Detection of Cyclospora cayetanensis Oocysts in Human Fecal Specimens by Flow Cytometry
http://www.100md.com
微生物临床杂志 2005年第5期
Microbiology Research Division, Bureau of Microbial Hazards, Food Directorate, Health Products and Food Branch, Health Canada
Animal Resources Division, Food Directorate, Health Products and Food Branch, Health Canada, Ottawa, Ontario, Canada
ABSTRACT
A diagnosis of cyclosporiasis typically involves stool examinations for the presence of Cyclospora oocysts by means of microscopy. In recent years, flow cytometry has been gaining in popularity as a novel method of detecting pathogens in environmental and clinical samples. The present study is an evaluation of a flow cytometric method for the detection and enumeration of Cyclospora oocysts in human fecal specimens associated with food-borne outbreaks of cyclosporiasis in Ontario, Canada. Flow cytometry results were generally very comparable to the original microscopy results for these specimens, in terms of both presence or absence of oocysts and relative oocyst concentrations. Of the 34 fecal specimens confirmed positive for Cyclospora by microscopy, 32 were also found positive by flow cytometry, and 2 others were considered equivocal. Of the eight fecal specimens reported to be negative by microscopy, two were found positive by flow cytometry and five others were considered equivocal. These two flow cytometry-positive samples and one of the equivocal samples were confirmed by microscopic reexamination, suggesting that flow cytometry may be more sensitive than microscopy. While the sample preparation time for flow cytometry is similar to or slightly longer than that for microscopy, the actual analysis time is much shorter. Further, because flow cytometry is largely automated, an analyst's levels of fatigue and expertise will not influence results. Flow cytometry appears to be a useful alternative to microscopy for the screening of large numbers of stool specimens for Cyclospora oocysts, such as in an outbreak situation.
INTRODUCTION
Cyclospora cayetanensis is a recently identified coccidian protozoan parasite of the human small intestine. While infections with C. cayetanensis have been reported worldwide, most of the prevalence data have come from studies done in Nepal, Haiti, Guatemala, and Peru, in which these infections are endemic.
The transmission stage of Cyclospora, known as an oocyst, is spherical and 8 to 10 μm in diameter. Unsporulated oocysts are shed into the environment with the feces of infected individuals. Oocysts undergo sporulation, becoming infective within a few days or weeks, depending upon environmental conditions. Infection may result when sporulated oocysts are ingested, primarily through contaminated water or food. The first outbreak of diarrhea due to infection with C. cayetanensis in North America was reported in Chicago, Ill., in 1990 and was associated with contaminated tap water (13). Numerous food-borne outbreaks and cases have also been reported in the United States and Canada in recent years (4, 5, 10, 11, 12). Approximately 90% of cyclosporiasis cases in the United States are thought to be food borne, amounting to an estimated 14,638 cases each year (14). Due to the required sporulation period outside of the host, person-to-person transmission of C. cayetanensis through the fecal-oral route is unlikely.
The actual infectious dose of Cyclospora oocysts is unknown, but based on data from outbreak investigations and extrapolations to other coccidians, it is generally thought to be relatively low. While the onset of symptoms has been reported to be abrupt in the majority of adult patients, the incubation period ranges from 1 to 11 days. Cyclosporiasis may result in profuse and prolonged diarrhea and a variety of other symptoms, including abdominal pain, nausea, vomiting, fatigue, fever, and loss of appetite. Illness is self-limiting but, untreated, may last several weeks. For most people, an infection with Cyclospora is treatable and does not result in life-threatening illness. Cyclosporiasis may be effectively treated with trimethoprim-sulfamethoxazole. There is some evidence that immunocompromised patients may develop more severe and prolonged symptoms.
Diagnosis usually involves microscopic examination of stool specimens for the presence of oocysts. Modified acid-fast methods have been widely used for diagnosis. However, C. cayetanensis oocysts stain variably with acid-fast methods, and this property has led to the use of other, more reliable stains, particularly safranin (20). Examination of wet mounts by bright-field, phase-contrast, or epifluorescence microscopy is becoming more widely used. Although monoclonal antibodies are not yet commercially available, epifluorescence microscopy may still be used, as C. cayetanensis oocysts are autofluorescent. Although there may be some variations in the intensity of autofluorescence observed among different laboratories, oocysts appear as bright blue circles when wet mounts are examined under UV light with an excitation filter within the range of 330 to 380 nm. Concentration of oocysts by sedimentation or flotation may improve detection by microscopic examination (8). Microscopy, however, is often a tedious and time-consuming procedure, potentially leading to analyst fatigue and false-negative results. In addition to microscopy, PCR has been found to be useful for diagnosis and detection (16). More recently, Varma et al. (19) developed a real-time PCR assay for the detection of Cyclospora oocysts.
In recent years, flow cytometry has been gaining in popularity as a novel method of detecting and enumerating Giardia cysts and Cryptosporidium oocysts in environmental and fecal samples (9, 15, 17). In recent studies, we found that flow cytometry was more sensitive than either conventional or immunofluorescence microscopy for the detection of Giardia sp. cysts in beaver fecal samples (6, 7). Similar findings have been reported for the detection of Cryptosporidium in SCID mice (1), seeded horse feces (3), and seeded human stool specimens (18). Further, because flow cytometry is an automated method, an analyst's levels of fatigue and expertise will not influence results as they may with microscopy. In addition to detection and enumeration, large-scale sorting could also be used in conjunction with flow cytometry to yield partially purified oocysts for research purposes, such as food-spiking and recovery experiments, viability determination, or molecular characterization. The present study involves an evaluation of the effectiveness of flow cytometry for the detection and enumeration of C. cayetanensis oocysts in human fecal specimens.
