Advanced Drug Delivery Reviews 55 (2003) 315–328

www.elsevier.com / locate / addr

Microfabricated drug delivery systems: from particles to pores

Sarah L. Tao a,b , Tejal A. Desai a,b , *
a
Department of Bioengineering, University of Illinois at Chicago, 851 S. Morgan Street, Chicago, IL 60607, USA
b
Department of Biomedical Engineering, Boston University, 44 Cummington Street, Boston, MA 02215, USA

Abstract

Microfabrication techniques which permit the creation of therapeutic delivery systems that possess a combination of
structural, mechanical, and perhaps electronic features may surmount challenges associated with conventional delivery of
therapy. In this review, delivery concepts are presented which capitalize on the strengths of microfabrication. Possible
applications include micromachined silicon membranes to create implantable biocapsules for the immunoisolation of
pancreatic islet cells—as a possible treatment for diabetes—and sustained release of injectable drugs needed over long time
periods. Asymmetrical, drug-loaded microfabricated particles with specific ligands linked to the surface are proposed for
improving oral bioavailability of peptide (and perhaps protein) drugs. In addition, microfabricated drug delivery systems
ranging from transdermal microneedles to implantable microchips will be discussed.
 2002 Elsevier Science B.V. All rights reserved.

Keywords: Silicon; Microtechnology; Microfabrication; Therapeutic; Drug delivery; Microparticles

Contents

1. Introduction ............................................................................................................................................................................ 316

2. Controlled release drug delivery systems................................................................................................................................... 316
3. Microfabrication technology .................................................................................................................................................... 317
4. Microneedles for transdermal drug delivery............................................................................................................................... 317
4.1. Microfabrication of silicon microneedle arrays ................................................................................................................... 317
4.2. Transdermal transport studies............................................................................................................................................ 318
4.3. Other microneedles .......................................................................................................................................................... 318
5. Implanted microchip for localized drug delivery ........................................................................................................................ 319
5.1. Microchip design ............................................................................................................................................................. 319
5.2. Irreversible metallic valves ............................................................................................................................................... 319
5.3. Reversible polymeric valves ............................................................................................................................................. 319
5.4. Current developments....................................................................................................................................................... 320
6. Bioadhesive microparticles for oral drug delivery ...................................................................................................................... 320
6.1. Silicon dioxide microparticles ........................................................................................................................................... 321
6.1.1. Fabrication of microparticle body ............................................................................................................................ 321
6.1.2. Fabrication of microparticle reservoir ...................................................................................................................... 321
6.1.3. Surface modification chemistry ............................................................................................................................... 321
6.2. Poly(methyl methacrylate) microparticles........................................................................................................................... 322

*Corresponding author. Tel.: 1 1-617-358-3054; fax: 1 1-617-353-6766.

E-mail address: tdesai@bu.edu (T.A. Desai).

0169-409X / 02 / $ – see front matter  2002 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0169-409X(02)00227-2
316 S.L. Tao, T. A. Desai / Advanced Drug Delivery Reviews 55 (2003) 315–328

6.2.1. Fabrication of microparticle body ............................................................................................................................ 322

6.2.2. Fabrication of microparticle reservoir ...................................................................................................................... 322
6.2.3. Surface modification chemistry ............................................................................................................................... 322
6.3. Release mechanism .......................................................................................................................................................... 323
6.4. Microparticle bioadhesion to intestinal epithelium .............................................................................................................. 323
6.4.1. Lectins .................................................................................................................................................................. 323
6.4.2. Caco-2 binding ...................................................................................................................................................... 323
6.5. Future developments ........................................................................................................................................................ 324
7. Nanoporous immunoisolating biocapsules ................................................................................................................................. 324
7.1. Allotransplantation ........................................................................................................................................................... 324
7.2. Biocapsule design ............................................................................................................................................................ 325
7.3. Biocompatibility and cytotoxicity studies ........................................................................................................................... 326
8. Conclusions ............................................................................................................................................................................ 326
Acknowledgements ...................................................................................................................................................................... 327
References .................................................................................................................................................................................. 327

