Towards bioreactor development with physiological motion control ...

Medical Engineering and Physics 39 (2017) 106–112

Contents lists available at ScienceDirect

Medical Engineering and Physics journal homepage: www.elsevier.com/locate/medengphy

Towards bioreactor development with physiological motion control and its applications Marcus Stoffel∗, Wolfgang Willenberg, Marzieh Azarnoosh, Nadine Fuhrmann-Nelles, Bei Zhou, Bernd Markert Institute of General Mechanics, RWTH Aachen University, Germany

a r t i c l e

i n f o

Article history: Received 25 April 2016 Revised 16 September 2016 Accepted 23 October 2016

Keywords: Bioreactor development Biomechanical modelling Remodelling laws Finite element method Experimental validation

a b s t r a c t In biomedical applications bioreactors are used, which are able to apply mechanical loadings under cultivation conditions on biological tissues. However, complex mechanobiological evolutions, such as the dependency between mechanical properties and cell activity, depend strongly on the applied loading conditions. This requires correct physiological movements and loadings in bioreactors. The aim of the present study is to develop bioreactors, in which native and artificial biological tissues can be cultivated under physiological conditions in knee joints and spinal motion segments. However, in such complex systems, where motions with different degrees of freedom are applied to whole body parts, it is necessary to investigate elements of joints and spinal parts separately. Consequently, two further bioreactors for investigating tendons and cartilage specimens are proposed additionally. The study is complemented by experimental and numerical examples with emphasis on medical and engineering applications, such as biomechanical properties of cartilage replacement materials, injured tendons, and intervertebral discs. © 2016 IPEM. Published by Elsevier Ltd. All rights reserved.

1. Introduction Bioreactors are important parts of any biomechanical and biomedical researches over the last several decades, since bioreactors could accelerate the studies of material properties during remodelling, degeneration and regeneration of biological soft tissues such as cartilage, intervertebral discs, and tendons under cultivation conditions. In studies of Waldman et al. [21] and Vunjak-Novakovic et al. [20] especially changes in material stiffness of cartilage were described after mechanical loadings in bioreactors. Moreover, a bioreactor with two degrees of freedom composed of combined compression and torsion loadings on spine motion segments was proposed by Illien-Jünger et al. [7]. Liu et al. [9] proposed a knee simulator without bioreactor environment for porcine knee joints to stimulate the native porcine cartilage with kinematics of human total knee replacements, scaled to porcine joint limits. In studies of Gao et al. [5], a knee joint bioreactor was proposed, which can subject compression and rolling motions to biological material. Sun et al. [19] described the necessity to apply physiological loadings to articular cartilage and proposed a bioreactor with compression and rolling movements. Combined com-

∗

Corresponding author. Fax: +49 241 8092231. E-mail address: [email protected] (M. Stoffel).

http://dx.doi.org/10.1016/j.medengphy.2016.10.010 1350-4533/© 2016 IPEM. Published by Elsevier Ltd. All rights reserved.

pression and shear loadings for the cultivation of cartilage replacement specimens were presented by Yusoff et al. [25]. In studies of Grad et al. [6], the high potential of bioreactor systems is described, which lies in the development of bioreactors exhibiting physiological joint motions for optimising tissue replacements. For this purpose, Grad et al. [6] use bioreactors with spheres, which apply compression and rolling on specimens. Especially in [26], the authors note to foresee that bioreactors in future will mimic in vivo mechanical loadings by complex mechanism systems. Consequently, straightforwardly, we discuss bioreactor systems exhibiting more than one degree of freedom. The necessity for combined loadings with different frequencies and repose times in cartilage were studied by Panadero et al. [13] as well. Numerical simulations as an evaluation method lead to the determination of material parameters and stress and strain distributions during the bioreactor cultivation. A comparison between calculated and measured results can lead to a validated numerical model, which offers the possibility to analyse e.g. stress peaks, damage, or remodelling inside the investigated tissue more precisely. In this way, injuries (Little et al. [8]) or mechanobiological couplings (Soukane et al. [14]) can be modelled. Due to the fact that combined tissue loading with compression and torsion seems to be an important criterion for creating the real physiological conditions, the development of these bioreactors is essential.

