Polymer Microneedles for Controlled-Release Drug Delivery
Jung-Hwan Park,1Mark G. Allen,2and Mark R. Prausnitz1,3,4
Received December 2, 2005; accepted January 11, 2006
Purpose. As an alternative to hypodermic injection or implantation of controlled-release systems, this
study designed and evaluated biodegradable polymer microneedles that encapsulate drug for controlled
release in skin and are suitable for self-administration by patients.
Methods. Arrays of microneedles were fabricated out of poly-lactide-co-glycolide using a mold-based
technique to encapsulate model drugsVcalcein and bovine serum albumin (BSA)Veither as a single
encapsulation within the needle matrix or as a double encapsulation, by first encapsulating the drug
within carboxymethylcellulose or poly-L-lactide microparticles and then encapsulating drug-loaded
microparticles within needles.
Results. By measuring failure force over a range of conditions, poly-lactide-co-glycolide microneedles
were shown to exhibit sufficient mechanical strength to insert into human skin. Microneedles were also
shown to encapsulate drug at mass fractions up to 10% and to release encapsulated compounds within
human cadaver skin. In vitro release of calcein and BSA from three different encapsulation formulations
was measured over time and was shown to be controlled by the encapsulation method to achieve release
kinetics ranging from hours to months. Release was modeled using the Higuchi equation with good
agreement (r2Q 0.90). After microneedle fabrication at elevated temperature, up to 90% of
encapsulated BSA remained in its native state, as determined by measuring effects on primary,
secondary, and tertiary protein structure.
Conclusions. Biodegradable polymer microneedles can encapsulate drug to provide controlled-release
delivery in skin for hours to months.
KEY WORDS: controlled-release drug delivery; microneedles; protein stability; transdermal drug
Conventional drug delivery using pills or injection is
often not suitable for new protein, DNA, and other therapies
(1,2). Devices for controlled release of such compounds have
been developed, which enable slow delivery over hours to
years. Controlled release is often achieved by encapsulating
drugs within biodegradable polymer matrices, from which
release is governed by drug diffusion and polymer erosion.
Decades of research on this topic have yielded clinical
products, such as the Lupron Depot, which delivers leupro-
lide acetate systemically for months (3), and the Gliadel
wafer, which administers carmustine locally to the brain for
days to weeks (4).
A limitation, however, of controlled-release systems is
that they typically require hypodermic needle injection of
polymeric microparticles or possibly surgical implantation of
macroscopic devices within the body. These painful and
invasive procedures are generally not suitable for self-
administration by patients and therefore are limited to use in
hospitals or clinics.
The goal of this study was to develop a minimally
invasive polymeric controlled-release system suitable for self-
administration without the pain or complexity of current
controlled-release devices. Rather than using a hypodermic
needle to introduce polymeric microparticles into the body,
we propose redesigning the microparticles to have the shape
of microneedles and thereby give these polymeric particles
the functionality of both needles and drug matrices for
controlled release (Fig. 1). By integrally forming these
microscopic needles onto a patch substrate, arrays of drug-
loaded microneedles could be inserted into the skin and worn
like a transdermal patch for slow release over time. An al-
ternative approach would involve intentionally separating the
patch base from the needles after insertion into the skin,
thereby leaving the drug-filled needles invisibly buried in the
skin for slow release. Because these microneedles are made
of FDA-approved, biodegradable polymer, they should safely
disappear after drug delivery is complete. Previous studies
have shown that microneedles are painless (5,6).
0724-8741/06/0500-1008/0#2006 Springer Science + Business Media, Inc.
Pharmaceutical Research, Vol. 23, No. 5, May 2006 (#2006)
1Wallace H. Coulter Department of Biomedical Engineering at
Georgia Tech and Emory University, Georgia Institute of Technol-
ogy, Atlanta, Georgia 30332, USA.
2School of Electrical and Computer Engineering, Georgia Institute
of Technology, Atlanta, Georgia 30332, USA.
3School of Chemical and Biomolecular Engineering, Georgia In-
stitute of Technology, Atlanta, Georgia 30332, USA.
4To whom correspondence should be addressed. (e-mail: prausnitz@
Microneedles have previously been proposed and devel-
oped for related applications. Solid microneedles have been
used to pierce the skin for increased permeability (7) as well
as to provide a substrate on which drug can be coated (8) or
encapsulated (9) for rapid release. Using this approach, a
range of compounds has been delivered to the skin, including
proteins, such as insulin and human growth hormone; genetic
material, including plasmid DNA and oligonucleotides; and
vaccines directed against hepatitis B and anthrax (10,11).
Hollow microneedles have also been developed for infusion
of drug solutions into the skin (12Y14). However, we believe
that this is the first study to address the use of microneedles
to encapsulate drug for controlled-release delivery (15).
Guided by previous microneedle studies, controlled-
release microneedles should measure hundreds of microns in
length and have a radius of curvature less than 10 mm at the tip
to ensure easy penetration into skin by manual insertion (16).
Microneedles of this size can penetrate past the skin’s outer
barrier of stratum corneum and deliver drug to the epidermis
and superficial dermis, where drug can diffuse rapidly for
local delivery to skin or systemic distribution via uptake by
dermal capillaries. Through the use of biodegradable poly-
mers such as poly-lactide-co-glycolide (PLGA), well-estab-
lished controlled-release mechanisms can be exploited to
control release from microneedles (17,18).
Microneedles of the proposed dimensions can be made
by adapting microfabrication technology (19). Although
microfabrication often involves lithography and etching of
silicon, the field is being expanded to include laser cutting,
molding, and other fabrication techniques to produce micro-
devices made of other materials, including metals and
polymers. By leveraging these technologies of the microelec-
tronics industry, methods to make microneedles should
provide reproducible mass production at disposable cost.
MATERIALS AND METHODS
Fabrication of Biodegradable Microneedles
Fabrication of Microneedle Master Structures and Molds
Microneedles were fabricated by first making master
structures using lithography-based methods, then creating
inverse molds of these master structures, and finally prepar-
ing replicate microneedles by melting biodegradable polymer
formulations into the molds. In this way, one master structure
could be used to make multiple molds, which could each be
used to make multiple replicates. Microneedles were fabri-
cated using two different geometries: beveled tip and tapered
cone. Methods to fabricate these master structures and molds
have been described in detail previously (20) and are sum-
Beveled-tip microneedle master structures were fabri-
cated out of SU-8 epoxy using standard UV-lithographic
techniques (20). SU-8 epoxy (SU-8 100; MicroChem, New-
ton, MA, USA) was coated onto a silicon wafer and
lithographically patterned into cylinders in the shape of the
desired needles. The space between the cylinders was filled
with a sacrificial polymer (PLGA 85/15, Sigma-Aldrich, St.
Louis, MO, USA) and a copper mask was patterned to
asymmetrically cover the tops of the epoxy cylinders and
some of the sacrificial polymer on one side of each cylinder.
Reactive ion etching (RIE; Plasma Therm, St. Petersburg,
FL, USA) partially removed the uncovered sacrificial layer
and asymmetrically etched the tips of the adjacent epoxy
cylinders. All remaining sacrificial polymer was removed by
ethyl acetate, leaving an array of epoxy cylinders with
asymmetrically beveled tips. This array of needles was coated
with poly(dimethylsiloxane) (PDMS; Sylgard 184, Dow
Corning, Midland, MI, USA), which was subsequently peeled
off to make an inverse mold.
Tapered-cone microneedles were fabricated using a
novel microlens technique (Y.-K. Yoon, J.-H. Park, and M.
G. Allen. Multidirectional UV lithography for complex 3-D
MEMS structures. J MEMS, in press). A chromium layer was
first deposited and patterned on a glass substrate to form an
array of circular dots of exposed glass. Isotropic wet etching
of the exposed glass was then performed to create concave
wells, which were filled with SU-8 epoxy cast on the surface.
The refractive index mismatch between glass and SU-8 epoxy
created an array of integrated microlenses. After soft-baking,
the SU-8 film was exposed from the bottom (i.e., through the
glass) to UV light, which passed through the microlenses to
form latent images in the SU-8 epoxy as ray traces from the
lenses. After development of the SU-8 epoxy, the resulting
tapered-cone microneedle master structure was used to make
an inverse PDMS mold.
Fabrication of Microneedles Encapsulating Drug
To prepare microneedles encapsulating drug for con-
trolled release, PDMS microneedle molds were first filled
with a model drug formulation and then filled with a PLGA
melt, which was allowed to cool and solidify. Three for-
mulations were used to achieve different timescales of
controlled release. For rapid release, the model drug was
Fig. 1. Controlled-release drug delivery using polymer microneedles.
Polymeric controlled release is often achieved by encapsulating drug
within microparticles, which are then injected into the body using a
hypodermic needle (shown on left). Polymer microneedles can
similarly be designed to encapsulate drug for controlled release, but
can be directly inserted into the skin without the need for
hypodermic injection (shown on right).
1009Polymer Microneedles for Controlled Release
directly encapsulated within the microneedles. For slower
release, drug was first encapsulated either within carboxy-
methylcellulose (CMC) or poly-L-lactide (PLA), which was
then encapsulated within microneedles. The process is
summarized in Fig. 2.
For the first formulation, calcein or Texas-Red-labeled
bovine serum albumin (BSA) powder (used as received from
Sigma-Aldrich or Molecular Probes, Eugene, OR, USA,
respectively) was suspended in acetonitrile (Sigma-Aldrich)
at a solids content of 10% (w/v) and then homogenized for
5 min at 10,000 rpm (PowerGen 700 homogenizer, Fisher
Scientific, Pittsburgh, PA, USA) to make drug microparticles.
The homogenized particles, with a broad size distribution over
the approximate range of 1Y100 mm, were filtered first through
a 30-mm filter, and then the filtrate was passed through a 1-mm
filter (nylon net filter, Millipore, Billerica, MA, USA). The
final solids cake containing particles 1Y30 mm in size was
redispersed in acetonitrile at a solids content >20% (w/v). The
resulting suspension was poured onto a PDMS microneedle
mold and placed in a vacuum chamber at j20 kPa for õ5 min.
This filled the mold with drug particles by first allowing the
then evaporate off the organic solvent. Residual particles
remaining on the surface of the mold were removed using
adhesive tape (Blenderm, 3M, St. Paul, MN, USA). As
described previously (20), the mold was then filled with
melted PLGA (PLGA 50/50, 1.2 dL/g, Birmingham Polymer,
Birmingham, AL, USA) in a vacuum oven at 135-C and
j70 kPa for 10Y20 min. After cooling, the resulting micro-
needles with encapsulated drug were manually removed from
To retard release from microneedles using a double-
encapsulation formulation approach (21), calcein was first
encapsulated within CMC microparticles, which were then
encapsulated within microneedles. A CMC solution was
prepared by dissolving 0.25 g of CMC sodium salt (reference
viscosity of 400Y800 cP @ 2% aqueous solution; Sigma-
Aldrich) in 9.6 mL of deionized (DI) water for 12 h on a 50-C
hot plate with stirring at 300 rpm. Then, 25 mg calcein
(Sigma-Aldrich) was dissolved in the CMC solution at a
calcein: CMC ratio of 1:10 (w/w). The resulting clear
solution was poured onto aluminum foil and dried to remove
water for 6 h under j50 kPa of vacuum. The resulting film
containing calcein dispersed in a solid CMC matrix was
pulverized by an agate mortar and pestle to form particles
measuring a few hundred microns to a few millimeters in size.
These large particles were dispersed in acetonitrile, homoge-
nized, and filtered to yield particles of 1Y30 mm in size, as
described above. The average diameter of the particles was 9.6
mm, with a standard deviation of 6.2 mm, determined by
analyzing scanning electron microscope (SEM) images. The
small CMC particles encapsulating calcein were finally loaded
into a PDMS mold, which was subsequently filled with PLGA,
as described above, to form PLGA microneedles, which
encapsulated CMC particles that further encapsulated calcein.
To slow release still more, a similar approach was used,
where calcein was first encapsulated within PLA micropar-
ticles, which were then encapsulated within microneedles.
Using the well-known double-emulsion technique to make
PLA microparticles (22), 50 mg of calcein was dissolved in 15
mL DI water and 0.2 g of PLA (L-PLA, 1.0 dL/g; Birming-
ham Polymer) was separately dissolved in 2 mL methylene
chloride (Sigma-Aldrich). Then, 200 ml of the calcein solution
was homogenized in 2 mL of PLA solution for 2 min at
15,000 rpm. The resulting water-in-oil emulsion was homoge-
nized in 50 mL of an aqueous solution of 0.1% polyvinyl
alcohol (Sigma-Aldrich) for 2 min at 10,000 rpm, which
produced a water-in-oil-in-water emulsion. After mixing for
3 h at 300 rpm, the methylene chloride was extracted into the
Fig. 2. Method to fabricate polymer microneedles that encapsulate
drug for controlled release. First, a suspension of drug particles is
filled into a microneedle mold. Evaporation of the solvent leaves
solid drug particles partially filling the mold. Pellets of biodegradable
polymer are then melted into the mold under vacuum. Cooling and
solidification of the polymer yields biodegradable polymer micro-
needles with encapsulated drug particles.