1. Introduction administration. However, these types of dosages are

not easily able to control the rate of drug delivery or
The application of micro- and nanotechnology to the target area of the drug and are often associated
the biomedical arena has tremendous potential in with an immediate or rapid drug release. Conse-
terms of developing new diagnostic and therapeutic quently, the initial concentration of the drug in the
modalities. Over the last several years, microfabrica- body peaks above the level of toxicity and then
tion technology has been applied to the successful gradually diminishes over time to an ineffective
development of a variety of health care-related level. The duration of therapeutic efficacy then
products including diagnostic (‘lab-on-a-chip’) sys- becomes dependent on the frequency of administra-
tems and techniques and apparatus for high through- tion, and half-life of the drug, and high dosages of
put screening of new drug candidates [1,2]. While the non-targeted drugs are often administered to achieve
majority of research has focused on the development an effective blood concentration [8]. In recent years,
of miniaturized diagnostic tools, researchers have increasingly sophisticated and potent drugs have
more recently concentrated on the development of been developed by the biotech industry. For many of
microdevices for therapeutic applications. Micro- these new protein-based and DNA-based com-
and nanofabrication techniques are currently being pounds, the therapeutic concentration range is often
used to develop implants that can record from, sense, small, toxicity is observed for concentration spikes,
stimulate, and deliver to biological systems. Mi- or the therapeutic concentration range varies with
cromachined neural prostheses, drug delivery mi- time, which renders traditional methods of drug
cropumps, tissue scaffolds, and stents [3–6] have all delivery ineffective [9]. An immense amount of
been fabricated using precision-based microtech- interest has been increasingly placed on controlled
nologies. Drug delivery remains an important chal- release drug delivery systems to maintain the thera-
lenge in medicine [7] and microfabrication tech- peutic efficacy of these drugs. There are a number of
niques may be used to develop novel drug delivery mechanisms that can provide such controlled release
devices with capabilities not possible with current of drugs, including transdermal patches, implants,
systems. This paper will review some of the current bioadhesive systems, and microencapsulation [7].
and future approaches that utilize microfabrication Newer drug delivery technologies currently attracting
technology for drug delivery. attention include inhaled and sustained release inject-
able peptide / protein drugs from biodegradable poly-
mers.
2. Controlled release drug delivery systems Nevertheless, the ideal drug delivery system has
not yet been attained. Therefore, interdisciplinary
Conventional dosage forms, such as oral delivery approaches are being sought. These alternative ap-
and injection, are the predominant routes for drug proaches should, at least conceptually, address those
 S.L. Tao, T. A. Desai / Advanced Drug Delivery Reviews 55 (2003) 315–328
unsolved issues that still engage the drug delivery
scientist. These include drug targeting (of peptides,
proteins, and DNA) to improve cancer chemotherapy
and cardiovascular treatment, improved control of
the rate of release from implanted dosage forms,
pulsatile drug delivery, closed-loop systems that can
sense a change in the bodies biochemistry leading to
modulated rate or extent of drug delivered, and the
ability to implant xenogeneic cells in humans to
more naturally control a range of diseases, including
diabetes and neurodegenerative conditions.
3. Microfabrication technology
The use of traditional microfabrication techniques,
the same processing techniques used to manufacture
microelectronic chips, is a recent and alternative
method of creating drug delivery platforms. Micro-
electronic process engineering was a discipline that
developed due to the rapid growth of the integrated
4.1. Microfabrication of silicon microneedle arrays
circuit industry. Traditionally, microelectromechani-
cal systems (MEMS) research has been used to
produce functional devices on the micron scale, such
as sensors, switches, filters, and gears, from silicon,
the dominant material used throughout the IC indus-
try. These devices are fabricated by the repeated
application of unit process steps such as thin-film
deposition, photolithography, and etching. Such mi-
crofabrication techniques allow for the precise con-
trol over surface microarchitecture, topography, and
feature size.
Although research on microfabricated devices for
biomedical applications (BioMEMS) has rapidly
expanded in recent years, relatively few researchers
have concentrated on therapeutic applications of
microfabrication technology such as drug delivery.
The use of microtechnology to tailor the size, shape,
reservoir number, reservoir volume, unidirectional
openings and surface characteristics of the drug
delivery vehicle in conjunction with appropriate
surface chemistry are potentially influential in the
area of controlled release.
4. Microneedles for transdermal drug delivery
One alternative to oral delivery and intravenous
317
injection is the administration of drugs across the
skin. This approach seeks to avoid any degradation
of the molecules in the gastrointestinal tract and
first-pass effects of the liver associated with oral
drug delivery as well as the pain of intravenous
injection [10–14]. It also offers the possibility to
continuously control the delivery rate over extended
periods of time [14]. However, conventional trans-
dermal drug delivery is severely hindered by the
outer 10–20 mm of skin, a barrier of dead tissue
called the stratum corneum [15]. The development of
microneedles for transdermal drug delivery came
about as an approach to enhance the poor permeabili-
ty of the skin by creating microscale conduits for
transport across the stratum corneum [14]. The
development of microneedles that are long and
robust enough to penetrate this layer of skin, but
short enough to avoid stimulating nerves has the
potential to make transdermal delivery of drugs more
effective [15].
By adapting microfabrication technology, three-
dimensional arrays of sharp-tipped microneedles can
be made for transdermal drug delivery [10,14,15]. To
fabricate microneedles, a deep reactive ion etching
process is commonly used. In this process, a
chromium masking material is deposited onto silicon
wafers and patterned into dots that have a diameter
approximately equal to that of the base of the desired
needles. When placed in the reactive ion etcher, the
wafers are exposed to carefully controlled plasma of
fluorine and oxygen, which causes a deep vertical,
etch and slight lateral underetching. The regions on
the wafer that are protected by chromium remain and
eventually form the microneedles. Etching is allowed
to proceed until the masks are undercut and fall off,
leaving behind an array of silicon spikes [14]. The
aspect ratio of the microneedles can be adjusted by
simply modifying the ratio of flow rates of SF 6 and
O 2 . Hollow silicon needles can also be fabricated
using deep reactive ion etching in an inductively
coupled reactive ion etcher. The deep etch creates
arrays of holes through the silicon wafer (the needle
lumen) and the microneedles are formed by reactive
ion etching around these holes [15].
Arrays of solid silicon microneedles have been
fabricated with individual needles measuring 150 mm
318 S.L. Tao, T. A. Desai / Advanced Drug Delivery Reviews 55 (2003) 315–328

in length, 80 mm in diameter at the base, with a
radius of curvature less than 1 mm [14]. Hollow
needles have also been microfabricated with similar
dimensions, but containing hollow bores anywhere
from 5 to 70 mm in diameter, depending on design
[15]. To test their durability, the solid needles were
inserted into skin with gentle pushing, an approxi-
mate force of 10 N. All but a few percent of the
microneedles remained intact. Even in these few
needles, only the top 5–10 mm was damaged
[14,15]. Additionally the array of microneedles could
also be removed without additional damage and
could also be reinserted into skin multiple times.
4.2. Transdermal transport studies
Quantification of transdermal transport of various
molecules with and without inserted microneedle
arrays was used to assess any increase in skin
permeability leaving (Fig. 1). Insertion of the mi-
croneedles increased permeability only 1000-fold
because the microneedles or the silicon plate may
have blocked access to the microscopic holes [14].
When the microneedles were removed after 10 s,
permeability increased by 10 000-fold [14]. Removal
after 1 h increased skin permeability by 25 000-fold
[14]. Elevated permeability after microneedle inser-
tion was found to remain at approximately the same
level for as long as 5 h [14]. Hollow microneedles
were also capable of insertion into the skin without
any extensive damage to the microneedles or skin
[15]. In addition, the improved design of these
needles increased skin permeability further still [15].
4.3. Other microneedles
More recently, the original microneedle design has
been further refined to provide better control over
drug delivery. Silicon microhypodermic needles have
been fabricated in combination with heat-controlled
bubble pumps [15]. Hollow metal microneedles have
been fabricated by defining molds in epoxy and
filling them by electrodepositing metal [15]. Similar-
ly, polysilicon microneedles have been fabricated
with reusable molds [15,16]. Polysilicon micronee-
dles are likely to be most cost effective and have the
potential to produce single use disposable platforms
[15]. These types of needles, combined with a
pressurized reservoir to generate a drug delivery
pump, have already been incorporated into a wear-
able drug infusion system to deliver insulin [17].
Furthermore, by modifying needle dimension and
design to incorporate multiple channels and ports,
optimized microhypodermic needles and micro-
probes can be developed for cellular, local tissue, or
systemic delivery [15].

