- Email: [email protected]
Serotypes of Escherichia coli isolated from healthy infants in Berlin, Germany and Melbourne, Australia
K.A. Bettelheim et al. / Comp. Immun. Microbiol. Infect. Dis. 26 (2003) 55±63
Comparative Immunology, Microbiology & Infectious Diseases 26 (2003) 55±63
55
www.elsevier.com/locate/cimid
Serotypes of Escherichia coli isolated from healthy infants in Berlin, Germany and Melbourne, Australia K.A. Bettelheim a,*, L. Beutin b, K. Gleier b, J.L. Pearce c, R.K.J. Luke c, S. Zimmermann b a
National Escherichia coli Reference Laboratory, Microbiological Diagnostic Unit, Public Health Laboratory, Department of Microbiology and Immunology, University of Melbourne, Melbourne, Vic. 3010, Australia b Robert Koch Institut, Division of Emerging Pathogens, Berlin, Germany c Department of Agricultural Sciences, Faculty of Science, Technology and Engineering, School of Life Sciences, La Trobe University, Melbourne, Vic. 3083, Australia Received 12 September 2001; received in revised form 20 December 2001; accepted 17 January 2002
Abstract The characterization of Escherichia coli strains isolated from healthy infants under one year of age with respect to O:H serotype, K1 and K5 antigens in two disparate parts of the developed world was the purpose of this investigation. A total of 450 strains were examined, 264 from Berlin and 186 from Melbourne. Of all the 220 different O:H serotypes found, 179 were only isolated once, 90 in Berlin and 89 in Melbourne. However, 30 of the 41 O:H serotypes (73.2%) found more than once were isolated in both centers. The most commonly identi®ed serotypes were found in both centers and included O1:H-; O1:H7; O2:H1; O2:H4; O2:H7; O4:H5; O6:H-; O6:H1; O15:H1; O18:H7; O25:H1; and O75:H-. Potentially pathogenic serotypes were found in both cities. Enteropathogenic E. coli (EPEC)-associated serotypes (O18:H7; O26:H-; O44:H34; O86:H-; O128:H2) were present in 11 cases and enterohaemorrhagic E. coli (EHEC)-associated types including O26:H11; O128:H2) were present in four cases. The distributions of serotypes found were similar in the two cities, strongly suggesting the wider applicability of these results. q 2002 Elsevier Science Ltd. All rights reserved. Keywords: Escherichia coli; Healthy infants; Serotypes; Berlin; Melbourne
ReÂsume Le but de cette enqueÃte eÂtait de de®nir les caracteÂristiques des souches d'Escherichia coli deÂceleÂes chez de jeunes enfants sains aÃgeÂs de moins d'un an en ce qui concerne le seÂrotype O:H, avec les * Corresponding author. Tel.: 161-3-8344-8180; fax: 161-3-8344-7833. E-mail addresses: [email protected] (K.A. Bettelheim), [email protected] (K.A. Bettelheim). 0147-9571/03/$ - see front matter q 2002 Elsevier Science Ltd. All rights reserved. PII: S 0147-957 1(02)00015-2
56
K.A. Bettelheim et al. / Comp. Immun. Microbiol. Infect. Dis. 26 (2003) 55±63
antigeÁnes K1 et K5 dans deux parties disparates du monde deÂveloppeÂ. Un total de 450 souches ont eÂte examineÂes. 264 souches de Berlin et 186 souches de Melbourne. Sur tous les 220 diffeÂrents seÂrotypes O:H trouveÂs, 179 ont eÂte deÂceleÂs une fois seulement, 90 aÁ Berlin et 89 aÁ Melbourne. Cependant, 30 des 41 seÂrotypes O:H (73.2%) trouveÂs plus d'une fois ont eÂte deÂceleÂs dans les deux centres. Les seÂrotypes identi®eÂs le plus souvent ont eÂte trouveÂs dans les deux centres et comprenaient O1:H-; O1:H7; O2:H1; O2:H4; O2:H7; O4:H5; O6:H-; O6:H1; O15:H1; O18:H7; O25:H1; et O75:H-. Potentiellement des seÂrotypes pathogeÁnes ont eÂte trouveÂs dans les deux villes. Les seÂrotypes causant E. coli enteÂropathogeÁne (EPEC) (O18:H7; O26:H-; O44:H34; O86:H-; O128:H2) eÂtaient preÂsents dans 11 cas et les types causant E. coli enteÂroheÂmorragique (EHEC) qui comprenaient O26:H11; O128:H2) eÂtaient preÂsents dans quatre cas. La distribution des seÂrotypes deÂceleÂs eÂtait similaire dans les deux villes, ce qui semble fortement indiquer la plus grande applicabilite de ces reÂsultats. q 2002 Elsevier Science Ltd. All rights reserved. Mots-cleÂ: Escherichia coli; Jeunes enfants sains; seÂrotypes; Berlin; Melbourne
1. Introduction It is a tribute to the ubiquity, variety and adaptability of the Escherichia coli species that their role in the human faecal ¯ora is still being explored, so long after the pioneering studies on neonatal faeces, of Escherich [1], which led to the characterization of these organisms. The normal human faecal ¯ora has been estimated at around 10 10 ±10 11 per gram [2,3]. The Enterobacteriaceae are present at about 10 8 ±10 9 per gram and E. coli are the most important enterobacterial species. Adults continuously acquire new types of E. coli with their food, which then set up a competitive situation with the resident ¯ora [4]. Neonates have comparatively high levels of Enterobacteriaceae. They acquire them at birth from their mothers [5] and to a lesser extent from other people in close contact with them. Thus, healthy infants under 1 year of age will carry an E. coli ¯ora re¯ecting the predominant types in the human population of their community at that time. Infants are also susceptible to intestinal infections with pathogenic E. coli. In the developing world, enteropathogenic E. coli (EPEC) and enterotoxigenic E. coli (ETEC) are still major causes of neonatal morbidity and mortality, although they are not major problems in developed countries like Australia and Germany. In the developed world, E. coli, especially those carrying the K1 capsular antigen, is a major cause of neonatal meningitis [6]. In addition, K1 and K5 belong to the K-types, which are over-represented among E. coli isolated from patients with cystitis and pyelonephritis in comparison to faecal E. coli [7]. Also infections by enterohaemorrhagic E. coli (EHEC), which are characterized by their ability to produce one or both Shiga-toxins (Stx), the adherence factor intimin (EAE) and Enterohaemolysin are major causes of infant morbidity and mortality. It has also been noted in a number of studies [8], that certain serotypes of E. coli are far more likely to be associated with some of these virulence factors, than other serotypes. The aim of this study was to characterize strains of E. coli isolated from healthy infants under one year of age with respect to O:H serotype, K1 and K5 capsular antigens in two disparate parts of the developed world. It is anticipated that such a study may also provide a baseline for the presence of the virulence factors, which characterize those E. coli, which
