Previous Article | Next Article ![]()
Journal of Clinical Microbiology, April 2003, p. 1694-1700, Vol. 41, No. 4
0095-1137/03/$08.00+0 DOI: 10.1128/JCM.41.4.1694-1700.2003
Copyright © 2003, American Society for Microbiology. All Rights Reserved.
Laboratory of Enteric Pathogens, National Public Health Institute, FIN-00300 Helsinki,1 Department of Food and Environmental Hygiene, University of Helsinki, FIN-00014 Helsinki, Finland2
Received 27 September 2002/ Returned for modification 15 November 2002/ Accepted 7 January 2003
|
|
|---|
|
|
|---|
In Finland, since the 1980s clinical microbiology laboratories have sent their L. monocytogenes findings to the Laboratory of Enteric Pathogens (LEP), National Public Health Institute (KTL); in 1994, such submission became obligatory. Also, physicians have been obligated to report any culture-confirmed cases of human listeriosis to the National Infectious Disease Registry since 1994.
From 1990 through 2001, the annual number of cases of invasive listeriosis in Finland varied between about 20 and 50 (M. Jahkola, unpublished data; http://www.ktl.fi/ttr). On the basis of only limited O serotyping of L. monocytogenes strains, listeriosis cases, excluding an outbreak caused by a rare serotype, 3a (26), have been thought to be sporadic.
Serotyping is a classic phenotypic tool for epidemiological studies (11, 12, 32). Thirteen serotypes of L. monocytogenes have been identified. However, most isolates belong to only three serotypes, 1/2a, 1/2b, and 4b (10, 24). Although the epidemiological benefit of serotyping is limited, it has provided rapid information for the screening of isolates during suspected outbreaks. Various genotyping methods have been used successfully in more detailed epidemiological studies (2, 13, 18, 22, 26, 28, 31). Pulsed-field gel electrophoresis (PFGE) is one of these genotyping methods, and it has proven to be highly discriminating and reproducible (1, 6, 22).
To study retrospectively the potential occurrence of infection clusters caused by L. monocytogenes in Finland since 1990, both systematic serotyping for O and H antigens and genotyping by PFGE were initiated in 1997 and have been continued prospectively since then. An electronic database library of PFGE profiles and serotypes was created to make possible prompt comparison of the PFGE profiles of different strains and rapid identification of infection clusters.
|
|
|---|
|
View this table: [in a new window] |
TABLE 1. Origins of L. monocytogenes isolates
|
PFGE. DNA from L. monocytogenes strains isolated from 1990 to 2000 was prepared as described earlier (27), with the following modifications. The strains were grown on blood agar overnight at 37°C and then for 17 to18 h at 37°C in BHI broth. Two milliliters of this broth culture was mixed with 5 ml of cold PIV buffer (10 mM Tris [pH 7.5], 1 M NaCl). The mixture was centrifuged at 3,000 rpm for 15 min at 4°C in a Midispin 2160 (LKB, Bromma, Sweden), and the cell pellet was suspended in 750 µl of cold PIV buffer. This cell suspension was mixed in equal parts with molten 2% low-melting-point agarose (SeaPlaque agarose; FMC BioProducts, Rockland, Maine), and the mixture was pipetted into plug molds. The plugs were incubated overnight at 37°C in a buffer containing 6 mM Tris-HCl (pH 7.5), 1 M NaCl, 100 mM EDTA, 0.5% Brij 58, 0.2% sodium deoksilate, and 0.5% sodium lauroylsarcosine and supplemented with 1 mg of lysozyme per ml and again overnight at 55 to 57°C in a buffer containing 0.5 M EDTA (pH 9.5) and 1% sodium lauroylsarcosine and supplemented with 0.3 mg of proteinase K per ml. The washing of the plugs and the conditions for restriction endonuclease digestion and PFGE were as described previously (25). Chromosomal DNA was digested overnight with 5 U of AscI (New England BioLabs Inc., Beverly, Mass.).
