Background

The HIV/AIDS epidemic is still a major public health problem, and the Southeast Asia region, where about a quarter of the world’s population live, is second only to sub-Saharan Africa in terms of disease burden [1]. In the 11 member states of the Southeast Asia region, there are an estimated 3.5 million people living with HIV [1]. Thailand is the only country with an HIV prevalence of over 1% among the adult population (aged 15–49 years), although a higher prevalence is reported among sex workers and their clients, men who have sex with men, people who inject drugs and transgender individuals.

Opportunistic infections still represent a serious threat for HIV-infected individuals, and among these, Cryptosporidium is recognized as a leading cause of prolonged, severe diarrheal disease, accounting for up to a third of diarrhea cases in HIV patients [2, 3]. Due to the lack of effective drugs to treat cryptosporidiosis, avoiding exposure to the parasite and maintaining immune competence are the only ways to prevent infection in these patients [4].

A recent global systematic review and meta-analysis of Cryptosporidium infection in HIV-infected people identified 106 studies with 43,218 patients from 36 countries examined [5]. Among the studies, nine were from Asia and the Pacific and eight from Thailand, where a pooled prevalence of 15.2% (95% CI: 8.8–21.6%) was inferred from 947 HIV-infected people [5].

Human cryptosporidiosis is mainly due to Cryptosporidium hominis and C. parvum, but many other species of potential zoonotic origin are known to infect both immunocompetent and immunocompromised people [3]. Therefore, information about the circulating parasite species and genotypes is essential in understanding the transmission dynamics of cryptosporidiosis. Moreover, studies have suggested that the clinical manifestations of cryptosporidiosis are influenced by the species or the genotypes involved [6, 7], therefore strengthening the value of informative genotyping data.

Thus far, molecular characterization of parasite isolates from HIV-infected people in Thailand has resulted in the identification of Cryptosporidium hominis, C. parvum, C. meleagridis, C. felis, C. canis and C. muris [8,9,10,11,12].

In this study, we investigated Cryptosporidium species and genotypes in 166 archived stool samples collected during 1999–2004 from 155 HIV-infected individuals at the Siriraj Hospital in Bangkok.

Methods

Study design

The investigation originally aimed at determining the etiologic cause(s) of diarrhea in HIV-infected patients [13]. Diarrhea was defined as the passage of at least two watery or loose stools per day, while chronic diarrhea as the continuous presence of diarrhea for more than 3 weeks. Patients attending the outpatient clinic at Siriraj Hospital, Bangkok and Bamrasnaradura Hospital, Nonthaburi, Thailand, were asked to provide stool samples. At the same time, HIV-infected patients without diarrhea, who were admitted to these hospitals because of other conditions, were also asked to provide a stool sample. Overall, 166 fecal samples from 155 HIV-infected patients, collected and archived between November 1999 and May 2004, were available for this study.

Microscopic examination of stools

Routine microscopic examination of fecal samples was performed at the Infectious Diseases and Tropical Medicine Laboratory, Department of Medicine, Faculty of Medicine Siriraj Hospital, as previously described [13]. All samples were examined by direct microscopy of a wet preparation, with or without iodine staining, and of sediments obtained by formol-ether extraction of about 2 g of feces. Sediments were analyzed by Ziehl-Neelsen staining for the detection of Cryptosporidium and Isospora oocysts, and by modified trichrome staining for the detection of Microsporidium spp. Entamoeba histolytica was not differentiated from the nonpathogenic Entamoeba dispar. Larvae of Strongyloides were detected by microscopy or by culture [13]. Culture was used to detect common enteric bacterial species, including Salmonella, Shigella, Vibrio and Campylobacter. Finally, Clostridium difficile toxin A was detected as reported [13].