K.A. Bettelheim et al. / Comp. Immun. Microbiol. Infect. Dis. 26 (2003) 55±63
57
Table 1 Characteristics of E. coli strains (n/a not applicable)
No. of strains studied No. of different O:H serotypes found No. of K1 strains No. of K5 strains No. of K? strains
Total
%
Berlin
%
Melbourne
%
450 220
n/a
264 135
n/a
186 124
n/a
76 26 348
16.9 5.8 77.3
42 10 212
15.9 3.8 80.3
34 16 136
18.3 8.6 73.1
are still important causes of morbidity and mortality amongst infants and children in the developed world. While there have been a number of studies on the carriage of E. coli by neonates and adults, the E. coli ¯ora of infants under one year of age in the post-neonatal period has been poorly documented. 2. Materials and methods 2.1. Strains of E. coli Faecal specimens from randomly selected infants aged under one year, who had not had any gastrointestinal symptoms for at least one week prior to the collection of the specimen and who were currently not taking any antibiotic therapy were collected in Berlin, Germany between 1986 and 1987 and Melbourne, Australia, during the early 1990's. About one gram of faeces was suspended in 10 ml Nutrient broth and a loop-full spread onto Endo-Agar (Berlin), or MacConkey agar (Melbourne). Up to 10 coliform E. coli-like colonies were selected after overnight incubation at 37 8C, such that representatives of each type of colonial morphology were selected. Generally one to four colonies per specimen was selected. 2.2. Characterization of the strains of E. coli The strains were characterized as E. coli on the basis of their reactions in triple-sugariron (TSI) agar (Oxoid), motility-indole-ornithine (MIO) medium (Oxoid), ONPG (Oxoid) and urease broth (Oxoid). Those strains, were considered to be E. coli if they produced acid in both the butt and slope of the TSI agar, and produced gas but did not produce H2S; if they produced indole and ornithine in the MIO medium; if they were ONPG-positive and urease negative. If they produced aberrant results in any of these tests they were subjected to additional tests. If they conformed to the de®nition of E. coli, they were then O and H serotyped using previously described methods [9,10]. Their ability to produce the K1 and K5 capsular antigens was determined using K1 and K5 speci®c phages [11,12]. K1 phages A±E and the K1 positive strain SC735 (O2:K1) were kindly donated by R.J. Gross, PHLS, London, UK. The K5 phages and the K5 positive strain Bi 8337-41 (O10:K5:H4) were from B. Jann, Max-Planck Institut fuÈr Immunbiologie, Freiburg, Germany.
58
K.A. Bettelheim et al. / Comp. Immun. Microbiol. Infect. Dis. 26 (2003) 55±63
Table 2 Serotypes of E. coli isolated from healthy infants in Berlin and Melbourne Serotype Ont:K?:HO6:K?:HO4:K?:H5 OR:K?:HO1:K1:H7 O18:K1:H7 O15:K?:H1 O1:K1:HOnt:K?:H4 O75:K5:HO2:K1:H7 OR:K?:H16 O6:K5:H1 O2:K1:H4 O2:K1:H1 Ont:K?:H1 O6:K?:H1 O17:K?:H18 O25:K5:H1 O77:K?:H18 OR:K?:HR Ont:K?:H10 O2:K?:H4 O162:H10 O11:K?:HO17:K?:HO125:H30 Ont:K?:H7 Ont:K?:H16 Ont:K?:H21 OR:K1:HOR:K1:H4 O8:K?:Hnt Total
Total no. of strains isolated
No. isolated in Berlin
No. isolated in Melbourne
24 16 13 11 9 9 9 8 7 7 6 6 5 5 4 4 4 4 3 3 3 3 3 3 2 2 2 2 2 2 2 2 2
20 10 4 8 4 6 7 5 4 5 3 3 2 3 2 2 3 3 1 1 1 2 2 2 1 1 1 1 1 1 1 1 1
4 6 9 3 5 3 2 3 3 2 3 3 3 2 2 2 1 1 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1
187
112
75
3. Results The general distribution of the strains studied and their characteristics is summarized in Table 1. The most common O:H serotypes found in each of these cities are given in Table 2. Of all the 220 different O:H serotypes isolated, 179 were only isolated once, 90 in Berlin and 89 in Melbourne and 30 of the 41 O:H serotypes (73.2%) found more than once were isolated in both centres. The numbers of potentially pathogenic E. coli serotypes found is listed in Table 3. There were no potentially Enteroinvasive E. coli (EIEC) serotypes found
K.A. Bettelheim et al. / Comp. Immun. Microbiol. Infect. Dis. 26 (2003) 55±63
59
Table 3 Potentially pathogenic serotypes of E. coli isolated from healthy infants in Berlin and Melbourne Serotype
O5:HO7:HO8:H9 O15:H11 O18:H7 O25:HO26:H11 O27:H7 O77:H18 O86:HO113:H21 O128:H2 O173:HTotal
No. isolated in Berlin
No. isolated in Melbourne
1 7 1 1 6 3 0 1 1 1 1 0 1
0 0 0 0 3 0 1 0 2 0 0 1 0
24
7
No. of potential EaggEC serotypes 7
No. of potential EHEC serotypes
No. of potential EPEC serotypes
1
9 1 3 1
11
No. of potential ETEC serotypes
1
1 1 3 1
1
1 1
1
4
12
1 7
but potentially Enteroaggregative (EaggEC), Enteropathogenic (EPEC), and Enterotoxigenic (ETEC) E.coli were found. Of the 76 isolates carrying the K1 antigen, the seven different serotypes, which were isolated singly in Berlin were O2:H6; O2:H-; O16:H-; O25:H-; O45:H7; O46:H31 and O117:H7. Those 14 isolated singly in Melbourne were O2:H39; O4:H10; O13:H4; O13:H17; O13:H45; O18:H-; O18:H44; O62:H48; O75:H7; O85:H-; O135:H10; O164:H31; OR:H-; OR:H31. The remaining K1 isolates are listed in Table 4. The 26 strains carrying the K5 antigen are listed in Table 5. Table 4 Serotypes of E. coli isolated from healthy infants in Berlin and Melbourne, carrying the K1 antigen Serotype O1.HO1.H7 O2.H1 O2.H4 O2.H7 O7.HO18.H7 O83.H33 OR.HOR.H4 OR.H7 Total
No. of strains isolated in Berlin
No. of strains isolated in Melbourne
Total no.