1010 Park, Allen, and Prausnitz
continuous phase, which solidified the discontinuous phase
into PLA microparticles encapsulating calcein. Micropar-
ticles of 1Y30 mm in size were isolated by filtration, loaded
into a PDMS mold, and encapsulated in PLGA micro-
needles, as described above.
Characterization of Microneedles Containing Drug
In Vitro Release Test in Saline
To measure release rates from microneedles, an array
containing 100Y200 needles encapsulating one of the formula-
tions of calcein or BSA was attached to the bottom or side of a
30-mL glass vial (All-Pak, Bridgeville, PA, USA) containing 5
or 10 mL of phosphate-buffered saline (PBS, pH 7.4, Sigma-
Aldrich) filtered using a 0.2-mm filter (Millipore). Glass vials,
PBS, and magnetic stir bars were autoclaved prior to use. The
vials were magnetically stirred at 300 rpm and incubated in a
37-C water bath. Periodically, a 100-ml aliquot of PBS was
sampled fromeach vial, replaced with fresh PBS,and analyzed
to determine the concentration of calcein or Texas Red-
labeled BSA by calibrated spectrofluorometry (QM-1, Photon
Technology International, South Brunswick, NJ, USA). Mea-
sured concentrations were converted into cumulative drug
released (Mt) by accounting for PBS volume. Total drug
content (M0) was determined by placing microneedles in 1 N
NaOH overnight at the end of each experiment to fully
degrade remaining PLGA and PLA and thereby release all
encapsulated drug (23). The resulting solution was returned
to pH 7.4 using HCl before analysis. M0was typically õ1 mg
for calcein only,BSAonly,andcalceininCMCmicroparticles
andõ0.1mg for calcein in PLA microparticles.
Drug release from microneedle formulations was mod-
eled using the Higuchi equation (24), which indicates that dif-
fusion-mediated release should be proportional to the square
root of time.
In this expression, D is the apparent diffusion coefficient
of the drug in the polymer matrix, t is time, and r is the radius
of the microneedle (50 mm) (25). This equation was fitted to
experimental data to yield D, which is the only unknown
(26). The fitting procedure used a least-squares method that
minimized the differences between experimental and theo-
In Vitro Release Test in Skin
To study the dissolution and release of drug in skin,
microneedles encapsulating calcein were inserted into full-
4-C. Refrigeration was used to avoid dehydration and degra-
dation of the skin. Recognizing that skin properties and drug
of 4-C and the body temperature of 37-C, we conducted this
experiment to qualitatively verify that our microneedles insert
and that encapsulated drug is released in the skin.
After 8 h, the needles were removed and any residual
calcein on the skin surface was cleaned off with wet tissue
paper. The spatial profile of calcein released in the skin was
then imaged by confocal microscopy (LSM 510; Zeiss,
Thornwood, NY, USA). Human cadaver skin was obtained
from the Emory University Body Donor Program with
approval from the Georgia Tech and Emory University
Institutional Review Boards.
Microneedle Failure Force Measurement
To determine the effect of calcein encapsulation on
microneedle mechanical properties, the microneedle failure
force was measured as described previously (16,20). Briefly,
stressYstrain curves were generated using a displacement-
force test station (Model 921A, Tricor Systems, Elgin, IL,
USA) while pressing an array of 35 microneedles against a
stainless steel surface at a rate of 1.1 mm/s until a preset
maximum load (19.6 N) was reached. Microneedles had a
base radius of 100 mm, tip radius of 12 mm, and height of 1
mm. Microneedle failure was indicated by a sudden drop in
applied force. After each test, microneedles were visually
inspected by microscopy to confirm that all microneedles had
deformed and failed uniformly. Failure force was determined
at calcein contents of 0, 2, and 10% prepared using a single-
Because encapsulation of drugs within microneedles
involves a brief exposure to a high-temperature polymer
melt, the encapsulation process could be damaging to drugs,
especially proteins. To assess possible damage, protein
stability was tested by measuring protein solubility, dynamic
light scattering, and circular dichroism (CD), using BSA as a
model protein (27). Because protein content in microneedles
is small, larger samples were generated by dispersing 500 mg
of homogenized BSA particles in a 27.5 g solution of 10%
(w/w) PLGA in acetonitrile and then pouring the suspen-
sion onto aluminum foil. A thin polymer film encapsulating
BSA particles was formed by evaporating off the acetoni-
trile for 5 h under j67 kPa vacuum. The film was cut into
3 ? 3-cm squares and placed on a 1 cm-thick PDMS film
to simulate the molding process. Samples prepared in this
way were placed in the oven at 135-C for predetermined
times. After cooling, PLGA samples were dissolved again in
acetonitrile, and the BSA particles were recovered by filtra-
tion, washed with methylene chloride, and dried for 6Y8 h
under j30 kPa vacuum. Each condition was tested in
As a first measure of stability, BSA samples were
dissolved in PBS to determine the fraction of BSA remaining
soluble after thermal exposure, which includes both native
BSA and reversibly denatured BSA. Insoluble aggregates
were removed by centrifugation at 23,000 ? g for 20 min (28).
Protein concentration in the supernatant was determined by
the Lowry protein assay (29).
The presence of soluble aggregates of BSA was detected
by dynamic light scattering (30). BSA particles were dis-
solved in PBS at a concentration of 800 mg/mL and filtered to
removed insoluble aggregates, and 45 ml of solution was
placed in a quartz cuvette (Proterion, now Wyatt Technolo-
gy, Santa Barbara, CA, USA) at room temperature for
1011Polymer Microneedles for Controlled Release
dynamic light scattering measurements (DynaPro-MS/X,
Proterion) using CONTIN analysis.
