Preparation of Decoquinate Solid Dispersion by Hot-Melt Extrusion as an Oral Dosage Form Targeting Liver-Stage Plasmodium Infection

ABSTRACT

Liver-stage Plasmodium in humans is an early stage of malarial infection. Decoquinate (DQ) has a potent multistage antimalarial activity. However, it is practically water insoluble. In this study, the hot-melt extrusion (HME) approach was employed to prepare solid dispersions of DQ to improve oral bioavailability. The DQ dispersions were homogeneous in an aqueous suspension that contained most DQ (>90%) in the aqueous phase. Soluplus, a solubilizer, was found compatible with DQ in forming nanoparticle formulations during the HME process. Another excipient HPMC AS-126 was also proven to be suitable for making DQ nanoparticles through HME. Particle size and antimalarial activity of HME DQ suspensions remained almost unchanged after storage at 4°C for over a year. HME DQ was highly effective at inhibiting Plasmodium infection in vitro at both the liver stage and blood stage. HME DQ at 3 mg/kg by oral administration effectively prevented Plasmodium infection in mice inoculated with Plasmodium berghei sporozoites. Orally administered HME DQ at 2,000 mg/kg to mice showed no obvious adverse effects. HME DQ at 20 mg/kg orally administered to rats displayed characteristic distributions of DQ in the blood with most DQ in the blood cells, revealing the permeability of HME DQ into the cells in relation to its antimalarial activity. The DQ dispersions may be further developed as an oral formulation targeting Plasmodium infection at the liver stage.

INTRODUCTION

Malaria is a mosquito-borne parasitic infection. Uninucleate sporozoites of Plasmodium (P) parasites are injected into a human host when Anopheles mosquitoes bite. The parasites rapidly invade liver parenchymal cells, where they mature into liver-stage schizonts, which burst to release uninucleate merozoites (1). Merozoites can infect red blood cells. The clinical symptoms of malaria are manifested when parasites invade and multiply inside human red cells. In P. vivax and P. ovale infections, the sporozoites differentiate into hypnozoites and liver schizonts. Schizonts are distinguishable from hypnozoites based on size and morphology. The hypnozoites can hide out in the liver and cause a relapse of malaria months or even years later. The liver stage presents the best opportunity for developing drugs for both prophylaxis and eradication of malaria (2, 3). In human primary hepatocytes, P. vivax completes development after day 9 whereas P. falciparum completes development after day 7, which opens a time window for antimalarial intervention (4).

Primaquine, tafenoquine (Arakoda), and atovaquone/proguanil (Malarone) are antimalarials targeting the liver stage of infection. Primaquine is used for interrupting the formation of both schizonts and hypnozoites (5). However, it is contradicted in patients with glucose-6-phosphate dehydrogenase (G6PD) deficiency. Tafenoquine, an analog to primaquine, is a drug newly approved by FDA for prophylaxis of malaria and for the radical cure (prevention of relapse) of vivax malaria (6, 7). However, tafenoquine use in people with G6PD deficiency or unknown G6PD status is also contraindicated (8). Atovaquone/proguanil (Malarone) is prescribed for the prevention of malaria while traveling but is ineffective against late, hypnozoite reactivation-related attacks (9).

Decoquinate (DQ) is a veterinary medicine for controlling the infection of coccidiosis in the digestive tract (10, 11). DQ is also active against other parasites (12–16). Meister et al. (12) identified DQ as a potent multistage plasmodium inhibitor. However, its potential as an antimalarial drug has been unexploited. DQ is practically insoluble in water and very slightly soluble in most organic solvents. Wang et al. (17, 18) showed that formulation and particle size reduction improve bioavailability and antimalarial efficacy of DQ. Compared to DQ microparticle suspension, a nanoparticle formulation orally dosed to mice showed a significant increase in DQ blood concentration (14.47-fold) and liver distribution (4.53-fold), and a dramatic improvement of antimalarial efficacy (15-fold). However, these were achieved, in addition to cosolvency and spray dry, by using extra procedures such as ultrasonication or high-pressure homogenization for particle size reduction.

Hot-melt extrusion (HME) technology is a novel approach for improvement of compound solubility and bioavailability in pharmaceutical formulation. The technology has an advantage of being solvent-free, continuous operation, and highly efficient. However, it was unknown whether this method was suitable for DQ formulation. Since only nanoparticle formulations can improve bioavailability and efficacy of DQ without changing molecular structure (17, 18), ideally, nanosized particles of the dispersions can be obtained by the optimization of hot-melt extrusion process. In addition, selection of polymer-surfactant combinations using solubility parameters is also important (19).

In this study, DQ formulations were made by the HME. The extruded solids were suspended in aqueous phase and determined to be nanosized particles comparable to the size of DQ nanoparticles described previously made by non-HME method (18). This was achieved without using additional mechanic approaches for particle size reduction such as ultrasonication or high-pressure homogenization. Drug dissolution, thermogravimetry (TG), differential scanning calorimetry (DSC), and X-ray diffraction (XRD) were applied to analyze HME DQ products. HME DQ had strong inhibitory effects on Plasmodium development and growth in vitro at the liver stage and the blood stage and was highly effective in vivo for causal prophylaxis of malarial infection in mice. Pharmacokinetic analysis of orally administered HME DQ in rats showed differential distributions of DQ between the plasma and the blood cells. The predominant accumulation of DQ inside the cells might contribute to its antimalarial efficiency.

RESULTS

DQ is thermodynamically stable and suitable for HME solid dispersion.

Evaluation of solid dispersion in different pH solutions depended on the physicochemical property of carriers (Table 1). HPMC AS is an enteric polymer that dissolves in water at pH ≧5.5 (20). The nonsalt form of DQ, an active pharmaceutical ingredient (API), is incorporated into a formulation with HPMC AS to target intestinal release of API. Thus, the formulation is suspended in a buffer with slightly base pH (7.4). Soluplus and VA64 are amphiphilic polymers employed to deliver a faster release of the poorly soluble API DQ. Thus, the solid dispersions with these excipients were suspended in a little strong acidic buffer with pH 3 to 7.

