Molecular Characterization of Giardia duodenalis in Children in Kenya

Background Giardia duodenalis is an important intestinal protozoan in humans worldwide with high infection rates occurring in densely populated and low resource settings. The parasite has been recorded to cause diarrhea in children. This study was carried out to identify G. duodenalis assemblages and sub-assemblages in children presenting with diarrhea in Kenya. Methods A total of 2112 faecal samples were collected from children aged ≤5 years and screened for the presence of Giardia cysts using microscopy. A total of 96 (4.5 %) samples were identified as Giardia positive samples and were genotyped using glutamate dehydrogenase (gdh), triose phosphate isomerase (tpi) and β-giardin loci. Results The three markers successfully genotyped 72 isolates and grouped 2 (1.4) isolates as Assemblage A, 64 (88.9) as Assemblage B and 7 (9.7 %) consisted of mixed infections with assemblage A and B. A further analysis of 50 isolates using GDH Polymerase Chain Reaction and Restriction Fragment Length Polymorphism (PCR-RFLP) categorized 2 assemblage A isolates as sub-assemblage AII while 6 and 14 assemblage B isolates were categorized into sub-assemblage BIII and BIV respectively. A mixed infection with sub-assemblage BIII and BIV was recorded in 28 isolates. Over half (55.6 %) of Giardia infections were recorded among the children between 13 to 48 months old. Conclusion This paper reports the first data on the assemblages and sub-assemblages of Giardia duodenalis in children representing with diarrhea in Kenya.


Background
Giardia duodenalis is a flagellated protozoan infecting humans and a wide range of animals worldwide, mainly transmitted through food and water contaminated with cysts [1,2]. In Asia, Africa, and Latin America, approximately 200 million people have symptomatic giardiasis with some 500,000 new cases being reported each year [3]. Previous studies on G. duodenalis had shown that the species comprises eight distinct genetic groups designated as assemblages A to H and which differ on the basis of host occurrence and genomic mutations [4,5]. All the assemblages have similar morphology and are indistinguishable using microscopy.
The genotyping of a large number of human Giardia isolates from different parts of the world revealed that humans are mainly infected with assemblage A or B with assemblage B being the most common [5]. Moreover these assemblages are found in numerous species of mammals and hence they are considered zoonotic. The assemblages C to H appear to be restricted to animals and are host specific, however occasionally assemblage C and D [6,7], E [8] and F [9] have been reported in humans.
Three sub-assemblages have been identified within Assemblage A and namely AI, AII and AIII [5,10]. The sub-assemblage AI is zoonotic, while subtype AII predominantly occurs in humans [11] and subtype AIII has solely been identified in animals (mainly wild ungulates) [12]. Within assemblage B, sub-assemblages BIII and BIV have been identified [13] and detected in humans, companion animals and wildlife. Studies searching for differences in clinical symptoms between people infected with assemblages A and B have reported varying results. Some studies reported a strong association between intermittent diarrhoea and assemblage A infection while persistent diarrhoea was strongly associated with assemblage B infection, while in others, children infected with assemblage A were more likely to be symptomatic compared with those infected with assemblage B [14,15].
The use of multi-locus genotyping approach using βgiardin, GDH, Tpi, SSU rRNA, ef1α, and variant surface protein [vsp] genes), is the preferred method for studying genetic variability in G. duodenalis from different hosts [5,6]. Moreover the use of primers based on Tpi marker detected more mixed infections with assemblage A and B than when general PCR primers were used [16,17]. In this paper, we report the detection and genetic variability of G. duodenalis in Human Immunodeficiency Virus infected and/or uninfected children presenting with diarrhoea in outpatient clinics at Mukuru informal settlement on the outskirt of Nairobi, Kenya and those admitted at the Paediatric ward at the Mbagathi district hospital in Nairobi.

Sampling and microscopy
Approximately 20 g of stool sample were collected from each child aged ≤5 years and who presented with diarrhea at the participating outpatient clinics and hospital. A total of 2112 faecal samples were collected and screened for the presence of Giardia cysts using microscopy. The stool was examined macroscopically for consistency, mucus and blood, and microscopically for the presence of ova, larvae, trophozoites or cysts of intestinal and extra-intestinal parasites through the formal-ether concentration method as described by Cheesebrough, (2005). Results of the parasitological survey are detailed in previously published study [18]. A total of 98(4.6 %) samples were identified as Giardia positive.

