A Fully Integrated Microneedle-Based Transdermal Drug Delivery System

A Fully Integrated Microneedle-based Transdermal Drug Delivery System

Niclas Roxhed

MICROSYSTEM TECHNOLOGY LABORATORY SCHOOL OF ELECTRICAL ENGINEERING ROYAL INSTITUTE OF TECHNOLOGY

ISBN 978-91-7178-751-4 ISSN 1653-5146 TRITA-EE 2007:046

Submitted to the School of Electrical Engineering KTH—Royal Institute of Technology, Stockholm, Sweden, in partial fulﬁllment of the requirements for the degree of Doctor of Philosophy Stockholm 2007 ii A Fully Integrated Microneedle-based Transdermal Drug Delivery System

The left picture on the front cover shows an integrated microneedle-based drug delivery system fabricated and used for experiments by the author. An array of microneedles is located on the other side of the device. The right picture on the cover shows a magniﬁed view of these side-opened hollow microneedles, likewise fabricated by the author. The needles are designed to penetrate skin tissue using a low insertion force. The needles are fabricated by deep reactive ion etching of silicon and the length of the needles is 400 μm.

Copyright 2007 by Niclas Roxhed All rights reserved to the summary part of this thesis, including all pictures and ﬁgures. No part of this publication may be reproduced or transmitted in any form or by any means, without prior permission in writing from the copyright holder. The copyrights for the appended journal papers belong to the publishing houses of the journals concerned. The copyrights for the appended manuscripts belong to their authors.

Printed by Universitetsservice US AB, Stockholm 2007. Thesis for the degree of Doctor of Philosophy at the Royal Institute of Technology, Stockholm, Sweden, 2007. ABSTRACT iii

Abstract

Patch-based transdermal drug delivery offers a convenient way to administer drugs without the drawbacks of standard hypodermic injections relating to issues such as patient acceptability and injection safety. However, conventional transdermal drug delivery is limited to therapeutics where the drug can diffuse across the skin barrier. By using miniaturized needles, a pathway into the human body can be established which allow transport of macromolecular drugs such as insulins or vaccines. These microneedles only penetrate the outermost skin layers, superficial enough not to reach the nerve receptors of the lower skin. Thus, microneedle insertions are perceived as painless. The thesis presents research in the field of microneedle-based drug delivery with the specific aim of investigating a microneedle-based transdermal patch concept. To enable controllable drug infusion and still maintain an unobtrusive and easy-to-use, patch-like design, the system includes a small active dispenser mechanism. The dispenser is based on a novel thermal actuator consisting of highly expandable microspheres. When actuated, the microspheres expand into a liquid reservoir and, subsequently, dispense stored liquid through outlet holes. The microneedles are fabricated in monocrystalline silicon by Deep Reactive Ion Etching. The needles are organized in arrays situated on a chip. To allow active delivery, the microneedles are hollow with the needle bore-opening located on the side of the needle. This way, the needle can have a sharp and well-defined needle tip. A sharp needle is a further requirement to achieve microneedle insertion into skin by hand. The thesis presents fabrication and evaluation of both the microneedle structure and the transdermal patch as such. Issues such as penetration reliability, liquid delivery into the skin and microneedle packaging are discussed. The microneedle patch was also tested and studied in vivo for insulin delivery. Results show that intradermal administration with microneedles give rise to similar insulin concentration as standard subcutaneous delivery with the same dose rate.

Niclas Roxhed, [email protected] Microsystem Technology Laboratory, School of Electrical Engineering KTH—Royal Institute of Technology, SE-100 44 Stockholm, Sweden iv A Fully Integrated Microneedle-based Transdermal Drug Delivery System CONTENTS v

Contents

Abstract iii

List of papers vii

1 Objectives and Overview 1 1.1Structure...... 1

2 Transdermal Drug Delivery 3 2.1Introduction...... 3 2.2 Conventional needle-based administration ...... 3 2.3Continuousdrugdeliveryandinfusionsystems...... 4 2.4 Traditional transdermal patches ...... 6 2.5 Alternative techniques for transdermal delivery ...... 7 2.5.1 Jetinjectors...... 7 2.5.2 Iontophoresis...... 8 2.5.3 Sonophoresis...... 9 2.5.4 Chemical penetration enhancers...... 10 2.5.5 Skinablation...... 10 2.5.6 Microneedles...... 11 2.6Intradermaldrugdelivery...... 11

3 Skin as a Barrier 13 3.1Skinanatomy...... 13 3.2Mechanicalpropertiesoftheskin...... 14 3.3Modelingtheskin...... 15 3.3.1 Skindeformation...... 16 3.3.2 Skinfracture...... 17 3.4Implicationsonmicroneedletechnology...... 18

4 Microneedles for Drug Delivery Applications 21 4.1Generalaspectsonmicroneedles...... 21 4.1.1 Microneedle types and applications ...... 21 4.1.2 Microneedlesfordrugdelivery...... 22 4.2MEMS...... 25 4.2.1 Generalfabricationtechniques...... 26 4.2.2 Deepreactiveionetching...... 26 vi A Fully Integrated Microneedle-based Transdermal Drug Delivery System

4.3Solidmicroneedlearrays...... 28 4.4Hollowmicroneedlearrays...... 33 4.5Concludingremarks...... 38

5 Microneedle-based Systems 39 5.1Vision...... 39 5.2Dosingsystems...... 40 5.2.1 Passivedelivery...... 40 5.2.2 Activedelivery...... 41 5.3Integratedmicroneedlesystems...... 42

6 Development of an Integrated Microneedle-based System 45 6.1Dosingandactuationunit...... 45 6.1.1 Design...... 45 6.1.2 Fabrication...... 46 6.2Ultra-sharphollowmicroneedles...... 46 6.2.1 Design...... 46 6.2.2 Fabrication...... 47 6.3Membrane-sealedmicroneedles...... 50 6.3.1 Design...... 51 6.3.2 Fabrication...... 51 6.4Integratedmicroneedlesystem...... 52 6.4.1 Design...... 52 6.4.2 Fabrication...... 52 6.5Results...... 53 6.5.1 Dosingunit...... 53 6.5.2 Ultra-sharpmicroneedles...... 53 6.5.3 Membrane-sealedmicroneedles...... 55 6.5.4 Integratedmicroneedlesystem...... 56 6.6Discussion...... 57

7 Summaries of the Appended Papers 59

8 Conclusions 61

Acknowledgements 63

References 65

Glossary 81

Paper reprints 83 LIST OF PAPERS vii

List of papers

The presented thesis is based on the following international reviewed journal papers:

1. A Compact, Low-cost Microliter-range Liquid Dispenser based on Expandable Microspheres N. Roxhed, S. Rydholm, B. Samel, W. van der Wijngaart, P. Griss and G. Stemme Journal of Micromechanics and Microengineering, vol. 16, pp. 2740–6, Dec. 2006.

2. AMethodforTaperedDeepReactiveIonEtchingusingaModiﬁedBoschPro- cess N. Roxhed,P.GrissandG.Stemme Journal of Micromechanics and Microengineering, vol. 17, pp. 1087–92, May 2007.

3. Penetration-enhanced Ultra-sharp Microneedles and Prediction on Skin Inter- action for Eﬃcient Transdermal Drug Delivery N. Roxhed, T. C. Gasser, P. Griss, G. A. Holzapfel and G. Stemme IEEE/ASME Journal of Microelectromechanical Systems, accepted for publication, tentative print, Aug. 2007.

4. Painless Drug Delivery through Microneedle-based Transdermal Patches featuring Active Infusion N. Roxhed,L.Nordquist,B.Samel,P.GrissandG.Stemme IEEE Transactions of Biomedical Engineering, accepted for publication, tentative print, July 2007.

5. Novel Microneedle Patches for Active Insulin Delivery are Eﬃcient in Main- taining Glycaemic Control: An Initial Comparison with Subcutaneous Admin- istration L. Nordquist, N. Roxhed,P.GrissandG.Stemme Pharmaceutical Research, vol. 24, pp. 1381–8, July 2007.

6. Membrane-sealed Hollow Microneedles and Related Administration Schemes for Transdermal Drug Delivery N. Roxhed,P.GrissandG.Stemme Submitted for journal publication. viii A Fully Integrated Microneedle-based Transdermal Drug Delivery System

The contribution of Niclas Roxhed to the diﬀerent publications:

1 major part of design, fabrication, experiments and writing 2 major part of design, all fabrication and experiments, major part of writing 3 major part of design, all fabrication, major part of experiments and writing 4 major part of design, all fabrication, major part of experiments and writing 5 major part of design, all fabrication, part of experiments and writing 6 major part of design, all fabrication, all experiments, major part of writing

The work has also been presented at the following international reviewed conferences:

7. Low Cost Device for Precise Microliter Range Liquid Dispensing N. Roxhed, S. Rydholm, B. Samel, W. van der Wijngaart, P. Griss and G. Stemme 17th IEEE Int. Conf. on Micro Electro Mechanical Systems, Maastricht, The Netherlands, Jan. 2004, pp. 326–9. 8. Microﬂuidic Dye Laser with Compact, Low-Cost Liquid Dye Dispenser S. Balslev, N. Roxhed, P. Griss, G. Stemme and A. Kristensen Proceedings of μTAS 2004 8th Int. Conf. on Miniaturized Systems for Chemistry and Life Sciences,Malm¨o, Sweden, Sept. 2004, pp. 375–7. 9. Generic Leak-free Drug Storage and Delivery for Microneedle-based Systems N. Roxhed,P.GrissandG.Stemme 18th IEEE Int. Conf. on Micro Electro Mechanical Systems,MiamiBeach,FL, U.S.A., Jan. 2005, pp. 742–6. 10. Reliable In-vivo Penetration and Transdermal Injection Using Ultra-sharp Hol- low Microneedles N. Roxhed,P.GrissandG.Stemme The 13th IEEE Int. Conf. on Solid-state Sensors, Actuators and Microsystems, Seoul, Korea, June 2005, pp. 213–6. 11. Compact, Seamless Integration of Active Dosing and Actuation with Micronee- dles for Transdermal Drug Delivery N. Roxhed,B.Samel,L.Nordquist,P.GrissandG.Stemme 19th IEEE Int. Conf. on Micro Electro Mechanical Systems, Istanbul, Turkey, Jan. 2006, pp. 414–7. 12. Tapered Deep Reactive Ion Etching: Method and Characterization N. Roxhed,P.GrissandG.Stemme The 14th IEEE Int. Conf. on Solid-state Sensors, Actuators and Microsystems, Lyon, France, June 2007, pp. 493–6. LIST OF PAPERS ix

In addition, the author has contributed to the following related work:

International reviewed journal papers

13. A Fast Passive and Planar Liquid Sample Micromixer J. Melin, G. Gimenez, N. Roxhed, W. van der Wijngaart and G. Stemme Lab on a Chip, vol. 4, pp. 214–9, Mar. 2004. 14. A Liquid-triggered Liquid Microvalve for On-chip Flow Control J. Melin, N. Roxhed, G. Gimenez, P. Griss, W. van der Wijngaart and G. Stemme Sensors and Actuators, B, vol. 100, pp. 463–8, May 2004.

International reviewed conference papers

15. Mechanically Bi-stable In-plane Switch with Dual-stiﬀness Actuators M. Sterner, N. Roxhed, G. Stemme and J. Oberhammer The 14th IEEE Int. Conf. on Solid-state Sensors, Actuators and Microsystems, Lyon, France, June 2007, pp. 1401–4. 16. Coplanar-waveguide Embedded Mechanically-bistable DC-to-RF MEMS Switches M. Sterner, N. Roxhed, G. Stemme and J. Oberhammer Proceedings of the IEEE MTT-S Int. Microwave Symposium, Honolulu, HI, U.S.A., June 2007, pp. 359–62. 17. Mechanically Tri-Stable SPDT Metal-Contact MEMS Switch Embedded in 3D Transmission Line M. Sterner, N. Roxhed, G. Stemme and J. Oberhammer Proceedings of the IEEE European Microwave Conference,Munich,Germany, Oct. 2007, to appear. Workshops

18. Miniaturized Drug Delivery System for Painless Trans- and Intradermal Injec- tions N. Roxhed, Invited talk Micro Dosing System Workshop, Munich, Germany, Oct. 2005.

The microneedle-based patch (papers 4 and 5) was also brought closer to the public by the full-page articles Pl˚aster kan bli ersättning för sprutor (Patches may replace syringes) published in the daily newspaper Upsala Nya Tidning, June 10, 2007, written by Ake˚ Spross; and Pl˚aster g˚ar lika bra (Patches are as good) published in the bi-monthly magazine Diabetes, i. 4, 2007, written by Ulla Ernström. x A Fully Integrated Microneedle-based Transdermal Drug Delivery System 1 OBJECTIVES AND OVERVIEW 1

1 Objectives and Overview

This thesis presents research in the ﬁeld of Microsystem Technology and speciﬁcally in the area of microneedle-based drug delivery. The objective of this thesis is to highlight the potential role of microneedles in achieving painless drug delivery of macromolecular drugs across the skin barrier. Furthermore, the concept of a microneedle-based transdermal patch is reviewed and discussed. This concept was further realized in form of a fully integrated microneedle system incorporating a drug reservoir as well as an electrically controlled dispensing mechanism.

1.1 Structure The structure of the thesis is as follows: chapter 2 provides an overview of transdermal drug delivery from classical transdermal patches to newer techniques currently in development. The beneﬁts and limitations of transdermal delivery are discussed and special emphasis is given to intradermal delivery, particularly for vaccination purposes. A brief introduction to the structure of the skin is given in chapter 3. The mechanical behavior of the skin is presented as well as possible ways to model this behavior. How the mechanical properties aﬀect microneedle-based delivery is also discussed. Chapter 4 introduces microneedle-based drug delivery and the concept of a microneedle-based patch. Methods of microfabrication are presented and the state-of-the- art in microneedle research is reviewed and commented. In chapter 5, the concept of microneedle-based patches is widen to also include active delivery capability. Requirements and possible approaches to such systems are discussed. Chapter 6 presents the development of a microneedle-based patch-like system with active delivery capability. Results from in vivo tests are presented and discussed. 2 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

If transdermal immunization works well, vaccination practice could be revolutionized

Stanley A. Plotkin in “Vaccines: past, present and future” Nature Medicine, vol 11(4), 2005 2 TRANSDERMAL DRUG DELIVERY 3

2 Transdermal Drug Delivery 2.1 Introduction Transdermal drug delivery means that a pharmaceutical compound is moved across the skin—the dermis—for subsequent systemic distribution. Hence, strictly seman- tically this does not only include the more commonly understood “patch”, but also traditional subcutaneous administration by means of a hypodermic needle and a syringe. Common to all methods of transdermal drug delivery, by this broad definition, is that the drug is passed through an artificial route into the body. The main advantage of this approach is that the drug is entered into the body undistorted without being passed through the body’s various defense systems. In contrast to oral administration (e.g. swallowing a pill), the most convenient way of drug administration, the transdermal route does not suffer from drug degradation in the gastrointestinal tract and reduced potency through first-pass metabolism (i.e. in the liver). In addition, oral-specific side-effects like liver damages are avoided, which are seen for example with common drugs like estradiol (estrogen) [1] or paracetamol [2].

2.2 Conventional needle-based administration Constituting the standard method of parenteral administration, the use of a hypodermic needle is an efficient way of delivering a drug. For example, a bolus delivery of a drug (e.g. a vaccine), may be completed within a minute using disposables with essentially no cost. Despite the effectiveness, this administration method has some major drawbacks. First, ordinary needles are associated with pain which cause problems with acceptance, and therefore compliance [3–5]. Second, the use of “sharps” raises concerns on device safety and safety for health care providers. This is of major concern especially in the developing countries [6, 7], where unsafe injections account for a significant portion of transmission of hepatitis B and C viruses and human immunodeficiency virus (HIV) [7]. For hepatitis B, as much as 20% of the viral transmissions are estimated to be caused in this way [8]. According to the World Health Organization

(WHO), more than 1.3 million deaths and costs of US °535 million are attributed annually to unsafe injection practice [9]. Third, and partly related to safety concerns, the use of “sharps” require trained personnel for administration and handling. Al- though patients can be trained in self-injections (as with diabetics or patients in daily 4 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

Toxic level

Therapeutic window Drug concentration Minimum eﬀective level

Time

Figure 1. Schematic drawing of drug concentration in blood plasma as function of time for diﬀerent administration schemes. Safe bolus dose (dash-dotted), unsafe bolus dose (dashed), controlled release by continuous infusion into the therapeutic window (solid). Redrawn from [10]. need of anticoagulant agents), this imposes a barrier for future potent drugs (such as nicotine conjugate vaccines) and excludes them from the over-the-counter (OTC) market. Besides increased cost of administrations, the need of trained personnel is also problematic especially in immunization of populations en masse,e.g.incaseof a pandemic inﬂuenza or bioterror threats [5].

2.3 Continuous drug delivery and infusion systems A bolus injection of drug into the body causes blood plasma concentrations to increase rapidly followed by an exponential decay as a result of metabolism and excretion. In pharmacokinetics this is usually modeled as a response of cascades of first-order linear systems. Thus, a bolus injection corresponds to an impulse response which as such lacks stationary levels. In pharmacology, therapeutic levels are defined by the range between the minimal effective dose and the toxic dose level (typically defined for 50% of the population). Hence, a therapeutic window is established which sets the limits on how a drug should be administered. A bolus injection may be designed to target the window but will inevitably fall short on keeping a certain therapeutic level. A continuous infusion on the other hand, corresponds to a step response which, at stable conditions, will reach a certain stable level. Keeping a certain level minimizes the risk of side-effects and insufficient medication. It is therefore desirable for most administration systems to resemble this effect, either by continuous infusion or by some kind of delay mechanisms that gradually releases the drug. Figure 1 illustrates the different administration schemes. Apart from the clear therapeutic advantage of controlled delivery, there are busi- nesslike reasons in achieving controlled-release devices. Through the tough requirements set up by authorities like the FDA (Food and Drug Administration), the cost of 2 TRANSDERMAL DRUG DELIVERY 5

Figure 2. The MiniMed Paradigm REAL-Time insulin pump and system for continuous glucose monitoring. Glucose is measured in vivo and wirelessly transmitted to the pump’s display system where the user is informed on the glycemic trends. Note, the system, in its current state, is not an automated feedback system. Reprinted with permission from Medtronic AB, Sweden.

introducing new drugs to the market is now estimated to more than US °200 million and the time from development to market may easily exceed 10 years [10]. In light of this, it is reasonable for pharmaceutical companies to try to extend the lifetime of an existing drug. A viable way to achieve this is to develop a new controlled-release formulation for the same drug. Demonstrating increased efficacy with a new release system over a single-dose formulation, either by an improved delivery device or just a new formulation for prolonged effect, might be measures enough to improve the product, keep down cost and maintain market position [10]. Typical solutions for prolonged action are sustained release agents or capsules that gradually dissolve in in vivo environments. However, since the effect of such techniques varies with the local environment and differs from patient to patient, a system where the drug is released independent of the patient is a more durable solution. A typical example, relevant to this thesis, where controlled release is beneficial and of outmost importance, is in diabetes management. To avoid long-term complications in diabetes, good glycemic control is an essential factor [11]. The use of a continuous infusion system via an insulin pump gives a more intensified therapy which, in turn, leads to improved glycemic control [11,12]. Figure 2 shows a representative example of a modern insulin pump. The pump is connected via a tube to a separate disposable infusion set. The infusion set, containing a hypodermic needle, is attached to the skin and typically replaced every third day. Through a user interface on the pump, the patient controls infusion speeds and insulin doses. Notably, apart from the pump and the infusion set, the system shown in figure 2 also includes an in vivo glucose monitoring device. Although this device, in its current state, only reports the current glucose value to the pump’s information display, the manufacturer most likely aims at a fully automated feedback system ultimately. Since control and safety requirements on insulin delivery are very high, market approval for such an “artificial pancreas” is an extensive procedure. 6 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

Drug Product name Clinical indication Manufacturer a Clonidine Catapres TTS Hypertension Boehringer Ingelheim

Estradiol Alora Hypoestrogenism Watson a Estraderm Novartis

Menostar Bayer a b Fentanyl Duragesic Pain relief Janssen

Lidocaine Lidoderm Post-shingles pain Endo

Methylphenidate Daytrona ADHD Shire

Nicotine Habitrol Smoking cessation Novartis a Nicoderm GlaxoSmithKline

Prostep Elan b Nicotrol McNeil

Nitroglycerine Minitran Angina 3M

Nitro-Dur Schering-Plough a Transderm Nitro Novartis

Oxybutynin Oxytrol Overactive bladder Watson

Selegiline Emsam Depression Somerset a Scopolamine Transderm Scop Motion sickness Novartis

Testosterone Androderm Hypogonadism Watson a b Testroderm Alza

aBased on Alza’s1 D-Trans technology. bSubsidiary of Johnson & Johnson.

Table 1. FDA-approved transdermal patches.

2.4 Traditional transdermal patches

While infusion pumps are reliable in achieving a preferred therapeutic delivery profile, the use of such a system (e.g. an insulin pump) is somewhat cumbersome, requires training, is costly, and requires a hypodermic needle-based infusion set. Transdermal patches, where the drug diffuses through the skin, offer a much more convenient way to administer a drug while still having the benefits of continuous drug release. Transdermal patches were introduced in the late 1970s, starting with a three- day patch to treat motion sickness. Since then, the market for drug administration through patches has been steadily increasing and by 2004 the annual U.S. market

value was more than °3 billion with several kinds of drug formulations available [13]. In 2001, 51 of 129 drug delivery products under clinical evaluation in the U.S. were transdermal or dermal systems [14]. Still, only about eleven drugs are presently available with transdermal patches, see table 1. The fundamental reason why so few drugs are used is that the barrier property of the skin limits the use of patches to therapeutics where the molecule size is small enough to diﬀuse through the skin at therapeutic rates. The drugs presently used in patches have molecular masses ranging from 162 u (nicotine) to 357 u (oxybutynin), giving a practical dose rate of 4–20 mg/day depending on the patch size. As a comparison, the mass of the insulin molecule is 5808 u and modern DNA-based vaccines, built of vectors with thousands of base-pairs, may have molecular weights in the order of hundreds of kilounits (kDa). Transdermal patches are generally divided into two categories based on their phys-

1Alza corp. is the leading provider of transdermal drug delivery technologies. In 2001, the

company was acquired by Johnson & Johnson in a stock-for-stock transaction worth US °10.5 billion. 2 TRANSDERMAL DRUG DELIVERY 7

Backing Rate-controlling membrane

Drug solution

Peel strip Adhesive

Figure 3. Schematic drawing of a reservoir-based transdermal patch. ical structure: reservoir-based and matrix-based. Reservoir-based patches hold the the drug in a solution (usually a liquid or a gel) in a separate compartment. The drug is released through a rate-controlling permeable membrane placed as an interface between the reservoir and the skin. Figure 3 shows schematic drawing of a reservoir-type patch. Matrix-based patches have a more simple design in which the drug is incorporated with the adhesive layer. There is no membrane that controls the release rate of the drug. Instead, the permeability of the skin governs the rate control. Matrix-based patches are easier to fabricate and thus the production cost is lower than for reservoir-based patches. On the other hand, reservoir-based patches oﬀer better control of the drug release but may raise safety concerns since a possible rupture of the membrane could result in a sudden release of the drug.

