Mortality in a Wood Turtle (Clemmys insculpta) Collection

Case Presentation

A reptile breeder with a collection of 130 turtles began to see high mortality rates among wood turtles (Clemmys insculpta) that had hatched between May and August 2000. The breeder had over 10 years of experience with this species, and historically 2–3% of the breeder's hatchling turtles did not survive to adulthood. In 2000, >40 of the hatchling turtles died between the months of September and October. Lethargy, inactivity, and periocular swelling were noted in several of the affected animals. Few hatchling turtles recovered after showing signs of disease; 5–10% of the adult turtles showed mild signs of illness, but none died.

Prior to the epizootic, the only major change in environmental conditions was the water supply. A new well had been installed in July 2000, and all areas of the facility were supplied with water from this well. No new animals had been introduced from outside the facility in the past year. Originally, 10–12 animals were housed together in tanks that were 15 in long, 20 in wide, and 4 in deep and contained 1 gal of water. The tanks were outdoors and above ground. The temperature remained between 69.8°F and 93.9°F between September 1 and September 30, 2000.¹ The breeder changed the tank water daily and routinely cleaned the tanks with a 5% bleach solution. When deaths began, the turtles were placed in separate containers. Turtles that became ill were treated by the owner with acriflavine (an antimicrobial) and were soaked repeatedly in bleach solution, but no improvement was seen.

On September 27, 2000, the breeder submitted 2 recently deceased wood turtles to the Department of Pathobiology at the College of Veterinary Medicine, University of Florida, for necropsy and histopathologic and microbiologic examination. Significant necropsy findings included moderate amounts of mucoid discharge within the nares of 1 turtle and mildly enlarged mottled pale tan to brown livers in both turtles. An impression smear of the liver was prepared and stained with Wright-Giemsa (Hareleco, EM Science, Gibbstown, NJ, USA) (Figure 1).

Based on necropsy and cytology results, surviving turtles were given 20 mg/kg metronidazole (Sidmark Laboratories, East Hanover, NJ, USA) orally once daily for 5 days. Five additional deaths occurred the week treatment was started; however, no other turtles showed signs of illness subsequently.

Cytologic Interpretation

The impression smear of the liver was moderately cellular with large amounts of necrotic debris and moderate numbers of RB Cs in the background. Low numbers of intact hepatocytes with abundant blue-gray mottled cytoplasm, moderately indistinct cell borders, round nuclei, somewhat diffuse chromatin, and single prominent nucleoli were observed individually and in cohesive clusters. Several hepatocytes contained large distinct cytoplasmic vacuoles. Low numbers of basophils, heterophils, and reactive macrophages with cytoplasmic vacuoles were noted. Frequent round unicellular organisms 27–32 μm in diameter with basophilic cytoplasm and small eccentrically located round nuclei approximately 5.5 μm in diameter were observed throughout the sample (Figure 1). The sample also contained several oval vacuolated cystlike structures approximately 16.5 μm wide by 30 μm long. A heterogeneous population of extracellular bacteria was noted. No intracellular organisms or neoplastic cells were observed. The round unicellular organisms were most consistent with amoeba trophozoites. Pathogenic amoebae, nonpathogenic amoebae, and some flagellates were considered as differential diagnoses because of the similar morphology of these organisms.²

Histologic Interpretation

Multifocal to focally extensive areas of hepatic necrosis were observed that occasionally extended to the capsular surface. Small to moderate aggregates of macrophages, heterophils, and lymphocytes were associated with the areas of necrosis. Scattered throughout the hepatic parenchyma, but especially common in the necrotic areas, were 10–15 μm in diameter round to slightly oval protozoal organisms. The organisms consisted predominately of trophozoites and were poorly stained in formalin-fixed sections stained with hematoxylin and eosin (Figure 2). The organisms had mildly to moderately vacuolated cytoplasm and nuclei containing chromatin plaques with a small, often indistinct endosome. Amoebae consistent with Entamoeba sp were observed in trichrome-stained liver samples, which had been emulsified and fixed overnight in polyvinyl alcohol before staining (not shown).Based upon these features and positive staining with periodic acid-Schiff (PAS) (Figure 3), amoebiasis was diagnosed.

Amoebae also were associated with mild to moderate diffuse necrotizing pneumonia, severe multifocal necrotizing enteritis, serositis, and adrenalitis. Cross sections of the head revealed mild to moderate conjunctivitis and rhinitis associated with large numbers of gram-positive bacterial rods detected with Brown-Brenn stain. Moderate to heavy growth of Aeromonas sobria was obtained from samples of the liver and coelomic cavity.