To identify possible changes in the ratio of a-helix/b-
sheet components of BSA structure, CD polarimetry mea-
surements were performed using BSA particles dissolved in
PBS at a concentration of 70 mg/mL (J700, Jasco, Easton,
MD, USA) (31). CD spectra were obtained over a wave-
length range of 400Y190 nm with a sensitivity of 20 mdeg and
a response time of 2 s.
Fabrication of Microneedles for Transdermal Drug Delivery
Fabrication of Microneedle Master Structures
The first step to make polymer microneedles for con-
trolled-release drug delivery involved fabricating master
structures using microelectromechanical systems (MEMS)
techniques. These master structures were then used to make
molds, which were in turn used to make replicate micro-
needles out of biodegradable polymers. Two different geom-
etries of microneedle master structures were fabricated out
of SU-8 epoxy using lithography-based methods. Representa-
tive beveled-tip microneedles are shown in Fig. 3A and have a
base radius of 50 mm, a tip radius of 5 mm, and a height of
600 mm. The needles are positioned in a 20 ? 6 array with a
center-to-center spacing between needles of 400 and 1400 mm.
The entire array occupies an area of 9 ? 9 mm. Geometric
parameters of beveled-tip microneedle arrays, such as needle-
to-needle spacing, needle base radius, and base shape, were
controlled by adjusting the size, shape, and spacing of the
lithography mask. The needle height was controlled by the
thickness of SU-8 photoresist casting and etching parame-
ters. The tip sharpness was controlled by the etching
Fig. 3. Microscopy images of microneedles. A section of an array of (A) bevel-tip microneedles
and (B) tapered-cone microneedles used as master structures (imaged by SEM). Making a mold
and using it to prepare polymer microneedles as described in Fig. 2 yielded (C) bevel-tip and
(D) tapered-cone microneedles made of PLGA and encapsulating calcein within their tips
(imaged by fluorescence and bright-field microscopy, respectively). Using a double-encapsula-
tion method produced microneedles that encapsulate microparticles that, in turn, encapsulate
calcein. (E) Cutting off the tip of a PLGA microneedle reveals the PLA microparticles within
(imaged by SEM). (F) A complete 20 ? 10 array of PLGA microneedles is shown (imaged by
1012 Park, Allen, and Prausnitz
Representative tapered-cone microneedles are shown in
andaheight of750 mm. Theneedles are positionedin a10 ?20
array with center-to-center spacing between needles of 400
and 800 mm. The entire array occupies an area of 9 ? 9 mm.
Needle-to-needle spacing, needle base radius, and base shape
were controlled by adjusting the size, shape, and spacing of the
lithography mask. Needle height, taper, and tip sharpness
were controlled by the optical properties and geometry of the
integrated lenses (see BMaterials and Methods^).
Polymer Microneedles Encapsulating Drug
Using PDMS molds created using the microneedle
master structures described above, PLGA microneedles were
fabricated with encapsulated calcein or BSA, which served as
model drugs. Figure 3C and D shows beveled-tip and
tapered-cone microneedles, respectively, with encapsulated
calcein. In this case, most calcein was entrapped near the tips
of microneedles to be sure that all of the model drug was
delivered, even if the microneedles only inserted partially
into the skin. Calcein was loaded as solid particles into the
microneedles at compositions up to 10% of needle mass.
Similar results were seen for encapsulation of BSA in
To facilitate slower release, the model drug was first
were then encapsulated within microneedles. Figure 3E shows
a PLGA microneedle with encapsulated PLA microparticles
containing 5Y10% calcein. The tip of the microneedle has
been cut off, thereby exposing the PLA microneedles within.
Although these microneedles are made of a solid polymer
matrix, a continuous PLGA domain surrounds discrete CMC
or PLA microdomains having the size and shape of the
microparticles that formed them. To supplement the highly
magnified views of microneedles, Fig. 3F shows the size and
shape of a complete PLGA microneedle array.
Controlled Release of Drug from Microneedles
Release from Microneedles Using a Single-Encapsulation
Polymer microneedles encapsulating drugs were devel-
oped to serve as a minimally invasive method of con-
trolled-release drug delivery that is similar to injectable
microparticulate systems already in clinical use, but does not
involve the pain and inconvenience of hypodermic needle
injection. To test this idea, controlled release from micro-
needles loaded with calcein was measured in vitro. Using a
single-encapsulation formulation (i.e., calcein was directly
entrapped within the PLGA matrix of the microneedles),
calcein release showed zero-order kinetics over a period of
4 h, after which 93% of encapsulated calcein was released
(Fig. 4A). Controlled release of BSA from a similar formula-
tion showed slower kinetics, where 80% of BSA was released
after 5 days (Fig. 4A).
These kinetics are consistent with release controlled by
drug diffusion through the microneedle polymer matrix, as
opposed to release due to polymer degradation and dissolu-
tion. First, significant degradation of polymer should not have
occurred over the timescale of hours to days. PLGA
degradation is known to occur over a timescale of months
(32). This is confirmed by microscopy analysis of micro-
needles discussed below, which shows that calcein was
released before significant needle degradation was observed.
Second, degradation-controlled release should be a weak
function of drug molecular size (17,18). In contrast, diffusion-
controlled release is a strong function of drug molecular size,
which is consistent with the order-of-magnitude difference in
release rates between calcein and BSA.
As a companion to quantitative release studies, micro-
needles were imaged by light microscopy during release and
degradation. Before release was initiated, calcein was encap-
sulated within microneedles, especially at the tips, as
indicated by the dark regions in the otherwise transparent
PLGA microneedles (Fig. 5A). After 9 h of release in PBS,
microneedles showed little change in geometry, indicating
that significant polymer degradation and dissolution did not
yet occur, but were largely devoid of encapsulated calcein, as
indicated by the white voids from which calcein was released
(Fig. 5B). This image is consistent with release measurements
in Fig. 4, which show release kinetics of hours. Subsequent
incubation in strong base to rapidly degrade microneedles
and fully release any residual entrapped calcein yielded
almost fully degraded microneedles (Fig. 5C).
Fig. 4. Cumulative release of model compounds from PLGA micro-
needles prepared using different formulations to control release
kinetics in vitro. (A) Release of calcein (P) and BSA (Í)
encapsulated within microneedles. The inset has an expanded time
axis to better show release over the initial 4 h. (B) Release of calcein
from microneedles that encapsulated (P) calcein only, (r) CMC
microparticles containing calcein, or (Í) PLA microparticles con-
taining calcein. Data are presented as average T SEM (n = 3Y5).
