Fisher John P. e.a. (ed.) Tissue Engineering

Подождите немного. Документ загружается.

mikos: “9026_c014” — 2007/4/9 — 15:51 — page 16 — #16

mikos: “9026_c015” — 2007/4/9 — 15:51 — page1—#1

Tissue Engineering

Bioreactors

Jose F. Alvarez-Barreto

Vassilios I. Sikavitsas

University of Oklahoma

15.1 Introduction.............................................. 15-1

15.2 Most Common Bioreactors in Tissue Engineering .... 15-2

Spinner Flask • Rotating-Wall Vessels • Perfusion

Chambers and Flow Perfusion Systems

15.3 Cell Seeding in Bioreactors.............................. 15-4

15.4 Bioreactor Applications in Functional Tissues......... 15-5

Tissues of the Cardiovascular System • Bone • Cartilage •

Anterior Cruciate Ligament and Tendons • Other Tissues

15.5 Design Considerations .................................. 15-13

15.6 Challenges in Bioreactor Technologies ................. 15-13

Acknowledgment................................................ 15-14

References ....................................................... 15-14

15.1 Introduction

The main goal of tissue engineering is the creation of artiﬁcial tissue having the ability to repair or simply

replace lost or damaged tissue. Common tissue engineering strategies involve the extraction of cells from

a small piece of tissue and their in vitro culture for later implantation using a carrier that allows the

formation of new tissue (Figure 15.1) [1]. Most approaches in this ﬁeld are based on common bioactive

factors, consisting of cells (generally stem cells, or progenitor cells), a scaffolding material, and growth and

differentiation factors [2]. The in vitro creation of an efﬁcient construct can be accelerated by applying

certain stimuli that can elicit speciﬁc responses to the cells. Stimulation can be done in two major ways:

chemically and electro/mechanically.

Chemical stimulation is carried out by using growth and differentiation factors speciﬁc for different

responses. Growth factors play a major role in cell division, matrix synthesis, and tissue differentiation

[3]. Examples of these proteins are: bone morphogenetic proteins (BMPs), which have been demonstrated

to induce the differentiation of mesenchymal stem cells into an osteoblastic lineage (BMP-2 and BMP-7),

and the vascular endothelial growth factor that greatly enhances angiogenesis [4,5]. The need for in vitro

mechanical stimulation in tissue engineering is drawn from the fact that most tissues function under

speciﬁc biomechanical environments in vivo. These environments play a key role in tissue remodeling and

regeneration. The stresses can be translated into different kinds of forces that range from load bearing to

hydrodynamic forces due to ﬂuid ﬂow [6]. Thus, the mechanochemical microenvironment that progenitor

15-1

mikos: “9026_c015” — 2007/4/9 — 15:51 — page2—#2

15-2 Tissue Engineering

Cell harvesting

from patient

Tissue

regeneration

and transplantation

Chemical and mechanical

stimulation (bioreactor)

Culture on

3-D scaffold

FIGURE 15.1 Basic steps in the process of tissue engineering: cells are harvested from the patient and proliferated

in a three-dimensional environment, followed by the application of mechanical and chemical stimuli in a bioreactor

prior to implantation.

cells growinto is expected to control the fate of these cells while undergoing differentiation toward different

lineages.

A bioreactor is generally deﬁned as a device capable of creating the proper environment for the creation

of a certain biological product [21]. Therefore, a bioreactor is described as a simulator, a device in which

biological as well as biochemical processes can be carried out. In tissue engineering, bioreactors are used to

impart certain forces that imitate different electromagnetic and mechanical stimuli occurring in the body.

However, these devices are not limited to the sole application of electromechanical stimuli; they must meet

other requirements in order to create grafts that, when implanted, will lead to the regeneration of damaged

organs. A bioreactor must efﬁciently transport nutrients and oxygen to the construct, maintaining an

appropriate concentration in solution. In most tissue engineering applications, a scaffold is seeded with

cells and supports the formation of extracellular matrix (ECM). Consequently, the bioreactor has to induce

a homogeneous cell distribution throughout these structures in order to generate a uniformly distributed

ECM. Tissue engineering bioreactors can be used for cell seeding and long-term cultures.

This chapter describes some of the most popular bioreactor designs available for the engineering of dif-

ferent tissues. Hydrodynamic conditions and transport phenomena considerations are addressed, as well

as applications to functional tissues and unique devices for speciﬁc applications. Design considerations,

challenges, and new directions are also presented.

15.2 Most Common Bioreactors in Tissue Engineering

The choice of a bioreactor to cultivate three-dimensional constructs depends upon the tissue to be engin-

eered and its functional biomechanical environment. Emulation of physiological conditions is the main

objective when developing these kinds of systems, and this issue has been addressed in different ways. The

incorporation of convective forces has become a common characteristic among most bioreactors. In this

section, we describe some of the most common bioreactors found in the engineering of several functional

tissues such as bone, cartilage, and cardiovascular applications, among others.

mikos: “9026_c015” — 2007/4/9 — 15:51 — page3—#3

Tissue Engineering Bioreactors 15-3

(a) (b)

(e)

FIGURE 15.2 Common bioreactors used in tissue engineering. (a) Static culture, (b) spinner ﬂask, (c) rotating wall,

(d) perfusion system, and (e) perfused column.

15.2.1 Spinner Flask

The spinner ