Survival of Infectious Cryptosporidium parvum Oocysts in Seawater and Eastern Oysters (Crassostrea virginica) in the Chesapeake Bay

Abstract

Oocysts of Cryptosporidium parvum placed in artificial seawater at salinities of 10, 20, and 30 ppt at 10°C and at 10 ppt at 20°C were infectious after 12 weeks. Those placed in seawater at 20 ppt and 30 ppt at 20°C were infectious for 8 and 4 weeks, respectively. These findings suggested that oocysts could survive in estuarine waters long enough to be removed by filter feeders such as oysters. Thereafter, 30 Eastern oysters, Crassostrea virginica, were collected with a dredge or with hand tongs at each of six sites within Maryland tributaries of the Chesapeake Bay in May and June and in August and September of 1997. Hemocytes and gill washings from all oysters were examined for the presence of Cryptosporidium oocysts and Giardia cysts by immunofluorescence microscopy utilizing a commercially available kit containing fluorescein isothiocyanate-conjugated monoclonal antibodies. Giardia was not detected by this method from any of the 360 oysters examined. Presumptive identification of Cryptosporidium oocysts was made in either hemocytes or gill washings of oysters from all six sites both times that surveys were conducted. In addition, during August and September, for each of the six sites, hemocytes from the 30 oysters were pooled and gill washings from the oysters were pooled. Each pool was delivered by gastric intubation to a litter of neonatal mice to produce a bioassay for oocyst infectivity. Intestinal tissue from two of three mice that received gill washings from oysters collected at a site near a large cattle farm and shoreline homes with septic tanks was positive for developmental stages of C. parvum. These findings demonstrate for the first time that oysters in natural waters harbor infectious C. parvum oocysts and can serve as mechanical vectors of this pathogen.

Marine waters worldwide have been found to be contaminated with domestic sewage from outfalls and storm water runoff (8). Fecal stages of both Cryptosporidium and Giardia spp., protozoan parasites of humans and animals, were found in such water near a bathing beach in Hawaii (10). Oocysts of Cryptosporidium parvum stored at 4°C in the dark in natural seawater remained viable, as determined by dye exclusion, for over a month (15). If oocysts could survive for a similar length of time in seawater at normal environmental temperatures, this would provide a sufficient period for potential exposure for humans and marine animals. We previously found that Eastern oysters are capable of removing oocysts of C. parvum from artificially contaminated seawater and of retaining them within hemocytes, on gills, and within the body for at least 1 month after exposure and that such oocysts residing in oyster tissues for 1 week are still infectious for mice (7). These findings suggested that if oocysts of C. parvum contaminated estuarine waters where oysters were harvested for human consumption, the oocysts could be taken up by the oysters, potentially posing a threat of infection to persons who eat raw oysters. However, it was not known how long oocysts would remain infectious in estuarine or marine waters or whether oysters filtered and retained oocysts in nature.

To determine the survival time of oocysts in seawater, we incubated oocysts in artificial seawater of various salinities and at various temperatures for up to 12 weeks, periodically determining their infectivities in mice. Satisfied that oocysts could survive for 12 weeks at 10°C in salinities favored by oysters, we conducted a study to determine whether oysters in natural waters actually harbored such oocysts. Oysters were collected twice from each of six sites in rivers emptying into the Chesapeake Bay where domestic sewage was discharged or runoff from cattle farms or residential septic tanks was likely to enter the water. Based on our previous findings (7), the hemocytes and gill washings from all oysters in the present study were examined for the presence of oocysts. Additionally, from each of the six sites, pooled hemocytes and pooled gill washings were administered to neonatal mice to determine whether oocysts were infectious for mammals. There are two genotypes of C. parvum, one transmitted from humans to humans and another transmissible among humans, cattle, mice, and possibly other animals (14). Thus, administering pooled hemocytes and gill washings to mice to determine infectivity, we knowingly may have underestimated the occurrence of the genotype that cycles only in humans, since that genotype does not infect mice.

