Clostridium perfringens Type A Food Poisoning

Transcription

1 INFECTION AND IMMUNITY, Jan. 1971, p Copyright 1971 American Society for Microbiology Vol. 3, No. I Printed in U.S.A. Clostridium perfringens Type A Food Poisoning II. Response of the Rabbit Ileum as an Indication of Enteropathogenicity of Strains of Clostridium perfringens in Human Beings DOROTHY H. STRONG, CHARLES L. DUNCAN, AND GIUSEPPE PERNA Food Research Institute, Departments of Food Science, Bacteriology, and Medicine, and University Health Services, University of Wisconsin, Madison Received for publication 22 June 1970 The effect of feeding human beings individual strains of Clostridium perfringens or culture filtrates thereof was examined. The strains selected for challenge included both those which had previously been shown to produce fluid accumulation in the ligated ileum or overt diarrhea when injected into the nonligated ileum of the rabbit, or had produced both, and those which did not regularly produce these responses. Challenge doses prepared by allowing each strain to grow in beef stew for 3 hr at 46 C resulted in a 61 % incidence of diarrhea when rabbit-positive cells were used. No diarrhea occurred among the subjects fed rabbit-negative strains prepared in a similar manner. The procedures employed in preparing the challenge dose appeared to influence the results obtained. When cell-free filtrates were fed, 4 of 15 persons consuming filtrates from rabbit-positive strains developed diarrhea. All subjects fed filtrates from rabbit-negative strains remained free from diarrhea. Serological tests were carried out to compare the identity of the strains of C. perfringens consumed by the subjects and those excreted in the feces. Heat resistance measured as Dloo values varied greatly among the rabbit-positive strains. Although significant progress concerning the lished between the accumulation of fluid in the etiology of human disease is frequently realized ileum of the rabbit and a diarrheal response in through the use of animal models, convincing monkeys (7). proof of any findings must ultimately depend on The purpose of this study was to test whether a demonstrated parallelism between the animal and positive response obtained after introduction of a the human response. In attempting to study the strain of C. perfringens or its cell extract or culture filtrate into the ileum of a young rabbit (5, 6) possible causative agent operating in the production of Clostridium perfringens food poisoning does denote a strain capable of producing the symptoms, various animals have been employed food poisoning syndrome in human volunteers. (11, 14, 18, 19, 23). The procedures for performing the rabbit test Recently, the use of ligated loops of intestine have been described elsewhere (8). has been proposed as an assay method for determining the enteropathogenicity of strains of C. term "rabbit-positive" is used to describe strains In this paper, as in the preceding report (7), the perfringens. Duncan et al. (8) employed this procedure, utilizing young rabbits as test animals. peatedly have been shown to produce in young of C. perfringens, or products thereof, which re- Almost simultaneously Hauschild et al. (12) described the application of the same technique in ileum or overt diarrhea subsequent to intraileal rabbits either fluid accumulation in the ligated lambs. The latter workers used a strain of C. perfringens which had been previously shown in their versely, "rabbit-negative" denotes the absence of injection of the nonligated gut, or both. Con- laboratory to have produced intestinal disturbance in five of six cases when fed to human beings such response in any regular pattern. MATERIALS AND METHODS (13). Their evidence supported the proposition that a positive response in a ligated loop of lamb Cultures. Eight rabbit-positive cultures of C. intestine indicated a strain of C. perfringens which perfringens type A were employed. These included the following, with the Hobbs serological type shown would produce symptoms commonly developed in parentheses: NCTC 8239 (H3), NCTC 8798 (H9), in C. perfringens food poisoning. NCTC (H12), and NCTC (H13), all More recently, a correlation has been estab- isolated and typed in the laboratory of Betty Hobbs 171

