In Vitro Antimalarial Activity of Azithromycin, Artesunate, and Quinine in Combination and Correlation with Clinical Outcome

Patient samples.Blood samples were obtained from adult men and nonpregnant women aged 20 to 65 years, recruited at the Hospital for Tropical Diseases, Mahidol University, Bangkok, Thailand, who presented with acute uncomplicated falciparum malaria, who were otherwise healthy, and who had parasite densities between 100 and 100,000 asexual P. falciparum parasites per μl (equivalent to ca. 0.002 to 2% parasitemia). Parasite samples with more than 1% parasitemia were diluted with uninfected red blood cells. Patients were not admitted to the study if any of the following criteria were present: pregnant and nursing women; signs and symptoms of severe malaria; mixed malaria infection on admission; malaria drug therapy or blood transfusions in the preceding 30 days; laboratory evidence or history of significant cardiovascular, hepatic, or renal functional abnormality; severe vomiting; reported alcohol or drug abuse; any clinically significant illness; or likelihood of requiring treatment with drugs not permitted by the protocol. Written informed consent was obtained from all study participants and the protocol was approved by the appropriate ethical review boards.

All of the 97 fresh parasite isolates obtained before the initiation of treatment from patients with microscopically confirmed P. falciparum monoinfections were tested in duplicate for their drug sensitivity to azithromycin (AZ), dihydroartemisinin (DHA), mefloquine (MQ), quinine (QN), and chloroquine (CQ). Checkerboard drug interaction assays were performed on four randomly selected samples with the combination of AZ with either DHA or QN.

Drug susceptibility testing.All samples were tested in the histidine-rich protein 2 (HRP2) in vitro drug susceptibility assay (12). Fresh P. falciparum parasite isolates were cultured in the presence of twofold serial dilutions of the antimalarial drugs AZ (781.25 to 50,000 ng/ml), DHA (0.15 to 9.38 ng/ml), MQ (3.22 to 206.27 ng/ml), QN (24.11 to 1,543.21 ng/ml), and CQ (16.14 to 1,033.06 ng/ml) at 1.5% hematocrit in complete RPMI 1640 with 0.5% Albumax (Albumax I; Gibco, Bangkok, Thailand) and 25 mg of gentamicin/liter without freezing, washing, dilution, the addition of serum, or preculturing. Checkerboard assays with the drug combinations were performed by diluting AZ (97.7 to 12,500 ng/ml) vertically and either QN (3.01 to 385.80 ng/ml) or DHA (0.02 to 2.34 ng/ml) horizontally on an 8×8 checkerboard. Stock solutions of the test drugs at 1 mg/ml were prepared in 70% ethanol and then diluted with distilled water to the test concentration. Serial twofold dilutions (seven concentrations and one drug-free control well) of the drugs (25 μl/well) were dispensed into standard 96-well microculture plates (Costar 3599) manually or by a semiautomated microdilution technique. The culture plates were then dried and stored at 4°C for up to 4 weeks. A total of 200 μl of cell medium mixture was added to each well, and the plates were incubated for 72 h in a gas mixture (5% CO₂, 5% O₂, 90% N₂) at 37.5°C. The plates were subsequently freeze-thawed twice to obtain complete hemolysis.

Parasite growth inhibition was quantified by using a highly sensitive HRP2 enzyme-linked immunosorbent assay (ELISA) based on two commercially available monoclonal antibodies (Immunology Consultants Laboratory, Inc., Newberg, OR) directed against P. falciparum-specific HRP2: MPFM-55A, an immunoglobulin M antibody used as the capture antibody, and MPFG-55P, a horseradish peroxidase-conjugated immunoglobulin G antibody, which was used as the indicator antibody. The ELISA was performed as previously described and allows for the testing of parasitized blood samples down to ca. 0.002% parasitemia (13). Spectrophotometric analysis was performed with an ELISA plate reader (SpectraMAX 340 microplate spectrophotometer; Molecular Devices, Sunnyvale, CA) at an absorbance maximum of 450 nm.

Data analysis.The 50 and 90% inhibitory concentrations (IC₅₀s and IC₉₀s, respectively) were calculated from optical density readings by nonlinear regression analysis. ICs were used to calculate fractional inhibitory concentrations (FICs) as previously described (2). Isobolograms were plotted to demonstrate synergism and/or antagonism for drug combinations. Activity correlations were calculated by nonparametric correlation analysis (Spearman). The significance for differences between two groups of continuous data was calculated by using the Mann-Whitney U test.

