To the Editor: With the extensive uptake of pertussis vaccine, morbidity and mortality related to pertussis have decreased significantly. The application of acellular vaccines (ACVs) occurred later in China than in developed countries; correspondingly, the pertussis resurgence in China also began later. The latest global genotype dynamic prediction model cannot explain the B. pertussis genotype change trend in China. Chinese B. pertussis strains gradually formed their own unique developmental branches. Here, a phylogenetic tree of prevailing B. pertussis strains from three different regions of China was visualized by EvolView (http://www.evolgenius.info/evolview.html). A total of 199 isolated B. pertussis strains were used for the phylogenetic analysis. The B. para-pertussis strain isolated from Tianjin was used as the out-group, and Chinese strain (CS), which is widely used as a vaccine strain for ACV production in China, was used as the reference strain. In particular, we analyzed the multilocus sequence typing (MLST) and macrolide-resistance-conferring gene mutations.
This study was approved by the medical ethics committee of Tianjin Second People's Hospital (No.  24). All enrolled patients provided signed informed consent. From 2018 to 2019, a total of 375 suspected pertussis patients in Tianjin Second People's Hospital provided samples for bacterial culture; all these patients either did not use macrolides or used them for <2 days. Strains were isolated mainly from nasopharyngeal swabs or nasopharyngeal aspirates. All isolates were cultured on charcoal agar (Oxoid Ltd., Altrincham, UK) supplemented with 10% sheep blood and 40 μg/mL cephalexin (Bordetella Selective Supplement; Oxoid Ltd., Altrincham, UK) for 3 days at 37 °C. Finally, 100 bacterial strains from 100 patients were successfully cultured and identified as B. pertussis by mass spectrometry. Genomic deoxyribonucleic acid (DNA) was isolated using the Promega Wizard® Genomic DNA Purification Kit (Promega, Madison, WI, USA) following the manufacturer's instructions. An Illumina HiSeqX sequencing platform was used for genomic sequencing. Reference-assisted chromosome assembly was made with Chromosomer v0.1.3 (https://pypi.org/project/chromosomer/) from the reference GCF_000212975.1 (CS, accession number NC_017223). Forty-nine B. pertussis strains from Xi’an and 50 from Shenzhen were downloaded from BioProject of PRJNA489102 and CNGB Sequence Archive (CNSA) of China National GeneBank DataBase with Accession Number CNP0001528. Variations from CS strain in the B. pertussis strains from the three regions were detected by using NucDiff v2.0.2 (https://pypi.org/project/NucDiff/).
The host epidemiological characteristics of the prevailing B. pertussis strains from Tianjin were analyzed. The cohort comprised 88 children (aged <18 years) and 12 adults (aged ≥18 years), of whom 46 were female and 54 were male. Most infections occurred in patients belonging to the 3 to 6-month-old (22% [22/100]) and 7 to 24-month-old (39% [39/100]) age groups. The enrolled participants were divided into three groups according to their ACV vaccination frequency: group I, comprising patients who completed all four doses of pertussis immunization (n = 12 [12%]); group II, comprising patients who were partially immunized with ACVs (n = 45 [45%]); and group III, comprising patients who had not been immunized with ACVs in the past 10 years (n = 43 [43%]).
A total of 13 B. pertussis strain MLST types, an amount significantly higher than any previously reported in China, were obtained in Tianjin. In order from most to least common, the MLST types were: fhaB3/fim2-1/fim3-1/prn1/ptxA1/ptxP1 (48%), fhaB3/fim2-1/fim3-2/prn1/ptxA1/ptxP1 (20%), fhaB3/fim2-1/fim3-4/prn1/ptxA1/ptxP1 (15%), fhaB3/fim2-1/fim3-2/prn4/ptxA1/ptxP3 (6%), fhaB3/fim2-1/fim3-4/prn4/ptxA1/ptxP3 (2%), fhaB3/fim2-1/fim3-1/prn4/ptxA1/ptxP3 (2%), and fhaB1/fim2-1/fim3-4/prn1/ptxA1/ptxP1, fhaB2/fim2-1/fim3-1/prn1/ptxA1/ptxP1, fhaB2/fim2-1/fim3-2/prn15/ptxA1/ptxP3, fhaB3/fim2-1/fim3-1/prn2/ptxA1/ptxP3, fhaB3/fim2-1/fim3-2/prn1/ptxA1/ptxP3, fhaB3/fim2-1/fim3-4/prn1/ptxA1/ptxP3, and fhaB3/fim2-1/fim3-2/prn11/ptxA1/ptxP3 (all 1%). Furthermore, using Cramer V.0.5 (https://www.statology.org/cramers-v-in-r/), we found that there was no significant difference in MLST type between different patient sexes (V value: 0.36) or different vaccination frequencies (V value: 0.30).