M. Stoffel et al. / Medical Engineering and Physics 39 (2017) 106–112

The aim of the present study is the development of four different bioreactors with different functions. By using uniaxial tension and compression bioreactor systems for tendon and cartilage the material parameters such as Young’s Modulus and Poisson’s Ratio can be determined. Afterwards, in the multi-axial knee and intervertebral disc bioreactors, the experiments will be performed to study other complex material properties such as the regeneration and degeneration processes of native or artificial tissues under physiological realistic conditions. Due to the complex mechanobiological evolutions in living tissues, it is worth to develop numerical models by means of the finite element method. This should support the understanding of loading histories, stress and strain distributions in knee joints and spine segments and, moreover, the remodelling effects.

2. Methods This study focuses on the knee joint and on spinal segments. A knee joint bioreactor was developed, which mimics the physiologically correct loading conditions in a knee joint, while the intervertebral disc (IVD) bioreactor is developed to study the de- and regeneration behaviours of the IVDs under long-term cyclic compression and torsion. However, the mechanical analysis of superposed stress distributions in different directions together with geometrical and physical nonlinearities becomes very complex in these two bioreactors. Therefore, there is a need for investigating single element samples with simplified geometry in a special type of bioreactor, especially if the pure mechanical effects are overlaid with remodelling phenomena. Experiments in the compression bioreactor, or in the tensile bioreactor as well, can be used for determining material properties under simpler boundary and loading conditions as in the knee joint bioreactor.

2.1. The tensile bioreactor The investigation of the mechanical behaviour of tendon tissues and tendon cells requires an appropriate in vitro cultivation environment. Fig. 1 shows the tensile bioreactor. A stepper motor is mounted, which drives a spindle being connected to the clamp de-

107

vice. The stepper motor works with the micro stepping technology and can perform 800 steps per revolution. The clamp device does not damage the clamped sample and also reliably prevents the sample from slipping out. Moreover, the clamping force is generated by the compression of a spring with a well defined length. Thus, in every experiment, the same clamping force is generated. All these described aspects of the clamping device are essential for reproducible experimental results. Hereby, the clamping device is limited to thin structures with a thickness of less than 3 mm. Despite this limitation, a wide range of different structures, such as whole tendons or cut specimens from the IVD, can be mounted in the bioreactor. A 9.81 N load cell (Althen AUMM-K1) is connected via a rod with a high tensile, but low shear strain resistance. Thus, the influence of lateral force on the load cells is minimised. Both, the load cell and the stepper motor are controlled and read out by a custom made printed circuit board (PCB). The PCB itself is controlled by a custom made software tool with a graphical user interface (GUI), which is running on a personal computer. With the aid of the software tool, it is possible to prepare experiments, to start and stop arbitrary stimulation cycles (e. g. periodical stimulation with sinoidal displacement shapes and integrated pauses, relaxation tests with pauses). During the experiments the bioreactor is placed in an incubator. In Fig. 1, the bioreactor is shown with a bovine tendon. A validation of this bioreactor has been presented in [24]. One interesting application of the bioreactor is the mechanical investigation and characterisation of shear forces inside of collagen containing materials. This is assumed to be an important factor, for example for the generation of tendon diseases, see [2] and [11]. However, it can be assumed that they are also an important factor in other types of tissue, such as in the IVDs. In order to provide a method for the measurement of shear forces inside of soft tissues, a special protocol has been developed and tested with tendon tissue. Therefore, a specimen with a rectangular shape was mounted in the bioreactor (initial gauge length 30 mm, width 12.1 mm) and elongated by 4 mm within 4 s. These parameters were chosen, as a strain of about 13 % and a velocity of 1 mm/s are in the physiological range [22]. The same testing procedure was repeated with an introduced cut, see Fig. 1. A comparison between the measured reaction forces reveals the acting shear stresses and strains [24].

