In Vivo and In Vitro Sensitivity of Fasciola hepatica to Triclabendazole Combined with Artesunate, Artemether, or OZ78

Abstract

Triclabendazole resistance is continually documented from livestock, and hence new treatment strategies for Fasciola hepatica infections are needed. We investigated the effect of triclabendazole combined with artesunate, artemether, or OZ78 compared to that of monotherapy against adult and juvenile F. hepatica in rats. In vitro experiments with triclabendazole and its sulfoxide and sulfone metabolites, each in combination with the peroxides, complemented our study. F. hepatica-infected rats were subjected to single drugs or drug combinations 3 to 4 weeks (juvenile flukes) and >8 weeks (adult flukes) postinfection. Negative binomial regressions of worm and egg counts were used to analyze dose-response relationships and whether the effects of drug combinations were synergistic or antagonistic. The in vitro assays were evaluated by means of viability scales based on fluke motility. Fifty percent effective dose values of 113.0, 77.7, 22.9, and 2.7 mg/kg of body weight were calculated for monotherapy with artesunate, artemether, OZ78, and triclabendazole, respectively, against adult F. hepatica. Likelihood ratio tests revealed synergistic interactions (P < 0.05) of combinations of triclabendazole (2.5 mg/kg) plus artesunate or artemether on adult worm burden. Antagonistic effects on the adult burden and egg output were observed when a lower triclabendazole dose (1.25 mg/kg) was combined with the artemisinins. No significant interactions (P = 0.07) were observed for OZ78 and triclabendazole combinations and between the triclabendazole effect and the effects of the other partner drugs on juvenile worms. Our in vitro studies of adult worms agreed with the in vivo results, while the in vitro analysis of juvenile worms revealed greater interactions than observed in vivo. In conclusion, single-agent triclabendazole demonstrated a more potent in vivo and in vitro fasciocidal activity than the experimental drugs artesunate, artemether, and OZ78. When combined, synergistic but also antagonistic effects depending on the doses administered were observed, which should be elucidated in more detail in future studies.

The food-borne trematodes Fasciola hepatica and F. gigantica are the causative agents of fascioliasis (fasciolosis). Fasciola hepatica parasitizes a wide spectrum of domestic and wild animals (e.g., sheep, cattle, rats, and deer), and it causes a huge economic loss of $3 billion annually to the agriculture sector worldwide through losses of milk and meat yields (26, 31). In addition, an estimated 90 million people are at risk of fascioliasis and up to 17 million individuals are infected (20).

Due to its excellent safety profile and the high activity against both juvenile and adult liver flukes, triclabendazole (Fasinex, Egaten) is the drug of choice for the treatment of human and veterinary fascioliasis. It is worrying, however, that triclabendazole resistance has been documented from Australian sheep farms since the mid-1990s and recently also from several western European countries (10). Furthermore, F. hepatica strains resistant to three alternative fasciocidal drugs (rafoxanide, closantel, and luxabendazole) have been isolated (11). Due to the rapid spread of resistance, the small arsenal of fasciocidal drugs, and the absence of effective vaccines in field conditions, there is a pressing need for the discovery and development of novel drugs or drug combinations (27). The use of drug combinations is an excellent strategy to avoid or delay drug resistance, since different drug targets are attacked simultaneously. Furthermore, drug combinations often are characterized by an increased activity and tolerability compared to that of monotherapy. Therefore, drug combinations are widely used in the treatment of infectious diseases (e.g., malaria, HIV, and tuberculosis), cancer, and chronic disorders (e.g., cardiovascular disease, diabetes, and pain management) (5, 8, 12, 30, 39). Combination chemotherapy has been tested in experimental studies using triclabendazole-susceptible and -resistant strains of F. hepatica. The combined treatment of triclabendazole plus clorsulon or luxabendazole was found to act synergistically. For example, the combination of triclabendazole and clorsulon, administered at one-fifth their recommended dosages, was highly effective against triclabendazole-resistant F. hepatica in sheep. This combination reduced 95% of the worms, while the single drugs achieved a worm burden reduction (WBR) of only approximately 30% (11, 28).

Recently, the semisynthetic artemisinins artemether and artesunate and the synthetic peroxide OZ78 were described to have excellent fasciocidal properties in rats (19, 21). Importantly, artemether and OZ78 cured a triclabendazole-resistant F. hepatica infection in the rat model (22). Contradictory results were obtained in sheep. While OZ78 failed to cure a chronic F. hepatica infection in sheep (16), artemether and artesunate achieved high worm burden reductions in naturally F. hepatica-infected sheep (18, 23).

The aim of the present investigation was to study the potential of the combined treatment of triclabendazole with artesunate, artemether, or OZ78 in F. hepatica-infected rats harboring juvenile or adult infections. Comparisons were made to monotherapies, and negative binomial regression was used for the analysis of egg burden and worm counts. In vitro studies complemented our work. We investigated the in vitro effect of triclabendazole and its sulfoxide and sulfone metabolites, each combined with the three peroxidic drugs (artesunate, artemether, and OZ78) against juvenile and adult F. hepatica flukes.