MATERIALS AND METHODS
Sample collection and microscopy. Thirty-four Cyclospora-positive human fecal specimens fixed in SAF (sodium acetate, 1.5% [wt/vol]; acetic acid, 2% [vol/vol]; formaldehyde, 4% [vol/vol]) were obtained from the Ontario Ministry of Health and Long-Term Care. These specimens were collected from individuals meeting the case definition during food-borne outbreaks of cyclosporiasis in 1998 (31 specimens) and 1999 (3 specimens) in Ontario. Eight Cyclospora-negative human fecal specimens collected in 1999 and associated with the same outbreak were similarly obtained. All of these fecal samples were initially examined for ova and parasites by the Parasitology Laboratory of the Ontario Ministry of Health and Long-Term Care. Auramine-rhodamine (acid-fast) staining was used as a screening procedure for all coccidian parasites, including Cyclospora. As the fluorescence of Cyclospora oocysts was generally not very intense with this method, their presence was confirmed by examination of wet mounts from formalin-ether sedimentation and by examination of smears stained with iron-hematoxylin (Ontario Ministry of Health and Long-Term Care, unpublished method). Iron-hematoxylin was used as a negative stain; Cyclospora oocysts appeared as 9- to 10-μm silvery balls which were readily apparent against a blue background (B. Yu, personal communication).
Sample preparation for flow cytometry. After receipt of the SAF-fixed fecal samples, Cyclospora oocysts were concentrated by using a sucrose gradient flotation procedure based on the method of Chaput (2), with the modification that only one milliliter of the suspended fecal sample was carefully added to the top of the gradient before centrifugation. The final pellet was resuspended in the small volume of phosphate-buffered saline remaining in the tube after aspiration of the supernatant, transferred to a 1.5-ml microcentrifuge tube, and centrifuged on an IEC Micromax for 10 min at 16,750 x g. The supernatant was aspirated, and the pellet was resuspended in 500 μl cold phosphate-buffered saline, transferred to a 5-ml round-bottom tube (Falcon; Becton Dickinson), and stored in the dark at 4°C until analyzed by flow cytometry. In the absence of commercially available anti-Cyclospora monoclonal antibodies, this flow cytometry method was based on the autofluorescence of Cyclospora oocysts. For each batch of samples prepared and analyzed by flow cytometry, one strongly positive sample (confirmed by microscopy) was similarly prepared and used as a control.
Flow cytometry. Samples were analyzed on a FACScan (Becton Dickinson, Mississauga, Ontario, Canada) equipped with an argon-ion laser operating at 488 nm and CellQuest software. The flow cytometer was calibrated by using CaliBRITE beads (Becton Dickinson, Mississauga, Ontario, Canada) with FACSComp software and QC3 beads (fluorescence reference standards) according to the manufacturer's recommendations (Bangs Laboratories, Inc., Fishers, IN). All samples were vortexed before and during acquisition on the flow cytometer. The entire volume (approximately 0.5 ml), or a maximum of 15,000 events, was analyzed for each sample. A strongly Cyclospora-positive sample was used to define the analysis gate (G8) as follows. Region 1 (R1) was set around the oocysts on a dual-parameter dot plot (right-angle light scatter versus autofluorescence). Backgating on a dual-parameter dot plot (forward light scatter versus right-angle light scatter), region 2 (R2) was set around the oocysts. Forward light scatter is a function of size, while right-angle light scatter is a function of complexity. Complexity refers to the internal granularity and surface roughness, whereby the amount of light deflected increases with complexity. Using a double-anchor gating strategy (G8 = R1 and R2), a dual-parameter dot plot (right-angle light scatter versus autofluorescence) gated on G8 was generated and displayed the number of oocysts in the analyzed sample. Samples having only 1 to 10 events in the analysis gate were considered equivocal in terms of the presence or absence of Cyclospora oocysts, while those with greater than 10 gated events were considered positive. While each fecal specimen was analyzed at least once, some specimens, including microscopy-negative samples, strongly positive samples used as controls, and some equivocal samples, were processed and analyzed twice. In some instances, the same tubes were analyzed more than once, and acquisition was stopped at fewer than 15,000 events in order to determine the influence of the number of events acquired on the total number of oocysts.
Confirmation of flow cytometry results. The presence of oocysts in a sample which was found to be strongly positive for Cyclospora by flow cytometry was confirmed by using a FACSCalibur system equipped with the sorting option (Becton Dickinson, Mississauga, Ontario, Canada). The analysis gate (G8) was used as the sort gate, and collection tubes were centrifuged for 30 min at approximately 10,000 x g. The supernatants were aspirated, and the pellets were pooled. One drop of the pooled suspension was transferred to each of two microscope slides by using a disposable transfer pipette. Coverslips were added, and the two slides were examined at a magnification of x400 on a Zeiss Axiophot epifluorescence microscope equipped with a UV filter system.
Samples which were originally reported to be negative by microscopy and which were subsequently found to be positive or equivocal by flow cytometry were reexamined by epifluorescence and differential interference contrast imaging at a magnification of x600 on a Nikon Eclipse E600 microscope.
