|Home | About | Journals | Submit | Contact Us | Français|
Mycobacterium tuberculosis is an intracellular pathogen that persists within macrophages of the human host. One approach to improving the treatment of tuberculosis (TB) is the targeted delivery of antibiotics to macrophages using ligands to macrophage receptors. The moxifloxacin-conjugated dansylated carboxymethylglucan (M-DCMG) conjugate was prepared by chemically linking dansylcadaverine (D) and moxifloxacin (M) to carboxymethylglucan (CMG), a known ligand of macrophage scavenger receptors. The targeted delivery to macrophages and the antituberculosis activity of the conjugate M-DCMG were studied in vitro and in vivo. Using fluorescence microscopy, fluorimetry, and the J774 macrophage cell line, M-DCMG was shown to accumulate in macrophages through scavenger receptors in a dose-dependent (1 to 50 μg/ml) manner. After intravenous administration of M-DCMG into C57BL/6 mice, the fluorescent conjugate was concentrated in the macrophages of the lungs and spleen. Analyses of the pharmacokinetics of the conjugate demonstrated that M-DCMG was more rapidly accumulated and more persistent in tissues than free moxifloxacin. Importantly, therapeutic studies of mycobacterial growth in C57BL/6 mice showed that the M-DCMG conjugate was significantly more potent than free moxifloxacin.
Tuberculosis (TB) still remains a leading cause of death among bacterial infections worldwide (22). The causative agent of TB, Mycobacterium tuberculosis, is a facultative intracellular pathogen that primarily persists within macrophages in the human host, the cells that are involved in dissemination of the infection (18). Intracellular bacilli are generally more difficult to treat because of the limited access of drugs to bacteria within macrophages, necessitating chronic treatment with high therapeutic doses for effective control and treatment of the disease (12). Moreover, TB treatment problems are exacerbated in AIDS patients undergoing chronic, combined therapies for human immunodeficiency virus infection and attendant opportunistic diseases (4). The increased prevalence of single- and multiple-drug-resistant forms of TB has further limited treatment options.
Many of these drug therapy problems could be attenuated or potentially eliminated through selective delivery of anti-TB drugs into infected macrophages, the primary site of infection. Unlike many other cell types, macrophages are known to express high levels of specific receptors on their plasma membrane that bind and internalize their specific target ligands through a variety of uptake mechanisms (6, 11, 19). For example, specific polysaccharide receptors (e.g., mannan receptors, glucan receptors, and galactan receptors) generally bind neutral polysaccharides of bacterial origin and internalize these ligands via specific receptor-mediated phagocytosis (7). Additionally, macrophage scavenger receptors bind anionic macromolecules and use phagocytosis for ligand uptake. These receptors have a high affinity for a wide spectrum of polyanionic molecules including negatively charged polysaccharides (e.g., dextran sulfate, heparin, bacterial lipopolysaccharides, and others), modified lipoproteins, and proteins (8). Chemical labeling of glucans with carboxy or sulfate groups can lead to their selective accumulation by tissue macrophages via scavenger receptor-mediated uptake (2, 21). The affinity and selectivity of these macrophage receptors may offer a unique opportunity for the selective delivery of anti-TB agents conjugated to macrophage receptor ligands. Such an approach might allow the high levels of anti-TB drugs to be concentrated in the main cellular reservoir of the tubercular bacilli, the macrophage, while minimizing exposure of other host tissues to high levels of potentially nonselective and/or toxic agents. In fact, it has been demonstrated recently in a murine TB model that a mannosyl-dextran conjugate of norfloxacin exerted a higher anti-TB effect than norfloxacin alone (17). Also, a p-aminosalicylic acid-bovine serum albumin antibiotic conjugate had superior efficacy compared to the free drug when tested in murine macrophages as well as a guinea pig TB infection model (10).
Previously, we demonstrated that chemical modification of glucans with carboxymethyl groups leads to their selective uptake by tissue macrophage scavenger receptors of the A type (ScR-A) (2, 11). In this study, we prepared a conjugate of the antibiotic moxifloxacin with carboxymethylglucan (CMG), investigated the targeted delivery of this conjugate to infected macrophages, and evaluated its antituberculosis activity. Here we show that the moxifloxacin-CMG conjugate has enhanced uptake into macrophages and increased antimycobacterial activity relative to the free drug.