5 4 2 3 3 6 6 2 1 1 2
3 5 2 2 3 0 3 0 1 1 0
8 9 4 5 6 6 9 2 2 2 2
35
20
55
60
K.A. Bettelheim et al. / Comp. Immun. Microbiol. Infect. Dis. 26 (2003) 55±63
Table 5 O:H serotypes of E. coli isolated from healthy infants in Berlin and Melbourne, which carry the K5 antigen Serotype O6.HO6.H1 O6.H32 O25.H1 O75.HO75.H28 O149.H16 Ont.HOnt.H1 Ont.H12 OR.H6 Total
No. of strains isolated in Berlin
No. of strains isolated in Melbourne
Total no.
0 2 0 1 5 1 0 0 0 0 1
3 3 1 2 2 0 1 1 2 1 0
3 5 1 3 7 1 1 1 2 1 1
10
16
26
4. Discussion This study highlights the diversity of E. coli serotypes normally carried by healthy infants. However, there appear to be certain serotypes, which are much more likely to be encountered than others. A characterization rate of 220 different serotypes from 450 isolates (Table 1) suggests that the mean number of each serotype isolated should be around two. The serotypes found can thus be divided into two groups, those isolated singly and those found twice or more frequently in this investigation. Of the latter group those found in both centers are considered especially signi®cant. Of the O-typeable serotypes, it is noteworthy that those belonging to O-groups O1, O2, O4, O6, O15, O17, O18, O25, O75, O77 and O162 (Table 2) are in a prominent position. These O-groups have regularly been reported in the literature as commonly isolated from human faeces, and therefore it is not surprising that serotypes carrying them are common in both the infants from Berlin and Melbourne. Many of these common serotypes were also found to carry the K1 or K5 antigens and the range of H antigens is also relatively limited to H1, H4, H5, H7, H10 and H16. The serotypes O1:K1:H7; O2:K1:H1; O2:K1:H4; O4:K?:H5; O6:K5:H1 and O75:K5:H-, which all occur near the top of Table 2 are some of the most common human isolates reported in the world literature. Thus, this study con®rms the view that certain serotypes of E. coli may be more likely to be adapted to the human intestine than other serotypes. These listed at the top of Table 2 may well be some of those. However, the fact that there were also found a number of O-untypeable strains, suggests that there may well be more such serotypes in our environment. A strong relationship between serotype and pathotype of E. coli has been extensively reported [8]. It was thus considered useful to determine whether any of the serotypes often associated with certain pathotypes were present among this collection of E. coli (Table 3). Of the 31 possibly pathogenic serotypes found, possible EAggEC or EPEC were the majority, with ETEC and EHEC serotypes being less frequently found. However, the
K.A. Bettelheim et al. / Comp. Immun. Microbiol. Infect. Dis. 26 (2003) 55±63
61
presence of such a number (6.8%) of potentially pathogenic E. coli serotypes in healthy infants needs to be given serious consideration and studies on the presence of speci®c virulence factors among these organisms are in progress to determine the range of types present. It has been found to be statistically signi®cant that all seven of the possible ETEC serotypes were found in Berlin (X 2 with Yates correction is 3.42). Similarly, it is statistically signi®cant (X 2 with Yates correction is 4.03), that only seven potentially pathogenic serotypes out of 186 were found in Melbourne, while 24 out of 264 were found in Berlin. Further studies on speci®c virulence factors should clarify the situation. The serotypes most commonly associated with the K1 antigen were restricted to ®ve O groups O1, O2, O7, O18 and O83 (Table 4). They represent some of the most commonly isolated human derived E. coli serotypes. All these serotypes were found in both cities, further con®rming their widespread nature. Similar results were also found among the K5 antigen-carrying serotypes (Table 5). Three serotypes O6:K5:H1, O25:K5:H1 and O75:K5:H- were found in both cities and these again comprise some of the more common human E. coli serotypes, con®rming the ubiquitous nature of some serotypes commonly associated with humans. These two K antigens have been especially associated with extraintestinal E. coli infections and strains with these K antigens were found to be signi®cantly more resistant to the alternate complement pathway than strains lacking them [13]. These studies demonstrate both the great diversity of E. coli serotypes present in the human intestinal tract, but also that certain serotypes appear to predominate regardless of geographical region. E. coli are organisms which are common inhabitants of the human intestinal tract, but also capable of being important human pathogens. It is important to build up a basic data set of the types of E. coli normally present in healthy humans, especially in infants and children, who are most vulnerable to E. coli infections. This would provide a greater understanding of the ecology of E. coli in the human intestinal tract, and their ability to become pathogens. Many virulence factors are plasmid/phage mediated and there appears to be a strong correlation between strains of certain serotypes carrying certain virulence factors. A recent example of this is ®nding that common ovine Shigatoxigenic