Prevalence of Non-Polio Enteroviruses in the Sewage of Guangzhou City, China, from 2013 to 2021

ABSTRACT Continuous surveillance of enteroviruses (EVs) in urban domestic sewage can timely reflect the circulation of EVs in the environment and crowds, and play a predictive and early warning role in EV-related diseases. To better understand the long-term epidemiological trends of circulating EVs and EV-related diseases, we conducted a 9-year (2013 to 2021) surveillance study of non-polio EVs (NPEVs) in urban sewage in Guangzhou city, China. After concentrating and isolating the viruses from the sewage samples, NPEVs were detected and molecular typing was performed. Twenty-one different NPEV serotypes were identified. The most isolated EVs were echovirus 11 (E11), followed by coxsackievirus (CV) B5, E6, and CVB3. EV species B prevailed in sewage samples, but variations in the annual frequency of different serotypes were also observed in different seasons, due to spatial and temporal factors. E11 and E6 were detected continuously before 2017, and the number of isolates was relatively stable during the surveillance period. However, after their explosive growth in 2018 and 2019, their numbers suddenly decreased significantly. CVB3 and CVB5 had alternating trends; CVB5 was most frequently detected in 2013 to 2014 and 2017 to 2018, while CVB3 was most frequently detected in 2015 to 2016 and 2020 to 2021. Phylogenetic analysis showed that at least two different transmission chains of CVB3 and CVB5 were prevalent in Guangzhou City. Our results show that in the absence of a comprehensive and systematic EV-related disease surveillance system in China, environmental surveillance is a powerful and effective tool to strengthen and further investigate the invisible transmission of EVs in the population. IMPORTANCE This study surveilled urban sewage samples from north China for 9 years to monitor enteroviruses. Samples were collected, processed, and viral identification and molecular typing were performed. We detected 21 different non-polio enteroviruses (NPEVs) with yearly variations in prevalence and peak seasons. In addition, this study is very important for understanding the epidemiology of EVs during the COVID-19 pandemic, as the detection frequency and serotypes of EVs in sewage changed considerably around 2020. We believe that our study makes a significant contribution to the literature because our results strongly suggest that environmental surveillance is an exceptionally important tool, which can be employed to detect and monitor organisms of public health concern, which would otherwise be missed and under-reported by case-based surveillance systems alone.

Seasonal distribution characteristics. Most EV-positive samples were obtained in summer months (Fig. 2). Overall, the number of EV-positive samples that were typed typically increased from May and peaked in June and July; however, in 2014, 2015, 2020, and 2021, the peak months were April (n = 38), December (n = 22), November (n = 6), and May (n = 2), respectively. Some variations in the seasonal prevalence of NPEVs were observed. As indicated in Fig. 2, the peaks in the number of isolates in 2014 and 2021 were brought forward to spring, while 2015, 2016, 2017, and 2018 had two peaks in summer and winter. Additionally, the number of NPEVs also increased in winter in 2020 and peaked in December. In 2018 and 2019, numerous NPEVs were isolated from May to July (90 and 106 isolates in total), and numerous (65 isolates in total) NPEVs were isolated from October to   December 2018. However, a few NPEVs were detected in 2020 (only 24 isolates), possibly due to the prevalence of COVID-19. Our results differ from those of other studies (18,19), in which the number of NPEVs in Guangzhou sewage samples did not all remain high during the summer and autumn. Epidemic characteristics of predominant EV serotypes. Notably, the distribution pattern of NPEV serotypes and the predominant serotypes also changed over the years. E6, E11, CVB5, and CVB3 were the most common serotypes during the study period (Table 1; Fig. 3), and E6 was isolated annually for 13 years. It reached the peak of strain isolation in 2010 (101 isolates) and was the most common serotype in 2010, 2013, and 2015 ( Fig. 3), and then gradually declined since 2019 (Fig. 4A). E11 was generally detected before 2017, and the number of isolates was relatively stable, with a peak in summer and autumn. However, after its explosive growth in 2018 and 2019 (accounting for 34.56% and 72.83% of the total number of annual isolates, respectively), E11 abruptly disappeared. It was not detected in 2020, and only one E11 strain was detected in 2021 (Fig. 4B). CVB5 was rarely detected in 2011 and 2012, but the number of CVB5 detected increased suddenly in 2014 and 2018 (accounting for 52.21% and 39.63% of the total number of isolates in the whole year, respectively) and then decreased again (Fig. 4C). In contrast, no CVB3 isolates were detected in 2018 and 2019, but the number of CVB3 isolates detected increased significantly from 2020 (n = 22) to 2021 (n = 67), accounting for 91.67% and 91.78% of the total number of isolates detected for the respective years (Fig. 4D).