MATERIALS AND METHODS

F. hepatica rat model.

Female Wistar rats (n = 222; age = 3 weeks; weight ∼100 g) were purchased from Harlan (Itingen, Switzerland). Rats were kept in groups of five in type-4 Makrolon cages under environmentally controlled conditions (temperature, ∼22°C; humidity, ∼70%; light/dark cycle, 12/12 h) with free access to water and rodent food (Rodent Blox from Eberle NAFAG; Gossau, Switzerland). F. hepatica metacercariae (Pacific Northwest wild strain) were purchased from Baldwin Aquatics (Monmouth, OR). After 1 week of adaptation, each rat was infected by oral gavage with 22 ± 2 metacercariae. The present study was authorized by the local veterinary agency based on national regulations (permission no. 2070).

Drugs.

Triclabendazole, triclabendazole-sulfoxide, and triclabendazole-sulfone were the products of Novartis Animal Health (Basel, Switzerland). Artesunate was obtained from Mepha AG (Aesch, Switzerland). Artemether was a gift of Dafra Pharma (Turnhout, Belgium). OZ78 was synthesized at the College of Pharmacy, University of Nebraska Medical Center (Omaha, NE). The drugs were freshly prepared as suspensions containing 7% (vol/vol) Tween 80 (Sigma-Aldrich, Buchs, Switzerland), 3% (vol/vol) ethanol 96% (Merck, Darmstadt, Germany), and tap water for the in vivo studies. For the in vitro experiments, stock solutions (3 and 10 mg/ml) were prepared in 100% (vol/vol) dimethylsulfoxide (DMSO).

In vivo drug treatment.

At least 4 rats per drug and dosage harboring adult F. hepatica (>8 weeks postinfection) were treated orally with triclabendazole, artesunate, artemether, or OZ78 alone. The drugs were given in the following dosages: artesunate and artemether, 200 and 100 mg/kg of body weight; OZ78, 100, 50, and 25 mg/kg; triclabendazole, 10, 5, 2.5, and 1.25 mg/kg (n = 44 rats in total). For the combination chemotherapy experiments, 4 rats were each given 1.25 mg/kg triclabendazole plus either artesunate (25, 50, 100 mg/kg), artemether (25, 50, 100 mg/kg), or OZ78 (25 and 50 mg/kg) (n = 32 rats in total). Finally, 4 rats each were treated with 2.5 mg/kg triclabendazole in combination with artesunate (6.25, 12.5, 25, 50, 100, and 200 mg/kg), artemether (6.25, 12.5, 25, 50, and 100 mg/kg), or OZ78 (12.5, 25, and 50 mg/kg) (n = 56 rats in total).

Four rats harboring juvenile F. hepatica (3 to 4 weeks postinfection) were treated each with single oral doses of either triclabendazole (2.5 and 5 mg/kg), artesunate (25 and 50 mg/kg), artemether (25 and 50 mg/kg), or OZ78 (25 and 50 mg/kg) (n = 32 rats in total). The following drug combinations were studied (4 rats per dosage): 2.5 mg/kg triclabendazole plus 25 mg/kg artesunate, artemether, or OZ78; 5 mg/kg triclabendazole plus 50 mg/kg artesunate, artemether, or OZ78 (n = 24 rats in total).

Untreated rats (n = 34) served as controls, as illustrated in Fig. 1. One week posttreatment, the number of F. hepatica worms per rat was determined by the examination of the bile duct and liver following the necropsy of rats (13).

FIG. 1.

Distribution of adult (A) and juvenile (B) F. hepatica worm counts in untreated control rats. The upper and lower limits of the boxes correspond to the interquartile range, the value in the middle to the median, and the limits of the whiskers correspond to the adjacent values, i.e., the largest value below the 25th percentile and the smallest value above the 75th percentile. Numbers in parentheses are the numbers of control animals in each experiment.

Egg excretion analysis.

A fecal sample was collected from each chronically F. hepatica infected rat (n = 132) 1 day before treatment and 7 days after treatment to calculate the egg burden reduction. Additionally, one fecal sample was collected from 85 of the 132 rats 2 months prior to treatment. The reason for including this additional analysis was to obtain information on the variation in egg output independently of drug effect. The rats were housed individually for 1 night to obtain > 2 g of stool. The FLOTAC double technique was used to analyze egg counts as described in previous publications (4, 7). A flotation solution prepared with 685 g ZnSO₄·7 H₂O (Merck, Darmstadt, Germany) and 685 ml deionized H₂O with a specific gravity of 1.35 was used. Each fecal sample was analyzed using a dilution ratio of 1:100 (multiplication factor of 20), resulting in a diagnostic sensitivity of 20 eggs per gram fecal sample (EPG).

In vitro drug assays.

Adult (>8 weeks postinfection) F. hepatica flukes were recovered from the central bile duct of rats and incubated for 2 h in RPMI 1640 culture medium (Gibco, NY) at 37°C in an atmosphere of 5% CO₂. Adult worms were placed in single wells of 12-well plates (Costar 3512) containing 3 ml RPMI 1640 and supplemented with 1% (vol/vol) antibiotics (50 μg/ml streptomycin and 50 U/ml penicillin; Gibco, NY) and 80 μg/ml hemin (17). Adult F. hepatica were incubated in the presence of either triclabendazole, triclabendazole-sulfoxide and triclabendazole-sulfone (15 μg/ml each), artesunate, artemether, or OZ78 (50 μg/ml each). For the drug combination studies of adult F. hepatica, triclabendazole or one of its two metabolites (15 μg/ml each) was combined with a peroxidic drug (50 μg/ml each).

