Simulated Respiratory Secretion for Use in the Development of Influenza Diagnostic Assays

Abstract

Many assays have been developed for the detection of influenza virus which is an important respiratory pathogen. Development of these assays commonly involves the use of human clinical samples for validation of their performance. However, clinical samples can be difficult to obtain, deteriorate over time, and be inconsistent in composition. The goal of this study was to develop a simulated respiratory secretion (SRS) that could act as a surrogate for clinical samples. To this end, we determined the effects major respiratory secretion components (Na+, K+, Ca2+, cells, albumin IgG, IgM, and mucin) have on the performance of influenza assays including both nucleic acid amplification and rapid antigen assays. Minimal effects on the molecular assays were observed for all of the components tested, except for serum derived human IgG, which suppressed the signal of the rapid antigen assays. Using dot blots we were able to show anti-influenza nucleoprotein IgG antibodies are common in human respiratory samples. We composed a SRS that contained mid-point levels of human respiratory sample components and studied its effect compared to phosphate buffered saline and virus negative clinical sample matrix on the Veritor, Sofia, CDC RT-PCR, Simplexa, cobas Liat, and Alere i influenza assays. Our results demonstrated that a SRS can interact with a variety of test methods in a similar manner to clinical samples with a similar impact on test performance.

Fig 1. The results of testing each component’s effect on the Liat assay.

H1N1pdm was diluted in each of the matrices containing each of the components and tested in triplicate in the Liat assay. The experiment was performed twice with the repeats shown in black and red. Each matrix was compared to PBS using the t-test. Bars indicate the mean of the three replicates for each repeat. Asterisks indicate a significant difference from PBS alone.

Fig 2. Results of testing virus spiked in PBS, SRS, and NCS with six assays.

Ct and antigen assay signal results were normalized as a percentage of the PBS result value. Data for high and low concentrations of virus are shown.

Fig 3

(A) Dot blot for the detection of IgG antibodies specific to the H1N1pdm NP. Membranes 1 to 3 were negative controls, 4 to 6 were human serum IgG, and 7 to 9 were monoclonal IgG. (B-C) Comparison of PBS, serum IgG, monoclonal IgG, and respiratory swabs with the Sofia and Veritor assays. H1N1pdm was diluted 10-fold from stock in PBS, serum IgG, or monoclonal IgG, absorbed on a clean swab, and tested. Also, a set of samples with virus diluted in PBS was absorbed on a freshly collected NPS and tested.