Introduction

Giardia duodenalis (synonyms G. lamblia and G. intestinalis) is one of the most common intestinal parasites worldwide and an important cause of waterborne and foodborne infections in humans (Adam 2001; Adams et al. 2004). Giardia infections are re-emerging with just under 200 million symptomatic cases and up to 500,000 new cases per year (Pires et al. 2015). They are a common cause of diarrhea in developing countries and poor communities in the developed world (Pires et al. 2015). The global distribution of the parasite is attributable to its resilient infectious cysts, which can persist for long periods in harsh environmental conditions (Feng and Xiao 2011). Infection by Giardia is initiated when cysts are ingested from contaminated water, food, or direct fecal–oral contact (Thompson and Monis 2012). After ingestion, excystment triggers release of motile trophozoites in the small intestine and encystation in the large intestine (Thompson and Monis 2012). However, in symptomatic hosts with diarrhea, there is no encystment and trophozoites are excreted which die soon after passage (Thompson and Monis 2012).

Giardia isolated from mammals has been divided into host-specific assemblages (A-H) which have been proposed as separate species (Thompson and Monis 2004). Assemblages A, referred to as G. duodenalis, and assemblage B, referred to as G. enterica, are zoonotic having a broad host range and infect predominantly humans, dogs, and cats (McGlade et al. 2003; Paz e Silva et al. 2012). Unlike A and B, assemblages C and D have strong host specificities and are most commonly found in dogs, foxes, and other canines; assemblage E in livestock (cattle, sheep, pigs) and assemblages F and G are commonly found in cats and rodents, while H was identified in seals and gulls (Adam 2001; Feng and Xiao 2011). Initially, allozyme along with phylogenetic analysis resulted in sub-clustering within assemblages A and B, namely, A-I, A-II, A-III, A-IV, and recently, B-I, B-II, B-III, and B-IV. This was derived from nucleotide sequence data of the glutamate dehydrogenase (gdh) gene. However, when compared to other genetic loci such as small subunit ribosomal RNA (SSU rRNA) and other housekeeping genes coding for β-giardin (bg), elongation factor 1 alpha (ef1α), and triosephosphate isomerase (tpi), this separation was not fully supported (Feng and Xiao 2011). These inconsistencies at the sub-assemblage level have resulted in the need of a multilocus genotyping approach to more accurately identify subtypes. Even though assemblages A and B have strong zoonotic potential, sub-assemblage A-I preferentially infects domestic animals and A-II is more common in humans (Sprong et al. 2009).

Cases of human giardiasis have been reported from Jamaica and other Caribbean countries, but data from the region are mostly limited to surveys and screening studies. For example, asymptomatic and symptomatic giardiasis was reported from 95 primary schoolchildren in Cuba at 57 and 43 %, respectively (Pelayo et al. 2008). The prevalence among a hospital cross-sectional population in Jamaica was 5.1 % (17 of 328) using microscopic and antigen detection methods (Lindo et al. 1998). Epidemiological studies on Giardia infections in dogs are largely lacking from the English-speaking Caribbeans. Dogs which may be infected with zoonotic assemblages of Giardia are important companion animals in the Caribbean and also occur in large numbers as strays due to the absence of leash laws (Brown et al. 2011). The prevalence of Giardia infection in dogs in Trinidad and Tobago was 25 % using PCR; however, these dogs were not infected with either assemblage A or B and zoonotic transmission was not considered likely (Mark-Carew et al. 2013).

The focus of this study was the identification and characterization of Giardia from humans and dogs in Jamaica using PCR and multilocus sequence analysis to gain insight into their zoonotic potential.

Materials and methods

Sample collection and microscopy

Two hundred and eighty five human fecal samples which were submitted for microscopic diagnosis to the Department of Microbiology at the University Hospital of the West Indies (UHWI) were randomly selected for the study. Two hundred and twenty five stool samples were randomly collected from dogs from 20 veterinary practices across Jamaica whether or not they presented with symptoms of Giardia infection. If transportation was delayed, samples were stored at 4 °C until transported to the University of the West Indies (UWI).

