Statistical approaches to developing a multiplex immunoassay for determining human exposure to environmental pathogens

• We developed and optimized an indirect capture 7-plex immunoassay using DOE.
• Animal-derived antibodies were used to confirm antigen coupling to the beads.
• FMM was used to establish cutoffs to distinguish exposed from non-exposed samples.
• A mean SNR ratio of 4.4 was observed for positive versus negative human samples.

There are numerous pathogens that can be transmitted through water. Identifying and understanding the routes and magnitude of exposure or infection to these microbial contaminants are critical to assessing and mitigating risk. Conventional approaches of studying immunological responses to exposure or infection such as Enzyme-Linked Immunosorbent Assays (ELISAs) and other monoplex antibody-based immunoassays can be very costly, laborious, and consume large quantities of patient sample. A major limitation of these approaches is that they can only be used to measure one analyte at a time. Multiplex immunoassays provide the ability to study multiple pathogens simultaneously in microliter volumes of samples. However, there are several challenges that must be addressed when developing these multiplex immunoassays such as selection of specific antigens and antibodies, cross-reactivity, calibration, protein-reagent interferences, and the need for rigorous optimization of protein concentrations. In this study, a Design of Experiments (DOE) approach was used to optimize reagent concentrations for coupling selected antigens to Luminex™ xMAP microspheres for use in an indirect capture, multiplex immunoassay to detect human exposure or infection from pathogens that are potentially transmitted through water. Results from Helicobacter pylori, Campylobacter jejuni, Escherichia coli O157:H7, and Salmonella typhimurium singleplexes were used to determine the mean concentrations that would be applied to the multiplex assay. Cut-offs to differentiate between exposed and non-exposed individuals were determined using finite mixed modeling (FMM). The statistical approaches developed facilitated the detection of Immunoglobulin G (IgG) antibodies to H. pylori, C. jejuni, Toxoplasma gondii, hepatitis A virus, rotavirus and noroviruses (VA387 and Norwalk strains) in fifty-four diagnostically characterized plasma samples. Of the characterized samples, the detection rate was 87.5% for H. pylori, and 100% for T. gondii assays and 89% for HAV. Further, the optimized multiplex assay revealed exposure/infection to several other environmental pathogens previously uncharacterized in the samples.