INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Isolates of the Mycobacterium avium-M. intracellulare complex (MAI) are the most common pathogens of disseminated bacterial infection in patients with AIDS (3). Unfortunately, members of the MAI are resistant to most drugs. Some new macrolide antibiotics such as clarithromycin and azithromycin are bactericidal against MAI. Their use has improved the therapy, but resistance still develops in as little as 4 months (3, 14, 33). Combinations of drugs are used in an effort to prevent emergence of resistant organisms (38, 41). Additive effects among antibiotics that improve therapy have been reported by multiple investigators, but synergistic effects have proven elusive (30, 31). Furthermore, drug interactions and toxicity limit the use of combination therapy (3, 14, 16). Consequently, there is a continuing need for new therapeutic agents.

Mycobacteria are relatively resistant to most antibiotics. Saprophytic soil-dwelling mycobacteria such as MAI are typically more resistant to antibiotics than obligate pathogens. These organisms must defend themselves against antibiotics and toxins in their natural habitat. It is thought that their cell walls developed very low permeability as a barrier defense against noxious agents in their environments (10). For example, the rates of diffusion of cephalosporins across the cell walls of mycobacteria are nearly four orders of magnitude lower than those of the outer membrane of Escherichia coli. The low permeability of mycobacterial cell walls explains the level of resistance (predicts the MICs) to cephalosporins and may be the main reason that many antibiotics are ineffective against MAI (32, 33, 35, 36).

The organization of the lipid hydrocarbons is thought to contribute to the barrier property of intact mycobacterial cell walls (35). Consequently, we hypothesized that agents that disorganize the surface lipids of mycobacteria might increase the efficacies of some antibiotics. This was not a new idea. Surfactants were known to modulate surface lipids of mycobacteria (5, 32). Cornforth and colleagues (11-13) reported that certain large hydrophobic nonionic surfactants had antimycobacterial activities and enhanced the activities of antibiotics in murine infection.

Stimulated by the work of Cornforth and Behling (11-13) and by the plasticities and the favorable toxicity profiles of poloxamers, we evaluated the activities of a series of poloxamers against MAI (19). We found that certain large hydrophobic poloxamers were bacteriostatic against fresh clinical isolates of MAI and produced synergistic effects with rifampin in broth culture. The most effective poloxamer, P331, was found to be a mixture of molecular species and to contain significant amounts of inactive and toxic impurities. Consequently, CRL-1072 was synthesized by an improved process by using supercritical fluid fractionation. The characterization and activity of CRL-1072 as an antimicrobic agent against M. tuberculosis have been reported previously (21, 22).

U937 is a nonadherent human monocytoid cell line that can be stimulated with phorbol myristyl acetate, retinoic acid, and vitamin D₃ to differentiate into macrophage-like cells that have the ability to kill ingested organisms due to increased levels of reactive oxygen and nitrogen synthesis and other mechanisms (17, 25, 37). Unstimulated U937 monocytes lack antimycobacterial mechanisms such as the expression of the natural resistance associated macrophage protein (NRAMP) and reactive oxygen but have surface complement, Fc, and scavenger receptors which enable phagocytosis of mycobacteria (8, 28, 42). We sought methods to study the effects of drugs on intracellular organisms in unstimulated U937 monocytic cells in order to discriminate between the effects of antibiotic-induced killing and macrophage-induced killing (2, 7, 34). We developed methods for the highly reproducible growth of M. tuberculosis strains (strains H37Rv and Erdman) and MAI in U937 cells (20-23).

The present studies were designed to evaluate the ability of CRL-1072 to enhance the activities of antibiotics against MAI in U937 cells and in murine infection. Limited pharmacokinetic and toxicity studies were conducted as a guide to dosing of the animals. CRL-1072 was found to produce synergistic killing of MAI with several antibiotics in each model studied.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

