free protein quantitation (LFQ intensity) to compare the relative abundance of the proteins (Fig. lc-e) . The reproducibility and correlation coefficient among the LFQ intensities of discovery proteomics data are illustrated in Fig. 2e and f.

[0061] Combining the data for small and large neoplastic cells resulted in the quantitation of 2049 proteins from the ITF and inner tumor. After excluding reverse sequences and those identified "only by site" entries and considering proteins with at least ten valid LFQ intensity values in at least one group (20 samples), 799 proteins were confidently identified (ITF and inner tumor) . For the tumor stroma dataset (ITF and inner tumor), 1733 proteins were quantified. After excluding reverse sequences and those identified "only by site" entries and considering proteins with at least eight valid values in at least one group (17 samples), 704 proteins were quantified.

[0062] The proteomic data analysis identified common and exclusive proteins from neoplastic islands (Fig. 2a) and tumor stroma (Fig. 2b) . The filtered dataset of the neoplastic islands from the ITF and inner tumor was subjected to statistical analysis using Student's t test (P value < 0.05), which resulted in 32 proteins with differential abundances. Similarly, paired Student's t test (P value < 0.05) indicated 101 proteins that were differentially expressed between the tumor stroma from the ITF and the inner tumor .

[0063] Unsupervised hierarchical clustering analysis of the identified proteins showed not only the proteomic diversity between samples from the ITF and the inner tumor, but also the variation within the ITF samples and the inner tumor samples for both neoplastic island (Fig. 2c) and tumor stroma proteomes (Fig. 2d) . The lower clustering identified in neoplastic islands compared with tumor stroma may be associated with different abilities to extract proteins from FFPE tissues and may also reflect the intrinsic tumor heterogeneity, as OSCCs are known for their biological variability, which leads to specific clinical behaviors, i.e., it has been observed that tumors at the same stage may present different clinical outcomes.

[0064] To investigate whether the biological processes could spatially separate neoplastic islands from tumor stroma, the inventors searched for biological processes in the Gene Ontology (GO) database that were enriched for proteins uniquely identified in each proteome. Cellular metabolic processes primarily represented the neoplastic island proteins, whereas cellular adhesion processes and protein cleavage processes overrepresented the tumor stroma proteins (Fig. 2g) , which indicate that proteome annotation discriminates neoplastic islands from adjacent stroma. Moreover, the analysis of the 601 significant proteins, which were significantly different between the neoplastic islands and the tumor stroma (Student's t test, P value < 0.05), also indicates metabolic processes overrepresented, among other annotations, for upregulated proteins of neoplastic islands ( Fig . 2g) .

[0065] We used linear regression to analyze the proteome LFQ dataset and clinic-pathological data to identify the proteins associated with patient features (Table 1) . The majority of proteins (ACTR2 , CSTB, LTA4H, PGK1, NDRG1 , FSCN1, ITGAV, and THBS2) significantly associated with clinical parameters showed lower expression in the ITF of the tumor stroma or neoplastic islands, with the exception of COL6A1, COL1A2 , S100A8, S110A9, and MB.

Prioritization of proteins for IHC and SRM-MS analysis.

[0066] The targeted proteins evaluated in the subsequent steps of verification using immunohistochemistry (IHC) in a 125-patient cohort and SRM in an independent 40-patient cohort were selected if they filled the following criteria: (1) only proteins with different protein abundances between the ITF and the inner tumor in the discovery phase (Student's t test, P value < 0.05); (2) only proteins that present a significant association with clinical characteristics of patients (Linear regression, P value < 0.05, R < -0.7 or 0.7 < R and R2 > 0.4) (Table 1); (3) only proteins with positive staining of squamous cell carcinoma in HNSCC in The Human Protein Atlas (https://www.proteinatlas.org/); and (4) only proteins not cited or cited only in limited studies related to oral cancer.

[0067] Cystatin-B (CSTB) , leukotriene A-4 hydrolase (LTA4H) , protein NDRG1 (NDRG1), and phosphoglycerate kinase 1 (PGK1) from the neoplastic island dataset and collagen alpha-1 (VI) chain (COL6A1), integrin alpha-V (ITGAV) and myoglobin (MB) from the tumor stromal dataset were prioritized (Fig. 3) . All these proteins, according to the specialized literature and to the domain predictions performed here, are nonclassically secreted.

IHC analysis of prioritized proteins.

