Background

Acute gastroenteritis (AG) outbreaks occur worldwide annually and can be caused by viruses, bacteria, and parasites. Norovirus and rotavirus are representative viruses causing infectious gastroenteritis outbreaks, and these diseases can affect everyone [1]. Sapovirus also causes viral gastroenteritis, and several sapovirus outbreaks have been reported [2]. In Yokohama, Japan, norovirus, rotavirus, and sapovirus outbreaks have occurred in various settings annually [3,4,5], and outbreaks due to these viruses have been repeatedly reported at the same facilities in Yokohama over several years.

Based on the major capsid (VP1) gene, norovirus strains can be classified into 10 genogroups: GI–GX [1, 6,7,8]. Most noroviruses detected in human infections belong to genogroups GI and GII, which contain nine and 26 genotypes, respectively [6,7,8]. Norovirus strains can also be divided into polymerase genogroups and genotypes using a partial region of the RNA-dependent RNA polymerase (RdRp) gene [6,7,8]. A recent mathematical model based on community transmission estimated that immunity to norovirus likely lasts 4–8 years [9]. Human challenge studies have shown poor cross-reactivity between GI and GII viruses and have reported that most repeat infections were due to genotypes that differed from those of previous infections [10, 11]. Genotype-specific herd immunity may influence norovirus outbreaks at the facility level [12]. Additionally, sapovirus is antigenically diverse and is classified into multiple genogroups and genotypes [2]. A case of reinfection with sapoviruses from different genogroups was recently reported, although protective immunity among humans to sapovirus infections remains unknown [13, 23, 24].

PCR for the partial RdRp and VP1 regions of norovirus strains

Some strains were analyzed for the partial RdRp and VP1 regions of norovirus. cDNA was synthesized from the extracted viral RNA with SuperScript III Reverse Transcriptase (Invitrogen, Carlsbad, CA, USA), and random hexamer primers (Takara Bio Inc., Shiga, Japan) were used as the PCR template. PCR was performed for amplification with TaKaRa Ex Taq DNA polymerase (Takara Bio Inc., Shiga, Japan). Additional file 1 lists the primers used for PCR. Our designed PCR primers were used under the following cycling conditions: 95 °C for 1 min, followed by 40 cycles of 95 °C for 15 s, 55 °C for 30 s, and 72 °C for 1 min. The other primers were used as described previously [23,24,25,26,27,28,29,30,31,32].

Data analysis

The nucleotide sequences of the purified PCR products (QIAquick PCR Purification Kit, Qiagen) were determined using the BigDye Terminator Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA) and a Genetic Analyzer 3130 or 3500 (Applied Biosystems) per the manufacturer’s instructions. The obtained data were used to determine the norovirus genotype using the web-based Norovirus Genoty** Tool, Version 2.0 software [33]. The obtained partial RdRp and VP1 region data were used to construct a phylogenetic tree on the basis of nucleotide sequences with the neighbor-joining method using MEGA 6 software (http://www.megasoftware.net/) with 1,000 bootstrap replicates. Genogroup/Genotype classification of norovirus were determined based on the latest 2019 reference [8]. The sequences reported herein were deposited in the DDBJ/GenBank/EMBL databases under accession numbers LC720153–LC720176.

Statistical analysis

Statistical analysis was performed using IBM SPSS Statistics Version 27 (IBM, Armonk, NY, USA). Student’s t-test or Welch’s t-test was used to determine significant differences between group means. P values < 0.05 were considered statistically significant.

Results

Outbreaks due to viral AG

During 10 consecutive 12-month periods starting in September 2007, 1459 AG outbreaks were reported in Yokohama. Table 1 summarizes the outbreaks by season. Almost AG outbreaks were due to person-to-person cases, although included some food poisoning cases. Outbreaks due to norovirus, sapovirus, RVA, or RVC were determined, and norovirus was dominant every season and consistently accounted for 87.4%–96.8% of the total outbreaks. These outbreaks occurred in various settings annually, but mainly in kindergarten/nursery schools (K/Ns), primary schools (PSs), and nursing homes for older adults (NHs). The distribution of settings in which the AG outbreaks occurred differed by causative virus (Additional file 2). Norovirus outbreaks occurred among all patient ages. RVA outbreaks occurred mainly in infants, and 57% occurred in K/Ns. Sapovirus outbreaks occurred mainly in children, and 53% occurred in PSs. All RVC outbreaks occurred in PSs. Additionally, we investigated the norovirus genotype distributions in K/Ns, PSs, and NHs by AG season (Fig. 1) and found that the dominant genotype in each season was altered in PSs.

