development and validation of a novel bioreactor system ... · system for load- and...

ARTICLE

Development and Validation of a Novel BioreactorSystem for Load- and Perfusion-Controlled TissueEngineering of Chondrocyte-Constructs

Ronny M. Schulz,1 Nico Wustneck,1 Corrinus C. van Donkelaar,2

Julia C. Shelton,3 Augustinus Bader1

1Department of Cell Techniques and Applied Stem Cell Biology,

Center of Biotechnology and Biomedicine, University of Leipzig, Germany;

telephone: þ49-341-97-31352; fax: þ49-341-97-31359;

e-mail: [email protected] of Biomedical Engineering, Eindhoven University of Technology,

Eindhoven, The Netherlands3Medical Engineering Division, School of Engineering and Materials Science,

Queen Mary, University of London, London, UK

Received 31 January 2008; revision received 15 April 2008; accepted 22 April 2008

Published online 2 May 2008 in Wiley InterScience (www.interscience.wiley.com). DO
I 10.1002/bit.21955
ABSTRACT: Osteoarthritis is a severe socio-economical dis-ease, for which a suitable treatment modality does not exist.Tissue engineering of cartilage transplants is the most pro-mising method to treat focal cartilage defects. However,current culturing procedures do not yet meet the require-ments for clinical implementation. This article presents anovel bioreactor device for the functional tissue engineeringof articular cartilage which enables cyclic mechanical loadingcombined with medium perfusion over long periods of time,under controlled cultivation and stimulation conditionswhilst ensuring system sterility. The closed bioreactor con-sists of a small, perfused, autoclavable, twin chamber culturedevice with a contactless actuator for mechanical loading.Uni-axial loading is guided by externally applied magneticfields with real-time feedback-control from a platformload cell and an inductive proximity sensor. This precisemeasurement allows the development of the mechanicalproperties of the cultured tissue to be monitored in real-time. This is an essential step towards clinical implementa-tion, as it allows accounting for differences in the cultureprocedure induced by patient-variability. This articledescribes, based on standard agarose hydrogels of 3 mm

Correspondence to: R.M. Schulz

Contract grant sponsor: European Commission 5th Framework Programme

Contract grant number: GRD1-2000-25394

Contract grant sponsor: European Funds for Regional Development

Contract grant number: #4212/03-12

Contract grant sponsor: German Ministry of Education and Research

Contract grant number: 0313836

Contract grant sponsor: German Research Foundation

Contract grant number: BA 1025/2-1

Contract grant sponsor: formel.1 programme of the Medical Faculty of Leipzig

Contract grant number: #55/2005; #97/2007

714 Biotechnology and Bioengineering, Vol. 101, No. 4, November 1, 2008

height and 10 mm diameter, the technical concept, imple-mentation, scalability, reproducibility, precision, andthe calibration procedures of the whole bioreactor instru-ment. Particular attention is given to the contactless loadingsystem by which chondrocyte scaffolds can be compressedat defined loading frequencies and magnitudes, whilstmaintaining an aseptic cultivation procedure. In a ‘‘proofof principle’’ experiment, chondrocyte seeded agarosegels were cultured for 21 days in the bioreactor system.Intermittent medium perfusion at a steady flow rate(0.5 mL/min) was applied. Sterility and cell viability (ds-DNA quantification and fluorometric live/dead staining)were preserved in the system. Flow induced shear stressstimulated sGAG (sulfated glycosaminoglycan) content(DMMB assay) after 21 days, which was confirmed byhistological staining of Alcian blue and by immunostainingof Aggrecan. Experimental data on mechanotransductionand long-term studies on the beneficial effects of combinedperfusion and different mechanical loading patterns onchondrocyte seeded scaffolds will be published separately.

Biotechnol. Bioeng. 2008;101: 714–728.

� 2008 Wiley Periodicals, Inc.

KEYWORDS: chondrocytes; cartilage; bioreactor; tissueengineering; biomedical engineering; physical stimulation

Introduction

Articular cartilage has a high load bearing capacity(Cohen et al., 1998) and a smooth joint surface with lowfriction due to its structure and composition (Buckwalter

� 2008 Wiley Periodicals, Inc.

and Mankin, 1998a). Native articular cartilage contains ahighly organized extracellular matrix (ECM) (Buckwalteret al., 1988) which is maintained by chondrocytes, theonly cell-type in cartilage (Buckwalter and Mankin, 1998b).Due to the limited regenerative capacity of articularcartilage, damage is often progressive. Therefore, theprevalence of osteoarthritis is high. Due to the socio-economical impact (Buckwalter, 2002), there is a pressingrequirement for innovative strategies to repair damagedhyaline cartilage (Hunziker, 2002). One promising approachis to replace damaged cartilage with tissue engineeredgrafts (Langer, 2000; Langer and Vacanti, 1993; Lindahlet al., 2001).

To create such grafts involves the in vitro cultivation ofchondrocytes (Kuo et al., 2006; Marlovits et al., 2006;Nesic et al., 2006) in a suitable matrix material (Behrenset al., 2006) in an appropriate culture system (Martin et al.,2004). These chondrocyte grafts in the shape of matrix-coupled autologous cartilage transplants (MACT) must bebiocompatible and have appropriate mechanical propertiesin order to be clinically useful (Andereya et al., 2006;Marcacci et al., 2005). Many approaches are now beingtaken to enhance the quality of these cartilage substitutesthrough the application of bioreactor devices with differentworking principles (Donkelaar and Schulz, 2008; Schulzand Bader, 2007). Bioreactor technologies should beautomated to become of economical interest, allow upscaling, ensure reliable, reproducible outcomes with lowcontamination risk and high product safety. Such auto-mated systems should be able to monitor, control andregulate operational (e.g., media supplementation, nutrientsupply, oxygen tension), mechanical, biochemical andbiophysical (e.g., pH, exchange of gases, humidity, andtemperature) environmental conditions, and perform as-eptically several bioprocess operations (e.g., scaffold seeding,feeding, and sampling). Ultimately, clinically used auto-mated bioreactors are equipped with feedback controlledoperation procedures to cope with expected patient-variability. The specific requirements for an extracorporealcultivation system can be distilled from the results of thepioneering experiments, as recently reviewed (Schulz andBader, 2007).

In addition to the above-mentioned global biotechno-logical needs, two key requirements for three-dimensional(3D) cartilage tissue engineering include the ability formechanical loading and the delivery of nutrients to andremoval of waste products. Application of mechanicalloading to chondrocytes is essential for developing cartilageconstructs with sufficient mechanical properties (LeBaronand Athanasiou, 2000). Dynamic mechanical stimulationof chondrocytes in vitro can modulate chondrocytemetabolism, improve ECM synthesis, prevent loss ofchondrocytic phenotype (Chowdhury et al., 2003; Davissonet al., 2002a; Lima et al., 2006; Mauck et al., 2000, 2003;Risbud and Sittinger, 2002), and enhance the load bearingcapacity of the graft (Kelly et al., 2006). Chondrocytesrapidly dedifferentiate in static culture and cease to

synthesize the appropriate matrix in the absence of ap-propriate mechanical loading (Domm et al., 2002; Vunjak-Novakovic et al., 1999).

Active tissue perfusion is often considered to ensure theconstant supply of nutrients and removal of waste products(Fermor et al., 2005; Galban and Locke, 1999; Grimshaw andMason, 2001; Hofstaetter et al., 2005; Sengers et al., 2005).Another beneficial effect of perfusion is that the fluid-flowinduced shear stress stimulates the expression of cartilagespecific matrix markers (Darling and Athanasiou, 2003;Heath and Magari, 1996).

A number of cultivation systems are geared towardsapplying mechanical stimuli by compressive loading(Buschmann et al., 1995; Cassino et al., 2007; Davissonet al., 2002a; Demarteau et al., 2003; Lee and Bader, 1997;Mauck et al., 2002; Torzilli et al., 1997), bending or tensilestress (De Witt et al., 1984; Fukuda et al., 1997; Millward-Sadler et al., 2000; Wright et al., 1997), shear stress(Frank et al., 2000; Jin et al., 2001; Waldman et al., 2003), orhydrostatic pressure (Carver and Heath, 1999a; Ikenoueet al., 2003; Parkkinen et al., 1993). Others have developedfluid flow culture systems that are intended to provide theappropriate biochemical environment, via direct mediumperfusion through cell seeded scaffolds (Davisson et al.,2002b; Kim et al., 1994; Pazzano et al., 2000; Sittinger et al.,1994; Wendt et al., 2003), by inducing fluid motion aroundthe scaffolds (e.g., using spinner flasks (Freed et al., 1994;Vunjak-Novakovic et al., 1999), orbital shakers (Martinet al., 1998), or rotating wall vessels (Freed et al., 1998;Obradovic et al., 2001), or which enable the constructs tobe transferred between separate perfusion and loadingsystems (Carver and Heath, 1999b). Surprisingly, only twodevices (Demarteau et al., 2003; Mizuno et al., 2002) allowcartilage substitutes to be cultivated in an environment inwhich both fluid-handling and compressive loading arecontrolled.

