Distribution and Genetic Variability of Sapoviruses in Africa

In this review, we describe the distribution and genetic diversity of sapoviruses detected among humans, animals and the environment in African countries. Databases were searched for studies conducted in African countries and published between Jan 2005 and Mar 2019. Only studies where RT- PCR was used for initial detection were included in the systematic review. We identified 27 studies from 14 African countries with 18 focused on human sapoviruses, two on animal sapoviruses and seven on sapoviruses observed in the environment. Samples. The overall estimated pooled prevalence of human sapovirus infections among symptomatic and asymptomatic individuals was similar at 5.0% (95% Confidence Interval (CI): 3.0–7.0) and 2.0% (95% CI: 1.0–3.0), respectively. In environmental samples sapovirus detection rates ranged from 0% to 90% while in animal studies it was 1.7% to 34.8%. Multiple causes of gastroenteritis, sensitivity of detection method used, diversity of sapovirus strains and rotavirus vaccine coverage rate are some of the factors that could have contributed to the wide range of sapovirus detection rates that were reported. The studies reported human genogroups GI, GII, and GIV, with genogroup GI being the most prevalent. Some potential novel strains were detected from animal samples. Most studies genotyped a small portion of either the capsid and/or polymerase region. However, this is a limitation as it does not allow for detection of recombinants that occur frequently in sapoviruses. More studies with harmonized genotyping protocols that cover longer ranges of the sapovirus genome are needed to provide more information on the genomic characterization of sapoviruses circulating in African countries. Further investigations on animal to human transmission for sapoviruses are needed as inter-species transmissions have been documented for other viruses.


Introduction
Sapoviruses, whose prototype is Sapporo virus, are single stranded, non-enveloped, positive sense RNA viruses that belong to the Caliciviridae family [1][2][3]. Their genome is 7.1 to 7.7 kb in length and has a polyadenylated 3' end that is essential for viral replication [4]. The 5' end is linked to VPg that plays a major role in initiation of translation [5,6]. The sapovirus genome consists of two to three open reading frames (ORF) [5]. ORF-1 encodes a large polyprotein that is cleaved into nonstructural proteins and major capsid protein, VP1 [1,2]. ORF-2 encodes minor structural protein VP2 [1]. The function of ORF-3, which is found in some human and bat strains, has not yet been defined [7]. Different methods have been used for the detection of sapovirus that include enzyme-linked immunosorbent assay (ELISA), electron microscopy, next generation sequencing and reverse transcription-PCR (RT-PCR) [3].

Search and Selection Strategy
A systematic literature review of sapovirus studies from African countries published in peer reviewed journals was performed. Electronic databases PubMed, Medline, and Google Scholar were searched using key words "sapovirus", "caliciviruses", "gastroenteritis", "Africa", and "name of country" alone and in various combinations. To ensure that studies that do not appear in major electronic databases were not missed, Rayyan [30] and Google Search were also used.
Studies from African countries published between January 2005 and March 2019 that described sapovirus infections among humans, animals and environmental samples were included in this review. Only studies where sapovirus initial detection was carried out by RT-PCR were included. For genotyping, all studies that used any region of the capsid and or polymerase were included.

Data Extraction and Analysis
The following variables were extracted from each of the selected studies where available: study reference, publication date, period and duration of the study, clinical symptoms, type of gastroenteritis (sporadic, outbreak), specimen type (human stool, animal stool, environmental sample), study population (symptomatic, asymptomatic), age group, diagnostic method used, prevalence, seasonality, co-infections, genogroups and genotypes identified. In assessing seasonality, only studies with duration of 12 months or more were included. We estimated the prevalence of sapovirus (the proportion of cases) for each study. We also estimated pooled sapovirus prevalence rates by age category and symptoms presentation with exact 95% confidence intervals (CI) using STATA version 15.1. In order to account for heterogeneity we also conducted random-effects meta-analyses to pool the prevalence of the sapovirus by age grouping and presentation of symptoms. We assessed the variability across studies using I 2 statistics and assessed for heterogeneity using forest plots (with prevalence estimates with 95% intervals. I 2 describes the percentage of total variation due to inter-study heterogeneity. STATA version 15.1 (SataCorp, College Park, Texas, USA) was used for the analysis.

Distribution of Human Sapovirus Infections
In Burkina Faso, a study conducted among symptomatic under-five-year-old children between 2009 and 2010 reported a sapovirus detection rate of 18% [38] and another one conducted between 2011 and 2012 reported a sapovirus detection rate of 10.3% [49].

<5
Real-time RT PCR a n/a n/a n/a

Age and Sapovirus Distribution
Among symptomatic and among asymptomatic participants the estimated pooled stool positivity rates were similar at 5.0% (95% CI 3.0-7.0) and 2.0% (95% CI 1.0-3.0), respectively ( Figure 3). While some studies reported only mild cases of diarrhea and vomiting, others reported more severe cases where individuals were hospitalized [35]. Only one study reported cases of bloody diarrhea associated with sapovirus/HIV co-infection [44].