Fig. 1. (A) Scanning electron micrograph of microneedles made by reactive ion etching technique. (B) Microneedle tips inserted across
epidermis. The underside of the epidermis is shown, indicating that the microneedles penetrated across the tissue and that the tips were not
damaged. Arrows indicate some of the microneedle tips. Reproduced with permission from Ref. [14].
 S.L. Tao, T. A. Desai / Advanced Drug Delivery Reviews 55 (2003) 315–328
5. Implanted microchip for localized drug
delivery
Microfabrication technology has also created a
new class of controlled release systems for drug
delivery based on programmable devices. These
devices are particularly intriguing due to their small
size, potential for integration with microelectronics
and their ability to store and release chemicals on
demand [18]. With the recent advancements in
biosensors and micromachining, implanted respon-
sive drug release systems are becoming more plaus-
ible.
5.1. Microchip design
The first experimental demonstration of a microch-
ip with potential application in drug delivery was
described in Nature [19]. The ultimate goal was to
develop a microfabricated device devoid of moving
parts, but with the ability to store and release
multiple chemical substances. The device was fabri-
cated by the sequential processing of a silicon wafer
using microelectronic processing techniques includ-
ing UV photolithography, chemical vapor deposition,
electron beam evaporation and reactive ion etching
[19]. The experimental prototype was a 17 mm 3 17
mm 3 310 mm square silicon device containing an
array of 34 square pyramidal reservoirs etched
completely through the wafer [18,19].
5.2. Irreversible metallic valves
The 25-nl reservoirs were sealed at one end by a
thin membrane of gold to serve as an anode in an
electrochemical reaction [18,19]. One other electrode
was placed on the device to serve as a cathode. The
reservoirs were filled through the open end with the
chemical to be released by either microsyringe
pumps or inkjet printing in conjunction with a
computer-controlled alignment apparatus [18,19].
The open ends of the reservoirs were then covered
with a thin adhesive plastic and sealed with water-
proof epoxy [18,19]. When submerged in an elec-
trolyte, ions form a soluble complex with the anode
material in its ionic form [18]. An applied electric
potential oxidizes the anode membrane, forming a
soluble complex with the electrolyte ions [18]. The
319
complex dissolves in the electrolyte, the membrane
disappears, and the chemical is released and allowed
to diffuse form the reservoir.
The time at which release occurs from each
individual reservoir is determined by the time at
which the reservoir’s anode membrane is removed
[18]. Each reservoir, or a group of reservoirs, may be
independently addressed by demultiplexing [18].
This allows each anode to have its own conducting
path and electric potential can be applied to any
given combination of reservoirs at any given time
[18,19]. However, the rate of release from the
reservoir is a function of the dissolution rate of the
materials in the reservoir and the diffusion rate of
these materials out of the reservoir [18]. Therefore,
the rate of release from an individual reservoir can
be controlled by proper selection of the materials
(e.g. pure drugs, or drugs with polymers) placed
inside the reservoir [18]. Using a material that
quickly dissolves once the reservoir is opened can be
used to achieve pulsatile release whereas a material
that dissolves slowly after the reservoir is opened can
be used to achieve sustained release (Fig. 2) [18].
5.3. Reversible polymeric valves
An alternative to the use of irreversible metallic
valves is a microchip using reversible polymeric
valves. The use of ‘artificial muscle’ valves in
conjunction with silicon micromachined drug release
structures can render a microchip responsive to a
patient’s therapeutic requirements and deliver certain
amount of a drug in response to a biological stimulus
[20]. ‘Artificial muscle’ refers to a chemomechanical
actuator consisting of a blend of a hydrogel and an
electronically conducting redox polymer [21]. The
redox polymer is sensitive to pH, applied potential,
and the chemical potential of its microenvironment
whereas the hydrogel provides a cross-linked net-
work of hydrophilic homo / copolymers that exhibit
dramatic swelling and shrinking upon changes in pH,
solvent, temperature, electric field, or ambient light
conditions [21]. By electropolymerizing these poly-
mers onto electrodes, reservoirs can be opened or
closed, and the drug compound released or retained,
via the swelling and shrinking processes of the
polymer system in response to electrochemical actua-
tion (Fig. 3) [20].
320 S.L. Tao, T. A. Desai / Advanced Drug Delivery Reviews 55 (2003) 315–328

Fig. 2. Photographs of a prototype microchip: the electrode-containing front side and the back side with openings for filling the reservoirs.
Scale bar 10 mm. Reproduced with permission from Ref. [18].

5.4. Current developments
Current developments on microchip systems con-
sist of integrating active components. Combining
battery clocks, reference electrodes and biosensors
with the microchip could provide a single package
for implantation. Another logical extension of the
controlled-release microchip would be the develop-
ment of a passive, polymer microchip that contains
no electronics, power sources, or microprocessors
[18]. Each of the reservoirs would be covered by a
cap of degradable material, or a nondegradable
material of known permeability for the drug, or left
uncapped. Time and rate of release from the reser-
voir would then be dictated by the degradation rate
of the cap or diffusion of the drug [18]. This type of
polymeric microchip device would have the addition-
al advantage of being biodegradable, and once
implanted for drug delivery applications need not be
removed.
6. Bioadhesive microparticles for oral drug
delivery
Oral drug delivery is one of the most preferred
methods of drug administration due to its non-inva-
sive nature. However, it is generally not a viable
method for peptide and protein delivery. The human
GI tract resists absorption of peptides, proteins, and
other large molecules until they are broken down
into smaller molecules. The acidic environment of
the stomach combined with an array of enzymes and
physical barriers in the intestines either destroy or
prevent absorption of nearly all macromolecules.