E. coli (STEC) serotypes and human isolates of the same serotype possess the same Stx2d toxin type [14]. This correlation suggests that there is a strong link not just between serotype and pathotype but between serotype and the subtypes of the virulence factors, which it might acquire and the type of host it might be able to colonize. A study carried out many years ago on strains of E. coli, which had previously been isolated from the faeces of healthy people and cattle, and which belonged to certain serotypes likely to be toxigenic [15,16] in some cases did carry such factors. The purpose of this study was speci®cally to establish a baseline of the carriage of E. coli serotypes carried by healthy infants in the developed world. Whilst a great diversity was noted, it was considered especially noteworthy that certain serotypes appeared to predominate in both centers studied. The extensive studies currently being undertaken, on STEC serotypes around the world and recently reviewed [17] show that while certain STEC serotypes found in animals, also cause disease in humans and other STEC serotypes seem to be more common in animals but only rarely isolated from humans. This paradox
62
K.A. Bettelheim et al. / Comp. Immun. Microbiol. Infect. Dis. 26 (2003) 55±63
may be due to these `animal-speci®c' serotypes not being able to adequately colonize the human intestinal tract. Only when the relationship between E. coli serotype and host of healthy individuals is established can the importance of other virulence factors be adequately addressed. There are many reports in the current literature on the carriage of STEC serotypes by healthy animals and by human patients, but there is a great paucity on the distribution of E. coli in general in the healthy human population. It is partially to remedy this situation that the present study was undertaken and it is hoped that it may stimulate further such investigations. Only when there is an understanding of the ecology of E. coli serotypes within the healthy human population, can the importance and role of E. coli carrying virulence factors be fully addressed, especially as such factors can frequently be lost or gained. Acknowledgements We wish to acknowledge with thanks the help from Mrs Alexandra Bettelheim in providing the French translation of the abstract. References [1] Escherich Th. Die Darmbakterien des Neugeborenen und des SaÈuglings. Fortschritte der Medizin 1885;3:515±22 see also pages 547±54. [2] Mitsuoka T, Hayakawa K. Die Faecal¯ora bei Menschen: I. Mitteilung: Die Zusammensetzung der Faecal¯ora der verschiedenen Altersgruppen. Zbl Bakt Paras Hyg, I Abt Orig A 1972;223:333±42. [3] Mitsuoka T, Hayakawa K, Kimura N. Die Faecal¯ora bei Menschen. III. Mitteilung: Die Zusammensetzung der faecal¯ora der verschiedenen Altersgruppen. Zbl Bakt Paras Hyg, I Abt Orig A 1975;232:499±511. [4] Bettelheim KA, Cooke EM, O'Farrell S, Shooter RA. The effect of diet on intestinal Escherichia coli. J Hyg Cambridge 1977;79:43±45. [5] Bettelheim KA, Lennox-King SMJ. The acquisition of Escherichia coli by newborn babies. Infection 1976;4:174±9. [6] Robbins JB, McCracken GH, Gotschlich EC, érskov F, érskov I, Hanson LA. Escherichia coli K1 capsular polysaccharide associated with neonatal meningitis. N Engl J Med 1974;290:1216±20. [7] Johnson JR. Virulence factors in Escherichia coli urinary tract infection. Clin Microbiol Rev 1991;4:80± 128. [8] Lior H. Classi®cation of Escherichia coli. In: Gyles CL, editor. Escherichia coli in domestic animals and humans, Wallingford, UK: CAB International, 1994. [9] Bettelheim KA, Thompson CJ. New method of serotyping Escherichia coli: implementation and veri®cation. J Clin Microbiol 1987;25:781±6. [10] Chandler ME, Bettelheim KA. A rapid method of identifying Escherichia coli H antigens. Zbl Bakt Paras Hyg, I Abt Orig A 1974;129:74±9. [11] Wullenweber M, Beutin L, Zimmermann S, Jonas C. In¯uence of some bacterial and host factors on colonization and invasiveness of Escherichia coli K1 in neonatal rats. Infect Immun 1993;61:2138±44. [12] Gross R, Cheasty T, Rowe B. Isolation of bacteriophages speci®c for the K1 polysaccharide antigen of Escherichia coli. J Clin Microbiol 1977;6:548±50. [13] Devine DA, Roberts AP. K1, K5 and O antigens of Escherichia coli in relation to serum killing via the classical and alternative complement pathways. J Med Microbiol 1994;41:139±44. [14] Ramachandran V, Hornitzky MA, Bettelheim KA, Walker MJ, Djordjevic SP. The common ovine Shiga toxin 2-containing Escherichia coli serotypes and human isolates of the same serotypes possess a Stx2d toxin type. J Clin Microbiol 2001;39:1932±7. [15] Bettelheim KA, Wilson MW, Shooter RA, O'Farrell SM. Studies on the enterotoxigenicity of environmental
K.A. Bettelheim et al. / Comp. Immun. Microbiol. Infect. Dis. 26 (2003) 55±63
63
Escherichia coli, belonging to serotypes normally considered enterotoxigenic. J Hyg, Cambridge 1980; 84:411±4. [16] Bettelheim KA, Wilson MW. The enterotoxigenicity of strains of Escherichia coli isolated from the faeces of healthy people and cattle. J Hyg Cambridge 1982;88:121±3. [17] Bettelheim KA. Role of non-O157 VTEC. J Appl Microbiol 2000;88:38S±50S.