VP1 sequence analysis of CVB5. In 2013 to 2021, a total of 282 CVB5 strains were isolated in Guangzhou sewage surveillance, and entire VP1 region nucleotide sequences of 849 nt were obtained, encoding 283 ammo acids. According to the preliminary sequence analysis and phylogenetic analysis, 111 representative sequences (GenBank number: OQ352015-OQ352125) from 2013 to 2021 were selected for analysis. The identity comparison showed that the sewage CVB5 sequences had 76.2% to 100% nucleotide identity and 94.6% to 100% amino acid identity, respectively. The nucleotide identity and amino acid identity of CVB5 and prototype strain Faulkner were 77.1% to 83.8% and 95.0% to 96.8%, respectively.

9-Year Prevalence of EVs in Sewage in Guangzhou
Microbiology Spectrum In addition to some genotype C strains in 2018, the sewage CVB5 strains from 2013 to 2021 belonged to genotype E (Fig. 5). VP1 sequence analysis of CVB3. In 2013 to 2021, a total of 148 CVB3 strains were isolated in Guangzhou sewage surveillance. The identity comparison showed that the nucleotide sequence identity of VP1 between CVB3 isolates from sewage surveillance was 78.9% to 100%, and the identity with the prototype strain of CB3 was 76.4% to 80.1%. Phylogenetic analysis was performed based on the 852-nt VP1 sequences, and 55 representative sequences (GenBank number: OQ351960-OQ352014) from 2013 to 2021 were selected for analysis. CVB3 isolated before 2019 belonged to genotype E, but it belongs to genotype C since 2020 (Fig. 6). Therefore, the phylogenetic tree indicates that a new CVB3 transmission chain may emerge in Guangzhou in 2020, leading to the prevalence of new genotypes.

DISCUSSION
To the best of our knowledge, this is the longest systematic study of EV prevalence in sewage in China, which reports the predominant NPEV serotypes circulating in environmental samples, emphasizing the important role of environmental surveillance in tracking EVs circulating in the population. Although a relatively uniform poliovirus surveillance program has been established worldwide to inform and evaluate vaccination and elimination programs, NPEV surveillance differs substantially between countries: for example, surveillance in the United States is based on voluntary reporting of EV typing data (20,21) whereas in Europe, NPEV surveillance typically focuses on EVs detected in severely infected patients hospitalized with neurological symptoms (22). Notably, Japan has adopted a sentinel-based surveillance system for NPEVs that captures patients with aseptic meningitis; hand, foot, and mouth disease; herpangina; and acute hemorrhagic conjunctivitis (23). In contrast, EV surveillance in China is mainly based on hand, foot, and mouth disease and AFP case surveillance, because these two are notifiable infectious diseases in China (24)(25)(26). However, most human EV infections are generally asymptomatic, making it difficult to trace the infections through disease surveillance. Therefore, environmental surveillance is a sensitive method to detect silently circulating viruses, which should be regarded as a supplementary assessment tool to track the prevalence of EVs in the population.
In this study, we reported an overview of NPEV prevalence in the sewage of Guangzhou city, China. During the 9-year surveillance study (2013 to 2021), four common circulating NPEV serotypes were identified: E11, CVB5, E6, and CVB3 (in descending order). In contrast, E20, E24, E1, CVA21, E30, CVA9, and CVA24 were only occasionally detected in sewage samples; these results were similar to those of previous studies (15). Our results are similar to those of other studies: members of EV-B were identified most frequently, followed by EV-A and EV-C. Members of EV-D are rarely reported in municipal wastewater (18,27,28). Because we used the supernatants obtained after centrifugation of sewage for virus isolation, viruses combined with sediment in sewage could not be detected. Moreover, some viruses may exist in sewage below the detection limit, so some EV transmission may not be detected. Moreover, in our previous study, the cell lines used were proven to be conducive to the isolation of EV-B (15). Due to the limited range of EV species observations in sewage surveillance study and the dominance of EV-B detections, the overall EV species distribution has not been well-described on a seasonal basis, but EV-B is still generally found to be the most prevalent in summer and autumn. Although the number of EV serotypes isolated is limited, the advantage of environmental surveillance is that it can use sewage samples to monitor large populations and can dynamically track the ecological changes of EV diversity in a region by monitoring these changes over time (18). In this study, E11, CVB5, E6, and CVB3 were the four most commonly detected serotypes. E11 is one of the most common causes of meningitis in the world; cases of E11 related to meningitis; AFP; and hand, foot and mouth disease have been detected in Shandong, Fujian, Yunnan, and other provinces in China (29,30). In this study, E11 was most frequently isolated from 2018 to 2019. From 2009 to 2017, the dominant transmission chain of E11 was A1 genotype, but there were also C3 and D5 genotypes. Since 2018, the dominant transmission chain of E11 was D5 genotype, which was consistent with the genotyping results of enterovirus infection cases in the same period (31). Simultaneously, an outbreak of neonatal E11 infection occurred in Shunde city, adjacent to Guangzhou city, in 2019 (4). Notably, we have reported that E11 sequences in clinical cases from 2017 to 2019 are closely related to E11 sequences continuously sampled from local sewage from 2018 to 2019 (31), which reflects that clinical infection is most likely caused by hidden circulation in the population, and E11 may spread and circulate among different cities. This indicates that our longterm sewage surveillance provides an early warning for possible large-scale outbreaks of E11 in Guangdong Province of China. However, due to the long-term prevalence or silent transmission of E11 in the environment and population in the past decade, this may lead to the improvement of the level of herd immunity. Therefore, it may lead to the decline of the prevalence of E11 in the population and environment, which is reflected in the decrease of the number of E11 detected in 2020 and 2021. Genotype E E11 was found in the domestic sewage isolates in 2021, but no case of infection with this genotype was reported in Guangzhou. It is necessary to continue to monitor E11 in cases and sewage (31).