Once the above conditions are met, the final require-ment before clinical application is feasible is to make theculture system compliant to cGMP regulations. Ideally,chondrocytes would be aseptically delivered to thedecentralized production system with the bioreactor thenremaining completely closed, and they would be processedthrough the entire tissue manufacturing (cell isolation,scaffold seeding or inoculation, sampling, proliferationand differentiation phases) according to cGMP regulationsusing a respective quality control/quality assurance (QC/QA) program until implantation when the resultantcartilage would be aseptically transferred from the deviceto the recipient within the operating theatre (Martin et al.,2004). During the past few years biomedical companies,for example, Aastrom Biosciences (Armstrong et al., 1997),Millenium Biologix (Smith et al., 2005), and Olympus(Hibino et al., 2005) have claimed and developed theso-called cartridge-based bioreactor systems for the cGMP-compliant manufacture of mainly autologous chondrocytetransplants as a tissue-engineered product for therapeuticuse in osteoarthritic patients (Donkelaar and Schulz, 2008;

Schulz et al.: Bioreactor System for Bioengineered Cartilage Grafts 715

Biotechnology and Bioengineering

Naughton, 2002). However, to the best of our knowledge,none of these industrial concepts allows mechanicalloading. Probably the most important reason for thisdrawback is that current concepts of load transmissionare based on pistons that are attached to non-sterileactuators on the one side, and connect to the sterilecultivation area on the other side. This is a potentialcontamination risk contradicting the present guidelines ofGMP.

We set out to develop a new bioreactor system thatallows controlled mechanical loading and simultaneousperfusion to be applied in a closed aseptic bioreactorequipped with an accurate, integrated mechanical loadingdevice which guarantees for product safety and systemsterility. These are the principal requirements to develop asystem that operates under GMP conditions. This bioreactoris to be used for the tissue engineering of articular cartilageequivalents in both research and clinical applications.The purpose of this technical article is to describe theresulting novel bioreactor design, which uses a magneticactuator for contact-free load transmission, controlled by anintegrated sensing concept for monitoring applied forceand construct deformation in real-time. This allows theuse of advanced feedback control for a fully automatedstimulation process under cGMP conditions. The techno-logical developments, calibration procedures and validationexperiments were specifically designed to expose a modelscaffold of chondrocyte-seeded agarose gels (3% (w/w);outer diameter (OD) of 10 mm, and height (H) of 3 mm) toa controlled regime of compressive deformation. Theapplicability of the bioreactor was evaluated in a 3-weekexperiment in which porcine articular chondrocytes in a3D agarose hydrogel are subjected to forced perfusion.Cell viability and deposition of extracellular matrix werecompared between perfused and control samples. Thebiological ‘‘proof of principle’’ of the complete bioreactorsystem in experiments on mechanotransduction and inlong-term studies on the beneficial effects of combined

Figure 1. Mechanical drawing (A), isometric sketch (B) and image of the bioreactor

in diameter (C). Annotations: I¼ vessel, II¼ lid, III¼ constructs, IV¼ locating stage, V¼


perfusion and mechanical loading lie in the scope ofsubsequent articles.

Materials and Methods

Bioreactor Device

Bioreactor Construction

The bioreactor (Fig. 1) is constructed from a 32 mm long,28 mm diameter polycarbonate (Makrolon, Bayer AG,Leverkusen, Germany) cylinder (I). A 30 mm long and15 mm diameter hole was centrally drilled. In the remaining2.0 mm depth, a 1.5 mm central hole of 7 mm diameter wasdrilled, which will be referred to as the lower chamber. Theremaining 0.5 mm polycarbonate wall is highly polished inorder to serve as an effective (98%) transparent surface toexamine constructs inside the bioreactor with an invertedmicroscope. The vessel (I) is closed with a cylindricalpolycarbonate lid (II) and a 1.5 mm Viton O-ring. Theinside of the lid contains a central hole 10 mm in diameterand 10.5 mm in depth to fit the mini-actuator (V).Constructs (III) are precisely located in the bioreactor withthe help of a perforated locating stage (IV) such that they arecentrally positioned and with the entire lower surface of theconstruct being situated 0.75 mm above the collar, coveringthe lower chamber. The inner wall of the location stagelimits the movement of the actuator such that the appliedcompressive strain by the mini-actuator (V) cannot exceed apredefined magnitude. The construct and the locating stagedivide the bioreactor into two chambers. The narrowerlower chamber serves as a pre-mixing compartment.It contains two diametrically opposed ports (1 mm OD,0.7 mm inner diameter (ID)) which can serve as fluid in- oroutlet. Two additional ports are located on the circumfer-ential wall of the upper chamber. One meets thecircumferential inner wall tangentially 6.5 mm above thecollar, the other is located radially 12.5 mm above the collar.

seeding procedure with chondrocyte-agarose constructs of 3 mm height and 10 mm

mini actuator.

Mini Actuator and Loading Plate

A cylindrical neodymium iron boron (NdFeB) magnet(Fig. 2A) (4 mm OD, 25 mm length) (S-04-25-N, WebcraftGmbH, Uster, Switzerland) is placed in a stainless steel tube(5 mm OD, 30.3 mm length) that is located inside thebioreactor. A PTFE cuff (5 mm ID, 10 mm OD, 20 mmlength) surrounds this cylinder at one end which is theninserted in the lid as described above. The opposite sideof the tube holds a stainless steel construct loading plate(15 mm OD, 1 mm thick). This plate is perforated with aseries of 0.5 mm diameter holes to enable medium transfer.The lower surface of this plate which is in contact with thetissue engineering construct is coated with an ultra-thinlayer of PTFE to prevent adhesion. This construction allowsfor a vertical displacement of the entire stainless steel shaft of0.9 mm, that is, its position varies between 0.15 mm above a3.0 mm cartilage construct, and can induce a maximumcompression of 0.75 mm, or 25% axial compression. Asmentioned before, the location stage can be used to limit themaximal imposable strain. The weight of the entire miniactuator is 7.5 g, the PTFE cuff weighs 2.6 g. The friction-coefficient (m) between PTFE and stainless steel is only0.04 under dry conditions, and less when lubricated. Evenwhen the horizontal force Fn reaches half the applied verticalstimulation force FS, the maximal frictional force Ff willreach maximal 2% of FS.

Ff ¼ mFn ¼ 0:041

2FS ¼ 0:02Fs (1)

Figure 2. Mechanical drawing (top row) and picture (bottom row) of the special mag

based on three small NE105 (B) magnets in a ferromagnetic iron assembly and two larger N

N (north) and S (south).

Actuating and Sensing Elements

Control Magnet System

The vertical motion of the stainless steel mini actuator withthe loading plate is controlled by a magnetic field inducedby external 1.1 Tesla NdFeB permanent magnets. Cyclicloading is applied by alternating the magnet orientationabove the bioreactor. Two rotating magnet holders wereconstructed, one of ferromagnetic iron housing three NE105control magnets (10 mm OD, 5 mm H) (Fig. 2B) and one ofparamagnetic aluminum housing two larger NE201 magnets(20 mm OD, 10 mm H) (Fig. 2C) (both IBS Magnet, Berlin,Germany). The orientation of the magnets is controlled by aDC miniature servo motor cell (6 V, 1.5 W) with a planetarydrive (available reduction ratios: 369:1, 84:1, 19:1) GP16 A,controlled by a power amplifier LSC 30/2 (all maxon motorGmbH, Sexau, Germany) connected to two controller areanetwork (CAN) open modules (CAN CBM DIO8 and CANCBM A04, esd GmbH), for regulating power on/off androtation velocity, respectively. This enables constantrotation of the magnet system for cyclic compression aswell as dynamic regulation of the rotation to imposecomprehensive loading patterns. Loading by the magnetscan be controlled by adjusting the distance between therotating magnet suspension and the mini-actuator. Anautomated force control system is implemented whichadjusts the position of the magnets by means of an electricalmini-slide SLTE-10-50-LS-G04 (stroke of 50 mm, stepwidth of <0.1 mm, repeat accuracy �0.05 mm) connected

netic loading actuator (A), and both embodiments of the control magnet suspensions

E201 (C) magnets in paramagnetic aluminum. The magnet polarization is indicated as



to the SFC-DC-VC-3-E-HO-IO controller with an imple-mented CANopen interface (both Festo AG & Co. KG,Esslingen, Germany).

