Impact of Formulation Parameters on In Vitro Release from Long-Acting Injectable Suspensions

Abstract

The development of long-acting injectable (LAI) suspension products has increased in recent years. A better understanding of the relationship between the physicochemical properties of these products and their in vitro as well as in vivo performance is expected to further facilitate their development and regulatory review. Using Depo-SubQ Provera 104® as the reference listed drug (RLD), four qualitatively and quantitatively (Q1/Q2) equivalent LAI suspensions with different formulation properties were prepared. Two recrystallization methods (solvent evaporation and anti-solvent) were utilized to obtain active pharmaceutical ingredient (API) with different properties and solid-state characterization was performed. In addition, two different sources of the major excipient were used to prepare the Q1/Q2 equivalent suspensions. Physiochemical characterization and in vitro release testing of the prepared Q1/Q2 equivalent suspension formulations and the RLD were conducted. In vitro drug release was dependent not only on the particle size, the morphology, and the crystallinity of the API, but also on the residual solvent in the API. The excipient source also affected the drug release rates.

1. Introduction

Long-acting injectable (LAI) formulations are intended to be injected intramuscularly or subcutaneously for extended drug release over a period ranging from weeks to several months. These products reduce dosing frequency and hence improve patient compliance. LAI formulations are commonly designed for the treatment of, but are not limited to, schizophrenia [1], opioid use disorder [2], human immunodeficiency virus (HIV) [3, 4] and contraception[5]. Based on the physical state prior to use, the formulations can be categorized as: 1) suspensions of crystalline drug suspended in appropriate media in prefilled syringes (e.g., Depo-Provera®); 2) powders (lyophilized pure drug or solid preparation) in vials or prefilled syringes, which are intended to be mixed with the required media to form suspensions (e.g., Vivitrol® and Perseris Kit®) or solutions (e.g., Eligard®) prior to use; and 3) solution preparations (e.g., Sublocade®). Additionally, there are different formulation technologies used for LAIs, including: recrystallization (to form crystalline drug suspensions); microencapsulation (microparticles or microspheres); and in situ forming implants (such as ATRIGEL®) [6]. For these LAIs, extended drug release is achieved through either poor drug solubility (e.g., suspensions of poorly soluble drugs in the crystalline state), or the use of a rate controlling polymer (e.g., poly(lactic-co-glycolic acid) (PLGA)) for microspheres as well as in situ forming implants. Currently marketed LAI drug products approved by the U.S. FDA- are listed in Table 1. Approved products that have been discontinued by the manufacturer are not included.

Table 1.

Long acting injectable products approved by the U.S. FDA^* [7].

