Antimycobacterial activities in vitro and in vivo and pharmacokinetics of dihydromycoplanecin A.

The in vitro activity of dihydromycoplanecin A (DHMP A), a new cyclic peptide antibiotic, was compared with those of antimycobacterial drugs such as streptomycin, isoniazid (INH), rifampin, and ofloxacin against several clinically isolated species of mycobacteria, including Mycobacterium tuberculosis, M. intracellulare, and M. kansasii. DHMP A demonstrated stronger activities than other drugs against all species of mycobacteria tested at concentrations of less than 0.0125 to 25 microgram/ml. A marked synergism between DHMP A and INH was demonstrated by the checkerboard technique against M. tuberculosis, M. intracellulare, and M. smegmatis, and the synergistic effect was observed by treatment of the culture of M. smegmatis with DHMP A for at least 3 h prior to treatment with INH. It was also shown that both absorption and excretion of INH in mice were faster than those of DHMP A. On the basis of these results, combination therapy with DHMP A and INH was successfully carried out in experimental tuberculosis in mice infected with M. bovis Ravenel. After a single intravenous administration of 10 mg of DHMP A per kg, its half-life in serum in mice was about 0.5 h and in dogs it was 5.5 h. A single oral administration to dogs of 12.5 mg/kg gave a peak of 5.0 micrograms/ml at 3 h. In these experiments, urinary recoveries within 48 h were 21.0% in mice and 25.2% in dogs. The tissue distribution level of DHMP A in mice after oral administration was in the order of liver greater than kidney greater than serum greater than spleen = lung. The 50% lethal doses of DHMP A for mice were more than 6,000 mg/kg orally and 1,840 mg/kg intraperitoneally.

Dihydromycoplanecin A (DHMP A) was primarily discovered as an active metabolite in the urine of mice and dogs administered mycoplanecin A (MP A), a major component of the novel family of antibiotics produced by a soil isolate classified as Actinoplanes awajinensis subsp. mycoplanecinus subsp. nov. (10).
MP A possesses a cyclic structure composed of 10 amino acid residues containing three kinds of proline analogs and a-ketobutyric acid, which can be differentiated from three other minor components, MPs B, C, and D, by high-pressure liquid chromatography (6,7). DHMP A has a-hydroxybutyric acid as an N-acyl group, which was derived by reduction of the ketone group of the oa-ketobutyric acid moiety in the parent compound, as shown in Fig. 1.
This family of antibiotics has strong activity against mycobacteria, including strains of Mycobacterium tuberculosis, M. intracellulare, and some other species of atypical mycobacteria (3,9).
The present paper deals with the antimycobacterial activities of DHMP A in vitro and evaluation of the activity in vivo against experimental tuberculosis in mice, its pharmacokinetics in mice and dogs, and acute toxicities in mice and rats. * Corresponding author.

MATERIALS AND METHODS
Antimicrobial agents tested. DHMP A was prepared by chemical reduction of MP A isolated from the culture broth of A. awajinensis in our laboratories, and its amorphous powder (purity, greater than 95%) was used throughout the experiments. Rifampin (RFP; Rifadin) and ofloxacin (OFLX) were purchased from Daiichi Pharmaceutical Co., Ltd., Japan, isoniazid (INH) and ethambutol (EB) were from Sigma Chemical Co., and streptomycin (SM) was from Meiji Seika Co., Ltd., Japan.
Susceptibility test. Mycobacterial strains used in this study were mainly clinically isolated from patients with pulmonary diseases at East Saitama Hospital, Saitama Prefecture, Japan.
The susceptibility test was performed by a broth dilution technique in Dubos medium supplemented with 10% bovine serum albumin (Eiken Co., Ltd., Japan). Each mycobacterial strain grown on 1% Ogawa medium was subcultured in the Dubos medium for 1 to 2 weeks at 37°C. A 40-,ul amount of 102-fold-diluted subculture with the same medium freshly prepared was inoculated into the test tubes containing serial twofold dilutions of the test drugs in 2 ml of Dubos medium and incubated at 37°C for 3 weeks. MICs were expressed as the lowest concentrations of drug at which no visible growth was observed.