Juvenile F. hepatica (3 to 4 weeks postinfection) were recovered from the liver of infected rats. The rat livers were pressed between transparent plastic films, and the flukes were collected using a binocular. The juvenile flukes were incubated in 48-well plates (Costar 3548, Corning, NY) in 1 ml supplemented RPMI 1640, with one F. hepatica in each well. The effects of single treatments of triclabendazole, triclabendazole-sulfoxide and triclabendazole-sulfone (30 μg/ml each), artesunate, artemether, and OZ78 (100 μg/ml each) were studied. For the drug combination experiments using juvenile F. hepatica, each triclabendazole derivative (30 μg/ml each) was combined with one of the semisynthetic artemisinins or OZ78 (100 μg/ml each).

Five to six flukes were examined per drug and drug combination in two (adult F. hepatica) and two or three (juvenile F. hepatica) independent experiments (n = 10 to 18 flukes per concentration), and the arithmetic mean was calculated. Six control flukes, without drug exposure, were included in each experimental set. The control well contained the highest concentration of DMSO used.

The viability of adult flukes was scored after 24, 48, and 72 h using the following scale: 3, normal movements; 2, reduced activities; 1, very weak activities, detected only by means of microscopic magnification (20×); and 0, death of worm (absence of movements for 2 min using a microscope [20×]). The viability of juvenile F. hepatica was graded as follows: 2, normal activity (microscopic magnification, ×20); 1, weak activity (microscopic magnification, ×80); and 0, death (absence of movements for 2 min under a microscope [×80]).

Statistical analysis.

For the statistical analyses we included three datasets with a total of 83 worm counts from previous experiments carried out with F. hepatica-infected rats in our laboratories (19, 21). This data set contained single-drug treatments of 20 rats harboring juvenile F. hepatica infections (3 to 4 week postinfection), 37 rats harboring adult F. hepatica infections (>8 weeks postinfection), and 26 control animals. Worm burdens were analyzed using negative binomial (NB) regression in STATA version 10 (Stata Corp., College Station, TX). For any rat in experiment i, ∼NB(ŷ,1 + αŷ), where y is the worm count, α is the overdispersion parameter (so that 1 + αŷ is the overall dispersion), and the expected worm count is ŷ = exp(xβ + y_i), where β measures the drug effect, x is the drug concentration, and γ_i is a fixed effect for the experiment. Additional fixed-effect terms were included into this model to estimate the effects of multiple drugs and drug interactions. Likelihood ratio tests were applied to assess statistical significance of these terms and the term synergism or antagonism used if significant deviations from the null hypothesis of independence were obtained (24).

Egg counts 7 days after treatment were analyzed using a similar model, but with the logarithm of the pretreatment egg count included in the model as an offset, so that the estimated drug effects refer to the comparison to pretreatment counts rather than to the comparison to the no-drug dose. The analysis of the log linearity of the relationship between drug dose and egg counts involved comparing negative binomial models to linear effects of drug dose with a reference model in which the effects of each drug were represented by separate terms for each dose level. Posttreatment egg counts and surviving worms were plotted on a square root scale, because the distribution is right skewed and egg counts included many zeros, which cannot be displayed if a logarithmic transformation is used.

All negative binomial models were fitted using the nbreg command in STATA v10.0.

RESULTS

A total of 305 F. hepatica-infected rats were studied, of which 212 rats harbored a chronic (>8 weeks postinfection) and 93 an acute (3 to 4 weeks postinfection) F. hepatica infection. The number of adult flukes in control rats (n = 43) varied from 1 to 12, with a median of 6 (Fig. 1A). In the absence of drug treatment, numbers of juvenile worms were similar to those of adult worms (Fig. 1). A range of 2 to 14 juvenile flukes with a median of 7 was observed in 17 control rats (Fig. 1B).

Effect of in vivo monotherapies against adult F. hepatica.

Worm burdens were recorded from 212 rats, of which 81 underwent monotherapy (Fig. 2). All drugs tested were effective in killing the flukes, with doses required to reduce worm burden by 50 and 95% given in Table 1. Triclabendazole has a 50% effective dose (ED₅₀) and ED₉₅ of 2.7 mg/kg (95% confidence limits [95% CL], 2.0, 4.4 mg/kg) and 11.7 mg/kg (95% CL, 8.4, 19.1 mg/kg), respectively. For OZ78, an ED₅₀ of 22.9 mg/kg (95% CL, 16.9, 35.4 mg/kg) and an ED₉₅ of 99.0 mg/kg (95% CL, 73.1, 153.2 mg/kg) were calculated. Artesunate and artemether achieve worm burden reductions of 50% administered at 113.0 mg/kg (95% CL, 69.9, 295.3) and 77.7 mg/kg (95% CL, 52.8, 146.8), respectively. Dosages of 488.6 mg/kg (artesunate) (95% CL, 302.1, 1276.1 mg/kg) and 335.7 mg/ kg (artemether) (95% CL, 228.2, 634.5 mg/kg) result in 95% worm burden reductions.