2.5 Alternative techniques for transdermal delivery 2.5.1 Jet injectors Jet injectors are hand-held devices that deliver a high-pressure liquid stream through a small nozzle orifice. The impact of the stream is high enough to penetrate the skin tissue and by controlling the magnitude, the liquid can be delivered to specific tissue depths, e.g. intradermal, subcutaneous or intramuscular. The major advantage of jet injectors is the extremely efficient way of drug administration where drug doses can be fired off sequentially, allowing up to 1000 subjects to be medicated per hour [15]. Jet injectors have been used in military and mass vaccination campaigns since the 1950s but their use was discontinued after an outbreak of hepatitis B in 1985 was linked to the use of jet injectors [16]. With the emergence of HIV, the use of jet injectors became too much of a risk and ceased definitely in 1997 as the U.S. Department of Defense announced that they would stop their use. During recent years, with renewed interest due to bioterror threats, a new gen- eration of safer jet injectors have been developed which use single-dose cartridges

and, for example disposable caps to eliminate the risk of cross-contamination be-

tween injections. Several systems are commercially available, e.g. Bioject ,Injex ,

ç ç Intraject , and J-Tip . A similar system called PMED also exists where the jet- injected substance has the form of a dry powder (PowderMed Ltd, a Pfizer Inc. subsidiary) and where the delivery specifically targets immuno-competent cells in the skin layer. Due to this efficient location of vaccine delivery (cf. sec 2.6), immune response has been achieved with 20–2500-fold lower doses as compared to conventional intramuscular delivery (using a needle and syringe) [17]. In addition, a big advantage of powder-based formulations of vaccines is that the so-called “cold-chain” could be 8 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

Electronics/drug

LED indicator

Activation button

Figure 4. Artistic drawing of Alza’s FDA-approved iontophoretic transdermal patch, E-Trans . The disposable system is approximately 7 cm long and incorporates a drug reservoir, a battery and an ASIC chip. eliminated [15]. The “cold-chain”, meaning an uninterrupted storage of drugs between 2–8 ◦C throughout the whole supply chain, imposes major complications and additional costs especially for vaccinations in more remote areas. While jet injectors are especially suited for mass vaccinations, injectors for insulin delivery have also been evaluated and commercialized (Medi-Jector ) [18, 19]. Like standard needle-based injections, jet injectors deliver a bolus shot and the injection can cause pain. Some injectors however, are speciﬁcally designed to cause small or no pain [19]. A system for continuous jet delivery using a piezoelectric actuator has also been proposed [20].

2.5.2 Iontophoresis Iontophoresis refers to the delivery of drugs across the skin by means of an electric field. By having two electrodes placed on the skin, drugs at the electrodes will start to migrate through the skin once a voltage is supplied to the electrodes. Once in the skin, the drug will be absorbed by the capillaries and systemically distributed. The current density is usually below 0.5 mA/cm2 in order not to cause the patient any discomfort. Three main physical mechanisms are involved in iontophoresis: charged species are driven from the electrodes as a result of the electric field, i.e. electrophore- sis; the flow of current increases the permeability of the skin; and, the established potential difference between the electrodes give rise to an electroosmotic flow [14]. Since electroosmosis occurs, uncharged species can be delivered as well. The efficiency of iontophoretic mass transport is closely related to the properties of the drug. Polarity, valency and mobility of the drug molecules are essential, with the mobility being tightly linked to the size of the molecule. As a consequence, iontophoresis is considered to be a feasible procedure for molecules of less than 7000 u. The flux of small molecules below 1000 u is typically 20–50 mg/day, whereas molecules above 5000 u generally leads to less than 1 mg/day [21]. However, effective transport at rates of 6 mg/day · cm−2 has been reported even for 12000 u-sized proteins [22]. There are several iontophoretic delivery systems commercially available. Most of them are electrode-based system with benchtop control units that require the physi- cian to add a drug to a reservoir on the electrode. In 2004, Vyteris Inc. (a Becton 2 TRANSDERMAL DRUG DELIVERY 9

Dickinson spin out) received an NDA (New Drug Application) approval from the FDA for a lidocaine-prefilled iontophoretic transdermal system used for local anal- gesia (pain relieve) in clinics. This system consists of a disposable drug-filled patch which is electrically connected to a reusable, portable and battery-operated, con- troller. In 2006, Alza corp. received an NDA on a fully integrated iontophoretic patch prefilled with fentanyl (pain relief). Based on Alza’s E-Trans platform, the disposable patch system, called Ionsysç , incorporates drug, battery, LEDs (Light Emitting Diodes) and an ASIC (Application Specific Integrated Circuit), see figure 4. The system is activated and operated by the patient and each activation renders a

40 ñg dose, delivered within 10 min. The ASIC ensures proper usage by the patient and compensates for physiological variations by controlling the iontophoretic current [23]. Another specific application field for iontophoresis is diabetes care. Although not used for drug delivery, non-invasive glucose monitoring by means of iontophoresis has been demonstrated and commercialized [24]. The GlucoWatch ,awrist-worn device, uses iontophoretic current to sample glucose from the interstitial fluid in the skin. By incorporating a replaceable enzymatic (glucose oxidase) glucose sensor in the device, the blood glucose of the patient can be estimated and continuously recorded. Delivery of insulin (5808 u) using iontophoresis has also been investigated, cf. [25] and references therein. However, basal rates of insulin needed by diabetics is 0.5– 1 mg/day and even this low basal drug input exceeds the theoretical flux predicted for iontophoretically delivered monomeric insulin on a 10 cm2 area [26].

2.5.3 Sonophoresis Another method to move drugs across the skin barrier is sonophoresis where skin is made permeable under influence of ultrasonic waves. The technique has been used frequently for over half a century, e.g. with hydrocortisone in combination with physical therapy of joint-related complications (arthritis) [13]. Traditionally, frequencies above 1 MHz have been used in order to reach and simultaneously stimulate tissues (joints and muscles) below the skin. However, during the past decade a considerable amount of research has been directed towards low-frequency sonophoresis (<100 kHz) shown to interact with the superficial skin tissue and increasing transdermal transport by orders of magnitude [27,28]. Thermal, chemical and mechanical alterations in the skin are considered to be the main transport enhancing mechanism in sonophoresis. For low-frequency sonophoresis, the formation and collapse of bubbles within the cells (cavitation) causes disruption of the skin tissue and is believed to be the predominant effect by which the method works [21]. Administration of many different drugs, e.g. insulin, low-molecular weight heparin, and vaccine, have been demonstrated feasible with low-frequency sonophoresis [28–30]. Typical acoustic intensities range from 0.25–1 W/cm2. In 2004, Sontra Medical Corp. (a spin off from R. Langer’s lab at MIT) received FDA approval for the first sonophoretic transdermal delivery system. The system, SonoPrep , is aimed for lidocaine administration (pain relief) and consists of a portable base unit connected to an ultrasonic horn that is pressed onto the area of skin to be treated. The company is also developing an non-invasive glucose monitoring system based on sonophoretic disruption of the skin [29, 31]. 10 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

While sonophoresis has been demonstrated to be an effective method to permeate skin, an integrated patch-like system, such as the one shown in figure 4, may be a more challenging task considering that it would need driving circuitry and a transducer delivering an acoustic power of about 1 W (i.e. 120 dB) to be therapeutically effective.

2.5.4 Chemical penetration enhancers

The delivery rate of conventional transdermal patches can be increased significantly by using chemical penetration enhancers in combination with the patch. The simplest form of penetration enhancement is the use of water. Hydration of skin tissue progressively increases permeability as water opens up the compact structure of the outermost skin layer [14,32]. The layer is also extremely hygroscopic as up to 500% of the dry weight can be absorbed within 1 h (by immersion) causing the layer thickness to increase 4–5 times [32]. Consequently, moisturizing factors like occlusive films or hydrophobic ointments (e.g. oily creams) also lead to increased skin permeability. Other penetration-enhancing chemicals work by diminishing the barrier property of the outermost skin layer. A great variety of chemicals are known to posses this capability. Some of the more common ones are surfactants (like Tween), fatty acids (like oleic acid), terpenes (e.g. eucalyptus oil) and solvents (e.g. ethanol) [33]. How- ever, to have significant penetration enhancing effect, the amount needed of these chemicals may easily become irritating for the skin and the use of them may therefore be physiologically incompatible [13]. Although chemical penetration enhancers alone have a limited effect on skin permeation, synergistic effects of enhancers with other transdermal delivery techniques (e.g. iontophoresis and sonophoresis) can promote the effectiveness of these methods considerably [34].

2.5.5 Skin ablation

The outermost layer of the skin, the main physical barrier of the skin, consists of dead keratinized cells. A straightforward approach to increase transdermal transport is therefore to simply remove this layer. A common approach among dermatologists and other professionals working with the skin is to use adhesive tape to remove or weaken the layer [35]. As an example, it has been determined that a doubling of the TEWL (TransEpidermal Water Loss), a measure of the skin permeability, occurs after approximately eleven successive tape strips with standard surgical tape [36]. Other techniques to remove the outermost skin layer include microjets of particles that cut through the layer [37] and thermal ablation methods making microconduits by burning away small micrometer-sized areas [38–41]. Since ablation only occurs on the superficial layer of the skin, these methods are reported to be painless. Of the thermal methods, different strategies are used to facilitate ablation, e.g. pulsed laser [38], arc discharge [39] or short-duration resistive heating [40, 41]. The latter is being commercialized by Altea Therapeutics Corp. Their PassPortç patch system resembles a classical transdermal patch but has integrated heater elements that are in contact with the skin. A separate handheld “applicator” activates the delivery by inducing a current into the patch’s heaters. The company is developing the system for analgesics, insulin and vaccines [41]. Like the transdermal methods mentioned 2 TRANSDERMAL DRUG DELIVERY 11 previously, thermal skin ablation has also been used to extract interstitial fluid for glucose monitoring [40].

2.5.6 Microneedles In recent years, attention has been drawn to a new type of delivery method where arrays of miniaturized needles are used to penetrate the skin layer. Since the needles are short, they do not reach the nerve-rich regions of the lower parts of the skin. As a consequence, the stimulus caused by microneedle insertion into the skin is weak and perceived as painless [42, 43]. By employing batch-fabrication techniques from the microelectronics industry, small-scale microneedles can be mass-produced with high precision and reproducibility in a cost-effective manner. Reviews of microneedles can be found through the following references: [44–46]. By combining microneedles with a patch-like structure, a system can be realized which essentially has all the favorable properties of a traditional transdermal patch, i.e. continuous release, ease-of-use, unobtrusiveness and painlessness. Unlike the standard patch, a microneedle-based patch enables delivery of virtually any macromolecular drug (including insulin and vaccine). Such a patch would not only offer a discreet and patient-friendly drug administration system, but also an efficient and possibly safe way to administer drugs with minimum involvement from health-care professionals. A number of Fortune 500 companies, as well as start-ups, are actively developing microneedle technology for transdermal drug delivery [13]. Most of them work with solid, non-hollow, needles. The remaining part of this thesis, except for parts of the next chapter, will pre- dominately focus on microneedles and microneedle-related systems for drug delivery across the skin.

2.6 Intradermal drug delivery Most of the methods presented in this chapter actually deliver the drug into the skin tissue itself, i.e. intradermally. However, they are referred to as transdermal techniques since the primary goal is to achieve systemic distribution. Hence, the skin itself merely functions as a buffer space for subsequent diffusive spread of the drug into the capillaries and further to the bloodstream. Skin as the delivery site may be advantageous since it, in the lower parts, contains a dense vascular network which can promote the uptake of the drug. In contrast to this, for certain applications the skin itself may be the targeted organ or a preferred site of delivery. A typical example is chemotherapy of certain skin lesions (e.g. cancerous skin). In case of cytotoxins, it is of outmost importance to deliver the drug locally, i.e. in the skin, and preferably nowhere else. Although feasible, intradermal delivery by means of a standard needle and a syringe is difficult and very hard to control. Hence, the transdermal, or more precisely intradermal, delivery methods discussed in this chapter are particularly useful in applications where the skin itself should be medicated. Another application, and possibly one of the most important, is immunization. The skin is populated with numerous highly potent resident antigen presenting cells (APCs) called Langerhans’ cells and dermal dendrocytes. These cells play a vital role 12 A Fully Integrated Microneedle-based Transdermal Drug Delivery System in induction of immune response. Once activated by antigens, the cells migrate from the skin to draining lymph nodes where the antigens are presented to B and T-cells, the white blood cells that initiate production of antibodies, identify pathogens and destroy infected cells [47]. The cells are dendritic cells, cells that feature branched projections, and reside as a network over the whole skin layer. Dendritic cells are the most potent of all APCs and developing immunization strategies that optimize antigen presentation by these cells is a rational approach to vaccine delivery [48]. Hence, targeted intradermal delivery to stimulate these cells is an attractive strategy for vaccination in order to elicit powerful immune responses [49,50]. A number of studies have specifically addressed intradermal vaccine delivery, and the results are promising. For example, in an effort to cover for the loss of half the U.S. supply of influenza vaccine, dose-sparing by intradermal administration was suggested as a viable approach [51]. It was demonstrated that a fifth of the conventional vaccine dose elicit similar or stronger immune responses to common influenza strains (H1N1, H3N2) following intradermal administration using a fine-gauge needle compared to conventional intramuscular administration. Recently, these results were confirmed and also shown to be better in some aspects (e.g. 10-fold dose-sparing for H1N1) [52]. Studies where targeted intradermal delivery techniques are employed show even better results. For example, delivery of ovalbumin (a model protein antigen) by precoated solid microneedles showed up to a 100-fold increase in immune response over intramuscular delivery by a standard hypodermic needle and the same dose [53]. A later study from the same group confirms a significant immune response difference between intradermal and intramuscular delivery over a range of different microneedle lengths [54]. As mentioned in section 2.5.1, jet injection of powder-based DNA vaccine into the skin required 20–2500-fold lower doses than usually administered to obtain the same effect [17]. In another study where a type of blunt microneedle array (microenhancers) precoated with DNA vaccine was used to scrape the skin, stronger and less variable immune responses were achieved compared to intramuscular needle injection [55]. Furthermore, intradermal delivery with the array required fewer immunizations to elicit full immunity. The WHO estimates that approximately 1 billion vaccinations are administered

annually [56]. The global vaccine market it estimated to US °10 billion in 2007 (where

80% stem from the developed world) and expected to grow to over °23 billion in 2012, largely as a result of introduction of cancer vaccines [57]. Thus, the potential for intradermal vaccinations and new delivery techniques that enhance the eﬀectiveness is enormous. The quotation in the beginning of this chapter can be seen in that respect. 3 SKIN AS A BARRIER 13

3 Skin as a Barrier

As reviewed in the previous chapter, one of the major limitation to successful transdermal drug delivery stem from the skin itself and its property of being an excellent physical barrier. While transdermal patches, passive or physically assisted, are limited by the dense tissue to deliver molecules of a certain size, methods that circumvent the skin barrier (i.e. ablative methods, jet injectors or microneedles) are not restricted by the size of the drug molecule. Nevertheless, the barrier property of the skin poses a challenge for these methods as well; less from an intercellular or chemical point of view, but instead from a mechanical perspective. Microneedles, as any other foreign microobjects, are hindered to enter the body by the skin’s though and extremely ﬂexible structure. Understanding how microneedles interact with the tissue help in designing needles, and to understand penetration mechanisms and intradermal liquid transport. This chapter will give an overview of the skin structure, its mechanical properties, and ways to model the skin from a mechanical perspective.

3.1 Skin anatomy The skin is the largest organ of the human body and has several functions. It is a physical barrier towards the environment, it regulates body temperature and ﬂuid loss, it conveys sensory information to the nervous system, and it processes immunologic information to the immune system. The skin can be divided into three main layers: the superﬁcial epidermis, dermis

and hypodermis, see ﬁgure 5. The epidermis is approximately 50–150 ñmthickand consists largely of constantly renewing, outward moving cells called keratinocytes. Apart for these cells, most of the antigen-presenting Langerhans’ cells are located in the epidermis. The outermost layer of the epidermis is the stratum corneum, a 10–

20 ñm thick layer of 15–30 stacked, dead, corniﬁed cells. These so-called corneocytes are ﬂat, hexagon-shaped and partly overlapping cells with a diameter of approximately

30 ñm. The cells are mechanically coupled to each other through special protein rivets (figure 6a) and together with stacked layers of lipids they form an interlinked mechanical scaffold [58]. The stratum corneum forms the major constituent of the water barrier in the skin [59]. The dermis represents the bulk of the skin and the predominant components are collagen fibers and a smaller amount of elastin. This fibrous network gives tensile strength and elasticity to the skin and also provides support for nerve and vascular 14 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

Sweat gland Hair shaft Stratum corneum 10–20 μm Living epidermis 30–130 μm

Dermis 1.1 mm Papillar region

Reticular Hypodermis 1.2 mm region

Figure 5. Cross-sectional illustration of human skin. From [60]. networks. In the upper, papillary, region of the dermis the collagen ﬁbers are small and loosely distributed. The deep, reticular region contains densely packed, bundled, collagen ﬁbers mainly running parallel to the skin surface and along certain directions, called Langer’s lines [59,61]. The dermis rests on the hypodermis (subcutis) which is composed of loose fatty connective tissue. Its thickness varies considerably over the surface of the body as well as between individuals [58].

3.2 Mechanical properties of the skin As already described by Langer in 1861, the skin is a highly anisotropic, heteroge- neous tissue subjected to prestress in vivo [62]. In addition, the skin is non-linear viscoelastic. The mechanical and structural properties of the skin vary significantly with age, skin type, hydration, body location, and between individuals [61,63]. Hence, general quantitative descriptions of the skin are very difficult to obtain. Qualitatively, looking at the skin as a whole, the mechanical response can be described as non-linear stiffening, i.e., a J-shaped stress-strain curve (figure 6b). This relationship is thought to stem from the collagen fibers in the dermis [65]. At low strains, the fibers are crimped and remain mechanically inactive while elastin accounts for the stiffness, giving rise to a fairly linear low-modulus response (0.1–2 MPa). At higher strains, collagen fibers straighten out and gradually start to bear load which cause a non-linear response. At even higher strains, all the fibers are stretched and the stress level increases rapidly (high-modulus response, 1–80 MPa) [61]. While this description can be seen to represent the general mechanical response, or the “bulk” property of the skin, individual skin layers have rather different properties. For interaction with small objects like microneedles, whose length is in the same order as the layer thicknesses, the influence of the individual layers becomes significant. Stratum corneum is generally regarded as the main physical barrier of the skin. 3 SKIN AS A BARRIER 15

Desmosome “rivet” 20

15 Corneocyte I II III

Stress / MPa Stress 10

0 0 0.2 0.4 0.6 0.8 1 Strain (a) (b)

Figure 6. (a) Coupling between the corniﬁed cells of the Stratum corneum. Redrawn from [58]. (b) Characteristic J-shaped stress versus strain relationship of skin. The regions I–III, represent low to high-modulus responses, respectively. Redrawn from [64].

The layer is relatively stiff compared to underlying tissues. Reported values from in vitro test of the elastic modulus range from 6 MPa up to 8900 MPa, largely dependent on the hydration level with less stiffness in hydrated cases [66–69]. Reports on isolated experiments on the viable epidermis are hard to find in the literature. While many experimental studies exist on skin (see e.g. [65] and references therein), it is often unclear which part of the skin that gives the measured behavior. Presumably, most studies measure the response of the dermis. However, one recent study by Hendriks et al. specifically deals with the mechanical properties of the different skin layers [70]. Describing skin as a two-layered hyperelastic Mooney material (cf. section 3.3.1) together with in vivo tests at small strains, they found the upper layer of the skin (epidermis and papillar dermis) to be orders of magnitude less stiff than the lower reticular dermis. With the reticular dermis having a (Mooney-Rivlin) material parameter of 0.16 MPa, this is well in the range of reported values for the stiffness of the dermis. However, the estimated stiffness parameter for the upper skin was only 0.11 kPa [70]. This indicates that the epidermis is extremely flexible and that the influence of the stiff stratum corneum, at small strains, is negligible.

3.3 Modeling the skin Despite the difficulties in quantifying the mechanical properties of skin, descriptive constitutive models are nevertheless useful to understand and estimate interactions with the tissue. Many of the early studies on skin tissue use Young’s modulus to de- scribe its mechanical properties (cf. [71] and references therein). However, as skin does not obey Hooke’s law (i.e. linear stress-strain relationship), it is rather meaningless to use Young’s modulus unless the exact strain level is specified. As biomechanics evolved into a scientific discipline in the 1960s, new more sophisticated models of soft tissue were developed [72]. Examples of such constitutive models, aimed to include some of the characteristics of skin, are those of Tong and Fung (anisotropy) [73] and Lanir (anisotropy and viscoelasticity) [74]. 16 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

3.3.1 Skin deformation Constitutive models pioneered by Fung et al. were followed by a large variety of different models to account for certain specific phenomenas observed in soft tissue (for reviews, see e.g. [72] or the thesis from Hendriks [71]). Of these often very theoretical models, multi-phase approaches may be of particular interest for microneedle-based drug delivery. Oomens et al. proposed such a model for skin, based on the hypothesis that skin behaves as a sponge-like material [75]. The model, which consisted of a porous solid (representing the fibrous network) in combination with a free movable fluid (interstitial fluid), represents a similar approach as those used in soil mechanics. When the sponge-like material is compressed, the fluid will be pressed away. As the pore size decreases upon compression this leads to increased viscous drag and a non- linear time-dependent deformation response [75]. For microneedles that are aimed to inject liquid into such a material, the compressed pores with high flow resistance impose a barrier for the injected liquid which limits the effective flow rate. The advanced models developed for skin and soft tissue in the last decades all require characteristic material parameters (which in general need to be measured) and tailored implementation into finite element analysis (FEA) software. However, if the anisotropic and viscoelastic properties of skin are neglected, i.e. a quasi-static isotropic case is considered, skin can be modeled as a hyperelastic material [71, 76]. Yet advanced enough to account for large non-linear behavior, such models are widely available in common FEA packages (e.g. ANSYS ) and may be sufficiently good when considering static deformation of skin. Two rather recent studies make use of this simplified approach to investigate interaction with skin. Hendriks et al. used a Mooney-Rivlin model to study epidermal and dermal interaction with personal care products such as electric shavers (supported by Philips). A Mooney-Rivlin material is described by a strain-energy density function, φ, i.e. the elastic energy stored in the material per unit volume, and the function used was:

φ = C10(I1 − 3) + C11(I1 − 3)(I2 − 3) (1) where I1 and I2 are the ﬁrst and second invariants of the Finger strain tensor and C10 and C11 are material constants. When modeling the skin as a whole, the parameters were determined to C10 =9.4kPaandC11 = 82 kPa [77]. When the skin was divided into an upper part (epidermis and papillary dermis) and a lower part (reticular dermis) the second term of eq. 1 was left out and C10 was determined to 0.11 kPa and 0.16 MPa, for the upper and reticular layer, respectively [70]. A similar approach was used by Shergold and Fleck who investigated skin penetration by jet injectors (sec. 2.5.1) [78]. Instead of the Mooney-Rivlin model, they used an Ogden-type strain-energy function [79] since that better describes the strain- hardening occurring in skin at large strains [76]. The function used was: μ φ 2 λα λα λα − = α2 ( 1 + 2 + 3 3) (2) where λi (i =1, 2, 3) are the principal stretches and μ and α material constants. From experiments, μ =0.11 MPa and α = 9 were found to characterize the mechanical behavior of human abdomen skin under tensile loading [80]. 3 SKIN AS A BARRIER 17

The model and data proposed by Shergold and Fleck was used in the work presented in Paper 3 of this thesis. This paper reports on simulated interaction between microneedles and the skin. Solving the deformation problem when microneedles are gradually propagated into a soft material (i.e. skin), a stress field in the material could be obtained. The stress field was then used to calculate the hydrostatic pressure, p, in the bulk material according to p =trace(σ), where σ is the Cauchy stress tensor [81]. The hydrostatic pressure represents liquid depleted and compressed regions in the solid matrix. By plotting the hydrostatic pressure distribution, regions that have a high resistance to fluid flow can be identified. This, in turn, helps to understand if and how liquid injected from microneedles can flow within the compressed tissue. As such, the method can be a useful tool in designing microneedles so that liquid can be injected with minimum resistance and without leakage to the skin surface.

