First identification of genotypes of Enterocytozoon bieneusi (Microsporidia) among symptomatic and asymptomatic children in Mozambique

Enterocytozoon bieneusi is a human pathogen with a broad range of animal hosts. Initially, E. bieneusi was considered an emerging opportunistic pathogen in immunocompromised, mainly HIV-infected patients, but it has been increasingly reported in apparently healthy individuals globally. As in other African countries, the molecular epidemiology of E. bieneusi in Mozambique remains completely unknown. Therefore, we undertook a study to investigate the occurrence and genetic diversity of E. bieneusi infections in children with gastrointestinal symptoms as well as in asymptomatic children in Mozambique. Individual stool specimens were collected from 1,247 children aged between 0 and 14 years-old living in urban and rural settings in Zambézia (n = 1,097) and Maputo (n = 150) provinces between 2016 and 2019. Samples were analysed for E. bieneusi by nested-PCR targeting the internal transcribed spacer (ITS) region of the rRNA gene. All positive amplicons were confirmed and genotyped. Penalised logistic regression (Firth) was used to evaluate risk associations. The overall prevalence of E. bieneusi in this children population was 0.7% (9/1,247). A 10-fold higher prevalence was found in Maputo (4.0%; 6/150) than in Zambézia (0.3%; 3/1,097). All E. bieneusi-positive samples were from children older than 1-year of age, and most (8/9) from asymptomatic children. Nucleotide sequence analysis of the ITS region revealed the presence of four genotypes, three previously reported (Peru11, n = 1; Type IV, n = 2, and S2, n = 2) and a novel genotype (named HhMzEb1, n = 4). Novel genotype HhMzEb1 was identified in both asymptomatic (75%, 3/4) and symptomatic (25%, 1/4) children from a rural area in Maputo province in southern Mozambique. Genotypes HhMzEb1, Peru11, S2, and Type IV belonged to the Group 1 that includes genotypes with low host specificity and the potential for zoonotic and cross-species transmission. Being infected by enteric protozoan parasites and no handwashing were identified as risk associations for E. bieneusi infection. This study reports the first investigation of E. bieneusi genotypes in Mozambique with the identification of three previously reported genotypes in humans as well as a novel genotype (HhMzEb1). Findings highlight the need to conduct additional research to elucidate the epidemiology of E. bieneusi in the country, especially in rural areas where poor hygiene conditions still prevail. Special attention should be paid to the identification of suitable animal and environmental reservoirs of this parasite and to the characterization of transmission pathways.

identified in both asymptomatic (75%, 3/4) and symptomatic (25%, 1/4) children from a rural area in Maputo province in southern Mozambique. Genotypes HhMzEb1, Peru11, S2, and Type IV belonged to the Group 1 that includes genotypes with low host specificity and the potential for zoonotic and cross-species transmission. Being infected by enteric protozoan parasites and no handwashing were identified as risk associations for E. bieneusi infection. This study reports the first investigation of E. bieneusi genotypes in Mozambique with the identification of three previously reported genotypes in humans as well as a novel genotype (HhMzEb1). Findings highlight the need to conduct additional research to elucidate the epidemiology of E. bieneusi in the country, especially in rural areas where poor hygiene conditions still prevail. Special attention should be paid to the identification of suitable animal and environmental reservoirs of this parasite and to the characterization of transmission pathways.