FIG. 2.

Effects of monotherapy on adult F. hepatica in rats. Shown are triclabendazole (A), artemether (B), artesunate (C), and OZ78 (D). Box plots are as described in the legend to Fig. 1. The dosages are arranged on a square-root scale. Black lines indicate the fitted dose-response curve assuming a linear effect of dose on the logarithm of the F. hepatica worm burden.

TABLE 1.

Estimates of effects of drug therapy against adult F. hepatica^a

Drug effect	Triclabendazole	Artemether	Artesunate	OZ78
Estimates of effect on worm burdens
N	24	17	16	22
Dose effect (β, kg/mg) (95% CL)	−0.26 (−0.35, −0.16)	−0.0089 (−0.0131, −0.0047)	−0.0061 (−0.0099, −0.0023)	−0.030 (−0.041, −0.020)
ED50 (mg/kg) (95% CL)	2.7 (2.0, 4.4)	77.7 (52.8, 146.8)	113.0 (69.9, 295.3)	22.9 (16.9, 35.4)
ED95(mg/kg) (95% CL)	11.7 (8.4, 19.1)	335.7 (228.2, 634.5)	488.6 (302.1, 1276.1)	99.0 (73.1, 153.2)
Estimates of effect on egg counts
Dose effect (β, kg/mg) (95% CL)	−0.54 (−0.85, −0.23)	−0.007 (−0.019, 0.005)	−0.011 (−0.023, 0.002)	−0.042 (−0.081, 0.002)
ED50 (mg/kg) (95% CL)	1.3 (0.8, 3.0)	102.7 (36.9, inf*)	65.8 (30.1, inf*)	16.6 (8.5, 356.2)
Significance tests
Log linearity of dose effect on worm burdens
Degrees of freedom	4	2	2	3
LRS (P value)	5.2 (0.3)	23.7 (0.00001)	0.1 (1.0)	9.7 (0.02)
Interaction of dose effect with log-linear effect of triclabendazole on worm burdens
Degrees of freedom		3	1	2
LRS (P value)		10.9 (0.012)	5.5 (0.019)	5.5 (0.07)
Dose effect on egg counts
Degrees of freedom	1	1	1	1
LRS (P value)	11.2 (0.0008)	1.28 (0.3)	3.0 (0.08)	5.3 (0.021)
Log linearity of dose effect on egg counts
Degrees of freedom	4	6	5	3
LRS (P value)	23.5 (0.0001)	6.5 (0.4)	5.9 (0.3)	8.1 (0.045)

^aN is the number of rats treated with monotherapy of the drug (across all doses); CL, confidence limits; ED₅₀is the dose required to kill 50% of the parasites; ED₉₅ is the dose required to kill 95% of the parasites; inf*, confidence limits includes no dose dependence. All estimates were derived from negative binomial regression models. Tests of log linearity were carried out by comparing a model to separate terms for each distinct drug dose, with a model that assumed the logarithm of the worm burden to decrease linearly with drug dose. LRS, likelihood ratio (χ²).

The estimate of the overdispersion parameter for this model was α = 0.33 (standard error, 0.11), indicating substantial overdispersion in the worm counts, justifying the use of negative binomial rather than Poisson models to analyze these data.

Tests of the log linearity of the dose effect confirmed that a linear relationship between the logarithm of the number of surviving worms and the drug dose gives a good fit to the data for triclabendazole and artesunate, but the dose responses estimated for artemether and OZ78 showed significant deviation from log linearity (Table 1).

Effect of in vivo monotherapies against juvenile F. hepatica.

The chemotherapy of juvenile F. hepatica was studied in four experiments, including 17 controls and 52 animals receiving monotherapy (Fig. 3). The effects of monotherapy with artesunate, artemether, and OZ78 were similar to those for adult worms, with highly statistically significant effects (likelihood ratio statistics [LRS], 1 degree of freedom of 9.6 [P = 0.002], 27.4 [P < 0.001], and 27.8 [P < 0.001], respectively). One hundred mg/kg of OZ78 cured all rats and 200 mg/kg artemether and artesunate resulted in 85.5% (95% CL, 64.1, 94.1%) and 56.4% (95% CL, 22.3, 75.5%) worm burden reductions, respectively. The effect of triclabendazole was less pronounced, with even the highest dose (5 mg/kg) having only a small effect against juvenile worms (WBR, 12.8; 95% CL, −36.8, 44.5%), even though this dose was effective in most animals against adult worms. Nevertheless, the negative binomial regression indicated a significant dose-response relationship (LRS, 8.0; P = 0.005).

FIG. 3.

Effects of monotherapy on juvenile F. hepatica in rats. Shown are triclabendazole (A), artemether (B), artesunate (C), and OZ78 (D). Box plots are as described in the legend to Fig. 1. The dosages are arranged on a square-root scale.

Effect of in vivo combination chemotherapy against adult F. hepatica.