Inductive Proximity Sensor SPos andPlatform Load Cell Sensor SForce

The aim of tissue engineering is to improve tissue propertieswith cultivation time (Kelly et al., 2006; Mauck et al., 2002).The possibility of monitoring the change of tissue propertiesin time would therefore be extremely valuable. It is possibleto derive this information from the relation between appliedload and associated strain. Dynamic protocols such as thoseused during tissue engineering are particularly suitable as theycan provide information on both the elastic and the viscousbehavior of the tissue. To enable this requires componentsto independently monitor the position of the loading plateand imposed load. The position of the loading plate in thebioreactor is detected in a contactless manner by an inductive,analog displacement sensor (BAW-M18MG-UAC80F, BalluffGmbH, Neuhausen, Germany), positioned below the bio-reactor. The sensor reports distances between 2 and 8 mm(repeat accuracy of �0.012 mm, response time of 1.5 ms)with the metal loading plate. The imposed load is determinedusing a platform load cell (SForce) PW4KRC3-MR (HBMGmbH, Darmstadt, Germany), maximal load 5 or 30 N,situated between the bioreactor and the fixed holdingstructure. Analog output signals of the displacement andforce sensors are digitized using CANopen analog inputmodules (CAN-CBM-AI4, esd GmbH, Germany).

Complete Bioreactor System

Figure 3A illustrates how the described bioreactor compo-nents are assembled. The bioreactor (BR) and proximitysensor (SPos) are coupled to the fixed platform load cell(SForce). The rotating magnet system which determines themechanical stimulation frequency (ASF) is directly linked to

Figure 3. The developed bioreactor system is composed of actuating and sensing co

cultivation device in a contact-free manner. The illustration depicts the concept of the es

measuring station for an individual bioreactor (B). The technical sketch depicts the adapta

were implemented into an individual perfusion environment (D). The models shown here

BR¼ bioreactor, AVD ¼ actuator vertical distance, ASF ¼ actuator stimulation frequency, S

can be seen in the online version of this article, available at www.interscience.wiley.com


the actor for its vertical positioning to control loadingmagnitude (AVD). This setup uncouples the actuatingsystem from the sensory system. In practice, the system canbe an individual bioreactor (Fig. 3B), or one with severalsystems running in parallel.

Figure 3C shows a station with six bioreactors where eachculture device is directly connected to its own individualinductive distance sensor (SPos) that itself is fixed to therack via a single load cell (SForce) with nominal load 30 N.The six control magnet suspensions are connected to anadjustable bar which is belt-driven by one motor. Bothactors (AVD, ASF) were directly connected to the rack, andthus did not affect the sensors. This principle minimizes theamount of sensor and actor components, thus increasingcost-effectiveness whilst allowing for appropriate monitor-ing of the mechanical conditioning of, for example, sixhydrogels descending from one batch. Figure 3D shows anexample of a station where six separated bioreactors areintegrated, and with a separate perfusion periphery con-sisting of its own pump, reservoir, tubing, and sterile filter.This particular station is used in the ‘‘proof of principle’’perfusion experiment explained below.

Control Program

Magnet rotation velocity (ASF) and position (AVD) control areautomated based on the output signals from the platformload cell (SForce) and the inductive proximity sensor (SPos).Figure 4 provides a block diagram of the closed-loopcontrol system. The control unit, a CBS637 embedded system(Cogent, Computer Systems Inc., Smithfield, RI), works as aCANopen node through a USB/CAN converter (PCAN USB-ISO, PEAK System Technik GmbH, Darmstadt, Germany).The control software provides continuous feedback of theprocess and allows at any time to adjust parameters of theotherwise fully automated process. Additional sensorsor actuators, for example pumps, flow sensors and theincubator, can easily be integrated in the system.

mponents for control of mechanical loading of three-dimensional constructs inside the

tablished control scheme (A) whereas the image shows the technical realization of a

tion of this monitoring principle into a rack for six separated bioreactors (C) that itself

are equipped with a manual motorized height adjuster. Abbreviations in this figure:

Pos ¼ sensor position measurement, SForce ¼ sensor force measurement. [Color figure

.]

Figure 4. The developed bioreactor control system was designed for a fully automated mechanical loading process. The graphics illustrate the modular hardware circuit

diagram (A) and the concept of process automatization (B) from the individual actuating and sensing elements to the user interface (Lintouch client). A screenshot (C) of this HMI

client depicts the control panel and the real-time data visualization of the applied force, displacement and direct perfusion flow rate to the cartilage construct in a diagram. TCP/

IP¼ transmission control protocol/Internet protocol, IPC¼ interprocess communication, CAN¼ controller area network, HMI¼ human machine interface. [Color figure can be

seen in the online version of this article, available at www.interscience.wiley.com.]

Sensor Calibration

Calibration experiments with the load (SForce) and positionsensor (SPos) focused on accuracy, reproducibility, andpotential zero point drift. Following standard calibration,the load cell was calibrated in the bioreactor system bypositioning standard weights (varying between 5 and 500 g)onto the locating stage or into the cultivation chamber of thebioreactor. Linear regression analyses between applied andmeasured loads correlate well (r2 ¼ 0.9989); the standarddeviation of force variations from the linear fit shows arelative full-scale error of 0.48%.

Similarly, the inductive proximity sensor SPos wascalibrated on the assembled bioreactor system. Using amicrometer head (164–161, Mitutoyo Messgerate GmbH,Neuss, Germany), with a measurement range of 0–50 mmand accuracy of 1.0 mm, the bioreactor was displaced severalmicrometers at a time over the whole range of 5–8 mm,while the positioning of the system was recorded with theinductive proximity sensor. The imposed and measuredpositions, averaged over eight measurements, correlatedlinearly (r2¼ 0.9963), with a relative full-scale error of1.04% (data not shown).

Prior to an experiment, the position sensor was calibratedusing the lower and upper reversal points of the actuator. Atthe lower reversal point it touched the inner wall of thelocating stage, which has an exactly known height. Theupper point equals the top of the container.

Subsequently, a 3% agarose construct (10 mm diameter,3 mm high) was installed and the bioreactor was filled withculture medium. By adjusting the vertical position of themagnet and using feedback by the force-sensor, the actuatorplate was positioned such that it was freely floating in theculture medium without touching the construct. Under thiscondition the force sensor was zeroed. Finally, the magnets

were lowered until the force sensor was loaded 0.02 N.Sample height was now determined from the actual positionand the previous position where the actuator touched theinner wall of the locating stage. The displacement sensor waszeroed at sample height.

Assessment of Drift

The fully calibrated system was permitted to run con-tinuously without process control intervention in order toquantify drift. Software instructions were issued withthe following input parameters: Maximal compressive strain15% s15 (0.45 mm), compressive load frequency 1 Hz,cyclic load duration 120 min, rest time 120 min. This load-rest cycle was run for 2 days followed by a rest period of2 days. This was done twice under a constant environment(temperature, relative humidity). No signs of drift of bothSForce and SPos sensors were observed after these 8 days.

Perfusion Experiment

Six parallel bioreactor devices are integrated in one system,each with its own closed loop perfusion circuit containing anindividual medium container, pump, flow sensor and sterilefilters (see Fig. 3C and D). Medium is delivered to the lowercompartment of each bioreactor via one of the inflow portsconnected to a 20 mL medium reservoir via gas permeablesilicone tubing (inner diameter 1 mm). A micro annulargear pump (MZR 2521) and a control unit (S-KD, bothHNP Mikrosysteme GmbH, Parchim, Germany) are used toregulate medium flow. The pump also serves as a deadvolume-free, self-closing valve. The flow rate is set at 0.5 mL/min, as monitored using a CMOS thermal mass flow sensor(ALS1430-24, Sensirion AG, Staefa, Switzerland). Medium



exiting the bioreactor is returned to the reservoir via thesame type of tubing through the radial outlet port in theupper chamber. The medium in the reservoir is oxygenated(95% air, 5% CO2) by means of additional tubing con-nection and a simple gassing pump. The entire systemis placed in an incubator (Thermo Electron, GmbH,Oberhausen, Germany) which records temperature, relativehumidity, carbon dioxide and oxygen partial pressures (seeFig. 4C).

