INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Haemophilus influenzae serotype b (Hib) has until recently been a leading cause of bacterial meningitis and other invasive bacterial infections in childhood worldwide; indeed, it is still the major cause of bacterial meningitis in Japan (10). The aim of the present study was to investigate the molecular epidemiology of clinically isolated Hib strains derived from pediatric meningitis patients in Japan. This was achieved by combining several molecular epidemiological methods including long PCR-ribotyping (18, 19), random amplified polymorphic DNA (RAPD) analysis (5), and restriction fragment length polymorphism (RFLP) analysis with the SmaI restriction enzyme and the pulsed-field gel electrophoresis (PFGE) apparatus (4, 6, 9).

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains. All Hib strains (30 strains) were isolated from the cerebrospinal fluid of children with meningitis, ranging from 3 months to 5 years and 3 months old (mean ± standard deviation, 2 years ± 16 months), between March 1992 and November 1995. Most of the strains were collected from the Kanto area (Tokyo Bay area). Of these, 19 were collected from Chiba Prefecture (four were collected from the Tokai area). All the bacterial strains isolated from the cerebrospinal fluid samples were confirmed as H. influenzae by conventional methods. They were biotyped by the methods described by Kilian (12), and capsular serotype b was confirmed with a commercial coagglutination kit (Phadebact Haemophilus test; Pharmacia & Upjohn Japan, Tokyo, Japan). The isolates were stored at 80°C, according to the procedures described by Gibson and Khoury (8), until required for use. The production of -lactamase was assayed with Cefinase discs (Becton Dickinson and Co., Cockeysville, Md.). Overnight cultures of the Hib isolates grown in brain heart infusion broth (Oxoid, Basingstoke, Hampshire, United Kingdom) supplemented with nicotinamide adenine dinucleotide (2.0 µg/ml) and hemin (10 µg/ml) at 37°C in 5% CO₂ were used to prepare DNA for the genotyping studies. Hib ATCC 10211 was used for PFGE-RFLP as a universal standard marker.

Long PCR-ribotyping and RAPD analysis. Genomic DNA for the PCR analyses was prepared with the SepaGene kit (Sanko Pharmaceutical Co., Tokyo, Japan) and quantified by photospectrometry. Long PCR-ribotyping was performed by the method described by Smith et al. (19). Briefly, 270 ng of each sample DNA was amplified with the 16S-H primer (5'-GGTATTGAGGAAGGTTGATGTGTTAATAGCACATC-3') and the 5S primer (5'-CATTACAGCGTTTCACTTCTGAGTTCGGTATGGTC-3') with a long-and-accurate (LA) PCR kit (TaKaRa LA PCR kit version 2; Takara Shuzo, Tokyo, Japan). The LA PCR-amplified products were digested with HaeIII (Takara Shuzo), electrophoresed in a 1.5% agarose gel with 1× TBE buffer (50 mM Tris-HCl, 50 mM boric acid, and 1 mM EDTA), and detected by ethidium bromide staining.

Chromosomal RFLP analysis by PFGE (PFGE-RFLP). Genomic DNA for the PFGE analysis was prepared by the procedure established by Hood (9) with some modifications. Disposable 100-µl scale plug molds (Bio-Rad) were used to prepare agarose plugs for each sample. Overnight cultures of the Hib isolates were resuspended at 5 × 10⁸ cells/ml in washing buffer A (100 mM NaCl, 20 mM Tris-HCl, 20 mM EDTA, pH 7.6). Fifty microliters of each cell suspension was used to prepare each plug. After treatment with proteinase K (2 mg/ml) overnight at 52°C, the sample plugs were incubated three times (for 30 min each time) with 2 ml of washing buffer B (50 mM Tris-HCl, pH 7.5) containing 2 mM phenylmethylsulfonyl fluoride at 50°C and then rinsed three times (for 30 min each time) with 2 ml of washing buffer B. One-eighth of each plug was sliced off and digested with 10 U of SmaI (Takara Shuzo). The gels were processed with the contour-clamped homogeneous electric field DR II PFGE apparatus (Bio-Rad) under the following electrophoresis conditions: 7 h at 170 V with an initial time of 5 s and a final time of 20 s, followed by 14 h at 170 V with an initial time of 5 s and a final time of 80 s. Electrophoresis was performed in 0.5× TBE buffer at 11°C. A 48.5-kb lambda DNA ladder (FMC BioProducts) was used as a molecular size marker. Hib ATCC 10211 DNA was used for PFGE-RFLP as a universal standard marker. After electrophoresis, the gels were stained with ethidium bromide and photographed under UV light at 302 nm.

