Combination Chemotherapy against Clonorchis sinensis: Experiments with Artemether, Artesunate, OZ78, Praziquantel, and Tribendimidine in a Rat Model

Abstract

Caused by the Chinese liver fluke Clonorchis sinensis, clonorchiasis is of growing public health importance. Treatment and control of the disease rely on a single drug, praziquantel, and little information regarding combination chemotherapy is available. Here, we evaluated the in vivo efficacy of praziquantel combined with artemether, artesunate, OZ78, and tribendimidine, as well as an artesunate-tribendimidine combination against C. sinensis, in a rat model. Data from previous experiments were included, and negative binomial regression analyses were carried out to determine dose-response relationships and to study the effect of drug combination. All drugs given in monotherapy were efficacious in killing the worms; doses of 16 and 70 mg/kg of body weight of artesunate, for example, reduced worm burden by 50% and 95%, respectively. Artemether and OZ78 (12.5 to 50 mg/kg) showed dose-dependent killing of worms but no significant drug interactions when given with 150 mg/kg praziquantel, suggesting independent additive effects. In contrast, artesunate and tribendimidine (12.5 to 50 mg/kg) showed synergistic interactions with 150 mg/kg praziquantel. When low doses of 3.1 and 6.25 mg/kg OZ78 and artemether, respectively, were combined with praziquantel (150 mg/kg) an increased worm survival, above the level observed with praziquantel monotherapy, was noted. A similar antagonism was seen when praziquantel (75 mg/kg) was combined with several of the companion drugs at various doses. In conclusion, in vivo efficacy of praziquantel, the artemisinins, OZ78, and tribendimidine against C. sinensis is confirmed, and combination chemotherapy with praziquantel produces synergistic and antagonistic effects depending on the doses administered. Further preclinical investigations are warranted.

The first documented use of combination chemotherapy, most likely for cancer, dates back to 1550 BC in Egypt (1). Today, combination chemotherapy is commonly used not only against cancer but also for treating human immunodeficiency virus infections, mycobacteria, malaria, and many other disorders such as acute migraines (4, 18, 22, 29). The key rationales for combining two or more drugs that act on different pathways are (i) to increase the potency of the chemotherapeutic response, (ii) to decrease toxicity, (iii) to reduce the probability of development of resistant mutants, and (iv) to have a broader spectrum of activity (1). In addition to enhanced therapeutic outcomes, fewer adverse events might occur with combination chemotherapy using compounds at lower doses than are used for monotherapy (18).

For helminth infections, combination chemotherapy has been widely used against nematodes, particularly in livestock production (1). However, little information is available regarding combination chemotherapy against food-borne trematode infections, including Clonorchis sinensis. This Chinese liver fluke is the causative agent of clonorchiasis, a disease of growing public health importance which is, however, often neglected. It is currently estimated that 601 million people are at risk of clonorchiasis, with more than 35 million infections (9, 15). Common manifestations in the chronic stages of severe infections include intrahepatic stones, biliary cirrhosis, pyogenic cholangitis, cholelithiasis, cholecystitis, pancreatitis, and cholangiocarcinoma (15).

Praziquantel is virtually the only drug for treating C. sinensis infections and has been recommended by the World Health Organization (WHO) for more than 20 years (23). Albendazole requires long treatment courses over multiple days and hence is rarely used (3, 8). We have shown that artemether and artesunate (the two most widely used semisynthetic derivatives of artemisinin), the synthetic peroxide OZ78, and the anthelmintic tribendimidine possess clonorchicidal properties (11-13, 26, 28). For example, a single 150-mg/kg of body weight oral dose of either artemether, artesunate, or tribendimidine resulted in worm burden reductions of 99 to 100% in rats harboring adult C. sinensis (11, 13). OZ78, at a single 300-mg/kg oral dose, achieved a worm burden reduction of 98.5% against adult C. sinensis in rats (12).

The aim of the present study was to elucidate whether combination chemotherapy using praziquantel plus either artemether, artesunate, OZ78, or tribendimidine and a combination of tribendimidine with artesunate act in additive, synergistic, or antagonistic manners. Comparisons against monotherapies were made. We employed a C. sinensis-rat model as in our previous work (11, 26).