Fig. 3. (A) A Ag /AgCl and IrO x valve electrode in the same micromachined drug delivery cavity. Both electrodes are 30 3 30 mm. (B)
SEM micrograph of ‘artificial muscle’ grown on TEM gold grid coated with poly-HEMA in holes (38.5 mm 3 38.5 mm) of a drug reservoir.
Reproduced with permission from Ref. [20].
 S.L. Tao, T. A. Desai / Advanced Drug Delivery Reviews 55 (2003) 315–328
This has led to the development of oral delivery
systems that can potentially enhance the delivery of
peptides utilizing mechanisms such as use of protec-
tive coatings [22], targeted delivery [23], permeation
enhancers [24], and protease inhibitors [25].
Bioadhesive drug delivery systems have also
generated considerable interest due to their potential
for prolonging the residence time at the site of drug
action or adsorption. Localization of the delivery
system at a given target site would intensify its
contact with the mucosal epithelial barrier, thereby
increasing the drug concentration gradient due to
intense contact [26,27]. Rather than having an im-
planted controlled-release microchip for local drug
delivery, such systems could also be fabricated for
specific targeting. By developing inorganic or poly-
meric reservoir-containing particles on the micron
scale, and grafting bioadhesive agents on their
releasing side, these particles could be adapted for
use as a bioadhesive controlled release oral drug
delivery system. This type of system could improve
the effectiveness of treatment by first, targeting and
localizing a drug at a specific site, inhibiting dilution
of the drug in body fluids, thereby helping to
maintain the drug concentration at the optimal
concentration between effective and toxic levels.
Micromachined platforms, when combined with
complementary approaches, may address some of the
shortcomings of current oral delivery systems for
peptides and proteins by combining several features
into a single drug delivery platform. First, one can
achieve control over the size and shape of the
delivery device. Unlike other spherical drug delivery
particles, microfabricated devices may be designed
to be flat, thin, and disc-shaped to maximize contact
area with the intestinal lining and minimize the side
areas exposed to the constant flow of liquids through
the intestines. The size of the particles can be
selected to be small enough to have good contact
with the undulations of the intestinal wall and large
enough to avoid endocytosis of the entire particle.
While endocytosis of nanoparticles has been pro-
posed as a method to enhance transport of large
molecules across the intestinal barrier, this process
can destroy the macromolecule. Secondly, one can
selectively attach bioadhesive agents onto the device
surface using relatively simple surface chemical
modification strategies. Finally, micromachining pre-
321
sents the opportunity to create multiple reservoirs of
desired size to contain not just one, but many drugs /
biomolecules of interest [28].
6.1. Silicon dioxide microparticles
Microparticles have been successfully fabricated
from silicon dioxide using microfabrication protocols
[28]. These particles are then easily modified by
silane chemistry to introduce to sites for the attach-
ment of biological molecules.
6.1.1. Fabrication of microparticle body
To fabricate the particle body an etch stop layer
was created by growing a thermal oxide under wet
conditions. Low-pressure chemical vapor deposition
was then used to deposit a sacrificial layer of poly-
crystalline silicon atop the thermal oxide by low-
pressure chemical vapor deposition. Next, a layer of
low temperature silicon dioxide (LTO) was depos-
ited to form the device layer. Negative lithography
was carried out to mask define the particle shape. A
reactive ion etch (RIE) with SF 6 and O 2 was used to
define the actual LTO particles and any remaining
photoresist was then removed in negative photoresist
remover.
6.1.2. Fabrication of microparticle reservoir
Positive lithography was carried out using infrared
(IR) backside alignment to define the wells. The
wafers were then time-etched in buffered oxide
etchant to carve out the wells. Any remaining
photoresist was removed. The welled microdevices
were then released into solution by etching the
sacrificial polysilicon layer with KOH. The KOH
solution was diluted with deionized water and fil-
tered to isolate the micro devices. As seen in Fig.
4A, the particles were fabricated with well-defined
features across the entire wafer. Released microparti-
cles are shown in Fig. 4B. The particles were
uniform and semi-transparent due to their poly-
crystalline nature.
6.1.3. Surface modification chemistry
Chemistries used to link biologically active mole-
cules, such as proteins, on to the releasing side of the
silicon dioxide microparticles have also been de-
veloped using traditional silane chemistry and car-
322 S.L. Tao, T. A. Desai / Advanced Drug Delivery Reviews 55 (2003) 315–328
Fig. 4. (A) An array of welled SiO 2 microparticles. (B) Released microparticles.
bodiimide coupling reagents. First, amine groups are
formed on the surface of unreleased microparticles
through hydroxylation of the silicon, followed by
silanization in 2% v / v APTES at room temperature
for at least 1.5 h. Biological molecules that contain
carboxyl groups were coupled with carbodiimide and
N-hydroxysuccinimide to form a reactive inter-
mediate ester that is susceptible to attack by amines.
This reaction can be applied to self-assemble a
monolayer of covalently coupled biological mole-
cules to the surface of the microparticles.
6.2. Poly(methyl methacrylate) microparticles
Such a system has also been developed based on
the polymer polymethylmethacrylate (PMMA).
PMMA is an ideal material for such a MEMS-based
bioadhesive system because it is biocompatible and
already used in many biomedical applications, is
commonly used as a resist in photolithographic
applications, and contains a functional methyl ester
group for potential surface modification. The current
prototype is a square particle 150 mm across and 3
mm thick containing square reservoirs 80 mm across
and 2 mm deep.
6.2.1. Fabrication of microparticle body
To fabricate these particles, PMMA was spun on
to clean silicon wafers. Positive lithography was
used to expose the area between the particles and
isolate the particle bodies. The unmasked area was
then etched completely through in the reactive ion
etcher using an O 2 plasma and any remaining resist
was removed. The thickness of the particles can be
modified by changing the spin rate or by spinning on
additional layers of PMMA. Additionally, altering
the masked area of the wafer easily changes the
shape and surface area of the particles.
6.2.2. Fabrication of microparticle reservoir
A second positive lithography was carried out to
expose the intended reservoir areas in the PMMA.
This area was then etched 1–2 mm deep using an
oxygen plasma and any remaining photoresist was
removed. The dimensions of the reservoir can be
altered by changing the masked area and their depth
can be modified by changing the time and / or flow
rate of plasma in the RIE (Fig. 5). By creating