Author summary
Enterocytozoon bieneusi is an obligate intracellular parasite that infects a wide range of vertebrate hosts. It is the most important etiological agent of human microsporidiasis. Most clinical and epidemiological studies conducted to date have focused on immunodeficient or immunosuppressed individuals including HIV+ patients and solid organ transplant recipients, as in those E. bieneusi infection causes life-threatening chronic diarrhoea. In contrast, latent microsporidia infections in immunocompetent individuals have received far less attention. Molecular epidemiological studies in humans and animals have revealed that E. bieneusi encompasses a very large diversity of genetic variants (genotypes) with marked differences in host specificity and even geographical distribution. In Mozambique, as in many other African countries, the epidemiology of E. bieneusi is completely unknown. Therefore, to identify the occurrence and genetic diversity of this pathogen in Mozambique stool samples were obtained from children, including apparently healthy and symptomatic, in Zambézia and Maputo provinces and tested for E. bieneusi by molecular methods. Results demonstrated the presence of E. bieneusi genotypes for the first time in Mozambique. Four genotypes were identified, three genotypes that have previously been reported in humans (Peru11, Type IV, and S2) and a novel genotype (HhMzEb1). Two of the genotypes Peru11, Type IV have also been frequently identified in animals indicating that potentially zoonotic E. bieneusi genotypes are inadvertently circulating in the surveyed populations. Additional population genetic studies are needed to elucidate the actual extent of the epidemiology and transmission dynamics of E. bieneusi in Mozambique.
wasting syndrome, and weight loss. Infections in immunocompetent subjects are often asymptomatic or result in self-limited diarrhoea and malabsorption [4,5]. Extraintestinal disorders and pathologies have also been reported, mainly in lung [6]. In addition, E. bieneusi infection in early childhood has been recently linked to impaired growth in children in low-resource settings including African countries such as Malawi, South Africa, or Tanzania [7,8]. Enterocytozoon bieneusi has a high genetic diversity and is capable of colonizing/infecting a broad spectrum of mammal and avian species. There are marked differences in host specificities and zoonotic potential among those genotypes [9][10][11]. Infections are acquired through ingestion of food and water contaminated with E. bieneusi spores, or through direct contact with faeces of infected persons and animals or with contaminated soils [2,12,13].
The highly polymorphic ribosomal internal transcribed spacer (ITS) of the ribosomal rRNA (rRNA) gene is the most widely used marker for assessing the genetic diversity within E. bieneusi [11]. Based on ITS nucleotide sequences nearly 500 genotypes have been validated according to current nomenclature standards that are distributed into 11 distinct phylogenetic groups [10,14]. Groups 1 and 2 comprise genotypes able to infect a broad range of mammalian species including humans and are, therefore, considered potentially zoonotic. On the other hand, groups 3-11 have strong host specificities and are considered to pose little or null zoonotic risk [11].
The Global Enteric Multicenter Study (GEMS) has provided important information on the aetiology and population-based burden of paediatric diarrheal diseases in sub-Saharan Africa and South Asia. GEMS included in their study enteric protozoan parasites Giardia duodenalis, Cryptosporidium spp., and Entamoeba histolytica, but not the microsporidia E. bieneusi [15]. However, in recent years there has been a steady increase in the studies aiming to improve our understanding of the epidemiology of E. bieneusi in Africa since this parasite is frequently reported in immunocompromised individuals in African countries (Table 1 and S1 Fig). Importantly, early microscopy-based studies documented E. bieneusi prevalences as high as 40-67% in apparently healthy people in Cameroon [16] and Nigeria [17]. These findings indicate that asymptomatic, chronic E. bieneusi infections could be more frequent than anticipated. Despite the undoubted progress achieved over the past years, the actual epidemiological situation of E. bieneusi in most African countries, including Mozambique, remains largely unknown. To fill this gap, the present study aims to investigate the occurrence and genetic diversity of E. bieneusi in symptomatic and asymptomatic paediatric populations in two of the most populous provinces of Central (Zambézia) and Southern (Maputo) Mozambique.

Ethics statement
Written informed consent was obtained from legal guardians of children voluntarily participating in this survey. All procedures involved in the study were approved by the Ethics

Study design
This study is part of an ongoing research collaborative effort involving the Spanish National Centre for Microbiology (SNCM) in Spain, the Centro de Investigação em Saúde de Manhiça

Collection of stool samples in Zambézia province
Zambézia (population: 5.1 million; total area: 103,478 km 2 ) is the second most-populous province of Mozambique, located in the central coastal region of the country. A prospective cross- sectional molecular epidemiological study was carried out between October 2017 and February 2019. All enrolled children were between 3 and 14 years old. A first set of samples were obtained from school children attending 18 public schools with each 35-2,111 (mean: 651) children in 10 of the 22 districts of Zambézia. Voluntary participants were provided with sampling kits (sterile polystyrene plastic flask with spatula and a unique identification number) to obtain individual stool samples. In addition, a second set of samples from children presenting with gastrointestinal complaints (abdominal pain, anal pruritus, bloating, constipation, diarrhoea, flatulence, loss of appetite, nausea, vomiting) seeking medical attention in health care centres (n = 6) and hospital settings (n = 1) from six districts of the province were collected. For both sets, an aliquot of each faecal sample was transferred to a TOTAL-FIX stool collection device (Durviz, Valencia, Spain) and shipped to SNCM in Majadahonda (Spain) for downstream molecular analyses.