Drug interaction studies on fresh clinical isolates.Four clinical isolates were successfully tested in checkerboard interaction studies with azithromycin and either dihydroartemisinin or quinine. With an overall mean FIC of 0.84, the interaction between dihydroartemisinin and azithromycin was classified as additive, with a tendency toward synergism (Fig. 1). The mean FICs were 0.84, 1.06, 0.73, and 0.72, with individual datum points ranging from 0.47 to 1.20. This similarly applied to the combination of quinine with azithromycin (Fig. 2). Here, the mean FIC was slightly lower at 0.77. The FICs were 0.85, 0.73, 0.68, and 0.84, with individual datum points ranging from 0.37 to 1.05. Drug combination ratios between dihydroartemisinin and azithromycin of 1:34 to 1:1,198 (mean, 1:547) resulted in the highest levels of synergy (i.e., the lowest FICs). For azithromycin and quinine, the optimal combination ratios were 1:4 to 1:124 (mean, 1:44).

• Open in new tab
• Download powerpoint

FIG. 1.

Isobolograms for the checkerboard assays with azithromycin and dihydroartemisinin against clinical field isolates of P. falciparum, showing additive to synergistic interactions. The mean FICs for the clinical isolates were 0.84, 1.06, 0.73, and 0.72 for the samples AZ10011003, AZ10011008, AZ10011017, and AZ10011022, respectively.

• Open in new tab
• Download powerpoint

FIG. 2.

Isobolograms for the checkerboard assays with azithromycin and quinine against clinical field isolates of P. falciparum, showing additive to synergistic interactions. The mean FICs were 0.85, 0.73, 0.68, and 0.84 for the samples AZ10011003, AZ10011008, AZ10011017, and AZ10011022, respectively.

Clinical correlates.The IC₅₀s for individual patient isolates were compared to the treatment outcome in the corresponding patients. In the low-dose quinine arm three RIII failures (early treatment failures) were seen. These failures had QN IC₅₀s (116.4, 103.3, and 93.2 ng/ml) that were consistently greater than the geometric mean (64.4 ng/ml). At the same time two of these three isolates were classified as highly resistant to mefloquine (i.e., IC₅₀ > 30 ng/ml), a drug that is widely used in Thailand and that is known for its cross-resistance with quinine (R = 0.44; P < 0.001 in the present study). In patients treated with quinine the quinine IC₅₀s were almost twice as high (114.8 ng/ml; P = 0.01) in isolates that later led to failures (n = 6; three RIs with 126.9 ng/ml and three RIIIs with 103.9) than in those which did not (63.5 ng/ml). In patients treated with artesunate combinations the IC₅₀ in the five patients who had recrudescences within the 28-day follow-up was 0.81 ng/ml (compared to 0.45 ng/ml for the geometric mean in patients who were cured; P > 0.05). Azithromycin IC₅₀s in isolates from patients who later developed recrudescences (2,953 ng/ml) were similar to those in cured patients (2,627 ng/ml; P > 0.05), suggesting little influence of the azithromycin in vitro drug sensitivity on clinical outcome.

Individual parasite clearance times (PCTs), i.e., the time from administration of the first dose of the study drug until complete parasite clearance, in the quinine arms were significantly (R = 0.53; P = 0.001) correlated with quinine IC₅₀s. With a correlation coefficient of R = 0.34 (P = 0.038) the relation between PCTs and DHA IC₅₀s was less significant. No correlation was seen between azithromycin IC₅₀s and the PCT in the quinine (R = −0.12; P > 0.05) or artesunate (R = −0.10; P > 0.05) arms.

Correlations.Individual ICs calculated for all drugs tested in parallel were correlated by nonparametric correlation analysis to determine cross-sensitivity and/or cross-resistance patterns between the drugs at the IC₅₀ level. Azithromycin IC₅₀s did not show any significant activity correlations with any of the other drugs, reflecting its unique chemical structure and mode of action among these drugs. Significant activity correlations were found between the IC₅₀s of DHA and MQ (R = 0.34; P = 0.003), as well as QN (R = 0.50; P < 0.001). Mefloquine activity at the IC₅₀ level correlated with the IC₅₀s of QN (R = 0.44; P < 0.001).