MEMS devices for drug delivery

Advanced Drug Delivery Reviews 128 (2018) 132–147

Contents lists available at ScienceDirect

Advanced Drug Delivery Reviews journal homepage: www.elsevier.com/locate/addr

MEMS devices for drug delivery☆ Hyunjoo J. Lee a, Nakwon Choi b,c, Eui-Sung Yoon b, Il-Joo Cho b,c,⁎ a b c

School of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon 34141, Republic of Korea Center for BioMicrosystems, Brain Science Institute, Korea Institute of Science and Technology (KIST), Seoul 02792, Republic of Korea Division of Bio-Medical Science & Technology (Biomedical Engineering), KIST School, Korea University of Science and Technology (UST), Daejeon 34113, Republic of Korea

a r t i c l e

i n f o

Article history: Received 20 May 2017 Received in revised form 6 September 2017 Accepted 2 November 2017 Available online 5 November 2017 Keywords: Brain infusion Blood-brain barrier (BBB) disruption Capsule endoscope Implantable drug delivery Microneedles Transdermal drug delivery

a b s t r a c t Novel drug delivery systems based on microtechnology have advanced tremendously, but yet face some technological and societal hurdles to fully achieve their potential. The novel drug delivery systems aim to deliver drugs in a spatiotemporal- and dosage-controlled manner with a goal to address the unmet medical needs from oral delivery and hypodermic injection. The unmet needs include effective delivery of new types of drug candidates that are otherwise insoluble and unstable, targeted delivery to areas protected by barriers (e.g. brain and posterior eye segment), localized delivery of potent drugs, and improved patient compliance. After scrutinizing the design considerations and challenges associated with delivery to areas that cannot be efficiently targeted through standard drug delivery (e.g. brain, posterior eye segment, and gastrointestinal tract), this review provides a summary of recent advances that addressed these challenges and summarizes yet unresolved problems in each target area. The opportunities for innovation in devising the novel drug delivery systems are still high; with integration of advanced microtechnology, advanced fabrication of biomaterials, and biotechnology, the novel drug delivery is poised to be a promising alternative to the oral administration and hypodermic injection for a large spectrum of drug candidates. © 2017 Elsevier B.V. All rights reserved.

Contents 1. 2.

3.

4.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Drug delivery to the brain . . . . . . . . . . . . . . . . . . . . . . . 2.1. Direct brain infusion . . . . . . . . . . . . . . . . . . . . . . 2.1.1. Design considerations, challenges, and opportunities. . . . 2.1.2. Recent advances and perspectives . . . . . . . . . . . . 2.2. Drug delivery to the brain through BBB disruption . . . . . . . . . 2.2.1. Design considerations, challenges, and opportunities. . . . 2.2.2. Recent advances and perspectives . . . . . . . . . . . . Ocular drug delivery . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Design considerations, challenges and opportunities . . . . . . . . 3.2. Recent advances . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1. Minimally invasive microneedles . . . . . . . . . . . . 3.2.2. Stimuli-responsive implants for on-demand drug delivery . 3.2.3. Batteryless actuation . . . . . . . . . . . . . . . . . . Gastrointestinal tract drug delivery – capsule endoscopy . . . . . . . . . 4.1. Design considerations, challenges, and opportunities . . . . . . . . 4.1.1. Anchoring mechanisms . . . . . . . . . . . . . . . . . 4.1.2. Trade-off between drug volume and on-chip functionalities 4.2. Recent advances . . . . . . . . . . . . . . . . . . . . . . . . 4.2.1. Legged mechanism . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

☆ This review is part of the Advanced Drug Delivery Reviews theme issue on “Microfluidic Devices for Drug Delivery Systems”. ⁎ Corresponding author. E-mail address: [email protected] (I.-J. Cho).

https://doi.org/10.1016/j.addr.2017.11.003 0169-409X/© 2017 Elsevier B.V. All rights reserved.

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . .

133 134 134 134 135 136 136 138 138 138 139 139 139 140 140 140 140 141 141 141

H.J. Lee et al. / Advanced Drug Delivery Reviews 128 (2018) 132–147

4.2.2. Other anchoring mechanisms . 4.2.3. Remaining problems . . . . . 5. Transdermal drug delivery. . . . . . . . . . 5.1. Challenges and opportunities . . . . . 5.2. Stimuli-responsive microneedles . . . 6. Implantable subcutaneous drug delivery . . . 6.1. Design considerations . . . . . . . . 6.2. Recent advances . . . . . . . . . . . 6.2.1. Biocompatibility and reliability 6.2.2. Perspective . . . . . . . . . 7. Conclusions . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

1. Introduction With the rapid development in novel pharmaceutical compounds and intervention, there is an increasing need for novel drug delivery systems that can address challenges associated with conventional drug delivery systems. For example, modern drug candidates include a wide spectrum of molecules including large biomolecules (e.g. peptides and proteins) with low bioavailability, small molecules with poor water solubility, and potent drugs with narrow therapeutic windows. Due to this wide range of physiochemical and pharmacokinetic properties of the modern drug candidates, conventional oral or intravenous administration might not be the most suitable route of administration. Evaluation of drug candidates and optimization of drug developments are thus limited by a lack of adequate delivery systems. In addition, with advances in diagnostic techniques, drug delivery is not limited only to wound healing and immunology but is applicable to cancer treatment, gene delivery, and insulin delivery [1], which requires targeted drug delivery with controlled release. There are also areas in our body such as the brain and the posterior eye segment that are significantly challenging sites to target through conventional intravenous administration due to the physical barriers (i.e. BBB and blood-retinal barrier) that separate the organs from blood circulation. One method to overcome these difficulties is to formulate nanoscale delivery vehicles to enable site-specific targeted drug delivery with a goal for oral delivery of proteins, peptides, and even imaging nanoprobes [2–4]. Nanoparticle-based drug delivery

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

133

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . .

141 142 142 142 142 143 144 144 144 144 144 144 144

enables targeted drug delivery to the desired regions but reliable encapsulation of drugs in the nanoparticle with desired release characteristics are challenging to achieve [5]. Microtechnology or Microelectromechanical Systems (MEMS) is another promising technology for developments of the novel drug delivery systems that overcome the current challenges and accommodate a vast variety of drug delivery applications. By offering miniaturization [4–7], integrations of multiple functions [8,9], and electromechanical control [10–12], microtechnology allows delivery of a wide range of drugs with high therapeutic efficacy. Also, microtechnology enables localized drug delivery to challenging areas in our body by means of alternative routes of administration. For instance, microneedle technology for transdermal drug delivery is one of the successful applications of microtechnology. Microneedle technology now in the third generation of development accommodates delivery of RNA and vaccine [13,14] and strives to achieve active control of drug release through smart triggering systems [8]. Using MEMS technology, hundreds of thin microneedles can be precisely manufactured in a single array to deliver drugs without incurring damage and pain. In addition, while the conventional drug delivery relies mostly on diffusion, MEMS micropump technology allows active control over drug release such as release rates and infusion volumes and provides means to continuously supply drugs through a reservoir [15]. Furthermore, the agile interface between a microtechnology component to electronic components greatly expands the efficiency and functionality of the novel drug delivery systems.

Table 1 Summary of MEMS devices utilized in novel drug delivery systems and the associated remaining challenges for systems targeting different body areas. Target area

MEMS devices

Brain

Implantable

Neural probes/microneedles [9,26,27,30–33,36–38]

Noninvasive

Ultrasound transducers [57–59,63,64] MEMS pumps, reservoir, polymeric inserts [76–83,85–90]

Eye

Implantable

Noninvasive GI

Oral

Skin

Implantable

Minimally invasive

Remaining challenges

Therapeutic contact lens [73] Capsule endoscope [100–114] Subcutaneous MEMS pumps, reservoir, polymeric inserts [137–144]

Transdermal microneedles [8,14,125–127,129]

• Biocompatibility • Systematic errors (leakage, clogging, biofouling) • • • • • • • • • • • • • • • • • • •

Multi-functionalities Miniaturization Higher control over ultrasound parameters Choric robustness Drug volume Removal On-demand activation Interaction between drugs and contact lens materials Preserving transparency Reliable anchoring and actuation mechanism Batteryless activation Biocompatibility Drug volume Removal Self-administrated release Closed-loop release On-demand release Multi-functionalities Chronic uses (clogging, biofouling)

134

H.J. Lee et al. / Advanced Drug Delivery Reviews 128 (2018) 132–147

Fig. 1. Schematic diagrams of various target organs and their associated physical barriers and targeting strategies using novel drug delivery systems enabled by microtechnology. The blood-brain barrier (BBB), stratum corneum, and retina-blood barrier are shown for brain, skin, and eye, respectively. All figures reprinted with permission [88,108,119,142,145].

Smart release through wireless transmission and accurate control of drug doses over multiple occasions are now possible [10]. Thus, advances in microfabrication technology and biocompatible materials have expanded the capability of drug delivery systems in terms of target sites, route of drug administration, patient compliance, a spectrum of deliverable drugs, and dosing procedures [16]. Not only to further advance existing drug delivery systems but also to accommodate the need for spatial-, temporal- and dosage-controlled release, it is important to understand the current challenges in the field. However, these challenges differ greatly depending on the drug candidates, target areas, and routes of administration. For instance, while the current challenge in targeting the brain and the posterior eye segment is intervention of the physical barriers, the recent challenge in developing advanced transdermal patch is less involved with overcoming stratum corneum (SC) barrier, but rather involves with advanced control of dosage duration (e.g. intermittently, as-needed, or over a long period). Thus, in this review, we first discuss the current challenges in targeting different areas of the body (Table 1) and discuss recent works that have addressed these problems (Fig. 1). 2. Drug delivery to the brain With a rapidly prevailing population ageing, there is an increasing need for effective treatments of neurodegenerative diseases. As an alternative to surgical resection, localized drug delivery to the brain is one of the important therapeutic means to treat brain diseases such as brain tumors and Alzheimer's diseases. However, unlike delivering drugs to most parts of our body, delivering drugs to the brain is not straightforward because of the existence of blood-brain barrier (BBB). BBB, the epithelial-like tight junctions within the brain capillary endothelia, serves as a physical and biochemical barrier to most molecules from entering the brain. Only a small fraction of small-molecule drugs (i.e. small molecules with the molecular weight below 400–500 Da and with high lipid solubility) can penetrate BBB in a pharmacologically significant amount [17]. Thus, the targetable brain diseases are limited in number. In addition, because of the high lipid solubility, these small molecules can penetrate through all biological membranes which consequently affects target specificity, increases the risk of side effects, and limits therapeutic efficacy. Therefore, various strategies spanning from direct injection to non-invasive BBB disruption [11,18–20], endogenous BBB transporter-mediated delivery [17,21], and transnasal route of administration [22,23] have been explored to overcome or bypass BBB (Fig. 1). In this review, we focus on two systems for drug delivery to the brain that

benefit from microtechnology: transcranial direct brain infusion and BBB disruption using focused ultrasound. 2.1. Direct brain infusion 2.1.1. Design considerations, challenges, and opportunities 2.1.1.1. Convection-enhanced diffusion. Direct brain infusion is invasive but is the most effective means to deliver a known concentration of therapeutic agents to a specific location in the brain. There are mainly three neurosurgical methods for human transcranial brain drug delivery: intracerebral implantation, intracerebroventricular (ICV) infusion, and convection-enhanced diffusion (CED) [24]. There are several factors that need to be considered in choosing the delivery method: desired infusion area, infusion speed, distribution of delivered drugs, area of tissue damage, and sensitivity to drug reflux. While intracerebral implantation and ICV infusion rely on diffusion of drugs from a concentrated source at the implanted sites (e.g. diffusion-based eluting polymer), CED uses external forces to push bulk flow of drugs at constant flow rate into the brain through a hollow cannula or needle [25]. Owing to the convective flow at the infusion site, CED allows for a homogeneous distribution of drugs over a longer distance (~ cm) regardless of the molecular size and the refluxed drugs are within the target volume. The minimum dimension of the cannula is limited by several factors such as fabrication technology and desired flow rate which is a function of viscosity and density of drug. For small animal experiments, CED is also the most dominant delivery method. A commercially available hypodermic needle is typically used to deliver agents such as genes, drugs, and neurotransmitters to a target region of the brain. Hypodermic needles with gauge numbers as small as 30 and 31 are commercially available if specially ordered which have the outer diameter of 311 μm and 260 μm, respectively. However, the size of the mouse brain is approximately 4 mm in depth (in the smallest dimension) and some brain regions of the mouse brain are only a few hundreds of micrometers in one dimension, which requires further miniaturization of the cannulae. Thus, microneedles [26,27] or neural probes with an integrated microfluidic channel [9,28] based on MEMS technology have been proposed for direct brain infusion for small animal experiments (Fig. 1). The first MEMS-based silicon microneedle for brain infusion was reported in 1997 where a boron-doped mask layer was used to fabricate a 15-μmthick 58-μm-wide microneedle [27]. Various techniques to integrate microfluidic channels without increasing the microneedle dimensions,