Parasitological diagnosis included the direct smear for the detection of Giardia trophozoites in fresh diarrheal stool samples and formalin–ether concentration on formed samples (Becker et al. 2011; Becker et al. 2013). Zinc sulfate flotation (specific gravity; 1.2–1.25), was used for the detection of Giardia cysts in stool samples collected from dogs (Dryden et al. 2005). Subsamples (∼3 g) of all 510 samples were concentrated and stored in 85 % ethanol for DNA extraction.

Molecular investigations

DNA extraction was done by transferring 20 mg or 200 μl of stool into a 1.5-ml tube to 1000 ml of stool lysis buffer, heated at 70 °C for 10 min and then vortexed for 3 min. The remaining extraction procedure was done using the QIAamp Fast DNA Stool Kit® (Qiagen, Hilden, Germany) as described previously and the extracted DNA was stored at −20 °C until PCR was performed (Babaei et al. 2011).

Conventional and nested PCR amplification of the SSU rRNA, tpi, bg, and gdh genes

Conventional PCR was used to screen all stool samples by amplifying fragments of the SSU rDNA using primers RH11 and RH 4 (Table 1). Concentration of PCR master mix included 50 μl of reaction volume, molecular grade H2O, 25 mM MgCl2, 20 mM dNTP mix, HOT FIREPol® (5U/μl),10× Buffer B, and 10 pmol/μl of forward and reverse primer dilutions (Solis BioDyne®, Estonia). Amplification was done under the following cycle conditions: 35 cycles (95 °C for 15 min; 95 °C for 1 min; 52 °C for 2 min, and 72 °C for 2 min with a final extension of 7 min). Samples positive by conventional PCR were subjected to nested PCR targeting the tpi, gdh, and bg loci using primers specified (Table 1). Conditions for primary and nested amplifications were identical: 35 cycles (94 °C for 1 min, 45 °C for 30 s, and 72 °C for 45 s) using an Eppendorf® thermocycler which included pre-heating at 95 °C for 10 min and final extension at 72 °C for 7 min. The primary primer sets were used to amplify the external products, which were used as templates for the nested PCR reaction. Products were analyzed using 2 % agarose gel electrophoresis and a GelRed™ stain (BioTrend, Cologne, Germany) and visualized using a UV transilluminator box.

Table 1 Primers used in conventional and nested PCR at the SSU rRNA, tpi, bg, and gdh loci
Full size table

DNA sequencing

Positive amplicons were purified using the illustra GFX PCR DNA and Gel Band Purification Kit® (illustra GFX, GE Healthcare, Austria) and stored at −20 °C until DNA sequencing was performed. Sequencing was done using the ABI BigDye® Terminator V3.1 sequencing kit and an ABI automated sequencer (Applied Biosystems, Vienna, Austria). Sequences were assembled to consensus sequences with the GeneDoc® editor (version 2.7.000) and then compared to reference sequences from GenBank® by multiple alignments using the CLUSTAL W® alignment software (Gelanew et al. 2007; Zheng et al. 2014).

Results

Formalin–ether concentration revealed a prevalence of 0.7 % (2/285) of Giardia infections among humans while zinc sulfate flotation revealed 2.7 % (6/225) prevalence among dogs. Molecular analysis also demonstrated that Giardia infections were notably higher in dogs (19.5 %; 44/225) than in humans (6.7 %; 19/285) based on conventional PCR targeting the SSU rDNA.

Nested PCR showed that the tpi gene proved to be highly sensitive for the detection of the Giardia assemblages with a 100 % (63/63) detection rate. At the bg locus, a slightly lower detection rate of 70 % (44/63) was seen while only 55.6 % (35/63) revealed PCR amplicons at the gdh locus.

Sequence analysis of all 63 positive samples comparing the three genetic loci identified only assemblage A in both humans and dogs in Jamaica. Sub-assemblage A-II was dominant in humans at a rate of 84.2 % (16/19) and in dogs at 70.5 % (31/44). Assemblage A-I accounted for 15.8 % (3/19) of Giardia infections in humans and 29.5 % (13/44) in dogs, respectively.

At the sub-assemblage level, nucleotide sequence data provided conflicting results when comparing sequences at different genetic loci. Four samples gave inconsistent findings when compared at the SSU rRNA locus. Assemblage A-I was identified in three samples and assemblage B in a single sample based on sequences of this gene. When compared to all other genetic loci, the same samples were consistently identified as sub-assemblage A-II with one exception where sub-assemblage A-I was identified at the bg locus.

