Advanced Drug Delivery Reviews 59 (2007) 631 – 644
`
`www.elsevier.com/locate/addr
`
`Nanosizing — Oral formulation development and
`biopharmaceutical evaluation☆
`Filippos Kesisoglou, Santipharp Panmai, Yunhui Wu ⁎
`
`Department of Pharmaceutical Research, Merck & Co., Inc., West Point, PA, USA
`
`Received 21 March 2007; accepted 10 May 2007
`Available online 25 May 2007
`
`Abstract
`
`Poor aqueous solubility represents a major hurdle in achieving adequate oral bioavailability for a large percentage of drug compounds in drug
`development nowadays. Nanosizing refers to the reduction of the active pharmaceutical ingredient (API) particle size down to the sub-micron
`range, with the final particle size typically being 100–200 nm. The reduction of particle size leads to a significant increase in the dissolution rate of
`the API, which in turn can lead to substantial increases in bioavailability. This review describes the principles behind nanosizing, the production
`and characterization of nanoformulations as well as the current experience with utilization of such formulations in vivo.
`© 2007 Elsevier B.V. All rights reserved.
`
`Keywords: Nanosizing; Nanoparticle; Nanosuspension; Bioavailability enhancement; Dissolution; Formulation
`
`Contents
`
`1.
`2.
`3.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`Introduction .
`Increasing dissolution rate through nanosization — theoretical aspects .
`.
`.
`.
`.
`.
`.
`Formulation development of nanoformulations
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`3.1.
`Selection of excipients
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`3.2.
`Characterization of nanoformulations .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`3.3.
`Process development at lab and commercial scales
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`3.3.1.
`Feasibility of nanosuspension .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`3.3.2. Nanosuspension for toxicology study .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`3.3.3. Nanoformulation for clinical applications .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`4. Biopharmaceutical evaluations of nanoformulations .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`Computational approach — predicting performance of nanoformulations
`4.1.
`4.2. Nanoformulations for toxicology studies .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`4.3. Nanoformulations for preclinical and clinical applications .
`.
`.
`.
`.
`.
`.
`4.3.1.
`In vitro studies
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`4.3.2.
`Preclinical studies .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`4.3.3.
`Clinical applications .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`5. Nanosuspensions for non-oral applications
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`632
`632
`633
`633
`633
`634
`634
`635
`636
`637
`637
`637
`639
`639
`639
`641
`641
`
`☆ This review is part of the Advanced Drug Delivery Reviews theme issue on “Drug solubility: How to measure it, how to improve it”.
`⁎ Corresponding author. Merck & Co., Inc., WP75B-310, 770 Sumneytown Pike, PO Box 4, West Point, PA 19486-0004, USA. Tel.: +1 215 652 6911; fax: +1 215
`993 2265.
`E-mail addresses: filippos_kesisoglou@merck.com (F. Kesisoglou), santipharp_panmai@merck.com (S. Panmai), yunhui_wu@merck.com (Y. Wu).
`
`0169-409X/$ - see front matter © 2007 Elsevier B.V. All rights reserved.
`doi:10.1016/j.addr.2007.05.003
`
`LUPIN EX. 1015
`Lupin v. iCeutica
`US Patent No. 8,999,387
`
`Page 1
`
`

`
`632
`
`F. Kesisoglou et al. / Advanced Drug Delivery Reviews 59 (2007) 631–644
`
`Future directions
`6.
`7. Conclusions .
`.
`References .
`.
`.
`.
`.
`
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`.
`.
`.
`
`. 642
`. 642
`. 643
`
`nanosized formulations through the different stages of drug
`development covering both formulation aspects such as excipient
`selection and assessment of bioperformance. Examples of
`utilization of nanosizing to increase oral absorption are provided
`and, when applicable, compared to other novel oral dosage forms.
`Finally, we discuss the advantages and limitations of nanosizing in
`terms of its applicability in the drug development process.