Comparative Immunology, Microbiology & Infectious Diseases 26 (2003) 55±63
55
www.elsevier.com/locate/cimid
Serotypes of Escherichia coli isolated from healthy infants in Berlin, Germany and Melbourne, Australia K.A. Bettelheim a,*, L. Beutin b, K. Gleier b, J.L. Pearce c, R.K.J. Luke c, S. Zimmermann b a
National Escherichia coli Reference Laboratory, Microbiological Diagnostic Unit, Public Health Laboratory, Department of Microbiology and Immunology, University of Melbourne, Melbourne, Vic. 3010, Australia b Robert Koch Institut, Division of Emerging Pathogens, Berlin, Germany c Department of Agricultural Sciences, Faculty of Science, Technology and Engineering, School of Life Sciences, La Trobe University, Melbourne, Vic. 3083, Australia Received 12 September 2001; received in revised form 20 December 2001; accepted 17 January 2002
Abstract The characterization of Escherichia coli strains isolated from healthy infants under one year of age with respect to O:H serotype, K1 and K5 antigens in two disparate parts of the developed world was the purpose of this investigation. A total of 450 strains were examined, 264 from Berlin and 186 from Melbourne. Of all the 220 different O:H serotypes found, 179 were only isolated once, 90 in Berlin and 89 in Melbourne. However, 30 of the 41 O:H serotypes (73.2%) found more than once were isolated in both centers. The most commonly identi®ed serotypes were found in both centers and included O1:H-; O1:H7; O2:H1; O2:H4; O2:H7; O4:H5; O6:H-; O6:H1; O15:H1; O18:H7; O25:H1; and O75:H-. Potentially pathogenic serotypes were found in both cities. Enteropathogenic E. coli (EPEC)-associated serotypes (O18:H7; O26:H-; O44:H34; O86:H-; O128:H2) were present in 11 cases and enterohaemorrhagic E. coli (EHEC)-associated types including O26:H11; O128:H2) were present in four cases. The distributions of serotypes found were similar in the two cities, strongly suggesting the wider applicability of these results. q 2002 Elsevier Science Ltd. All rights reserved. Keywords: Escherichia coli; Healthy infants; Serotypes; Berlin; Melbourne
ReÂsume Le but de cette enqueÃte eÂtait de de®nir les caracteÂristiques des souches d'Escherichia coli deÂceleÂes chez de jeunes enfants sains aÃgeÂs de moins d'un an en ce qui concerne le seÂrotype O:H, avec les * Corresponding author. Tel.: 161-3-8344-8180; fax: 161-3-8344-7833. E-mail addresses: [email protected] (K.A. Bettelheim), [email protected] (K.A. Bettelheim). 0147-9571/03/$ - see front matter q 2002 Elsevier Science Ltd. All rights reserved. PII: S 0147-957 1(02)00015-2
56
K.A. Bettelheim et al. / Comp. Immun. Microbiol. Infect. Dis. 26 (2003) 55±63
antigeÁnes K1 et K5 dans deux parties disparates du monde deÂveloppeÂ. Un total de 450 souches ont eÂte examineÂes. 264 souches de Berlin et 186 souches de Melbourne. Sur tous les 220 diffeÂrents seÂrotypes O:H trouveÂs, 179 ont eÂte deÂceleÂs une fois seulement, 90 aÁ Berlin et 89 aÁ Melbourne. Cependant, 30 des 41 seÂrotypes O:H (73.2%) trouveÂs plus d'une fois ont eÂte deÂceleÂs dans les deux centres. Les seÂrotypes identi®eÂs le plus souvent ont eÂte trouveÂs dans les deux centres et comprenaient O1:H-; O1:H7; O2:H1; O2:H4; O2:H7; O4:H5; O6:H-; O6:H1; O15:H1; O18:H7; O25:H1; et O75:H-. Potentiellement des seÂrotypes pathogeÁnes ont eÂte trouveÂs dans les deux villes. Les seÂrotypes causant E. coli enteÂropathogeÁne (EPEC) (O18:H7; O26:H-; O44:H34; O86:H-; O128:H2) eÂtaient preÂsents dans 11 cas et les types causant E. coli enteÂroheÂmorragique (EHEC) qui comprenaient O26:H11; O128:H2) eÂtaient preÂsents dans quatre cas. La distribution des seÂrotypes deÂceleÂs eÂtait similaire dans les deux villes, ce qui semble fortement indiquer la plus grande applicabilite de ces reÂsultats. q 2002 Elsevier Science Ltd. All rights reserved. Mots-cleÂ: Escherichia coli; Jeunes enfants sains; seÂrotypes; Berlin; Melbourne
1. Introduction It is a tribute to the ubiquity, variety and adaptability of the Escherichia coli species that their role in the human faecal ¯ora is still being explored, so long after the pioneering studies on neonatal faeces, of Escherich [1], which led to the characterization of these organisms. The normal human faecal ¯ora has been estimated at around 10 10 ±10 11 per gram [2,3]. The Enterobacteriaceae are present at about 10 8 ±10 9 per gram and E. coli are the most important enterobacterial species. Adults continuously acquire new types of E. coli with their food, which then set up a competitive situation with the resident ¯ora [4]. Neonates have comparatively high levels of Enterobacteriaceae. They acquire them at birth from their mothers [5] and to a lesser extent from other people in close contact with them. Thus, healthy infants under 1 year of age will carry an E. coli ¯ora re¯ecting the predominant types in the human population of their community at that time. Infants are also susceptible to intestinal infections with pathogenic E. coli. In the developing world, enteropathogenic E. coli (EPEC) and enterotoxigenic E. coli (ETEC) are still major causes of neonatal morbidity and mortality, although they are not major problems in developed countries like Australia and Germany. In the developed world, E. coli, especially those carrying the K1 capsular antigen, is a major cause of neonatal meningitis [6]. In addition, K1 and K5 belong to the K-types, which are over-represented among E. coli isolated from patients with cystitis and pyelonephritis in comparison to faecal E. coli [7]. Also infections by enterohaemorrhagic E. coli (EHEC), which are characterized by their ability to produce one or both Shiga-toxins (Stx), the adherence factor intimin (EAE) and Enterohaemolysin are major causes of infant morbidity and mortality. It has also been noted in a number of studies [8], that certain serotypes of E. coli are far more likely to be associated with some of these virulence factors, than other serotypes. The aim of this study was to characterize strains of E. coli isolated from healthy infants under one year of age with respect to O:H serotype, K1 and K5 capsular antigens in two disparate parts of the developed world. It is anticipated that such a study may also provide a baseline for the presence of the virulence factors, which characterize those E. coli, which
