Increase in Genetic Diversity of Haemophilus influenzae Serotype b (Hib) Strains after Introduction of Hib Vaccination in The Netherlands
http://www.100md.com
微生物临床杂志 2005年第6期
Laboratory for Vaccine-Preventable Diseases, National Institute of Public Health and the Environment, Bilthoven, The Netherlands
Netherlands Reference Laboratory for Bacterial Meningitis, Department of Medical Microbiology, Academic Medical Centre, Amsterdam, The Netherlands
Applied Maths, Sint-Martens-Latem, Belgium
ABSTRACT
Recently, there has been an increase in The Netherlands in the number of cases of invasive disease caused by Haemophilus influenzae serotype b (Hib). To study a possible change in the Hib population that could explain the rise in incidence, a multiple-locus variable number tandem repeats analysis (MLVA) was developed to genotype H. influenzae isolates. The MLVA enabled the differentiation of H. influenzae serotype b strains with higher discriminatory power than multilocus sequence typing (MLST). MLVA profiles of noncapsulated H. influenzae and H. influenzae serotype f strains were more heterogeneous than serotype b strains and were distinct from Hib, although some overlap occurred. The MLVA was used to genotype a collection of 520 H. influenzae serotype b strains isolated from patients in The Netherlands with invasive disease. The strains were collected from 1983 from 2002, covering a time period of 10 years before and 9 years after the introduction of the Hib vaccine in the Dutch national vaccination program. MLVA revealed a sharp increase in genetic diversity of Hib strains isolated from neonates to 4-year-old patients after 1993, when the Hib vaccine was introduced. Hib strains isolated from patients older than 4 years in age were genetically diverse, and no significant change in diversity was seen after the introduction of the vaccine. These observations suggest that after the introduction of the Hib vaccine young children no longer constitute the reservoir for Hib and that they are infected by adults carrying genetically diverse Hib strains.
INTRODUCTION
Haemophilus influenzae is a respiratory pathogen that causes a wide spectrum of infections in humans (22). H. influenzae carrying a polysaccharide capsule have been an important cause of serious invasive diseases such as meningitis and septicemia in children. The majority of cases of invasive disease are caused by H. influenzae serotype b (Hib), and this has led to the introduction of nationwide vaccination with Hib polysaccharide conjugate vaccines in many developed countries. Vaccination has resulted in impressive reductions in Hib disease and reduced carriage of Hib among both vaccinated and nonvaccinated individuals. In The Netherlands, where vaccination was introduced in 1993, the incidence of invasive Hib disease among children younger than 5 years of age dropped from 28.7 per 100,000 in 1992 to 0.8 in 2002 (32). Similar reductions of Hib disease have been observed in most other countries where vaccination has been implemented in the national vaccination programs and has almost led to eradication of invasive Hib disease among children.
Despite nearly complete vaccine coverage, a small number of fully vaccinated children in The Netherlands have experienced invasive Hib infection. In 2002 the number of cases of vaccine failures in Dutch children aged 0 to 4 years increased by a factor of 3 compared to the years 1996 to 2001 (24). Although the number of cases of invasive Hib disease was quite low, with only 16 cases among 0 to 4 year olds, 14 of these (81%) occurred in fully vaccinated children. A more impressive rise has been observed in the United Kingdom, where the incidence among children in the age group of 0 to 4 years started to rise in 1999 to reach 3.7 per 100,000 in 2002 (23). The rise in disease in the United Kingdom has been associated with several factors. First, relatively fast waning immunity in the age group 1 to 4 years has resulted in lower vaccine-induced antibody levels (8, 23, 29). Second, the use of combination vaccines containing acellular pertussis vaccine has been associated with reduced immunogenicity of the Hib component (13). To compensate for the reduced anti-Hib titers, a catch-up campaign designed to boost immunity in children aged 6 months to 4 years of age was implemented in the United Kingdom (30). In contrast to the United Kingdom, the vaccination schedule in The Netherlands contains a booster vaccination at 11 months of age. Furthermore, only whole-cell pertussis vaccine has been used in combination with the Hib vaccine. This suggests that waning immunity may not be the cause of the rise of vaccine failures in The Netherlands. Also, the total number of cases of invasive Hib disease among all age groups has increased significantly and is back to the level seen in 1996, suggesting that there has been an increase in the circulation of Hib in The Netherlands, too (32).
The various reports that discuss the increase of invasive Hib in the United Kingdom focus on immune responses and do not mention a change in the properties of the bacteria as a possible cause of the increased number of cases. The reason for ignoring this possibility probably lies in the fact that the bacterial component used for vaccination, the capsular polysaccharide, is a very simple component that is unlikely to change. Alterations in the Hib polysaccharide would almost certainly result in the loss of the reactivity of the bacteria with the sera used for capsular typing. However, subtle changes in capsule structure or expression could go unnoticed if only serotyping is performed. Furthermore, changes may not be restricted to the capsule but could be related to other virulence factors. This may result in an altered composition of the circulating Hib population in which particular clones are successful in evading the barrier of the vaccine induced immunity. We describe here the genotypic characterization by MLVA and MLST of Dutch Hib strains collected in the period from 1983 to 2002, comprising a 10-year period before the introduction of the Hib vaccine in 1993 and 9 years after nationwide Hib vaccination of newborns.
MATERIALS AND METHODS
Bacterial strains. A total of 574 H. influenzae strains were included in the present study. Of this collection, 536 strains were serotype b, 7 were serotype f, and 31 were nontypeable. All strains were collected during the years from 1983 to 2002 by The Netherlands Reference Laboratory for Bacterial Meningitis from patients with invasive disease in The Netherlands. Of the 520 Hib strains, 349 were isolated from cerebrospinal fluid or from cerebrospinal fluid and blood, 161 were from blood only, and 12 were from other materials such as synovial fluid. From each year ca. 20 Hib strains isolated from children in the age group 0 to 4 year and 10 strains from patients older than 4 years, mainly adults, were included. The selected strains contained all available Hib strains isolated from 1996 to 2002 and randomly selected strains from the time period before 1996.
Bacterial growth and preparation of lysates. Cultures stored at –80°C were streaked onto tryptic soy chocolate agar plates, cultured overnight at 35°C in an atmosphere of 95% air-5% CO2, and visually inspected for purity. In the few cases where cultures were contaminated, a single colony was restreaked to obtain a pure culture. A loopful of cells from the pure culture was suspended in 500 μl of TE (10 mM Tris-HCl, 1 mM EDTA [pH 8]) and heated for 10 min at 95°C to lyse the cells, and then the lysate was stored at –20°C until use in PCR.
MLST. For multilocus sequence typing (MLST), the typing scheme proposed by Meats et al. (17) was used to analyze the H. influenzae strains in the present study. During the setup of the MLST in our lab the published PCR conditions proved unsuitable when using lysates as a source of DNA for MLST. Therefore, a number of new primers were designed (Table 1), and the PCR conditions were adapted to facilitate a multiplex format to amplify all seven gene fragments in a single reaction. Optimized multiplex PCRs were performed in 15-μl volumes in 96-well PP-PCR plates (Greiner Bio-One, Frickenhausen, Germany) in Applied Biosystems 9700 PCR machines (Applied Biosystems, Foster City, Calif.). A total of 5 μl of 1:10 diluted heat-treated H. influenzae lysate was added to a 10-μl mixture containing 7.5 pmol of each primer and 7.5 μl of HotStar Taq mastermix (Qiagen, Hilden, Germany). The PCR program used was 15 min at 95°C, followed by 30 cycles of amplification that consisted of 30 s at 95°C, 30 s at 50°C, and 1 min at 72°C, and a final step of 10 min at 72°C. PCR products were purified by adding 6 μl of ExoSAP-IT (USB Corp., Cleveland, Ohio), followed by subsequent incubations for 15 min at 37°C and 15 min at 80°C. All MLST genes were sequenced from the multiplex PCR mixture in 14 separate sequence reactions by using 1 μl of purified PCR product and 5 pmol of PCR primer per 20-μl sequence reaction. Sequence reactions were performed with the ABI Prism BigDye Terminator cycle sequencing kit (v3.1; Applied Biosystems, Foster City, Calif.) and analyzed on an ABI 3700 DNA sequencer. If a sequence reaction failed to yield an unambiguous MLST DNA sequence, PCR and sequencing of this particular gene were repeated. In that case, PCR was performed in a 10-μl volume with 0.5 μl of lysate and 5 pmol of each of the two primers and 5 μl of HotStar Taq mastermix using the above-mentioned PCR cycling protocol. Subsequently, sequence reactions were performed on 1 μl of the unpurified PCR product according to the above-described protocol. The sequence trace files were imported into a Bionumerics 4 database (Applied Maths, Sint-Martens-Latem, Belgium) and aligned and trimmed by using a script. After editing the MLST sequences, another set of scripts was used to assign allele numbers and sequence types (STs) using the data from the H. influenzae MLST scheme available on the MLST website (http://www.mlst.net/). Allele sequences and MLST profiles that were not present in the database were assigned provisional allele numbers starting at 800 and provisional STs starting at 500.