Eighty-eight rats were given combinations of triclabendazole with one of the three peroxidic drugs (artesunate, artemether, or OZ78) as presented in Fig. 4. Triclabendazole (1.25 mg/kg) on its own had rather little effect on the worm burden (worm burden reduction was estimated as 13.5% [95% CL, −35.5, 44.8%], allowing for experimental variation). In contrast, 2.5 mg/kg triclabendazole on its own had a moderate effect on worm burden (estimated as 58.8%; 95% CL, 20.9, 78.5%). The administration of 2.5 mg/kg triclabendazole to rats showed huge variations in the treatment outcome: while approximately half of the rats were cured, no effect was seen in the remaining rats.

FIG. 4.

Effects of combination chemotherapy with 1.25 mg/kg (A to C) and 2.5 mg/kg (D to F) triclabendazole on adult F. hepatica in rats. Shown are artemether (A and D), artesunate (B and E), and OZ78 (C and F). Box plots are as described in the legend to Fig. 1. The dosages are arranged on a square-root scale.

Combinations with a dose of 1.25 mg/kg triclabendazole appeared to slightly inhibit the effects of both artemether and artesunate. The formal analysis of interactions in the negative binomial regressions to test the effects of the drug combinations, including the data for all doses of these drugs, indicated significant deviations from the null hypothesis of the independence of drug effects for the combination of artemether (P = 0.012) and artesunate (P = 0.019) with triclabendazole (Table 1). This suggests that combinations of artemether or artesunate with the higher dose of triclabendazole (2.5 mg/kg) increases worm burden reduction over those expected for a hypothesis of independence. However, the estimates of effect sizes in these regressions are imprecise.

For OZ78 there was, overall, no significant interaction with triclabendazole. The dose-response curve of OZ78 combined with 1.25 mg/kg triclabendazole was similar to that for its effect as monotherapy. However, there was a shift toward a significant interaction (P = 0.07) at lower dosages, with 25 mg/kg OZ78 showing no killing on its own (WBR estimate, adjusting for experimental variation of −7.1% [95% CL, −60.6, 28.6%]) but substantial killing in combination with both dosages of triclabendazole (1.25 mg/kg plus 25 mg/kg; WBR, 87.3%; 95% CL, 45.8, 97.0%). Using 2.5 mg/kg triclabendazole plus 25 mg/kg OZ78, all worms were killed.

Effect of in vivo combination chemotherapy against juvenile F. hepatica.

Twenty-four rats received combinations of triclabendazole and the peroxidic drugs. In each case 4 rats were treated with 2.5 or 5 mg/kg triclabendazole plus either 25 or 50 mg/kg artemether, artesunate, or OZ78. Twenty-five or 50 mg/kg of artesunate and artemether on their own resulted in low worm burden reductions (0 to 30%) against juvenile F. hepatica. OZ78 at 50 mg/kg achieved a modest worm burden reduction of 50%.

The combinations with 2.5 mg/kg triclabendazole plus 25 mg/kg of the peroxidic drugs did not show higher worm burden reductions compared to those of the single-drug treatments. Higher worm burden reductions compared to those of monotherapy were observed when 5 mg/kg triclabendazole was combined with 50 mg/kg artemether, artesunate, or OZ78. Nevertheless, there were no significant interactions between the triclabendazole effect and the effects of the other partner drugs on juvenile worms, since the numbers of rats tested were small and low doses of triclabendazole were used, but this does not mean that interactions would be detected in larger experiments with higher doses.

Egg excretion analysis.

Egg counts were recorded from 132 rats harboring adult F. hepatica infections 1 day prior to drug treatment and 7 days posttreatment. Each of these rats tested positive in shedding F. hepatica eggs 1 day before treatment. Additionally, the egg excretion of 85 of the 132 rats were sampled 2 months prior drug treatment and also at the time of treatment. The median egg count of 2 months prior to treatment was 6,600 EPG (interquartile range [IQR], 3,780 to 9,880), and the median EPG the day before treatment was 6,014 (IQR, 3,674 to 8,962), indicating very little change in average egg counts during this period, with a statistically nonsignificant difference (P = 0.4, two-sided Wilcoxon test). It follows from this that average reductions in egg load during this 2-month period are negligible and can be ignored relative to the drug effects.

Posttreatment egg counts were clearly correlated with surviving worms (Spearman correlation, 0.58; P < 0.0001) (Fig. 5), but there was considerable scatter, with many rats with surviving worms producing very low or no egg counts. Thirty-six (27.3%) of the rats tested had negative posttreatment egg counts, while 64 (48.4%) had 0 worms; however, only 25 had neither eggs nor worms.

FIG. 5.

Association of posttreatment adult worm burdens and egg counts of F. hepatica in rats. The worm burdens and egg counts are arranged on a square-root scale.