`
`2. Increasing dissolution rate through nanosization —
`theoretical aspects
`
`1. Introduction
`
`Advances in combinatorial chemistry, biology and genetics in
`the recent years have led to a steady increase in the number of
`drug candidates under development. Due to the phospholipidic
`nature of cell membranes, a certain degree of lipophilicity is
`oftentimes a requirement for the drug compound, not only to be
`absorbed through the intestinal wall following oral administra-
`tion but possibly also to exert its pharmacological action in the
`target tissue. While high lipophilicity is advantageous in terms of
`compound permeability,
`it
`intrinsically translates into poor
`aqueous solubility. Since the first step in the oral absorption
`process is dissolution of the drug compound in the gastrointes-
`tinal lumen contents, poor aqueous solubility is rapidly becom-
`ing the leading hurdle for formulation scientists working on oral
`delivery of drug compounds [1].
`Nanosizing refers to the reduction of the active pharmaceutical
`ingredient (API) particle size down to the sub-micron range.
`While reduction of particle size has been employed in pharma-
`ceutical industry for several decades, recent advances in milling
`technology and our understanding of such colloidal systems have
`enabled the production of API particles of 100–200 nm size in a
`reproducible manner. The sub-micron particles are stabilized with
`surfactants or polymers in nanosuspensions which can be further
`processed into standard dosage forms, such as capsules or tablets,
`suitable for oral administration. These nanoformulations offer
`increased dissolution rates for drug compounds and complement
`other technologies used to enhance bioavailability of insoluble
`compounds (BCS Class II and IV) such as solubility enhancers
`(i.e. surfactants), liquid-filled capsules or solid dispersions of
`drugs in their amorphous state.
`The advantages of nanoformulations in oral drug delivery have
`been demonstrated in vitro in dissolution testing and in vivo in both
`preclinical studies as well as clinical trials. Nanocrystalline API has
`been shown to dramatically increase the rate of dissolution in vitro,
`improve bioavailability, reduce variability and alleviate positive
`food effects for orally administered drug molecules. As seen in
`Table 1, there are currently five pharmaceutical products that utilize
`nanocrystalline API to achieve their drug delivery goals. The goal
`of this review is to cover the theoretical background and practical
`aspects behind utilization of nanosizing as a means to improve oral
`bioavailability of drug compounds. We discuss development of
`
`Table 1
`Current marketed pharmaceutical products utilizing nanocrystalline API
`
`The solid API dissolution rate is proportional to the surface
`area available for dissolution as described by the Nernst–
`Brunner/Noyes–Whitney equation [2–4]:
`
`
`¼ A D
`Cs− Xd
`V
`
`dX
`dt
`
`h
`
`ð1Þ
`
`where dX/dt = dissolution rate, Xd = amount dissolved, A = par-
`ticle surface area, D = diffusion coefficient, V = volume of fluid
`available for dissolution, Cs = saturation solubility, h = effective
`boundary layer thickness.
`Based on this principle, API micronization has been
`extensively used in the pharmaceutical industry to improve
`oral bioavailability of drug compounds. It is evident that a
`further decrease of the particle size down to the sub-micron
`range will further increase dissolution rate due to the increase of
`the effective particle surface area [5]. For example in the case of
`aprepitant, the nanocrystal dispersion of 120-nm particle size
`exhibits a 41.5-fold increase in surface area over the standard
`5 μm suspension [6]. Furthermore, as described by the Prandtl
`equation, the diffusion layer thickness (h) will also be decreased
`thus resulting in an even faster dissolution rate [7].
`In addition to the dissolution rate enhancement described above,
`an increase in the saturation solubility of the nanosized API is also
`expected [8], as described by the Freundlich–Ostwald equation:
`
`
`ð2Þ
`S ¼ Sl exp
`
`2gM
`rqRT
`
`where S=saturation solubility of the nanosized API, S∞=satu-
`ration solubility of an infinitely large API crystal, γ is the crystal-
`
`Product
`
`Drug compound
`
`Indication
`
`Company
`
`Nanoparticle technology
`
`RAPAMUNE®
`EMEND®
`TriCor®
`MEGACE® ES
`Triglide™
`
`Sirolimus
`Aprepitant
`Fenofibrate
`Megestrol acetate
`Fenofibrate
`
`Immunosuppressant
`Antiemetic
`Treatment of hypercholesterolemia
`Appetite stimulant
`Treatment of hypercholesterolemia
`
`Wyeth
`Merck
`Abbott
`PAR Pharmaceutical
`First Horizon Pharmaceutical
`
`Elan Drug Delivery Nanocrystals®
`Elan Drug Delivery Nanocrystals®
`Elan Drug Delivery Nanocrystals®
`Elan Drug Delivery Nanocrystals®
`SkyePharma IDD®-P technology
`
`Page 2
`
`

`
`F. Kesisoglou et al. / Advanced Drug Delivery Reviews 59 (2007) 631–644
`
`633
`
`medium interfacial tension, M is the compound molecular weight,
`r is the particle radius, ρ is the density, R is a gas constant and T is
`the temperature.