E6 is also one of the most frequently detected EV serotypes in sewage. The environmental surveillance of sewage in Shandong province has proved that E6 has been the main serotype for some years (32,33). Our study also showed that E6 was the most common serotype in sewage in 2013 and 2015. However, since the outbreak of COVID-19 in December 2019, many countries have actively taken preventive and control measures to limit the spread of severe acute respiratory syndrome coronavirus 2 (the virus that causes COVID-19), such as wearing masks, washing hands frequently, and maintaining social distance, which has greatly reduced crowd gathering and communication, thereby, reducing the incidence of many other infectious diseases (34,35). Therefore, the detection rate of most EVs in sewage, including E6 and E11, decreased significantly in 2020 and 2021.
CVB3 and CVB5 exhibit a trend of alternating popularity. CVB5 was detected in large numbers in 2013 to 2014 and 2017 to 2018, while CVB3 was most frequently detected in 2015 to 2016 and 2020 to 2021, and CVB3 was the serotype with the highest detection rate during 2020 to 2021, which may influence the level of herd immunity in the population. Among CVBs, CVB5 is the most common enterovirus serotype, which may cause human infection (36)(37)(38). Phylogenetic analysis showed that the global CVB5 could be divided into five genotypes of A to E. In China, most CVB5 strains belong to genotype E. For example, the CVB5 strains in sewage from 2013 to 2021 in our study mainly belong to the genotype E, but it is relatively scattered abroad (38). In China, CVB5 infection has led to more and more reports of severe hand, foot, and mouth disease, sporadic central nervous system disease, and aseptic meningitis (36,39). Genotype C is usually associated with sporadic cases and outbreaks, including cases in Shandong Province in 2005 and 2009 (36,40). In 2009 and 2012, CVB5 also caused aseptic meningitis outbreaks in Henan, Yunnan, Guangdong, and Changchun (38). Another study on sewage surveillance in Zhoukou City, Henan Province in 2018 showed that the nucleotide identity of CVB5 from AFP case surveillance and environmental surveillance was 94.7% to 99.6%, indicating that CVB5 from AFP case surveillance and environmental surveillance belonged to the same genotype and highly correlated (41). However, China has not included CVB5 infection in the disease surveillance system, so it is difficult to assess the disease burden comprehensively. Therefore, environmental surveillance provides important and useful information for case surveillance (41).
The results of environmental surveillance in Shandong Province and Guangdong Province of China showed that CVB3 was a serotype with high detection rate (14,15). The wide variety of transmission routes and high proportion of silent infection make the prevention and control of CVB3 infection more difficult, which is one of the reasons for the prevalence of CVB3 in Guangdong Province. CVB3 of genotype E is widely distributed around the world, including Asia, Europe, America, and Oceania, indicating that it may have strong transmission capacity. In the past, CVB3 of genotype D was the main prevalence in China, while CVB3 of other genotypes was only prevalent in some countries and regions (42). Interestingly, in our study, CVB3 isolated before 2019 belonged to genotype E closely related to the isolates from Shandong Province. However, since 2020, from the perspective of evolutionary relationship, genotype C has been closely related to French isolates. Thus, we conclude that it may be an imported genotype. At present, our research on CVB3 is insufficient, and environmental surveillance still needs to be strengthened.
In conclusion, outbreaks of diseases caused by EVs with severe consequences occur frequently worldwide, and both neonates and the elderly are more susceptible to EVs than other age groups. In the absence of a comprehensive and systematic EV disease surveillance system, environmental surveillance is a powerful and effective tool to investigate the silent circulation of EVs in the population. It is difficult to master the prevalence of EVs in the community caused by asymptomatic infections through disease surveillance, which reflects the significance of long-term environmental surveillance.