DNA extraction and molecular typing for Cryptosporidium

DNA was extracted from approximately 200 mg (or 200 μl for watery or loose samples) of stool by using the FastPrep120 apparatus and the FastDNA kit (MP Biomedicals, Santa Ana, USA). For species identification, a nested PCR assay was used to amplify a ~590 bp fragment of the small subunit rRNA (18S rDNA) gene, as described [14]. For subtyping at the glycoprotein 60 (gp60) locus, primers Gp15start (5′-ATG AGA TTG TCG CTC ATT ATC-3′) and Gp15end (5′-TTA CAA CAC GAA TAA GGC TGC-3′) were used in primary PCR to amplify a ~950 bp fragment from all samples. For the nested reaction, primers AL3531 (5′-ATA GTC TCC GCT GTA TTC-3′) and AL3534 (5′-GCA GAG GAA CCA GCA TC-3′) [15] were used for samples classified as C. hominis or C. parvum, and primers ATGFmod (5′-GAG ATT GTC GCT CGT TAT CG-3′) and GATR2 (5′-GAT TGC AAA AAC GGA AGG-3′) [16] were used for samples classified as C. meleagridis. For samples classified as C. canis, C. felis and C. suis, combinations of the nested PCR primers reported above were tested. PCR conditions, for both primary and secondary amplification, were as follows: initial denaturation at 94 °C for 3 min, followed by 40 cycles of denaturation at 94 °C for 30 s, annealing at 50 °C for 1 min and extension at 72 °C for 1 min, followed by a final extension step at 72 °C for 7 min. Negative (no template) and positive (DNA from the C. parvum Moredun strain) controls were included in each experiment. PCR was performed using 25 μl of 2× GoTaqGreen (Promega, Madison, USA), 10 pmol of each primer, 5.0 μl of DNA, and nuclease-free water up to a final volume of 50 μl. Reactions were performed on a Perkin Elmer 9700 apparatus (Life Technologies, Carlsbad, USA). Aliquots of PCR reactions (10 μl) were loaded on 1.5% agarose gel stained with ethidium bromide. PCR products were purified using spin columns (QiaQuick PCR purification kit, Qiagen, Milan, Italy). Bidirectional sequencing of the PCR products was performed by a commercial company (BMR Genomics, Padua, Italy). Sequences were edited and assembled using the software package SeqMan v.7.1 (DNASTAR, Madison, USA). BLASTn searches (http://blast.ncbi.nlm.nih.gov/Blast.cgi) against the GenBank database were used to identify Cryptosporidium at the species and subtype levels. Novel gp60 sequences were deposited in the GenBank database under the accession numbers MK331714-MK331719.

Results

Epidemiological and clinical data

Among the 155 patients, 90 were male and 65 were female, with a median age of 34.5 years (range, 20–65 years) for both genders. The information collected from patients regarded the consistency of stool, duration of diarrhea, number of stool passed per day, presence of mucus and CD4+ T-lymphocyte count (Additional file 1: Table S1). Diarrhea lasting less than 3 weeks was observed in 52 patients, chronic diarrhea characterized 98 patients, while three patients had no diarrhea. No data about diarrhea were available for the remaining five patients. Concerning the presence of mucus, 85 had mucus and 53 patients did not have mucus in their stools; no data were available for the remaining 17 patients. Finally, the CD4+ T-lymphocyte count was determined in 113 patients, and the median count was 12 cell/mm3 (range 1–331); all patients, except two, had a CD4+ count of less than 200 cell/mm3.

Detection of enteric pathogens

Microscopic analysis and culture methods showed that 130 of 155 (83.9%) patients had at least one detectable pathogen in the stool (Table 1). Mixed infections with two or three pathogens were detected in 34 (21.9%) and six (3.9%) patients, respectively (Table 1).

