An integrated nationwide genomics study reveals transmission modes of typhoid fever in China

Abstract

Typhoid fever is a life-threatening disease caused by Salmonella enterica serovar Typhi, resulting in a significant disease burden across developing countries. Historically, China was very much close to the global epicenter of typhoid, but the role of typhoid transmission within China and among epicenter remains overlooked in previous investigations. By using newly produced genomics on a national scale, we clarify the complex local and global transmission history of such a notorious disease agent in China spanning the most recent five decades, which largely undermines the global public health network.

Keywords: Salmonella Typhi; exploiting genomics; pandemic clone; transmission; typhoidal fever.

Fig 1

Incidence of typhoid fever in China. (A) Annual incidence (number of cases per 100,000 persons per year) of typhoid fever in China during 2004–2017. (B) The average incidence of typhoid fever in each province of China during 2008–2017. (C) Spearman correlation of climate, social, and economic factors (y-axis) with the incidence of typhoid fever (x-axis). Each dot indicates a province of China. The data for the plots are appended in Table S1.

Fig 2

Distinct genotypic and antimicrobial patterns of S. Typhi between regions. The S. Typhi isolates are grouped into primary clades (A) and subclades (B) by the Genotyphi typing scheme. According to the geographic source, the comparison is made between different parts of China, i.e., eastern (n = 144), northern (n = 38), northwestern (n = 83), southern (n = 459), and between various regions over the world, i.e., China (n = 731), Africa (n = 1,074), Asia (excluding China, n = 1,443), and Europe (n = 1,942). (C) Comparison of carriage rate of antimicrobial-related genetic determinants.

Fig 3

“Long-cycle” transmission of four dominant clades in China. (A) Maximum clade credibility tree (reconstructed using BEAST2) with the antimicrobial-associated metadata. The Chinese S. Typhi isolates from four regions are highlighted by boxes in different colors. The branches comprising international isolates are collapsed into triangles in gray. The right columns indicate the presence of plasmid and gyr mutations as well as the detailed genotypes. (B) Transmission routes and time inferred by BEAST analysis. Two kinds of arrows show transmission events from abroad and between provinces. The numbers near the arrows indicate the estimated year of the putative transmission events. (C) Bayesian skyline plots show the historical changes in the effective population size of the four clades. Lines and shadings indicate the medians and the 95% HPD intervals of estimated effective population sizes. The x-axis shows the time, and the y-axis indicates the population size.

Fig 4

“Short-cycle” (or travel-associated) transmission of S. Typhi. (A) Histogram of cgSNP distance for the Chinese S. Typhi isolates collected from the same province (left y-axis) and different provinces (right y-axis). The x-axis indicates the number of cgSNPs. (B) Putative transmission events across provinces in China. Each line indicates a pair of S. Typhi isolates with pairwise distance <5 cgSNPs, indicating a putative transmission event. (C) Comparison of cgSNP distance between 4.3 isolates derived from different reference genomes as illustrated by a scatter plot and a violin plot, respectively. In the scatter plot, each dot indicates a pair of isolates, with its x-axis indicating the cgSNP distance calculated using the genome of strain CT18 (belonging to 3.2) as the reference and its y-axis using strain ERL052042 (accession number GCF_001088345, belonging to 4.3) as the reference. In the violin plot, the blue and green violins illustrate the distribution of the two sets of distance; and the red indicates their difference. (D) Example from the Chinese 4.3 isolate 458 (collected from Shanghai city) as well as the isolates of its closest genome match. Using CT18 strain as the reference, the left ML tree shows isolate 458 was one cgSNP distant from the international isolates (from the UK and Pakistan). Using ERL052042 as the reference, the right ML tree shows isolate 458 was one to three cgSNPs distant from these international isolates, but most of the international isolates still have the identical cgSNP profile, i.e., they cannot be distinguished from each other by their pairwise cgSNPs. (E) Histogram of cgSNP distance for 4.3 isolates collected from different countries and continents. The x-axis indicates the number of cgSNPs. (F) Genomic relatedness of 4.3 isolates worldwide as revealed by their pairwise cgSNP distance. The heatmaps show the numbers of close cgSNP matches (i.e., <5 cgSNPs) between countries and between continents. Countries with the same shade color belong to the same continents.

Fig 5

In vitro phenotypic comparison between 4.3.1 and other clades. Cellular invasion assays in the U937 cell line (A) and THP-1 cell line (B). The y-axis scale represents the log₁₀ value of the ratio of detected bacteria at 12 h when compared with the initial bacteria that successfully invaded. In vitro tolerance assays with acid (pH = 2) (C) and desiccation (D). The y-axis scale represents the percentage of detected bacteria compared to the initial inoculation. Four groups of clinical isolates (Table S4) were chosen based on their lineage origin and availability in the current collection. The unpaired two-tailed t test was performed to assess the difference between groups with the calculated P value.