and Benjamin (21). Serotyping. ..... Benjamin, J., S. Leaper, R. J. Owen, and M. B. Skirrow. 1983. Description of ... Prescott, J. F., and D. L. Munroe. 1982.
APPLIED AND ENVIRONMENTAL MICROBIOLOGY, June 1985, p. 1507-1510
Vol. 49, No. 6
0099-2240/85/061507-04$02.00/0 Copyright C) 1985, American Society for Microbiology
Serotyping of Campylobacterjejuni, Campylobacter coli, and Campylobacter laridis from Domestic and Wild Animals OLAV ROSEF,'2 GEORG KAPPERUD,13* SABINE LAUWERS,4 AND BJ0RN GONDROSEN' Department of Food Hygiene, Norwegian College of Veterinary Medicine, Oslo 1,1 Norwegian Defence Microbiological Laboratory, Oslo 4,3 and The Food Inspection Service in Aust-Agder, 4800 Arendal,2 Norway; and Infectious Disease Unit, Free University of Brussels, Brussels, Belgium4 Received 13 November 1984/Accepted 5 March 1985
By using 50 unabsorbed antisera, we were able to serotype 272 (65.7%) of 414 thermotolerant campylobacters from wild and domestic animals, on the basis of heat-stable antigens identified by means of passive hemagglutination. Forty-two serotypes were recognized. The pattern of serotypes detected in the various animal species was compared to human clinical isolates by using the Czekanowski index (proportional similarity index). The highest degree of similarity to the clinical isolates was observed for the poultry isolates, followed by strains from wild birds, flies, and pigs (in order of decreasing similarity). The serotypes recovered most frequently from poultry (LAU 1 and LAU 2) were also most prevalent in Norwegian patients. In contrast, serotype LAU 35/44, the predominant porcine serotype, was never recovered from human clinical specimens. Flies captured in chicken farms and in piggeries harbored serotypes which were also commonly seen in chickens and pigs, respectively. Nine of the strains included in this study could not be ascribed to any defined species. All of these were resistant to nalidixic acid and did not produce H2S.
During the past few years, considerable progress has been made in the understanding of the importance of Campylobacter jejuni and Campylobacter coli (synonyms, thermotolerant campylobacters, Campylobacter fetus subsp. jejuni) as human pathogens (3, 4, 15, 20). Although many aspects concerning the epidemiology of Campylobacter enteritis remain to be clarified, the significance of several animal species as reservoirs for these bacteria has been established (3, 15, 20). The introduction of effective methods for the differentiation of isolates, based on biochemical and serological criteria, has enabled further insight to be gained into the epidemiology of Campylobacter enteritis (9, 11, 21). Serological characterization by means of heat-labile or heat-stable antigens has become the most useful typing method. In Norway, an extensive indigenous reservoir of thermotolerant campylobacters has been detected among wild and domestic animals (7, 16-19). However, the contribution of each animal species to human infection is largely unknown. The purpose of this study was to serotype thermotolerant campylobacters from wild and domestic animals in Norway to contribute to the understanding of the epidemiology of Campylobacter enteritis in this country.
specified elsewhere (7, 16-19). All isolates were stored at -70°C in heat-inactivated horse serum with 17% glycerol. Biotyping. All strains were tested for hippurate hydrolysis, H2S production, and susceptibility to nalidixic acid as described previously (17). These parameters enabled allocation to C. jejuni biotype 1, C. jejuni biotype 2, C. coli, or Campylobacter laridis (synonym, "nalidixic acid-resistant thermophilic campylobacters") by the method of Skirrow and Benjamin (21). Serotyping. All strains were serotyped on the basis of thermostable antigens identified by means of the passive hemagglutination technique as described by Lauwers et al. (10). The procedure followed has been detailed previously (6). The serotyping was accomplished by using 50 unabsorbed rabbit antisera, which were prepared at the University Hospital of St. Pierre, Microbiology Laboratory, in Brussels. Antigen preparation, sensitization of erythrocytes, and hemagglutination were carried out at the Department of Food Hygiene, Norwegian College of Veterinary Medicine, Oslo, Norway. Similarity index. The similarity between the frequency distributions of Campylobacter serotypes among different animal species, was estimated by using the Czekanowski index (proportional similarity index [PS]) (5): PS = 1 - 0.5 ELpi - qil, where pi and qi represent the proportion of strains belonging to serotype i out of all strains serotyped from animal species P and Q, respectively. The values for PS range from 1 for the highest possible similarity to zero for distributions with no common serotypes. PS is an objective estimate of the area of intersection between two frequency distributions. This index is intuitively and mathematically meaningful even when there are empty cells in one or both of the distributions being compared.