TABLE 1.

Physical characterization of the DQ HME formulations (F1–F15)^a

Formulations	PS (nm)	PDI	Drug load (%)	% of drug in AS	Drug/formulation (g/g)	Medium and pH
F1	1011	0.202	10	93.10	18.62/180	PBS (pH 7.4)
F2	1097	0.055	20	97.20	38.88/160	PBS (pH 7.4)
F3	4385	0.220	10	97.70	19.54/180	PBS (pH 7.4)
F4	2232	0.185	20	98.10	39.24/160	PBS (pH 7.4)
F5	898	0.255	20	94.70	28.41/120	PBS (pH 7.4)
F6	762	0.204	20	93.70	28.11/120	PBS (pH 7.4)
F7	588	0.235	25	93.72	34.68/113	PBS (pH 7.4)
F8	227	0.108	25	98.92	36.60/113	PBS (pH 7.4)
F9	820	0.233	25	99.32	49.66/150	PBS (pH 7.4)
F10	627	0.238	25	98.20	29.46/120	HCl (pH 4.0)
F11	891	0.343	25	84.40	42.20/150	HCl (pH 4.0)
F12	505	0.352	10	95.70	19.14/180	Saline (≈pH 5.5)
F13	345	0.184	20	91.25	36.50/160	Saline (≈pH 5.5)
F14	317	0.235	20	93.60	28.08/121	Saline (≈pH 5.5)
F15	209	0.154	25	96.52	48.26/150	Saline (≈pH 5.5)

^aFormulations were suspended in the solutions indicated in the right and stirred at room temperature for at least 24 h for evaluation. Decoquinate (DQ) in AS measured by high-performance liquid chromatography and divided by drug load gave the drug percentage (%) present in AS. The ratio of drug/formulation components (excipients) represent how much added drug was fused with formulation particles through hot-melt extrusion (HME) process. AS, aqueous suspension; PDI, polydispersity index; PS, particle size.

After overnight shaking at room temperature, particle size and drug content suspended in the buffer were measured. Among hundreds of formulations prepared and evaluated, 15 formulations with small particle size and high DQ content in the aqueous phase were selected and are shown in Table 1. The particle sizes ranged from 207 to 1,097 nm. Later preparation of F8 had a particle size (227 nm) much smaller than that of the early preparation. However, particle stability had been only assessed for the early preparation (Table 1; see Fig. 3A). F15 (207 nm) had the smallest particles. The aqueous suspension of F8 and F15 appeared to be homogeneous and remained to be stable for at least 1 week without particle size change, precipitation, agglomeration, and cloudiness. The amount of DQ suspended in aqueous solution determined by high-performance liquid chromatography (HPLC) was greater than 90% of total DQ loaded in all formulations except for F11 (Table 1).

FIG 3.

The stability of particle size and functionality of HME formulations of F8 and F15. (A) The stability of nanoparticles of F8 and F15. The stock suspensions of F8 in PBS (pH 7.4) and F15 in saline solution (slightly acidic) were stored at room temperature and the stability of the particles observed periodically for more than 12 months. The formulations were measured in triplicate after dilution with sterile water. (B) The stability of antimalarial activity of F8 and F15. Inhibition of Plasmodium berghei was examined in vitro 79 days after preparation of the formulations. HepG2 liver cells were incubated with sporozoites of P. berghei expressing firefly luciferase. The formulations were diluted to series of concentrations and each diluted drug solution added to the culture in triplicates (n = 3). After 48 h, parasite inhibition was measured by the detection of luminescent intensity. The experiments were repeated at least three times.

TG analysis of DQ showed that the amount of DQ was reduced to 99% by weight at 250.5°C, indicating that the compound was very stable, no thermal decomposition occurred, and the minimal loss could be counted as water molecules (Fig. 1A). DSC graphs showed that HME DQ formulation (F15) lost the peaks of pure DQ (Fig. 1B). The physical mixture of F15 also did not retain entire peaks of pure DQ, which could be due to the interaction of DQ with other components as temperature increased in the analytical procedure. The results indicated that the extruded material had completely lost DQ peaks when it was melted or fused with excipients through HME process. Three large batches of F8 preparation were analyzed by DSC (Fig. 1C) and by XRD (Fig. 1D). In DSC plots, all three peaks of large preparations were reproducibly generated and significantly different from that of the physical mixture of F8, suggesting an extrudate with reproducibility and stability. In XRD analysis, a diffraction pattern plots intensity (y axis) against the angle of the detector, 2θ (x axis). The peaks of three different lots shown in XRD analysis (Fig. 1D) were quite minor compared to those of pure DQ and DQ in physical mixture of F8, indicating the fusion of most DQ with the excipients through HME process, with existence of very small amount of crystalline material (≤3%) compared to the typical peaks of pure DQ representing the crystalline form. The DSC and XRD patterns of material prepared by HME process may help understand the quality of the final products (21, 22).

FIG 1.

Thermogravimetric (TG) analysis of thermostability of pure decoquinate (DQ) and differential scanning calorimetry (DSC) and X-ray diffractometer (XRD) patterns of hot-melt extrusion (HME) DQ. (A) TG graph of pure DQ. (B) DSC. Red: reference standard of pure DQ. Blue: HME processed F15 formulations. Green: HME processed vehicle components of F15 (no DQ). Pink: physical mixture of F15 (no HME). (C) Graphs of DSC for hot-melt extrudates of F8 in three different lots of scale up preparation (top 3 samples) and for physical mixture of F8 without HME (a sample at bottom). (D) Graphs of XRD for the same samples used for DSC in C plus the sample of pure DQ (the last graph). Notice that the peaks shown in y axis with the scale up samples of three different lots are quite minor compared to the peaks of DQ in physical mixture (F8) and pure DQ.

In vitro dissolution rates.