Cell Isolation, Construct Preparation, andCulture Conditions

Unless otherwise stated chemicals were obtained fromSigma–Aldrich (Taufkirchen, Germany). Chondrocyteswere isolated and seeded into agarose hydrogels aspreviously described (Schulz et al., 2006). Briefly, articularcartilage samples were aseptically dissected from themetacarpophalangeal joint from porcine feet (n¼ 3)obtained from the local abattoir. The isolation of thechondrocytes was performed by mechanically mincing thecartilage samples, followed by enzymatic digestion with2 mg/mL Collagenase A in Dulbecco’s modified Eagle’smedium (DMEM) (Biochrom KG, Berlin, Germany) at378C for 24 h. After enzymatic isolation, the primary porcinechondrocytes were counted and the viability was determinedusing Trypan-blue. Viability greater than 95% was usuallyachieved. Base medium containing sodium pyruvate(110 mg/L), L-glutamine (580 mg/L), fetal bovine serum(10%) (Invitrogen, Karlsruhe, Germany), ascorbic acid(0.05 mg/mL) and gentamycin (50 mg/mL; PAA Labora-tories, Pasching, Austria) was used for the culturing. For cellseeding, 1 mL of chondrocyte suspension (3� 107 viablecells) was thoroughly mixed with 10 mL of 3.3% agarosetype VII in a pipette. Individual 250 mL aliquots of theagarose chondrocyte suspensions (containing 0.75� 106

primary cells) were seeded into individual wells of a mouldwith a thickness of 3 mm and a diameter of 10 mm. After10 min of gelling scaffolds were transferred into separatewells, filled with 2 mL medium, and incubated under static,free swelling conditions for 24 h at 378C, 5% CO2. Theconstructs were then either placed in the bioreactor to besubjected to perfusion or maintained as controls in Petridishes with the identical medium volume (20 mL) as used inthe perfusion system. For 21 days, an hourly intermittentperfusion regime was applied in the bioreactor with a flowrate of 0.5 mL/min for 10 min followed by a rest period of50 min. The chambers were disconnected at days 7 and 21, toaseptically remove the constructs for subsequent analysis.Three independent experiments were carried out.

Cell Viability and Cell Content

Cell viability in the agarose gels was assessed using a calceinAM (green) and ethidiumhomodimer (red) fluorometricstaining kit (Mobitec, Gottingen, Germany). The number ofchondrocytes per construct was indirectly determined by


quantification of DNA content from aliquots of papaindigests of the hydrogels using the ds-DNA Quantitations Kit(Mobitec). To obtain papain digests, the agarose specimenswere mixed with 750 mL papain digestion buffer (5 mML-cystein, 5 mM EDTA, 100 mM Na2HPO4) and melted at708C for 1 h, cooled down to 458C, followed by a terminal168h digestion step in 5 mL papain-solution (50 mg) and10 mL agarase-solution (1.66 U). Afterwards, all sampleswere stored at �208C until analysis. After preparation of aLambda-DNA-standard in a range of 0.02–2 mg, the 100 mLsamples were incubated with 100 mL PicoGreen dye for5 min and measured in a fluorescence spectrometer at480 nm excitation and 520 nm emission. To convertbetween DNA content and cell number, 7.7 pg DNA perchondrocyte was assumed.

Determination of Sulfated Glycosaminoglycan Content

Samples of each of the papain digests were also analyzed forthe proteoglycan content by quantifying the sGAG contentusing the DMMB (1,9 dimethyl-methyleneblue) dye bind-ing assay (Roche, Basel, Switzerland). The absorbance wasdetermined in a photometer at 595 nm, the concentration ofsGAG was extrapolated from a master curve based onchondroitin sulfate with a range of 10–100 mg/mL. Allsamples and standards were done in duplicate. sGAG wasmeasured in mg/mL and reported as weight per DNA orweight per wet weight of construct.

Histological Analysis and Immunocytochemistry

Agarose gels were shock frozen, embedded in Tissuetek andsliced with a cryotom (both Leica, Bensheim, Germany).The 6 mm thick sections were processed for histologicalanalysis with Alcian blue staining to verify proteoglycanaccumulation. Immunostaining of aggrecan was performedaccording to a two step indirect method. The sections weretreated with monoclonal mouse antibody against aggrecan(1:200; DPC Biermann, Bad Nauheim, Germany). Afterwashing with PBS, the secondary antibody of PODconjugated goat anti-mouse IgG (1:50; Jackson ImmunoResearch, Cambridgeshire, UK) was added for 1 h at 378C.Immunostaining used AEC (3-amino-9-ethyl-carbazol)substrate.

Statistical Analyses

All results were expressed as mean� standard error of themean of at least three separate experiments (n¼ 3) andanalyzed with statistical software (Origin, Friedrichsdorf,Germany). The data from different animals are pooledtogether. For pairwise comparisons between the cultureconditions collected data were analyzed by one-way analysisof variance (ANOVA) using the Bonferroni method andstatistical significance was accepted at P< 0.05.

Results

Maximum Stimulation Force, Frequency, Strain Rate,and Sample Geometry

The determination of maximal applicable uni-axial forces totissue grafts within this bioreactor construction depends onthe magnetic flux density B(s) of the utilized magnets andtheir material properties, that is, remanence and geometry asformulated in Equation (2) and illustrated in Figure 5A.

BðsÞ ¼ Br

2

L þ sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiR2 þ ðL þ sÞ2

q � sffiffiffiffiffiffiffiffiffiffiffiffiffiffiR2 þ s2

p

264

375 (2)

Figure 5B shows the resultant metered values ofcompressive forces F (N) or rather pressures p (kPa)depending on the distance s (mm) between two variants ofencapsulated loading plate magnets and both variants ofexternally arranged magnet suspension. For the standardagarose gel with a diameter of 10 mm a force of 1 Ncorresponds to a pressure of 12.7 kPa. The cylindrical discmagnet S-04-25 in the actuator equipped with NE105 andNE201 external control magnets can deliver forces between0 N and maximum forces of 2.0 and 2.8 N, depending onthe vertical distance (AVD) between the magnets and theactuator. Due to the high proportional influence of magnetgeometry on the resulting repulsion force another mini-actuator prototype containing a larger (NE105) magnet wasutilized in the measuring station. The determined maximumrepulsion forces against both control magnets, NE105 andNE201, were found at 7.1 and 9.6 N, respectively. Theachievable maximum uni-axial stimulation force, that is,minimum proximity, is spatially limited by the controlmagnet suspension at 2 mm (NE 105) and 3 mm (NE201),respectively.

Figure 5. Magnetic flux density B(s) (mT) on the axis of the implemented NdFeB cir

pressures p (kPa) generated between the encapsulated magnet of the mini actuator and both

dependence (B). Both continuous lines were measured for the actuator containing the s

when NE105 was encapsulated into the PTFE cuff of the actuator. The pressure p is indicated

(10 mm OD).

Mechanical stimulation frequency is defined by therotation velocity of the magnet disc which ranges between 0and 6.66 rotations per second, and the number of magnets inthe disc, that is, 2 for NE201 and 3 for NE105. Hence, forthe NE105 magnet system, maximal stimulation frequencyequals 6.66� 3¼ 20.0 Hz. Rotation velocity of the controlmagnet system (ASF) is applied by a DC miniature servomotor in combination with one of three fitting planetarygear heads GP16A, which have specific reduction ratios of369:1, 84:1, and 19:1, enabling 0.33, 1.4, and 6.66 revolutionsper second under maximum power, respectively.

The bioreactor was used to stimulate individual hydrogelswith a construct dimension of 3 mm thickness and adiameter of 10 mm since this geometry is widely used instudies dedicated to investigating chondrocyte mechano-biology (Chowdhury et al., 2003; Lee and Bader, 1997;Mauck et al., 2003). The modular setup of the system and thescalability of the components allowed the existing vessel, lid,locating stage, and mini actuator to be easily adapted tothese agarose construct sizes. Such adaptations can be madeat any time, depending on for instance the geometry of theconstruct or the actuator magnet, without the necessity tomodify the core bioreactor setup.