Microbiological assay of DHMP A. The antibacterial activity in serum, urine, feces, and tissue homogenates was determined by the conventional paper disk-agar diffusion method on nutrient agar (Eiken Co., Ltd., Japan) with Micrococcus luteus PCI1001 as the test organism. A stock solution of DHMP A was prepared at a concentration of 100 p.g/ml in 1/15 M phosphate buffer (pH 7.0) containing 10% methanol, from which three drug concentrations, 12.5, 6.25, and 3.13 ,ug/ml, were prepared with either serum from an experimental animal (serum samples) or 1/15 M phosphate buffer, pH 7.0 (for urine, tissue homogenates, and feces), as a diluent.
The lowest limit of detection of this method was 0.2 p±g/ml, and its precision was ±5% for drug concentrations ranging from 1 to 100 ptg/ml. Blood samples were drawn from the carotid artery of mice and from the antecubital vein of dogs. They were allowed to clot, and the sera were used for bioassay. Fecal samples extracted with 50% aqueous methanol were centrifuged. The supernatant was used for bioassay. Detection of DHMP A or its active metabolite in biological fluids. The ethyl acetate extracts from serum, urine, feces, and tissue homogenates were concentrated and applied on a thinlayer chromatography plate (Art 5715, F254 silica gel plate; Merck Co., Ltd., Federal Republic of Germany) and developed with a solvent system of CHCl3-methanol (10:1) for bioautography of the active principle. The Rf of DHMP A in this system was 0.4. Distribution of DHMP A and INH in organs of biological fluids in mice. Distribution of DHMP A in mice was determined as follows. Mice received an oral dose of 100 mg of DHMP A per kg and were sacrificed at 1, 2, 4, 6, 8, or 24 h after administration. The lungs, livers, kidneys, and spleens from five mice in each group were pooled after blotting each organ with filter paper to remove residual blood and homogenized in an eightfold volume of 0.9% saline. DHMP A in the supernatant was determined by the microbiological assay described above. Distribution of INH in mice was also determined as follows. Five mice in each group received a subcutaneous dose of 25 mg of INH per kg and were sacrificed at 0.5, 1, 2, or 4 h after administration. Tissue homogenates were prepared by the method described above. INH in biological fluids or tissue homogenates was deter-mined fluorometrically by the method described by Scott and Wright (8).
Combination effects of DHMP A and other antituberculous drugs. Combination effects of DHMP A and other antituberculous drugs, such as RFP, INH, SM, and EB, were examined by the checkerboard titration method. An overnight culture of M. smegmatis ATCC 607 in nutrient broth supplemented with 2% glycerol or a 2-week culture of M. tuberculosis and M. intracellulare on 1% Ogawa medium at 37°G was suspended in fresh Dubos medium, followed by application of 20 ,ul of inoculum (about 5 x 104 CFU) into 2 ml of Dubos medium containing DHMP A and other antituberculous drugs in a serial twofold dilution ranging from 0 to 100 ,ug/ml. MICs in every combination were determined after 1 or 2 weeks of incubation at 37°C. The synergistic effect of DHMP A and INH on A. smegmatis ATCC 607 was analyzed by incubating the culture with either of the drugs for various periods of time (DHMP A, 0, 3, 7, and 20 h; INH, 0, 1, 3, and 7 h) and removing the drug by centrifugation at 5,000 rpm for 15 min, followed by determination of the MIC for the rest of this pretreated culture (5 x 104 CFU/ml) in Dubos medium at 37°G for 7 days. The drug concentration in each pretreatment was equivalent to the MIC against the test organism, 0.2 ,ug/ml for DHMP A and 12.5 ,ug/ml for INH.
Experimental animals. Strain ICRIJCL male mice, weighing 20 to 22 g; strain ddY male mice, weighing 22 to 24 g; and strain NMRI male and female mice, weighing 20 to 22 g, were used. Strain Fischer male and female rats weighing 150 to 200 g were used. Male beagle dogs weighing 10 to 15 kg were used. The dogs were fasted for 15 to 18 h prior to administration of the drug.
