Journal of Controlled Release 219 (2015) 2–7
`
`Contents lists available at ScienceDirect
`
`Journal of Controlled Release
`
`j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / j c o n r e l
`
`Controlled Drug Delivery: Historical perspective for the next generation
`Yeon Hee Yun, Byung Kook Lee, Kinam Park ⁎
`
`Purdue University, Departments of Biomedical Engineering and Pharmaceutics, West Lafayette, IN 47907, U.S.A.
`
`a r t i c l e
`
`i n f o
`
`a b s t r a c t
`
`Article history:
`Received 24 June 2015
`Received in revised form 29 September 2015
`Accepted 2 October 2015
`Available online 9 October 2015
`
`Keywords:
`Drug delivery
`History
`Physicochemical barriers
`Biological barriers
`
`The modern day drug delivery technology is only 60 years old. During this period numerous drug delivery
`systems have been developed. The first generation (1950–1980) has been very productive in developing many
`oral and transdermal controlled release formulations for clinical applications. On the other hand, the second
`generation (1980–2010) has not been as successful in generating clinical products. This is in large part due to
`the nature of the problems to overcome. The first generation of drug delivery technologies dealt with physico-
`chemical problems, while the second struggled with biological barriers. Controlled drug delivery systems can
`be made with controllable physicochemical properties, but they cannot overcome the biological barriers. The
`third generation (from 2010) drug delivery systems need to overcome both physicochemical and biological bar-
`riers. The physicochemical problems stem from poor water solubility of drugs, large molecular weight of peptide
`and protein drugs, and difficulty of controlling drug release kinetics. The biological barriers to overcome include
`distribution of drug delivery systems by the body rather than by formulation properties, limiting delivery to a
`specific target in the body. In addition, the body's reaction to formulations limits their functions in vivo. The pros-
`perous future of drug delivery systems depends on whether new delivery systems can overcome limits set by
`human physiology, and the development process can be accelerated with new ways of thinking.
`© 2015 Elsevier B.V. All rights reserved.
`
`1. Drugs and drug delivery systems
`
`Drug delivery systems exist to provide a more effective way to deliv-
`er drugs. The most important ingredient in any formulation is the drug.
`All other ingredients, collectively known as excipients, in a formulation
`are used to make the drug more effective. Once in a while, a newly
`developed drug becomes a blockbuster drug, i.e., the annual sales
`exceed $1 billion. The blockbuster drugs during the last few years in-
`clude those treating hypercholesterolemia (e.g., Lipitor and Crestor),
`acid reflux (e.g., Nexium), arthritis (e.g., Humira, Enbrel, and Remicade),
`depression (Seroquel, Cymbalta, and Zyprexa), and asthma (Advair and
`Singular). Of these, Seroquel is unique in formulation as it employs a
`sustained release technology for once-a-day delivery of quetiapine.
`Quite often, sustained release versions of drug formulations are devel-
`oped for product lifecycle management [1]. Thus, the sustained release
`technology is important to make existing drugs more effective.
`When a new drug is developed, it is usually formulated into a sim-
`plest possible dosage form that is effective in treating the intended
`disease. Different drugs have different physicochemical and biological
`properties, necessitating different formulations. This point is made
`here by comparing oral and parenteral routes of administration.
`Table 1 lists some of the drug properties that need to be considered
`for finding suitable delivery systems. Since oral delivery is the most
`
`⁎ Corresponding author at: Purdue University, Weldon School of Biomedical
`Engineering, 206 S. Martin Jischke Drive, West Lafayette, IN 47907, USA.
`E-mail address: kpark@purdue.edu (K. Park).
`
`convenient and widely used route of drug administration, it is the first
`to consider. Some drugs, however, have very poor water solubility or
`very poor permeability across the cells, making it difficult to develop
`oral formulations. In addition, a recent breed of biotech drugs, such as
`peptides, proteins, and nucleic acids, is much larger than the traditional
`small molecular drugs. They are usually delivered by parenteral routes
`due to their large size, limited stability, and short half life.