Identification of VNTR loci and MLVA design. MLVA is increasingly used as a typing method for bacterial species (5-7, 10, 11, 25). For H. influenzae VNTR analysis has been described by van Belkum et al. (31). Therefore, we initially repeated the VNTR PCRs on the 18 VNTR loci described in van Belkum's study and included 10 additional tandem repeat loci. Analysis showed that only two loci previously described by Van Belkum et al. (31) VNTR5-2 and VNTR6-1 and two newly identified 12-bp tandem repeat loci VNTR12-1 and VNTR12-2 yielded limited, but sufficient, variation and were used to compose the MLVA profile. The majority of the other loci were either too variable in composition or not present in all strains from the test panel. To study the in vitro stability of the MLVA loci, we subcultured a Hib strain, a noncapsulated H. influenzae strain, and an H. influenzae strain Rd through daily serial passage for a month. The composition of the VNTR loci was determined, and no alterations were observed, indicating that the number of repeats is relatively stable. However, this does not exclude rapid changes if strains were passaged through human hosts.
With the exception of VNTR5-2, the VNTR loci of the MLVA set were located in open reading frames in the H. influenzae Rd strain (Table 2). VNTR5-2 is located upstream from the gene encoding for ribosome-binding factor (rbfA) which is essential for efficient processing of the 16S rRNA. Variation in the number of repeats may alter the level of expression of rfbA. VNTR6-1 is located within tonB, a gene required to energize outer membrane transport reaction. VNTR12-1 is located in a "predicted coding region" with unknown function, and VNTR12-2 is found within the hsdS gene which is involved in the type I restriction modification system of the bacterial cell.
MLVA. Primers were designed for two newly identified loci: VNTR12-1 and VNTR12-2. Furthermore, primers for the earlier described VNTR5-2 and VNTR6-1 loci (31) were redesigned to yield robust PCRs. The MLVA primers used to type the H. influenzae strains from this collection are listed in Table 1. VNTR PCRs were performed in 20-μl volumes in Applied Biosystems 9700 PCR machines. We added 1 μl of 1:10-diluted heat-treated H. influenzae lysate to a mixture containing 8 pmol of 5' 6-carboxyfluorescein-labeled (FAM) forward primer, 8 pmol of unlabeled reverse primer, and 10 μl of HotStar Taq mastermix. VNTR loci were amplified in separate PCRs by using the following program: 15 min at 95°C, followed by 17 cycles of amplification that consisted of 30 s at 95°C, 1 min at 54°C, and 1 min at 72°C, with a final step of 30 min at 68°C to ensure complete terminal transferase activity of the Taq DNA polymerase. After PCR, samples were diluted 1:200 in water, and 2 μl of the diluted samples was mixed with 10 μl of GeneScan 500 ROX-Marker (Applied Biosystems). After heat denaturation for 5 min at 95°C, fragments were separated on an ABI 3700 DNA sequencer by using the standard GeneScan module with filterset D. The GeneScan data were analyzed in the GenoTyper software (Applied Biosystems) to perform sizing and to calculate the number of repeats in the PCR fragments. The data with calculated number of repeats were imported into the Bionumerics software package for further cluster analysis. The data obtained with the number of repeats found at the four different VNTR loci in the order VNTR5-2, VNTR6-1, VNTR12-1, and VNTR12-2 were used to make up the MLVA profile. The accuracy of the sizing and the translation of the fragment sizes into repeat numbers was confirmed by sequencing a number of VNTR PCR fragments.
Data analysis. The MLST and MLVA profiles were clustered in the Bionumerics software by using a categorical coefficient and a graphing method called a minimum spanning tree as described before (25). In the minimum spanning tree the priority rules to first link types that have the highest number of single locus variants was chosen. For MLST a maximum neighbor difference of 2 was used to create complexes.
For calculation of the genetic diversity and discrimination index the Simpson's index of diversity was used as follows:
where ni is the number of strains belonging to ith type and N is the total number of strains in the sample population (9, 26).
In the present study, we compared diversity indices from various samples, inferring conclusions that the diversity of one sample is larger than of the other. Because these indices are based on limited sample sizes, they are subject to statistical uncertainty. Hence, in order to be able to validate the observed differences, it was necessary to calculate their statistical significance as measured by P values. We used Monte-Carlo simulations to obtain these values (18). It is well known that, for an observed sample with size N, the class sizes for all types ni (i = 1 ... p) follow a multinomial distribution function, characterized by the class probabilities i. These probabilities i are unknown but can be estimated from the sample by calculating i as ni/N. During the Monte-Carlo simulation process, a large number (10,000) of randomly generated synthetic data sets with sample size N is generated, all having the same multinomial distribution of strains over the different types as dictated by the probabilities i. In order to obtain a P value for the difference in diversity indices between two samples, this process is performed for both samples, and the fraction of simulations that violate the hypothesis about the difference in diversity yields the P value.
RESULTS
MLST of Hib strains. MLST can be considered the gold standard for molecular typing of a number of bacterial species (2-4, 12). Recently, an MLST typing scheme for H. influenzae has been published (17), and we have adapted the MLST protocol by redesigning the primers and creating a multiplex PCR in which all seven MLST fragments were amplified in a single PCR. Use of the multiplex approach saves time and cost. In addition, the redesigned primers enabled the use of crude H. influenzae lysates for PCR. The multiplex PCR seemed to be robust, amplifying sufficient PCR product of all MLST loci with the exception of the atpG locus, which often was insufficiently amplified. For future use of the multiplex MLST PCR, a new primer set will have to be designed for amplification of the atpG locus. The collection of strains used for the MLST was composed of 241 Hib strains. A total of 61 strains were from the prevaccination era and dated from 1989 to 1993, and the remaining 180 strains were isolated in the years from 1994 to 2002, after the introduction of the Hib vaccine. Of the strains from the prevaccination era only 2 were from patients older than 4 years, and 59 of the strains from the postvaccination group were from patients older than 4 years. There were 29 different STs among the 241 strains (Fig. 1A). Of the 29 STs, 17 represented previously unidentified allele combinations (http://haemophilus.mlst.net/) and were therefore designated provisional STs. The majority of the strains had ST6 (76.3%) and, with the exception of nine strains, all other strains (96%) had MLST profiles that differed in only one or two loci from the ST6 profile and thus belonged to the same clonal complex. There was no apparent expansion of particular STs with the exception of ST190. This ST was found in 12 strains isolated in the postvaccination era and was absent in the prevaccination collection. The ST190 was present in 8 of the 180 (4.4%) of the strains from children aged 0 to 4 years and in 4 of the 61 (6.6%) adult patients older than 40 years. Four of the ST190 strains were isolated in 2001 and 2002 from children with vaccine failures. ST190 differs from ST6 in 12 nucleotides in the sequence of the mdh gene only.
The diversity index by MLST of the 241 strains was 41.4%. There were no apparent differences between the MLST profiles of Hib strains from the postvaccination period compared to those isolated before the introduction of the vaccine. The Simpson's index of diversity for strains isolated from 0 to 4 year olds during the postvaccination era was higher (47.9%) than that in the prevaccination strains (39.4%). However, Monte-Carlo analysis showed that this difference was statistically not significant (P = 0.14).
MLVA of Hib. The results obtained with MLST typing suggested that the Hib strains isolated during the postvaccination period were more diverse than those of the prevaccination period, albeit statistical significance could not be reached. To investigate these findings more thoroughly, we expanded the strain collection to 574 isolates and analyzed the strains by MLVA. The strain collection was composed 536 Hib strains, 7 serotype f strains, and 31 noncapsulated H. influenzae strains. Of the 574 Hi isolates, 16 Hib strains yielded incomplete MLVA profiles, and these strains were omitted from further analysis.
Among the 558 H. influenzae strains that yielded complete MLVA profiles in the present study, 68 different MLVA types were observed. Although some overlap occurred, the MLVA profiles of the serotype f and noncapsulated strains were mostly distinct from the Hib strains (Fig. 1B). In a number of the H. influenzae strains the VNTR PCR products were composed of the repeat flanking regions only and contained no repeats. In these cases the number of repeats was designated 0 in the MLVA profile. For VNTR5-2 this happened in a single strain (0.2%), for VNTR6-1 it happened in 59 strains (10.6%), for VNTR12-1 it happened in 3 strains (0.5%), and in VNTR12-2 it occurred in 27 strains (4.8%). The VNTR5-2 locus had the lowest degree of diversity (DI = 13.7%) and locus VNTR12-2 the highest (DI = 64.9%) (Table 3).
Of the 520 Hib strains with complete MLVA profiles, 312 were isolated in the prevaccination period from 1983 to 1993 and 208 strains in the years from 1994 to 2002, after the introduction of the Hib vaccine in the Dutch national immunization program. There were 45 different MLVA types (Table 4) among the 520 Hib strains (diversity index DI = 75.9%), 7 MLVA types among 7 Hif strains, and 27 among the 31 noncapsulated H. influenzae strains. The predominant MLVA types among the Hib strains were MT40 (44.2%), MT19 (17.9%), and MT22 (9.0%). These MLVA types were not found among the Hif and noncapsulated strains from the present study. However, a single noncapsulated strain had MLVA profile MT40 and, after further analysis, this strain turned out to be a so-called Hib-minus strain, a Hib strain that does carry the genetic information for capsule biosynthesis but has lost its ability to express the capsule. There was no apparent emergence, loss, or expansion of predominant MLVA types among the Hib strains isolated in the postvaccination period.