3.3.2 Skin fracture The existing literature on soft tissue penetration indicates that deep penetration involves cracking of the tissue followed by reversible deformation when the penetrator is removed [76]. Thus, following the general principle in fracture mechanics of elastic materials (Griﬃth’s theory [82]), fracture can be described as a result of work due to a displacement increment δl of a generalized load Q:

Qδl = δWC + δΦ(3) where δWC is the work required to create a crack and δΦ is the stored strain energy in the solid upon opening the crack. If the crack occurs as a result of tensile stress perpendicular to the crack plane (the most common way of crack formation), this is called a mode-I crack. The work required to create a mode-I crack of length 2a is determined by mode-I fracture toughness of the material, JIC. δWC is then given by:

δWC =2aJICδl (4) where 2aδl is the new surface created by the crack [76]. Shergold and Fleck investigated crack formation in human skin and rubber-like solids [76, 80]. They found that the penetration mechanism in soft solids depend on the penetrator tip geometry [80]. A sharp-tipped penetrator penetrates by the formation and wedging open of a planar mode-I crack. A flat-tipped hollow penetrator on the other hand, penetrates the solid by the growth of a mode-II crack (shear stress parallel to the crack plane). It was further found that the penetration pressure needed for a flat-tipped penetrator was several times larger than that for a sharp-tipped penetrator of the same diameter. It was argued that a flat-tipped penetration give rise to a similar kind of crack-tip blunting as in a trouser-tear test [76]. This is consistent with differences in reported data from different kinds of measurements on skin fracture. Reported value of skin fracture toughness (on rat) from a trouser-tear test is 26.9±2.7 kJ/m2 [83]. In contrast, reported fracture toughness from scissor cutting test (on human skin), which better represents a mode-I crack, is 1.7±0.6 kJ/m2 [84], i.e. an order of magnitude less than for the tear test. Specifically for microneedle application, Davis et al. made a thorough analysis of the forces needed for microneedle insertion into skin in relation to the needle 18 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

strength [85]. Using ﬂat-tipped hollow microneedles with outer diameters between ñ 60 ñm and 160 m, they found that the insertion force into human skin varies approximately linear with the interfacial area of the needle. By equating the applied work to insert the needle with the work needed to pierce the tissue (i.e. neglecting the second strain-energy term, of eq. 3), the fracture toughness was determined to 30.1±0.6 kJ/m2 [85]2. As pointed out by Purslow, however, for soft materials such as skin, energy storage and deformation preceding a crack can in general not be neglected [86].

3.4 Implications on microneedle technology To achieve physiologically relevant delivery rates, microneedle-based drug delivery is preferably made with arrays of needles over a certain area. To insert microneedle arrays with a large number of needles into the skin without using a special insertion tool (e.g. a high-velocity plunger), the insertion force needed to pierce the tissue has to be minimized. An example from the previous chapter may highlight the need of a low insertion force. Consider one of microneedle arrays for intradermal delivery mentioned in section 2.6. The 2 cm2 (low needle-density) array contains 280 non-hollow microneedles [54]. A reasonable force, applied manually on an area of 2 cm2 on, for example the lateral part of the upper arm, would be in the order of 10 N (i.e. ∼1 kg). Hence, distributed over the microneedles, the insertion force per needle needs to be below 35 mN (∼3.5 g)3. In addition, since the patch is aimed to deliver drugs, it should penetrate regardless of skin type or age of the subject, or the present humidity. All these factors may change the needed penetration force considerably. Thus, it is reasonable to believe that safety margins of 3–5 times may be required, which, in turn, leaves only a few millinewtons for the insertion! As demonstrated experimentally by Davis et al., penetration of microneedles into the skin is strongly linked to the the interfacial area between the microneedle and the skin [85]. A small interfacial area gives a low insertion force. In other words, a sharp needle, where the interfacial area is minimal, will penetrate the tissue better than a blunt needle. Following the findings of Shergold and Fleck, a sharp-tipped punch penetrates skin by a mode-I crack, whereas a flat-tipped hollow punch penetrates by a mode-II ring crack at a higher insertion force [80]. It is suggested that the latter is linked to higher fracture toughness [76]. The higher toughness value (30.1 kJ/m2)estimated by Davis et al [85] for flat-tipped hollow microneedles, compared to the low toughness value (1.7 kJ/m2 [84]) reported for mode-I-like rupture, supports this theory. In addition, and as a direct consequence of a ring-crack formation, a hollow needle with the bore opening facing the skin will punch out a piece of the tissue. For hollow microneedles aimed to deliver liquid, this leads to blockage of the fluid path and

2Davis et al. use the interfacial area of the microneedle as the newly created surface area. Hence, the created surface is independent of the thickness of the pierced material and not as usually described, a product between the crack length and the depth of the crack, i.e. 2aδl as written in eq. 4. 3The microneedle array referred to in this example is usually applied to the skin by an “impact applicator” [54]. 3 SKIN AS A BARRIER 19 hinders effective delivery. Both Shergold and Fleck, and Davis, report on this type of tissue coring [80,87]. Based on the findings regarding the crack type, microneedles with minimal insertion force should penetrate the tissue by the formation of mode-I crack. To obtain this crack type, a sharp-tipped microneedle is required. In the example given above, 280 non-hollow microneedles are used to achieve an appropriate drug delivery rate (in a small animal, though). It can be argued that an array of hollow microneedles, aimed for convective drug delivery, require far fewer needles to achieve therapeutic delivery rates. Thus, since fewer needles can be used, the insertion force of the needles does not necessarily need to be minimized. This is basically true. However, if a small number of needles are used, the delivery rate per needle needs to be higher than in the case of many needles. While a high flow rate may lead to significant flow resistance in micrometer-sized needle bores [88], the main fluid dynamical limitation lies in the tissue. A microneedle inserted into skin will cause a large deformation of the tissue around the insertion area. As a consequence, the tissue will be highly compressed, which leads to a concurrent reduction of the fluidic permeability in the tissue. As simulated in paper 3 of this thesis, the hydrostatic pressure, which is linked to the permeability, can be very high in stressed regions. Experimental studies on microneedle insertion into tissue confirm the occurrence of compressed, less permeable, tissue ahead of the needle [89]. By partially retracting the needle after insertion, and thus relieving the compressed tissue, the flow resistance is decreased [89–91]. A large flow resistance in the tissue does not limit flow rate per se. A large infusion pressure can indeed compensate for viscous losses in the tissue. However, a large fluidic pressure at the interface between the needle outlet and the tissue may create fluidic paths along the needle shaft and to the skin surface. In other words, leakage may occur. Hence, for a given microneedle geometry and tissue deformation, there is a maximum pressure with which liquid can be injected through the needle and into the skin. This, in combination with the fluidic resistance defines the maximal achievable infusion flow rate. Therapeutic infusion rates range from microliters per hour to several milliliters per hour depending on drug formulation and concentration. These flow rates have been reported in literature, but only in vitro and after precisely controlled retraction of the microneedles [89–91]. By using an array of many needles, the flow rate is distributed, which allows the tissue to absorb the liquid without leakage to the skin surface. In summary, to achieve therapeutic delivery rates with microneedles without leakage, the delivery has to be made over a certain area using many needles. In turn, to allow insertion by hand, an array with many microneedles requires a sharp needle where the insertion force is low. 20 A Fully Integrated Microneedle-based Transdermal Drug Delivery System 4 MICRONEEDLES FOR DRUG DELIVERY APPLICATIONS 21

4 Microneedles for Drug Delivery Applications

This chapter gives an overview of microneedles for drug delivery applications. The concept of miniaturized needles is presented and deﬁned. Speciﬁc requirements for microneedles aimed for transdermal drug delivery are discussed and the scope is de- limited. Some of the basic microfabrication methods used to fabricate microneedles are introduced and microneedles for drug delivery presented so far are reviewed and commented.

4.1 General aspects on microneedles Following conventional terminology, a microneedle is a needle with representative parts (e.g. diameter) on the micrometer length scale. However, this deﬁnition is rather bold as it includes most of the standard hypodermic needles used in medical practice. Although there are many examples of “microneedles” with lengths of a few millimeters described in the literature, a common understanding of microneedles is that the length of the needle is shorter than 1 mm. What can be said is that microneedles are signiﬁcantly smaller than ordinary needles, especially concerning the length.

4.1.1 Microneedle types and applications A classification for microneedles usually used in literature is based on the fabrication process: in-plane or out-of-plane microneedles. In-plane microneedles (figure 7a) are fabricated with the shaft being parallel to substrate surface. The advantage of this arrangement is that the length of the needle can be very accurately controlled. A disadvantage is that it is difficult to fabricate two-dimensional arrays. Out-of-plane microneedles (figure 7b) on the other hand, protrude from the substrate and are straightforward to fabricate in arrays. Instead, the length and high aspect-ratios become significant challenges in the fabrication of these kind of needles. Another useful point of distinction is whether the microneedles are solid or hollow. Hollow needles with a needle bore,orlumen, allow an active liquid transport through the microneedle. Microneedles have been used in many different applications, ranging from neuro- stimulation to gene delivery into individual cells. A common goal is to create a pathway to an object by physically circumventing some kind of barrier. In most applications this barrier is the skin. The rationale of using microneedles, as opposed 22 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

(a) (b)

Figure 7. (a) 6 mm long, hollow, in-plane microneedle. From Talbot and Pisano [92]. (b) Individ- ually addressable, 1.5 mm long, solid, out-of-plane microneedles used as electrodes. From Campbell et al. [93]. to macroscale devices, is motivated either by the size of the target or the benefit of piercing in a minimally invasive manner. One of the earliest reported microneedles in the scientific literature was an out- of-plane silicon needle array featuring 100, 1.5 mm long, needles on an area of 4.2 mm × 4.2 mm (figure 7b) [93]. These extremely slender needles were used as electrical electrodes and designed to stimulate the visual cortex of the brain in order to regain sight. Related to this application, in-plane, microneedle probes have been used for activity recording and cellular chemostimuli of brain tissue [94,95]. Solid, out- of-plane, microneedles have been used to penetrate the stratum corneum to facilitate EEG (Electroencephalogram) measurements for anesthesia monitoring [96,97]. Here,

arrays of 200 ñm long needles were used to circumvent the electrically insulating layer of the skin. Similar microneedle probes have also been used for diagnostic purposes, where the needles were used for impedance measurements of skin lesion in order to de- tect skin cancer [98]. The technique is currently being commercialized by SciBase AB and expected to reach market in 2007–2008 [99]. Another application for microneedles is sampling of body ﬂuids. Resembling the proboscis of a mosquito, Oka et al. fabricated a millimeter-long, jagged, hollow in-plane microneedle for blood collection [100]. Sampling of interstitial ﬂuid through capillary action has been demonstrated with ar-

rays of 350 ñm long, hollow, out-of-plane microneedles [101]. Microneedles have also been fabricated for microdialysis, where a hollow in-plane needle equipped with a semi-permeable membrane ﬁlters the sampled liquid [102]. Although other application ﬁelds exist for microneedles, the vast majority of published microneedles concern drug delivery in various forms. The following sections will give an overview of drug delivery microneedles, related fabrication techniques and design principles.

4.1.2 Microneedles for drug delivery The concept of an array of miniaturized needles for drug delivery purposes essentially dates back to 1976 and a patent (ﬁled 1971) from Gerstel and Place at Alza corp. [103]. In this patent, a drug delivery device featuring miniaturized projections 4 MICRONEEDLES FOR DRUG DELIVERY APPLICATIONS 23

Figure 8. The original concept of a microneedle-based transdermal patch from the patent of Gerstel and Place 1976 [103]. The device contains microneedles (12), a drug reservoir (16), adhesives (31) and optionally a rate-controlling membrane.

(i.e. microneedles) and a drug reservoir is claimed. The needles are small enough to penetrate only the stratum corneum and can be either solid or hollow. Delivery from the device may occur through diffusion or through convection by applying a force to the backing of the reservoir. Figure 8 shows a drawing from the original document illustrating the device. Although not fully to the point, some predecessors to Gerstel and Place’s patent exist, especially for vaccine delivery. For example, a disc-like device featuring an array of 1–2 mm long needles to permeate skin prior to vaccine appliance was patented by Rosenthal in 1952 [104] and a vaccine-filled patch-like device with short scarifying needles was patented by Kravitz and Lettvin in 1957 [105]. Although the concept of miniaturized needles for drug delivery was presented earlier, it was not until the 1990s that the technique was tested experimentally. A reason for this was that microfabrication techniques, evolving strongly at that time, enabled these micrometer-sized needles to be precisely fabricated in a potentially cost-effective manner. The first reported study on microneedles for transdermal drug delivery came 1998 by Henry et al. [106]. This work led by Allen, with a background in microfabrication and MEMS (Microelectromechanical systems), and Prausnitz, with a background at Alza corp. and in drug delivery research, demonstrated a four orders of magnitude increase in permeability of human skin after insertion of an array of

150 ñm long, solid silicon, out-of-plane microneedles. Given the governing goal to deliver a substance across the skin for subsequent systemic distribution, and the means (microneedles), several possible strategies can be employed to accomplish this. The simplest way, as also proposed by the early vaccinations strategies mentioned above, is to perforate the skin with microneedles and then apply the drug onto the skin for subsequent diffusive spread into the body. The drug can be applied to the skin surface as a gel or through a medicated patch to achieve prolonged release. Another way is to precoat the microneedles with the drug before they are inserted into the skin. A third option is to fabricate the microneedles in a biodegradable material that incorporates the drug. When the needles are inserted into the skin, the needles dissolve and the drug is subsequently released. If the microneedles are hollow, the drug can be actively injected into the tissue. Hollow needles can also be used with passive, diffusion-driven, delivery. In that case, the 24 A Fully Integrated Microneedle-based Transdermal Drug Delivery System needles merely functions as controlled and sustained paths (channels) into the body. The achievable dose for precoated and drug-embedded needles is naturally limited to the amount that the needles can bear. For moderately sized microneedle arrays, it is difficult to embed more than 1 mg. This may be sufficient for certain highly potent drugs (e.g. vaccines) but requires tailored drug formulations to be used. Diffusion- based methods are restricted to delivery rates obtained by the diffusive flux. Together with physiological elimination (excretion and metabolism), the small rates achieved by diffusion through micrometer-sized holes in the skin also limits the use to very potent drugs. Considerably higher delivery rates can be achieved with hollow microneedles. However, as discussed in section 3.4, the dense tissue enforces a practical limit on the delivery rate for this method as well. To maximize the delivery rate, a rational strategy for all the mentioned methods is to distribute the delivery over several microneedles. That is, by using an array of needles over a larger skin area, it exposes a larger area of the drug which promotes further diffusion to the capillaries. For hollow needles and injected drugs, delivery over a larger area lets the skin tissue absorb the liquid without being saturated, thus enabling higher infusion rates without leakage. In-plane microneedles are difficult to fabricate in two-dimensional arrays and are therefore less suited for general drug delivery applications. There may still be other applications for this type of needle, e.g. for body fluid sampling or localized delivery of small amounts into deeper tissue. However, as ordinary hypodermic needles decrease

in size (31G needles, 260 ñm in diameter, are already commercially available [107]), the advantage of single microfabricated needles over small size cannulas is not obvious. Guided by a vision of microneedle-based drug delivery, a few but basic requirements for microneedles can be deﬁned:

Suited to the purpose: That microneedles should work in in vivo environments is rather obvious, but nevertheless needs to be emphasized. There has been a tendency, particularly in the microfabrication community, to use specialized fabrication techniques to make microneedles without demonstrating basic feasibility. Furthermore, and probably due to the lack of proper resources, some of the test methods used to validate functionality (when that is the case) are based on rather inadequate models, e.g. penetration into dead animalic muscle tissue or through thin solid ﬁlms. For microneedles to function properly, the needles need to have a certain length and a certain sharpness, and they should be fabricated in a material which can withstand the forces of matter. 4 MICRONEEDLES FOR DRUG DELIVERY APPLICATIONS 25

Batch compatible: As a minimally invasive medical device, microneedles for drug delivery will need to be disposable, single-use, devices to gain acceptance in medical practice. Competing with standard needles and syringes at minute 4

costs of approximately US °0.1 [56] , microneedles need to be produced in a cost-eﬀective manner. Hence, for general drug delivery applications such as vaccinations or insulin delivery, fabrication methods need to be batch compatible to be commercially competitive. Although a signiﬁcantly higher price is likely to be accepted seeing the advantages of a patch-like and painless delivery method, the additional cost over alternative techniques has to be reasonable.

Biocompatible: Microneedles are designed to be inserted into human tissue and as such they need to be compatible with the local environment, both in terms of toxicity and intended function. The duration of contact with the tissue range from minutes to days at the most (insulin infusion-sets are typically changed every third day). Hence, if the material is non-toxic in the short term, malfunctioning due to biological host response through e.g. biofouling is unlikely to occur. For hollow microneedles there is a risk of blocking the needle bore with cored tissue during insertion into the skin. Concerning toxicity, well-known, bioinert, materials such as titanium, stainless steel or gold, or biodegradable polymers such as PLGA (polylactic-co- glycolic acid), may be used with conﬁdence as microneedle material. In the scientiﬁc literature, it is often questioned whether silicon (the traditional microengineering material) can be used as microneedle material. Although biocompatibility has not been investigated for silicon microneedles inserted into skin, recent studies show that silicon as a material is biocompatible in prolonged (days to weeks) implantations [108–110]. These results indicate that silicon needles inserted into skin for shorter durations are unlikely to cause harmful toxicological reactions.

Since the first studies on microneedle-based drug delivery in 1998, there has been an increasing amount of activity in the field. Most of the work concerns microneedle fabrication techniques, but as the field has become more mature, detailed studies on delivery mechanisms and preclinical evaluations have become more frequent. Be- fore reviewing the current state of the art on microneedle-aided drug delivery, the technology platform used to fabricate microneedles needs to be introduced.

4.2 MEMS Microelectromechanical Systems (MEMS) or Microsystem Technology (MST) refers to devices with sub-millimeter features. MEMS extend the fabrication techniques developed at the microelectronics industry to add mechanical structures onto microdevices. As such, MEMS devices can be made to interact with the surroundings, control ﬂuidic ﬂows or simply be used as small-scale mechanical devices. Typical MEMS devices are

4Estimated cost at UNICEF for disposable needle and syringe including safety disposal; i.e. the costisbasedonveryhighvolumes. 26 A Fully Integrated Microneedle-based Transdermal Drug Delivery System various sensors, e.g. pressure, flow, acceleration sensors; microfluidic systems such as ink-jet printheads or chemical analysis systems, or micromechanical devices such as micromirror arrays or microswitches. Characteristic attributes of MEMS fabrication are miniaturization, parallelization and integration. Miniaturization allows fabrication of compact and energy-efficient, fast-responding, devices. Parallelization refers to batch fabrication methods inherited from the microelectronics industry in which thousands or millions of devices are con- currently produced. Integration refers to monolithic integration of electronics and to packaging techniques. With its heritage from microelectronics fabrication, the traditional MEMS material is highly purified crystalline silicon, the same material as used in electronic chips such as memory or processors. Although silicon is relatively cheap (considering the purity), many MEMS devices do not need the characteristic properties of silicon for its function. In many cases, polymer materials offer sufficiently good alternatives to silicon. Particularly for disposable devices, polymer replication technologies such as hot embossing or injection molding are attractive alternatives to produce microdevices at a much lower price per unit.

4.2.1 General fabrication techniques Typical MEMS fabrication techniques include very precisely controlled deposition and etching of materials. By utilizing differences in selectivity to the etchant between different types of materials, structures can be formed in a controlled manner. The structures to be fabricated are defined by a two-dimensional pattern. This pattern is transferred from an original photomask to a photosensitive film on a substrate by photolithography. The substrate is typically a silicon wafer with a thickness

of 300–700 ñm. Also in case that polymer replicated microdevices are desired, a common method is to fabricate the master device in silicon due to the precision achievable through silicon micromachining. Once a structure is defined on the substrate, materials can be etched with respect to each other. By consecutively redepositing, patterning and etching materials, intricate three-dimensional structures may be created. Common techniques to add material to the substrate are spin coating, physical vapor deposition (e.g. evaporation or sputtering) or chemical vapor deposition (CVD). Etching may be accomplished through wet etching (dipping into liquid solution) or dry, plasma-based, etching. In plasma-based etching a gas is excited into a reactive state, enabling reactions between the gas and the substrate to take place. By controlling the gas pressure, the relative amount of ions over reactive radicals can be adjusted, which in turn affects the degree of isotropy of the etch. An electric field (bias) may accelerate the ions and further increase the directivity of the etch. Such an anisotropic plasma-based etch is referred to as Reactive Ion Etching (RIE).

4.2.2 Deep reactive ion etching Although RIE generates anisotropic etch proﬁles, the etch rate is relatively low and it is diﬃcult to maintain high aspect ratios over etch depths of more than a few micrometers. In the mid 1990s, techniques were established to etch deeply (hundreds 4 MICRONEEDLES FOR DRUG DELIVERY APPLICATIONS 27

Etch mask (e.g. SiO2)

1st cycle anisotropic etch

1st cycle isotropic passivation

2nd cycle anisotropic etch, etc.

Scalloping Silicon substrate

Figure 9. Schematic drawing of DRIE through the Bosch process. The basic cyclic principle is shown in the foreground. Etched trenches after 80 cycles are illustrated in the background. of micrometers) into silicon while still maintaining high aspect ratios and straight sidewalls. The primary deep reactive ion etching (DRIE) technique, commonly called the Bosch process, was patented by Laermer and Schilp at Robert Bosch GmbH in 1994 [111]. In what can almost be seen as a paradigm shift in MEMS fabrication, DRIE through the Bosch process quickly became widespread key technology which enabled high aspect ratio structures (HARS) of up to 50:1 to be fabricated. Etch rates in

current etchers are about 10–30 ñm/min (depending on the etched area) and predicted 5 to increase to 50 ñm/min in a few years [112,113] . This can be compared to etch rates

of about 1 ñm/min for traditional wet etch processes. With the etch rates achievable with modern etchers, the process is now used in very high-volume fabrication of low- cost devices such as CMOS image sensors for mobile phones [113, 114]. In relation to this, traditional microelectronics fabrication equipment suppliers, such as market leading Applied Materials, are now entering the market of DRIE systems. Particularly for out-of-plane microneedles (in silicon or as silicon master for replication), DRIE oﬀer good options of fabricating high aspect-ratio structures. Especially hollow microneedles can have needle bores that are several hundred micrometers long with aspect ratios in the order of 30:1. By alternating between anisotropic etching and isotropic etching (i.e. unbiased continuous etching), intricate three-dimensional features can be made, including well-deﬁned and sharp needle tips. The basic philosophy in DRIE is to protect (passivate) the sidewalls during etching so that the etch only proceeds vertically into the bulk material. In the Bosch process this is achieved by repetitively switching between an RIE etching step and polymer

5 ñ Laermer [112] predicts 50 ñm/min in 2015. Equipment supplier Alcatel [113] predicts 50 m/min in 2010. However, the latter is likely to be under best-case conditions, i.e. an exposed etch area <1% and moderate etch depths. 28 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

(a) (b) (c)

Figure 10. (a–b) KOH-etched microneedles in silicon for DNA delivery into cells. From [121,122]. (c) Illustration from Reed and Weiss’ patent of hollow barbed needles [123]. deposition step which uniformly coats and passivates the sidewalls. The polymer layer, usually a Teﬂon-like layer polymerized from C4F8, is only a few tens of nanometer thick and is quickly sputtered away from faces perpendicular to the biased ﬁeld when the etch cycle begins. Because of this alternating etch scheme, a scalloping pattern is generated on the sidewalls. Figure 9 shows a schematic drawing of the process. DRIE through e.g. the Bosch process is designed to be truly anisotropic; i.e. etching only in one direction, thus yielding vertical sidewalls. In some cases however, non-vertical sidewalls may be the desired choice for a certain structure. Two relevant examples where tapered sidewalls may be required are processes involving electrical contacting of 3D-structures (e.g. through-wafer vias) or mold insert fabrication; the former to allow conformal metal coating and the latter to obtain a certain release angle needed for proper demolding. There are a few techniques reported in the literature on achieving tapered sidewalls with deep plasma-based etching [115–120]. However, these are all based on very specialized processes (e.g. special mask designs [115–117], special etch gas composi- tions [118, 119] or special lithography [120]). In paper 2 of this thesis, a method is described to achieve controlled tapering of the sidewalls using DRIE. The advantage of this method is that it is based on the well-characterized Bosch process and can be implemented in standard DRIE systems. The basic principle of the method is to continuously switch between dry isotropic etching and anisotropic etching, thus real- izing a stair-like sidewall which can be made relatively smooth if the switching occurs frequently. The method was originally developed to increase yield of a microneedle fabrication process.

