Final Report: Understanding Risk Factors to Cryptosporidium parvum: Studies in Gnotobiotic Pigs

EPA Grant Number: R826138
Title: Understanding Risk Factors to Cryptosporidium parvum: Studies in Gnotobiotic Pigs
Investigators: Ward, Lucy A.
Institution: The Ohio State University
EPA Project Officer: Hiscock, Michael
Project Period: February 20, 1998 through February 19, 2001 (Extended to December 19, 2002)
Project Amount: $332,084
RFA: Drinking Water (1997) RFA Text |  Recipients Lists
Research Category: Drinking Water , Water


The overall objective of this research project was to increase our knowledge and understanding of risk factors for the development of Cryptosporidiosis, the disease caused by the waterborne protozoan pathogen, Cryptosporidium parvum. Thus, we used the gnotobiotic pig model of Cryptosporidiosis to: (1) determine the infectivity and virulence of different Cryptosporidium strains known to infect humans; (2) assess the impact of host age and immune status on Cryptosporidiosis; and (3) characterize the host's immunological responses to Cryptosporidium.

Summary/Accomplishments (Outputs/Outcomes):

Initial studies were done using the Grafton Compton Human I (GCH1) C. parvum isolate (originating from an AIDS patient and provided to our laboratory by Dr. Saul Tzipori, Tufts University School of Veterinary Medicine, Grafton, MA) and the Ohio (OH) C. parvum isolate (originating from an immune-competent adult laboratory worker with clinical cryptosporidiosis at OARDC, Wooster, OH). In collaboration with Dr. Lihua Xiao at the Centers for Disease Control and Prevention (CDC), Chamblee, GA, we were able to obtain and test several additional C. parvum isolates that infect humans, including two additional BoG2 C. parvum, six human genotype 1 (HuG1) C. parvum, six C. meleagridis, three C. canis, and two C. felis isolates. We successfully passaged most isolates and currently are maintaining viable populations of three BoG2 strains (GCH1, OH, B7038), five HuG1 strains (H2132, H2265, H2576, H3438, H5016), and four C. meleagridis strains (Av2567, Av4500, Av4501, Av4509).

Molecular characterization of the parental GCH1 and OH isolates was done in collaboration with Drs. Saul Tzipori and Giovanni Widmer at Tufts University School of Veterinary Medicine. Genetic analyses of the C. parvum oocystwall protein (COWP) and beta-tubulin locus of GCH1 and OH showed both strains to be bovine genotype 2 or BoG2 C. parvum. Indepth analysis of the tubulin locus of OH demonstrated a homogeneous bovine type, which was unusual in that most BoG2 C. parvum isolates from humans are heterogeneous at this locus, including GCH1 (Widmer, personal communication). Interestingly, the homogeneous bovine type beta-tubulin locus was lost after several passages of OH in pigs and calves. We subsequently derived single oocyst clones of GCH1 and OH using micromanipulation techniques to isolate a single oocyst of each strain, which subsequently was inoculated into gnotobiotic pigs (Wang, 2000). Routine molecular analysis of these clones and all other isolates has continued in collaboration with Dr. Xiao at the CDC. Subgenotyping of the HuG1 isolates suggests that strains H2132 and H2265, which were derived from U.S. waterborne disease outbreaks at geographically separated sites (one from Nebraska and one from North Carolina), are the same. Genotyping/subgenotyping studies also have shown that laboratory cross-contamination of strains can occur (see C. canis studies below), despite the use of very strict quality control steps during the purification and handling of different isolates and strains within the laboratory.

Infectivity of (Human-Derived) BoG2 C. parvum, HuG1 C. parvum, C. meleagridis, C. canis, and C. felis in Gnotobiotic Pigs

BoG2 C. parvum. The minimum infective dose (MID) for our parental GCH1 C. parvum in neonatal (1 to 4 days old) gnotobiotic pigs is less than or equal to 5 oocysts, which results in 25 percent diarrhea and 100 percent oocyst shedding for more than 10 days. The median (50 percent) diarrhea dose (DD50) is 730 oocysts, which results in mild to moderate diarrhea for 5 to 10 days. Doses of qreater than or equal to 108 GCH1 oocysts were 100 percent lethal, with onset of oocyst shedding and diarrhea observed by 24-36 hours post-inoculation (PI); thus, the median lethal dose (LD50) for parental GCH1 was estimated at 107 oocysts. The MID of parental OH C. parvum in neonatal pigs similarly is low (< 5 oocysts), but its DD50 was approximately 1 log higher (6,900 oocysts), and 108 OH oocysts were lethal to only 20 percent of inoculated animals; thus, the LD50 for parental OH also is at least one log higher than that of parental GCH1. Single oocyst clones of GCH1 and OH C. parvum subsequently were derived from our parental stock (Wang, 2000; Pereira, 2001; Pereira, et al., 2001). Like their parental stock, the MID for each clone was low, and in fact, shown to be a single oocyst (Wang, 2000; Pereira, et al., 2002). However, when freshly propagated (< 3 weeks of age) oocysts of each clone were used for pig inoculums, both the DD50 and LD50 were shown to be much lower at less than or equal to 5 oocysts for both GCH1 and OH. Furthermore, variability in the clinical responses of gnotobiotic pigs to different pools of (freshly propagated) C. parvum clones was minimal to nonexistent, suggesting that our clones were a genetically homogeneous (clonal) population of parasites (as compared to the parental stock, in which variability in clinical responses to different pools of oocysts was much greater). Finally, unlike our parental GCH1 and OH, which showed significant beta-tubulin gene sequence variability within and between oocyst pools generated by high-dose passages (see Widmer, et al., Applied Environmental Microbiology 1998;64:4477-4481), we have failed to detect any beta-tubulin gene sequence variability within or between all subsequent (low-dose) passages of our clones.