MATERIALS AND METHODS Bacterial strains. A total of 414 thermotolerant campylobacters was subjected to serological typing. All strains were isolated from wild and domestic animals in Norway (7, 16-19). The strains fell into seven categories (Table 1): (i) 207 strains from poultry, (ii) 66 strains from slaughter pigs, (iii) 59 strains from wild-living birds, (iv) 42 strains from houseflies captured in two piggeries and one chicken farm, (v) 26 strains from dogs and cats (B. Gondrosen, T. Knaevelsrud, and K. Dommarsnes, Acta Vet. Scand., in press), (vi) 12 strains from sheep, and (vii) two strains from dairy cows. Isolation procedures and criteria for identification of strains have been
*
RESULTS The serotyping scheme employed in this work enabled us to type 272 (65.7%) of 414 thermotolerant campylobacters isolated from domestic and wild animals in Norway (Table 1). The highest percentage of serotypable strains was re-
Corresponding author. 1507
1508
ROSEF ET AL.
wild-living birds (40.7%). Of the typable strains from all sources, 199 reacted in only one antiserum, 61 strains reacted in various pairs of antisera, whereas five strains agglutinated in three different sera. The remaining seven strains were multiagglutinable. Altogether 42 serotypes were distinguished. The most common serotype was LAU 2, which made up 15.4% of the typable strains, followed by LAU 35/44 and LAU 1 with 13.2% and 10.7%, respectively. An additional five serotypes occurred regularly, each making up about 4 to 6% of the typable strains.
Relative dominance of serotypes. Within each animal group, the degree to which isolates were concentrated in relatively few, dominant serotypes varied considerably (Table 1). The highest concentration of dominance was observed with the porcine isolates, 43.9% of which belonged to serotype LAU 35/44. Likewise, serotypes LAU 1 and LAU 2, taken together, constituted 32.4% of the total number of poultry isolates. The dominant serotypes among houseflies were LAU 6 and LAU 35/44, each of which made up 16.7%. The strains isolated from wild-living birds were more evenly distributed than was the case for the animals mentioned above. The remaining animal species were too sparsely
represented to justify comparisons. Relationship between serotypes and Campylobacter spp. The 414 thermotolerant campylobacters serotyped in this study comprised 225 C. jejuni biotype 1, 34 C. jejuni biotype 2, 136 C. coli, and 10 C. laridis. The remaining nine strains could not be ascribed to any defined species. Eight strains were resistant to nalidixic acid and were negative for hippurate hydrolysis and H2S production. One strain was resistant to nalidixic acid, hydrolyzed hippurate, and was negative for H2S. The relative prevalence of C. jejuni, C. coli, and C. laridis among the animal species concerned, has been reported previously (7, 16-19). The percentages of serotypable strains were as follows: C. jejuni biotype 1, 67.1%; C. jejuni biotype 2, 64.7%; C. coli, 64.7%; and C. laridis, 60.0%. C. jejuni biotype 1 encompassed strains belonging to 24 different serotypes, whereas C. jejuni biotype 2, C. coli, and C. laridis fell into 9, 20, and 3 serotypes, respectively. Of the 42 serotypes detected, 12 were specific for C. jejuni biotype 1, 5 were specific for C. jejuni biotype 2, and 12 were specific for C. coli. No serotype was specific for C. laridis. Ten serotypes comprised both C. jejuni and C. coli isolates, and in three of these, C. laridis isolates were also found. DISCUSSION It is now known that many animal species may harbor Campylobacter serotypes that also occur in human patients suffering from Campylobacter enteritis (10, 12-14). Some of the same serotypes which are common among human clinical isolates have been shown to be prevalent in domestic animals, including those used for food production, such as poultry and cattle. The present study provides further evidence to support these results and adds several new species to the list of animals previously shown to harbor serotypes also found in man (Table 1). The results of serotyping 172 thermotolerant campylobacters from human cases of gastroenteritis in Norway have been previously reported (6). To assess the epidemiology of campylobacteriosis in this country, the serotype profiles of human and animal Campylobacter isolates were compared. A simple measure of similarity between two frequency distributions is provided by the Czekanowski index. However, the nontypable (NT) strains, which encompass an unknown number of serotypes not yet defined by the sero-
APPL. ENVIRON. MICROBIOL.