In 2001, the shorter protocol described by Graves and Swaminathan (17) for the preparation of genomic DNA was brought into use, with slight modifications. Bacterial cells were suspended in 2 ml of CBS (100 mM Tris, 100 mM EDTA, pH 8.0) to an optical density at 450 nm of 0.7 to 0.8. When the cell suspension containing lysozyme was mixed with low-melting-point agarose, 2% SeaPlaque agarose and 0.2 mg of proteinase K/ml were used without 1% sodium dodecyl sulfate. Chromosomal DNA was digested for 4 h with 10 U of AscI.
Electrophoresis was performed at 210 V with 1.0% Pronadisa D-5 agarose gels (Hispanlab, Madrid, Spain) by using a CHEF Mapper or CHEF-DR system (Bio-Rad Laboratories, Richmond, Calif.). Running conditions for AscI-digested DNA were 1 to 28 s for 10 h, followed by 28 to 30 s for 10 h. Low-range PFG markers (New England BioLabs Inc.) were used as molecular weight standards. The gels were visualized on a UV transilluminator and were photographed by using AlphaImager 1220 (Alpha Innotec Corporation, San Leandro, Calif.). The TIFF images were analyzed by using BioNumerics software (Applied Maths, Kortrijk, Belgium) and were normalized by using the low-range PFG marker standards on each gel. Any difference between two PFGE profiles was considered sufficient to distinguish these profiles. The different PFGE profiles were marked with numbers based on the coding agreed upon with the National Veterinary and Food Research Institute (L. Rantala, S. Lukinmaa, A. Siitonen, and T. Honkanen-Buzalski, Abstr. ISOPOL XIV Int. Symp. Problems Listeriosis, p. 155, 2001). Similarity values were calculated by the unweighted pair-group method with arithmetic averages and the Dice coefficient by using BioNumerics software.
Groups of PFGE types. BioNumerics software was used to compare all PFGE types with PFGE types containing five or more strains. When the similarity value was over 80% and the number of fragment differences between the profiles was three or less, the PFGE type was regarded as being closely related to the one with which it was compared (40). Groups of closely related types were designated with the letter G or with the letters GT when a type containing at least five strains was not related to any other type; the group number indicates the PFGE type with which all of the other PFGE types were compared.
Statistical methods. Fisher's exact two-tailed test (Epi-Info 6.04 software; World Health Organization, Geneva, Switgerland, and Centers for Disease Control and Prevention, Atlanta, Ga.) was used for statistical analysis. A P value of <0.05 indicated statistical significance.
|
|
|---|
|
View this table: [in a new window] |
TABLE 2. Annual distributions of serotypes of L. monocytogenes strains isolated from infections in human subjects
|
![]() ![]() View larger version (48K): [in a new window] |
FIG. 1. Cases of L. monocytogenes infections by month from 1990 to 2001. The serotypes of the strains (a) and the five most common groups (G1, GT5, G11, G21, and G71) of closely related PFGE types (b) are indicated by colors. The seven mother-child pairs are indicate by numbers (1 to 7).
|
|
View this table: [in a new window] |
TABLE 3. Annual distributions of L. monocytogenes PFGE types
|
![]() View larger version (18K): [in a new window] |
FIG. 2. Dendrogram of closely related PFGE types of the five most common groups (G1, GT5, G11, G21, and G71).
|
Group GT5 (27 strains, of type 5) formed clusters in 1997, 1998, and 2000 (Fig. 1b), and four strains in this group were pregnancy associated.
Strains belonging to group G11 (32 strains) formed clusters in 1990, 1993, 1994, 1995, 1997, and 2001 (Fig. 1b). In 1990, 1991, 1992, and 1993, a total of eight strains were pregnancy associated; four of these cases were caused by PFGE type 68 and four were caused by PFGE type 11. All four cases caused by PFGE type 11 in 1994 occurred in the Helsinki metropolitan area.