HuG1 C. parvum (aka C. hominis). We have six HuG1 isolates available for study (H2132, H2265, H2546, H2576, H3438, H5016) and successfully have derived single oocyst clones from three of these. The MID of HuG2 isolates for the gnotobiotic pig also is a single oocyst when freshly propagated oocysts are used for preparing inoculums. The DD50 for the HuG1 isolates similarly was low when using freshly propagated oocysts (10-100 oocysts depending on the strain). However, the severity of diarrheal disease is significantly less than that observed with BuG2 infections (Pereira, et al., 2002). Mostly, only moderate disease is observed in pigs inoculated with qreater than or equal to 107 HuG1 oocysts. Thus, the LD50s for the HuG1 C. parvum strains we possess have not been determined, as such strains appear to be nonlethal in this animal model.

C. meleagridis. We successfully have propagated four C. meleagridis strains (Av2567, Av4500, Av4501, Av4509) in neonatal gnotobiotic pigs using an initial infective dose of 104 oocysts. Verification of these strains as C. meleagridis has been confirmed by restricted fragment length polymorphism (RFLP) and sequence analysis. We have infected pigs with as few as 5 oocysts (freshly propagated) and thus predict the MID for C. meleagridis, also to be a single oocyst. As observed with C. parvum, the DD50 is low (10 to 100 oocysts) and the severity of diarrheal disease is more similar to that elicited by HuG1 C. parvum and significantly less severe than that observed for BoG2 C. parvum. Only mild to moderate disease has been observed at doses of 105 C. meleagridis oocysts. The LD50s of C. meleagridis have not been determined, as such strains appear to be nonlethal in this animal model.

C. canis. Two 1-day-old pigs were given approximately 104 immunomagnetic-bead (IMB) purified oocysts extracted from C. canis-infected human feces (K9-2562 strain). A productive infection was detected on fecal smears by the ultraviolet-acid fast (UV-AF) method (Nielsen and Ward, 1999) at 12 and 18 days (average 14.6 + 2.2 days) post infection (PI). However, molecular analyses (genotyping) of the UV-AF positive pig feces revealed only BoG2 C. parvum, even though the C. canis genotype could be reisolated from the human feces used to prepare the pig inoculants (L. Xiao, personal communication). Thus, we concluded that at least one of the pig inoculants had been inadvertently contaminated with BoG2 C. parvum during laboratory preparation, as our GCH1 C. parvum had been purified from calf feces in the same laboratory space the previous day and its MID is a single oocyst. Thus, two more 1 day-old pigs were given approximately 104 IMB-purified K9-2562 oocysts extracted from the same human stool sample. A productive infection was detected by UV-AF at 20 days PI for 1-day duration in one pig. Consequently, both pigs were killed at 22 days PI to check for subclinical infection, but polymerase chain reaction (PCR) analyses of intestinal contents have been negative. Two new C. canis isolates were obtained from human fecal samples (K9-4495 and K9-4496) and 5 x 104 IMB-purified oocysts per isolate per pig used for inoculation. This time, neither pig developed a productive infection as determined by UV-AF analysis of feces by 15 days PI. Consequently, each pig was given a second inoculum of 105 IMB-purified C. canis oocysts and monitored for 2 additional weeks, at which time the experiment was terminated. PCR analysis of each pig’s feces/intestinal content has been negative.

C. felis. Two C. felis strains (F6486 and F7055) obtained from human patients were tested for infectivity in gnotobiotic pigs. One mL of each human stool sample containing C. felis was treated with peracetic acid and inoculated into 1 day-old gnotobiotic pigs. Neither inoculum elicited productive infection as determined by UV-AF fecal analysis by 10 days PI, at which time a second inoculum (prepared from 4 mL of each stool sample and peracetic acid treated) was given to each pig. A productive infection, as determined by UV-AF fecal analysis, was not observed following the second inoculation, and the experiment was terminated 2 weeks later. PCR analysis of the collected feces/intestinal contents has been negative.

