Perspectives On Animal Research

Volume 1, Supplement

An Evaluation of Ten Randomly-Chosen Animal Models of Human Disease

Bacterial Endocarditis (Opossums) (1)

Description of the Model:

Rowlands described the induction of Streptococcus viridans endocarditis in opossums as a model of human bacterial endocarditis.(1) Noting that other researchers had observed spontaneously-occurring endocarditis in captive opossums, he reasoned that opossums were unusually susceptible to endocarditis. Endocarditis had been induced in other animals, but only after surgical damage to heart valves or the creation of a large arteriovenous shunt. Rowlands explained:

Because of the apparent natural susceptibility of opossums to bacterial infection, it seemed reasonable to attempt induction of bacterial endocarditis in adult opossums. This paper reports on the successful induction of bacterial endocarditis in unmodified captive opossums with a single injection of Streptococcus viridans.(2)

Each of the test animals received a single 1-ml injection of 52 million S. viridans organisms, which had been originally isolated from an opossum with endocarditis.

Criterion I: Concordance between the Animal Model and the Human Disease Clinical Presentation:

Rowlands concluded, "Naturally occurring and induced bacterial endocarditis in opossums is similar to that in humans both with regard to the morphology of the individual lesions and their distribution on heart valves."(1) However, the natural history of S. viridans endocarditis in opossums differs dramatically from that in humans. Sherwood et al., studying spontaneous endocarditis in captive opossums, noted, "In most instances, these opossums showed no evidence of disease until they died suddenly 3 weeks to several months after capture."(3) Three of seven animals with culture-positive valvular vegetations had alpha-hemolytic streptococci. The researchers did not do cultures to determine if these were S. viridans.

In contrast to the rapidly fatal infection in previously healthy opossums, human S. viridans endocarditis has an indolent course and occurs in patients who have almost always had previous valvular damage. The condition is termed "subacute bacterial endocarditis," reflecting its chronic course. This was the leading cause of human endocarditis prior to the antibiotic era.(4) Patients presented with a history of low-grade fever and progressive wasting over many weeks. Today, S. viridans endocarditis is uncommon, in part because of improved antibiotic treatment of patients with rheumatic heart disease and the use of prophylactic antibiotics for surgical and dental procedures in patients with damaged heart valves.

Rowlands noted that S. viridans endocarditis tended to be left-sided in opossums, as it is in humans. In humans, the left-sided valves are more likely to be damaged in rheumatic heart disease, and this accounts for predisposition of the left heart valves to S. viridans vegetations. However, the experimental endocarditis occurred in opossums that presumably had heart valves normal for their species.

Pathogenesis:

The animals were obtained from the wild because little was known about the care of captive opossums. Consequently, their ages and past medical histories were unknown. There was no way to determine if environmental pollutants predisposed these animals to infection, or if old or previously ill animals were more likely to get caught in the traps.

The control animals, which were not inoculated with bacteria, developed other infections. Rowlands observed, "The distribution of myocarditis and meningitis ... suggests that these changes are not specifically related to bacterial endocarditis ... glomerulonephritis occurred with an appreciable frequency in both experimental and control groups ..."(2) Indeed, four of the twelve control animals, which were not inoculated with bacteria, died during the experiment. This is a high mortality rate, presumably much higher than occurs in free opossums. Two of these four animals had valvular lesions. Because there was a high rate of significant diseases in all the animals, including the control subjects, there were other factors besides endocarditis that influenced the clinical course of these animals.

One possible explanation for the high rate of endocarditis and other causes of mortality among the control animals was that the animals were under great stress. Formerly free-roaming animals were placed in small cages with 3.06 square feet of floor space. They were housed singly to avoid contagious spread of disease, thus depriving them of social contact. The effect of handling, if even for a single injection, may have been profound.(5) The effect of stress on the immune system's ability to fight disease is well documented in animals and in humans.(6,7) For example, Angrist et al. tumbled rats up to about 25,000 times in up to 35 sessions, surgically joined animals together, and exposed rats to cold and to high altitude simulation and cold, and they found an increased incidence of valvular lesions, including infective vegetations.(8) Rowlands himself noted, "Both hemodynamic changes and stress have been thought to play significant roles in the initiation of valvular injury. An effort was made to minimize these factors in our experiments."(2)

Another possible factor that may have contributed to high mortality among the control animals was inadequate diet. Sherwood et al. found that the cat food diet "...seems adequate for maintenance of the animals..." but acknowledged that "...little is known about the dietary requirements of the opossum."(3) Whatever the causes, there were conditions other than S. viridans inoculation that predisposed these animals to endocarditis and other causes of death.