Table 1 Pathogens detected in the stools of 155 HIV-infected individuals
Full size table

Protozoans detected by microscopy included Cryptosporidium spp. in 104 patients, Isospora spp. in four patients, Giardia spp. in three patients, and Entamoeba histolytica/dispar in two patients. Spores of microsporidian species were identified in 14 patients. Infection by helminths was found in 23 patients, with identification of Strongyloides larvae in 14 patients, of Opisthorchis viverrini eggs in six patients, of hookworm eggs in two patients and of Ascaris eggs in one patient. Mycobacterium tuberculosis was found in three patients. Enteric bacterial pathogens were found in 17 patients, including Campylobacter spp. in 10 patients, Salmonella spp. in two patients, Vibrio spp. in two patients and Aeromonas spp. in one patient. Finally, the Clostridium difficile toxin A was detected in three patients.

Molecular identification of Cryptosporidium species

PCR amplification of the 18S rDNA gene fragment identified 92 of the 155 (59%) samples as positive for Cryptosporidium spp. (Additional file 1: Table S1). There was a good correspondence between microscopy and PCR: 90 of the 104 (86.5%) microscopically positive samples were PCR positive, whereas only two of the 63 (4%) microscopically negative samples were PCR positive.

Sequencing of the PCR products identified C. hominis as the most prevalent species (n = 42), followed by C. meleagridis (n = 20), C. canis (n = 12), C. felis (n = 7), C. suis (n = 6) and C. parvum (n = 5) (Additional file 1: Table S1).

Genotyping of Cryptosporidium at the gp60 locus

All samples positive by 18S rDNA PCR were tested with different combination of primers to amplify fragments of the gp60 gene. No amplification was observed for samples classified as C. felis, C. canis and C. suis.

For C. hominis, PCR amplification was positive for 29 of the 42 isolates (69%), and sequencing revealed the presence of nine different subtypes in allelic families Ia, Ib, Id, Ie and If (Table 2). Subtype IeA11G3T3 was the most prevalent (n = 13), but allelic family Ia was the most diverse, with four subtypes. In particular, we identified subtypes IaA18R3, IaA19R3 and IaA20R3 in three, one and two isolates, respectively, as well as a new subtype (IaA16R3) in four isolates, which showed the highest similarity (98%) to an isolate from Ghana (KM538999). Finally, subtype IfA12G1 was detected in three isolates, whereas subtype IbA9G3, subtype IdA17 and a new subtype, IdA11, were each found in single isolates.

Table 2 Distribution of gp60 subtypes in C. hominis, C. meleagridis and C. parvum isolates from HIV-infected patients in Thailand
Full size table

For C. meleagridis, amplification yielded positive results for 14 of the 20 isolates (70%). Sequencing revealed the presence of allelic families IIIb (n = 13), IIIe (n = 2) and IIIg (n = 1) (Table 2). Subtypes were defined following a proposed nomenclature [16]. Accordingly, in allelic family IIIb, subtypes IIIbA19G1R1 (100% identity to KJ216111), IIIbA20G1R1 (100% identity to AB539720), IIIbA21G1R1b (100% identity to KJ210618), IIIbA22G1R1c (100% identity to KJ210614), IIIbA23G1R1b (100% identity to KJ210609), IIIbA23G1Rc (a new subtype with two nucleotide differences in the non-repeated region) and IIIbA24G1R1 (a new subtype with a single nucleotide difference in the non-repeated region), were identified. In allelic family IIIe, a new subtype, IIIeA22G1R1 (closely related to IIIeA21G2R1; KU852728), was found in two isolates. Finally, a new subtype in allelic family IIIg, IIIgA19G3R1, was found in a single isolate.

For C. parvum, amplification was successful for 4 out of 5 isolates, and sequencing revealed the presence of a single subtype, IIoA16G1 (100% identity to JN867335).

Cryptosporidium species and subtypes in multiple samples from the same patients

Additional stool samples (n = 11) collected at different time points from 10 patients were studied. Of the five C. hominis cases, four were confirmed at the species level, and three at the subtype (IeA11g3T3) level; the remaining samples were PCR negative. The single C. meleagridis case was confirmed at both the species and subtype (IIIbA20G1R1) level. The two cases involving C. canis were confirmed at the species level, whereas the two cases involving C. felis were PCR negative (Additional file 2: Table S2).