C57BL/6 male mice (6 to 8 weeks old) were housed under pathogen-free conditions until infection with M. tuberculosis and then transferred to and maintained in a biosafety level 3 physical containment facility. All procedures with animals were approved by the Bioethical Committee of the State Research Center of Virology and Biotechnology Vector (IRB 00015) or the Center for Biologics Evaluation and Research's Institutional Animal Care and Use Committee.
CMG was obtained from Croda USA, Inc., NJ (Cromoist CMG) as a 2% aqueous solution (degree of carboxymethylation = 75%) and was purified by dialysis and lyophilized. The freeze-dried CMG was used in the synthesis of the dansylated CMG (DCMG) and moxifloxacin-conjugated DCMG (M-DCMG). For a detailed description of the synthesis of the moxifloxacin conjugate, see the supplemental material.
J774 macrophage cells were incubated at 4°C or 37°C, 5% CO2, and 100% humidity in triplicate in 12-well Linbro plates for 1 or 2 h on cover glasses (0.5 × 106 cells/ml/well). The cells were incubated in RPMI 1640 medium supplemented with 10% fetal calf serum and 2 mM l-glutamine in the presence or absence of different doses of M-DCMG and excessive concentrations (100-fold greater than the conjugate) of unlabeled ligands for the scavenger receptor CMG or dextran sulfate (DS). Unlabeled CMG, DS, and conjugate were added into the wells in a minimal volume (20 μl/ml of medium). After the incubations, cell monolayers were washed three times with Hanks solution, and the cover glasses were transferred into petri dishes with 2 ml of Hanks solution for further analysis using a fluorescence microscope. Unless otherwise indicated, three independent cell culture experiments were done.
Fluorescence microscopy using Axioplan imaging (Zeiss) under eyepiece magnification of ×10, objective magnification of ×40, and filter sets 0 to 4 was used to measure the fluorescence of the dansylated conjugate. Photography and processing of pictures were carried out with the program Aviovision (version 3). Monochromatic images were used for analysis. The results were expressed in conventional units (range, 1 to 255 conventional units) as the fluorescence levels of individual average cells.
To evaluate quantitatively the uptake of the preparations by macrophages in vitro, cells were incubated with the conjugate and excess concentrations of CMG and DS as described above and then washed four times with phosphate-buffered saline (PBS), lysed with a solution containing 10 mM Tris-HCl (pH 8) and 0.1% Triton X-100, and transferred into reader plates. Fluorescence was measured with a LabSystems fluorescence reader at an excitation wavelength of 380 nm and emission wavelength of 525 nm using a calibration curve constructed at M-DCMG concentrations of 1, 10, 100, and 1,000 ng/ml (200 μl/well).
M-DCMG or moxifloxacin was administered intravenously (i.v.) to C57BL/6 male mice in 0.1 ml of Hanks solution at the doses of 100 and 10 μg/kg of body weight, respectively. After 10 min, the cardiolung complex was removed and placed into ice-cold Hanks solution. A suspension of the cells from bronchoalveolar lavage fluid was obtained in RPMI 1640, and adherent and nonadherent cells were isolated by culturing the cells at 37°C, 5% CO2, and 95% humidity in 12-well Linbro plates for 2 h on cover glasses. After the incubations, adherent cells (macrophages) were washed three times with Hanks solution, and the cover glasses were transferred into petri dishes with 2 ml of Hanks solution for further analysis under a fluorescence microscope. Nonadherent cells were concentrated by centrifugation, resuspended, and then analyzed under a fluorescence microscope similarly to macrophage monolayers. Three independent experiments were done with three mice in each group; two samples were taken from each mouse. Adherent and nonadherent splenocytes were isolated and analyzed in a similar manner.
