Factors Causing Difficulties in Uniformity of Results Among Testing Facilities in Microbiologic Monitoring of Laboratory Animals
Toshio Itoh
Deputy Director, ICLAS Monitoring Center
Central Institute for Experimental Animals
Kawasaki, Japan
It is important to select test items for harmonization of microbiologic monitoring of laboratory animals, but selection alone is not sufficient. These tests will not be effective unless the methods, including sampling, testing techniques, and expression of results, are uniform and the items and methods are considered as a set. I believe that many factors are involved in disparate results among different testing facilities. At the Central Institute for Experimental Animals (CIEA), we have encountered discrepancies in results among testing facilities involving Pasteurella pneumotropica, Clostridium piliforme (Tyzzer's organism), and Hantavirus.
DISCREPANCIES IN RESULTS
Pasteurella pneumotropica
P. pneumotropica is a pathogen that has been found in the respiratory tract of mice and rats, but until recently, it has not been considered important as a pathogen. In culture tests, discrepancies in results have often been observed. To test for this organism, we coat swabs from the trachea and conjunctiva on horse blood agar plates and ultimately identify suspected colonies by Gram staining, testing of characteristics by the API system, and DNA testing by polymerase chain reaction (PCR). Identification is often difficult, however, because there are many analogous bacteria with slightly different properties from this organism. Even
when kits are used, difficulties in accurate identification often arise in testing facilities with little experience or in animal facilities with small scale testing.
In one facility in a medical school where P. pneumotropica was listed in the quarantine protocol for the barrier facility, tests were performed by PCR because of difficulty in identification by the conventional method. A swab from the conjunctiva was cultured in a liquid medium for bacterial growth, and the specific sequence of 16SrDNA of this organism was then detected by PCR. This method was used for quarantine because good results had been obtained in an experimental study and live animals could be tested. However, when the method was used in actual quarantine, the number of positive animals immediately increased and the introduction of new animals became difficult. When we tested some of the animals that were positive in the first test, we obtained consistently negative results. It is evident that results of bacteriologic tests using cultures will show discrepancies if there are differences in sampling sites, test methods, and identification criteria.
We also have encountered prolems in antibody testing, which we use for microbiologic monitoring of laboratory animals. The ICLAS Monitoring Center typically uses enzyme-linked immunosorbent assay (ELISA) as a screening test followed by the indirect immunofluorescent antibody (IFA) test or another method (such as the immunoblot method, hemagglutination inhibition test, or neutralization test) on serum suspected of being positive. When the sample is positive in both tests, it is evaluated as positive.
Clostridium piliforme (Tyzzer's organism)
Tyzzer's disease, which is characterized by diarrhea, focal necrosis in the liver, or death, has recently occurred in rats and rabbits at several breeding facilities in Japan. Sporadic cases of laboratory animals positive for this organism have also appeared in Japan.
Because it is difficult to culture this organism, we commonly detect the infection by using an antibody test such as complement fixation (CF), IFA, or ELISA. Generally, in antibody tests on mice and rats, the CF method is less sensitive and results in many false negatives, whereas ELISA is very sensitive and seldom results in false-positive cases.
At the ICLAS General Assembly held in May 1999, it was reported that a breeder of laboratory animals who first obtained antibody-positive test results from a testing facility and destroyed the animals later learned that the test results were incorrect, which caused a major loss. We also had tested these sera and had obtained negative results, and I have heard that these samples were also found to be negative when tested by an American testing facility.
The ICLAS Monitoring Center distributes an ELISA antibody test kit for Tyzzer's disease. Serum samples suspected of being antibody positive are sent to the Center where a confirmation test using IFA is performed. According to the
confirmation test, approximately one fourth of the specimens have actually been positive. I suspect that major problems have not occurred in Japan concerning serologic checking of this organism because this confirmation test is performed.
With the usual serologic test methods for Tyzzer's disease, ELISA and IFA, there appear to be no marked differences in use among testing facilities. However, a detailed investigation revealed differences in strains, antigen preparation, and evaluation of results. It had been found previously that false positives occurred even in Tyzzer's disease tests by ELISA. False-positive reactions can be divided into nonspecific reactions that occur in systems and specific reactions due to cross-reactions with organisms having common antigens. The following example pertains to the latter type
As shown in Table 1 and Table 2, we checked serum samples that were ELISA positive and IFA negative for Tyzzer's organism. All of these rat sera reacted with Clostridium spiroforme and most rabbit sera reacted with Clostridiurn clostridiforme, the two clostridial species that are components of these animals ' intestinal flora. Results indicated that cross-reactions with C. spiroforme in rats and C. clostridiforme in rabbits are one of the causes of false-positive reactions
TABLE 1 Cross-Reaction in ELISAa for Detection of Antibody to Clostridium piliforme (Tyzzer's organism) in Rat Serab
Tyzzer's Organism |
Clostridium clostridiforme |
Clostridium spiroforme |
||||
Sample no. |
ELISA |
IFAa |
ELISA |
IFA |
ELISA |
IFA |
A - 1 |
