ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, Apr. 1975, p. 421-425 Copyright 0 1975 American Society for Microbiology Vol. 7, No. 4 Printed in U.S.A. Effects of Minocycline and Other s on Fusobacterium necrophorum Infections in Mice NYDIA A. KUCK Lederle Laboratories Division, American Cyanamid Company, Pearl River, New York 10965 Received for publication 17 October 1974 Several antibiotics were evaluated in model infections produced in mice with each of two strains of Fusobacterium necrophorum. In one model, local abscesses occurred at the site of subcutaneous injection; in another intra-abdominal abscesses were produced when the organisms were injected into the peritoneal cavity. Treatment with effective antibiotics prevented the formation of abscesses or minimized the size of the lesions. Several treatment schedules were used. Minocycline was the most active antibiotic of the seven agents tested against both strains and in both models. Clindamycin was equal to minocycline against one strain with certain multiple dose treatment schedules and less active with others. Protective effects in mice were achieved with serum levels of minocycline and clindamycin that appear to be clinically achievable. Doxycycline was less active than minocycline, and tetracycline was relatively ineffective, as were cephalexin, ampicillin and penicillin G. Infections with anaerobic organisms are being reported with increasing frequency since isolation methods have been improved. Fusobacteria, either alone or with other bacteria, have been associated with metastatic abscesses affecting abdominal and pelvic cavities, the liver, lung and pleural spaces, as well as septicemia (2, 3, 6, 8). The high mortality rate due to Fusobacterium necrophorum was reported as early as 1935 (3). More recently, Gelb and Seligman reported that among the major anaerobic species associated with human infection is Fusobacterium (Sphaerophorus) necrophorum (5). Until recently, model infections in animals suitable for the evaluation of therapeutic agents for anaerobic infections have been lacking. In 1974 Hill et al. published a report on a mouse infection consisting of progressive intrahepatic abscesses produced by injection of single and combined species of nonsporulating anaerobic bacteria (7). Single species produced disease in a small percentage of animals; the percentage was increased by the injection of combined species. Antibacterial agents were not evaluated in this report. Later in 1974, Wilkins and Smith reported on the efficacy of antibacterial agents in a mouse model in which disease resulting in death was produced by a single strain of F. necrophorum isolated from a case of sheep foot rot (9). In this report we describe mouse models in which minocycline was the most effective of 421 seven agents against either local abscesses or metastatic intra-abdominal abscesses produced by either of two strains of F. necrophorum isolated from human clinical cases. MATERIALS AND METHODS Cultures. The two strains of F. necrophorum, no. 606 and 1260, were clinical isolates received from Vera L. Sutter, Veterans Administration (Wadsworth) Center, Los Angeles, Calif. All cultures were prepared in fresh media. Broth cultures were made in fluid thioglycollate medium (135-C, Baltimore Biological Laboratories) supplemented with 0.5% yeast extract, hemin (5 Ag/ml) and menadione (1 ug/ml), except where indicated. Trypticase soy agar (BBL) or Mueller-Hinton agar supplemented with 5% rabbit blood was used for plate cultures. All cultures were incubated at 37 C in GasPak jars. s. Minocycline HCI and tetracycline HCI are products of the Lederle Laboratories Division, American Cyanamid Co. Doxycycline hyclate and penicillin G (1,650 U/mg) were obtained from Pfizer, Inc; clindamycin HCI from the Upjohn Co.; ampicillin trihydrate from Beecham Pharmaceuticals; and cephalexin monohydrate from Eli Lilly and Co Ṁice. Female Swiss mice, strain CD-1 (from Charles River Breeding Laboratories, Inc.), weighing 18 to 20 g were used in these studies. Infection. After preliminary tests showed that the cultures produced abscesses in mice from which the organisms could be isolated in pure culture, the following procedure was established for the preparation of inocula for infection. Purulent material from a well-isolated abscess was transferred aseptically to