Fig. 1. Left: Tensile bioreactor with clamped sample. Right: Specimens made of tendon tissue. Due to the different lengths of the damages, the actual acting shear forces can be quantified.

108


Fig. 2. Compression bioreactor with measurement results of increasing Young’s modulus during cultivation.

2.2. The compression bioreactor Here, the compression bioreactor is used to investigate remodelling effects occurring in cell-seeded cellular samples stimulated for four weeks under vertical compression. Experiments with collagen type I implants, which was seeded with human chondrocytes, was performed [27]. In Fig. 2, this bioreactor is shown, which was proposed in similar form in [17] and which is extended here by the measurement technique described in Section 2.1. At first, the components of the reaction chamber and the punch were sterilised separately using the plasma sterilisation and then assembled together. The construction of the reaction chamber ensured an aseptic condition for the specimens during the whole experiment. The whole set-up is placed in an incubator keeping the experiments at body temperature. Gas filters on top of the chamber ensured a 5% CO2 atmosphere inside the chamber, and the medium was changed manually several times during the experiment. Up to four cylindrical specimens with 2 mm thickness and 10 mm diameter can be placed in the chamber and four bioreactors of this type can be used in an incubator at the same time to assure a statistically usable number of results. During the experiment, specimens were firstly loaded with a preload to ensure the contact between specimens and the punch, and then loaded with cyclic compression with a constant frequency. To trace the evaluation of mechanical properties of the specimens under stimulation, specimens were removed from one bioreactor weekly and tested with unconfined compression and relaxation experiments. With the help of the numerical model proposed in Section 2.5, material parameters such as Young’s modulus and relaxation parameters were identified. On the right hand side of Fig. 2 a module with slots for four cylindrical specimens is shown. The evolution of mechanical properties of the implants was validated with histological investigations [12]. 2.3. The knee joint bioreactor To investigate cell-seeded implants under reproducible, physiological conditions, a knee joint bioreactor was developed, which is presented in Fig. 3, where the knee joint module is inserted

into an MTS 859 MiniBionix® II testing machine. The aim is to find out, how the remodelling effect differs under physiological loading, when the implants are preconditioned in the compression bioreactor or directly placed in the knee joint. To enhance the investigations in the compression bioreactor, equal implants and experimental times of four weeks are intended. The artificial knee parts are manufactured by means of 3D-plots out of P430 ABS material filled with casting resin or alternatively produced out of acrylic glass. The source data for these geometries were CT images of healthy patients. Also the position of femur and tibia to each other is taken from CT images. The position of femur and tibia to each other is taken from CT images. Structural measures define the positions at assembling. The joint parts are coated with an artificial cartilage replacement material [10,15,23] as a mixture made of polyvinyl pyrrolidone (PVP) and PVA in order to obtain more cartilage-like mechanical properties as in hydrogels (Fig. 4). Due to small pores and small surface irregularities at the top of the artificial knee parts in combination with the production process of the hydrogel, an artificial cartilage coating with native, small barbs emerges. The production process of the hydrogel requires several freezing and unfreezing cycles until a material stiffness similar to native cartilage is obtained. A flexible latex sleeve closes the bioreactor and is clamped by an elastic band which is placed in a slot outside the acrylic glass and the base plate. The nutrient medium can be removed and replaced manually in a repose time. A heating for providing 37 ◦ C body temperature is located at the bottom of the module with temperature control; gas exchange is provided at the top. The components of the bioreactor were sterilised separately and assembled in a bench. The motions, which are applied to the knee joint, are according to the German code ISO 14243, Fig. 3, which simulates human walking cycles. Fig. 3 shows the coordinate axes, which indicate the six possible degrees of freedom in terms of displacements and rotations. The femur is loaded (1) vertically against x3 -direction, causing a compression in the knee joint. Additionally, an axial rotation (2) around x3 (tibial internal-external rotation) from −5.72◦ to +1.87◦ , simulating the axial rotation in a knee joint while stretching, and a rotation (3) of the femur around the x1 -axis from 0.0° to +58.0◦ , simulating flexion and extension, is applied. The tibia


109

Fig. 3. Material testing machine with knee joint module with membrane and human walking cycles.