Virulence of BoG2 C. parvum and HuG1 C. parvum in Gnotobiotic Pigs

HuG1 and BoG2 C. parvum. Comparative pathogenesis studies using BoG2 C. parvum isolates GCH1 and OH, and HuG1 C. parvum isolates H2132, H2265, H3438, and H34396 were done to establish the virulence of these Cryptosporidium in the gnotobiotic pig model (Pereira, 2001; Pereira, et al., 2002). Preliminary HuG1 and BoG2 dual infection studies also were performed to determine whether HuG1 and BoG2 infections could co-exist in the host (alternatively, one genotype would out-compete the other) and whether genotypes could exchange genetic material. Findings from these studies are summarized in the sections that follow and provide scientific evidence to support designation of the new species C. hominis for the HuG1 C. parvum (Morgan-Ryan, et al., 2002).

Clinical Disease and Oocyst Shedding. Oocyst shedding was evaluated by UV-AF. The onset to oocyst shedding (or prepatent period) consistently was longer for HuG1 C. parvum (7 to 14 days) compared to BoG2 C. parvum (4 to 7 days). The length of oocyst shedding (or patent period) also was consistently longer for HuG1 C. parvum (13 to 26 days) compared to BoG2 C. parvum (5 to 16 days). Feces were scored daily from 0-3, where 0 = normal, 1 = pasty, 2 = semi-liquid, and 3 = liquid. Diarrhea was deemed present at scores qreater than or equal to 2. The diarrheal disease in pigs during the first week post-onset of oocyst shedding consistently was less severe with HuG1 C. parvum strains (daily diarrhea scores ranged from 0.63 to 1.89) compared to BoG2 C. parvum (daily diarrhea scores ranged from 1.63 to 2.78). Uninfected control pigs had an average daily fecal score of 0.48±0.8 standard error of the measurement (SEM). BoG2 C. parvum also induced moderate to severe dehydration, weight loss, protruding vertebrae, and sunken eye appearance in pigs during the patent (shedding) period, whereas HuG1 C. parvum elicited little to no dehydration with minimal to no weight loss during the patent period.

Morphologic Changes in the Intestinal Tissues. Microscopic examination of Hematoxylin and Eosin (H&E) stained tissue sections showed BoG2 C. parvum parasites throughout the duodenum, jejunum, ileum, and colon. BoG2 C. parvum parasites were detectable in the intestinal tissues of the pigs killed during the prepatent (3 days PI [DPI]) and patent (0-21 days post-onset [DPO]) periods. By comparison, HuG1 C. parvum parasites of strain H2132 and H2265 were seen only during the patent period and in the ileum and colon, and no human C. parvum parasites have been found in the duodenum and upper jejunum. However, at the onset of shedding, consistently more HuG1 C. parvum parasites per villous can be observed in the ileum compared to BoG2 C. parvum (greater infection intensities). BoG2 C. parvum elicits moderate lymphoid hyperplasia, with scattered inflammatory cell (neutrophils) infiltrates at the onset of shedding in the duodenum at 7 DPO in the jejunum and ileum and 7 and 21 DPO in the colon. With HuG1 C. parvum, lymphoid hyperplasia generally is observed only during the patent period, and only mild lymphocytic infiltrates with rare inflammatory cells are seen at this time in the ileum and colon. Mild to moderate villous attenuation with epithelial sloughing was seen at the onset of shedding and 7 DPO in the duodenum, and 21 DPO in the colon with BoG2 C. parvum, whereas HuG1 C. parvum elicited only mild villous/mucosal attenuation, which generally was restricted to the ileum and colon during the patent period.

Dual Infection Studies With BoG2 and HuG1 C. parvum. One day-old pigs were inoculated with 10 GCH1 oocysts and 8 x 104 H2265 oocysts, and killed at selected DPO of shedding to evaluate tissues and fecal contents. Onset of oocyst shedding, as determined by UV-AF analysis of fecal smears, was 6.3±0.8 days PI (consistent with BoG2 C. parvum prepatent period) and continued for 21 days (consistent with HuG1 C. parvum patent period). Pigs developed relatively severe diarrhea (average daily fecal score 2.9±0.1) the first week post-onset of shedding accompanied by dehydration and weight loss (consistent with BoG2 infection). Repeated molecular analysis of the small and large intestinal contents collected at DPO 2, 8, and 22 revealed GCH1 DNA only, except for a single PCR/RFLP of upper small intestinal content at DPO 22, which showed the human genotype. Genetic recombinations of GCH1 and H2265 have not been found. Thus, because our BoG2 C. parvum appeared to out-compete our HuG1 C. parvum, a second experiment was performed, in which HuG1 infection was established followed by inoculation with BoG2 C. parvum. In this experiment, 1 day-old pigs were inoculated with 5 x 103 H5016 oocysts, followed 9 days later by inoculation with 5 x 103 GCH1 oocysts. Onset of oocyst shedding as determined by UV-AF analysis of fecal smears was 8.8±1.4 days PI (consistent with HuG1 C. parvum prepatent period), which continued for 7±0.7 days. Pigs were killed at DPO 1, 7, and 11, and 23 days PI. This time, repeated molecular analysis of the intestinal contents collected at necropsy revealed only human genotype (bovine genotype was not isolated) and no genetic recombinations were found.