Changes in genetic diversity of the Hib strains after introduction of the vaccine. MLST of a limited number of Hib isolates suggested that the genetic diversity of the Hib population had increased after the introduction of vaccination against Hib. The genetic diversity of the Hib population before and after vaccination was analyzed after the number of Hib isolates was extended and by using MLVA, a typing method with a higher resolving power. Figure 2 shows the genetic diversity based on MLVA profiles of Hib strains isolated from 0 to 4 year olds and that of strains isolated from patients older than 4 years. Strains from 0 to 4 year olds isolated in the period before the introduction of the vaccine show a limited genetic diversity with a Simpson's index of diversity of 66%. However, the genetic diversity of the Hib strains from the same age group isolated after the introduction of the vaccine had increased considerably reaching a Simpson's index of diversity of 84%. In contrast, the genetic diversity of the strains isolated from patients older than 4 years of age did not change after the introduction of the vaccine and had a Simpson's index of diversity of 78%.
In order to determine whether the observed increase in genetic diversity was significant a Monte-Carlo analysis was performed, and this showed that the observed increase in genetic diversity calculated from the MLVA data was highly significant (P < 0.01). Limiting the data set to the same as that used in the MLST analyses still yielded a significant genetic difference between isolates form the prevaccination period and those from the postvaccination period (P = 0.02). This indicates that the lower resolving power of MLST rather than the small sample size caused the lack of significance in the Monte-Carlo analysis of the MLST data. The lower resolving power of MLST was obvious from the fact that the discrimination index of the MLST was 40.4% for the 239 Hib strains tested, whereas the MLVA on the same set of strains yielded a discrimination index of 81.13%.
When the MLVA data were further stratified in age groups the results suggested that the postvaccination increase in diversity was also present in Hib strains from the age group of 4- to 19-year-old patients and that there was a slight decrease in the diversity in strains from the group of patients older than 19 years (Fig. 3). However, the number of strains from the stratified group is too low to ensure the statistical significance of these results. Stratification by time period using 3-year intervals showed that in the strains from neonates to 4-year-old patients the diversity sharply increased after 1993, the year in which Hib vaccination was introduced in the Dutch national vaccination program (Fig. 4). Such increase was not seen in strains isolated from patients older than 4 years.
DISCUSSION
In The Netherlands the number of cases of vaccine failures had tripled in 2002 compared to the years from 1996 to 2001, albeit that in 2003 the number of Hib vaccine failures in The Netherlands remained approximately the same as in 2002 (24; data not shown). In the present study we addressed the issue of increased incidence of invasive Hib disease in fully vaccinated Dutch children. In addition, we wanted to study the impact of the introduction of the Hib vaccine in 1993 on the composition of the Hib population and therefore analyzed a large number of strains collected during the years from 1983 to 2002.
The results obtained by MLST analyses of Hib strains collected during the prevaccination period and during the postvaccination period indicated that the increase of Hib cases among vaccinees was not caused by Hib with a particular genotype. This was confirmed by MLVA typing, showing that strains isolated from the vaccine failure patients did not comprise a group with distinct MLVA profiles. Since there was no apparent emergence of loss of particular dominant MLST or MLVA types, we could find no evidence that there had been a clonal expansion of vaccine escape variants. However, in the MLST or MLVA typing loci are characterized that are not related to the virulence of Hib, and this does not exclude the existence of escape variants. To address this question, important virulence genes such as those coding for the polysaccharide capsule need to be analyzed.
For the analysis of Hib strains included in the present study, the discrimination index of MLVA (75.9%) was considerably larger than that of MLST (40.4%), making MLVA more suitable to type Hib strains than MLST. In some organisms, such as Hib, the changes in the housekeeping genes may be too slow to afford enough discrimination by MLST. Possibly, MLST based on more variable genes known to play a role in virulence may resolve this problem. For Listeria monocytogenes and Bordetella pertussis, this has proven to be a successful approach (33, 34). MLVA of Hib isolates revealed an increased genetic diversity of Hib isolates of the postvaccination period compared to those of the prevaccination period. Hib strains isolated from neonates to 4-year-old children with invasive Hib disease in the prevaccination era showed only limited genetic variation. In contrast, Hib strains from patients older than 4 years, mainly adults, were genetically diverse. Strains from neonates to 4-year-old children isolated after the introduction of the vaccine had a high diversity index and were even slightly more diverse than the strains from patients older than 4 years. The strains from the group of older patients had the same degree of diversity before and after the introduction of the Hib vaccine. This finding suggests that, prior to the introduction of the vaccine Hib, young children were infected by Hib strains with limited genetic diversity present in other young children. After the Hib vaccine was introduced children were mainly infected by genetically diverse Hib strains from adults. Possibly, an increase of Hib disease in adults may have been caused by an increased circulation of Hib within this age group. Several studies have reported an impressive decrease in Hib carriage among children after the introduction of the Hib vaccine (1, 19, 27, 28). Although there is no published experimental data to support this, it seems safe to presume that a similar decrease in carriage occurred after the introduction of the Hib vaccine in The Netherlands. In a recent study, McVernon et al. (14) reported that no asymptomatic carriage among children attending day-care nurseries in the United Kingdom sampled in 1997 and 2002 could be detected. In contrast, prevalence of Hib colonization was nearly 4% in children sampled in 1992, the year in which nationwide Hib vaccination was introduced in the United Kingdom. This is a remarkable finding considering there has been a considerable resurgence of invasive Hib disease among children in the United Kingdom from 1999 onward (13, 29, 30). In another recent report McVernon et al. (16) described the increase of Hib disease in patients aged 15 years and older after 1998 concurrent with the increase seen in fully vaccinated children. Although they mention that the reverse may occur as well, these authors theorized that, because transmission usually is from child to parent, the increase in the incidence of pediatric Hib disease has caused an increase in adults. However, this would suggest that the children still constitute the major reservoir for Hib infection and contradicts the observation that no carriage was detected in children and our observation that in the postvaccination period the genetic diversity of Hib isolates from children reached an even higher level than that of the Hib isolates from adults.
The observed abrupt increase in genetic diversity of Hib after the introduction of the vaccine may have been caused by unmasking highly susceptible individuals. Musser et al. showed that the bulk of invasive Hib disease is caused by a few clones or lineages but that there is abundant genetic variation (20, 21). Apparently, some clones are more successful in colonizing healthy young children compared to other, less abundant, Hib clones. This may account for a limited number of genotypes found in of cases of invasive disease. A minority of children may be highly susceptible to infection with Hib and even less successful, genetically diverse Hib clones may colonize these individuals and cause disease. In the bulk of the cases of invasive disease occurring prior to the nationwide Hib vaccination, disease caused by these minor types would go unnoted. However, when the circulation of the dominant Hib types and the number of cases of invasive disease was greatly reduced due to vaccination, disease in highly susceptible children may have continued to occur and eventually made up the majority of the Hib cases. As a result mainly highly susceptible children would get invasive disease, and disease would be caused by genetically diverse strains. Although this theory fits with the observed increased genetic diversity it does not explain why the total number of cases has been rising. However, if the highly susceptible children are infected mainly from the genetically diverse reservoir present in older people, an increase in Hib circulation in adults could result in an increase of Hib colonization and disease.
An increase in vaccine failures has only been reported in the United Kingdom and in The Netherlands. In the United Kingdom the lack of a booster vaccination or the decline in Hib colonization may have caused waning immunity and/or loss of herd immunity, leading to a reduced protection of children 1 to 4 years in age but not in the children under 1 year of age (15). Analysis of the increase seen in Dutch invasive Hib disease shows that, in contrast to the United Kingdom, there is no difference between the increases in the age groups of children under 1 year of age and in those 1 to 4 years of age (Fig. 5). In The Netherlands there has been an increase since 1999 in all age groups, particularly in persons older than 4 years. This indicates that the increases seen in both countries may have, at least in part, different causes. There have been two major differences between the Hib vaccinations in United Kingdom and The Netherlands. First, the Dutch Hib vaccination program includes a booster at 11 months of age which may prevent waning immunity among children. This could explain why children aged 1 to 4 years are protected as well as those younger than 1 year. It will be interesting to genotype Hib strains from the United Kingdom by MLVA to determine whether the increase in genetic diversity seen in the Dutch strains can also be found among the United Kingdom Hib strains. The second difference between the vaccination schedules of the two countries is that no acellular pertussis vaccine has been used in combination with the Hib vaccine in The Netherlands. However, an acellular pertussis vaccine will be introduced into the Dutch national vaccination program on January 2005 to replace the whole-cell vaccine, and this may result in increased incidence of invasive Hib disease. Ongoing surveillance will be required to monitor the effect of the change in the vaccine mixture on the incidence and manifestations of Hib disease, as well as on the composition of the Hib population.