F8 which contains HPMC AS-216 was compared with F7 composed of HPMC AS-912 as a carrier. Both neutral media (phosphate-buffered saline [PBS], pH 6.8) and acidic media (0.1 N HCl and 10 mM sodium dodecyl sulfonate [SDS]) were used for dissolution tests (Fig. 2A and B). F8 obviously had a higher release rate of DQ than F7 in acidic media and in neutral media after 90 min, suggesting that AS-216 is more suitable for fast release of DQ. In addition, it took 9 h for F8 to release 80% of DQ in acidic dissolution media but F8 released 80% of DQ in only 2 h in pH 6.8 media. This feature might be beneficial for intestinal absorption.

FIG 2.

In vitro dissolution rates of accumulated DQ; 0.1 N HCl/10 mM SDS and PBS (pH 6.8) were used as dissolution medium for F7 and F8. (A) Drug release in acidic medium for F7 and F8. (B) Drug release in PBS (pH 6.8) for F7 and F8. (C) Drug release in acidic medium for F14 and F15. Samples taken at each time point indicated on the graphs according to the time course of specific formulations and DQ measured by high-performance liquid chromatography. Each solid dispersion sample was cut in equal amount, and dissolution tests done in triplicates (HCl media) or in six replicates (PBS pH 6.8). Each experiment was repeated at least three times. Data were plotted by using Prism program.

Both F14 and F15 had a drug load of 25% and Soluplus as a main carrier. However, F15 was much more efficient than F14 in releasing DQ. F15 had dissolution rate plateaued in 60 min whereas F14 needed 180 min to achieve the plateau (Fig. 2C). The reason could be that F15 had PEG 6000 as a plasticizer, which helped not only lower the melting temperature during HME process but also the DQ release after the formation of HME formulations (23).

Stability of HME DQ.

F8 suspended in PBS (pH 7.4) and F15 in saline solution (slightly acidic) were selected to observe the stability of DQ HME formulations. No significant changes in particle size were found for both formulations over a period of 1 year after storage at 4°C (Fig. 3A). Their antimalarial activity was evaluated in hepatic cells (HepG2) infected by P. berghei against atovaquone and primaquine. Both formulations had potent inhibitory effects (half-maximal inhibitory concentration [IC₅₀] <0.5 nM), and the potency remained roughly unchanged 79 days after the formulations were prepared (Table 2 and Fig. 3B).

TABLE 2.

Inhibitory effects of DQ HME formulations on Plasmodium berghei ANKA expressing luciferase at the liver stage^a

Antimalarials	PQ IC50 (μM)		AQ IC50 (nM)
	Mean (n = 3)	SD	Mean (n = 3)	SD
Day 01	1.882	0.584	9.497	2.239
Formulation/days tested	F8 IC50 (nM)		F15 IC50 (nM)
Day 01	0.342	0.073	0.349	0.054
Day 30	0.168	0.086	0.195	0.035
Day 64	0.432	0.034	0.401	0.012
Day 70	0.312	0.060	0.296	0.023
Day 79	0.171	0.064	0.176	0.015

^aHepG2 cells were cultured and infected by the sporozoites (SPZ) isolated from salivary glands of infected mosquitoes. Aqueous suspensions of F8 and F15 were stored at 4°C and tested on different days after they were made to assess the stability of the formulations. The results are expressed by the half-maximal inhibition concentration (IC₅₀). In vitro potency of HME DQ against liver-stage Plasmodium between day 1 and all other days was not significantly different (P > 0.05). AQ, atovaquone; PQ, primaquine.

DQ associated with HME formulation (F8) was intact upon being exposed to high temperature (T), high humidity and strong light for 10 days. Pure DQ was not resistant to strong acid and alkaline after being exposed for 2 days whereas F8 DQ remained intact upon being exposed to strong acid and alkaline for 10 days (Fig. S1 in supplemental material). After three large-scale (kilograms) preparations of F8 by HME were exposed to high temperature, high humidity, strong light, strong acid, and strong base for 5 and 10 days, respectively, there was no loss of DQ in the recovery tests. Therefore, DQ in F8 formulation when prepared in large scale was also resistant to all tested conditions (supplemental Table 1).

Safety assessment of HME DQ.

F15 HME DQ was assessed in in vitro cytotoxicity assay (supplemental Table S2). The therapeutic index was calculated based on in vitro potency of DQ against P. falciparum (IC₅₀ = 2.51 nM) and the cellular toxicity of DQ (CC₅₀ > 100 μM). F15 DQ prepared by HME was basically nontoxic to HepG2 cells.

A single dose of F15 DQ at 1,000 and 2,000 mg/kg was given by intragastric gavage to NIH mice (supplemental Fig. S2). Saline and excipients of F15 (no DQ) were in each of the control groups (supplemental Table S3). No toxicity or abnormal behaviors were found in all mice. At the end of the experiments, all mice were alive and active. The body weights were increased but not significantly different among all groups during the observation period. Male group had apparently more increase in body weight than female group (supplemental Fig. S3).

Distribution of HME DQ in the blood.

After intragastric administration of F15 to rats, plasma DQ began to rise at 1 h, peaked at 3 h, and started to fall at 8 h (Fig. 4A). DQ concentrations in the plasma were proportionally much lower than those in the blood cells at 1 and 3 h (Fig. 4B). After intravenous injection of nanoparticle formulation of DQ to rats, DQ peaked in both the plasma and the blood at 30 min and then began to drop considerably (Fig. 4C). However, the DQ curve representing plasma concentration was much lower than that of the blood DQ concentration during the time from 2 min to 50 h. The results clearly indicated the majority of DQ accumulated in the blood cells. When oral route of DQ HME formulation (F15) was compared to intravenous injection of DQ nanoparticle formulation, at 1 h after administration, DQ concentrations in the blood distribution were quite similar between the two (Fig. 4D).

FIG 4.

Pharmacokinetics of HME DQ in Sprague-Dawley rats. Sample preparation and DQ measurements (liquid chromatography-mass spectrometry) were described in Materials and Methods. (A) DQ concentrations in the plasma. DQ at 50 mg/kg of F15 was administered to rats (n = 4) by intragastric gavage, and DQ was measured from samples collected at different time points. (B) DQ concentrations in the plasma and in the whole blood from rats (n = 3) given DQ nanoparticle formulation by intravenous injection. (C) DQ concentrations in the plasma and blood cells at 1 h and 3 h following F15 HME formulation administered to rats (n = 4) by intragastric gavage. (D) Comparison of DQ concentrations in the plasma and in the blood cells 1 h after F15 given to rats (n = 4) by intragastric administration with those after nanoparticle formulation given to rats (n = 3) intravenously. All experiments were repeated at least twice.

