Introduction

Cryptosporidium infections are increasingly recognised as a cause of diarrhoea in immunodeficient patients, elderly persons and children worldwide (Neill et al. 1996; Hunter and Nichols 2002; Huang et al. 2004). Cryptosporidiosis which is usually self-limiting in immunocompetent individuals can be severe and life threatening in immunocompromised ones, particularly in AIDS patients and in those with primary immunodeficiencies (PID) (Winkelstein et al. 2003; Cacciò et al. 2005; Thompson et al. 2005). Patients with secondary immunodeficiencies, i.e. those undergoing immunosuppressive treatment before or after transplantation or undergoing chemotherapy during cancer treatment, constitute another group at high risk of acute cryptosporidiosis (Campos et al. 2000; Abdo et al. 2003; Hong et al. 2007). Among different PID, patients with CD40 ligand deficiency (CD40L) and patients with combined immunodeficiency syndromes or interferon gamma deficiency are particularly prone to Cryptosporidium infection (Kocoshis et al. 1984; Levy et al. 1997; Wolska-Kuśnierz et al. 2007). Clearance of cryptosporidiosis is impaired in patients with genetic defects of cell-mediated immunity in whom colonisation of the biliary tract can occur predisposing to the development of sclerosing cholangitis, cirrhosis and cholangiocarcinoma (McLauchlin et al. 2003; Winkelstein et al. 2003).

Detection of Cryptosporidium infection in such patients by microscopy, using Ziehl–Neelsen staining of faecal smears and direct or indirect immunofluorescent assays (IFA), is often insufficient. Polymerase chain reaction (PCR)-based methods are more sensitive than microscopy in detecting parasites in patients with immunodeficiencies especially during asymptomatic periods (McLauchlin et al. 2003). Additionally, molecular studies are useful to understand the epidemiology of human cryptosporidiosis as they enable an objective and reliable identification of Cryptosporidium species/genotypes (Cacciò et al. 2005).

In Poland, Cryptosporidium detection is not routinely performed in clinical practise and only single cases of human cryptosporidiosis have been reported up to date (Siński et al. 1988; Golab et al. 2007; Wolska-Kuśnierz et al. 2007). Unrecognised Cryptosporidium infections in immunocompromised patients can be severe and even lead to death (i.e. after the exaggeration of infection following the haematopoietic stem cell transplantation (HSCT) in PID patients) as well as increase the risk of hospital outbreaks (O’Donoghoue 1995; McLauchlin et al. 2003). In this paper we report on the occurrence of Cryptosporidium and on the molecular analysis of isolates in three groups of patients: (1) immunocompetent children and adults with symptoms of gastroenteritis, (2) children with different PID and (3) patients with secondary immunodeficiencies, i.e. adults and children with chronic diarrhoeas undergoing immunosuppressive treatment.

Materials and methods

Clinical cases

During a 5-year period (2002–2007), 103 samples were collected from 35 patients in three hospitals: the Children’s Memorial Health Institute in Warsaw, the Children’s Hospital in Otwock and the Institute of Oncology in Warsaw. Patients were divided into three groups on the basis of their immunological status. The first group consisted of 13 immunocompetent children and one adult who suffered from recurrent abdominal pains and/or recurrent or chronic diarrhoea. The second group consisted of 17 children with different PID (six with CD40L deficiencies, six with severe combined immunodeficiencies (SCID), two with X-linked lymphoproliferative syndrome, one with primary lymphopenia CD40 and two with immunodeficiencies of unknown background). Patients in this group were either asymptomatic or showed signs of sclerosing cholangitis and hepatitis and diagnosis for Cryptosporidium was performed due to regular screening recommendation for PID. The third group consisted of patients with secondary immunodeficiencies and chronic diarrhoea and included two adults undergoing cancer treatment and two children after liver and/or kidney transplantation.

Faecal samples and, in one case of PID with sclerosing cholangitis and chronic hepatitis, bile and liver tissue samples were stored at +4°C for no longer than 5 days prior to DNA extraction.

