Developed over five years, the study had the participation of researchers from two Unicamp units and national and international institutions
Researchers identified a correlation between the abundance of proteins present in tumor tissue and saliva with the progression of mouth cancer.
The discovery appears as a parameter capable of anticipating or predicting the progression of the disease – whether there is the presence or absence of metastasis in the cervical lymph node, for example –, in addition to overcoming the limitations of clinical and imaging tests used in the clinic and guiding the choice the ideal type of treatment for each patient.
The study began in the discovery phase through proteomic analysis of different areas of the tumor tissue using 120 microdissected samples and in the verification phase the prognostic signatures were confirmed in approximately 800 tissue samples using the immunohistochemistry technique. – localization of antigens in tissues, exploiting the principle of specific binding of antibodies to antigens in biological tissue – and 120 saliva samples from patients with the disease by target-based or directed proteomics.
“The data set led us to have a robust and very promising result in defining the severity of the disease. In addition to suggesting potential markers of the disease in a first phase, we also verified these markers in a second phase of the research, which gives more reliability to the findings, showing that these markers are efficient for classifying patients with cervical lymph node metastasis,” he said. Adriana Franco Paes Leme, researcher at the National Biosciences Laboratory (LNBio), at the National Center for Research in Energy and Materials (CNPEM), and corresponding author of an article published in Nature Communications. about the study.
The work, supported by FAPESP, was conducted at CNPEM in partnership with the Cancer Institute of the State of São Paulo (Icesp) and the Faculty of Dentistry of Piracicaba (FOP) at Unicamp, the Institute of Mathematical and Computing Sciences at the University of São Paulo (USP), the Computing Institute (IC) at Unicamp and the Faculty of Dentistry at the State University of Western Paraná, among other national and international institutions.
Oral cancer, also called squamous cell carcinoma (SCC), is the most common type of malignant tumor of the head and neck. It has high prevalence and mortality, with around 300 new cases diagnosed per year worldwide and 145 deaths. Although it is relatively easy to detect through mouth sores identified by dentists, the diagnosis is generally made when the disease is already at an advanced stage.
“The study took five years until we reached this discovery. It was divided into two phases. In the first, we use discovery-based proteomics, when we identify and quantify proteins from tumor tissues. In the second phase of the study, analyzes were carried out using immunohistochemistry and also target-based or directed proteomics – where we knew exactly which proteins we needed to quantify”, said Paes Leme.
Proteomics is the study of a set of proteins in a sample, whether in tissue or cell, for example, where it is possible to identify, quantify, determine modifications, localize, evaluate activity and protein interactions.
Bioinformatics
In the first phase, the researchers mapped the proteins in oral cancer tissue using laser microdissection and proteomics and correlated them with the patients' clinical characteristics. This evaluation allowed the identification of several proteins, such as CSTB, NDRG1, LTA4H, PGK1, COL6A1, ITGAV and MB, with different abundance patterns depending on the tumor area evaluated and association with important clinical outcomes.
In the second phase, after identifying and quantifying the proteins in the 120 tumor tissue samples, the researchers used two strategies to verify the proteins.
“In one strategy, we evaluated the abundance of selected proteins in independent tissue samples from patients using antibodies through immunohistochemistry. Another strategy was to use patients’ saliva, in which we monitored these same pre-selected targets,” Paes Leme told FAPESP Agency.
She explains that the fluid was chosen since the cancer lesion is located in the mouth, where neoplastic cells could secrete proteins.
“Saliva is a promising source of markers, in addition to being a fluid obtained through non-invasive collection. To this end, the proteins in the saliva of 40 patients were checked and, to obtain greater reliability of the result in this phase of the study, the analyzes were carried out in technical triplicates”, he said.
After analysis of patient saliva samples, the researchers used bioinformatics and machine learning techniques to arrive at the prognostic signature – checking which proteins or peptides selected in the first phase could separate patients with and without cervical lymph node metastasis.
“In addition, we also had valuable information about the clinical evolution of patients who voluntarily participated in the study, through the donation of saliva samples,” said Paes Leme.
From this result, it was possible to define the signature of three specific peptides of LTA4H, COL6A1 and CSTB, capable of classifying patients with and without metastasis in cervical lymph nodes, with great potential to help clinicians overcome the limitations of exams and guide strategies personalized treatment.
The team of scientists is carrying out new research that aims to act in a translational and accessible way in the construction of biosensors to detect this prognostic signature in patients' saliva. Currently, peptides can be identified and quantified by mass spectrometry and proteomics analysis, techniques that are costly and uncommon in clinics and hospitals.
“We want to develop a simpler, cheaper and more accessible method for healthcare professionals to assess the progression of the disease using tests that can be carried out in the dentist's office, doctor's office or in clinical laboratories. In the work we have just published, it was possible to identify this prognostic signature using mass spectrometry. Our idea is to develop a biosensor focused on using this prognostic signature, so that it has clinical use, and guides the definition of treatment”, said Paes Leme.
The article Combining discovery and targeted proteomics reveals a prognostic signature in oral cancer (doi: 10.1038/s41467-018-05696-2), by Carolina Moretto Carnielli, Carolina Carneiro Soares Macedo, Tatiane De Rossi, Daniela Campos Granato, César Rivera, Romênia Ramos Domingues, Bianca Alves Pauletti, Sami Yokoo, Henry Heberle, Ariane Fidelis Busso-Lopes, Nilva Karla Cervigne, Iris Sawazaki-Calone, Gabriela Vaz Meirelles, Fábio Albuquerque Marchi, Guilherme Pimentel Telles, Rosane Minghim, Ana Carolina Prado Ribeiro, Thaís Bianca Brandão, Gilberto de Castro Jr, Wilfredo Alejandro González-Arriagada, Alexandre Gomes, Fabio Penteado, Alan Roger Santos-Silva, Márcio Ajudarte Lopes, Priscila Campioni Rodrigues, Elias Sundquist, Tuula Salo, Sabrina Daniela da Silva, Moulay A. Alaoui-Jamali, Edgard Graner, Jay W. Fox, Ricardo Della Coletta and Adriana Franco Paes Leme, can be read at www.nature.com/articles/s41467-018-05696-2.