Use of a Recombinant Strain of Mycobacterium avium Expressing β-Galactosidase To Evaluate the Activities of Antimycobacterial Agents inside Macrophages

Abstract

A reliable and low-cost method that enables rapid screening of the activity exerted by new antimicrobial agents on intracellularly growing Mycobacterium avium has been developed. To this aim, a recombinant (lacZ) strain of M. avium expressing the Escherichia coli β-galactosidase gene was used to evaluate, in murine macrophages, the susceptibility of M. avium to common antimycobacterial agents. β-Galactosidase levels, measured in the presence of each of the antibiotics tested, were closely correlated with the number of CFU recovered from the M. avium lacZ strain-infected macrophages.

Mycobacterium avium is a ubiquitous opportunistic pathogen which rarely causes disease in healthy subjects, while it is one of the major causes of disseminated bacterial infections in AIDS patients (5). Although highly active antiretroviral therapy has reduced the incidence of opportunistic infections in immunocompromised subjects, localized or disseminated M. avium infections are still frequent in human immunodeficiency virus-positive patients (3, 7). Chemotherapy of M. avium infection is hindered by the natural resistance of this microorganism to most of the common antimycobacterial drugs (6). Moreover, treatment of AIDS patients usually requires administration of at least two or three different drugs to minimize the emergence of acquired resistance of the infectious M. avium strain during the therapy itself. For these reasons, the identification of new drugs to be used in therapy is of paramount importance.

A low correlation between the efficacy of new drugs against M. avium, assessed by in vitro tests, and their therapeutic effectiveness has been demonstrated (8). Because mycobacteria are intracellular parasites, a better correlation has been obtained by evaluating antimicrobial susceptibility in M. avium-infected macrophages (8).

Conventional methods for determining drug susceptibility are time-consuming, because the number of viable bacteria within the eucaryotic cells after antibiotic treatment is assessed by plating multiple dilutions of cell homogenates and waiting 2 to 3 weeks for the colonies to grow on the solid media. We previously described a recombinant strain of M. avium (the lacZ strain) expressing, as the reporter gene, the Escherichia coli β-galactosidase gene placed under the control of the Mycobacterium bovis BCG heat shock protein 60 promoter (hsp60) (2). The aim of the present study was to evaluate the use of such a recombinant strain of M. avium for rapid susceptibility testing of antimycobacterial agents within murine macrophages.

Relationship between β-galactosidase activity and CFU number.

To evaluate the relationship between β-galactosidase levels and CFU number, the lacZ strain was grown in complete Middlebrook 7H9 broth until the log phase (optical density at 600 nm of 0.4), and twofold dilutions of bacterial suspension were prepared. An aliquot of each dilution was plated on Middlebrook 7H11 agar, while the remainder was sonicated (Ultrasonic Processor XL; Hertz Systems, Farmingdale, N.Y.) at 1-min pulser-on, 15-s pulser-off intervals for a total of 20 min on ice and tested for β-galactosidase activity. Briefly, the reaction mixture was created with Z buffer (0.06 M Na₂HPO₄, 0.04 M NaH₂PO₄, 0.01 M KCl, 0.001 M MgSO₄, 0.05 M β-mercaptoethanol [pH 7.0]) containing 3 mM o-nitrophenyl-β-d-galactopyranoside (ONPG) (Sigma Chemical Co., St. Louis, Mo.) and various amounts of bacterial lysate. The formation of a yellow product was monitored at 30°C and 420 nm in a Perkin-Elmer Lamda 5 spectrophotometer (Norwalk, Conn.). Enzyme activity was expressed as variation of absorbance (ΔA) per minute per milliliter of bacterial lysate. Enzyme levels were proportional to cell concentration in samples containing as few as 10⁶ bacilli (Fig. 1). No β-galactosidase activity was detectable either in homogenates obtained from the M. avium wild type or from M. avium transformed with pROLHYG 60lacZ without the hsp60 promoter (2).

FIG. 1.

Relationship between CFU number and β-galactosidase activity of the M. avium lacZ strain. At different bacterial concentrations, enzyme activity was assayed and expressed as ΔA per minute per milliliter of bacterial homogenate. Error bars show the standard error of the mean of three independent experiments.

Determination of MIC and MBC in vitro.

In order to assess if the recombinant strain exhibited the same susceptibility to three antimycobacterial drugs with different mechanisms of action, as exhibited by the parental strain, MICs and minimal bactericidal concentrations (MBCs) of amikacin (AMK), levofloxacin (LVX), and ethambutol (EMB) in broth were determined by macrodilution susceptibility testing (10). For both M. avium and the lacZ strain, MICs, defined as the lowest concentration of drug that inhibited 99% of bacterial growth, were as follows: 1 μg/ml for AMK and LVX and 1.5 μg/ml for EMB. For both bacterial strains, MBCs, defined as the concentration that reduced the bacterial inoculum by 1 log₁₀, were as follows: 1 μg/ml for AMK and LVX and 3 μg/ml for EMB. When a drug known to be inactive against M. avium (isoniazid) was used in the drug susceptibility test, both parental and recombinant strains grew equally well at concentrations as high as 8 μg/ml, which were usually active against other mycobacterial species (MIC for Mycobacterium tuberculosis, 0.02 to 0.2 μg/ml). These results suggested that electroporation procedures and the genetic manipulation employed to obtain the recombinant strain did not affect the susceptibility of the microorganism to the drugs.