Detection of Cryptosporidium oocysts by microscopy

Faecal or bile specimens from patients were collected once, twice or more times and examined for Cryptosporidium oocysts by a modified Ziehl–Neelsen acid-fast technique (Henriksen and Pohlenz 1981) and by IFA using a commercial kit (Merifluor Cryptosporidium/Giardia, Meridian Diagnostics, Cincinnati, OH, USA) and following the manufacturer’s instructions. Oocysts were concentrated prior to IFA using modified Sheather’s sucrose flotation (Bajer et al. 2002).

Molecular typing of Cryptosporidium spp.

Oocysts disruption and DNA purification were carried out using two methods: (1) extraction from whole faeces using the FastPrep 120 instrument and the FastDNA Spin Kit (Bio 101 Inc., Carlsbad, CA, USA) as described by da Silva et al. (1999) and (2) extraction from concentrated or unconcentrated samples using Genomic Mini AX Stool kit (A & A Biotechnology, Gdynia, Poland). Concentration was performed by modified Sheather’s sucrose flotation or by overnight decantation in phosphate-buffered saline (pH 7.2) at +4°C. Purified DNAs were stored at −20°C until analysis. The same methods were applied to the bile and liver biopsy specimens.

Amplification of an N-terminal fragment of the Cryptosporidium oocyst wall protein (COWP) gene was performed using the previously described nested PCR protocol (Spano et al. 1997; Pedraza-Diaz et al. 2001a). In the primary PCR, primers BCOWPF (5′-ACCGCTTCTCAACAACCATCTTGTCCTC-3′) and BCOWPR (5′-CGCACCTGTTCCCACTCAATGTAAACCC-3′) were used to produce a ∼769-bp fragment. In the nested PCR reaction, primers Cry15 (5′-GTAGATAATGGAAGAGATTGTG-3′) and Cry9 (5′-GGACTGAAATACAGGCATTATCTTG-3′) were used to produce a ∼550-bp fragment.

Amplification of a fragment of the beta-tubulin gene was performed using a nested PCR protocol. In the primary reaction, the full-length gene was amplified as described by Cacciò et al. (1999) whereas in the nested reaction the forward primer (5′-ACCAGATTGGTGCTAAATTC-3′) and the reverse primer (5′-CTGCAAAATACGATCTGGGA-3′) were used to amplify a ∼550-bp fragment.

Cryptosporidium species were first identified by a PCR–restriction fragment length polymorphism (RFLP) method (Spano et al. 1997; Xiao et al. 2000). The nested PCR COWP amplicons were digested with the endonuclease RsaI and the resulting fragments were separated by electrophoresis in 2% agarose gel.

COWP and beta-tubulin PCR products were purified and sequenced on both strands and sequences were edited and assembled using the software SeqMan 7.0. BLAST comparisons were run against the GenBank database (http://www.ncbi.nlm.nih.gov/BLAST).

Results

Prevalence of Cryptosporidium infection in three groups of patients

In the first group, six out of 14 (42.9%) immunocompetent patients were positive for Cryptosporidium on the basis of nested PCR, whereas Ziehl–Neelsen staining revealed no positive samples in this group.

In the second group, five of 17 children with PID were determined as Cryptosporidium positive (29.4%) on the basis of Ziehl–Neelsen staining of faecal smears and nested PCR. Nested PCR detected twice as many positive samples as Ziehl–Neelsen technique in specimens from infected individuals.

In the third group of four patients with secondary immunodeficiencies, only one Cryptosporidium infection was confirmed, solely on the basis of nested PCR, in an individual undergoing immunosuppressive chemotherapy.

Taken together, 12 Cryptosporidium infections were detected in 35 patients (34.3%). Cryptosporidium DNA was detected in the single bile sample but not in the liver tissue biopsy sample collected from a patient with PID who suffered from sclerosing cholangitis (Fig. 1).