K.A. Bettelheim et al. / Comp. Immun. Microbiol. Infect. Dis. 26 (2003) 55±63
57
Table 1 Characteristics of E. coli strains (n/a not applicable)
No. of strains studied No. of different O:H serotypes found No. of K1 strains No. of K5 strains No. of K? strains
Total
%
Berlin
%
Melbourne
%
450 220
n/a
264 135
n/a
186 124
n/a
76 26 348
16.9 5.8 77.3
42 10 212
15.9 3.8 80.3
34 16 136
18.3 8.6 73.1
are still important causes of morbidity and mortality amongst infants and children in the developed world. While there have been a number of studies on the carriage of E. coli by neonates and adults, the E. coli ¯ora of infants under one year of age in the post-neonatal period has been poorly documented. 2. Materials and methods 2.1. Strains of E. coli Faecal specimens from randomly selected infants aged under one year, who had not had any gastrointestinal symptoms for at least one week prior to the collection of the specimen and who were currently not taking any antibiotic therapy were collected in Berlin, Germany between 1986 and 1987 and Melbourne, Australia, during the early 1990's. About one gram of faeces was suspended in 10 ml Nutrient broth and a loop-full spread onto Endo-Agar (Berlin), or MacConkey agar (Melbourne). Up to 10 coliform E. coli-like colonies were selected after overnight incubation at 37 8C, such that representatives of each type of colonial morphology were selected. Generally one to four colonies per specimen was selected. 2.2. Characterization of the strains of E. coli The strains were characterized as E. coli on the basis of their reactions in triple-sugariron (TSI) agar (Oxoid), motility-indole-ornithine (MIO) medium (Oxoid), ONPG (Oxoid) and urease broth (Oxoid). Those strains, were considered to be E. coli if they produced acid in both the butt and slope of the TSI agar, and produced gas but did not produce H2S; if they produced indole and ornithine in the MIO medium; if they were ONPG-positive and urease negative. If they produced aberrant results in any of these tests they were subjected to additional tests. If they conformed to the de®nition of E. coli, they were then O and H serotyped using previously described methods [9,10]. Their ability to produce the K1 and K5 capsular antigens was determined using K1 and K5 speci®c phages [11,12]. K1 phages A±E and the K1 positive strain SC735 (O2:K1) were kindly donated by R.J. Gross, PHLS, London, UK. The K5 phages and the K5 positive strain Bi 8337-41 (O10:K5:H4) were from B. Jann, Max-Planck Institut fuÈr Immunbiologie, Freiburg, Germany.
58
K.A. Bettelheim et al. / Comp. Immun. Microbiol. Infect. Dis. 26 (2003) 55±63
Table 2 Serotypes of E. coli isolated from healthy infants in Berlin and Melbourne Serotype Ont:K?:HO6:K?:HO4:K?:H5 OR:K?:HO1:K1:H7 O18:K1:H7 O15:K?:H1 O1:K1:HOnt:K?:H4 O75:K5:HO2:K1:H7 OR:K?:H16 O6:K5:H1 O2:K1:H4 O2:K1:H1 Ont:K?:H1 O6:K?:H1 O17:K?:H18 O25:K5:H1 O77:K?:H18 OR:K?:HR Ont:K?:H10 O2:K?:H4 O162:H10 O11:K?:HO17:K?:HO125:H30 Ont:K?:H7 Ont:K?:H16 Ont:K?:H21 OR:K1:HOR:K1:H4 O8:K?:Hnt Total
Total no. of strains isolated
No. isolated in Berlin
No. isolated in Melbourne
24 16 13 11 9 9 9 8 7 7 6 6 5 5 4 4 4 4 3 3 3 3 3 3 2 2 2 2 2 2 2 2 2
20 10 4 8 4 6 7 5 4 5 3 3 2 3 2 2 3 3 1 1 1 2 2 2 1 1 1 1 1 1 1 1 1
4 6 9 3 5 3 2 3 3 2 3 3 3 2 2 2 1 1 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1
187
112
75
3. Results The general distribution of the strains studied and their characteristics is summarized in Table 1. The most common O:H serotypes found in each of these cities are given in Table 2. Of all the 220 different O:H serotypes isolated, 179 were only isolated once, 90 in Berlin and 89 in Melbourne and 30 of the 41 O:H serotypes (73.2%) found more than once were isolated in both centres. The numbers of potentially pathogenic E. coli serotypes found is listed in Table 3. There were no potentially Enteroinvasive E. coli (EIEC) serotypes found
K.A. Bettelheim et al. / Comp. Immun. Microbiol. Infect. Dis. 26 (2003) 55±63
59
Table 3 Potentially pathogenic serotypes of E. coli isolated from healthy infants in Berlin and Melbourne Serotype
O5:HO7:HO8:H9 O15:H11 O18:H7 O25:HO26:H11 O27:H7 O77:H18 O86:HO113:H21 O128:H2 O173:HTotal
No. isolated in Berlin
No. isolated in Melbourne
1 7 1 1 6 3 0 1 1 1 1 0 1
0 0 0 0 3 0 1 0 2 0 0 1 0
24
7
No. of potential EaggEC serotypes 7
No. of potential EHEC serotypes
No. of potential EPEC serotypes
1
9 1 3 1
11
No. of potential ETEC serotypes
1
1 1 3 1
1
1 1
1
4
12
1 7
but potentially Enteroaggregative (EaggEC), Enteropathogenic (EPEC), and Enterotoxigenic (ETEC) E.coli were found. Of the 76 isolates carrying the K1 antigen, the seven different serotypes, which were isolated singly in Berlin were O2:H6; O2:H-; O16:H-; O25:H-; O45:H7; O46:H31 and O117:H7. Those 14 isolated singly in Melbourne were O2:H39; O4:H10; O13:H4; O13:H17; O13:H45; O18:H-; O18:H44; O62:H48; O75:H7; O85:H-; O135:H10; O164:H31; OR:H-; OR:H31. The remaining K1 isolates are listed in Table 4. The 26 strains carrying the K5 antigen are listed in Table 5. Table 4 Serotypes of E. coli isolated from healthy infants in Berlin and Melbourne, carrying the K1 antigen Serotype O1.HO1.H7 O2.H1 O2.H4 O2.H7 O7.HO18.H7 O83.H33 OR.HOR.H4 OR.H7 Total
No. of strains isolated in Berlin
No. of strains isolated in Melbourne
Total no.
