|Home | About | Journals | Submit | Contact Us | Français|
Between 1998 and 2009, Bordetella pertussis clinical isolates were collected during three periods, i.e., 1998 to 2001 (n = 102), 2004 to 2005 (n = 154), and 2007 to 2009 (n = 140), from nine countries with distinct vaccination programs, i.e., Denmark, Finland, France, Germany, The Netherlands, Norway, Poland, Sweden, and the United Kingdom. Pulsed-field gel electrophoresis (PFGE) analysis was performed according to standardized recommendations for epidemiological typing of B. pertussis. There were 81 different PFGE profiles, five of which (BpSR3, BpSR5, BpSR10, BpSR11, and BpSR12) were observed in 61% of the 396 isolates and shown to be predominant in almost all countries. The major profile, BpSR11, showed a decreasing trend from 25% to 30% in 1998 to 2005 to 13% in 2007 to 2009, and there were increases in BpSR3 and BpSR10 from 0% and 8% to 21% and 22%, respectively. One difference between these profiles is that BpSR11 contains isolates harboring the fim3-2 allele and BpSR3 and BpSR10 contain isolates harboring the fim3-1 allele. The total proportion of the five predominant profiles increased from 44% in 1998 to 2001 to 63% in 2004 to 2005 to 70% in 2007 to 2009. In conclusion, common PFGE profiles were identified in B. pertussis populations circulating in European countries with different vaccination programs and different vaccine coverages. These prevalent isolates contain the novel pertussis toxin promoter ptxP3 allele. However, there is evidence for diversifying selection between ptxP3 strains characterized by distinct PFGE profiles. This work shows that, even within a relatively short time span of 10 years, successful isolates which spread through Europe and cause large shifts in B. pertussis populations may emerge.
Whooping cough (pertussis) is caused by the Gram-negative bacterium Bordetella pertussis. The disease is still a health problem throughout the world even though vaccination against it has been used for 50 years. Despite a long history of vaccination with high coverage in children, a resurgence of pertussis has been observed in many countries, including, Australia, Canada, France, Germany, The Netherlands, Norway, Poland, and the United States (1–10).
In Europe, countries use different vaccination strategies against pertussis (11) (Table 1). Recently, in all European countries (except Poland) whole-cell pertussis vaccines have been replaced by acellular vaccines in which only 1 to 5 protein components (in different formulations) of B. pertussis are included: pertussis toxin (PT), pertactin (PRN), filamentous hemagglutinin (FHA), fimbrial protein type 2 (FIM2), and fimbrial protein type 3 (FIM3) (12). In many European countries, differences in these proteins have been found between B. pertussis vaccine strains and circulating isolates (13–19). A new, possibly more virulent, B. pertussis lineage (designated the ptxP3 lineage) has been recently described (20). The ptxP3 lineage now predominates in several European countries (18, 20–22) and is also found in Australia (23). Moreover, the emergence of B. pertussis isolates not expressing the vaccine antigen PRN was described in France (22, 24). Clinical isolates deficient in PRN were also reported in Italy and Japan (25, 26). For more effective vaccination programs to prevent pertussis, it is important to monitor changes in bacterial populations and to study the impact of these changes on the incidence of disease.
Our previous EUpertstrain I and II studies have shown that the recent epidemics in Europe are associated with clonal expansion of certain B. pertussis strains (27–29). EUpertstrain is the acronym for a project funded by European Commission Fifth Framework Research Program. It has since continued with the aim to characterize European B. pertussis isolates. The EUpertstrain I project was initiated in 2001, and the five participating countries were Finland, France, Germany, The Netherlands, and Sweden. The EUpertstrain II project was supported by GlaxoSmithKline (Rixensart, Belgium) and Sanofi Pasteur MSD (Lyon, France). In addition to the five countries mentioned above, Denmark, Poland, and the United Kingdom joined the project. In the EUpertstrain I and II studies, B. pertussis clinical isolates from eight European countries were analyzed by standardized pulsed-field gel electrophoresis (PFGE) analysis. The isolates were collected during the two periods from 1998 to 2001 and from 2004 to 2005. The 255 isolates produced 59 different PFGE profiles, which were clearly separated from those obtained from vaccine strains. The PFGE profile BpSR11 was found to be predominant, with 30% to 50% in all countries except Denmark (10%) and Poland (0%).