RESULTS
Using flow cytometry, oocysts could be separated according to their autofluorescence, size, and complexity, and a cluster representing Cyclospora oocysts could be clearly observed on the dot plots of positive samples. Dual-parameter dot plots of representative Cyclospora-positive (sample no. 5) and Cyclospora-negative (sample no. 41) samples are shown in Fig. 1.
There was a close correlation between the original microscopy results and those obtained by flow cytometry in the present study, in terms of the presence or absence of oocysts (Table 1). Of the 34 fecal specimens confirmed positive for Cyclospora by microscopy, 32 were also found positive by flow cytometry. Two specimens (sample no. 27 and 34) were considered equivocal in terms of the presence or absence of Cyclospora oocysts by flow cytometry, since the number of gated events in both was less than 10. Similarly, both of these specimens were reported as low-positive specimens (1+) by microscopy. The relative concentrations of oocysts in the samples, as determined by microscopy and flow cytometry, also showed good agreement, with the 3+ or higher samples generally showing a large number of gated events in the flow cytometry analysis (range, 36 to 13,089). Further, the four samples designated rare by microscopy also had a relatively small number of events in the analysis gate (range, 16 to 40). Of the eight fecal specimens reported negative by microscopy, two were found positive by flow cytometry (sample no. 38 and 39), five were considered equivocal, since they had less than 10 events in G8, and one was considered negative. Microscopic reexamination confirmed the presence of Cyclospora oocysts in sample no. 38 and 39 and in one of the equivocal samples (sample no. 35). Samples run in duplicate generally demonstrated relatively small standard deviations in the mean numbers of oocysts. In addition, whenever a sample analysis was repeated, with an acquisition of fewer than 15,000 events, the number of oocysts was proportional to the original analysis (results not shown).
The identification of events in the analysis gate (G8) as C. cayetanensis oocysts was confirmed by sorting, followed by microscopy. Numerous Cyclospora oocysts were observed on each of the two microscope slides prepared from the sorted events. Oocysts appeared as bright blue circles of the appropriate size and shape under UV light. Bright-field microscopy revealed a granular internal appearance for the oocysts.
DISCUSSION
While there was a close correlation between the flow cytometry results obtained in the present study and the original microscopy results, in terms of presence or absence of oocysts, the sensitivity of flow cytometry appeared to be higher, as two of the samples determined to be negative by microscopy were found to be positive by flow cytometry and five others were equivocal. Although clinical histories were not available, all samples determined to be negative by microscopy were collected from individuals involved in the Cyclospora outbreaks; therefore, it was feasible that some of them may have represented false-negative results. In fact, the two flow cytometry-positive samples mentioned above and one equivocal sample were subsequently confirmed by microscopic reexamination. The higher sensitivity of flow cytometry is likely due largely to the considerably larger sample volume analyzed (up to 0.5 ml). There was also a good correlation between flow cytometry and microscopy in terms of relative oocyst concentrations. The clinical significance of oocyst concentrations is questionable, however, since it is as yet unclear whether Cyclospora oocysts are shed continuously or intermittently in infected individuals.
Samples having only 1 to 10 events in the analysis gate were considered equivocal in terms of the presence or absence of Cyclospora oocysts. This high degree of stringency was required in the interpretation of data due to the fact that no specific monoclonal antibodies against Cyclospora oocysts are currently available. Instead, the natural autofluorescence of the oocysts was used in the gating strategy. Specifically, a double-anchor gating strategy was used: events were analyzed for right-angle light scatter, autofluorescence intensity, and forward light scatter. Each of these three intrinsic parameters must be satisfied in order for an event to be included in the analysis gate. The specificity of the gating analysis was confirmed in the present study through a combination of sorting and microscopy. Nevertheless, it was possible that a small proportion of debris in the fecal suspensions could exhibit characteristics of size, internal complexity, and autofluorescence similar to those of Cyclospora oocysts and could result in false-positive flow cytometry results. In addition, electronic artifacts might occasionally be observed on flow cytometry dot plots. The stringency used in the number of gated events required for a positive result therefore allowed for a greater degree of confidence in reporting positive results in the present study. More work is required to further improve the specificity of this flow cytometric method. Due to the small numbers of gated events (oocysts) present, not all samples reported as equivocal in Table 1 could be confirmed microscopically; therefore, these might be considered presumptive positive only. Alternative methods of detection, particularly PCR, may be useful in confirming equivocal samples such as these.
Samples which were processed and run in duplicate demonstrated relatively small standard deviations in the mean numbers of oocysts (Table 1), indicating good reproducibility of results with this method. Further, whenever a sample analysis was repeated, with an acquisition of fewer than 15,000 events, the numbers of oocysts were always proportional to the original analysis, suggesting that meaningful results can be achieved with fewer acquired events if necessary (i.e., very dilute fecal samples).
While the sample preparation time for flow cytometry may be similar to or even longer than that for microscopy, depending upon the concentration and staining procedures used, the time it takes to analyze a sample by flow cytometry is considerably shorter than the time it takes to analyze a sample by microscopy. Sample analysis took only minutes, whereas microscopic examination is often a very time-consuming procedure. As a result, a larger number of samples could be analyzed by flow cytometry in a relatively short period of time. More importantly, as the method is largely automated, the results are not influenced by an analyst's levels of fatigue and expertise, as they may be with microscopy. While stool specimens are generally not examined for Cyclospora oocysts unless specifically requested, the results of the present study suggest that flow cytometry may be a useful alternative to microscopy in the screening of large numbers of fecal specimens for the presence of Cyclospora oocysts. During an outbreak situation, for example, when the number of requested tests increases dramatically, the benefits of flow cytometry as a detection method may become even more apparent. While the cost of a flow cytometer may be somewhat prohibitive to smaller diagnostic laboratories, flow cytometry core facilities are often available at larger diagnostic laboratories, particularly those that do hematology work.