Active Ingredient	Proprietary Name	Route	Applicant Holder	Treatment	Approval Date	Strength/Efficacy duration
Aqueous Crystalline Drug Suspensions
Aripiprazole	Abilify Maintena Kit	IM	Otsuka	Schizophrenia	2013	300 mg/vial, 1 month 400 mg/vial, 1 month
					2014	300 mg, 1 month 400 mg, 1 month
Aripiprazole Lauroxil	Aristada	IM	Alkermes	Schizophrenia	2015	441 mg/1.6 ml, 1 month 662 mg/2.4 ml, 1 month 882 mg/3.2 ml, 6 weeks or 1 month
					2017	1064 mg/3.9 ml, 2 months
	Aristada Initio Kit	IM	Alkermes	Schizophrenia	2018	675 mg/2.4 ml, 1 month
Medroxyprog esterone Acetate	Depo-Provera	IM	Pfizer	Contraception	Prior To 1982	400 mg/ml, 3 months
					1992	150 mg/ml, 3 months
	Depo-SubQ Provera 104	SC	Pfizer	Contraception	2004	104 mg/ml, 3 months
Methylprednis olone Acetate	Depo-Medrol	IM, IA, etc.	Pfizer	Inflammation, etc.	Prior to 1982	20 mg/ml, 40 mg/ml and 80 mg/ml, variable duration:1 week to 4 weeks
Olanzapine Pamoate	Zyprexa Relprevv	IM	Eli Lilly	Schizophrenia	2009	210 mg/vial, 2 weeks 300 mg/vial, 2 weeks 405 mg/vial, 4 weeks
Paliperidone Palmitate	Invega Trinza	IM	Janssen	Schizophrenia	2015	273 mg/0.875 ml, 410 mg/1.315 ml, 546 mg/1.75 ml and 819 mg/2.625 ml: 3 months.
	Invega Sustenna	IM	Janssen	Schizophrenia	2009	39 mg/0.25 ml, 78 mg/0.5 ml, 117 mg/0.75 ml, 156 mg/ml, and 234 mg/1.5 ml: for monthly maintenance based on initial treatment regimen.
Microspheres
Exenatide Synthetic	Bydureon	SC	AstraZeneca	Type 2 diabetes	2012	2 mg/vial, 1 week
	Bydureon Pen	SC	AstraZeneca	Type 2 diabetes	2014	2 mg, 1 week
	Bydureon Bsice	SC	AstraZeneca	Type 2 diabetes	2017	2 mg/0.85 mL, 1 week
Leuprolide Acetate	Lupron Depot	IM	AbbVie	Prostatic cancer	1995	3.75 mg, 1 month
					1995	22.5 mg/vial, 3 months
					1997	30 mg/vial, 4 months
					2011	45 mg/vial, 6 months
	Lupron Depot-Ped	IM	AbbVie	Central precocious puberty	1993	7.5 mg/vial, 1 month
					1994	11.25 mg/vial;15 mg/vial/: 1 month
					2011	11.25 mg/vial; 30 mg/vial: 3 months
Minocycline HCl	Arestin	Peridontal	OraPharma	Periodontitis	2001	1 mg Base, variable
Naltrexone	Vivitrol	IM	Alkermes	Alcohol dependence	2006	380 mg/vial, 1 month
Octreotide Acetate	Sandostatin LAR	SC	Novartis	Acromegaly	1998	10 mg base/vial; 20 mg base/vial; 30 mg base/vial: four weeks
Pasireotide Pamoate	Signifor LAR	IM	Novartis	Acromegaly; Cushing’s disease	2014	20 mg base/vial; 40 mg base/vial; 60 mg base/vial: 4 weeks
					2018	10 mg base/vial; 30 mg base/vial: 4 weeks
Risperidone	Risperdal Consta	IM	Janssen	Schizophrenia	2018	10 mg base/vial; 30 mg base/vial: 4 weeks
					2007	12.5 mg/vial, 2 weeks
Triamcinolone acetonide	Zilretta	IM	Flexion	Osteoarthritis pain	2017	32 mg/vial; 3 months
Triptorelin Pamoate	Trelstar	IM	Allergan	Prostate cancer	2000	3.75 mg base/vial, 4 weeks
					2001	11.25 mg base/vial, 12 weeks
					2010	22.5 mg base/vial, 24 weeks
	Triptodur Kit	IM	Arbor	Central precocious puberty	2017	22.5 mg/base/vial, 24 weeks
In-Situ Forming Implants
Buprenorphine	Sublocade	SC	Indivior	Opioid disorder	2017	100 mg/0.5 ml; 300 mg/1.5 ml: 1 month
Degarelix acetate	Firmagon	SC	Ferring	Prostate cancer	2008	80 mg base/vial; 120 mg base/vial: 28 days
Doxycycline Hyclate	Atridox	Periodontal	Tolmar	Subgingival application	1998	50 mg
Lanreotide Acetate	Somatuline Depot	SC	Ipsen	Acromegaly; Neuroendocri ne Tumors	2007	60 mg base /0.2 ml, 4 weeks 90 mg base /0.3 ml, 4 weeks 120mg base/0.5 ml, 4 weeks
Leuprolide Acetate	Eligard	SC	Tolmar	Prostate cancer	2002	7.5 mg/vial, 1 month; 22.5 mg mg/vial, 3 months
					2003	30 mg/vial, 4 months
					2004	45 mg/vial, 6 months
Risperidone	Perseris Kit	SC	Indivior	Schizophrenia	2018	90 mg; 120 mg: 1 month
Implants
Goserelin Acetate	Zoladex	SC	Tersera	Prostatic cancer	1989	3.6 mg Base/28 days
					1996	10.8 mg Base/12 weeks

^*Discontinued products are not listed; oil-based products are not listed.

The focus of the current research is on LAI suspensions described in the first category of Table 1: crystalline drug suspensions in aqueous media. Among these LAI suspensions, six products are used for the treatment of schizophrenia and the other two are contraceptive products. Four of these LAI products were approved in the past three years. Additionally, a new drug application for a once monthly regimen containing two (cabotegravir [8] and rilpivirine [9]) LAI suspensions to be used in the treatment of HIV was recently submitted to the U.S. FDA [10]. This medication was approved this year by Health Canada under the brand name Cabenuva™ [11].