Macrolide antibiotics (e.g., erythromycin, azithromycin, and clindamycin) are effective drugs against pertussis. Macrolide-resistance-conferring mutations were concentrated in the 23s rRNA (rrn) gene A2047G; there were 84 (84%) A2047G stains. Mutations in these mutant strains included gene mutations 2361/BPTD_RS11675 (81/100), 3130/BPTD_Rs15455 (1/100), and 2071/BPTD_Rs1025 (2/100). Strains with the 2361/BPTD_RS11675 gene mutation were distributed in the ptxP1 (75/85) and ptxP3 (6/15) lineages. Strains with the 3130/BPTD_Rs15455 and 2071/BPTD_Rs1025 gene mutations were distributed in the ptxP1 lineage. Thus, the A2047G strains exist not only in the ptxP1 lineage but also in the ptxP3 lineage.
The evolutionary history of B. pertussis isolates was inferred by using RAxML v8.2.12 (https://evomics.org/learning/phylogenetics/raxml/), with the parameters 862 singel nucleotide polymorphism (SNP) and >10 samples, to select strains for phylogenic analysis. Consequently, 199 isolated B. pertussis strains were used for phylogenetic analysis, including 100 strains from Tianjin, 49 from Xi’an, and 50 from Shenzhen, all reported after ACV uptake. The resulting phylogenetic tree was visualized by EvolView. The sequences of B. pertussis isolates from Tianjin are relatively evenly distributed throughout the development tree, and their clustering is not as obvious as those of the isolates from Xi’an or Shenzhen [Figure 1]. Moreover, the correlation between strains from Tianjin is not as close as that of those from Shenzhen in 2018. Most of the sequences of B. pertussis strains from Xi’an are distributed at both ends of the developmental tree and can be divided into five clusters. The largest of these clusters is closer to the CS strain than are the clusters of strains from the other two regions. The sequences of strains from Shenzhen can be divided into three clusters, one of which is large and dense and has little homology with any other branches. This cluster is located on its own branches at the root of the developing tree, forming a clear group, with obvious correlation between different strains in this cluster. On the development tree, they are separated from the strains from the other two regions and are far away from the CS strain. The SNP analysis revealed 864 identical mutations compared with the CS strain in the isolates from the three regions. Overall, the sequences from Xi’an, Tianjin, and Shenzhen contained 2831, 1857, and 1754 mutations, respectively. Sequences from Tianjin had the lowest number (123) of unique mutations among the three regions.
Our study has some limitations. First, it is a comparison of strains in only three regions of China; although these three regions are all big cities with populations of tens of millions and high population density, there may still be certain deviations. Second, there is no expanded study on the relationship between ptxP3 and A2047G.
In conclusion, the evolution of B. pertussis strains after ACV uptake differed among three regions of China, and the mismatch between the prevailing strains in different regions and the CS strain may explain the current pertussis resurgence. The identical mutations observed in the prevailing B. pertussis strains from these three regions may be the result of positive selection; these mutations may provide direction for improving the pertussis vaccine in China. With the onslaught of global communication, it is uncertain which strains will be the dominant strains thereafter. More research on the prevailing B. pertussis strains in regions of China with different economic levels should be conducted.
We thank Katie Oakley, Ph.D., from Liwen Bianji (Edanz) (www.liwenbianji.cn) for editing the English text of a draft of this manuscript.
This work was supported by Tianjin Key Medical Discipline (Specialty) Construction Project (No. TJYXZDXK059B).
Conflicts of interest
1. Lefrancq N, Bouchez V, Fernandes N, Barkoff AM, Bosch T, Dalby T, et al. Global spatial dynamics and vaccine-induced fitness changes of Bordetella pertussis
. Sci Transl Med 2022;14:eabn3253. doi: 10.1126/scitranslmed.abn3253.
2. Yao K, Deng J, Ma X, Dai W, Chen Q, Zhou K, et al. The epidemic of erythromycin-resistant Bordetella pertussis with limited genome variation associated with pertussis resurgence in China. Expert Rev Vaccines 2020;19:1093–1099. doi: 10.1080/14760584.2020.1831916.
3. Zhang S, Xu Y, Zhou Z, Wang S, Yang R, Wang J, et al. Complete genome sequence of Bordetella pertussis CS, a Chinese pertussis vaccine strain. J Bacteriol 2011;193:4017–4018. doi: 10.1128/JB.05184-11.
4. Xu Z, Wang Z, Luan Y, Li Y, Liu X, Peng X, et al. Genomic epidemiology of erythromycin-resistant Bordetella pertussis in China. Emerg Microbes Infect 2019;8:461–470. doi: 10.1080/22221751.2019.1587315.
5. Wu S, Hu Q, Yang C, Zhou H, Chen H, Zhang Y, et al. Molecular epidemiology of Bordetella pertussis and analysis of vaccine antigen genes from clinical isolates from Shenzhen, China. Ann Clin Microbiol Antimicrob 2021;20:53. doi: 10.1186/s12941-021-00458-3.