`Assuming a molecular weight of 500, ρ = 1 g/mL and a γ
`value of 15–20 mN m- 1 for the crystal-intestinal fluid inter-
`facial tension, the above equation would predict an approxi-
`mately 10–15% increase in solubility at a particle size of
`100 nm. However a more significant increase in solubility ap-
`pears to occur in reality e.g. Muller and Peters reported an
`increase of 50% in the solubility of an insoluble antimicrobial
`compound when the particle size was reduced from 2.4 μm to
`800 or 300 nm [8]. This increase in solubility leads to a further
`increase in dissolution rate and, as a result, nanosuspensions
`often achieve significantly higher exposure levels compared to
`suspensions of micronized API, even when the same surfactants
`are used. Finally, the increase in surface wetting by the sur-
`factants in the nanosuspension formulations most likely results
`in a further enhancement of the dissolution rates compared to
`micronized suspensions.
`
`3. Formulation development of nanoformulations
`
`Compared with formulation efforts using traditional processes
`such as wet-granulation (WG), roller-compaction (RC), or direct
`compression (DC), development of nanoformulations is one of
`the more complex formulation works. Not only must the drug
`particles be rendered into nanosized domains via technically
`demanding processes, but
`they must also be stabilized and
`formulated rigorously to retain the nature and properties of the
`nanoparticles. This review will focus on Elan's nanomilling
`technology for oral formulation applications. Before delving in, a
`snapshot of other nanoparticle technologies is provided.
`For the purposes of this discussion,
`the definition of
`“nanoparticles” will be confined to crystalline particles with a
`monolithic core. There are two main approaches to making
`nanoparticles: ‘top down’ and ‘bottom up’ technologies [9,10].
`The ‘top down’ approach is by far the more popular; it will be
`referred to as ‘nanosizing’. The approach basically relies on
`mechanical attrition to render large crystalline particles into
`nanoparticles. Examples of the ‘top down’ approach include
`Elan's NanoCrystal® wet-milling technology [11] and Sky-
`ePharma's Dissocubes® high-pressure homogenization technol-
`ogy [9,12]. The ‘bottom up’ approach relies on controlled
`precipitation/crystallization [10]. The process involves dissolving
`the drug in a solvent and precipitating it in a controlled manner to
`nanoparticles through addition of an anti-solvent (usually, water).
`This technology is available from DowPharma (Midland, MI,
`USA) and BASF Pharma Solutions (Florham Park, NJ, USA). A
`hybrid approach is also feasible. Baxter's NANOEDGE®
`technology employs both ‘bottom up’ and ‘top down’ approaches
`through microprecipitation and homogenization [9].
`The focus and examples of this review will be based on the
`application of the NanoCrystal® technology to the development
`of nanoformulations. However, most of the discussion on
`properties and characterization of nanoparticles, selection of
`stabilizers, and considerations in nanoformulation development
`is relevant to the other technologies.
`
`3.1. Selection of excipients
`
`Formulation of nanosuspension requires a careful selection of
`stabilizers. Stabilizers are needed to stabilize the nanoparticles
`against inter-particle forces and prevent them from aggregating. At
`the nanometer domain, attractive forces between particles, due to
`dispersion or van der Waals forces, come into play [13]. This
`attractive force increases dramatically as the particles approach
`each other, ultimately resulting in an irreversible aggregation. To
`overcome the attractive interaction, repulsive forces are needed.
`There are two modes of imparting repulsive forces or energetic
`barriers to a colloidal system — steric stabilization and electrostatic
`stabilization. Steric stabilization is achieved by adsorbing polymers
`onto the particle surface. As the particles approach each other, the
`osmotic stress created by the encroaching steric layers acts to keep
`the particles separate. Electrostatic stabilization is obtained by
`adsorbing charged molecules, which can be ionic surfactants or
`charged polymers, onto the particle surface. Charge repulsion
`provides an electrostatic potential barrier to particle aggregation.
