Receptor binding properties, cell tropism and transmission of influenza A virus

Abstract

The first influenza pandemic of the 21st century was caused by the influenza A (H1N1) 2009 virus (A(H1N1)pdm09) that emerged from a swine-origin source. Although human infections with swine-origin influenza have been reported intermittently in the past decade, none went on to cause a pandemic or sustained human-to-human transmission. In previous pandemics, specific residues in the receptor binding site of the haemagglutinin (HA) protein of influenza have been associated with the ability of the virus to transmit between humans. In this thesis the effect of mutations at residue 227 in HA on cell tropism and transmission of A(H1N1)pdm09 is described. In A(H1N1)pdm09 and previous seasonal H1N1 viruses this residue is glutamic acid (E), whereas in swine influenza it is alanine (A). Using human airway epithelium, a differential cell tropism of A(H1N1)pdm09 compared to A(H1N1)pdm09 E227A and swine influenza was shown suggesting this residue may alter the sialic acid conformer binding preference of the HA. Furthermore, multi-cycle viral growth of both A(H1N1)pdm09 E227A and swine influenza was found to be attenuated in comparison to A(H1N1)pdm09 in human airway epithelium. However this altered tropism and viral growth in human airway epithelium did not abrogate respiratory droplet transmission of A(H1N1)pdm09 E227A in ferrets. This suggests that acquisition of 227E was not solely responsible for the ability of A(H1N1)pdm09 to transmit between humans. Because the work with the E227A mutant showed that small differences in cell tropism that may affect influenza virus transmissibility could be detected in human airway cells, a receptor binding assay was developed for laboratory surveillance using commercial human airway epithelium cultures, to screen for zoonotic influenza strains of particular concern for human health. To further investigate adaptations by influenza virus for infection of the human host, the cell tropism determined by the HA protein of an avian H7N7 and human H7N3 isolate was compared. Binding to non-ciliated human airway cells was increased for the human isolate. This human case of H7N3 infection yielded two isolates from different sites (eye and throat) from the same patient on the same day. A whole genome sequencing assay was designed for H7 isolates and both eye and throat isolate were fully sequenced. One synonymous nucleotide change was found in the NS gene segment and one synonymous and two non-synonymous nucleotide changes were found in the PB2 gene segment. Comparison of the non-synonymous changes in the protein sequence of PB2 to available avian and human virus PB2 sequences revealed that the substitutions in the eye isolate were comparatively uncommon. Interestingly, these changes resulted in an increased viral growth in human airway epithelial cells at 32ºC when compared to the throat isolate, a phenomenon which was not observed at 37ºC. Finally, the use of a lung model maintained by the ex-vivo lung perfusion (EVLP) technique for study of virus infection was tested. This technique allows the use of both human and porcine lungs up to 24 hours after abstraction and is a potential model for respiratory pathogens and novel treatments. Porcine lungs were infected with A(H1N1)pdm09. Physiological and virological parameters were measured in two separate experiments and infection was demonstrated by increased viral loads in samples obtained at late time points after infection.