ACKNOWLEDGMENTS
We extend our sincere thanks to Billy Yu, Parasitology Department, Laboratories Branch, Ontario Ministry of Health and Long-Term Care, for providing the stool specimens and microscopy results used in the present study. We also thank Michele Bergeron and Nadia Soucy, National HIV Immunology Laboratory, Health Canada, Ottawa, for performing the cell sorting.
REFERENCES
Arrowood, M. J., M. R. Hurd, and J. R. Mead. 1995. A new method for evaluating experimental cryptosporidial parasite loads using immunofluorescent flow cytometry. J. Parasitol. 81:404-409.
Chaput, M. P. 1999. Recovery and purification of Cyclospora cayetanensis oocysts using sucrose gradients. FDA laboratory information bulletin, vol. 15, no. 4, p. 4163. U.S. Food and Drug Administration, Washington, D.C.
Cole, D. J., K. Snowden, N. D. Cohen, and R. Smith. 1999. Detection of Cryptosporidium parvum in horses: thresholds of acid-fast stain, immunofluorescence assay, and flow cytometry. J. Clin. Microbiol. 37:457-460.
DeGraw, E., S. Heber, and A. Rowan. 1997. Update: outbreaks of cyclosporiasis—United States and Canada, 1997. Morb. Mortal. Wkly. Rep. 46:521-523.
Dixon, B. R. 2003. The prevalence and control of foodborne protozoan parasites, p. 31-76. In B. W. Blais (ed.), Current challenges in food microbiology. Research Signpost, Kerala, India.
Dixon, B. R., M. Parenteau, C. Martineau, and J. Fournier. 1997. A comparison of conventional microscopy, immunofluorescence microscopy and flow cytometry in the detection of Giardia lamblia cysts in beaver fecal samples. J. Immunol. Methods 202:27-33.
Dixon, B. R., J. Bussey, L. Parrington, M. Parenteau, R. Moore, J. Jacob, M.-P. Parenteau, and J. Fournier. 2002. A preliminary estimate of the prevalence of Giardia sp. in beavers in Gatineau Park, Quebec, using flow cytometry, p. 71-79. In B. E. Olson, M. E. Olson, and P. M. Wallis (ed.), Giardia: the cosmopolitan parasite. CAB International, Wallingford, United Kingdom.
Eberhard, M. L., N. J. Pieniazek, and M. J. Arrowood. 1997. Laboratory diagnosis of Cyclospora infections. Arch. Pathol. Lab. Med. 121:792-797.
Ferrari, B. C., and D. Veal. 2003. Analysis-only detection of Giardia by combining immunomagnetic separation and two-color flow cytometry. Cytometry 51A:79-86.
Herwaldt, B. L. 2000. Cyclospora cayetanensis: a review, focusing on the outbreaks of cyclosporiasis in the 1990s. Clin. Infect. Dis. 31:1040-1057.
Herwaldt, B. L., M.-L. Ackers, et al. 1997. An outbreak in 1996 of cyclosporiasis associated with imported raspberries. N. Engl. J. Med. 336:1548-1556.
Herwaldt, B. L., M. J. Beach, et al. 1999. The return of Cyclospora in 1997: another outbreak of cyclosporiasis in North America associated with imported raspberries. Ann. Intern. Med. 130:210-220.
Huang, P., J. T. Weber, D. M. Sosin, P. M. Griffin, E. G. Long, J. J. Murphy, F. Kocka, C. Peters, and C. Kallick. 1995. The first reported outbreak of diarrheal illness associated with Cyclospora in the United States. Ann. Intern. Med. 123:409-414.
Mead, P. S., L. Slutsker, V. Dietz, L. F. McCaig, J. S. Bresee, C. Shapiro, P. M. Griffin, and R. V. Tauxe. 1999. Food-related illness and death in the United States. Emerg. Infect. Dis. 5:607-625.
Moss, D. M., and M. J. Arrowood. 2001. Quantification of Cryptosporidium parvum oocysts in mouse fecal specimens using immunomagnetic particles and two-color flow cytometry. J. Parasitol. 87:406-412.
Orlandi, P. A., and K. A. Lampel. 2000. Extraction-free, filter-based template preparation for rapid and sensitive PCR detection of pathogenic parasitic protozoa. J. Clin. Microbiol. 38:2271-2277.
Power, M. L., S. R. Shanker, N. C. Sangster, and D. A. Veal. 2003. Evaluation of a combined immunomagnetic separation/flow cytometry technique for epidemiological investigations of Cryptosporidium in domestic and Australian native animals. Vet. Parasitol. 112:21-31.
Valdez, L. M., H. Dang, P. C. Okhuysen, and C. L. Chappell. 1997. Flow cytometric detection of Cryptosporidium oocysts in human stool samples. J. Clin. Microbiol. 35:2013-2017.
Varma, M., J. D. Hester, F. W. Schaefer, M. W. Ware, and H. D. A. Lindquist. 2003. Detection of Cyclospora cayetanensis using a quantitative real-time PCR assay. J. Microbiol. Methods 53:27-36.
