Multiple Sources of Genetic Diversity of Influenza A Viruses during the Hajj

Abstract

Outbreaks of respiratory virus infection at mass gatherings pose significant health risks to attendees, host communities, and ultimately the global population if they help facilitate viral emergence. However, little is known about the genetic diversity, evolution, and patterns of viral transmission during mass gatherings, particularly how much diversity is generated by in situ transmission compared to that imported from other locations. Here, we describe the genome-scale evolution of influenza A viruses sampled from the Hajj pilgrimages at Makkah during 2013 to 2015. Phylogenetic analysis revealed that the diversity of influenza viruses at the Hajj pilgrimages was shaped by multiple introduction events, comprising multiple cocirculating lineages in each year, including those that have circulated in the Middle East and those whose origins likely lie on different continents. At the scale of individual hosts, the majority of minor variants resulted from de novo mutation, with only limited evidence of minor variant transmission or minor variants circulating at subconsensus level despite the likely identification of multiple transmission clusters. Together, these data highlight the complexity of influenza virus infection at the Hajj pilgrimages, reflecting a mix of global genetic diversity drawn from multiple sources combined with local transmission, and reemphasize the need for vigilant surveillance at mass gatherings.IMPORTANCE Large population sizes and densities at mass gatherings such as the Hajj (Makkah, Saudi Arabia) can contribute to outbreaks of respiratory virus infection by providing local hot spots for transmission followed by spread to other localities. Using a genome-scale analysis, we show that the genetic diversity of influenza A viruses at the Hajj gatherings during 2013 to 2015 was largely shaped by the introduction of multiple viruses from diverse geographic regions, including the Middle East, with only little evidence of interhost virus transmission at the Hajj and seemingly limited spread of subconsensus mutational variants. The diversity of viruses at the Hajj pilgrimages highlights the potential for lineage cocirculation during mass gatherings, in turn fuelling segment reassortment and the emergence of novel variants, such that the continued surveillance of respiratory pathogens at mass gatherings should be a public health priority.

Keywords: Hajj; epidemiology; evolution; influenza; phylogeny; transmission.

FIG 1

Phylogenetic trees of concatenated coding regions for A(H1N1) and A(H3N2) viruses. Publicly available vaccine strains were included as outgroups and are shown in green. For the viruses sampled at the Hajj pilgrimages, AU and SA denote Australia and Saudi Arabia (pilgrim's country of origin), respectively, the first digit indicates the sequence number, followed by the year of sampling (2013 to 2015), and the virus subtype is given in parentheses (H1 or H3). The A(H1N1) (A) and A(H3N2) (B) phylogenies were rooted using the A/California/04/2009 and A/Perth/16/2009 sequences, respectively. Statistical support for individual nodes was estimated from 1,000 bootstrap replicates, and values are represented as percentages on the nodes (values of >70% are shown). Scale bars are proportional to the number of nucleotide substitutions per site.

FIG 2

Phylogenetic relationships of the PB2, PB1, PA, HA, NP, NA, M, and NS gene segments of A(H3N2) viruses, as indicated. A virus with a likely reassortant history, SA/1/2013(H3), is highlighted in orange, and publicly available vaccine strains were included as outgroups and are shown in green. Phylogenies were rooted using the A/Perth/16/2009 sequence. Statistical support for individual nodes was estimated from 1,000 bootstrap replicates, and values are represented as percentages on the nodes (values of >70% are shown). Scale bars are proportional to the number of nucleotide substitutions per site.

FIG 3

Phylogenetic trees of HA coding sequences for A(H1N1) and A(H3N2) viruses sampled globally from 2013 to 2015. All full-length HA coding gene sequences were downloaded from GenBank and GISAID, with duplicate sequences removed. The A(H1N1) (A) and A(H3N2) (B) phylogenies were rooted using the A/California/04/2009 and A/Perth/16/2009 sequences, respectively. Seasonal vaccine strains are labeled in gray, and antigenic groups are annotated. Branch colors represent the years of collection, and the Hajj samples are shown in black. Black dotted lines highlight the phylogenetic positions of samples collected from the Hajj pilgrimages, while proposed introduction events are labeled with Roman numerals (i to xii). Scale bars are proportional to the number of nucleotide substitutions per site. Bootstrap support values of >50% are shown for relevant nodes.

FIG 4

Phylogenetic trees of the HA coding genes for A(H1N1) and A(H3N2) viruses circulating in Saudi Arabia and neighboring countries. Introductions are labeled as described in the legend of Fig. 2. The A(H1N1) (A) and A(H3N2) (B) phylogenies were rooted using A/California/04/2009 and A/Perth/16/2009 sequences, respectively. Seasonal vaccine strains are labeled in gray, and antigenic groups are annotated. Branch colors reflect the years of sample collection, and black represents viruses collected in this study. Black dotted lines highlight the phylogenetic positions of samples collected from the Hajj pilgrimages, while proposed introduction events are labeled with Roman numerals (i to xii). Scale bars are proportional to the number of nucleotide substitutions per site. Bootstrap support values of >60% are shown for relevant nodes.

FIG 5

Diagrammatic representation of the genetic differences between potential transmission groups and the distribution of minor variants. Each panel represents a potential transmission cluster. Boxes represent each of the coding genes, and the numbers below the boxes refer to the nucleotide positions in the coding genes. The y axis displays the nucleotide frequencies, with ticks corresponding to 0, 50, and 100% variant frequency. Nucleotides present are identified according to the following color scheme: green, A; red, U; blue, C; black, G. Open symbols indicate nucleotide fixation events between samples, and closed symbols represent positions of minor variants.