Sub-assemblage A-II was predominant among humans and dogs in Jamaica with clear evidence of genetic polymorphism within the tpi gene in this sub-assemblage.

Discussion

This study revealed prevalence of Giardia infections among humans and dogs in Jamaica of 0.7 and 2.7 %, respectively, based on microscopy. Conventional PCR yielded prevalence of 6.7 % in humans and 19.5 % in dogs. Further, multilocus sequencing at the tpi, bg, and gdh loci demonstrated that only assemblage A was found in both humans and dogs.

Generally, the prevalence was similar to a study previously conducted at the University Hospital of the West Indies (UHWI), which reported a prevalence of 1.8 % using microscopy and 5.2 % using antigen detection (Lindo et al. 1998). The UHWI is an urban tertiary referral hospital and has reported a general decline in intestinal parasitic infections, which was attributed to improved sanitation and living standards (Rawlins 1982). However, the true prevalence of Giardia infections in Jamaica, including also asymptomatic infections, is unknown, and it was last estimated at 6.3 % in a prospective community based cross-sectional study (Rawlins et al. 1991). In neighboring Cuba, the prevalence of infection was estimated at between 6.02 and 9.34 % but can be as high as 54.8 % in children in day care (Cañete et al. 2012; Jerez Puebla et al. 2015; Puebla et al. 2014). Asymptomatic and symptomatic Giardia infections were reported in Cuba at 57 and 43 %, respectively, from primary school children (Pelayo et al. 2008). Prevalence of human Giardia infections in developing countries ranges between 8 and 30 % in children (Dib et al. 2008; Johnston et al. 2010; Lalle et al. 2009; Sánchez-Vega et al. 2006; Solarczyk et al. 2010; Tungtrongchitr et al. 2010; Usluca et al. 2010). Among adults, infection rates were 25.1 % in pregnant women in Mexico and up to 14 % in African refugees to the USA (Manzardo et al. 2008; Robertson et al. 2010).

Despite the low prevalence reported from many countries, Giardia infection has been suspected to be underreported due to the intermittent shedding of cysts in stool and low sensitivity of microscopy when a single sample is examined. In the USA, the total number of unreported cases was estimated at 20,000 annually with an incidence of 7.4–7.6 cases per 100,000 population (Yoder et al. 2010). Molecular screening, in the current study, estimated a prevalence of 6.7 % in humans which was similar in Australia (7.6 %), and slightly lower in South Korea (2.5 %), Germany (1.5 %), the USA (1.4 %), and the UK (1.3 %) (Church et al. 2010; Huh et al. 2009; Read et al. 2002; Sagebiel et al. 2009). While these estimates based on multilocus sequencing revealed increased prevalence over microscopy, it should be noted that variations in DNA extraction methods, PCR inhibitors, and whether multicopy or single copy genes are targeted will affect the accuracy of molecular methods in estimating the true prevalence of Giardia infections (Thompson and Ash 2016).

The infection rates in dogs were 2.7 % based on microscopy and 19.5 % using PCR. High prevalence rates in dogs were also reported from the USA and Europe (19 and 24.78 %, respectively) (Epe et al. 2010). In Trinidad and Tobago, 25 % of dogs in one study were positive for G. duodenalis using PCR (Mark-Carew et al. 2013).

Interestingly, multilocus sequence analysis revealed assemblage A as the only assemblage identified in both humans and dogs in Jamaica, where sub-assemblage A-II accounted for most infections (84.2 and 70.5 % in humans and dogs, respectively) and assemblage A-I for the rest (19.8 % in humans and 29.5 % in dogs). No other Giardia assemblages were identified. The findings mirror those of four studies from Mexico, which reported only assemblage A in both humans and animal samples including dogs (Cedillo-Rivera et al. 2003; Eligio-Garcia et al. 2008; Lalle et al. 2005; Ponce-Macotela et al. 2002). However, direct comparisons should be done with caution as many of these studies are based on amplification of a single gene, and there are differences in methods used such as the absence of sequencing (Eligio-Garcia et al. 2008; Lebbad and Svard 2008). In contrast, assemblage B was consistently more common than assemblage A in four studies from Cuba occurring in 70.4 % of cases in one study (Puebla et al. 2016) assemblage B was also the predominant species among humans in Leon, Nicaragua, with only the non-human-associated assemblages C and D found in dogs (Lebbad et al. 2008; Puebla et al. 2016; Puebla et al. 2014). A predominance of assemblage A has been reported from several countries, and in humans, particularly sub-assemblage A-II seems to be linked to symptomatic infections; however, predominance and also virulence of either assemblage A or B in humans seems to be geographically very different and also to depend on other factors such as age, diet, and immune and socioeconomic status of the study population (Feng and Xiao 2011).

