INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The Prevotella intermedia group bacteria are black-pigmented, gram-negative anaerobic rods that are commonly found as members of the polymicrobial flora in various infections in humans (8, 24). The P. intermedia group includes two phenotypically indistinguishable species, P. intermedia and Prevotella nigrescens (29). Moreover, several authors have found isolates that biochemically resemble P. intermedia and P. nigrescens (4, 20, 24, 28). These include a recently described, weakly pigmented, lipase-negative species, Prevotella pallens (21). The primary site of isolation of P. intermedia, P. nigrescens, and P. pallens is the oral cavity. P. intermedia is associated with periodontal disease (9, 22), whereas P. nigrescens and P. pallens can also be detected in samples from periodontally healthy subjects (7, 9, 20, 22). In addition to occurring in the oral cavity, all three species occur in extraoral infections (6, 20, 24).

Although several antibiotics, such as metronidazole, azithromycin, and -lactam antibiotics combined with -lactamase inhibitors, are still generally active against pigmented Prevotella species (35), resistance to tetracyclines and penicillins is increasing (32). Penicillin resistance due to -lactamase production (32) is presently common in pigmented Prevotella species (3, 15, 17, 19, 31). The various frequencies of -lactamase production of P. intermedia group isolates (3, 19) may be explained by differences in the geographic locations and sampling sites of the isolates. Knowledge about the frequency of -lactamase production as well as the susceptibility patterns of P. intermedia, P. nigrescens, and P. pallens is limited, since in most previous studies species differentiation was not performed or only a limited number of isolates were tested. Furthermore, there are no reports on the genetic heterogeneity of -lactamase-producing isolates.

The aims of the present study were to investigate the -lactamase production of P. intermedia, P. nigrescens, and P. pallens isolates and their susceptibilities to six antimicrobial agents and to compare the frequencies of -lactamase production by P. intermedia and P. nigrescens isolates from oral samples and various extraoral infections originating from two geographic locations between 1985 and 1995. In addition, the clonality of -lactamase-positive and -negative isolates and the clustering of genotypes were studied by using arbitrarily primed PCR (AP-PCR) for DNA fingerprinting.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial isolates. The material comprised 73 P. intermedia isolates, 84 P. nigrescens isolates, and 14 P. pallens clinical isolates included in our previous studies (20, 22-24). A total of 171 isolates were recovered from oral and extraoral samples obtained from 164 subjects between 1985 and 1995. The isolates originated from two geographic locations: Wadsworth Anaerobic Bacteriology Laboratory of the Veterans Affairs Medical Center, Los Angeles, Calif., and the Anaerobe Reference Laboratory, National Public Health Institute, and the Institute of Dentistry, University of Helsinki, both in Helsinki, Finland (Table 1).

-Lactamase test. -Lactamase production was assessed by using a chromogenic cephalosporin disk test (Biodisk, Solna, Sweden). Bacteroides fragilis ATCC 25285 was included as a positive control. The ² test was used to determine the statistical significance of the differences in the frequency of -lactamase production between the isolates originating from different sampling sites or geographic locations and at different time periods.

Susceptibility testing. The MICs of penicillin G, amoxicillin-clavulanate (2:1), metronidazole, cefoxitin, azithromycin, and trovafloxacin were determined by the Wadsworth agar dilution method according to procedures outlined by the National Committee for Clinical Laboratory Standards (NCCLS) (25). An inoculum of 10⁵ CFU was delivered with a multipoint inoculator to supplemented brucella-based, laked blood agar (BBL, Cockeysville, Md.). The plates were incubated at 37°C for 48 h in jars filled with a mixed gas (85% N₂, 5% CO₂, and 10% H₂). The appearance of growth was compared with that of the control plate, and the MIC for each was defined as the lowest concentration of antimicrobial agent resulting in no growth, one discrete colony, or multiple tiny colonies. The MICs were determined by one reader (J.M.), and the reference strains B. fragilis ATCC 25285, Bacteroides thetaiotaomicron ATCC 29741, P. intermedia ATCC 25611, and P. nigrescens ATCC 33563 were included as controls in each test run.

AP-PCR amplification. Clonal analysis of the P. intermedia, P. nigrescens, and P. pallens isolates (171 clinical isolates and the type strains P. intermedia ATCC 25611 and P. nigrescens ATCC 33563) was performed by using AP-PCR with the primer OPA-03 (5'-AGTCAGCCAC-3'; Operon Technologies, Alameda, Calif.), which is highly discriminative for these species (23). DNA extraction and PCR amplification were performed as previously described (23).