[0068] IHC analysis was performed using 125 FFPE OSCC cases for neoplastic island proteins and 96 FFPE OSCC cases for tumor stroma proteins. For the IHC analysis, an adopted score system was used to differentiate the staining among the two regions, the ITF and the inner tumor, in a blinded and independent manner by three pathologists (kappa = 0.706) . Protein abundance varied according to the score created on a scale of 0 to 6, which generated a staining scale for all cases ( Fig . 4 ) .

[0069] The IHC of the cases (approximately 14 cases for each selected protein) used in the discovery phase showed similar differential protein abundances and locations between the ITF and the inner tumor (Table 1) for both the neoplastic islands and tumor stroma, despite the distinct dynamic range of the techniques. Further, the number of IHC cases was increased to 125 cases of neoplastic islands and 96 cases of tumor stroma, and most of the IHC results are similar to those for the original 14 IHC cases. However, with this significant increase in the size cohort, LTA4H, PGK1, and ITGAV staining were identified with a slight variation within the lower and higher scores in each region, either ITF or inner tumor. In addition, in the IHC analysis of neoplastic island proteins (Fig. 4a), LTA4H and PGK1 showed peripheral staining in neoplastic cells and were also detected in cells in the tumor stroma, such as inflammatory cells. Increased CSTB and NDRG1 expression was identified in the inner tumor, according to the MS discovery results, with staining only inside neoplastic cells of the OSCC, but in some cases, CSTB and NDRG1 were also detected in the adjacent normal epithelium. In turn, for tumor stroma proteins (Fig. 4b) , higher COL6A1 and MB expression was identified in the ITF according to the MS discovery analysis. Staining for COL6A1, ITGAV, and MB proteins indicated a preferential localization in the tumor stroma, with staining also identified in neoplastic cells. Positive staining in muscle was detected for MB. Taken together, nuclear and cytoplasmic staining was identified for all proteins; however, the marked proteins were more pronounced or more specific to the cell type in which they were identified. For the specificities of the antibodies employed here, please see Figure 10.

Integrating IHC data with clinic-pathological parameters.

[0070] In the 125- and 96-patient cohorts, IHC combined with clinic-pathological characteristics strengthened the prognostic values for the selected proteins from the OSCC samples .

[0071] Cross-tabulation and the chi-square test indicated significant associations between the clinic-pathological parameters and CSTB, PGK1, COL6A1, and ITGAV expression. Kaplan-Meier survival analysis in the OSCC patients indicated the association of lower abundance of CSTB and NDRG1 in the ITF with local relapse and second primary tumor, respectively, and higher abundance of PGK1 and ITGAV in the ITF with loco regional relapse and lymph node relapse, respectively (Fig. 5) . CSTB, PGK1, and ITGAV were significant for the 5-year disease-specific survival and disease-free survival, while NDRG1 was only significant for the 5-year disease-free survival (Table 2) .

Table 2 Kaplan-Meier survival analysis of OSCC patients according to the levels of protein expression obtained by IHC analysis.

HR hazard ratio, Cl confidence interval, ITF invasive tumor front. CSTB was associated with local relapse in Cox multivariate analysis with 3-year disease-free survival rate of 82% for 44 patients, thus confirming this protein as an independent prognostic marker for OSCC patients. Also, this analysis showed that the lower abundance of CSTB in the ITF was associated to a higher risk of developing local recurrence (HR 0.1224, 95% Cl 0.0153-0.9801, P value = 0.0478, Cox multivariate analysis) .

SRM-MS of prioritized proteins in saliva of OSCC patients. [0072] Saliva samples from 40 OSCC patients were obtained to monitor peptides derived from CSTB, NDRG1, LTA4H, PGK1, COL6A1, ITGAV, and MB based on the previously described criteria for protein selection. The samples were divided into two groups: patients without lymph node metastasis (NO) and patients with lymph node metastasis (N+) .

[0073] Among the seven proteins investigated in saliva, six proteins, including CSTB, LTA4H, PGKI, NDRG1 , COL6A1, and ITGAV, showed lower abundances in the saliva of the patients with lymph node metastasis (N+) than the patients without lymph node metastasis (NO) (Mann-Whitney U test, with P values adjusted for multiple comparisons using the Benj amini-Hochberg FDR method, adj . P value < 0.05, Fig. 6a, c) . Another six proteins investigated presented the same results HEMO, NGAL, LSR, KLK1 , PLEC, KLK1 (Table 3) .