1013Polymer Microneedles for Controlled Release
Release from Microneedles Using Double-Encapsulation
To achieve slower release, calcein was first encapsulated
within CMC microparticles, which were then encapsulated
within PLGA microneedles. Using this approach, calcein
release from microneedles showed steady release over 4 days
(Fig. 4B). This demonstrates that double encapsulation using
CMC can slow release kinetics by more than an order of
magnitude. Some calcein may not have been encapsulated
within CMC, which could explain the burst effect seen during
the first hours of release.
To achieve slow release over a still longer time, calcein
was first encapsulated within PLA microparticles, which were
then encapsulated within PLGA microneedles. Using this
approach, calcein release from microneedles was much
slower, exhibiting an initial burst (Fig. 4B), followed by slow
release over a 2-month period (discussed below). Although
not examined in this study, the double-encapsulation ap-
proach lends itself to achieving release over other timescales,
which can be controlled by formulation of drugs in micro-
particles with different polymer compositions using well-
known methods (33).
Modeling Controlled Release from Microneedles
Controlled release from polymer matrix systems can be
modeled using the Higuchi equation describing Fickian
diffusion, which predicts that drug release increases linearly
with the square root of time [see Eq. (1)]. This relationship
should be valid at times following the initial burst-effect
release of nonencapsulated drug and at times before polymer
degradation can play a role and the apparent diffusion
coefficient is no longer constant. During this release period,
the apparent diffusion coefficient of drug inside the micro-
needles is the only unknown in the Higuchi equation and was
determined by nonlinear regression using experimental data
for controlled release of calcein and BSA from microneedles
with single- and double-encapsulation formulations shown in
Fig. 4. Fits of Eq. (1) to these experimental data are shown
in Fig. 6. The corresponding correlation coefficients were
r2= 0.92, 0.90, 0.99, and 0.96 for calcein, calcein entrapped
in CMC microparticles, calcein entrapped in PLA micro-
particles, and BSA,respectively.
Based on these fitted equations, the apparent diffusion
coefficient of calcein in PLGA microneedles was 1.2 ?
10j10cm2/s (Fig. 6A). The diffusion coefficient of free calcein
in water has previously been calculated to be 5.0 ? 10j6cm2/s
(13), which is four orders of magnitude greater. Thus, calcein
encapsulation within PLGA microneedles significantly
slowed calcein diffusion for controlled release. The apparent
diffusion coefficient of BSA in PLGA microneedles was 3.0 ?
10j12cm2/s (Fig. 6D), which is five orders of magnitude
smaller than free BSA diffusivity in water, 5.9 ? 10j7cm2/s
(13). This is again consistent with release controlled by
reduced BSA diffusivity in the PLGA matrix.
Encapsulation of calcein in CMC and PLA micro-
particles reduced the apparent diffusivity still further. CMC
encapsulation reduced the diffusivity to 4.7 ? 10j12cm2/s
(Fig. 6B), which is more than an order of magnitude lower
than without CMC. Encapsulation in PLA microparticles
further reduced the diffusion coefficient by another two
orders of magnitude to a value of 6.2 ? 10j14cm2/s.
Microneedle Insertion and Controlled Release in Skin
Controlled Release from Microneedles into Skin
Because encapsulated microneedles are envisioned for
use in skin, we assessed the ability of microneedles to insert
into skin and release drug. In the first experiment, an array of
microneedles with encapsulated calcein was inserted into
human cadaver skin and then removed after 9 h. Imaging the
Fig. 5. Microscopic images of microneedles during controlled
release and degradation in vitro (imaged by bright-field microscopy).
(A) Initially, microneedles encapsulated calcein, as indicated by dark
regions in the otherwise transparent PLGA microneedles. (B) After
incubation in PBS for 9 h, almost all calcein was released from the
microneedles, as indicated by the white voids. (C) Incubation in
concentrated NaOH rapidly degraded the microneedles and fully
released any residual calcein.
1014 Park, Allen, and Prausnitz
skin surface by fluorescence microscopy revealed fluores-
cence at each site of microneedle insertion, indicating that
calcein was released into the skin (Fig. 7A). In a companion
experiment, skin was treated in the same way and then
imaged by confocal microscopy (Fig. 7B). This revealed little
fluorescence at the skin surface and intense fluorescence
deeper in the skin, with a peak fluorescence at 125Y160 mm
below the surface, which is just below the dermalYepidermal
junction and near the dermal capillary bed (34). This
observation further demonstrates that polymer microneedles
can release encapsulated compounds within the skin. Because
calcein was loaded into the tips of these microneedles, these
images also indicate that the microneedles did not insert to
their full 750-mm length during this in vitro hand insertion. If
desired, insertion at higher velocity, with vibration, or with
skin under tension (as found on the body in vivo) can
increase insertion depth by reducing skin deflection during
Mechanics of Microneedle Insertion into Skin
The ability of microneedles to insert into skin also
depends on needle mechanical properties. Although metal
microneedles used in other studies are extremely strong,
polymer microneedles require special attention to mechani-
cal strength. Our previous study of polymer microneedles
(without encapsulated drug) showed that microneedles made
of PLGA and other polymers could be designed with
sufficient strength for reliable insertion without breaking
(20). Because the encapsulation of drug could weaken
microneedles, the force required to cause failure of PLGA
microneedles was measured at 0, 2, and 10% loading with
calcein. As shown in Fig. 8, microneedles without encapsu-
lated drug failed at a force of 163 T 10 mN per needle.
Previous measurements and calculations have shown that
needles of the same geometry insert into human skin with a
force of 45 mN per needle (16). The safety factorVdefined as
the ratio of failure force to insertion forceVis therefore 3.6,
which means that these microneedles insert into the skin with
a force much less than the failure force.