MATERIALS AND METHODS

C. sinensis-rat model.

Metacercariae of C. sinensis from freshwater fish (Pseudorasbora parva) caught in Heng county, Guangxi province, People's Republic of China, which is an area of endemicity, were isolated as described in previous publications (12, 26) and transferred on ice to Switzerland. Female Wistar rats (n = 135; age, 5 weeks; weight, ∼100 g) were purchased from Harlan (Itingen, Switzerland). Rats were kept in groups of five in Makrolon cages in environmentally controlled conditions (temperature, ∼25°C; humidity, ∼70%; 12-h light/12-h dark cycle) with free access to water and rodent diet. Rats were acclimatized for several days before infection by oral gavage with 40 to 55 C. sinensis metacercariae each. All animal studies were approved by the cantonal veterinary authority (permission no. 2070).

Drugs.

Artemether was obtained from Kunming Pharmaceutical Cooperation (Kunming, People's Republic of China). Artesunate was the product of Mepha AG (Aesch, Switzerland). Tribendimidine was obtained from Shandong Xinhua Pharmaceutical Company Limited (Zibo, People's Republic of China). Praziquantel was purchased from Shanghai no. 6 Pharmaceutical Company (Shanghai, People's Republic of China). OZ78 was synthesized at the College of Pharmacy, University of Nebraska Medical Center (Omaha, NE). Drugs were prepared in suspensions in 7% (vol/vol) Tween 80 and 3% (vol/vol) ethanol shortly before oral administration.

Treatment and evaluation of drug efficacy.

Treatment was administered to rats harboring adult C. sinensis worms (4 to 7 weeks postinfection). First, for monotherapy experiments, five groups of four rats each were treated orally with artemether, artesunate, praziquantel, OZ78, and tribendimidine administered alone. The drugs were given in the following dosages: artemether and artesunate, 25 mg/kg; OZ78 and tribendimidine, 50 mg/kg; praziquantel, 150 mg/kg. Drug dosages were chosen to be close to the dose required to kill 50% of C. sinensis worms (EC₅₀), as determined from the initial dose-finding studies (11-13, 26), which are briefly summarized below. Second, for the combination therapy experiments, 20 groups of three or four rats each were treated with praziquantel (75 and 150 mg/kg) plus simultaneously either artemether (6.25, 12.5, and 25 mg/kg), artesunate (6.25, 12.5, and 25 mg/kg), OZ78 (3.1, 6.25, 12.5, 25, and 50 mg/kg), or tribendimidine (12.5, 25, and 50 mg/kg) in two sets of experiments. Finally, 20 rats were given tribendimidine (25 and 50 mg/kg) plus artesunate (6.25, 12.5, and 25 mg/kg).

One week posttreatment, rats were necropsied, and flukes were removed from the bile ducts and counted for each rat separately. Four and six rats, also infected with adult C. sinensis but left untreated, served as control groups and were examined in the same way.

Data from previous experiments.

For the statistical analyses we included results obtained from four previous experiments carried out with C. sinensis-infected rats in our laboratories. In these studies 61 rats served as controls and 144 rats were treated with either artesunate, artemether, OZ78, praziquantel, or tribendimidine. A preliminary study tested the effect of combined chemotherapy in 28 C. sinensis-infected rats. In a first study, the efficacy of OZ78 in 12 C. sinensis-infected rats was tested, with animals treated orally with OZ78 at single doses of 75, 150, and 300 mg/kg. Five infected but untreated rats served as controls (12). In a second study, consisting of several subsets of experiments, the in vivo efficacy of artemether, artesunate, praziquantel, and tribendimidine in 89 C. sinensis-infected rats was determined. Four different regimens of tribendimidine (37.5, 75, 150, and 300 mg/kg), three different regimens of artemether and artesunate (37.5, 75, and 150 mg/kg), and two different regimens of praziquantel (75 and 150 mg/kg) were employed. Forty-three rats served as controls (26). In the third study, tribendimidine (75 and 150 mg/kg) was tested in eight C. sinensis-infected rats (our unpublished findings). Finally, 63 C. sinensis-infected rats were treated with single oral doses of either tribendimidine or praziquantel (both drugs administered at 150 mg/kg), either artemether or artesunate (both drugs given at 75 mg/kg), or a combination of two of these drugs as follows: artesunate or artemether (30 mg/kg) plus praziquantel (150 mg/kg) or tribendimidine (50 or 75 mg/kg) or tribendimidine (50 mg/kg or 75 mg/kg) plus praziquantel (150 or 175 mg/kg). Thirteen C. sinensis-infected rats served as controls (27).

Statistical analysis.