Table 3. Proteins with differential abundance in oral cancer

[0074] At the peptide level, we also determined that the majority of peptides showed statistically significant lower abundances in the N+ patients than in the NO patients (Mann-Whitney U test, with P values adjusted for multiple comparisons using the Benj amini-Hochberg FDR method, adj . P value < 0.05, Fig. 6b, d) .

[0075] Moreover, the inventors performed an analysis of the relationship between the tissue and saliva based on the two MS techniques employed, DDA and SRM (Fig. 6e) . Most evaluated proteins exhibited lower abundance at the site of ITF and in saliva, which correlated with a poor prognosis. However, the overall results indicate that the abundance of proteins in saliva and its association with prognosis (N+ and NO) is not necessarily associated with the proximity of the altered oral epithelium (Fig. 6f) .

Saliva prognostic_signatures_distinguish_regional metastasis .

[0076] The low expression of the proteins LTA4H, PGK1, NDRG1, COL6A1, and ITGAV in the saliva samples is associated with lymph node metastasis and advanced clinical staging ( crosstabulation and chi-square test, P value < 0.05) .

[0077] Further, through the strategies of machine learning described (Fig. 7a), we evaluated the predictive power of individual and groups of peptides and proteins to distinguish the patient with lymph node metastasis (N+) from the patient without lymph node metastasis (NO) (Fig. 7b-e, Table 4) . The groups SI: (Pep8_LTA4H, Pepl2_CSTB) , S2 :

( Pep8_LTA4H, Pep9_COL6Al, Pepl2_CSTB) , S3: (Pep8_LTA4H, Pep9_COL6Al ) , and S4: (LTA4H) are the most relevant signatures (Si) considering accuracy and AUC (Fig. 7d; Table 4 , Panel 1 ) . [0078] The inventors determined that the signatures with the highest accuracies did not have the highest AUC values, which may be explained by the class imbalance between NO and N+, further confirmed through the oversampling analysis. The AUC of the peptide level is considerably higher than that of the protein level, 82.8% ( S2 ) compared with 73.9% ( S4 ) . Only the signature S4 (LTA4H) was selected at the protein level, with an AUC of 73.9%, as other signatures have AUCs less than 62.5%. Further, balancing the training subsets also increased the overall prediction performance (Fig. 7b, e, Table 3, Panel 1) .

[0079] Furthermore, the signatures SI and S2 at the peptide level and S4 at the protein level are the best candidates for both types of cross-validation, using imbalanced and balanced classes. In addition, the S2 is the best signature to discriminate NO and N+ of OSCC.

[0080] It is also one object of the present invention use the antibodies disclosed herewith in two different methods for biomarker detection:

[0081] 'Absolute' quantitation method of OSCC markers using saliva and detected by mass spectrometry (MS) .

[0082] Point-of-care biosensor of OSCC markers using saliva and detected by an electrochemical response.

'Absolute' quantitation method of OSCC markers using saliva and detected by mass spectrometry (MS) .

[0083] Mass spectrometry (MS) has recently expanded in strategies based in Discovery and Targeted based proteomics. The identification of biomarkers by Discovery based mass spectrometry has the goal to identify the largest number of proteins without previous knowledge of the sample. In targeted based, the goal is to quantify a small group of previously known proteins. Besides Selected Reaction Monitoring, SRM, in MS has appeared as a promising strategy capable of quantifying potential biomarkers with high reproducibility, sensibility, and accuracy simultaneously, dynamic linear range (as a result of the efficiency of the quadrupoles and detector ions) and specificity (due to the two steps of selection performed in serial and the transitions that are monitored in relative high speed) .

[0084] Triple-quadrupole mass spectrometers (QqQ-MS) suit perfectly the high standard requirements for targeted clinical proteomics. Its architecture composed by three quadrupoles operating in tandem as mass filter (Ql), collision cell (Q2) and mass filter (Q3) enables respectively the isolation of the targeted peptide ion, followed by its collision-induced dissociation (CID) and the selection of product ions that will pass to the detector. Notably, there is an interdependence between each precursor ion and its dissociation products that ensures greater analyte selectivity. Together, these pairs of precursor and product ion m/ z are called mass transitions. Under defined settings of targeted SRM acquisition only analyte ions having the specified mass transition (precursor/product ion pair) reach the detector. This results in a high specificity and minimizes interferent detection, what ultimately leads to increased signal-to-noise and improved sensitivity. This acquisition mode is referred to as Selected Reaction Monitoring (SRM) , or Multiple Reaction Monitoring (MRM) when multiple transitions are monitored over a chromatographic gradient .

[0085] Importantly, MS-signal of peptide ions is not inherently quantitative. Therefore, the presence of internal standards (IS) is required when the concentration of analytes needs to be accurately measured. Stable isotope-labelled synthetic peptides are often used as IS serving as a reference for 'absolute' quantification of the target peptides. Since the isotopic labelling results in a mass shift, the surrogate peptides can be spiked into the sample (ID-MS) and resolved from the endogenous analyte by specific mass transitions. The relationship between the MS-signal and known IS concentration is then used to determine the concentration of the endogenous peptide. Besides, reference peptides are added to the sample, known as iRT, to control the quality of the MS runs, evaluating changes in retention time and peptide intensity among runs. Alternatively, a calibration curve can be monitored between the samples permitting the interpolation of the analyte's concentration.

Point-of-care biosensor of markers using saliva and detected by an electrochemical response.

[0086] In recent years, there is an increased interest in the improvement of biomarker immobilization strategies in nanomaterials to develop cancer detection biosensors. In addition, biosensors have a high potential for simultaneous detection of multiple biomarkers.

[0087] The incorporation of nanomaterials in biosensors has been pointed out as advantageous technological progress since such a strategy allows improving its analytical performance. Nanomaterials play several roles in biosensors with respect to their optical, thermal and electronic properties. In addition, nanomaterials have higher surface area and dimensional similarity to biomolecules, enabling the development of functional hybrid systems, providing high stability and detection sensitivity. In addition to improved device performance, emerging nano-biosensor technologies allow the construction of a biomarker amplification system to facilitate detection in low disturbance regimes.

[0088] Among the types of nanomaterials, metal-organic structures (MOFs) are an attractive class of crystalline materials consisting of metal ions interconnected by organic spacers which have a high surface area, high porosity, and several available active sites. In biological sensors, this type of material stands out due to its ability to detect biomoleculesl 6. The interaction between biomarkers and MOFs occurs mainly through surface modifications. Antigen- antibody interactions for biomolecule immobilization show better specificity and detection sensitivity. Thus, the immobilization of biomarkers in MOFs can be detected by an electrochemical response.

[0089] Different types of biosensors based on electrochemical effects have been developed in recent years from MOFsl8, but there are few analytical applications of biomarker / MOFs biosensors for cancer diagnosis and prognosis. Recently, MOF known as porous zeolitic imidazolate-8 (ZIF-8) has been used to detect sarcosine, a prostate cancer biomarker. In that work, ZIF-8 was synthesized and doped with platinum nanoparticles (Pt @ ZIF- 8) . The porous structure of ZIF-8 helped stabilize platinum nanoparticles while maintaining their high catalytic activity. The modified Pt @ ZIF-8 biosensor showed a good response to sarcosine due to its unique structure and morphology. The ZIF-8 structure is well known for its satisfactory electrochemical response in biosensors and outstanding chemical stability. In addition, the affinity of its organic binder (2-methylimidazole) with common protein functional groups (e.g., hydroxyl, carboxylic acid, and amine) can facilitate crystal nucleation and biomarker immobilization .

[0090] However, studies have focused on single analyte analysis, and simultaneous analysis of multiple biomarkers is of great importance in cancer diagnosis and prognosis. The use of multiple biomarkers may improve sensitivity for disease detection, but it is necessary to maintain the specificity of interaction between the biomarker and the MOF. MOFs may provide different types of interaction with the biomarker, and the strategy commonly used to maintain binding specificity is to employ several antibodies, each specific for an antigen (biomarker) .

[0091] The main requirements considered in the interaction of antibodies with MOFs are related to orientation control, minimization of chemical modification of the antibody and reduction of non-specific surface adsorption. Recently, it has been shown that ZIF-8 structures doped with gold nanorod and conjugated with antibodies showed high efficacy in preserving biosensor bio-recognition capabilities .

[0092] The present point-of-care biosensor use saliva in order to allow the detection of the following biomarkers CSTB, LTA4H, PGKI, NDRG1 , COL6A1, HEMO, NGAL, KLK1 , PLEC, and/or ITGAV for oral cancer to be more easily measured and rapidly interpreted resulting in early diagnosis and in the increase of patients' survival. Saliva is a promising source of biomarkers because it is in contact with the active lesion, non-invasive fluid, inexpensive, and easy to collect. The proposed biosensor allows the immobilization of multiple biomarker candidates of oral cancer in metal- organic frameworks (MOFs) nanomaterials.

'Absolute' protein quantitation using Selected Reaction Monitoring

[0093] Protein digestion and synthesis of highly purified proteotypic peptides : The protocols for saliva quantification and tryptic digestion is well known for a person skilled in the art.

[0094] Synthetic peptides isotopically labeled for SRM assays were acquired, in which the cysteine residues will be carbamidomethylated and the C-terminal of lysine and arginine peptides will be uniformly labeled with 13C and 15N. These synthetic peptides will be used as references since they present the same physical-chemical characteristics of the corresponding endogenous peptide, without the isotopic label.

[0095] The highly purified heavy labeled synthetic peptides are necessary because they allow the absolute quantification of the targets as well as to determine and confirm the identity of the endogenous peptide with more confidence (by comparing the intensity of the transitions between the light and heavy peptide) . In addition, it helps to normalize the quantity of the endogenous peptide (considering that the variations of intensity and retention time that occur for the isotopic heavy labeled peptide are also true for the light peptide form) .

[0096] SRM method: Four prototypic peptides were selected for 'absolute' quantification as predictive markers of OSCC in human saliva ( LTYTAEVSVPK and DLSSHQLNEFLAQTLQR, LTA4H; GLEQLLVGGSHLK, COL6A1; HDELTYF, CSTB) .

[0097] Clinical saliva samples will be spiked with stable isotope-labelled synthetic peptides (>98% purity; MRM proteomics) and run intercalated with a blank matrix and calibration curves prepared with a pool of saliva spiked with the IS spanning over 80-0.15625fmol/ug total protein. Limit of detection (LOD) and limit of quantification (LOQ) will be determined. Peak areas of the monitored transitions will be integrated and peptide intensity determined for each channel. Linear regression of calibration curves will be used to interpolate the concentration of the endogenous peptides .

[0098] The calibration curves and saliva samples will be analyzed by SRM in a triple quadrupole Xevo TQ-XS (Waters, Milford, MA, USA) equipped with an electrospray source (IonKey, Waters, Milford, MA, U.S.A) and controlled by the software MassLynx (version 4.2) . An aliquot containing 1 pg of the peptide mixture, will be separated in the analytical column BEH Shield C18 IonKey (10 cm x 150 pm ID packed with 1,7-pm C18 particles, Waters, USA) heated at 40oC. Peptides will be maintained at 4oC inside the autosampler and injected in the column by Acquity UPLC-Class M LC autosampler (Waters, Milford, MA, USA) . The instrument will be operated by the programed SRM mode (scheduled retention time, 3 min window and cycle time of 3s), with the resolution unit of 0,7 Th FWMH on mass analyzers Q1 and Q3. The mass spectrometer was operated in the positive mode in ESI, capillary temperature at 150°C, source at 30V, capillary voltage at 3,5kV. Cone voltage adjusted to 50V, cone gas flow at 150L/h, nanoflow gas at 0,4 barr and collisional energy (CE) was determined by software Skyline. The dwell time for all the monitored peptides was calibrated automatically by the software MassLynx (version 4.2), with the minimum of 0.010s, corresponding approximately to 10 or more acquisition points per LC peak. Internal peptide standards labeled with isotopes in the C-terminal portion of the peptide, at lysine or arginine, +8 or +10 Da, respectively, (Thermo Scientific, Sigma-Aldrich) , was used to validate and normalize the intensity of the endogenous peptides based on chromatographic profiles and fragmentation properties.

[0099] In addition, the calibration curves, saliva samples and system suitability control were run in a blocking and randomization scheme to control the stability of the system and to prevent any bias during the mass spectrometry analysis. The running order of the samples was prepared considering the two groups of interest, NO and N+, including technical replicates, in R environment.

[00100] Data Analysis : The analysis of the proteomic data obtained from SRM will be performed by Skyline software that will be used to visualize and inspect the chromatographic peaks obtained by SRM data. Both peptides, light and heavy counterparts, was verified regarding the data quality observing the alignment in elution time, co-elution of all three transitions, relative intensity regarding correlation with the spectral library (dotp close to 1), relative intensity regarding correlation of the transitions from light and heavy peptides (rdotp close to 1), proximity to the predicted retention time and reproducibility in terms of retention time and intensity among technical replicates. Besides, the reproducibility based on the sample group was evaluated by Pearson correlation.

[00101] The targeted data was analyzed by the MSstats tool for statistical analysis and relative quantification of peptides and proteins, quality control of samples and statistical tests for significant changes in the abundance of proteins in regards to mixed linear models. The tool is available as a tool in Skyline or through packages available in R. With the results, we expect to validate SRM assay for clinical application to be used for biomarker (prognostic signature) quantification in the clinical routine.

Protein quantitation using Point-of-care biosensor

[00102] The point-of-care biosensor was constructed using conventional photolithography in S1O2 substrates from Cr/Au (20/20 nm) interdigitated electrodes (IDEs) . Then the IDEs were cleaned in a warm ultrasonic bath (37°C) in the following sequence, 10 min in acetone, 10 min VLSI acetone, 10 min VLSI isopropanol, and 10 min in deionized water. After that, the IDEs were dried under nitrogen flow. Following the cleaning of the IDEs, they were rinsed in Piranha solution (H2S04/H202, 60:40 (v/v)) for 20 min at 65°C to obtain -OH in the surface. The IDEs were rinsed with deionized water and methanol. The zeolitic imidazole framework (ZIF-8) is the MOF used as an active layer. Methanolic stock solutions of Zn (NO3) 2 · 6H2O (25 mM) and 2-methylimidazole (50 mM) were first prepared. The deposition was performed by dipping the substrates into a fresh mixture of 10 ml Zn(N0 ₃ ) ₂ stock solution and 10 ml 2-methylimidazole stock solution for 6 to 10 hours, under magnetic stirring, and at room temperature to grow ZIF-8 film. Then the fabricated films were washed with methanol and dried under nitrogen flow.

[00103] Morphological images of the surface were obtained using a high-resolution field emission SEM (FEI Quanta 650 FEG, Thermo Scientific) with a 5 kV acceleration voltage. The crystal structure of the coated ZIF-8 thin film was measured using XRD (X'Pert PRO MRD XL, Panalytical) .

[00104] To analyze the sensibility of the ZIF-8 coated interdigitated electrodes (ZIF-8 IDE) it was performed an electrochemical analysis to detect the interaction between the proteins thioredoxin-1 (Trx-1) and the cytoplasmic domain of ADAM-17 protein (Cyto) , once the Trx-1 was previously characterized as a partner of Cyto. Furthermore, the mutant form ADAM-17cyto F730A (Cyto-F) show less interaction with Trx-1 and can be used as a negative control. The I-V characteristics of the ZIF-8 IDEs were obtained by the semiconductor characterization system (4200-SCS, Keithley) . The measurements were conducted at atmospheric pressure and at room temperature.

[00105] The concentrations of 0.5, 2.0, 4.0, and 8.0 mM of Trx-1, Cyto, and Cyto-F were prepared in PBS solution and kept under 4°C. After the I-V characterization, 20 pL of 0.5 mM Trx-1 solution was dropped over the ZIF-8 IDE active area for 15 min. Then the substrate was rinsed in PBS solution and dried with nitrogen flow. The electrochemical characterization was performed from 105 Hz to 0.01 Hz using an impedance analyzer (SI1260, Solartron) at ambient conditions. In sequence, 20 L of 0.5 mM Cyto or Cyto-F solution was dropped over the ZIF-8 IDE active area with immobilized Trx-1 protein for 15 min. Then the substrate was rinsed in PBS solution and dried with nitrogen flow. The impedance analysis was performed in the same conditions as described before.

[00106] All the different concentrations of the proteins were analyzed following the described procedure to obtain a Bode and Phase Angle plots. Moreover, the analysis was also performed using only PBS to detect the solvent effect.

[00107] Figure 9 shows the surface SEM images respectively, of the pristine IDE (a) and the IDE modified with ZIF-8 (b) . The surface images also show a granular tetrahedrally coordinated structure for ZIF-8 (c) , supporting the assumption that the intergranular pores are not blocked, a vital requirement to ensure free access of the proteins or antibodies in further experiments. In Figure 2d the XRD pattern of ZIF-8 is shown. The presence of high intensity (110), (200), (211), (220), (310), and (222) peaks verified the formation of ZIF-8, on which the (200) orientation is predominant.