Sapovirus and Coinfections
Sapoviruses often occurred with other enteric bacteria like Cryptosporidium, E. coli and Shigella [36,44] and other viruses that included rotavirus, norovirus and adenovirus [49]. One study showed that children with diarrhea and sapovirus infection were likely to be infected with more than one enteric pathogen [44]. Three studies enrolled participants with known HIV status [33,42,44]. A study in Cameroon reported higher detection of enteric viruses among asymptomatic children who were HIV negative compared to asymptomatic adults who were HIV positive [33]. One study in South Africa reported that 37.5% (3/8) of children co-infected with sapovirus and HIV died [44]. These co-infected children were also likely to have bloody stools with mixed pathogen infections. A study from Kenya did not indicate any susceptibility to sapovirus infections due to an individuals' HIV status [42].

Seasonality of Human Sapovirus Infections
Sixty seven percent (12/18) of the studies conducted lasted 12 months or more and were used to assess the seasonality of the infections [35][36][37][38]40,[43][44][45][46][47][48]50]. These studies were three from Southern Africa, four from East Africa, two from North Africa and three from West Africa. Most of the studies did not show any clear seasonal patterns as infections occurred all year round. However, in North Africa (Tunisia) one of the two studies showed a winter peak [45] but the overall prevalence in the study was low at 0.8%. One study conducted in Burkina Faso, West Africa, showed a peak in the cold dry season [39]. The study in Malawi, Southern Africa showed a slight peak in the rainy season [35].

Sapovirus Distribution in the Environment
Seven studies from three countries, Kenya, South Africa, and Tunisia reported on sapovirus strains that were isolated from environmental samples, which comprised wastewater, river water and sewage [21,[53][54][55][56][57][58] (Table 3).These studies were conducted between 2003 and 2010 and ranged in duration from nine to twelve months. All of the studies were conducted in countries where human sapovirus studies had also been conducted allowing comparison of genotypes isolated from the environment to those circulating in the communities. In one study, sapovirus was not detected in sewage or wastewater [57]. Primers used in this study targeted the RdRp region. However, another study showed that sapovirus detection rate was as high as 90% in rural river water and sewage [53] and these used primers that targeted the RdRp-VP1 junction. In some instances treated wastewater and river water had high concentrations of sapoviruses with prevalence reaching more than 50% [56]. The reported data showed that not only were sapoviruses detected in higher concentrations but there was a wide diversity of sapovirus strains circulating in river water and water bodies in Africa. Three studies from South Africa and two from Tunisia had molecular data available where genotyping was done using either the partial capsid or partial RdRp region [21,[54][55][56]58]. Based on the phylogenetic analysis the sapovirus strains clustered in three genogroups GI (GI.1, GI.2, GI.3.GI.5, GI.6, GI.7); GII (GII.1, GII.2, GII.3, GII.4, GII.5, GII.8) and GIV, (Figure 4 and Figure S2). Genogroups GI and GII were isolated in both countries but GIV was only isolated in Tunisia. Two sapovirus strains isolated in Tunisia could not be assigned to any genotype [58] and these were identified using primers targeting partial RdRp.

Animal Sapovirus Infections
Two animal sapoviruses studies from Ethiopia and Tanzania [52,59] were reviewed ( Table 4). The animals that were studied included pigs [52], spotted hyenas, African lion and bat eared fox [59]. The detection rate of sapovirus ranged from 1.7% among pigs in Ethiopia [52] to 34.8% among spotted hyenas in Tanzania [59]. In both studies, sequencing was done using primers that target the RdRp region. In Ethiopia, one of the sapovirus strains isolated from pigs [52] was identified as genogroup GIII and another one could not be assigned ( Figure S3). Sapovirus strains detected in spotted hyenas, African lions, and bat eared fox from Tanzania clustered in monophyletic group and these could be a potential new group.

Animal Sapovirus Infections
Two animal sapoviruses studies from Ethiopia and Tanzania [52,59] were reviewed ( Table 4). The animals that were studied included pigs [52], spotted hyenas, African lion and bat eared fox [59]. The detection rate of sapovirus ranged from 1.7% among pigs in Ethiopia [52] to 34.8% among spotted hyenas in Tanzania [59]. In both studies, sequencing was done using primers that target the RdRp region. In Ethiopia, one of the sapovirus strains isolated from pigs [52] was identified as genogroup GIII and another one could not be assigned ( Figure S3). Sapovirus strains detected in spotted hyenas, African lions, and bat eared fox from Tanzania clustered in monophyletic group and these could be a potential new group.