typing scheme, need special attention. Since objective subdivision is impossible at the present stage, two different approaches were applied. (i) The NT strains were regarded as one homogenous serotype; this will, evidently, lead to overestimation of similarity when both distributions being compared contain NT strains. (ii) The NT strains were considered as a spectrum of distinct serotypes; since it is probable that at least some overlap may occur in the NT group, this approach will tend to underestimate similarity. The Czekanowski index was calculated on the basis of each of these interpretations, and the resultant estimates are both presented in Table 2. Regardless of the interpretation adopted, the highest similarity to human clinical isolates was obtained for the poultry isolates, followed by strains from wild birds, flies, and pigs TABLE 1. Serotyping of campylobacters from domestic and wild animals Serotype
LAU 1 LAU 2 LAU 3 LAU 4 LAU 6 LAU 8 LAU 14 LAU 15 LAU 16 LAU 18 LAU 19 LAU 20 LAU 25 LAU 27 LAU 32 LAU 35 LAU 37 LAU 40 LAU 43 LAU 44 LAU 51 LAU 3/16 LAU 5/8 LAU 6/32 LAU 8/11 LAU 8/35 LAU 14/47 LAU 35/44 LAU 8/11/33 LAU 14/PEN 19 LAU 17/PEN 17 LAU 38/PEN 21 LAU 14/25/PEN 19 LAU 37/40/PEN 40 LAU 40/PEN 20/32 PEN 15 PEN PEN PEN PEN PEN
Poultry
25
Others
4
42 6 2 2
3
1
2
1
7
3 3 1 2
1 1c
1 2 13 11 1 1 2
2 1
4 2
1
1 1
3 4
7 11
9
3
1 1
1
1
2 1
1 7
1 29
1 1
1
2
2 1 1
2 1 2 2
17 19 20 21 27
Multiagglutinable Nontypable
No. of isolates from: Wild birdsa Flies Pigs
2
le
1 1 1 3
S
4b 4
60 35 6 16 25e a Isolates from 29 seuagulls, 17 crows, 8 puffins, 3 pigeons, and 2 owls. b Isolates from dogs. Isolates from cats. Isolate from a shee ,P. eIsolates from 11 shtteep, 8 dogs, 4 cats, and 2 cows.
'
d
VOL. 49, 1985
CAMPYLOBACTER SEROTYPING
1509
TABLE 2. Similarity between the frequency distributions of Campylobacter serotypes isolated from man and animals Humans
Poultry
Similarity indexa Wild birds
Flies
Pigs
1.000/1.000 0.432/0.233 0.350/0.151 0.289/0.146 0.282/0.083
0.330/0.040 0.309/0.166 0.312/0.070
0.227/0.084 0.333/0.091
0.380/0.237
1.000/1.000
Source
Humans Poultry Wild birds Flies Pigs
a Estimated by the Czekanowski index (5). Two different values, based on dissimilar interpretations of the nontypable strains, are presented (see the text).