CRL-1072. The chemistry and synthesis of poloxamers have been described previously (9, 29). Poloxamers are nonionic surfactants composed of chains of hydrophilic polyoxyethylene (POE) and hydrophobic polyoxypropylene (POP) in the configuration


where a and b are integers approximating 4 and 60, respectively. CRL-1072 was synthesized to produce a mean molecular mass of POP chains of 3,500 Da each and POE chains of 200 Da each. Previous investigations had demonstrated that these chain lengths produce optimal antimycobacterial activity in broth culture (19). The poloxamers used in previous studies, P331 and CRL8131, had similar chain lengths but differed in their purities (19, 21-23). The synthetic method was optimized to reduce the variation in the lengths of the chains about these means. The material was then subjected to supercritical fluid fractionation to remove low-molecular-mass impurities and was named CRL-1072. CRL-1072 was formulated at 30 mg/ml in a vehicle consisting of 2% Tween 80-1% ethanol for all of the present studies. [¹⁴C]CRL-1072 was synthesized by the same method by CytRx Corporation, Norcross, Ga. It had a specific activity of 274 µCi/g and a purity of 99.7%. For pharmacokinetic studies, weighed blood, tissue, urine, or fecal samples were desiccated and combusted to release CO₂. The ¹⁴C label was then measured in a scintillation counter. Most of the injected radioactivity (70 to 80%) could be accounted for in these studies.

Mycobacterial strains. MAI reference strain TMC 724 (ATCC 25291) and clinical isolate ATCC 49601 were obtained from the American Type Culture Collection (Manassas, Va.). They were grown in 7H9 broth, harvested at the logarithmic phase, washed with saline, sonicated to disperse clumps, and matched in turbidity to a no. 1 McFarland suspension prior to freezing in aliquots at 70°C. Thawed aliquots were diluted in saline and were plated on 7H11 agar to determine the CFU counts of stock suspensions.

Macrophage assay of intracellular growth of MAI. Antimicrobic sensitivity studies conducted with cells have been reported to have greater predictive value for clinical infection than conventional broth assays (39). A standardized human U937 monocytoid cell infection model was developed to test drugs and combinations with CRL-1072 (21-23). U937 (ATCC CRL-1593) (40) was maintained by in vitro passage in RPMI 1640 medium with 10% fetal bovine serum and gentamicin at 50 µg/ml (growth medium). The cells were expanded in antibiotic-free growth medium, washed, and suspended in antibiotic-free RPMI 1640 medium containing mycoplasma-free 2% human type AB serum (assay medium). For infection studies, 10⁷ U937 monocytoid cells were mixed with sonicated suspensions of MAI containing 10⁸ CFU (in 0.1 ml) in 5 ml of assay medium. Phagocytosis was allowed to occur for 3 h with gentle mixing at 37°C, and the cells were washed with assay medium six times. The cells were then diluted to 10⁶ cells/ml and were plated at 1 ml per well in a 24-well Costar plate. After appropriate addition of drugs in triplicate for each concentration, the plates were incubated at 37°C in 5% CO₂. Fresh medium (1.0 ml/well) was added on day 3.

Acute i.v. toxicity study of CRL-1072 in C57BL/6 mice. The acute intravenous (i.v.) toxicity study was conducted with healthy C57BL/6 mice (Jackson Laboratories, Bar Harbor, Maine) that weighed 18 to 30 g and that were approximately 6 to 8 weeks old at the time of testing. After an acclimation period of at least 3 days, each animal received a single i.v. injection of 25 to 200 mg of CRL-1072 or vehicle equivalent per kg of body weight into the tail vein. Equal numbers of males and females were treated with each dose. Animals were observed for 14 days after treatment for weight loss, hunched posture, ruffled fur, or other signs of distress.