2 172 STRONG, DUNCAN, AND PERNA INFEC. IMMUN. and originally obtained from food or feces associated with human food poisoning. In addition, the following strains were tested: E-13 (H16) provided by H. E. Hall of the U.S. Public Health Service, 027 (serological type unknown) recovered in this laboratory from beef liver, and (serological type unknown) and (H10) isolated from foods associated with two food poisoning incidents in Madison, Wis. Rabbit-negative strains included NCTC 8247, also procured from B. Hobbs; F42 isolated at the U.S. Public Health Service laboratory in Cincinnati, Ohio; FD1 provided by Stanley Harmon of the Food and Drug Administration in Washington, D.C.; and 215b isolated in this laboratory. The original isolations of the rabbit-negative strains NCTC 8247 (H4), F42 (Hobbs type VII), and FD1 (not typable by Hobbs sera) were made from food poisoning-associated food or feces. Strain 215b (H18) was originally recovered from beef liver. Preparation of challenge dose. The cell suspensions were prepared as previously described (7). Two different carriers were used in presenting the challenge dose to the subjects. In the first series of tests, for each subject, 20 ml of cell suspension, providing as an average 3.3 X 1010 total viable cells and 2.5 X 108 spores, was added to 80 ml of chocolate-flavored dairy drink contained in a commercial-type wax carton. The cell suspension was added just before the consumption of the dairy drink by the subject. The dairy drink was at refrigerator temperature. In a second series of experiments, the carrier was commercially canned beef stew. Cans (7.5 oz) of the stew were placed in a incubator at 46 C in sufficient numbers so that one can was available for each subject for each test. In the first experiments in this series, the required number of cans of beef stew was removed from the incubator, each was flamed with alcohol and opened, the inoculum (average, 2.5 X 1010 total cells; 7.8 X 107 spores) was added to the contents of the can, and the mixture was stirred with a sterile wooden tongue depressor. Immediately thereafter the subject consumed the stew together with buttered bread and milk or coffee, or both. For all later experiments, the inoculum was added aseptically to the warm stew, with subsequent incubation of the stew at 46 C for 3 hr. The average counts at the time of feeding were 4.4 X 1010 total cells and 3.0 X 108 spores. Additional cartons of chocolate dairy drink or cans of beef stew were prepared exactly as were those consumed by the subjects. These were used to determine total viable cell and spore counts. For all count determinations, peptone water (0.1%) was used as the diluent, and pour plates were made with TSN (BBL) agar. Spore counts were made after heatshocking at 75 C for 20 min. Anaerobic incubation was at 46 C, under a 901% nitrogen, 10% carbon dioxide mixture. When cell-free preparations were substituted for viable cultures as the challenge dose, for each test, each volunteer was fed 8 g of lyophilized concentrated culture filtrate prepared from a single strain of C. perfringens. To prepare the 48-hr-old culture filtrates (concentrated culture supernatant fluids) for use as challenge feedings, the procedures described earlier were utilized (6). For feeding, the lyophilized preparation was resuspended in 100 ml of chocolate dairy drink. Submitted to the test were filtrates of rabbitpositive strains NCTC 8239, NCTC 8798, NCTC 10239, and NCTC 10240, rabbit-negative strains NCTC 8247, FD1, F42, and 215b, and uninoculated medium subjected to all of the procedures utilized in preparing the culture filtrates. Management of the subjects. All subjects, both male and female, were at least 21 years of age. Each person was examined by the cooperating gastroenterologist who also supervised hematological studies. Each prospective subject presented fecal specimens at the State Hygiene Laboratory for examination for parasites. Only when the physician had cleared the individual was he eligible to become an experimental subject. Presenting the challenge dose to the subjects was routinely done late on Tuesday afternoon. The subjects were requested to eat their evening meal as soon as possible after having consumed the challenge dose. In most instances, on the preceding Monday morning each subject submitted a preliminary fecal specimen. A second fecal sample was obtained during the diarrheal episode if such developed (DS); if diarrhea did not develop, the subject submitted a fecal sample from the first defecation to occur 24 hr after consuming the challenge dose (24-hr S). A final fecal sample was obtained on the 6th day after the ingestion of the