`
`2. History of drug delivery technologies
`
`Before 1950, all drugs were made into pill or capsule formulations
`that released the loaded drug immediately upon contact with water
`without any ability to control the drug release kinetics. In 1952, Smith
`Klein Beecham introduced the first sustained release formulation that
`was able to control the drug release kinetics and achieve 12-h efficacy
`[2]. The technology, known as the Spansule technology, allowed control
`of the drug release kinetics at a predetermined rate. In the early days
`when the new controlled drug delivery technology began, various
`terms were introduced to describe newer formulations having minor
`differences each other. Controlled release formulations included those
`with sustained release, timed release, extended release, and others. Of
`these, the term “sustained release” has been used more widely than
`any other names. These terms, however, are used interchangeably
`nowadays. After several decades of advances in drug delivery technolo-
`gies, the small differences in the functions that different names entail
`have become unnecessary.
`
`http://dx.doi.org/10.1016/j.jconrel.2015.10.005
`0168-3659/© 2015 Elsevier B.V. All rights reserved.
`
`ALKERMES EXHIBIT 2023
`Amneal Pharmaceuticals LLC v. Alkermes Pharma Ireland Limited
`IPR2018-00943
`
`Page 1 of 6
`
`

`

`Y.H. Yun et al. / Journal of Controlled Release 219 (2015) 2–7
`
`3
`
`Table 1
`Drugs with different properties requiring different delivery systems.
`
`Drugs
`
`Small molecules
`
`BCS class I
`
`BCS class II
`
`BCS class III
`
`BCS class IV
`
`↓
`Molecular weight
`↑
`Water solubility
`↑
`Cell permeability
`↕
`Half life
`Oral
`Main delivery route
`BCS: Biopharmaceutics Classification System
`↓: Low ↑: High ↔: Acceptable ↕: Individual variation from low to high.
`
`↓
`↓
`↑
`↕
`Oral
`
`↓
`↑
`↓
`↕
`Oral
`
`↓
`↓
`↓
`↕
`Oral
`
`Large molecules
`
`Peptides
`
`↑
`↔
`↓
`↓
`Parenteral
`
`Proteins
`
`↑
`↔
`↓
`↓
`Parenteral
`
`Nucleic acids
`
`↑
`↔
`↓
`↓
`Parenteral
`
`The history of controlled drug delivery field is described in Table 2.
`Most of the fundamental understanding on the drug release mecha-
`nisms, especially oral and transdermal dosage forms, was obtained
`during the first generation (1G) of development from 1950 to 1980.
`This period identified four drug release mechanisms that accelerated
`development of numerous oral and transdermal controlled release
`formulations. The most widely used mechanisms were dissolution-
`controlled and diffusion-controlled systems. Osmosis-based formula-
`tions gained a transient popularity, but the number of products based
`on osmosis is orders of magnitude smaller than those with the other
`two. The ion-exchange mechanism distinguishes itself from the others,
`but it has not been useful without combining with diffusion-controlled
`mechanism. Even today, many oral once-a-day formulations are devel-
`oped based on the dissolution- or diffusion-controlled mechanism.
`Since oral delivery is the most convenient mode of drug administration,
`oral sustained release formulations will continue to flourish.
`Unlike 1G drug delivery formulations, the second generation (2G)
`technologies have been less successful, as measured by the number of
`clinical products produced. One of the reasons for this is that the 2G
`technologies deal with more difficult formulations. For example, inject-
`able depot formulations made of biodegradable poly(lactic-co-glycolic
`acid) (PLGA) are designed to deliver peptide and protein drugs for a
`month or longer. Most depot formulations have a difficult time control-
`ling the initial burst release, which often releases 50% of the total drug in
`the first day or two [3]. During the 2G period, pulmonary delivery
`systems for insulin have been also developed. Pulmonary insulin deliv-
`ery system was developed, but its lower bioavailability required deliv-
`ery of several times more drug than required by parenteral injection.