Discussion
Viral gastroenteritis is a public health concern, with high morbidity and mortality particularly in the under 5 years age group [60]. Sapoviruses are recognized as one of the causative agents of viral gastroenteritis. Therefore, it is important to understand their distribution and molecular characterization in each region and country to contribute towards prevention and management strategies. There is no treatment for sapovirus disease, and prolonged shedding of the virus even after cessation of symptoms has previously been reported [20]. In this review, the pooled sapovirus prevalence among symptomatic cases, 5.0% (95% CI 3.0-8.0), and among asymptomatic, 2.0% (95% CI 1.0-3.0), were not statistically different. However, the detection rates observed in the reviewed African studies is consistent with what has been reported in other countries [9,19,60,61]. Since sapovirus has been identified as one of the causative agents of gastroenteritis [3], one would have expected rates among symptomatic individuals to be significantly higher than those among asymptomatic. One contributing factor could be the multiple causes for gastroenteritis that include bacteria and other viruses predominantly rotavirus and norovirus [61]. This was reported by one study from Nigeria that isolated rotavirus and norovirus strains and no sapovirus from symptomatic individuals [50]. In the reviewed studies, a wide range of sapovirus stool positivity rates varied from 0% in Nigeria, West Africa [50] to 19% in Angola, Southern Africa [34]. These ranges were similar to those reported in other countries outside Africa: Thailand 1.1% [19], Nicaragua 17% [29]; Peru 12.4% [2] and Philippines 7% [8].
The differences observed in sapovirus prevalence rates may also be due the sensitivity of the detection methods used in these studies. RT-PCR methods for the detection of sapoviruses have limitations even though they have higher sensitivity compared to other methods such as ELISA [3,4]. Amplification that targets the more conserved region between the polymerase and capsid has been shown to detect diverse genogroups compared to those that use either capsid or polymerase region only [3,8]. Most of the reviewed studies [38,42,49] that reported sapovirus detection rate of at least 10% were where the initial detection method used primers that targeted the conserved RdRp-VP1 region. Primers targeting the RdRp-VP1 region are able to amplify multiple sapovirus genotypes [4]. One study where primers targeting this region were used reported only six positive sapovirus cases belonging to the five sapovirus genotypes [42]. Another study in South Africa, using a similar method identified at least thirteen different sapovirus genotypes [48], indicating the sensitivity of this method in amplification of sapovirus. However, limitations were observed in some of the studies where primers targeting the partial RdRp were used to amplify the sapovirus strains [35,41,49,58,59]. This often resulted in misclassification of the sapovirus strains or inability to confidently declare new genotypes. For example, a study conducted in Malawi [35] initially identified the isolated sapovirus strain as a GIII, and further analysis classified this as a GII.1 ( Figure S2). Some sapovirus studies conducted in Tunisia identified strains that did not cluster with known sapovirus genotypes [58,59]. While these could be potential new genotypes they could not be classified as such since new genotypes are normally defined based on the complete VP1 sequence [3]. These reports indicate the need for standardized and more sensitive methods for the detection of sapovirus.
The coverage rate of rotavirus vaccination may also contribute to the differences in sapovirus prevalence among the reviewed African countries. It has been reported that where rotavirus vaccination coverage is high, norovirus and sapovirus infections are prominent [29] and with rotavirus vaccination coverage of greater than 95%, sapovirus infections were the second major cause of viral gastroenteritis following norovirus [29]. This information underscores the need to understand the burden of sapovirus in each country especially in countries where rotavirus vaccine coverage is high. Between 1998 and 2014 when most of the reviewed studies were conducted, rotavirus vaccination coverage among African countries ranged from 0% to more than 95% [62,63] and this can contribute to the differences in the observed sapovirus detection rates among symptomatic individuals. It was observed that detection rates of sapovirus infections also differed in regions of the same country [43], where the same factors would contribute to prevalence.
Globally, gastroenteritis continues to affect many people, especially children and the elderly. The pooled prevalence on sapovirus detection among symptomatic children under 5 years was 5.0% (95% CI 3.0-8.0), which was comparable to those reported in other regions, and confirmed sapoviruses as one of the many causes of gastroenteritis among children under 5 years [8,29,60]. Though some of the studies included all age groups, data on the impact of sapovirus infections among the elderly were scarce. Gastroenteritis is also a common illness of diverse etiology among the elderly especially those in nursing homes [64]. Understanding the impact of sapovirus to this health problem can contribute to its management and prevention among this age group.
The findings of generally milder symptoms of diarrhea, vomiting, and dehydration that were associated with sapovirus infections [49,65] is comparable to what other studies have described in other regions outside Africa [29,60,66]. However, severe sapovirus symptoms are not uncommon, and have often been associated with GI infections [29]. A study from South Africa [44], reported more severe symptoms that included those in which bloody stool and sapovirus was detected in 11.4% (9/79) of those who died, with three of them being HIV infected. The GI.2 was also the most common strain in this cohort and could have contributed to the severe symptoms. Moreover, lack of access to sanitation was observed to increase rates of sapovirus infection especially among HIV-infected children. Control measures for the prevention of sapovirus infections should focus more on maintaining good personal, food and environmental hygiene.