`Typically, the use of steric stabilization alone is sufficient to
`stabilize the nanoparticles and prevent irreversible aggregation.
`However, enough attractive force between particles may still
`remain to cause a loose and reversible flocculation. To circumvent
`flocculation, steric stabilization is often combined with electrostatic
`stabilization for additional repulsive contribution.
`Common pharmaceutical excipients that are suitable for use as
`polymeric stabilizers include the cellulosics, such as hydroxy-
`propylcellulose (HPC) and hydroxypropylmethylcellulose
`(HPMC), povidone (PVP K30), and pluronics (F68 and F127)
`[11,14–16]. The molecular weights of these polymers are usually
`between 50 k and 100 kDa. The chains should be long enough to
`provide a steric layer, but not too big to slow down dissolution.
`The surfactant stabilizers can be non-ionic, such as polysorbate
`(Tween 80), or anionic, such as sodium laurylsulfate (SLS) and
`docusate sodium (DOSS). Cationic surfactants are typically not
`used as stabilizers for oral formulation due to their antiseptic
`properties. Smaller surfactant molecules can also stabilize
`nanoparticles, but are usually more prone to Ostwald ripening
`and particle growth. Several groups have reported the use of the
`above stabilizers in their work [11]. Also, surfactants often help in
`the wetting and dispersion of the drug particles which are usually
`very hydrophobic. In marketed products based on Elan's
`NanoCrystal® technology, stabilizers such as HPMC E3,
`Povidone, HPC-SL, DOSS, and SLS have been used.
`Nanosuspensions are typically converted to a solid dosage
`form for clinical formulations. Prior to drying, redispersants need
`to be added to the nanosuspension to ensure complete redisper-
`sion of nanoparticles into their primary, pre-drying state [17].
`Sugars, such as sucrose, lactose, and mannitol, are commonly
`used as redispersants in oral formulations. The sugar molecules
`serve as “protectants” and prevent nanoparticles from aggregating
`as they are concentrated during drying [17].
`
`3.2. Characterization of nanoformulations
`
`A broad range of characterization tools and techniques exists for
`nanosuspensions. Muller et al. [12] have provided a comprehensive
`
`Page 3
`
`

`
`634
`
`F. Kesisoglou et al. / Advanced Drug Delivery Reviews 59 (2007) 631–644
`
`coverage of these characterization tools. Techniques for charac-
`terizing nanoparticles can be generalized into two sub-groups.
`The first group deals with attributes and properties of single
`nanoparticles, such as their particle size and surface charge (zeta
`potential). Particle crystallinity, dissolution, and surface coverage
`also fall in this category. The second group measures bulk
`properties, such as the viscosity. Redispersibility testing is
`additionally used to evaluate the redispersion of solid nanofor-
`mulations in relevant media, such as water and gastric fluid.
`The most basic property of a nanoparticle is its size. Various
`methods are available for particle size measurement [18]. A
`popular technique is laser light scattering, which allows quick
`determination of the particle size and distribution. Many models
`are available from Horiba [Irvine, CA, USA], Malvern [Worcester-
`shire, UK], and Microtrac [Montgomeryville, PA, USA] etc. One
`model that has been frequently used is the Horiba LA-910, which
`can measure from 50 nm to 1000 μm. Typically, only a few drops
`of nanosuspensions are required for a measurement (note that
`dilution into an aqueous medium is necessary). Useful information
`includes the mean values, along with the D10, D50, and D90 (D90
`means that 90% of the particles, by volume, are below the given
`size). The particle size distribution of milled nanosuspensions is
`typically narrow, with a coefficient of variation (CV) of about 25 to
`40%. For comparison, the CVof latex standard particles, which are
`relatively monodisperse, is generally 5 to 10% whereas the CV of
`un-milled APIs, which are more polydisperse, is typically N100%.
`Another fundamental property of nanoparticles is their surface
`charge. Surface charges can arise from (i) ionization of the particle
`surface or (ii) adsorption of ions (such as surfactants) onto the
`surface. Typically, the surface charge is assessed through measure-
`ments of the zeta potential. Zeta potential is the potential at the
`hydrodynamic shear plane and can be determined from the particle
`mobility under an applied electric field [13]. The mobility will
`depend on the effective charge on the surface. Zeta potential is also
`a function of electrolyte concentration. Examples of dilution media
`− 4 M. Various models are
`are aqueous KCl solutions, e.g., 10
`available from Brookhaven [Worcestershire, UK], Horiba [Irvine,
`CA, USA], Malvern [Worcestershire, UK], Matec [Northborough,
`MA, USA] etc. The addition of anionic surfactants typically leads
`to a more negative zeta potential value. Zeta potential values in the
`−15 mV to −30 mVare common for well-stabilized nanoparticles.