LAI suspensions achieve extended release through the formation of drug depots from which poorly soluble drugs gradually undergo dissolution followed by absorption into the systemic circulation. The drug particles are in the form of micro or nano sized crystals dispersed in the injection media. The particle size of the drug crystals is critical to the in vitro and in vivo performance. According to the Noyes-Whitney equation [12], the specific surface area of the drug particles is inversely proportionally to the dissolution rate. There are two LAI paliperidone palmitate nano-suspension products: 1) Invega Sustenna® (approved in 2009); and 2) Trinza® (approved in 2015). Both LAI suspensions are manufactured using Nano-crystal® technology. The dosing frequency of Invega Sustenna® and Trinza® are monthly and every three months, respectively [13]. The increase in the duration of efficacy of the three-month formulation is considered to be a result of increased particle size compared to the one-month formulation. There have been several reports on the impact of different aspects (including particle size, stabilizers) on formulation performance for LAI nanosuspensions [14, 15]. However, there have been few reports on the effect of formulation parameters on LAI micro size drug suspensions. In addition, source variation of major excipients has been shown to have significant impact on drug release from tablets [16], semisolids [17], intrauterine systems [18] and PLGA microspheres [19]. However, it is not clear if source variation of major excipients (such as the suspending agent) will have any impact on the performance of suspension formulations.

To better understand the critical parameters influencing the performance of LAI suspensions, a reliable in vitro release testing method is necessary. Although there are some U.S. FDA recommendations to use USP apparatus 2 (e.g., Aripiprazole for intramuscular suspension, extended release) and USP apparatus 4 (e.g., Depo-Provera) [20] to perform in vitro release testing, there is currently no in vitro release guidance available to the pharmaceutical industry for Depo-SubQ Provera 104® or other complex parenterals.

The earliest approved LAI suspensions are the two trimonthly contraceptive medroxyprogesterone acetate (MPA) products: 1) Depo Provera® (intramuscular injection with two strengths: 400 mg/ml and 150 mg/ml) approved prior to 1982 and in 1992, respectively; and 2) Depo-SubQ Provera 104® (subcutaneous injection with a strength of 104 mg/0.65 ml) approved in 2004. Depo-SubQ Provera 104® came off patent in 2020 [21] and it is anticipated that generic versions of this product may be developed in the near future. Therefore, Depo-SubQ Provera 104® was selected as the reference listed drug (RLD) for this study and formulations were prepared to be qualitatively (Q1) and quantitatively (Q2) equivalent to each other. Medroxyprogesterone acetate is a derivative of progesterone with poor water solubility and a LogP of 4.1 [22]. In the RLD, the crystalline MPA particles are suspended in an appropriate media and the formulation composition is listed in the prescribing information of the RLD [23].

In the current research, formulation parameters of LAI suspensions such as particle size and excipient sources of the major suspending agent (polyethylene glycol (PEG)) were investigated. Crystalline MPA with different particle size was obtained through recrystallization (both antisolvent and solvent evaporation methods). Formulations that were Q1 and Q2 equivalent to each other but with manufacturing differences were prepared. An in vitro release testing method with discriminatory capability was developed for these LAI suspensions. Physiochemical characterization and in vitro release testing of the Q1/Q2 equivalent formulations was performed. The impact of particle size and excipient source on the in vitro release of the LAI suspensions was investigated.

2. Materials and methods

2.1. Materials

Medroxyprogesterone acetate (Micronized, USP grade), polyethylene glycol (PEG) 3350 (NF-USP), Tween 80 (NF grade), methionine (USP grade), methylparaben (NF grade), and propylparaben (NF grade) were purchased from Spectrum Chemical Manufacturing Corp. (New Brunswick, NJ, USA). Sodium phosphate monobasic monohydrate and sodium phosphate dibasic dodecahydrate were purchased from Millipore-Sigma (Burlington, MA, USA). PVP K17 PF and PEG 3350 (Kollisolv®) were gifts from BASF (Ludwigshafen, Germany). Sodium chloride and sodium dodecyl sulfate (SDS) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Unless otherwise specified, all materials were of analytical grade.

2.2. HPLC analysis of MPA

The concentration of MPA was determined using an Agilent 1290 HPLC system with a UV detector set at 244 nm. The mobile phase was a mixture of acetonitrile and water (65/35, v/v). A C18 column (Kinetex®, 250 × 4.6 mm, 5 μm, Phenomenex®) was used with a flow rate at 1.5 mL/min. The column temperature was set at 30°C and the injection volume was 50 μL. The chromatographs were analyzed using Agilent OpenLAB CDS ChemStation.