In vitro antimalarial potency of the DQ HME formulations.

Compared to chloroquine and artemisinin, F15 DQ exhibited excellent inhibitory activity against both the strains of P. falciparum 3D7 (Fig. 5A) and P. falciparum Dd2 (Fig. 5B) at the blood stage. The IC₅₀ values derived from the same experiments represented by Fig. 5A and B are summarized in Table 3. The antimalarial potency of both nanoformulation DQ and HME formulation DQ was slightly less than that of DMSO dissolved DQ. The differences may be due to drug availability of less exposure. For all Plasmodium species evaluated, DQ IC₅₀ values of F8 and F15 were below 5 nM (Table 4), which were highly potent in contrast to those of chloroquine and artemisinin.

FIG 5.

In vitro potency of HME DQ against Plasmodium falciparum at the blood stage. Antimalarial activity of F15 formulation was evaluated in cell culture of human red blood cells infected by the strains of P. falciparum parasites, chloroquine sensitive (3D7) and multidrug resistant (Dd2). Nanoformulations of DQ, standard DQ dissolved in DMSO, chloroquine, and artemisinin were also assessed in parallel. After 72 h, fluorescent signals generated by the binding of SYBR green I dye to Plasmodium DNA were measured. (A) Dose-response curves generated from inhibition of P. falciparum 3D7. (B) Dose-response curves from inhibition of P. falciparum Dd2. Each concentration of diluted HME DQ was tested in 5 replicates. The experiments were repeated multiple times (>3).

TABLE 3.

Inhibition of Plasmodium falciparum growth in infected human erythrocytes^a

Test agents		IC50 (nM)
	Particle size	Chloroquine-sensitive 3D7			Multidrug-resistant Dd2
	Mean (μm)	Mean	SD	P value	Mean	SD	P value
Artemisinin	n/a	7.900	2.260		15.910	3.730
Chloroquine	n/a	32.060	11.030		364.500	45.080
DMSO dissolved DQ	n/d	1.410	0.390		2.630	1.880
Nanoformulation DQ	0.21	2.100	1.330		3.110	0.860
HME DQ (F15)	0.47	2.390	1.230		3.530	0.740
Artemisinin verusus HME DQ				0.021			0.005
DMSO DQ versus Nano DQ				0.437			0.708
DMSO DQ VS HME DQ				0.259			0.483

^aF15 DQ was more potent than artemisinin and chloroquine against two different P. falciparum strains in the in vitro test (P < 0.05). F15 DQ had inhibitory potency comparable to DMSO dissolved DQ and nano-DQ (P > 0.05). Each value represents triplicate and the experiment repeated multiple times; n/d, not detectable; n/a, not applicable.

TABLE 4.

Inhibitory effects of the DQ HME formulations on different strains of P. falciparum expressed as the half-maximal inhibition concentration (IC₅₀)^a

Formulations/drugs	F8			F15			Chloroquine			Artemisinin
IC50 (nM)	Mean	n	SD	Mean	n	SD	Mean	n	SD	Mean	n	SD
Pf3D7	2.09	5	0.69	2.51	5	0.54	57.45	3	0.13	23.12	3	0.81
PfDd2	3.41	3	0.29	2.53	5	0.82	243.90	3	6.08	26.01	3	0.85
Pf803	2.65	8	0.42	1.89	3	0.40	957.20	3	132.20	32.00	3	0.85
PfB74F6 (PS)	2.33	3	0.75	1.47	3	0.15	79.20	3	4.45	14.47	3	3.57
PfB74D6 (PS)	2.68	3	0.79	1.70	3	0.64	78.15	3	7.03	25.44	3	1.77
PfB64G11 (PS)	2.21	3	0.48	0.45	3	0.11	30.38	3	0.14	6.66	3	1.54
PfB74C4 (PS)	4.07	4	0.98	0.32	3	0.13	59.29	3	0.74	26.41	3	0.60
PfB12D9 (PS)	3.47	4	0.61	1.64	3	0.50	63.30	3	4.03	24.12	3	1.80

^aPf: Plasmodium falciparum; 3D7, chloroquine-sensitive strain; Dd2, multiple drug resistance; 803, chloroquine resistance; PS, patients’ blood samples.

Antimalarial efficacy of the DQ HME formulations in mice.

Images generated by in vivo imaging system (IVIS) at 72 h postinoculation of the sporozoites showed the efficacy of different doses of DQ of F15 for causal prophylaxis of malaria of mice infected by sporozoites (SPZ) of P. berghei ANKA parasites (Fig. 6A). As shown in the images, control mice treated by vehicle components of F15 had infection signals in abdominal area, produced by luciferase-substrate reaction, within 48 h of SPZ infection (24 and 44 h). Then, at 72 h postinfection, the signals were spread to the whole body because in rodents, the parasites of this species burst to release uninucleate merozoites after 48 h. The mice died by day 21 after SPZ inoculation. Atovaquone and primaquine at therapeutic doses were given to infected mice, and they had a complete protection. F15 DQ was shown to be highly effective against Plasmodium infection at the liver stage with a dose-effect relationship. F15 DQ at 1 mg/kg did not provide complete protection, and two out of five mice got infected and died. One of these two mice developed parasitemia as high as 60% (Fig. 6B). None of the animals in groups of DQ at 3 and 10 mg/kg were infected, and therefore, they were fully protected. Survival rates were calculated on day 60 postinfection (Fig. 6C). At doses of 1, 3, and 10 mg/kg, the protective rates were 60%, 100%, and 100%, respectively. Consistent with the results of parasitemia in Fig. 6B, the mice that received F15 DQ at 3 mg/kg had a full protection. The data further supported that HME DQ at 3 mg/kg was sufficient in providing full protection for mice, which indicated that the development of P. berghei ANKA SPZ was blocked in the liver. Four experiments were repeated, and comparable results were obtained.