5 4 2 3 3 6 6 2 1 1 2
3 5 2 2 3 0 3 0 1 1 0
8 9 4 5 6 6 9 2 2 2 2
35
20
55
60
K.A. Bettelheim et al. / Comp. Immun. Microbiol. Infect. Dis. 26 (2003) 55±63
Table 5 O:H serotypes of E. coli isolated from healthy infants in Berlin and Melbourne, which carry the K5 antigen Serotype O6.HO6.H1 O6.H32 O25.H1 O75.HO75.H28 O149.H16 Ont.HOnt.H1 Ont.H12 OR.H6 Total
No. of strains isolated in Berlin
No. of strains isolated in Melbourne
Total no.
0 2 0 1 5 1 0 0 0 0 1
3 3 1 2 2 0 1 1 2 1 0
3 5 1 3 7 1 1 1 2 1 1
10
16
26
4. Discussion This study highlights the diversity of E. coli serotypes normally carried by healthy infants. However, there appear to be certain serotypes, which are much more likely to be encountered than others. A characterization rate of 220 different serotypes from 450 isolates (Table 1) suggests that the mean number of each serotype isolated should be around two. The serotypes found can thus be divided into two groups, those isolated singly and those found twice or more frequently in this investigation. Of the latter group those found in both centers are considered especially signi®cant. Of the O-typeable serotypes, it is noteworthy that those belonging to O-groups O1, O2, O4, O6, O15, O17, O18, O25, O75, O77 and O162 (Table 2) are in a prominent position. These O-groups have regularly been reported in the literature as commonly isolated from human faeces, and therefore it is not surprising that serotypes carrying them are common in both the infants from Berlin and Melbourne. Many of these common serotypes were also found to carry the K1 or K5 antigens and the range of H antigens is also relatively limited to H1, H4, H5, H7, H10 and H16. The serotypes O1:K1:H7; O2:K1:H1; O2:K1:H4; O4:K?:H5; O6:K5:H1 and O75:K5:H-, which all occur near the top of Table 2 are some of the most common human isolates reported in the world literature. Thus, this study con®rms the view that certain serotypes of E. coli may be more likely to be adapted to the human intestine than other serotypes. These listed at the top of Table 2 may well be some of those. However, the fact that there were also found a number of O-untypeable strains, suggests that there may well be more such serotypes in our environment. A strong relationship between serotype and pathotype of E. coli has been extensively reported [8]. It was thus considered useful to determine whether any of the serotypes often associated with certain pathotypes were present among this collection of E. coli (Table 3). Of the 31 possibly pathogenic serotypes found, possible EAggEC or EPEC were the majority, with ETEC and EHEC serotypes being less frequently found. However, the
K.A. Bettelheim et al. / Comp. Immun. Microbiol. Infect. Dis. 26 (2003) 55±63
61
presence of such a number (6.8%) of potentially pathogenic E. coli serotypes in healthy infants needs to be given serious consideration and studies on the presence of speci®c virulence factors among these organisms are in progress to determine the range of types present. It has been found to be statistically signi®cant that all seven of the possible ETEC serotypes were found in Berlin (X 2 with Yates correction is 3.42). Similarly, it is statistically signi®cant (X 2 with Yates correction is 4.03), that only seven potentially pathogenic serotypes out of 186 were found in Melbourne, while 24 out of 264 were found in Berlin. Further studies on speci®c virulence factors should clarify the situation. The serotypes most commonly associated with the K1 antigen were restricted to ®ve O groups O1, O2, O7, O18 and O83 (Table 4). They represent some of the most commonly isolated human derived E. coli serotypes. All these serotypes were found in both cities, further con®rming their widespread nature. Similar results were also found among the K5 antigen-carrying serotypes (Table 5). Three serotypes O6:K5:H1, O25:K5:H1 and O75:K5:H- were found in both cities and these again comprise some of the more common human E. coli serotypes, con®rming the ubiquitous nature of some serotypes commonly associated with humans. These two K antigens have been especially associated with extraintestinal E. coli infections and strains with these K antigens were found to be signi®cantly more resistant to the alternate complement pathway than strains lacking them [13]. These studies demonstrate both the great diversity of E. coli serotypes present in the human intestinal tract, but also that certain serotypes appear to predominate regardless of geographical region. E. coli are organisms which are common inhabitants of the human intestinal tract, but also capable of being important human pathogens. It is important to build up a basic data set of the types of E. coli normally present in healthy humans, especially in infants and children, who are most vulnerable to E. coli infections. This would provide a greater understanding of the ecology of E. coli in the human intestinal tract, and their ability to become pathogens. Many virulence factors are plasmid/phage mediated and there appears to be a strong correlation between strains of certain serotypes carrying certain virulence factors. A recent example of this is ®nding that common ovine Shigatoxigenic