Susceptibility testing in the intracellular milieu.

To evaluate the use of the lacZ strain for rapid drug susceptibility testing under intracellular conditions, the susceptibility of the recombinant strain to three selected drugs was tested by assessing both β-galactosidase levels and CFU count after 7 days of growth in murine macrophages. Briefly, murine spleen cells were seeded in 24-well plates at a density of 2 × 10⁶ cells per cm². After incubation for 7 days at 37°C in humidified air containing 5% CO₂, nonadherent cells were removed, and each well was infected with about 5 × 10⁶ bacilli. Phagocytosis was allowed to occur for 3 h at 37°C. Extracellular bacteria were removed by gentle repetitive washes, and the number of bacteria phagocytized was determined by lysing the monolayers by hypotonic shock in control wells. Lysates from six wells were pooled, and the number of live lacZ strain cells per ml of macrophage lysate was evaluated by plating 10-fold dilutions on solid medium. Different amounts of each antimycobacterial agent were added to the remaining wells to obtain the following final concentrations: AMK; 2, 4, and 8 μg/ml; LVX; 1, 2, and 4 μg/ml; and EMB; 3, 12, 24, and 48 μg/ml. Wells were incubated in complete Iscove's medium (Sigma Chemical Co.) supplemented with 10% fetal calf serum (HyClone Europe, Ltd., Cramlington, Holland). After 7 days of incubation, the medium was discarded, and the monolayers were lysed. Lysates from 12 to 36 wells were pooled, centrifuged, and resuspended in 1 ml of deionized water. An aliquot (0.1 ml) was used to count intracellular bacilli, while the remainder was sonicated to disrupt mycobacterial cells. The crude homogenates were assayed for β-galactosidase activity as described above. Enzyme activity was expressed as ΔA per minute per well, dividing the enzyme activity per milliliter of homogenate by the number of wells used. The MIC and MBC were established by plating serial dilutions of the macrophage lysates onto 7H11 agar medium and counting the colonies after 14 days of incubation at 37°C. The MIC and MBC were defined as described above considering as the inoculum the phagocytized bacterial population ([5 × 10⁵] ± 0.5 CFU/well). For all antimicrobial agents tested, a good correlation between the reduction in β-galactosidase levels and CFU counts was obtained at increasing concentrations of each drug (Fig. 2) (Pearson correlation indices: AMK, R = 0.87, P = 0.0027; LVX, R = 0.97, P = 0.027; EMB, R = 0.95, P = 0.012). As indicated in Fig. 2, a reduction of at least 10 times in β-galactosidase activity in comparison to the M. avium lacZ strain grown intracellularly for 7 days without drugs (0 μg/ml) corresponded at least to the MICs evaluated intracellularly (AMK, 2 μg/ml; LVX, 1 μg/ml; EMB, 12 μg/ml). Paradoxically, at EMB concentrations higher than the MIC, a slight increase in both the CFU number and β-galactosidase activity of the M. avium lacZ strain was observed (Fig. 2). A similar phenomenon was described for other antibiotics, such as vancomycin and teicoplanin, for which, after the saturation point, the amount of drug able to penetrate phagocytic cells was reported to inversely correlate with the extracellular drug concentration (9).

FIG. 2.

Comparison of the β-galactosidase activities of the recombinant M. avium lacZ strain and CFU count in murine macrophages in the presence of three different antimycobacterial drugs. ○, β-Galactosidase activity expressed as ΔA per minute per well; ⧫, CFU per well. The 0-μg/ml point corresponds to the enzyme activity and CFU of the M. avium lacZ strain grown intracellularly for 7 days in the absence of antibiotics. Error bars show the standard error of the mean of three independent experiments.

Rapid tests for assessing in vitro susceptibility of several mycobacterial species to antimicrobial drugs have previously been described based on the use of recombinant strains expressing the eukaryotic luciferase reporter gene (4). However, only one report demonstrated the validity of these tests for M. tuberculosis and M. bovis BCG during growth inside macrophages (1). So far, the possibility of using a recombinant strain of M. avium expressing a reporter gene to assess drug susceptibility inside macrophages has not been reported. Unlike conventional methods based on the colony count on solid medium after 2 to 3 weeks of incubation at 37°C, the proposed method makes it possible to evaluate drug susceptibility within the last day of the incubation period with the tested drug.

The rapidity of the test and the low cost of the substrate for β-galactosidase activity measurement make the test particularly suitable for screening large numbers of new molecules for their activity against M. avium in an intracellular milieu.