DNA extraction
A total of 98 samples identified positive for Giardia through microscopy were processed for extraction of genomic DNA using QiAmp® DNA stool Mini kit (Qiagen, Crawley, West Sussex, UK) and following the manufacturers protocol. The resulting DNA was aliquoted and stored at −20°C until further needed.
Amplification and restriction digestion of the GDH gene A fragment of the GDH gene of Giardia (432 bp) was amplified by seminested PCR using the primers GDHeF, GDHiR and GDHiF as previously described (Read et al. [6]). The resulting products were visualized on 1.5 % agarose gels stained with ethidium bromide. The resulting secondary PCR products were analysed by restriction (RFLP) through digestion with the restriction endonucleases Nla IV and RsaI separately to distinguish sub-assembages AI, AII, BIII and BIV [6]. The resulting profiles were visualized on 2 % high resolution grade agarose stained with ethidium bromide.

Amplification of Tpi gene
A fragment of approximately 605 bp of the tpi gene was obtained using the external primers AL3543 and AL3546, and internal primers AL3544 and AL3545 from primary PCR [19]. This was followed by two separate specific nested PCRs for assemblage A [16] that gave expected amplicons of 373 bp and assemblage B [17] that showed amplicons of approximately 400 bp.
Amplification and sequencing of β-giardin gene The primary fragment of Giardia β-giardin gene was amplified as described [10] and followed by the amplification of a secondary fragment of 511 bp using a nested PCR. The resulting amplicon were identified using a 2 % high resolution grade agarose stained with ethidium bromide.
The reaction mixtures containing the correct size fragment of 511 bp were purified using QIAquick PCR purification kit (Qiagen GmbH, Hilden, Germany) following the manufacturer's protocol. The resulting DNA was quantified using Nanodrop and the eluates that had concentrations of 50 ng/μl were prepared for sequencing. A total of 41 PCR products were sequenced in both directions using forward primer bGiarF 5′-GAACGAGATCG AGGTCCG-3′and reverse primer bGiarR 5′CTCGACG AGCTTCGTTGTT-3′ [12]. The DNA sequencing was carried out using the ABI Big Dye terminator sequencing kit. All pairs of sequences obtained were edited and consensus sequence generated using CLC DNA workbench 6.1 (CLC Bio, www.clcbio.com). Each consensus sequence from individual isolates was used for the identification of β-giardin assemblages and sub-assemblages.

Phylogenetic analysis
The resulting sequences were blasted using the basic local alignment search tool (BLAST) (http://blast.ncbi.nlm.nih.gov/Blast.cgi) to determine genetic relatedness of individual assemblages with sequences in GenBank. Multiple sequence alignment of the representative Giardia isolates and reference sequences of various assemblages was done using ClustalX 2.1 [19]. To assess the extent of genetic diversity of Giardia species in samples and their evolutionary relationships to other Giardia assemblages and sub-assemblages, a phylogenetic analysis was carried out using the software package MEGA version 5.05 [20]. Representative β-giardin gene sequences from each major G. duodenalis assemblages AI, AII, BIII, BIV and D with GenBank accession numbers X85958, AY072723, AY072725, AY072726, AY072727 and AY545648 were used as reference. Giardia ardeae (GenBank accession number AF069060) was used as out-group.

Ethics statement
The study was approved by the Kenya National Ethical Review Committee (SSC No. 1579). All parents and/or guardians of participating children were informed of the study objectives and voluntary written consent was sought and obtained before inclusion. A copy of the signed consent was filed and stored in password protected cabinets at KEMRI.

Multi-locus PCR
Microscopy showed 98 stool samples to be positive for Giardia of which 80(83 %) were positive using the PCR tests targeting GDH, Tpi and β giardin loci while sixteen were negative. The expected 432 bp GDH gene fragment was amplified in 73/96 (76 %) samples, however only 50 samples gave strong PCR products for RFLP analysis. The RFLP classified two isolates as assemblage A and 48 as assemblage B ( Table 1). The Assemblage A DNA showed RFLP patterns of 70, 80, 90 120 bp, typical for sub-assemblage AII with the NlaIV enzyme. Among the 48 assemblage B isolates, six were identified as subassemblage BIII, 14 as sub-assemblage BIV and 28 showed patterns of both BIII and BIV ( Table 2). The Tpi gene was amplified in 63 samples of which one was grouped as assemblage A, 56 as assemblage B and six DNA samples showed presence of both mixed assemblage A and B.