AP-PCR fingerprinting analysis. The amplified DNA fragments were separated in a 1% agarose gel containing 0.5 µg of ethidium bromide per ml and 0.5× Tris-borate-EDTA buffer at 100 V for 4 h. A 1-kb ladder (Gibco BRL, Life Technologies, Inc., Gaithersburg, Md.) was used in the outermost lanes of each gel as a molecular size marker and as a reference for normalization of different gels. The gel pictures were digitized and analyzed with the Taxotron program package (Taxolab, Institut Pasteur, Paris, France). The error margin was set to 5%, and clustering was performed by using the unweighted pair-group method with mathematical averages included in the package. Only the bands with medium to high intensity were included in the computer analysis, since the weak bands were not always reproducible.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Reference strains. The control strains B. fragilis ATCC 25285 and B. thetaiotaomicron ATCC 29741 were both -lactamase positive, and the MICs were within acceptable ranges (25), as follows: 16 to 32 and 16 to 32 µg/ml for penicillin G, 0.5 and 0.5 to 1 µg/ml for amoxicillin-clavulanate, 8 and 32 µg/ml for cefoxitin, 0.5 and 1 to 2 µg/ml for metronidazole, 4 to 8 and 8 to 16 µg/ml for azithromycin, and 0.25 and 0.5 µg/ml for trovafloxacin, respectively. The reference strains P. intermedia ATCC 25611 and P. nigrescens ATCC 33563 were -lactamase negative, and the MICs varied within two dilutions between the test runs, 0.031 and 0.016 µg/ml for penicillin G, 0.031 to 0.06 and 0.016 µg/ml for amoxicillin-clavulanate, 0.125 to 0.25 and 0.06 to 0.125 µg/ml for cefoxitin, 1 to 2 and 0.25 to 0.5 µg/ml for metronidazole, 0.125 to 0.25 and 0.125 µg/ml for azithromycin, and 1 and 1 µg/ml for trovafloxacin, respectively.

-Lactamase production. -Lactamase activity was detected in 21 of 73 (28.8%) P. intermedia isolates in 24 of 84 (28.6%) P. nigrescens isolates, and in 8 of 14 (57.1%) P. pallens isolates (Table 2). No difference in the frequencies of -lactamase production by P. intermedia and P. nigrescens was detected between isolates from extraoral sites and those from oral sites or between P. intermedia isolates originating from the United States and those originating from Finland. However, -lactamase production was detected more frequently in P. nigrescens isolates recovered from the United States than in those recovered from Finland (² test, P = 0.026). P. intermedia isolates obtained from 1991 to 1995 were significantly more often -lactamase positive than those isolated from 1985 to 1990 (² test, P = 0.038), whereas no time-related difference was observed for P. nigrescens. All P. pallens isolates were recovered from the oral cavities of Finnish subjects during the 1990s, and they were not included in the comparisons of the source and time of recovery.

View this table: [in this window] [in a new window]   TABLE 2.   Frequency of -lactamase production in P. intermedia, P. nigrescens, and P. pallens isolates, presented by site of infection or recovery, geographic location, and time of isolation

Susceptibility. The MICs of penicillin G were in the range of 0.016 to 16 µg/ml for P. intermedia and P. nigrescens and in the range of 0.06 to 32 µg/ml for P. pallens (MICs at which 90% of the isolates were inhibited [MIC₉₀s], 8, 8, and 16 µg/ml, respectively) (Table 3). According to the interpretive categories of the NCCLS (25), 73% of P. intermedia isolates, 76% of P. nigrescens isolates, and 43% of P. pallens isolates were susceptible (MIC, 0.5 µg/ml) to penicillin G; 1, 4, and 0%, respectively, were intermediate (MIC, 1 µg/ml) to penicillin G; and 26, 20, and 57%, respectively, were resistant (MIC, 2 µg/ml) to penicillin G. The isolates of each species could be separated into two categories according to the MICs and -lactamase production. One category included -lactamase-positive isolates for which the MICs were higher (0.5 µg/ml for P. intermedia and P. nigrescens and 2 µg/ml for P. pallens) and the other one included -lactamase-negative isolates for which the MICs were lower (0.06 µg/ml for P. intermedia, 0.03 µg/ml for P. nigrescens, and 0.125 µg/ml for P. pallens). The MIC breakpoint separating -lactamase-positive and -lactamase-negative isolates was 0.5 µg/ml. Penicillin G MIC₉₀s for -lactamase-positive isolates (16, 8, and 32 µg/ml) were 8 to 10 dilution steps higher than those for -lactamase-negative isolates (0.031, 0.031, and 0.125 µg/ml) (Table 4).