The overall egg count distributions of pre- and posttreatment egg counts of the 132 rats are given in Table 2 . The median of pretreatment counts was 4,902.9 (IQR, 2,929.8 to 7,798.8). A median of 83.5 EPG (IQR, 0 to 844.4) was calculated posttreatment. Each drug was able to reduce the egg excretion. The ratios of posttreatment to pretreatment egg counts reached from 0.001 to 1.951 (99.9 to −95.1% egg burden reduction). An EPG ratio of 1.951 (standard deviation [SD], 0.627), 0.01 (SD, 0.012), 0.032 (SD, 0.062), and 0.007 (SD, 0.014) was recorded for triclabendazole administered at 1.25, 2.5, 5, and 10 mg/kg. The treatment of rats with 100 and 200 mg/kg artesunate or artemether resulted in a posttreatment-to-pretreatment egg count ratio of 0.017 to 0.013 (SD, 0.023 to 0.021) and 0.351 to 0.066 (SD, 0.292 to 0.104), respectively. The EPG ratio decreases from 1.454 (SD, 0.598) to 0.229 (SD, 0.376) to 0.003 (SD, 0.003) with increasing OZ78 dosages of 25, 50, and 100 mg/kg, respectively.

TABLE 2.

Overall egg count distributions^a

Treatment status	Median	Lower quartile	Upper quartile
Pretreatment	4,902.9	2,929.8	7,798.8
Posttreatment	83.5	0	844.4
Ratio	0.016	0	0.475

^aUnits are eggs per gram of feces.

The ED₅₀s of triclabendazole, artesunate, artemether, and OZ78 calculated on the basis of the egg counts are 1.3 mg/kg (95% CL, 0.8 to 3.0 mg/kg), 65.8 mg/kg (95% CL, 30.1 to ∞ mg/kg), 102.7 mg/kg (95% CL, 36.9 to ∞ mg/kg), and 16.6 mg/kg (95% CL, 8.5 to 356.2 mg/kg). For all drugs, estimates of ED₅₀ on egg counts were similar to those for the effect on worm burdens (e.g., ED₅₀ of 2.7 mg/kg for triclabendazole and 22.9 mg/kg for OZ78). Significance tests of the linear effect of drug dose on egg counts demonstrated dose dependence for OZ78 and triclabendazole but not for artemether or artesunate (Table 1), despite the very low ratios of posttreatment-to-pretreatment egg counts in many of the rats treated with artemether or artesunate.

The combination of 2.5 mg/kg triclabendazole plus the peroxidic drugs achieved a posttreatment/pretreatment egg count ratio mean of 0.025 (egg burden reduction [EBR], 97.5%) with a range of 0.001 (EBR, 99.9%) to 0.138 (EBR, 86.2%), which is in line with 2.5 mg/kg triclabendazole single dosage (EBR, 99.0%). In more detail, triclabendazole (2.5 mg/kg) applied together with artesunate (6.25 to 200 mg/kg), artemether (6.25 to 100 mg/kg), or OZ78 (12.5 to 50 mg/kg) reduced the mean F. hepatica egg output to about 96.7, 98.7, and 97.0%, respectively. The effectiveness of the egg reduction of combinations with a triclabendazole dosage of 1.25 mg diminished the mean ratio to 0.67 (EBR, 33%) with a range of 0.007 (EBR, 99.3%) to 1.266 (EBR, −26.6%), which is considerably lower than the ratio of 1.951 at the 1.25-mg/kg dosage of triclabendazole alone. On the other hand, artesunate and artemether treatments (100 mg/kg) were more effective than in combination with 1.25 mg/kg triclabendazole. Tests of deviation from linearity were significant for OZ78 and for triclabendazole, indicating that the dose-response relationship for these drugs was not well modeled by the assumption of log linearity, with a substantial difference between egg counts in rats treated with 1.25 mg/kg triclabendazole combinations (which had little effect on egg counts) and those receiving combinations with 2.5 mg/kg triclabendazole, which had very low posttreatment egg counts.

Effects of in vitro chemotherapy.

Temporal drug effects (monotherapy and combination chemotherapy) on the viability on adult F. hepatica in vitro are presented in Fig. 6. Adult F. hepatica controls (n = 26) showed normal activities during the entire observation period. Flukes incubated in the presence of peroxides showed a fast decrease of movements, and after 72 h 33% (artemether) to 75% (artesunate) of worms had died. Differences in the effect of the triclabendazole derivatives on F. hepatica were observed. While triclabendazole showed only moderate activity on adult worms in vitro, incubation with the sulfone and sulfoxide metabolite resulted in decreased movement and the death of 50 to 67% of flukes.

FIG. 6.

In vitro effects of monotherapy and combination chemotherapy on adult F. hepatica flukes. Single-drug incubation (continuous lines): 15 μg/ml benzimidazole derivatives (•); triclabendazole (1; TBZ), triclabendazole-sulfoxide (2; TBZ-SX), and triclabendazole-sulfone (3; TBZ-SO); 50 μg/ml peroxidic drugs (□); artemether (a), artesunate (b), and OZ78 (c). Combinations (dotted lines): 15 μg/ml triclabendazole (1), triclabendazole-sulfoxide (2), or triclabendazole-sulfone (3) plus either 50 μg/ml artemether (a), artesunate (b), or OZ78 (c). The adult control worms (n = 26) are described with a continuous line without symbols. The limits of the whiskers correspond to the standard error of the mean values per time point.