Collection of stool samples in Maputo province
The study in the Maputo province was conducted in the Manhiça district (population: 0.2 million; total area: 2,373 km 2 ) that is located approximately 80 km north of the capital city Maputo in Southern Mozambique. The Manhiça district has been under a Health and Demographic Surveillance System (HDSS) from CISM since 1996, allowing the link between demographic and clinical data of the population [18].
Stool samples were collected from children under 5 years of age under an ongoing surveillance of diarrheal diseases conducted by the CISM since September 2015 in the Manhiça District Hospital (MDH) and other health peripheral facilities [18,19]. Children under 5 years of age presenting to the peripheral health facilities with moderate-to-severe and less-severe diarrhoea were recruited. Definition of diarrhoeal cases and inclusion criteria to participate in the research study were as previously described [20]. In addition, starting from December 2016, 1 to 3 community controls (asymptomatic children free of diarrhoea episode > 14 days) matched to the index case by age, sex and neighbourhood were identified (using the HDSS databases) and enrolled within 14 days after said index case was enrolled. Stools were collected for both cases and controls and sent to the laboratory for processing. Additionally, stools from children presenting to the sentinel health facility complaining with less-severe diarrhoea and those children presenting to the Xinavane Rural Hospital (XRH), were enrolled starting in April 2017 up to December 2018. Stool samples were collected in sterile flasks, transported to the laboratory, and stored at -80˚C without any preservatives for further testing of enteric pathogens as previously described [21].

Data collection
Epidemiological and clinical information was collected in a standardized questionnaire provided as part of the sampling kits and labelled with the same identification number. The questionnaire was completed by researchers interviewing participating children and their caretakers at the time of sampling. The questionnaire included basic demographic characteristics (age, sex, place of living), and potential risk factors including contact with livestock or companion animals, type of drinking water, and information regarding the defecation place. There were some differences between Manhiça and Zambézia questionnaires.

DNA extraction and purification
In stool samples from Zambézia, DNA was isolated from 200 mg of faecal material using the QIAamp Fast DNA stool mini Kit and the QIAcube robot (Qiagen, Hilden, Germany) following the manufacturer's instructions except that treatment of the samples with the lysis buffer was carried out for 10 minutes at 95˚C. Purified DNA samples (200 μL) were stored at -20 ºC for downstream molecular analysis. In Manhiça, DNA was manually extracted and purified using the QIAamp DNA Stool Mini Kit from 200 mg of faecal material following the manufacturer's instructions, with the same modification for the lysing step. Purified DNA samples (200 μL) were shipped to the SNCM for molecular testing.

Molecular detection of Enterocytozoon bieneusi
To detect E. bieneusi, a nested PCR was conducted to amplify a 390-bp fragment including the entire ITS as well as portions of the flanking large and small subunit of the rRNA gene [22]. Primary and secondary reactions (50 μL) were carried out using the outer primer set EBITS3 (5´-GGTCATAGGGATGAAGAG-3´) and EBITS4 (5´-TTCGAGTTCTTTCGCGCTC-3´), and the inner primer set EBITS1 (5´-GCTCTGAATATCTATGGCT-3´) and EBITS2.4 (5´-A TCGCCGACGGATCCAAGTG-3´), respectively. Reaction mixtures consisted of 1 μL of template DNA, 200 nM of each primer, 2.5 units of MyTAQ DNA polymerase (Bioline GmbH, Luckenwalde, Germany), and 5 μL of MyTAQ Reaction Buffer containing 5 mM dNTPs and 15 mM MgCl 2 . Cycling conditions for the primary PCR were as follows: after denaturation at 94˚C for 3 min, samples were subjected to 35 cycles of amplification (denaturation at 94˚C for 30 s, annealing at 57˚C for 30 s, and elongation at 72˚C for 40 s), followed by a final extension at 72˚C for 10 min. Conditions for the secondary PCR were identical to the primary PCR except only 30 cycles were carried out with an annealing temperature of 55˚C. Negative (no DNA template) and positive (E. bieneusi PCR-positive samples, genotypes Ebcar2 and EbpA) controls were included in all PCR runs. PCR products were resolved on 2% D5 agarose gels (Conda, Madrid, Spain) stained with Pronasafe (Conda).