Fig. 1
figure 1

Electrophoretic separation of COWP amplification product after digestion with the endonuclease RsaI. Lanes M molecular weight markers (50-bp ladder and marker V from Boerhinger), lane 1 C. parvum in immunocompetent child B.D. (group 1), lane 5 C. parvum in patient J.W., Institute of Oncology (group 3) (fragments of 413, 106 and 34 bp), lanes 2–4 C. meleagridis in patient E.N. with PID (fragments of 372, 147 and 34 bp), lanes 6 and 7 C. hominis in stool (6) and bile sample (7) from patient H.Ch. with PID (fragments of 284, 129, 106 and 34 bp)

Full size image

Distribution of Cryptosporidium species in the three groups of patients

All six Cryptosporidium isolates derived from immunocompetent patients (group 1) were Cryptosporidium parvum according to PCR–RFLP analysis and to sequence analysis of COWP products (Fig. 1). The homology to the Cryptosporidium parvum homologous sequence (GenBank accession number AF266273) ranged between 99.78% and 100%.

The same result was obtained for the isolate derived from the patient of the Institute of Oncology (group 3) where PCR–RFLP analysis and sequencing of COWP product showed the presence of C. parvum (100% homology to GenBank AF266273).

The highest diversity of Cryptosporidium species was observed in the second group, i.e. in children with PID. Among five positive cases, three different parasite species were identified by PCR–RFLP analysis and sequencing of COWP amplicons (Fig. 1). The prolonged infection in patient E.N., who had CD40L deficiency, was caused by Cryptosporidium meleagridis (100% homology with GenBank sequence AF266266). This species assignment was further confirmed by sequencing of a beta-tubulin gene fragment that revealed full homology with C. meleagridis isolates of human origin (GenBank accession numbers AF323574 and AF323575).

Cryptosporidium hominis was identified in stool and bile samples in patient H.Ch. who had primary lymphopenia CD40. Sequence analysis of the COWP fragment revealed 100% homology with GenBank sequence AF248741 (Fig. 1). Cryptosporidium DNA was not detected in liver specimen taken from this patient. In two other PID children, both with CD40L deficiency, PCR–RFLP and sequencing of the COWP fragment revealed the presence of C. parvum (with 99.79–100% homology with GenBank accession number AF266273). Of note, one of the patients was the brother of patient E.N. who was infected with C. meleagridis. Species identification was further confirmed by sequencing of beta-tubulin gene fragment that revealed 100% homology with C. parvum (GenBank accession number Y12615).

For the last patient in this group (patient M.W., with SCID and symptoms of cholestasis), it was impossible to determine the Cryptosporidium species involved.

Duration of Cryptosporidium infection and resolution of symptoms

In immunocompetent patients (group 1) and in those with secondary immunodeficiency (group 3), the parasite was detected only once during the sampling period. For the six children in group 1, the treatment was successful and all of them recovered. The course of C. parvum infection in the adult patient undergoing chemotherapy was not monitored.

In the group of children with PID (group 2), all five infections were chronic, lasting from 3 to 29 months despite applied treatment (Table 1). Two patients within this group died, one due to complication after HSCT (C. parvum infection) and the second (H.Ch.) due to B-cell lymphoma infiltration of the gall bladder.

Table 1 Duration of Cryptosporidium infection in children with PID
Full size table

C. meleagridis infection in patient E.N. was monitored both by microscopy and PCR and, despite different treatment methods, lasted for up to 29 months. After a third and successful HSCT, the patient recovered as confirmed by negative results of microscopy and nested PCR on the COWP gene. The other C. parvum-positive child also recovered after successful HSCT (Table 1).

Discussion

The aim of the present study was to investigate Cryptosporidium infection in three groups of patients at high risk including immunocompetent individuals with symptoms of cryptosporidiosis, children with PID and individuals with secondary immunodeficiencies and symptoms of gastroenteritis. Using two sensitive PCR-based methods, an overall prevalence of 34.3% of Cryptosporidium infection was observed in these patients. Of three groups of patients, the highest prevalence was found in children with symptoms of cryptosporidiosis and in those with PID and cholangitis (43% and 29%, respectively). One case of Cryptosporidium infection was also identified in a patient from the Institute of Oncology (group 3). Similar high rates of Cryptosporidium infection in immunocompetent children with chronic diarrhoea were confirmed recently in Poland (Golab et al. 2007) and in other countries worldwide (Ajjampur et al. 2007; Llorente et al. 2007; ten Hove et al. 2007; Sanad and Al-Malki 2007). High prevalence of Cryptosporidium infection in individuals with primary and secondary immunodeficiencies, exceeding even 82% in patients with AIDS and chronic diarrhoea, is commonly found in clinical practise (Winkelstein et al. 2003; Sanad and Al-Malki 2007).