challenge dose (6-day S). Total viable cell and spore counts for C. perfringens were made on all fecal specimens, and the results were expressed as counts per gram (dry weight) of fecal material. Fecal specimens were not obtained when the concentrated culture filtrates from NCTC strains 8247, 8798, 10240, and strain 215b were fed. Individuals served as subjects once every 14 days. For each feeding test, the subjects were divided into groups of four or five members. Each person within a group received an identical challenge dose. Each subject completed and returned a chart which described his response to the challenge dose. In total, 131 feeding tests were made. Serological typing of C. perfringens isolated from postchallenge fecal specimens. To assess whether the C. perfringens cells detected in the fecal specimens collected after oral ingestion of the challenge dose by the subject (i.e., the DS or 24-hr-S fecal samples) belonged to the strain which had been recently consumed, a limited number of serological typing reactions were carried out. Specific antisera were employed to confirm the identity of the C. perfringens isolates. In addition to using the antisera of the strain most recently fed, antisera for strains consumed in earlier feedings were employed, to test for possible "carry-over" of the strains. For controls, similar serological testing was carried out on strains isolated from feces of subjects before initial challenge. The serological identification of the strains was determined by the agglutination test (9). r' Heat resistance determinations. High-heat resistance of spores of C. perfringens has, in some instances, appeared to be a characteristic associated with the

3 VOL. 3, 1971 C. PERFRINGENS TYPE A FOOD POISONING. II 173 food poisoning strains. However, it has been reported that both heat-sensitive and heat-resistant strains may cause food poisoning (22). It was, therefore, of interest to determine whether the rabbit-positive and rabbit-negative strains, which were fed to the subjects during the course of these experiments, differed in this regard. The Dloo value (time in minutes for inactivation of 90% of the spores at 100 C) of spores of each strain was determined. Methods previously described (4) were used, except that TSN (BBL) instead of SPS (Difco) agar was used to determine viable counts. In the present experiments, the time required for the contents of the thermal death time tubes to reach 100 C ("come-up time") was 42 sec. RESULTS AND DISCUSSION The results presented in Table 1 indicated that, of the persons challenged by rabbit-positive viable cells of C. perfringens, 57% reported one or more of the symptoms commonly associated with C. perfringens food poisoning. Similarly, 21% of persons fed rabbit-negative cultures reported such symptoms. Closer inspection of the data suggests that the responses of the human subjects, when fed rabbitpositive or rabbit-negative strains of C. perfringens, were somewhat more divergent than first appeared (Tables 1 and 2). Occurrence of diarrhea is probably the most objective of the symptoms. Based upon this criterion, 26 or 45% of the persons tested showed signs of C. perfringens food poisoning when fed rabbit-positive strains, whereas only 1 of 29 or 3.4% of those tested responded similarly when rabbit-negative strains were fed. In the latter instance, the one person reporting diarrhea indicated that the symptom did not appear until 31 hr after the culture had been consumed. In known C. perfringens food TABLE 1. Relationi of response inl the rabbit ileum to response in human subjects subsequenzt to inigestion of challenzge doses of viable cells of Clostridium perfringens Strain no. NCTC 8239 NCTC 8239 NCTC 8798 NCTC 8798 NCTC 8798 NCTC NCTC NCTC NCTC NCTC E NCTC 8247 NCTC 8247 FD1 F42 F42 215b 215b Rabbit reaction' Carrier for cellsb pep- BS-1 BS-I BS %0 tone water Total count 6.6 X X X X X X X X X X X X X X X X X X X X X 10 No. of cells fed Spores 7.8 X X l0 1.5 X X X X X X X X X X X X X X X X X X 109 Human subjects No. positivec/no. at test 2/5 3/3 1/5 4/5 3/4 3/4 4/4 3/4 4/4 3/5 1/5 0/3 1/5 No. diarrhea/no. at test a (rabbit-positive) denotes strains of C. perfrinzgens, and products thereof, which have repeatedly been shown to produce fluid accumulation in the ligated ileum of young rabbits or overt diarrhea subsequent to intraileal injection of the nonligated gut, or both. Conversely, (rabbit-negative) indicates the absence of these responses in any regular pattern. b Abbreviations:, Chocolate dairy drink inoculated immediately before being fed;, previous to feeding, cells grown for 3 hr in canned beef stew; BS-I, canned beef stew inoculated immediately before being fed. Positive here defined as any person reporting symptoms of any type, after ingestion of challenge dose. 2/5 3/3 0/5 3/5 3/4 4/4 1/5 0/5 0/3 0/5