Dendrograms. The genomic DNA fingerprinting patterns produced by PFGE-RFLP were analyzed with Molecular Analyst Fingerprinting Plus software, version 1.12 (Bio-Rad) on a Microsoft Windows 95 operating system to generate dendrograms (17). The dendrogram was constructed with the unweighted pair group method with averages clustering algorithm with the Jaccard coefficient (S_J) with band positions.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Phenotypes. Table 1 shows the profiles of the clinically isolated Hib strains, including the locations of the hospitals at which they were isolated, the ages of the patients at the onset of Hib meningitis, RAPD data, and the results of the -lactamase production tests. Among the 30 Hib strains, 22 (73.3%) were of biotype I, 6 (20.0%) were of biotype II, and 2 (6.7%) were of biotype IV. There were no biotype III, V, VI, or VII strains. Despite the fact that only four strains were isolated from the Tokai area, two of the six biotype II Hib strains were isolated from this area. Figure 1 shows the locations of the hospitals at which the different strains were isolated. Fifteen strains were -lactamase producers.

View larger version (34K): [in this window] [in a new window]   FIG. 1.   Geographic location of a case study area. The geographic distribution pattern of 30 pediatric meningitis patients in this study is plotted on maps of Japan and Chiba Prefecture.

Long PCR-ribotyping and RAPD analysis of Hib strains. The long PCR-ribotyping method, which was originally developed to identify nontypeable H. influenzae, showed no RFLPs among the Hib isolates (all restriction patterns after HaeIII digestion were the same [data not shown]). RAPD analysis showed a number of unique bands characteristic for a particular biotype with both primer sets. When primer A was used (Fig. 2A), the RAPD pattern exhibited a 600-bp band with biotype I or IV strains, while when primer B was used (Fig. 2B), the RAPD pattern included a 380-bp band specific for biotype II strains. Strain 9, isolated from Toyohashi City, showed a unique band pattern compared with the other strains with biotype I; however, this strain also fit the 600- and 380-bp band rule described above.

View larger version (91K): [in this window] [in a new window]   FIG. 2.   RAPD analysis of 30 Hib isolates from pediatric meningitis patients. M, molecular size marker (100-bp DNA ladder). Lane numbers correspond to the strain numbers in Table 1. (A) RAPD analysis with primer A, 5'-TGCCCGTCGT-3'; (B) RAPD analysis with primer B, 5'-GTAGACCCGT-3'.

Chromosomal RFLP analysis by PFGE (PFGE-RFLP). The chromosomal RFLP patterns obtained after SmaI digestion of the 30 Hib strains isolated are shown in Fig. 3. The computer-generated graphic pattern and dendrogram also are shown in Fig. 3. Five biotype II strains (no. 5, H93029; no. 6, H93130; no. 15, H94014; no. 22, H94207; and no. 24, H95048, which all exhibited a 380-bp band in their RAPD pattern as shown in Fig. 2B), no. 9 (H93175), and no. 14 (H93225) formed a cluster which includes three of the strains from the Tokai area. All of the strains isolated from Kisarazu County in Chiba Prefecture, which were biotype I strains (no. 7, H93152; no. 10, H93184; no. 13, H93223; no. 20, H94084), were included in the other large cluster.