In vivo activity of DHMP A. (i) Efficacy of DHMP A on experimental tuberculosis in mice. Eighty male ddY mice were inoculated intravenously with 3 x 106 CFU of M. tuberculosis H37Rv. They were divided into eight groups (10 mice in each). Treatment by oral administration of DHMP A, MP A, and RFP started on day 3 after challenge and continued for 21 days. All mice were sacrificed on day 31 after challenge. The lungs and spleens from five randomly selected mice in each group were separately homogenized in 1/15 M phosphate buffer (pH 7.0) in saline to make a 10% homogenate; 0.1-ml portions were poured onto 1% Ogawa medium, and viable colonies were counted after incubation for 4 weeks at 37°C.
(ii) Efficacy of combination therapy (DHMP A with INH) on a model of lethal infection with M. bovis Ravenel. Four groups of ddY male mice (10 mice in each group) were inoculated intravenously with 0.5 x 107 to 2.5 x 107 CFU of M, bovis Ravenel. Each group of mice received DHMP A or INH alone, a combination of both drugs, or saline as a control. DHMP A was orally administered to the mice at a dose of 4 mg in a 0.2-ml suspension in distilled water per mouse after 5 days of infection, and 4 h later, INH was given subcutaneously at a dose of 0.1 mg in 0.2 ml of saline per mouse, and treatment was continued once a day for 3 weeks. The efficacy was assessed by percent survival and by body weight gain over 40 and 100 days, respectively, after challenge. VOL. 32, 1988 (iii) Efficacy of combination therapy (DHMP A with INH) on experimental tuberculosis. The regimen for M. tuberculosis R-KM infection was similar to that described above, except that treatment started on day 12 after challenge.
Oral administration. DHMP A was dissolved in cold water at low doses between 12.5 and 200 mg/kg (DHMP A is more soluble in cold water than in hot water) or suspended in 0.5% carboxymethyl cellulose at high doses of more than 200 mg/kg. RFP was dissolved in 25% aqueous propylene glycol. The total volume of the sample solution for each dose was 0.2 ml in mice by oral gavage and 1 ml/kg of body weight in dogs by oral intubation.
Intravenous administration. The solution of DHMP A injected into mice was an emulsion of 1 ml of medium-chainlength triglyceride (Nissin Oil Co., Ltd., Japan) and 100 mg of polyoxyethylene glycol laurate ester (MYL-10) in 9 ml of 0.9% physiological saline. The vehicle for dogs was 15% ethanol in polyoxyethylene glycol 400.
Intraperitoneal administration. Intraperitoneal administration of the drug was carried out after suspension in 0.5% carboxymethyl cellulose.
Urinary and fecal excretion. After administration of the antibiotics, urine and feces samples from three animals confined to a metabolic cage were collected at appropriate intervals and pooled every 24 h. Pharmacokinetic analysis. Pharmacokinetic analysis of DHMP A concentrations in serum after oral and intravenous administration was done by nonlinear least-squares program NONLIN 74 (5) with the following equations: area under the curve = dose/k /V and t1/2 = 0.693/k1, where kel is the first-order elimination rate constant, V is the apparent volume of distribution, and t1/2 is the half-life.
Acute toxicities. Acute toxicities (50% lethal doses [LD50s]) of the drug were determined in mice and rats (20 animals in each group) by oral and intraperitoneal administration. LD50s in each experiment were calculated by the method described by Litchfield and Wilcoxon (4).

RESULTS
In vitro antimycobacterial activities. Antimycobacterial activities of DHMP A in Dubos medium were the best among the antituberculous drugs tested against M. tuberculosis, M.
intracellulare, M. gordonae, and M. nonchromogenicum. effect was observed for the combination of DHMP A with INH, and the minimal FIC index between them was 0.250 to 0.375. No antagonistic effect was observed between DHMP A and the other three drugs (Fig. 2). A similar synergistic effect of DHMP A with INH was also apparent on clinically important mycobacteria, such as M. tuberculosis H37Rv IFM 2029 and M. intracellulare IFM 2073 (Fig. 3).