`This, in turn, resulted in unexpected side effects that, along with other
`factors, caused withdrawal of the product from the market [4]. In an
`alternative approach, various self-regulated insulin delivery systems
`
`Table 2
`History of drug delivery technology from 1950 to the present and the technology
`necessary for the future.
`
`1950
`
`1980
`
`2010
`
`2040
`
`Year
`
`1st Generation
`Basics of controlled release
`Oral delivery
`Twice-a-day, once-a-day
`Transdermal delivery
`Once-a-day, once-a-week
`
`Drug release mechanisms
`Dissolution
`Diffusion
`Osmosis
`Ion-exchange
`
`2nd Generation
`Smart delivery systems
`Zero-order release
`First-order vs zero-order
`Peptide and protein delivery
`Long-term depot using
`biodegradable polymers
`Pulmonary delivery
`Smart polymers and hydrogels
`Environment-sensitive
`Self-regulated release
`(working only in vitro)
`
`Nanoparticles
`Tumor-targeted delivery
`Gene delivery
`
`Successful control of
`physicochemical properties of
`delivery systems
`
`Inability to overcome
`biological barriers
`
`3rd Generation
`Modulated delivery systems
`Poorly soluble drug delivery
`Non-toxic excipients
`Peptide and protein delivery
`Delivery for >6 months
`Control of release kinetics
`Non-invasive delivery
`Smart polymers and hydrogels
`Signal specificity and
`sensitivity
`Fast response kinetics
`(working in vivo)
`Targeted drug delivery
`Non-toxic to non-target
`cells
`Overcoming blood-brain
`barrier
`Need to overcome both
`physicochemical and
`biological barriers
`
`have been developed over the years [5–8]. Self-regulated insulin
`delivery systems work reasonably well in the laboratory setting, but
`they lose the function soon after implanted in vivo. The last decade of
`the 2G period (i.e., 2000~2010)has focused on tumor-targeted drug de-
`livery using nanoparticles. The seemingly promising nanoparticle
`approaches based on small animal models have not been successful in
`numerous clinical trials [9,10]. The limited successes of the 2G technol-
`ogies need careful analysis to make the current 3G technologies
`prepared for eventual clinical applications.
`
`3. Differences between 1G and 2G drug delivery technologies
`
`Development of more clinical products based on the 3G technolo-
`gies, which are still under development, requires understanding why
`most of the 2G technologies have not been translated into clinical prod-
`ucts. Huge successes of the 1G technology are mainly based on the oral
`and transdermal drug delivery systems. In these formulations, adjusting
`in vitro drug release kinetics has a direct effect on the in vivo pharmaco-
`kinetics. For oral and transdermal systems, the relationships between
`in vitro drug release kinetics and in vivo bioavailability are fairly well
`understood. Once the in vitro–in vivo correlation (IVIVC) of a formula-
`tion is established, other formulations using different mechanisms can
`be easily produced with an expectation that the new systems will be
`as effective as the reference formulation [11,12]. For most drug delivery
`systems developed in the 1G period, mainly for oral and transdermal
`delivery, understanding the physicochemical properties (e.g., in vitro
`drug release kinetics) was enough for developing clinically useful
`formulations. No particular biological barriers were identified for
`those formulations, except for the inability to overcome the limited
`gastrointestinal (GI) transit time and the different absorption properties
`by different segments in the GI tract (i.e., absorption window) of oral
`formulations.
`The drug delivery systems developed during the 2G period dealt
`with more difficult problems. The technologies developed during the
`2G period are listed in Table 2. Various oral controlled release formula-
`tions were developed to achieve zero-order release, but the zero-order
`release achieved in various in vitro dissolution systems did not result
`in maintenance of the constant drug concentration in vivo, mainly due
`to the variations in the drug absorption properties along the GI tract.