4.3 Solid microneedle arrays One of the ﬁrst microneedle arrays for drug delivery, although not transdermal, was presented 1993 by Dizon et al. [121]. The array, featuring pyramidal-shaped silicon spikes at densities of thousands per square centimeter (ﬁgure 10a–b), represents one of the most basic designs of microneedles. The needles are etched in potassium hy- 4 MICRONEEDLES FOR DRUG DELIVERY APPLICATIONS 29

(a) (b)

Figure 11. (a) The ﬁrst out-of-plane microneedles for transdermal drug delivery applications. The

solid silicon needles are 150 ñm long and made of silicon. From [106] (b) Electroplated nickel-iron microneedles. From [128]. droxide (KOH) solution and the geometry is deﬁned by controlled undercutting of the etch mask in combination with the anisotropic etch rates in monocrystalline silicon. Through the controlled etch (with intersecting crystal planes), the needles have an extremely sharp apex with a tip radius below 100 nm [124]. The array was used to transfect cells by coating the needles with foreign DNA before pressing the array onto cell cultures. Successful delivery of DNA into tobacco leaf cells [124] and animal nematode cells were demonstrated [122]. The overall goal of this research was targeted delivery of anti-restenosis drugs into coronary arteries. By incorporating the solid spikes onto stents, the spikes could penetrate compressed arterial plague as well as the elastic lamina, and thus enable a viable path for local drug therapy [125]. In relation to this work, the group also demonstrated barbed spikes designed to secure the needle in the elastic tissue [121, 123]. Similar to Gerstel and Place’s patent, these barbed spikes were also suggested as hollow to facilitate convective drug delivery (ﬁgure 10c) [123]. Lately, the concept of KOH-etched microneedles has been investigated further by other research groups [126, 127]. As mentioned earlier, the real interest of microneedles for transdermal drug delivery applications began in 1998. Henry et al. demonstrated four orders of magnitude increase in permeability for calcein and BSA (bovine serum albumin) through human

epidermis in vitro after penetration with a microneedle array [106]. The 150 ñmlong microneedles were fabricated in silicon using DRIE and featured sharp tips with a

tip radius below 1 ñm (ﬁgure 11a). The authors report that approximately 10 N was used to penetrate the epidermis (placed on a layer of dermis) with an array of 20

by 20 microneedles, pitched 150 ñm apart. In a separate study by Kaushik et al.,

following a blind-trial on twelve human individuals, the insertion of these 150 ñm long needles into skin was rated as painless [42]. The authors also demonstrated solid metal (NiFe) microneedles and showed feasibility of a less costly microneedle material (figure 11b) [128, 129]. In 2004, Chabri et al. made a similar study as the one made by Henry et al. but for delivery of genes [130]. Using solid silicon microneedles like the ones described by Griss et al. [96], Chabri et al. confirmed the several-orders-of- magnitude increase also for large gene vectors. In 2001, the first delivery results using microneedle arrays in vivo were pub- 30 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

(a) (b) (c)

Figure 12. Macroﬂux’ microneedle array. (a) Array of 330 ñm long microneedles made from a titanium foil. Scale bar: 1 mm. From [53]. (b) Desmopressin coated tip of a microneedle. Scale bar:

50 ñm. From [131]. (c) Illustration of the array mounted on an adhesive backing. The inset shows how the patch is applied to the skin with an impact applicator [132]. From [131, 133]. lished. A research group at Alza corp. reported successful intradermal delivery of oligodeoxynucleotide (20 bases) into hairless guinea pigs using a microneedle array made of stainless steel [134]. The array, called Macroﬂux (now a separate spin-out

from Alza), was fabricated by etching the needle contour through a 30 ñm stainless steel foil followed by a 90◦ raise, to realize out-of-plane projections (ﬁgure 12) [135]. 2

Usinga2cm array featuring 480 microneedles, 430 ñm long, up to 14 mg/day of the substance could be delivered. It was found that a conﬁguration with the needles inserted into the skin, with the drug in close contact to the array, gave a better drug uptake than if the drug was placed on bare microneedle-perforated skin. The microneedle array was manually inserted to the skin using ﬁnger force. In a later study, authors from the same group presented results on intradermal delivery of a vaccine-like substance (ovalbumin) in vivo [53]. Using precoated mi-

croneedles (now made of titanium and 330 ñm long, figure 12a), it was shown that ñ intradermal delivery of 1–80 ñg elicited up to 100-fold higher (for a 1 gdose)immune responses than that of conventional intramuscular doses. In these experiments, 1–2 cm2 large microneedle arrays were applied to the skin by an impact applicator (figure 12c) [132] and removed after 5 s. In addition to the above-mentioned work, the group has also presented studies on delivery of desmopressin (a synthetic hormone) [131] and more recently an extensive study on the influence of delivery parameters such as depth of delivery, delivered dose and needle density, on the overall immune response [54]. Most notable, it was

found that the shortest, most tolerable, microneedles (225 ñm) in a high-density configuration (725 needles/cm2) rendered a similar immune response as that of longer 2 needles (600 ñm) at lower densities (140 needles/cm ). Similar to Macroflux’ study on delivery parameters for coated microneedles, Gill and Prausnitz recently presented studies on different coating methods for microneedles [136, 137]. The authors tested 20 different coating formulations using various drugs and microneedle geometries. Similar to Macroflux’ needles, the needles were fabricated in-plane by laser cutting of a stainless steel sheet and manually raised out of the plane. In 2002, a research group at Becton, Dickinson and Co. (the leading company of 4 MICRONEEDLES FOR DRUG DELIVERY APPLICATIONS 31

(a) (b)

Figure 13. (a) BD’s 50–200 ñm long microenhancer array by Mikszta et al. The array, etched in silicon, was used to scrape the skin to achieve delivery of DNA vaccine. From [55]. (b) 3M’s

Microstructured Transdermal System (MTS). 250 ñm long molded polymer needles. Scale bar:

100 ñm. From [138]. conventional needles) reported successful delivery of DNA vaccine in vivo by scraping the skin with a precoated “microenhancer array” [55]. This array featured blunt, 50–

200 ñm long, microneedles and was fabricated by wet etching of silicon (ﬁgure 13a). The authors report that immune responses following administration with the array were both higher and less variable than intramuscular and intradermal delivery using conventional methods. Even more notable, 100% seroconversion (i.e. immunity) was achieved after only two immunizations, whereas conventional techniques elicited 40–50% seroconversion and required more immunizations for full seroconversion. In a separate test on human subjects, the subjects scored the pain of the scraping procedure as being mild but more painful than scraping using sandpaper. In a later study on anthrax vaccine delivery, the microenhancer array was replaced by a plastic replica [139]. In this study the immune response was reported to be similar to that of intramuscular delivery and lower than that of intradermal delivery using a single hollow microneedle. Another multinational company working with microneedle arrays is 3M. The company’s Microstructured Transdermal System (MTS) consists of an array featuring

250 ñm long pyramidal-shaped polymer microneedles (ﬁgure 13b) and is being tested for vaccine delivery [138]. The company holds more than 20 unique microneedle- related patents including designs, fabrication methods, and various insertion tools. Commercial activity from a number of other companies can be noted as well. For example, Procter & Gamble has numerous patents concerning microneedle designs and fabrication methods (cf. e.g. [140] and references therein). In 2005, some of these patents were acquired by the drug delivery technology company Corium International Inc. [141]. Another example is LifeScan Inc., a diagnostics company within the John- son & Johnson group, which also holds several microneedle patents (cf. e.g. [142]). Among other smaller companies working with solid microneedles are Apogee Tech- nologies Inc. and Valeritas LLC (former BioValve). In 2002, Park et al. introduced biodegradable polymer microneedles [143]. These needles, made by vacuum casting of PLGA or PGA (polyglycolic acid) in a silicone mold, were demonstrated to possess mechanical robustness and capable of penetrating human epidermis in vitro (ﬁgure 14a) [144, 145]. Needle master structures were 32 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

(a) (b)

Figure 14. Biodegradable microneedles by Park et al. (a) Beveled PLGA microneedles inserted in vitro through human epidermal tissue . The needles are approximately 400 ñm long. From [144]. (b) Drug incorporated microneedle. The picture shows a cut oﬀ tip of a PLGA microneedle incorporating PLA microparticles encapsulating calcein. From [147]. fabricated in photoresist SU-8 through lithography. Beveled tips were achieved by ion etching [145] or by employing a special photomask featuring lenses that rendered

tapered needles with a tip-sharpness down to 5 ñm [146]. The fabricated needles ñ had lengths between 400 ñm and 1500 m. Following in vivo insertion force tests described by Davis et al. [85] on hydrated skin, Park et al. reports measured microneedle

insertion force of 37 mN with a tip 20 ñm in diameter [146]. In most cases, needle fracture forces were found to be at least two times higher than insertion forces. PGA, the strongest material tested, showed the largest margins [146]. Park et al. also introduced the concept of biodegradable microneedles incorporating drugs [144]. Model drugs (BSA or calcein) were incorporated in two ways; either directly in the polymer matrix, or by ﬁrst encapsulating the drug into PLA (polylactic acid) microspheres to obtain an even slower release of the drug in the tissue (ﬁgure 14b) [147]. Slow release, from hours to months, of drugs from gradually dissolving microneedles were demonstrated in skin in vitro. Following tests on mechanical strength, the authors concluded that up to 10% drug inclusion is possible until the strength of the needle becomes critically low to still enable skin insertion. Based on that, it was estimated that an array of 1000 needles potentially could deliver a drug mass of 1 mg [147]. To maintain a low thermal budget during fabrication of the needles incorporating temperature-sensitive drugs, Park et al. also developed a particle-based molding technique for microneedles [148]. Miyano et al. demonstrated a similar approach. Instead of polymer needles, the authors fabricated 0.2–2 mm long microneedles in sugar (maltose) with up to 10% inclusion of a model drug [149]. The advantage of sugar microneedles, the authors claim, is that any remaining needle waste can be easily disposed by dissolving the needles in hot water [150]. To improve manufacturability, the authors have also fabricated the needles in polyethylene glycol (PEG) [150]. Recently, Kolli and Banga made a thorough study on maltose-made microneedles, including in vivo tests [151]. Although these needles did not include a drug, the

tetrahedrally-shaped, 500 ñm needles (ﬁgure 15a) are both robust enough and sharp enough to be inserted by hand. In 2004, Martanto et al. presented the ﬁrst systematic study on insulin delivery 4 MICRONEEDLES FOR DRUG DELIVERY APPLICATIONS 33

(a) (b)

Figure 15. (a) Sharp, 500 ñm microneedles made of maltose by Kolli and Banga. From [151].

(b) 150 ñm replicated microneedles in COC (cyclic olefin copolymer) by Trautmann et al. From [153]. in vivo using a microneedle array [152]. Similar to Macroflux’ microneedles, arrays of 105, 1 mm long, needles were fabricated by laser cutting of a stainless steel foil and manual bending. After microneedle insertion using a high-velocity applicator, insulin was administered topically onto the insertion site. Significantly reduced blood glucose levels were reported which were similar to reductions achieved by subcutaneous delivery of 0.05–0.5 units of insulin. In contrast to Macroflux’ results, the authors measured higher drug concentrations in tests where the microneedle array was removed directly after the insertion as compared to tests where the array was left in the tissue. In 2005, Trautmann et al. demonstrated polymer replicated microneedles using vacuum casting [153]. Although this had already been shown by Park et al. (see above), the significance of this work was that the authors presented drug delivery

using a polymer microneedle array on human skin in vivo. The 150 ñmlongmi- croneedles (ﬁgure 15b) were replicated from a sharp silicon microneedle master and used as substitute to pain-causing lancets in allergy prick tests. Recently, Han et al. demonstrated another example of sharp replicated polymer microneedles. The needles, mounted on a roller and aimed for cosmetic applications, were fabricated in polycarbonate by hot embossing [154]. A master needle array was achieved by assem- bling rows of sharp in-plane-fabricated microneedles into a jig.

4.4 Hollow microneedle arrays In contrast to solid microneedles, hollow needles offer the possibility of active injection of the drug into the tissue. The apparent advantage of this is that a considerably larger amount of drug can be delivered for a given time, thus opening for applications where relatively large amounts are needed to obtain a therapeutic effect. Additionally, pressure-driven delivery adds the possibility to precisely steer the flow rate and to obtain a more controlled delivery. The first hollow out-of-plane microneedles was presented by McAllister et al. in 1999 [155]. By combining the fabrication process of solid silicon microneedles shown

in figure 11a with the Bosch process to form a needle bore, 150 ñm long hollow microneedles and microtubes could be fabricated (figure 16a). The paper also describes the fabrication of hollow metal microneedles (figure 16b) which were produced by a 34 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

(a) (b)

Figure 16. (a) The ﬁrst hollow out-of-plane microneedles by McAllister et al. The silicon needles ñ are 150 ñm long and fabricated through DRIE using the Bosch process. (b) 150 m long hollow electroplated nickel-iron microneedles. From [155].

(a) (b)

Figure 17. Hollow silicon microneedles by Stoeber and Liepmann. The 200 ñm needles are fabricated by isotropic dry etching and DRIE (lumen). (a) Pointed tip. (b) Flat tip. From [157]. lost-mold technique, where the needles were electroplated. Solid silicon needles were used as the mold insert to facilitate the bore. These nickel-iron needles had a bore-

opening of 10 ñmindiameterandwereshowntopenetrate epidermal tissue in vitro. A similar approach of fabricating hollow metal microneedles was later demonstrated by Kim et al. who electroplated needles on sacriﬁcial solid microneedle arrays made of SU-8 [156]. In 2000, Stoeber and Liepmann presented another type of hollow silicon micronee-

dles [157, 158]. The fabrication of these 200 ñm long needles starts by etching the needle bores from the backside of the silicon wafer using DRIE. The needle structures are then etched from the frontside of the wafer by isotropic dry etching (ﬁgure 17). In a later study, the needles were tested by delivery of methyl nicotinate (a vasodi- lating agent) into human subjects [43]. Microneedle chips featuring a few needles were mounted on a standard 1 ml syringe and pressed against the subject’s volar

forearm while injecting. It was estimated that approximately 1 ñl (0.1 M concentration) was injected during the 30 s administration period. Pharmacodynamical results showed a significant increase in blood flux after delivery through pointed microneedles (figure 17a) while the increase for flat-tipped microneedles (figure 17b) was not statistically significant. However, the onset was significantly faster for both needle types as compared to the topical-delivery control. All the eleven subjects reported that the ad- 4 MICRONEEDLES FOR DRUG DELIVERY APPLICATIONS 35

(a) (b)

Figure 18. 210 ñm, cross-shaped, hollow, side-opened, silicon microneedles by Griss and Stemme.

(a) Needles with a 50 ñm long base shaft. (b) Needles without a base shaft. From [88].

(a) (b)

Figure 19. (a) 350 ñm long silicon microneedle by Gardeniers et al., etched by combining DRIE and wet etching. From [160]. (b) Microneedles made in PMMA by Moon and Lee, using LIGA- techniques. From [161]. ministration was painless, conﬁrming the observations by Kaushik et al. [42]. In 2001, Abbott Laboratories (a Fortune 100 health care company) ﬁled a patent [159] almost exactly describing the same microneedle process as that of Stoeber and Liepmann. To avoid clogging of microneedles bores during insertion into skin, Griss and Stemme developed and demonstrated a concept of side-opened microneedles [88]. By combining anisotropic DRIE and isotropic dry etching of silicon, a three-dimensional needle structure was formed that intersects with the needle bore at the shaft of the

needle (figure 18). The 210 ñm needles were demonstrated to have a relatively low fluidic resistance and were capable of penetrating aluminum foil without being damaged. Side-opened hollow microneedles for drug delivery applications are being commercialized by Debiotech SA, Switzerland. In 2003, Gardeniers et al. presented tetrahedrally-shaped hollow microneedles etched in silicon by combining DRIE and wet etching by KOH (figure 19a) [160]. The advantage of this approach is that the needles become sharp as a result of intersecting crystal planes while the tetrahedral shape ensures mechanical robustness.

Following skin insertion of 350 ñm needles, a 2.3-fold increase in transepidermal water was measured, showing that the stratum corneum had been disrupted. The authors also report on insulin delivery to a small number of diabetic rats using a microneedle array connected to an insulin pump delivering 1 insulin-unit per hour. After 6 h de- 36 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

(a) (b)

Figure 20. (a) 500 ñm hollow nickel microneedles by Davis et al. From [168]. (b) Cross-section of one of the needles inserted into cadaver skin. This remarkable picture from Davis’ thesis, illustrates

well the huge strains and tissue compaction associated with needle insertion. It also shows that the ñ actual penetration depth of the 500 ñm needle is about 100 m. From [87]. livery, blood glucose levels had decreased by 80%, which was similar to subcutaneous delivery. In a separate study, the same group also presented a concept of microneedles featuring electrodes to monitor proper skin insertion [162]. In addition, the method of combining wet etching and DRIE to facilitate beneﬁcial microneedle-tip geometries was recently used to fabricate molds for hollow microneedles made of photoresist SU-8 [163]. Hollow silicon microneedles fabricated by combining DRIE and KOH- etching is being commercialized by NanoPass Ltd., Israel. The company has entered collaboration with GlaxoSmithKline on microneedle-based vaccine delivery and joint patents [164]. Apart from the needles, the company has patented tools to aid microneedle insertion into skin—for example, in order to achieve a certain insertion angle [165,166]. Moon and Lee presented a similar microneedle design to that of Gardeniers et al. (ﬁgure 19b) [161]. Instead of silicon, the needles were fabricated in polymer PMMA (polymethyl methacrylate) using LIGA-techniques [167] with inclined x-ray exposure.

900 ñm needles were demonstrated to cause bleeding after insertion to the back of a human hand while insertion into a finger tip did not cause any bleeding. Following the first study of hollow metal microneedles by McAllister et al. [155] (figure 16b), Davis et al. developed the method further [168]. Instead of using a mold insert, the authors fabricated sacrificial molds through laser ablation of Mylar

sheets, yielding 500 ñm long tapered through-holes of a Gaussian shape. Hollow microneedles with a certain wall thickness were then electroplated from the mold, which was then sacrificed (figure 20). The microneedles were then thoroughly tested in various aspects. As mentioned in sec. 3.3.2, the insertion mechanics in relation to the needle strength was investigated [85]. The authors found the margin of safety to be much larger than 1, thus showing the needles to be sufficiently strong to be inserted into skin. The insertion force in vivo on human subjects was measured to 0.1–3 N per needle depending on the tip area. Although an accurate measurement method was used, the subjects’ skin had been immersed into warm water prior to the experiment, something that according to literature alters the mechanical properties of skin considerably (cf. e.g. [67–69]). The microneedles were also tested for insulin delivery to diabetic rats [168]. Although insulin had been delivered with microneedles earlier, 4 MICRONEEDLES FOR DRUG DELIVERY APPLICATIONS 37

(a) (b)

Figure 21. (a) 400 ñm long, ultra-sharp, side-opened microneedles by Roxhed et al. (b) Magnified view of the microneedle tip. The tip-radius is below 100 nm. From paper 3 of this thesis. the significance of this study was that the actual delivered amount was measured, i.e. the pharmacokinetic effects. The authors used passive, diffusion-driven, delivery by placing a insulin-filled chamber on top of the inserted needle array. The array (with 16 needles) was inserted into the skin with high-velocity plunger. Results showed a plasma concentration of 0.4 ng/ml after a 4 h delivery period and concurrent reduction in blood glucose. The authors also noted a discrepancy in the obtained pharmacodynamic effect which led them to suggest that the pharmacodynamic response could be stronger at the delivery site (being near the capillary loops) than in the subcutaneous space. In 2005, Teo et al. [169] presented an experimental study including both in vitro

and in vivo tests using 150 ñm cylindrically-shaped microneedles similar to the microtubes presented earlier by McAllister (see above). The authors reported a 10–20 times increase in permeability through needle-penetrated full-thickness skin in vitro. However, in vivo trials on insulin delivery to diabetic rats showed no difference in measured blood glucose compared to the controls. It was argued that to achieve delivery in vivo, the needles had to be both sharper and longer. The side-opened microneedles presented by Griss and Stemme (figure 18) was developed further by Roxhed et al. Paper 3 of this thesis presents ultra-sharp side- opened microneedles (first presented in 2005 [170]), designed to reliably penetrate skin tissue without the use of any special insertion procedure (e.g. tooling or pre- treatments). In relation to the previous design, the aspect ratio was increased by

making the silicon needles considerably longer to 400 ñm (figure 21). The fact that the side opening offers the possibility of a sharp and well-defined tip was emphasized by changing the geometry from cross-shaped to circular. The governing idea was that a pointed tip better penetrates tissue than cross-shaped blades, or, as described by Shergold and Fleck (cf. sec. 3.3.2), to achieve skin fracture through a mode-I crack (tensile) rather than a mode-II crack (shear). It was demonstrated that the new design had superior penetration characteristics, compared to a cross-shaped needle, and the insertion force was estimated to be below 10 mN per needle. Successful liquid injection into human skin in vivo on several test sites was also reported. In a separate study by Nordquist et al. (paper 5 of this thesis), the needles were used to actively infuse insulin into diabetic rats. The significance of this study was that it reported on 38 A Fully Integrated Microneedle-based Transdermal Drug Delivery System pharmacokinetic effects following active, convective, drug delivery. This administration mode was compared to passive (e.g. diffusion-driven) delivery using microneedles, topical delivery of microneedle-perforated skin, conventional subcutaneous delivery and intravenous delivery. It was found that active intradermal insulin delivery using the microneedles yielded similar insulin concentrations as standard subcutaneous delivery of the same infusion rate. In all the reported tests with this microneedle array, insertion was made manually using finger force and without pretreatment of the insertion site.

4.5 Concluding remarks After nearly a decade of microneedle research, a few trends can be noted. First, as the ﬁeld has become more mature, more relevant and more adequate experimental evaluations are being performed, including in vivo trials. Second, the lengths of the needles are longer. Third, the material choice is more diverse and polymer needles are gaining more ground. To achieve skin penetration, some groups make use of an impact applicator or a special insertion device. While hollow microneedles have become more frequent in recent years, several groups actively work and develop delivery techniques for solid microneedles. 5 MICRONEEDLE-BASED SYSTEMS 39

5 Microneedle-based Systems

In this chapter the concept of microneedle-based drug delivery is further elaborated. Methods on transporting drugs in combination with microneedles are discussed and eﬀorts on system level are reviewed.

