Clinical application of a multiplex genetic pathogen detection system remaps the aetiology of diarrhoeal infections in Shanghai

Abstract

Background: Culture-based diagnostic methods cannot achieve rapid and precise diagnoses for the identification of multiple diarrhoeal pathogens (DPs). A high-throughput multiplex genetic detection system (HMGS) was adapted and evaluated for the simultaneous identification and differentiation of infectious DPs and a broad analysis of DP infection aetiology.

Results: DP-HMGS was highly sensitive and specific for DP detection compared with culture-based techniques and was similar to singleplex real-time PCR. The uniform level of sensitivity of DP-HMGS for all DPs allowed us to remap the aetiology of acute diarrhoeal infections in Shanghai, correcting incidences of massively underdiagnosed DP species with accuracy approaching that of sequencing-based methods. The most frequent DPs were enteropathogenic Escherichia coli, rotavirus and Campylobacter jejuni. DP-HMGS detected two additional causes of infectious diarrhoea that were previously missed by routine culture-based methods: enterohemorrhagic E. coli and Yersinia enterocolitica. We demonstrated the age dependence of specific DP distributions, especially the distributions of rotavirus, intestinal adenovirus and Clostridium difficile in paediatric patients as well as those of dominant bacterial infections in adults, with a distinct "top 3" pattern for each age group. Finally, the multiplexing capability and high sensitivity of DP-HMGS allowed the detection of infections co-induced by multiple pathogens (approximately 1/3 of the cases), with some DPs preferentially co-occurring as infectious agents.

Conclusions: DP-HMGS has been shown to be a rapid, specific, sensitive and appropriate method for the simultaneous screening/detection of polymicrobial DP infections in faecal specimens. Widespread use of DP-HMGS is likely to advance routine diagnostic and clinical studies on the aetiology of acute diarrhoea.

Keywords: DP composition; Diarrhoeal pathogens (DPs); Faecal specimens; High-throughput multiplex genetic detection system (HMGS); Polymicrobial infection; Rapid screening.

Fig. 1

DP-HMGS was optimized for robust detection of DPs in a multiplex reaction. a Before optimization of the DP-HMGS assay, eighteen targets (from left to right: hum_RNA, S. typhimurium, S. enteritidis, EIEC, Shigella, HADV, EHEC, HASV, ETEC, Vibrio, C. difficile, C. jejuni, norovirus, hum_DNA, EPEC, EAEC, rotavirus, IC) exhibited relatively large differences in signal intensity. Notably, the specific peaks for Y. enterocolitica and E. coli O157 were not detected by the DP-HMGS assay prior to optimization. b After optimization, all 20 targets (from left to right: hum_RNA, S. typhimurium, S. enteritidis, EIEC, Shigella, HADV, EHEC, HASV, Y. enterocolitica, ETEC, Vibrio, C. difficile, C. jejuni, E. coli O157, norovirus, hum_DNA, EPEC, EAEC, rotavirus, IC) reached intermediate signal levels. Notably, the specific peaks for Y. enterocolitica and E. coli O157 were now visible at 165 bp and 218 bp

Fig. 2

Optimized DP-HMGS maintained high sensitivity for the simultaneous detection of all specific viral and bacterial pathogens. The detection limits of the DP-HMGS assay were determined by amplifying tenfold diluted plasmids containing equal amounts of 17 pathogen and 3 quality control templates at a 1 × 10⁵, b 1 × 10⁴, and c 1 × 10³ copies/μL in the DP-HMGS assay. d ddH₂O was used as the negative control. Note that the pathogen-defining DNA targets (from left to right: S. typhimurium, S. enteritidis, EIEC, Shigella, HADV, EHEC, HASV, Y. enterocolitica, ETEC, Vibrio, C. difficile, C. jejuni, E. coli O157, norovirus, EPEC, EAEC, rotavirus) all generated specific peaks. The quality controls Hum_RNA, Hum_DNA and IC produced specific peaks at 106 bp, 233 bp and 313 bp, respectively

Fig. 3

DP-HMGS robustly detected multiple DP signatures without interference between multiple signatures. a The plasmids of S. enteritidis, HADV and EHEC combined at different proportions (1 × 10³ copies, 1 × 10⁵ copies, and 1 × 10⁴ copies, respectively) produced specific peaks at 120 bp, 145 bp and 152 bp with low, high and midrange signal intensities (2600 rfu, 20,000 rfu and 6000 rfu, respectively). b The combined plasmids of three DPs with 7 negative control pathogen templates consistently showed the specific peaks of S. enteritidis, HADV and EHEC with no interference. Notably, the 313-bp peak for the IC was generated in each DP-HMGS reaction. The individually tested S. enteritidis (c), HADV (d), and EHEC (e) produced specific peaks at 120 bp, 145 bp, and 152 bp, respectively, with corresponding intensities of 2600 rfu, 20,000 rfu and 6000 rfu. Notably, the 313-bp peak for the IC was generated in each DP-HMGS reaction

Fig. 4

Workflow for faecal specimen processing. A schematic workflow chart depicting the processing of 613 faecal specimens for analysis, comparing conventional methods and DP-HMGS

Fig. 5

The uniformly high sensitivity of DP-HMGS corrected previous conclusions regarding the major causes of infectious diarrhoeas. a Comparisons of culture-based methods versus DP-HMGS with regard to sequencing-based detection of 13 diarrhoeal bacterial pathogens were conducted. The sensitivity of DP-HMGS for the detection of Vibrio (P = 0.000), S. typhimurium (P = 0.000), Shigella (P = 0.005), C. difficile (P = 0.000), C. jejuni (P = 0.000), EPEC (P = 0.000), ETEC (P = 0.000), EIEC (P = 0.003) and E. coli O157 (P = 0.008) was significantly lower than that of culture-based detection. Notably, the sensitivity of DP-HMGS was at least twice that of culture-based methods for most bacterial species and as virtually identical to that of sequencing. b Uniform sensitivity of DP-HMGS in detecting both bacterial and viral DPs corrected the ratios of bacterial versus viral causes of diarrhoeas, demonstrating the overwhelming importance of bacterial infections. c The epidemiological contribution of individual DP species was ranked based on the frequency of positive detection rates observed using conventional methods (top) and DP-HMGS (bottom). Notably, DP-HMGS corrected the epidemiological outcomes attained by combining the results of conventional methods, which were profoundly skewed by the low and highly variable sensitivity of culture-based methods in detecting bacterial DPs

Fig. 6

The DP frequency distribution exhibited significant age variation. a The positive detection rate and the trends in the distribution of total bacteria and total viruses. b The positive detection rate and the trends in the distribution of C. difficile, rotavirus and adenovirus. c The positive detection rate and the trends in the distribution of Vibrio, S. typhimurium, EPEC and ETEC. d The frequency of the predominant DPs in different age groups. Note that the P-values of the DPs selected in b were all less than 0.05. The Y-axis represents the positive DP detection rates, while the X-axis represents the different age groups

Fig. 7

DP-HMGS application revealed an unexpectedly high incidence of multifactorial diarrhoea. a Infection profiles of 17 DPs; the different colours represent single infection, double infection, triple infection, quadruple infection and quintuple infection. b The positive detection rates of single and polymicrobial infections of 17 DPs are shown. c Analysis of the correlation between the positive detection rates of single and polymicrobial infections and age groups for 17 DPs. Darker and lighter colours in the histogram indicate single and co-pathogen detection, respectively