smaller wells, a series of multiple wells can be
etched into the particles to create separate reservoirs
for a combination of drugs or permeation enhancers.
Since the PMMA is adherent to the surface of silicon
by linkage to the native oxide layer, the wafer was
soaked in basic solution to break this bonding and
immediately release the particles.
6.2.3. Surface modification chemistry
Heterogeneous modification of PMMA with N-
lithioethylenediamine as the aminolyzing agent leads
to amination of its ester groups [29,30]. This reaction
layer of amine sites tethered to the PMMA backbone
by stable amide groups. Amine groups were placed
on the surface of the releasing side of PMMA
microparticles by reacting unreleased microparticles
with N-lithioethylenediamine for 10 min. Biological
molecules were linked to these amine sites using the
same traditional carbodiimide coupling reagents used
for silicon dioxide microparticles.
 S.L. Tao, T. A. Desai / Advanced Drug Delivery Reviews 55 (2003) 315–328
Fig. 5. Arrays of 150 mm 3 150 mm PMMA particles with (A) 50 mm 3 50 mm, (B) 80 mm 3 80 mm, (C) 100 mm 3 100 mm, and (D)
multiple 28 mm 3 28 mm wells.
6.3. Release mechanism
These reservoirs can be filled with pico- to
nanoliters of a polymeric solution with microinject-
ors. Water quickly evaporates from these reservoirs
leaving behind the drug contained in polymer which
acts as a timed-release plug. Using a specific type of
polymer predetermines the time and rate of release of
drug from the reservoir; for example, a hydrogel that
swells in response to a specific pH, solvent or
temperature or a polymer with a known dissolution
rate. Different polymers with various dissolution
rates can be used in each reservoir to obtain con-
trolled release of separate compounds.
6.4. Microparticle bioadhesion to intestinal
epithelium
The current microparticle model is designed to
specifically target the intestinal epithelium. The
microparticles were first rendered bioadhesive by
forming avidin-coupled surfaces using the chemistry
previously described. Commercially available
biotinylated lectins were then attached to the mi-
croparticle surface by taking advantage of the strong
interaction and affinity of avidin and biotin in nature.
323
6.4.1. Lectins
Lectins are a class of carbohydrate binding pro-
teins or glycoproteins of non-immune origin. They
have been deemed a ‘second generation’ bioadhesive
because they bind to cell-surface glycoconjugates in
a complementary way analogous to ligand–receptor
interactions. Tomato lectin is of special interest in
intestinal targeting due to its stability in low pH
environments and its non-toxicity. Furthermore, it
has been shown to bind selectively to the small
intestine epithelium [31], resist digestion in the
alimentary canal of rats and bind to rat intestinal villi
without disruption of the lectin integrity [32].
6.4.2. Caco-2 binding
In vitro studies using a Caco-2 model of the
intestinal epithelium were used to demonstrate the
specific binding of the tomato lectin-conjugated
PMMA particles. The tomato lectin-conjugated par-
ticles showed, on average, a marked increase in
binding almost five times greater than the binding of
unmodified particles over an entire period of 2 h
(Fig. 6). Furthermore, the modified particles once
bound, remained bound whereas only 20% of origi-
nally bound unmodified particles remained bound
after 20 min. Currently, in vivo studies are being
324 S.L. Tao, T. A. Desai / Advanced Drug Delivery Reviews 55 (2003) 315–328
Fig. 6. Binding of unmodified and lectin-conjugated PMMA
microparticles to Caco-2 monolayers.
7.1. Allotransplantation
conducted to determine the bioadhesive properties of
these particles in the gastrointestinal tract of rats.
6.5. Future developments
By replacing the molecule attached to the mi-
croparticles, an array of cells and tissues can be
targeted. For example, if the lectin is substituted with
an antibody that selectively binds to tumor cells in
the colon, the microparticles could actively seek out
cancerous masses in the colon and deliver anticancer
drugs directly. This would allow the high concen-
tration of drug to be locally delivered while keeping
the systemic concentration at a low level. Moreover,
microfabricated polymer particles with the same
targeting abilities, but small enough for injection,
could be developed for direct delivery into the
circulatory system.
7. Nanoporous immunoisolating biocapsules
Diabetes mellitus (DM) represents a serious medi-
cal problem. In the US alone, it is the third leading
cause of death. While the majority of patients have
type 2 diabetes, about 10% of all patients diagnosed
with DM are insulin-dependent (type 1). In both
cases, disease is caused by decreased circulating
concentrations of insulin and decreased response of
peripheral tissue to insulin (insulin resistance) [33].
The disease manifests itself as hyperglycemia. In-
sulin remains the mainstay of virtually all type 1 DM
and many type 2 DM patients and in most cases is
administered subcutaneously. However, the kinetics
of insulin administered by this route do not mimic
the normal rapid rise and decline of insulin secretion
in response to ingested nutrients.
Efforts to address the short-comings of current
subcutaneous administration of insulin, including the
use of complex multidose regimens, have led to the
development of other dosage forms and routes of
administration such as ‘needleless’ injectors, con-
stant infusion pumps, and inhaled insulin. These
newer approaches still suffer from the same general
issue plaguing current subcutaneous administration.
A potentially useful approach, which has proven
effective in only a handful of cases, is the allotrans-
plantation of islets or whole pancreases from a
suitable human donor into a diabetic recipient.
Researchers in Canada recently reported successful
transplantation of islet cells in type 1 DM patients
[34]. Although the potential complications of im-
munosuppressive therapy were reduced by avoiding
the use of glucocorticoids, each transplant required
two harvestings of islet cells from organ donors.
Moreover recipients are still required to take immune
suppressing drugs for the rest of their lives. These
immunosuppressive drugs are toxic and have po-
tential adverse side effects, including cancer. For this
reason, an islet or pancreas allotransplant is normally
carried out only in conjunction with a kidney trans-
plant, for which immunosuppression is required in
any case.
Because of the toxicity of immune suppressing
drugs, and the shortage of organ donors, islet and
pancreas allotransplantation appears to hold limited
promise as a cure for diabetes. A method then is
required to sequester the islets from the body’s
immune system which is able to recognize and reject
these xenogeneic cell grafts. For the past 20 years,
investigators have focused on a range of microen-
capsulation methods most commonly involving so-
dium alginate and another polycationic substance
such as polylysine. These materials have been used
in an attempt to create a semipermeable membrane
 S.L. Tao, T. A. Desai / Advanced Drug Delivery Reviews 55 (2003) 315–328 325