Amplification and sequencing of COWP and beta-tubulin gene fragment revealed the presence of three Cryptosporidium species in the patients in this study. The most common was C. parvum which was detected in ten samples. All six positive children in group 1 were infected with this species. On the contrary, in other studies in immunocompetent children, the most common species (>80%) was C. hominis (Gatel et al. 2006; Ajjampur et al. 2007). However, C. parvum infection was detected only once during the sampling period and the children showed full recovery. A larger diversity of Cryptosporidium was observed in group 2, i.e. in children with different PID. In this group, C. parvum was detected in two cases and C. hominis and C. meleagridis were identified in the remaining two cases (Wolska-Kuśnierz et al. 2007). In other studies, up to eight different Cryptosporidium species/genotypes were identified in immunocompetent and immunodeficient children, including the mouse and cervine genotypes of C. parvum as well as C. meleagridis, Cryptosporidium felis, Cryptosporidium canis and Cryptosporidium muris (Pedraza-Diaz et al. 2001b; Xiao et al. 2001; Soba et al. 2006; Ajjampur et al. 2007; Gatel et al. 2007; Llorente et al. 2007). In contrast with the first group of patients, those within group 2 suffered from disseminated infection with symptoms of inflammation of the bile duct and gall bladder. Two patients in this group died, one due to complication after HSCT and the other one due to cholangiocarcinoma which was very likely caused by prolonged inflammation of bile ducts caused by cryptosporidiosis (Levy et al. 1997; Wolska-Kuśnierz et al. 2007). Successful amplification of Cryptosporidium DNA from a bile sample confirmed the extra-intestinal dissemination of C. hominis infection in this fatal case. Sclerosing cholangitis associated with Cryptosporidium infection is a well-known clinical manifestation found in PID patients, i.e. among patients with the X-linked hyper-IgM syndrome (Levy et al. 1997; Winkelstein et al. 2003). Severe and disseminated course of Cryptosporidium infections in PID was also previously described in UK (McLauchlin et al. 2003). HSCT is a method of choice in treatment of cryptosporidiosis in patients with PID, giving them opportunity for recovery even from an advanced disseminated infection (Dimicoli et al. 2003).

An adult patient, who was receiving immunosuppressive treatment at the Institute of Oncology, was infected with C. parvum and suffered from prolonged and severe diarrhoea.

In summary, the course of cryptosporidiosis was more severe among patients with PID resulting in disseminated infection and sclerosing cholangitis. Patients with secondary immunodeficiency suffered from acute cryptosporidiosis and among immunocompetent children mild infection was observed.

Chronic Cryptosporidium infections were monitored in children with PID and lasted from 3 to 29 months (Table 1), confirming their inability to clear the infection despite a range of treatment methods (Wolska-Kuśnierz et al. 2007). Cryptosporidium infections were cleared at last after successful HSCTs in two patients. One case of chronic infection was caused by C. meleagridis and thus represents, to the best of our knowledge, the first report of a long-lasting C. meleagridis infection in humans.

In agreement with a previous study (McLauchlin et al. 2003), our data have shown that microscopy is less sensitive than nested PCR in the detection of the parasite in asymptomatic cases. No positive cases were detected by Ziehl–Neelsen staining and IFA in immunocompetent children and in a patient from the Institute of Oncology (groups 1 and 3), probably due to the low intensity of infection. In children with PID, five were Cryptosporidium positive according to PCR assay and only four of these were also positive by microscopy. However, only half of all samples positive by PCR were also positive by microscopy. Therefore, molecular diagnostic tools should be considered the method of choice in screening possible Cryptosporidium infections of PID. Stool samples from such patients should be checked regularly, at least three times per year, even in the absence of diarrhoea or other symptoms.