INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Meningococcal disease occurs in the form of sporadic cases, local outbreaks, and epidemics and has a high mortality. Much attention has been paid to methods for typing meningococcal strains in order to trace and control outbreaks and elucidate the global epidemiology of the bacterium, and this has generated a considerable variety of typing methods. The "gold standard" method is isoenzyme electrophoresis (ET typing) (16). This method, which compares electrophoretic polymorphisms in multiple enzymes, defines epidemic clones of Neisseria meningitidis, provides an estimate of genetic similarity, and has provided the basis for inference of the genetic population structure of N. meningitidis. In the clinical setting, a typing system based on serogrouping (capsular polysaccharides), serotyping and subtyping (class 1, 2, and 3 outer membrane proteins), and sulfonamide resistance is more frequently used (6). Although this method may be a useful guide to the clonal affinities of meningococcal strains when applied locally, on a global basis serogroup and, to a lesser extent, serotype are poorly correlated with genetic relatedness (3). More recently, genetic methods, such as whole-genome DNA fingerprinting (12, 18), random amplification of polymorphic DNA (21), repetitive element-based PCR (20), and restriction fragment length polymorphism (RFLP) analysis of PCR products (7, 10), have been successfully applied in epidemiological studies of N. meningitidis. Whole-genome DNA fingerprinting, though useful for tracing outbreak strains, is unsuitable for classification because the restriction pattern is too complex, but other DNA-based methods that generate simpler patterns are more promising.

We have previously described a method (PCR amplicon restriction endonuclease analysis [PCR-AREA]) for PCR-based RFLP analysis of the highly polymorphic chromosomal dhps gene, which determines resistance or sensitivity to sulfonamides in N. meningitidis (10). To evaluate PCR-AREA as a classification method, and in order to study the clonal distribution of meningococcal strains in the Norwegian county of Telemark, we have used PCR-AREA, in combination with DNA fingerprinting and serogroup, serotype and subtype, and sulfonamide resistance determinations, to compare 42 strains isolated from patients with meningococcal disease. This represents all primary cases of meningococcal disease in Telemark from January 1987 to March 1995.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Samples. Meningococci were isolated from cerebrospinal fluid (CSF) (16 cases), blood cultures (23 cases), or throat and nasal swabs (3 cases) of 42 patients with invasive meningococcal disease. For the three patients where meningococci were isolated from throat samples, the following criteria led to the diagnosis of invasive meningococcal disease. Patient 1 had petechial bleeding on the arms, legs, back, and face, fever, a high sedimentation rate and C-reactive protein level, normal blood pressure and peripheral circulation, no neck stiffness, and a diagnosis of benign meningococcemia. Patient 5 was admitted to the hospital with symptoms of fulminant meningococcal septicemia (unconsciousness, petechiae, and low blood pressure) and was diagnosed with fulminant meningococcal septicemia. Patient 25 had fever, petechial bleeding, vomiting, poor peripheral circulation, a stiff neck, high C-reactive protein, 6800 cells/ml in the CSF, and a diagnosis of meningococcal meningitis and septicemia.

Extraction of DNA for PCR and chromosomal DNA fingerprinting. DNA was purified by lysozyme-EDTA-Triton X-100 lysis and phenol-chloroform extraction as previously described (12).

PCR. A 634-bp segment of the chromosomal dhps gene was amplified with the primers NM6 (CGC CAT CAA TTC GGG CAA ATG; nucleotides 711 to 734) and NM7 (TTG GCA GGC AGG GTT TGA; nucleotides 1324 to 1344) (4). The 100-µl PCR mixtures contained 200 ng of purified N. meningitidis DNA, 500 ng of each primer, 1.5 U of Taq polymerase (Promega, Madison, Wis.), 50 mM KCl, 10 mM Tris-HCl (pH 9.0), 0.1% Triton X-100, 0.2 mM deoxynucleoside triphosphates, and 1.5 mM MgCl₂. Reactions were run on an Omnigene thermal cycler (Hybaid, Teddington, United Kingdom) with the following program: 94°C, 60 s; 55°C, 60 s; 72°C, 60 s; 40 cycles.