Visvesvara, G. S., H. Moura, E. Kovacsnace, S. Wallace, and M. L. Eberhard. 1997. Uniform staining of Cyclospora oocysts in fecal smears by a modified safranin technique with microwave heating. J. Clin. Microbiol. 35:730-733.(Brent R. Dixon, Jeff M. B)
Animal Resources Division, Food Directorate, Health Products and Food Branch, Health Canada, Ottawa, Ontario, Canada
ABSTRACT
A diagnosis of cyclosporiasis typically involves stool examinations for the presence of Cyclospora oocysts by means of microscopy. In recent years, flow cytometry has been gaining in popularity as a novel method of detecting pathogens in environmental and clinical samples. The present study is an evaluation of a flow cytometric method for the detection and enumeration of Cyclospora oocysts in human fecal specimens associated with food-borne outbreaks of cyclosporiasis in Ontario, Canada. Flow cytometry results were generally very comparable to the original microscopy results for these specimens, in terms of both presence or absence of oocysts and relative oocyst concentrations. Of the 34 fecal specimens confirmed positive for Cyclospora by microscopy, 32 were also found positive by flow cytometry, and 2 others were considered equivocal. Of the eight fecal specimens reported to be negative by microscopy, two were found positive by flow cytometry and five others were considered equivocal. These two flow cytometry-positive samples and one of the equivocal samples were confirmed by microscopic reexamination, suggesting that flow cytometry may be more sensitive than microscopy. While the sample preparation time for flow cytometry is similar to or slightly longer than that for microscopy, the actual analysis time is much shorter. Further, because flow cytometry is largely automated, an analyst's levels of fatigue and expertise will not influence results. Flow cytometry appears to be a useful alternative to microscopy for the screening of large numbers of stool specimens for Cyclospora oocysts, such as in an outbreak situation.
INTRODUCTION
Cyclospora cayetanensis is a recently identified coccidian protozoan parasite of the human small intestine. While infections with C. cayetanensis have been reported worldwide, most of the prevalence data have come from studies done in Nepal, Haiti, Guatemala, and Peru, in which these infections are endemic.
The transmission stage of Cyclospora, known as an oocyst, is spherical and 8 to 10 μm in diameter. Unsporulated oocysts are shed into the environment with the feces of infected individuals. Oocysts undergo sporulation, becoming infective within a few days or weeks, depending upon environmental conditions. Infection may result when sporulated oocysts are ingested, primarily through contaminated water or food. The first outbreak of diarrhea due to infection with C. cayetanensis in North America was reported in Chicago, Ill., in 1990 and was associated with contaminated tap water (13). Numerous food-borne outbreaks and cases have also been reported in the United States and Canada in recent years (4, 5, 10, 11, 12). Approximately 90% of cyclosporiasis cases in the United States are thought to be food borne, amounting to an estimated 14,638 cases each year (14). Due to the required sporulation period outside of the host, person-to-person transmission of C. cayetanensis through the fecal-oral route is unlikely.
The actual infectious dose of Cyclospora oocysts is unknown, but based on data from outbreak investigations and extrapolations to other coccidians, it is generally thought to be relatively low. While the onset of symptoms has been reported to be abrupt in the majority of adult patients, the incubation period ranges from 1 to 11 days. Cyclosporiasis may result in profuse and prolonged diarrhea and a variety of other symptoms, including abdominal pain, nausea, vomiting, fatigue, fever, and loss of appetite. Illness is self-limiting but, untreated, may last several weeks. For most people, an infection with Cyclospora is treatable and does not result in life-threatening illness. Cyclosporiasis may be effectively treated with trimethoprim-sulfamethoxazole. There is some evidence that immunocompromised patients may develop more severe and prolonged symptoms.
Diagnosis usually involves microscopic examination of stool specimens for the presence of oocysts. Modified acid-fast methods have been widely used for diagnosis. However, C. cayetanensis oocysts stain variably with acid-fast methods, and this property has led to the use of other, more reliable stains, particularly safranin (20). Examination of wet mounts by bright-field, phase-contrast, or epifluorescence microscopy is becoming more widely used. Although monoclonal antibodies are not yet commercially available, epifluorescence microscopy may still be used, as C. cayetanensis oocysts are autofluorescent. Although there may be some variations in the intensity of autofluorescence observed among different laboratories, oocysts appear as bright blue circles when wet mounts are examined under UV light with an excitation filter within the range of 330 to 380 nm. Concentration of oocysts by sedimentation or flotation may improve detection by microscopic examination (8). Microscopy, however, is often a tedious and time-consuming procedure, potentially leading to analyst fatigue and false-negative results. In addition to microscopy, PCR has been found to be useful for diagnosis and detection (16). More recently, Varma et al. (19) developed a real-time PCR assay for the detection of Cyclospora oocysts.
In recent years, flow cytometry has been gaining in popularity as a novel method of detecting and enumerating Giardia cysts and Cryptosporidium oocysts in environmental and fecal samples (9, 15, 17). In recent studies, we found that flow cytometry was more sensitive than either conventional or immunofluorescence microscopy for the detection of Giardia sp. cysts in beaver fecal samples (6, 7). Similar findings have been reported for the detection of Cryptosporidium in SCID mice (1), seeded horse feces (3), and seeded human stool specimens (18). Further, because flow cytometry is an automated method, an analyst's levels of fatigue and expertise will not influence results as they may with microscopy. In addition to detection and enumeration, large-scale sorting could also be used in conjunction with flow cytometry to yield partially purified oocysts for research purposes, such as food-spiking and recovery experiments, viability determination, or molecular characterization. The present study involves an evaluation of the effectiveness of flow cytometry for the detection and enumeration of C. cayetanensis oocysts in human fecal specimens.
MATERIALS AND METHODS
Sample collection and microscopy. Thirty-four Cyclospora-positive human fecal specimens fixed in SAF (sodium acetate, 1.5% [wt/vol]; acetic acid, 2% [vol/vol]; formaldehyde, 4% [vol/vol]) were obtained from the Ontario Ministry of Health and Long-Term Care. These specimens were collected from individuals meeting the case definition during food-borne outbreaks of cyclosporiasis in 1998 (31 specimens) and 1999 (3 specimens) in Ontario. Eight Cyclospora-negative human fecal specimens collected in 1999 and associated with the same outbreak were similarly obtained. All of these fecal samples were initially examined for ova and parasites by the Parasitology Laboratory of the Ontario Ministry of Health and Long-Term Care. Auramine-rhodamine (acid-fast) staining was used as a screening procedure for all coccidian parasites, including Cyclospora. As the fluorescence of Cyclospora oocysts was generally not very intense with this method, their presence was confirmed by examination of wet mounts from formalin-ether sedimentation and by examination of smears stained with iron-hematoxylin (Ontario Ministry of Health and Long-Term Care, unpublished method). Iron-hematoxylin was used as a negative stain; Cyclospora oocysts appeared as 9- to 10-μm silvery balls which were readily apparent against a blue background (B. Yu, personal communication).
Sample preparation for flow cytometry. After receipt of the SAF-fixed fecal samples, Cyclospora oocysts were concentrated by using a sucrose gradient flotation procedure based on the method of Chaput (2), with the modification that only one milliliter of the suspended fecal sample was carefully added to the top of the gradient before centrifugation. The final pellet was resuspended in the small volume of phosphate-buffered saline remaining in the tube after aspiration of the supernatant, transferred to a 1.5-ml microcentrifuge tube, and centrifuged on an IEC Micromax for 10 min at 16,750 x g. The supernatant was aspirated, and the pellet was resuspended in 500 μl cold phosphate-buffered saline, transferred to a 5-ml round-bottom tube (Falcon; Becton Dickinson), and stored in the dark at 4°C until analyzed by flow cytometry. In the absence of commercially available anti-Cyclospora monoclonal antibodies, this flow cytometry method was based on the autofluorescence of Cyclospora oocysts. For each batch of samples prepared and analyzed by flow cytometry, one strongly positive sample (confirmed by microscopy) was similarly prepared and used as a control.
Flow cytometry. Samples were analyzed on a FACScan (Becton Dickinson, Mississauga, Ontario, Canada) equipped with an argon-ion laser operating at 488 nm and CellQuest software. The flow cytometer was calibrated by using CaliBRITE beads (Becton Dickinson, Mississauga, Ontario, Canada) with FACSComp software and QC3 beads (fluorescence reference standards) according to the manufacturer's recommendations (Bangs Laboratories, Inc., Fishers, IN). All samples were vortexed before and during acquisition on the flow cytometer. The entire volume (approximately 0.5 ml), or a maximum of 15,000 events, was analyzed for each sample. A strongly Cyclospora-positive sample was used to define the analysis gate (G8) as follows. Region 1 (R1) was set around the oocysts on a dual-parameter dot plot (right-angle light scatter versus autofluorescence). Backgating on a dual-parameter dot plot (forward light scatter versus right-angle light scatter), region 2 (R2) was set around the oocysts. Forward light scatter is a function of size, while right-angle light scatter is a function of complexity. Complexity refers to the internal granularity and surface roughness, whereby the amount of light deflected increases with complexity. Using a double-anchor gating strategy (G8 = R1 and R2), a dual-parameter dot plot (right-angle light scatter versus autofluorescence) gated on G8 was generated and displayed the number of oocysts in the analyzed sample. Samples having only 1 to 10 events in the analysis gate were considered equivocal in terms of the presence or absence of Cyclospora oocysts, while those with greater than 10 gated events were considered positive. While each fecal specimen was analyzed at least once, some specimens, including microscopy-negative samples, strongly positive samples used as controls, and some equivocal samples, were processed and analyzed twice. In some instances, the same tubes were analyzed more than once, and acquisition was stopped at fewer than 15,000 events in order to determine the influence of the number of events acquired on the total number of oocysts.
Confirmation of flow cytometry results. The presence of oocysts in a sample which was found to be strongly positive for Cyclospora by flow cytometry was confirmed by using a FACSCalibur system equipped with the sorting option (Becton Dickinson, Mississauga, Ontario, Canada). The analysis gate (G8) was used as the sort gate, and collection tubes were centrifuged for 30 min at approximately 10,000 x g. The supernatants were aspirated, and the pellets were pooled. One drop of the pooled suspension was transferred to each of two microscope slides by using a disposable transfer pipette. Coverslips were added, and the two slides were examined at a magnification of x400 on a Zeiss Axiophot epifluorescence microscope equipped with a UV filter system.
Samples which were originally reported to be negative by microscopy and which were subsequently found to be positive or equivocal by flow cytometry were reexamined by epifluorescence and differential interference contrast imaging at a magnification of x600 on a Nikon Eclipse E600 microscope.
RESULTS
Using flow cytometry, oocysts could be separated according to their autofluorescence, size, and complexity, and a cluster representing Cyclospora oocysts could be clearly observed on the dot plots of positive samples. Dual-parameter dot plots of representative Cyclospora-positive (sample no. 5) and Cyclospora-negative (sample no. 41) samples are shown in Fig. 1.
