Acute and subacute oral toxicity of artemisinin-hydroxychloroquine sulfate tablets in beagle dogs

Abstract Artemisinin-hydroxychloroquine sulfate tablets (AH) are regarded as a relatively inexpensive and novel combination therapy for the treatment of various forms of malaria, particularly aminoquinoline drugs-resistant strains of Plasmodium falciparum. Our aim was to conduct acute and subacute oral toxicity studies in non-rodents to obtain more nonclinical data on the safety of AH. Acute toxicity evaluation was performed in beagle dogs at single doses of 230, 530, 790, 1180, 2660, and 5000 mg/kg. Beagle dogs at doses of 0, 56, 84, and 126 mg/kg were used to assess subacute toxicity for 14 days. The approximate lethal dose range for acute oral administration of AH in dogs is found to be 790–1180 mg/kg, and toxic symptoms prior to death include gait instability, limb weakness, mental fatigue, tachypnea, and convulsion. Repeated doses of AH in dogs caused vomiting, soft feces, decreased activity, anorexia, and splenic red pulp vacuolation. Of note, AH could reduce body weight gain and prolong the QTc interval of individual dogs. Therefore, the no-observed-adverse-effect level (NOAEL) and lowest-observed-adverse-effect level (LOAEL) of oral administration of AH for 14 days in dogs are determined to be 84 mg/kg and 126 mg/kg, respectively.


Introduction
Malaria is an acute febrile disease caused by parasites and spreads to people through the bites of infected female Anopheles mosquitoes. According to the World Health Organization (WHO) world malaria report (WHO 2021), nearly half of the world's population was at risk of malaria in 2020. It is estimated that there were an estimated 241 million cases of malaria worldwide and 627 000 malaria deaths in 2020. The WHO African Region carries a disproportionately high share of the global malaria burden and is home to 95% of malaria cases and 96% of malaria deaths. The antimalarial situation is more precarious due to a convergence of threats, ranging from the Corona Virus Disease 2019  and Ebola outbreaks to flooding and other humanitarian emergencies, which have led to disruptions in malaria services in several high-burden African countries. The best available treatment, particularly for P. falciparum malaria, is artemisinin-based combination therapy (ACT), which combines artemisinin or its derivatives with a partner drug. In recent years, the emergence of antimalarial drug resistance in East Africa has been of great concern. Drugs compatible with artemisinins, such as piperaquine phosphate, pyrimethamine, mefloquine, and amodiaquine, are widely used in treating malaria but have developed resistance (Menard andDondorp 2017, Pasupureddy et al. 2019). Although artemisinin and its derivatives are also resistant in the border areas of Cambodia and Thailand (Muller et al. 2019, van der Pluijm et al. 2019, they are still available on the condition that the paired drugs are effective locally. ACTs are still efficacious in countries in the WHO African Region, thus there should be no immediate impact for patients. Therefore, we hope to select an antimalarial drug that is not extensively used and compatible with artemisinin to form a new effective therapy with broad application prospects.
Hydroxychloroquine sulfate is a derivative of chloroquine, belonging to the 4-aminoquinolines, with similar effects to chloroquine, such as treating malaria, regulating immunity, antibacterial (Ben-Zvi et al. 2012). But chloroquine has serious drug resistance and side effects, leading to limited clinical application (Marques et al. 2014). Hydroxychloroquine, which is relatively less toxic than chloroquine, is currently one of the first-line drugs for the treatment of rheumatoid arthritis, but it is rarely used as an antimalarial in clinic (Hu et al. 2017, Jorge et al. 2018. Recently, we developed artemisininhydroxychloroquine sulfate tablets (AH) to treat malaria. Our previous studies have demonstrated that the combination of artemisinin and hydroxychloroquine sulfate has many benefits, including improving the therapeutic efficacy of malaria, decreasing the dose of both, reducing the toxic side effects of hydroxychloroquine, and postponing the progress of single drug resistance (unpublished data). In addition to being antimalarial, hydroxychloroquine and artemisinin are also antivirals with broad activity, such as anti-flavivirus and anti-coronavirus (Andreani et al. 2020, Cao et al. 2020, Liu et al. 2020, Wang et al. 2020. Recent reports showed they have been used in many clinical trials to treat COVID-19 (Nicol et al. 2020, Lewis et al. 2021, Li et al. 2021a. Hence, as a combination of hydroxychloroquine and artemisinin, AH has great potential in treating many diseases, particularly coronavirus infection. Artemisinin is generally viewed as a safe drug, but some adverse drug reactions of hydroxychloroquine sulfate may occur in clinical use, including retinopathy, cardiomyopathy, neuromuscular disease, myopathy, etc. (Al-Bari 2015). We had previously reported the studies of acute and subchronic toxicity of AH in rodents (Li et al. 2022). As a series of preclinical safety investigations, the present studies were conducted in non-rodents to obtain comprehensive preclinical data on the safety of AH. The results could provide scientific information for clinical trials of AH in the future.