H.J. Lee et al. / Advanced Drug Delivery Reviews 128 (2018) 132–147

such as buried channel technology [29] and glass cover on silicon technology [26], have been developed to minimize the dimension and thus the damage in the brain tissue. The devices for direct brain infusion have been used for the experiments with animals in the preclinical stage. 2.1.1.2. Biocompatibility and systematic errors. Despite the advances in the fabrication techniques, there are still critical challenges that need to be resolved such as biocompatibility and systematic complications such as clogging, reflux, and leakage. During the insertion of a microneedle, there are initial bleeding, death of neurons along the insertion path, and a microglial response, followed by an astrogliotic reaction during the period of implantation. In addition, due to the brain micromotion and the mechanical mismatch between the implant and the brain, there is a persistent friction. Thus, over the past decade, new materials such as SU-8 and parylene have been exploited as flexible substrates of the microneedles to minimize the chronic inflammation [30,31]. However, microneedles based on these relatively soft materials are often thicker than rigid silicon microneedles to ensure insertion without buckling; thus, initial damage incurred during the insertion still remains a problem. In addition to the biocompatibility, the current systems are also susceptible to systematic errors associated with delivering drugs through a cannula. Some of these errors are more prone for miniaturized cannulae; due to the small dimensions of the microfluidic channels and outlet ports, microneedles are more susceptible to clogging, which often requires much more endeavor for removal, cleaning, and re-insertion, and leads to irreversible deterioration. 2.1.1.3. Perspective for additional functionalities. There are also additional features or functions that are desired to achieve more efficient and powerful drug delivery to the brain, such as active control of drug release, sampling, and delivery of multiple drugs in situ [32,33]. The current CED-based system is more suitable for acute drug delivery because the implanted cannula must be connected to the external fluidic driving

135

system via tubing. Thus, temporal control of drug delivery is restricted. Another problem is the difficulty of predicting the exact dosage delivered to the target site due to dead volume; dead volume arises due to the presence of a reservoir, adsorption of the drugs to channel walls, and leakage at an outlet along the insertion path [26]. Thus, simply monitoring the decrease in the drug volume at an input reservoir does not necessarily indicate actually delivered volume to the target site. Lastly, with the current delivery system, it is challenging to deliver multiple drugs or different concentrations of drugs in situ through single implantation because the tubing must be changed without introducing any air bubbles [32]. These aforementioned problems and additional features associated with direct brain infusion are not only restricted to small animal experiments but also map directly to large primates and humans. For example, demyelination and loss of cells were observed in the primate brain during the autopsy [24], which raised concern about long-term effects of direct brain infusion for humans. Therefore, it is highly required to discuss the challenges and improvements in microneedle technology to apply the direct infusion to therapeutic applications as well as animal experiments. 2.1.2. Recent advances and perspectives 2.1.2.1. Recent advances. Recent works focused on the active control of drug release in the brain. For example, one method of functional brain mapping is to trace responses of injected neuronal tracer such as Mn2 + ions upon electrical stimulation using manganese-enhanced MRI, which requires a timed release of a small volume of Mn2+ synchronized with the electrical stimulation. Recently, Huang, et al. reported a conductive nanogel-based neural interface that achieved both temporal and spatial controlled release of Mn2+ [34]. Mn2+-encapsulated conductive nanostructural networks were coated on a neural probe with multiple gold microelectrodes; by applying biphasic pulse on the

Fig. 2. Microneedles with integrated microfluidic channels for brain infusion: (a) silicon-parylene hybrid structure with openings at the tip for the regrowth of neural tissues to enhance biocompatibility [40], (b1) microneedle integrated with U-shaped channels and nanopores at the tip for microdialysis [36], (b2) push-pull microneedle to sample neurotransmitters [37], (b3) droplet-based sampling microfluidics [38], (c1) fluidic channel integrated with a refillable, multi-drug chamber (scale bar: 5 mm) [33], and (c2) neural probe integrated with microfluidic channel and interfaced with a PDMS chip which consists of 3 input channels, a micromixer, and 1 output channel [32]. All figures reprinted with permission.

136

H.J. Lee et al. / Advanced Drug Delivery Reviews 128 (2018) 132–147

microelectrodes, both the electrical stimulation for activation of neural circuits and the release of Mn2+ for tracing the circuit were achieved. The authors reported an increase in both released concentration of Mn2 + and diffusion volume by adjusting frequencies of the biphasic pulse from 20 Hz to 120 Hz. As a similar application of the neuronal tracing, Fekete, et al. have reported a hybrid drug-delivery system where a neuronal tracer was injected through a microfluidic channel but drug delivery was actuated through iontophoresis rather than by pressure [35]. The iontophoresis, which ejects charged molecules by generating voltage gradients, was achieved by placing an electrode near an outlet of the microfluidic channel and another electrode outside of brain tissue. In contrast to diffusion-based delivery where no current was applied, the authors reported successful delivery of neuronal tracers at various current levels and injection times. In both reports, no neuronal damage due to the application of electrical currents were observed. Since the current delivery systems are mostly limited to delivery of a single drug, there are on-going efforts to design a drug delivery system that allows infusion of multiple drugs. McCall, et al. have reported a system which consisted of four fluidic reservoirs connected to four separate microfluidic channels [33]. Different drugs could be loaded using a blunt syringe into a reservoir (Fig. 2(c1)). In addition, since the drugs were delivered through thermally actuated pumps via the metal electrodes located at the bottom of a reservoir, this system was operational without a bulky tubing and an external driving module. In contrast, Shin, et al. have integrated a microfluidic mixer to accommodate three different drugs [32]. Since three inputs were combined through the mixer that was merged into a single microfluidic channel, this system enabled delivery of drugs at different concentrations in situ at a single implantation (Fig. 2(c2)). 2.1.2.2. Perspectives. There are few additional functionalities such as sampling and enhancement of biocompatibility which would further improve the current drug delivery systems. For example, accurate sampling of fluids around the cannula would be an approach to confirm successful delivery of drugs and monitor a dosage as well as to monitor changes in neurochemical concentrations affected by the delivered drugs or the brain functions. There are largely two methods for sampling cerebrospinal fluid (CSF): microdialysis and push-pull. Microdialysis uses semi-permeable membranes at the target site to deliver drugs and to retrieve fluids; due to the small size of pores (typically sub nm), microdialysis relies only on concentration gradients of the small molecules at the interfacial membrane between brain tissue and a sampling buffer and consequently is diffusion-limited. In contrast, push-pull uses two separate cannulae with an open gap in between. Principle of pushpull perfusion relies on the exchange of constituents between perfusate and brain at the tip; At the same time a perfusate is pumped into the inflow (or push) cannula, the purfusate now mixed with constituents in the brain is immediately withdrawn at the same flow rate through the outflow (or pull) cannula. Thus, push-pull perfusion is applicable for charged and larger molecules and is free of the diffusion limit. Both microdialysis and push-pull microneedles have recently been developed to sample CSF and analyze neurochemicals quantitatively [36,37]. While Lee, et al. have utilized deep reactive ion etching (DRIE) through a porous aluminum oxide layer to fabricate a nanoporous membrane for microdialysis (Fig. 2(b1)), Chae, et al. have fabricated two parallel microfluidic channels for the push-pull operations (Fig. 2(b2)). Recently, Petit-Pierre, et al. have proposed an interesting idea of applying droplet microfluidic technology to the push-pull method for sampling CSF [38]. The CSF that was retrieved through a pull microchannel was immediately segmented in nanoliter droplets by a perfluoromethyldecaline carrier phase (Fig. 2(b3)). Due to this isolation of the retrieved solution immediately after the push-pull operation, the sampled liquid was affected by neither the diffusion through a membrane nor the Taylor dispersion in the sampling tube. Therefore, a high temporal resolution of 170 ms for sampling was reported. The capability to deliver drugs and to acquire samples simultaneously at an identical

site serves as an important function for next generation drug delivery systems because a closed-loop control of drug release which depends on the concentration of neurotransmitters in the brain is now possible. In addition, there are largely two groups of efforts to minimize the immune response of the implanted microneedles: physical and biochemical approaches. Physical efforts involve an adjustment of mechanical structures and properties by using soft materials [28] that have Young's modulus closer to that of the brain (~a few kPa) [39]. Recently, Liu, et al. have modified both the mechanical structure and properties of microneedles to enhance biocompatibility [40]. The body and the sharp tip of the microneedles were composed of silicon which were connected through a 250-μm-long parylene structure with opening gaps (Fig. 2(a)); the silicon-parylene hybrid structure provided a flexible section between the implant and the brain while the openings in the structure allowed for the regrowth of surrounding neural tissue. The authors demonstrated a reduced reactivity of both astrocytes and microglial around the implant. Another approach to enhance biocompatibility is to deliver anti-inflammatory drugs either through a microfluidic channel or active sites along the microneedle [40,41]. The same authors who fabricated the aforementioned hybrid structure also delivered minocycline, which is known to suppress the microglia activation, through the openings and reported a reduced microglial activity. Another recent work by Boehler, et al. designed an actively controlled release system for chronic applications which released an anti-inflammatory drug (i.e. dexamethasone) to actively intervene with the foreign body reaction [41]. The drug was loaded on a layer of a conducting polymer coated selectively on electrodes. By releasing the drug through cyclic voltammetry (CV) on a weekly basis, the authors observed reduced inflammation with a larger neuronal population near the implant over 12 weeks of implantation. Through the 12-week study, the authors have also verified their hypothesis that the anti-inflammatory treatment beyond an initial healing phase of 6 weeks was still effective in enhancing the biocompatibility. 2.2. Drug delivery to the brain through BBB disruption There are increasing efforts towards delivering drugs to the brain via blood stream by overcoming BBB. One promising strategy is to use focused ultrasound (FUS) to transiently disrupt or breach BBB by applying a concentrated acoustical energy to a focal spot in the brain. Upon sonication, the induced BBB disruption lasts for several hours, after which the BBB disruption recovers to the baseline [42]. Thus, when a drug along with an ultrasound contrast agent (i.e. microbubble) is administrated through intervascular injection, the drug circulating in the blood stream enters the brain only at the location where BBB is disrupted by sonication. The clinical need for this technology is expected to be tremendous because not only the FUS technology enables spatial- and temporal-controlled delivery to the brain, but also this technology is non-invasive and applicable to a wide spectrum of drugs including DNAs [43], genes [44], small molecules [45], hydrophilic molecules [46], drug-loaded microbeads [47], and macromolecules [48]. Moreover, early studies on the BBB disruption using ultrasound demonstrated that no apparent damages on surrounding parenchyma were induced [49–51]. In addition to targeting the brain, the FUS technology has been applied for drug delivery to other parts of the body through transdermal [52,53], transcorneal [54], and gastrointestinal [55] routes of administration. 2.2.1. Design considerations, challenges, and opportunities 2.2.1.1. BBB disruption. The effect of ultrasound on BBB has been reported as early as 1956 but the idea of delivering drugs through BBB disruption was not technologically possible. Because ultrasound attenuates greatly in the skull due to its high mechanical impedance, delivering focused acoustic energy through skull was challenging. With the advancement in the ultrasound transducer technology, non-invasive BBB disruption

H.J. Lee et al. / Advanced Drug Delivery Reviews 128 (2018) 132–147

through intact skull became possible. By combining the ultrasound contrast agents with FUS, BBB disruption became more effective and repeatable. Thus, the threshold for BBB disruption was lowered which resulted in less risk of overheating in the skull and higher spatial selectivity. As a result, the FUS-mediated drug delivery is now ready for clinical trials. The first clinical trial for delivering chemotherapy to patients with brain tumors started in 2015 at Sunnybrook Health Sciences Center in Toronto, Canada. For more in-depth historical developments and pre-clinical achievements, readers are referred to the following comprehensive reviews by Vykhodtseva, et al. [19], Arayl, et al. [11], and Dasgupta, et al. [20]. 2.2.1.2. Sonication parameters and microbubble properties. For a specific target drug, there are two sets of parameters that can be adjusted in the FUS-mediated drug delivery: sonication parameters (e.g. frequency, intensity, pulse repetition frequency (PRF), burst length, and duration) and microbubbles properties (e.g. size, concentration, and composition). Over the years, a large range of parameters has been reported for successful disruption of BBB. For example, the reported center frequency and burst length span over several orders of magnitude: 28 kHz–8.9 MHz and a few periods ~ 100 ms, respectively. The threshold for BBB disruption tends to decrease with larger pressure amplitude, lower frequency, longer duration, and larger microbubbles while mixed trends were observed for burst length, PRF, and concentration of the contrast agent [11]. One of the reasons for the difficulty in finding an optimal set of parameters is because the exact mechanism is still unknown; modeling