REFERENCES
Barbour, M. L., R. T. Mayon-White, C. Coles, D. W. Crook, and E. R. Moxon. 1995. The impact of conjugate vaccine on carriage of Haemophilus influenzae type b. J. Infect. Dis. 171:93-98.
Dingle, K. E., F. M. Colles, D. R. A. Wareing, R. Ure, A. J. Fox, F. E. Bolton, H. J. Bootsma, R. J. L. Willems, R. Urwin, and M. C. Maiden. J. 2001. Multilocus sequence typing system for Campylobacter jejuni. J. Clin. Microbiol. 39:14-23.
Enright, M. C., N. P. J. Day, C. E. Davies, S. J. Peacock, and B. G. Spratt. 2000. Multilocus sequence typing for characterization of methicillin-resistant and methicillin-susceptible clones of Staphylococcus aureus. J. Clin. Microbiol. 38:1008-1015.
Enright, M. C., and B. G. Spratt. 1998. A multilocus sequence typing scheme for Streptococcus pneumoniae: identification of clones associated with serious invasive disease. Microbiology 144:3049-3060.
Farlow, J., D. Postic, K. L. Smith, Z. Jay, G. Baranton, and P. Keim. 2002. Strain typing of Borrelia burgdorferi, Borrelia afzelii, and Borrelia garinii by using multiple-locus variable-number tandem repeat analysis. J. Clin. Microbiol. 40:4612-4618.
Farlow, J., K. L. Smith, J. Wong, M. Abrams, M. Lytle, and P. Keim. 2001. Francisella tularensis strain typing using multiple-locus, variable-number tandem repeat analysis. J. Clin. Microbiol. 39:3186-3192.
Frothingham, R., and W. A. Meeker-O'Connell. 1998. Genetic diversity in the Mycobacterium tuberculosis complex based on variable numbers of tandem DNA repeats. Microbiology 144:1189-1196.
Heath, P. T., R. Booy, H. Griffiths, E. Clutterbuck, H. J. Azzopardi, M. P. Slack, J. Fogarty, A. C. Moloney, and E. R. Moxon. 2000. Clinical and immunological risk factors associated with Haemophilus influenzae type b conjugate vaccine failure in childhood. Clin. Infect. Dis. 31:973-980.
Hunter, P. R., and M. A. Gaston. 1988. Numerical index of the discriminatory ability of typing systems: an application of Simpson's index of diversity. J. Clin. Microbiol. 26:2465-2466.
Klevytska, A. M., L. B. Price, J. M. Schupp, P. L. Worsham, J. Wong, and P. Keim. 2001. Identification and characterization of variable-number tandem repeats in the Yersinia pestis genome. J. Clin. Microbiol. 39:3179-3185.
Liu, Y., M. A. Lee, E. E. Ooi, Y. Mavis, A. L. Tan, and H. H. Quek. 2003. Molecular typing of Salmonella enterica serovar Typhi isolates from various countries in Asia by a multiplex PCR assay on variable-number tandem repeats. J. Clin. Microbiol. 41:4388-4394.
Maiden, M. C. J., J. A. Bygraves, E. Feil, G. Morelli, J. E. Russell, R. Urwin, Q. Zhang, J. J. Zhou, K. Zurth, D. A. Caugant, I. M. Feavers, M. Achtman, and B. G. Spratt. 1998. Multilocus sequence typing: a portable approach to the identification of clones within populations of pathogenic microorganisms. Proc. Natl. Acad. Sci. USA 95:3140-3145.
McVernon, J., N. Andrews, M. P. Slack, and M. E. Ramsay. 2003. Risk of vaccine failure after Haemophilus influenzae type b (Hib) combination vaccines with acellular pertussis. Lancet 361:1521-1523.
McVernon, J., A. J. Howard, M. P. Slack, and M. E. Ramsay. 2004. Long-term impact of vaccination on Haemophilus influenzae type b (Hib) carriage in the United Kingdom. Epidemiol. Infect. 132:765-767.
McVernon, J., N. A. Mitchison, and E. R. Moxon. 2004. T helper cells and efficacy of Haemophilus influenzae type b conjugate vaccination. Lancet Infect. Dis. 4:40-43.
McVernon, J., C. L. Trotter, M. P. Slack, and M. E. Ramsay. 2004. Trends in Haemophilus influenzae type b infections in adults in England and Wales: surveillance study. Br. Med. J. 329:655-658.
Meats, E., E. J. Feil, S. Stringer, A. J. Cody, R. Goldstein, J. S. Kroll, T. Popovic, and B. G. Spratt. 2003. Characterization of encapsulated and noncapsulated Haemophilus influenzae and determination of phylogenetic relationships by multilocus sequence typing. J. Clin. Microbiol. 41:1623-1636.
Metropolis, N., and S. Ulam. 1949. The Monte Carlo method. J. Am. Stat. Assoc. 44:335-341.
Mohle-Boetani, J. C., G. Ajello, E. Breneman, K. A. Deaver, C. Harvey, B. D. Plikaytis, M. M. Farley, D. S. Stephens, and J. D. Wenger. 1993. Carriage of Haemophilus influenzae type b in children after widespread vaccination with conjugate Haemophilus influenzae type b vaccines. Pediatr. Infect. Dis. J. 12:589-593.
Musser, J. M., J. S. Kroll, D. M. Granoff, E. R. Moxon, B. R. Brodeur, J. Campos, H. Dabernat, W. Frederiksen, J. Hamel, G. Hammond, et al. 1990. Global genetic structure and molecular epidemiology of encapsulated Haemophilus influenzae. Rev. Infect. Dis. 12:75-111.
Musser, J. M., J. S. Kroll, E. R. Moxon, and R. K. Selander. 1988. Evolutionary genetics of the encapsulated strains of Haemophilus influenzae. Proc. Natl. Acad. Sci. USA 85:7758-7762.
Peltola, H. 2000. Worldwide Haemophilus influenzae type b disease at the beginning of the 21st century: global analysis of the disease burden 25 years after the use of the polysaccharide vaccine and a decade after the advent of conjugates. Clin. Microbiol. Rev. 13:302-317.
Ramsay, M. E., J. McVernon, N. J. Andrews, P. T. Heath, and M. P. Slack. 2003. Estimating Haemophilus influenzae type b vaccine effectiveness in England and Wales by use of the screening method. J. Infect. Dis. 188:481-485.
Rijkers, G. T., P. E. Vermeer-de Bondt, L. Spanjaard, M. A. Breukels, and E. A. Sanders. 2003. Return of Haemophilus influenzae type b infections. Lancet 361:1563-1564.
Schouls, L. M., van der Heide, H. G., L. Vauterin, P. Vauterin, and F. R. Mooi. 2004. Multiple-locus variable-number tandem repeat analysis of Dutch Bordetella pertussis strains reveals rapid genetic changes with clonal expansion during the late 1990s. J. Bacteriol. 186:5496-5505.
Simpson, E. H. 1949. Measurement of diversity. Nature 163:688.
Takala, A. K., J. Eskola, M. Leinonen, H. Kayhty, A. Nissinen, E. Pekkanen, and P. H. Makela. 1991. Reduction of oropharyngeal carriage of Haemophilus influenzae type b (Hib) in children immunized with an Hib conjugate vaccine. J. Infect. Dis. 164:982-986.
Takala, A. K., M. Santosham, J. Almeido-Hill, M. Wolff, W. Newcomer, R. Reid, H. Kayhty, E. Esko, and P. H. Makela. 1993. Vaccination with Haemophilus influenzae type b meningococcal protein conjugate vaccine reduces oropharyngeal carriage of Haemophilus influenzae type b among American Indian children. Pediatr. Infect. Dis. J. 12:593-599.
Trotter, C. L., J. McVernon, N. J. Andrews, M. Burrage, and M. E. Ramsay. 2003. Antibody to Haemophilus influenzae type b after routine and catch-up vaccination. Lancet 361:1523-1524.
Trotter, C. L., M. E. Ramsay, and M. P. Slack. 2003. Rising incidence of Haemophilus influenzae type b disease in England and Wales indicates a need for a second catch-up vaccination campaign. Commun. Dis. Public Health 6:55-58.
van Belkum, A., S. Scherer, W. van Leeuwen, D. Willemse, L. van Alphen, and H. Verbrugh. 1997. Variable number of tandem repeats in clinical strains of Haemophilus influenzae. Infect. Immun. 65:5017-5027.
Van der Ende, A., L. Spanjaard, and J. Dankert. 2003. Bacterial meningitis in The Netherlands: 31th annual report of The Netherlands Reference Laboratory for Bacterial Meningitis, p. 1-50. The Netherlands Reference Laboratory for Bacterial Meningitis, University of Amsterdam, Amsterdam, The Netherlands.
Van Loo, I. H., K. J. Heuvelman, A. J. King, and F. R. Mooi. 2002. Multilocus sequence typing of Bordetella pertussis based on surface protein genes. J. Clin. Microbiol. 40:1994-2001.