`Drug absorption is controlled by the biological barrier, in addition to
`the drug release kinetics from oral formulations. More importantly,
`maintaining the constant drug concentration in the blood is not neces-
`sary, as long as the drug concentration is above the minimal therapeuti-
`cally effective concentration [13]. The 2G period also introduced
`sustained release formulations of peptide/protein drugs after implanta-
`tion in the body [14,15]. The drug release from a formulation in vivo
`depends not only on the formulation properties, but also on the biolog-
`ical environment surrounding the implanted formulation. This makes
`prediction of the drug release kinetics in vivo, and thus, bioavailability,
`more difficult. Simply put, IVIVC has not been found for most parenteral
`formulations of biotech drugs, making it difficult to predict the in vivo
`bioavailability from the in vitro release profiles, especially for long-
`
`Page 2 of 6
`
`

`

`4
`
`Y.H. Yun et al. / Journal of Controlled Release 219 (2015) 2–7
`
`Table 3
`Barriers to overcome by the 3G drug delivery systems.
`
`Delivery technology
`
`Poorly water-soluble drug delivery
`
`Formulation barriers
`• New excipients for increasing drug solubility
`
`Peptide/protein/nucleic acid
`delivery
`
`Targeted drug delivery using
`nanoparticles
`
`• Control of drug release kinetics
`• Control of drug loading
`• Control of therapeutic period
`• Control of nanoparticle size, shape, surface chemistry, functionality,
`and flexibility
`• Surface modification with ligands
`• Stimuli-sensitive delivery systems
`
`Self-regulated drug delivery
`
`• Signal specificity and sensitivity
`•
`Fast responsive kinetics
`• Ability to stop drug release
`
`Biological barriers
`• Non-toxic to the body
`• No drug precipitation in the blood
`•
`IVIVC
`•
`Long-term delivery up to a year
`• Non-invasive delivery
`• Controlling biodistribution through altering vascular
`extravasation, renal clearance, metabolism, etc.
`• Navigating microenvironment of diseased tissues to reach
`target cells
`• Crossing endothelial barriers (e.g., blood–brain barrier)
`• Crossing mucosal barriers
`•
`Functional inside the bod
`•
`Functional over the lifetime of drug delivery
`
`term depot formulations [16]. Furthermore, there are no standard
`in vitro drug release test methods that can reliably predict in vivo
`pharmacokinetic profiles [17]. The difficulty of predicting in vivo behav-
`ior of drug delivery systems is aggravated for self-regulated insulin
`delivery systems. Upon introduction to the body, modulated insulin
`delivery systems fail to function after a day or two due to the interfer-
`ence with proteins and cells present in the body [18]. Recent uses of
`nanotechnology for tumor-targeted drug delivery are another casualty
`of inadequate understanding of the effects of the body on drug delivery
`systems [13]. In short, the difficulty faced by the 2G drug delivery
`systems is mainly due to the inability of the drug delivery systems to
`overcome biological barriers.
`
`4. The 3G drug delivery technologies
`
`The limited success of the 2G drug delivery technologies is, in large
`part, due to their inability to overcome the body responses after drug
`delivery systems are administered by parenteral route. The current
`drug delivery systems, however smart they may have been constructed,
`are not able to deal with challenges posed by the biological environment
`which is not-well understood and unpredictable. For the 1G formula-
`tions, controlling physicochemical properties, such as water solubility
`and cell permeability, were adequate enough to establish IVIVC. The
`3G drug delivery technologies will have to be advanced much beyond
`the 2G technologies to overcome both physicochemical and biological
`barriers. As a brief review of the 2G technologies above indicates, under-
`standing and overcoming the biological barriers, in addition to physico-
`chemical barriers, is the key for success. Some of the barriers to
`overcome for developing successful 3G drug delivery systems are listed
`in Table 3. There are many other drug delivery systems that need to
`be developed during the 3G period. The four areas in Table 3 are
`discussed here solely to emphasize the importance of understanding
`and overcoming biological barriers.