Flukes incubated with the combination of triclabendazole plus artesunate, artemether, or OZ78 were slightly less affected than adult F. hepatica exposed to the peroxidic drugs alone (Fig. 6, row 1, graphs a to c). Lower mortality rates were calculated for the triclabendazole combinations (mortality rates of 8 to 25%) than for the single peroxidic drugs.

F. hepatica exposed to triclabendazole-sulfoxide and triclabendazole-sulfone combinations showed a rapid decrease in viability, and after 48 h only minimal activities were recorded. Adult flukes incubated with triclabendazole-sulfoxide and triclabendazole-sulfone combinations showed significantly less movement than worms exposed to single drugs, and the majority of these worms (67 to 100%) were killed 72 h postincubation (Fig. 6, rows 2 and 3, graphs a to c).

In Fig. 7, the effects of combination chemotherapy and single drugs on juvenile F. hepatica are depicted. Juvenile control flukes showed a viability 72 h postincubation similar to that at the beginning of the assay. Incubation with single drugs revealed clear differences in the activities on juvenile F. hepatica. Artesunate showed the highest activity, resulting in the death of all juvenile flukes after 72 h (Fig. 7, rows 1 to 3, graph b). A moderate effect (decreased viability) was observed with artemether, OZ78, and triclabendazole (Fig. 7, row 1, graphs a and c). On the other hand, triclabendazole-sulfone- and triclabendazole-sulfoxide-incubated worms behaved like controls (Fig. 7, rows 2 and 3, graphs a to c). Figure 7 illustrates that with the exception of a triclabendazole-sulfone-artesunate combination (which had an activity similar to that of artesunate alone [Fig. 7, row 3, graph b]), all combinations tested showed a superior activity compared to that of incubation with single drugs. For example, 24 h postincubation with artesunate and artemether in combination with triclabendazole, flukes showed only minimal activities, and approximately 90% of the flukes had died after 48 h (Fig. 7, row 1, graphs a and b). None of the juvenile F. hepatica survived in the presence of any of the triclabendazole-sulfoxide combinations tested (Fig. 7, row 2, graphs a to c).

FIG. 7.

In vitro effects of monotherapy and combination chemotherapy on juvenile F. hepatica flukes. Single-drug incubation (continuous lines): 30 μg/ml benzimidazole derivatives (•); triclabendazole (1; TBZ), triclabendazole-sulfoxide (2; TBZ-SX), and triclabendazole-sulfone (3; TBZ-SO); 100 μg/ml peroxidic drugs (□); artemether (a), artesunate (b), and OZ78 (c). Combinations (dotted lines): 30 μg/ml triclabendazole (1), triclabendazole-sulfoxide (2), or triclabendazole-sulfone (3) plus either 100 μg/ml artemether (a), artesunate (b), or OZ78 (c). The juvenile control worms (n = 15) are described with a continuous line without symbols. The limits of the whiskers correspond to the standard error of the mean values per time point.

DISCUSSION

Cases of triclabendazole resistance are continuously documented from livestock, and hence new strategies for the treatment of F. hepatica infections are needed (10). Today, drug combinations are a popular tool to decrease the selection pressure and increase parasitological cure rates to reduce toxicity (11). In addition, drug combinations are less likely to produce adverse events as reduced dosages are used. Artemisinin combinations therefore might have a better safety profile than monotherapies. For example, toxicological effects were observed with artesunate in F. hepatica-infected rats at doses of 200 mg/kg and above, and neurotoxicity or embryotoxicity have been described in rodents, dogs, and monkeys (9, 19).

The present study investigated the effect on worm and egg burden of triclabendazole, artesunate, artemether, and OZ78 monotherapy and combinations in rats harboring acute and chronic F. hepatica infections. In addition, drug effects on juvenile and adult F. hepatica organisms were monitored in vitro.

This study confirmed that single oral doses of artesunate, artemether, OZ78, and triclabendazole are active against F. hepatica in rats, with the lowest ED_50/95 values observed for the reference drug triclabendazole. Somewhat higher activities were observed with OZ78 compared to those of the artemisinins in vivo, but an opposite tendency was observed in vitro, which might be explained by the low bioavailability and half-lives of artemether and artesunate (15, 37). In vivo, OZ78, artemether, and artesunate all showed reductions in burdens of juvenile flukes similar to those of adults, while in vitro the concentrations of these drugs needed to kill the juvenile flukes were about twice those that killed the adults. On the other hand, triclabendazole was more effective against adult than juvenile F. hepatica flukes in vivo; for instance, at a dosage of 5 mg/kg it achieved worm burden reductions of 56.8 and 12.8% against adult and juvenile flukes, respectively. Although triclabendazole has been well studied against immature and mature F. hepatica in sheep and cattle (2, 14, 32, 38), to our knowledge thorough stage specifity studies of rats have not been carried out to date. This stage specificity appears to arise because of the reduced in vitro sensitivity of the juvenile flukes to the metabolites of triclabendazole, into which the drug is converted rapidly in vivo (10). In vitro, triclabendazole sulfoxide- and sulfone-treated adult flukes were killed, while juveniles behaved like the controls. In contrast, triclabendazole was much more effective against juvenile than adult F. hepatica flukes. In a previous study, immature flukes also were found to be more sensitive to triclabendazole than adults in vitro, while triclabendazole-sulfoxide exhibited a delayed effect on juvenile flukes (1).