FIG 6.

In vivo efficacy of HME DQ against P. berghei at the liver stage. (A) The causal prophylaxis of F15 formulation in mice infected with P. berghei. The parasite expressing luciferase was monitored by the in vivo image system. VC control represents the excipient components of F15; atovaquone (ATOV; 10 mg/kg) and primaquine (PQ; 30 mg/kg) used as positive controls; F15 DQ at 1, 3, and 10 mg/kg was tested for inhibition of P. berghei at the liver stage. (B) Parasitemia indicated Plasmodium infection at the blood stage 72 h after sporozoite (SPZ) inoculation and the monitor continued periodically until 60 days after SPZ injection. (C) Survival rates summarized were on day 60 after the sporozoite inoculation.

More animal studies were conducted to assess antimalarial efficacy of causal prophylaxis for more HME formulations at a dose of 5 mg/kg DQ. F2 and F12 had protection rates of 80% whereas F8, F9, F11, and F15 provided 100% protection rates (Table 5). Although two mice got infected, one in the F11 and one in the F15 group, the DQ HME formulations had the ability to clear up parasites and enable the mice to survive up to day 60.

TABLE 5.

The causal prophylaxis of DQ HME formulations in female NIH-Swiss mice^a

Formulations or compounds	Dose (mg/kg)	n	Day 12		Day 60
			Infected	Protected rate	Survived	Survival rate
Saline	0	5	5	0	0	0
PQ	20	5	2	60	4	80
F2(DQ)	5	5	2	60	4	80
F8(DQ)	5	5	3	40	5	100
F9(DQ)	5	5	2	60	5	100
F11(DQ)	5	5	1	80	5	100
F12(DQ)	5	5	1	80	4	80
F15(DQ)	5	5	1	80	5	100

^aSelected DQ HME formulations (DQ dose of 5 mg/kg) were given to mice by intragastric gavage at day 1 (the day before sporozoite [SPZ] inoculation), day 0 (the same day), and day 1 (the day after) of SPZ inoculation. Blood smears from tail vein were examined at different days (days 7, 12, and 19). Survival rates were summarized at day 60.

DISCUSSION

DQ is known to be practically water insoluble. Previous studies have demonstrated that nanoparticle formulations of DQ for oral route significantly improved the bioavailability and efficacy of DQ. In contrast, large particles such as the microsized DQ formulation had some improvement compared to pure DQ but were much less effective (17–18). However, the methods used to make nanoparticles of DQ formulation of this kind were complex, used a large quantity of solvents, and needed extra procedures such as ultrasonic or high-pressure homogenization to reduce particle size. As a result, it would be difficult and costly to scale up or produce a large quantity. HME has been adopted by the pharmaceutical formulation from the plastic industry. One of the main applications of HME is to convert poorly water-soluble drugs or compounds to amorphous solid dispersion or solid solution to improve the solubility and bioavailability and therefore augment the bioactivity. Successful formulation examples through the HME approach have been described including those of itraconazole, artesunate, and clotrimazole (24–26). However, there are no previous reports by others in the literature of making DQ nanoparticles by HME without extra procedures.

The polymers evaluated for creating DQ HME formulations are generally recognized as safe for clinical use. The use of polymers and surfactants should be as minimal as possible to allow maximal drug load but sufficient to hold poorly water-soluble API DQ in the glassy amorphous matrix that is shelf stable. Stability of the formulations in resisting humidity, heat, light, acid, and base is assessed and described in the supplementary materials (Table S1). In vitro and in vivo toxicity of F15 was tested, and there was no evidence of obvious adverse effects. Therapeutic index from the in vitro tests was remarkably good (Table S2).

To achieve a goal of obtaining desired formulations for DQ by HME, finding appropriate excipients and corresponding validation parameters is challenging but essential. Many experiments have been conducted to pretest numerous different polymers and surfactants. Some formulations were found to be in nanosized particles or particles in low-micrometer range (Table 1). DQ is thermodynamically stable and suitable for HME formulation preparation (Fig. 1A and supplement Fig. S1). Analytical results of DSC and XRD further suggested that a physical interaction or fusion of DQ into the polymers HPMC AS (F8) or Soluplus (F15) had occurred through HME process.

In vitro dissolution rates of both F7 and F8 were quite different between acidic medium and pH 6.8 PBS regardless of the subtypes of HPMC AS used. It is interesting because this may indicate that the formulations had the ability to hold the API in the acidic gastric cavity and then release it in pH 6.8 fluid close to intestinal environment, which may help increase oral bioavailability. HPMC AS has been identified, among many materials assessed, as the most effective at inhibiting drug precipitation in supersaturated solutions and maintaining drug supersaturation in the gastrointestinal (GI) milieu (27). Based on physicochemical property of DQ HME formulations, HPMC AS-126 was more compatible with DQ than other subtypes of HPMC AS (716 and 912).

Unlike the dissolution data of F7 and F8, in the case of F14 and F15, most of the API in acidic medium were released within 2 h. If the released drug, presumably in the gastric cavity (acidic), could enter the intestine before crystallization, then it might be absorbed. Soluplus was used as a major carrier of these formulations. The ideal type of solid dispersion for increasing dissolution is solid glassy solution that is thermodynamically stable as the API and excipients are completely miscible with each other over the whole composition range (28). Soluplus is a polymeric solubilizer with an amphiphilic chemical structure, which was particularly developed for solid solutions (29). Extruded Soluplus-based oral solid dosage forms improved the dissolution of a poorly water-soluble drug (30). Soluplus used as a DQ carrier generated homogeneous nanoparticles, which might also improve the dissolution rates of DQ. PEG used as a plasticizer for F15 may also help accelerate the dissolution rate (24).

The prepared HME DQ extrudates suspended in aqueous solutions (Table 1) appeared to be homogeneous and well dispersed in the water without visible particles, precipitates, or floating materials. The nanoparticle sizes of F8 and F15 were not only comparable to those of nanoparticle formulations prepared by different methods (18, 31) but also quite stable (Fig. 3A). Except for one sample (F11), drug content in aqueous suspension for all formulations accounted for greater than 90% of the drug load (Table 1). High content of DQ in aqueous suspension suggested that a high percentage of drug had been incorporated into excipients. Nanosized particles present in the form of drug/excipient complex were a good indication of drug accessible to the intestinal surface.