137

the behavior of microbubbles in an acoustic field is complex, especially in a capillary [20]. Among the known bioeffects induced by FUS, heating and inertial cavitation due to the collapse of microbubbles leading to the rapid opening of BBB (i.e. sonoporation) have been dismissed through experimental studies. No measurable increase in temperature and no wideband acoustic emission due to inertial cavitation were observed. Also, the effects of sonication were not transient but rather lasted for over several hours. Thus, the only remaining known bioeffect seems to be stable cavitation, which can stimulate brain endothelial cells via three potential means: radiation force acting on the bubble, oscillation of microbubbles in the acoustic field, and microstreaming of fluid around the bubbles. However, the difference in the threshold for BBB opening and inertial cavitation has been reported to be small. Due to this small margin, studies have reported the observation of both stable and inertial cavitation which accompanied mild capillary and erythrocyte extravasation. Thus, recently there have been efforts in designing microbubbles that would favor stable cavitation and also in integrating drugs into the microbubble for more efficient delivery of drugs to target regions [56]. 2.2.1.3. Unmet needs. Since the FUS-mediated drug delivery systems are still in the early clinical stage, there is still a large number of challenges and thus opportunities for technological innovations, especially for researchers in the field of microtechnology. Currently, commercially available ultrasound transducers based on piezoelectric materials are used for BBB disruption, which are fixed in terms of frequency, radius of curvature, and aperture [11,20]. Thus, customized or adaptive adjustments

Fig. 3. (a) MR image immediately after BBB disruption by pulsed ultrasound using the implanted ultrasound transducer (top) with a contrast-enhanced image (bottom) [57]. (b) the 16element pMUT array with a resonant frequency of 781 kHz developed for in vitro neuromodulation [63]. All figures reprinted with permission.

138

H.J. Lee et al. / Advanced Drug Delivery Reviews 128 (2018) 132–147

of the level of disruption for different patients and different target drugs are currently not possible, which limits full exploration of large parameter space of sonication. Since there is a natural trade-off between the frequency, attenuation, and focal spot size, operators are often limited to one frequency with a fixed focal spot size [47,48]. For example, for a low-frequency transducer, the focal spot size might not be spatially specific enough. In addition, the commercially available transducers are suitable for human cases but are often too bulky for animal experiments. Because use of ultrasound for drug delivery is relatively new, there is still an increasing need for miniaturized transducers for small animal experiments [57–59]. For example, pre-clinical screening of newly developed drugs using animal models are still required and the safety of long-term exposure still needs to be validated. Another important issue in the FUS-mediated drug delivery is that the ultrasound beam inside the brain cannot be visualized and thus requires usage of an imaging system such as magnetic resonance imaging (MRI) [46]. However, since drug administration would often require multiple sessions and MRI is bulky and costly, a better solution is to develop an MRI-free system that still accurately predicts the location of the sonication. Lastly, the exact mechanism of BBB disruption by ultrasound should be continuously researched to provide insights into the optimal ultrasound parameters and microbubble properties. 2.2.2. Recent advances and perspectives Medical ultrasound imaging is the major application for the ultrasound transducers which require both transmit and receive modes. Although small in number, recently, there have been early research works on developing miniaturized ultrasound transducer arrays using MEMS technology for sonication only. Another impetus for these efforts is applications of FUS technology in non-invasive neuromodulation [60,61] and cell ablation using high-intensity FUS (HIFU) [62]. Recently, Goldwirt, et al. have demonstrated a successful drug delivery in a primate model where a custom-built implantable air-backed 1 MHz transducer was used [57] (Fig. 3(a)). Although the transducer was less bulky than the commercially available transducers with a size of 10-mm in diameter, the transducer was a flat single element without any mechanical focusing which resulted in unfocused ultrasound with a large focal spot. Another work on devising a customized ultrasound transducer array was the development of a light-weight PZT-based transducer array designed for portable transdermal drug delivery systems [58,59]. A 12.7-mm-diameter PZT-4 disk was encapsulated in cymbal-shaped capping to amplify the displacement by a factor of 40, which was packed in a 6-cm-wide 3-by-3 array. Another interesting application of bioMEMS technology for FUS-mediated brain sonication is in vitro investigation of the effects of ultrasound. Recently, Ko, et al. have reported an in vitro system that resolved both of these problems [63]. A 781-kHz piezoelectric micromachined ultrasonic transducer (pMUT) array was custom-designed and fabricated, which contained 16 elements placed in a linear fashion. By directly placing cultured cells on the backside of the pMUT array, sonication of the cultured cells at multiple locations at the same time was possible (Fig. 3(b)). This MEMS-based in vitro system could provide a means to investigate the effects of ultrasound in cell levels. This work is an example of excellent interdisciplinary research where the application of microtechnology has advanced the overall system to a great extent. Because the FUS-mediated drug delivery system is still at the clinical stage, there is a large room for technological advancements in terms of transducers and interfacial circuits. For example, capacitive micromachined ultrasound transducer (CMUT) arrays which has an advantage of design freedom in terms of frequency and array shapes are excellent candidates for future ultrasound-mediated drug delivery systems [64]. The integration of these custom-built transducers with lowpower, compact integrated circuits (ICs) and an advanced control system for locating the target region could one day allow a fully portable

drug-delivery system. Patients can receive sonication for drug delivery at home simply by wearing a sonication headset and swallowing a pill. Thus, with miniaturization of the sonication system, we can anticipate a completely non-invasive wearable system that will allow on-demand drug delivery. 3. Ocular drug delivery The eye is an interesting yet complex organ for drug delivery; the anterior segment is exposed and thus accessible externally through a thin layer of the cornea while the posterior segment interfaces with the internal vascular network through inner and outer blood-retinal barriers (Fig. 1). Possible routes of administration for intraocular drug delivery include topical, systemic, intravitreal, and periocular routes; the topical route using eye drops and ointments is currently the most common for treating ocular diseases in the anterior eye segment. However, because of a continuous turnover of tears, clearance through nasolacrimal ducts, and low permeability of the corneal layer, less than 5% of a dose is absorbed into the eye [65]. Thus, the frequency of instilment or the concentration of drugs must be increased to reach the therapeutically effective dosage. Thus, despite the accessibility, targeting the anterior segment is still challenging. In addition, it is even harder to target the posterior segment because of the physiological blood-retinal barrier and anatomical inaccessibility and most ocular diseases that lead to visual impairments occur in the posterior segment. Therefore, implantable drug delivery systems are attractive alternatives in targeting diseases in the posterior segment. Intraocular implants bypass or penetrate the barriers to increase the efficiency of drug delivery and there are a large number of potential implantable sites: vitreous space, pars plana, peribulbar, and intrascleral space. Implantable intraocular drug delivery systems are especially useful for ocular diseases such as cytomegalovirus (CMV) retinitis and proliferative vitreoretinopathy (PVR) which require repeated drug injections and surgeries. In addition, these systems are suitable for treatments of chronic diseases with no other effective treatments, such as age-related macular degeneration (AMD), macular edema, and retinitis pigmentosa [66]. Lastly, through localized delivery, implantable intraocular drug delivery systems allow administration of a large spectrum of drugs including potent antiangiogenic agents and biomolecules. For general details on ocular drug delivery, readers are referred to [66–72]. These implantable systems have been tested with animals in the preclinical stage. 3.1. Design considerations, challenges and opportunities Since the introduction of the first polymeric insert developed for sustained drug release in the 1800s, a large number of polymeric implants such as vitreal implants, intrascleral discs, and scleral plugs was developed and commercialized [66]. For example, Vitrasert (Bausch & Lomb, USA) was developed in the 1990s to treat opportunistic viral retinitis in human immunodeficiency virus infection and acquired immune deficiency syndrome (HIV/AIDS) patients over 8 months. While the surgery for these implants is not complex, there is always a risk of complication accompanied with the surgery. Thus, soft contact lenses that satisfy both medication needs and refraction correction were explored as an alternative non-invasive means. Various methods of incorporating drugs into contact lenses without compromising transparency were explored including soak and absorption, surface immobilization of drugs, and molecular imprinting [73]. Bandage contact lenses that deliver antibiotic and anti-inflammatory medication have been commercialized for effective managements of ocular trauma and post-surgery conditions. Few examples include Pure Vision (balafilcon A, Bausch + Lomb), Acuvue 2 (etafilcon A, Vistakon Inc.), Acuvue Oasys (senofilcon A, Vistakon Inc.), and Air Optix Night & Day (lotrafilcon A, Alcon) [73]. However, due to the concerns over diffusion of preservative (e.g. benzalkonium chloride) on the ocular surface, pharmacokinetics and stability of the drug, and safety issues on the cornea, commercialization

H.J. Lee et al. / Advanced Drug Delivery Reviews 128 (2018) 132–147

139

Fig. 4. Novel intraocular drug delivery systems: (a1) individually switchable two-channel microneedle arrays [79], (a2) microneedle array that releases drug as it dissolves [80], (b1) MEMS implantable reservoir and cannula with a check valve [82], (b2) multi-drug polymeric transscleral inserts developed for non-invasive ocular drug delivery [88], (c1) selfpowered drug delivery systems using a triboelectric nanogenerator [89], and (c2) magnetically actuated drug-delivery system [90]. All figures reprinted with permission.

of drug-loaded contact lenses were not as successful as initially anticipated. The ultimate objective of intraocular drug delivery is to deliver drugs more than two years with improved ocular drug bioavailability, minimal complications, and maximal patient compliance [74]. These unmet needs can be fulfilled by microtechnology which enables miniaturization of biocompatible materials and integration of active control components. Thus, microtechnology-based intraocular drug delivery systems, such as minimally invasive microneedle systems, MEMSbased drug delivery systems with on-demand drug release, and advanced bioresponsive polymeric devices for self-regulated drug delivery, are actively researched. Design factors for intraocular drug delivery systems include volume of the implants, site of implantation, materials (biodegradability), and actuation methods (passive or active), and reliability. As for any implantable systems, there is a trade-off between the amount of a drug that can be loaded and size of the implants. The size of the implant is typically designed to be less than a few millimeters in order to enhance patient comfort level and minimize the complexity of surgery. However, the volume of loadable drug also decreases. Since the therapeutic effective drug concentration is fixed, the frequency of replacements or refills of the reservoir would increase. However, this implies an increase in the implant size and therefore reduces patient compliance. The material of the implant is also an important parameter which affects release characteristics. For instance, degradable polymeric devices that release a drug through dissolution do not require additional surgical removal, but suffer from an initial burst of the drug, which results in an overdose of the drug in the eye over a short period of time. Biocompatibility and reliability are categorized into another important set of problems; complications such as vitreous hemorrhage and rhegmatogenous retinal detachment were observed in 13 out of 110 eyes after implantation [75]. Thus, the site of implantation is another design factor as the degree of invasiveness and the associated risks of complications differ.

3.2. Recent advances 3.2.1. Minimally invasive microneedles Because of minimally invasive nature of microneedles that penetrate through only a thin layer of physical barrier, various microneedles such as drug-coated microneedles and hollow microneedles were fabricated and tested for intrascleral and suprachoroidal drug delivery [76–78]. Recently, Valdés-Ramírez, et al. reported a multiplexed, switchable microneedle-based intraocular drug delivery system [79]; the system consisted of an actuator composed of a conductive polymer, a PDMS drug reservoir, and an electrode array to control actuation. Application of suitable redox potentials induced a reversible volume change of the conductive polymer which led to expulsion of drugs from the reservoir. Individually addressable two channels allowed delivery of multiple drugs at single implantation (Fig. 4(a1)). However, the system required a separate reservoir and a delivery cannula for each drug and thus fluidic interfaces were not optimal for multiplexed delivery of a larger number of drugs. There is also a movement towards fabricating a dissolvable microneedle array. Thakur, et al. developed a dissolving polymeric microneedle based on a polyvinylpyrrolidone (PVP)/hydrogel formulation [80] (Fig. 4(a2)). Drug-loaded PVP rapidly dissolved in biological tissues due to its high water solubility. In addition to the fact that no additional surgery is required for the dissolvable system, dissolvable polymeric microneedle systems offer an advantage of enhanced safety; a broken small part of the rigid microneedles during implantation could lead to a serious complication.