Sequence analysis
All amplicons of the expected size were directly sequenced in both directions using the internal primer set in 10 μL reaction mixes using Big Dye chemistries and an ABI 3730xl sequencer analyser (Applied Biosystems, Foster City, CA). Sequencing data were viewed using Chromas Lite version 2.1 software (https://technelysium.com.au/wp/chromas/) to generate consensus sequences. The Blast tool (http://blast.ncbi.nlm.nih.gov/Blast.cgi) was used to compare those sequences with reference sequences deposited at the National Center for Biotechnology Information (NCBI). Enterocytozoon bieneusi genotypes were determined using the established nomenclature system based on ITS nucleotide sequence [14]. Sequences obtained in the present study were deposited in the GenBank database under accession numbers MN845065 to MN845068.

Phylogenetic analysis
Nucleotide sequences obtained in this study and E. bieneusi nucleotide sequences for all genotypes previously identified in human and animals in Africa as well as appropriate reference sequences to include all E. bieneusi groups retrieved from GenBank were aligned with the Clustal W algorithm using MEGA X [23]. Phylogenetic inference was carried out by the Neighbor-Joining (NJ) method as previously described [24], genetic distance was calculated with the Kimura parameter-2 model using MEGA X [23].

Statistical analysis
We analysed the data using EpiData 4.2.0 (EpiData Association, Odense, Denmark) and Stata software, versions 15 (STATA Corp., College Station, Texas, US). We calculated odds ratios (OR) for associations; a probability (P) value < 0.05 was considered evidence of statistical significance. We examined for possible confounders (change of > 20% in OR) and interactions. As both populations included selection by symptomatic and asymptomatic populations, we included "a priori" the symptomatic variable in the multivariable model. As the event was rare, and some factors (e.g. G. duodenalis) were found in all cases, we used a penalised regression (Firth regression) selecting risk factors with a P-value � 0.2 from the univariable analysis, using Akaike's information criterion (AIC) and Bayesian information criterion (BIC) to determine selection and evaluate the final model [25].