Humans with normal heart valves almost never acquire S. viridans endocarditis, despite the fact that S. viridans may be introduced following dental procedures, tonsillectomy, use of water jet, or chewing food.(9) Thus, there were fundamental differences between these opossums and humans. These "normal" opossums were at risk for infection, but the "normal" human is not.


Criterion II: Citations

There were nine references in Science Citation Index from 1970 to 1988 to the article by Rowlands et al.(2) (See Appendix A.) None of the nine articles used human subjects or tissues.


Criterion III: Historical Impact

No clinical review article in the Science Citation Index cited the original article by Rowlands et al.(2) We looked at two recent articles to see if this animal model appeared to be of clinical value. Sussman et al. reported 12 cases of S. viridans endocarditis in humans, and their discussion included a review of the literature.(9) They cited no animal research. Oikawa and Kaye had 109 citations in their 1986 review entitled, "Endocarditis: epidemiology, pathophysiology, management, and prophylaxis".(10) There were two papers cited that involved animal research. One assessed the susceptibility of rabbits with experimentally induced sterile endocardial thrombi to S. viridans infection. The other involved in vitro studies with traumatized rabbit valves and human plasma. There was no reference to the opossum endocarditis model.


Conclusions:

Experimental endocarditis in opossums bears a superficial resemblance to human endocarditis in its predilection for left-sided heart valves and a similarity in the gross appearance of the vegetations. Nevertheless, there are fundamental differences between the experimental disease and the human disease in both pathogenesis and natural history. Therefore, it is highly unlikely that the opposum model can provide meaningful insights into human endocarditis. Furthermore, because of differences in clinical presentation, it would be hazardous for clinicians to use studies with this model to guide the treatment of endocarditis in man. In humans, S. viridans endocarditis is a chronic disease that almost always involves previously damaged heart valves, while endocarditis in the opossum model has an acute course and apparently involves healthy heart valves. The lack of citations in the clinical literature evidently reflects the difficulty in applying this model to the understanding or treatment of human bacterial endocarditis.

References

1. Rowlands DT: Bacterial endocarditis, Model No.4, in Jones TC Hackel DB, Migaki G (eds): Handbook: Aninial Models of Human Disease, Fascicle 1. Washington, DC, Registry of Comparative Pathology, Armed Forces Institute of Pathology, 1972.

2. Rowlands DT, Vakilzadeh J, Sherwood SF, LeMay JC: Experimental bacterial endocarditis in the opossum (Didelphis Virginia). Am J Pathol 1970;58:295-304.

3. Sherwood BF, Rowlands DT, Hackel DB, LeMay JC: Bacterial endocarditis, glomenslonephritis, and atnyloidosis in the opossum. Am J Pathol 196853:115-126.

4. Braunwald E, Isselbacher KJ, Petersdorf RG, Wilson JD, Martin JB, Fauci AS (eds): Harrison's Principals of Internal Medicine 11th Ed. New York, McGraw-Hill, 1987.

5. Marsh JT, Rasmussen AF: Response of adrenals, thymus, spleen and leukocytes to shuttle box and confinement stress. Proc Soc Exp Biol Med 1960;104:180-183.

6. Gross W: Effect of social stress on occurrence of Marek's disease in chickens. Amer J Vet Res 1972;33:2275-2279.

7. Mechanic DM: Medical Sociology, 2nd Ed. New York, The Free Press 1978.

8. Angrist AA, Oka M, Nakao K, Marquiss J: Studies in experimental endocarditis. Am J Pathol 1960;36:181-199.

9. Sussman JI, Tenenbaum MJ, Greenspan J, Tyburski MB, Kanzer BF: Viridans streptococcal endocarditis: Clinical, microbiological, and echocardiographic correlations. J Infect Dis 1986;154:597-603.

10. Oikawa JH, Kaye D: Endocarditis: Epidemiology, pathophysiology, management, and prophylaxis. Cardiovasc Clin 1986;16:335-357.