3.2.2. Stimuli-responsive implants for on-demand drug delivery Intraocular drug delivery systems with active control enable on-demand drug delivery which is controlled internally through either physiological signals or external stimuli, such as temperature, electrical field, magnetic field, and osmotic pressure [15]. The first on-demand MEMSbased intraocular drug delivery system was reported by Lo, et al. which

140

H.J. Lee et al. / Advanced Drug Delivery Reviews 128 (2018) 132–147

consisted of a refillable PDMS reservoir, suture tabs, and cannula with a one-way check valve [81]. The reliability of the refillable PDMS reservoir was limited to 12 times after which a leakage was observed through an injection port. The same authors demonstrated the chronic usage of this system where no blockage of the cannula due to biofouling was observed for over 6 months after implantation [82] (Fig. 4(b1)). However, because the drug delivery was actuated manually, this system suffered from imprecise control over infused volumes. This problem was overcome by devising a mini drug pump that was actuated by hydrolysis [83]. With low power consumption and precise control over the infusion volume, the proposed intraocular drug delivery system was more practical for clinical applications. Now, there is also a commercially available minipump (MicroPump™ System, Replenish) which is designed for clinical use with a lifetime of up to five years. This system contained a fluidic flow sensor, bi-directional telemetry for wireless programming, and a refillable reservoir using a proprietary 31-guage needle [84]. Zhang, et al. demonstrated a flexible ocular iontophoretic drug delivery device that was implanted under an eyelid [85]. Placement of this device under the eyelid offered several advantages including simplified surgery and stable, conformal contact to the target region using a planar electrode. Moreover, the device consisted of a simple PDMS cup with poly-3,4-ehtylenedioxy-thiophene (PEDOT) electrodes for actuation which was simple and batch-fabricated. Since iontophoresis can increase the temperature near the electrodes, thermal effects were also monitored in vivo where 37.8 °C was determined by the authors as the maximum safe temperature. However, this system did not contain drugs and thus required an external application of a liquid drug; once the drug was administrated under the eyelid, iontophoresis drove the drugs into vitreous space. Furthermore, because of the wiring to provide an electrical interface to the PEDOT electrodes, this system was not completely implantable. Another interesting area for ocular drug delivery where the MEMS technology could contribute is polymeric implants composed of various biocompatible materials such as hydrogel. For more details, readers are referred to the following recent work [86–88]. 3.2.3. Batteryless actuation If there are no reliability issues in the implant, lifetime of the implant is limited by either the total volume of drugs or capacity of small size batteries. Implantable batteries not only increase the footprint of the implant but also require frequent removal of the implant due to the limited energy capacity. There are currently two research fronts that aim to resolve this battery problem: wireless power transmission (or actuation) and energy harvesting. Energy harvesting from physiological, mechanical, electrical and thermal processes in the human body is possible but has not been applied for drug delivery systems. Song, et al. developed a triboelectric nanogenerator which harvested energy from human body movements and interfaced the nanogenerator to an electrochemical pressure pump embedded in an implantable intraocular drug delivery system (Fig. 4(c1)) [89]. A PDMS microtube was implanted at the sclera and allowed for successful delivery of 50 μL of microparticles to the anterior segment of porcine eyes ex vivo. However, because the self-powered triboelectric nanogenerator must be attached to the body with largest movements and because it does not generate high enough power for wireless transmission, this unit was connected to the implant through long wiring. For clinical applications, a more practical solution for the connection should be proposed. Pirmoradi, et al. demonstrated an intraocular drug delivery system without a battery by actuating a pump magnetically [90]. A PDMS chamber was fabricated with a top membrane containing iron oxide nanoparticles with a 100 × 100 μm2 aperture for drug release (Fig. 4(c2)). A 64-fold increase in the infused volume was observed in the presence of 213 mT magnetic field when compared with the spontaneous release through the aperture without the magnetic actuation. The limitation of this system is rapid depletion of the magnetic response of the small magnets and thus chronic functionality must be further

investigated. In addition, magnetic actuation is not compatible with two major medical imaging modalities: computerized tomography (CT) and MRI. 4. Gastrointestinal tract drug delivery – capsule endoscopy Unlike other parts of the body that require surgery for implantation, drug delivery to a specific location of the gastrointestinal (GI) tract can be achieved non-invasively by oral intake of drugs. However, gastrointestinal absorption along the GI tract depends greatly on many physiological, pathological, and pharmacological factors, such as regions, gastric emptying time, intestinal motility, and drug formulations [91, 92]. Thus, targeted and efficient drug delivery using passive oral medication is challenging. Another possible means to deliver drugs in the GI tract is through the use of traditional endoscopes. However, the endoscope that enters either through the oral cavity or the anus does not provide access to the full GI tract. In addition, the traditional endoscope has low patient compliance. In contrary, the swallowable capsule endoscope is highly patient compliant and provides access to the entire GI tract [93]. With the first introduction of a commercially available capsule endoscope in 2001, three small bowel capsules and one esophageal capsule have been approved by U.S. Food and Drug Administration (FDA) while one colonic capsule is now available in Europe and Japan [10]. By integrating the drug delivery function to the capsule endoscope, spatial-, temporal-, and dosage-controlled drug delivery in the GI tract can be achieved. This targeted on-demand drug release not only enables direct treatments of GI diseases but also provides an efficient route of administration by targeting the regions with the highest drug absorption. For example, the small intestine is an attractive site for drug administration because of its high drug absorption ascribed to large surface area. Moreover, drug delivery to different regions of the GI tract is an important field of study for the pharmaceutical industry because understanding the different absorption rates across the GI tract provides useful information in determining the dosage form of a drug for sustained release [94]. Some of the passive devices have been utilized in human studies for colon targeting [95]. Since the intubation process can disturb the normal physiological function of the GI tract, the swallowable capsules can be utilized as an evaluation tool to study absorption characteristics for drug developments. In addition to the capsule endoscope, there is an increasing number of works on developing GI patches which are completely passive systems that rely on various matrices of polymeric layers to achieve timed release in the GI tract. For details on the GI patches, the readers are referred to a comprehensive review on this topic [96]. 4.1. Design considerations, challenges, and opportunities 4.1.1. Anchoring mechanisms The environment of the GI tract has been a major limitation factor in the design of the capsule endoscopes. Firstly, the small capsule size, which is typically ~ 11 mm in diameter and ~ 26 mm in length with a volume of 3 cm3, is determined by the minimum diameter of the GI tract (2–3 cm) and comfortable swallowable size. The second design factor is the control mechanism of the capsule movement: passive or active. Without active control, the capsule moves by the natural peristaltic movement of the GI tract and thus does not require additional space in the capsule for an on-chip battery. However, accurate control of the capsule position over time is not plausible with the passive movement and thus certain areas could be overlooked during diagnostic examinations. Therefore, the active control of the capsule movement back and forth along the GI tract has been recently explored. As capsule endoscope technology has reached its mature stage of development, recent research has focused on integrating additional functionality to the capsule endoscope, such as biopsy, surgical interventions, and drug delivery [10]. The drug delivery functionality

H.J. Lee et al. / Advanced Drug Delivery Reviews 128 (2018) 132–147

141

Fig. 5. Micromotion and anchoring mechanisms for capsule endoscope: (a) microrobot based on legging mechanism where the bow-shaped legs extend largely and resist peristaltic pressure along the GI tract [108], (b) schematic of inchworm-like locomotion that enables capsule endoscope to crawl along the GI tract [109], and (c) wheel-based movement [107]. All figures reprinted with permission.

requires additional considerations in terms of a capsule design such as anchoring mechanism, volume of a loaded drug, and release parameters including release rates and the frequency of dosage. Anchoring is often achieved through the thin mechanical legs that protrude from the capsule. The anchoring allows not only the navigation of the capsule against the natural peristaltic movement of the GI tract but also localized delivery. The delicate walls of the GI tract, the different sizes of organs along the GI tract (i.e. stomach vs. small intestine), and limited size of the capsule, however, impose challenges in the leg design such as length, shape, and actuation mechanisms of the legs. 4.1.2. Trade-off between drug volume and on-chip functionalities Therapeutically effective drug dosage and the formulation of drugs (powder vs. liquids) are another important design factors in the capsule endoscope because of the limited capsule size. Due to the electronic components such as battery, antenna, and integrated circuits, typically only 20–30% of the volume is available for drug loading. The storage size and formulation will determine the number of dosage and capability to provide multi-site delivery. The volumetric ratio between a drug reservoir and the capsule (Rdc) is thus an important parameter that imposes trade-off in the capsule design. Another design consideration is the actuation mechanism of drug release. Since there are many factors that affect rates of GI absorption such as gastric emptying rate, formulation of drugs, intestinal motility and metabolism [91], full control over the dosage, time of delivery, total amount of drug, and release rates is often desired. The actuation of the drug release can be implemented in an either passive or active manner. Passive drug release driven by diffusion is prone to leakage, retention of drugs, and low release rates. In addition, due to a lack of an external driving force, the passive drug release relies largely on the dynamic environmental conditions at target locations such as fluid availability. However, with the passive system, a high Rdc (~35%) can be achieved due to a small number of active circuit components. Conversely, the active system provides a wide range of control over the drug release but suffers from low Rdc. The active release system often utilizes mechanical actuation by pistons and stretchable films to expel the drug. To overcome the low Rdc issue, other activation mechanisms such as wireless power transfer [97,98] and non-mechanical actuation such as

chemically triggered release [99] and magnetic actuation through external magnets [100,101] have been reported. 4.2. Recent advances 4.2.1. Legged mechanism MEMS technology plays a pivotal role in capsule endoscope integrated with the drug delivery capability, especially in the propulsion and anchoring functions. Various mechanical systems have been proposed that achieve efficient and safe propulsion and anchoring through the usage of multiple legs [102–104], inchworm-like locomotion [105,106], and micropatterned wheels [107]. Followed by the first work in 2004 on the use of mini-legs actuated by shape memory alloy (SMA) actuators [104], Quirini, et al. demonstrated a 40-mm-length four-legged system which allowed propulsion of the capsule through the GI tract [103]. By designing the anchors with a C-shape, damage of the anchors to the delicate wall was minimized. However, this system suffered from three limitations: high power consumption due to the usage of a micromotor, relatively short length of the legs (40 mm), and potential collision of the capsule into the wall. To minimize the impact of the collision, Valdastri, et al. proposed a system with twelve legs which were evenly distributed along the capsule body [102]. These legs not only provided higher control over the propulsion but also prevented the capsule from impacting the wall of the GI tract. However, the length of the legs was still limited to 30–35 mm since it was challenging to construct a long thin leg with high stability and controllability. To overcome this limitation, Woods, et al. have proposed bow-shaped legs that could extend as far as 71.25 mm stably through an extra support near the capsule body [108] (Fig. 5(a)). 4.2.2. Other anchoring mechanisms Similar to the legged mechanism, earth-worm like locomotive mechanism has also been explored using anchor-and-pull motion based on an SMA actuator [105] and a piezo actuator [106]. Chen, et al. have recently demonstrated a capsule that combined both the legged and locomotion mechanisms; the earth-worm like locomotion was used for propulsion and the eight spiral-legs were used for anchoring [109] to benefit from the both systems (Fig. 5(b)). Another proposed