Assemblages C and D are considered host-specific in dogs and accounted for up to 76.5 % of cases in a study based on SSU rRNA gene sequencing in Italy (Berrilli et al. 2004). Mark-Carew et al. (2013) reported assemblages C and D at 15.4 and 80.8 %, respectively, among dogs in a study from the Caribbean islands of Trinidad and Tobago. In a recent study from Germany, high genetic diversity was seen in dogs presented to local veterinarians, including assemblages A, B, C, and D, with assemblages C, C/D, or D identified in 75.6 % of samples (93/123) (Pallant et al. 2015). However, in another study from Germany, similar to our study which included asymptomatic dogs, it also reported only very few host-specific strains (9 % assemblage C, 0 % assemblage D), while 60 % had assemblage A and another 27.3 % had mix of A and C (Leonhard et al. 2007). Generally, this seems to depend to a very high degree on the specific study setting (country, symptomatic/asymptomatic, pet/stray dogs Mexico).

The detection of a single circulating assemblage in dogs in the current study may be attributed to the cross-sectional nature of sample collection. Samples were not specifically collected from symptomatic animals presenting with gastrointestinal disturbances. As such, the predominance of assemblage A could indicate these dogs acted primarily as carriers. On the other hand, screening samples from dogs presenting with gastrointestinal disturbances might have shown a predominance of host-specific assemblages C and D, in addition to using species-specific primers. Moreover, Giardia infections are most commonly reported in young pups, while the vast majority of samples were collected from adult dogs (Alves and Santos 2016).

The predominance of sub-assemblage A-II in both humans and dogs was also unexpected. Sub-assemblages A-I and A-II differ in host preference. In accordance with our results, A-II has been most commonly reported from humans; however, in animals, A-I seems to be predominant (Sprong et al. 2009; Xiao and Fayer 2008).

While the high prevalence of Giardia infections in dogs coupled with the overlap in assemblages and sub-assemblages with the human infections strongly suggests a zoonotic transmission it does not provide definitive evidence. Nevertheless, zoonotic transmission is highly possible in Jamaica due to the large number of free-roaming and uncared for dogs on the island which has no leash laws. Robust epidemiological evidence for zoonotic transmission from dogs to humans was seen in north eastern India where in addition to shared assemblages there was a significant correlation between the presence of a Giardia positive dog in the household and the prevalence of human infections (Traub et al. 2004). What also has to be taken into account is what has been described as “reverse zoonotic transmission” or anthropozoonotic transmission (Thompson 2013; Thompson and Ash 2016). This has been demonstrated in transmission between humans and wildlife such as non-human primates in Uganda and muskoxen in the Canadian Arctic (Johnston et al. 2010; Kutz et al. 2008).

Nested PCR at the tpi locus has proven to be highly sensitive and specific for detection and identification of Giardia spp. Inconsistencies when the gdh and bg genes were targeted—with a slightly lower amplification yield has been reported previously (Thompson and Ash 2016). It has been reported that the Giardia-specific SSU rRNA PCR has a detection limit of 10 pg DNA/microliter while the primers targeting the tpi locus have a detection limit of 1000 pg/microliter (Jaros et al. 2011). Moderate detection rates using the gdh locus were also seen in a study on Giardia infections in domestic cats (Jaros et al. 2011).

Multilocus sequence analysis is increasingly being used for characterization of Giardia isolates and is necessary to understand the distribution of different genotypes and the infection dynamics in and between different hosts. The findings from this study, demonstrating a predominance of Giardia assemblage A in both humans and dogs in Jamaica, suggest that zoonotic transmission of the parasite is likely occurring in this tropical endemic focus.