The purpose of this EUpertstrain III study was to investigate the possible effects of different vaccination strategies on the population structure of B. pertussis and to identify changes which have occurred in the bacterial population during the last 10 years in countries where acellular vaccines have been in use. In this study, we used the same method, PFGE analysis, to identify and trace profiles represented in material collected from seven countries participating in the EUpertstrain III study. For comparison, EUpertstrain I and II culture collections were also included.
(The results of this study were presented in part as a poster at the 9th International Bordetella Symposium, Baltimore, MD, 30 September to 3 October 2010.)
The pertussis vaccination programs used in nine countries participating in the EUpertstrain I to III studies are shown in Table 1.
In order to make data comparable, the criteria for selection of isolates in the EUpertstrain III project were the same as those used in EUpertstrain I and II projects. The selection criteria included the following: B. pertussis isolates from different areas and not from local outbreaks, equal number of isolates from partially vaccinated and unvaccinated individuals, and isolates from individuals <5 years of age. The age distribution of the 140 individuals from whom B. pertussis was isolated in the period from 2007 to 2009 was as follows: 33 were ≤1 month, 20 were 1 to 2 months, 8 were 2 to 3 months, 22 were 4 to 6 months, 11 were 7 to 12 months, 22 were 1 to 5 years, 14 were 6 to 18 years, and 10 were older than 18 years. Of the 24 individuals older than 5 years, 4 were from Denmark, 3 were from Finland, 4 were from France, 7 were from Norway, and 6 were from the United Kingdom.
In the EUpertstrain I project, a total of 102 isolates were collected from children in five countries, i.e., Finland, France, Germany, The Netherlands, and Sweden, in the period from 1998 to 2001. In the EUpertstrain II project, a total of 154 isolates were collected from children in the five countries participating in the EUpertstrain I project plus Denmark, Poland, and the United Kingdom in the period from 2004 to 2005. In the EUpertstrain III project, 140 isolates were collected in seven countries, i.e., Finland, France, The Netherlands, Sweden, Denmark, United Kingdom, and Norway, in the period from 2007 to 2009. The numbers of isolates collected among the nine countries during the three study periods are shown in Table 2.
All isolates collected during the three study periods were analyzed by pulsed-field gel electrophoresis (PFGE) at the Swedish Institute for Communicable Disease Control (SMI) according to the standardized recommendations for typing of B. pertussis (30), with the modifications described in reference 31. The PFGE analyses of all isolates tested were performed by the same person (A.A.) at SMI. The PFGE method has been proven to be stable and has high resolution. The profiles were analyzed by using BioNumerics software version 4.61 (Applied Maths, NV, Belgium). Initially major PFGE groups were determined as described by Khattak and Matthews (32), i.e., PFGE groups I, II, III, and IV (19, 22, 27). Later, PFGE groups were defined as distinct DNA band patterns if they differed by at least one band and were designated BpSR1, BpSR2, BpSR3, etc., for those isolates first detected in Sweden (31). Isolates first identified in a country other than Sweden, such as Finland, were designated BpFINR1, BpFINR2, etc. A cluster analysis was performed using BioNumerics software with the unweighted-pair group method using the arithmetic average algorithm with 2% band tolerance and 1.5% optimization settings. For the purpose of comparability, the same band tolerance and optimization settings used in the previous EUpertstrain study (28) were used. Strains 18323 (PFGE cluster I), Tohama I (PFGE cluster II), Bp134 (PFGE cluster III), B902 (PFGE cluster IVα), FR743 (PFGE cluster IVβ), FIN12 (PFGE cluster IVγ), and FR287 (PFGE cluster V) were included in the dendrogram as reference strains (19, 27).
Fisher's exact test or the chi-square test was used for statistical comparison of the frequencies of five PFGE profiles between different age groups. P values (two sided) of <0.05 were considered significant.
Altogether, 396 isolates collected in the three study periods produced 81 distinct PFGE profiles, and most of them are included in cluster IV (Fig. 1). In the 102 isolates belonging to the collection from the period from 1998 to 2001, 33 profiles were identified. In the 154 isolates from 2004 to 2005, 36 profiles could be found. Among the 140 isolates collected from 2007 to 2009, 29 profiles were detected. Seventeen profiles were common in the three study periods.