2.3. Drug solubility

2.3.1. Aqueous solubility of MPA

The drug solubility in different aqueous media was tested. In brief, approximately 1 mg samples of drug were weighed into 1 mL glass vials and 500 μL of media (refer to Table 3) were added to achieve the target concentration of 2 mg/mL. The samples were placed in a water shaker bath at 37°C and rotated at 100 rpm overnight (18~24 hr). The samples were filtered via a Duropore™ 96-well filter plate with a hydrophilic PVDF membrane and a pore size of 0.45 μm. The samples were diluted (if necessary) and injected into a HPLC for analysis. The drug solubility of all prepared MPA suspensions as well as the RLD was also determined. All the experiments were performed in triplicate.

Table 3.

Drug solubility in different aqueous media (n=3, mean ± SD).

No.	Different pH values	Solubility (MQ/mL)
1	50 mM PBS pH 3.0	0.23±0.01
2	50 mM Acetate Buffer pH 4.0	0.32±0.04
3	50 mM PBS pH 6.0	0.23±0.01
4	50 mM PBS pH 7.4	0.27±0.04
5	50 mM Tris pH 9.0	0.35±0.07
6	Milli-Q Water	0.39±0.08
Different concentrations of SDS
7	0.1% SDS in pH 7.4 PBS	41.73±0.72
8	0.25% SDS in pH 7.4 PBS	102.22±4.92
9	0.5% SDS in pH 7.4 PBS	264.08±20.55
10	1% SDS in pH 7.4 PBS	545.98±4.12
Same concentration of excipient as in the RLD
11	0.3% Tween 80 in pH 7.4 PBS	6.34±0.45
12	2.88% PEG03350 in pH 7.4 PBS	0.63±0.01
13	0.5% PVP K-17PF in pH 7.4 PBS	0.33±0.02
In suspension formulations
14	RLD	10.01±0.26
15	F-A	7.19±0.11
16	F-B	10.79±0.04
17	F-C	7.53±0.03
19	F-D	10.04±0.04

2.3.2. Solubility of MPA in different organic solvents

To obtain different particle size through the recrystallization process, drug solubility in several commonly used organic solvents and in hydro-organic mixtures (refer to Table 4) was investigated at room temperature. In brief, approximately 10 mg of MPA was weighed into 2 mL HPLC glass vials and stirred with a magnetic stirrer. 100 μL aliquots of the organic solvent were added to the vials in a stepwise fashion until all the drug was dissolved based on visual observation. Once the drug was dissolved no more solvent was added. If the drug was not fully dissolved in 1 mL of the organic solvent, then this solvent was abandoned as a potential suitable solvent and the test was ended. Following the solubility test, solid drug obtained from either solution samples via solvent evaporation, or from suspension samples via centrifuge, was dried at 40°C under vacuum. The dried solid drug was further characterized using powder X-ray diffraction (PXRD) and differential scanning calorimetry (DSC) to examine any physical form change following solubility testing.

Table 4.

Observed solubility of MPA in different organic solvents.

Solvent	Solubility at room temperature mg/mL	Solubility at 50°C mg/mL
Methanol	< 18.09	> 18.09
Ethanol	< 20	> 20
Isopropanol	< 18.24	> 18.24
Acetone	> 25.03	--*
Acetonitrile	< 25.05	> 25.05
Ethyl acetate	< 24.15	> 24.15
Tetrahydrofuran	> 34.97	--*
Methanol/H2O (1/1)	< 10**	< 10
Ethanol/H2O (1/1)	< 10**	< 10
Acetone/H2O (1/1)	< 10**	< 10

^*Since the solubility at room temperature is sufficiently high, solubility at higher temperature (50°C) was not tested.

2.4. Recrystallization of MPA

To obtain different particle size of MPA, recrystallization of MPA was conducted using two different methods based on the solubility studies in different organic solvents.

Antisolvent method

MPA had good solubility in acetone and poor solubility in water (Table 2). Accordingly, acetone was used as the good solvent and water was used as the anti-solvent to perform recrystallization of MPA. In brief, 1.5 g of MPA was added into a bottle with a magnetic stirrer followed by addition of 35 mL of acetone. Upon the complete dissolution of MPA in acetone, 35 mL of water was pumped into the solution at 1 mL/min, while stirring at 300 rpm. The recrystallized MPA solid was collected by removing the supernatant liquid. The residual solvent was removed by drying at 40°C under vacuum. The dry solid was then passed through a 45 μm sieve and stored for further use.

Table 2.