Phylogenetic analysis
The β-giardin locus was amplified in 60 samples of which 41 gave strong DNA products for sequencing. The targeted β-giardin sequence fragment was sequenced in thirty two isolates while nine gave short sequences that could not be analyzed. Based on the resulting sequences 30 G. duodenalis were categorized as assemblage B, and only two isolates M669 and M1021n were identified as assemblage A. These two isolates had been identified as of mixed infections with subassemblages AII and BIII through GDH PCR-RFLP analysis. Phylogenetic analysis grouped the isolates into three main clusters namely cluster I which contained assemblages B, with the majority of the isolates clustering within this clade (Fig 1). Cluster II, contained assemblage B with isolates (MB108, M1070, MB026, M1391, M011, M377) clustering with sub assemblages BI, BII, and BIII. Assemblage A (two isolates) clustered distinctly from the reference assemblage A isolates.

Distribution of Assemblages and sub-assemblages
The distribution of 72 assemblages successfully genotyped among the inpatients was 5 (7 %) and 67 (93 %) from outpatients of which 44 (61 %) were isolated from males and 28 (39 %) from female patients (Table 1). Giardiasis was more prevalent among the children between 13 to 48 months old with 56 (77.8 %) compared to 16 (22.2 %) the 0-12 and 49-60 months old (Table 1). We recorded seven assemblage B infections from HIV infected children. The only assemblage A isolate was from a HIV negative male child. The RFLP of the GDH locus is discussed above. All the mixed infection with