422 KUCK tubes of broth and streaked on blood agar plates. A set of cultures was incubated anaerobically and another set aerobically. After 24-h incubation, the cultures were checked for purity, and 1 ml of the anaerobic broth culture was transferred to 100 ml of fresh thioglycollate broth. After 18 to 20 h of incubation in a GasPak jar, the culture was diluted in broth to contain approximately the required number of bacteria per milliliter by measurements of optical density. The viable count was estimated from colonies developing from duplicate 0.1-ml samples of 10-fold dilutions of the culture spread on blood agar plates after 48-h anaerobic incubation. The diluted culture was used immediately to infect mice by injecting 0.5 ml either subcutaneously at the base of the tail or intraperitoneally. Untreated control mice were infected at the beginning and at the end of the infecting period. During the course of this study no significant increase in virulence was detected. Treatment. The antibiotics were dissolved or suspended in 0.2% aqueous agar. A dose was contained in 0.5 ml and administered by gavage or injected subcutaneously. In each test, five to ten mice were treated at each dose level. The data from two or three separate tests were pooled for the estimation of median effective doses by probit analysis (4). In vitro antibiotic susceptibility. Minimal inhibitory concentration values were determined by the agar dilution method. Twofold dilutions of drug were incorporated in brain heart infusion agar supplemented with 5% rabbit blood. The agar surfaces were inoculated with 10-2 dilutions of 24-h-old broth cultures of the organism applied with Steer multiple inocula replicator. After 48-h incubation in a GasPak, the lowest concentration of drug showing no macroscopic growth was recorded as the minimal inhibitory concentration. RESULTS In vitro susceptibility to antibiotics. Both strains of F. necrophorum were susceptible to relatively low concentrations of the seven antibiotics in an agar dilution test (Table 1). F. necrophorum strain 606 subcutaneous infections. The injection of approximately 1 x 106 to 5 x 106 viable units of strain 606 subcutaneously at the base of the tail resulted in lesions at the site of injection in 95 to 100% of TABLE 1. In vitro susceptibility of Fusobacterium necrophorum strains to antibiotics Minimal inhibitory concentration (Mg/ml) Strain 606 Strain 1260 Minocycline 0.03 0.003 Doxycycline 0.06 0.007 Tetracycline 0.25 0.06 Clindamycin 0.007 0.003 Ampicillin 0.03 0.03 Cephalexin 0.50 0.06 Penicillin 0.03 0.007 ANTIMICROB. AGENTS CHEMOTHER untreated mice. Small lesions could be detected macroscopically in 3 or 4 days and developed to approximately 1 to 2 cm in 10 days. These lesions then generally healed but occasionally progressed, causing other subcutaneous or internal lesions. Death sometimes occurred due to either massive internal infection or to a single large abscess blocking the intestinal or urinary system. Among the treated mice lesions were absent, similar to the controls, or substantially smaller. Because the size of lesions varied among the mice treated with a particular dose level, the effectiveness of the dose was determined on the basis of the number of mice in the group without macroscopic lesions 10 days after infection. After 20 days, there was generally no significant change in the number of completely protected mice. Some of the lesions that were small at 10 days were large at 20 days, whereas other lesions healed without progressing to the large lesions seen in the untreated controls. Cultural examination of abscesses from representative mice from each group indicated a pure-culture infection. Effective drugs produced a dose response to the subcutaneous infection (Table 2). Median effective doses could be determined. Results are summarized in Table 3. Although the different treatment schedules represent separate groups of tests, it appears that a single oral dose administered immediately after infection was less effective than multiple treatments, but there was no advantage in administering nine doses instead of five doses. In each case minocycline and clindamycin were the most active drugs. They were two or three times more effective than doxycycline and at least four times more effective than tetracycline, ampicillin, and cephalexin. Minocycline and clindamycin were also effective when treatment was initiated 24 h after infection and administered once daily for 4 days. When a single subcutaneous dose was administered immediately after infection, minocycline again was the most active drug. It was at least three times more effective than the six other antibiotics tested. F. necrophorum strain 606 intraperitoneal infection. An intraperitoneal dose of 5 x 106 to 10 x 10 viable units of strain 606 resulted in infection of 95 to 100% of untreated mice. At necropsy 10 to 14 days after infection, untreated mice had multiple abscesses of various sizes involving several sites including peritoneal membrane, mesentery, intestinal wall, gonads, and liver. Spleens were large and sometimes grossly enlarged. Such mice survived up to 4 months. Among the treated mice, necropsy