5.1 Vision A microneedle-based drug delivery system may feature all the favorable properties that made the classical transdermal patch a success. Like the ordinary patch, the system would be easily attached to, for instance, the upper arm and worn for a shorter time while medicating. The advantages of such a system are:

– Pain-free administration – Easy to use—OTC-compliant

– Discreetness

– Continuous release – Controlled release

– Safer handling

Pain-free administration: Microneedles with a length of a few hundred micrometers, only penetrates the superﬁcial layers of the skin where the density of nerve receptors is low. As a consequence, insertion of microneedles into skin is perceived as painless. Easy to use: Like an ordinary transdermal patch, an envisioned system can be applied by the patient himself virtually without any training. However, to achieve this, special insertion tools and procedures are highly unwanted. Hence, the insertion force of the microneedles needs to be low and the insertion procedure needs to be reliable and robust. If this is achieved, it is reasonable to believe that the system, for certain medication, can be sold over the counter (OTC). Discreetness: Incorporating a microneedle-array with a planar and compact dosing system yields a patch-like, unobtrusive device that can be discreetly worn under clothing. Continuous release: An unobtrusive device may be worn for longer times, thus enabling continuous and sustained delivery at therapeutic levels. 40 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

10–30 mm

Delivery mechanism

Drug reservoir 1–3 mm

Skin tissue

Figure 22. Conceptual drawing of a microneedle-based drug delivery system.

Controlled release: Drug release through a separate mechanism allows the release- rate to be precisely controlled. This may be accomplished through integration of passive elements, e.g. flow restrictors or membranes, or active devices. Active dosing systems offer the possibility to modulate the delivery in time and in amplitude. Even more advanced, active elements permit the use of closed-loop systems. Safer handling: Microneedles protruding a few hundred micrometers from a surface pose a far less risk of accidental needle sticks than hypodermic needles do. Since microneedles do not reach into the blood, the risk of transmission of blood-borne pathogens is also further reduced. Figure 22 shows a conceptual drawing of a microneedle-based system including a drug reservoir and a delivery mechanism. In principle, this is almost identical to the system proposed in Gerstel and Place’s patent from 1976 (figure 8). However, Gerstel and Place’s patch does not include a dispenser mechanism. Constrained to a patch-like system of small form factor, MEMS technology offers a range of solutions by which a dispenser mechanism may be realized. In the next section, delivery methods suitable for microneedle-based systems will be discussed.

5.2 Dosing systems Depending on whether the microneedles are hollow or not, different modes of delivery are possible. Non-hollow needles are usually limited to passive, diffusion-driven methods. Additional techniques may enhance transport, e.g. through iontophoresis as demonstrated in conjunction with microneedles by Choi et al. [171]. Hollow microneedles can used to actively inject a drug into the tissue. Active, in this context, means that the drug is forced into the tissue by a certain external pressure at a certain flow rate.

5.2.1 Passive delivery The simplest method of delivery from drug reservoir into the tissue is by pure diffusion. In an idealized case, there is an unbroken liquid path between the reservoir and the interstitial fluid in the tissue. In such a case, diffusion occurs in a continuous (aqueous) media and the concentration in the tissue is merely a balance between diffusivity and physiological elimination. As the interfacial area of microneedle bore-openings or conduits in the skin are very small, and the diffusivity in the dense tissue matrix 5 MICRONEEDLE-BASED SYSTEMS 41

One-Shot Pump 2 μmthick p p Pressurized gas Thermoplastic membrane

Heat Heat Heat

Localized microheater detonator detonator Air-bursting Liquid

biochip Gas bursting after trigger μ Disposable Liquid sample on biochip 0.1 mthick (a) (b) (c)

Figure 23. (a) Schematic drawing of a air-burst detonator by Hong et al. From [184]. (b) Schematic drawing of a one-shot pump using expandable microspheres by Griss et al. From [185]. (c) Expandable microspheres, analogy with a spring system. From paper 1. is low, diffusion is very slow. Thus, for drugs that are eliminated in the body, the resulting concentration is very low. However, for delivery of very potent drugs (e.g. vaccines), low amounts of the drugs (e.g. micrograms) may be sufficient to obtain a therapeutic effect. Passive, diffusion-driven delivery of insulin using hollow microneedles has been studied and quantified by Davis et al. [168] and Roxhed et al. (papers 4 and 5).

5.2.2 Active delivery A straightforward way to accomplish delivery is to let the user of the system apply a force that renders flow through the∝ microneedles. In the same way as a normal syringe, this can be realized by pressing a piston or a flexible membrane into a drug- filled cavity with an outlet hole. However, to obtain controlled delivery, the actuation force needs to be separated from the user to an independent dispensing mechanism. Although there are a large variety of methods for propelling liquids, only few are suited for integration with a patch-like, disposable, device. For such a system, properties like low cost, compactness, robustness and low power consumption are essential. Compact micropumps (reviewed in [172]) offer good performance and controlla- bility but are likely too expensive to be used in disposables for low-cost applications. Less complex mechanisms that propel liquids can be realized by resembling the piston- into-cavity function of a standard syringe. Various small-scale expansion actuators may be used in such a configuration. A simple method is to use thermopneumatic gas expansion to propel the liquid [173–176]. However, as expansion is proportional to the temperature rise of the gas, either a large gas volume (less compact) or a high temperature (high energy consumption) is needed to propel the liquid. Another method is to use phase-change actuators like paraffin actuators [177,178] which are strong but difficult to make compact; or gas generating actuators such as by heating azobisisobu- tyronitrile [179] or electrolysis of water [180,181]. Thermally responsive hydrogels may be used as actuator for liquids and offer expansion above 100% [182, 183]. However, hydrogels expand due to a temperature decrease, and thus requires the integration of a costly cooling element. A way to reduce the amount of external energy needed to dispense liquid is to store energy in the device already at the fabrication stage. An example of this are air- 42 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

(a) (b)

Figure 24. (a) Microneedle array bonded to a flexible silicone reservoir by Stoeber and Liepmann. (b) Schematic drawing of the operation principle. From [190]. bursting detonators presented by Hong et al. who stored pressurized gas in membrane- sealed cavities on-chip for liquid propulsion [184, 186]. By fusing the membrane, the gas was released and acted as driving force for a liquid sample (figure 23a). A similar approach for moving liquid in microfluidic channels was demonstrated by Griss et al. who used thermally expandable microspheres for propulsion (figure 23b) [185]. The spheres contain highly pressurized hydrocarbons encapsulated in a thermoplastic shell. When heated, the shells fuse and the condensed hydrocarbons inside the spheres expand, approaching atmospheric pressure. During the process, the spheres undergo a dramatic volume change of up to 60 times their original volume (figure 23c). When heating ceases, the spheres remain expanded. Although a lot of energy is pre-stored in the spheres, the shells need to be heated to 70 ◦Ctosoften which, in turn, may require a considerable amount of energy. Roxhed et al. (paper 1 in this thesis) usedthesamespheres,mixedinglycerine, to realize a low-cost liquid dispenser. A flexible membrane separated the expandable glycerine paste from the delivery liquid and a volume expansion of 200% was shown

to dispense 101 ñl of liquid. Samel et al. developed a thermally responsive solid material by dispersing the expandable microspheres into a PDMS (polydimethylsiloxane) matrix [187]. Single- use microﬂuidic pumps and valves were demonstrated and integrated on chip as well as fabricated at wafer level scale [188,189].

5.3 Integrated microneedle systems Although a relatively large amount of microneedles have been presented in the last decade, there are only very few studies reporting on microneedles integrated with a delivery system. As mentioned earlier, one of the key advantages of microneedles is that they allow integration into a patch-like device. Thus, beneﬁts like ease-of-use, discreetness and compactness can be achieved. A possible explanation to the limited work in this area is that microneedle-based systems essentially should be of low cost, thus the integration of complex on-chip pumping would raise the unit cost to unacceptable levels. Although cost is of outmost concern for microneedle technology (given that ordinary needle administration is cheap), existing disposable products on the market, e.g. Alza’s E-Trans (ﬁg- ure 4) incorporating electronics, shows that more complex devices can be acceptable. 5 MICRONEEDLE-BASED SYSTEMS 43

Microneedle array

Reservoir Liquid Unexpanded composite

PCB

Released liquid

Expanded composite

(a) (b)

Figure 25. (a) Microneedle array integrated with an electrically controllable dispenser by Roxhed et al. (b) Schematic drawing of the operation principle. From paper 4.

Relevant for microneedle-based systems, Debiotech’s MEMS-based micropump [191] (based on the classical van-Lintel pump [192]) is marketed as being a disposable device [193] and could in principle be integrated with microneedles. However, as indicated earlier, for low-cost applications, it is reasonable to believe that the delivery mechanism should be simple and of very low cost. So far, three specific studies have been presented where microneedles are integrated with liquid dispensing mechanisms. Stoeber and Liepmann demonstrated a very basic concept where an array of hollow silicon microneedles (figure 17) was capped with a silicone reservoir (figure 24a) [190]. After filling the reservoir, liquid could be ejected through the needles by manually pressing on the flexible reservoir (figure 24b). Injection of dye into skinless chicken breast was demonstrated with the device. A more advanced approach was presented by Zahn et al. who integrated a displacement pump with an in-plane polysilicon microneedle (figure 7a) [194]. Although an in-plane needle, less relevant for general drug delivery purposes, was used, the study demonstrated successful monolithic integration and dispensing from an on-chip

micropump. Liquid pumping at 7.2 ñl/h was reported. Roxhed et al. integrated an array of ultra-sharp silicon microneedles (ﬁgure 21) with an electrically controllable dispenser (paper 4). The dispenser consisted of a printed circuit board (PCB), a reservoir cavity and the thermally responsive PDMS developed by Samel et al. [187] (ﬁgure 25). By controlling the power to heater elements on the PCB, the expansion of the thermally responsive material was controlled which,

in turn, determined the ﬂow rate from the dispenser. Flow rates from 0.1 ñl/h to

300 ñl/h was achieved (paper 4). The system was also successfully tested in vivo by ñ insulin delivery to diabetic rats at rates of 2 ñl/h and 4 l/h (papers 4 and 5). 44 A Fully Integrated Microneedle-based Transdermal Drug Delivery System 6 AN INTEGRATED MICRONEEDLE-BASED SYSTEM 45

6 Development of an Integrated Microneedle-based System

This chapter presents the development of a fully integrated microneedle-based drug delivery system introduced already in the previous chapter. First, a liquid dosing and actuation concept is discussed in more detail and the work that led to paper 1 is presented. Second, the fabrication of ultra-sharp microneedles that led to paper 2 and 3 is presented in more detail. Paper 2 reports on a method to achieve tapered sidewall in DRIE using the Bosch process and isotropic etching. Third, a concept of membrane-sealed hollow microneedles to achieve a closed-package system is introduced. This work eventually led to paper 6. Finally, the work and evaluation of the integrated microneedle system that led to papers 4 and 5 is presented. The chapter ends with a discussion on the system and its components.

6.1 Dosing and actuation unit The concept of expandable microspheres to propel liquid, introduced by Griss et al. (cf. sec. 5.2.2), was developed further and used to realize a dispenser device (paper 1). The main objective for this work was to realize a compact, precise and low-cost dosing unit suitable to be used in conjunction with microneedles similar to what is illustrated in ﬁgure 22.

6.1.1 Design

The dispenser was designed to deliver 100 ñl, as this was seen to be a relevant volume similar to those used in medical practice when administering drugs such as insulin or vaccines. To obtain dosing precision, a design was implemented where the expandable microspheres expand into a liquid-filled reservoir cavity of a predefined volume. Since the reservoir has an outlet hole, the stored volume will by ejected by the expansion. Figure 26 illustrates the operation principle. Since the expandable microspheres expand up to 60 times after heating, the dispenser can be made compact in the sense that the volume of the actuator is small in relation to the actuated volume. However, to enable handling and efficient heating of the microspheres, they need to be incorporated into a matrix which, in turn, reduces 46 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

Heater Liquid Supporting ring Container Membrane Expandable paste

70 ◦C

Figure 26. Schematic drawing of the dispensing mechanism.

the relative volume expansion. For actuating 100 ñl, glycerine was used as the matrix, yielding a paste-like mixture which in contrast to water does not evaporate when heated. The paste was separated from the delivery liquid by a thin elastic membrane.

6.1.2 Fabrication The dispenser was fabricated entirely using materials and processes with the potential of rapid scale-up and low-cost fabrication. A demonstrator device was fabricated

which consisted of a plastic liquid reservoir, a 200 ñm thick vinyl membrane, expandable paste, a plastic ring that defines the volume of the expandable paste and a printed circuit board featuring heater elements. The expandable paste was made by mixing Expancel microspheres (a commercially available blowing agent sold in cubicmeter volumes) and glycerine. During assembly of the device, the membrane is clamped into the reservoir by the ring and the paste is then screen-printed into the ring (figure 26). The device is sealed off by the PCB (being in contact with the paste)

and ﬁlled by reducing the pressure in the reservoir. 35 ñl of the expandable paste was used in the device.

6.2 Ultra-sharp hollow microneedles Following the work of side-opened hollow microneedles by Griss and Stemme [88], in vivo tests using cross-shaped microneedles (figure 27a), similar to those presented by Griss and Stemme, gave non-conclusive results pointing towards that reliable penetration through the skin tissue was a problem. In contrast, an earlier microneedle design by Griss et al. [96] (for EEG electrodes) had shown good penetration results in preclinical trials on human skin in vivo [97, 195]. A significant difference between the two designs was the aspect ratio and the tip geometry, which in the latter case was circular with a clear pointed tip. Working from the hypothesis that tip geometry has a crucial influence on penetration, a new microneedle was designed with the aim of allowing reliable skin penetration through insertion by hand.

6.2.1 Design A circular microneedle design, already suggested in the paper from Griss et al. [88], was adopted since it featured a well-deﬁned pointed tip, proven earlier to be successful. 6 AN INTEGRATED MICRONEEDLE-BASED SYSTEM 47

(a) (b)

Figure 27. Artistic drawings of side-opened hollow microneedles. (a) Cross-shaped microneedle. (b) New circular-shaped microneedle.

Apart from the changes in shape, the aspect ratio was increased and the microneedle was made longer. The disadvantages of increased aspect ratio are that the fabrication becomes more challenging and that the needles become more sensitive to fracture due to bending. Figure 27b shows an artistic drawing of the design. As with the previous microneedles, the structural material is monocrystalline silicon which offers superior mechanical properties and process precision, compared to other materials. The microneedle structure can be etched using anisotropic DRIE in combination with isotropic dry etching. A first estimate on process parameters can be derived from the design assuming ideal etch behavior (i.e. fully isotropic and anisotropic). For example, the height of the needle tip can be determined through geometrical relations between the isotropic etch lengths d1, d2 and d3. Figure 28a shows a drawing on the needle profile with the different etch lengths indicated. Fig- ure 28b shows the geometry of the needle lumen aligned with an offset to facilitate a side-opening on the needle. The microneedles were grouped into 5 × 5 square arrays with an inter-needle pitch

of 500 ñm. Each such array, with a 1 mm wide surrounding blank area (for handling purposes), forms a microneedle chip.

6.2.2 Fabrication

The microneedles were fabricated on 600 ñm thick, 100 mm wafers using only two photo masks: one for the microneedle bores, patterned on one side of the wafer; and one for the outer proﬁle of the needle, patterned on the other side of the wafer. The needles were etched in an ICP (Inductively Coupled Plasma) system where DRIE through the Bosch process and isotropic dry etching using SF6 was used. Silicon dioxide (SiO2) was used as the masking material. 48 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

d 3 d2 d1

Etch mask

α Shaft diameter Needle lumen

d 3 d2

β Upper diameter d d1 d2 3

Needle profile

(a) (b)

Figure 28. (a) Schematic drawing of the microneedle proﬁle with isotropic etch lengths di indicated. (b) Schematic drawing of the cross-section of the microneedle viewed from the top.

Microneedle lumen

The needle bores measure 40 ñminwidthandweredesignedtobeetchedtoadepth

of 530 ñm down in the substrate using DRIE. At that depth and aspect ratio, direct microscopy-based measurements to determine the etched depth are not possible. Also, since the etch rate vary with the etched depth (so-called aspect ratio dependent etching, ARDE [196]) but also between process runs (due to contamination of the etch chamber), extrapolation and calibration (from destructive measurements) to find the right etch time is not an accurate method. Instead, the depth was controlled through monitor structures on the wafer. By utilizing ARDE, larger structures will etch completely through the wafer and stop at the SiO2 masking layer on the other side of the wafer. As a result, larger structures form transparent holes in the wafer which are clearly visible when shining light at the wafer. By calibrating a series of monitor structures of different size to the desired depth of the needle bore, the depth of the bores was accurately controlled within ±2%. As the microneedle bore is designed to intersect with the outer profile of the needle, the bores need to be protected during etching of the outer needle structure. This was achieved by growing a layer of SiO2 in the bores.

Microneedle shape After the fabrication of the bores, the other side of the wafer was etched and the outer shape of the needle was structured. The etch consists of ﬁve parts: three isotropic etch steps alternated with two anisotropic etch steps. Isotropic etches determine the curvatures of the needle shape, and anisotropic steps regulate the height of the needle and the shaft. Figure 29 illustrates the diﬀerent etch steps. After the last etch step, the SiO2 mask resides as a cap on the needle structure, attached to the tip through a connection only a few micrometers wide. To remove the mask, an additional layer of SiO2 is grown on the wafer so that silicon faces intersect 6 AN INTEGRATED MICRONEEDLE-BASED SYSTEM 49

Step 1: isotropic etch Step 2: anisotropic etch Step 3: isotropic etch

Step 4: anisotropic etch

Step 5: isotropic etch

Figure 29. Illustration of the microneedle shape etch process showing the five different etch steps. At the last step, the outer profile intersects with the needle bore etched from the backside of the wafer prior to the steps illustrated in this figure. at the tip, yielding a needle apex of extreme sharpness. This process of sharpening tips, first demonstrated by Marcus et al. [197, 198], was followed by stripping excess SiO2, including the mask, in hydrofluoric acid (HF).

Yield and tapered etching From an etch-technical perspective, the microneedle shape etch is rather challenging since a major part of the substrate material is to be removed, leaving only sparsely spaced structures on the wafer. This leads to a very poor etch uniformity over the wafer which, in turn, decreased the yield of the fabrication process. Large exposed silicon areas in the center of etch chamber depletes the plasma and causes inhomogeneities in the plasma density, which leads to a lower etch rate in the center of the wafer than on the edge. To compensate for this, the size of the needle etch mask was radially increased over the wafer. In addition, to make the design less sensitive to non-uniform etch behavior, the needle bore openings were aligned facing the center of a needle chip and the needles located in the corners of an array were non-hollow. Another undesired fabrication effect was that needles tend to be negatively tapered (figure 30a). The effect is particularly apparent when etching free standing mesa structures, such as out-of-plane microneedles, and is due to a certain angular distribution of incident ions and backscattering. The effect can be reduced by decreased ion energy (less bias) to the expense of decreased etch rate but is difficult to completely avoid. However, this effect can also be compensated for. By incorporating short isotropic etch steps during anisotropic etching, a positively tapered structure is formed. This method was employed in the needle process and the inherited negative tapering was compensated by positive tapering through the method. Figure 30 shows pictures of the microneedles with and without using the method. 50 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

(a) (b)

Figure 30. (a) Microneedle with negative tapering. (b) Slightly positively tapered microneedles as a result of using the method to achieve tapered sidewalls presented in paper 2.

AnisotropicIsotropic Etch mask

Figure 31. Schematic drawing showing the basic principle of switching between anisotropic and isotropic etching to achieve tapered sidewalls.

In paper 2 of this thesis, the method is generalized and proposed as a potentially suitable method to achieve controlled tapering using DRIE. The governing principle is to alternate between isotropic dry etching and anisotropic etching through the Bosch process. By frequent relatively smooth sidewalls was achieved an sidewall angles up to 36◦(relative straight vertical) were demonstrated. Figure 31 illustrates the basic etch scheme.

6.3 Membrane-sealed microneedles The envisioned microneedle patch illustrated in ﬁgure 22 includes a valve mechanism that encloses the stored liquid from the environment. The purpose of this valve is to prevent the drug from leaving the device unintentionally (through device handling or evaporation) and to protect the drug from contamination through the microneedle openings. A method to achieve the same eﬀect as with a valve is to seal the microneedle bore- opening with temporary membranes that can be removed at the time of the delivery. 6 AN INTEGRATED MICRONEEDLE-BASED SYSTEM 51

Burst pressure Electrochemical Insertion force

F Physiological saline

Skin tissue

Figure 32. Schematic drawing of three possible opening methods for membrane-sealed hollow microneedles. From the left: Burst opening by applying pressure, electrochemical opening in the skin in presence of e.g. interstitial ﬂuid, opening by rupture due to forces involved during skin penetration.

The advantage of this approach over a separate valve is reduced dead volume by utilizing the needle bores, and that the membranes can add functional features useful for the delivery.

6.3.1 Design

Figure 32 illustrates three possible methods of opening the membranes before delivery is initiated: opening by bursting the membranes by applying pressure through the needle bore, stimulated corrosion by applying a small voltage, or opening by the forces involved during microneedle insertion into tissue (e.g. shear force). While the ﬁrst method will cause all membranes of a microneedle array to open (assuming the same pressure at all membranes), the two latter methods are conditioned by the local environment around the needle. Corrosion can only occur in the presence of an electrolyte and therefore the membrane needstobeintheskinandincontactwith e.g. interstitial ﬂuid to enable opening. Opening through rupture requires that the tissue exerts some force on the membrane. Hence, these two methods may be used to trigger delivery based on the condition that the needles are inserted into the tissue.

6.3.2 Fabrication

To achieve sealed microneedles, a metal layer was deposited onto the microneedles before the needles were fully stripped of the SiO2 layer used to protect the microneedle bore. Since the SiO2 layer, at that stage, covers the bore openings, it serves as temporary support for the deposited metal. After the deposition, the supporting oxide is removed by HF vapor which leaves microneedles where the bore opening is covered by a metal membrane. Gold was chosen as the membrane material because of its bioinert properties and likelihood of being biocompatible. The thickness of the membrane ranged from 170– 470 nm. 52 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

6.4 Integrated microneedle system

As mentioned in the previous chapter, an integrated patch-like microneedle system with active dosing capability has some clear advantages, both in allowing larger amounts of drugs to be infused, as well as oﬀering a simple and discreet solution. The goal of this work was to realize and test such a novel and fully integrated system.

6.4.1 Design

With a dosing unit preferably being of low cost and without the need of micromachining precision, and a more costly microneedle array, hybrid integration of these components is a rational approach to achieve a system. Microneedle arrays were integrated with a similar dispensing device, as mentioned earlier. Instead of using a glycerine paste as the expandable material, a solid expandable composite was used where expandable microspheres was mixed into PDMS. This approach, developed by Samel et al. (sec. 5.2.2), has the advantage that the expandable material can be included into the device as an unstructured layer, thus eliminating the need for special conﬁnement. As a solid hydrophobic material, the need for separating membrane to the delivery liquid is also avoided. The disadvantage is that the expansion capability is reduced since the microspheres are bound in an elastic layer. However, the expansion capability is still well over 100% [187] which was considered good enough to realize a compact device. Like for the glycerine-based dispenser, the solid expandable material expands into a liquid reservoir and ejects stored liquid (cf. ﬁgure 25b).