MATERIALS AND METHODS
Sewage sample collection. We obtained totally 432 raw sewage samples from January 2013 to December 2021 from the primary sedimentation tank at the Liede wastewater treatment plant located in the northern part of Guangzhou City, China. Four samples (1 L per sample) were collected from the inlets of the primary sedimentation tanks on a routine basis each month. Samples were collected, transported, and refrigerated in the laboratory within 2 h for sample processing.
Virus concentration. We used negatively charged membrane filters (mixed cellulose ester membrane filters; Advantec Co. Ltd., Tokyo, Japan) and sonication methods to concentrate viruses in sewage samples (43). The collected sewage samples were centrifuged at 4°C and 3,000 rpm for 30 min to precipitate impurities in the water samples. Subsequently, MgCl 2 (0.05 M) was added to the supernatant, and the pH was adjusted to 3.5 to 4.0 with HCl. The filter papers and the negatively charged membranes were installed on the filter in turn, and then the positive pressure pump was connected to slowly filter the sewage sample, so that the virus was adsorbed on the negatively charged membrane. To elute the viruses, the virus-adsorbed filters were cut into pieces and sonicated for 1 min in 10 mL of 3% beef extract solution (pH 9.6) (Meilunbio, Dalian, China). The filter was eluted twice to obtain the first and second eluents, and then the pH was adjusted to 7.0 with HCl. Following centrifugation at 4°C and 3,000 rpm for 30 min, the supernatant was aspirated and passed through a 0.22 mm syringe filter (Merck Millipore, USA) to remove bacteria and fungi.
Virus isolation. Human rhabdomyosarcoma and laryngeal epidermoid carcinoma cells used for the virus isolation in our experiments were obtained from National Polio laboratory at the Chinese Center for Disease Control and Prevention. The cells were cultured in 24-well plates and six parallel wells were set, of which 200 mL of the first and 200 mL of second eluents were pipetted into four and two wells, respectively. A total of 200 mL of each concentrated eluate was used to inoculate standard monolayers of cells and cultured in a 37°C, 5% CO 2 incubator. We recorded wells with cytopathic lesions by microscopy for 7 consecutive days. Independent molecular typing of positive culture supernatants was conducted (44).
RNA extraction and molecular typing. Viral RNA was extracted from the collected cultures using a QIAamp Viral RNA minikit (Qiagen, Valencia, CA, USA), and nested reverse transcription-PCR (RT-PCR) was performed according to a previously developed method (45). Briefly, cDNA was synthesized by using a Qiagen OneStep RT-PCR kit (final volume, 20 mL) with 1 mL each of cDNA primers (AN32, AN33, AN34, and 9-Year Prevalence of EVs in Sewage in Guangzhou Microbiology Spectrum AN35). Following incubation at 37°C for 60 min, 2.5 mL cDNA was then used in the first PCR (PCR1: final volume, 25 mL) with a Hot Start PCR Master Mix kit (Invitrogen, USA) with 1 mL (each) of primers 224 and 222, targeting a highly conserved motif in the viral protein (VP)3 and VP1 regions, respectively. The first-round PCR (PCR1) was carried out under the following conditions: 95°C for 30 s, 42°C for 30 s, and 60°C for 45 s. After 40 cycles of amplification, an additional elongation step at 60°C for 6 min was done, and then 2.5 mL of PCR1 products was used as a template in the second-round PCR with 1 mL (each) of primers AN88 and AN89, targeting a partial VP1 region. After 40 cycles of amplification (95°C for 30 s, 40°C for 20 s, and 72°C for 15 s) and an additional elongation step at 60°C for 6 min, the PCR products were analyzed on 1.5% agarose gels, and the positive products (approximately 350 bp long) were purified using a QIAquick PCR purification kit (Qiagen) and sent for sequencing using either primer AN88 or AN89. The sequences were analyzed with the Basic Local Alignment Search Tool server at the National Center for Biotechnology Information, and the serotype of each isolate was determined according to a previously described molecular typing method (46).
Sequence data sets and phylogenetic analysis. Partial VP1 sequences of enterovirus strains from the study were compared with sequences of representative strains available in GenBank (Table 2). BioEdit 7.0.9.1 software was used for nucleotide and amino acid identity analysis (36). These sequences were analyzed by the neighbor-joining method implemented in the MEGA-X program. The credibility of phylogenetic relationships was evaluated through 1,000 bootstrap replications (47).
Data availability. The complete VP1 sequences of 55 CVB3 strains and 111 CVB5 strains described in this study have been deposited in the GenBank database under accession numbers OQ351960 to OQ352014 and OQ352015 to OQ352125.