Example Stimulation Protocol

Force and position were recorded for the 3% agaroseconstruct in the bioreactor, with three- (Fig. 6A) and two-magnet embodiments (Fig. 6B), rotating with uniformvelocity of 0.33 and 0.5 Hz, respectively, which leads to aperiodic loading time of 1 s. The recorded curves forthe NE105 magnet arrangement show that at repulsivestimulation forces of 2 N the agarose construct reachesthe maximum displacement of 450 mm (i.e., 15% strain)that is allowed by the inserts at a stimulation frequency of1 Hz. The compression profiles differ between the two- and

cular disc magnets at a distance s (mm) (A). The forces F (N) of repulsion along with

types of external control magnets (NE105 and NE201) show a typically strong proximal

tandard circular disc magnet (S-04-25-N) whereas both dotted lines were metered

for a surface area of 78.54 mm2 which corresponds to the geometry of the agarose gels



Figure 6. The diagram shows the values of force and position as recorded in real-time with the three NE105 (A) or two NE201 (B) magnets. The vertical distance between

the bioreactor device and both external permanent magnet arrangements was adjusted to reach a repulsive force of 2 N to a standard agarose gel. The recorded force profile

(gray line), along with the corresponding actual position of the loading plate (black line), as well as the theoretical position profile without a hydrogel in the device (dotted)

are shown.

three-magnets systems, wherein the two NE205 magnetsgenerate a more trapezoidal profile.

The dotted line in these figures shows the displacementof the actuator without the presence of an agarose construct.The difference between these lines is determined bythe properties of the tissue under evaluation. In fact, therecorded curves can be displayed as a stress–strain curve(Fig. 7) from which material properties of the constructcan be derived. Differences in stiffness (slope), nonlinearbehavior (increasing slope with strain) and viscoelasticity(hysteresis) between gels with different percentages ofagarose are apparent from the example (Fig. 7).

Figure 7. Stress–strain curves for standard hydrogels (10 mm OD, 3 mm H)

made of three agarose concentrations (2.5%, 3.0%, and 3.5%) recorded in real-time by

the control system of the bioreactor device. A maximal vertical force of 1.7 N

(s¼ 21.25 kPa) was applied 5 times at a stimulation frequency of 1 Hz to each type

of agarose gels.


The different material properties used in this experimentimitate the well-known fact that dynamic deformationalloading significantly increases the Poisson’s ratio (Kellyet al., 2006) and Young’s modulus (Lima et al., 2006; Maucket al., 2002) of chondrocyte-seeded agarose hydrogels inlong-term studies.

The metered stress–strain curves show the exemplarycourse of five complete loading and unloading cycles.Obviously, during the tissue engineering process, thecartilage tissue develops and as a result the materialproperties increase with time. This is represented byincreasing slopes of these curves. Hence, this bioreactorsystem allows the development of the mechanical propertiesof tissue engineered cartilage to be monitored in real-time.

Cell Viability, Cell Content, and sGAG Accumulationin Static and Perfused Constructs

Figure 8 depicts the microscopic images of the fluorometricviability kit and the staining for cartilage specific extra-cellular matrix components in agarose constructs assessedby Alcian blue histology and by immunostaining forproteoglycan and aggrecan at 1 and 3 weeks of static wellplate culture and perfusion culture in the novel bioreactorsystem, respectively.

The live/dead images of Figure 8 (top row) demonstratethat viable primary porcine chondrocytes were dispersedthroughout the specimens and displayed a phenotypicspherical morphology under the absence of flattened cellsin both culture conditions at both examination days.Furthermore, nearly 100% of the chondrocytes survivedafter 3 weeks in all specimens. Thus, increased cellapoptosis—for example, caused by system non-sterility—was not detectable in samples that were cultured in well

Figure 8. Cell viability, histological and immunohistochemical evaluation of chondrocyte seeded 3% agarose constructs that were cultivated as control group in standard

well plates (columns 1 and 3) or were perfused in the bioreactor device (columns 2 and 4) and subsequently examined at days 7 and 21. Analysis of chondrocyte viability in agarose

gels as determined by fluorometric viability assay (top row, A–D, 50�). Assessment of cartilage specific matrix markers by histological staining with Alcian blue (middle row, E–H,

200�) and by immunostaining of the most abundant proteoglycan Aggrecan (lower row, J–M, 100�) in cryosections of hydrogels. [Color figure can be seen in the online version of

this article, available at www.interscience.wiley.com.]

plates as well as in the bioreactor device with the outlinedperfusion environment.

The relative changes of cell numbers in the hydrogels forthe respective culture condition at both examination days (see

Table I. Results of the biochemical assays for the chondrocyte-seeded agarose c

culture in the bioreactor.

7 days in vitro

Well plate Biore

Cell number 761.000� 29.000y 1.156.000

sGAG (mg) 91.14� 11.80y 91.51

sGAG/wet weight (mg/mg) 0.38� 0.05y 0.38

sGAG/DNA (mg/mg) 15.16� 2.18,y 10.02

Asterisks () indicate P< 0.05 versus other culture condition at the same timethe same culture condition.

Values are given as mean� standard error of the mean (SEM).

Table I) were obtained by quantifying the ds-DNA contentand the cell amount showed to be 1.52 times in perfusedagarose constructs after 7 days, whereas dynamic cultured gelshad fewer cells after 3 weeks in vitro when compared to

onstructs at 1 and 3 weeks of static well plate culture and dynamic perfusion

21 days in vitro

actor Well plate Bioreactor

� 72.000 1.495.000� 216.000 1.248.000� 172.000

� 8.92y 242.08� 24.87 384.55� 32.02

� 0.05y 1.01� 0.10 1.60� 0.13

� 1.3y 20.50� 1.78 39.00� 8.03

point, and daggers (y) at day 7 indicate P< 0.05 versus the day 21 value for



hydrogels of well plate control. Static cultured constructsexhibited no increase in DNA content by day 7, followed bya substantial increase (97%) in DNA content between weeks1 and 3, resulting in a net increase superior to that seen in theother cultivation condition. Perfused agarose constructs hadgreatly significant increases in DNA content by day 7 (54%)followed by a slightly insignificant increase (8%) during thesubsequent 2 weeks. At day 21, agarose constructs of staticand perfused culture had reached approximately 2.0 and1.66 times the original seeding density, with statisticallysignificant differences among these scaffolds.

No differences were observed in the distribution of thecells or in the deposition of articular cartilage specific matrixmarkers between Alcian blue (middle row, Fig. 8) andAggrecan immunostaining (lower row, Fig. 8) in specimenstaken after 7 days culture. Both construct types containeduniformly distributed cells surrounded by a less intensepericellular matrix. Specimens of both culture modelsacquired after 21 days displayed more intense matrixstaining surrounding single chondrocytes and cell clusters.However, enhanced staining of pericellular matrix compo-nents was observed in cryosections of perfused agarose gels.

To determine the efficiency of both culture models themeasured total accumulated sGAG was expressed per ds-DNA content and construct wet weight (see Table I). Theabsolute sGAG content in static and perfused agarose gels wassimilar after 7 days, while significantly increased sGAG valueswere detected in hydrodynamic stimulated specimens after3 weeks when compared to static constructs. For static andperfused cultures, sGAG deposition substantially increasedbetween weeks 1 and 3 2.66- and 4.2 fold, respectively. Atweek 3, perfused agarose gels exhibit 1.59 times higherabsolute sGAG content than well plate control constructs.Due to equal agarose seeding volumes in both groups thenormalization of sGAG against wet weight demonstratedsimilar ratios for both conditions after 1 and 3 weeks.

The patterns of matrix accumulation and cell contentchanges resulted in small differences in sGAG/DNA ratioamong cultivation conditions at both time points. Similarto the pattern of sGAG deposition, the relation of sGAG/DNA demonstrates for both cultivation modes a continuedsubstantial increase in accumulated extracellular matrixbetween days 7 and 21. At day 7, static agarose gels hadsignificantly higher sGAG/DNA than the perfusion group,while at day 21 perfused hydrogels had a substantiallyincreased sGAG/DNA ratio. During the terminal 2 weeksof static and perfused cultivation the sGAG/DNA ratioincreased 1.35- and 3.89-fold, respectively. At the finalexamination time point gels of the perfusion group had a1.90 times higher sGAG/DNA content than the staticallycultured well plate gels.

Discussion

This article presents a new bioreactor for articular cartilagetissue engineering that allows for highly controlled


mechanical loading and tissue perfusion while ensuringsystem sterility or rather product safety. The combination ofthese features in one bioreactor system is new and openspossibilities to culture 3D tissue substitutes such as MACTs,controlled by the development of the viscoelastic tissueproperties. This article explains the construction of thebioreactor, the implemented sensors, and the automatedcontrol processes in detail. Subsequently, calibration of thesystem was performed. As an example of the applicability,we finally showed compression profiles in different agarosegels, a widely used carrier material for tissue engineering ofarticular cartilage (Buschmann et al., 1995; Kelly et al., 2006;Lee and Bader, 1997; Lima et al., 2006; Mauck et al., 2000).