To obtain pharmacokinetic data, samples of liver and blood serum were obtained at different time points after i.v. administration of 0.1 ml of M-DCMG or moxifloxacin at doses of 100 and 10 μg/kg, respectively, to C57BL/6 mice. Although 10-fold-higher concentrations of M-DCMG were used in these studies since the conjugate contained 11% moxifloxacin by weight, these levels of the compounds represented essentially equivalent drug concentrations. These samples were frozen at −60°C for later high-pressure liquid chromatography (HPLC) analysis. Control samples were obtained from animals given 0.1 ml of 0.9% NaCl. Before HPLC analysis, 100 μl of plasma was mixed with 10 μl of 1.8 M HClO4 in an Eppendorf tube, homogenized with vigorous shaking, centrifuged at 3,000 × g for 5 min, and loaded onto the column. Livers were perfused with phosphate-buffered saline, and a 33% water homogenate was prepared from the right lobe of the liver, frozen at −20°C, and stored for 3 to 4 days until the samples were tested. After thawing, the liver samples were homogenized, and 500 μl of each homogenate was mixed with 2.5 ml of 1 M HClO4, homogenized again, centrifuged at 3,000 × g for 5 min. The pH of supernatants was adjusted carefully to 3.7 to 4.2 with 10 M KOH using vigorous stirring and constant cooling of the samples. Protein was precipitated using potassium perchlorate with centrifugation at 3,000 × g for 5 min. Twenty-microliter portions of the samples were loaded onto the column. The protein contents in the samples were determined using the method of Lowry et al. (9).
Reverse-phase ion-paired HPLC was used to determine moxifloxacin concentrations in biological specimens. Chromatography was performed using Nucleosil C18 5-μm column (250 mm with a 2.0-mm inside diameter [ID]) with a C18 Vydac precolumn (30 to 40 μm, 30 mm with a 2.0-mm ID). Mobile phase A consisted of 20% (vol/vol) acetonitrile and 5 mM tetrabutylammonium phosphate, pH 3.7; mobile phase B contained 50% acetonitrile and 5 mM tetrabutylammonium phosphate, pH 3.7. The gradient profile consisted of elution with buffer A for 1 min with further transition to 70% buffer B for 9 min, elution under isocratic conditions for 5 min, and elution with buffer A for 5 min. Fluorimetric detection was at an excitation wavelength of 296 nm and an emission wavelength of 504 nm. Calibration was carried out at concentration range between 40 ng/ml and 3.0 μg/ml. For this purpose, different amounts of moxifloxacin were added to blood plasma and processed by the procedure described above. Linearity was achieved over all intervals of concentrations used, and the resulting regression coefficient was 0.9996.
The pharmacokinetic parameters [constant of elimination, kel; constant of the pharmacokinetic phase, α(kα); clearance total, Clt; volume of distribution, Vd; volumes of distribution of the central (V0) and peripheral (Vp) compartments; time of half-elimination, t1/2; area under the curve “concentration-time,” AUC; mean residence time, MRT; and mean residence times in the central (MRT0) and peripheral (MRTp) compartments] were calculated with equations of the noncompartmental method of statistical moments using M-IND software developed in the State Scientific Center of Prophylactic Medicine, Moscow, Russia (24).
M. tuberculosis strain Erdman was obtained from the Tarasevich Institute of Standardization and Control (Moscow, Russia) and deposited in the Collection of Cultures of Microorganisms, State Research Center of Virology and Biotechnology Vector (V-1006; Koltsovo, Russia). Mycobacteria were grown under standard conditions at 37°C in Middlebrook 7H9 broth or 7H11 agar medium (Difco Laboratories) containing 10% bovine albumin-dextrose-catalase.
Mycobacterial suspensions were diluted in modified 7H9 broth at a concentration of 1 × 106 CFU/ml. The inoculum size and its purity were verified by plating serial dilutions of the bacterial suspension in triplicate onto 7H11 agar plates supplemented with 10% oleic acid-bovine albumin-dextrose-catalase enrichment medium (BBL, Cockeysville, MD) and incubating at 37°C in ambient air for 3 weeks.
For the intravenous TB challenge studies, C57BL/6 mice were subdivided into seven groups: a control group consisting of 30 mice infected intravenously through a caudal vein with 105 cells of M. tuberculosis strain Erdman suspended in 0.3 ml of PBS; three moxifloxacin-treated groups of 20 mice each, infected analogously to the control and treated daily i.v. 7 days later over the next 7 or 15 days with moxifloxacin at a dose of 5, 50, or 200 μg/kg of body weight; and three M-DCMG-treated groups of 20 mice each treated with M-DCMG at the same doses (5, 50, and 200 μg/kg of the conjugate) and time schedule as the moxifloxacin-treated groups. The suboptimal dosages of moxifloxacin were chosen to maximize the differences between the antibiotic- and the conjugate-treated groups (25).