PCR-AREA. PCR product (20 µl) was digested with 30 U of CfoI (Promega, Madison, Wis.) in a volume of 25 µl for 1 h at 37°C. The restriction fragments were separated by electrophoresis for 5 h at 400 V on an 8% polyacrylamide gel and visualized by ethidium bromide staining and UV transillumination (10).

DNA fingerprinting. Purified N. meningitidis DNA (25 µg) was digested with HindIII, and the fragments were separated on 4% polyacrylamide gels and visualized by ethidium bromide staining and UV transillumination as previously described (12).

Serotyping. The serotype and subtype were determined by dot blot analysis as described by Wedege et al. (19), except that signal detection was done with 3-amino-9-ethyl carbazole and hydrogen peroxide and the strips were photocopied after development.

Serogrouping. Serogrouping was performed by slide agglutination with N. meningitidis agglutination sera for serogroups A, B, C, Y, and W135 from Murex Diagnostics (Dartford, United Kingdom) according to the manufacturer's specifications.

Sulfonamide resistance determination. The MIC of sulfonamide was determined by an agar dilution method. A suspension of one or two freshly grown N. meningitidis colonies was adjusted to a turbidity of 0.5 McFarland units by visual reference to a standard. Twenty-five-microliter samples of the suspension were innoculated onto agar plates containing 2× serial dilutions of sulfathiazole from 2,000 to 1 mg/liter in NM medium. The plates were incubated for 24 h at 37°C in a 10% CO₂ atmosphere. NM medium is an enriched variant of HTM medium (8) prepared as follows: 38 g of Mueller-Hinton medium and 5 g of yeast extract (Mast, Merseyside, United Kingdom) were dissolved in 1 liter of distilled water and autoclaved at 121°C. The medium was cooled to 48°C, and 10 ml of 27% glucose, 15 ml of 0.1% hematin, and 1 ml of solution E4 [0.2% cocarboxylase (Sigma catalog no., C-8754), 0.4% NAD, 10% L-glutamine, 26% L-cysteine, and 0.5% Fe(NO₃)₃ · 9H₂O] were added. After the addition of sulfathiazole, the medium was dispensed into petri dishes in 25-ml aliquots. Isolates were classified as resistant, intermediate, or sensitive according to the following system: isolates for which the MICs were 16 mg/liter were considered sensitive; isolates for which the MICs were 31 but 62 mg/liter were considered intermediate; and isolates for which the MICs were 250 mg/liter were considered resistant.

Statistical methods. Statistical analyses were performed with the chi-square test or with Fisher's exact test where small numbers precluded use of the chi-square test.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Clonal analysis by PCR-AREA and DNA fingerprinting. PCR-AREA of the 42 isolates revealed 11 different CfoI restriction patterns (classes 1 to 11) for the 634-bp NM6-NM7 PCR product (Fig. 1). Within gels, pattern comparison could be achieved quickly and unambiguously (Fig. 2). A tentative class assignment could be achieved by comparing gels, but final assignments were always made by comparison of lanes on a single gel. Fifteen isolates belonged to class 1, eight isolates belonged to class 2, six isolates belonged to class 3, three isolates belonged to class 6, and two isolates belonged to each of classes 4, 5, and 8. Classes 7, 9, 10, and 11 contained only one isolate each.

View larger version (51K): [in this window] [in a new window]   FIG. 1.   PCR-AREA of N. meningitidis isolates from Telemark showing the 11 different CfoI digestion patterns. Lane numbers correspond to the class number. The size standards (Std) are U.S. Biochemicals PCR molecular size markers, with fragment sizes of 1,000, 700, 525, 500, 400, 300, 200, 100, and 50 bp.