Zhang, W., B. M. Jayarao, and S. J. Knabel. 2004. Multi-virulence-locus sequence typing of Listeria monocytogenes. Appl. Environ. Microbiol. 70:913-920.(Leo M. Schouls, Arie van )
Netherlands Reference Laboratory for Bacterial Meningitis, Department of Medical Microbiology, Academic Medical Centre, Amsterdam, The Netherlands
Applied Maths, Sint-Martens-Latem, Belgium
ABSTRACT
Recently, there has been an increase in The Netherlands in the number of cases of invasive disease caused by Haemophilus influenzae serotype b (Hib). To study a possible change in the Hib population that could explain the rise in incidence, a multiple-locus variable number tandem repeats analysis (MLVA) was developed to genotype H. influenzae isolates. The MLVA enabled the differentiation of H. influenzae serotype b strains with higher discriminatory power than multilocus sequence typing (MLST). MLVA profiles of noncapsulated H. influenzae and H. influenzae serotype f strains were more heterogeneous than serotype b strains and were distinct from Hib, although some overlap occurred. The MLVA was used to genotype a collection of 520 H. influenzae serotype b strains isolated from patients in The Netherlands with invasive disease. The strains were collected from 1983 from 2002, covering a time period of 10 years before and 9 years after the introduction of the Hib vaccine in the Dutch national vaccination program. MLVA revealed a sharp increase in genetic diversity of Hib strains isolated from neonates to 4-year-old patients after 1993, when the Hib vaccine was introduced. Hib strains isolated from patients older than 4 years in age were genetically diverse, and no significant change in diversity was seen after the introduction of the vaccine. These observations suggest that after the introduction of the Hib vaccine young children no longer constitute the reservoir for Hib and that they are infected by adults carrying genetically diverse Hib strains.
INTRODUCTION
Haemophilus influenzae is a respiratory pathogen that causes a wide spectrum of infections in humans (22). H. influenzae carrying a polysaccharide capsule have been an important cause of serious invasive diseases such as meningitis and septicemia in children. The majority of cases of invasive disease are caused by H. influenzae serotype b (Hib), and this has led to the introduction of nationwide vaccination with Hib polysaccharide conjugate vaccines in many developed countries. Vaccination has resulted in impressive reductions in Hib disease and reduced carriage of Hib among both vaccinated and nonvaccinated individuals. In The Netherlands, where vaccination was introduced in 1993, the incidence of invasive Hib disease among children younger than 5 years of age dropped from 28.7 per 100,000 in 1992 to 0.8 in 2002 (32). Similar reductions of Hib disease have been observed in most other countries where vaccination has been implemented in the national vaccination programs and has almost led to eradication of invasive Hib disease among children.
Despite nearly complete vaccine coverage, a small number of fully vaccinated children in The Netherlands have experienced invasive Hib infection. In 2002 the number of cases of vaccine failures in Dutch children aged 0 to 4 years increased by a factor of 3 compared to the years 1996 to 2001 (24). Although the number of cases of invasive Hib disease was quite low, with only 16 cases among 0 to 4 year olds, 14 of these (81%) occurred in fully vaccinated children. A more impressive rise has been observed in the United Kingdom, where the incidence among children in the age group of 0 to 4 years started to rise in 1999 to reach 3.7 per 100,000 in 2002 (23). The rise in disease in the United Kingdom has been associated with several factors. First, relatively fast waning immunity in the age group 1 to 4 years has resulted in lower vaccine-induced antibody levels (8, 23, 29). Second, the use of combination vaccines containing acellular pertussis vaccine has been associated with reduced immunogenicity of the Hib component (13). To compensate for the reduced anti-Hib titers, a catch-up campaign designed to boost immunity in children aged 6 months to 4 years of age was implemented in the United Kingdom (30). In contrast to the United Kingdom, the vaccination schedule in The Netherlands contains a booster vaccination at 11 months of age. Furthermore, only whole-cell pertussis vaccine has been used in combination with the Hib vaccine. This suggests that waning immunity may not be the cause of the rise of vaccine failures in The Netherlands. Also, the total number of cases of invasive Hib disease among all age groups has increased significantly and is back to the level seen in 1996, suggesting that there has been an increase in the circulation of Hib in The Netherlands, too (32).
The various reports that discuss the increase of invasive Hib in the United Kingdom focus on immune responses and do not mention a change in the properties of the bacteria as a possible cause of the increased number of cases. The reason for ignoring this possibility probably lies in the fact that the bacterial component used for vaccination, the capsular polysaccharide, is a very simple component that is unlikely to change. Alterations in the Hib polysaccharide would almost certainly result in the loss of the reactivity of the bacteria with the sera used for capsular typing. However, subtle changes in capsule structure or expression could go unnoticed if only serotyping is performed. Furthermore, changes may not be restricted to the capsule but could be related to other virulence factors. This may result in an altered composition of the circulating Hib population in which particular clones are successful in evading the barrier of the vaccine induced immunity. We describe here the genotypic characterization by MLVA and MLST of Dutch Hib strains collected in the period from 1983 to 2002, comprising a 10-year period before the introduction of the Hib vaccine in 1993 and 9 years after nationwide Hib vaccination of newborns.
MATERIALS AND METHODS
Bacterial strains. A total of 574 H. influenzae strains were included in the present study. Of this collection, 536 strains were serotype b, 7 were serotype f, and 31 were nontypeable. All strains were collected during the years from 1983 to 2002 by The Netherlands Reference Laboratory for Bacterial Meningitis from patients with invasive disease in The Netherlands. Of the 520 Hib strains, 349 were isolated from cerebrospinal fluid or from cerebrospinal fluid and blood, 161 were from blood only, and 12 were from other materials such as synovial fluid. From each year ca. 20 Hib strains isolated from children in the age group 0 to 4 year and 10 strains from patients older than 4 years, mainly adults, were included. The selected strains contained all available Hib strains isolated from 1996 to 2002 and randomly selected strains from the time period before 1996.
Bacterial growth and preparation of lysates. Cultures stored at –80°C were streaked onto tryptic soy chocolate agar plates, cultured overnight at 35°C in an atmosphere of 95% air-5% CO2, and visually inspected for purity. In the few cases where cultures were contaminated, a single colony was restreaked to obtain a pure culture. A loopful of cells from the pure culture was suspended in 500 μl of TE (10 mM Tris-HCl, 1 mM EDTA [pH 8]) and heated for 10 min at 95°C to lyse the cells, and then the lysate was stored at –20°C until use in PCR.
MLST. For multilocus sequence typing (MLST), the typing scheme proposed by Meats et al. (17) was used to analyze the H. influenzae strains in the present study. During the setup of the MLST in our lab the published PCR conditions proved unsuitable when using lysates as a source of DNA for MLST. Therefore, a number of new primers were designed (Table 1), and the PCR conditions were adapted to facilitate a multiplex format to amplify all seven gene fragments in a single reaction. Optimized multiplex PCRs were performed in 15-μl volumes in 96-well PP-PCR plates (Greiner Bio-One, Frickenhausen, Germany) in Applied Biosystems 9700 PCR machines (Applied Biosystems, Foster City, Calif.). A total of 5 μl of 1:10 diluted heat-treated H. influenzae lysate was added to a 10-μl mixture containing 7.5 pmol of each primer and 7.5 μl of HotStar Taq mastermix (Qiagen, Hilden, Germany). The PCR program used was 15 min at 95°C, followed by 30 cycles of amplification that consisted of 30 s at 95°C, 30 s at 50°C, and 1 min at 72°C, and a final step of 10 min at 72°C. PCR products were purified by adding 6 μl of ExoSAP-IT (USB Corp., Cleveland, Ohio), followed by subsequent incubations for 15 min at 37°C and 15 min at 80°C. All MLST genes were sequenced from the multiplex PCR mixture in 14 separate sequence reactions by using 1 μl of purified PCR product and 5 pmol of PCR primer per 20-μl sequence reaction. Sequence reactions were performed with the ABI Prism BigDye Terminator cycle sequencing kit (v3.1; Applied Biosystems, Foster City, Calif.) and analyzed on an ABI 3700 DNA sequencer. If a sequence reaction failed to yield an unambiguous MLST DNA sequence, PCR and sequencing of this particular gene were repeated. In that case, PCR was performed in a 10-μl volume with 0.5 μl of lysate and 5 pmol of each of the two primers and 5 μl of HotStar Taq mastermix using the above-mentioned PCR cycling protocol. Subsequently, sequence reactions were performed on 1 μl of the unpurified PCR product according to the above-described protocol. The sequence trace files were imported into a Bionumerics 4 database (Applied Maths, Sint-Martens-Latem, Belgium) and aligned and trimmed by using a script. After editing the MLST sequences, another set of scripts was used to assign allele numbers and sequence types (STs) using the data from the H. influenzae MLST scheme available on the MLST website (http://www.mlst.net/). Allele sequences and MLST profiles that were not present in the database were assigned provisional allele numbers starting at 800 and provisional STs starting at 500.
