Comparison of SEC and AF4 analytical tools for size estimation of typhoid Vi polysaccharides

Vi capsular polysaccharides (Vi-CPSs) and Vi conjugates are used as antigens in vaccines to protect against Salmonella typhi. Their physico-chemical properties and immunogenicity have to be controlled to ensure vaccine efficiency. Polysaccharide molecular weight is one of the critical characteristics that must be monitored during the manufacturing process to ensure product quality. The common analytical technique used for this purpose is size exclusion chromatography (SEC) although the polysaccharide sometimes interacts with the column phase. In this study, we propose an alternative technique to SEC for the characterization of the Vi polysaccharide size and molecular weight: asymmetrical flow field-flow fractionation (AF4). We compared both techniques by analyzing five solutions of Vi-CPSs injected five times each. One study was performed on the native Vi-CPS and the other on the depolymerized product. Significant differences were found by comparing the results obtained using both techniques, especially when analyzing native Vi-CPSs. Statistical processing of the data enabled us to highlight the causes of the differences in values and therefore to better identify the limitations and the advantages of each analytical tool. It was demonstrated that in SEC, results were based on a non-representative part of the sample for native Vi-CPS analysis, due to its interaction with the stationary phase. This was not the case for the analysis of the depolymerized Vi-CPS as the results were comparable between the techniques. Sample preparation also appears to be a very important factor. Statistical analyses showed that the important variability of the AF4 results in native Vi-CPS analysis was due to the series effect and not the repeatability effect. Again, for the analysis of the depolymerized product, no such difference was observed. Consequently, depending on the nature of the analyte, AF4 could be a more relevant technique for size estimation.