MATERIALS AND METHODS

Source of oocysts.

Oocysts of C. parvum AUCP-1 were purified from calf feces and stored for less than 2 weeks in deionized water at 4°C as described previously (6, 11).

Survival of oocysts in seawater.

To determine how long oocysts of C. parvum retained infectivity in seawater, they were bioassayed in 5- to 7-day-old BALB/c mice as described previously (7). Forty 1.6-ml microcentrifuge tubes were prepared, each containing 3 × 10⁶ purified C. parvum oocysts (6, 11) suspended in 1 ml of water. Ten tubes contained deionized water. Of the 30 remaining tubes, 10 each contained artificial seawater (Forty Fathoms Bio-Crystals Marine Mix; Marine Enterprises, Baltimore, Md.) with 10, 20, and 30 ppt of salt, respectively. All tubes were capped and incubated at a constant temperature in circulating water baths in the dark (model 9101; Polyscience, Inc., Niles, Ill.). Five tubes of each salinity were incubated at 10°C, and the remaining five of each salinity were incubated at 20°C. One tube for each variable was removed 1, 2, 4, 8, and 12 weeks after incubation, and 150 μl containing 1.5 × 10⁵ oocysts was aspirated and administered by gastric intubation to each of five mouse pups within a litter. Oocysts remaining in each tube were examined microscopically and counted with the aid of a hemacytometer to ensure that 150,000 oocysts had been administered to each mouse. Mice were euthanized 96 h later by CO₂ asphyxiation, the ileum was removed from each mouse, and histologic specimens were prepared. Oocysts were considered infectious if developmental stages of C. parvum were found in epithelial cells by microscopy.

Recovery of oocysts from artificially exposed oysters.

To determine the level of exposure detectable by the methods we would employ for examination of naturally infected oysters, the following experiment was conducted. Twenty-four oysters were obtained from a retail seafood market. Six were placed into each of three 38-liter aquaria containing artificial seawater (salinity, 12 ppt) at 20°C, and six were processed immediately as described below. Oysters were exposed to 1,000, 100, or 10 oocysts per oyster in each of the three aquaria, respectively, and removed from the aquaria 72 h after exposure. Oysters were shucked, and gills were removed with scissors and placed in a capped centrifuge tube with 5 ml of phosphate-buffered saline. Tubes were shaken for 15 s with the aid of a Vortex-Genie (Scientific Industries, Inc., Bohemia, N.Y.), gills were removed, and 200 μl of the gill washing was placed on a glass microscope slide to dry; two slides were prepared from each oyster. Slides were stained with MeriFluor fluorescein-labeled anti-Giardia and anti-Cryptosporidium monoclonal antibodies (Meridian Diagnostics, Cincinnati, Ohio), and the entire area was examined with an epifluorescence microscope equipped with a fluorescein isothiocyanate-Texas Red dual-wavelength filter.

Oyster collection sites.

At each of six sites chosen from Maryland Department of Natural Resources and Maryland Department of the Environment (MDE) charts of oyster bars, 35 market-size oysters (>7.5 cm wide) were collected with a dredge or with hand tongs from vessels after sounding or trial dredge tows to locate the bar (Table 1). Sites 1 to 3 were adjacent to wastewater outfalls, and sites 4 to 6 were near cattle farms. A portion of the shoreline at site 4 had approximately 12 homes with septic tank sewage systems. Oysters were collected between 14 May and 2 June 1997 and again between 5 August and 19 August 1997, except for site 6, from which oysters were collected on 18 September (Table 1). At sites 1 to 5 oysters were collected from research vessels. At site 6, on the Wicomico River, oysters were collected from a private vessel. Sites 1, 2, and 6 were closed to commercial harvesting of oysters; sites 3, 4, and 5 were open.

TABLE 1.

