New tumour markers for pancreatic cancer based on the altered glycosylation of serum glycoproteins
- Autores: Adrià Duran Sidera
- Directores de la Tesis: Rosa Peracaula Miró (dir. tes.), Esther Llop Escorihuela (codir. tes.)
- Lectura: En la Universitat de Girona ( España ) en 2023
- Idioma: inglés
- Tribunal Calificador de la Tesis: Cristina Esteva Font (presid.), Pere Boadas Vaello (secret.), Neus Martínez Bosch (voc.)
- Programa de doctorado: Programa de Doctorado en Biología Molecular, Biomedicina y Salud por la Universidad de Girona
- Materias:
- Texto completo no disponible (Saber más ...)
- Resumen
Pancreatic cancer (PaC) is the deadliest of all tumours, with a 5-year survival below 12% and a mortality/incidence ratio of 94.5%. One of the main reasons behind this dismal prognosis is the diagnosis of the disease at late stages, usually when metastasis have already occurred and no effective therapies are available. Nowa- days, there is no accurate tumour marker for the detection of PaC. Only the use of the carbohydrate antigen 19-9 (CA19-9) is clinically validated for the manage- ment and recurrence evaluation of the disease, but lacks specificity to be used as a diagnostic test. For this reason, the discovery of novel biomarkers able to detect PaC in early stages, when current therapies are still effective, is of utmost signifi- cance for the research and medical communities.
Altered glycosylation, a common feature of cancer, stands as a potential source for developing new tumour markers. However, a specific glycan signature to diagnose PaC has not been described. As glycosylation is one of the main post-translational modification of proteins, we hypothesise that the combinatorial analysis of tu- mour-associated glycan structures on overexpressed PaC proteins could outper- form the sensitivity and specificity of current methodologies. In this regard, we have studied the glycosylation pattern of two neo-/over-expressed proteins in PaC, mesothelin (MSLN) and regenerating islet-derived protein 1 (REG1), and we have assessed the potential of their glycoforms as PaC tumour markers.
On one hand, MSLN expression was assessed by western blotting (WB) on PaC cell lines protein lysates and conditioned media. MSLN was observed in the seven studied cell lines, with higher amounts on the secreted media. MSLN N-glycosyl- ation was characterised in those cell lines with higher expression (Capan-2, AsPC- 1 and HPAF-II). PNGaseF digestion, peptide mapping and total glycoprotein mass analysis showed that the three MSLN N-glycan sites were occupied. The structural characterisation of fluorescently labelled MSLN N-glycans by UPLC (in a HILIC column) combined with several exoglycosidase digestions revealed the expression of complex type sialylated structures, most of which were core fucosyl- ated and highly branched. Next, the study of immunopurified MSLN N-glycan determinants using lectins indicated the potential to use the lectins AAL (fucose recognition), PhoSL (core fucose recognition), PHA-E (bisected GlcNAc recogni- tion) and PHA-L (ß1,6-antenna recognition) to detect specific glycan moieties on MSLN.
MSLN expression was also assessed in pancreatic tissues corresponding to tumours or healthy adjacent structures. MSLN expression was detected in 77.4% of tu- mours, while just one non-tumour sample (10%) expressed the protein. Then, we developed a methodology for the quantification of specific MSLN glycoforms in biological samples, as tissue lysates or blood serum. The method consisted in the quantification of MSLN protein levels by an enzyme-linked immunosorbent assay (ELISA) followed by the identification of specific core fucosylated MSLN (Cf- MSLN) glycoforms in an enzyme-linked lectin assay (ELLA) using the PhoSL lec- tin. Application of this methodology in a cohort of patients including PaC, chronic pancreatitis and healthy individuals showed a significant decrease in Cf-MSLN abundancy in PaC samples. Moreover, combination of this biomarker with CA19- 9 enhanced its potential to be used as a diagnostic tumour marker, yielding a sen- sitivity and specificity of 93.3% and 81.2%, respectively.
On the other hand, REG1 expression was also assessed by WB on PaC cell lines and pancreatic tissues. While no expression was observed in cultured PaC cell lines, 90.9% of non-tumour tissues and 60.5-65.8% of PaCs expressed REG1. The analysis of REG1 O-glycosylation through WB with lectins on immunopurified REG1 showed that, while the lectin PNA (galactose recognition) could only detect control samples, the lectins SNA (2,6-sialic acid recognition) and VVL (GalNAc recognition) were only reactive in tumour samples. However, the development of an ELLA assay using these lectins as detection probes to quantify REG1 gly- coforms was not possible due to the lack of analytical sensitivity. To overcome this limitation, probably caused by the use of lectins, we performed an in-depth anal- ysis of REG1 glycoforms by mass spectrometry (MS). MALDI-Tof evaluation of the total glycoprotein and its O-glycopeptide obtained after trypsin digestion showed up to 16 different structures, which included the Tn antigen, core 1, core 2 and core 4 O-glycans. Next, we developed a LC-MS application consisting of a hydrophilic interaction liquid chromatography (HILIC) separation of glycopep- tides followed by their detection in a triple quadrupole under a selected reaction monitoring (SRM) data acquisition to detect the most relevant glycoforms in com- plex biological samples. This workflow has already been successful in analysing a control and a PaC serum sample, and has shown the increased expression of sialyl- T and sialyl-Tn antigens on REG1A in malignant conditions. However, these are still preliminary results, and validation in a cohort of patients is required.
To sum up, we have been able to perform a comprehensive analysis of MSLN and REG1 glycosylation in PaC. In addition, we have developed new strategies to quantify their tumour-associated glycoforms. While the reduction of Cf-MSLN in PaC patients holds a promising potential to detect PaC, the quantification of spe- cific REG1 glycoforms still needs to be tested in more serum samples. Overall, this work has provided candidate MSLN and REG1 glycoforms to be used as novel PaC biomarkers, which will require further validation in large cohorts of patients.