MICs of DHMP
The effect of pretreatment with either of the two drugs on their synergistic effect was examined. The MIC of DHMP A was not affected even when the bacterial cells were pretreated with INH at 12.5 ,ug/ml (the MIC) for 7 h, but that of INH was lowered from 12.5 to 3.13 ,ug/ml by pretreatment concentration was observed within 30 min after administration of 25 mg/kg subcutaneously. It was also rapidly eliminated from all of the tissues and became undetectable at 4 h. Absorption of DHMP A (100 mg/kg) given orally was slower than that of INH, and peak concentrations in liver and other tissues, including serum, were observed at 2 and 4 h after administration, respectively. The tissues with the highest concentrations were liver and kidney, followed by serum, spleen, and lung (Fig. 5). The active principle in all tissues tested was identified to be DHMP A itself as a sole active compound by thin-layer chromatography on silica gel. Concentrations of DHMP A in all of the tissues tested gradually decreased, but the level in serum remained at 0.5 ,ug/ml even after 8 h (Fig. 5). These results combined with the results of pretreatment studies shown in Fig. 4 suggested that the two drugs might not coexist in tissues if they were administered simultaneously and the in vivo efficacy of combining DHMP   Fig. 7 for details.
given subcutaneously at a dose of 0.1 mg per mouse; treatment continued once a day for 3 weeks. Efficacy was assessed by percent survival for 100 days after challenge. Mean survival days (± standard deviation) of four groups (not treated or treated with DHMP A alone, INH alone, or a combination of DHMP A and INH) were 20.9 ± 1.4, 52.7 + 7.6, 39.0 ± 6.2, and 89.2 ± 5.4, respectively (Fig. 7).
In the second experiment, combination therapy with DHMP A and INH was examined in experimental tuberculosis in mice infected with M. tuberculosis R-KM. The regimen was similar to that in the first experiment described above, except that treatment started on day 12 after challenge. Treatment continued for 4 weeks, but no detectable reduction in viable units of the pathogen in the lung and spleen was noted compared with those of the nontreated group (Fig. 8).
Pharmacokinetics in mice. (i) Intravenous administration.
After administration of 10 mg of DHMP A per kg, its t1/2,3 in (ii) Oral administration. A single oral administration (50 mg/kg) of DHMP A revealed a peak level in serum of 10.0 ,ug/ml at 2 to 4 h, and its urinary recovery was 21.5% (Fig.  10).
Pharmacokinetics in dogs. The t1/2 in serum after a single intravenous DHMP A dose of 10 mg/kg was 5.5 h, and its urinary excretion within 48 h was 37% (Fig. 9). Single oral DHMP A doses of 12.5 and 25 mg/kg gave peaks of 5.0 and 9.0 ,ug/ml, respectively, in serum at 3 h after administration (Fig. 10). Recoveries of the antibiotic in urine within 48 h were 25.2 and 22.4%, respectively.
LD50s. Both mice and rats tolerated an oral DHMP A dose of 6,000 mg/kg in 0.5% carboxymethyl cellulose suspension. No toxic signs were observed in the animals, and there was a normal increase in body weight over the 14-day observa-  A marked synergistic effect was observed in the combination of DHMP A with INH, and the minimal FIC index was 0.250 to 0.375 against M. tuberculosis as well as M. intracellulare. When bacterial cells were exposed to either DHMP A or INH for various periods of time and then to the other drugs, a synergistic effect was recognized only when the cells were treated with DHMP A 3 h before INH. Comparative studies on tissue distributions of DHMP A and INH in mice revealed that both absorption and excretion of INH were faster than those of DHMP A. Therefore, mice infected with M. bovis Ravenel 5 days before were administered DHMP A orally and then INH 4-h-delayed spiking once daily, and therapy was continued for 3 weeks. The mean survival of the group treated with DHMP A and INH was 89 days, compared with 21, 39, and 53 days for the nontreated, INH alone, and DHMP A alone groups, respectively. Thus, the combination effect of these two drugs was clearly demonstrated in a lethal infection of experimental tuberculosis in mice. But it was not so effective in the second experiment of experimental tuberculosis, in which treatment was started on day 12 after challenge. After oral and intravenous administration of DHMP A in mice and dogs, the recoveries in urine suggested that absorption of the antibiotic after oral administration might be more than 80% of the given dose and that it was excreted mainly through the biliary tract. Although a higher dose of DHMP A than of RFP was necessary to protect mice from experimental tuberculosis, its lower toxicity than that of RFP reported in the literature (1) and broader antimycobacterial spectrum, as well as its higher level of oral and intravenous administration in dogs, suggested the possible clinical usefulness of DHMP A alone as well as in combination therapy with INH not only in the treatment of tuberculosis but also in the treatment of diseases caused by atypical mycobacteria.