Oyster collection sites in the Chesapeake Bay tributaries, with water quality data and fecal coliform counts

Site no.	River	Proximity to outfall, farm, or residential area (nm)a	Name of oyster bar	Latitude/longitudeb	Water temp (°C)		Salinity (ppt)		Coliform counte
					M–Jc	A–Sd	M–J	A–S	M–J	A–S
1	Choptank	Cambridge outfall (0.1)	Shoal Creek	38°34.45′/76°03.39′	17.0	26.0	7.0	11.8	1.0	3.6
2	Severn	Annapolis outfall (<1)	Chinks Point	38°57.93′/76°27.60′	18.8	26.0	7.3	12.0	23.0	9.1
3	Miles	St. Michaels outfall (1)	Ash Craft	38°47.74′/76°12.57′	16.0	25.0	8.0	13.8	9.1	3.6
4	Wye	Cattle farm (<6), residential (<1)	Bryan	38°53.47′/76°10.42′	17.0	26.0	7.4	11.9	1.0	3.6
5	Potomac	Cattle farm drainage (<1)	Cedar Point	38°09.04′/76°29.87′	18.1	28.5	10.2	14.5	3.6	1.0
6	Wicomico	Dairy farm drainage (<1)	Benton Lease	38°15.00′/75°50.35′	21.0	25.0	7.9	12.9	3.6	43.0

^anm, nautical miles. 

^bAs recorded from a Global Positioning System monitor. 

^cM–J, May-June collection period. 

^dA–S, August-September collection period. 

^eMPN/100 ml of seawater. 

At each site during May and June, oysters were collected and placed in mesh bags with identifying labels. Bags were held in insulated coolers aboard the vessels, refrigerated at the end of each day, and taken to the U.S. Department of Agriculture laboratory in Beltsville, Md., for processing within 5 days or fewer of collection. During the August collection period, oysters from five sites were processed aboard the research vessel immediately following collection. Benthic water temperature and salinity measurements were obtained from all sites (Table 1). Fecal coliform counts were obtained for water at each site within one month preceding our collecting of oysters. Counts presented in the form of the most probable number (MPN) were obtained from the MDE (Baltimore, Md.). Counts were derived from triplicate three-tube dilutions of water utilizing procedures accepted by the National Shellfish Sanitation Program for the examination of seawater and shellfish (1).

Examination of oysters for Giardia and Cryptosporidium spp.

Thirty live oysters from each site were selected for examination. To determine whether cysts or oocysts were taken into the body and phagocytosed by hemocytes, a 20 gauge needle attached to a syringe was inserted into the adductor muscle through a hole drilled between the valves. Approximately 3 to 5 ml of hemolymph was aspirated into a syringe, and 200 μl was placed on a glass microscope slide. Hemocytes were allowed to settle and attach to the glass. Slides were dried overnight before staining and examination. To determine whether oocysts had collected on the gills, oysters were shucked, gills were removed, and the same procedures described above were followed for examination of gills from artificially exposed oysters. Positive control slides for C. parvum were prepared by drying a 200-μl aqueous suspension containing approximately 2 × 10⁵ C. parvum oocysts purified from calf feces on glass microscope slides. Positive control slides for Giardia were prepared by smearing unpurified bovine feces containing Giardia cysts on glass microscope slides and then drying them. Control slides were then processed and examined in the same manner as slides with hemocytes and gill washings. Fluorescing ovoid bodies 8 to 12 μm in diameter were considered a positive finding of Giardia cysts. Fluorescing round bodies 4.5 to 5.5 μm in diameter with a distinctly brightly staining peripheral wall were presumptively considered to be C. parvum oocysts. Confirmative identification for C. parvum cannot be made without species-specific molecular markers or infectivity tests in mammals.

Bioassay for infectious oocysts.