Fig. 4. Knee joint module inside the material testing machine with membrane finite element simulation results, (a) Applied compressive stresses, (b) Inner normal stresses in x3 -direction, (c) Inner transverse shear stresses S13 , (d) Inner transverse shear stresses S23 .

slides (4) actively, horizontally in x2 -direction (anterior-posterior motion) to adjust the contact points in the joint caused by the distance of the native rotation axis to the given flexion/extension axis and to recreate the natural sliding. Furthermore, a free sliding (5) in x1 -direction and a rotation (6) around the x2 -axis is possible. The axial load (1) ranges from −167.6 N to −2600.0 N, if the joint is loaded by 100%. The percentage of loading can be adjusted individually, as well as the stretching velocity and the number of load cycles. Also repose times can be taken into account. Inertia effects are neglected by the ISO. The load is actually measured by an axial-torsional load transducer, which is a component of the MTS

testing machine. For future investigations, an add-on by pressure measurements directly under the implants is planed. Accompanied to experiments, numerical simulations are carried out to investigate normal and shear stress distributions in the artificial coating of the knee parts. According to the experimental set-up, the geometry is extracted out of CT images, discretisated and imported to Abaqus [1]. The artificial knee parts are treated as rigid bodies. For the cartilage a C3D8R element type with homogeneous solid properties is used. Initially, pure isotropic, elastic material behaviour is used. The boundaries correspond to the experimental set-up.

110


Fig. 5. Intervertebral disc bioreactor with simulated and measured results with healthy and injured intervertebral discs.

2.4. The intervertebral disc bioreactor The IVD bioreactor is presented in Fig. 5. It consists of a closed cylindrical chamber with 60 cm height and 30 cm diameter. The whole chamber can be coupled to a material testing machine (MTS) in order to transmit compression forces and torsion moments to the internal structure of the bioreactor. Inside the bioreactor, a clamping device with screws for the fixation of the motion segment is located. The ends of the clamping device are perforated to allow a circulation of nutrient medium around the specimen. Connections for gas exchange are provided and at the bottom of the IVD-bioreactor, a heat exchanger for ensuring body temperature is installed. In the experiments the aim is to study strength carrying capacities of injured discs after several physiological loading cycles. This procedure can be carried out with a comparative study between healthy and injured discs. Additional ligaments, which are attached naturally to the spine, can be studied also separately in the tensile bioreactor. For example, fresh sheep lumbar motion segments L1– L2 are applied for the displacement controlled experiments up to 30% strain. The exact geometry of the experimental specimen is extracted from CT-Data. The invisible disc is interpolated by filling the space between the vertebrae. After discretisation, the geometry is imported into Abaqus [1] for FE computation via an user material subroutine. The element used for the modelling of the deformable ring in the IVD is a standard solid continuum tetrahedral element (C3D4) and a solid continuum tetrahedral element (C3D10) is applied for the modelling of the core. In order to work with an appropriate discretisation, a study of convergence is performed. The studies of convergence have shown that the solution is not changed significantly by reducing the element size to less than 0.1 mm. Furthermore, the vertebrae are represented as rigid bodies. In the applied model, the elastic stiffness tensor is treated anisotropically, which leads to parameter identification procedures described in [18].

lation. The finite element model can then be used to study the specimen’s behaviour under modified boundary and loading conditions. Modelling the multi-phase material of biological soft tissues yields a model with many material parameters, which are not easy to verify. Therefore, the basis for the numerical model is a phenomenological material law for cartilage replacement materials proposed in [16] and refined in [3], which leads to viscoelastic models of Maxwell- and Poynting–Thomson types with relaxation components in form of a Taylor series depending on the volume strain. A term for taking remodelling phenomena into account was added in [17] accounting for an increasing Young’s modulus E1rem in radial direction of the specimen. The stress rate tensor is then expressed by 4