Identification of VNTR loci and MLVA design. MLVA is increasingly used as a typing method for bacterial species (5-7, 10, 11, 25). For H. influenzae VNTR analysis has been described by van Belkum et al. (31). Therefore, we initially repeated the VNTR PCRs on the 18 VNTR loci described in van Belkum's study and included 10 additional tandem repeat loci. Analysis showed that only two loci previously described by Van Belkum et al. (31) VNTR5-2 and VNTR6-1 and two newly identified 12-bp tandem repeat loci VNTR12-1 and VNTR12-2 yielded limited, but sufficient, variation and were used to compose the MLVA profile. The majority of the other loci were either too variable in composition or not present in all strains from the test panel. To study the in vitro stability of the MLVA loci, we subcultured a Hib strain, a noncapsulated H. influenzae strain, and an H. influenzae strain Rd through daily serial passage for a month. The composition of the VNTR loci was determined, and no alterations were observed, indicating that the number of repeats is relatively stable. However, this does not exclude rapid changes if strains were passaged through human hosts.
With the exception of VNTR5-2, the VNTR loci of the MLVA set were located in open reading frames in the H. influenzae Rd strain (Table 2). VNTR5-2 is located upstream from the gene encoding for ribosome-binding factor (rbfA) which is essential for efficient processing of the 16S rRNA. Variation in the number of repeats may alter the level of expression of rfbA. VNTR6-1 is located within tonB, a gene required to energize outer membrane transport reaction. VNTR12-1 is located in a "predicted coding region" with unknown function, and VNTR12-2 is found within the hsdS gene which is involved in the type I restriction modification system of the bacterial cell.
MLVA. Primers were designed for two newly identified loci: VNTR12-1 and VNTR12-2. Furthermore, primers for the earlier described VNTR5-2 and VNTR6-1 loci (31) were redesigned to yield robust PCRs. The MLVA primers used to type the H. influenzae strains from this collection are listed in Table 1. VNTR PCRs were performed in 20-μl volumes in Applied Biosystems 9700 PCR machines. We added 1 μl of 1:10-diluted heat-treated H. influenzae lysate to a mixture containing 8 pmol of 5' 6-carboxyfluorescein-labeled (FAM) forward primer, 8 pmol of unlabeled reverse primer, and 10 μl of HotStar Taq mastermix. VNTR loci were amplified in separate PCRs by using the following program: 15 min at 95°C, followed by 17 cycles of amplification that consisted of 30 s at 95°C, 1 min at 54°C, and 1 min at 72°C, with a final step of 30 min at 68°C to ensure complete terminal transferase activity of the Taq DNA polymerase. After PCR, samples were diluted 1:200 in water, and 2 μl of the diluted samples was mixed with 10 μl of GeneScan 500 ROX-Marker (Applied Biosystems). After heat denaturation for 5 min at 95°C, fragments were separated on an ABI 3700 DNA sequencer by using the standard GeneScan module with filterset D. The GeneScan data were analyzed in the GenoTyper software (Applied Biosystems) to perform sizing and to calculate the number of repeats in the PCR fragments. The data with calculated number of repeats were imported into the Bionumerics software package for further cluster analysis. The data obtained with the number of repeats found at the four different VNTR loci in the order VNTR5-2, VNTR6-1, VNTR12-1, and VNTR12-2 were used to make up the MLVA profile. The accuracy of the sizing and the translation of the fragment sizes into repeat numbers was confirmed by sequencing a number of VNTR PCR fragments.
Data analysis. The MLST and MLVA profiles were clustered in the Bionumerics software by using a categorical coefficient and a graphing method called a minimum spanning tree as described before (25). In the minimum spanning tree the priority rules to first link types that have the highest number of single locus variants was chosen. For MLST a maximum neighbor difference of 2 was used to create complexes.
For calculation of the genetic diversity and discrimination index the Simpson's index of diversity was used as follows:
where ni is the number of strains belonging to ith type and N is the total number of strains in the sample population (9, 26).
In the present study, we compared diversity indices from various samples, inferring conclusions that the diversity of one sample is larger than of the other. Because these indices are based on limited sample sizes, they are subject to statistical uncertainty. Hence, in order to be able to validate the observed differences, it was necessary to calculate their statistical significance as measured by P values. We used Monte-Carlo simulations to obtain these values (18). It is well known that, for an observed sample with size N, the class sizes for all types ni (i = 1 ... p) follow a multinomial distribution function, characterized by the class probabilities i. These probabilities i are unknown but can be estimated from the sample by calculating i as ni/N. During the Monte-Carlo simulation process, a large number (10,000) of randomly generated synthetic data sets with sample size N is generated, all having the same multinomial distribution of strains over the different types as dictated by the probabilities i. In order to obtain a P value for the difference in diversity indices between two samples, this process is performed for both samples, and the fraction of simulations that violate the hypothesis about the difference in diversity yields the P value.
RESULTS
MLST of Hib strains. MLST can be considered the gold standard for molecular typing of a number of bacterial species (2-4, 12). Recently, an MLST typing scheme for H. influenzae has been published (17), and we have adapted the MLST protocol by redesigning the primers and creating a multiplex PCR in which all seven MLST fragments were amplified in a single PCR. Use of the multiplex approach saves time and cost. In addition, the redesigned primers enabled the use of crude H. influenzae lysates for PCR. The multiplex PCR seemed to be robust, amplifying sufficient PCR product of all MLST loci with the exception of the atpG locus, which often was insufficiently amplified. For future use of the multiplex MLST PCR, a new primer set will have to be designed for amplification of the atpG locus. The collection of strains used for the MLST was composed of 241 Hib strains. A total of 61 strains were from the prevaccination era and dated from 1989 to 1993, and the remaining 180 strains were isolated in the years from 1994 to 2002, after the introduction of the Hib vaccine. Of the strains from the prevaccination era only 2 were from patients older than 4 years, and 59 of the strains from the postvaccination group were from patients older than 4 years. There were 29 different STs among the 241 strains (Fig. 1A). Of the 29 STs, 17 represented previously unidentified allele combinations (http://haemophilus.mlst.net/) and were therefore designated provisional STs. The majority of the strains had ST6 (76.3%) and, with the exception of nine strains, all other strains (96%) had MLST profiles that differed in only one or two loci from the ST6 profile and thus belonged to the same clonal complex. There was no apparent expansion of particular STs with the exception of ST190. This ST was found in 12 strains isolated in the postvaccination era and was absent in the prevaccination collection. The ST190 was present in 8 of the 180 (4.4%) of the strains from children aged 0 to 4 years and in 4 of the 61 (6.6%) adult patients older than 40 years. Four of the ST190 strains were isolated in 2001 and 2002 from children with vaccine failures. ST190 differs from ST6 in 12 nucleotides in the sequence of the mdh gene only.
The diversity index by MLST of the 241 strains was 41.4%. There were no apparent differences between the MLST profiles of Hib strains from the postvaccination period compared to those isolated before the introduction of the vaccine. The Simpson's index of diversity for strains isolated from 0 to 4 year olds during the postvaccination era was higher (47.9%) than that in the prevaccination strains (39.4%). However, Monte-Carlo analysis showed that this difference was statistically not significant (P = 0.14).
MLVA of Hib. The results obtained with MLST typing suggested that the Hib strains isolated during the postvaccination period were more diverse than those of the prevaccination period, albeit statistical significance could not be reached. To investigate these findings more thoroughly, we expanded the strain collection to 574 isolates and analyzed the strains by MLVA. The strain collection was composed 536 Hib strains, 7 serotype f strains, and 31 noncapsulated H. influenzae strains. Of the 574 Hi isolates, 16 Hib strains yielded incomplete MLVA profiles, and these strains were omitted from further analysis.
Among the 558 H. influenzae strains that yielded complete MLVA profiles in the present study, 68 different MLVA types were observed. Although some overlap occurred, the MLVA profiles of the serotype f and noncapsulated strains were mostly distinct from the Hib strains (Fig. 1B). In a number of the H. influenzae strains the VNTR PCR products were composed of the repeat flanking regions only and contained no repeats. In these cases the number of repeats was designated 0 in the MLVA profile. For VNTR5-2 this happened in a single strain (0.2%), for VNTR6-1 it happened in 59 strains (10.6%), for VNTR12-1 it happened in 3 strains (0.5%), and in VNTR12-2 it occurred in 27 strains (4.8%). The VNTR5-2 locus had the lowest degree of diversity (DI = 13.7%) and locus VNTR12-2 the highest (DI = 64.9%) (Table 3).
Of the 520 Hib strains with complete MLVA profiles, 312 were isolated in the prevaccination period from 1983 to 1993 and 208 strains in the years from 1994 to 2002, after the introduction of the Hib vaccine in the Dutch national immunization program. There were 45 different MLVA types (Table 4) among the 520 Hib strains (diversity index DI = 75.9%), 7 MLVA types among 7 Hif strains, and 27 among the 31 noncapsulated H. influenzae strains. The predominant MLVA types among the Hib strains were MT40 (44.2%), MT19 (17.9%), and MT22 (9.0%). These MLVA types were not found among the Hif and noncapsulated strains from the present study. However, a single noncapsulated strain had MLVA profile MT40 and, after further analysis, this strain turned out to be a so-called Hib-minus strain, a Hib strain that does carry the genetic information for capsule biosynthesis but has lost its ability to express the capsule. There was no apparent emergence, loss, or expansion of predominant MLVA types among the Hib strains isolated in the postvaccination period.