For the aerogenic-challenge experiments, C57BL/6 mice were infected by the aerosol route with M. tuberculosis Erdman suspended in phosphate-buffered saline with 0.04% Tween 80 at a concentration known to deliver about 200 CFU in the lungs over a 30-min exposure period in a Middlebrook chamber (Glas-Col, Terre Haute, IN). At 10 days postchallenge, mice (five per group) were given i.v. either CMG (0.2 or 5 mg/kg), M-DCMG (0.2 or 5 mg/kg), or moxifloxain (0.2, 10, or 100 mg/kg) for 14 days.
To assess the bacterial growth in vivo, mice were sacrificed, and the lungs, liver, and spleens were removed aseptically and homogenized separately in 5 ml of 0.04% Tween 80-PBS. The number of viable organisms was determined by serial 10-fold dilutions and subsequent inoculation onto 7H11 agar plates. The numbers of CFU in the infected organs were determined after 14 to 21 days of incubation at 37°C in sealed plastic bags.
One-way analysis of variance by the Dunnett test or unpaired, two-tailed t test was applied to evaluate the significance of differences between groups.
To characterize the activity of macrophage-targeted drug conjugates, dansylated moxifloxacin-CMG was prepared. The dansyl group was added to the conjugate as a fluorescent tag so that uptake of the conjugate could be assessed. The uptake of the conjugate by J774 macrophage cells and the effect of standard competitive ligands for scavenger receptors upon M-DCMG uptake were evaluated using fluorescence microscopy of the cells and fluorescence photometry of cell lysates. Fluorescence microscopy showed that incubation of the cells in the presence of 5 μg/ml of M-DCMG for 1 h resulted in a significant increase of the cellular fluorescence compared to controls (Fig. (Fig.1).1). Increasing the concentration of the conjugate up to 50 μg/ml resulted in further increases in the fluorescence of the cells (Table (Table1).1). Interestingly, the addition of the unlabeled competing ligand dextran sulfate (Fig. (Fig.1,1, Table Table1)1) or CMG to the incubation medium at concentrations exceeding the conjugate concentration by 100-fold resulted in decreases of the cellular fluorescence to control levels.
Comparison by fluorimetry of the intracellular accumulation of the conjugate, relative to the free antibiotic, is shown in Fig. Fig.2.2. For these studies, since moxifloxacin represented 11% of the total conjugate weight, equivalent drug doses were achieved by adding 10-fold-elevated concentrations of M-DCMG. As seen in Fig. Fig.2,2, more M-DCMG than free moxifloxacin accumulated in the J774 macrophages. This fluorimetric analysis also confirmed the fluorescence microscopy data; accumulation of M-DCMG was almost completely prevented by addition of CMG or dextran sulfate into the incubation medium in concentrations exceeding conjugate concentration by 100-fold. In contrast, the fluorescence of the cells in the presence of moxifloxacin was not changed by the addition of CMG or dextran sulfate (Fig. (Fig.2).2). Overall, these data strongly suggest that M-DCMG can be efficiently taken up by macrophages in vitro via scavenger receptors.
To evaluate the cellular specificity of the conjugate uptake, the M-DCMG contents in lung and spleen macrophages, compared to appropriate nonadherent cells, were determined by fluorescence microscopy. For these studies, suspensions of splenocytes and bronchoalveolar cells were obtained 10 min after i.v. administration of moxifloxacin or M-DCMG. Adherent and nonadherent cells were isolated and analyzed under a fluorescence microscope. In animals injected with Hanks solution or with moxifloxacin, the fluorescence of bronchoalveolar nonadherent cells and alveolar macrophages (adherent cells) was minimal. Conversely, in mice given M-DCMG, the fluorescence of alveolar macrophages was, respectively, 2.7 and 2.1 times greater than the fluorescence of macrophages from Hanks solution- or moxifloxacin-injected mice (Fig. (Fig.3).3). Importantly, alveolar macrophages accumulated more than twofold-higher amounts of M-DCMG than did nonadherent cells.
The pharmacokinetics of the conjugate and free moxifloxacin were compared in the C57BL/6 murine model. The “concentration-time” curve for M-DCMG concentration changes in blood serum was nonlinear, so that an α phase (first 9 min) and a β phase could be easily distinguished (Fig. (Fig.4A).4A). By contrast, the pharmacokinetics for moxifloxacin were substantially more linear (Fig. (Fig.4B4B).