Table 1 Number of virus-associated outbreaks due to acute gastroenteritis in Yokohama, Japan
Full size table
Fig. 1
figure 1

Distribution of norovirus genotypes in a K/Ns, b PSs, and c NHs by season. Distributions from 2007–2008 to 2014–2015 season were confirmed in our previous study (ref. 3). GII.other includes undetermined genotypes for GII. K/Ns: Kindergarten/nursery schools; PSs: Primary schools; NHs: Nursing homes for the aged

Full size image

Repeated AG outbreaks reported at the same facilities

During the study period, 1099 facilities reported AG outbreaks. Of these facilities, 227 reported multiple outbreaks at the same facility, accounting for approximately 20% of all facilities. Table 2 summarizes the repeated AG outbreaks for each setting. Repeated AG outbreaks occurred in 94 K/Ns, 95 PSs, 29 NHs, 4 welfare facilities, 3 hospitals, and 2 other settings, with higher proportions in K/Ns (94/353) and PSs (95/255). The most frequent occurrences for each setting were 10 outbreaks at a K/N (KNo5-7), 7 at a PS (PSo5-3), 5 at an NH (NHo5-1), 3 at a welfare facility (WF3-1), 2 in a hospital (HP2-1–HP2-3), and 2 in other settings (OT2-1 and OT2-2; Additional file 3). No reports of multiple outbreaks at the same facility were confirmed in restaurants, junior or senior high schools, or universities. We investigated the viruses and norovirus genotypes detected in these 227 facilities (Additional file 3), and 76.2% (173/227) had only norovirus outbreaks. Figure 2 summarizes the detected virus combinations.

Table 2 Number of facilities where repeated AG outbreaks were reported during 10 AG seasons
Full size table
Fig. 2
figure 2

Combinations of AG viruses or norovirus genotypes detected in each setting. a AG outbreaks occurred twice. b AG outbreaks occurred more than three times. Other settings are one child consultation center and one foster home. Other viruses are sapovirus, RVA, and/or RVC. K/Ns: Kindergarten/nursery schools; PSs: Primary schools; NHs: Nursing homes for the aged

Full size image

Figure 2a and Table 3 describe the 149 facilities where two outbreaks occurred at the same facility. Norovirus + norovirus combinations (“norovirus + norovirus”) occurred in 120 facilities, and different genogroup or genotype combinations occurred more frequently than the same combinations. Twenty-three facilities had the same genotype combinations. Of these, 11 NHs, 2 K/Ns, and 1 hospital had different genogroup II genotype 4 (GII.4) variant combinations. Seventy-one facilities had different norovirus GII genotype combinations; these combinations were the most common among the 120 facilities that had “norovirus + norovirus”. K/Ns had different norovirus GII genotype combinations more frequently than other settings (Fig. 2a). Fifteen GII genotype combinations were detected, and GII.2 + GII.4 and GII.3 + GII.4 were the most frequent (Table 3). Twenty-six facilities had different genogroup combinations (GI + GII), and PSs had different genogroup combinations more frequently than other settings (Fig. 2a). Other than “norovirus + norovirus”, 15 facilities had norovirus + sapovirus combinations, 11 had norovirus + RVA combinations, 1 had a norovirus + RVC combination, and 2 had sapovirus + RVA combinations.

Table 3 Interval between first and second outbreak for the facilities with two repeated AG outbreaks
Full size table

For these 149 facilities, we investigated the combination types and intervals between the first and second outbreaks (Table 3). The mean interval between the first and second outbreaks was 32.3 ± 22.5 months for all combinations, and the longest interval was 107 months for GII.4 + GII.17. For “norovirus + norovirus”, the mean interval was longer for the same combinations (36.1 ± 22.8 months) than for different genogroup or genotype combinations (32.0 ± 22.4 months). The mean interval was longer for norovirus combinations (32.8 ± 22.5 months) than for other virus combinations (30.2 ± 22.1 months). However, neither were significantly different between groups (P > 0.05).

Of 78 facilities that each had more than three outbreaks, all reported at least one norovirus outbreak (Fig. 2b). Furthermore, of these facilities, 53 reported only norovirus outbreaks. Six had the same norovirus genotype, GII.4, and all six were NHs (NH3-1–NH3-4, NH4-1, and NHo5-1). At NH3-2, three outbreaks were all the same variant type (GII.4 Sydney 2012; Additional file 3). Twenty-three facilities had different GII genotypes, and 24 had both GI and GII. Of these 47 facilities, 28 included the same genotypes. The remaining 25 facilities reported noroviruses and other AG virus combinations. Of these, eight included the same norovirus genotypes. Two facilities (KNo5-6, KNo5-7) had two RVA outbreaks, and two (PS4-9, PS4-12) had two sapovirus outbreaks. These outbreaks occurred with different RVA or sapovirus genotypes (data not shown).