6.4.2 Fabrication

Dispensing devices were fabricated by bonding a silicon wafer featuring wet-etched cavities (liquid reservoirs)6 to a wafer-sized PCB covered by a spin-coated layer of the PDMS-based expandable composite. At the locations of the cavities, the PCB had predefined copper heater elements, fabricated using conventional PCB fabrication methods. After bonding, the wafer stack was diced to achieve single dispensing devices. Figure 33 illustrates the different structural layers. Integrated devices were assembled by gluing the microneedle array on the dispensing unit by adding a droplet of low viscosity cyanoacrylat to one of the corners of the microneedle chip. To prevent undesired filling of adhesive into the reservoir, a perimeter in form of an etched trench acted as a stop for capillary filling [199,200]. The fabricated devices measured 10 mm × 10 mm × 1.7 mm and were capable of

storing and dispensing 12 ñl of liquid. Figure 25a shows a photograph of one such device.

6Silicon as the liquid reservoir material was only chosen because it was simple to fabricate with facilities at hand. The liquid reservoir layer does not need the high precision of silicon micromachining and can preferably be fabricated using polymer replication techniques. 6 AN INTEGRATED MICRONEEDLE-BASED SYSTEM 53

Microneedle array

Drug reservoir Capillary stop trench

Expandable composite

Electrical heater (PCB)

Figure 33. Exploded view of the microfabricated drug delivery system.

6.5 Results 6.5.1 Dosing unit The fabricated dispensing device was tested and three main properties were investigated: the delivery precision, the temperature of the delivered liquid and the ability to dispense against a counter pressure. Following subsequent tests with an actuation power of 1 W during 150 s, a mean

dispensed volume of 101 ñl with a standard deviation of 3.2% was measured. The temperature of the delivery peaked at 59 ◦C. However, the peak temperature was reached before heating ceased which shows that the expanding material, as expansion proceeds, has an insulating effect towards the delivery liquid. Since the expanding microspheres incorporates a partly condensed hydrocarbon, the actuator is relatively strong. Measurements showed that the dispenser could deliver against a counter pressure of 75 kPa. However, at delivery against a counter pressure, the dispenser cannot fill the complete reservoir and the total dispensed volume is less than the reservoir volume. In addition to these tests, the dispenser was successfully tested in a microfluidic dye laser application and used to maintain a regenerating flow at a low flow rate. It

was demonstrated that by changing the actuation power, the ﬂow rate could be varied ñ between 1 ñl/h and 2400 l/h.

6.5.2 Ultra-sharp microneedles After making improvements in the design to account for nonuniformities in the etch process of the microneedles, the yield was increased to above 90% with 68% of the wafer area being processed. On a 100 mm wafer this resulted in more than 250 microneedle chips per wafer. Each chip featured a needle array that occupied 25% of the chip area and consisted of 25 needles, of which 21 were hollow. After using the sharpening procedure of sacriﬁcial oxidation, the tip radius of the needle was below 100 nm (ﬁgure 21b). 54 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

(a) (b)

Figure 34. (a) Microscope picture showing a sustainable pattern in living human skin after dye injection through the microneedle array. From paper 3. (b) Picture showing a rash formation on fatty abdomen tissue after microneedle administration of an allergen.

The microneedle array (figure 21 and 30) was tested an investigated on dry living humanskin.Itwasdemonstratedthatthearraycouldbeinsertedandusedto deliver liquid into the skin at clinically relevant test sites such as the upper arm and the shoulder. Insertion of the array was done by hand by pressing the needle array (mounted on a holder) onto the skin surface using moderate finger-force. Figure 34 shows pictures of human skin after insertion and injection through the microneedle array. The penetration characteristics of the microneedles were evaluated by monitoring the insertion progression into human skin in vivo by measuring the electrical impedance between a counter electrode and the microneedle array. Two different methods were used to insert the array: controlled propagation of the chip array using a large force, and insertion by controlled increase of the applied force. The circular- shaped microneedles were compared to cross-shaped needles (figure 27a) and a flat silicon chip without needles that served as control. Figure 35 shows the results of the measurements, and for both insertion methods tested, the circular-shaped microneedle array demonstrated completely different impedance values as compared to the cross-shaped needle. The cross-shaped needle, on the other hand, showed similar values as the non-penetrating control. From figure 35a, it can be noted that the resistance decrease ceases after the circular-shaped

microneedle array has been moved approximately 430 ñm into the skin. This distance

corresponds well to the length of the microneedles (400 ñm) and it is reasonable to believe that this represents full insertion of the array into the tissue. With that being the case, the resistance decrease preceeding this change can be seen as gradual insertion of the needles into the skin. Figure 35b shows the impedance and the phase as a function of applied force on the array during insertion. From the phase data, representing the level of resistive coupling over capacitive coupling, it can be noted that the phase reaches a relatively stable value at approximately 250 mN. This indicates that penetration through the skin occurs below this force. Distributed over the array, this force corresponds to 10 mN per needle. 6 AN INTEGRATED MICRONEEDLE-BASED SYSTEM 55

2 10 2 Circular 10 Circular Cross−shaped Cross-shaped Ω Dummy Dummy 1 10 Impedance / M 0

Ω 10

0 100 200 300 400 500 600 1 Force / mN 10

−20 Resistance / M

−30 ° −40

Phase / −50

−60 0 10 −70 0 100 200 300 400 500 600 700 800 900 0 100 200 300 400 500 600 Chip displacement / μm Force / mN (a) (b)

Figure 35. (a) Electrical resistance between diﬀerent microneedle chips and a counter electrode (placed on the skin) during succesive insertion into human skin in vivo. (b) Electrical impedance and phase at 100 Hz between diﬀerent microneedle chips and the counter electrode during insertion into human skin in vivo by gradually increasing the applied force used to insert the chip. From paper 3.

6.5.3 Membrane-sealed microneedles The fabricated membrane-sealed microneedles were tested based on their suitability to be opened at the time of delivery. The methods illustrated in figure 32 were evaluated by means of some basic tests. The ability to burst the membranes was tested on microneedle arrays with different membrane thicknesses. Results showed that 170 nm thick gold membranes burst when an air pressure of more than 120 kPa was applied to the back of the chip. Thicker membranes burst at higher pressures. Figure 36a shows the tip part of a cross-shaped microneedle where the membrane has been burst. In vitro test on opening the membranes electrochemically in a saline solution similar to interstitial fluid present in the skin, showed that a 170 nm membrane is

(a) (b) (c)

Figure 36. Tip part of gold-membrane sealed microneedles where the membrane (170 nm thick) has been opened by various methods. (a) Cross-shaped needle opened by bursting the membrane. (b) Cross-shaped where the membrane has been removed electrochemically in a saline solution. (c) Circular-shaped needle where the membrane has ruptured as a result of skin insertion in vivo. From paper 6. 56 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

250 180 160 * 200 IU/ml)

140 IU/ml) µ ∝ ∝ µ 120 * 150 100 * 80 * * * 100 60 * * * 40 50 20 Insulin concentration ( Insulin concentration ( 0 0 0309015021003090150210 time (min) time (min) (a) (b)

Figure 37. Insulin concentration in blood plasma of diabetic rats during administration (0–

180 min) with diﬀerent methods. (a) Results between diﬀerent microneedle-aided methods. Methods

ñ • are denoted as follows. : Microneedle-patch 4 ñl/h, : Microneedle-patch 2 l/h, : Microneedle- ◦ patch 0 ñl/h (i.e. passive), : Topical on microneedle-perforated skin. (b) Active microneedle-based infusion compared to standard methods, denoted as follows. : Subcutaneous 2 ñl/h, :Micro-

◦ ñ ∗ needle-patch 2 ñl/h, : Intravenous 2 l/h (diluted to 10 ). In both graphs, denotes a statistically signiﬁcant change (P<0.05) compared to the ﬁrst measurement in the same group. From paper 5. opened within 2 min. Figure 36b shows a needle opening after such a test was made. In the tests a voltage of approximately 1 V was used. Opening by insertion force was tested in vivo using circular-shaped microneedles sealed by 170 nm gold membranes. Out of ten tested microneedle arrays, only two single needles had their membrane intact after that the array had been inserted into skin. As with the other in vivo tests, the microneedle array was inserted by hand. Figure 36c shows a picture of a needle membrane after skin insertion.

6.5.4 Integrated microneedle system As for the microneedle array, injection of a dye marker in vivo intohumanskinwas also tested with the integrated microneedle patch. In contrast to the tests of the needle array, liquid dispensing was now made on-chip with the dispensing technology developed. As for the needle array alone, clear dye marks were visible in the skin showing that liquid is delivered into the tissue. However, at the infusion rate used in

these experiments (60 ñl/h), leakage of the dye to the skin surface could clearly be observed after the device had been removed. In a more comprehensive study, devices were tested by insulin administration to

diabetic rats. To ensure leak-free delivery, a low infusion rate of 2–4 ñl/h was used. At these rates, leakage could not be observed. Active infusion using the device was compared to passive delivery with the device (i.e. the patch was attached but not actuated), to topical delivery through microneedle-perforated skin, to subcutaneous delivery, and to intravenous delivery. Both pharmacodynamic and pharmacokinetic effects were measured throughout the 3 h administration period and for 1 h post administration. Figure 37 shows the blood plasma concentration of insulin following the different administration methods. Active delivery using the microneedle system caused significant increases in in-

sulin levels. As could be expected, infusion at 4 ñl/h caused higher insulin levels

than at 2 ñl/h. However, neither passive delivery with an unactuated microneedle 6 AN INTEGRATED MICRONEEDLE-BASED SYSTEM 57 patch nor topical delivery showed any significant increases of insulin levels during the measurement period. Compared to standard subcutaneous delivery using a hypodermic needle, active delivery with the microneedle patch at the same dose rate gave rise to similar insulin levels (figure 37b). All microneedle-aided methods caused a decrease in blood glucose (data not shown, cf. paper 5). Consistent with the increase in insulin levels, active delivery through the microneedle patch resulted in lower glucose levels than diffusion-based methods.

6.6 Discussion The results show feasibility of a minimally-invasive patch-like system for active administration of macromolecular drugs. Although the system in its current state can deliver a signiﬁcant drug volume, some issues remain and deserve more attention. The most generic and thus most relevant issue concerns the infusion rate. While

2 ñl/h may be a suﬃcient rate in a rodent model, higher rates are generally needed in human applications. Although infusion was made on a 2 × 2mm2 area and a larger microneedle array, say containing 400 needles on 1 cm2 area, would result in the more

reasonable rate 40 ñl/h, it is to the expense of a larger needle array which, in turn, increases the unit cost. Nevertheless, the low infusion rate possible highlights a problem with microneedle-based infusion in general. Namely, that the local circumstances at the interface between the microneedle and the extremely flexible tissue is complicated and poses a barrier for efficient infusion. It is known that tissue compaction occurs ahead of microneedles inserted into skin tissue and that this has a rate-limiting effect on infusion [89]. It has also been demonstrated that the rate can be increased by retracting the microneedle, thus relieving the stress; or by using hyaluronidase, an FDA-approved adjuvant and enzyme that breaks down collagen and increases tissue permeability [90, 91]. However, a precise, micrometer-sized retraction is difficult to achieve in practice and the use of an enzymatic adjuvant increases cost and complicates storage as well as handling. While tissue compaction ahead of the microneedle-tip is a factual challenge for tip- opened microneedles, infusion through side-opened needles should be less affected by this limitation. The simulations made in paper 3 on needle-tissue interaction further support this assumption. Nevertheless, the experimental results show that leakage occurs already at moderate to low infusion rates and it is reasonable to believe that the force exerted by the tissue on the needle is too weak to seal-off possible leakage paths to the skin surface. Another explanation would be that the needle is simply not completely inserted into the tissue, yielding leakage for a side-opened needle. However, the in vivo results clearly demonstrates differences between infusion-based (also at different rates) and topical administration methods, showing that active infusion is indeed the case. Assuming that the extremely flexible and viscoelastic nature of living skin prevents an effective seal to be formed around the needle, a few measures can be taken to limit this effect. A na¨ıve approach is to make a longer needle. Like for a hypodermic needle, a very long needle will penetrate deep enough to close any leakage paths to the surface. However, a longer needle may cause pain and is in general less preferred since 58 A Fully Integrated Microneedle-based Transdermal Drug Delivery System it complicates fabrication and weakens robustness. A more sophisticated approach to ensure leak-free delivery also at higher rates is to design the needle in a way that prevents leakage; for example, by providing some kind of element that seals the tissue around the bore-opening, e.g. a barb as proposed by others [201]. Another issue concerning the current system relates to safety. Although nanometer- sharp microneedles fabricated in silicon demonstrates superior penetration properties over softer and less formable materials, the outermost tip-region is brittle and may break in the skin. However, the mass of such tip-residues is only a few picograms and hence far from reaching toxic levels. Similar reasoning also applies for microneedle membranes. A nanometer thick membrane covering a bore-opening has a mass of a

few nanograms. This can be compared to natural concentrations of 70–90 ñg/g skin, found in skin below normal gold finger rings [202]. In relation to this, it should also be noted that skin is a constantly growing organ and parts of possible residues will naturally grow away. As an example, the turn-over time of the epidermis is 30–60 days [203]. The dispensing technology used in the microneedle system has benefits through its simple design and potential of low-cost fabrication. However, the dispenser uses 200–250 mW for actuation which accumulates to a considerable amount of energy for actuation periods of hours. It should be mentioned that no special concerns have been taken to optimize the actuator in terms of heat transfer and that such measures are likely to reduce the power consumption substantially. For example, as the PCB is a fairly good heat conductor, a lot of the generated heat will simply dissipate from the back of the device. Nonetheless, the expandable microspheres need to be heated to 60– 70 ◦C to expand. A configuration with smaller compartments actuated sequentially is therefore a more energy-efficient solution. Since the actuator is relatively strong and irreversible, a low flow-rate, continuous release can be achieved by including flow-restricting elements. Taken together, the fabricated microneedle-based system is far from optimal. How- ever, it demonstrates the basic behavior of active and controlled transdermal drug delivery of macromolecular drugs using a painless and unobtrusive patch-like device. As such, it is an example of a future administration form where drugs such as insulin or vaccines can be delivered in a convenient, discreet and patient-friendly manner. 7 SUMMARIES OF THE APPENDED PAPERS 59

7 Summaries of the Appended Papers

Paper 1: A Compact, Low-cost Microliter-range Liquid Dispenser based on Expand- able Microspheres This paper presents a new low-cost liquid dispenser for the dispensing of microliter to milliliter volumes. The dispensing mechanism is based on a thermal actuator where highly expandable microspheres expand into a liquid reservoir consequently displacing any stored liquid. All device components are made from low-cost materials and the fabrication process has the potential for high volume batch manufacturing. The device utilizes the property of the expandable microspheres to form a heat insulating layer between the heat source and the delivered liquid. Paper 2: A Method for Tapered Deep Reactive Ion Etching Using a Modiﬁed Bosch Process This paper presents a method for etching tapered sidewalls in silicon using deep reactive ion etching. The method is based on consecutive switching between anisotropic etching using the Bosch process and isotropic dry etching. By controlling the etch depths of the anisotropic and isotropic etch sessions, the sidewall angle can be controlled over a relatively large range. Tapered sidewalls are useful in microfabrication processes such as metal coating of 3D-structures (e.g. for electrical connections or vias), mold tool fabrication or as a tool to compensate for reentrant etching. The method was tested and characterized by etching basic test structures in silicon wafers. Paper 3: Penetration-enhanced Ultra-sharp Microneedles and Prediction on Skin In- teraction for Eﬃcient Transdermal Drug Delivery This paper presents penetration-enhanced hollow microneedles and an analysis on the biomechanical interaction between microneedles and skin tissue. The aim of this work was to fabricate microneedles that reliably penetrate skin tissue without using any penetration enhancers or special insertion tools used in previous studies. The microneedles are made in silicon and feature ultra-sharp tips and side-openings. The microneedle chips were experimentally tested in vivo by injection of dye markers. To further investigate penetration, the insertion progression and insertion force was monitored by measuring the electrical impedance between microneedles and a counter electrode on the skin. The microneedle design was also tested using a novel simulation approach and compared to other previously published microneedles designs. The purpose of this 60 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

specific study was to investigate interaction mechanisms between a microneedle and the skin tissue. This study is used to predict how the skin deforms upon insertion and how microneedles can be used to create a leak-free liquid delivery into the skin. Paper 4: Painless Drug Delivery through Microneedle-based Transdermal Patches featuring Active Infusion This paper presents the first microneedle-based transdermal patch with integrated active dispensing functionality. The electrically controlled system consists of a low-cost dosing and actuation unit capable of controlled release of liquid in the microliter range at low flow-rates, and of minimally invasive, side- opened, microneedles. The system was tested in vivo by insulin administration to diabetic rats. Active infusion of insulin was compared to passive, diffusion- driven delivery. Paper 5: Novel Microneedle Patches for Active Insulin Delivery are Efficient in Maintaining Glycaemic Control: An Initial Comparison with Subcutaneous Ad- ministration This study was designed to validate painless intradermal delivery via a patch- like microneedle array. Diabetes was induced by an intravenous injection of streptozotocin in adult male Sprague Dawley rats. Plasma insulin and blood glucose were measured before, during and after subcutaneous or intradermal microneedle-infusion of insulin (0.2 IU/h) under Inactin anaesthesia. Paper 6: Membrane-sealed Hollow Microneedles and Related Administration Schemes for Transdermal Drug Delivery This paper presents fabrication and testing of membrane-sealed hollow microneedles. This novel concept offers the possibility of a sealed microneedle- based transdermal drug delivery system in which the drug is stored and protected from the environment. Sealed microneedles were fabricated by covering the tip openings of out-of-plane silicon microneedles with thin gold membranes. In this way a leak-tight seal was established which hinders both contamination and evaporation. To allow drug release from the microneedles, three different methods of opening the seals were investigated: burst opening by means of pressure; opening by applying a small voltage in the presence of physiological saline; and opening as a result of microneedle insertion into the skin. 8 CONCLUSIONS 61

8 Conclusions

This thesis presents a fully integrated microneedle-based transdermal drug delivery system and development work towards the same.

The system: – demonstrates feasibility of a patch-like device featuring painless, hollow, side- opened microneedles and an electrically controllable liquid dispenser loaded with adrug. – can be appropriately applied to skin tissue by hand without any aids. – is capable of delivering insulin intradermally at different flow rates and to cause relevant pharmacodynamic response in a rodent model. – is capable of delivering liquid intradermally into human skin in vivo. The microneedles: – require a very low insertion force to be inserted into skin, potentially allowing arrays with hundreds of such microneedles to be inserted into skin by hand without any aids. – can penetrate human skin in vivo at clinically relevant sites. – can be hermetically sealed through thin membranes and be opened at the time of delivery by applying pressure or inserting the sealed needles into skin tissue. – can be fabricated with high process yield. The liquid dispenser: – can function as delivery mechanism in microneedle-based systems and be fabricated using low-cost materials and processes. – is capable of delivering liquid at different flow rates by controlling the power to the device. In addition, a method for etching tapered structures using DRIE was presented and characterized. The method is useful for compensating for reentrant etching of mesa structures such as out-of-plane microneedles. 62 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

A very little kid ACKNOWLEDGEMENTS 63

Acknowledgements

The present work is a result of the help, guidance, encouragement and assistance of many people. In particular, I would like to thank the following. Göran Stemme for accepting me as his student, sharing his clear-minded and innovative thinking as well as giving me full and encouraging support. Patrick Griss for his supervision, skillful ability to communicate spirit and for initiating the project which funded this work. Wouter van der Wijngaart for introducing me to MEMS research and explaining how it works. Björn Samel and Lina Nordquist for fruitful collaboration and fun times together. Prof. Gerhard Holzapfel and Dr. Christian Gasser for showing a strong interest in my work and accepting joint research. Kjell Norén for his skillful help with all kinds of experimental setups and fine mechanics. The Swedish state through The Swedish Foundation for Strategic Research (SSF) for funding this work. People that helped out with various experiments, especially Dr. Sten Stemme but also Prof. Bernt Lindelöf who made a true impression in my life. People in the lab who generously shared their knowledge, especially Cecilia and Matteo. Special thanks also to Magnus Lindberg for being patient with all my requests concerning the ICP. Sjoerd Haasl and Joachim Oberhammer for always keeping their doors open and their willingness to discuss both technical and other matters. My father Rune and brother Dan for teaching me true craftsmanship—a skill that many researchers lack—and quite possibly a reason this thesis has such a specific title. The people that played a role for me entering this field in the first place. My parents, Heide and Rune, who always supported and encouraged my eagerness to learn. René for introducing me to the field of engineering and explaining to me how things work in the early days. My other brothers and sister for showing me other sides to life. All the excellent teachers I have had during the years that have supported and inspired me to learn more: Barbro Olsson; all the old-school teachers at Thorildsplans Gymnasium, special thanks to Peter Morgan and Ulla Sandberg; teachers at KTH who inspired me to continue into research, especially GP, Peter Fuks and Ulf Ringström. My friends Kristofer, Kalle and Fredrik for all the discussions relating to art and intellectual enterprises in general. Peter R. and Mange L. who inspired me through their determination and sharp-minded thinking. And finally, Evan for adding a fifth dimension to my life, and Vynn for her endless support, ability to endure and indulge my hopeless, but very passionate, behavior. 64 A Fully Integrated Microneedle-based Transdermal Drug Delivery System REFERENCES 65

References

[1] M. P. Cramer and S. R. Saks, “Translating safety, eﬃcacy and compliance into economic value for controlled release dosage forms,” PharmacoEconomics, vol. 5, no. 6, pp. 482–504, 1994.

[2] C. L. Sheen, J. F. Dillon, D. N. Bateman, K. J. Simpson, and T. M. Macdonald, “Paracetamol toxicity: epidemiology, prevention and costs to the health-care system,” QJM, vol. 95, no. 9, pp. 609–19, 2002.

[3] D. R. Owens, B. Zinman, and G. Bolli, “Alternative routes of insulin delivery,” Diabetic Medicine, vol. 20, no. 11, pp. 886–98, 2003.

[4] M. Korytkowski, L. Niskanen, and T. Asakura, “FlexPen(R): Addressing issues of conﬁdence and convenience in insulin delivery,” Clinical Therapeutics, vol. 27, no. Supplement 2, pp. S89–S100, 2005.

[5] M. M. Levine, “Can needle-free administration of vaccines become the norm in global immunization?” Nat. Med., vol. 9, no. 1, pp. 99–103, 2003.

[6]B.Aylward,J.Lloyd,M.Zaﬀran,R.McNair-Scott,andP.Evans,“Reduc- ing the risk of unsafe injections in immunization programmes: ﬁnancial and operational implications of various injection technologies,” Bull. World Health Organ., vol. 73, no. 4, pp. 531–40, 1995.

[7] L. Simonsen, A. Kane, J. Lloyd, M. Zaﬀran, and M. Kane, “Unsafe injections in the developing world and transmission of bloodborne pathogens: a review,” Bull. World Health Organ., vol. 77, pp. 789–800, 1999.

[8] A. Kane, J. Lloyd, M. Zaﬀran, L. Simonsen, and M. Kane, “Transmission of hepatitis B, hepatitis C and human immunodeﬁciency viruses through unsafe injections in the developing world: model-based regional estimates,” Bull. World Health Organ., vol. 77, pp. 801–7, 1999.

[9] M. A. Miller and E. Pisani, “The cost on unsafe injections,” Bull. World Health Organ., vol. 77, pp. 808–11, 1999.

[10] B. D. Ratner, A. Hoﬀman, F. Schoen, and J. Lemons, Eds., Biomaterials Sci- ence, 2nd ed. Elsevier, 2004. 66 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

[11] “The eﬀect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. The Diabetes Control and Complications Trial Research Group,” N. Engl. J. Med., vol. 329, no. 14, pp. 977–86, 1993.