4 174 STRONG, DUNCAN, AND PERNA INFEC. IMMUN. poisoning outbreaks, the symptoms have been described as appearing 6 to 24 hr after consumption of the peccant food (1). Therefore, the question arises as to whether the diarrhea experienced by the person consuming the rabbit-negative strain was associated with the experimental procedure. Table 2 also shows that the subjects fed rabbitpositive strains suffered more frequently and more severely from abdominal pain and reported other symptoms in more instances than did those who consumed rabbit-negative strains, thus supporting the ligated ileal loop test in rabbits as a means of distinguishing human food poisoning strains of C. perfringens. In all cases, with the one exception noted above, symptoms appeared in the subjects within 24 hr of administration of the test dose. Abdominal pain persisted for an average of 4.9 hr after onset and diarrhea persisted for 5.7 hr. Another factor which apparently entered into the response of human beings to ingestion of large numbers of C. perfringens was the procedure employed in preparing the challenge culture. The most effective procedure appeared to be that of placing the viable cells in beef stew and allowing them to grow for 3 hr at 46 C. Seventy-four per cent of the persons fed rabbit-positive strains reported food poisoning symptoms, whereas none of those fed rabbit-negative strains experienced any untoward effects. Furthermore, 19 of the 31 persons (61 %) showing positive responses had diarrhea in degrees varying from mild to severe. No one in the group fed rabbit-negative strains grown in the beef stew exhibited diarrhea. The results obtained when cell-free concentrated filtrates were fed to the subjects are shown in Table 3. Four of 15 subjects developed diarrhea when culture filtrates from rabbit-positive strains were fed, whereas no diarrhea was reported by the 16 persons who ingested filtrates prepared from rabbit-negative strains. One subject fed filtrate from a rabbit-positive strain vomited. These results support the concept of an enteropathogenic factor being responsible for C. perfringens food p g symptoms (3, 6). A consideration of total viable cell and spore TABLE 2. Symptoms reported by subjects subsequenit to consuiminlg viable cells of ClostridiuJm perfrinigents Strain no. and carrier' NCTC 8239, NCTC 8239, NCTC 8798, NCTC 8798, BS-I NCTC 8798, NCTC 10239, NCTC 10239, BS-I NCTC 10239, NCTC 10240, NCTC 10240, E13, 68900, 027, 79394, NCTC 8247, NCTC 8247, FD1, F42, F42, 215b, 215b, BS Rabbit reactionb Abdominal pain Mild Moderate Severe 2, 4 2, 7.5 1, 6 2, 9 1, 6 1, 16 1, 6 3, 9 1, 4 1, , 5.5 1,2 1, 3 1, 5 1,9 1, 13 1, 20 2, , 11 1,12 1, 3.5 1, 9 Symptomsc Diarrhea Mild jmoderatel Severe 2, 2.5 1,12.5 1,19 2, 15 3, 16 1, 22 1,6 1,3 1,6 2, 6 1, 16 1, 31 1, l, 13 2, 10 1, 11 3, 12 1, 4 1, 9.5 Nausea 1, 6 2, 8 1, 1, 12 Headache 1, 22 1,2 1, 16 1, 14.5 a See footnote b of Table 1. I See footnotes a of Table 1. c First value records number of cases; second value is the average number of hours after consumption of the culture at which symptom first occurred. Flatus

5 VOL. 3, C. PERFRINGENS TYPE A FOOD POISONING TABLE 3. Response of human subjects to culture filtrates prepared from selected strains of Clostridium perfrinigensa No. of Rabbit1 subjects Source strain for filtrate responseb diarrheal/ total no. challenged io'of--- D2 TOTAL CELL COUNT [j] SPORE COUNT H AVERAGE VALUE NCTC 8239 NCTC /3 NCTC NCTC NCTC 8247 FD1 F42 215b Uninoculated medium a Challenge doses presented in chocolate dairy drink. I See footnote a of Table 1 for explanation of symbols. counts of C. perfringens in the feces of the experimental subjects reveals little which is definitive. Figure 1 illustrates the average and the range of the viable cell and spore counts which were encountered. When the total cell and spore counts for the preliminary fecal sample (which was obtained previous to the feeding of each challenge dose of viable cells of C. perfringens) were taken as a base line, in 19 of 21 tests there was some increase in average total counts for the samples (DS or 