Amoxicillin Treatment of Experimental Acute Otitis Media Caused by

advantageous for the bacterium during amoxicillin treatment of acute otitis media, a total of 70 Sprague-Dawley rats were challenged with a susceptible recipient strain or a genetically similar resistant transformant strain.


Antibiotic therapy was started on day 3 after inoculation, and the animals were monitored by daily otomicroscopy and analysis of bacterial samples from middle ear effusions obtained on day 8, the last day of observation. The animals were also sacriAced on days 4 and 8 and after 2 months for morphological examination. Compared with the susceptible recipient strain, recovery from infections caused by the resistant transformant strain was delayed, and the late structural changes were more severe in the animals challenged with the latter strain.


The results of the study indicate that chromosomal alterations mediating a relatively low level of resistance to beta-lactams may be advantageous for H. in\xeuenzae during antibiotic treatment of a local infection in the rat, and the alterations may occur without any signiAcant loss of virulence. Since the Nrst occurrence of ampicillin resistance in Hae- mophilus in)uenzae in the early 1970s (17), the prevalence of H.


in)uenzae strains with reduced susceptibility to beta-lactam antibiotics has increased worldwide. About 10 to 30% of all H. in)uenzae strains are beta-lactamase producers, with the rate being slightly higher among type b strains than among other, non-type b strains (4, 12, 14).


Through alterations pri- marily in the penicillin-binding proteins (PBPs), a non-beta- lactamase-mediated or chromosomal resistance (2, 7) has also evolved in up to 5% of the H. in)uenzae strains (26). Most strains with altered PBPs so far reported are nontypeable.


They are genetically diverse and are of respiratory origin (19). Data on the virulence of these strains are sparse, but it has been suggested that they may have a lower virulence potential. In the infant rat model of bacteremia, a reduced mortality and a reduced magnitude of bacteremia were observed after intra- peritoneal inoculation of ampicillin-resistant transformant, nontypeable H.


in)uenzae strains (28). However, both the site of infection and the in vivo proliferation were questioned. It is possible that deNcient proliferation is not a major disadvantage for the bacteria in respiratory tract infections, since H.


in)u- enzae cells grown in animals rapidly undergo a change to a state of limited cell division similar to that found in the late stationary phase in vitro (3), and the efNcacies of beta-lactams may be reduced (27). Furthermore, nasopharyngeal carriage in combination with frequent exposure to antibiotics and an in- completely developed immune system, conditions found in young children, could be a basis for selection. To address the questions of whether nontypeable H.


in)u- enzae strains with chromosomal resistance to beta-lactams may be less virulent in a local infection system and whether the chromosomal alterations could be advantageous for the bac- terium during treatment with beta-lactams, a fully susceptible nontypeable H. in)uenzae strain was compared with a geneti- cally similar transformant strain with reduced susceptibility to beta-lactams in the rat model for acute otitis media (AOM) during and after therapy with amoxicillin. MATERIALS AND METHODS Bacteria and media.


Four wild-type and two transformant nontypeable H. in)uenzae strains were used. Wild-type strains 3655 (biotype II; recipient strain with full susceptibility to beta-lactams) and 1161 (biotype III; control strain for lipopolysaccharide [LPS] analysis) were kindly donated by R.


Munson, St. Louis, Mo. Wild-type strains 4700 and 4089 (both biotype III; donor strains with re- duced susceptibilities to beta-lactams) were clinical isolates from the Depart- ment of Medical Microbiology, Lund University, Malmo¨ General Hospital, Malmo¨, Sweden.


For transformation of strains 3655/4700 and 3655/4089, see below. The clinical isolates were identiNed as H. in)uenzae by colony morphol- ogy, lack of hemolysis on blood agar, growth requirements for NAD and hemin, fermentation pattern, and symbiotic growth around a staphylococcal streak on blood agar (13).


The biotypes of the clinical isolates and transformant strains were determined as described by Kilian (13). The strains did not agglutinate with Murex typing sera a through f (Murex Diagnostics Ltd., Dartford, England), and no beta-lactamase production was detected with a chromogenic cephalosporin (nitroceNn; PDM Diagnostics Discs; AB Biodisk, Solna, Sweden). The organisms were stored in glycerol broth at \x3 70°C, and all cultures were initially inoculated from these frozen stocks onto solid media and grown at 37°C in an atmosphere with 5% CO 2 .


The media used were chocolate agar, chocolate agar supplemented with 1% enrichment (an in-house mixture corresponding to IsoVitaleX [BBL]) and 1% hemin (Sigma Chemical Co., St. Louis, Mo.) or brain heart infusion (BHI) broth or agar (Difco Laboratories, Detroit, Mich.) supple- mented with NAD and hemin (Sigma), each at 10 \x8 g/ml, and, if indicated, ampicillin (Sigma) at 50 \x8 g/ml. The inocula for middle ear challenge were prepared as described earlier (18).


Only freshly prepared stationary-phase bacteria at a concentration of 10 3 , 10 5 , or 10 7 CFU/ml were used. Viable counts of the bacterial suspensions were regularly determined at the time of middle ear challenge. Bacterial samples from the middle ears were obtained and cultured as described previously (18).