142

H.J. Lee et al. / Advanced Drug Delivery Reviews 128 (2018) 132–147

mechanism that provided movements along the GI tract was based on contact locomotion through the use of wheels to produce tank-like motion [107] (Fig. 5(c)). By patterning a few-hundreds-μm-wide PDMS micro-treads with an aspect ratio of 1:2 on the surface of the wheel, sufficient traction was achieved on various types of tissues such as the stomach and liver tissues in vivo. In addition, active drug release also requires a means to expel the drugs from a reservoir in the capsule. MEMS technology has been also utilized in pushing drugs by microthrust [110], piezoactuation [111], pneumatic control [112], and stretchable piston membrane [113]. 4.2.3. Remaining problems Unfortunately, because of the mechanical complexity and high power requirement, none of the reported capsule endoscopes equipped with the mechanical systems are in the clinical stage. Nevertheless, there have been continuous efforts towards minimizing the volume required for anchoring and release activation by moving the control unit outside the capsule such as magnetic actuation through the use of external magnets [100,114]. In addition, developments of a thin-film battery [115] for the implant and exploration of energy harvesting strategies [116] provide a good prospect for drug delivery using capsule endoscopes. Moreover, components to actuate the drug release such as microheaters and mechanical parts could benefit from MEMS technology [93]. 5. Transdermal drug delivery Drug administration through the skin is an attractive alternative to oral delivery and hypodermic injection because of several distinctive advantages. Firstly, a transdermally delivered drug distributes systemically through rich vascular networks in the dermis and is not subject to the first-pass metabolism in the liver as in the case of oral delivery. In addition, transdermal drug delivery allows long-term drug administration through the use of biocompatible patches. Nicotine patches and analgesia-loaded microneedles are a few examples where administration over an extended period of time (e.g. at least one day) is preferred. Intravenous (IV) injection can deliver larger volumes of drugs but the mobility of patients is heavily restricted in comparison to transdermal patches. Among the novel drug delivery systems, transdermal drug delivery is perhaps most widely researched, developed, and accepted due to low patient compliance. Not only several transdermal drugs have received the US FDA approval [117], but also a number of transdermal delivery systems are available in the market [118]. From a historical perspective, transdermal drug delivery had evolved over three generations of developments with hallmark achievements in each generation: 1) selection of transdermal drugs, 2) enhancements of skin permeability and transdermal transport, and 3) direct targeting of stratum corneum

using microneedles, thermal ablation, microdermabrasion, and ultrasound [117]. 5.1. Challenges and opportunities Although the history of microneedle technology is relatively long compared with that of other novel drug delivery systems, there are still unresolved challenges and desired features that require technological innovations. Transdermal microneedles are designed to penetrate stratum corneum but not deeply enough to reach nerve endings. Thus, secure positioning of the microneedles on the skin is important especially regarding drug administration over an extended period of time. Therefore, there is still a continuous effort to optimize structures of microneedles such as tip shape, length, and spacing. For instance, Seong, et al. have recently reported a bullet-shape microneedle array with water-swellable tips; upon insertion, the tips swell due to an uptake of biofluid and provide a strong mechanical interlock to the neighboring soft tissues [119]. Completely dissolvable microneedle arrays serve as another important area of interest in the transdermal drug delivery especially for vaccine delivery [120,121]. With the advancement in the polymeric fabrication and biotechnology, delivery of a wide range of drug candidates including proteins, DNAs, and virus-based vaccines is now also actively investigated [30,122]. During the early stage of development, the infused drugs by microneedles were diffusion-limited because of passive drug release. In order to enhance drug penetration and release rates, external forces were applied such as electric field, ultrasound, and mechanical forces. These external forces can also be used as external stimuli for on-demand drug release [123]. If the stimuli are internally-generated physiological signals, a closed-loop or self-regulated drug delivery can be achieved; activation of drug release is determined by changes in the physiological signals through the use of bioresponsive vesicles [124]. This smart closed-loop approach would become useful for diabetes patients as well as patients who require on-demand administration of analgesia over an extended period of time. Possible internal physiological signals include temperature, pH, glucose level, and enzyme concentrations. Design considerations for these systems include response rates of the closed-loop system, sensitivity to physiological signals, resolutions of the detectable signals, on/off ratios, local and systemic side effects due to biocompatibility, and the volume of drugs required to match long-term dosage of specific therapeutics. 5.2. Stimuli-responsive microneedles There is an increasing number of works on developing stimuli-responsive microneedles that apply either internal or external stimuli to achieve on-demand and self-regulatory drug delivery. Teodorescu, et al. have recently developed a flexible transdermal patch that was

Fig. 6. Schematics of on-demand and self-regulated drug delivery using smart microneedles: (a) light-triggered drug delivery system using near-infrared (NIR) light [8]and (b) closed-loop insulin injection according to the glucose level in the intravascular network [130]. All figures reprinted with permission.

H.J. Lee et al. / Advanced Drug Delivery Reviews 128 (2018) 132–147

143

Fig. 7. Subcutaneous implantable drug delivery systems: (a) photo of encapsulated electrolysis pump [135], (b) photo of biological responses to the implanted device over 28 days [143], (c) MEMS mechanical components such as driving gears and rotors fabricated using soft hydrogel materials [144], and (d) a biocompatible implant to deliver in vivo antibody delivery (scale bar: 750 μm) [142]. All figures reprinted with permission.

photo-thermally triggered; by applying continuous-wave near-infrared (NIR), an increase in temperature was induced which release drugs loaded in microneedles through dissolution [125]. Photo-controlled delivery has an advantage of high spatial and temporal resolutions. However, thermal damage to the skin must be carefully designed. Thus, the authors conducted immunohistology of skin tissues at the implant site to empirically determine the maximum light intensity allowed for drug release. Another work by Chen, et al. reported a patient-controlled transdermal analgesia system which was also triggered by NIR [8] (Fig. 6(a)). Exposure to NIR resulted in an increase in temperature up to 50 °C at which lidocaine-loaded microneedles started dissolving. Programmed drug release was demonstrated by controlling the on/off cycles of the NIR laser. Gulati, et al. also implemented a switchable nicotine patch which used a carbon nanotube (CNT) membrane as a switching layer [126]. The permeability of the CNT membrane was changed when DC bias (~600 mV) was applied. However, because the system relied on diffusion, the on/off ratio was low and the response time was in an order of hours. Di, et al. have also proposed a wearable, tensile strain-triggered drug delivery system which consisted of a stretchable elastomer and microgel depots containing drug-loaded nanoparticles [127]. In this stretch-triggered system, a tensile stress activated the drug release by increasing the surface area for diffusion. Self-regulated drug release delivers therapeutic agents based on the physiological signals sensed through biofunctional materials of the microneedle platform itself. The idea of closed-loop drug release was first demonstrated by Huang, et al. [128] where real-time measurements of glucose level in a PDMS microfluidic chip controlled the release of insulin. While this closed-loop system used separate sensing and drug release units which were connected through an external circuitry, sensing and actuation units can be integrated in a single microneedle tip without any external circuitry by using bioresponsive vesicles. For example, in recently proposed smart insulin delivery systems for diabetic patients, a change in glucose level within vascular and lymph capillaries triggered the release of insulin by use of glucose-responsive vesicles [129], H2O2-sensitive polymersomes [130], and antibody-loaded pH-sensitive dextran nanoparticles [131]. Hu, et

al. designed H2O2-responsive vesicles that facilitated oxidation of glucose to gluconic acid and generated H2O2 which in turn dissociated vesicles to allow for the release of insulin (Fig. 6(b)) [130]. Although the released amount of insulin was not constant at each pulsation, the response time was fast enough for the insulin concentration to follow a 10-min-period cyclic change in glucose concentrations. Since the closed-loop systems are in the early stage of development, there are many areas for innovation in terms of increasing sensitivity to physiological signals, improving repeatability of released drug dosage, and maximizing the capacity to load therapeutically effective dosage enough for long-term administration.

6. Implantable subcutaneous drug delivery Microfabrication plays an essential role in implementing microdevices for transdermal drug delivery and subcutaneous implants, such as microneedles, micropumps, reservoirs, and encapsulation [132]. While transdermal drug delivery is minimally invasive, self-administrable, inexpensive, and highly patient compliant, subcutaneous implants are invasive with low patient compliance but provide long-term administration in an order of years. Because of the difference in the nature of invasiveness, these two systems face different technological challenges and thus are discussed separately in this review. Although implantation under the skin is subject to the issue of patients' compliance and acceptance, there is still a need for implantable subcutaneous drug delivery to administrate drugs over an extended period of time such as insulin injection and cancer therapy. Long-term administration reduces the need for frequent needle injections and thus dramatically improves the long-term patient comfort of patients. The first report on the multiwell silicon-based drug-release device in 1999 [133] was followed by another diffusion-based system, a leuprolide acetate implant, which was evaluated over a period of 12 months by Alza Corporation [134]. Although small in number, there has been a steady number of reports on the subcutaneous drug delivery systems [135, 136] (Fig. 7(a)). For details on the individual components of the

144

H.J. Lee et al. / Advanced Drug Delivery Reviews 128 (2018) 132–147

implantable systems such as micropumps and reservoir designs, readers are referred to [12,132]. 6.1. Design considerations Similar to the design of implants for the posterior segments of eyes and the capsule endoscopes, the volume of the implant is the most important design criterion for subcutaneous implants. The size of the implant is directly related to patients' compliance and comfort, complexity of surgery, capacity for drug loading, and available areas for active circuit components. The volume of loaded drug is an imperative design parameter since it determines a maximum number of deliverable dosage and thus the duration of implantation. Implementation of refillable reservoirs is another option to increase the number of dosages but is potentially subject to clogging and infection. Similar to the capsule endoscope and eye implants, active actuation of drug release is possible through magnetic, piezoelectric pumping, and electrothermal activation [136]. In addition, diffusion-based passive drug release was also explored in terms of various regulating means such as membranes, polymer matrices, microchannel implants, and osmotic pumps. Moreover, there is a continuous effort in applying wireless technology to minimize the active components in the implant to secure space for drug storage [137–139]. Similar to the drug delivery implants for the posterior segment of eyes, long-term biocompatibility and robustness are the major challenges for subcutaneous implants. Studies conducted by van Dijk, et al. on complications of implantable pumps for intraperitoneal insulin infusion showed 10–17% of people reported haematoma, cutaneous erosion, pain or infection at surgical sites [140]. Moreover, 5–17% of patients reported the need for replacements due to technical failure of the implantable pumps. A recent study on the implantable pump for subcutaneous insulin delivery in healthy cats showed that one out of nine implantable pumps malfunctioned immediately after the implantation and only 50% of pumps delivered the correct amount of drugs during the four consecutive boluses [141]. Faulty parts in an implant imply surgical removal and thus it is critical to ensure biocompatibility of the implants and to improve the reliability of the driving components in the subcutaneous implants. 6.2. Recent advances 6.2.1. Biocompatibility and reliability There are continuously evolving reports on new implantable subcutaneous microdevices [142–144] with movements towards wireless actuation [137–139]. Liu, et al. devised a refillable MEMS drug delivery device which consisted of electrodes for actuation, a drug reservoir, and a cannula [143]. Metal electrodes for the actuation were fabricated on a silicon wafer while the 10-mm-long, 10-mm-wide, and 2-mmthick drug reservoir was fabricated using an SU-8 photoresist. Successful delivery of 50 μL of adrenaline formulation was observed through an increase in blood pressure after an injection in six mice, respectively. The long-term biocompatibility associated with drug delivery was investigated on ten microdevices through in vivo mice experiments. After 28 days of implantation, the devices were fully encapsulated with fibrous tissue (Fig. 7(b)). Chin, et al. have recently demonstrated an implantable MEMS (iMEMS) drug delivery system that ensured biocompatibility and improved reliability [144]. By fabricating a set of versatile three-dimensional multilayers for biocompatible hydrogel devices, multilayered moving parts such as multi-reservoir single gear, gate valves, and toothed rotors were all integrated in hydrogels (Fig. 7(c)). Poly(ethylene glycol) (PEG)-based hydrogel used in this work offered several distinctive advantages such as biocompatibility, a non-fouling property, and flexibility for conformal contact and less damage. The versatile fabrication also enabled embedment of iron nanoparticles in precise locations to achieve remote magnetic actuation of the mechanical

components. Authors demonstrated successful in vivo delivery of doxorubicin in mice for over 10 days. The presented additive manufacturing of biocompatible materials is a promising enabling microtechnology for the next-generation subcutaneous drug delivery systems. 6.2.2. Perspective In order to accommodate a vast range of drug candidates such as macromolecules, genes, and vaccines, there is a trend towards developing a system that can accommodate drugs with low bioavailability. Recently, Lathuiliere, et al. have developed a bioactive cellular implant that delivered recombinant anti-amyloid-β antibodies in the subcutaneous tissue as immunotherapy against Alzheimer's disease [142]. The implantation of genetically engineered cells allowed for continuous production of monoclonal antibodies, and the implanted cells secreted sustained level of the therapeutic antibodies in the plasma (Fig. 7(d)). Long-term survival of myogenic cells in the encapsulated subcutaneous implant for over 10 months was demonstrated in two mouse models of Alzheimer's disease. Combining this encapsulated cell technology with novel drug delivery systems is a promising alternative technology for treatments of diseases with no satisfactory therapies. 7. Conclusions This review provides a comprehensive outlook of current challenges in novel drug delivery systems that are enabled by microtechnology and a summary of recent advances in drug delivery systems that have attempted to overcome these challenges. As each organ imposes different anatomical and physiological challenges, these challenges and recent advancements have been discussed in terms of target areas (Table 1). Through miniaturization, integration of multiple functions, and electromechanical actuation, MEMS technology has served as the emerging technology to achieve spatiotemporal- and dosage-controlled drug delivery for a wide range of target areas. Thin microneedles or neural probes have enabled localized drug delivery to various regions of the brain with a high spatial resolution while microneedle arrays have been widely used for intraocular and transdermal drug delivery. Implantable systems that provide on-demand or self-regulated drug delivery are now possible through the use of internal and external stimuli for intraocular, gastrointestinal, and subcutaneous drug delivery. For implantable systems, gaining patients' compliance and comfort through implementation of highly robust system is the most challenging hurdle. Regardless of the areas, there are common issues associated with the implant devices, such as total volume of drug loaded in a reservoir, frequency of reimplantation, passive or active activation and power transfer. With further technological innovations in each of the drug delivery systems and matching the appropriate techniques with a wide spectrum of novel drug candidates, the prospect for novel drug delivery systems enabled by MEMS or microtechnology is apparently high for clinical translation. Acknowledgements This work was supported through the National Research Foundation of Korea (NRF) funded by the Ministry of Science and ICT (NRF2016R1A2B2008691). This work was supported by the Brain Research Program through the National Research Foundation of Korea (NRF) funded by Ministry of Science and ICT (NRF-2017M3C7A1028854) and (NRF-2016M3C7A1904343). References [1] A.C. Hunter, S.M. Moghimi, Smart polymers in drug delivery: a biological perspective, Polym. Chem. 8 (2017) 41–51. [2] B. Mishra, B.B. Patel, S. Tiwari, Colloidal nanocarriers: a review on formulation technology, types and applications toward targeted drug delivery, Nanomedicine 6 (2010) 9–24.