24-hr S) secured subsequent to the feeding. Compared in the same way, the spore counts were always somewhat higher in the time shortly after the subjects had consumed the suspensions of cells. Six days after the feeding of the respective challenge doses to each group of subjects, the fecal total cell and spore counts for C. perfringens (6-day S) presented no regular picture when the counts were compared with those of the preliminary sample. There was no comparable increase in viable cell or spore counts after the feeding of filtrates as challenge doses. When the counts of total viable cells and spores of C. perfringens per gram (dry weight) of feces for those subjects within a group who developed diarrhea were singled out and compared with the average for all members of the group, again no regular pattern appeared. The tests carried out with Hobbs antisera for comparing the identity of strains of C. perfringens fed with those recovered from the postchallenge fecal specimens of the subjects indicated the following. (i) When strain NCTC 8239 (H3) was fed, the strain was detected in two of three of the fecal samples (DS or 24-hr S) obtained from the three subjects postchallenge; detections for the other F-(0_ - z 0106~~~~~~~~1 105-~- 0) k o VI( )- ( u ' V rwt IJ Z I _3 _c v_ m'.i O NJ O :M PS STr^AINS 6-DAY CONTRO(.S STRAINS - FIG. 1. Ranige antd average offecal counits of Clostridium perfrinigenis per gram of dry weight of feces. Abbreviationis: PS, sample obtainied previous to challenige feedinig; DS, diarrhleal sample; 24-hir S, sample obtainied subsequentt to conisumptiont of challenge dose wheni no diarrhlea occurred; 6-day S, sample obtainzed 6 days after conzsumptiont of challenge dose. All challenige doses were C. perfrintgenis viable cells except where nioted to be cell filtrates. See footniote a of Table I for additionzal defintitionl of symbols. strains were, respectively, NCTC 8247 (H4), NCTC 8798 (H9) 3/4, NCTC (H12) 3/4, and NCTC (H13) 4/4. (ii) At the outset of the feeding experiments, before the subjects had consumed any of the test cultures, strains serologically identical with NCTC 8247 and NCTC 8798 were each detected in the feces of one person; subsequently, when these strains were fed, the feces of these two persons were negative for the respective strains. (iii) In only one case did a challenge strain (NCTC 8798) apparently persist in the digestive tract for 2 weeks, to appear subsequently after the feeding of another strain. (iv) All cross-agglutinations with specific strains were negative. (v) In some cases (NCTC 8239, one instance; NCTC 10239, 2 instances; NCTC 10240, one instance), isolates picked from plates inoculated with samples of feces heated at 75 C for 20 min yielded positive agglutination tests, whereas those isolates obtained from unheated feces were negative, thus suggesting good sporulation in the intestines. The failure to detect strain NCTC 8247 in the feces of subjects subsequent to challenge may be explained by the fact that, although viable cells were fed in the magnitude of 1010 cells, only

6 176 STRONG, DUNCAN, AND PERNA INFEC. IMMUN. 102 spores were present. Since vegetative cells appear to be more affected by acidity than spores (8a), it may be that this strain did not survive passage through the stomach. Table 4 indicates the results obtained when D,oo values were determined as a measure of heat resistance for the rabbit-positive and the rabbitnegative strains of C. perfringens. For rabbitpositive strains, the D1oo values varied from 0.70 to These results point out that strains producing a response in the ileum of the rabbit may be of varied heat resistance. Fewer observations were made on rabbit-negative strains, but the data available suggest that such strains are characterized by a low degree of heat tolerance. However, evidence has been presented by Sutton and Hobbs (22), among others, that C. perfringens food poisoning may be caused by heat-sensitive as well as heat-stable strains of C. perfringens. During 1968 and for the period of January to June 1969, the attack rate for persons exposed to the possibility of contracting C. perfringens food poisoning was stated to be, overall, 33.9 and 56.7%, respectively (16, 17). In the present experiments, attack rates obtained for the selected population which was fed rabbit-positive strains were essentially identical (57%) with the 1969 figure. Under conditions most closely resembling a "natural" outbreak (i.e., viable