Changes in genetic diversity of the Hib strains after introduction of the vaccine. MLST of a limited number of Hib isolates suggested that the genetic diversity of the Hib population had increased after the introduction of vaccination against Hib. The genetic diversity of the Hib population before and after vaccination was analyzed after the number of Hib isolates was extended and by using MLVA, a typing method with a higher resolving power. Figure 2 shows the genetic diversity based on MLVA profiles of Hib strains isolated from 0 to 4 year olds and that of strains isolated from patients older than 4 years. Strains from 0 to 4 year olds isolated in the period before the introduction of the vaccine show a limited genetic diversity with a Simpson's index of diversity of 66%. However, the genetic diversity of the Hib strains from the same age group isolated after the introduction of the vaccine had increased considerably reaching a Simpson's index of diversity of 84%. In contrast, the genetic diversity of the strains isolated from patients older than 4 years of age did not change after the introduction of the vaccine and had a Simpson's index of diversity of 78%.
In order to determine whether the observed increase in genetic diversity was significant a Monte-Carlo analysis was performed, and this showed that the observed increase in genetic diversity calculated from the MLVA data was highly significant (P < 0.01). Limiting the data set to the same as that used in the MLST analyses still yielded a significant genetic difference between isolates form the prevaccination period and those from the postvaccination period (P = 0.02). This indicates that the lower resolving power of MLST rather than the small sample size caused the lack of significance in the Monte-Carlo analysis of the MLST data. The lower resolving power of MLST was obvious from the fact that the discrimination index of the MLST was 40.4% for the 239 Hib strains tested, whereas the MLVA on the same set of strains yielded a discrimination index of 81.13%.
When the MLVA data were further stratified in age groups the results suggested that the postvaccination increase in diversity was also present in Hib strains from the age group of 4- to 19-year-old patients and that there was a slight decrease in the diversity in strains from the group of patients older than 19 years (Fig. 3). However, the number of strains from the stratified group is too low to ensure the statistical significance of these results. Stratification by time period using 3-year intervals showed that in the strains from neonates to 4-year-old patients the diversity sharply increased after 1993, the year in which Hib vaccination was introduced in the Dutch national vaccination program (Fig. 4). Such increase was not seen in strains isolated from patients older than 4 years.
DISCUSSION
In The Netherlands the number of cases of vaccine failures had tripled in 2002 compared to the years from 1996 to 2001, albeit that in 2003 the number of Hib vaccine failures in The Netherlands remained approximately the same as in 2002 (24; data not shown). In the present study we addressed the issue of increased incidence of invasive Hib disease in fully vaccinated Dutch children. In addition, we wanted to study the impact of the introduction of the Hib vaccine in 1993 on the composition of the Hib population and therefore analyzed a large number of strains collected during the years from 1983 to 2002.
The results obtained by MLST analyses of Hib strains collected during the prevaccination period and during the postvaccination period indicated that the increase of Hib cases among vaccinees was not caused by Hib with a particular genotype. This was confirmed by MLVA typing, showing that strains isolated from the vaccine failure patients did not comprise a group with distinct MLVA profiles. Since there was no apparent emergence of loss of particular dominant MLST or MLVA types, we could find no evidence that there had been a clonal expansion of vaccine escape variants. However, in the MLST or MLVA typing loci are characterized that are not related to the virulence of Hib, and this does not exclude the existence of escape variants. To address this question, important virulence genes such as those coding for the polysaccharide capsule need to be analyzed.
For the analysis of Hib strains included in the present study, the discrimination index of MLVA (75.9%) was considerably larger than that of MLST (40.4%), making MLVA more suitable to type Hib strains than MLST. In some organisms, such as Hib, the changes in the housekeeping genes may be too slow to afford enough discrimination by MLST. Possibly, MLST based on more variable genes known to play a role in virulence may resolve this problem. For Listeria monocytogenes and Bordetella pertussis, this has proven to be a successful approach (33, 34). MLVA of Hib isolates revealed an increased genetic diversity of Hib isolates of the postvaccination period compared to those of the prevaccination period. Hib strains isolated from neonates to 4-year-old children with invasive Hib disease in the prevaccination era showed only limited genetic variation. In contrast, Hib strains from patients older than 4 years, mainly adults, were genetically diverse. Strains from neonates to 4-year-old children isolated after the introduction of the vaccine had a high diversity index and were even slightly more diverse than the strains from patients older than 4 years. The strains from the group of older patients had the same degree of diversity before and after the introduction of the Hib vaccine. This finding suggests that, prior to the introduction of the vaccine Hib, young children were infected by Hib strains with limited genetic diversity present in other young children. After the Hib vaccine was introduced children were mainly infected by genetically diverse Hib strains from adults. Possibly, an increase of Hib disease in adults may have been caused by an increased circulation of Hib within this age group. Several studies have reported an impressive decrease in Hib carriage among children after the introduction of the Hib vaccine (1, 19, 27, 28). Although there is no published experimental data to support this, it seems safe to presume that a similar decrease in carriage occurred after the introduction of the Hib vaccine in The Netherlands. In a recent study, McVernon et al. (14) reported that no asymptomatic carriage among children attending day-care nurseries in the United Kingdom sampled in 1997 and 2002 could be detected. In contrast, prevalence of Hib colonization was nearly 4% in children sampled in 1992, the year in which nationwide Hib vaccination was introduced in the United Kingdom. This is a remarkable finding considering there has been a considerable resurgence of invasive Hib disease among children in the United Kingdom from 1999 onward (13, 29, 30). In another recent report McVernon et al. (16) described the increase of Hib disease in patients aged 15 years and older after 1998 concurrent with the increase seen in fully vaccinated children. Although they mention that the reverse may occur as well, these authors theorized that, because transmission usually is from child to parent, the increase in the incidence of pediatric Hib disease has caused an increase in adults. However, this would suggest that the children still constitute the major reservoir for Hib infection and contradicts the observation that no carriage was detected in children and our observation that in the postvaccination period the genetic diversity of Hib isolates from children reached an even higher level than that of the Hib isolates from adults.
The observed abrupt increase in genetic diversity of Hib after the introduction of the vaccine may have been caused by unmasking highly susceptible individuals. Musser et al. showed that the bulk of invasive Hib disease is caused by a few clones or lineages but that there is abundant genetic variation (20, 21). Apparently, some clones are more successful in colonizing healthy young children compared to other, less abundant, Hib clones. This may account for a limited number of genotypes found in of cases of invasive disease. A minority of children may be highly susceptible to infection with Hib and even less successful, genetically diverse Hib clones may colonize these individuals and cause disease. In the bulk of the cases of invasive disease occurring prior to the nationwide Hib vaccination, disease caused by these minor types would go unnoted. However, when the circulation of the dominant Hib types and the number of cases of invasive disease was greatly reduced due to vaccination, disease in highly susceptible children may have continued to occur and eventually made up the majority of the Hib cases. As a result mainly highly susceptible children would get invasive disease, and disease would be caused by genetically diverse strains. Although this theory fits with the observed increased genetic diversity it does not explain why the total number of cases has been rising. However, if the highly susceptible children are infected mainly from the genetically diverse reservoir present in older people, an increase in Hib circulation in adults could result in an increase of Hib colonization and disease.
An increase in vaccine failures has only been reported in the United Kingdom and in The Netherlands. In the United Kingdom the lack of a booster vaccination or the decline in Hib colonization may have caused waning immunity and/or loss of herd immunity, leading to a reduced protection of children 1 to 4 years in age but not in the children under 1 year of age (15). Analysis of the increase seen in Dutch invasive Hib disease shows that, in contrast to the United Kingdom, there is no difference between the increases in the age groups of children under 1 year of age and in those 1 to 4 years of age (Fig. 5). In The Netherlands there has been an increase since 1999 in all age groups, particularly in persons older than 4 years. This indicates that the increases seen in both countries may have, at least in part, different causes. There have been two major differences between the Hib vaccinations in United Kingdom and The Netherlands. First, the Dutch Hib vaccination program includes a booster at 11 months of age which may prevent waning immunity among children. This could explain why children aged 1 to 4 years are protected as well as those younger than 1 year. It will be interesting to genotype Hib strains from the United Kingdom by MLVA to determine whether the increase in genetic diversity seen in the Dutch strains can also be found among the United Kingdom Hib strains. The second difference between the vaccination schedules of the two countries is that no acellular pertussis vaccine has been used in combination with the Hib vaccine in The Netherlands. However, an acellular pertussis vaccine will be introduced into the Dutch national vaccination program on January 2005 to replace the whole-cell vaccine, and this may result in increased incidence of invasive Hib disease. Ongoing surveillance will be required to monitor the effect of the change in the vaccine mixture on the incidence and manifestations of Hib disease, as well as on the composition of the Hib population.