Repeated AG outbreaks during the same season at the same facilities

Forty-four facilities experienced repeated AG outbreaks during the same season: 19 K/Ns, 20 PSs, 4 NHs, and 1 welfare facility. Table 4 summarizes these outbreaks. Four facilities (KNo5-4, KNo5-6, KNo5-7, and PS4-2) experienced repeated AG outbreaks in two seasons, and four facilities (KNo5-5, KNo5-7, PS4-13, and PSo5-3) had three AG outbreaks in the same season. Most of these 44 facilities had combinations of different norovirus genotypes or different viruses. Thirteen facilities had norovirus + RVA and/or sapovirus combinations, 12 had different GII genotype combinations, 10 had different norovirus genogroup combinations, and 1 had different GI genotype combinations. Seven facilities had combinations of the same norovirus genotypes (GI.3, GII.2, GII.3, GII.4, or GII.6), and PSs had more than other settings. PS4-2 had the same genotype combinations of GII.6 in the 2013–2014 season and GI.3 in the 2014–2015 season. Overall, the norovirus genotypes that predominated during each season at each setting were mostly related to repeated outbreaks during the same season (Table 4, Fig. 1). PS4-9 had sapovirus + sapovirus, and these genotypes differed.

Table 4 Repeated AG outbreaks during the same season
Full size table

Repeated norovirus outbreaks due to the same genotype at the same facilities

During the 10-year period, 44 facilities (15 K/Ns, 19 PSs, 9 NHs, and 1 hospital) had norovirus outbreaks of the same genotype (or the same variant type for GII.4; Additional file 4). KNo5-7 (GII.2, GII.6) and PS4-2 (GI.3, GII.6) each experienced two outbreaks of the same genotype combinations. NHo5-1, KNo5-1, and NH3-2 each had three outbreaks of the same genotype (GII.4 Den Haag 2006b, GII.4 Sydney 2012, and GII.4 Sydney 2012, respectively). These outbreaks were treated as two combinations of the 1st + 2nd outbreaks and the 2nd + 3rd outbreaks. Therefore, 44 facilities experienced 49 combinations of the same genotype.

Table 5 summarizes the average interval between these outbreaks by genotype and by setting. The most prevalent genotype was GII.4, followed by GII.2, GII.6, GII.3, GII.14, and GI.3; these genotypes have been dominant in Yokohama (Fig. 1). The mean interval between outbreaks per combination was 31.2 ± 26.8 months for all outbreaks. By genotype, the mean intervals were 39.9 ± 30.4 months for non-GII.4 genotype outbreaks and 18.6 ± 12.1 months for GII.4 outbreaks, and a statistically significant difference was observed (P < 0.05). The longest average interval was 47.2 months for GII.3, followed by GII.2, GII.6, GII.14, GII.4, and GI.3. Among GII.4 variants, the average intervals were 23.6 months for GII.4 Den Haag 2006b and 16.9 months for GII.4 Sydney 2012. Thus, the average intervals differed by genotype. By setting, the mean intervals were longer for K/Ns and PSs (36.4 ± 26.2 months and 35.7 ± 30.3 months, respectively) than for NHs (16.8 ± 12.1 months), and a statistically significant difference was observed (P < 0.05).

Table 5 Average interval between outbreaks by the same genotype at the same facility
Full size table

To clarify the relationship between the strains in each combination, we conducted sequencing analysis using a highly conserved N-terminal shell region. The analysis showed 85.8%–100.0% nucleotide sequence identity between the strains of each combination (Additional file 4). For seven of 49 combinations, the nucleotide sequences showed 100% identity. The outbreaks of eight combinations occurred in the same season, and of these, the nucleotide sequences of three combinations showed 100% identity. We further analyzed the sequences of the partial RdRp and VP1 regions for the strains of 12 combinations, which showed 100% identity or occurred in the same season. Table 6 compares the strains in each combination and lists the epidemiological information for each outbreak.