capable of blocking immune molecules such as IgG,

cytokines, and cell-secreted antigens from reaching
the encapsulated xenogeneic islet cells while allow-
ing glucose and insulin to freely diffuse through the
barrier [35]. However, this approach has proven
generally unsuccessful due to mechanical rupture of
the membrane, biochemical instability, incompatibili-
ty with islet cell heterogeneity, and broad pore size
distributions [35–39]. When the barrier between the
xenogeneic cells and the external bioenvironment is
compromised, these foreign cells are subject to
various endogenous cells and antibodies as well as
complement and a host of cytokines such as tumor
necrosis factor, all of which can inflict cell damage.
As a result, the use of polymeric microcapsules for
allotransplantation has been unsuccessful clinically
in the absence of immunosuppression [35,37,39].

7.2. Biocapsule design

Microfabrication techniques have been applied to

create a biocapsule for effective immunoisolation of
transplanted islet cells for treatment of diabetes [40].
The fabrication of nanochannels in the membrane Fig. 7. Micrograph of a biocapsule membrane with 24.5-nm
structure consists of two steps: (1) surface mi- pores.
cromachining nanochannels in a thin film on the top
of a silicon wafer, and (2) releasing the membrane chemically, and mechanically stable and retrievable.
by etching away the bulk of the silicon wafer It is also expected that improved dynamic response
underneath the membrane. These nanopore mem- of islets can be obtained due to the limited mem-
branes (Fig. 7) are designed to allow the permeabili- brane thickness (Fig. 8) compared with the thickness
ty of glucose, insulin, and other metabolically active of conventional polymeric membranes prepared from
products, while at the same time, preventing the alginate and polylysine (100–200 mm) [41]. It is
passage of cytotoxic cells, macrophages, antibodies, crucial that rapid secretion kinetics, particularly in
and complement. The membranes are bonded to a
capsule that houses the pancreatic islet cells. Because
the difference in the size of insulin, which must be
able to pass freely through the pores and the size of
IgG immunoglobulins, which must be excluded, is
only a matter of a few nanometers, the highly
uniform pore size distribution provided by mi-
cromachine membranes is essential for effective
immunoisolation and therapeutic effect.
Control of pore sizes in the tens of nanometers has
recently been suggested as probably the most realis-
tic way to achieve immunoisolation [41,42]. The use
of unconventional biomaterials such as silicon and
silicon dioxide provides a means to encapsulate
pancreatic islet cells in devices that are thermally, Fig. 8. Insulin secretory profile through differing pore sizes.
326 S.L. Tao, T. A. Desai / Advanced Drug Delivery Reviews 55 (2003) 315–328
the first phase of insulin release, be maintained over
time to provide physiologic feedback control of
blood glucose concentrations.
Another key feature of an implanted biocapsule
system is the role of neovascularization. The mem-
branes proposed for testing have outer openings of 2
by 2 mm while the inner diffusion channels have a
pore size of between 10 and 30 nanometers. Studies
have shown that neovascularization at the mem-
brane–tissue interface occurs in membranes having
pore sizes large enough to allow complete penetra-
tion by host cells (0.8–8 mm) [41]. Thus, it is
expected that neovascularization can occur at the
large openings while not penetrating into the
nanometer pores. This phenomenon has two key
advantages: (1) the ability to rapidly deliver insulin
into the blood stream through new blood vessel
growth while (2) limiting pore clogging or fouling.
7.3. Biocompatibility and cytotoxicity studies
Preliminary biocompatibility and cytotoxicity tests
have been carried out by examining the cell mor-
phology, growth, and function of test cell lines
placed in contact with arrays of membranes, with
promising preliminary results. The biocompatibility
was evaluated via direct contact tests by cultivating
several different cell lines such as macrophages,
fibroblasts, and HeLa cells, as well as isolated
primary islets of Langerhans both on the wafer
surface and within the porous wafer pockets. All
cells were seeded on silicon culture wafers, observed
via light microscope, stained for cell viability and
functionality, and counted with a hemocytometer. All
cell types had normal growth characteristics, mor-
phology, and greater than 90% viability [43,44].
Results indicate that the insulin secretion by
encapsulated islets and subsequent diffusion through
the biocapsule membrane channels is similar to that
of unencapsulated islets for both 3 mm and 78 nm
pore sized membranes, with insulin diffusion though
the membrane occurring within 10 min of stimula-
tion. Fig. 8 shows the typical insulin release profile
in response to stimulatory (16.7 mM) glucose
medium over 1 h under static incubation for 78-, 66-
and 18-nm pore-sized membranes. This profile indi-