`
`4.1. Delivery of poorly water-soluble drugs
`
`Poor water solubility of drugs was one of the most important prob-
`lems in drug development, and it still remains to be true today. Discus-
`sion on poorly soluble drugs requires understanding of the meaning of
`drug solubility. Table 4 shows the descriptive terms used in U.S.
`Pharmacopeial and National Formulary to indicate approximate drug
`solubilities in water. The term “poorly soluble” is commonly used to de-
`scribe drugs that belong to the “practically insoluble” category. For these
`drugs the aqueous solubility is 0.1 mg/mL or less, i.e., 100 μg/mL or less.
`Many new drug candidates are poorly water soluble, and thus, a large
`portion of the candidate drugs are not translated into clinically useful
`formulations. Analysis of 200 orally administered drug products
`showed that practically insoluble drugs account for almost 40% of the
`
`total drugs [19]. Delivering these drugs effectively through the GI tract
`for therapeutically effective bioavailability remains an important issue.
`The dissolution rate of practically insoluble drugs may be so slow that
`dissolution takes longer than the GI transit time resulting in therapeuti-
`cally unacceptable bioavailability [20].
`Technologies to dissolve poorly soluble drugs in water have been
`studied for decades, and some of the methods are listed in Table 5.
`Poorly soluble drugs have inherently low water solubility, and thus,
`suitable excipients are added to increase the solubility by using
`surfactants, polymer micelles, hydrotropic agents, complexing agents
`(e.g., cyclodextrins and proteins), cosolvents, and lipid formulations
`(e.g., self-emulsifying systems) [21–23]. For weakly acidic or basic
`drugs, pH can be controlled to increase the drug solubility. Alternative
`to increasing the drug solubility, drug dissolution kinetics can be
`enhanced through selecting appropriate polymorph, making solid
`dispersions (i.e., maintaining amorphous structure of the drug using
`polymers), reducing drug particle size, and increasing wetting with
`surfactants. Of these, the solid dispersion approach has been widely
`used for its ease of preparation and efficacy [24–26]. Making drug
`nanocrystals has also been frequently used, as the increase in bioavail-
`ability by increasing the drug crystal surface resulted in improved bio-
`availability [23]. The surface area increases proportionally as the
`decrease in the size of drug particles. The drug solubility is an inherent
`property and so it should not change as the dissolution kinetics in-
`creases. But increasing the dissolution kinetics can result in improved
`bioavailability of oral formulations. Enhanced dissolution of the drug
`can produce the dissolved drug in sufficient quantity fast enough to
`replace those drugs that have been absorbed from the GI tract, thereby
`improving bioavailability.
`The problem of poor water solubility becomes even more serious for
`intravenous formulations. For example, many anticancer drugs are
`extremely poorly water soluble, e.g., b1 μg/mL, and thus they are usually
`dissolved in organic solvents. Paclitaxel and docetaxel are good exam-
`ples of poorly soluble drugs making injectable formulations difficult.
`There are various injectable formulations of paclitaxel: Taxol utilizing
`Cremophor® EL [27], Abraxane® based on paclitaxel-albumin complex
`
`Table 4
`Solubility definitions.
`
`Descriptive terms
`
`Parts of solvent required
`for 1 part of solute
`
`Solubility range
`
`mg/mL
`
`%
`
`Very soluble
`Freely soluble
`Soluble
`Sparingly soluble
`Slightly soluble
`Very slightly soluble
`Practically insoluble
`
`Less than 1
`From 1 to 10
`From 10 to 30
`From 30 to 100
`From 100 to 1,000
`From 1,000 to 10,000
`10,000 and over
`
`N1000
`100 ~ 1000
`33 ~ 100
`10 ~ 33
`1 ~ 10
`0.1 ~ 1
`≤0.1
`
`N100
`10 ~ 100
`3.3 ~ 10
`1 ~ 3.3
`0.1 ~ 1
`0.01 ~ 0.1
`≤0.01
`
`Page 3 of 6
`
`

`

`Y.H. Yun et al. / Journal of Controlled Release 219 (2015) 2–7
`
`5
`
`Table 5
`Methods to improve drug dissolution.