Worm burdens were analyzed using negative binomial (NB) regression in STATA version 10 (Stata Corp., College Station, TX). For any rat in experiment i, y ∼ NB(ŷ,1 + αŷ), where y is the worm count and α is the overdispersion parameter, so that 1 + αŷ is the overall dispersion and the expected worm count is given by ŷ = exp(xβ + γ_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 added into this model to estimate effects of multiple drugs and drug interactions. Statistical significance of these terms was tested using likelihood ratio tests.

RESULTS

A total of 368 rats were infected with C. sinensis in different experiments. In the 71 control rats the number of worms per rat varied between 14 and 49, with a median of 30. There was some variation between experiments in these numbers (Fig. 1), justifying adjustment for experimental variation in the analyses.

FIG. 1.

Distributions of C. sinensis worm counts in nontreated control rats. The upper and lower limits of the boxes correspond to the interquartile ranges, the values in the middle are the medians, and the limits of the whiskers correspond to the adjacent values, i.e., the largest values below the 25th percentile and the smallest values above the 75th percentile. Figures in parentheses are the numbers of control animals in each experiment. Experiments 9 and 11 comprised dose finding studies in which there were no untreated animals.

Effect of monotherapies.

Overall, 165 rats underwent monotherapy (Fig. 2a to e). The negative binomial regression applied to the data of the controls and the animals treated with a single drug indicated that all drugs were efficacious in killing the worms, with EC₅₀ and EC₉₅ values given in Table 1.

FIG. 2.

Effects of monotherapy on C. sinensis worm burdens. (a) Praziquantel; (b) artemether; (c) artesunate; (d) OZ78; (e) tribendimidine. Box plots are as in Fig. 1. The dosages are arranged on a square root scale. (In panels b and c, there was a dose of 30 μg/kg between the 25- and 37-μg/kg doses.) Black lines indicate the fitted dose-response curve assuming a linear effect of dose on the logarithm of the C. sinensis worm burden.

TABLE 1.

Estimates of drug effects against adult C. sinensis harbored in rats

Drug	Monotherapy				Combination chemotherapy (significance tests)
	No. of rats treated	β (95% CI)	EC50 (mg/kg) (95% CI)	EC95 (mg/kg) (95% CI)	Log linearity of dose effecta		χ2 (P) for:
					Degrees of freedom	χ2 (P)	Dose effect in the presence of 75 mg/kg praziquantel	Interaction of dose effect with 150 mg/kg praziquantel
Praziquantel	38	−0.007 (−0.009 to −0.006)	95.0 (76.4-125.5)	410.4 (330.0-542.6)	2	3.7 (0.2)
Artemether	23	−0.030 (−0.037 to −0.023)	22.8 (18.5-29.6)	98.5 (80.1-128.0)	3	14.9 (0.002)	4.5 (0.035)	2.1 (0.15)
Artesunate	23	−0.043 (−0.052 to −0.033)	16.2 (13.3-20.8)	70.2 (57.6-89.8)	3	15.6 (0.001)	4.5 (0.034)	6.5 (0.01)
OZ78	16	−0.015 (−0.019 to −0.010)	47.3 (36.3-68.0)	204.4 (156.7-293.8)	3	2.7 (0.4)	26.4 (<0.001)	0.1 (0.752)
Tribendimidine	65	−0.016 (−0.019 to −0.014)	42.7 (36.9-50.5)	184.4 (159.7-218.1)	4	5.6 (0.2)	2.8 (0.095)	5.7 (0.02)

^aTests of log linearity were carried out by comparing a model with separate terms for each distinct drug dose with a model that assumed that the logarithm of the worm burden decreases linearly with drug dose. All estimates were derived from NB regression models.

Praziquantel given at 95 mg/kg (95% confidence interval [CI]: 76 to 126 mg/kg) and 410 mg/kg (95% CI: 330 to 543 mg/kg) is estimated to achieve worm burden reductions of 50% and 95%, respectively. Artesunate was found to be slightly superior to artemether (EC₅₀, 16 mg/kg versus 23 mg/kg; EC₉₅, 70 mg/kg versus 99 mg/kg). The synthetic peroxide OZ78 and tribendimidine displayed similar efficacies (EC₅₀s, 47 mg/kg and 43 mg/kg, respectively; EC₉₅s, 204 mg/kg and 184 mg/kg, respectively).

The estimate of the overdispersion parameter (α) of the model including main effects of the dose of each drug was 0.34 (standard error 0.05), indicating substantial overdispersion in the worm counts, justifying the use of NB rather than Poisson models to analyze the data.