Test compound
AH were produced in a GMP facility of Artepharm Co., Ltd (Meizhou, China). The specification of AH is 250 mg, each tablet contains 50 mg artemisinin and 64.6 mg hydroxychloroquine sulfate. The batch lot was 20170901, and the content of artemisinin and hydroxychloroquine sulfate in each tablet was 96.2% and 97.4% of the labeled amount, respectively.

Animals
Male and female beagle dogs were purchased from Najing Bigdoor Bioscience Inc. (Nanjing, China). The laboratory animal production license number was SCXK (Suzhou) 2016-0007. Dogs were housed individually in a stainless steel cage under conventional conditions. Environmental conditions were maintained at room temperature 20-26 C, relative humidity 40-70%, under 12 h light/dark cycles. The animals were acclimatized under laboratory conditions for 2 weeks before dosing. All animal experiments were carried out in the laboratory of the New South Center of Safety Evaluation for Drugs of Guangzhou University of Chinese Medicine (China animal use license number: SYXK (Guangdong) 2018-0014) following the principles of good laboratory practice (GLP) of the China Medical Products Administration. The study protocols were approved by the Institutional Animal Care and Use Committee (IACUC) and the use of animals was based on the 3 R principle (reduction, replacement, and refinement).

Animal selection rationale
Since different species of animals have their own characteristics, responses and biotransformation to the same test substance may vary. The current pharmaceutical guidelines recommend that acute and subacute toxicity studies shall be carried out in two species of mammals, generally one rodent and one non-rodent (FDA 1996, ICH 2009, EMA 2010. Therefore, rats and dogs were selected to study the acute and subacute toxicity of AH, and the results of rats have been published (Li et al. 2022). In addition, beagle dogs are usually the preferred non-rodent animals and are widely used in various toxicological tests of pharmaceuticals due to their small size, docile personality, accessible training, few genetic diseases, good consistency, and repeatability in response to experimental conditions.

Acute toxicity study
Six dogs (7-8 months old, weighing 8.59-9.16 kg, 3 males and 3 females) were randomly allocated into six dosages sorted by random numbers, one dog in each group. According to the results from a pilot toxicity study, the dose of AH for 100% death in rats was 10 g/kg, so we presumed that the equivalent lethal dose of AH to dogs could be approximately 2.66 g/kg based on the equivalent body surface area conversion method. Using 2.66 g/kg as a baseline, a 50% increment method was used to extend upward and downward to design a sequence table of 9 doses, including 5000, 3990,2660,1770,1180,790,530,350, and 230 mg/kg. Within this dose range, five of the dogs were administrated a single oral dose of 5000 (male), 2660 (male), 1180 (female), 530 (male), and 230 mg/kg (female) to measure the lowest lethal dose and the highest non-lethal dose. Then the remaining one dog (female) was administered with the dose between the two lethal doses to obtain an approximate lethal dose range or not receive treatment if no dog died at 5 doses.
The dogs were thoroughly observed for mortality and visible toxicity signs for 6 hours following dose administration on Day 0, and thereafter twice daily for 14 days for appearance, behavioral activity, gland secretion, respiration, stool properties, etc. Individual body weights were measured on Day 0, 1, 3, 7, 14, and when the death occurred during the study period. Food consumption was evaluated after each feeding. Food consumption was assessed 30 min after each feeding, which was divided into four grades according to the proportion of the residue amount to the feeding amount: good (no residue), poor (less than 1/3), very poor (1/3-2/3), and extremely poor (more than 2/3). Hematology, serum biochemistry, and electrocardiography (ECG) examinations were determined at pre-dose and Day 1, 14 after dosing. The general gross autopsy was performed on all dogs, including the dead dogs during the study period and the survival dogs at termination. Tissues would be processed for histopathological examination if they show any abnormality.