DQ concentrations in the plasma of rats after F15 was orally dosed were much lower than those in the blood cells (Fig. 4A and B). The differential distributions were further supported by a parallel experiment in which the nanoparticle formulations of DQ were prepared by a different approach (non-HME) (18) and administered to rats by intravenous injection, and the concentration curve in the plasma was much lower than that of the curve in the whole blood (Fig. 4C). The amounts of DQ in the blood cells were proportionally much more than those in the plasma whether the drug was dosed by an oral route with F15 or by an intravenous injection of nanosuspension of DQ (Fig. 4D). This could be due to the nature of lipophilicity of the molecule. In addition to enhanced solubility and dissolution rate, lipophilicity may also contribute to absorption by helping penetration of drug molecules through phospholipid bilayers of enterocytes. After absorption, the drug molecules are transported in the bloodstream and interact with the blood cells. In this event, lipophilicity may also play a significant role in transcellular transport of drug molecules across the membrane of red blood cells.

It is known that the blood cells, especially the red blood cells, are the host of Plasmodium parasites. Drug molecules could act more efficiently if they could penetrate these cells. When Plasmodium parasites mature into liver-stage schizonts, they burst to release uninucleate merozoites (1). Merozoites invade and multiply in the red blood cells. Antimalarial agents need to be associated with the host cells to kill or suppress the parasites. Thus, predominate distribution of DQ in the blood cells rather than in the plasma is in favor of growth inhibition of the parasites.

In vitro and in vivo antimalarial activity of DQ HME formulations at the liver stage was evaluated by using P. berghei ANKA, a rodent species genetically engineered to express firefly luciferase. The IC₅₀ values of both F8 and F15 were below 0.5 nM, which represented highly potent antimalarial activity compared to the IC₅₀ values of primaquine and atovaquone. F15 DQ at 3 mg/kg was found to be the effective dose of complete protection of mice infected by P. berghei ANKA. Several other DQ formulations made by HME were also assessed. Twelve days after SPZ inoculation, there were two mice infected, one in F11 group and the other in F15 group (Table 5). When examined 19 days after SPZ infection, the parasitemia in these two mice disappeared. Interestingly, although these two mice failed the liver-stage protection, they survived the blood stage infection. According to the pharmacokinetic data, HME DQ absorbed by the GI largely disappeared from the plasma after 48 h, some of which presumably was diffused to tissue distribution such as the liver (17) and might have reentered the bloodstream after 48 h. If this were the case, it could explain the recovery of the two infected mice. Furthermore, the overwhelming distribution of cell associated DQ (Fig. 4) might also contribute to the inhibition and the clearance of the parasites.

Antimalarial activity of DQ HME formulations at the blood stage was assessed by using P. falciparum, a human species that causes severe malaria. HME DQ at concentration of lower nanomoles was shown to be effective at inhibiting strains of P. falciparum 3D7 (chloroquine sensitive) and P. falciparum Dd2 (multidrug resistant) (Dd2) (Table 3 and Fig. 5) and all other strains evaluated (Table 4). Other human species of Plasmodium causing malaria such as P. vivax and P. ovale could be assessed according to the method of Alison Roth (4). These experiments need to be done to establish the coverage of DQ effectiveness for inhibiting malaria at the liver stage.

In vitro toxicity experiments showed that the therapeutic index of F15 was remarkably high, and the intragastric dose of up to HME DQ 2,000 mg/kg to mice showed no signs of toxicity (see supplemental materials). This information may be valuable for the reference of further efficacy evaluation in nonhuman primate (monkey) models and clinical trials of DQ HME formulations.

HME technology offers a solvent-free, noise-free, and dust-free friendly environment and can be designed as continuous manufacturing processes with better product quality assurance and drug development efficiency (32). There have been successful examples of making oral dosage formulations for practically water-insoluble compounds (33). Postextrusion processing equipment can be adapted to manage the extruded shape, making it amenable to downstream processing into a dosage form such as tablets or capsules. The development of DQ HME formulations as an oral dosage form can be useful for the prevention of Plasmodium infections and the treatment of malaria at the liver stage.

MATERIALS AND METHODS

Decoquinate (batch number: 130802; molecular weight: 417.53) was purchased from Genebest Pharmaceutical Co., Ltd. (Zhejiang Province, China); propranolol standard was from Sigma Chemical Co. (China branch). Polyvinyl caprolactam-polyvinyl acetate-polyethylene glycol graft copolymer (Soluplus), vinylpyrrolidone-vinyl acetate copolymers (Kollidon VA 64), poloxamer 188 (Kolliphor P188), and polyoxyl 40 hydrogenated castor oil (Kolliphor RH 40) were gifts given by or purchased from BASF (Germany). Hydroxypropyl methylcellulose (AFFINISOL HPMC HME 15LV, HME 4M) and hydroxypropyl methylcellulose acetate succinate (AFFINISOL HPMCAS-716, AS-912, and AS-126) were given as gifts by Dow Pharma & Food Solutions (USA). Dimethylaminoethyl methacrylate copolymer (EUDRAGIT EPO) was purchased from EUDRAGIT (Rohm, Germany). Polyethylene glycol glyceryl laurate (Gelucire 44/14) and polyethylene glycol glyceryl stearate (Gelucire 50/13) were from Gattefosse (France). Polyethylene glycols 6000 (PEG 6000) was purchased from Sigma-Aldrich (St. Louis, MO, USA). All other chemicals used were of analytic grade.

Sprague-Dawley male rats (200–250 g) purchased from Southern Medical University, China, were utilized for pharmacokinetic study. The Animal Care and Use Committee (Guangzhou Institute of Biomedicine and Health) approved the protocol. Animals were kept for 1 week prior to experiments, maintained on standard animal chow and water ad libitum, in a condition-controlled room (23 ± 1°C, 30–70% relative humidity, a minimum of 10 exchanges of room air per hour and a 12-h light/dark cycle).

