Antibody isotype analysis of malaria-nematode co-infection: problems and solutions associated with cross-reactivity

Background: Antibody isotype responses can be useful as indicators of immune bias during infection. In studies of parasite co-infection however, interpretation of immune bias is complicated by the occurrence of cross-reactive antibodies. To confidently attribute shifts in immune bias to the presence of a co-infecting parasite, we suggest practical approaches to account for antibody cross reactivity. The potential for cross-reactive antibodies to influence disease outcome is also discussed. Results: Utilising two murine models of malaria-helminth co-infection we analysed antibody responses of mice singly- or co-infected with Plasmodium chabaudi chabaudi and Nippostrongylus brasiliensis or Litomosoides sigmodontis. We observed cross-reactive antibody responses that recognised antigens from both pathogens irrespective of whether crude parasite antigen preparations or purified recombinant proteins were used in ELISA. These responses were not apparent in control mice. The relative strength of cross-reactive versus antigen-specific responses was determined by calculating antibody titre. In addition, we analysed antibody binding to periodate-treated antigens, to distinguish responses targeted to protein versus carbohydrate moieties. Periodate treatment affected both antigen-specific and cross-reactive responses. For example, malaria-induced cross-reactive IgG1 responses were found to target the carbohydrate component of the helminth antigen, as they were not detected following periodate treatment. Interestingly, periodate treatment of recombinant malaria antigen Merozoite Surface Protein-1(19) (MSP-1(19)) resulted in increased detection of antigen-specific IgG2a responses in malaria-infected mice. This suggests that glycosylation may have been masking protein epitopes and that periodate-treated MSP-1(19) may more closely reflect the natural non-glycosylated antigen seen during infection. Conclusions: In order to utilize antibody isotypes as a measure of immune bias during co-infection studies, it is important to dissect antigen-specific from cross-reactive antibody responses. Calculating antibody titre, rather than using a single dilution of serum, as a measure of the relative strength of the response, largely accomplished this. Elimination of the carbohydrate moiety of an antigen that can often be the target of cross-reactive antibodies also proved useful.