Subacute toxicity study
Forty dogs of both gender, 7-8 months old and weighing 7.14-8.71 kg, were randomly assigned to a control group and AH 56, 84, 126 mg/kg groups (defined as low dose, middle dose, high dose group, respectively). Each group consisted of 5 females and 5 males. The dose levels of AH were selected based on its intended adult clinical dose and the results from a previous 14-day subacute toxicity study in rats (Li et al. 2022). The maximum proposed daily dose of AH for adults is 1880 mg. According to the equivalent body surface area conversion method, if the adult weight is 60 kg, the clinical dose is 31.3 mg/kg, comparable to 53.8 mg/kg for dogs. The low dose of 56 mg/kg is close to the clinical equivalent dose. The middle and high doses were 84 and 126 mg/kg according to the principle of 1.5 times equal ratio. So the low, medium, and high doses were 1.8, 2.7, and 4.0 folds of the clinical dose, respectively. In addition, the dose multiple was chosen to be consistent with that in the subacute toxicity study in rats, which dose levels were 146, 219, 328, and 429 mg/kg, respectively. The obvious toxicities of AH to rats were observed at doses of 328 mg/kg and above.
Dogs in treated groups were orally administrated with AH once per day. After treatment for 14 days, six dogs (3 males and 3 females) of each group were sacrificed, and the remaining dogs were continuously observed for 56 days and sacrificed after the recovery period. The day of the first administration was considered as D 0 , and the day of anatomy after administration, i.e., the first day of recovery period was rD 1 .
Animals were observed twice daily for clinical signs, such as appearance, behavior, secretions and excretions, respiration, and other toxicity symptoms. Body weight and temperature were measured using a TJ60KY electronic platform scale (G&G, China) and an MC-141W electronic thermometer (Omron, Japan) at least twice weekly throughout the administration period and once weekly during the recovery period. After each feeding, food consumption was evaluated in the same manner as described in the acute toxicity study.
All animals were anesthetized intravenously with pentobarbital sodium (Hua Xia Chemical, China) at a dose of 30 mg/kg and euthanized by acute hemorrhage. A complete necropsy was conducted after the terminal blood collection. The weights of the brain, liver, heart, kidneys, adrenal glands, spleen, thyroid, thymus, uterus, ovaries, testis, and epididymides were measured and used to calculate the organ index (absolute organ weight/body weight). Sternal bone marrow was harvested to prepare a bone marrow smear with fetal bovine serum. After Wright's staining, cell morphological changes were observed under a CX41 biomicroscope (Olympus, Japan). Histopathological examination was performed in the following tissues: brain (cerebrum, cerebellum, brain stem), hypophysis, spinal cord (cervical, thoracic, lumbar), thymus, thyroid, parathyroid, salivary gland, eyeball, optic nerve, trachea, esophagu, cholecysts, aorta, femur, heart, lung, liver, kidneys, adrenal gland, spleen, pancreas, stomach, duodenum, jejunum, ileum, colon, cecum, rectum, lymph node (mesenteric lymph node and superficial lymph node), bladder, sternum, ischiadic nerve, vagina, uterus, cervix, ovaries, oviduct, breast, testis, epididymides, prostate, seminal vesicle, skeletal muscle, and skin. The testis was fixed with improved Davidson fixative solution, and the eyeball was fixed with Davidson fixative solution. The remaining tissues were fixed with 10% neutral buffered formalin. These tissues were trimmed, paraffin-embedded, sectioned, stained with hematoxylin and eosin, and finally assessed by an experienced pathologist with a BX63 biomicroscope (Olympus, Japan).
Furthermore, we conducted a accompanied toxicokinetic (TK) test in the subacute toxicity study of AH in dogs, which was previously published (Li et al. 2021b).