For in vivo efficacy experiments, NIH female mice, 7 weeks old, were raised at least for 7 days after arrival. The protocol was approved by the Ethical Committee for Animal Care, CAS-Lamvac. Mice were provided with food and water ad libitum during quarantine with one mouse per cage under temperature between 18°C and 26°C and a relative humidity of 34% to 68% in a 12/12h light-dark cycle.

All animal studies followed “Guides for the Care and Use of Laboratory Animals” (the Chinese State Council’s Laboratory Animal Management Regulations (Revised March 1, 2017).

Preparation of DQ formulations by hot-melt extrusion.

HME machine (Pharma 11; Thermo Fisher Scientific, Germany) was employed. The selection of formulation carriers was based on the calculation of Hildebrand solubility parameters between DQ and the carriers (19). The solubility parameter difference less than 7 MPa^1/2 between the two was considered acceptable miscibility (34, 35). Preliminary experiments were done by a simple hot-melt method to assess the homogeneity of DQ in different polymers and the appropriate ratios of drug to polymers (36).

To process HME formulation, DQ was well blended with excipients and fed into the corotating extruder. The melting temperatures were set to 120–160°C, which was well below DQ’s melting point (241.0–245.0°C). The screw speed was initially 50 rpm and then turned up to 150 rpm or adjusted to the speed according to the pressure and torque numbers. The melt is extruded through a shape-forming outlet. Upon rapid cooling at room temperature, a desirable solid single phase is produced. The extrudates (solid stick) were cut into small pieces by scissors, suspended in aqueous phase, and stirred for 24 h. The aqueous suspensions of HME extrudates were evaluated via in vitro and in vivo experiments. Particle size (diameters) and zeta-potential were measured by using a particle size distribution analyzer (Zetasizer Nano ZSE, Malvern, UK).

Preparation of DQ nanoparticle formulations.

Two-hundred milligrams of DQ (Sigma), 1,000 mg egg yolk l-α-phosphatidylcholine (PC) (Sigma), 200 mg Kolliphor HS 15 (BASF, Germany), and 1,250 mg cholesterol (Sigma) were weighed, and ethanol was added to dissolve all components. The ethanol was removed with a rotary evaporator, and the sample was transferred to a glass container to eliminate all the residual ethanol. Glucose solution (2%) was added, and the sample was ultrasonicated. The particles in the sample were reduced to size smaller than 300 nm. After freeze-drying, the powder was stored or resuspended for in vitro or in vivo experiments (31).

HPLC and liquid chromatography-mass spectrometry/mass spectrometry.

DQ was quantitated at the wavelength of 260 nm by HPLC (Agilent 1260, USA). Diamonsil C18 column (250 × 4.6 mm, 5 μm; Beijing Dikma Technologies, Inc., China) was used for DQ isolation in isocratic mobile phase containing 80% ethanol and 20% water (0.1% formic acid contained in water and ethanol, respectively).

DQ from animal samples was analyzed by liquid chromatography-mass spectrometry (LC-MS). MS2 Turnover type oscillator from IKA Works (Guangzhou, China) and a 5415R high-speed tabletop centrifuge (Eppendorf AG, Germany) were applied to treat samples. The LC-MS system includes LC-10ADvp Pump (Shimadzu, Japan), MPS3C automatic sampler (Gerstel Autosampler, Germany) and API3000 triple quadrupole tandem mass spectrometer (AB Co., USA). The chromatographic column Plus C8 (2.1 × 50 mm, 3.5 μm; Agilent ZORBAX Eclipse) was set at 30°C and the injector at 15°C. A mobile phase run at a flow rate of 600 μL/min contains methanol with 0.1% formic acid and water containing 0.1% formic acid (vol/vol, methanol: H₂O = 90:10). Water was purified using a Millipore (Little Rock, AK, USA) laboratory ultra-pure water system (0.2-μm filter).

TG, DSC, and XRD.

TG can assess the thermostability of molecules by indicating the temperature point at which the molecule starts to decompose. DSC monitors changes of physical properties of a sample, associated with temperature against time, which provides information to set up specific HME process parameters to ensure consistent product quality. XRD can be used for the study of crystal structures. When an X-ray is shined on a crystal, it diffracts in a pattern characteristic of the structure. The diffraction pattern can be obtained from a powder of the material. Crystalline material gives sharp diffraction peaks.

All the protocols followed the general rules. The detailed settings for each of these analyses were previously described (23, 31). Thermostability of pure DQ was measured by TG (Model: TG209F1; NETZSCH, Germany). Pure DQ, F15 formulation (HME), vehicle components of F15 processed by HME, and physical mixture of F15 (None-HME processed) were analyzed by DSC (204F1, NETZSCH, Germany). Additionally, DSC and X-ray diffractometer (Empyrean) were employed to analyze the large preparations of three different lots of HME formulation (F8), pure DQ, and physical mixture of F8.

In vitro dissolution test.

Small pieces of HME solid extrudates of all formulations (Table 6) were weighed, roughly 50 mg of each sample, and placed in a dissolution medium of 0.1 N hydrochloric acid (HCl) mixed with 10 mM SDS, which was added, prior to test, to the vessel of apparatus, RC-6 dissolution rate apparatus (Tianjin). PBS with pH 6.8 was also used as a dissolution medium for drug release of some formulations. Drug release was assessed according to USP, run in six replicates, and conducted at 37°C and 50 rpm of the paddle speed with vessel volume of 900 mL medium (23). One milliliter of dissolution medium was taken out at each time point according to each individual formulation, and then 1 mL blank medium was added back to the vessel. The media containing released drugs were passed through a 0.45-μm filter for DQ measurement by HPLC.

TABLE 6.