Encapsulation of 2% calcein in needles of the same
geometry lowered the failure force to 91 T 30 mN per needle,
which indicates that encapsulation weakened the needles, but
still maintained a safety factor of 2.0. Encapsulation probably
weakened the microneedles because calcein particles are
mechanically weaker than PLGA and poor adhesion between
calcein particles and the PLGA matrix provided sites for
Fig. 6. Modeling of controlled release from PLGA microneedles prepared using different
formulations: (A) calcein only, (B) calcein encapsulated in CMC microparticles, (C) calcein
encapsulated in PLA microparticles, and (D) BSA only. Data were obtained from Fig. 4. The
Higuchi equation (Eq. 1) was fitted to the data, as shown by the solid line. For these fits, the t =
0 data points were excluded to eliminate release due to the initial burst effect. In (C), only data
up to 100 h were fitted because the effects of polymer degradation became significant at later
times. Values of effective diffusivity and correlation coefficients resulting from these fits are
presented in the text. The x-axes are presented as the square root of time because the Higuchi
equation predicts a linear dependence of release on the square root of time for diffusion-
1015 Polymer Microneedles for Controlled Release
Encapsulation of 10% calcein lowered the failure force
further to 40 T 2 mN per needle, which reduced the safety
factor to 0.89. Because this value is less than 1.0, it means
that these microneedles mechanically fail before inserting
into skin. However, the microneedles used for this study were
relatively blunt (12-mm tip radius) and long (1-mm length).
Sharpening tip radius is known to decrease insertion force
and shortening needle length is known to increase failure
force (16,20). Thus, it is possible that redesigned PLGA
microneedles with 10% drug capsulation might be made
strong enough for insertion into skin without failure.
Protein Stability During Microneedle Fabrication
Encapsulation in microneedles exposed drugs to the high
temperature of melted PLGA, which could be damaging,
especially to temperature-sensitive proteins. The possibility
of damage was reduced by selecting a polymer that can be
melted at a relatively low temperature (135-C), minimizing
the exposure time to elevated temperature to 10Y20 min, and
keeping encapsulated drug in the solid state, which reduces
conformational mobility and thereby increases protein sta-
bility (37). To assess possible protein damage, the effect of
exposing BSA to melted PLGA under encapsulation con-
ditions was measured after different exposure times using
three different protein stability assays. BSA was chosen as a
model protein, although additional studies will be needed to
assess stability of other proteins.
First, irreversible denaturation that led to insoluble
protein aggregates was assessed by measuring aqueous BSA
solubility after various treatments. Exposure to the solvent-
based processes used during encapsulation, but without any
thermal exposure, did not affect BSA solubility (0 min
thermal exposure in Fig. 9A). Exposure to elevated temper-
ature for 10, 20, or 30 min lowered protein solubility by 10,
18, and 32%, respectively, presumably due to irreversible
aggregation. After 1 h at elevated temperature, essentially all
protein was denatured.
To determine if additional protein aggregates existed
among the soluble fraction of BSA, we further tested samples
by dynamic light scattering. As shown in Fig. 9B, a 10-min
exposure to elevated temperature had no significant effect,
whereas 20- and 30-min exposures caused increasing levels of
aggregation that formed particles measuring tens of nano-
meters in size.
Fig. 7. Fluorescence microscopy images of calcein delivered into
human cadaver skin using polymer microneedles. (A) A 100-needle
array of PLGA microneedles containing õ1 2g of calcein was inserted
into full-thickness cadaver skin for 9 h. After removing the needles,
the skin surface was imaged en face byfluorescencemicroscopy,which
shows calcein delivery into the skin at the site of each microneedle
insertion. (B) Confocal microscopy of the site of one needle insertion
shows calcein release within the skin below the skin surface, which is
expected for release of calcein encapsulated primarily at the needle
tip. Optical section depths are (B1) 20 mm, (B2) 55 mm, (B3) 90 mm,
(B4) 125 mm, (B5) 160 mm, (B6) 195 mm below the skin surface.
Fig. 8. Mechanical strength of polymer microneedles as a function of
calcein encapsulation. The force-per-needle required to fracture an
array of 35 microneedles decreased with increasing calcein content.
Microneedles had a geometry of 12-mm tip radius, 100-2m base
radius, and 1-mm length. The dashed line indicates the expected
force required for insertion of a microneedle of these dimensions into
skin (16). Data are presented as average T SEM (n = 7).
1016 Park, Allen, and Prausnitz
Finally, circular dichroism measurements were made to
detect the presence of intermediate species by measuring the
ratio of a-helix/b-sheet (27). As shown in Fig. 9C, there was
little difference in the spectra of BSA exposed to thermal
treatments for up to 30 min, indicating little or no formation
of intermediate species. In contrast, a positive control of BSA
incubated in aqueous solution for 25 min at 67-C exhibited a
large spectral shift corresponding to significant conforma-
tional changes. This observation is consistent with the
expectation that keeping protein in the solid state increases
stability during thermal processing.
Advantages of Controlled Release
from Polymer Microneedles
A limitation of conventional controlled-release polymer
formulations is that they often require surgical implantation
or hypodermic needle injection (1,2). This limits the ability of
such systems to be self-administered and generally requires
the time and expense of trained clinical personnel. Conven-
tional transdermal patches provide up to 1 week of controlled
delivery, but only for the small subset of drugs that can cross
skin at useful rates (38). Thus, the possibility of a controlled-
release, microneedle-based delivery method could provide a
significant advance that captures the ease of use afforded by a
patch and the versatile controlled-release properties of
biodegradable polymer systems.
Microneedles can be painlessly inserted into the skin in a
minimally invasive manner that lends itself to self-adminis-
tration by patients (5). As designed in this study, the intact
microneedle patch could be left in place on the skin for an
extended period and later removed. Alternatively, the patch
could be designed to intentionally break off the needles,
leaving them invisibly embedded in the skin, so that the patch
backing could be discarded. In this scenario, microneedles
would need to be designed to remain anchored within the
skin to prevent unintentional expulsion. By using FDA-
approved materials, such as PLGA and CMC, controlled-
release microneedles are likely to be safe. By leveraging
advanced microelectronics industry fabrication technology,
microneedles are likely to be mass-produced at disposable
costs (e.g., US$0.10 per patch) (11).
Controlled-release microneedles require no power supply
or sophisticated controllers, which also reduces cost and
complexity. This contrasts with most other minimally invasive
transdermal delivery methods under development (38). More-
over, intersubject variations in skin diffusional barrier
properties are less important because delivery is largely
controlled by polymer properties of diffusivity and degrada-
tion rate. In this study, in vitro release kinetics ranging from
hours to months were demonstrated using single- and double-
encapsulation methods, which should be broadly applicable
to many drugs.
Limitations of Controlled Release
from Polymer Microneedles
Despite many advantages, controlled-release micronee-
dles have limitations that constrain their possible applica-
tions. One disadvantage of the microneedle fabrication
method used in this study is that it involves melted polymer
that exposes encapsulated drugs to elevated temperature.