cates that insulin and glucose diffusion occur at
sufficiently high rates through the microfabricated
Fig. 9. IgG diffusion through microfabricated biocapsules of three
different pore sizes.
membrane to ensure nutrient exchange for encapsu-
lated islet cells. These experiments show that no
diffusion barrier is formed by the membrane for
glucose and insulin, while taking into account the
effect of rotation on mass transfer. In addition,
microfabricated biocapsule membranes can be tailor
made to attain desired IgG diffusion kinetics (Fig.
9). At the same time, the deselection of IgG requires
absolute pore dimensions below 18 nm. It is noted
that the percent of IgG diffusion (concentration of
IgG that passes through the membrane) was less than
0.4% after 24 h and 2% after over 150 h through the
18-nm membranes [44–46].
It may be possible to design nanopore membranes
which achieve a more constant rate of drug delivery,
avoiding the ‘burst effect’. By precisely controlling
pore size, pore length and pore density, the nanopore
membrane fitted into a polymeric capsule suitable for
subcutaneous implantation can serve as a diffusion
barrier for a variety of biological drugs.
8. Conclusions
The race to find effective diagnostic and therapeu-
tic tools is under way, as scientific and engineering
disciplines uncover and elucidate more about the
human pathologic condition than ever before. Al-
though we are getting closer to the clinical applica-
tion of intelligent drug delivery devices, many
challenges remain for the future. The convergence of
microtechnology and biology will lead to new ap-
proaches in drug delivery and may provide advan-
 S.L. Tao, T. A. Desai / Advanced Drug Delivery Reviews 55 (2003) 315–328
tages over existing technologies. By focusing efforts
at the microscale, we have the unique ability to
engineer control over the cellular environment, lead-
ing to novel ways in which we can control molecular
delivery and cell / tissue interactions. The future
challenge lies in assembling and applying our collec-
tive knowledge to develop functional and clinically
relevant therapeutic delivery devices.
Acknowledgements
Funding is gratefully acknowledged from The
Whitaker Foundation, NSF ECS9820829, NSF
Career, and iMEDD, Inc. Also, special thanks to
those who have contributed to this work: Lara Leoni,
Mike Lubeley, Chris Bonner, and Aamer Ahmed of
UIC; colleagues from iMEDD, Inc.; and Professors
Derek Hansford and Mauro Ferrari from OSU.
References
[1] P.L. Gourley, Semiconductor microlasers: a new approach to
cell-structure analysis, Nat. Med. 2 (1996) 942–944.
[2] J. Drews, Drug discovery: a historical perspective, Science
287 (2000) 1960–1964.
[3] D.J. Anderson, K. Najafi, S.J. Tanghe, D.A. Evans, K.L.
Levy, J.F. Hetre, X. Xue, J.J. Zappia, K.D. Wise, Batch-
fabricated thin-film electrodes for stimulation of the central
auditory system, IEEE Trans. Biomed. Eng. 36 (1989) 693–
704.
[4] J. Evans, D. Liepmann, A.P. Pisano, Planar laminar mixer,
Proc. IEEE MEMS Workshop 10 (1997) 96–101.
[5] J. Deutsch, D. Motlagh, B. Russell, T.A. Desai, Fabrication
of microtextured membranes for cardiac myocyte attachment
and orientation, J. Biomed. Mater. 53 (2000) 267–275.
[6] M.L. Reed, C. Wu, J. Kneller, S. Watkins, D.A. Vorp, A.
Nadeem, L.E. Weiss, K. Rebello, M. Mescher, A.J.C. Smith,
W. Rosenblum, M.D. Feldman, Micromechanical devices for
intravascular drug delivery, J. Pharm. Sci. 87 (1998) 1387–
1394.
[7] D.D. Breimer, Future challenges for drug delivery, J. Con-
trolled Release 62 (1999) 3–6.
[8] J.T. Santini, A.C. Richards, R.A. Scheidt, M.J. Cima, R.S.
Langer, Microchip technology in drug delivery, Ann. Med.
32 (2000) 377–379.
[9] S.S. Davis, L. Illum, Drug delivery systems for challenging
molecules, Int. J. Pharm. 176 (1998) 1–8.
[10] D.L. Polla, A.G. Erdman, W.P. Robbins, D.T. Markus, J.
Diaz-Diaz, R. Rizq, Y. Nam, H.T. Brickner, Microdevices in
medicine, Annu. Rev. Biomed. Eng. 2 (2000) 551–576.
327
[11] J. Hadgraft, R.H. Guy (Eds.), Transdermal Drug Delivery:
Developmental Issues and Research Initiatives, Marcel
Dekker, 1989.
[12] E.W. Smith, H.I. Maibach (Eds.), Precutaneous Penetration
Enhancers, CRC Press, 1995.
[13] B.G. Amsden, M.F.A. Goosen, Transdermal delivery of
peptide and protein drugs: an overview, AIChE J. 41 (1995)
1972–1997.
[14] S. Henry, D.V. McAllister, M.G. Allen, M.R. Prausnitz,
Microfabricated microneedles: a novel approach to transder-
mal drug delivery, J. Pharm. Sci. 87 (1998) 922–925.
[15] D.V. McAllister, M.G. Allen, M.R. Prausnitz, Microfabri-
cated microneedles for gene and drug delivery, Annu. Rev.
Biomed. 2 (2000) 289–313.
[16] D. Lieppmann, A.P. Pisano, B. Sage, Microelectromechanical
systems technology to deliver insulin, Diabetes Technol.
Ther. 1 (1999) 469–476.
[17] J.D. Zhan, A.A. Peshmukh, A.P. Pisano, D. Liepmann,
Continuous on-chip micropumping through a microneedle,
in: Int. 14th Conf. MEMS, 2001, pp. 503–506.
[18] J.T. Santini, A.C. Richards, R. Scheidt, M.J. Cima, R.
Langer, Microchips as controlled drug-delivery devices,