The prepared compositionally equivalent MPA suspensions

Formulation	API used	Source of PEG3350
F-A	As received	Spectrum (media 1)
F-B	Recrystallized using antisolvent method (acetone:water=1:1)	Spectrum (media 1)
F-C	As received	BASF (media 2)
F-D	Recrystallized using solvent evaporation method (THF)	Spectrum (media 1)

Solvent evaporation method

MPA has the highest solubility in THF among all the solvents screened. 2.5 g of MPA was added into a bottle with a stirrer and 20 mL of THF was added to completely dissolve the drug. The drug solution was then left for a week to evaporate at room temperature. The residual organic solvent was further removed by vacuum drying at 40°C. The dry solid was collected and stored for further use.

2.5. Preparation of the Q1/Q2 equivalent MPA suspensions

For the preparation of Q1/Q2 equivalent MPA suspensions, the formulation space is limited to the drug particle size, as well as the excipient sources of the major polymer in the formulations since the formulation composition and ratio are fixed. Q1/Q2 equivalent MPA suspensions were prepared based on the RLD label formula [23]. Suspending media containing all of the required amounts of excipients were prepared. To investigate the impact of polymer source of the major suspending agent PEG3350, two suspending media were prepared with the only difference being the PEG source. Media 1 and media 2 were prepared using PEG from Spectrum Chemical and BASF (Kollisolv®), respectively. Formulations A, B and D were prepared using suspending media 1 with the MPA as received, recrystallized MPA from acetone-water antisolvent method, and recrystallized MPA from THF evaporation method, respectively. Formulation C was prepared using the suspending media 2 with the MPA as received. Formulations A, B, C, and D were designated as F-A, F-B, F-C, and F-D, respectively. The MPA suspensions were prepared via a simple mixing method. In brief, the MPA powder was added into the suspending media and then mixed using a magnetic stirring bar at 600 rpm until the formation of uniform suspension (i.e., 1 hour F-A, B and C; 18 hours for F-D).

2.6. Drug content and uniformity of the prepared MPA suspensions

Drug content and uniformity is crucial to the product quality of suspensions. 10 μL of the formulations was diluted in 60% acetonitrile to fully dissolve. The samples were injected into a UPLC for analysis. The experiments were performed in triplicate. The relative standard deviation (RSD) of the drug content was used to indicate the uniformity of the formulations.

2.7. Powder X-Ray Diffraction (PXRD)

The drug powder from different samples were characterized using an X-ray diffractometer (D2 phaser, Bruker AXS Inc., Madison, WI, USA) with Cu-kα radiation. The voltage was set at 30 kV and the current was 10 mA. All scans were performed from 5° to 45° at 2° ranges with a step increment of 0.02°/step and 0.35 sec/step.

2.8. Thermal analysis

To examine the physical form of the solid drug, DSC was performed using a TA Q1000 calorimeter (TA instruments, New Castle, DE, USA) equipped with a refrigerated cooling accessory. The samples were equilibrated at 30 °C and then heated up at 10 °C/min to 300 °C. Nitrogen gas was used for purging at a flow rate of 50 mL/min. To investigate the residual solvents in the API, thermogravimetric analysis was performed on a TA Q500 (TA instruments, New Castle, DE, USA). The samples were loaded to the platinum pan, equilibrated at 30°C and heated to 300 °C at a ramp rate of 10 °C/min. All collected data were analyzed using TA universal analysis software.

2.9. Particle size and size distribution

The particle size and size distribution of the prepared MPA suspensions and RLD were determined using an AccuSizer 780 AD (NICOMP, Santa Barbara, CA, USA). In brief, approximately 10 μL of 16 ug/mL MPA suspension was injected into the particle sizing system. Dv10, Dv50 and Dv90, as well as the Span values ((Dv90-Dv10)/Dv50) of the samples were analyzed. All the measurements were performed in triplicate. For comparison, the particle size and size distribution of the Q1/Q2 equivalent MPA suspensions were also analyzed using an Olympus BX51 polarized light microscopy (PLM) (Olympus America Inc., New York, USA) with an ImageJ software (NIH, Bethesda, MD, USA). The samples were spread on a glass slide and dispersed with one drop of mineral oil. Cover slips were placed on top of the dispersed samples. The microscopy images were acquired at 200 magnification using a SPOT Imaging software.

2.10. Morphology of the prepared suspensions

The morphology of the RLD and the prepared MPA suspensions were characterized using a scanning electron microscope, a FEI Nova NanoSEM 450 unit (FEI Company, Hillsboro, Oregon, USA). Samples were mounted on carbon taped aluminum stubs and sputter coated with gold for 3 min at 6 mA with a deposition rate of 6 nm/min before imaging.