Virulence of C. meleagridis in Gnotobiotic Pigs

Clinical Disease and Shedding. Onset to oocyst shedding as detected by UV-AF was 6.5±0.3 days PI and continued for 21.4±2.6 days. Diarrhea was mild to moderate the first week post-shedding with an average daily score of 1.7±0.3 (uninfected, age-matched control pigs average daily fecal score = 0.48±0.8). Little to no dehydration with minimal to no weight loss was observed in pigs during the patent (oocyst shedding) period.

Morphologic Changes in the Intestinal Tissues. Microscopic evaluation of H&E stained tissues collected at necropsy showed low numbers of parasites throughout the small intestine (1-10 parasites per 100 linear microns) and moderate (11-20 parasites/100 µm) to high (> 20 parasites/100 µm) numbers in the large colon throughout the patent period. Villous/mucosal attenuation generally was mild throughout the infected regions of the small and large intestines. Moderate lymphoid hyperplasia also was noted throughout most sections of the infected intestines.

Impact of Host Age and Immune Status on Cryptosporidiosis

Host Age. The infectivity and virulence of our parental BoG2 C. parvum isolates (GCH1 and OH) were investigated in adolescent (3-5 week-old) gnotobiotic pigs and compared to that found for neonatal gnotobiotic pigs. When freshly propagated parental GCH1 or OH oocysts were used for inoculation, the MID similarly was low at less than or equal to 5 oocysts. However, older pigs shed oocysts (as detected by UV-AF) for a slightly shorter period than neonatal pigs (6.9±0.8 days versus 8.0+0.1 days) and the severity of disease markedly was less (average daily fecal score 1.4±0.1 versus 2.2+0.2 in neonatal pigs) the first week post-shedding. In addition, no deaths were observed in the older pigs, even at doses of 108 to 109 oocysts (or ~100X greater than the neonatal pig LD50). The DD50 was higher in older pigs at ~105 parental GCH1 or OH oocysts, compared to ~103 to 104 oocysts for neonates. The histologic changes in hematoxylin and eosin-stained tissue sections also were less pronounced than those observed for neonatal pigs, with only low to moderate numbers of parasites (1 to 20 parasites/100 µm) associated with mild to moderate villous/mucosal attenuation, and moderate lymphoid hyperplasia noted throughout the intestine during the patent period.

Immune Status. It was readily apparent from our earlier studies that the status of the pig's immune system would not factor into the MID. Thus, we focused on the role of the intact immune system in disease expression and parasite clearance. Immunosuppression of pigs was achieved through daily oral dosing, with dexamethazone (dex) from 1 to 6 days of age, followed by prednisolone (pred) from 1 to 4 weeks of age. Twelve pred/dex treated and 12 untreated control pigs were orally inoculated with 10-20 cloned GCH1 C. parvum oocysts that had been stored at 4°C for 3 months. Onset to shedding (prepatent period) was shorter in the dex/pred treated versus untreated GCH1-inoculated pigs (4.8±0.4 versus 7.4±0.3 days) and their duration of shedding period (or patent period) was longer (9.3±2.3 versus 6.9±1.4 days). However, no difference in clinical disease was observed between dex/pred treated and untreated GCH1-inoculated pigs, with their average daily fecal score during the first week post-shedding being 2.2±0.3 versus 2.4±0.2, respectively. It was of interest to note that the onset to shedding in the untreated GCH1-inoculated pigs was longer using the aged GCH1 oocysts than that observed when using freshly propagated oocysts (7.4±0.3 days versus 5.4±0.68 days, Pereira, et al., 2002). In addition, inoculation of 10 to 20 "aged" H2265 oocysts (stored at 4°C for 3 months) into either dex/pred treated or untreated 1 day-old pigs failed to elicit productive infection as determined by UV-AF analysis of daily fecal smears, and histologic evaluation of these pigs' tissues, as well as PCR analysis of their intestinal contents collected at necropsy, were negative.