Table 6 Comparison of norovirus strains in each combination and epidemiological information for each outbreak
Full size table

Figure 3 shows the results of the phylogenetic analysis based on the nucleotide sequences of the partial RdRp and VP1 regions. P-types, defined as genotypes of the RdRp region, were consistent between outbreaks of each combination, and the phylogenetic tree showed 92.7%–100.0% nucleotide sequence identity between strains of each combination (Fig. 3a). The VP1 region showed 77.3%–99.9% nucleotide sequence identity and 87.2–100.0% deduced amino acid sequence identity between strains of each combination. Notably, the phylogenetic tree showed that the y14-V1098-4 of GI.3 was separated from y15-V1215-1 and other GI.3 strains (Fig. 3b). y14-V1098-4 and y15-V1215-1 had 12.8% amino acid differences and were different variants of GI.3, which was determined by a 5% cutoff for amino acid differences [34]. Additionally, when the shell and protruding domain of the GI or GII genotypes were defined based on the GI.1 strain (M87661) or the GII.4 strain (X86557) by temporary alignment of the VP1 amino acid sequences, most amino acid differences in each combination were in the protruding domain (Table 6). Notably, most antigenic differences in GII.4 variants map specifically on five major antigenic sites (A, C, D, E, and G) located on the P2 sub-domain in the protruding domain [35, 36]. Amino acid differences in five major antigenic sites between the strains in each combination related to GII.4 were observed in three of the five combinations.

Fig. 3
figure 3

Phylogenetic analysis of norovirus. Phylogenetic trees based on a the partial RdRp and b the VP1 nucleotide sequences. The tree was constructed with the neighbor-joining method using MEGA 6 software (http://www.megasoftware.net/) with 1,000 bootstrap replicates. The percentage of bootstrap support is indicated at each node (values < 95% are omitted). The scale bar represents the number of substitutions per site. Norovirus strains for which the genes were determined in this study are denoted in bold typeface. The y13-V1056-4 strain of the partial RdRp region was omitted because the determined length was too short to construct a phylogenetic tree

Full size image

Most outbreaks due to the same genotype in the same season occurred within 2 months. The epidemiological information suggests that these were not original infections, but infections that spread to distinct groups (Table 6).

Discussion

Viruses associated with diarrhea are reported worldwide annually and are public health problems. Among these viruses, noroviruses are the most common causes of gastrointestinal disease outbreaks [1]. In our study, norovirus was consistently predominant in all settings throughout 10 seasons, whereas sapovirus, RVA, and RVC did not exceed 10% in any season. The distribution of norovirus genotypes in K/Ns, PSs, and NHs was different, similar to our previous study [3]. The infants attending K/Ns have immature immunity and are in closer contact with adults than school children, therefore they may also be affected by the dominant genotype in adults among various genotypes. School children have a wider sphere of activity than infants, come in contact with more various genotypes. Conversely, transmission in NHs mainly occurs through person-to-person transmission through helpers or visitors. It is consider that the genotype detected in NHs was homogeneous because they have limited mobility and live in a confined space. Thus, the distribution of genotype is likely dependent upon the facility, especially the patient age [3].

To consider herd immunity at the facility level, we clarified the status of repeated AG outbreaks at the same facilities in Yokohama, Japan. When considering herd immunity at the facility level, the influence of different population, such as transition of people in facilities over time, cannot be avoided. However, in this study, we conducted analyses without considering these influences because such factors cannot be fully captured in these facilities. PSs and K/Ns had larger proportions of repeated outbreaks than other settings. Thus, younger people such as infants and children may be more susceptible to various viral genotypes because they have not acquired sufficient immunity [3].

Among 227 facilities that reported repeated AG outbreaks, 173 (76.2%) had only norovirus outbreaks. Of these, different genotype combinations occurred more frequently than the same genotype combinations, which is consistent with other reports in which most repeat infections were due to a different genotype than those of previous infections [10, 11]. Additionally, for facilities that experienced two norovirus outbreaks, the average interval between outbreaks was longer for groups with the same combinations than for groups with different genogroup or genotype combinations. Our results support those of a previous report that found norovirus genotype-specific herd immunity may influence norovirus outbreaks at the facility level [12]. Moreover, the average interval between norovirus combination outbreaks was longer than that of other virus combinations, possibly because cross-reactivity between allogeneic viruses is higher than that between heterologous viruses.

Multiple AG outbreaks occurred during the same season at the same facilities and most of these facilities had different norovirus genotypes or virus combinations, suggesting that herd immunity was maintained at the facility level for at least the same season. An early challenge study conducted on volunteers demonstrated poor cross-reactivity among viruses from different norovirus genogroups [10]. Recent reports showed that repeat infections with the same genotype were rare for a certain time period, possibly because of the immune protection obtained from initial infection of the same genotype [11, 34]. This study showed that most reinfections occurred with viruses of different immunotypes. Our results support this conclusion and show that all outbreaks due to different genotypes that occurred during the same season at the same facility were all caused by different immunotypes. Our epidemiological information suggested that most outbreaks due to the same norovirus genotypes in the same season may have been caused by the spread of infection to distinct groups. However, in kindergartens, nursery schools, and primary schools in Japan, children typically enroll every April; thus, new groups of children with different susceptibilities may have partially influenced these results.