Of the 81 distinct profiles, 5 (BpSR11, BpSR10, BpSR3, BpSR5, and BpSR12) were found in 21 to 91 isolates, representing 61% of the total isolates studied (Table 2). Thirty-two profiles were found in 2 to 9 isolates, representing 28% of the isolates studied, and 44 profiles were found only in 1 isolate, representing 11% of the total isolates studied. Although unique profiles were detected in all countries except Denmark and Norway, most of the unique profiles were found only in 1 or 2 isolates.
The five common profiles BpSR11, BpSR10, BpSR3, BpSR5, and BpSR12 belonged to clusters IVβ, IVα, IVα, IVβ, and IVγ, respectively, in the previous EUpertstrain II study (28). However, with the use of the new version of BioNumerics software in the present EUpertstrain III study, the profiles BpSR3 and BpSR5 were found in a different place but did belong to the cluster IV (Fig. 1).
Table 2 shows the 5 PFGE profiles representing 15% to 90% of the isolates from each country with the exception of Poland. The broad distribution of these profiles suggests that these isolates spread rapidly through Europe.
The most common profile, BpSR11, belonging to cluster IVβ, was observed in the whole study period, representing 23% of the total isolates. This profile was found in proportions of 5% to 45% in the period from 1998 to 2001, 10% to 50% in the collection of the period from 2004 to 2005 (Poland excluded), and 0% to 35% in the collection of the period from 2007 to 2009. It has been shown that isolates with BpSR11 (cluster IVβ) were characterized by the following combination of alleles: ptxP3 prn2 ptxC2 tcfA2 fim3-2 (21).
The second most common PFGE profile, BpSR10, belonging to cluster IVα, was also observed in the whole study period, representing 13% of the total isolates (Table 2). This profile was found in proportions of 0% to 25% in the period from 1998 to 2001, 5% to 20% in 2004 to 2005 (Poland excluded), and 12% to 35% in 2007 to 2009 period. The isolates with BpSR10 harbored an allele combination of ptxP3 prn2 ptxC2 tcfA2 fim3-1.
The third most common PFGE profile, BpSR3, falling into cluster IV, was found only in the period from 2004 to 2009, representing 11% of the total isolates studied. This profile was seen in proportions of 0% to 25% in 2004 to 2005 (Poland excluded) and 5% to 50% in 2007 to 2009. The isolates with BpSR3 harbored an allele combination of ptxP3 prn2 ptxC2 tcfA2 fim3-1.
The fourth most common PFGE profile, BpSR5, also falling into cluster IV, was found in the whole study period, representing 8% of the total isolates studied. This profile was seen in proportions of 4% to 15% in the period from 1998 to 2001, 0% to 22% in 2004 to 2005 (Poland excluded), and 0% to 35% in 2007 to 2009 period. The isolates with BpSR5 harbored an allele combination of ptxP3 prn2 ptxC2 tcfA2 fim3-2.
The fifth most common PFGE profile, BpSR12, belonging to cluster IVγ, was also found in the whole study period, representing 5% of the total isolates tested. This profile was seen in proportions of 0% to 10% in the period from 1998 to 2001, 0% to 25% in 2004 to 2005 (Poland excluded), and 0% to 10% in 2007 to 2009. The isolates with BpSR12 harbored an allele combination of ptxP3 prn2 ptxC2 tcfA2 fim3-2.
The distribution of the 5 common PFGE profiles in the three study periods is shown in Fig. 2. The major profile, BpSR11, showed a decreasing trend from 25% in the period from 1998 to 2001 and 30% in the period from 2004 and 2005 to 13% in the period from 2007 to 2009. There was an increase in BpSR3 from 0% in the period from 1998 to 2001 and 8% in the period from 2004 to 2005 to 22% in the period from 2007 to 2009 and in BpSR10 from 8% in the period from 1998 to 2001 period and 10% in the period from 2004 to 2005 to 21% in the period from 2007 to 2009. The proportions of other profiles also showed a decreasing trend, from 56% in the period from 1998 to 2001 and 38% in the period from 2004 to 2005 to 30% in the period from 2007 to 2009.