Infection of mice with MAI. Acute infections with two strains of MAI (strains TMC 724 and ATCC 49601) were initiated in beige mice (C57BL/6; bj/bj; Jackson Laboratories). Four- to 6-week-old female mice were infected i.v. with 10⁶ CFU by modifications of previously published protocols (15). Infection with MAI TMC 724 results in the death of untreated beige mice. Infection of mice with MAI 49601 produced a nonlethal infection that was followed by CFU count determination only at the end of 4 weeks by plating organ homogenates on 7H11 agar. Beige mice lethally infected with TMC 724 were treated from day 1 postinfection for 4 weeks. Those infected with ATCC 49601 were left untreated for 7 days so that the bacteria could multiply within organs. These animals had 10⁶ to 10⁷ CFU in their lungs, spleens, and livers at the start of treatment. This pretreatment bacterial load was designed to investigate the bactericidal effects of drugs or their combinations on rapidly progressive established infections (26). Drug treatment was begun on day 8 and lasted for 4 weeks. There was a 2- to 3-day interval between the last day of treatment and killing of the mice. The mice were killed at the end of 4 weeks (strain TMC 724-infected mice) or day 40 (strain ATCC 49601-infected mice), and organ homogenates were plated out on 7H11 agar to enumerate the surviving organisms.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Evaluation of CRL-1072 in combination with antibiotics in cell culture. The MIC of CRL-1072 for MAI growing in broth was found to be 100 µg/ml. The MIC of CRL-1072 for MAI growing in U937 cells was lower, 5.0 µg/ml. A still lower concentration, 0.125 µg/ml, enhanced the bactericidal action of clarithromycin in human U937 monocytoid cell culture (Fig. 1). In the presence of CRL-1072, clarithromycin was found to be bactericidal at concentrations at which it was only bacteriostatic when used alone. The MIC was reduced 16-fold, from 5 to 0.31 µg/ml, by CRL-1072. This produced an FIC of 0.09 for the combination, indicating synergy. Further studies demonstrated that 0.1 µg of CRL-1072 per ml was the lowest dose that could produce synergistic effects with clarithromycin against MAI in cell culture.

View larger version (18K): [in this window] [in a new window]   FIG. 1.   Effects of CRL-1072 and clarithromycin on growth of MAI in human macrophages. U937 monocytoid cells (107 cells/ml) were infected with 108 CFU of a clinical isolate of MAI (strain ATCC 49601) for 4 h, washed, and plated at 106 U937 cells/ml/well. Multiple doses of clarithromycin were added in triplicate on day 0 alone (closed circles) or with CRL-1072 at concentrations of 0.125 µg/ml (open circles) or 0.25 µg/ml (squares). The results on day 5 are shown as mean ± standard deviation log CFU. The horizontal line indicates the baseline day 0 CFU.

View larger version (40K): [in this window] [in a new window]   FIG. 2.   Effects of CRL-1072 with antimycobacterial drugs on growth of MAI in macrophages. U937 monocytoid cells were infected with MAI (strain ATCC 25291 or TMC 724). Five antibiotics were added in triplicate experiments on day 0, each at a sub-MIC of 5 µg/ml with (black bars) or without (white bars) 0.1 µg of CRL-1072 per ml. Results on day 5 are shown as mean ± standard deviation log CFU. The horizontal line indicates the baseline day 0 CFU.

Pharmacokinetic and toxicity studies. An acute i.v. toxicity study of CRL-1072 was carried out with C57BL/6 mice. Mice were injected i.v. with doses of CRL-1072 ranging from 25 to 200 mg/kg. The maximum dose of CRL-1072 at which all animals survived was 125 mg/kg. Beginning at 1 h after injection of 125, 150, 175, or 200 mg of CRL-1072 per kg, the mice demonstrated lethargy or inactivity, hunched posture, increased respiratory rate, squinting of the eyes, and an unkempt appearance. All mice that received 200 mg/kg died between 2 and 3 h after injection. Mice that died following injection of lower doses did so between 3 and 24 h after administration. No animals died after 24 h postinjection.

Evaluation of i.v. CRL-1072 in combination with antibiotics against nonlethal MAI infection in mice. Beige mice were infected with a clinical isolate of MAI (strain ATCC 49601) that causes a nonlethal infection (24, 26). Clarithromycin doses were chosen as suboptimal (50 or 100 mg/kg) and optimal (200 mg/kg) on the basis of studies reported in the literature (24, 26). Treatment was started on day 8 after infection in order to determine the effects of the drugs on an established infection with rapidly growing organisms. CRL-1072 significantly enhanced the bactericidal activities of both doses of clarithromycin in the lungs and spleens of mice (P < 0.05 by Student's t test) (Fig. 3). The CFU counts in the lungs of mice treated with the combination were nearly 2 logs lower than those in the lungs of mice treated with clarithromycin alone at both the 100- and 200-mg/kg doses. The CFU counts in the spleens of these mice treated with the combination were reduced by over 1 log. CRL-1072 by itself had no significant effect.