Discussion

This study initially aimed at determining the etiologic agents of diarrhea in HIV-infected people in Thailand. Microscopic analyses of stools (Table 1) showed a very high prevalence of single and mixed infections, which were sustained by a variety of enteric pathogens, in agreement with other studies [3, 5]. Infections due to Cryptosporidium spp. were particularly common, and this prompted a molecular study to identify the parasite species and subtypes and contribute to a better understanding of transmission routes.

The high prevalence of C. hominis (about 45%) is in agreement with previous studies conducted in this country [8,9,10,11,12] and, more generally, with findings from low income regions, where anthroponotic transmission prevails over zoonotic transmission [17,18,19]. Typing at the gp60 locus showed extensive genetic variability among C. hominis isolates, with nine subtypes belonging to allelic families Ia, Ib, Id, Ie and If, known to be associated with human cryptosporidiosis in low income countries [18, 20,21,22].

The other Cryptosporidium species detected in this study are zoonotic, and infect different animal hosts, including livestock (C. parvum), avian species (C. meleagridis), cats (C. felis), dogs (C. canis) and pigs (C. suis). The circulation of these parasite species among HIV-infected people in Thailand has been previously documented [8,9,10,11,12]. On the other hand, the few molecular studies on animal cryptosporidiosis identified C. canis in two dogs, C. felis in two cats, C. meleagridis in a pigeon [23] and C. parvum in dairy cows [10, 24] and water buffaloes [25]. The existence of zoonotic reservoirs is therefore plausible, yet human-to-human transmission of zoonotic species cannot be excluded. Subtyping data are needed to better understand the extent of zoonotic transmission.

The majority of the C. meleagridis subtypes identified in this study had been previously reported in patients who had travelled, immediately prior to the infection, to Thailand, Vietnam and/or India [16], or had acquired infection in Japan and China [26, 27]. Zoonotic transmission of C. meleagridis to humans has been demonstrated in several studies [16, 28], and specific gp60 subtypes were found to be shared by humans and birds [29], including the IIIbA22G1R1b and IIIbA22G1R1c subtypes identified in the present study.

Similar to C. meleagridis, the subtype IIoA16G1 identified in four C. parvum samples was previously described in a single patient from Sweden, who was reported to have traveled to Thailand before developing the symptoms [30]. This suggests that specific subtypes of different species play a more important role than others in defined regions of the world, underlying the complex epidemiology of cryptosporidiosis [18].

Mixed infections were found in about a quarter of the patients, and in one third of patients infected with Cryptosporidium. This is in line with findings from other studies [31], and should be taken into account when attributing severity and duration of diarrheal disease to specific pathogens.

This study has several limitations. First, the samples were collected at a time when antiretroviral therapies were not widely available for HIV-infected patients; it is possible that the relative prevalence of diarrhea-causing pathogens, including Cryptosporidium, has changed after the implementation of free-of-charge therapy for HIV-infected patients in Thailand. The retrospective nature did not allow the collection of further samples, or of clinical and epidemiologic data. It is therefore difficult to contextualize the results of this retrospective study and assessed rigorously for their current relevance. However, the temporal dynamics of cryptosporidiosis has not been carefully investigated, and genotyping data, even if from archived samples, are important. Secondly, a not negligible proportion (13.5%) of the samples was PCR negative, possibly due to a negative impact of the prolonged storage of stools on DNA integrity. Thirdly, and despite the use of combinations of published primers, no amplification of the gp60 gene could be obtained for C. felis, C. canis and C. suis isolates, which accounted for about one quarter of infections. Subtyping methods for these species should be developed.

Conclusions

This retrospective study highlights the complex epidemiology of Cryptosporidium infection in the highly vulnerable population of HIV-infected individuals. Extensive genetic heterogeneity was found among C. hominis and C. meleagridis isolates, with many subtypes previously identified in travelers returning from Thailand or other Asiatic countries.