2.11. In vitro release testing

Enhancer cells (surface area: 4 cm², Agilent Technologies, USA) were used with USP apparatus 2 equipped with 200 ml flat bottom dissolution vessels to perform the in vitro release testing of the MPA suspensions. 50 μL of the MPA suspension formulations were loaded into the enhancer cells (depth: 0.4 mm). A Whatman™ GF/D glass microfiber filter membrane (2.7 μm pore size and 2.5 cm in diameter) was pretreated in water and the surface water was removed using a Kimwipe. The membrane was then placed on top of the sample and the cells was assembled per the manufacturer’s manual. 150 mL of pH 7.4 phosphate buffered saline (PBS) with 1% w/v of SDS was used as the release media. The agitation speed was set at 120 rpm and the temperature at 37°C. At pre-determined time intervals, 1 mL samples of the release medium were withdrawn and replenished with fresh media.

2.12. Statistical analysis

The data analysis was performed using OriginPro 2017 software (OriginLab Corporation Wellesley Hills, MA, USA). The data are presented as average ± the standard deviation (SD).

3. Results and Discussion

3.1. Solubility of MPA

Aqueous solubility

The solubility of MPA in both aqueous and organic solvents was investigated to facilitate formulation design. As shown in Table 3, the aqueous solubility of MPA at different pH values is low (less than 0.4 μg/mL) and is pH independent. The influence of Tween 80 as well as the polymers (PEG3350 and PVP K-17PF) on the drug solubility was investigated at the same concentrations as they are present in the RLD formulation. There was no solubilizing effect of the PVP (0.5% w/v) on the drug. PEG3350 was able to increase the MPA solubility by ~2-fold. MPA solubility increased more than 20-fold in the presence of 0.3% w/v Tween 80. The drug solubility in the RLD and the prepared formulations was approximately10 μg/mL, and this can be largely attributed to the presence of Tween 80. Drug solubility in different concentrations (0.1% w/v to 1% w/v) of SDS in PBS buffer was investigated to help determine appropriate in vitro release testing conditions.

Solubility in organic solvents

To screen for good solvents for MPA recrystallization, solubility in commonly used organic solvents was investigated at both room temperature and 50°C. THF followed by acetone were the best solubilizers of MPA (over 25 mg/mL at room temperature, Table 4), and accordingly were selected for the recrystallization process. Both acetone and THF were used to recrystallize the MPA via the antisolvent method with water as the antisolvent. Reproducible particle size with high yield was achieved when using the acetone-water (1:1, v:v) solvent/antisolvent system. On the contrary, the yield was much lower when using the THF-water (1:1, v:v) solvent/antisolvent system due to the extremely high solubility of MPA in THF. Accordingly, the solvent evaporation method using pure THF was also utilized for the API recrystallization.

3.2. Characterization of recrystallized MPA

PLM, PXRD and thermal analysis

The MPA as received and the recrystallized MPA solid were visualized using PLM (Fig. 1). As observed in Fig. 1, the recrystallization process resulted in crystalline solid but with different particle size depending on the processing solvents (antisolvent and evaporation). No amorphous solids were observed. PXRD confirmed that the MPA was in the crystalline state following the recrystallization process, (Fig.2) However, the samples prepared via the THF solvent evaporation method had much higher crystalline intensity. In addition, DSC and TGA thermal analysis (Fig.3) showed that the recrystallized MPA had similar melting points. The raw API, and the recrystallized API using the antisolvent and solvent evaporation methods had residual solvent concentrations of 0.20%, 0.17% and 0.17% w/w, respectively.

Fig. 1.

Representative PLM images of: A) API as received; B) API recrystallized using the antisolvent method (acetone:water=1:1, v:v); and C) API recrystallized using the solvent evaporation method (THF).

Fig. 2.

PXRD profiles of MPA powder following different processes.

Fig. 3.

Thermal analysis including DSC and TGA profiles of recrystallized MPA powder using: A) API recrystallized using the antisolvent method (acetone:water=1:1, v:v); and B) API recrystallized using the solvent evaporation method (THF).

3.3. Characterization of MPA suspensions

Content uniformity and pH

The drug content for all of the formulations achieved the target of 160 mg/mL with good drug content uniformity (RSD<5%). All of the prepared MPA suspensions had a pH of 6.4, similar to the RLD.

