Identification of Persistent Subtypes of Campylobacter jejuni by Pulsed-Field Gel Electrophoresis in Finland

Introduction

C ampylobacter species are the most common bacterial gastrointestinal pathogens in humans in the European Union and other developed countries. In Finland, 3500–4000 Campylobacter cases are reported annually to the National Infectious Disease Register (www3.ktl.fi). We have previously investigated the Finnish Campylobacter strains in a population-based study on serotype level (Nakari et al., 2010), but a more discriminatory typing method is needed for identification of common persisting national subtypes of Campylobacter jejuni. Pulsed-field gel electrophoresis (PFGE) is one of the most discriminatory subtyping methods that can be applied to Campylobacter isolates (Nielsen et al., 2000). This study describes the genetic diversity of C. jejuni strains isolated from Finnish patients and compares strains isolated from domestically acquired and foreign-travel-related infections. The longitudinal nature of the study enabled the investigation of temporal diversity of the PFGE types and the identification of potentially persistent C. jejuni genotypes, especially among strains causing domestic infections.

Materials and Methods

The strains (n=508) originated from a collection of 1407 (622 domestic and 785 travel-related) previously serotyped (Nakari et al., 2010) C. jejuni strains. They were collected during two 1-year periods (July 2002–June 2003 and July 2004–June 2005) in nine clinical microbiology laboratories across Finland. The strains were regarded as sporadic, as no outbreaks were notified by the hospital districts. A strain was defined as domestic if the patient had no history of foreign travel within 10 days before the onset of symptoms or 17 days before the specimen was taken. A large representative set of the strains was systematically selected for PFGE typing based on serotype, hospital district, sampling date, and travel destination. In all, 508 C. jejuni strains, 311 domestic and 197 travel-related, were typed by PFGE using SmaI. Additionally, strains representing the most common SmaI profiles and strains that were not digested by SmaI were typed using KpnI. PFGE was performed as previously described, and the PFGE profiles were compared with the previously identified profiles (Hakkinen et al., 2009) and named accordingly.

Results and Discussion

The strains were divided into 240 SmaI subtypes: 311 domestic strains into 86 and 197 travel-related strains into 172. Of these, 18 profiles were shared between the two groups. The most common SmaI subtypes were S7 (14%), S5 (6%), S64 (5%), and S54 (4%) (Table 1). These subtypes accounted for 45% (139/311) of the domestic strains and 3% (5/197) of the travel-related strains. Among the domestic strains, 15 SmaI subtypes were detected during both sampling periods, accounting for 69% (n=215) of the strains. Of the travel-related isolates, 18% (n=36) belonged to 17 SmaI subtypes that were detected during both sampling periods.

Subtyping by KpnI was applied to the strains belonging to the most common SmaI subtypes (n=144) and those that were not digested by SmaI (n=38). The SmaI subtypes S7, S64, and S54 were divided into several KpnI subtypes, whereas all 28 S5 strains had an identical KpnI profile K27 (Table 1). Of the 38 strains not digested by SmaI, 21 domestic strains had an identical KpnI profile K36. Seven SmaI/KpnI combination types with more than five strains were identified (Table 1). Of these common subtypes, S7/K1, S7/K5, and S5/K27 were found in one or two travel-related strains, whereas all strains of S7/K2, S7/K47, S64/K19, and S54/K10 were of domestic origin.

The strains were typed retrospectively. When the strains were typed, it was seen that the majority of S5/K27 had been from the second sampling period (Table 1), and a cluster of 11 S5/K27 strains had occurred in Western Finland in July 2004. All S7/K47 strains were from the first sampling period (Table 1), isolated between July and September 2002. Of the seven S7/K47 strains, six were isolated in Western Finland. Subtypes S7/K1, S7/K5, S7/K2, S64/K19, and S54/K10 were equally distributed between the two sampling periods among the domestic strains (Table 1). Subtypes S7/K1, S7/K5, S7/K2, S5/K27, S64/K19, and S54/K10 were considered persistent subtypes and S7/K47, a transient subtype.

The reliability of PFGE typing has been questioned due to alterations in the PFGE patterns caused by temperate bacteriophages and exchange of DNA between strains (de Boer et al., 2002; Barton et al., 2007). On the other hand, recurrent genotypes in a defined area have been previously reported in Finland (Hänninen et al., 2000), and long-term persistence of PFGE subtypes has been reported in other countries. Indistinguishable SmaI subtypes were found in human isolates over a 12-year period in Japan (Yabe et al., 2010), and five of the six predominant SmaI/KpnI combination types in Sweden in 2000 were still predominant in 2003 (Lindmark et al., 2009).

We have previously reported that Pen 2 is common in Finland in both domestically acquired infections and travel-related infections (Nakari et al., 2010). The diversity of the PFGE subtypes among travel-related Pen 2 strains was much higher than among Pen 2 isolates of domestic origin. PFGE pattern S5/K27 accounted for 65% of the 40 domestic Pen 2 strains but was found in only 2 of the 28 travel-related strains. This suggests that the sources for the travel-related Pen 2 infections are diverse, whereas the domestic strains could be related to a specific source. Pen 2 was the most common serotype in cattle in a Finnish slaughterhouse survey (Hakkinen et al., 2007) but was less commonly detected in domestic broilers (Perko-Mäkelä et al., 2002). Also, S5/K27 was found in cattle samples in 2003, together with S7/K1, S7/K2, S7/K5, and S64/K19 (Hakkinen et al., 2009). Two previously reported closely related SmaI subtypes, S7 and S74 (Hakkinen et al., 2009), were combined into one, S7. These two subtypes differed only in the size of one fragment and could no longer be separated when a larger number of strains were analyzed. Cattle are potential sources for human infections caused by the nationally persistent common subtypes, as long-term excretion of the same C. jejuni genotypes by cattle has been reported (Kwan et al., 2008; Hakkinen and Hänninen, 2009).

Of the six common persistent subtypes, S7/K1, S7/K5, S7/K2, S64/K19, and S54/K10 were identified in chicken samples in 2003, but only S7/K1 and S64/K19 were temporally associated with human infections (Hakkinen et al., 2009). In general, the strains isolated from broilers seem to vary from year to year, as shown by multi-locus sequence typing data (de Haan et al., 2010). Although chickens and cattle appear to be sources for part of the domestic human infections caused by the persistent genotypes, it is evident that other sources also exist. K36, the KpnI subtype accounting for 55% of the human strains not digestable by SmaI, was identified in a strain isolated from mink in 2009 (unpublished results from M. Hakkinen). Other potential sources should be investigated to find additional reservoirs for these genotypes, as the infections in humans, chickens, and cattle could also be related to a common environmental source. The overlap of C. jejuni genotypes between humans and environmental sources is much lower compared with the overlap between human strains and food-production animals (Petersen et al., 2001; Lindmark et al., 2004). Thus, obtaining a representative set of environmental samples would require extensive sampling.

In conclusion, six common persistent SmaI/KpnI C. jejuni genotypes were identified among strains isolated from patients on a nationwide scale. These genotypes were only occasionally identified among travel-related strains, suggesting that they are true national subtypes. Five of them have been previously identified in cattle, and two in chickens with a temporal association with human infections. Cattle may be a more likely potential source of these persistent subtypes, as long-term excretion of Campylobacter strains by cattle has been reported, whereas isolates from chickens show more temporal variation.