[12] N. Sulli and B. Shashaj, “Long-term beneﬁts of continuous subcutaneous insulin infusion in children with Type 1 diabetes: a 4-year follow-up,” Diabetic Medicine, vol. 23, no. 8, pp. 900–6, 2006.

[13] M. R. Prausnitz, S. Mitragotri, and R. Langer, “Current status and future potential of transdermal drug delivery,” Nat. Rev. Drug. Discov., vol. 3, pp. 115–24, 2004.

[14] B. Barry, “Novel mechanisms and devices to enable successful transdermal drug delivery,” Eur. J. of Pharmaceut. Sci., vol. 14, pp. 101–14, 2001.

[15] E. L. Giudice and J. D. Campbell, “Needle-free vaccine delivery,” Adv. Drug Deliv. Rev., vol. 58, no. 1, pp. 68–89, Apr 2006.

[16] J. Canter, K. Mackey, L. S. Good, R. R. Roberto, J. Chin, W. W. Bond, M. J. Alter, and J. M. Horan, “An outbreak of hepatitis B associated with jet injections in a weight reduction clinic,” Arch. Intern. Med., vol. 150, no. 9, pp. 1923–7, Sep 1990.

[17] M. J. Roy, M. S. Wu, L. J. Barr, J. T. Fuller, L. G. Tussey, S. Speller, J. Culp, J. K. Burkholder, W. F. Swain, R. M. Dixon, G. Widera, R. Vessey, A. King, G. Ogg, A. Gallimore, J. R. Haynes, and D. H. Fuller, “Induction of antigen-speciﬁc CD8+ T cells, T helper cells, and protective levels of antibody in humans by particle-mediated administration of a hepatitis B virus DNA vaccine,” Vaccine, vol. 19, no. 7-8, pp. 764–78, Nov 2000.

[18] D. L. Bremseth and F. Pass, “Delivery of insulin by jet injection: recent observations,” Diabetes Technol. Ther., vol. 3, no. 2, pp. 225–32, 2001.

[19] S. Mitragotri, “Current status and future prospects of needle-free liquid jet injectors,” Nat. Rev. Drug Discov., vol. 5, pp. 543–8, 2006.

[20] A. Arora, I. Hakim, J. Baxter, R. Rathnasingham, R. Srinivasan, D. A. Fletcher, and S. Mitragotri, “Needle-free delivery of macromolecules across the skin by nanoliter-volume pulsed microjets,” Proc. Natl. Acad. Sci. U.S.A,vol. 104, no. 11, pp. 4255–60, 2007.

[21] A. Naik, Y. N. Kalia, and R. H. Guy, “Transdermal drug delivery: overcoming the skin’s barrier function,” Pharm. Sci. Tech. Today., vol. 3, no. 9, pp. 318–26, Sep 2000.

[22] J. Czares-Delgadillo, A. Naik, A. Ganem-Rondero, D. Quintanar-Guerrero, and Y. N. Kalia, “Transdermal delivery of cytochrome C–A 12.4 kDa protein–across intact skin by constant-current iontophoresis,” Pharm. Res., vol. 24, no. 7, pp. 1360–8, Jul 2007. REFERENCES 67

[23] J. A. Subramony, A. Sharma, and J. B. Phipps, “Microprocessor controlled transdermal drug delivery,” Int. J. Pharm., vol. 317, no. 1, pp. 1–6, Jul 2006.

[24] J. A. Tamada, S. Garg, L. Jovanovic, K. R. Pitzer, S. Fermi, and R. O. Potts, “Noninvasive glucose monitoring: comprehensive clinical results. cygnus research team,” JAMA, vol. 282, no. 19, pp. 1839–44, Nov 1999.

[25] N. Kanikkannan, J. Singh, and P. Ramarao, “Transdermal iontophoretic delivery of bovine insulin and monomeric human insulin analogue,” J. Cont. Rel., vol. 59, no. 1, pp. 99–105, May 1999.

[26] Y. N. Kalia, A. Naik, J. Garrison, and R. H. Guy, “Iontophoretic drug delivery,” Adv. Drug. Deliv. Rev., vol. 56, no. 5, pp. 619–58, Mar 2004.

[27] S. Mitragotri, D. Blankschtein, and R. Langer, “Ultrasound-mediated transdermal protein delivery,” Science, vol. 269, no. 5225, pp. 850–3, Aug 1995.

[28] S. Mitragotri and J. Kost, “Low-frequency sonophoresis: a review,” Adv Drug Deliv Rev, vol. 56, no. 5, pp. 589–601, Mar 2004.

[29] J. Kost, “Ultrasound-assisted insulin delivery and noninvasive glucose sensing,” Diabetes Technol Ther, vol. 4, no. 4, pp. 489–97, 2002.

[30] A. Tezel, S. Paliwal, Z. Shen, and S. Mitragotri, “Low-frequency ultrasound as a transcutaneous immunization adjuvant,” Vaccine, vol. 23, no. 29, pp. 3800–7, May 2005.

[31] “Sontra Medical Corp., www.sontra.com,” Jul 2007.

[32] H. Zhai and H. I. Maibach, “Occlusion vs. skin barrier function,” Skin. Res. Technol., vol. 8, no. 1, pp. 1–6, Feb 2002.

[33] A. C. Williams and B. W. Barry, “Penetration enhancers,” Adv. Drug. Deliv. Rev., vol. 56, no. 5, pp. 603–18, Mar 2004.

[34] S. Mitragotri, “Synergistic eﬀect of enhancers for transdermal drug delivery,” Pharm. Res., vol. 17, no. 11, pp. 1354–9, Nov 2000.

[35]S.J.Bashir,A.L.Chew,A.Anigbogu,F.Dreher,andH.I.Maibach,“Phys- ical and physiological eﬀects of stratum corneum tape stripping,” Skin. Res. Technol., vol. 7, no. 1, pp. 40–8, Feb 2001.

[36] H. Dickel, T. M. Bruckner, S. M. Erdmann, J. W. Fluhr, P. J. Frosch, J. Grabbe, H. Lﬄer, H. F. Merk, C. Pirker, H. J. Schwanitz, E. Weisshaar, and J. Brasch, “The ”strip” patch test: results of a multicentre study towards a standardization,” Arch. Dermatol. Res., vol. 296, no. 5, pp. 212–9, 2004.

[37] T. Herndon, S. Gonzalez, T. R. Gowrishankar, R. Anderson, and J. Weaver, “Transdermal microconduits by microscission for drug delivery and sample acquisition,” BMC Medicine, vol. 2, no. 1, p. 12, 2004. 68 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

[38] J. S. Dover, G. J. Hruza, and K. A. Arndt, “Lasers in skin resurfacing,” Semin. Cutan. Med. Surg., vol. 19, no. 4, pp. 207–20, Dec 2000. [39] P. D. Gadiraju, J.-H. Park, J. W. Lee, M. R. Prausnitz, and M. G. Allen, “Micro-ablation of skin by arc-discharge jet ejection for transdermal drug delivery,” in Transducers’07 The 14th Int. Conf. on Solid-state Sensors, Actuators and Microsystems, Lyon, France, 2007, pp. 1947–50. [40] A. P. Gadre, A. J. Nijdam, J. A. Garra, A. H. Monica, M. C. Cheng, C. Luo, Y. N. Srivastava, T. W. Schneider, T. J. Long, R. C. White, M. Paranjape, and J. F. Currie, “Fabrication of a ﬂuid encapsulated dermal patch using multilayered SU-8,” Sens. Actuators A, Phys., vol. A114, no. 2-3, pp. 478–85, 2004. [41] “PassPort patch, www.alteatherapeutics.com,” Jul 2007. [42] 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., vol. 92, pp. 502–4, 2001. [43] 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. Tech., vol. 11, no. 11, pp. 152–6, 2005. [44] M. R. Prausnitz, “Microneedles for transdermal drug delivery,” Adv. Drug Deliv. Rev., vol. 56, pp. 581–7, 2004. [45] M. L. Reed and W.-K. Lye, “Microsystems for drug and gene delivery,” Proc. IEEE, vol. 92, no. 1, pp. 56–75, 2004. [46] A. L. Teo, C. Shearwood, K. C. Ng, J. Lu, and S. Moochhala, “Transdermal microneedles for drug delivery applications,” Mater. Sci. Eng. B, vol. 132, no. 1-2, pp. 151–4, 2006. [47] J. Banchereau and R. M. Steinman, “Dendritic cells and the control of immunity,” Nature, vol. 392, no. 6673, pp. 245–52, Mar 1998. [48] C. Condon, S. C. Watkins, C. M. Celluzzi, K. Thompson, and L. D. Falo, “Dna- based immunization by in vivo transfection of dendritic cells,” Nat. Med.,vol.2, no. 10, pp. 1122–8, Oct 1996. [49] S. Babiuk, M. Baca-Estrada, L. A. Babiuk, C. Ewen, and M. Foldvari, “Cuta- neous vaccination: the skin as an immunologically active tissue and the challenge of antigen delivery,” J. Contr. Rel., vol. 66, no. 2-3, pp. 199–214, May 2000. [50] S. A. Plotkin, “Vaccines: past, present and future,” Nat Med, vol. 11, no. 4, pp. S5–11, 2005. [51] R. T. Kenney, S. A. Frech, L. R. Muenz, C. P. Villar, and G. M. Glenn, “Dose sparing with intradermal injection of inﬂuenza vaccine,” N. Engl. J. Med.,vol. 351, no. 22, pp. 2295–301, Nov 2004. REFERENCES 69

[52] J. B. Alarcon, A. W. Hartley, N. G. Harvey, and J. A. Mikszta, “Preclinical evaluation of microneedle technology for intradermal delivery of inﬂuenza vaccines,” Clin. Vaccine Immunol., vol. 14, no. 4, pp. 375–81, Apr 2007.

[53] J. A. Matriano, M. Cormier, J. Johnson, W. Young, M. Buttery, K. Nyam, and P. Daddona, “Macroﬂux microprojection array patch technology: A new and eﬃcient approach for intracutaneous immunization,” Pharm. Res., vol. 19, no. 1, pp. 63–70, 2002.

[54] G. Widera, J. Johnson, L. Kim, L. Libiran, K. Nyam, P. E. Daddona, and M. Cormier, “Eﬀect of delivery parameters on immunization to ovalbumin following intracutaneous administration by a coated microneedle array patch system,” Vaccine, vol. 24, no. 10, pp. 1653–64, Mar 2006.

[55] 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 targeted epidermal delivery,” Nat. Med.,vol.8, no. 4, pp. 415–9, Apr 2002.

[56] D. U. Ekwueme, B. G. Weniger, and R. T. Chen, “Model-based estimates of risks of disease transmission and economic costs of seven injection devices in sub-Saharan Africa,” Bull. World Health Organ., vol. 80, no. 11, pp. 859–70, 2002.

[57] “Vaccine market to top °23b, market analysis, united press int.” Feb 9 2007.

[58] L. P. O. Norl´en, “The skin barrier — structure and physical function,” Ph.D. dissertation, Karolinska Institute, Stockholm, Sweden, 1999.

[59] M. H. Ross, W. Pawlina, and G. I. Kaye, Histology : A Text and Atlas,4thed. Lippincott Williams & Wilkins, 2003.

[60] “www.oup.co.uk/oxed/children/oise/pictures/humans/skin/,” June 2007.

[61] R. C. Haut, Biomechanics of soft tissue, 2nd ed. New York: Springer, 2002, ch. 11, pp. 228–53.

[62] K. Langer, “On the anatomy and physiology of the skin i-iv,” Sitzungsbericht der Akademie der Wissenschaften in Wien, 1861. Translated by T. Gibson and reprinted: Br. J. Plast. Surg., vol. 31, pp. 3–8, 93–106, 185–99, 273–8, 1978.

[63] R. Reihsner, B. Balogh, and E. J. Menzel, “Two-dimensional elastic properties of human skin in terms of an incremental model at the in vivo conﬁguration,” Med. Eng. Phys., vol. 17, pp. 304–13, 1995.

[64] C. H. Daly, “Biomechanical properties of dermis,” J. Invest. Dermatol.,vol.79 Suppl 1, pp. 17s–20s, Jul 1982.

[65] P. A. Payne, “Measurement of properties and function of skin,” Clin. Phys. Physiol. Meas., vol. 12, pp. 105–29, 1991. 70 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

[66] A. Park and C. Baddiel, “Rheology of stratum corneum I.A. molecular inter- pretation of the stress-strain curve,” J. Soc. Cos. Chem., vol. 23, pp. 3–12, 1972.

[67] Y. S. Papir, K.-H. Hsu, and R. H. Wildnauer, “The mechanical properties of stratum corneum : I. The eﬀect of water and ambient temperature on the tensile properties of newborn rat stratum corneum,” Biochim. Biophys. Acta Gen. Subj., vol. 399, no. 1, pp. 170–80, 1975.

[68] A. Hendley, R. Marks, and P. Payne, “Measurement of forces for point indenta- tion of the stratum corneum in vivo: the inﬂuence of age, sex, site delipidisation and hydration,” Bioengineering and the Skin, vol. 3, pp. 234–40, 1982.

[69] K. S. Wu, W. W. van Osdol, and R. H. Dauskardt, “Mechanical properties of human stratum corneum: Eﬀects of temperature, hydration, and chemical treatment,” Biomaterials, vol. 27, no. 5, pp. 785–95, 2006.

[70] F. M. Hendriks, D. Brokken, C. W. Oomens, D. L. Bader, and F. P. Baaijens, “The relative contributions of diﬀerent skin layers to the mechanical behavior of human skin in vivo using suction experiments,” Med. Eng. Phys., vol. 28, pp. 259–66, 2006.

[71] F. M. Hendriks, “Mechanical behaviour of human epidermal and dermal layers in vivo,” Ph.D. dissertation, TU Eindhoven, Eindhoven, The Netherlands, 2005.

[72] J. D. Humphrey, “Continuum biomechanics of soft biological tissues,” Proc. R. Soc. Lond. A, vol. 459, pp. 3–46, 2003.

[73] P. Tong and Y. C. Fung, “The stress-strain relationship for the skin,” J. Biomech., vol. 9, no. 10, pp. 649–57, 1976.

[74] Y. Lanir, “Constitutive equations for ﬁbrous connective tissues,” J. Biomech., vol. 16, no. 1, pp. 1–12, 1983.

[75] C. W. Oomens, D. H. van Campen, and H. J. Grootenboer, “A mixture approach to the mechanics of skin,” J. Biomech., vol. 20, no. 9, pp. 877–85, 1987.

[76] O. A. Shergold and N. A. Fleck, “Mechanics of the deep penetration of soft solids with application to the injection and wounding of skin,” Proc.R.Soc. Lond. A, vol. 460, pp. 3037–58, 2004.

[77] F. M. Hendriks, D. Brokken, J. T. W. M. van Eemeren, C. W. J. Oomens, F. P. T. Baaijens, and J. B. A. M. Horsten, “A numerical-experimental method to characterize the non-linear mechanical behaviour of human skin,” Skin Res. Technol., vol. 9, no. 3, pp. 274–83, Aug 2003.

[78] O. A. Shergold, N. A. Fleck, and T. S. King, “The penetration of a soft solid by a liquid jet, with application to the administration of a needle-free injection,” J. Biomech., vol. 39, no. 14, pp. 2593–602, 2006. REFERENCES 71

[79] R. W. Ogden, “Large deformation isotropic elasticity — on the correlation of theory and experiment for incompressible rubberlike solids,” Proc. R. Soc. Lond. A, vol. 326, pp. 565–84, 1972. [80] O. A. Shergold and N. A. Fleck, “Experimental investigation into the deep penetration of soft solids by sharp and blunt punches, with application to the piercing of skin,” J. Biomech. Eng., vol. 127, pp. 838–48, 2005. [81] O. Coussy, Poromechanics. John Wiley & Sons, 2004. [82] A. A. Griffith, “The phenomena of rupture and flow in solids,” Phil. Trans. R. Soc. Lond. A, vol. 221, pp. 163–98, 1921. [83] P. P. Purslow, “Measurement of the fracture toughness of extensible connective tissues,” J. Mater. Sci., vol. 18, no. 12, pp. 3591–8, 1983. [84] B. P. Pereira, P. W. Lucas, and T. Swee-Hin, “Ranking the fracture toughness of thin mammalian soft tissues using the scissors cutting test,” J. Biomech., vol. 30, no. 1, pp. 91–4, 1997. [85] 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., vol. 37, pp. 1155–63, 2004. [86] P. P. Purslow, “Fracture of non-linear biological materials: some observations from practice relevant to recent theory,” J. Phys. D: Appl. Phys., vol. 22, no. 6, pp. 854–6, 1989. [87] S. P. Davis, “Hollow microneedles for molecular transport across skin,” Ph.D. dissertation, Georgia Institute of Technology, GA, USA, 2003. [88] P. Griss and G. Stemme, “Side-opened out-of-plane microneedles for microfluidic transdermal liquid transfer,” IEEE ASME J. Microelectromech. Syst., vol. 12, no. 3, pp. 296–301, 2003. [89] W. Martanto, J. S. Moore, T. Couse, and M. R. Prausnitz, “Mechanism of fluid infusion during microneedle insertion and retraction,” J. Contr. Rel.,vol. 112, no. 3, pp. 357–61, 2006. [90] W. Martanto, J. S. Moore, O. Kashlan, R. Kamath, P. M. Wang, J. M. O’Neal, and M. R. Prausnitz, “Microinfusion using hollow microneedles,” Pharm. Res., vol. 23, no. 1, pp. 104–13, 2006. [91] P. M. Wang, M. Cornwell, J. Hill, and M. R. Prausnitz, “Precise microinjection into skin using hollow microneedles,” J. Investig. Dermatol., vol. 126, no. 5, pp. 1080–7, 2006. [92] N. H. Talbot and A. P. Pisano, “Polymolding: two wafer polysilicon micromolding of closed-flow passages for microneedles and microfluidic devices,” in Technical Digest. Solid-State Sensor and Actuator Workshop, Hilton Head Is- land, SC, USA, 1998, pp. 265–8. 72 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

[93] P. K. Campbell, K. E. Jones, R. J. Huber, K. W. Horch, and R. A. Normann, “A silicon-based, three-dimensional neural interface: manufacturing processes for an intracortical electrode array,” IEEE Trans. Biomed. Eng., vol. 38, no. 8, pp. 758–68, Aug 1991. [94] S. L. BeMent, K. D. Wise, D. J. Anderson, K. Najafi, and K. L. Drake, “Solid- state electrodes for multichannel multiplexed intracortical neuronal recording,” IEEE Trans. Biomed. Eng., vol. 33, no. 2, pp. 230–41, Feb 1986. [95] J. Chen, K. D. Wise, J. F. Hetke, and S. C. Bledsoe, “A multichannel neural probe for selective chemical delivery at the cellular level,” IEEE Trans. Biomed. Eng., vol. 44, no. 8, pp. 760–9, Aug 1997. [96] P. Griss, P. Enoksson, H. Tolvanen-Laakso, P. Merilainen, S. Ollmar, and G. Stemme, “Micromachined electrodes for biopotential measurements,” IEEE ASME J. Microelectromech. Syst., vol. 10, no. 1, pp. 10–6, 2001. [97] P. Griss, H. Tolvanen-Laakso, P. Merilainen, and G. Stemme, “Characterization of micromachined spiked biopotential electrodes,” IEEE Trans. Biomed. Eng., vol. 49, no. 6, pp. 597–604, 2002. [98] P. Aberg,˚ P. Geladi, I. Nicander, J. Hansson, U. Holmgren, and S. Ollmar, “Non-invasive and microinvasive electrical impedance spectra of skin cancer — a comparison between two techniques,” Skin Res. Techl., vol. 11, no. 4, pp. 281–6, 2005. [99] “SciBase AB, www.scibase.se,” Jul 2007. [100] K. Oka, S. Aoyagi, Y. Arai, Y. Isono, G. Hashiguchi, and H. Fujita, “Fabrication of a micro needle for a trace blood test,” Sens. Actuators A, Phys., vol. 97-98, pp. 478–85, Apr. 2002. [101] E. V. Mukerjee, S. D. Collins, R. R. Isseroff, and R. L. Smith, “Microneedle array for transdermal biological fluid extraction and in situ analysis,” Sens. Actuators A, Phys., vol. A114, no. 2-3, pp. 267–75, 2004. [102] J. D. Zahn, D. Trebotich, and D. Liepmann, “Microdialysis microneedles for continuous medical monitoring,” Biomed. Microdevices, vol. 7, no. 1, pp. 59–69, Mar 2005. [103] M. S. Gerstel and V. A. Place, “U.S. patent no. 3,964,482: Drug delivery device,” 1976. [104] S. R. Rosenthal, “U.S. patent no. 2,619,962: Vaccination appliance,” 1952. [105] H. Kravitz and N. Lettvin, “U.S. patent no. 2,817,336: Means for vaccinating,” 1957. [106] S. Henry, D. V. McAllister, M. G. Allen, and M. R. Prausnitz, “Microfabricated microneedles: a novel method to increase transdermal drug delivery,” J. of Pharmaceutical Sci., vol. 87, pp. 922–5, 1998. REFERENCES 73

[107] “Becton, Dickinson and Co., www.bd.com/products,” Jul 2007. [108] G. Kotzar, M. Freas, P. Abel, A. Fleischman, S. Roy, C. Zorman, J. M. Moran, and J. Melzak, “Evaluation of MEMS materials of construction for implantable medical devices,” Biomaterials, vol. 23, no. 13, pp. 2737–50, 2002. [109] G. Voskerician, M. S. Shive, R. S. Shawgo, H. von Recum, J. M. Anderson, M. J. Cima, and R. Langer, “Biocompatibility and biofouling of MEMS drug delivery devices,” Biomaterials, vol. 24, no. 11, pp. 1959–67, 2003. [110] L. Ferrara, A. Fleischman, D. Togawa, T. Bauer, E. Benzel, and S. Roy, “An in vivo biocompatibility assessment of MEMS materials for spinal fusion monitoring,” Biomed. Microdevices, vol. 5, no. 4, pp. 297–302, 2003. [111] F. Laermer and A. Schilp, “German Pat. DE-4241045: Verfahren zum anisotropen tzen von Silicium,” 1994. [112] F. Laermer, “personal discussion,” Jun 2005. [113] M. Puech, J. M. Thevenoud, J. M. Gruffat, N. Launay, P. Godinat, and O. L. Barillec, “Achievements and perspectives of the DRIE technology for the microsystems market,” in Transducers’07. The 14th Int. Conf. on Solid-state Sen- sors, Actuators and Microsystems, vol. 1, Lyon, France, 2007, pp. 77–80. [114] A. Badihi, “Ultrathin wafer level chip size package,” IEEE Trans. Adv. Packag., vol. 23, no. 2, pp. 212–4, 2000. [115] T. A. Chou and K. Najafi, “Fabrication of out-of-plane curved surfaces in Si by utilizing RIE lag,” in 15th IEEE Int. Conf. on Micro Electro Mechanical Systems, Las Vegas, NV, USA, 2002, pp. 145–8. [116] T. Bourouina, T. Masuzawa, and H. Fujita, “The MEMSNAS process: microloading effect for micromachining 3-D structures of nearly all shapes,” IEEE ASME J. Microelectromech. Syst., vol. 13, no. 2, pp. 190–9, 2004. [117] M. P. Rao, M. F. Aimi, and N. C. MacDonald, “Single-mask, three-dimensional microfabrication of high-aspect-ratio structures in bulk silicon using reactive ion etching lag and sacrificial oxidation,” Appl. Phys. Lett., vol. 85, no. 25, pp. 6281–3, 2004. [118] R. F. Figueroa, S. Spiesshoefer, S. L. Burkett, and L. Schaper, “Control of sidewall slope in silicon vias using SF6/O2 plasma etching in a conventional reactive ion etching tool,” J. Vac. Sci. Tech. B, vol. 23, no. 5, pp. 2226–31, 2005. [119] S.-B. Jo, M.-W. Lee, S.-G. Lee, E.-H. Lee, S.-G. Park, and B.-H. O, “Characterization of a modified Bosch-type process for silicon mold fabrication,” J. Vac. Sci. Tech. A, vol. 23, no. 4, pp. 905–10, 2005. [120] B. Morgan, X. Hua, T. Iguchi, T. Tomioka, G. S. Oehrlein, and R. Ghodssi, “Substrate interconnect technologies for 3-D MEMS packaging,” Microelectronic Engineering, vol. 81, no. 1, pp. 106–16, 2005. 74 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