Components of DQ HME formulations^a

Formulation	Soluplus (g)	VA 64 (g)	HPMC (g)	Others (g)	DQ (g)
1	30		130 (15LV)	20 (PEG6000)	20
2	27		113 (15LV)	20 (PEG6000)	40
3	30		130 (4M)	20 (PEG6000)	20
4	27		113 (4M)	20 (PEG6000)	40
5			90 (15LV)	15 (P188), 15 (RH 40)	30
6			105 (15LV)	15 (RH 40)	30
7	28		70 (AS-912)	15 (PEG6000)	37
8	28		70 (AS-126)	15 (PEG6000)	37
9	25	105		14 (PGGS), 6 (P188)	50
10				120 (EPO)	30
11		30		120 (EPO)	50
12	30	130		20 (PEG6000)	20
13	27	113		20 (PEG6000)	40
14	74	37		10 (PGGS)	30
15	87	43		20 (PEG6000)	50

^aSoluplus, polyvinyl caprolactam-polyvinyl acetatepolyethylene glycol graft copolymer; VA 64, vinylpyrrolidone/vinylacetate copolymer; HPMC, hydroxypropyl methyl cellulose; 15LV, HPMC HME 15LV; 4M, HPMC HME 4M; AS-912, HPMCAS-912; AS-126, HPMCAS-126; PEG 6000, polyethylene glycol 6000; P188, Poloxam 188 (Kolliphor P188); RH40, polyoxyl 40 hydrogenated castor oil; PGGS, polyethylene glycol glyceryl stearate (Gelucire 50/13); EPO, dimethylaminoethyl methacrylate copolymer (EUDRAGIT EPO).

Pharmacokinetics of hot-melt extruded DQ.

(i) Dosing and sample collection. DQ HME formulation suspended in sterile saline solution was administered at a dose of DQ 20 mg/kg (1.00–1.25 mL) to rats (n = 4) by intragastric gavage. At different time points, blood samples (200 μL) were collected via the tail vein and placed in tubes containing heparin, which were immediately centrifuged at 3,329 × g, 6 min, 4°C. Plasma fraction (upper phase) and blood cells (bottom portion) were separately collected and stored at −20°C until LC-MS analysis. In a separate assessment, DQ nanoparticle formulation was given to rats (n = 3) by intravenous injection at a drug dose of 10 mg/kg (100–125 μL).

(ii) Standard curve and sample preparation.

DQ was dissolved in ethanol at a concentration of 50 μg/mL and diluted to a series of gradient concentrations. Each standard solution 10 μL was mixed with 50 μL blank plasma by vortex for 3 min (Scilogex, USA). A protein precipitation solution (150 μL) was then added to each sample and was mixed by vortex for 5 min. A mixture of ethanol: acetonitrile (1:1) was used protein precipitation solution containing propranolol as internal standard diluted at 1 μg/mL which. The samples were then centrifuged at 16,000 × g for 60 min at 4°C. The supernatant (100 μL) was transferred to a new container, and DQ was measured by LC-MS/MS. Final standard concentrations were 5, 10, 20, 50, 100, 200, 500, and 1,000 ng/mL. For DQ analysis of sample plasma (50 μL), ethanol (10 μL) was added and processed as for DQ standard curve preparation.

(iii) Blood cell sample preparation.

Blood cells (50 mg/test) from plain animals were mixed with 10 μL of a series of each DQ standard solution for a standard curve preparation. After mixing well, a 250-μL precipitation solution containing internal standards was added to the mixture and homogenized for 5 min in the grinder machine (JXFSTPRP-32 automatic sample rapid grinder). The rest of procedures were the same as above for plasma.

In vitro antimalaria activity of the DQ HME formulations.

The in vitro assay for testing antimalarial activity at the blood stage was performed as described previously (37). A series of diluted HME formulation DQ (10 μL) were cocultured with human red blood cells (hRBCs) (90 μL) having 0.5% parasitemia (P. falciparum) and 2% hematocrit under supply of 5% CO₂ at 37°C for 72 h in 96-well plate. After incubation, lysis buffer containing SYBR green I dye was added to each well and DNA of the parasites in late-ring or early-trophozoite stages was measured as previously described hRBCs for in vitro test were obtained from normal subjects and patients with informed consent. Half-maximal inhibitory concentration (IC₅₀) values were obtained from results analyzed by Prism software.

To evaluate hepatic stage antimalarial activity of DQ HME formulations, SPZ of P. berghei parasites expressing firefly luciferase were isolated from salivary glands of mosquitoes fed by the blood of infected mice. HepG2 cells (10⁴/well) in 96-well plate were cultured overnight and then the SPZ (1.5 × 10⁴) added to each well. The plate was sealed and centrifuged at 200 × g for 8 min to allow the SPZ to settle. After incubated for 4 h, the culture medium was removed, and the cells washed three times to remove the nonsettled SPZ. The concentrations of HME formulation DQ (F8 and F15) were determined by HPLC. The concentrations of HME formulation DQ (F8 and F15) were determined by HPLC. DMEM complete medium (100 μL) containing each diluted HME DQ was added to each well and cultured for 48 h. Luminescent intensity was determined by inhibition of liver-stage development assay to reflect intracellular proliferation of luciferase-expressing P. berghei parasites inside HepG2 liver cells (18, 38).

In vivo antimalarial efficacy of the DQ HME formulations.

Each mouse was inoculated with 5 × 10⁴ SPZ of P. berghei ANKA (200 μL) by intravenous injection through the tail vein at day zero. The DQ HME formulations were given in three consecutive days (-1, 0, 1) in the saline suspension (150 μL) to animals by intragastric gavage. F15 were dosed at DQ 1.0, 3.0, and 10.0 mg/kg. In screening experiments (Table 5), HME formulations were given at 5.0 mg/kg DQ. Atovaquone at 10 mg/kg and primaquine at 30 mg/kg as positive controls and vehicles of F15 were used as negative controls. IVIS was performed 24, 48, and 72 h after the injection of P. berghei SPZ to monitor the Plasmodium infection at the liver stage as described previously (18, 39). Detection for the blood stage infection began after 2 days of SPZ injection and was continuously monitored by examining thin film under microscope with oil immersion (40). Parasitemia represented the number of infected red blood cells per 1,000 red blood cells as erythrocyte-infected rate (‰). Animal survival rates were summarized on day 60 postinfection.