(in order of decreasing similarity). It is notable that the serotypes recovered most frequently from Norwegian patients (LAU 1 and LAU 2), were also most prevalent in poultry. A high degree of similarity between human and poultry isolates has also been observed in other countries (1, 12, 13). In contrast, a majority of the porcine isolates, all of which were C. coli, belonged to serotypes not yet described in Norwegian patients. Serotype LAU 35/44, the dominant porcine serotype, was never recovered from human clinical specimens. The data presented by Munroe et al. (13) and Banffer (1) are similar, whereas Lior et al. (12), who used a serotyping scheme based on heat-labile antigens, concluded that all of their porcine strains belonged to the same serotypes isolated from human cases. The number of isolates from dogs, cats, sheep, or cows was insufficient to justify epidemiological conclusions. However, dogs, cats, and sheep were all found to harbor serotypes previously recovered from human cases of campylobacteriosis. In this and similar studies, it has been assumed that epidemiological affinity between animal species is proportional to the similarity between the serotype distributions of the species being compared. However, this tacit assumption is not necessarily correct. Since the factors responsible for virulence in C. jejuni and C. coli are unknown, it is quite possible that many animal isolates may not be pathogenic, even though they are serologically identical to clinical isolates. The epidemiological interpretation of the present results is further confounded by the fact that more than 50% of the Norwegian cases of Campylobacter enteritis were contracted in foreign countries (8). Hence, definite conclusions concerning the relative importance of the various animal reservoirs in Norway, are not warranted at the present stage. Effective virulence assays, capable of screening large numbers of isolates, are required to evaluate the pathogenicity of individual isolates and thereby enhance our understanding of the epidemiology of campylobacteriosis. Of the strains isolated from flies captured in piggeries, 66.7% belonged to two serotypes (LAU 6 and LAU 35/44), these being also common in swine. Likewise, 50% of the strains from flies inhabiting a chicken farm belonged to serotypes also seen in chickens. This supports the suggestion made in a previous report (19) that flies become infected with the types of campylobacters which prevail in the environments in which they live. The serotype distributions recorded among poultry and pigs were each characterized by a different set of dominant serotypes, indicating that some host preference may exist among the Campylobacter serotypes. In this light, it is not surprising that the serotypes detected among wild birds were more evenly distributed, since avian wildlife constitutes a taxonomically and ecologically heterogenous group. Resistance to nalidixic acid has been occasionally found in strains of C. jejuni (22), a circumstance which indicates that
this property may be transferable (2). It is possible, therefore, that the atypical strains reported in this study may represent nalidixic acid-resistant variants of C. jejuni and C. coli. It is also possible that these variants may belong to the new group of hippurate-negative, nalidixic acid-resistant campylobacters recently described by Walder et al. (22). LITERATURE CITED 1. Banffer, J. R. J. 1983. Biotype and serotype of Campylobacter strains isolated from patients, pigs and chickens in the region of Rotterdam, p. 99. In A. D. Pearson, M. B. Skirrow, B. Rowe, J. R. Davies, and D. M. Jones (ed.), Campylobacter II. Public