4

4

4

S˙ =B0 E˙ + B1 E˙ − D1 S+ B2 E˙ , 4

(2.1)

4

where B0 , B2 , S, E denote the elasticity tensor, the elasticity tensor with one Young’s modulus E1rem as a remodelling function, second Piola–Kirchhoff stress tensor, and Green–Lagrange strain tensor, re4

spectively. The elasticity tensor B1 accounts for strain rate dependency and includes the components η

B1i jkl = η0 E˙ i j1 if i = k and j = l (otherwise B1i jkl = 0 )

(2.2)

as a function of the strain rate and with material parameters η0 4

and η1 . The components of the relaxation tensor D1 which is driven by diffusive processes in the water-saturated material, is expressed by

D1i jkl = D0 + D1 v if i = k and j = l (otherwise D1i jkl = 0 ) (2.3) with D0 , D1 , and v denoting the initial value of relaxation, a material parameter, and the volume strain, respectively. For covering also remodelling phenomena, i.e. a strengthening effect of the specimen after up to four weeks cultivation time, a free energy for compressible Neo-Hookean materials [4] is introduced and extended here by means of the last term accounting for the energy in single fibres by

2.5. Numerical model

1 1 = λ log2 (J ) + μ I1C − 3 − μ log J + S11 dE11

In order to identify material parameters, compression and relaxation tests are conducted followed by a finite element simu-

with J = det F and I1C = C : I. Here, C and F denote right Cauchy– Green tensor and deformation gradient, respectively, while λ and

2

2

(2.4)


μ are the Lamé constants. It is assumed that fibre growth takes place in horizontal direction, as it was studies in histological sections [17]. So, assumed to develop in only one elastic nent B21111 of the stiffness tensor. The modulus is then expressed by

E˙ 1rem (a ) = k a E1lim − E1rem

confirmed in preliminary the change in stiffness is parameter in the compoevolution of the Young’s

(2.5)

with an accumulated free energy

a =

dt

(2.6)

and k and E1lim as a material parameter and a limit value for E1rem , respectively. The initial value of E1rem is zero. In this phenomenological model, the accumulated free energy is treated as a stimulus for a pronounced cell activity, which results in a stiffening of the material. 3. Results To characterise the mechanical properties of different kinds of biological soft tissue replacement materials the compression bioreactor is developed. During long time stimulation over four weeks, a remodelling phenomenon can be observed in Fig. 2. In this case, four bioreactors were used parallelly for cultivation durations over one, two, three, and four weeks. In this way, an evolution of the Young’s modulus E1rem in radial direction in the collagen sample was observed, see [12]. To investigate the same cell-seeded samples under reproducible, physiological conditions, a knee joint bioreactor was constructed. So far experiences in bioreactor systems, cell cultivation and sterility over several weeks acquired by the compression bioreactor were used. In Fig. 4, a finite element mesh of the surfaces of the artificial knee parts is shown corresponding to the samples in Fig. 4. At one instance of time, in the four pictures on the right hand side of Fig. 3 the compressive stress acting on the condyle (a), the normal stress in x3 -direction (b), and two transverse shear stresses ((c), (d)) are indicated (units in MPa). By means of these simulations, stress peaks can be identified, which can lead to perforated cartilage material in the experiments. To study the influence of shear stresses in the injured tendons, four different tendons were provided with fissures of different lengths. Measurements with injured tendons, see Fig. 1, are carried out with a uniaxial tensile bioreactor. Experiment type C is done with a specimen exhibiting a fissure with the half of the intact sample width as in type A. However, the measured reaction force is higher than the half of the reaction force of experiment type A. This effect is directly related to the shear forces acting inside of the investigated tendon tissue. Thus, it can be concluded that the bioreactor is suitable to observe shear effects inside of tendon tissue. To investigate the biomechanical response of healthy and injured IVDs under long-term cultivation, the cellular activity and gas exchange can be measured and controlled in the IVDbioreactor. In Fig. 5, a compression and relaxation test with a healthy sheep motion segment in the IVD-bioreactor is shown. Accompanied to the measurements, a finite element simulation was carried out, followed by a material parameter identification. Following this method, predictions of strength carrying capacities of injured IVDs, e.g. decreasing stiffness and relaxation due to a removed nucleus pulposus in Fig. 5, are conducted. 4. Discussion In Fig. 2, both types of expected remodelling phenomena are measured in the experiments. Due to the fact that arthrotic cells