E. coli (STEC) serotypes and human isolates of the same serotype possess the same Stx2d toxin type [14]. This correlation suggests that there is a strong link not just between serotype and pathotype but between serotype and the subtypes of the virulence factors, which it might acquire and the type of host it might be able to colonize. A study carried out many years ago on strains of E. coli, which had previously been isolated from the faeces of healthy people and cattle, and which belonged to certain serotypes likely to be toxigenic [15,16] in some cases did carry such factors. The purpose of this study was speci®cally to establish a baseline of the carriage of E. coli serotypes carried by healthy infants in the developed world. Whilst a great diversity was noted, it was considered especially noteworthy that certain serotypes appeared to predominate in both centers studied. The extensive studies currently being undertaken, on STEC serotypes around the world and recently reviewed [17] show that while certain STEC serotypes found in animals, also cause disease in humans and other STEC serotypes seem to be more common in animals but only rarely isolated from humans. This paradox
62
K.A. Bettelheim et al. / Comp. Immun. Microbiol. Infect. Dis. 26 (2003) 55±63
may be due to these `animal-speci®c' serotypes not being able to adequately colonize the human intestinal tract. Only when the relationship between E. coli serotype and host of healthy individuals is established can the importance of other virulence factors be adequately addressed. There are many reports in the current literature on the carriage of STEC serotypes by healthy animals and by human patients, but there is a great paucity on the distribution of E. coli in general in the healthy human population. It is partially to remedy this situation that the present study was undertaken and it is hoped that it may stimulate further such investigations. Only when there is an understanding of the ecology of E. coli serotypes within the healthy human population, can the importance and role of E. coli carrying virulence factors be fully addressed, especially as such factors can frequently be lost or gained. Acknowledgements We wish to acknowledge with thanks the help from Mrs Alexandra Bettelheim in providing the French translation of the abstract. References [1] Escherich Th. Die Darmbakterien des Neugeborenen und des SaÈuglings. Fortschritte der Medizin 1885;3:515±22 see also pages 547±54. [2] Mitsuoka T, Hayakawa K. Die Faecal¯ora bei Menschen: I. Mitteilung: Die Zusammensetzung der Faecal¯ora der verschiedenen Altersgruppen. Zbl Bakt Paras Hyg, I Abt Orig A 1972;223:333±42. [3] Mitsuoka T, Hayakawa K, Kimura N. Die Faecal¯ora bei Menschen. III. Mitteilung: Die Zusammensetzung der faecal¯ora der verschiedenen Altersgruppen. Zbl Bakt Paras Hyg, I Abt Orig A 1975;232:499±511. [4] Bettelheim KA, Cooke EM, O'Farrell S, Shooter RA. The effect of diet on intestinal Escherichia coli. J Hyg Cambridge 1977;79:43±45. [5] Bettelheim KA, Lennox-King SMJ. The acquisition of Escherichia coli by newborn babies. Infection 1976;4:174±9. [6] Robbins JB, McCracken GH, Gotschlich EC, érskov F, érskov I, Hanson LA. Escherichia coli K1 capsular polysaccharide associated with neonatal meningitis. N Engl J Med 1974;290:1216±20. [7] Johnson JR. Virulence factors in Escherichia coli urinary tract infection. Clin Microbiol Rev 1991;4:80± 128. [8] Lior H. Classi®cation of Escherichia coli. In: Gyles CL, editor. Escherichia coli in domestic animals and humans, Wallingford, UK: CAB International, 1994. [9] Bettelheim KA, Thompson CJ. New method of serotyping Escherichia coli: implementation and veri®cation. J Clin Microbiol 1987;25:781±6. [10] Chandler ME, Bettelheim KA. A rapid method of identifying Escherichia coli H antigens. Zbl Bakt Paras Hyg, I Abt Orig A 1974;129:74±9. [11] Wullenweber M, Beutin L, Zimmermann S, Jonas C. In¯uence of some bacterial and host factors on colonization and invasiveness of Escherichia coli K1 in neonatal rats. Infect Immun 1993;61:2138±44. [12] Gross R, Cheasty T, Rowe B. Isolation of bacteriophages speci®c for the K1 polysaccharide antigen of Escherichia coli. J Clin Microbiol 1977;6:548±50. [13] Devine DA, Roberts AP. K1, K5 and O antigens of Escherichia coli in relation to serum killing via the classical and alternative complement pathways. J Med Microbiol 1994;41:139±44. [14] Ramachandran V, Hornitzky MA, Bettelheim KA, Walker MJ, Djordjevic SP. The common ovine Shiga toxin 2-containing Escherichia coli serotypes and human isolates of the same serotypes possess a Stx2d toxin type. J Clin Microbiol 2001;39:1932±7. [15] Bettelheim KA, Wilson MW, Shooter RA, O'Farrell SM. Studies on the enterotoxigenicity of environmental
K.A. Bettelheim et al. / Comp. Immun. Microbiol. Infect. Dis. 26 (2003) 55±63
63
Escherichia coli, belonging to serotypes normally considered enterotoxigenic. J Hyg, Cambridge 1980; 84:411±4. [16] Bettelheim KA, Wilson MW. The enterotoxigenicity of strains of Escherichia coli isolated from the faeces of healthy people and cattle. J Hyg Cambridge 1982;88:121±3. [17] Bettelheim KA. Role of non-O157 VTEC. J Appl Microbiol 2000;88:38S±50S.