Results
A total of 1,247 children aged between 0 and 14 years-old were recruited to participate in this study in the Manhiça District in Maputo (n = 150) and Zambézia province (n = 1,097) in Mozambique. The main socio-demographic and epidemiological features of these paediatric populations are summarized in Table 2. In Zambézia, 807 asymptomatic schoolchildren and 290 children presenting with gastrointestinal complaints seeking for medical care were investigated in 10 and six districts of the province, respectively (Table 2 and S1 and S2 Tables). The overall male/female ratio was 1.0, and children in the age group 6 to 10 years-old represented 60% of the surveyed individuals. Three out of four children investigated lived in rural areas.
In the rural Manhiça district, 109 asymptomatic children and 41 children presenting with gastrointestinal complaints (11 with moderate-to-severe diarrhoea and 30 with less severe diarrhoea) and seeking for medical care were investigated (Table 2 and S3 and S4 Tables). In this subpopulation, the male/female ratio was 1.1, and children in the age group 12-23 months-old represented 40.7% of the surveyed individuals. Most children (62.7%) were from the area covered by the MDH. All these children lived in rural areas.
The main socio-demographic, epidemiological, and genotypes identified in each of the E. bieneusi infections identified in the present study are shown in Table 3. In the Zambézia province, E. bieneusi was found in 3 asymptomatic children. Two of them lived in a rural area, attended the same school, but were infected by different genotypes (Peru11 and S2) of the parasite. Both children were co-infected with G. duodenalis, Blastocystis sp., and Strongyloides spp. The third child was infected by the Type IV genotype of E. bieneusi and was co-infected with G. duodenalis. None of the symptomatic children investigated in this province tested positive to E. bieneusi.
In Maputo province, E. bieneusi was found in six male children, one presenting with moderate-to-severe diarrhoea and 5 asymptomatic. Five out of the six cases were younger than 2 years old. The S2 and Type IV genotypes were found in a single boy each, one of them coinfected with Rotavirus. Among the 4 boys infected with the novel genotype, HhMzEb1, one was HIV positive, presented with moderate-to-severe diarrhoea, and was co-infected with Cryptosporidium meleagridis. Another boy with this genotype was co-infected with C. parvum, and a third one was co-infected with Rotavirus. All six E. bieneusi-positive samples were coinfected with G. duodenalis.
Univariable analysis (Table 4) of the combined data showed that significant risk associations included infections with G. duodenalis or Cryptosporidium spp. and age, but handwashing was protective. In the Zambézia dataset alone, G. duodenalis was a significant association. In addition, travel, Blastocystis sp., and absence of a latrine were associated with presence of E. bieneusi (P value < 0.5), but confidence intervals crossed 1, with only 3 positive cases. In the Maputo dataset there were no significant associations. In the final multivariable model using the combined dataset, after adjusting for symptoms, the risk associations for E. bieneusi Phylogenetic analysis revealed that the novel genotype HhMzEB1 clustered within Group 1 when its relationship with other genotypes of human and animal origin previously reported in Africa was evaluated (Fig 1).