`Viscosity is one of the more prominent bulk properties. For a
`nanosuspension, whose viscosity can vary dramatically, depend-
`ing on the extent of flocculation, it is helpful to determine the
`viscosity as a function of shear rate. Either a controlled-stress or
`a controlled-strain rheometer can be used (note that the yield
`stress can only be determined with the former design). Several
`models are available from TA Instruments [New Castle, DE,
`USA], Malvern [Worcestershire, UK], and Brookfield [Mid-
`dleboro, MA b USA]. Typically, measurements can be made
`using the cone-and-plate geometry. The working shear rate range
`is from 0.01 to 1000 s- 1. The viscosity ranges from 1 cP for water
`or dilute nanosuspensions to 1000 cP or greater for concentrated
`nanosuspensions. Newtonian behavior (constant viscosity
`across the usual range of shear rates) is typical of well-stabilized
`nanosuspensions while shear-thinning (decreasing viscosity
`with increasing shear rate) is inherent to flocculating systems.
`
`3.3. Process development at lab and commercial scales
`
`3.3.1. Feasibility of nanosuspension
`At the early stage in development, the API is usually in tight
`supply, e.g., even an amount of 100 mg may be hard to come by.
`Therefore, it is crucial that
`the feasibility of a nanomilled
`suspension can be assessed with as little API as possible.
`Typically, the feasibility work can be carried out at the small
`scale using 100 mg or less of API. The Nanomill® System [Elan
`Drug Discovery, King of Prussia, PA] can be employed. The
`working capacity of the smallest chamber is 10 mL, and a very
`small suspension volume can be evaluated. The nanomilling
`process involves the high shearing of drug suspensions in the
`presence of grinding media, as described by Merisko-Liversidge
`et al. [11]. The milling media is a highly cross-linked polystyrene
`resin (500-μm beads). Selected stabilizers can be screened, such
`as the cellulosics and pluronics. The batch milling time is
`normally within a few hours at a mill speed of 5000 rpm. The
`particle size of the milled suspension can be checked at the
`completion of milling. Drug suspensions with terminal mean
`particle size in the 100- to 250-nm range are generally deemed
`feasible and can be considered for preclinical pharmacokinetic
`evaluation (the ‘success rate’ in reaching the described mean
`particle size range, based on oral bioavailability enhancement is
`around 80% to 90% in our experience). The milled material can
`be recovered from the suspension + media mixture. At
`this
`juncture, only short-term physical stability (e.g. 24 to 48 h)
`needs to be demonstrated, mainly to cover the duration of the
`animal study. If oral bioavailability enhancement is achieved and
`further development is warranted, then additional formulation
`development and optimization work can be conducted.
`While particle size and morphology of the starting API are of
`less concern if nanosizing is to be employed in formulation
`development, the chemical form of the API needs be considered
`prior to laboratory testing. Typically, the neutral form is the
`preferred starting form. While there has been an example of a
`salt form-containing nanoformulation, such as Par Pharmaceu-
`tical's MEGACE® ES with its acetate salt, pharmaceutical salts
`are generally not preferred. Possible liabilities of the salt form
`are (i) risk of disproportionation, e.g., an HCl salt disproportio-
`nating to the free base form during nanomilling, (ii) risk of
`aggregation due to charge-based interactions in the small
`intestine, such as those with bile salts, and (iii)
`rapid
`solubilization and turnover of nanoparticles of a salt form into
`larger particles of the neutral form due to the pH changes in the
`gastrointestinal (GI) tract. Furthermore, the API's solubility
`should be low to minimize the potential for Ostwald ripening,
`and the most stable form in water should be used. APIs with
`ionizable groups and pKa between 2 and 7 (e.g. physiological
`pH range) run the risk of charge-based interactions even if not
`presented as a salt. The typical starting particle size of the API is
`between a few microns and a hundred microns. Larger starting
`materials are acceptable at the feasibility stage, but run the risk
`of clogging the nanomill at the larger processing scale (in which
`case, an API pre-milling step is usually employed).