Discussion
Giardia duodenalis is among the most common intestinal protozoa and also the most frequent parasitic agent of gastroenteritis especially in the developing countries [21]. This study provides, for the first time in Kenya, data on prevalence and genetic diversity of G. duodenalis isolates from children in Kenya.
The genotyping results show that all Giardia infections in this population are due to G. duodenalis assemblages A and B. This confirms the results of a number of studies performed elsewhere [22]. Distribution of different assemblages differs among and within countries, as surveys in several countries showed a diverse prevalence of assemblages A and B [5]. Here we have shown that children in urban informal settlement in Nairobi, predominantly carry Giardia assemblage B, which conforms to reports from several other regions of the world [23][24][25][26][27][28]. Giardia Assemblage B displays high cyst excretion pattern, which in combination with oral-faecal transmission, may contribute to its elevated prevalence rates and broad distribution [29]. On the other hand, studies carried out in Germany, Uganda, Egypt and Portugal, reported a predominance of assemblage A [30][31][32][33][34]. Although both major G. duodenalis assemblages A and B have been found in humans throughout the world, their propensity to cause disease might vary.
The predominance of one G. duodenalis assemblage over another in a particular area has been attributed to biological as well as geographical factors and, in certain endemic areas; all infections due to Giardia in humans appear to involve just one assemblage [35]. The reasons behind the geographic variation in the predominance of the Giardia assemblages are still unclear. It may be explained by the difference in the dynamics of transmission. It has been known that assemblage A is most often responsible for zoonotic transmission with wide range of animals acting as reservoir hosts. Although assemblage B is most likely transmitted from human to human, it Distribution of Giardia sub-assemblages in relation to gender, age, patient type, HIV status and seasons, and distribution of the sub-assemblages among children presenting with different clinical symptoms. The sub-assemblages were identified through PCR-RFLP of the β-giardin gene has been reported in some animals and may represent a zoonotic potential as well [5,25,35,36]. In this study 76 % of the samples amplified successfully with GDH primers. Most samples analysed at this locus were identified as assemblage B, except 2 that were assemblage A. These isolates were further identified as sub assemblages AII, and both occurred as mixed infections with sub-assemblage BIII. Results of previous studies elsewhere have shown that humans are mostly infected with AII, although AI is also seen in some  Fig. 1 Evolutionary relationships of G. duodenalis isolated from selected test samples. The evolutionary history was inferred using the Neighbor-Joining method [44]. The optimal tree with the sum of branch length = 1.10022 is shown. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (2000 replicates) are shown next to the branches (Felsenstein et al., 1985). The evolutionary distances were computed using the p-distance method (Nei M et al., 2000) and are in the units of the number of base differences per site. The analysis involved 50 nucleotide sequences. All positions containing gaps and missing data were eliminated. There were a total of 289 positions in the final dataset. Evolutionary analyses were conducted in MEGA5 [20]. The samples are coded according to where they were recruited from and patient number. M1070 refers to Mukuru patient (outpatient) number 1070, MB108 refers to Mbagathi patient (inpatient) number 108 studies, while animals are mostly infected with AI, with AII being occasionally reported [22,25,37]. The AI subassemblage and the B assemblage, regardless of the B sub-assemblages, have a broad host range, including pets, wildlife, and livestock while the AII sub-assemblage is more limited to human subjects [3]. Thus, it is possible the AII infections reported in this study were anthroponotic, while the B infections could have been either zoonotic or anthroponotic. Human to human transmission of Giardia infection in the study area could have been exacerbated by the water shortage, and poor sanitary conditions in the slum areas, which has a direct effect on hygiene.
Our study identified both sub-assemblages BIII and BIV in the population, with BIV being commonly isolated. Sprong [11] reported that in Africa, infection with G. duodenalis assemblage B, sub-assemblage BIII was more prevalent than infection with sub-assemblage BIV, whereas this differed from findings in North-America where more infections were associated with subassemblage BIV, and only few with sub-assemblage BIII, with more balanced distribution being found in Europe and Australia [5]. This however differs with our findings, where BIV is more prevalent. Our study however agrees with findings from Thailand where Assemblage B, subassemblage BIV was found to be the most common in preschool children [38].
Occurrence of mixed infections in human cases of giardiasis involving various assemblages appears to be more common than previously thought [16,17]. Tpi assemblage specific primers have proved reliable enough to detect mixed assemblages in the presence of a few copies of the Giardia genomes [8,17,39]. Co-infection by both Giardia assemblage A and B which was observed in 6 cases has been previously reported in Ethiopia and Rwanda [9,25]. Co-infections with other rare assemblages have also been observed in Ethiopia, where mixed infections with A+ F were reported. In our study mixed infections with sub-assemblage BIII and BIV were frequently observed in 28(56 %) of cases, while AII&BIII was observed in 2 cases. Remarkably, mixtures between BIII and BIV have been previously commonly reported [11]. The occurrence of mixed infections by several assemblages/sub-assemblages of G. duodenalis reflects the complex circulation of the parasite in the environment and the exposure of the study population to multiple sources [40].
Phylogenetic analysis of the isolates after bi-directional sequencing of the β giardin gene showed that the assemblage B test isolates, formed two clusters. This could be attributed to genetic variation between reference sequences and the test samples, which after comparison of base pair position with the reference B assemblages revealed sequence profile variation within our isolates. A high degree of polymorphism in assemblage B has been observed in other studies [10,28,41], and has been further investigated by cloning [42,43]. This feature has been attributed to mixed subtype infections or allelic sequence divergence, or a combination of both. Assemblage B Kenyan isolates, formed two sub-grouping (Assemblage B, Assemblage B, cluster II). This could be attributed to genetic variation between reference sequences (AY072726.1, AY072725.1). Comparison of base pair position between, reference B assemblages, revealed sequence profile variation within the test isolates from the GenBank.
There was good agreement between assignment of assemblages at all three loci, with assemblage swapping (i.e., different assemblages at different loci in the same isolate) not being observed in any of the isolates. Assemblage swapping has been reported by other investigators [10,21,26] and has been attributed to recombination between assemblages or mixed assemblage infection.

Conclusion
The study provides some preliminary data on assemblage and sub-assemblage distribution of G. intestinalis in the country and highlighted that Giardia assemblages A and B are prevalent in children in Kenya, with a predominance of assemblage B. These findings suggest that anthroponotic transmission could be a dominant transmission route for giardiasis in Kenya, though there is need to explore the possibility of zoonotic transmission.