`
`Enhancing drug solubility
`
`Enhancing dissolution kinetics
`
`Using surfactant micelles
`Using polymer micelles
`Using hydrotropic agents
`Using complexing agents
`Using cosolvents
`Using self-emulsifying systems
`Controlling pH
`
`Selecting appropriate polymorph
`Making amorphous forms (solid dispersions)
`Reducing particle size (nanocrystals)
`Adding surfactant for better wetting
`
`[23,28,29], and Genexol® utilizing PEG-PLA polymer micelle [30].
`Taxotere, delivering docetaxel, a derivative of paclitaxel, is dissolved in
`polysorbate 80 which is suspected to cause hypersensitivity [31,32].
`Cremophor EL, an excipient used to increase the solubility of paclitaxel,
`can cause serious hypersensitivity reactions and kill patients if the
`patient is not properly preconditioned [27]. Development of new drug
`delivery systems for poorly soluble drugs without using organic solvent
`is important for bringing promising new drug candidates to clinical
`applications and more effective use of existing drugs.
`
`4.2. Peptide/protein/nucleic acid delivery
`
`Macromolecular drugs, such as peptides, proteins, and nucleic acids,
`are usually delivered by parenteral administration. They are too big to
`cross the intestinal epithelium, i.e., to be absorbed from the GI tract
`[33]. A number of attempts have been made to protect them from the
`harsh acidic condition of the stomach by enteric coating, and from
`enzymatic degradation by adding enzyme inhibitors. These attempts,
`however, do not address the real issue that proteins cannot be absorbed
`without enzymatic degradation into small molecules [34,35]. It has been
`suggested that nanoparticles can be translocated across M cells in
`Peyer's patches and enterocytes in the villus part of the intestine, but
`the extent of particle absorption has been controversial [36]. The
`absorbed amount is too low and too irreproducible to have therapeutic
`significance. Thus, these macromolecular drugs are mainly delivered by
`parenteral routes. Recently, new approaches have been attempted to
`deliver them by non-invasive, or minimally invasive means, such as pul-
`monary, nasal, and transdermal delivery [37].
`Macromolecular drugs usually have very short half-lives, ranging
`from minutes to hours, and thus, sustained release for months
`requires depot formulations. There are more than a dozen depot formu-
`lations that are administered by parenteral routes. They include
`Zoladex® Depot (goserelin acetate), Lupron Depot® (leuprolide
`acetate), Sandostatin LAR® Depot (octreotide acetate), Nutropin
`Depot® (somatropin), Trelstar® (triptorelin pamoate), Suprefact®
`Depot (Buserelin acetate), Somatuline® Depot (lanreotide), Arestin®
`(minocycline HCl), Eligard (leuprolide acetate), Risperdal® CONSTA®
`(risperidone), Vivitrol® (naltrexone), Ozurdex® (dexamethasone),
`and Bydureon® (exenatide). The fact that there are only a handful of
`depot formulations, as compared with thousands of oral sustained
`release formulations, indicates the difficulty associated with developing
`parenteral depot formulations. The majority of these formulations have
`the initial burst release, resulting in the initial peak blood concentration
`much larger (up to 100 times) than the therapeutically effective
`concentration at the steady state (i.e., after drug concentration at the
`steady state after the initial peak). Thus, it is urgently required to im-
`prove the technology of controlling the drug release profiles. The ability
`of controlling drug release kinetics becomes even more important as the
`drug loading increases. Depot formulations designed to have longer
`duration need higher drug loading. Thus, patient-friendly depot formu-
`lations must have higher drug loading with controllable drug release
`kinetics for a long-period of time, up to 1 year, or even longer.