VOL. 7, 1975 revealed either complete protection or various degrees of abscess involvement. Thus, as in the case of the subcutaneous infection, the effectiveness of treatment was based on the number of mice with no macroscopic lesions 14 days after infection. Cultural examination of abscesses from various sites in treated and untreated mice revealed a pure-culture infection. An effective antibiotic produced a dose response to the intraperitoneal infection (Table 4). Median effective doses are summarized in Table 5. Although the treatment schedules represent different groups of tests, it appears that nine oral doses were more effective than five oral doses. Again minocycline was the most active. Clindamycin was about equal to minocycline when nine doses were administered but only one-third to one-half as active when five doses were used. With the nine-dose schedule, minocycline was twice as active as doxycycline and five times as active as ampicillin and at TABLE 2. Effect ofgraded doses of antibiotics against Fusobacterium necrophorum 606 subcutaneous infection in mice No. of mice protected/no. tested at oral dose (mg/kg per dose) of:a 256 64 16 4 Minocycline 25/30 14/29 6/30 8/30 Doxycycline 15/30 3/30 5/30 2/29 Tetracycline 5/30 2/30 2/30 3/30 Clindamycin 27/30 9/30 5/30 4/30 Ampicillin 5/30 2/30 0/10 0/10 Cephalexin 3/30 2/10 1/10 1/10 a Pooled data from three tests. 59/60 untreated mice had subcutaneous lesions 10 days after infection. Treatment once daily 1, 2, 3 and 4 days after infection. TABLE 3. F. NECROPHORUM INFECTIONS 423 TABLE 4. Effects of graded subcutaneous doses of antibiotics on Fusobacterium necrophorum strain 606 intraperitoneal infection in mice No. of mice protected/no. tested at single subcutaneous dose (mg/kg) of:a 256 128 64 32 16 Minocycline 15/15 11/15 8/15 3/15 0/10 Doxycycline 9/15 6/15 4/15 0/5 Tetracycline 4/15 0/15 0/5 Clindamycin 4/15 4/15 0/10 Ampicillin 5/15 3/15 0/5 Cephalexin 0/15 0/5 Penicillin G 4/15 1/5 1/5 apooled data from three tests. 30/30 untreated infected mice had intra-abdominal abscesses 14 days after infection. Dose administered immediately after infection. TABLE 5. Comparison of the activities of antibiotics against Fusobacterium necrophorum 606 intraperitoneal infection in mice Median effective dose (mg/kg/dose)a Oral doses Subcutane- 5 Doses (0h - 2 9 Doses (0 h 4 ousdh)oe ( days) days) Oh Minocycline 23 (14-38) 23 (10-46) 65 (48-87) Doxycycline -256 67 (28-220)P 152 (98-225)c Tetracycline > 256 > 256 > 256 Clindamycin 63 (48-82)C 18 (8-33)d > 256 Ampicillin > 256 132 (59_453)c > 256 Cephalexin > 256 > 256 > 256 Penicillin G - - > 256 aeach treatment regimen represents separate groups of tests. Median effective doses were determined from pooled data from two to three separate tests. Figures in parentheses are 95% confidence limits. 'Significantly different at P < 0.05. CP < 0.01. dnot significantly different from minocycline. Comparison of the activities of antibiotics against Fusobacterium necrophorum 606 subcutaneous infection in mice Median effective dose (mg/kg/dose)a Oral doses" Subcutaneous dose 1(Oh) 5(Oh -2days) 9(Oh-4days) 4(+1_-+4days) (Oh) Minocycline 350 (160-1,200) 88 (60-140) 75 (58-97) 45 (25-84) 88 (54-150) Doxycycline >1,024 270 (193-540)c 2256 2256 >256 Tetracycline > 1,024 > 512 > 256 > 256 > 256 Clindamycin 540 (285-1,400)d 101 (65-200)d 86 (65-114)d 61 (37-100)d >256 Ampicillin > 1,024 > 512 > 256 > 256 > 256 Cephalexin > 1,024 > 512 > 256 > 256 > 256 Penicillin G > 256 a Median effective doses were determined from pooled data from two to three separate tests. Figures in parentheses are 95% confidence limits. b Each treatment schedule represents separate groups of tests. c Significantly different at P < 0.01. d Not significantly different from minocycline.