cells allowed to grow for 3 hr in beef stew held at 46 C), the attack rate when the criterion for illness was diarrhea was 61.0% for subjects fed rabbit-positive cells and 0% for those fed rabbit-negative strains. When diarrhea occurred in members of the test groups fed viable cells, the range of time of ap- TABLE 4. Heat resistance (Dloo value) of strains of Clostridium perfringens which have demonstrated different responses in the ileum ofthe rabbit Strain no. Rabbit D1oo value reactiona (min) NCTC NCTC NCTC NCTC E NCTC 8247 NTb FD F b 0.97 a See footnote a of Table 1 for explanation of symbols. b Not tested. Strain did not produce spores in sufficient numbers to permit test. pearance was 2.5 to 22 hr, with one exception, noted above, in which the period of incubation was 31 hr. Despaul (1) stated that the time for appearance of symptoms after consumption of food contaminated with large numbers of viable C. perfringens cells is 6 to 24 hr. Hauschild and Thatcher (13) reported diarrhea to occur between 4 and 9 hr after cell ingestion and to continue for periods of 4 to 21 hr. The four subjects demonstrating diarrhea subsequent to the ingestion of concentrated culture filtrates reported the appearance of the symptoms in 2 to 2.5 hr after consumption of the challenge. In the laboratory study done by Hauschild and Thatcher (13), certain food poisoning symptoms, typical of those induced by the presence of large numbers of C. perfringens in food, were induced in five of six human volunteers subsequent to feeding 4 X 109 to 6 X 109 vegetative cells of strain S-79. For feeding, the cells were suspended in cooked milk and the subjects consumed an additional 100 ml of milk, apparently followed immediately by their eating lunch. In our experiments, the effects produced by feeding the viable cultures to the subjects seemed to be enhanced if considerable amounts of food were eaten at the time the cultures were presented. A logical explanation for this observation would appear to be the buffering action of food on the acidity of the stomach. Strains of C. perfringens have been shown to be susceptible in some degree to the action of acid in the range of the ph of the stomach (8a, 10). In our laboratory, it has not been possible to induce diarrhea in rabbits, a species known to maintain low ph value in the empty stomach, by oral challenge (5). Obviously, the fate of the cells in the stomach is closely related to the sequential effect in the gut. Interpretation of data relative to the number of total viable cells or spores of C. perfringens, or both, in the feces of "normal" people and of persons who have experienced diarrhea as a consequence of consumption of large numbers of the organisms in food is difficult. Variation exists between individuals and for a single individual at different times. The average viable counts we obtained for total cells and for spores of C. perfringens in feces during the prechallenge dose period were the same or higher than those that have been reported for the general population by other investigators. The range of average total cell counts for groups for all experiments was 1.5 X 105 to 9.2 X 107 and for spores was 1.1 x 104 to 2.3 x 107 per g of dry feces. These values should be considered approximate since the effect of heat shock, and hence the relationship between total counts and spore counts, cannot be definitively estab-

7 VOL. 3, 1971 C. PERFRINGENS TYPE A FOOD POISONING. II lished; nevertheless, the figures presumably reflect, with some accuracy, the concentration of viable organisms which existed. Sutton (21) reported for an assorted human population a median of 3.7 X 104 and a mean of 1.40 X 105 as the viable count for C. welchii (perfringens) per g of feces. Earlier (20) he had found that participation in communal feeding and poor hygienic conditions influenced the carrier-rate for heat-resistant C. perfringens in human feces. In a later study, Sutton and Hobbs (22) stated that fecal median counts for C. welchii (perjringens) of 103 per g of feces are within normal limits. In a series of 57 samples of feces from normal human intestines, Nakagawa and Nishida (15) found a variation of from 103 or less to 106 or more cells of C. perfringens per ml of feces. Hauschild and Thatcher (13) reported a concentration of 104 to 3 X 104 spores of C. perfringens per g of feces of subjects before cell ingestion. Presumably, the values reported in the literature were based on moist fecal material