`
`4.2.1. The initial burst release from PLGA depot formulations
`Examples of pharmacokinetic profiles of two clinically used depot
`products are shown in Fig. 1. Each PK profile can be divided into two
`regions: the initial burst release region (red arrows in Fig. 1) and the
`therapeutically effective region (green arrows in Fig. 1). The Y axis in
`Fig. 1 is in the log scale, and the peak concentration in the initial burst
`region is about 100 times larger than the concentrations that are in
`the therapeutically effective range. This observation brings a few ques-
`tions. First, is it really necessary to have 100 times higher drug concen-
`tration in the first day or two than the known therapeutically effective
`drug concentration? Second, does the initial burst release play any
`role in the efficacy of the drug at the steady state? There is no scientific
`reason to justify that the initial burst release is necessary for the thera-
`peutic effect. The initial burst release is simply an outcome of the emul-
`sion methods for microparticle production available a few decades ago.
`Some may argue that the initial peak concentration in blood may be
`necessary for therapeutic efficacy. This, however, cannot be true,
`because it implies that daily injection of the same drug without the
`peak concentration should not work. This, of course, is not the case. It
`is the drug concentrations in the therapeutically effective region that
`is important. Controlling the initial burst release is still not easy, but im-
`proved understanding on the emulsion methods and recent develop-
`ment of new microfabrication processes have made it possible to
`reduce or eliminate the initial burst release.
`
`4.3. Targeted drug delivery using nanoparticles
`
`Nanoparticle-based drug delivery systems have been used
`extensively for the last few decades. A search in SciFinder using
`“drug delivery nanoparticle” resulted in 19,950 references during
`1995–2014 (Fig. 2). Of these, 57% is associated with the term “target”
`for targeted drug delivery or targeting. Clearly, the majority of the stud-
`ies on nanoparticle-based drug delivery have been focused on targeted
`drug delivery, mainly on tumor-targeted drug delivery.
`The initial excitement on nanoparticulate drug delivery systems
`arose from the ability of producing nanoparticles in various size and
`shape, and the ability to control the physicochemical and surface prop-
`erties to make smart nanoparticles. Many of these systems have worked
`well in the laboratory where cell culture systems were used for testing
`drug delivery. The systems also worked reasonably well in small animal
`models, mostly xenograft mouse models. The nanoparticle systems
`showing promising results in those models have not been translated
`into clinical studies [38,39]. The current nanoparticles cannot control
`their fate after intravenous administration. The so-called “targeting”
`by nanoparticles is a misleading concept, because the current nanopar-
`ticles cannot find their way to an intended target, but are simply distrib-
`uted throughout the body by the blood circulation [40]. Only a very
`small fraction of the total administered nanoparticles end up at the
`target site, mostly by chance. The concept of the enhanced permeability
`and retention (EPR) effect is frequently cited whenever nanoparticles
`are used for drug delivery to tumors. However, most studies have not
`quantitatively measured the actual amount of drugs reaching the target
`tumor, and thus, there is no quantitative information on the role of the
`EPR effect in targeted drug delivery. The tumors grown in mice are
`usually 1 ~ 2 mm which are similar in size as the liver, but only a
`small fraction, in the range of about 1% of the total administered dose,
`of the so-called targeted nanoparticles end up at the tumors, while the
`majority ends up at the liver [41]. For nanoparticle systems to become
`a clinically effective tool for targeted drug delivery, they may have to
`be designed differently from those showing potential in small animal
`studies. The observations made in mice, which have only a few millili-
`ters of blood, may not be extended to human with 5 l of blood. Further-
`more, the size ratio of a tumor in a mouse is usually much larger than
`that of a tumor in a human. This massive scale differences need to be
`considered when experimental animal models are used and their data
`are analyzed.