H.J. Lee et al. / Advanced Drug Delivery Reviews 128 (2018) 132–147 [3] H. Gensler, R. Sheybani, P.-Y. Li, R. Lo, E. Meng, An implantable MEMS micropump system for drug delivery in small animals, Biomed. microdevices 14 (2012) 483–496. [4] R. Fernandes, D.H. Gracias, Self-folding polymeric containers for encapsulation and delivery of drugs, Adv. Drug Deliv. Rev. 64 (2012) 1579–1589. [5] J. Shi, A.R. Votruba, O.C. Farokhzad, R. Langer, Nanotechnology in drug delivery and tissue engineering: from discovery to applications, Nano Lett. 10 (2010) 3223. [6] Y.-C. Kim, J.-H. Park, M.R. Prausnitz, Microneedles for drug and vaccine delivery, Adv. Drug Deliv. Rev. 64 (2012) 1547–1568. [7] G. Kook, S.W. Lee, H.C. Lee, I.-J. Cho, H.J. Lee, Neural probes for chronic applications, Micromachines 7 (2016) 179. [8] M.-C. Chen, H.-A. Chan, M.-H. Ling, L.-C. Su, Implantable Polymeric Microneedles with Phototriggerable Properties as a Patient-Controlled Transdermal Analgesia SystemJournal of Materials Chemistry B 2017. [9] H.J. Lee, Y. Son, J. Kim, C.J. Lee, E.-S. Yoon, I.-J. Cho, A multichannel neural probe with embedded microfluidic channels for simultaneous in vivo neural recording and drug delivery, Lab Chip 15 (2015) 1590–1597. [10] F. Munoz, G. Alici, W. Li, A review of drug delivery systems for capsule endoscopy, Adv. Drug Deliv. Rev. 71 (2014) 77–85. [11] M. Aryal, C.D. Arvanitis, P.M. Alexander, N. McDannold, Ultrasound-mediated blood–brain barrier disruption for targeted drug delivery in the central nervous system, Adv. Drug Deliv. Rev. 72 (2014) 94–109. [12] A. Nisar, N. Afzulpurkar, B. Mahaisavariya, A. Tuantranont, MEMS-based micropumps in drug delivery and biomedical applications, Sensors Actuators B Chem. 130 (2008) 917–942. [13] K. van der Maaden, W. Jiskoot, J. Bouwstra, Microneedle technologies for (trans) dermal drug and vaccine delivery, J. Control. Release 161 (2012) 645–655. [14] K. Ita, Dermal/transdermal delivery of small interfering RNA and antisense oligonucleotides-advances and hurdles, Biomed Pharmacother 87 (2017) 311–320. [15] M.N. Yasin, D. Svirskis, A. Seyfoddin, I.D. Rupenthal, Implants for drug delivery to the posterior segment of the eye: a focus on stimuli-responsive and tunable release systems, J. Control. Release 196 (2014) 208–221. [16] H. Li, Y. Yu, S. Faraji Dana, B. Li, C.-Y. Lee, L. Kang, Novel engineered systems for oral, mucosal and transdermal drug delivery, J. Drug Target. 21 (2013) 611–629. [17] W.M. Pardridge, Drug transport across the blood–brain barrier, J. Cereb. Blood Flow Metab. 32 (2012) 1959–1972. [18] S. Mitragotri, Healing sound: the use of ultrasound in drug delivery and other therapeutic applications, Nat. Rev. Drug Discov. 4 (2005) 255–260. [19] N. Vykhodtseva, N. McDannold, K. Hynynen, Progress and problems in the application of focused ultrasound for blood–brain barrier disruption, Ultrasonics 48 (2008) 279–296. [20] A. Dasgupta, M. Liu, T. Ojha, G. Storm, F. Kiessling, T. Lammers, Ultrasound-mediated drug delivery to the brain: principles, progress and prospects, Drug Discov. Today Technol. 20 (2016) 41–48. [21] F. Fang, D. Zou, W. Wang, Y. Yin, T. Yin, S. Hao, B. Wang, G. Wang, Y. Wang, Noninvasive approaches for drug delivery to the brain based on the receptor mediated transport, Mater. Sci. Eng. C 76 (2017) 1316–1327. [22] L. Illum, Is nose-to-brain transport of drugs in man a reality? J. Pharm. Pharmacol. 56 (2004) 3–17. [23] Z.N. Warnken, H.D. Smyth, A.B. Watts, S. Weitman, J.G. Kuhn, R.O. Williams, Formulation and device design to increase nose to brain drug delivery, J. Drug Delivery Sci. Technol. 35 (2016) 213–222. [24] W.M. Pardridge, The blood-brain barrier: bottleneck in brain drug development, NeuroRx 2 (2005) 3–14. [25] M.S. Fiandaca, M.S. Berger, K.S. Bankiewicz, The use of convection-enhanced delivery with liposomal toxins in neurooncology, Toxins 3 (2011) 369–397. [26] H.J. Lee, Y. Son, D. Kim, Y.K. Kim, N. Choi, E.-S. Yoon, I.-J. Cho, A new thin silicon microneedle with an embedded microchannel for deep brain drug infusion, Sensors Actuators B Chem. 209 (2015) 413–422. [27] J. Chen, K.D. Wise, J.F. Hetke, S. Bledsoe, A multichannel neural probe for selective chemical delivery at the cellular level, IEEE Trans. Biomed. Eng. 44 (1997) 760–769. [28] Z. Fekete, A. Pongrácz, Multifunctional soft implants to monitor and control neural activity in the central and peripheral nervous system: a review, Sensors Actuators B Chem. 243 (2017) 1214–1223. [29] M.J. de Boer, R.W. Tjerkstra, J. Berenschot, H.V. Jansen, G. Burger, J. Gardeniers, M. Elwenspoek, A. van den Berg, Micromachining of buried micro channels in silicon, J. Microelectromech. Syst. 9 (2000) 94–103. [30] A. Altuna, E. Bellistri, E. Cid, P. Aivar, B. Gal, J. Berganzo, G. Gabriel, A. Guimerà, R. Villa, L.J. Fernández, SU-8 based microprobes for simultaneous neural depth recording and drug delivery in the brain, Lab Chip 13 (2013) 1422–1430. [31] N. Vasylieva, S. Marinesco, D. Barbier, A. Sabac, Silicon/SU8 multi-electrode microneedle for in vivo neurochemical monitoring, Biosens. Bioelectron. 72 (2015) 148–155. [32] H. Shin, H.J. Lee, U. Chae, H. Kim, J. Kim, N. Choi, J. Woo, Y. Cho, C.J. Lee, E.-S. Yoon, Neural probes with multi-drug delivery capability, Lab Chip 15 (2015) 3730–3737. [33] J.G. McCall, R. Qazi, G. Shin, S. Li, M.H. Ikram, K.-I. Jang, Y. Liu, R. Al-Hasani, M.R. Bruchas, J.-W. Jeong, Preparation and implementation of optofluidic neural probes for in vivo wireless pharmacology and optogenetics, Nat. Protoc. 12 (2017) 219–237. [34] W.-C. Huang, Y.-C. Lo, C.-Y. Chu, H.-Y. Lai, Y.-Y. Chen, S.-Y. Chen, Conductive nanogel-interfaced neural microelectrode arrays with electrically controlled insitu delivery of manganese ions enabling high-resolution MEMRI for synchronous neural tracing with deep brain stimulation, Biomaterials 122 (2017) 141–153. [35] Z. Fekete, E. Pálfi, G. Márton, M. Handbauer, Z. Bérces, I. Ulbert, A. Pongrácz, L. Négyessy, Combined in vivo recording of neural signals and iontophoretic injection of pathway tracers using a hollow silicon microelectrode, Sensors Actuators B Chem. 236 (2016) 815–824.