There was a close correlation between the original microscopy results and those obtained by flow cytometry in the present study, in terms of the presence or absence of oocysts (Table 1). Of the 34 fecal specimens confirmed positive for Cyclospora by microscopy, 32 were also found positive by flow cytometry. Two specimens (sample no. 27 and 34) were considered equivocal in terms of the presence or absence of Cyclospora oocysts by flow cytometry, since the number of gated events in both was less than 10. Similarly, both of these specimens were reported as low-positive specimens (1+) by microscopy. The relative concentrations of oocysts in the samples, as determined by microscopy and flow cytometry, also showed good agreement, with the 3+ or higher samples generally showing a large number of gated events in the flow cytometry analysis (range, 36 to 13,089). Further, the four samples designated rare by microscopy also had a relatively small number of events in the analysis gate (range, 16 to 40). Of the eight fecal specimens reported negative by microscopy, two were found positive by flow cytometry (sample no. 38 and 39), five were considered equivocal, since they had less than 10 events in G8, and one was considered negative. Microscopic reexamination confirmed the presence of Cyclospora oocysts in sample no. 38 and 39 and in one of the equivocal samples (sample no. 35). Samples run in duplicate generally demonstrated relatively small standard deviations in the mean numbers of oocysts. In addition, whenever a sample analysis was repeated, with an acquisition of fewer than 15,000 events, the number of oocysts was proportional to the original analysis (results not shown).
The identification of events in the analysis gate (G8) as C. cayetanensis oocysts was confirmed by sorting, followed by microscopy. Numerous Cyclospora oocysts were observed on each of the two microscope slides prepared from the sorted events. Oocysts appeared as bright blue circles of the appropriate size and shape under UV light. Bright-field microscopy revealed a granular internal appearance for the oocysts.
DISCUSSION
While there was a close correlation between the flow cytometry results obtained in the present study and the original microscopy results, in terms of presence or absence of oocysts, the sensitivity of flow cytometry appeared to be higher, as two of the samples determined to be negative by microscopy were found to be positive by flow cytometry and five others were equivocal. Although clinical histories were not available, all samples determined to be negative by microscopy were collected from individuals involved in the Cyclospora outbreaks; therefore, it was feasible that some of them may have represented false-negative results. In fact, the two flow cytometry-positive samples mentioned above and one equivocal sample were subsequently confirmed by microscopic reexamination. The higher sensitivity of flow cytometry is likely due largely to the considerably larger sample volume analyzed (up to 0.5 ml). There was also a good correlation between flow cytometry and microscopy in terms of relative oocyst concentrations. The clinical significance of oocyst concentrations is questionable, however, since it is as yet unclear whether Cyclospora oocysts are shed continuously or intermittently in infected individuals.
Samples having only 1 to 10 events in the analysis gate were considered equivocal in terms of the presence or absence of Cyclospora oocysts. This high degree of stringency was required in the interpretation of data due to the fact that no specific monoclonal antibodies against Cyclospora oocysts are currently available. Instead, the natural autofluorescence of the oocysts was used in the gating strategy. Specifically, a double-anchor gating strategy was used: events were analyzed for right-angle light scatter, autofluorescence intensity, and forward light scatter. Each of these three intrinsic parameters must be satisfied in order for an event to be included in the analysis gate. The specificity of the gating analysis was confirmed in the present study through a combination of sorting and microscopy. Nevertheless, it was possible that a small proportion of debris in the fecal suspensions could exhibit characteristics of size, internal complexity, and autofluorescence similar to those of Cyclospora oocysts and could result in false-positive flow cytometry results. In addition, electronic artifacts might occasionally be observed on flow cytometry dot plots. The stringency used in the number of gated events required for a positive result therefore allowed for a greater degree of confidence in reporting positive results in the present study. More work is required to further improve the specificity of this flow cytometric method. Due to the small numbers of gated events (oocysts) present, not all samples reported as equivocal in Table 1 could be confirmed microscopically; therefore, these might be considered presumptive positive only. Alternative methods of detection, particularly PCR, may be useful in confirming equivocal samples such as these.
Samples which were processed and run in duplicate demonstrated relatively small standard deviations in the mean numbers of oocysts (Table 1), indicating good reproducibility of results with this method. Further, whenever a sample analysis was repeated, with an acquisition of fewer than 15,000 events, the numbers of oocysts were always proportional to the original analysis, suggesting that meaningful results can be achieved with fewer acquired events if necessary (i.e., very dilute fecal samples).
While the sample preparation time for flow cytometry may be similar to or even longer than that for microscopy, depending upon the concentration and staining procedures used, the time it takes to analyze a sample by flow cytometry is considerably shorter than the time it takes to analyze a sample by microscopy. Sample analysis took only minutes, whereas microscopic examination is often a very time-consuming procedure. As a result, a larger number of samples could be analyzed by flow cytometry in a relatively short period of time. More importantly, as the method is largely automated, the results are not influenced by an analyst's levels of fatigue and expertise, as they may be with microscopy. While stool specimens are generally not examined for Cyclospora oocysts unless specifically requested, the results of the present study suggest that flow cytometry may be a useful alternative to microscopy in the screening of large numbers of fecal specimens for the presence of Cyclospora oocysts. During an outbreak situation, for example, when the number of requested tests increases dramatically, the benefits of flow cytometry as a detection method may become even more apparent. While the cost of a flow cytometer may be somewhat prohibitive to smaller diagnostic laboratories, flow cytometry core facilities are often available at larger diagnostic laboratories, particularly those that do hematology work.