Tests of 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 praziquantel, OZ78, and tribendimidine. However, the dose responses estimated for artemether and artesunate showed significant deviation from log linearity (Table 1). Inspection of the data (Fig. 2) suggests that this is because the best-fitting dose-response curves for these drugs would be steeper around the EC₅₀ than the log-linear fitted line, i.e., more like a step function. There is no suggestion that the deviation from linearity invalidates the EC₅₀s given in Table 1.

Effect of combination chemotherapy.

A total of 132 C. sinensis-infected rats were given combination chemotherapy. Thirty-two of these, in one experiment, received a dose of 75 mg/kg of praziquantel together with another drug, with each of the other drugs tested at two distinct doses (four rats at each dose). For three of the other drugs (artemether, artesunate, and tribendimidine), worm burdens were higher than those observed following monotherapy in a dose-dependent manner (Fig. 3). For example, while rats treated with a single oral dose of 75 mg/kg praziquantel harbored a mean of 18 worms, a mean of 52 C. sinensis worms were recovered from animals treated with 75 mg/kg of praziquantel combined with 25 mg/kg artesunate (Fig. 3). The effect was statistically significant for artemether and artesunate (Table 1). In contrast, OZ78 showed a statistically significant dose-dependent killing effect when coadministered with this dose of praziquantel; e.g., rats treated with 75 mg/kg praziquantel combined with 25 mg/kg OZ78 showed a mean of only four worms (Fig. 3 and Table 1).

FIG. 3.

Effects of combination therapy with 75 mg/kg praziquantel on C. sinensis worm burdens. (a) Artemether; (b) artesunate; (c) OZ78; (d) tribendimidine. Box plots are as in Fig. 1.

Seventy rats received 150 mg/kg praziquantel with various doses of one of the other drugs (Fig. 4). At 150 mg/kg praziquantel, doses of 12.5 to 30 mg/kg of artemether and 6.25 to 50 mg/kg OZ78 behaved as anticipated, showing dose-dependent killing of C. sinensis and achieving high worm burden reductions, but the lowest doses of artemether and OZ78 seemed to increase worm survival above the level observed with 150 mg/kg praziquantel alone. For example, rats treated with 3.25 mg/kg OZ78 and 150 mg/kg praziquantel harbored a mean of 28 worms/rat compared to less than 4 worms when 6.25 mg/kg OZ78 was given in combination with 150 mg/kg praziquantel. Higher doses of artesunate and tribendimidine resulted in worm burdens similar to those in animals treated with 150 mg/kg praziquantel alone. In view of the small number of observations, formal tests of the linearity of the dose effects were not carried out. Further NB models were used to test whether the overall dose dependence for the other drugs in the presence of 150 mg/kg praziquantel was similar to that in its absence (Table 1). These analyses excluded the 75-mg/kg praziquantel treatments, where there was a clear deviation from independence of effects. Artemether and OZ78 showed no significant interaction of the drug effect with that of 150 mg/kg praziquantel, suggesting independent multiplicative effects on the worm burden. Artesunate and tribendimidine showed significant interactions with 150 mg/kg praziquantel (Table 1): combinations with praziquantel showed higher efficacy than expected if the drug effects were independent (Fig. 4).

FIG. 4.

Effects of combination therapy with 150 mg/kg praziquantel on C. sinensis worm burdens. (a) Artemether; (b) artesunate; (c) OZ78; (d) tribendimidine. Box plots are as in Fig. 1.

A further 30 rats were treated with combinations of tribendimidine and artesunate, of which 17 received various doses of artesunate combined with 50 mg/kg of tribendimidine (Fig. 5). In these animals there was also a clear indication that underdosing with a companion drug resulted in a reduction in drug efficacy, with 6.25 mg/kg of artesunate again tending to give slightly higher worm burdens than those in animals treated with tribendimidine only. Overall, the NB models indicated a highly significant deviation from linearity in the interaction between these two drugs (likelihood ratio [χ²] = 71.4; P < 0.001), confirming that the effects cannot be assumed to act independently.

FIG. 5.

Effects of combination therapy with 50 mg/kg tribendimidine on C. sinensis worm burdens. Box plots are as in Fig. 1.

DISCUSSION

Since the advent of praziquantel more than 30 years ago (5, 17), discovery and development research pertaining to trematocidal drugs has been scarce (2, 9). Indeed, treatment and control of most trematode infections in humans heavily relies on just one drug, praziquantel. This is a dangerous situation should resistance emerge. In the case of clonorchiasis a low cure rate has been observed in patients from Vietnam after administration of praziquantel (19). In the absence of significant funding for research and development of novel trematocidal drugs, combination chemotherapy with existing drugs warrants investigation.