In the bioreactor, the tissue substitute is located on a stageat the boundary between two compartments in a cylindricalbioreactor. The smaller lower compartment serves as adistributor for incoming medium and supplements, thelarger upper chamber of the vessel consists of two outflowports and contains the most culture medium and thereforefulfils a nutritional role during static culture periods (Schulz,2003). This system can be used to force medium perfusionthrough the construct as investigated in the presented‘‘proof of principle’’ experiment.

In this study we examined the effects of intermittentperfusion at a single flow rate of 0.5 mL/min on the viabilityand the extracellular matrix deposition in chondrocytesseeded into agarose constructs, while static cultured hydro-gels act as control group.

The designed bioreactor device with the perfusionsystem allowed for automated cultivation and automated,intermittent forced medium perfusion of 3D cell-loadedhydrogel scaffolds for a mid term duration of 3 weeks whilstmaintaining system sterility. Hence, product safety duringculturing is assured in this system.

Moreover, in this work, positive effects of direct perfusionon both cell proliferation and cell differentiation wereobserved. Our results show initially increased cell numbersin perfused hydrogels after 7 days in vitro. Absolute sGAGcontents at day 7 are comparable between static andperfused conditions. The chondrocytes in the perfusedspecimens appear to stop this early proliferation after 7 daysof cultivation and proceed to differentiation, since nofurther increase in ds-DNA was detectable after 7 days,whereas a significant increase in sGAG accumulation wasdetermined between weeks 1 and 3 of dynamic cultivation.

Similar increases in sGAG synthesis and ds-DNA contentwere demonstrated by other groups with well-acceptedperfusion bioreactors in which medium flow is directlyforced through a three-dimensional scaffold (Davissonet al., 2002b; Pazzano et al., 2000). Nevertheless, a contraryoutcome of a perfusion experiment was reported by Mizunoet al. (2001) when bovine articular chondrocyte loadedcollagen sponges were perfused at 0.33 mL/min for 15 days.Currently, the mechanism of the effect of directed perfusionon chondrocyte differentiation is not well understood;differences between reports may for instance relate to thetype of scaffold used. However, it is well accepted that flow-

through bioreactor apparatuses provide a controlled cultureenvironment that can promote the development of tissueengineered articular cartilage constructs. It was shown thatsuch a direct perfusion bioreactor system enhances masstransfer and ensures homogenous distribution of nutrients,supplements and oxygen, which are likely to improvechondrocyte metabolism and viability (Grimshaw andMason, 2000; Ysart and Mason, 1994).

Currently, none of these well-acknowledged directperfusion devices offer the (bio)-technological opportunityfor the implementation of an additional mechanicalstimulus, which is also known to have a significant effecton metabolism of articular chondrocytes embedded inscaffolds (Kim et al., 1995; Sah et al., 1989). This enhancesthe perspectives of the presented bioreactor design, which, inaddition to direct perfusion, has an innovative non-contactloading mechanism integrated in its upper compartment.This construction ensures a closed system and thereforeaseptic operation, in the identical way as presented in thepresent article. The promising concept of a scaffoldloading technology based on magnetic repulsion was alsoutilized by Schumann (Schumann, 2004). Drawbacks oftheir setup and that of others, such as the inability tosimultaneously handle culture medium and apply (complexor simple) compression regimes, have been solved in thepresent design.

In addition, the present bioreactor design is the firstsystem that combines non-contact loading with a reliablereal-time measurement of effective load and displacement,therewith enabling the monitoring of tissue development inreal-time while culturing, and without compromising thesterility of the graft in the test system. This offers significantbenefits over other mechanical testing devices used in tissueengineering studies where the mechanical properties aregenerally assessed off-line in samples that are taken out ofthe experiment at various time-points of culture. This makesthe present bioreactor particularly useful as part of a QC/QAprogram where the end-product after the cultivationprocedure is required to meet predefined conditions.

The bioreactor system is capable of automaticallycontrolling a complex regime of loading to individual softtissues. Stimulation frequencies up to 4 Hz are possible, withforces up to 2.8 N. Stronger magnets can increase maximalvertical forces to 9.6 N, without major revisions of theperipheral equipment. For our test specimen geometry,2.8 and 9.6 N forces are equivalent to pressure values ofapproximately 36 and 122 kPa, respectively. In addition, avariety of sample thicknesses and strain rates can be usedwith our device by changing the geometry of the locatingstage. This flexibility in the conditions is eminent inscientific research where specific geometries may be requiredto answer scientific questions.

Alternatively, the fully automated control system digitizesin real time all related bioprocess activities via a userinterface (Lintouch client), including time scheme and flowrate of medium perfusion, the regimes of mechanicalconditioning and other incubator settings. Thus, it allows

the bioreactor system also to be used in standard productionfacilities or clinical settings.

Measuring the individual progress of the mechanicalproperties in time is anticipated to become a key feature inthe production of patient-specific MACTs. The behavior ofautologous cells embedded in a scaffold differs per patientdue to variations in cell pool, -age, -constitution, etc. Hence,the response to their biochemical and mechanical environ-ments cannot be predicted. To cope with such differentresponses requires thorough feedback systems during theculturing. The present bioreactor allows for such real-timefeedback.

The geometry of our agarose test specimens withapproximately 0.8 cm2 is significantly smaller than therecommended 4–10 cm2 for clinically useful MACT grafts(Behrens et al., 2004). Constructs of the size used in thisstudy are standard in pre-clinical, in vitro research. Smallerconstructs allow multiple experiments to run in parallel withcells from single biopsies. To accommodate for the largerMACT scaffolds only requires up-scaling the individualculture chambers with the respective components and thesize of the external magnet.

An additional extension of the bioreactor system that weare currently investigating is feedback control of the culturemedium. Automated medium supplementation or replen-ishment upon depletion or particular nutrients may be donewith similar feedback control systems as the one described inthis article for mechanical loading. To replenish mediumrequires implementation of miniaturized pumps and valves,control requires input from for instance sensors for pH, pO2,pCO2, glucose- and lactate concentrations. This majorbiotechnological challenge will be presented in subsequentarticles.

Nomenclature

m

Sch

frictional coefficient (N)

FS
stimulation force (N)
Ff
frictional force (N)
Fn
horizontal force (N)
B(s)
magnetic field on axis of round magnet (mT)
Br
remanence (T)
R
magnet radius (m)
L
magnet length (m)
s
distance (m)
p
pressure (kPa)
F
force (N)
s15
compressive strain of 15% on 3 mm thick gels (equivalent to a
displacement of 450mm) (m)

Abbreviations

BR

ulz

bioreactor

AVD
actor vertical distance
ASF
actor stimulation frequency
SPos
inductive proximity sensor
et al.: Bioreactor System for Bioengineered Cartilage Grafts 725


SForce

726

platform load cell sensor

ID
inner diameter
OD
outer diameter
H
height
3D
three-dimensional
ECM
extracellular matrix
cGMP
current good manufacturing practices
MACT
matrix-coupled autologous cartilage transplant
QC/QA
quality control/quality assurance
NdFeB
neodymium iron boron
TCP/IP
transmission control protocol/Internet protocol
IPC
interprocess communication
CAN
controller area network
HMI
human machine interface
The authors would like to thank Dr. Chowdhury, Professor Lee,

Professor Bader (Queen Mary University of London, Department

of Engineering, UK), Dr. Oomens, Professor Baaijens (Eindhoven

University of Technology, Department of Biomedical Engineering,

The Netherlands), and the entire consortia of the 5th European

Framework Programme project ‘‘IMBIOTOR-Intelligent Mini Bio-

reactor’’ for their technical recommendations and their helpful multi-

disciplinary advice. We are grateful for the support of MEng (FH)

Mueller and Dipl Eng (FH) Meißner, and Professor Pretschner

(Leipzig University of Applied Sciences, Department of Electrical

Engineering and IT, Germany) with the technical implementation

of the novel magnetic concept and intelligent control strategy. No

benefits in any form have been received or will be received from a

commercial party related directly or indirectly to the subject of this

article.