`Another consideration is the possibility of shear-induced API
`form conversion or amorphous drug formation. Milling speed is
`
`Page 4
`
`

`
`F. Kesisoglou et al. / Advanced Drug Delivery Reviews 59 (2007) 631–644
`
`635
`
`often the cause of formation of amorphous form of the API,
`which can lead to enhanced solubility and Ostwald ripening.
`Oftentimes, poor physical stability can be attributed to
`amorphous form formation. One remedy would be to mill at a
`lower speed. Generation of heat by the milling process can also
`result in form conversion. The mill is typically jacketed to
`minimize the temperature rise. The crystallinity (and form) of
`the milled API can be checked by XRPD. This is accomplished
`by spinning down the nanoparticles via ultracentrifugation and
`performing a measurement on the moist sediment. Crystalline
`peaks of the API should be identifiable on top of the broad
`amorphous band of the polymeric stabilizer.
`
`3.3.2. Nanosuspension for toxicology study
`The next step in formulation development is geared toward
`supporting toxicology studies. An early-phase toxicology study
`typically spans 2 weeks to 3 months. The API requirement could
`range from a few hundred grams to a few kilograms. It is usually
`acceptable to formulate as a liquid nanosuspension at this stage
`[as opposed to a solid nanoformulation, which is preferred in a
`clinical setting (see Section 3.3.3)]. In the following sub-
`sections, various aspects in formulating and manufacturing
`nanosuspensions for toxicology studies are discussed: from
`composition and process to storage and manufacturing logistics.
`
`3.3.2.1. Stabilizer selection. A nanosuspension formulation
`for toxicology studies needs to be stable as well as be
`processable at the target drug concentration. Hence, formulation
`screening work should continue until a sufficiently stable
`nanosuspension can be identified. The work can be conducted at
`the small scale, for example, with the Nanomill® system and at
`low drug concentrations. A reasonable stability protocol would
`be to mill various nanosuspensions and monitor the particle size
`of these suspensions at 5 °C and ambient, starting from a day to
`a few weeks. A two- to three-week time is adequate to identify a
`sufficiently stable nanosuspension. Typically, the search should
`produce a few feasible polymeric stabilizers, such as HPMC or
`HPC. A suitable working polymer:drug ratio is from 0.05:1 to
`0.5:1. After a stable composition has been identified, the next
`step is to scale-up the drug loading to the target concentration.
`Normally, the target will be at least 100 mg/mL to meet the
`needs for toxicology studies. At these higher concentrations, the
`main issue is usually flocculation. Physical stability of more
`concentrated suspensions generally falls in line with those of
`dilute suspensions. The major consequences of flocculation are
`two-fold. The first is the larger effective particle size with
`reduced surface area for dissolution. The second is the viscosity
`increase (but this is generally minor and not problematic). In
`many cases, flocculation can be minimized by raising the level
`of anionic surfactant, such as SLS or DOSS, which helps
`improve wetting and electrostatic stabilization. Care should be
`taken not to add excessive surfactant as this can result in
`enhanced solubility and Ostwald ripening. To characterize a
`nanosuspension formulation, only a small quantity is needed, on
`the order of tens of grams of API. This low API requirement is
`advantageous, given the limited supply of API available at this
`point in the development process.
`
`One concept that could prove useful in the selection of
`polymeric stabilizer is that of surface coverage. In principle, to
`fully provide steric stabilization, the polymeric stabilizers must
`fully adsorb onto the surfaces of the nanoparticles. While
`nanosuspensions can and have been formulated successfully,
`little attention is paid to the whereabouts of the stabilizer.
`Knowledge of the adsorption isotherm may help provide
`additional insights into the formulation efforts. Panmai and
`Deshpande [19] described a convenient method for determining
`the adsorption isotherm of a nanosuspension, which involves
`the determination of the fractions of the stabilizer that are bound
`to the drug surface and unbound in solution for a given polymer
`concentration. A drug example was given using HPC-SL and
`PVP K29/32 as stabilizers (mean particle size = 100 nm). First, a
`series of nanosuspensions were prepared for different stabilizers
`and at different amounts (ranging from 0.05:1 to 0.5:1 stabilizer:
`drug). Then, the nanosuspensions were ultracentrifuged to settle
`the nanoparticles,
`leaving a clear supernatant, which was
`assayed for the concentration of the unbound polymer. Through
`mass balance, the fractions of bound and unbound polymers
`were calculated. The resulting adsorption isotherm clearly
`showed a monotonic adsorption and surface saturation for HPC-
`SL. On the other hand, there was a virtual lack of surface
`adsorption for PVP K29/32. The greater affinity of HPC-SL is
`likely due to its greater hydrophobicity than that of PVP.