In the present study we used two artemisinin derivatives, OZ78, and tribendimidine alone or in combination with praziquantel, employing a C. sinensis-rat model. Artemether and artesunate are the most potent antimalarials currently available and are widely used in artemisinin-based combination therapy (16). The artemisinins and synthetic peroxides (i.e., OZ78) also possess trematocidal properties in vivo, i.e., against schistosomes (20, 24), C. sinensis (12, 13, 26, 28), and the liver fluke Fasciola hepatica (6, 10). Finally, tribendimidine, which has been registered in China in early 2004 for use against common soil-transmitted helminth infections (25), also shows activity against the intestinal trematode Echinostoma caproni (7), C. sinensis, and Opisthorchis viverrini (11, 26, 28).

Our study confirmed that single oral doses of artemether, artesunate, OZ78, and tribendimidine are active against C. sinensis harbored in rats, with the lowest EC₅₀ and EC₉₅ values observed for artesunate.

A number of issues are worth discussing with regard to the combination chemotherapy experiments presented here. First and somewhat surprisingly, when praziquantel (150 mg/kg) was given simultaneously with low doses of either artemether or OZ78 (3.1 and 6.25 mg/kg), an increased worm survival (above the level for a single dose of 150 mg/kg praziquantel) was observed. Moreover, halving the dose of praziquantel to 75 mg/kg and combining with different doses of artemether, artesunate, and tribendimidine also resulted in antagonistic effects, i.e., higher worm burdens, in a dose-dependent manner compared to that in rats treated with 75 mg/kg of praziquantel alone. On the other hand, this phenomenon was not observed with the 75-mg/kg praziquantel-OZ78 combinations. These antagonistic effects cannot be explained at the moment, and future studies should assess whether pharmacokinetic factors or biological mechanisms are involved.

Second, treatment of either artemether, artesunate, or OZ78 at dosages of 12.5 and 25 mg/kg combined with praziquantel at 150 mg/kg resulted in almost complete worm burden reductions. The test of interaction was highly significant for the artesunate-praziquantel (150 mg/kg) combinations, suggesting synergism between these two drugs, while our model showed no interactions for combined treatment with praziquantel plus either artemether or OZ78.

Third, when praziquantel (150 mg/kg) was combined with OZ78 at a low dose of 6.25 mg/kg, a worm burden reduction exceeding 80% was obtained, in contrast to low observed worm burden reductions with artesunate-praziquantel or artemether-praziquantel combinations at this dose (150 mg/kg praziquantel and 6.25 mg/kg artemether or artesunate). The greater bioavailability and longer half-life of OZ78 compared to the artemisinins in rats (21) might explain the differences in clonorchicidal efficacies at these dose combinations.

Finally, praziquantel (150 mg/kg) combined with tribendimidine (12.5 to 50 mg/kg) and tribendimidine combined with artesunate (25 mg/kg) showed a highly significant deviation from linearity in the interaction between these drugs. For example, while a worm burden reduction of below 50% was achieved with praziquantel (150 mg/kg) plus tribendimidine (50 mg/kg), a worm burden reduction in excess of 80% was found when praziquantel was administered with one-half the tribendimidine dose (25 mg/kg). The lack of dose-response relationship cannot be explained at the moment. However, our findings point to the presence of an interaction between tribendimidine and the companion drugs. Our statistical models suggest a highly synergistic rather than an additive effect for the tribendimidine-praziquantel combination.

In conclusion, synergistic but also antagonistic effects were observed by treating C. sinensis-infected rats with different drug combinations. Future studies should also test the efficacy of different drug combinations not only against the adult stages but also against the juvenile stages of C. sinensis and other food-borne trematodes. A recent study, for example, has demonstrated the stage-specific susceptibility of the artemisinins, with adult C. sinensis showing highest susceptibility to these drugs (26). Additionally, preclinical studies should address the safety and pharmacokinetic and pharmacology profiles of the drug combination. Finally, clinical investigations of patients infected with C. sinensis using artemisinins and tribendimidine can be envisaged, particularly in areas where clonorchiasis and malaria or soil-transmitted helminthiasis coexist. It is suggested that artemisinin-based combination therapies and tribendimidine, at the doses commonly employed against malaria and soil-transmitted helminth infections, be used in a first step. If activity of these drugs can be confirmed in humans infected with C. sinensis, combination treatment trials could be planned, with the ultimate goal of developing options for treatment of clonorchiasis that are safe and efficacious.