111

from surgeries were used in the experiments, the samples exhibit a higher stiffness at the beginning (see Fig. 2), which decreases after one week. A regeneration of the material takes place after the second week and reaches a Young’s modulus of 1 MPa, which is realistic for cartilage materials. Furthermore, in Fig. 2, the Young’s modulus of a control sample is indicated, which was not stimulated during the whole cultivation duration. With this experience, these remodelling studies are continued to study the influence of superposed normal and shear stresses in the knee joint bioreactor under physiological conditions with cellular material. The finite element results in Fig. 4 show the stress distribution which occurs in the cartilage replacement material during cyclic loading without failure of the coating. A comparison between experimental and numerical studies in combination with a viscoelastic material model shall lead to an optimisation of the replacement material. The effect of supporting shear stresses around damaged zones was also made in the tensile bioreactor with tendon experiments, see Fig. 1. Compared to a measurement, the advantage of the finite element simulation in Fig. 5 is, that an investigation with a missing disc core can be carried out without cutting the fibre ring. In an experimental study, the fibre ring must be injured to remove the core. 5. Conclusions Four types of bioreactors were presented in this study. The focus of this investigation was to perform physiologically realistic motions during bioreactor cultivations. Following this method, complex motions can be created in knee joint and intervertebral disc bioreactors while material properties can be determined under simplified loading conditions in compression and tensile bioreactors. This strategy follows the engineering principle to determine material properties by material tests before going ahead to structural elements under arbitrary loading situations. The proposed combination of four bioreactors allows to investigate body elements such as tendons and cartilage separately and to study the behaviour of complete body parts in special bioreactors as well. Conflict of interests There are no competing interests to declare, no external fundings, and no ethical approval required. References [1] Abaqus. 6.14, dassault systèmes; 2014. [2] Archambault JM, Elfervig-Wall MK, Tsuzaki M, Walter HW, Banes A. Rabbit tendon cells produce MMP-3 in response to fluid flow without significant calcium transients. J Biomech 2002;35:303–9. [3] Azarnoosh M, Stoffel M, Markert B. Anisotropic viscohyperelastic behavior of intervertebral discs: Modeling and experimental validation. PAMM 2014;14:91–2. [4] Bonet J, Wood RD. Nonlinear Continuum Mechanics for Finite Element Analysis. Cambridge University Press; 1997. [5] Gao J, Zhang C, Liu H, Gao L, Sun M, Dong X. A roller-loading bioreactor system for researching cartilage mechanobiology. Procedia Environ Sci 2011;8:197–201. [6] Grad S, Loparic M, Peter R, Stolz M, Aebi U, Alini M. Sliding motion modulates stiffness and friction coefficient at the surface of tissue engineered cartilage. Osteoarthritis Cartilage 2012;20:288–95. [7] Illien-Jünger S, Gantenbein-Ritter B, Grad S, Lezuo P, Ferguson SJ, Alini M, et al. The combined effects of limited nutrition and high-frequency loading on intervertebral discs with endplates. Spine 2010;35(19):1744–52. [8] Little JP, Adam CJ, Evans JH, Pettet GJ, Pearcy MJ. Nonlinear finite element analysis of anular lesions in the l4/5 intervertebral disc. J Biomech 2007;40:2744–51. [9] Liu A, Jennings LM, Ingham E, Fisher J. Tribology studies of the natural knee using an animal model in a new whole joint natural knee simulator. J Biomech 2015;48:3004–11.