View this table: [in this window] [in a new window]   TABLE 4.   MIC50s, MIC90s, and MIC ranges of selected -lactam antibiotics for -lactamase-positive and -lactamase-negative P. intermedia, P. nigrescens, and P. pallens isolates

AP-PCR. The 21 -lactamase-positive P. intermedia isolates were of 20 AP-PCR types, the 24 -lactamase-positive P. nigrescens isolates were of 23 AP-PCR types, and the 8 -lactamase-positive P. pallens isolates were of 7 AP-PCR types. The numbers of AP-PCR types identified among the 54 -lactamase-negative P. intermedia isolates, the 60 -lactamase-negative P. nigrescens isolates, and the 6 -lactamase-negative P. pallens isolates were 51, 51, and 3, respectively. Figure 1 shows a dendogram of AP-PCR patterns of P. intermedia, P. nigrescens, and P. pallens displayed similar clustering (trees not shown). In two cases, -lactamase-positive and -lactamase-negative P. nigrescens isolates were of the same AP-PCR type. The P. intermedia and P. pallens isolates having the same AP-PCR type were always similar in -lactamase tests. The -lactamase-positive isolates of all three species were distributed into several clusters, and no correlation between the AP-PCR patterns and -lactamase production was observed. In addition, no correlation between the site of isolation or geographic location of the isolates and their AP-PCR patterns was found. The isolates having the same AP-PCR type were recovered from unrelated subjects, except for two P. intermedia isolates from two sites from a single subject, two P. nigrescens isolates from a married couple, and four P. pallens isolates from two mother-child pairs. The AP-PCR types of P. intermedia, P. nigrescens, and P. pallens displayed no overlapping.

View larger version (28K): [in this window] [in a new window]   FIG. 1.   Dendogram and list indicating -lactamase (-lact) production of the P. intermedia (P.i.) isolates (73 clinical isolates and the type strain, ATCC 25611). Extraoral isolates are underlined. The WAL isolates originated in the United States, and all other isolates originated in Finland. The -lactamase-positive isolates were scattered in several clusters, and characteristics shared by isolates belonging to the same cluster could not be found.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

-Lactamase production was a common feature of P. intermedia, P. nigrescens, and P. pallens isolates in the present study. The prevalence of -lactamase production in the P. intermedia group, including P. nigrescens (29%), is well in accordance with several earlier studies (4, 5, 15, 31). However, there are also studies in which higher (3) or lower (19) prevalences have been reported. The discrepancies in the frequencies may result from differences in the sources of the isolates and in the histories of antibiotic therapy of the subjects. Moreover, in most earlier studies P. intermedia and P. nigrescens were not identified to species level and they may have been unevenly distributed in the material. In the present study -lactamase-producing isolates were found with equal frequencies in P. intermedia and P. nigrescens, which disagrees with a recent study by Bernal et al. (4), in which -lactamase production was suggested to be a feature more common in P. nigrescens than in P. intermedia. However, in the study by Bernal et al. (4), only the -lactamase-positive isolates were identified to species level. The P. pallens isolates used for the present study were relatively more often -lactamase positive than the P. intermedia and P. nigrescens isolates, which agrees well with an earlier study by Könönen et al. (19), in which, contrary to P. pallens (formerly "P. intermedia/nigrescens-like organism"), most P. nigrescens isolates obtained from young children were -lactamase negative. P. intermedia was not included in the study by Könönen et al. (19), since the bacterium is not usually found in periodontally healthy children or adults (7, 9, 22). In the present study, no difference in the frequency of -lactamase production between isolates from extraoral and oral sites or between the P. intermedia isolates from different geographical locations was detected, whereas -lactamase-producing P. nigrescens isolates were detected statistically more often among isolates from U.S. subjects than among those from Finnish subjects. Most of the isolates from U.S. subjects were recovered from extraoral infections, whereas the isolates from the Finnish subjects originated from both healthy and diseased sites. Unfortunately, the antibiotic history of the subjects was not available, and it is possible that there is a difference in the use of antibiotics between these subject groups. Recent penicillin exposure is known to increase the prevalence of penicillin-resistant microbial populations (17). An increase in the penicillin resistance of pigmented Prevotella over the past 10 to 15 years has been reported (32), and an increase in the frequency of -lactamase production was seen in the present study among P. intermedia isolates but not among P. nigrescens isolates.