in contrast to our values which are expressed in terms of dry fecal material. Questions which might arise concerning other aspects of our study include the possible development of immunity in our subjects, many of whom were tested eight times, and the psychological effect of suggesting to the patients symptoms which they might encounter. Dische and Elek (2) observed earlier that repeated feedings of the same subject with viable cultures of C. welchii (perfringens) did not mitigate the symptoms occurring at the time of the second or third challenge dose. Therefore, they concluded that "... there was no indication that clinical immunity had developed" (reference 2, p. 73). Our findings supported those of Dische and Elek. It is difficult to assess the effect of suggestibility in regard to the symptoms reported by the subjects after consuming challenge doses of viable C. perfringens. It is possible that subjective symptoms reported were not associated with the ingestion of the test cultures. For this reason, those instances in which diarrhea occurred are regarded as especially significant. Finally, the correlation of results obtained in these human feeding studies with those acquired throughout the series of studies in which rabbits and monkeys have been utilized deserves comment. A summary of the findings resulting from the injection of the ligated loops and the nonligated ilea of rabbits with viable cells, cell extracts, or culture filtrates for selected strains of C. perfringens was presented in an earlier publication (6). Suffice it to state here that both viable cells and cell-free products of strains NCTC 8239, NCTC 8798, NCTC 10239, NCTC 10240, E13, 68900, and 027 produced, with significant 177 regularity, an accumulation of fluid in ileal loops, or overt diarrhea, or both, subsequent to challenge in young rabbits. Viable cells of the same strains produced diarrhea in 57% of the 35 monkeys tested (strain not tested; 7) and in 45% of 58 persons in the current study. Culture filtrates prepared for each of four strains (NCTC 8793, NCTC 10239, NCTC 10240, and NCTC 8239) produced, overall, diarrhea in 8 of 20 monkeys and vomiting in 11 of 20 monkeys (7). The concentrated filtrates obtained from the same four strains produced diarrhea in 4 of 15 human subjects and vomiting in 1. Of the rabbit-negative strains, NCTC 8247, F42, and FD1 also were first isolated from sources (i.e., food or feces) associated with C. perfringens food poisoning. Neither viable cultures of these strains nor, in a limited number of trials, culture filtrates produced a response in rabbit tests or in any of 20 tests in monkeys (7). In human beings, only 1 somewhat doubtful positive response was recorded in 22 challenges with viable cells, and an additional 12 challenges made with culture filtrates prepared from the rabbit-negative strains were also negative. Viable cells or filtrates of strain 215b, which was originally isolated from beef liver, did not induce a positive response in rabbits, monkeys, or human beings. The observations concerning the rabbit-negative strains of C. perfringens suggest that the factor responsible for food poisoning has, in some cases, apparently been permanently lost during repeated subculturing in the laboratory. Results obtained with other strains indicate that the production of the diarrhea-inducing factor may occur irregularly and thus these strains would seem to be intermediate between the rabbit-negative and the rabbit-positive groups. ACKNOWLEDGMENTS Published with the permission of the Director of the Research Division of the College of Agricultural and Life Sciences, University of Wisconsin, Madison. This study was supported by Public Health Service grant 5- R01-FD from the Consumer Protection and Environmental Health Service and by contributions from memnber industries of the Food Research Institute. The antisera were provided through the generous cooperation of Betty Hobbs. She further cooperated with us by determining the serological type of certain of our strains. We gratefully acknowledge this assistance. LITERATURE CITED 1. Despaul, J. E The gangrene organism-it food poisoning agent. J. Amer. Diet. Ass. 49: Dische, F. E., and S. D. Elek Experimental food poisoning by Clostridium welchii. Lancet 2: Duncan, C. L Clostridium perfringenis food poisoning. J. Milk Food Technol. 33: Duncan, C. L., and D. H. Strong Improved medium for sporulation of Clostridiuin perfringents. Appl. Microbiol. 16:82-89.