View larger version (20K): [in this window] [in a new window]   FIG. 3.   Effects of i.v. CRL-1072 on activity of clarithromycin against a nonlethal MAI infection in beige mice. Beige mice in groups of six were infected i.v. with 106 MAI (strain ATCC 49601) on day 0. Beginning on day 8, they were treated with clarithromycin alone or in combination with CRL-1072. Clarithromycin was given by gavage daily for 5 days/week at the doses shown for a total of 20 doses. CRL-1072 was injected i.v. at a dose of 1.0 mg/kg on the same days. The day 7 pretreatment CFU counts are shown by the horizontal line. Results for lungs (dashed lines) and spleens (solid lines) removed on day 40 are shown as mean ± standard deviation log CFU for six mice per group. The symbols are clarithromycin alone (open circles), CRL-1072 alone (no symbols), and clarithromycin plus CRL-1072 (closed circles).

View larger version (45K): [in this window] [in a new window]   FIG. 4.   Effects of i.v. CRL-1072 on activity of rifampin against a nonlethal MAI infection in beige mice. Groups of six beige mice, infected as described in the legend to Fig. 3, were administered rifampin (25 mg/kg) by oral gavage as five daily doses per week for 4 weeks from day 8 (total 20 doses), CRL-1072 on 3 alternate days per week for 4 weeks (total of 12 i.v. doses of 1 mg/kg each), or both. Day 40 CFU counts (mean ± standard deviation log counts; n = 6 mice per group) are shown.

Evaluation of CRL-1072 in combination with clindamycin against lethal MAI infection in mice. Beige mice were infected i.v. with MAI ATCC 25291 and were treated with clindamycin orally alone, CRL-1072 i.v. alone, both drugs, or neither drug. Half (50%) of the mice treated with CRL-1072 and most (75%) of the mice untreated or treated with clindamycin alone died of infection. The combination of clindamycin and CRL-1072 protected 100% of the mice from death. The CFU counts, determined by plating homogenates of organs from both dead and killed mice, demonstrated that CRL-1072 enhanced the activity of clindamycin by reducing the bacterial load by nearly 3 logs in the lungs and over 4 logs in the spleens (P < 0.01 by Student's t test) (Fig. 5). The combination was bactericidal, whereas either agent alone was, at best, bacteriostatic.

View larger version (37K): [in this window] [in a new window]   FIG. 5.   Effects of i.v. CRL-1072 on activity of clindamycin against lethal infection with MAI. Beige mice were infected i.v. with 106 MAI (strain TMC 724) and were treated from day 1 postinfection with clindamycin alone (bars with diagonal stripes), CRL-1072 alone (bars with cross-hatching), both drugs (black bars), or neither drug (white bars). Clindamycin was given by gavage at 50 mg/kg/dose as five daily doses per week for 4 weeks (total = 20 doses). CRL-1072 was given at 5 mg/kg/dose i.v. as five daily doses per week (total = 20 doses). Results for the lungs and spleens at the time of death or killing on day 30 are shown as mean ± standard deviation log CFU (n = 6 mice per group). The horizontal line shows the day 0 CFU counts.

Evaluation of orally administered CRL-1072 alone and with clindamycin against lethal infection. Beige mice were infected i.v. with MAI ATCC 25291, which causes a lethal infection, and were treated orally with clindamycin, CRL-1072, both drugs, or neither drug. All (100%) of the clindamycin-treated mice, 75% of the untreated mice, and 50% of the CRL-1072-treated mice died by day 28 postinfection. However, 100% of mice given both CRL-1072 and clindamycin survived through 28 days. The CFU counts determined for the organs of all surviving and dead mice between days 25 and 28 are shown in Fig. 6. The combination of CRL-1072 and clindamycin produced significantly lower CFU counts in both the lungs and the spleens compared to those produced by monotherapy with either agent (P < 0.05 by Student's t test).