Discussion
Little information is currently available on the occurrence and distribution of enteric parasites in Mozambique [26][27][28], although the GEMS project has significantly improved our understanding of the epidemiology of diarrhoea-associated with parasites G. duodenalis, Cryptosporidium spp., and E. histolytica [15,29]. However, there is almost no information regarding Microsporidia, and only one study that aimed to investigate the prevalence of intestinal parasite infections in an informal settlement in Beira, Mozambique has reported E. bieneusi [27]. They detected E. bieneusi by real-time PCR in 9.3% (28/301) of the individuals tested and no genotyping was provided. Therefore, this is to our knowledge the first molecular epidemiological study describing occurrence and genetic diversity of E. bieneusi in Mozambique. The overall prevalence of E. bieneusi infection in the paediatric (0-14 years of age) populations from Mozambique tested in this study was 0.7% (9/1,247). Infections were more commonly identified in asymptomatic children in rural settings. The E. bieneusi infection rates found in our paediatric populations in Zambézia (0.3%) and Maputo (4%) were lower than the prevalence reported in Beira in Sofala province (9%) [27], indicating potential differences in the geographical distribution of the parasite in Mozambique. The disparity could also be associated with differences in socioeconomic characteristics of the studied population as the study in Beira was conducted in an informal settlement in an area that is frequently flooded and not connected to a sewage system after a local hospital noticed high number of diarrhoea cases in this settlement. Prevalence in Mozambique (0.7%) is lower to those previously documented in immunocompetent children populations (range: 4-9%) in other African countries including the Democratic Republic of São Tomé and Principe [30], Nigeria [31], South Africa [32], and Uganda [33], and much lower than those found in similar paediatric populations (range: 20-67%) in Cameroon [16], Nigeria [17], and Tanzania [34] (see Table 1). However, prevalence is influenced by the age group selected in the study and the presence of symptoms, so comparisons among studies must keep those differences in mind.
It is noteworthy to indicate that all E. bieneusi-positive children were also co-infected by other pathogens including two children in Manhiça that were HIV-positive. Indeed, one to three additional enteric parasites (G. duodenalis, Cryptosporidium spp., Blastocystis sp., and Strongyloides spp.) were found and the strongest risk associations were with G. duodenalis and Cryptosporidium spp. These data clearly depict a highly endemic scenario where polyparasitism is common as previously reported in the city of Beira [27].
Our study is limited by the small number of positive samples, thus the wide confidence intervals for our results. Furthermore, detailed information on handwashing (use of soap or not) and latrine use (toilet or outside latrine) were unavailable in the Zambézia dataset, and missing values for water source and animal livestock in the Maputo dataset implied limited exploration of these associations, especially given the low numbers of cases. Nucleotide sequence analysis on the ITS region of the nine positive cases revealed the presence of four genotypes, three previously reported (Peru11, Type IV, and S2) and a novel genotype (named HhMzEb1). This constitutes the first report of these genotypes in Mozambique. Genotypes Type IV and S2 were both identified in Zambézia and Maputo, while Peru11 was only identified in Zambézia and novel genotype HhMzEb1 only in Maputo. All genotypes identified in the present study, including novel HhMzEb1, belonged to Group 1 (Fig 1) that includes E. bieneusi genotypes with low host specificity that are found not only in humans but also in domestic and wild animals worldwide [10, 11,14]. This is the second report of genotype S2 in humans, that had only been previously reported in Malawi from three children (2 HIVpositive and one HIV-negative) and eight HIV-positive adults [35]. Genotype Type IV (also reported previously as K, Peru2, BEB5, CMITS1, BEB-var, PtEB III) has widely been found in humans in different regions throughout the world [5,[36][37][38] Table). In addition, Type IV has also been reported in a wide range of animal hosts including nonhuman primates, cattle, cats, dogs, rodents, birds, snakes, and black bears (S5 Table) as well as in water samples [44][45][46][47]. There are no previous reports of Peru11 (also reported previously as Peru12) in humans in Africa, but it has been found in humans in Asia and South America (S5 Table). In Africa, Peru11 has been found in a baboon in Kenya [48]. In addition, there are also reports of this genotype in other parts of the world in non-human primates, cats, raccoons, rabbits, rodents, and birds (S5 Table).
In this study, no mixed infections by different E. bieneusi genotypes were noticed using ITS-PCR and Sanger sequencing. However, it should be noted that co-infections seem a common finding in human populations in endemic areas. Indeed, intra-isolate diversity has been demonstrated in children with diarrhoea in Uganda using microsatellite markers (in addition to ITS) and subsequent cloning and sequencing of PCR products [49]. These data preclude us to conclusively state that mixed infections are not present in the surveyed Mozambican population as the methodology used in this study was not chosen to address mixed infections. The identification of potentially zoonotic genotypes Peru11 and Type IV, commonly found not only in humans but also in domestic and wild animals worldwide, indicates the potential for zoonotic or cross-species transmission. There are few studies in the African continent that include molecular characterization of E. bieneusi in humans and in animals. However, environmental (water) samples should be also considered as a source of infection by microsporidia, as E. bieneusi has been demonstrated to be involved in waterborne outbreaks of gastrointestinal disease [50]. Our findings emphasize the need of further studies to explore risk associations as these were limited in our study by the low prevalence and missing information. Studies could include water as well as animals that may be in contact with human populations in Mozambique to understand the modes of transmission of E. bieneusi, and, by analogy, of other diarrhoea-causing enteric protist species. Phylogenetic relationships among Enterocytozoon bieneusi genotypes identified in this study. All genotypes identified in humans and animals in Africa, and genotypes to cover all groups of E. bieneusi were included for comparative purposes. Analyses were inferred by a Neighbor-Joining method of the entire ITS of rRNA gene based on genetic distances calculated by the Kimura two-parameter model (MEGA X software). All nucleotide sequences include host information with the GenBank accession number in parenthesis. Nucleotide sequences determined in this study are identified with black circles before the genotype name. White circles indicate genotypes identified in Africa. https://doi.org/10.1371/journal.pntd.0008419.g001

Conclusions
This is the first molecular study of E. bieneusi in Mozambique. The parasite was primarily found in asymptomatic children in Maputo (Manhiça district) and Zambézia province. Molecular characterization detected a novel genotype (HhMzEb1) and three known genotypes (Type IV, Peru11 and S2) in diarrheal and healthy Mozambican children. The identification of genotypes previously described in animals suggests potential zoonotic and anthroponotic transmission. There is a lack of information of E. bieneusi in the African continent in humans, animal hosts and environmental samples. Under the One Health perspective, further genotyping studies are needed to better understand the epidemiology of this parasite.