References

Andereya S, Maus U, Gavenis K, Muller-Rath R, Miltner O, Mumme T,

Schneider U. 2006. First clinical experiences with a novel 3D-collagen

gel (CaReS) for the treatment of focal cartilage defects in the knee.

Z Orthop Ihre Grenzgeb 144(3):272–280.

Armstrong RD, Maluta J, Roecker DW. 1997. Apparatus and method for

maintaining and growth biological cells. US patent US6238908.

Behrens P, Bosch U, Bruns J, Erggelet C, Esenwein SA, Gaissmaier C,

Krackhardt T, Lohnert J, Marlovits S, Meenen NM, Mollenhauer J,

Niethard FU, Noth U, Perka C, Richter W, Schafer D, Schneider U,

Steinwachs M, Weise K. 2004. Indications and implementation of

recommendations of the working group ‘‘Tissue Regeneration and

Tissue Substitutes’’ for autologous chondrocyte transplantation (ACT).

Z Orthop Ihre Grenzgeb 142(5): 529–539.

Behrens P, Bitter T, Kurz B, Russlies M. 2006. Matrix-associated autologous

chondrocyte transplantation/implantation (MACT/MACI)–5-year fol-

low-up. Knee 13(3):194–202. Epub 2006 Apr 24.

Buckwalter JA. 2002. Articular cartilage injuries. Clin Orthop Relat Res 402:

21–37.

Buckwalter JA, Mankin HJ. 1998a. Articular cartilage: Degeneration and

osteoarthritis, repair, regeneration, and transplantation. Instr Course

Lect 17:487–504.

Buckwalter JA, Mankin HJ. 1998b. Articular cartilage: Tissue design

and chondrocyte-matrix interactions. Instr Course Lect 47:477–

486.

Buckwalter JA, Hunziker EB, Rosenberg L, Coutts R, Adams M, Eyre D.

1988. Articular cartilage: Composition and structure. In: Woo SL,

Buckwalter JA, editors. Injury and repair of the musculoskeletal soft

tissues. Park Ridge: American Academy of Orthopaedic Surgeons.

p 405–425.

Biotechnology and Bioengineering, Vol. 101, No. 4, November 1, 2008

Buschmann MD, Gluzband YA, Grodzinsky AJ, Hunziker EB. 1995.

Mechanical compression modulates matrix biosynthesis in chondro-

cyte/agarose culture. J Cell Sci 108(Pt 4):1497–1508.

Carver SE, Heath CA. 1999a. Increasing extracellular matrix production in

regenerating cartilage with intermittent physiological pressure. Bio-

technol Bioeng 62(2):166–174.

Carver SE, Heath CA. 1999b. Influence of intermittent pressure, fluid flow,

and mixing on the regenerative properties of articular chondrocytes.

Biotechnol Bioeng 65(3):274–281.

Cassino TR, Anderson R, Love BJ, Huckle WR, Seamans DK, Forsten-

Williams K. 2007. Design and application of an oscillatory compression

device for cell constructs. Biotechnol Bioeng 98(1):211–220.

Chowdhury TT, Bader DL, Shelton JC, Lee DA. 2003. Temporal regulation

of chondrocyte metabolism in agarose constructs subjected to dynamic

compression. Arch Biochem Biophys 417(1):105–111.

Cohen NP, Foster RJ, Mow VC. 1998. Composition and dynamics of

articular cartilage: Structure, function, and maintaining healthy state.

J Orthop Sports Phys Ther 28(4):203–215.

Darling EM, Athanasiou KA. 2003. Articular cartilage bioreactors and

bioprocesses. Tissue Eng 9(1):9–26.

Davisson T, Kunig S, Chen AC, Sah RL, Ratcliffe A. 2002a. Static and

dynamic compression modulate matrix metabolism in tissue engi-

neered cartilage. J Orthop Res 20(4):842–848.

Davisson T, Sah RL, Ratcliffe A. 2002b. Perfusion increases cell content and

matrix synthesis in chondrocyte three-dimensional cultures. Tissue Eng

8:807–816.

De Witt MT, Handley CJ, Oakes BW, Lowther DA. 1984. In vitro response

of chondrocytes to mechanical loading. The effect of short term

mechanical tension. Connect Tissue Res 12(2):97–109.

Demarteau O, Jakob M, Schafer D, Heberer M, Martin I. 2003. Develop-

ment and validation of a bioreactor for physical stimulation of engi-

neered cartilage. Biorheology 40(1–3):331–336.

Domm C, Schunke M, Christesen K, Kurz B. 2002. Redifferentiation of

dedifferentiated bovine articular chondrocytes in alginate culture under

low oxygen tension. Osteoarthritis Cartilage 10(1):13–22.

Donkelaar CC, Schulz RM. 2008. Review on patents for mechanical

stimulation of articular cartilage tissue engineering. Recent Patents

Biomed Eng 1(1):1–12.

Fermor B, Weinberg JB, Pisetsky DS, Guilak F. 2005. The influence of

oxygen tension on the induction of nitric oxide and prostaglandin E2

by mechanical stress in articular cartilage. Osteoarthritis Cartilage

13(10):935–941.

Frank EH, Jin M, Loening AM, Levenston ME, Grodzinsky AJ. 2000.

A versatile shear and compression apparatus for mechanical stimula-

tion of tissue culture explants. J Biomech 33(11):1523–1527.

Freed LE, Vunjak-Novakovic G, Marquis JC, Langer R. 1994. Kinetics of

chondrocyte growth in cell-polymer implants. Biotechnol Bioeng 43:

597–604.

Freed LE, Hollander AP, Martin I, Barry JR, Langer R, Vunjak-Novakovic G.

1998. Chondrogenesis in a cell-polymer-bioreactor system. Exp Cell

Res 240(1):58–65.

Fukuda K, Asada S, Kumano F, Saitoh M, Otani K, Tanaka S. 1997. Cyclic

tensile stretch on bovine articular chondrocytes inhibits protein kinase

C activity. J Lab Clin Med 130(2):209–215.

Galban CJ, Locke BR. 1999. Effects of spatial variation of cells and

nutrient and product concentrations coupled with product inhibition

on cell growth in a polymer scaffold. Biotechnol Bioeng 64(6):633–

643.

Grimshaw MJ, Mason RM. 2000. Bovine articular chondrocyte function in

vitro depends upon oxygen tension. Osteoarthritis Cartilage 8(5):386–

392.

Grimshaw MJ, Mason RM. 2001. Modulation of bovine articular chon-

drocyte gene expression in vitro by oxygen tension. Osteoarthritis

Cartilage 9(4):357–364.

Heath CA, Magari SR. 1996. Mini-review: Mechanical factors affecting

cartilage regeneration in vitro. Biotech Bioeng 50:430–437.

Hibino H, Katayama S, Irie H, Inoue H, Mizuno H, Hakamatsuka Y, Saito

Y, Koyanagi H. 2005. Cell culture system, apparatus for checking

cultured cells and apparatus for culturing cells. US patent

US2005158846.

Hofstaetter JG, Wunderlich L, Samuel RE, Saad FA, Choi YH, Glimcher MJ.

2005. Systemic hypoxia alters gene expression levels of structural

proteins and growth factors in knee joint cartilage. Biochem Biophys

Res Commun 330(2):386–394.

Hunziker EB. 2002. Articular cartilage repair: Basic science and clinical

progress. A. review of the current status and prospects. Osteoarthritis

Cartilage 10(6):432–463.

Ikenoue T, Michael CD, Trindade MC, Lee MS, Lin EY, Schurman DJ,

Goodman SB, Smith RL. 2003. Mechanoregulation of human articular

chondrocyte aggrecan and type II collagen expression by intermittent

hydrostatic pressure. J Orthop Res 21:110–116.

Jin M, Frank EH, Quinn TM, Hunziker EB, Grodzinsky AJ. 2001. Tissue

shear deformation stimulates proteoglycan and protein biosynthesis in

bovine cartilage explants. Arch Biochem Biophys 395(1):41–48.

Kelly TA, Ng KW, Wang CC, Ateshian GA, Hung CT. 2006. Spatial and

temporal development of chondrocyte-seeded agarose constructs in

free-swelling and dynamically loaded cultures. J Biomech 39(8):1489–

1497.

Kim YJ, Sah RL, Grodzinsky AJ, Plaas AH, Sandy JD. 1994. Mechanical

regulation of cartilage biosynthetic behavior: Physical stimuli. Arch

Biochem Biophys 311(1):1–12.