`
`Page 4 of 6
`
`

`

`6
`
`Y.H. Yun et al. / Journal of Controlled Release 219 (2015) 2–7
`
`Fig. 1. Examples of pharmacokinetic profiles of Nutropin Depot (A) and Trelstar (B) (obtained from the packaging inserts). The red arrow indicates the PK region resulting from the initial
`burst release of a drug, and the green arrow indicates the PK region of the therapeutically effective drug concentrations.
`
`Nanoparticles may have unexpected benefits, even though the antic-
`ipated targeting has not been observed yet. The nanoparticles, with
`suitable surface modification, may alter the biodistribution, which in
`turn, may alter the toxicity profiles of the same drug. In fact, reducing
`the toxicity, or the side effects, of the drug through engineering
`nanoparticle formulations may be a better way of utilizing the unique
`properties of nanoparticles. Doxil®, the PEGylated liposome formula-
`tion, is a case in point. It was approved by the U.S. Food and Drug
`Administration not because of its improved drug efficacy, but because
`of its reduced cardiotoxicity [42]. Considering the difficulties in
`translating the targeting ability observed in mouse models to clinical
`applications, one could consider utilizing nanoparticle formulations for
`reducing the toxicity. This can be achieved not only by altering the
`biodistribution, but also by increasing the water solubility without
`using toxic organic solvents. Good examples of this approach are
`Abraxane® and Genexol® as described above. Formulations without or-
`ganic solvents, such as Cremophor EL or polysorbate, are certainly more
`desirable, especially when the resulting therapeutic efficacy is about the
`same [43].
`
`4.4. Self-regulated drug delivery
`
`Self-regulated drug delivery, in particular, self-regulated insulin
`delivery, remains one of the most important technologies to develop.
`Imagine that millions of diabetes patients can take care of their glucose
`
`Fig. 2. The number of articles on nanoparticle drug delivery systems published from 1995
`to 2014. In SciFinder, the research topic of “drug delivery nanoparticle” was used for the
`initial search to find more than 30,000 references containing the concept. The search
`was further refined using the research topic of “target”.
`
`level for months with one injection of self-regulated insulin delivery
`system, instead of multiple daily injections of insulin. There are several
`self-regulated insulin delivery systems developed over the years which
`work well in the laboratory setting [5–8,44,45]. As soon as they are
`introduced inside the body, however, their function decreases by
`hours. The glucose sensor, which is essential in detecting the varying
`glucose level, becomes less efficient due to protein adsorption and cell
`adhesion, and the insulin delivery module becomes less efficient after
`each cycle [18,46,47]. It has been several decades since the concept of
`self-regulated insulin delivery started, but the progress has been slow.
`This is also mainly due to the biological barriers that the body poses to
`the implanted device [48]. Unless the biological barriers are understood
`and the new delivery systems are designed to overcome those, develop-
`ment of self-regulated insulin delivery system will remain as a concept
`for a while. The biological barriers to overcome include maintaining
`glucose sensor specificity and sensitivity in the biological milieu.
`Another key requirement is to build an actuator that releases a right
`amount of insulin fast with automatic turn-off function [13].
`
`5. Perspective of the future
`
`Significant advances in drug development, along with better and
`early diagnostics for preventive medicine, have helped extend human
`life expectancy. This, in turn, requires development of more drugs for
`various diseases, such as coronary artery disease, diabetes mellitus,
`chronic pain, chronic lower respiratory disease, Alzheimer's disease,
`and Parkinson's disease. Finding drugs for these diseases is the first
`and most important step. The drug delivery systems can make drug
`candidates with poor water solubility into therapeutically effective
`drug formulation, and drug candidates with short half-lives into
`sustained release formulations. The drug delivery technologies will
`have valuable contributions to the development of