[00108] Figure 10 shows the impedance analysis for IDE modified with ZIF-8 in the presence of different proteins, in three concentrations (0.5, 2.0 and 4.0 mM) . The IDE modified with ZIF-8 showed a very high impedance (|Z| > 1011 W at 10 mHz) and a highly capacitive character. After the deposition of Trx protein a decrease in impedance is observed. At 4.0 mM an increase in impedance was observed due to the higher concentration of Trx protein. On the other hand, the higher the concentration of the Trx protein, the lower was the impedance after Cito deposition, which can demonstrate an unexpected interaction between Trx and Cito observed by the inventors . Such unexpected interaction led the inventors to develop the present invention. Regarding Cito F protein, no significant difference from Trx is observed at 0.5 mM. However, with an increase in concentration (2.0 and 4.0 mM) , the Cito F impedance is very close to PBS solution.

References

- Almangush, A. et al . A simple novel prognostic model for early stage oral tongue cancer. Int. J. Oral Maxil lofac . Surg. 44 , 143-150 (2015) ,

Aragao A. Z . et al. Identification of novel interaction between ADAM17 (a disintegrin and metalloprotease 17) and thioredoxin-1. J Biol Chem. Dec 14 ; 287 (51) : 3071-82 (2012) .

Brandwein-Gensler, M, et al. Oral squamous cell- carcinoma: histologic risk assessment, but not margin status, is strongly predictive of local disease-free and overall survival. Am. J. Surg. Pathol. 29, 167-178 (2005) .

- Brandwein-Gensler, M. et al . Validation of the histologic risk model in a new cohort of patients with head and neck squamous cell carcinoma. Am. J. Surg. Pathol. 34, 1 (2010) ,

- Costa, L. C. M. C. et al . Expression of epithelial-mesenchymal transition markers at the invasive front of oral squamous cell carcinoma. J. Appl . Oral Sci, 23, 169-178 (2015) ,

Hier, M., Mlynarek, A., Kowalski, L.

P. & Alaoui-Jamali , M A. Recurrent oral cancer: current and ernerging therapeutic approaches . Front . Pharmacol, 3, 149

- Edge, S. B. & Compton, C. C. The American Joint Committee on Cancer: the 7th edition of the AJCC cancer staging manual and the future of TNM. Ann. Surge Oncol. 17, 1471-1474 (2010)

- Ferlay, J. et al . Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012. Int. J. Cancer 136, E359-E386 (2015) .

- Ganly, I., Patel, S. & Shah, J. Early stage squamous cell cancer of the oral tongue-clinicopathologic features affecting outcome. CancerllB, 101-111 (2012) .

Granato D.C. et al. Thioredoxin-1 Negatively Modulates ADAM17 Activity Through Direct Binding and Indirect Reductive Activity. Antioxid Redox Signal. Sep 10; 29(8) : 717-734 (2018) ,

Mehanna, H, et al. Appendix—PET-CT surveillance versus neck dissection in advanced head and neck cancer. N. Engl . J. Med. 374, 1-11 (2016) ,

- Romanska, H. M, et al. Loss of CD151/Tspan24 from the complex with integrin a3b! in invasive front of the tumour is a negative predictor of disease-free survival in oral squamous cell carcinoma. Oral Oncol. 49, 224-229 (2013) .

- Sawazaki-Calone, I. et al . The prognostic value of histopathological grading systems in oral squamous cell carcinomas Oral Di s . 21, 755-761 (2015) ,

3cullv, C. & Oral squamous cell carcinoma overview. Oral Oncol. 45, 301-308 (2009) .

Sharma, M. , San, P. , Snarma, S. & Radhakrislman, R, Molecular changes in invasive front of oral cancer, J. Oral Maxi11 ofac , Pa thol , 17 , 240 (2013) .

- Tsai, J. H. & Yang, J. Epithelial-mesenchymal plasticity in carcinoma metastasis. Genes Dev. 27, 2192-2206 (2013) ,

- van der Waal, I. Are we able to reduce the mortality and morbidity of oral cancer; some considerations. Med. Oral Patol. Oral Clr. Bucal 18, e33-e37 (2013) .

[00109] It is understood that the examples and embodiments described herein are for illustrative purposes only and that various modifications or changes in light of thereof will be suggested to a person skilled in the art and are included within the spirit and purview of this application and scope of appended claims. All publications, patent applications and patents cited herein are hereby incorporated by reference in their entirety for all purposes.