Kim YJ, Bonassar LJ, Grodzinsky AJ. 1995. The role of cartilage streaming

potential, fluid flow and pressure in the stimulation of chondrocyte

biosynthesis during dynamic compression. J Biomech 28(9):1055–

1066.

Kuo CK, Li WJ, Mauck RL, Tuan RS. 2006. Cartilage tissue engineering: Its

potential and uses. Curr Opin Rheumatol 18(1):64–73.

Langer R. 2000. Tissue engineering. Mol Ther 1(1):12–15.

Langer R, Vacanti JP. 1993. Tissue engineering. Science 260(5110):920–

926.

LeBaron RG, Athanasiou KA. 2000. Ex vivo synthesis of articular cartilage.

Biomaterials 21(24):2575–2587.

Lee DA, Bader DL. 1997. Compressive strains at physiological frequencies

influence the metabolism of chondrocytes seeded in agarose. J Orthop

Res 15(2):181–188.

Lima EG, Bian L, Mauck RL, Byers BA, Tuan RS, Ateshian GA, Hung CT.

2006. The effect of applied compressive loading on tissue-engineered

cartilage constructs cultured with TGF-beta3. Conf Proc IEEE Eng Med

Biol Soc 1:779–782.

Lindahl A, Brittberg M, Peterson L. 2001. Health economics benefits

following autologous chondrocyte transplantation for patients with

focal chondral lesions of the knee. Knee Surg Sports Traumatol

Arthrosc 9(6):358–363.

Marcacci M, Berruto M, Brocchetta D, Delcogliano A, Ghinelli D, Gobbi A,

Kon E, Pederzini L, Rosa D, Sacchetti GL, Stefani G, Zanasi S. 2005.

Articular cartilage engineering with Hyalograft C: 3-year clinical results.

Clin Orthop Relat Res 435:96–105.

Marlovits S, Zeller P, Singer P, Resinger C, Vecsei V. 2006. Cartilage repair:

Generations of autologous chondrocyte transplantation. Eur J Radiol

57(1):24–31. Epub 2005 Sep 26.

Martin I, Padera RF, Vunjak-Novakovic G, Freed LE. 1998. In vitro

differentiation of chick embryo bone marrow stromal cells into carti-

laginous and bone-like tissues. J Orthop Res 16(2):181–189.

Martin I, Wendt D, Heberer M. 2004. The role of bioreactors in tissue

engineering. Trends Biotechnol 22(2):80–86.

Mauck RL, Soltz MA, Wang CC, Wong DD, Chao PH, Valhmu WB, Hung

CT, Ateshian GA. 2000. Functional tissue engineering of articular

cartilage through dynamic loading of chondrocyte-seeded agarose gels.

J Biomech Eng 122(3):252–260.

Mauck RL, Seyhan SL, Ateshian GA, Hung CT. 2002. Influence of seeding

density and dynamic deformational loading on the developing struc-

ture/function relationships of chondrocyte-seeded agarose hydrogels.

Ann Biomed Eng 30(8):1046–1056.

Mauck RL, Nicoll SB, Seyhan SL, Ateshian GA, Hung CT. 2003. Synergistic

action of growth factors and dynamic loading for articular cartilage

tissue engineering. Tissue Eng 9(4):597–611.

Millward-Sadler SJ, Wright MO, Davies LW, Nuki G, Salter DM. 2000.

Mechanotransduction via integrins and interleukin-4 results in altered

aggrecan and matrix metalloproteinase 3 gene expression in normal,

but not osteoarthritic, human articular chondrocytes. Arthritis Rheum

43(9):2091–2099.

Mizuno S, Allemann F, Glowacki J. 2001. Effects of medium perfusion on

matrix production by bovine chondrocytes in three-dimensional col-

lagen sponges. J Biomed Mater Res 56(3):368–375.

Mizuno S, Tateishi T, Ushida T, Glowacki J. 2002. Hydrostatic fluid

pressure enhances matrix synthesis and accumulation by bovine chon-

drocytes in three-dimensional culture. J Cell Physiol 193(3):319–

327.

Naughton GK. 2002. From lab bench to market: Critical issues in tissue

engineering. Ann NY Acad Sci 961:372–385.

Nesic D, Whiteside R, Brittberg M, Wendt D, Martin I, Mainil-Varlet P.

2006. Cartilage tissue engineering for degenerative joint disease. Adv

Drug Deliv Rev 58(2):300–322.

Obradovic B, Martin I, Padera RF, Treppo S, Freed LE, Vunjak-Novakovic

G. 2001. Integration of engineered cartilage. J Orthop Res 19(6):1089–

1097.

Parkkinen JJ, Ikonen J, Lammi MJ, Laakkonen J, Tammi M, Helminen HJ.

1993. Effects of cyclic hydrostatic pressure on proteoglycan synthesis in

cultured chondrocytes and articular cartilage explants. Arch Biochem

Biophys 300(1):458–465.

Pazzano D, Mercier KA, Moran JM, Fong SS, DiBiasio DD, Rulfs JX,

Kohles SS, Bonassar LJ. 2000. Comparison of chondrogensis in

static and perfused bioreactor culture. Biotechnol Prog 16(5):893–

896.

Risbud MV, Sittinger M. 2002. Tissue engineering: Advances in in vitro

cartilage generation. Trends Biotechnol 20(8):351–356.

Sah RL, Kim YJ, Doong JY, Grodzinsky AJ, Plaas AH, Sandy JD. 1989.

Biosynthetic response of cartilage explants to dynamic compression.

J Orthop Res 7(5):619–636.

Schulz RM. 2003. Aspects of design in an intelligent mini bioreactor with

perfusion and mechanical stimulation; 22.10.2003; Leipzig.

Schulz RM, Bader A. 2007. Cartilage tissue engineering and bioreactor

systems for the cultivation and stimulation of chondrocytes. Eur

Biophys J 36(4–5):539–568.

Schulz RM, Hohle S, Zernia G, Zscharnack M, Schiller J, Bader A, Arnold K,

Huster D. 2006. Analysis of extracellular matrix production in artificial

cartilage constructs by histology, immunocytochemistry, mass spectro-

metry, and NMR spectroscopy. J Nanosci Nanotechnol 6(8):2368–

2381.

Schumann D. 2004. Methoden zur Optimierung von Tissue Engineering

Produkten auf dem Wege zur Reparatur osteochondraler Defekte.

Regensburg: Universitat Regensburg. p 220.

Sengers BG, Heywood HK, Lee DA, Oomens CW, Bader DL. 2005. Nutrient

utilization by bovine articular chondrocytes: A combined experimental

and theoretical approach. J Biomech Eng 127(5):758–766.

Sittinger M, Bujia J, Minuth WW, Hammer C, Burmester GR. 1994.

Engineering of cartilage tissue using bioresorbable polymer carriers

in perfusion culture. Biomaterials 15(6):451–456.

Smith TJN, Sydney MP, Rupert H. 2005. Automated tissue engineering

system. US patent US2006141623.

Torzilli PA, Grigiene R, Huang C, Friedman SM, Doty SB, Boskey AL, Lust

G. 1997. Characterization of cartilage metabolic response to static and

dynamic stress using a mechanical explant test system. J Biomech 30(1):

1–9.

Vunjak-Novakovic G, Martin I, Obradovic B, Treppo S, Grodzinsky AJ,

Langer R, Freed LE. 1999. Bioreactor cultivation conditions modulate

the composition and mechanical properties of tissue-engineered car-

tilage. J Orthop Res 17(1):130–138.

Waldman SD, Spiteri CG, Grynpas MD, Pilliar RM, Kandel RA. 2003.

Long-term intermittent shear deformation improves the quality

of cartilaginous tissue formed in vitro. J Orthop Res 21(4):590–

596.

Wendt D, Marsano A, Jakob M, Heberer M, Martin I. 2003. Oscillating

perfusion of cell suspensions through three-dimensional scaffolds



enhances cell seeding efficiency and uniformity. Biotechnol Bioeng

84(2):205–214.

Wright MO, Nishida K, Bavington C, Godolphin JL, Dunne E, Walmsley S,

Jobanputra P, Nuki G, Salter DM. 1997. Hyperpolarisation of cultured

human chondrocytes following cyclical pressure-induced strain:


Evidence of a role for alpha 5 beta 1 integrin as a chondrocyte

mechanoreceptor. J Orthop Res 15(5):742–747.

Ysart GE, Mason RM. 1994. Responses of articular cartilage explant cultures

to different oxygen tensions. Biochim Biophys Acta 1221(1):15–

20.

development and validation of a novel bioreactor system ... · system for load- and...

Documents