Immunological Responses to C. parvum

Cellular Immunity: Cytokine Responses. We investigated the correlation between intestinal IL-12 (p40), IFNgamma, TNFalpha, IL-10, and IL-6 mRNA levels and C. parvum pathogenesis (Pereira, 2001; Pereira, et al., submitted 2003 ). Briefly, 1 day-old gnotobiotic pigs were inoculated orally with approximately 5 x 106 bovine (GCH1 or OH) C. parvum oocysts, and portions of their bowels were collected for morphologic evaluation and cytokine mRNA activity at selected times PI. Diarrhea developed in all pigs by 5 days PI and was resolved by 16 days PI. At 3 days PI, small intestinal cytokine mRNA (predominantly IFN-g, TNF-alpha, and occasionally IL-12 and IL-10) levels of infected pigs were increased, whereas colonic cytokine activity was no different than negative controls. Histologically, heavy parasite infection rates, especially in the upper small bowel, with villous atrophy and neutrophilic, lymphocytic infiltrates were observed but no significant colonic histologic changes were seen. At 10 days PI, small intestinal cytokine mRNA levels remained elevated in the infected pigs, with slight elevation in colonic cytokine mRNA levels. Histologically, moderate villous attenuation and mucosal inflammation continued to be seen throughout the small bowel, although parasites mostly were observed in the ileum and colon. At 3 weeks PI, cytokine mRNA levels remained increased in colon and ileum (predominantly TNF-alpha) and decreased in duodenum (except IFN-gamma and TNF-alpha). Histologically, small intestinal tissues showed mild mucosal attenuation with no parasites, and colonic tissues showed prominent lymphoid hyperplasia and moderate parasite infection rates. At 7 weeks PI, bowel segments were histologically similar to negative controls (except for notable lymphoid hyperplasia in infected pigs) and all intestinal cytokine mRNA levels were similar to negative controls. A significant correlation between time, tissue, and cytokine mRNA expression was observed, which appeared dependent on parasite presence or absence and not severity of tissue pathology. Analysis of systemic tissues (spleens) from these same pigs revealed a markedly different cytokine mRNA expression pattern compared to intestinal tissues, with all cytokine mRNAs increasing immediately PI and remaining elevated throughout the entire study period.

Because of the significant differences we observed between the pathogenesis of human and bovine genotypes of C. parvum in the pig model (Pereira, et al., 2002), we developed a real-time reverse transcriptase (RT)-PCR assay to quantitate expression of IL-1, IFNg, IL-4, IL-10, and IL-12 mRNA in the tissues of pigs inoculated with low dose HuG1 or BoG2 C. parvum. In these studies, 1 day-old pigs were inoculated with 10 to 100 oocysts of either BoG2 (GCH1 and OH) or HuG1 (H2132 and H2265) C. parvum and portions of their bowels collected for cytokine mRNA analysis at 3 days PI, 7-10 days PI (or 0-2 DPO of oocyst shedding), 14-17 days PI, 21-28 days PI, and 49-50 days PI. Unlike the previous cytokine study (Pereira, et al., 2003, submitted), no significant correlation was found between time, tissue, and cytokine mRNA level, and no significant differences in cytokine mRNA expression were found between the BoG2-infected and HuG1-infected pigs. However, it was of interest to note that the Th2-helper cytokines, IL-4 and IL-10, were expressed only by BoG2-infected pigs post-shedding.

Humoral Immunity. Using an isotype-specific ELISA developed by Dr. Karol Sestak at Tufts University School of Veterinary Medicine, the humoral immune responses of gnotobiotic pigs inoculated with high doses of BoG2 C. parvum isolates (parental GCH1 and OH) were examined. In summary, we found 23 percent of neonatal gnotobiotic pigs developed Cryptosporidium-specific IgM serum antibodies, while only 2 percent developed IgG2 and IgA serum antibodies at 6-10 days PI. By 21 days PI, 26 to 34 percent of pigs developed detectable low titer (geometric mean titer [GMT] = 1.8-2.4) IgM, IgG1, and IgG2 serum antibodies. No IgA serum antibodies were detected at this time. By 28 days PI, 75-100 percent of pigs had developed higher titer IgM, IgG1, and/or IgG2 serum antibodies (GMT = 8 to 26) but no serum IgA antibodies. By 49 days PI, all BoG2-inoculated pigs (100 percent) had Cryptosporidium-specific IgG1 and IgG2 (GMT = 80 and 112), and one-half (50 percent) had Cryptosporidium-specific IgA (GMT = 4.5). However, antibodies rarely were detected in the intestinal contents from these same pigs, with very low IgM, IgG, and/or IgA antibody titers (GMT 1.2–1.6) detectable in 8-9 percent of all pigs at 6-10 days PI and 21 days PI only.