Although this may not pose difficulties for some small-
molecule drugs, this study showed that 10% of encapsulated
Fig. 9. Stability of BSA after encapsulation within PLGA using
thermal exposures similar to those used to fabricate microneedles.
(A) BSA solubility in water decreased with increasing length of
exposure to the PLGA melt at 135-C. (B) Dynamic light scattering
shows that the size distribution of BSA molecules did not change
after a 10-min exposure to the PLGA melt, but BSA aggregation was
evident after longer exposures. (C) Circular dichroism shows that
BSA spectra did not change significantly after exposure of solid-state
BSA to the PLGA melt, but BSA in aqueous solution was damaged.
1017 Polymer Microneedles for Controlled Release
BSA was irreversibly aggregated after a 10-min exposure to
the polymer melt and that longer exposures led to more
extensive denaturation of primary and secondary structure.
Thus, applications involving delivery of heat-sensitive drugs
are possible, but may lead to partial denaturation. Efforts in
our laboratory are underway to develop different methods to
fabricate polymer microneedles that do not involve elevated
temperatures that damage proteins.
Polymer microneedles for controlled-release delivery are
also constrained by needle mechanical properties. Although
polymer microneedles can be designed to be strong enough
to reliably insert into skin, the addition of encapsulated drug
can weaken them. This study found that microneedles with a
2% drug loading retained sufficient mechanical strength, but
needles with a 10% loading did not. Although redesign of
needle geometry should increase the maximum drug loading
that retains needle strength, there is clearly an upper limit,
which constrains the maximum possible dose that can be
Perhaps the greatest shortcoming of controlled-release
microneedles is the limited dose that can be administered.
Because drug is encapsulated within microneedles and micro-
canbeadministeredis likelytobelessthan 1mg.Thisestimate
is based on the calculation that one microneedle has a mass on
the order of 10 mg. Thus, a 100-needle patch with 2% drug
loading contains 20 mg of drug. Better needle design may
permit a patch with 100Y1000 needles loaded with 10% drug,
which corresponds to 100Y1000 mg. Coating the surface of
microneedles could further increase dose.
Although the maximum dose constrains applications,
controlled-release delivery of up to 1 mg has a number of
candidate drugs on the market, with more likely to be
approved in the future. For example, interferon a-2A for
hepatitis C, interferon b-1A for multiple sclerosis, and
erythropoietin for anemia have doses of 33 mg/week, 132
mg/week and 100 mg/day, respectively (39). Thus, controlled
release of 1 mg of drug would last for 30 weeks, 7 weeks, and
10 days, respectively. Vaccine delivery presents another
compelling opportunity. For example, hepatitis B vaccine
and influenza vaccine require antigen doses of just 10 and 45
mg, respectively. Moreover, the rich dendritic cell population
in the skin has been shown to increase immune response to
vaccines (40), which provides a further motivation for
microneedle-based vaccine delivery to the skin.
Polymeric controlled-release drug delivery is a powerful
technology with demonstrated clinical utility. To remove the
need for administration by clinical personnel, this study
developed a microneedle-based controlled-release device
designed for self-administration at home. Using this ap-
proach, drug can be encapsulated within biodegradable
polymer microneedles for controlled release in the skin using
fabrication methods designed for inexpensive mass produc-
tion. Drug-release kinetics ranging from hours to months
were controlled by encapsulating drugs directly within the
PLGA microneedle matrix or encapsulating drug within
CMC or PLA microparticles, which were then encapsulated
within the needle. Effective microneedle design required
drug loadings of less than 10% to maintain needle mechan-
ical strength, limited exposure to elevated temperature
during needle processing to maintain protein stability, and
selection of drugs with total doses less than 1 mg, due to the
inherently small size of microneedles. Overall, this study
demonstrates the feasibility of using polymer microneedles
for controlled-release drug delivery in skin.
We thank Jin-Woo Park, Hak-Jun Sung, and Ping Ming
Wang for helpful discussions and Gary Meek for photograph-
ing Fig. 3F. This work was supported in part by the National
Institutes of Health. J.-H. P., M. G. A., and M. R. P. are
members of the Microelectronics Research Center, and
J.-H. P. and M. R. P. are members of the Institute for
Bioengineering and Bioscience and the Center for Drug
Design, Development and Delivery at Georgia Tech.
1. R. Langer. Drug delivery and targeting. Nature 392:5Y10 (1998).
2. H. Rosen and T. Abribat. The rise and rise of drug delivery. Nat.
Rev. Drug Discov. 4:381Y385 (2005).
3. P. Periti, T. Mazzei, and E. Mini. Clinical pharmacokinetics of
depot leuprorelin. Clin. Pharmacokinet. 41:485Y504 (2002).
4. H. Brem, S. Piantadosi, P. C. Burger, M. Walker, R. Selker,
N. A. Vick, K. Black, M. Sisti, S. Brem, and G. Mohr, et al.
Placebo- controlled trial of safety and efficacy of intraoperative
controlled delivery by biodegradable polymers of chemotherapy
for recurrent gliomas. The Polymer-Brain Tumor Treatment
Group. Lancet 345:1008Y1012 (1995).
5. S. Kaushik, A. H. Hord, D. D. Denson, D. V. McAllister, S.
Smitra, M. G. Allen, and M. R. Prausnitz. Lack of pain
associated with microfabricated microneedles. Anesth. Analg.
6. J. A. Mikszta, J. B. Alarcon, J. M. Brittingham, D. E. Sutter, R.
J. Pettis, and N. G. Harvey. Improved genetic immunization via
micromechanical disruption of skin-barrier function and tar-
geted epidermal delivery. Nat. Med. 8:415Y419 (2002).
7. S. Henry, D. McAllister, M. G. Allen, and M. R. Prausnitz.
Microfabricated microneedles: a novel method to increase
transdermal drug delivery. J. Pharm. Sci. 87:922Y925 (1998).
8. M. Cormier, B. Johnson, M. Ameri, K. Nyam, L. Libiran, D. D.
Zhang, and P. Daddona. Transdermal delivery of desmopressin
using a coated microneedle array patch system. J. Control.
Release 97:503Y511 (2004).
9. T. Miyano, Y. Tobinaga, T. Kanno, Y. Matsuzaki, H. Takeda,
M. Wakui, and K. Hanada. Sugar micro needles as transdermic
drug delivery system. Biomed. Microdevices 7:185Y188 (2005).
10. M. R. Prausnitz. Microneedles for transdermal drug delivery.
Adv. Drug Deliv. Rev. 56:581Y587 (2004).