During the 10-year study period, the same facilities experienced outbreaks of the same norovirus genotype combination by GII.4, GII.2, GII.6, GII.3, GII.14, and/or GI.3. These genotypes have been dominant in Yokohama and have been detected every season, although the frequency of detection fluctuates [3, 26]. The average interval between these outbreaks was 31.2 months during the study period. The average intervals between outbreaks differed by genotype and were 39.9 months for non-GII.4 genotypes cases and 18.6 months for GII.4 cases, and significantly differed between groups (P < 0.05). The average intervals were shorter for GI.3 and GII.4 than for other genotype combinations. For two combinations of GI.3, the intervals may have been much shorter because one was an expanded infection to distinct age groups, and the other was a rare case due to a different variant type. Although previous studies assumed that individual genotypes represent strains with similar phenotypes, recent studies have shown that evolution in some genotypes, such as GII.4, is sufficient to generate mutant clusters with new ligand-binding characteristics and antigenic properties [38, 39]. Because y14-V1098-4 and y15-V1215-1 had 12.8% amino acid differences in VP1 and the differences concentrated in the protruding domain, where there are predicted antigenic sites, the strain of the second outbreak (y15-V1215-1) in PS4-2 may have had different phenotypes from the strain of the first outbreak (y14-V1098-4).

Non-GII.4 genotypes sustain a low number of intragenotypic variants with a limited number of amino acids differences, even if they occur decades apart, whereas GII.4 produces the most intragenotypic variants [34]. Antigenicity changes in GII.4 were reported to be associated with amino acid substitutions in the protruding domains of VP1 proteins; additionally, even within each GII.4 variant, amino acid changes occurred in VP1, despite the presence of evolutionary constraints [26, 40]. VP1 sequence analysis of some GII.4 strains in this study revealed that most amino acid substitutions were detected in the protruding domain. Notably, there were amino acid differences in some combinations related to GII.4 in five major antigenic sites located on the P2 sub-domain. These antigenicity changes may have facilitated escape from herd immunity, leading to successive outbreaks in the short term with GII.4 variants. Additionally, an age-related decline in immune function is partially responsible for the increased prevalence of infectious diseases [41]. Most outbreaks that occurred in NHs were GII.4, which might have contributed to the shorter average interval between outbreaks.

Sakon et al. reported that genotype-specific herd immunity in infants and young children lasts for at least a few years, thereby influencing the endemic norovirus genotype in the next season [12]. The average intervals between outbreaks were 36.4 months for K/Ns and 35.7 months for PSs, suggesting that genotype-specific herd immunity lasts several years, which is consistent with the findings of Sakon et al. Additionally, our data indicated that herd immunity may influence endemic norovirus genotypes in the next season, especially in PSs (Fig. 1).

Recombination frequently occurs in the open reading frame (ORF)1/ORF2 overlap and is associated with antigenic shift [42]. Although we conducted sequence analyses of the RdRp region for the 12 strain combinations to consider the possibility of recombination, the P-type of each combination was consistent.

RVA vaccination provides protection against severe RVA. It does not confer sterilizing immunity, but may have indirect protective effects for unimmunized individuals as a result of others being immunized [43]. Antibodies to norovirus may protect against certain genotypes of norovirus infection, and vaccination is a way to actively acquire antibodies. Norovirus vaccinations may alleviate public health problems similar to RVA vaccinations. However, norovirus vaccine development has many difficulties and limitations, partly because of the limited availability of norovirus cell cultures, complexity of protective immunity against norovirus, antigenic variation among and within genogroups and genotypes, and unknown effects of pre-exposure history [44, 45]. We believe that our study is informative for public health, but is limited geographically to Yokohama, Japan. Comprehensive studies of facility-level herd immunity remain scarce. Norovirus genotype trends differ depending on year, area, and age group, and steady global surveillance and further studies are needed to influence future vaccine policy decisions.

Conclusions

We determined the statuses of repeated AG outbreaks at the same facilities in Yokohama, Japan. Most of these facilities experienced a combination of noroviruses. Our data indicate that herd immunity at the facility level was maintained for at least the same season. Norovirus genotype-specific herd immunity was maintained for an average of 31.2 months at the facility level during the study period, and the intervals differed depending on genotype. However, we conducted our analyses without considering the influence of different populations within each facility, so the results should be interpreted with caution.