A second isotype-specific ELISA, developed in parallel to Dr. Sestak's work, was utilized in-house to compare the serum antibody responses of gnotobiotic pigs that were: (1) C. parvum-infected and dex/pred-treated (immunosuppressed) versus untreated; (2) C. parvum-infected with high versus low inoculating dose; and (3) C. parvum-infected with HuG1 (H2132 and H2265) versus BoG2 (GCH1 and OH) strains. First, all untreated C. parvum-infected pigs developed Cryptosporidium-specific serum antibodies by 3 weeks PI, whereas more than one-half (57 percent) of the dex/pred-treated pigs failed to develop Cryptosporidium-specific serum IgG, and more than one-quarter (29 percent) failed to develop detectable Cryptosporidium-specific serum IgM or IgA by 5 weeks PI. The serum GMTs of dex/pred-treated C. parvum-infected pigs also were lower than untreated C. parvum-infected pigs at this time (4, 12, and 15 versus 710, 70, and 39 for IgG, IgM, and IgA, respectively). Second, the dose of C. parvum used to infect the pigs was found to influence the subsequent antibody responses with low doses (< 10 oocysts) of GCH1 or OH, generally eliciting a slower and lower serum IgG response and a slower but higher serum IgM and IgA response, compared to high doses (> 10,000 oocysts) at 3 to 6 weeks PI (980, 47, and 151 versus 237, 27, and 88 for IgG, IgM, and IgA, respectively). High doses of HuG1 and BoG2 C. parvum isolates or low doses of BoG2 isolates elicited serum antibody responses in all (100 percent) pigs tested, whereas low doses of HuG1 C. parvum did not (~20 percent did not respond). Third, the C. parvum genotype influenced the antibody response, with HuG1 C. parvum eliciting lower IgG, IgM, and IgA responses than low doses of BoG2 C. parvum (35, 11, and 33 versus 80, 47, and 151 for IgG, IgM, and IgA, respectively). However, we have not excluded the possibility that our coating antigen (BoG2 C. parvum oocysts) has biased our results towards lower titers for HuG1 C. parvum-infected pigs because of known phenotypic variability between the immunodominant surface glycoproteins of HuG1 and BoG2 C. parvum (see below).

Finally, sera collected from selected GCH1-inoculated pigs were used to investigate the antigenic (phenotypic) relatedness between HuG1 and BoG2 C. parvum surface glycoproteins 40/15 (gp40/15) (Sestak, et al., 2002). The gp15 was found to be immunodominant for both HuG1 and BoG2 C. parvum isolates. Lower genetic sequence similarity between HuG1 and BoG2 gp40/15 corresponded to gp15 protein differences as detected by Western Blot. Moreover, gp15 was found to contain immunodominant epitopes. Deglycosylation of C. parvum proteins resulted in decreased ability of gp15, gp23, and gp900 to react with homologous polyclonal antibodies, suggesting that these proteins also express carbohydrate epitopes. These findings suggest there is high phenotypic variability between C. parvum HuG1 and BoG2 isolates at the level of gp15.

Ultraviolet (UV)-Acid Fast (AF) Method. Our laboratory reported on the enhancement of the AF stain procedure by viewing AF stained specimens under fluorescent light microscopy using a 540-560 nm (rhodamine) excitation filter. This procedure increased the sensitivity of the AF stain on fecal smears to levels comparable with a commercially available immunofluorescence assay (IFA) product (MERIFLOUR, Meridian Diagnostics, Inc., Cincinnati, OH) (Nielsen and Ward, 1999).

Extracting PCR Quality DNA From Frozen Fecal Specimens. A relatively rapid, inexpensive method for extracting PCR quality DNA from frozen fecal specimens, which allows detection of a single C. parvum oocyst, was described and published (Ward and Wang, 2001). This procedure involves the use of glassmilk (silica suspension) to directly extract C. parvum DNA from feces, and thus, is useful for specimens in which the oocyst integrity is disrupted, such as frozen samples. We further showed that, when combined with nested PCR, our method repeatedly allowed for detection of a single C. parvum organism in frozen fecal materials (Ward and Wang, 2001). Interestingly, the pattern of fecal oocyst shedding in experimentally inoculated pigs clearly was dependent on the assay used to monitor the feces: fecal oocyst shedding commenced earlier (by 1-5 days) and lasted longer (by 0 to 20 days) by PCR versus UV-AF, and only PCR (not UV-AF) detected fecal oocyst shedding at low infective doses.

Quantitative Competitive (QC)-PCR Assay. A standard curve QC-PCR assay was developed to determine the number of C. parvum organisms in the intestinal tissues (duodenum, jejunum, ileum, and colon) of gnotobiotic pigs following low dose inoculations (Wang, 2000). We showed that the number of C. parvum organisms per microgram of intestinal DNA significantly varied by intestinal segment (p<0.016) and 1 day PI (p<0.001). Following single BoG2 C. parvum oocyst inoculation, the greatest number of C. parvum organisms per microgram of intestinal DNA was found in the colon at 10 days PI (5.6 x 104 organisms), and the earliest but lowest detectable number of organisms/microgram DNA was in the ileum at 5 days PI (9 organisms). Parasitic DNA was not detectable in intestinal tissues at 0 and 3 days PI. The QC-PCR proved useful for monitoring developing intestinal parasitemia in the gnotobiotic piglet model of cryptosporidiosis, and was more rapid and sensitive than conventional microscopic screening of H&E-stained tissue sections.