Drug solubility in the MPA suspensions

As observed in Table 3 above, PEG3350 and Tween 80 can increase MPA solubility. Therefore, the drug solubility in the final formulations was determined. The solubility of MPA in the final formulations increased more than 20-fold, mainly because of the Tween 80 present in the formulations (refer to Table 3). The drug solubility in the RLD, and in formulations F-A, F-B, F-C and F-D was 10.01±0.26, 7.19±0.11, 10.79±0.04, 7.53±0.03 and 10.04±0.04 μg/mL, respectively. The drug had comparable solubility in the RLD, and in formulations F-B and F-D, which was higher than in formulations F-A and F-C. These solubility differences may be due to differences in residual solvent in the raw API and the recrystallized drug as well as in their crystal habits. From the TGA results, the API as received (used in formulations F-A and F-C), recrystallized in acetone/water antisolvent system (used in F-B) and recrystallized in THF (used in F-D) have minimal residual solvent (lower than 0.20% w/w). The solvent used in the raw API as received was unclear. In the API recrystallization process, acetone and THF were used, which are both good solvents for MPA. Although, the amount of residual solvent was approximately 0.16 to 0.17 % w/w this small amount of solvent may contribute to an increase in drug solubility. In addition, the drug solubility differences in the various formulations could also result from different morphologies of the drug in the final formulation.

Particle size and size distribution

Drug particle size and crystallinity in the final formulation may change during the manufacturing process. The drug particles in the prepared MPA suspensions and in the RLD were characterized using PLM, an AccuSizer as well as SEM. Compared to the PLM images in Figs. 1 and 4, the drug particle size in all formulations was reduced following preparation. The size reduction may be due to solubilization in the suspending media as well as shearing during the preparation process. The PLM images showed that the drug suspensions remained in the crystalline state (Fig.4).

Fig. 4.

Representative PLM images of the RLD as well as the prepared Q1/Q2 equivalent MPA suspensions (200-fold magnification).

The particle size obtained using PLM (number weighted) is consistent with the particle size determined using the AccuSizer (volume weighted) (Fig.5). Formulation F-B and the RLD had the largest mean particle sizes (~20 μm). Formulations F-A, F-C and F-D showed similar particle size and the smallest mean particle size (~15 μm). The RLD, and formulations F-A and F-C had lower span values (1.04, 0.94, and 1.01, respectively) when compared to formulation F-D (1.24) and formulation F-B (1.38). Formulations F-B and F-D have broader particle size distributions, whereas the other formulations were relatively more monodisperse (Fig.4). The commercial raw API as received has a relatively narrow particle size distribution. However, during the recrystallization process in either THF or the acetone/water mixture, the drug crystals may grow resulting in wider particle size distributions. For example, the Dv10 value of F-B is similar to the other prepared formulations, whereas the Dv90 value of F-B is much higher than the other formulations.

Fig. 5.

The particle size and size distribution of the RLD and the prepared Q1/Q2 equivalent MPA suspensions using: A) PLM; and B) AccuSizer. All data are presented as mean ± SD (n=3).

Morphology

The morphology of the RLD and the prepared MPA suspensions was investigated using SEM. The RLD and formulation F-B had larger particles but with different shape. The particles in the RLD were of round shape whereas those in formulation F-B were plate-like and elongated. Formulations F-A and F-C were prepared using the same raw API and therefore the drug particles in both formulations were similar - irregular shape with smooth edges. The drug particles in F-D had a similar shape to those in formulation F-B, elongated plate-like shape with sharp edges. This could be due to the processing methods used for the API treatment. The more spherical morphology of the API used in the RLD and in formulations F-A and F-C may be due to applied micronization or milling processes. The API was used as received in formulations F-A and F-C. Whereas the API recrystallized from acetone/water (used in F-B) and THF (used in F-D) did not undergo further treatment, resulting in crystals with sharper edges.