The kαs of the conjugate and of moxifloxacin were calculated as blood serum concentrations changed during the first 9 min after i.v. injection. The kel, as well as other pharmacokinetic parameters characterizing conjugate and moxifloxacin elimination were calculated for the terminal period of the “concentration-time” curve (12 to 180 min). The kα (the overall distribution of the compound) for the conjugate was greater than that for free moxifloxacin (Table (Table2).2). This result implies that M-DCMG was distributed from the circulation to the tissues more rapidly than free moxifloxacin. The V0 of M-DCMG was about twofold smaller than total body water (80% of body weight). In contrast, free moxifloxacin had a V0 comparable to the total body water of an animal. Hence, this parameter for M-DCMG was nearly two times less than that for moxifloxacin. Interestingly, the Vp of the conjugate was seven times higher compared to the Vp value for the free antibiotic. Again this result suggests that the conjugate was more widely distributed to the peripheral tissues than the free drug. Overall, comparison of the kα, V0, and Vp parameters for M-DCMG to those of the free antibiotic suggests an accelerated accumulation of the conjugate in the tissues. Substantially shorter MRT0 and much longer MRTp also support this conclusion. The parameters of elimination (t1/2, kel) for M-DCMG show a tendency (though statistically nonsignificant) for the delayed elimination of the conjugate from the animal, compared to the free antibiotic. This delayed elimination of the conjugate may contribute to the longer M-DCMG retention in the tissues.
Injections with either the conjugate or moxifloxacin began 7 days after i.v. infection of animals with M. tuberculosis strain Erdman and continued for the next 14 days, and the antituberculosis activities of these preparations were evaluated by assessing the bacterial burdens in relevant organs at 14 days after the initiation of therapy (Fig. (Fig.5).5). Three different drug doses (5, 50, and 200 μg/kg) were investigated in this study. In previous murine studies, the therapeutic effects of moxifloxacin were seen at higher 50- to 100-mg/kg treatment levels (3, 13, 15, 25). At the 200-μg/kg dose of moxifloxacin, no significant decreases in mycobacterial CFU were detected in the organs of TB-infected animals, compared to untreated controls. In contrast, a dose-dependent therapeutic effectiveness was seen in M-DCMG-treated mice. Splenic CFU were significantly decreased in the 50-μg/kg group. Importantly, significant reductions of 1.5 to 2 log10 CFU were detected in all tested organs of infected mice given the 200-μg/kg doses of moxifloxacin bound to CMG.
The effectiveness of the drug conjugate was further evaluated using an aerogenic-TB-challenge model. In the initial aerogenic-infection study, mice were given i.v. injections for 14 days of either the moxifloxacin conjugate (200 μg/kg), the carboxymethyl glucan carrier (200 μg/kg), or increasing doses of moxifloxacin beginning 10 days after challenge with 200 CFU of virulent M. tuberculosis. For this study, only lung CFU values were assessed because dissemination of TB to other organs can be variable during the first 3 weeks after a low-dose pulmonary infection. As seen in Fig. Fig.6A,6A, moxifloxacin was not active at the 200-μg/kg dose but showed significant antituberculosis activity at 10 mg/kg and was extremely effective when given at 100 mg/kg. In contrast, a significant reduction in mycobacterial growth (0.7 log10) was seen in mice injected with 200 μg/kg of the moxifloxacin conjugate; the effectiveness of the conjugate at this dose was statistically equivalent to the activity of moxifloxacin given at 10 mg/kg. Interestingly, injection of the CMG ligand at the 200-μg/kg dose also significantly reduced mycobacterial growth (57%, P < 0.05) compared to untreated controls.