Our preliminary findings obtained with drug combinations on adult flukes in vitro somewhat correspond to the findings observed with adult infections in vivo. The combinations with the triclabendazole metabolites plus peroxidic drug (Fig. 6, rows 2 and 3, graphs a to c) against mature F. hepatica in vitro were more effective than the single drugs, and this was supported by in vivo experiments combining 2.5 mg/kg triclabendazole with artesunate or artemether. The almost-complete elimination of worms was achieved with dosages of these two semisynthetic artemisinins as low as 12.5 mg/kg when combined with 2.5 mg/kg triclabendazole (Fig. 4D and E). This finding might be explained by the independent mechanism of action of the drugs. While artemether enters the parasite through oral ingestion and causes substantial disruption to the gut, triclabendazole mainly affects the tegument (29, 35). Furthermore, artemether and OZ78 also act against triclabendazole-resistant F. hepatica (22).

On the other hand, combinations with 1.25 mg/kg triclabendazole appeared to slightly inhibit the effect of artemether or artesunate; i.e., somewhat higher worm burdens were observed than those achieved with artemisinin monotherapy (Fig. 4A and B). Similar antagonistic effects have been observed when the artemisinins were combined with low doses of praziquantel (75 mg/kg) and administered to Clonorchis sinensis-infected rats (24). The rather contradictory results obtained following small modifications in the triclabendazole dose (1.25 versus 2.5 mg/kg) cannot be explained at the moment and warrant further investigation, such as pharmacokinetic studies.

Interestingly, we observed no overall significant interaction between triclabendazole and OZ78 when administered to rats infected with mature flukes. Unlike the artemisinins, OZ78 effects did not depend on the dosage of triclabendazole, and it also could cure rats in combination with 1.25 mg/kg triclabendazole with a dose response similar to that of monotherapy with OZ78. Combinations with OZ78 showed a trend toward a synergistic effect (however, likelihood ratio statistics showed no significance; P = 0.07) at lower dosages of OZ78; for example, 25 mg/kg of OZ78 was more effective in combination with both dosages of triclabendazole used than when applied alone.

In vitro, all drug combinations had more pronounced effects against immature than adult flukes (Fig. 6 and 7), and the in vitro combination chemotherapy data suggest synergistic effects on juveniles. However, serial drug dilutions would have been necessary to construct isobolograms to detect true synergistic or antagonistic effects (3). Furthermore, while our in vitro study was based on periodic phenotypic evaluation, more sophisticated approaches such as calorimetric measurement (25), which continuously measures the energy release of the worm, are required to evaluate drug-worm interactions in more detail. On the other hand, in vivo most of the combinations, like the monotherapies, were less effective against juveniles than against mature flukes, and we found little evidence for synergy. Exceptions were combinations with 5 mg/kg triclabendazole and 50 mg/kg peroxidic drug, which killed juvenile worms, despite the ineffectiveness of monotherapies at these doses against them. We did not use dosages higher than 5 mg/kg in our studies, since a 10-mg/kg triclabendazole dose has been shown to achieve worm burden reductions of 85 to 100% (14 and unpublished observations). Nonetheless, it would be interesting to evaluate intermediate triclabendazole doses.

Egg excretion analyses are helpful to confirm the presence of F. hepatica and other helminth infections and to monitor infection intensities. Fasciola hepatica eggs were detected in all 132 fecal samples examined prior to drug treatment, and posttreatment egg counts correlated significantly with surviving worms. It appears that the treatment reduced the worm burden and therefore minimized the influence of the crowding effect (a decreased egg output as a function of higher worm burden), which was previously documented in F. hepatica-infected rats (36).

However, there is a substantial loss of precision in using egg counts as indicators of worm burdens. There are several reasons why the direct counting of worms gives more-precise estimates of dose-response relationships. Some rats with surviving worms produced very few or no eggs, probably because drug treatment can inhibit egg production even when it does not kill the worms. It has been demonstrated that triclabendazole affects spermatogenic and vitelline cells of F. hepatica and artemether inhibits the egg production of F. gigantica in vitro (6, 33, 34). We also recorded rats characterized by no worms but egg-positive fecal samples. This could be because the period of 7 days between treatment and dissection was too short to eliminate all eggs from the rats' bodies. In addition, since rats were housed in groups of five, we might have detected transient eggs, which were orally taken up from egg-positive fecal pellets present in the cage.

In conclusion, we confirmed the promising fasciocidal properties of the peroxidic drugs artesunate, artemether, and OZ78. It is encouraging that enhanced drug effects were observed using combinations of artesunate, artemether, and OZ78 plus triclabendazole in F. hepatica-infected rats. However, the great variations in dose response following slight titrations in the doses are striking. Further experiments, such as pharmacokinetic and pharmacodynamic studies following combination chemotherapy in rats, studies using triclabendazole-resistant F. hepatica strains, or combination trials in larger animals (e.g., sheep), are warranted, which might further strengthen our knowledge of the fasciocidal properties of triclabendazole-peroxide drug combinations.