145

[36] W. Lee, T. Ngernsutivorakul, O.S. Mabrouk, J.-M.T. Wong, C.E. Dugan, S.S. Pappas, H.J. Yoon, R.T. Kennedy, Microfabrication and in vivo performance of a microdialysis probe with embedded membrane, Anal. Chem. 88 (2016) 1230. [37] U. Chae, H. Shin, H.J. Lee, J. Lee, N. Choi, Y.J. Lee, S.H. Lee, J. Woo, Y. Cho, E.-S. Yoon, A New MEMS Neural Probe System Integrated with Push-Pull Microfluidic Channels and Biosensors for Real-Time Monitoring of Neurochemicals, Micro Electro Mechanical Systems (MEMS), 2016 IEEE 29th International Conference on, IEEE 2016, pp. 329–332. [38] G. Petit-Pierre, A. Bertsch, P. Renaud, Neural probe combining microelectrodes and a droplet-based microdialysis collection system for high temporal resolution sampling, Lab Chip 16 (2016) 917–924. [39] M.A. Green, L.E. Bilston, R. Sinkus, In vivo brain viscoelastic properties measured by magnetic resonance elastography, NMR Biomed. 21 (2008) 755–764. [40] B. Liu, E. Kim, A. Meggo, S. Gandhi, H. Luo, S. Kallakuri, Y. Xu, J. Zhang, Enhanced biocompatibility of neural probes by integrating microstructures and delivering anti-inflammatory agents via microfluidic channels, J. Neural Eng. 14 (2017), 026008. [41] C. Boehler, C. Kleber, N. Martini, Y. Xie, I. Dryg, T. Stieglitz, U. Hofmann, M. Asplund, Actively controlled release of dexamethasone from neural microelectrodes in a chronic in vivo study, Biomaterials 129 (2017) 176–187. [42] F.-Y. Yang, Y.-S. Lin, K.-H. Kang, T.-K. Chao, Reversible blood–brain barrier disruption by repeated transcranial focused ultrasound allows enhanced extravasation, J. Control. Release 150 (2011) 111–116. [43] Y. Negishi, M. Yamane, N. Kurihara, Y. Endo-Takahashi, S. Sashida, N. Takagi, R. Suzuki, K. Maruyama, Enhancement of blood–brain barrier permeability and delivery of antisense oligonucleotides or plasmid DNA to the brain by the combination of bubble liposomes and high-intensity focused ultrasound, Pharmaceutics 7 (2015) 344–362. [44] S. Wang, T. Kugelman, A. Buch, M. Herman, Y. Han, M.E. Karakatsani, S.A. Hussaini, K. Duff, E.E. Konofagou, Non-invasive, focused ultrasound-facilitated gene delivery for optogenetics, Sci. Rep. 7 (2017). [45] R.D. Airan, R.A. Meyer, N.P. Ellens, K.R. Rhodes, K. Farahani, M.G. Pomper, S.D. Kadam, J.J. Green, Noninvasive targeted transcranial neuromodulation via focused ultrasound gated drug release from nanoemulsions, Nano Lett. 17 (2017) 652–659. [46] R.D. Airan, C.A. Foss, N.P. Ellens, Y. Wang, R.C. Mease, K. Farahani, M.G. Pomper, MR-guided delivery of hydrophilic molecular imaging agents across the bloodbrain barrier through focused ultrasound, Mol. Imaging Biol. 19 (2017) 24–30. [47] Z. Jin, Y. Choi, S.Y. Ko, J.O. Park, S. Park, Experimental and simulation studies on focused ultrasound triggered drug delivery, Biotechnol. Appl. Biochem. 64 (2016) 134–142. [48] Y. Chau, W.L.L. Suen, H.Y. Tse, H.S. Wong, Ultrasound-enhanced penetration through sclera depends on frequency of sonication and size of macromolecules, Eur. J. Pharm. Sci. 100 (2017) 273–279. [49] H. Ballantine Jr., E. Bell, J. Manlapaz, Progress and problems in the neurological applications of focused ultrasound, J. Neurosurg. 17 (1960) 858–876. [50] L. Bakay, T. Hueter, H. Ballantine, D. Sosa, Ultrasonically produced changes in the blood-brain barrier, A.M.A. Arch. Neurol. Psychiatry 76 (1956) 457–467. [51] A.H. Mesiwala, L. Farrell, H.J. Wenzel, D.L. Silbergeld, L.A. Crum, H.R. Winn, P.D. Mourad, High-intensity focused ultrasound selectively disrupts the blood-brain barrier in vivo, Ultrasound Med. Biol. 28 (2002) 389–400. [52] B.E. Polat, D. Hart, R. Langer, D. Blankschtein, Ultrasound-mediated transdermal drug delivery: mechanisms, scope, and emerging trends, J. Control. Release 152 (2011) 330–348. [53] C.M. Schoellhammer, B.E. Polat, J. Mendenhall, R. Maa, B. Jones, D.P. Hart, R. Langer, D. Blankschtein, Rapid skin permeabilization by the simultaneous application of dual-frequency, high-intensity ultrasound, J. Control. Release 163 (2012) 154–160. [54] P. Hariharan, M. Nabili, A. Guan, V. Zderic, M. Myers, Model for porosity changes occurring during ultrasound-enhanced transcorneal drug delivery, Ultrasound Med. Biol. 43 (2017) 1223–1236. [55] C.M. Schoellhammer, A. Schroeder, R. Maa, G.Y. Lauwers, A. Swiston, M. Zervas, R. Barman, A.M. DiCiccio, W.R. Brugge, D.G. Anderson, Ultrasound-mediated gastrointestinal drug delivery, Sci. Transl. Med. 7 (2015) 310ra168-310ra168. [56] C.-H. Fan, C.-Y. Ting, Y.-C. Chang, K.-C. Wei, H.-L. Liu, C.-K. Yeh, Drug-loaded bubbles with matched focused ultrasound excitation for concurrent blood– brain barrier opening and brain-tumor drug delivery, Acta Biomater. 15 (2015) 89–101. [57] L. Goldwirt, M. Canney, C. Horodyckid, J. Poupon, S. Mourah, A. Vignot, J.-Y. Chapelon, A. Carpentier, Enhanced brain distribution of carboplatin in a primate model after blood–brain barrier disruption using an implantable ultrasound device, Cancer Chemother. Pharmacol. 77 (2016) 211–216. [58] E. Park, J. Werner, N.B. Smith, Ultrasound mediated transdermal insulin delivery in pigs using a lightweight transducer, Pharm. Res. 24 (2007) 1396–1401. [59] E. Maione, K.K. Shung, R.J. Meyer, J.W. Hughes, R.E. Newnham, N.B. Smith, Transducer design for a portable ultrasound enhanced transdermal drugdelivery system, IEEE Trans. Ultrason. Ferroelectr. Freq. Control 49 (2002) 1430–1436. [60] Y. Tufail, A. Matyushov, N. Baldwin, M.L. Tauchmann, J. Georges, A. Yoshihiro, S.I.H. Tillery, W.J. Tyler, Transcranial pulsed ultrasound stimulates intact brain circuits, Neuron 66 (2010) 681–694. [61] M. Fini, W.J. Tyler, Transcranial focused ultrasound: a new tool for non-invasive neuromodulation, Int. Rev. Psychiatry (2017) 1–10. [62] H.-S. Yoon, S. Vaithilingam, K.K. Park, A. Nikoozadeh, K. Firouzi, J.W. Choe, R.D. Watkins, H.K. Oguz, M. Kupnik, K.B. Pauly, A 1-MHz 2-D CMUT array for HIFU thermal ablation, AIP Conference Proceedings, AIP Publishing 2017, p. 050003.

146

H.J. Lee et al. / Advanced Drug Delivery Reviews 128 (2018) 132–147

[63] K. Ko, J.-H. Lee, H.J. Lee, S.-J. Oh, Y.E. Chun, T.S. Kim, C.J. Lee, E.-S. Yoon, K.-S. Yun, I.-J. Cho, Micromachined ultrasound transducer array for cell stimulation with high spatial resolution, Micro Electro Mechanical Systems (MEMS), 2015 28th IEEE International Conference on, IEEE 2015, pp. 651–654. [64] J. Chen, Capacitive micromachined ultrasonic transducer arrays for minimally invasive medical ultrasound, J. Micromech. Microeng. 20 (2010), 023001. [65] E.M. Del Amo, A. Urtti, Current and future ophthalmic drug delivery systems: a shift to the posterior segment, Drug Discov. Today 13 (2008) 135–143. [66] T. Yasukawa, Y. Ogura, E. Sakurai, Y. Tabata, H. Kimura, Intraocular sustained drug delivery using implantable polymeric devices, Adv. Drug Deliv. Rev. 57 (2005) 2033–2046. [67] P. Bansal, S. Garg, Y. Sharma, P. Venkatesh, Posterior segment drug delivery devices: current and novel therapies in development, J. Ocul. Pharmacol. Ther. 32 (2016) 135–144. [68] R.R. Joseph, S.S. Venkatraman, Drug delivery to the eye: what benefits do nanocarriers offer? Nanomedicine 12 (2017) 683–702. [69] P.W. Morrison, V.V. Khutoryanskiy, Advances in ophthalmic drug delivery, Ther. Deliv. 5 (2014) 1297–1315. [70] N. Kuno, S. Fujii, Recent advances in ocular drug delivery systems, Polymer 3 (2011) 193–221. [71] A. Guzman-Aranguez, B. Colligris, J. Pintor, Contact lenses: promising devices for ocular drug delivery, J. Ocul. Pharmacol. Ther. 29 (2013) 189–199. [72] A. Patel, K. Cholkar, V. Agrahari, A.K. Mitra, Ocular drug delivery systems: an overview, World J. Pharmacol. 2 (2013) 47. [73] H.P. Filipe, J. Henriques, P. Reis, P.C. Silva, M.J. Quadrado, A.P. Serro, Contact lenses as drug controlled release systems: a narrative review, Rev. Bras. Oftalmol. 75 (2016) 241–247. [74] N. Haghjou, M. Soheilian, M.J. Abdekhodaie, Sustained release intraocular drug delivery devices for treatment of uveitis, J. Ophthalmic Vis. Res. 6 (2011) 317. [75] J.I. Lim, R.A. Wolitz, A.H. Dowling, H.R. Bloom, A.R. Irvine, D.M. Schwartz, Visual and anatomic outcomes associated with posterior segment complications after ganciclovir implant procedures in patients with AIDS and cytomegalovirus retinitis, Am J. Ophthalmol. 127 (1999) 288–293. [76] J. Jiang, J.S. Moore, H.F. Edelhauser, M.R. Prausnitz, Intrascleral drug delivery to the eye using hollow microneedles, Pharm. Res. 26 (2009) 395–403. [77] J. Jiang, H.S. Gill, D. Ghate, B.E. McCarey, S.R. Patel, H.F. Edelhauser, M.R. Prausnitz, Coated microneedles for drug delivery to the eye, Invest. Ophthalmol. Vis. Sci. 48 (2007) 4038–4043. [78] S.R. Patel, A.S. Lin, H.F. Edelhauser, M.R. Prausnitz, Suprachoroidal drug delivery to the back of the eye using hollow microneedles, Pharm. Res. 28 (2011) 166–176. [79] G. Valdés-Ramírez, J.R. Windmiller, J.C. Claussen, A.G. Martinez, F. Kuralay, M. Zhou, N. Zhou, R. Polsky, P.R. Miller, R. Narayan, Multiplexed and switchable release of distinct fluids from microneedle platforms via conducting polymer nanoactuators for potential drug delivery, Sensors Actuators B Chem. 161 (2012) 1018–1024. [80] R.R.S. Thakur, I.A. Tekko, F. Al-Shammari, A.A. Ali, H. McCarthy, R.F. Donnelly, Rapidly dissolving polymeric microneedles for minimally invasive intraocular drug delivery, Drug Deliv. Transl. Res. 6 (2016) 800–815. [81] R. Lo, P.-Y. Li, S. Saati, R. Agrawal, M.S. Humayun, E. Meng, A refillable microfabricated drug delivery device for treatment of ocular diseases, Lab Chip 8 (2008) 1027–1030. [82] R. Lo, P.-Y. Li, S. Saati, R.N. Agrawal, M.S. Humayun, E. Meng, A passive MEMS drug delivery pump for treatment of ocular diseases, Biomed. Microdevices 11 (2009) 959. [83] P.-Y. Li, J. Shih, R. Lo, S. Saati, R. Agrawal, M.S. Humayun, Y.-C. Tai, E. Meng, An electrochemical intraocular drug delivery device, Sensors Actuators A Phys. 143 (2008) 41–48. [84] I.D. Rupenthal, Sector overview: ocular drug delivery technologies: exciting times ahead, ONdrugDelivery 54 (2015) 7–11. [85] Y. Zhang, Y. Chen, X. Yu, Y. Qi, Y. Chen, Y. Liu, Y. Hu, Z. Li, A flexible device for ocular iontophoretic drug delivery, Biomicrofluidics 10 (2016), 011911. [86] Y. Shikamura, Y. Yamazaki, T. Matsunaga, T. Sato, A. Ohtori, K. Tojo, Hydrogel ring for topical drug delivery to the ocular posterior segment, Curr. Eye Res. 41 (2016) 653–661. [87] S. Yu, X. Zhang, G. Tan, L. Tian, D. Liu, Y. Liu, X. Yang, W. Pan, A novel pH-induced thermosensitive hydrogel composed of carboxymethyl chitosan and poloxamer cross-linked by glutaraldehyde for ophthalmic drug delivery, Carbohydr. Polym. 155 (2017) 208–217. [88] N. Nagai, H. Kaji, H. Onami, Y. Ishikawa, M. Nishizawa, N. Osumi, T. Nakazawa, T. Abe, A polymeric device for controlled transscleral multi-drug delivery to the posterior segment of the eye, Acta Biomater. 10 (2014) 680–687. [89] P. Song, S. Kuang, N. Panwar, G. Yang, D.J.H. Tng, S.C. Tjin, W.J. Ng, M.B.A. Majid, G. Zhu, K.T. Yong, A self-powered implantable drug-delivery system using biokinetic energy, Adv. Mater. 29 (2017) (in press). [90] F.N. Pirmoradi, J.K. Jackson, H.M. Burt, M. Chiao, On-demand controlled release of docetaxel from a battery-less MEMS drug delivery device, Lab Chip 11 (2011) 2744–2752. [91] C.R. Craig, R.E. Stitzel, Modern Pharmacology with Clinical Applications, Lippincott Williams & Wilkins, 2004. [92] I.R. Wilding, Site-specific drug delivery in the gastrointestinal tract, Crit. Rev. Ther. Drug Carrier Syst. 17 (2000). [93] C. Mc Caffrey, O. Chevalerias, C. O'Mathuna, K. Twomey, Swallowable-capsule technology, IEEE Pervasive Comput. 7 (2008). [94] H.N. Stevens, C.G. Wilson, P.G. Welling, M. Bakhshaee, J.S. Binns, A.C. Perkins, M. Frier, E.P. Blackshaw, M.W. Frame, D.J. Nichols, Evaluation of Pulsincap™ to provide regional delivery of dofetilide to the human GI tract, Int. J. Pharm. 236 (2002) 27–34.