424 KUCK least ten times more active than tetracycline and cephalexin. When a single subcutaneous dose was administered immediately after infection, minocycline was again the most active. It was twice as active as doxycycline and at least four times as active as clindamycin, tetracycline, ampicillin, cephalexin, and penicillin. F. necrophorum strain 1260 infections. The infections with strain 1260 required 1 x 107 to 5 X 107 viable units to develop lesions. The subcutaneous lesions were generally smaller and tended to heal more readily than those produced by strain 606. In the intraperitoneal infection, the degree of involvement among the untreated controls varied and was often observed as small abscesses on the liver, although abdominal and pelvic lesions were also observed. There was a tendency to heal spontaneously. The effectiveness of treatment was determined on the basis of complete protection. Graded response was produced with effective drugs (Table 6). Minocycline was the most effective antibiotic against both the subcutaneous and intraperitoneal infections with strain 1260 (Table 7). Lower doses were required for protection against this strain than for the strain 606 infections. Clindamycin was only one-eighth to one-fourth as active an minocycline; doxycycline about one-tenth as active, and tetracycline, ampicillin, and cephalexin less than one-sixteenth as active as minocycline. Serum concentrations in mice. The serum concentrations in mice following single oral doses of 128 mg of four antibiotics per kg are shown in Table 8. Minocycline produced the highest peak level and was well maintained, as was doxycycline. Tetracycline produced relatively low concentrations. Since the protective effects of minocycline were generally associated with doses of less than 128 mg/kg, it can be TABLE 6. Effect ofgraded doses of antibiotics against Fusobacterium necrophorum 1260 intraperitoneal infections in mice No. of mice protected/no. tested at oral dose (mg/kg per dose) of:a 256 64 16 4 1 Minocycline 9/10 4/10 3/10 1/5 Doxycycline 6/10 2/10 2/10 0/5 Tetracycline 0/10 1/10 Clindamycin 9/10 6/10 2/10 0/5 Ampicillin 1/5 1/5 Cephalexin 0/5 0/5 _ a Pooled data from two separate tests. 29/30 untreated mice had multiple abscesses 14 days after infection. Treatment immediately after infection and twice daily for 2 days after infection. ANTIMICROB. AGENTS CHEMOTHER. TABLE 7. Comparison of the activities of antibiotics on infections produced with Fusobacterium necrophorum 1260 in mice Median effective dose (mg/kg/dose)" (5 oral doses) Subcutaneous Intraperitoneal infection infection Minocycline 9 (4-23) 11 (3-35) Doxycycline >256 105 (47-400)" Tetracycline > 256 > 256 Clindamycin 78 (29-153) 48 (22-100)b Ampicillin > 256 > 256 Cephaleyin > 256 > 256 a Median effective doses were determined from pooled data from two separate tests. Figures in parentheses are 95% confidence limits. b Significantly different from minocycline at P < 0.01. TABLE 8. Serum concentrations in mice after administration of single oral dose of antibiotic Single oral dose Serum concn (ug/ml)a (128 mg/kg) lb 3 6 24 Minocycline 12 5.2 2.8 0.22 Doxycycline 7.4 6.4 5.9 0.22 Tetracycline 2.5 0.83 0.33 <0.10 Clindamycin 1.8 0.51 0.18 <0.15 a Each serum sample was obtained from the pooled heart blood from five mice. The sera were assayed by the disk plate method. Bacillus cereus var. mycoides was the test organism for the tetracyclines, and Sarcina lutea was the test organism for clindamycin. b Hours after dose. concluded that protection was afforded with maintained levels of less than 4,ug/ml, which are readily achieved in man with much lower doses. Clindamycin produced the lowest blood levels of the four drugs; protection in mice was associated with serum concentrations that are also readily achievable in man. DISCUSSION In a recent report concerning the serious problems that intra-abdominal abscesses presented in the clinic, Altemeier et al. emphasized that anaerobes are of greater frequency and importance than aerobes (1). Antibacterial agents have been evaluated for efficacy against aerobic bacteria by both in vitro and in vivo methods, but until the recent report of Wilkins and Smith (9) only in vitro methods were used for evaluation against anaerobic bacteria. Our mouse models offer a practical method for the evaluation of antibacterial agents against an anaerobic infection. Two models are described.