In Vitro Viability Assay. A novel flow cytometric viability assay was developed to evaluate the in vitro effect of four common food fermenting lactic acid bacteria (LAB) derived from human intestines (Lactobacillus acidophilus, L. reuteri, Bifidobacterium breve, and B. longum) on the oocyst (infective) stage of C. parvum (Foster, et al., 2003). Compared to broth controls, Lactobacillus supernatants significantly reduced oocyst viability up to 81 percent, whereas Bifidobacterium supernatants reduced viability only 10-37 percent. These results suggest the presence of antimicrobial substance(s) against the oocyst stage of C. parvum in the supernatants of L. reuteri and L. acidophilus broth cultures.

In Vitro Infectivity Assay. A cell culture infectivity assay was developed to investigate the in vitro effect of our human intestinal L. acidophilus and L. reuteri on C. parvum infectivity (Glass, et al., 2002). Supernatants from these LAB were shown to significantly reduce the infectivity of fresh and aged BoG2 C. parvum oocysts seeded onto HCT8 cell monolayers by 20 to 43 percent. Oocyst infectivity was reduced nearly two-fold when LAB supernatants were combined compared to individual supernatants. These studies lend further support to the hypothesis that production of anti-microbial compounds by LAB may be an effector mechanism against C. parvum. In addition, a positive correlation between oocyst viability (as determined by the flow cytometric assay) and infectivity (in cell culture) was found whose value was determined by the age of oocyst pool. With fresh BoG2 and HuG1 C. parvum oocyst pools (< 3 weeks old), the relationship essentially was 1:1 (r = 1.0) regardless of the strain used. However, with older oocyst pools (1 to 3 months old), the correlation was less and varied by both strain and oocyst age (r = 0.77 to 0.94).


The following conclusions have been drawn from the research:

(1) Human-derived HuG1 C. parvum, BoG2 C. parvum, and C. meleagridis isolates readily infect neonatal gnotobiotic pigs, whereas human isolates of C. canis and C. felis do not. The neonatal gnotobiotic pig is thus a very appropriate model for comparative studies of human-derived HuG1 C. parvum, BoG2 C. parvum, and C. meleagridis. Additional C. canis and C. felis isolates should be tested before determining the appropriateness of the gnotobiotic pig model for these species, as viability of our test samples prior to pig inoculation was unknown.

(2) Comparative infection studies using our freshly propagated HuG1 C. parvum, BoG2 C. parvum, and C. meleagridis strains in neonatal gnotobiotic pigs suggest differences exist between their median diarrheal and lethal doses but not their MID (see (12) below).

(3) The MID of C. parvum and C. meleagridis for the susceptible, naïve host is extremely low (a single viable oocyst) and not altered by the age or immune status of the host or by the strain of parasite.

(4) The virulence of C. parvum in the gnotobiotic pig model (defined by onset of oocyst shedding, severity of diarrheal disease, number of oocysts shed, duration of oocyst shedding) is dependent on the C. parvum strain (parental versus clone; HuG1 versus BoG2), the dose or number of oocysts ingested, as well as the age and immune status of the pig.

(5) The use of single-oocyst "cloned" C. parvum isolates, which are maintained by low-dose propagation methods, appears to eliminate both genetic and biological behavior variability that often is observed within and between different pools of the same "uncloned" C. parvum strain propagated by traditional high-dose methods.

(6) HuG1 C. parvum isolates significantly differ from BoG2 C. parvum isolates in prepatent periods, clinical disease, intestinal site of replication, and types and severity of lesions induced in the gnotobiotic pig model, whereas only minimal differences in virulence exist for different strains within the same genotype.

(7) There is high phenotypic variability between HuG1 and BoG2 C. parvum isolates at the level of surface glycoprotein 15 (gp15).

(8) Dual infection studies failed to produce genetic recombinants of HuG1 and BoG2 C. parvum. In fact, infecting pigs simultaneously with an HuG1 and BoG2 C. parvum strain proved extremely difficult. When HuG1 and BoG2 C. parvum were administered together, the BoG2 isolate generally out-competed the HuG1 isolate, even when 1,000 times more HuG1 oocysts were given. Furthermore, once an infection with one genotype was established, superimposing infection by the other genotype was not possible, regardless of the strain or dose given.

(9) Taken collectively, our studies provide additional support for designation of the new species, C. hominis for HuG1 C. parvum.

(10) Comparative morphology and cytokine studies in pig's infected with high dose BoG2 C. parvum showed a strong positive correlation between time, tissue, and cytokine mRNA expression, which was dependent on the parasite presence/absence and not the severity of tissue pathology. No significant differences in cytokine mRNA expression were found between HuG1 versus BoG2 C. parvum-infected pigs. These studies suggest that C. parvum infection, regardless of infecting strain or its virulence, elicits mucosal cytokine responses in gnotobiotic pigs that significantly contribute to parasite clearance and recovery.