8 178 STRONG, DUNCAN, AND PERNA INFEC. IMMUN. 5. Duncan, C. L., and D. H. Strong Experimental production of diarrhea in rabbits with Clostridium perfrinigens. Can. J. Microbiol. 15: Duncan, C. L., and D. H. Strong Ileal loop fluid accumulation and production of diarrhea in rabbits by cellfree products of Clostridiumn perfrinigens. J. Bacteriol. 100: Duncan, C. L., and D. H. Strong Clostridium perfringens type A food poisoning. I. Response of the rabbit ileum as an indication of enteropathogenicity of strains of Clostridium perfringens in monkeys. Infec. Immun. 3: Duncan, C. L., H. Sugiyamiia, aniid D. H. Strong Rabbit ileal loop response to strains of Clostridiumn perfrinigens. J. Bacteriol. 95: a. Fischer, L. H., D. H. Strong, and C. L. Duncan Resistance of Clostridium perfringenis to varying degrees of acidity during growth and sporulation. J. Food Sci. 35: Hall, H. E., R. Angelotti, K. H. Lewis, and M. J. Foter Characteristics of Clostridiumn perfrinigenis strains associated with food and food-borne disease. J. Bacteriol. 85: Hauschild, A. H. W., R. Hilschimer, and F. S. Thatcher Acid resistance and infectivity of food-poisoning Clostridium perfringenis. Can. J. Microbiol. 13: Hauschild, A. H. W., L. Niulo, and W. J. Dorward Experimental enteritis with food poisoning and classical strains of Clostridium perfrinigenis type A in lambs. J. Infec. Dis. 117: Hauschild, A. H. W., L. Niilo, and W. J. Dorward Clostridiumii perfrinigenis type A infection of ligated intestinal loops in lambs. Appl. Microbiol. 16: Hauschild, A. H. W., and F. S. Thatcher Experimental food poisoning with heat-susceptible Clostridiuni perfrinigens type A. J. Food Sci. 32: Hobbs, B. C., M. E. Smith, C. L. Oaklev, G. H.Warrack, and J. C. Cruickshank Clostr-idiu,ni welchii food poisoning. J. Hyg. 51: Nakagawa, M., and S. Nishida Heat rcsistance and toxigenicity of Clostridiuni perfrinigenis in normal intestines of Japanese. Jap. J. Microbiol. 13: National Communicable Disease Center Foodborne Outbreaks Annual Summary. U.S. Public Health Service, Atlanta, Ga. 17. National Communicable Disease Center Foodborne Outbreaks January-June. U.S. Public Health Service, Atlanta, Ga. 18. Nygren, B Phospholipase C-producing bacteria and food poisoning. Acta Pathol. Microbiol. Scand. Suppl Satterlee, L. D., and H. W. Wailker Response of white mice to cells and culture constituents of Clostridiuni perfringens. Appl. Microbiol. 18: Sutton, R. G. A Distribution of heat-resistant Clostridium welchii in a rural area of Australia. J. Hyg. 64: Sutton, R. G. A Enumeration of Clostridium welchii in the feces of varying sections of the human population. J. Hyg. 64: Sutton, R. G. A., and B. C. Hobbs Food poisoning caused by heat sensitive Clostridium welchii. A report of five recent outbreaiks. J. Hyg. 66: Weiss, K. F., D. H. Strong, and R. A. Groomn Mice anid monkeys as assay animals for Clostridiunm perfrin1genis food poisoning. Appl. Mircobiol. 14: Downloaded from on July 7, 2018 by guest