To determine whether oysters harbored infectious C. parvum oocysts, hemolymph from 30 oysters collected in August and September at each of the sites was pooled; gill washings from these oysters were also pooled. After centrifugation at 1,500 × g for 15 min, supernatant was aspirated and pelleted hemocytes or gill wash debris from each site was administered to a litter of three to seven neonatal BALB/c mice by gastric intubation with a 26 gauge gavage needle. Mice were euthanized by CO₂ asphyxiation 96 h later, and the ileum from each was collected and examined by bright-field microscopy for developmental stages of C. parvum.

RESULTS

Survival of oocysts in seawater.

Oocysts of C. parvum held for 12 weeks at 10°C in fresh water and artificial seawater at concentrations of 10, 20, and 30 ppt remained infectious for mice (Table 2). Those held at 20°C were infectious at salinities of 0 and 10 ppt for 12 weeks, 20 ppt for 4 weeks, and 30 ppt for 2 weeks (Table 2). These test salinities overlapped those of waters where oysters were found. The test temperatures overlapped those for the May and June collection period but fell below those recorded for the August and September period.

TABLE 2.

Survival of oocysts of C. parvum AUCP-1 after 1, 2, 4, 8, and 12 weeks of incubation in artificial seawater at different salinities and temperatures

Salinity (ppt)	Survival of oocystsa after:
	1, 2, and 4 wks at:		8 wks at:		12 wks at:
	10°C	20°C	10°C	20°C	10°C	20°C
0	5	5	5	5	5	5
10	5	5	5	5	5	5
20	5	5	5	4	5	0
30	5	5	5	0	5	0

^aValues represent the number of mice (of five inoculated) with developmental stages of C. parvum in intestinal tissues 96 h after oral inoculation with incubated oocysts. 

Recovery of oocysts from artificially exposed oysters.

Of 24 oysters, 6 control oysters were negative for oocysts. Of the 6 oysters each exposed to 10, 100, or 1,000 oocysts and examined 72 h later for oocysts in gill washings 3, 6, and 6 oysters were found to be positive for oocysts, respectively. The number of oocysts observed by immunofluorescence microscopy in each of two 200-μl aliquots of gill washing from each oyster is shown in Table 3.

TABLE 3.

Number of oocysts detected by immunofluorescence microscopy in each of two 200-μl aliquots of gill washings from 24 oysters (6 control oysters and 6 oysters each in three artificial seawater aquaria) 72 h after exposure to 10, 100, or 1,000 C. parvum oocysts

Oyster no.	No. of oocysts detected with the following no. of oocysts seeded per oyster in each of three aquaria
	Controls	10	100	1,000
1	0, 0	1, 1	8, 5	46, 26
2	0, 0	2, 0	6, 3	28, 22
3	0, 0	2, 0	3, 1	17, 15
4	0, 0	0, 0	2, 1	11, 7
5	0, 0	0, 0	1, 0	8, 6
6	0, 0	0, 0	1, 0	2, 2

Oyster collection sites.

Measurements of water salinity at oyster bars obtained during collections in May and June ranged from 7.0 to 10.2 ppt, and water temperatures ranged from 16 to 21°C (Table 1). Those in August measured 11.8 to 14.5 ppt and 25 to 28.5°C, respectively (Table 1). Fecal coliform counts obtained from water collected at each site varied greatly between sites and between times of collection at the same site (Table 1).

Examination of oysters for Giardia and Cryptosporidium.

Although Giardia cysts were detected in control specimens with MeriFluor reagents and fluorescence microscopy, objects resembling Giardia cysts were not detected in any hemocytes or gill washings from any of the oysters collected.

Bodies fitting the criteria for C. parvum oocysts were found in some oysters at all sites. Among these sites, the percentage of oysters with oocysts in gill washings and/or hemocytes ranged from 16.7 to 60% in the spring and from 6.7 to 86.7% in the summer (Table 4). Oysters found positive for oocysts in gill washings were not always positive for oocysts in hemocytes and vice versa. No clear pattern was apparent between oysters harboring oocysts at sites impacted by wastewater outfall and those impacted by animal agriculture runoff.