Group G21 (29 strains) formed small clusters in 1991, 1992, 1997, 1998, 1999, and 2000 (Fig. 1b). Five strains were pregnancy associated; three of these cases were caused by PFGE types 72, 56, and 69 and two (a mother-child pair) were caused by PFGE type 21.
Group G71 (34 strains) formed clusters in 1997, 1998, and 1999 (Fig. 1b). The clusters were formed by serotype 3a, PFGE type 71. Serotype 1/2a, PFGE type 71, caused only one case, which was pregnancy associated, in 1992.
In group G61 (13 strains), PFGE type 61 (9 strains) formed a cluster of five cases in 1999 (Table 3). Four of these patients and one patient in 1998 were from the greater Tampere area. Also, PFGE type 65 (five strains), which belongs to group GT65, formed a cluster of three cases in 1997 in the Helsinki metropolitan area. However, two of these cases involved a mother-child pair.
The 32 pregnancy-associated cases were caused by 17 different PFGE types. From 1990 to 1993, the pregnancy-associated cases were mostly (8 cases) caused by serotype 4b, PFGE type 68 or 11, each belonging to group G11 (Table 3). After 1993, cases caused by PFGE types in group G11 were not detected. PFGE type 24, which belonged to group GT24, was significantly associated with listeriosis in male subjects (9 of 180 male subjects [5%] versus 0 of 132 female subjects) (P = 0.012). Otherwise, PFGE types were distributed evenly between both genders and among different age groups.
|
|
|---|
In order to gain new information on infection clusters and on the distributions of subtypes of L. monocytogenes, 314 clinical isolates from an 11-year period in Finland were analyzed by PFGE and serotyping.
It should be mentioned that not all strains in Finland during the 4-year period from 1990 to 1993 were available for this study, since before 1994 physicians and clinical laboratories notified KTL and submitted strains to LEP only on a voluntary basis. However, the strains studied represented almost 60% of all 119 listeriosis cases diagnosed in the early 1990s (Jahkola, unpublished). Also, the proportion of pregnancy-associated cases might have been underestimated, since information on pregnancy was not always required in the laboratory notes.
From 1990 through 2001, the most common L. monocytogenes serotypes were 1/2a and 4b, accounting, respectively, for 53 and 27% of the 314 isolates from cases of human listeriosis. Since 1990, the number of cases caused by serotype 4b has been fairly constant, at about 4 to 7 cases per year, except in 1997, when the number of cases was 16, and in 1999, when it was 10. However, the number of listeriosis cases caused by serotype 1/2a has increased; therefore, the percentage share of this serotype has also increased. These results support findings in the United Kingdom (30), Denmark (14), Switzerland (33), and Sweden (23) suggesting that serotype 1/2a is replacing serotype 4b in human infections.
In the present study, when an electronic database library was constructed by using BioNumerics software, any difference between two PFGE types was considered sufficient to distinguish two different PFGE types. Therefore, PFGE types containing at least five strains were compared with all other PFGE types to approximate whether they were closely related to any other type according to the criteria of Tenover et al. (40). These criteria are stringent and generally appropriate for studies of strains collected over a short period. However, in the present study, grouping of closely related types collected over a long period yielded more information on clusters. For example, PFGE type 1 was the most prevalent single type, with 37 strains. It belonged to a major cluster, group G1, representing 23% of all 314 strains studied. Furthermore, 27 strains of PFGE type 5 were not related to any other type and formed several clusters over a period of 3 years.
The most prevalent single type, PFGE type 1, alone was previously associated with a vacuum-packed cold smoked rainbow trout product, and it caused a small infection cluster of febrile gastroenteritis in 1997 (31). In addition, Johansson et al. (20) found isolates with an indistinguishable (L. Rantala, personal communication) PFGE profile in retail ready-to-eat vacuum-packed fish products from four different producers in Finland in 1996. However, in order to draw conclusions from this kind of genotyping data, results should be linked to available epidemiological data. It was shown by Autio et al. (4) that similar strains were found in different types of food from different food processors and even in different countries. Therefore, identical PFGE profiles for isolates from food and a patient do not prove, without epidemiological data, that the food isolate caused the infection.