For the minor profiles, BpSR173 was found in France, Germany, The Netherlands, and Sweden (at frequencies of 2%, 9%, 2%, and 2%, respectively), and BpSR19 was found in Finland, Sweden, and The Netherlands (at frequencies of 2%, 3%, and 10%, respectively). The profile BpSR257 was found only in Norway and Finland (at frequencies of 5% and 14%, respectively), BpSR7 was found only in Finland and Sweden (at frequencies of 5% and 5%, respectively), BpSR13 was found only in Denmark and Sweden (at frequencies of 5% and 8%, respectively), and BpSR243 was found only in Germany and Denmark (at frequencies of 3% and 9%, respectively).
Isolates from Finland, France, The Netherlands, and Sweden were available from all three periods. In addition to the change in proportions of the 5 major profiles, a shift in the minor or unique PFGE profiles was also observed with time. Of the 7 minor profiles detected among Finnish isolates in the period from 1998 to 2001, one was found in the period from 2004 to 2005 and none was found in 2007 to 2009 period. Of the 6 minor profiles detected among French isolates in the period from 1998 to 2001, none was detected in the two later periods.
Among the 13 Polish isolates, there were 7 distinct PFGE profiles identified from isolates collected in the period from 2004 to 2005. None of the isolates belonged to the five common profiles. Two profiles, together representing four isolates, were unique. The remaining 5 were minor profiles, and some of them had been recovered in Sweden and Germany as well.
Table 3 shows proportions of the five predominant PFGE profiles in three different age groups of subjects included in the study period from 2007 to 2009. No difference was noticed in frequencies of the five PFGE profiles between infants ≤2 months of age (unvaccinated) and those in older age groups. Furthermore, we did not observe a relationship between the vaccination programs (Table 1) used and the distribution of PFGE profiles among the participating countries. Furthermore, changes from the whole-cell vaccine to the acellular vaccine did not seem to have a direct influence on the distribution of profiles. It should be noted, however, that none of the isolates from Poland, where the whole-cell vaccine is still in use, belonged to the 5 most common profiles found in countries where the acellular vaccine was used, although there were too few isolates for final conclusions to be made.
In this study, we analyzed and compared B. pertussis isolates collected from nine European countries with different vaccination programs during a 10-year period. All isolates were tested by the same method at the same institution. This type of analysis allows the identification of highly fit isolates which have increased capacity to spread in an immunized population.
Indeed, we identified common PFGE profiles dominant in Europe. BpSR11, BpSR10, BpSR3, BpSR5, and BpSR12 represent examples of the dynamic change observed in B. pertussis populations. The B. pertussis strain BpSR11 was isolated first in France around 1996 (19), in Sweden in 1997 (16), and later in Finland in 1999 (13). Of the 20 Dutch isolates collected in the period from 1998 to 2001, 6 were found to be BpSR11. Isolates sharing the profile increased in frequency, and it has since become the most common profile in the three countries. During the 10-year study period, frequencies of BpSR11 increased from 26% in 1998 to 2001 to 30% in 2004 to 2005 and then decreased. In 2007 to 2009, only 13% of isolates tested were BpSR11. In contrast to the profile BpSR11, BpSR3 was not detected in any of isolates included in the period from 1999 to 2001. In the period from 2004 to 2005, BpSR3 increased to 8%. However, in the period from 2007 to 2008, BpSR3 became the most common profiles, representing 22% of isolates tested. Furthermore, data on the change in unique profiles observed in each country also support that the B. pertussis population is dynamic.