REFERENCES
Barbour, M. L., R. T. Mayon-White, C. Coles, D. W. Crook, and E. R. Moxon. 1995. The impact of conjugate vaccine on carriage of Haemophilus influenzae type b. J. Infect. Dis. 171:93-98.
Dingle, K. E., F. M. Colles, D. R. A. Wareing, R. Ure, A. J. Fox, F. E. Bolton, H. J. Bootsma, R. J. L. Willems, R. Urwin, and M. C. Maiden. J. 2001. Multilocus sequence typing system for Campylobacter jejuni. J. Clin. Microbiol. 39:14-23.
Enright, M. C., N. P. J. Day, C. E. Davies, S. J. Peacock, and B. G. Spratt. 2000. Multilocus sequence typing for characterization of methicillin-resistant and methicillin-susceptible clones of Staphylococcus aureus. J. Clin. Microbiol. 38:1008-1015.
Enright, M. C., and B. G. Spratt. 1998. A multilocus sequence typing scheme for Streptococcus pneumoniae: identification of clones associated with serious invasive disease. Microbiology 144:3049-3060.
Farlow, J., D. Postic, K. L. Smith, Z. Jay, G. Baranton, and P. Keim. 2002. Strain typing of Borrelia burgdorferi, Borrelia afzelii, and Borrelia garinii by using multiple-locus variable-number tandem repeat analysis. J. Clin. Microbiol. 40:4612-4618.
Farlow, J., K. L. Smith, J. Wong, M. Abrams, M. Lytle, and P. Keim. 2001. Francisella tularensis strain typing using multiple-locus, variable-number tandem repeat analysis. J. Clin. Microbiol. 39:3186-3192.
Frothingham, R., and W. A. Meeker-O'Connell. 1998. Genetic diversity in the Mycobacterium tuberculosis complex based on variable numbers of tandem DNA repeats. Microbiology 144:1189-1196.
Heath, P. T., R. Booy, H. Griffiths, E. Clutterbuck, H. J. Azzopardi, M. P. Slack, J. Fogarty, A. C. Moloney, and E. R. Moxon. 2000. Clinical and immunological risk factors associated with Haemophilus influenzae type b conjugate vaccine failure in childhood. Clin. Infect. Dis. 31:973-980.
Hunter, P. R., and M. A. Gaston. 1988. Numerical index of the discriminatory ability of typing systems: an application of Simpson's index of diversity. J. Clin. Microbiol. 26:2465-2466.
Klevytska, A. M., L. B. Price, J. M. Schupp, P. L. Worsham, J. Wong, and P. Keim. 2001. Identification and characterization of variable-number tandem repeats in the Yersinia pestis genome. J. Clin. Microbiol. 39:3179-3185.
Liu, Y., M. A. Lee, E. E. Ooi, Y. Mavis, A. L. Tan, and H. H. Quek. 2003. Molecular typing of Salmonella enterica serovar Typhi isolates from various countries in Asia by a multiplex PCR assay on variable-number tandem repeats. J. Clin. Microbiol. 41:4388-4394.
Maiden, M. C. J., J. A. Bygraves, E. Feil, G. Morelli, J. E. Russell, R. Urwin, Q. Zhang, J. J. Zhou, K. Zurth, D. A. Caugant, I. M. Feavers, M. Achtman, and B. G. Spratt. 1998. Multilocus sequence typing: a portable approach to the identification of clones within populations of pathogenic microorganisms. Proc. Natl. Acad. Sci. USA 95:3140-3145.
McVernon, J., N. Andrews, M. P. Slack, and M. E. Ramsay. 2003. Risk of vaccine failure after Haemophilus influenzae type b (Hib) combination vaccines with acellular pertussis. Lancet 361:1521-1523.
McVernon, J., A. J. Howard, M. P. Slack, and M. E. Ramsay. 2004. Long-term impact of vaccination on Haemophilus influenzae type b (Hib) carriage in the United Kingdom. Epidemiol. Infect. 132:765-767.
McVernon, J., N. A. Mitchison, and E. R. Moxon. 2004. T helper cells and efficacy of Haemophilus influenzae type b conjugate vaccination. Lancet Infect. Dis. 4:40-43.
McVernon, J., C. L. Trotter, M. P. Slack, and M. E. Ramsay. 2004. Trends in Haemophilus influenzae type b infections in adults in England and Wales: surveillance study. Br. Med. J. 329:655-658.
Meats, E., E. J. Feil, S. Stringer, A. J. Cody, R. Goldstein, J. S. Kroll, T. Popovic, and B. G. Spratt. 2003. Characterization of encapsulated and noncapsulated Haemophilus influenzae and determination of phylogenetic relationships by multilocus sequence typing. J. Clin. Microbiol. 41:1623-1636.
Metropolis, N., and S. Ulam. 1949. The Monte Carlo method. J. Am. Stat. Assoc. 44:335-341.
Mohle-Boetani, J. C., G. Ajello, E. Breneman, K. A. Deaver, C. Harvey, B. D. Plikaytis, M. M. Farley, D. S. Stephens, and J. D. Wenger. 1993. Carriage of Haemophilus influenzae type b in children after widespread vaccination with conjugate Haemophilus influenzae type b vaccines. Pediatr. Infect. Dis. J. 12:589-593.
Musser, J. M., J. S. Kroll, D. M. Granoff, E. R. Moxon, B. R. Brodeur, J. Campos, H. Dabernat, W. Frederiksen, J. Hamel, G. Hammond, et al. 1990. Global genetic structure and molecular epidemiology of encapsulated Haemophilus influenzae. Rev. Infect. Dis. 12:75-111.
Musser, J. M., J. S. Kroll, E. R. Moxon, and R. K. Selander. 1988. Evolutionary genetics of the encapsulated strains of Haemophilus influenzae. Proc. Natl. Acad. Sci. USA 85:7758-7762.
Peltola, H. 2000. Worldwide Haemophilus influenzae type b disease at the beginning of the 21st century: global analysis of the disease burden 25 years after the use of the polysaccharide vaccine and a decade after the advent of conjugates. Clin. Microbiol. Rev. 13:302-317.
Ramsay, M. E., J. McVernon, N. J. Andrews, P. T. Heath, and M. P. Slack. 2003. Estimating Haemophilus influenzae type b vaccine effectiveness in England and Wales by use of the screening method. J. Infect. Dis. 188:481-485.
Rijkers, G. T., P. E. Vermeer-de Bondt, L. Spanjaard, M. A. Breukels, and E. A. Sanders. 2003. Return of Haemophilus influenzae type b infections. Lancet 361:1563-1564.
Schouls, L. M., van der Heide, H. G., L. Vauterin, P. Vauterin, and F. R. Mooi. 2004. Multiple-locus variable-number tandem repeat analysis of Dutch Bordetella pertussis strains reveals rapid genetic changes with clonal expansion during the late 1990s. J. Bacteriol. 186:5496-5505.
Simpson, E. H. 1949. Measurement of diversity. Nature 163:688.
Takala, A. K., J. Eskola, M. Leinonen, H. Kayhty, A. Nissinen, E. Pekkanen, and P. H. Makela. 1991. Reduction of oropharyngeal carriage of Haemophilus influenzae type b (Hib) in children immunized with an Hib conjugate vaccine. J. Infect. Dis. 164:982-986.
Takala, A. K., M. Santosham, J. Almeido-Hill, M. Wolff, W. Newcomer, R. Reid, H. Kayhty, E. Esko, and P. H. Makela. 1993. Vaccination with Haemophilus influenzae type b meningococcal protein conjugate vaccine reduces oropharyngeal carriage of Haemophilus influenzae type b among American Indian children. Pediatr. Infect. Dis. J. 12:593-599.
Trotter, C. L., J. McVernon, N. J. Andrews, M. Burrage, and M. E. Ramsay. 2003. Antibody to Haemophilus influenzae type b after routine and catch-up vaccination. Lancet 361:1523-1524.
Trotter, C. L., M. E. Ramsay, and M. P. Slack. 2003. Rising incidence of Haemophilus influenzae type b disease in England and Wales indicates a need for a second catch-up vaccination campaign. Commun. Dis. Public Health 6:55-58.
van Belkum, A., S. Scherer, W. van Leeuwen, D. Willemse, L. van Alphen, and H. Verbrugh. 1997. Variable number of tandem repeats in clinical strains of Haemophilus influenzae. Infect. Immun. 65:5017-5027.
Van der Ende, A., L. Spanjaard, and J. Dankert. 2003. Bacterial meningitis in The Netherlands: 31th annual report of The Netherlands Reference Laboratory for Bacterial Meningitis, p. 1-50. The Netherlands Reference Laboratory for Bacterial Meningitis, University of Amsterdam, Amsterdam, The Netherlands.
Van Loo, I. H., K. J. Heuvelman, A. J. King, and F. R. Mooi. 2002. Multilocus sequence typing of Bordetella pertussis based on surface protein genes. J. Clin. Microbiol. 40:1994-2001.
Zhang, W., B. M. Jayarao, and S. J. Knabel. 2004. Multi-virulence-locus sequence typing of Listeria monocytogenes. Appl. Environ. Microbiol. 70:913-920.(Leo M. Schouls, Arie van )