Statistical analysis
The data were presented as mean and standard deviation (mean ± SD). The statistical analyses were performed using SPSS 19.0 software. If the variance was homogeneous, the difference between groups was evaluated by the one-way analysis of variance (ANOVA) test. If not, the Kruskal-Wallis non-parametric test was used. If the one-way ANOVA test showed a significant difference between groups, the analysis was continued by the multiple comparison procedure of Dunnett's test. If the Kruskal-Wallis non-parametric test was significant, the analysis was continued by the the Mann-Whitney U test. The data of body weight and temperature were analyzed by repeated-measures ANOVA. Differences were considered statistically significant when p 0.05.

Acute toxicity assessment
Beagle dogs at doses of 5000, 2660, and 1180 mg/kg started to show gait instability, limb weakness, mental fatigue, tachypnea, and convulsion at 21, 27, and 37 min post-dose, and then died at 33, 35, and 53 min post-dose, respectively. While dogs at doses of 530 and 230 mg/kg doses survived. Based on the death of animals, 1180 and 530 mg/kg were the lowest lethal dose and the highest non-lethal dose, so the dose of 790 mg/kg between the two was administered to the remaining one dog. Except for the dogs at doses of 790 and 530 mg/kg vomited once at 3 and 5.17 h after administration, respectively, no other toxic symptoms were found on the day of administration and 14-day observation period in all surviving dogs. Therefore, the approximate lethal dose range for a single oral administration of AH to beagle dogs was from 790 mg/kg to 1180 mg/kg.
The body weight of all surviving dogs increased overtime during the observation period (Figure 1), and the food consumption of each feeding was good. As shown in Supplementary Tables S1 and S2, on Day 1, the AST, ALT, and QT values in the dog at 790 mg/kg were slightly increased compared with pre-dose (AST: 30.8 ! 81.7, ALT: 33.46 ! 64.22, QT: 197 ! 239) but decreased to the average level on Day 14 (AST: 28.1, ALT: 25.72,QT: 209). There was no apparent change in hematology, serum biochemistry, and ECG in the dogs at doses of 230 and 530 mg/kg throughout the study (Supplementary Tables S1, S2, and S3). No gross pathological signs were observed in any organs of all animals during necropsy.

Subacute toxicity assessment
3.2.1. Clinical signs, body weight, food consumption, and body temperature No mortality occurred in any group throughout the experiment. During the administration period, vomiting was observed in 1 male at low dose and 3 males at middle dose, soft feces was found in 2 females at middle dose and 1 male at high dose, and decreased activity was noted in 2 females at high dose. All the above symptoms were mild and disappeared within two weeks after stopping administration.
As shown in Figure 2, there was no significant difference in body weight among the groups during the administration and recovery periods. Notably, compared with D 0 , the body weight gain of the dogs in the control group was at least 0.5 kg after 14-day administration, 1 male and 1 female at low dose and 2 females at high dose gained less than 0.5 kg in body weight (low dose group: 0.040 kg, 0.084 kg; high dose group: 0.300 kg, 0.231 kg), while they gained more than 0.7 kg after previous 21-day acclimation (low dose group: 1.27 kg, 0.72 kg; high dose group: 1.422 kg, 0.762 kg), suggesting that AH may reduce body weight gain of individual dogs, but it can gradually increase and return to average weight after drug withdrawal.
During the administration, 4 females at high dose presented poor food consumption but quickly recovered in a week without administration. No significant difference in the body temperature of dogs was observed among the groups throughout the study period.