View larger version (20K): [in this window] [in a new window]   FIG. 6.   Effects of oral CRL-1072 and clindamycin on the CFU in lungs and spleens of beige mice lethally infected with MAI. Mice were infected i.v. with 106 MAI (strain TMC 724). They were treated from day 1 postinfection with clindamycin alone (50 mg/kg) given as five daily oral doses per week for 3 weeks (total = 15 doses), CRL-1072 (5 mg/kg) given in saline by oral gavage as five daily oral doses per week for 3 weeks (total = 15 doses), or both. CFU counts were determined for lungs and spleens of mice that died after day 25 or that were killed at day 28 after infection (mean ± standard deviation; n = 6 mice per group). The horizontal line shows the day 0 baseline CFU.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Previous studies reported that copolymers like CRL-1072 are able to disorganize lipids on the surfaces of mycobacteria and increase the level of uptake of antibiotics (4, 5, 22). The present studies were based on the hypothesis that such agents could potentiate the activities of antibiotics against MAI, even though they have little or no activity as single agents. The results surpassed our expectations. The concentration of CRL-1072 required to enhance the activities of antibiotics against MAI in human cells was 0.1 µg/ml, which was 40 times lower than its MIC as a single agent in cell culture and 1,000 times lower than its MIC in broth. Studies with labeled drug showed that high concentrations of CRL-1072 are localized in the liver and spleen. This is consistent with the suggestion that the copolymer becomes localized in macrophages. The concentrating effect of macrophages might explain the increased effect of the agent against intracellular organisms compared to the effect against those growing in broth.

Members of MAI are resistant to antibiotics via two main mechanisms: the natural permeability barriers of their cell walls and mutations acquired as a result of sublethal exposure. Clinically, mutations conferring resistance to clarithromycin may arise in as little as 4 months (14). Combination therapy is used in an effort to enhance bactericidal activity and reduce the number of organisms in which mutations may occur. The present data suggest that CRL-1072 might be of value in combating the emergence of resistant organisms in two ways. First, its ability to increase the bactericidal effect of clarithromycin and other known antimycobacterial agents might reduce the opportunity for mutation. Second, CRL-1072 has an ability to make members of MAI sensitive to some antibiotics to which they are naturally resistant. This was most clearly demonstrated with streptomycin. Clindamycin, another example, is a broad-spectrum antibiotic that is commonly used to treat AIDS patients, but it is not considered effective against MAI. Clindamycin proved to be bactericidal for MAI when it was administered in combination with CRL-1072 in vitro and in vivo.

The i.v. administration of antimycobacterial drugs has value in experimental studies, but it is not desirable clinically. Poloxamers are typically not absorbed well following oral administration. However, Krahenbuhl et al. (27) reported that they can be effective agents for the treatment of experimental toxoplasmosis following oral administration. Araujo and Slifer (1) extended these studies and reported that they potentiate the activities of drugs for the treatment of lethal toxoplasmosis. Consequently, we evaluated oral administration of CRL-1072 alone and in combination with clindamycin for the treatment of lethal MAI infection in mice. The results demonstrated that treatment with the combination of CRL-1072 with clindamycin resulted in 100% survival and significant reductions in CFU counts in the lungs and spleens, whereas treatment with clindamycin alone resulted in no survival. The effect was smaller than that produced by i.v. injection. However, it demonstrates that the absorption of CRL-1072 following oral administration can be sufficient to produce a therapeutic effect.

Earlier development of a nonionic surfactant as a drug that could be used to enhance the efficacies of antibiotics against mycobacteria was halted because of toxicity (12). Poloxamer surfactants are much less toxic than the Triton derivatives used by earlier investigators. All mice survived i.v. injections of doses of 125 mg/kg. A 4-week intravenous toxicity study was conducted in compliance with Good Laboratory Practices regulations (Protocol FRC 530) by Frederick Research Center, Frederick, Md. (13a). The no-effect dosage of CRL-1072 in the 28-day study with mice was greater than 25 mg/kg/day. Synergistic enhancement of the bactericidal effects of clarithromycin and rifampin were observed by using three doses of 1 mg/kg per week. This dosing regimen was based on studies on the efficacies of the drugs in human U937 monocytoid cells and pharmacokinetic studies with mice. A concentration of 0.1 µg/ml was sufficient to produce synergistic effects with clarithromycin, rifampin, amikacin, clindamycin, and streptomycin in human U937 cells. We calculated that the doses used to treat mice would produce concentrations greater than this in tissue. These data suggest that CRL-1072 has an acceptable toxicity profile for further development.