(11) Inoculation of high (qreater than or equal to 105 oocysts) or low (< 100 oocysts) doses of BoG2 C. parvum elicits Cryptosporidium-specific serum antibody responses in all (100 percent) of neonatal gnotobiotic pigs by 3 (high dose) to 5 (low dose) weeks PI. Inoculation of high doses of HuG1 C. parvum oocysts also elicited serum antibody responses in all pigs tested, whereas inoculation of low doses did not (~20 percent did not respond). Infection of gnotobiotic pigs with our BoG2 C. parvum isolates typically resulted in development of greater and earlier antibody responses than our HuG1 C. parvum isolates. However, all antibody responses were slow to develop (3-4 weeks PI for high doses and 5-6 weeks PI for low doses) and of low magnitude compared to that observed for other enteric pathogens, such as rotavirus. These findings begin to provide insight as to why hosts often are susceptible to reinfection with this parasite.

(12) In vitro studies clearly show that the age of the oocysts dictates the infectivity of a C. parvum strain in cell culture. Viability and infectivity correlate 1:1 when fresh oocyst are used (<3 weeks old) regardless of the strain used. In contrast, the correlation between viability and infectivity will vary in oocyst pools >3 weeks old, depending on both the strain and actual age of the oocyst pool. Thus, the MID of a C. parvum strain in an animal also is likely to be influenced by the age of the oocyst pool being used for inoculation.

(13) Because of the extremely low MID of Cryptosporidium, genotyping/subgenotyping of Cryptosporidium isolate(s) before and after each animal/human inoculation should be routine practice in laboratories doing dose-response or genetic studies for quality assurance/quality control purposes (eliminates erroneous results because of inadvertent cross-contamination of isolates).

Journal Articles on this Report : 6 Displayed | Download in RIS Format

Other project views: All 41 publications 8 publications in selected types All 6 journal articles
Type Citation Project Document Sources
Journal Article Foster JC, Glass MD, Courtney PD, Ward LA. Effect of Lactobacillus and Bifidobacterium on Cryptosporidium parvum oocyst viability. Food Microbiology 2003;20(3):351-357. R826138 (Final)
  • Abstract: Science Direct Abstract
  • Other: Ohio-State PDF
  • Journal Article Morgan-Ryan, UM, Fall A, Ward LA, Hijjawi N, Sulaiman I, Fayer R, Thompson RCA, Olson M, Lal A , Xiao L. Cryptosporidium hominis n. sp (Apicomplexa: Cryptosporidiidae) from Homo sapiens. Journal of Eukaryotic Microbiology 2002;49(6):433-440. R826138 (2001)
    R826138 (Final)
  • Abstract from PubMed
  • Journal Article Nielsen CK, Ward LA. Enhanced detection of Cryptosporidium parvum in the acid-fast stain. Journal of Veterinary Diagnostic Investigation 1999;11(6):567-569. R826138 (1999)
    R826138 (2000)
    R826138 (2001)
    R826138 (Final)
  • Abstract from PubMed
  • Other: JVDI PDF
  • Journal Article Pereira SJ, Ramirez NE, Xiao L, Ward LA. Pathogenesis of human and bovine Cryptosporidium parvum in gnotobiotic pigs. Journal of Infectious Diseases 2002;186(5):715-718. R826138 (Final)
  • Abstract from PubMed
  • Other: UChicago PDF
  • Journal Article Sestak K, Ward LA, Sheoran A, Feng X, Akiyoshi DE, Ward HD, Tzipori S. Variability among Cryptosporidium parvum genotype 1 and 2 immunodominant surface glycoproteins. Parasite Immunology 2002;24(4):213-219. R826138 (2001)
    R826138 (Final)
  • Abstract from PubMed
  • Full-text: InterScience Full Text
  • Other: InterScience PDF
  • Journal Article Ward LA, Wang Y. Rapid methods to isolate Cryptosporidium DNA from frozen feces for PCR. Diagnostic Microbiology and Infectious Disease 2001;41(1-2):37-42. R826138 (2000)
    R826138 (2001)
    R826138 (Final)
  • Abstract from PubMed
  • Supplemental Keywords:

    dose-response, drinking water, pathogens, immunology., RFA, Health, Scientific Discipline, Water, Environmental Chemistry, Health Risk Assessment, Chemistry, Risk Assessments, Biochemistry, Drinking Water, cryptosporidium parvum oocysts, pathogens, public water systems, risk factors, microbial risk assessment, waterborne disease, exposure and effects, disinfection byproducts (DPBs), exposure, dose response, community water system, gnotobiotic pigs, human exposure, susceptibility, treatment, dietary ingestion exposures, drinking water contaminants, infectivity, water treatment

    Progress and Final Reports:

    Original Abstract
  • 1998
  • 1999 Progress Report
  • 2000 Progress Report
  • 2001 Progress Report