`Furthermore, the minimum ratio of HPC-SL to drug to ensure
`surface coverage is around 0.12 to 1. This result is very much in
`line with the common working ratios of 0.1:1 to 0.2:1 for
`stabilizer:drug. The value is expected to change with the
`stabilizer and the particle size. Hence, this approach may be
`used to select a polymeric stabilizer on a more rational basis.
`
`3.3.2.2. Milling process. A conventional media mill would be
`needed to process the amount of API and suspension volume
`required for toxicology studies. Assuming a drug concentration
`of 100 to 200 mg/mL for a typical toxicology study (in dogs or
`rats) the required suspension volume is often greater than 5 L.
`One example of a suitable media mill is the Dynomill [Glen
`Mills, Inc., Clifton, NJ], with chamber sizes of 300 mL and
`600 mL. Inside the milling chamber is a shaft with a series of
`impellers, which provide the high-shear agitation (up to
`4000 rpm). Larger mills, such as the Netzsch mills [Netzsch
`Inc., Exton, PA], which come in 2-L, 10-L, and 60-L chamber
`sizes, also exist to handle even larger volume requirements. To
`supply a large volume, the mill is configured in the recirculation
`mode. For example, a 600-mL chamber, charged with milling
`media and suspension, can be connected to a vessel of 5 to 10 L.
`The suspension then flows into the milling chamber, undergoes
`intense media grinding, and exits the mill through a small gap.
`The milling media are strained by the gap and retained within
`the mill. The mill and the vessel are jacketed to control the
`temperature. The inlet suspension is around 5 to 10 °C while the
`outlet product can be anywhere from 15 to 30 °C, or even
`higher, depending on the mill speed and product viscosity. The
`overall milling time scales according to the residence time in the
`chamber. The required supply can typically be prepared within a
`12- to 48-h total run time. Once milling is complete, the product
`
`Page 5
`
`

`
`636
`
`F. Kesisoglou et al. / Advanced Drug Delivery Reviews 59 (2007) 631–644
`
`is harvested. The process yield is generally very high, being
`greater than 95% yield.
`It is perhaps worthwhile noting some of the key variables in
`media milling. Milling speed is one of the most important
`operating parameters. Typically,
`the milling tip speed falls
`between 5 and 15 m/s [20]; the actual impeller or shaft speed (in
`the thousands of rpms) of a particular mill will vary with its
`design and size. The media loading is also important and is
`generally around 80% to 90% of the maximum fill volume,
`which permits milling to the 100- to 200-nm range. As for the
`suspension properties, the milling time generally decreases with
`increasing suspension viscosity, due to higher shear forces and
`energy input, and with increasing drug loading, due to higher
`particle densities in the grinding zone [20].
`
`3.3.2.3. Manufacturing logistics. The use of nanosuspensions
`in toxicology studies can be complicated by the manufacturing
`logistics involved, which if not well managed, may result in a
`delay in the development program. Since the duration of the
`study can be quite long, e.g., up to 3 months, API and
`suspension volume requirements for the studies are large. Also,
`the doses are normally high, from 100 to 1000 mg/kg. These
`two factors, not to mention the non-routine processing, virtually
`rule out the option of a daily preparation. As a result, the
`nanosuspension supply needs to be prepared in bulk (or at least
`roll-outs) and analytically tested prior to the initiation of the
`toxicology studies. With careful planning of the manufacture
`well in advance, the supply can typically be turned around in a
`few weeks — a time-span that should be acceptable to the
`Project Team.
`
`In addition to supply,
`3.3.2.4. Nanosuspension storage.
`stability and storage are additional considerations for long-
`term usage of nanosuspensions. As mentioned above,
`the
`toxicology studies can be up to 3 months long, or perhaps
`longer, a time-frame which can pose concerns with respect to
`both