Ophthalmology, ECG, and urinalysis
No treatment-related changes were observed in ophthalmological examination and urinalysis during the study.
The results of ECG parameters determined at the end of administration are listed in Table 1. The Pa value in males of the high dose group was significantly higher than that in the control group, but the values of individuals were within the normal ranges, thus it was not considered to have toxicological meaning. A minor yet significant increase of QTc was found in males of the high dose group compared to the vehicle control group. After analyzing the individual QTc values of all groups, it was found that the QTc values of 1 female at low dose and 1 female at high dose were 364 ms and 361 ms, respectively, which were more prolonged than that in the control group (288-325 ms) and their QTc values in the previous acclimation (285 ms and 298 ms), and no obvious change was noted in the QTc values of other dogs. Therefore, we presumed that AH might prolong the QTc of individual dogs, but it can be returned to typical values after 28 days of recovery (Supplementary Table S4).

Hematology and serum biochemistry
As shown in Table 2 and Supplementary Table S5, no statistically significant changes were observed in hematological parameters throughout the experiment.
For the coagulation parameter examined (Table 3 and  Supplementary Table S6), after 14 days of AH administration, a significant increase of TT in males at middle dose was toxicologically irrelevant due to the absence of a dose-response relationship, and the significant reduction in Fbg in males at high dose also had no toxicological significance because the difference was within historical values of control dogs.
The biochemical data are summarized in Table 4 and  Supplementary Table S7. At the end of administration, CHOL and Ca significantly increased in females at middle and high dose, LDH significantly decreased in females at middle dose and males at low dose, Na þ significantly increased in males at high dose, K þ significantly decreased in males at low and high dose. The above changes lacked dose correlation or  were within historical values of control dogs, thus considered of no toxicological meaning.

Bone marrow smear, organ index, and histopathology
At the end of the administration and recovery period, macroscopic and bone marrow smear examination of dogs in control and all treated groups revealed no abnormalities. A significant increase of testis index in males at low dose was found at the end of the administration, which was not considered toxicologically meaningful since there was no doserelated response (Supplementary Table S8).