Discussion

Infection with some Entamoeba species can cause fatal disease in captive populations of reptiles. The major pathogenic amoeba in reptiles is E invadens.^2–5 Epidemics associated with E invadens have been reported in snakes³ and red-footed tortoises.⁴ Several amoeba species have been isolated from reptiles; however, a specific diagnosis cannot be made by morphologic examination of tissue sections or cytologic smears. The 2 common amoeba species isolated from humans are E dispar and E histolytica. Differentiation between the nonpathogenic E dispar and the pathogenic E histolytica can be made using polymerase chain reaction,^6–8 DNA hybridization,⁹ immunohistochemistry,¹⁰ or ELISA to detect parasite antigens¹¹ or antibody¹² against E histolytica. Any of these methods could be adapted to diagnose E invadens in reptiles, but none of them are commercially available. Immunohistochemistry using a polyclonal antibody against E invadens has been reported in snakes.² In the past, differentiation between E invadens and other species of amoebae has been based on host range and temperature sensitivity in culture.³

Amoebic infections in human beings occur worldwide, although there is increased prevalence of the disease in developing countries.¹³,¹⁴ More than 500 million people are thought to be infected with either E dispar or E histolytica.¹⁴ Approximately 40 million people infected with E histolytica develop disease,¹⁴ and 70,000 die annually.¹³ Ninety percent of people with intestinal amoebiasis develop dysentery or bloody diarrhea. Other clinical forms of intestinal amoebiasis include fulminating colitis, amoebic appendicitis, and amoebae in the colon.¹³ The most common extraintestinal form of amoebiasis is amoebic liver abscessation.¹³,¹⁴ This disease is characterized by severe radiating pain, cough, and fever.¹³ Other signs that may be associated with the disease include anorexia, nausea, vomiting, and diarrhea.¹³

In snakes, pathogenic and nonpathogenic amoebic infections are most common in the southeastern United States.⁵ Amoebiasis in reptiles is associated with anorexia, depression, listlessness, diarrhea, and dehydration.^3,4 In these affected turtles, the classic clinical signs associated with amoebiasis were not observed in the weeks before death, however, the large number of amoebae in the livers, the history of high mortality, and the resolution of the epizootic with metronidazole treatment pointed to amoebiasis as the primary cause of disease. Diarrhea may have been present but not observed by the breeder; however, only one-third of humans with amoebic liver abscesses have concurrent diarrhea.¹³,¹⁴

Pathogenic amoebae can spread rapidly through a reptile collection. Cysts shed in the feces of infected animals contaminate communal water or food and are ingested by other animals.³ Curators can inadvertently carry the cysts from one area to another on cleaning equipment and clothing.³ Cysts can develop into trophozoites and can invade host tissues if favorable conditions exist in the host.⁴ In experimental infections in snakes, E invadens caused disease at 25°C but not at 10°C or 30°C.⁴ Also, amoebae are more likely to invade tissues when the availability of nutrients (particularly carbohydrates) in the gastrointestinal tract is low.⁴ The trophozoites penetrate the mucus of the colon, the intestinal epithelium, and the lamina propria.^13,14 Pathogenic amoebae can migrate to the liver via the portal circulation.¹³,¹⁴ This invasion may lead to hepatic tissue necrosis, abscess formation, and death of the animal. Rupture of amoebic liver abscesses and dissemination of the infection prior to death have been reported in people.^13,14

The bacteria isolated from the liver and coelomic cavity of affected turtles were probably nonpathogenic. A sobria is a gram-negative, facultative anaerobic bacillus that is ubiquitous in fresh and brackish water.¹⁵ This organism also was cultured from the well water at the facility. Each turtle had continuous access to a pool of water in the tank, and the bacteria could have been isolated from any turtle in the facility. A sobria is a putative pathogen of human beings and fish¹⁵ and has been cultured from feces and blood of people with gastroenteritis and septicemia.¹⁵ However, Koch's postulate has not been satisfied for this organism.¹⁵

Oral metronidazole is an effective medical treatment for amoebiasis; however, it is often difficult to identify the disease prior to the loss of several specimens in a collection. Typically, several animals die before a postmortem diagnosis is made.^3,4 Cytologic examination of a liver aspirate is a simple, relatively noninvasive technique that can be used to diagnose hepatic amoebiasis before death. A combination of Romanowsky-type and PAS stains is highly recommended to aid in identification of the organisms. Formalin fixation causes the organisms to appear 2 times smaller in histologic samples compared with cytologic samples.