112


[10] Ma R, Xiong D, Miao F, Zhang J, Peng Y. Novel PVP/PVA hydrogels for artiular cartilage replacment. Mater Sci Eng C 2009;29:1979–83. [11] Magra M, Maffulli N. Matrix metalloproteases: A role in overuse tendinopathies, vol. 39. BMJ Publishing Group Ltd and British Association of Sport and Exercise Medicine; 2005. p. 789–91. [12] Nebelung S, Gavenis K, Lüring C, Zhou B, Mueller-Rath R, Stoffel M, et al. Simultaneous anabolic and catabolic responses of human chondrocytes seeded in collagen hydrogels to long-term continuous dynamic compression. Ann Anat 2012;194(4):351–8. [13] Panadero JA, Lanceros-Mendez S, Gomez-Ribelles JL. Differentiation of mesenchymal stem cells for cartilage tissue engineering: individual and synergetic effects of three-dimensional environment and mechanical loading. Acta Biomater 2016;33:1–12. [14] Soukane DM, Shirazi-Adl A, Urban JPG. Computation of coupled diffusion of oxygen, glucose and lactic acid in an intervertebral disc. J Biomech 2007;40:2645–54. [15] Stammen JA, Williams S, Ku DN, Guldberg RE. Mechanical properties of a novel PVA hydrogel in shear and unconfined compression. Biomaterials 2001;22:799–806. [16] Stoffel M, Weichert D, Müller-Rath R. Modeling of articular cartilage replacement materials. Arch Mech 2009;61(1):1–19. [17] Stoffel M, Yi JH, Weichert D, Zhou B, Nebelung S, Müller-Rath R, et al. Bioreactor cultivation and remodelling simulation for cartilage replacement material. Med Eng Phys 2012;34:56–63.

[18] Strampe M, Stoffel M, Weichert D, Sellei RM, Pape HC. Study of the biomechanical behaviour of structurally stable/unstable motion segments of the sheep spine. PAMM 2012;12:97–8. [19] Sun M, Lv D, Zhang C, Zhu L. Culturing functional cartilage tissue under a novel bionic mechanical condition. Med Hyp 2010;75(6):857–9. [20] Vunjak-Novakovic G, Martin I, Obradovic B, Treppo S, Grodzinsky AJ, Langer R, Freed LE. Bioreactor cultivation conditions modulate the composition and mechanical properties of tissue-engineered cartilage. J Orthop Res 1999;17:130–8. [21] Walsh AJL, Lotz JC. Biological response of the intervertebral disc to dynamic loading. J Biomech 2004;37:329–37. [22] Wang T, Gardiner B, Lin Z, Rubenson J, Kirk T, Wang A, et al. Bioreactor design for tendon/ligament engineering. Tissue Eng Part B Rev 2013;19(2):133–46. [23] Watase M, Nishinari K. Large deformation of hydrogels of poly (vinyl alcohol) hydrogel, agarose and kappa-carrageenan. Makromol Chem 1985;186:1081–6. [24] Willenberg W, Azarnoosh M, Stoffel M, Markert B. Experimental and numerical investigation of tendons and tendon cells. PAMM 2016;16:113–14. [25] Yusoff N, Osman NAA, Pingguan-Murphy B. Design and validation of a bi-axial loading bioreactor for mechanical stimulation of engineered cartilage. Med Eng Phys 2011;33(8):782–8. [26] Zhao J, Griffin M, Cai J, Li S, Bulter PEM, Kalaskar DM. Bioreactors for tissue engineering: An update. Biochem Eng J 2016;109:268–81. [27] Zhou B, Stoffel M, Weichert D, Rath B. Theoretical and experimental investigation of cartilage replacement material for medical applications. PAMM 2012;12:95–6.