Genetic transformation of H. in\xeuenzae. High-molecular-weight bacterial DNA from nontypeable strains H.


in)uenzae 4700 and 4089 was prepared by a * Corresponding author. Mailing address: Department of Medical Microbiology, Lund University, Malmo¨ General Hospital, S-205 02 Malmo¨, Sweden. Phone: 46 40 331350.


Fax: 46 40 336234. 1979 on May 21, 2012 by guest http://aac.asm.org/ Downloaded from modiNcation (25) of the method described by Moxon et al. (22).


The DNA was sheared by repeated pipetting through a Nne syringe before being used for transformation. Nontypeable H. in)uenzae 3655 cells made competent by the aerobic-anaerobic incubation procedure of Goodgal and Herriot (6), as modiNed by Cameron (1), were transformed with approximately 1 \x8 g of sheared chromo- somal DNA.


The transformation mixtures were plated onto supplemented BHI agar plates, selecting for ampicillin resistance. Transformants were picked and reinoculated onto the same BHI agar plates described above to obtain single colonies, which in turn were colony puriNed once. Determination of MICs.


The MICs of penicillin V, amoxicillin, cefaclor, ce- furoxime, and cefotaxime were determined by an agar dilution method with supplemented chocolate agar. Antibiotics were added to the agar at twofold dilutions in freshly prepared solutions. The organisms were grown overnight in supplemented BHI broth and diluted in phosphate-buffered saline (pH 7.2), a Nnal inoculum of 10 4 CFU/spot was applied to the surface of the agar in 1- to 2- \x8 l quantities with a multipoint inoculator, and the plates were incubated overnight at 37°C in an atmosphere with 5% CO 2 .


Staphylococcus aureus ATCC 29213, Escherichia coli ATCC 25922, Enterococcus faecalis 29212, and H. in)uenzae NCTC 8468 were included as quality control strains. The lowest antibiotic con- centration which completely inhibited bacterial growth was recorded as the MIC (Table 1).


As an extra control, the MICs of benzylpenicillin, ampicillin, cefu- roxime, and cefotaxime were determined by the E-test (AB Biodisk). Interex- perimental variations in MICs were 1 twofold dilution. Growth rate.


For comparison of growth rates of wild-type and transformant strains, the different H. in)uenzae strains (except strain 1161) were plated onto chocolate agar. After 12 h of incubation, Nve colonies were transferred to 25 ml of supplemented BHI broth in 200-ml Oasks and were grown at 37°C with 200-rpm agitation overnight.


Finally, 100 \x8 l of each bacterial suspension was inoculated into 300 ml of supplemented BHI broth in 1-liter Oasks. The cultures were grown under the conditions described above. Samples were taken every 30 min, and the optical density at 620 nm was determined.


The experiment was carried out three times. LPS preparation and analysis. LPS was puriNed by a small-scale modiNcation (5) of the hot phenol-water extraction method (34) from wild-type strains 1161, 3655, 4700, and 4089 and transformed strains 3655/4700 and 3655/4089 grown overnight on chocolate agar plates.


Electrophoretic separation was done by tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis procedure de- scribed by Scha¨gger and von Jagow (29). The LWM electrophoresis calibration kit (Pharmacia Biotech Nordic AB, Sollentuna, Sweden) was used as the source of molecular weight standards, and the bacterial LPS was visualized by using the Instaview silver staining kit (BDH Laboratory Supplies, Poole, England). A whole-cell enzyme-linked immunosorbent assay was carried out to deter- mine the immunological properties of the LPSs from the strains listed above, as described elsewhere (11).


The panel of monoclonal antibodies reactive against different epitopes of H. in)uenzae LPS was kindly donated by A. Lindberg, Karolinska Institute, Huddinge, Sweden.


In brief, bacteria grown overnight were harvested from chocolate agar plates and were suspended in phosphate-buffered saline to an optical density at 600 nm of 0.2. Each well of the microtiter plate was coated with 100 \x8 l of the bacterial suspension at 4°C overnight. The plate was blocked, and after each step the plate was rinsed.


After incubation with mono- clonal antibodies in culture supernatant (100 \x8 l/well), a 1:1,000 dilution of horseradish peroxidase-conjugated rabbit anti-mouse immunoglobulins (Dako a/s, Glostrup, Denmark) was added. Finally, tetramethylbenzidine and potassium citrate were added as substrate. The process was halted with 1 M H 2 SO 4 , and the optical density values were read at 450 nm.


Animals and surgical procedures. To study the effects of amoxicillin treatment on the outcome of AOM after challenge with nontypeable H. in)uenzae strains with and without chromosomal changes mediating reduced susceptibility to beta- lactams, 70 clinically and otomicroscopically healthy male Sprague-Dawley rats (weight, 250 to 300 g) were used.


Whenever operated on or inspected under the otomicroscope, the animals were anesthetized with chloral hydrate (Apoteksbo- laget, Lund, Sweden), which was administered intraperitoneally. AOM was in- duced by instilling approximately 0.05 ml of the bacterial suspension with a Nne needle through the bony wall of the right bulla directly into the middle-ear cavity, with the bulla being reached through a ventral midline incision after blunt dissection of the soft tissue as described elsewhere (18). The diagnosis of AOM required direct visualization of opaque Ouid behind the tympanic membrane, and the effects of