[95] J.M. Hebden, P.J. Gilchrist, A.C. Perkins, C.G. Wilson, R.C. Spiller, Stool water content and colonic drug absorption: contrasting effects of lactulose and codeine, Pharm. Res. 16 (1999) 1254–1259. [96] R.S. Shawgo, A.C.R. Grayson, Y. Li, M.J. Cima, BioMEMS for drug delivery, Curr. Opinion Solid State Mater. Sci. 6 (2002) 329–334. [97] Z. Jia, G. Yan, Y. Shi, B. Zhu, A wireless power transmission system for an active capsule endoscope for colon inspection, J. Med. Eng. Technol. 36 (2012) 235–241. [98] A. Abid, J.M. O'Brien, T. Bensel, C. Cleveland, L. Booth, B.R. Smith, R. Langer, G. Traverso, Wireless power transfer to millimeter-sized gastrointestinal electronics validated in a swine model, Sci. Rep. 7 (2017). [99] M.N. Antipina, G.B. Sukhorukov, Remote control over guidance and release properties of composite polyelectrolyte based capsules, Adv. Drug Deliv. Rev. 63 (2011) 716–729. [100] S. Yim, M. Sitti, Design and rolling locomotion of a magnetically actuated soft capsule endoscope, IEEE Trans. Robot. 28 (2012) 183–194. [101] A.W. Mahoney, J.J. Abbott, 5-DOF manipulation of an untethered magnetic device in fluid using a single permanent magnet, robotics: science and systems, citeseer, 2014. [102] P. Valdastri, R.J. Webster III, C. Quaglia, M. Quirini, A. Menciassi, P. Dario, A new mechanism for mesoscale legged locomotion in compliant tubular environments, IEEE Trans. Robot. 25 (2009) 1047–1057. [103] M. Quirini, A. Menciassi, S. Scapellato, C. Stefanini, P. Dario, Design and fabrication of a motor legged capsule for the active exploration of the gastrointestinal tract, IEEE/ASME Trans. Mechatron. 13 (2008) 169–179. [104] A. Menciassi, C. Stefanini, S. Gorini, G. Pernorio, B. Kim, J. Park, P. Dario, Locomotion of a legged capsule in the gastrointestinal tract: theoretical study and preliminary technological results, Engineering in Medicine and Biology Society, 2004. IEMBS’04, 26th Annual International Conference of the IEEE, IEEE 2004, pp. 2767–2770. [105] B. Kim, S. Lee, J.H. Park, J.-O. Park, Design and fabrication of a locomotive mechanism for capsule-type endoscopes using shape memory alloys (SMAs), IEEE/ASME Trans. Mechatron. 10 (2005) 77–86. [106] B. Kim, S. Park, C.Y. Jee, S.-J. Yoon, An earthworm-like locomotive mechanism for capsule endoscopes, Intelligent Robots and Systems, 2005.(IROS 2005), 2005 IEEE/RSJ International Conference on, IEEE 2005, pp. 2997–3002. [107] L.J. Sliker, X. Wang, J.A. Schoen, M.E. Rentschler, Micropatterned treads for in vivo robotic mobility, J. Med. Devices 4 (2010), 041006. [108] S.P. Woods, T.G. Constandinou, Wireless capsule endoscope for targeted drug delivery: mechanics and design considerations, IEEE Trans. Biomed. Eng. 60 (2013) 945–953. [109] W. Chen, G. Yan, Z. Wang, P. Jiang, H. Liu, A wireless capsule robot with spiral legs for human intestine, Int. J. Med. Rob. Comput. Assisted Surg. 10 (2014) 147–161. [110] P. Xitian, L. Hongying, W. Kang, L. Yulin, Z. Xiaolin, W. Zhiyu, A novel remote controlled capsule for site-specific drug delivery in human GI tract, Int. J. Pharm. 382 (2009) 160–164. [111] D. Sinha, Wireless actuation of piezoelectric coupled micromembrane using radio frequency magnetic field for biomedical applications, J. Appl. Phys. 121 (2017) 134501. [112] H. Dehghani, C.R. Welch, A. Pourghodrat, C.A. Nelson, D. Oleynikov, P. Dasgupta, B.S. Terry, Design and preliminary evaluation of a self-steering, pneumatically driven colonoscopy robot, J. Med. Eng. Technol. 41 (2017) 223–236. [113] J. Cui, X. Zheng, W. Hou, Y. Zhuang, X. Pi, J. Yang, The study of a remote-controlled gastrointestinal drug delivery and sampling system, Telemed. E Health 14 (2008) 715–719. [114] F. Munoz, G. Alici, W. Li, A magnetically actuated drug delivery system for robotic endoscopic capsules, J. Med. Devices 10 (2016), 011004. [115] D. She, M. Tsang, J. Kim, M. Allen, Immobilized electrolyte biodegradable batteries for implantable MEMS, solid-state sensors, actuators and microsystems (TRANSDUCERS), 2015 Transducers-2015 18th International Conference on, IEEE 2015, pp. 494–497. [116] P. Nadeau, D. El-Damak, D. Glettig, Y.L. Kong, S. Mo, C. Cleveland, L. Booth, N. Roxhed, R. Langer, A.P. Chandrakasan, Prolonged energy harvesting for ingestible devices, Nat. Biomed. Eng. 1 (2017) 0022. [117] M.R. Prausnitz, R. Langer, Transdermal drug delivery, Nat. Biotechnol. 26 (2008) 1261–1268. [118] S.M. Bal, Z. Ding, E. van Riet, W. Jiskoot, J.A. Bouwstra, Advances in transcutaneous vaccine delivery: do all ways lead to Rome? J. Control. Release 148 (2010) 266–282. [119] K.-Y. Seong, M.-S. Seo, D.Y. Hwang, E.D. O'Cearbhaill, S. Sreenan, J.M. Karp, S.Y. Yang, A self-adherent, bullet-shaped microneedle patch for controlled transdermal delivery of insulin, J. Control. Release (2017) (in press). [120] K. Tsioris, W.K. Raja, E.M. Pritchard, B. Panilaitis, D.L. Kaplan, F.G. Omenetto, Fabrication of silk microneedles for controlled-release drug delivery, Adv. Funct. Mater. 22 (2012) 330–335. [121] S.P. Sullivan, D.G. Koutsonanos, M. del Pilar Martin, J.W. Lee, V. Zarnitsyn, S.O. Choi, N. Murthy, R.W. Compans, I. Skountzou, M.R. Prausnitz, Dissolving polymer microneedle patches for influenza vaccination, Nat. Med. 16 (2010) 915–920. [122] P.C. DeMuth, Y. Min, B. Huang, J.A. Kramer, A.D. Miller, D.H. Barouch, P.T. Hammond, D.J. Irvine, Polymer multilayer tattooing for enhanced DNA vaccination, Nat. Mater. 12 (2013) 367–376. [123] A. Alexander, S. Dwivedi, T.K. Giri, S. Saraf, S. Saraf, D.K. Tripathi, Approaches for breaking the barriers of drug permeation through transdermal drug delivery, J. Control. Release 164 (2012) 26–40. [124] J. Yu, Y. Zhang, A.R. Kahkoska, Z. Gu, Bioresponsive transcutaneous patches, Curr. Opin. Biotechnol. 48 (2017) 28–32.

H.J. Lee et al. / Advanced Drug Delivery Reviews 128 (2018) 132–147 [125] F. Teodorescu, G. Quéniat, C. Foulon, M. Lecoeur, A. Barras, S. Boulahneche, M.S. Medjram, T. Hubert, A. Abderrahmani, R. Boukherroub, Transdermal skin patch based on reduced graphene oxide: a new approach for photothermal triggered permeation of ondansetron across porcine skin, J. Control. Release 245 (2017) 137–146. [126] G.K. Gulati, T. Chen, B.J. Hinds, Programmable carbon nanotube membrane-based transdermal nicotine delivery with microdialysis validation assay, Nanomedicine 13 (2017) 1–9. [127] J. Di, S. Yao, Y. Ye, Z. Cui, J. Yu, T.K. Ghosh, Y. Zhu, Z. Gu, Stretch-triggered drug delivery from wearable elastomer films containing therapeutic depots, ACS Nano 9 (2015) 9407–9415. [128] C.-J. Huang, Y.-H. Chen, C.-H. Wang, T.-C. Chou, G.-B. Lee, Integrated microfluidic systems for automatic glucose sensing and insulin injection, Sensors Actuators B Chem. 122 (2007) 461–468. [129] Y. Ye, J. Yu, C. Wang, N.Y. Nguyen, G.M. Walker, J.B. Buse, Z. Gu, Microneedles integrated with pancreatic cells and synthetic glucose-signal amplifiers for smart insulin delivery, Adv. Mater. 28 (2016) 3115–3121. [130] X. Hu, J. Yu, C. Qian, Y. Lu, A.R. Kahkoska, Z. Xie, X. Jing, J.B. Buse, Z. Gu, H2O2-responsive vesicles integrated with transcutaneous patches for glucose-mediated insulin delivery, ACS Nano 11 (2017) 613–620. [131] C. Wang, Y. Ye, G.M. Hochu, H. Sadeghifar, Z. Gu, Enhanced cancer immunotherapy by microneedle patch-assisted delivery of anti-PD1 antibody, Nano Lett. 16 (2016) 2334–2340. [132] C.L. Stevenson, J.T. Santini, R. Langer, Reservoir-based drug delivery systems utilizing microtechnology, Adv. Drug Deliv. Rev. 64 (2012) 1590–1602. [133] J.T. Santini, M.J. Cima, R. Langer, A controlled-release microchip, Nature 397 (1999) 335–338. [134] J.C. Wright, S.T. Leonard, C.L. Stevenson, J.C. Beck, G. Chen, R.M. Jao, P.A. Johnson, J. Leonard, R.J. Skowronski, An in vivo/in vitro comparison with a leuprolide osmotic implant for the treatment of prostate cancer, J. Control. Release 75 (2001) 1–10. [135] H. Gensler, R. Sheybani, P.-Y. Li, R. Lo, S. Zhu, K.-T. Yong, I. Roy, P.N. Prasad, R. Masood, U.K. Sinha, Implantable MEMS drug delivery device for cancer radiation

[136] [137]

[138] [139] [140]

[141]

[142]

[143]

[144]

[145]

147

reduction, Micro Electro Mechanical Systems (MEMS), 2010 IEEE 23rd International Conference on, IEEE 2010, pp. 23–26. M. Ochoa, C. Mousoulis, B. Ziaie, Polymeric microdevices for transdermal and subcutaneous drug delivery, Adv. Drug Deliv. Rev. 64 (2012) 1603–1616. R. Jegadeesan, K. Agarwal, Y.-X. Guo, S.-C. Yen, N.V. Thakor, Wireless power delivery to flexible subcutaneous implants using capacitive coupling, IEEE Trans. Microwave Theory Tech. 65 (2017) 280–292. A. Cobo, R. Sheybani, H. Tu, E. Meng, A wireless implantable micropump for chronic drug infusion against cancer, Sensors Actuators A Phys. 239 (2016) 18–25. W.-J. Jo, S.-K. Baek, J.-H. Park, A wireless actuating drug delivery system, J. Micromech. Microeng. 25 (2015), 045014. P.R. van Dijk, S.J. Logtenberg, K.H. Groenier, J.W. Haveman, N. Kleefstra, H.J. Bilo, Complications of continuous intraperitoneal insulin infusion with an implantable pump, World J. Diabetes 3 (2012) 142. E. Zini, I. Padrutt, K. Macha, A. Riederer, M. Pesaresi, T. Lutz, C. Reusch, Use of an implantable pump for controlled subcutaneous insulin delivery in healthy cats, Vet. J. 219 (2017) 60–64. A. Lathuilière, V. Laversenne, A. Astolfo, E. Kopetzki, H. Jacobsen, M. Stampanoni, B. Bohrmann, B.L. Schneider, P. Aebischer, A subcutaneous cellular implant for passive immunization against amyloid-β reduces brain amyloid and tau pathologies, Brain 139 (2016) 1587–1604 aww036. Y. Liu, P. Song, J. Liu, D.J.H. Tng, R. Hu, H. Chen, Y. Hu, C.H. Tan, J. Wang, J. Liu, An invivo evaluation of a MEMS drug delivery device using Kunming mice model, Biomed. Microdevices 17 (2015) 6. S.Y. Chin, Y.C. Poh, A.-C. Kohler, J.T. Compton, L.L. Hsu, K.M. Lau, S. Kim, B.W. Lee, F.Y. Lee, S.K. Sia, Additive manufacturing of hydrogel-based materials for next-generation implantable medical devices, Sci. Robot. 2 (2017) eaah6451. A. Weltman, J. Yoo, E. Meng, Flexible, penetrating brain probes enabled by advances in polymer microfabrication, Micromachines 7 (2016) 180.