View larger version (76K): [in this window] [in a new window]   FIG. 2.   Typical PCR-AREA gel showing classes 1, 2, 4, and 5. The lanes are numbered according to the class. The standard (St) is a mixture of PCR products and restriction fragments. Fragment sizes are 645, 217, 100, 83, 74, and 62 bp.

Phenotypic analyses. Isolates were further classified by serogroup, serotype, and sulfonamide resistance. Twenty-eight isolates belonged to serogroup B, 11 isolates belonged to serogroup C, 2 isolates belonged to serogroup Y, and 1 isolate belonged to serogroup W135. The serogroups, serotypes, and subtypes are summarized in Table 2. Twenty-six isolates were sulfonamide resistant, nine isolates were sulfonamide sensitive, and seven isolates had an intermediate level of resistance. There were several clear correlations of phenotypic characteristics. All serogroup B isolates that were either serotype 15 or subtype P1.7,16 (14 isolates) were also sulfonamide resistant. All six B:16 isolates were sulfonamide sensitive. Ten of 11 serogroup C isolates were serotype 2a or 2b, subtype P1.2,5, and sulfonamide resistant.

Association between indistinguishable isolates. Five groups of two to three indistinguishable isolates could be identified (Table 4). They are designated by their PCR-AREA classes and DNA fingerprint patterns. These isolates had identical serogroups, serotypes, serosubtypes, and sulfonamide resistances, with the exception of isolate 22, which differed from other isolates in group 3-c1 in being nonsubtypeable (Table 1). There was a maximum of 21 months between cases of disease caused by indistinguishable isolates. We are not aware of any social connection between these patients, except in the case of isolates 35-1 and 35-2 (3-c2), which were from siblings who both fell sick in the space of 24 h. However, the cases of disease caused by strains of group 2-b1 and 2-b6 were sufficiently close in time and space that a connection is to be suspected.

Correlation of PCR-AREA class with sulfonamide MIC. Sulfonamide resistance in N. meningitidis is caused by alterations of the chromosomal dhps gene (15), which is the target for PCR-AREA. The PCR-AREA pattern is therefore expected to correlate with sulfonamide resistance. There is a nearly exact correlation between the PCR-AREA patterns and the sulfonamide MICs for the 31 isolates of classes 1, 2, 3, and 8; these strains are closely related, with a generally high degree of phenotypic homogeneity. When the genetically heterogeneous classes 4, 5, and 6 are examined, we see that only class 4 is homogeneous with respect to the MIC of sulfonamide. Thus, similar PCR-AREA patterns in unrelated strains do not predict sulfonamide MICs with any degree of exactitude (Table 1).

Temporal distribution of clones. Figure 3 shows the occurrence of isolates belonging to the three major PCR-AREA classes during the 8 years of the study. Class 1 was dominant in 1987 and since then has made a moderate but persistent contribution to cases of meningococcal disease in Telemark. Class 2 was first observed in 1988 and persisted until 1992, after which it disappeared (a single class 2 case has since been observed, in January 1997 [data not shown]). Class 3 was the most temporally confined clone, with all cases occurring in 1991 and 1992. 

View larger version (19K): [in this window] [in a new window]   FIG. 3.   Temporal distribution of PCR-AREA classes from 1987 to 1995.

Geographical distribution of clones. The population of Telemark (163,000) is concentrated in the coastal urban-industrial area of Grenland (population, 89,000). The remainder of the county is predominantly rural and thinly populated, with a population of 74,000. Thirty-one of the 42 cases in this study were associated with the Grenland area (the incidence for the study period was 35 per 100,000), while 13 cases were associated with the rural part of the county (the incidence for the study period was 17.5 per 100,000). Two cases had affinities to both rural and industrial parts of the county.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Usefulness of PCR-AREA. PCR-AREA was developed in order to address the limitations of DNA fingerprinting (12) in classification. The complex band pattern generated by DNA fingerprinting allows precise identification of the strain present in a patient among strains isolated from contacts, but exhaustive cross-comparison of patterns in more than a handful of strains is unachievable. The PCR-AREA restriction pattern of 8 to 15 CfoI (HhaI) fragments is simpler and should allow exhaustive cross-comparison of strains. However, the reduced information density of PCR-AREA patterns is expected to reduce its precision relative to DNA fingerprinting and to give rise to more "false calls" of strain identity.