11. M. Prausnitz, J. Mikszta, and J. Raeder-Devens. Microneedles.
In E. Smith and H. Maibach (ed.), Percutaneous Penetration
Enhancers, CRC Press, Boca Raton, FL, 2005, pp. 239Y255.
12. J. G. E. Gardeniers, R. Luttge, J. W. Berenschot, M. J. de Boer,
Y. Yeshurun, M. Hefetz, R. van’t Oever, and A. van den Berg.
Silicon micromachined hollow microneedles for transdermal
liquid transport. J. MEMS 6:855Y862 (2003).
13. D. V. McAllister, P. M. Wang, S. P. Davis, J.-H. Park, P. J.
Canatella, M. G. Allen, and M. R. Prausnitz. Microfabricated
needles for transdermal delivery of macromolecules and nano-
particles: fabrication methods and transport studies. Proc. Natl.
Acad. Sci. USA 100:13755Y13760 (2003).
14. R. K. Sivamani, B. Stoeber, G. C. Wu, H. Zhai, D. Liepmann,
and H. Maibach. Clinical microneedle injection of methyl
nicotinate: stratum corneum penetration. Skin Res. Technol.
1018Park, Allen, and Prausnitz
15. J. H. Park, S. P. Davis, Y. K. Yoon, M. R. Prausnitz, and
M. G. Allen. Micromachined biodegradable microstructures.
In The 16th Annual International Conference on Micro Electro
Mechanical Systems, IEEE, Piscataway, NJ, 2003, pp. 371Y
16. S. P. Davis, B. J. Landis, Z. H. Adams, M. G. Allen, and M. R.
Prausnitz. Insertion of microneedles into skin: measurement and
prediction of insertion force and needle fracture force. J.
Biomech. 37:1155Y1163 (2004).
17. W. M. Saltzman. Drug Delivery: Engineering Principles for Drug
Therapy, Oxford University Press, New York, 2001.
18. D. G. Kanjickal and S. T. Lopina. Modeling of drug release from
polymeric delivery systemsVa review. Crit. Rev. Ther. Drug
Carr. Syst. 21:345Y386 (2004).
19. M. J. Madou. Fundamentals of Microfabrication: The Science of
Miniaturization, CRC Press, Boca Raton, FL, 2002.
20. J.-H. Park, M. G. Allen, and M. R. Prausnitz. Biodegradable
polymer microneedles: fabrication, mechanics and transdermal
drug delivery. J. Control. Release 104:51Y66 (2005).
21. H. K. Lee, J. H. Park, and K. C. Kwon. Double-walled
microparticles for single shot vaccine. J. Control. Release 44:
22. J.-P. Benoit, H. Marchais, H. Rolland, and V. V. Velde. Bio-
degradable microspheres: advances in production technology. In
S. Benita (ed.), Microencapsulation, Marcel Dekker, New York,
1996, pp. 35Y72.
23. R. K. Gupta, A. C. Chang, P. Griffin, R. Rivera, Y. Y. Guo, and
G. R. Siber. Determination of protein loading in biodegradable
polymer microspheres containing tetanus toxoid. Vaccine
24. Y. W. Chien (ed.), Novel Drug Delivery Systems: Fundamentals,
Developmental Concepts, Biomedical Assessments, Marcel Dek-
ker, New York, 1991.
25. R. W. Baker and H. K. Lonsdale. Controlled release: mecha-
nisms and rates. In A. C. Tanquary and R. E. Lacey (ed.),
Controlled Release of Biologically Active Agents. Advances in
Experimental Medicine and Biology, Vol. 47, Plenum Press, New
York, 1974, pp. 15Y72.
26. N. Faisant, J. Siepmann, and J. P. Benoit. PLGA-based micro-
particles: elucidation of mechanisms and a new, simple mathe-
matical model quantifying drug release. Eur. J. Pharm. Sci.
27. D. Bulone, V. Martorana, and P. L. San Biagio. Effects of
intermediates on aggregation of native bovine serum albumin.
Biophys. Chem. 91:61Y69 (2001).
28. W. R. Liu, R. Langer, and A. M. Klibanov. Moisture-induced
aggregation of lyophilized proteins in the solid state. Biotechnol.
Bioeng. 37:177Y184 (1991).
29. O. H. Lowry, N. J. Rosebrough, A. L. Farr, and R. J. Randall.
Protein measurement with the Folin phenol reagent. J. Biol.
Chem. 193:265Y275 (1951).
30. B. J. Berne and R. Pecora. Dynamic Light Scattering: With
Applications to Chemistry, Biology, and Physics, Dover Pub-
lications, Mineola, NY, 2000.
31. G. D. Fasman (ed.). Circular Dichroism and the Conformational
Analysis of Biomolecules. Plenum, New York, 1996.
32. L. Lu, C. A. Garcia, and A. G. Mikos. In vitro degradation of
thin poly(DL-lactic-co-glycolic acid) films. J. Biomed. Mater.
Res. 46:236Y244 (1999).
33. K. Park (ed.), Controlled Drug Delivery: Challenges and Strate-
gies. American Chemical Society, Washington, DC, 1997.
34. N. A. Monteiro-Riviere. Comparative anatomy, physiology, and
biochemistry of mammalian skin. In D. W. Hobson (ed.),
Dermal and Ocular Toxicology, CRC Press, Boca Raton, FL,
1991, pp. 3Y71.
35. P. J. Rousche and R. A. Normann. A method for pneumatically
inserting an array of penetrating electrodes into cortical tissue.
Ann. Biomed. Eng. 20:413Y422 (1992).
36. M. Yang and J. D. Zahn. Microneedle insertion force reduction
using vibratory actuation. Biomed. Microdevices 6:177Y182
37. T. Arakawa, S. J. Prestrelski, W. C. Kenney, and J. F. Carpenter.
Factors affecting short-term and long-term stabilities of proteins.
Adv. Drug Deliv. Rev. 46:307Y326 (2001).
38. M. R. Prausnitz, S. Mitragotri, and R. Langer. Current status
and future potential of transdermal drug delivery. Nat. Rev.
Drug Discov. 3:115Y124 (2004).
39. Physicians’ Desk Reference. Thomson PDR, Montvale, NJ, 2005.
40. S. Mitragotri. Immunizationwithoutneedles. Nat. Rev. Immunol.
1019Polymer Microneedles for Controlled Release