Angew. Chem. Int. Ed. Engl. 39 (2000) 2396–2407.
[19] J.T. Santini, M.J. Cima, R. Langer, A controlled-release
microchip, Nature 397 (1999) 335–338.
[20] L. Low, S. Seetharaman, K. He, M.J. Madou, Microactuators
toward microvalves for responsive controlled drug delivery,
Sensors Actuators B Chem. 67 (2000) 149–160.
[21] M. Madou, J. Florkey, From batch to continuous manufactur-
ing of microbiomedical devices, Chem. Rev. 100 (2000)
2679–2692.
[22] M. Saffran, G.S. Kumar, D.C. Neckers, J. Pena, R.H. Jones,
J.B. Field, Biodegradable azopolymer coating for oral deliv-
ery of peptide drugs, Biochem. Soc. Trans. 18 (1990) 752–
754.
[23] J.W. Fara, R.E. Myrback, D.R. Swanson, Evaluation of
oxprenolol and metoprolol Oros systems in the dog: com-
parison of in vivo and in vitro drug release, and of drug
absorption from duodenal and colonic infusion sites, Br. J.
Clin. Pharmacol. 19 (1985) 91S–95S.
[24] A. Fasano, S. Uzzau, Modulation of intestinal tight junctions
by Zonula occludens toxin permits enteral administration of
insulin and other macromolecules in an animal model, J.
Clin. Invest. 99 (1997) 1158–1164.
[25] U.I. Schwarz, T. Gramatte, J. Krappweis, R. Oertel, W.
Kirch, P-glycoprotein inhibitor erythromycin increases oral
bioavailability of talinolol in humans, Int. J. Clin. Pharmacol.
Ther. 38 (2000) 161–167.
[26] C.M. Lehr, Lectin-mediated drug delivery: the second gene-
ration of bioadhesives, J. Controlled Release 65 (2000)
19–29.
[27] G. Poncel, J. Irache, Specific and non-specific bioadhesive
particulate systems for oral delivery to the gastrointestinal
tract, Adv. Drug Deliv. Rev. 34 (1998) 191–219.
[28] A. Ahmed, C. Bonner, T.A. Desai, Bioadhesive mi-
crodevices with multiple reservoirs: a new platform for oral
drug delivery, J. Controlled Release 81 (2002) 291–306.
328 S.L. Tao, T. A. Desai / Advanced Drug Delivery Reviews 55 (2003) 315–328
[29] B. Karandikar, J. Puschett, K. Matyjaszewski, Homogeneous
and heterogeneous modification of poly(methyl methacrylate)
with ethylene diamine, Polym. Prep. Am. Chem. Soc. Div.
Polym. Chem. 30 (1989) 250–251.
[30] A.C. Henry, T.J. Tutt, M. Galloway, Y. Davidson, C.S.
McWhorter, S. Soper, R. McCarley, Surface modification of
poly(methyl methacrylate) used in the fabrication of mi-
croanalytical devices, Anal. Chem. 72 (2000) 5331–5337.
[31] B. Carreno-Gomez, J.F. Woodley, A.T. Florence, Studies on
the uptake of tomato lectin nanoparticles in everted gut sacs,
Int. J. Pharm. 183 (1999) 7–11.
[32] D. Kilpatrick, A. Pusztai, G. Grant, C. Graham, S. Ewen,
Tomato lectin resists digestion in the mammalian alimentary
canal and binds to intestinal villi without deleterious effects,
FEBS Lett. 185 (1985) 299–305.
[33] S.N. Davies, D.K. Granner, The Pharmacological Basis of
Therapeutics, McGraw-Hill, 1996.
[34] S.S. Davies, Biomedical applications of nanotechnology—
implications for drug targeting and gene therapy, Trends
Biotechnol. 15 (1997) 217–224.
[35] R.P. Lanza, D.K. Cooper, Xenotransplantation and cell
therapy: progress and controversy, Mol. Med. Today 5
(1999) 105–106.
[36] C.K. Colton, E.S. Avgoustiniatos, Bioengineering in de-
velopment of the hybrid artificial pancreas, ASME J.
Biomech. Eng. 113 (1991) 152–170.
[37] P. Lacy, O.D. Hegre, A. Gerasimidi-Vazeou, F.T. Gentile,
K.E. Dionne, Maintenance of normoglycemia in diabetic
mice by subcutaneous xenograft of encapsulated islets,
Science 254 (1991) 1728–1784.
[38] P. Soon-Shiong, M. Otterie, G. Skjak-Braek, O. Smidsrod, R.
Heintx, R.P. Lanza, T. Espevik, An immunologic basis for
the fibrotic reaction to implanted microcapsules, Transplant.
Proc. 23 (1991) 758–759.
[39] R.P. Lanza, J.L. Hayes, W.L. Chick, Encapsulated cell
technology, Nat. Biotechnol. 14 (1996) 1107–1111.
[40] T.A. Desai, W.H. Chu, J.K. Tu, G.M. Beattie, A. Hayek, M.
Ferrari, Microfabricated immunoisolating biocapsules,
Biotechnol. Bioeng. 57 (1998) 118–120.
[41] M. Brissova, I. Lacik, A.C. Powers, A.V. Anilkumar, T.
Wang, Control and measurement of permeability for design
of microcapsule cell delivery system, J. Biomed. Mater. Res.
39 (1998) 61–70.
[42] T. Wang, I. Lacik, M. Brissova, A.V. Anilkumar, A. Prokop,
D. Hunkeler, R. Green, K. Shahrokhi, A.C. Powers, An
encapsulation system for the immunoisolation of pancreatic
islets, Nat. Biotechnol. 15 (1997) 358–362.
[43] T.A. Desai, D. Hansford, M. Ferrari, Characterization of
micromachines membranes for immunoisolation and bio-
separation applications, J. Biomed. Microdev. 4132 (1999)
1–11.
[44] T.A. Desai, W.H. Chu, G. Rasi, P.S. Vallebona, E. Guarino,
M. Ferrari, Microfabricated biocapsules provide short-term
immunoisolation of insulinoma xenografts, Biomed. Mi-
crodev. 1 (1999) 131–181.
[45] T.A. Desai, D.J. Hansford, L. Leoni, M. Essenpreis, M.
Ferrari, Nanoporous anti-fouling silicon membranes for
implantable biosensor applications, Biosens. Bioelectron. 15
(2000) 453–462.
[46] L. Leoni, T. Boriarski, T.A. Desai, Characterization of
nanoporous membranes for immunoisolation: diffusion prop-
erties and tissue effects, J. Biomed. Microdev. 4 (2002)
131–139.