Health Laboratory Service, London, England. 2. Benjamin, J., S. Leaper, R. J. Owen, and M. B. Skirrow. 1983. Description of Campylobacter laridis, a new species comprising the nalidixic acid resistant thermophilic Campylobacter (NARTC) group. Curr. Microbiol. 8:231-238. 3. Blaser, M. J., D. N. Taylor, and R. A. Feldman. 1983. Epidemiology of Campylobacter jejuni infections. Epidemiol. Rev. 5:157-176. 4. Butzler, J. P., and M. B. Skirrow. 1979. Campylobacter enteritis. Clin. Gastroenterol. 8:737-765. 5. Feinsinger, P., E. E. Spears, and R. W. Poole. 1981. A simple measure of niche breadth. Ecology 62:27-32. 6. Kapperud, G., J. Lassen, S. Lauwers, and 0. Rosef. 1984. Serotyping and biotyping of Campylobacter jejuni and Campylobacter coli from sporadic cases and outbreaks in Norway. J. Clin. Microbiol. 19:157-160. 7. Kapperud, G., and 0. Rosef. 1983. Avian wildlife reservoir of Campylobacter fetus subsp. jejuni, Yersinia spp., and Salmonella spp. in Norway. Appi. Environ. Microbiol. 45:375-380. 8. Lassen, J., and G. Kapperud. 1984. Epidemiological aspects of enteritis due to Campylobacter spp. in Norway. J. Clin. Microbiol. 19:153-156. 9. Lauwers, S., and J. L. Penner. 1984. Serotyping Campylobacter jejuni and Campylobacter coli on the basis of thermostable antigens, p. 51-59. In J. P. Butzler (ed.), Campylobacter infection in man and animals. CRC Press, Inc., Boca Raton, Fla. 10. Lauwers, S., L. Viaes, and J. P. Butzler. 1981. Campylobacter serotyping and epidemiology. Lancet i:158-159. 11. Lior, H. 1984. Serotyping of Campylobacterjejuni and C. coli by slide agglutination based on heat-labile antigenic factors, p. 61-76. In J. P. Butzler (ed.), Campylobacter infection in man and animals. CRC Press, Inc., Boca Raton, Fla. 12. Lior, H., D. L. Woodward, J. A. Edgar, L. J. Laroche, and P. Gill. 1982. Serotyping of Campylobacterjejuni by slide agglutination based on heat-labile antigenic factors. J. Clin. Microbiol. 15:761-768. 13. Munroe, D. L., J. F. Prescott, and J. L. Penner. 1983. Campylobacter jejuni and Campylobacter coli serotypes isolated from chickens, cattle, and pigs. J. Clin. Microbiol. 18:877-881. 14. Pearson, A. D., M. B. Skirrow, B. Rowe, J. R. Davies, and D. M. Jones (ed.). 1983. Campylobacter II, p. 81-102. Public Health Laboratory Service, London, England. 15. Prescott, J. F., and D. L. Munroe. 1982. Campylobacterjejuni
1510
ROSEF ET AL.
enteritis in man and domestic animals. J. Am. Vet. Med. Assoc. 181:1524-1530. 16. Rosef, 0. 1981. Campylobacter fetus subsp. jejuni as a surface contaminant of fresh and chilled pig carcasses. Nord. Veterinaermed. 33:535-538. 16a.Rosef, O., B. Gondrosen, and G. Kapperud. 1984. Campylobacter jejuni and Campylobacter coli as surface contaminants of fresh and frozen poultry carcasses. Int. J. Food Microbiol. 1:205-215. 17. Rosef, O., B. Gondrosen, G. Kapperud, and B. Underdal. 1983. Isolation and characterization of Campylobacter jejuni and Campylobacter coli from domestic and wild mammals in Norway. Appl. Environ. Microbiol. 46:855-859.
APPL. ENVIRON. MICROBIOL.
18. Rosef, O., and G. Kapperud. 1982. Isolation of Campylobacter fetus subsp. jejuni from faeces of Norwegian poultry. Acta Vet. Scand. 23:128-134. 19. Rosef, O., and G. Kapperud. 1983. House flies (Musca domestica) as possible vectors of Campylobacter fetus subsp. jejuni. AppI. Environ. Microbiol. 45:381-383. 20. Skirrow, M. B. 1982. Campylobacter enteritis-the first five years. J. Hyg. 89:175-184. 21. Skirrow, M. B., and J. Benjamin. 1980. Differentiation of enteropathogenic Campylobacter. J. Clin. Pathol. 33:1122. 22. Walder, M., K. Sandstedt, and J. Ursing. 1983. Phenotypic characteristics of thermotolerant Campylobacter from human and animal sources. Curr. Microbiol. 9:291-296.