In a second aerogenic-infection study, the potency of a higher dose of the conjugate was evaluated. Eleven days after an aerogenic challenge with M. tuberculosis Erdman, mice were intravenously injected with either increasing doses of moxifloxacin, the moxifloxacin conjugate (5 mg/kg), or the CMG carrier (5 mg/kg). The moxifloxacin growth inhibition results from the second aerogenic-challenge experiment were similar to the initial data; a reduction of greater than a 2 log10 in lung bacterial CFU was seen at the highest drug concentration, while much less antituberculosis activity was detected at the 0.2- and 10-mg/kg moxifloxacin doses. Treatment with 5 mg/kg conjugate resulted in a significant reduction in mycobacterial growth (0.82 log10) in the lung, compared to naive controls. In addition, the level of lung bacterial CFU detected in the 5-mg/kg conjugate group was significantly lower (5.32 ± 0.02 log10, P < 0.05) than the CFU value for the 10-mg/kg moxifloxacin-treated mice (5.72 ± 0.04 log10). These data again demonstrate the enhanced potency of the moxifloxacin conjugate relative to the free drug. Surprisingly, injection of the CMG carrier at the elevated 5-mg/kg dose had no effect on growth of the mycobacteria in the lungs. Importantly, the lung mycobacterial CFU were significantly lower in mice treated with the moxifloxacin conjugate at 5 mg/kg compared to mice given the same dose of CMG (0.81 log10 difference, P < 0.01).
The medical importance of killing intracellular bacteria persisting in macrophages has led to the development of antibiotic conjugates that can be targeted directly into these cells (17). We and others have demonstrated that chemical modification of glucans with carboxymethyl groups leads to their selective uptake by tissue macrophages mediated by ScR-A (2, 16). The present study was designed to determine the potential benefit of using a conjugate of DCMG and the antibiotic moxifloxacin as a potential approach for treating TB. In our experiments, moxifloxacin was bound to DCMG and the targeted delivery and potency of this conjugate were tested in vitro and in vivo. Fluorescence-microscopic data and fluorimetric results showed that the conjugate accumulates in J774 macrophages much more effectively than free moxifloxacin. Since excess concentrations of CMG and DS, known ligands to ScR-A, inhibited uptake of M-DCMG but had no effect upon free antibiotic accumulation, M-DCMG likely enters macrophages through scavenger receptors (2, 20). Our in vivo data are also consistent with macrophage targeting of the M-DCMG conjugate. Fluorescence microscopy demonstrated that, as early as 10 min after i.v. administration of the conjugate, the fluorescence of alveolar macrophages was significantly increased, whereas fluorescence of other cells in the bronchoalveolar lavage fluid did not change. Conversely, the fluorescence of these cells was not significantly increased 10 min after injection of moxifloxacin at an antibiotic dose equivalent to M-DCMG.
Most importantly, the in vivo chemotherapy studies clearly demonstrated that the moxifloxacin bound to conjugate was significantly more effective than free moxifloxacin. Using an intravenous-infection model, nearly 2 log10 reductions in organ mycobacterial CFU were detected after treatment with 200-μg/kg doses of moxifloxacin in the conjugate form. At the same concentration, free moxifloxacin was not active against the tuberculous infection. In two separate experiments, the moxifloxacin conjugate was also more effective than the free drug (per mg) in mice that had been challenged by the aerosol route with a low dose of virulent M. tuberculosis. A greater-than-80% reduction in lung bacterial burden was detected in mice given the moxifloxacin conjugate for 2 weeks following an aerogenic infection. Overall, the greater therapeutic effectiveness of the moxifloxacin bound to conjugate compared to free moxifloxacin indicates that the conjugate has an enhanced capacity to reach and maintain effective concentrations at the sites of infection. It is not unexpected that the M-DCMG conjugate is less effective in mice infected by the aerosol route than in mice challenged intravenously. It has been previously shown that M. tuberculosis is much more virulent for mice when given by the respiratory route compared to the intravenous route (14). Also, although the increased efficacy of the conjugate is likely primarily due to its targeting to the macrophage, the known immunologically enhancing effect of CMG may contribute to the potency of M-DCMG, especially at lower doses (1, 5, 23). Additional studies are needed to precisely define the antimycobacterial role of CMG in the moxifloxacin conjugate.
In summary, these results demonstrate that a prodrug generated by conjugating moxifloxacin to CMG has enhanced antituberculosis activity relative to the free drug. Based on these results, further testing and characterization of the CMG-moxifloxacin conjugate is warranted and consideration should be given to development of CMG conjugates with other antituberculosis medications.
This work was supported by grant BTEP/ISTC 39/2174p.
We are thankful to Robert N. Comber of Croda USA, Inc., for the kind gift of Cromoist CMG.
†Supplemental material for this article may be found at http://aac.asm.org/.