The patients in the 1997 diarrheal cluster described by Miettinen et al. (31) had no known underlying diseases. PFGE type 1, causing the symptoms, was common in sporadic cases from 1994 onward but not before 1994. This finding could reflect a change in dietary habits, i.e., vacuum-packed fish products being more favored in diets; this PFGE type might be more pathogenic than other PFGE types; or this PFGE type simply might be more common than others in fish products.
In some countries, as in France from 1987 to 1997, the United States from 1989 to 1993, and the United Kingdom from 1983 to 1996, the incidence of L. monocytogenes infections has decreased (3, 16, 39). In Finland, since 1990, the incidence has been quite stable, the average being 20 cases per year, except in 1997 (47 cases), 1998 (43 cases), and 1999 (45 cases). An outbreak affecting 25 people, caused by serotype 3a, PFGE type 71, explains in part the high incidences in 1998 and 1999 (26). The findings of the present study also suggested that the outbreak had already started in 1997. In 1998 and 1999, several cases were also caused by PFGE types 1 and 5 of serotype 1/2a. When the outbreak caused by serotype 3a, PFGE type 71, began in 1997, the numbers of cases caused by various PFGE types of serotype 4b (34% of the total cases) and serotype 1/2a (55%) were also unusually high.
In our study, some L. monocytogenes strains of different serotypes displayed indistinguishable PFGE profiles. The serotypes (1/2a and 3a) of these strains belonged to the same flagellar H antigen group (AB); in addition, they have been reported to belong to the same genetic subgroup (7, 18, 32). In these cases, serotyping and PFGE were repeated for the same colony, and the identical results obtained confirmed the findings. However, only one restriction enzyme, AscI, which is commonly used for L. monocytogenes in PFGE (4, 6, 17), was used; the use of other enzymes and another genotyping method might have been more discriminatory for these strains.
From 1990 to 1993, the pregnancy-associated cases were most commonly caused by serotype 4b, PFGE type 68 or 11, group G11. However, no specific serotype or PFGE type could be concluded to be associated with pregnancy after 1993. Also, no particular PFGE types could be connected with gender or age group. The only exception was PFGE type 24, which was significantly associated with gender: all nine patients were men.
In this study, a new database library of the PFGE profiles of different strains was constructed by using BioNumerics software. In this new electronic library, PFGE profiles can be compared with each other more rapidly; therefore, clusters can be detected more rapidly and at an early stage. This database library also makes possible continuous surveillance of invasive L. monocytogenes infection clusters in Finland. After the library was constructed, a bilateral computer-based network for comparison of PFGE profiles of L. monocytogenes isolates from human subjects, food, and food production environments was created (Rantala et al., Abstr. ISOPOL XIV Int. Symp. Problems Listeriosis) by the National Veterinary and Food Research Institute and the KTL. This kind of national cooperation during a suspected outbreak will probably help in recognizing sources of infections. It will also enable authorities to track down food production plants that need to improve production hygiene for their food products or to give direct information on food hygiene to specific risk groups of consumers. During the 1990s, clusters of cases of the same PFGE type or clusters of groups of closely related types were seen every year, indicating the need for and the importance of timely typing of human L. monocytogenes strains. Furthermore, on a larger scale, an electronic network of PFGE profiles for human L. monocytogenes strains has been in use as PulseNet since 1996 in the United States (38), where it has demonstrated its value in the early recognition of outbreaks and the rapid identification of their sources.
The skillful technical assistance of Jari Aho, Sirkku Ekström, Tarja Heiskanen, Liisa Immonen, Ritva Taipalinen, and Sirkku Waarala is gratefully acknowledged. We also thank Tiina Autio (DVM) for scientific advice.
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»