A Multiplex Label-Free Approach to Avian Influenza Surveillance and Serology

Abstract

Influenza serology has traditionally relied on techniques such as hemagglutination inhibition, microneutralization, and ELISA. These assays are complex, challenging to implement in a format allowing detection of several types of antibody-analyte interactions at once (multiplex), and troublesome to implement in the field. As an alternative, we have developed a hemagglutinin microarray on the Arrayed Imaging Reflectometry (AIR) platform. AIR provides sensitive, rapid, and label-free multiplex detection of targets in complex analyte samples such as serum. In preliminary work, we demonstrated the application of this array to the testing of human samples from a vaccine trial. Here, we report the application of an expanded label-free hemagglutinin microarray to the analysis of avian serum samples. Samples from influenza virus challenge experiments in mallards yielded strong, selective detection of antibodies to the challenge antigen in most cases. Samples acquired in the field from mallards were also analyzed, and compared with viral hemagglutinin inhibition and microneutralization assays. We find that the AIR hemagglutinin microarray can provide a simple and robust alternative to standard methods, offering substantially greater information density from a simple workflow.

Fig 1. Strong responses to polyclonal anti-HA antiserum are readily observable on an AIR hemagglutinin microarray.

(a) 1% BSA control. (b) Anti-H7 polyclonal antiserum (A/Netherlands/219/2003, H7N7), 1:80 dilution (1.3%) in 1% BSA. Spots showing substantially increased brightness indicate binding to immobilized H7. In both cases, antigens were arrayed in square patterns as indicated by the yellow boxes in (a); a mouse IgG Fc domain was included as negative control (red boxes). Slight differences in spot intensity in the control (a) are due to differences in deposition efficiency for different antigens or controls. Specific antigens used in these experiments are indicated in Table 2.

Fig 2. Representative array response to a dilution series of H7 polyclonal antiserum.

Cross-reactivity (non-H3 response) is shown only at the highest concentration tested, but was negligible throughout. Error on each point represents the square root of the sum of squares between the analyte and control group standard deviations for the HA isoforms; N = 1, n = 10.

Fig 3. Titration plots for selected challenge samples.

Samples 1502 and 1473 demonstrated highly specific and robust responses to the challenge antigen on the microarray, while samples 1468 and 1493 produced weaker and less specific responses. Error on each point represents the square root of the sum of squares between the analyte and control group standard deviations for the HA isoforms; N = 1, n = 10.

Fig 4. Comparison of HA microarray and selected microneutralization (MN) and virus isolation (VI) results for field samples (mallard).

a) AIR microarray data. Average differences in spot thickness (normalized relative to the strongest response for each antigen) for chips treated with 18% field serum are reported relative to control. Color coding (green = low to red = high) is scaled relative to this maximum response. b) MN results (numbers indicate antibody titer); two samples highlighted in yellow tested positive by VI. Empty cells indicate an antibody titer of < 20.