VOL. 7, 1975 The subcutaneous infection offers the advantage of permitting the evaluation of activity without necropsy. On the other hand, the intraperitoneal infection, with the 606 strain at least, required lower doses of drug for effect, thus permitting the detection of activity of drugs of lower potency, as in the case of ampicillin (Table 5). The relative efficacies of the drugs were in general similar in the two models. In neither model did the in vivo activity reflect the in vitro potency of the antibiotic. The pharmacodynamics of an agent in the animal are important. Thus in the mouse, particularly against a localized chronic infection, the maintained serum concentration and the concentration of the drugs at the site of infection is to be considered as well as the inherent susceptibility of the infecting organism. On a dosage basis, minocycline was clearly the most active drug of the seven agents tested in both models and against both strains of F. necrophorum. On the basis of serum concentration, clindamycin was more active than minocycline. With both drugs, however, the doses associated with protective effects produced serum concentrations readily achievable in man with lower dosages. Although less frequently than Bacteroides fragilis, Fusobacterium is encountered in clinical infections. As yet we have no suitable animal model for the study of antibiotic effects against Bacteroides infection. Our models with Fusobacterium, however, demonstrated the effectiveness of clindamycin and the superiority F. NECROPHORUM INFECTIONS 425 of minocycline over doxycycline and tetracycline against an anaerobic infection in which abscess simulating the clinical disease were involved. More recently we observed that metronidazole was also effective in our models. ACKNOWLEDGMENTS I am grateful to Caroline Clancy and Gary Wilson for technical assistance and to Emil Pelcak and Penelope Smith for the assay of serum samples. LITERATURE CITED 1. Altemeier, W. A., W. R. Culbertson, W. D. Fullen, and C. D. Shook. 1973. Intra-abdominal abscesses. Am. J. Surg. 125:70-79. 2. Bartlett, J. G., and S. M. Finegold. 1974. Anaerobic infections of the lung and pleural space. Am. Rev. Respir. Dis. 110:56-77. 3. Chow, A. W., and L. B. Guze. 1974. Bacteriodaceae bacteremia: clinical experience with 112 patients. Medicine 54:93-126. 4. Finney, D. J. 1964. The maximum likelihood solution, p. 48-64. In Probit analysis, 2nd ed. University Press, Cambridge. 5. Gelb, A. F., and S. J. Seligman. 1970. Bacteroidaceae bacteremia. J. Am. Med. Assoc. 212:1038-1041. 6. Gorbach, S. L., and J. G. Bartlett. 1974. Anaerobic infections (part II). N. Engl. J. Med. 290:1237-1245. 7. Hill, G. B., S. Osterhout, and P. C. Pratt. 1974. Liver abscess production by non-spore-forming anaerobic bacteria in a mouse model. Infect. Immun. 9:599-603. 8. Sabbaj, J., V. L. Sutter, and S. M. Finegold. 1972. Anaerobic pyogenic liver abscesses. Ann. Intern. Med. 77:629-638. 9. Wilkins, T. D., and L. Ds. Smith. 1974. Chemotherapy of an experimental Fusobacterium (Sphaerophorus) necrophorum infection in mice. Antimicrob. Agents Chemother. 5:658-662.