ACKNOWLEDGMENTS
We extend our sincere thanks to Billy Yu, Parasitology Department, Laboratories Branch, Ontario Ministry of Health and Long-Term Care, for providing the stool specimens and microscopy results used in the present study. We also thank Michele Bergeron and Nadia Soucy, National HIV Immunology Laboratory, Health Canada, Ottawa, for performing the cell sorting.
REFERENCES
Arrowood, M. J., M. R. Hurd, and J. R. Mead. 1995. A new method for evaluating experimental cryptosporidial parasite loads using immunofluorescent flow cytometry. J. Parasitol. 81:404-409.
Chaput, M. P. 1999. Recovery and purification of Cyclospora cayetanensis oocysts using sucrose gradients. FDA laboratory information bulletin, vol. 15, no. 4, p. 4163. U.S. Food and Drug Administration, Washington, D.C.
Cole, D. J., K. Snowden, N. D. Cohen, and R. Smith. 1999. Detection of Cryptosporidium parvum in horses: thresholds of acid-fast stain, immunofluorescence assay, and flow cytometry. J. Clin. Microbiol. 37:457-460.
DeGraw, E., S. Heber, and A. Rowan. 1997. Update: outbreaks of cyclosporiasis—United States and Canada, 1997. Morb. Mortal. Wkly. Rep. 46:521-523.
Dixon, B. R. 2003. The prevalence and control of foodborne protozoan parasites, p. 31-76. In B. W. Blais (ed.), Current challenges in food microbiology. Research Signpost, Kerala, India.
Dixon, B. R., M. Parenteau, C. Martineau, and J. Fournier. 1997. A comparison of conventional microscopy, immunofluorescence microscopy and flow cytometry in the detection of Giardia lamblia cysts in beaver fecal samples. J. Immunol. Methods 202:27-33.
Dixon, B. R., J. Bussey, L. Parrington, M. Parenteau, R. Moore, J. Jacob, M.-P. Parenteau, and J. Fournier. 2002. A preliminary estimate of the prevalence of Giardia sp. in beavers in Gatineau Park, Quebec, using flow cytometry, p. 71-79. In B. E. Olson, M. E. Olson, and P. M. Wallis (ed.), Giardia: the cosmopolitan parasite. CAB International, Wallingford, United Kingdom.
Eberhard, M. L., N. J. Pieniazek, and M. J. Arrowood. 1997. Laboratory diagnosis of Cyclospora infections. Arch. Pathol. Lab. Med. 121:792-797.
Ferrari, B. C., and D. Veal. 2003. Analysis-only detection of Giardia by combining immunomagnetic separation and two-color flow cytometry. Cytometry 51A:79-86.
Herwaldt, B. L. 2000. Cyclospora cayetanensis: a review, focusing on the outbreaks of cyclosporiasis in the 1990s. Clin. Infect. Dis. 31:1040-1057.
Herwaldt, B. L., M.-L. Ackers, et al. 1997. An outbreak in 1996 of cyclosporiasis associated with imported raspberries. N. Engl. J. Med. 336:1548-1556.
Herwaldt, B. L., M. J. Beach, et al. 1999. The return of Cyclospora in 1997: another outbreak of cyclosporiasis in North America associated with imported raspberries. Ann. Intern. Med. 130:210-220.
Huang, P., J. T. Weber, D. M. Sosin, P. M. Griffin, E. G. Long, J. J. Murphy, F. Kocka, C. Peters, and C. Kallick. 1995. The first reported outbreak of diarrheal illness associated with Cyclospora in the United States. Ann. Intern. Med. 123:409-414.
Mead, P. S., L. Slutsker, V. Dietz, L. F. McCaig, J. S. Bresee, C. Shapiro, P. M. Griffin, and R. V. Tauxe. 1999. Food-related illness and death in the United States. Emerg. Infect. Dis. 5:607-625.
Moss, D. M., and M. J. Arrowood. 2001. Quantification of Cryptosporidium parvum oocysts in mouse fecal specimens using immunomagnetic particles and two-color flow cytometry. J. Parasitol. 87:406-412.
Orlandi, P. A., and K. A. Lampel. 2000. Extraction-free, filter-based template preparation for rapid and sensitive PCR detection of pathogenic parasitic protozoa. J. Clin. Microbiol. 38:2271-2277.
Power, M. L., S. R. Shanker, N. C. Sangster, and D. A. Veal. 2003. Evaluation of a combined immunomagnetic separation/flow cytometry technique for epidemiological investigations of Cryptosporidium in domestic and Australian native animals. Vet. Parasitol. 112:21-31.
Valdez, L. M., H. Dang, P. C. Okhuysen, and C. L. Chappell. 1997. Flow cytometric detection of Cryptosporidium oocysts in human stool samples. J. Clin. Microbiol. 35:2013-2017.
Varma, M., J. D. Hester, F. W. Schaefer, M. W. Ware, and H. D. A. Lindquist. 2003. Detection of Cyclospora cayetanensis using a quantitative real-time PCR assay. J. Microbiol. Methods 53:27-36.
Visvesvara, G. S., H. Moura, E. Kovacsnace, S. Wallace, and M. L. Eberhard. 1997. Uniform staining of Cyclospora oocysts in fecal smears by a modified safranin technique with microwave heating. J. Clin. Microbiol. 35:730-733.(Brent R. Dixon, Jeff M. B)