TABLE 4.

Number of oysters (n = 30) positive for Cryptosporidium oocysts in gill washings and hemocytes at each Chesapeake Bay site during May and June and August and September collections as determined by immunofluorescence microscopy, size measurements, and staining patterns^a

Site	No. of oysters positive
	May and June collection			August and September collection
	Gill washings	Hemocytes	Totalb (%)	Gill washings	Hemocytes	Totalb (%)
1	0	5	5 (16.7)	3	4	6 (20.0)
2	10	2	12 (40.0)	0	26	26 (86.7)
3	7	8	13 (43.3)	1	1	2 (6.7)
4	10	11	18 (60.0)	11	2	12 (40.0)
5	10	5	12 (40.0)	4	18	21 (70.0)
6	10	2	11 (36.7)	3	1	4 (13.3)

^aSite names are shown in Table 1. 

^bTotal number of positive oysters. 

Bioassay for infectious C. parvum oocysts from oysters.

Of all the mice that received either pooled hemocytes or pooled gill washings from oysters at sites 1 to 6, developmental stages of C. parvum were found in the ileal epithelial cells of only two of three mice that received pooled gill washings from site 4 (data not shown).

DISCUSSION

In a previous study involving oysters, no oocysts were found within hemocytes or on gills of 80 unexposed control oysters and none were found in histologic sections prepared from 20 other controls, suggesting that these oysters harvested commercially from the Tred Avon River had not been contaminated with C. parvum (7). However, results from that same study indicated that oysters in an aquarium containing artificial seawater could readily remove oocysts from the water and retain them in hemocytes, on gills, and within the lumena of the stomach, intestine, and digestive diverticula (7). Furthermore, oocysts held within the oysters retained infectivity for mice for at least a week after exposure, suggesting that if any mammal, including humans, ingested raw oysters harboring such oocysts they could become infected with C. parvum (7). This was the first report of a bivalve mollusk potentially serving as a mechanical vector for a coccidian parasite of humans (7). Oysters were found not to be alone in this ability to filter out and sequester oocysts in their bodies. A subsequent study with freshwater clams, Corbicula fluminea, exposed to oocysts of C. parvum under similar artificial conditions demonstrated that they also filtered oocysts from the water and retained them in an infective state (9).

The present study has moved these observations from the laboratory into the natural habitat. First, we demonstrated that oocysts can survive for relatively long periods in seawater at salinities and temperatures overlapping those in which oysters live. Then, we determined that oocysts that could be detected in gill washings (within 72 h after exposure) of half the oysters exposed to as few as 10 oocysts and in gill washings of all those exposed to 100 or more oocysts. Then, we detected oocysts in hemocytes and gill washings from oysters harvested from natural waters at all six sites examined. Finally, we recovered oocysts of C. parvum from naturally exposed oysters and demonstrated that they were infectious in neonatal mice. It is important to emphasize that the oysters examined in the present study were from sites selected because of their close proximity to possible sources of contamination (Table 1), with three of the sites open to commercial harvesting and three closed. However, the site that yielded infectious oocysts (site 4) was an open site, and water at that site had a low coliform count within the month before oysters were collected (Table 1), suggesting that a high fecal coliform count may not necessarily be a good indicator for the presence of C. parvum.