PCR-AREA pattern and sulfonamide resistance. PCR-AREA targets the meningococcal chromosomal dhps gene, which determines sulfonamide resistance in N. meningitidis. In the strains studied here there is a high degree of correlation between the PCR-AREA pattern and the level of sulfonamide resistance. This is, however, primarily due to the fact that the collection is dominated by clonal isolates with high degrees of phenotypic and genotypic similarity. In the nonclonal classes 4, 5, and 6, the sulfonamide MICs correlate poorly with the PCR-AREA patterns. This may be because a single PCR-AREA pattern can encompass both resistant and sensitive dhps alleles or because other genetic loci may play a role in determining the MIC of sulfonamide.

Clonal structure of strains from Telemark. One of the most striking, and at first sight puzzling, aspects of our results is that while homogeneous (clonal) PCR-AREA classes (classes 1, 2, 3, and 8) are typically large (15, 8, 6, and 2 members, respectively), heterogeneous (nonclonal) classes (classes 4, 5, and 6) are small (2, 2, and 3 isolates, respectively). This observation may be accounted for by the proposed genetic structure of meningococcal populations (13, 17). Meningococcal populations have a basically panmictic structure, where free exchange of genetic material between strains results in a very low degree of linkage disequilibriumthat is to say, genetic markers associate randomly. Upon this panmictic structure is imposed an element of clonality caused by the spread of epidemic strains. The spread of such strains exceeds the rate of genetic recombination, so that the strains are genetically homogeneous. Our results conform well to this model. Members of classes 1, 2, 3, and 8 are epidemic strains, belonging to the clonal part of the meningococcal population, and are genetically homogeneous and numerous. Members of classes 4, 5, and 6 are nonepidemic strains belonging to the panmictic population; they have similarity in the dhps gene, due presumably to transformational spread of dhps alleles through the meningococcal population, but are otherwise dissimilar. The small sizes of these classes reflect the great variety of dhps sequences present in the meningococcal population (4).

Geographical distribution of clones. The geographical distributions of classes 1 (B:15; ET5-like), 2 (C: 2a; ET37-like), and 3 (B:16) differ significantly. Classes 2 and 3 are confined to cases associated with the Grenland urban district. Class 3 appears to have arisen in this district and to have subsequently died out without spreading to other parts of the county. Class 2 (ET37-like) has been endemic in Norway for more than 10 years, so its confinement to the Grenland district is difficult to explain. Class 1 resembles ET5, which has been present in Norway since the 1970s and should thus have had ample time to achieve a homogeneous distribution, so its predilection for the rural areas of Telemark is similarly difficult to account for. Environmental factors or patchy distribution and a tendency for the social networks that facilitate transmission to be exclusively rural or urban are possible explanations.

Indistinguishable isolates. Five small groups of two to three indistinguishable isolates were identified in this study. Such isolates are probably epidemiologically related. Although cases of disease caused by indistinguishable isolates occurred within 2 years of each other, we found no noticeable geographical clustering, nor are we aware of any links between cases, with the single exception of two coprimary cases occurring in the same family. Although it is likely that a thorough investigation would reveal chains of common contacts between patients with disease caused by identical isolates, such an investigation is undesirable because of the burden of guilt which might be placed on these contacts. Cases caused by indistinguishable isolates represent failures of the objective of the Telemark Meningococcal Project (9, 11), which is to prevent secondary cases by directed chemoprophylaxis given to carriers of the disease-causing strain among a patient's contacts identified by DNA fingerprinting. Such failures probably arise because of difficulties in identifying and contacting all relevant contacts.