View larger version (26K): [in this window] [in a new window]   FIG. 3.   Computer-generated graphic patterns and dendrogram by PFGE-RFLP analysis of 30 Hib isolates from pediatric meningitis patients. Strain numbers with isolate locations in Table 1 and H. influenzae ATCC 10211 are shown. A molecular size scale (upper right) and a band-based Jaccard similarity coefficient (SJ [%]) scale (upper left) are also shown.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

There is a difference in the incidence of invasive Hib infectious diseases between the United States and Japan. Uehara et al. reported that the average annual incidence of H. influenzae meningitis per 100,000 children less than 5 years of age in Japan in 1994 was 4.0 cases (20), a much lower incidence than that of 24.0/100,000 reported by the National Bacterial Meningitis Reporting System before the start of the Hib vaccination program in the United States (1). The incidence of Hib meningitis in Japan has been gradually increasing in recent years, and this may pose a future risk of outbreaks of invasive Hib disease. Since Hib vaccines are not currently available in Japan, the difference in the incidence of Hib meningitis between Japan and the United States prior to the use of Hib vaccines may be explained by the empiric therapy used for upper respiratory infections in Japanese children, which is based on prophylactic antibiotic usage. It has been recognized that the control of invasive diseases caused by Hib requires routine immunization. We are as yet unable to control invasive Hib disease well, despite early diagnosis and appropriate treatment, because of the lack of a vaccination program. Introduction of the Hib conjugate vaccine to Japan is therefore necessary. Of the Hib meningitis strains isolated during this study, 73% were of biotype I. The Hib strains isolated from healthy carriers were mainly of biotypes I, II, and IV. Biotype I strains were the most frequently isolated, reflecting the biotype distribution of the strains of Hib responsible for invasive diseases. In Japan, more than half of the invasive strains of Hib are of biotype I (16).

In this study, PFGE-RFLP analysis of Hib DNA after digestion with SmaI demonstrated several genotype groups. This technique may therefore be useful in clarifying the strain(s) endemic in certain regions. Epidemiological studies of Hib infections commonly use phenotyping (serotyping, biotyping, and analysis of the major outer membrane protein patterns) (3, 15) or ribotyping methods (14, 18, 19). Recently, epidemiological studies of Hib strains isolated from vaccinated patients have been performed. Falla et al. reported that there was no difference in the distribution of cap b genotypes [i.e., b(S), b(G), b(V), or b(O)] between Hib strains causing disease in Hib vaccine recipients and those isolated from nonrecipients (7), suggesting that the cap b gene subtypes do not explain the virulence of Hib strains. van Alphen et al. reported a reduction in the incidence of disease caused by all clonal groups of Hib associated with widespread infant vaccination in Finland (21).

PFGE analysis is currently considered the most reliable and practical tool for molecular epidemiologic analyses of bacterial infections (2, 4, 6, 9, 11). As demonstrated here, additional RAPD analysis can increase the accuracy of these analyses. PCR-based RAPD analysis is increasingly being used for molecular epidemiologic applications, such as the subtyping of Escherichia coli isolates (5). Although RAPD can provide RFLP data very rapidly, the resolution and reproducibility of these data are somewhat limited. Nevertheless, RAPD can yield valuable preliminary molecular epidemiologic information during bacterial outbreaks while more time-consuming PFGE analyses are performed. The major limitations of PFGE analysis include requirements for technical skill and expensive apparatus and the time taken to complete the analysis (4 to 7 days). Kits for performing amplified fragment length polymorphism (Keygene, N.V., Wageningen, The Netherlands) recently have been developed and are now commercially available (22), and this will improve the resolutions for molecular epidemiological analysis.

Another area of interest in the study of Hib invasive diseases is the existence and contribution of b H. influenzae strains. Loss of capsular b expression has frequently been demonstrated under both in vivo and in vitro conditions. Identification of b strain carriers among healthy children is therefore important, since b H. influenzae strains or non-b-serotyped H. influenzae strains may be capable of gaining virulence by natural DNA transfer between Haemophilus strains (13). More attention must be paid to genetic studies of Hib diseases caused by b and non-b-serotyped H. influenzae strains after vaccination to explain and reduce vaccine failures. We are currently planning to expand our Hib PFGE-RFLP study both nationwide and within smaller communities to allow better understanding of the routes of transmission of Hib and b and non-b-serotype H. influenzae infections.