Our examination of a single 200-μl aliquot of hemolymph or gill washing per oyster underestimates the number of positive specimens in oysters with low numbers of oocysts, as demonstrated in Table 3, in which oysters 2 and 3 exposed to 10 oocysts and oysters 5 and 6 exposed to 100 oocysts would have been considered negative if duplicate aliquots had not been examined. Because oysters in some sites may be exposed to oocysts only occasionally or intermittently, we probably further underestimate the prevalence of exposure. The number of oocysts in gill washings did not always correspond with the number found in hemocytes from the same oyster. For example, from the August and September collection at site 2, oocysts were not detected in gill washings from 30 oysters, whereas oocysts were detected in hemocytes from 26 of these oysters (Table 4). Differences observed in numbers of oocysts detected in gill washings versus numbers detected in hemocytes may reflect differences in the time elapsed from exposure to oocysts and the time oysters were collected. Shortly after exposure one might expect to find more oocysts in gill washings than in hemocytes. After a period of time, if no subsequent exposure had occurred, the oocysts would be taken into the body and would be more numerous in hemocytes and less numerous on the gills. Those oysters continually exposed to oocysts or exposed shortly before harvest might have a more even distribution of oocysts in the hemocytes and on gills.

The finding of oocysts infectious for mice in oysters at only one of the six sites might greatly underestimate the actual number of sites where infective oocysts were present. The use of new molecular tools has suggested that C. parvum may consist of two strains or subspecies with possibly two distinct transmission pathways. One isoenzyme pattern or genotype has been associated with oocyst isolates transmitted only from humans to humans, whereas other oocyst isolates, with a second isoenzyme pattern or genotype, have been associated with bovines, but this genotype is also capable of infecting humans and mice. Isoenzyme typing differences were observed for phosphoglucomutase and hexokinase (2). Gene or regional differences were reported for randomly amplified polymorphic DNA analysis of an nonspecified region (13), restriction fragment length polymorphism (RFLP) analysis of a repetitive DNA sequence (3); the 18S rRNA gene (4), RFLP analysis of an oocyst wall protein gene (16), the bifunctional dihydrofolate reductase-thymidylate synthase gene (17), and the thrombospondin-related adhesion protein (TRAP-C2) gene (14). Based on genetic polymorphism at the TRAP-C2 locus, oocysts of genotype 1 were recovered only from humans and were not infectious for calves or mice (14), whereas oocysts of genotype 2 were recovered from bovines and a subset of humans who had direct exposure to bovines or ingested items possibly contaminated with bovine feces. Furthermore, oocysts of genotype 2 were infectious for calves and mice. It is possible that oocysts observed in oysters at sites 1 to 3, 5, and 6 were infective oocysts of genotype 1 and therefore did not demonstrate infectivity for mice in our bioassay and that oocysts from oysters at site 4 were of genotype 2 and therefore were infective for mice. If enough oocysts can be recovered from oysters in future collections they will be subjected to genotypic analysis. It is also possible that the noninfectious oocysts represented species other than C. parvum, a situation that also would be clarified by genotyping.

Because bivalve mollusks such as oysters and clams are filter feeders, feeding on plankton and a variety of microflora, and are also important seafood, often eaten raw by humans, they pose a potential public health risk when they concentrate pathogenic bacteria and viruses from polluted waters. Salmonella typhi has historically been the most significant microbial contaminant with respect to shellfish-borne epidemics (12). However, Vibrio vulnificus, Campylobacter jejuni, hepatitis A virus, and Norwalk virus have also been transmitted to humans who have eaten contaminated shellfish (5, 12). Vibrio, Pseudomonas, Achromobacter, and Cytophaga/Flavobacterium were the predominant genera of organisms found in eastern oysters in two harvesting sites in the Chesapeake Bay (12). Although no human cases of cryptosporidiosis have been linked to ingestion of raw shellfish, the present study clearly indicates that the potential for such transmission exists and that ultimately the source of this organism is also a source of the other pathogens, namely, natural waters contaminated with human and animal feces.

Giardia cysts were not detected in hemocytes or gill washings from oysters at any of the collection sites, although they were detected in control specimens by the same techniques applied to detection in oyster specimens. Lack of detection suggests that (i) cysts were not present because oysters did not remove them from the water, or they were not present in the water, or (ii) cysts were present but in numbers below our level of detection or were masked from detection by unknown factors in the hemolymph or gill washing milieu.