[121] R. Dizon, H. Han, A. G. Russell, and M. L. Reed, “An ion milling pattern transfer technique for fabrication of three-dimensional micromechanical structures,” IEEE ASME J. Microelectromech. Syst., vol. 2, no. 4, pp. 151–9, 1993. [122] S. Hashmi, P. Ling, G. Hashmi, M. Reed, R. Gaugler, and W. Trimmer, “Ge- netic transformation of nematodes using arrays of micromechanical piercing structures,” Biotechniques, vol. 19, no. 5, pp. 766–70, Nov 1995. [123] M. L. Reed and L. E. Weiss, “U.S. patent no. 5,312,456: Micromechanical barb and method for making the same,” May 1994. [124] W. Trimmer, P. Ling, C.-K. Chin, P. Orton, R. Gaugler, S. Hashmi, G. Hashmi, B. Brunett, and M. Reed, “Injection of DNA into plant and animal tissues with micromechanical piercing structures,” in 8th IEEE Ann. Int. Workshop on Micro Electro Mechanical Systems, Nagoya, Japan, 1995, pp. 111–5. [125] M. L. Reed, C. Wu, J. Kneller, S. Watkins, D. A. Vorp, A. Nadeem, L. E. Weiss, K. Rebello, M. Mescher, A. J. Smith, W. Rosenblum, and M. D. Feldman, “Micromechanical devices for intravascular drug delivery,” J. Pharm. Sci., vol. 87, no. 11, pp. 1387–94, Nov 1998. [126] T. H. Mitsuhiro Shikida and K. Sato, “Fabrication of a hollow needle structure by dicing, wet etching and metal deposition,” J. Micromech. Microeng., vol. 16, no. 10, pp. 2230–9, 2006. [127] N. Wilke and A. Morrissey, “Silicon microneedle formation using modiﬁed mask designs based on convex corner undercut,” J. Micromech. Microeng., vol. 17, no. 2, pp. 238–44, 2007. [128] D. V. McAllister, “Microfabricated needles for transdermal drug delivery,” Ph.D. dissertation, Georgia Institute of Technology, GA, USA, 2000. [129] 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 nanoparticles: Fabrication methods and transport studies,” Proc. Natl. Acad. Sci. U.S.A, vol. 100, no. 24, pp. 13 755–60, 2003. [130] F. Chabri, K. Bouris, T. Jones, D. Barrow, A. Hann, C. Allender, K. Brain, and J. Birchall, “Microfabricated silicon microneedles for nonviral cutaneous gene delivery,” Br.J.Dermat., vol. 150, no. 5, pp. 869–77, May 2004. [131] 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. Contr. Rel., vol. 97, no. 3, pp. 503–11, Jul 2004. [132] J. C. Trautman, R. L. Keenan, A. P. Samiee, W. Q. Lin, M. Cormier, J. Matri- ano, and P. E. Daddona, “U.S. patent no. 20020123675: Apparatus and method for piercing skin with microprotrusions,” 2002. [133] W. Lin, M. Cormier, and P. E. Daddona, Celluar Drug Delivery: Principles and Practice. Humana Press, 2004, ch. 14, pp. 277–85. REFERENCES 75

[134] W. Lin, M. Cormier, A. Samiee, A. Griﬃn, B. Johnson, C. L. Teng, G. E. Hardee, and P. E. Daddona, “Transdermal delivery of antisense oligonucleotides with microprojection patch (Macroﬂux) technology,” Pharm. Res., vol. 18, no. 12, pp. 1789–93, 2001. [135] M. J. N. Cormier, A. P. Neukermans, B. Block, F. T. Theeuwes, and A. A. Amkraut, “European Patent No. EP0914178: Device for enhancing transdermal agent delivery or sampling,” 1999. [136] H. S. Gill and M. R. Prausnitz, “Coated microneedles for transdermal delivery,” J. Contr. Rel., vol. 117, no. 2, pp. 227–37, Feb 2007. [137] ——, “Coating formulations for microneedles,” Pharm. Res.,vol.24,no.7,pp. 1369–80, Jul 2007. [138] J. Raeder-Devens, “Microstructured Transdermal Systems (MTS), 3M Drug Delivery Systems, http://solutions.3m.com/3MContentRetrievalAPI/ BlobServlet?locale=en WW&univid=1114280334026&fallback=true& assetType=MMM Image&blobAttribute=ImageFile,” Nov. 2005. [139] J. A. Mikszta, V. J. Sullivan, C. Dean, A. M. Waterston, J. B. Alarcon, J. P. Dekker, J. M. Brittingham, J. Huang, C. R. Hwang, M. Ferriter, G. Jiang, K. Mar, K. U. Saikh, B. G. Stiles, C. J. Roy, R. G. Ulrich, and N. G. Harvey, “Protective immunization against inhalational anthrax: a comparison of minimally invasive delivery platforms,” J. Infect. Dis., vol. 191, no. 2, pp. 278–88, Jan 2005. [140] V. V. Yuzhakov, F. F. Sherman, G. D. Owens, and V. Gartstein, “U.S. patent no. 2005/0209565: Intracutaneous microneedle array apparatus,” 2005. [141] “MicroCor Delivery System, www.coriumgroup.com,” Aug. 2007. [142] L. Olson, “U.S. patent no. 2003/0009113: Micro-needles and methods of man- ufacture and use thereof,” Jan. 2003. [143] J.-H. Park, M. G. Allen, and M. R. Prausnitz, “Polymer microneedles for transdermal drug delivery,” in Proc.Int.Symp.Control.Rel.Bioact.Mater.,Seoul, Korea, 2002. [144] J.-H. Park, S. Davis, Y.-K. Yoon, M. G. Allen, and M. R. Prausnitz, “Micro- machined biodegradable microstructures,” in 16th IEEE Int. Conf. on Micro Electro Mechanical Systems, Kyoto, Japan, 2003, pp. 371–4. [145] J.-H. Park, M. G. Allen, and M. R. Prausnitz, “Biodegradable polymer microneedles: Fabrication, mechanics and transdermal drug delivery,” J. Contr. Rel., vol. 104, no. 1, pp. 51–66, May 2005. [146] J.-H. Park, Y.-K. Yoon, S.-O. Choi, M. R. Prausnitz, and M. G. Allen, “Tapered conical polymer microneedles fabricated using an integrated lens technique for transdermal drug delivery,” IEEE Trans. Biomed. Eng., vol. 54, no. 5, pp. 903– 13, 2007. 76 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

[147] J.-H. Park, M. G. Allen, and M. R. Prausnitz, “Polymer microneedles for controlled-release drug delivery,” Pharm. Res., vol. 23, pp. 1008–19, 2006.

[148] J.-H. Park, S.-O. Choi, R. Kamath, Y.-K. Yoon, M. G. Allen, and M. R. Prausnitz, “Polymer particle-based micromolding to fabricate novel microstructures,” Biomed. Microdevices, vol. 9, no. 2, pp. 223–34, Apr 2007.

[149] 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, vol. 7, no. 3, pp. 185–8, Sep 2005.

[150] T. Miyano, T. Miyachi, T. Okanishi, H. Todo, K. Sugibayashi, T. Uemura, N. Takano, and S. Konishi, “Hydrolytic microneedles as transdermal drug delivery system,” in Transducers’07. The 14th Int. Conf. on Solid-state Sensors, Actuators and Microsystems, Lyon, France, 2007, pp. 355–8.

[151] C. Kolli and A. Banga, “Characterization of solid maltose microneedles and their use for transdermal delivery,” Pharm. Res., Jun 2007.

[152] W. Martanto, S. P. Davis, H. R. Holiday, J. Wang, H. S. Gill, and M. R. Prausnitz, “Transdermal delivery of insulin using microneedles in vivo,” Pharm. Res., vol. 21, no. 6, pp. 947–52, 2004.

[153] A. Trautmann, F. Heuck, C. Mueller, P. Ruther, and O. Paul, “Replication of microneedle arrays using vacuum casting and hot embossing,” in Transduc- ers’05. The 13th Int. Conf. on Solid-state Sensors, Actuators and Microsystems, vol. 2, Seoul, Korea, 2005, pp. 1420–3.

[154] M. Han, D.-H. Hyun, H.-H. Park, S. S. Lee, C.-H. Kim, and C. Kim, “A novel fabrication process for out-of-plane microneedle sheets of biocompatible polymer,” J. Micromech. Microeng., vol. 17, no. 6, pp. 1184–91, 2007.

[155] D. V. McAllsiter, F. Cros, S. P. Davis, L. M. Matta, M. R. Prausnitz, and M. G. Allen, “Three-dimensional hollow microneedles and microtube arrays,” in Transducers’99. The 10th Int. Conf. on Solid-state Sensors and Actuators, Sendai, Japan, 1999, pp. 1098–101.

[156] K. Kim, D. S. Park, H. M. Lu, W. Che, K. Kim, J.-B. Lee, and C. H. Ahn, “A tapered hollow metallic microneedle array using backside exposure of su-8,” J. Micromech. Microeng., vol. 14, no. 4, pp. 597–603, 2004.

[157] B. Stoeber and D. Liepmann, “Fluid injection through out-of-plane microneedles,” in 1st Ann. Int. IEEE-EMBS Special Topic Conf. on Microtechnologies in Medicine and Biology, 2000, pp. 224–28.

[158] ——, “Arrays of hollow out-of-plane microneedles for drug delivery,” IEEE ASME J. Microelectromech. Syst., vol. 14, no. 3, pp. 472–9, 2005.

[159] S. T. Cho, “U.S. patent no. 2002/0193754: Microneedles for minimally invasive drug delivery,” 2002. REFERENCES 77

[160] H. Gardeniers, R. Luttge, E. Berenschot, M. de Boer, S. Yeshurun, M. Hefetz, R. van’t Oever, and A. van den Berg, “Silicon micromachined hollow microneedles for transdermal liquid transport,” IEEE ASME J. Microelectromech. Syst., vol. 12, no. 6, pp. 855–62, 2003.

[161] S. J. Moon and S. S. Lee, “A novel fabrication method of a microneedle array using inclined deep x-ray exposure,” J. Micromech. Microeng., vol. 15, no. 5, pp. 903–11, 2005.

[162] P. Loeters, R. Duwel, F. Verbaan, R. Luttge, D. van den Berg, J. Bouwstra, and A. van den Berg, “Measuring the insertion of microfabricated microneedles into skin with a penetration sensor,” in Proc. μTAS 2004 8th Int. Conf. on Miniaturized Systems in Chemistry and Life Sciences, vol. 1, Malm¨o, Sweden, 2004, pp. 497–9.

[163] R. Luttge, E. J. W. Berenschot, M. J. de Boer, D. M. Altpeter, E. X. Vrouwe, A. van den Berg, and M. Elwenspoek, “Integrated lithographic molding for microneedle-based devices,” IEEE ASME J. Microelectromech. Syst., vol. 16, no. 4, pp. 872–84, 2007.

[164] “Current Patents Gazette, iss. 0648,” Dec. 2006.

[165] M. Hefetz, G. Fruchtman, and Y. Levine, “WO patent no. 03/074102 A2: De- vices and methods for transporting ﬂuid across a biological barrier,” 2003.

[166] Y. Yeshurun, M. Hefetz, Y. Sefi, Y. Levine, and G. Lavi, “WO patent no. 2006/054280 A2: System and methods for delivering fluid into flexible biological barrier,” May 2006.

[167] M. J. Madou, Fundamentals of Microfabrication, 2nd ed. CRC press, 2002.

[168] S. P. Davis, W. Martanto, M. G. Allen, and M. R. Prausnitz, “Hollow metal microneedles for insulin delivery to diabetic rats,” IEEE Trans. Biomed. Eng., vol. 52, no. 5, pp. 909–15, 2005.

[169] M. Teo, C. Shearwood, K. C. Ng, J. Lu, and S. Moochhala, “In vitro and in vivo characterization of MEMS microneedles,” Biomed. Microdevices,vol.7, no. 1, pp. 47–52, 2005.

[170] N. Roxhed, P. Griss, and G. Stemme, “Reliable in-vivo penetration and transdermal injection using ultra-sharp hollow microneedles,” in Transducers’05. The 13th Int. Conf. on Solid-state Sensors, Actuators and Microsystems,vol.1, Seoul, Korea, 2005, pp. 213–6.

[171] S.-O. Choi, J.-H. Park, Y. Choi, Y. Kim, H. S. Gill, Y.-K. Yoon, M. R. Praus- nitz, and M. G. Allen, “An electrically active microneedle array for electro- poration of skin for gene delivery,” in Transducers’05. The 13th Int. Conf. on Solid-state Sensors, Actuators and Microsystems, vol. 2, Seoul, Korea, 2005, pp. 1513–6. 78 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

[172] D. J. Laser and J. G. Santiago, “A review of micropumps,” J. Micromech. Microeng., vol. 14, no. 6, pp. R35–64, 2004. [173] X. Yang, C. Grosjean, Y.-C. Tai, and C.-M. Ho, “A MEMS thermopneumatic silicone rubber membrane valve,” Sens. Actuators A, Phys., vol. 64, no. 1, pp. 101–8, 1998. [174] A. Wego, H.-W. Glock, L. Pagel, and S. Richter, “Investigations on thermopneumatic volume actuators based on PCB technology,” Sens. Actuators A, Phys., vol. A93, no. 2, pp. 95–102, 2001. [175] C. G. Cooney and B. C. Towe, “A thermopneumatic dispensing micropump,” Sens. Actuators A, Phys., vol. A116, no. 3, pp. 519–24, 2004. [176] W. van der Wijngaart, D. Chugh, E. Man, J. Melin, and G. Stemme, “A low- temperature thermopneumatic actuation principle for gas bubble microvalves,” IEEE ASME J. Microelectromech. Syst., vol. 16, no. 3, pp. 765–74, 2007. [177] P. Selvaganapathy, E. T. Carlen, and C. H. Mastrangelo, “Electrothermally actuated inline microfluidic valve,” Sens. Actuators A, Phys., vol. A104, no. 3, pp. 275–82, 2003. [178] R. Boden, M. Lehto, U. Simu, G. Thornell, K. Hjort, and J.-A. Schweitz, “A polymeric paraffin actuated high-pressure micropump,” Sens. Actuators A, Phys., vol. 127, no. 1, pp. 88–93, 2006. [179] C.-C. Hong, S. Murugesan, S. Kim, G. Beaucage, J.-W. Choi, and C. Ahn, “A functional on-chip pressure generator using solid chemical propellant for disposable lab-on-a-chip,” Lab Chip, vol. 3, no. 4, pp. 281–6, 2003. [180] S. Bohm, W. Olthuis, and P. Bergveld, “An integrated micromachined electrochemical pump and dosing system,” Biomed. Microdevices, vol. 1, no. 2, pp. 121–30, 1999. [181] J. W. Munyan, H. V. Fuentes, M. Draper, R. T. Kelly, and A. T. Woolley, “Electrically actuated, pressure-driven microfluidic pumps,” Lab Chip,vol.3, no. 4, pp. 217–20, 2003. [182] C. Yu, S. Mutlu, P. Selvaganapathy, C. H. Mastrangelo, F. Svec, and J. M. Frechet, “Flow control valves for analytical microfluidic chips without mechanical parts based on thermally responsive monolithic polymers,” Anal. Chem., vol. 75, pp. 1958–61, 2003. [183] A. Richter, D. Kuckling, S. Howitz, T. Gehring, and K.-F. Arndt, “Electronically controllable microvalves based on smart hydrogels: magnitudes and potential applications,” IEEE ASME J. Microelectromech. Syst., vol. 12, no. 5, pp. 748–53, 2003. [184] C.-C. Hong, J.-W. Choi, and C. H. Ahn, “An on-chip air-bursting detonator for driving fluids on disposable lab-on-a-chip systems,” J. Micromech. Microeng., vol. 17, no. 2, pp. 410–7, 2007. REFERENCES 79

[185] P. Griss, H. Andersson, and G. Stemme, “Expandable microspheres for the handling of liquids,” Lab Chip, vol. 2, no. 2, pp. 117–20, May 2002. [186] C.-C. Hong, J.-W. Choi, and C. H. Ahn, “Disposable air-bursting detonators as an alternative on-chip power source,” in 15th IEEE Int. Conf. on Micro Electro Mechanical Systems, Las Vegas, NV, USA, 2002, pp. 240–3. [187] B. Samel, P. Griss, and G. Stemme, “A thermally responsive PDMS composite and its microfluidic applications,” IEEE ASME J. Microelectromech. Syst., vol. 16, no. 1, pp. 50–7, 2007. [188] B. Samel, V. Nock, A. Russom, P. Griss, and G. Stemme, “A disposable lab-on-a-chip platform with embedded fluid actuators for active nanoliter liquid handling,” Biomed. Microdevices, vol. 9, no. 1, pp. 61–7, Feb 2007. [189] B. Samel, J. Chretien, R. Yue, P. Griss, and G. Stemme, “Wafer-level process for single-use buckling-film microliter-range pumps,” IEEE ASME J. Microelec- tromech. Syst., vol. 16, no. 4, pp. 795–801, 2007. [190] B. Stoeber and D. Liepmann, “Design, fabrication, and testing of a MEMS syringe,” in Technical Digest. Solid-State Sensor and Actuator Workshop, Hilton Head Island, SC, USA, 2002. [191] D. Maillefer, S. Gamper, B. Frehner, P. Balmer, H. van Lintel, and P. Re- naud, “A high-performance silicon micropump for disposable drug delivery systems,” in 14th IEEE Int. Conf. on Micro Electro Mechanical Systems,Inter- laken, Switzerland, 2001, pp. 413–7. [192] H. T. G. van Lintel, F. C. M. van De Pol, and S. Bouwstra, “A piezoelectric micropump based on micromachining of silicon,” Sens. Actuators, vol. 15, no. 2, pp. 153–67, Oct 1988. [193] “Debiotech SA, Insulin Nanopump, www.debiotech.com,” Aug 2007. [194] J. D. Zahn, A. Deshmukh, A. P. Pisano, and D. Liepmann, “Continuous on-chip micropumping for microneedle enhanced drug delivery,” Biomed. Microdevices, vol. 6, no. 3, pp. 183–90, 2004. [195] P. Griss, “Micromachined interfaces for medical and biochemical applications,” Ph.D. dissertation, Royal Institute of Technology, Stockholm, Sweden, 2002. [196] A. A. Ayon, R. Braff, C. C. Lin, H. H. Sawin, and M. A. Schmidt, “Characterization of a time multiplexed inductively coupled plasma etcher,” J. Electrochem. Soc., vol. 146, no. 1, pp. 339–49, 1999. [197] R. B. Marcus, T. S. Ravi, T. Gmitter, K. Chin, D. Liu, W. J. Orvis, D. R. Ciarlo, C. E. Hunt, and J. Trujillo, “Formation of silicon tips with < 1nm radius,” Appl. Phys. Lett., vol. 56, no. 3, pp. 236–8, 1990. [198] T. S. Ravi, R. B. Marcus, and D. Liu, “Oxidation sharpening of silicon tips,” J. Vac. Sci. Tech. B, vol. 9, no. 6, pp. 2733–7, 1991. 80 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

[199] P. F. Man, C. H. Mastrangelo, M. A. Burns, and D. T. Burke, “Microfabricated capillarity-driven stop valve and sample injector,” in 11th IEEE Ann. Int. Workshop on Micro Electro Mechanical Systems, Heidelberg, Germany, 1998, pp. 45–50. [200] J. W. Kwon, S. Kamal-Bahl, and E. S. Kim, “Film transfer and bonding technique to cover lab on a chip,” in Transducers’05. The 13th Int. Conf. on Solid- state Sensors, Actuators and Microsystems, vol. 1, Seoul, Korea, 2005, pp. 940– 3. [201] F. Neftel, V. Schneider, and N. Schneeberger, “European patent no. 1669100: Micro-needle,” 2006. [202] C. F. Shaw, Metals and Their Compounds in the Environment. Weinheim, Germany: VCH, 1991, pp. 931–8. [203] K. M. Halparin, “Epidermal ”turnover time” — a re-examination,” Br. J. Dermat., vol. 86, no. 1, pp. 14–9, 1972. GLOSSARY 81

Glossary

APC Antigen Presenting Cell: a cell that presents antigens to white blood cells in order to elicit an immune response.

ARDE Aspect Ratio Dependent Etching: an eﬀect in DRIE (and RIE) where the etch rate lowers at high aspect ratios (> 5:1) due to gas transport limitations. The eﬀect is also known as RIE-lag.

DRIE Deep Reactive Ion Etching: a plasma-based etch technique to etch anisotrop- ically and deep into a substrate.

FDA Food and Drug Administration: U.S. authority that regulates the U.S. market on food, drugs, cosmetics and medical devices.

MEMS Microelectromechanical Systems: a technology discipline of engineering at the micrometer length scale, possibly including electrical and mechanical features. Also referred to as Microsystem Technology (MST).

NDA New Drug Application: the application needed to get a new drug approved in the U.S. by the FDA.

OTC Over-the-counter: name for drugs that are sold without a prescription.

PCB Printed Circuit Board: an epoxy/glass ﬁber-based substrate with electrically conducting copper lines used to interconnect electronic components.

WHO World Health Organization: U.N. organ that acts as a coordinating authority on international public health.

Anisotropy exhibiting properties (as etch rate or skin elasticity) with diﬀerent values when measured in diﬀerent directions.

Aspect ratio the ratio of an object’s longer dimension to its shorter dimension, e.g. height to width.

Bolus a large dose of medication given at a single moment.

Bore a cylindrical hole made by or as if by boring. Used in this context for the hole in a hollow microneedle. 82 A Fully Integrated Microneedle-based Transdermal Drug Delivery System

Dendritic cells antigen presenting cells (APC) that have branched projections, thus covering a larger area. Dermis the main layer of the skin. Epidermis the outermost layer of the skin containing several distinct regions. The stratum corneum is the outermost layer of the epidermis. In vitro refers to testing in a tube, or generally, in a controlled environment outside a living organism. In vivo refers to action in or on living tissue. In-plane microneedle a microneedle fabricated parallel to the substrate. Intradermal in the skin. Isotropy exhibiting the same properties (for example etch rate) in all directions. Langerhans’ cells dendritic antigen presenting cells (APC) present in the epidermal layer of the skin and representing the body’s second line of defence. Lumen the cavity of a tubular organ or part. Used in this context for the hole in a hollow microneedle and synonymous with bore. Microneedle a needle considerbly smaller than a standard hypodermic needle, especially in terms of length. Out-of-plane microneedle a microneedle fabricated so that the needle protrude from the substrate. Stratum corneum the corniﬁed outermost layer of the epidermis. Subcutaneous under the skin. Transdermal patch a medicated patch (plaster) to be placed on the skin for subsequent drug release (e.g. a nicotine patch). Transdermal across the skin. A Fully Integrated Microneedle-based Transdermal Drug Delivery System

Paper reprints