Discussion
Chloroquine is dangerous and potentially lethal in overdose, causing severe acute toxicity, including coma, acute respiratory distress syndrome, hypotension, cardiac arrhythmias, and fatal cardiac arrest (Watson et al. 2020). Although the median lethal dose (LD 50 ) for chloroquine and hydroxychloroquine in humans has not been determined, the most frequently reported lethal dose in adults is 3-4 g. The described toxidromes are similar for hydroxychloroquine and chloroquine. The difference between hydroxychloroquine and chloroquine is only one hydroxyl group, but animal toxicological studies demonstrated that hydroxychloroquine is about 40% less toxic than chloroquine (Della Porta et al. 2020, Doyno et al. 2021. Artemisinin and its derivatives are usually considered safe drugs with a wide therapeutic dose range. Although dose-dependent neurotoxicities, cardiovascular toxicities, and gastrointestinal side effects of artemisinin have been reported in animal and human studies, large clinical studies and meta-analyses have not shown that artemisinin has serious side effects (Efferth and Kaina 2010, Marques et al. 2014, Li et al. 2019, Whegang Youdom et al. 2019, Shibeshi et al. 2021. The reported LD 50 of oral artemisinin in rats and mice was 5576 mg/kg and 4228 mg/kg, respectively (Zhang and Yan 2005). Previously, we have reported the LD 50 of acute oral administration of AH in rats was 3119 mg/kg, and the dying rats presented dyspnea, spasticity, and convulsion (Li et al. 2022). In the present acute toxicity study of AH, death occurred in dogs at single oral doses of 1180, 2660, and 5000 mg/kg, with toxic symptoms prior to death including gait instability, limb weakness, mental fatigue, tachypnea, convulsion, suggesting that a high dosage of AH also has neurotoxic effects on dogs and can be lethal. In vitro and in vivo studies suggest that currently deployed artemisinins have neurotoxic potential. There are also reports of human hearing loss, ataxia, and tremor caused by artemisinins (Medhi et al. 2009, Li and Hickman 2011). Hydroxychloroquine has a wide range of neuropsychiatric toxicity with varying incidence. The toxicity of hydroxychloroquine to the central nervous system includes headache, dizziness, balance disorders, seizures, paresthesia, and so on (Doyno et al. 2021). Furthermore, Boulware et al. (2020) reported that adults taking hydroxychloroquine for postexposure prophylaxis of COVID-19 had a higher incidence of neurological reactions compared to placebo (5.4% vs 3.7%), such as irritability, dizziness, or dizziness. Thus neurotoxicity should be of concern if AH is used to treat COVID-19 in the future.
The subacute toxicity study showed that AH caused vomiting, soft feces, decreased activity, and poor appetite in dogs, which could be due to the common gastrointestinal side effects of artemisinin and hydroxychloroquine as nausea, anorexia, vomiting, diarrhea, etc. However, the cases of toxicity symptoms were few and mild, and they can be quickly recovered after drug withdrawal. Although there was no statistically significant difference in body weight of dogs in all Table 5. Histopathological findings of dogs after 14 days of administration with AH (n ¼ 6). treated groups, a few individual dogs had no or slight weight gain during administration. The reason might be partly related to gastrointestinal side effects of AH.
Hydroxychloroquine has been documented to have potential cardiotoxicity, affecting ventricular myocytes by blocking the potassium outflow channel, resulting in QT interval prolongation, which is associated with increases in both arrhythmic and non-arrhythmic mortality. Hence, it is recommended to examine ECG regularly before and after hydroxychloroquine in clinical use (Bonow et al. 2020, Chorin et al. 2020, Gasperetti et al. 2020, Oscanoa et al. 2020. The clinical reading of the QT interval normally takes into account a correction for the heart rate (QTc). In the present studies, a prolonged QT interval was observed in the dog with single oral administration of AH at 790 mg/kg, and the QTc interval was also prolonged in a few individual dogs after repeated doses of AH, both possibly due to the cardiotoxic effect of hydroxychloroquine. However, the above was only modest QT or QTc prolongation and did not cause severe cardiotoxicity.
Drug-induced phospholipidosis (DIPL) refers to a lysosomal storage disorder characterized by excessive accumulation of phospholipids in the kidney, liver, brain, lung, cornea, and other organs after long-term treatment with cationic amphipathic drugs in animals and humans (Breiden and Sandhoff 2019). Its typical pathological features are foamy or vacuolated changes of macrophages or various parenchymal cells under the optical microscope. Hydroxychloroquine has been shown to induce phospholipid accumulation in clinical application and can be used for the treatment of antiphospholipid antibody syndrome (Edwards et al. 1997, Espinola et al. 2002, Sperati and Rosenberg 2018. The pathological damage to the liver and kidney caused by DIPL has been observed in the previous subacute toxicity of AH in rats. A similar pathological finding of splenic red pulp vacuolation possibly induced by DIPL also appeared in dogs at the high dose in the present subacute study. In addition, according to the accompanied TK results of AH in the subacute toxicity study (Li et al. 2021b), the exposure of hydroxychloroquine in dogs increased in a linear correlation with increasing doses after the first and last administration, and there was a mild accumulation in dogs following repeated doses of AH. The TK results were basically consistent with the toxic findings of subacute toxicity of AH in dogs.

Conclusion
In summary, the approximate lethal dose range for acute oral administration of AH in dogs is found to be 790-1180 mg/kg, and toxic symptoms prior to death include gait instability, limb weakness, mental fatigue, tachypnea, and convulsion. In the current subacute toxicity study, repeated doses of AH in dogs caused vomiting, soft feces, decreased activity, anorexia, and splenic red pulp vacuolation. Of note, AH could reduce body weight gain and prolong the QTc interval of individual dogs. Therefore, the no-observed-adverse-effect level (NOAEL) and lowest-observed-adverse-effect level (LOAEL) of oral administration of AH for 14 days in dogs are determined to be 84 mg/kg and 126 mg/kg, respectively, which are about 2.7-fold and 4.0-fold of its clinical intended dosage. Further toxicity studies should explore the effects of longer-term administration of AH in rodents and non-rodents.