There is an indication that different weather conditions can affect sapovirus infection. Some studies have reported seasonal patterns of sapovirus infections with most infections from sporadic cases occurring in the cold season [3,67]. However, other studies have reported infections that occurred all year round [35,45]. The reviewed African countries fall in regions of different weather conditions and these differences are sometimes observed within a country. A study in Tunisia showed that infections peaked in winter, similar to studies in Japan and Iran [67,68]. However, another study in Tunisia did not show any seasonal patterns [47]. Furthermore, a study conducted in Burkina Faso showed infection peaks in the cold and dry season [49]. Nonetheless, this study only lasted for 10 months [49]. Prevention and management efforts that include, among others, long-term surveillance programs for all common etiological agents of gastroenteritis in each country can provide the necessary information for its control and management.
Sapovirus as an RNA virus undergoes frequent mutations leading to the emergence of new strains that have altered antigenicity and can persist over time leading to more disease outbreaks [3]. Among the eight studies that conducted molecular characterization of sapoviruses, the detected genogroups, GI, GII and GIV have also been identified in outbreaks in other continents [4,67,68]. On average GIV was the most prevalent genogroup isolated mainly in South Africa between 2009 and 2013 [48]. In other regions of USA, Japan, and Europe, this strain was most prevalent in 2007 [3]. Initially, classification of sapovirus was not standardized leading to some strains being wrongly classified. An example is a study conducted in 2005 in Malawi among hospitalized children, which reported isolation of eight strains of sapovirus, and all of them were classified as GIII [35]. GIII is commonly isolated in pigs [1,18] and its presence in humans has not been reported elsewhere. Sapoviruses like most RNA viruses evolve through recombination to improve their fitness and evade the host immune system. The use of whole genome sequences or those that cover the ORF1/ORF2 junction are needed to detect the presence of recombinants. Studies that provide sapovirus whole genome analyses are needed in African countries to add to the body of knowledge in this area, especially the discovery of novel strains.
Sapovirus strains have also been detected in environmental samples of river water, treated and untreated sewage and very often strains isolated in the environment reflect those circulating in the surrounding communities [69]. In these studies, sapovirus strains detected in the wastewater bodies were similar to those identified on other continents [22,54,69]. There was correlation in the frequency of sapovirus detections in the environment and the prevalence of human infections, for example, in Tunisia where human sapovirus studies reported very low detection rates of 0.8% and 1% [45,47]. Studies conducted on sewage samples did not detect any sapovirus [57]. Another study conducted in Tunisia in waste water identified potential new genotypes belonging to genogroups II [58] and similar genotypes had previously been detected among children in USA [70]. In South Africa, genotypes that were isolated in human studies were also detectable in river water, treated and untreated wastewater [48,55,56]. The stability of sapoviruses in different environmental conditions influences their transmission back into the community. In the different countries that were reviewed, living conditions such as access to clean water and cultural practices in food handling vary in different parts of a country, and can contribute to the different patterns observed in the distribution of sapoviruses.
While most mammalian sapoviruses belong to particular genogroups not infecting humans, strains that genetically resemble those of humans have been identified [4]. Thus, these could possibly serve as reservoirs for human infections or vice-versa [5]. Some of the factors that have been suggested to enhance animal to human transmission include broad host range and infection of wild animals [71]. As shown in Table 1, sapoviruses have a wide host range including wild animals, making them a potential candidate for human-animal transmission.
In conclusion, this review showed that human sapoviruses circulating in African countries are genetically diverse, with stool positivity rates varied from 0% to 19%. Sapovirus detection rates in environmental samples ranged from 0% to 90% while in animal studies it was 1.7% to 34.8%. Potential new sapovirus strains have also been identified and these can contribute to broader understanding of sapovirus evolution. From the reviewed data, it is evident that studies in Africa can contribute to the discovery of novel sapovirus strains especially in areas with a high population of animals. Moreover, due to emerging viruses or mutations, more studies need to be undertaken that characterize longer sections of the genome.
Supplementary Materials: The following are available online at http://www.mdpi.com/1999-4915/12/5/490/s1, Figure S1: A phylogenetic tree of human sapovirus using 183 bp of partial RdRp region, Figure S2  Trust, the U.K. government. The funders had no role in the study design, data collection, and decision to publish, or in the preparation of the manuscript. The content is solely the responsibility of the authors and does not necessarily represent the official views of the funding agencies.