0.523 |
− |
− |
− |
1.868 |
+ |
2 |
0.435 |
− |
− |
− |
1.096 |
+ |
3 |
0.704 |
− |
− |
− |
0.815 |
+ |
4 |
0.619 |
− |
− |
− |
1.254 |
+ |
5 |
0.577 |
− |
0.332 |
− |
1.743 |
+ |
6 |
0.600 |
− |
1.084 |
+ |
0.509 |
+ |
B - 1 |
0.691 |
− |
− |
− |
1.946 |
+ |
2 |
0.401 |
− |
− |
− |
1.747 |
+ |
3 |
0.757 |
− |
− |
− |
1.348 |
+ |
C - 1 |
0.301 |
− |
0.334 |
+ |
2.408 |
+ |
2 |
0.309 |
− |
− |
− |
1.910 |
+ |
3 |
0.348 |
− |
0.626 |
+ |
1.751 |
+ |
4 |
0.379 |
− |
0.409 |
+ |
1.582 |
+ |
D - 1 |
0.699 |
− |
0.315 |
− |
2.358 |
+ |
2 |
0.820 |
− |
0.656 |
− |
2.554 |
+ |
3 |
1.237 |
− |
− |
− |
1.517 |
+ |
aELISA, enzyme-linked immunosorbent assay; IFA, immunofluorescent antibody. bELISA: OD492 value of 0.3 or higher was regarded as positive; IFA: Antibody titer 1:10 or higher was regarded as positive. |
TABLE 2 Cross-reaction in ELISAa for Detection of Antibody to Clostridium piliforme (Tyzzer's Organism) in Rabbit Serab
Tyzzer's Organism |
Clostridium clostridiforme |
Clostridium spiroforme |
||||
Sample no. |
ELISA |
IFAa |
ELISA |
IFA |
ELISA |
IFA |
E - 1 |
1.934 |
− |
− |
− |
2.523 |
+ |
F - 1 |
0.320 |
− |
0.499 |
+ |
− |
− |
G - 1 |
1.109 |
− |
1.161 |
+ |
1.069 |
− |
H - 1 |
0.449 |
− |
0.378 |
+ |
0.599 |
+ |
I - 1 |
0.409 |
− |
1.151 |
+ |
3.000 |
+ |
J - 1 |
0.495 |
− |
1.585 |
+ |
− |
− |
K - 1 |
0.812 |
− |
1.722 |
+ |
− |
− |
aELISA, enzyme-linked immunosorbent assay; IFA, immunofluorescent antibody. bELISA: OD492 value of 0.3 or higher was regarded as positive; IFA: Antibody titer 1:10 or higher was regarded as positive. |
for C. piliforme in ELISA. Results also indicated that cross-reactions in ELISA could be differentiated by IFA. As long as a whole bacterial eluant is used as the antigen, it is difficult to avoid cross-reactions with other bacteria that have common antigens. For this reason, we use several methods concomitantly for each test item in antibody tests. From the previous ICLAS report, it appears that differences in the antigen strains used and in the testing systems caused the discrepancy in the results.
Hantavirus Testing by Serology
Hantavirus is a zoonotic agent for which rats are the reservoir. About 20 years ago, an outbreak of Hantavirus infection occurred in laboratory animal facilities in Japan. More than 100 animal caretakers and researchers were infected, and one died. Thereafter, thorough testing was undertaken, contaminated facilities were disinfected, and all animals were replaced. Contamination of animal experimentation facilities was eliminated in a short time. A description of our experience with antibody testing of this virus follows.
Antibody testing of this virus was possible in several facilities in Japan at the time of the outbreak. The ICLAS Monitoring Center sent technicians to Fort Detrick in the United States with an introduction from Dr. Allen, who is here today, and also to the testing facility for this virus in a university in Japan. We introduced an antibody testing technique using IFA, which currently, along with immunoblot and ELISA methods, is available.
Several years ago, we received a report from the university testing facility that rats in the laboratory animal facility of a medical school were Hantavirus
antibody positive. All animal experiments were immediately stopped and the animals destroyed. Later we were asked to test the samples, and the results were negative. Testing was performed using basically the same method except for a slight modification to inhibit nonspecific reactions; however, the results were different.
The results in the testing facility of the university were considered positive at an antibody titer of 1:16. They took the position of stating only that there was a positive reaction and deferring to the client to decide whether there was an infection, even though they were experts in this field. When we obtained results showing that antibody titers were in the 1:20 to 1:100 range (so-called low antibody titer) in the IFA test, we reported these results to the client while continuing the testing process. Inasmuch as there was little possibility of real infection, we collected blood samples after 1 week for retesting because generally in contaminated facilities, both the prevalence and antibody titer increase when retesting is performed. We have performed Hantavirus antibody tests on several thousand samples a year and have found only about 10 samples showing such low antibody titers. However, these samples were all found to be negative using immunoblot analysis.
The discrepancy between our test results and those of the university appeared to be caused by differences in the basic position of the testing facility when submitting the test results to the client. There are two different positions: One is simply to hand over the test results, and the other is to consider countermeasures after obtaining the results and assuring highly accurate test results. We naturally take the latter position.
CONCLUSION
The reasons for discrepancies in culture results include difficulty in accurate identification of isolates and differences in sample collection sites and tests for identification of bacteria. For the several test methods available in antibody tests, it is necessary to confirm the advantages and disadvantages of each method and select the method based on an overall evaluation of the factors. The position taken by the testing facility concerning its level of responsibility when submitting test results will also influence the results.
I am well aware that test methods are in a constant state of development, test facilities take pride in their techniques, and it is very difficult to achieve uniform test methods. I want to stress that simply listing the test items is not sufficient for international harmonization of microbiologic testing of laboratory animals. The test items must include the test method as well as a recommendation of usefulness. Even when there are no major differences in test methods, there are cases in which different results are presented to the client. These differences probably depend on the philosophy adopted by each testing facility.