Penicillin-resistant, pigmented Prevotella isolates, occasionally including P. intermedia and P. nigrescens isolates, have been reported in several previous studies (6, 15, 17, 19). Penicillin-resistant isolates of P. intermedia and P. nigrescens as well as of P. pallens were also detected in the present study. All penicillin-resistant isolates were -lactamase producers, and the MIC (0.5 µg/ml) separating -lactamase-positive isolates from -lactamase-negative isolates was well in line with the value observed earlier (0.5 to 1 µg/ml) (6, 15, 17, 19). Interestingly, according to the MICs (0.5 µg/ml), one -lactamase-positive P. intermedia isolate and four P. nigrescens isolates of the present study belonged to the susceptible category of the latest NCCLS standard (25), which emphasizes the necessity of screening the pigmented Prevotella species for -lactamase.

All P. intermedia, P. nigrescens, and P. pallens isolates in the present study were susceptible to amoxicillin-clavulanate and cefoxitin, although these agents were relatively more active against -lactamase-negative isolates than against -lactamase-positive isolates. Good activity of cefoxitin and amoxicillin-clavulanate against pigmented Prevotella has also been reported in earlier studies (3, 6, 10). Consistent with the present finding, 100% susceptibility of pigmented Prevotella species to metronidazole has been observed in several previous studies (1, 14, 34). However, one metronidazole-resistant, pigmented Prevotella isolate, from a sample taken from a subject with pleuropulmonary infection, has been found (6). In the present study, we detected MIC₉₀s of metronidazole two dilution steps higher for P. intermedia than for P. nigrescens. Although the MICs for P. intermedia isolates were far below the breakpoint recommended for metronidazole (16 µg/ml) (25), a gradual development of resistance is possible, so resistance should be monitored. Azithromycin and trovafloxacin also were highly active against all isolates in the present study, which agrees well with earlier reports on Bacteroides and Prevotella species (14, 30, 35). However, the MIC₉₀s of azithromycin (0.125 µg/ml for P. intermedia and P. pallens and 0.06 µg/ml for P. nigrescens) were lower than those reported in earlier studies, which included only a limited number of P. intermedia and P. nigrescens isolates and no P. pallens isolates (13, 26). Other investigators (1, 14, 30, 34) have reported good in vitro activity of trovafloxacin, a recently described fluoroquinolone, against pigmented Prevotella and other gram-negative anaerobes. Conversely, the activity of the older fluoroquinolones against anaerobes has generally been poor or modest (11, 34).

P. intermedia, P. nigrescens, and P. pallens species are genetically highly diverse (20, 22, 23, 28). Therefore, finding of isolates with identical AP-PCR types from different subjects strongly suggests a common source of the isolates. Consistent with an earlier study by Paquet and Mouton (27), no association between the AP-PCR patterns and the sampling sites or geographic locations of the present isolates was found. Furthermore, there was no correlation between -lactamase production and AP-PCR patterns. This agrees with the results of the study by Könönen et al. (18) of Prevotella melaninogenica, a pigmented indigenous oral anaerobe, indicating that isolates of the same ribotype occasionally show variation in -lactamase production. Penicillin resistance in the P. intermedia group is due to a constitutively produced broad-spectrum -lactamase, which is related but not identical to -lactamases produced by the B. fragilis group organisms (2, 33). There are reports on conjugal transfer of the genes coding for -lactamase from P. intermedia to other Prevotella and Bacteroides species (12, 33). The transfer seems to be linked to transfer of the tet(Q) gene, which codes for tetracycline resistance (12, 33). The genetic heterogeneity of the present -lactamase-producing P. intermedia and P. nigrescens isolates suggests that, in addition to the overgrowth of the resistant clones during antibiotic therapy, acquisition of the resistance genes by transfer and mutation may have a role in the increase of penicillin resistance in these species. Therefore, studies on similarity of genes coding for -lactamase are needed to evaluate the significance of horizontal gene transfer in these species.

In conclusion, -lactamase production was common among the P. intermedia, P. nigrescens, and P. pallens isolates of the present study. -Lactamase-producing isolates were genetically heterogeneous and originated from both oral and extraoral sites as well as from different geographic locations, the United States and Finland. Because of the enzymatic hydrolysis of penicillin due to the -lactamase production of these species and because of the horizontal transfer of penicillin resistance determinants from P. intermedia to different species (12), therapy with penicillin may not be optimal in infections involving P. intermedia group bacteria. Fortunately, several other antibiotics (amoxicillin-clavulanate, cefoxitin, metronidazole, azithromycin, and trovafloxacin) proved to be highly active against these species.