3.4. In vitro drug release testing of the MPA suspensions

It is challenging to develop an appropriate in vitro release testing method since there has been no guidance to date on such LAI suspensions. The in vitro release testing of the prepared formulations and the RLD was performed using USP apparatus 2 with enhancer cells. The developed in vitro release testing method was able to differentiate the Q1/Q2 equivalent MPA suspensions and the RLD (Fig.7). The RLD and formulations F-C and F-D showed the fastest release rates among all the formulations, and they were not statistically different. Formulation F-B had the slowest release rate, which was considered to be due to the larger drug particle size in this formulation compared to any of the other formulations. Despite their similar drug particle size (~15 μm), the release rates of formulations F-A, F-C and F-D varied: with formulations F-C and F-D showing similar release rates, which were much faster than formulation F-A. The only difference between formulations F-A and F-C is that the major excipient, PEG3350, used in these two formulations was from two different sources. Previous studies [17, 18, 24] have shown that the source of major excipients in a formulation may result in significant differences in their in vitro and/or ex vivo performance. The only difference between formulations F-A and F-D was the API preparation used in these formulations. The API used in formulation F-A was used as received, whereas the API used in formulation F-D was recrystallized using the solvent (THF) evaporation method. Despite the similar particle size of formulations F-A and F-D, formulation F-D showed a faster release rate compared to formulation F-A. This may be due to the different residual solvents in the API, even though the amounts were minimal. The type of residual solvent in the API as received (for F-A) was not clear. Based on the certificate of analysis (COA) of the raw API, several residual solvents are listed: N,N-dimethylformamide, methanol, methylene chloride, THF, pyridine, acetic acid and ethanol. All of the residual solvents passed the test for the residual solvent limit required by the US Pharmacopeia and the amounts are unknown in the COA. Therefore, it is unclear which solvent(s) in the API as received. However, for the API recrystallized from THF (used in F-D), residual THF may be responsible for the higher release rate since the drug solubility in THF was the highest (Table 2). The API recrystallized using THF (F-D) showed the highest crystallinity, which would normally be expected to result in slower drug release. However, this formulation showed one of the fastest drug release rates suggesting the dominance of the THF residual solvent effect over the drug crystallinity.

Fig. 7.

The release profiles of the RLD and the prepared Q1/Q2 equivalent MPA suspensions. All data are presented as mean ± SD (n=3).

Table 5 below, summarized the physicochemical properties and release rates of the final suspension formulations.

Table 5.

Summary of the physicochemical properties of the raw API and the final suspensions.

Formulations	API size (μm)	API crystallinity	Residual(w/w)	Solubility (μg/mL)	Morphology	PEG3350 source*	Release rate
F-A	15	weak	0.20%, unclear**	7.19±0.11	Irregular, round edge	1	Medium
F-B	20	weak	0.16%, Acetone	10.79±0.04	Plate like, sharp edge	1	Slowest
F-C	15	weak	0.20%, unclear**	7.53±0.03	Irregular, round edge	2	Fastest
F-D	15	strong	0.16%, THF	10.04±0.04	Plate like, sharp edge	1	Fastest
RLD	20	weak	unknown	10.01±0.26	Irregular, spherical	Unknown	Fastest

^*1 and 2 represent Spectrum and BASF, respectively.

^**Based on the certificate of the analysis of the raw API, several different residual solvents are listed: N,N-dimethylformamide, methanol, methylene chloride, THF, pyridine, acetic acid and ethanol. However, the specific amount of each is unclear.

Different formulation variables such as particle size and size distribution, morphology, crystallinity, excipient sources and residual solvent have different impacts on the in vitro release rate. However, these effects may or may not be observed in vivo since the physiological conditions (presence of enzymes, absence of surfactants, etc.) may alter the drug release behavior in vivo. Future research will focus on the in vivo performance testing (i.e., pharmacokinetics) of these formulations as well as investigation of possible in vitro-in vivo correlation. Together with the research reported here this will help providing a roadmap for the development of generic LAI suspension formulations.

4. Conclusions

API with different particle size and size distribution can be achieved through different recrystallization methods (solvent evaporation and antisolvent methods). Suspensions with larger drug particle size resulted in slower drug release rates. However, the particle size effect appears to become less significant when other factors such as residual solvent, as well as drug crystallinity and morphology are involved. Different sources of major excipients may also cause different drug release rates and therefore care must be taken during formulation design. Although the RLD has large particle size, the drug release rate was faster compared to two of the in-house prepared Q1/Q2 equivalent formulations (one with smaller drug particle size and another with similar drug particle size). This may be due to excipient source differences as well as differences in drug particle morphology of the formulations. A discriminatory in vitro release testing method was developed for the long-acting injectable suspensions, which will be valuable for formulation screening as well as quality control of this type of product.

Fig. 6.

SEM images of the RLD and the prepared Q1/Q2 equivalent MPA suspensions. (Upper panels are 5,000-fold magnification and lower panels are 25,000-fold magnification)

Abbreviations:

LAI

long-acting injectable

MPA

medroxyprogesterone acetate

PLGA

poly(lactic-co-glycolic acid)

API

active pharmaceutical ingredient

Q1/Q2 equivalent

qualitatively and quantitatively equivalent

HIV

human immunodeficiency virus

IM

intramuscular

SC

subcutaneous

RLD

reference listed drug

PEG

polyethylene glycol

PVP

polyvinylpyrrolidone

PXRD

powder X-Ray diffraction