With respect to the 5 antigens potentially included in acellular pertussis vaccines, BpSR11 and BpSR3 or BpSR10 differ only in the fim3 gene. Thus far, four fim3 alleles have been identified (33–35). With the exception of fim3-4, which carries a silent mutation, all alleles have single-nucleotide polymorphisms (SNPs) that alter the amino acid sequences (35). The difference between the products of the fim3-1 and fim3-2 alleles is a replacement of alanine by glutamic acid (35). Moreover, all strains used for production of whole-cell or acellular vaccines analyzed harbor the fim3-1 allele (33, 34). In The Netherlands, when 704 B. pertussis isolates collected from 1949 to 2010 were studied, only fim3-1 and fim3-2 were detected (36). Before 1996, all isolates had fim3-1. The allele fim3-2 was first detected in 1996, increased in frequency to 62% in 2002 and then gradually decreased in frequency. Since 2009, all isolates tested have been fim3-1. In contrast to the Dutch data mentioned above, the allele fim3-2 was rarely detected in Poland when isolates collected in 1995 to 2005 were genotyped (A. Lutyńska et al., unpublished data). Recently, the emergence of fim3-2 allele in the United States has been reported (37). Of 661 B. pertussis isolates from 1935 to 2009, all except one harbored fim3-1 or fim3-2. The allele fim3-2 was first detected in 1994, and since then its frequency has steadily increased. During 2006 to 2009, about 82% of isolates tested carried fim3-2. Since humans are the only host of B. pertussis, the selective pressure for the change in fim3 allele frequency is most likely due to human population immunity. It remains to be shown what drives the change, i.e., natural infection, vaccine, or both.
Recently, the emergence of isolates with a novel allele (ptxP3) for the B. pertussis toxin promoter was described (20). The ptxP3 isolates were found to produce more pertussis toxin (20) and have spread worldwide (18, 20, 21, 23). When 138 isolates with the major PFGE profiles (BpSR11-cluster IVβ, BpSR10-cluster IVα, BpSR12-cluster IVγ, BpSR5-cluster IV, or BpSR3-cluster IV) identified in the present study were analyzed, they all contained the ptxP3 allele (data not shown), suggesting that these isolates have increased fitness and capacity to spread among immunized populations. As yet it is not clear if and to what extent the ptxP3 allele contributed to the increased fitness of ptxP3 strains. For example, the ptxP3 allele may be linked to other loci which affect strain properties. Further research is required to resolve this issue.
In France, where the surveillance of clinical isolates has been performed since 1990 and where acellular vaccines were introduced in 1998, B. pertussis isolates which do not express PT, FHA, or PRN have been isolated (22). These isolates were detected sporadically even during the prevaccine era. However, in France, 7 years after the introduction of acellular pertussis vaccines, the prevalence of PRN deficient isolates had increased in frequency (22, 24), and patients infected by PRN-deficient isolates were found to have typical pertussis symptoms. The increased prevalence of PRN-deficient isolates has been also reported in Japan, where the acellular vaccines were introduced in the early 1980s (25). When isolates belonging to the study period from 2007 to 2009 were analyzed, PRN-deficient isolates were detected among those isolates from France, Norway, and Sweden. All the PRN-deficient isolates were associated with the profile BpSR11, BpSR10, or BpSR3 (reference 22 and data not shown).
In this study, the vaccination status was available for only some of the subjects, and the number of clinical isolates included in each study period was limited. Therefore, it is not possible to directly compare B. pertussis isolated from (partially) vaccinated individuals with that from unvaccinated individuals. However, according to the vaccination programs used in each country, infants younger than 2 months of age can be considered unvaccinated. It seems that there were no significant differences between the isolates from infants too young to be vaccinated and those from older children and adults. Further comparative studies based on a larger number of isolates are needed.
Representative samples are important for the interpretation of results and conclusions. In this study, most of the countries have followed the selection criteria described. It is possible for some countries that due to the limited number of strains collected, the isolates selected might not be representative. However, in countries such as Finland, France, The Netherlands, and Sweden, where the serial collection of isolates was done and tested, the frequencies in the major profiles observed were similar to those found in the present study. Moreover, a recent Polish study reported that there were 59 PFGE profiles identified among 110 isolates from 1995 to 2005 (38). Most of these profiles were different from those observed in this study, confirming that the trends/proportions observed in this study were representative.
In conclusion, common PFGE profiles were identified in B. pertussis populations circulating in European countries with different vaccination programs, and these prevalent profiles contain the novel ptxP3 allele. Further, there is evidence for diversification within ptxP3 isolates characterized by distinct PFGE profiles but corresponding to the same PFGE group IV. This work supports the view that the B. pertussis population is adapting even within the relatively short time span covered by this investigation. Further analyses may elucidate how B. pertussis is able to resurge in vaccinated populations according to the vaccine type and the vaccine coverage.
This study was supported by GlaxoSmithKline (Rixensart, Belgium) and Sanofi Pasteur MSD (Lyon, France).
Published ahead of print 21 November 2012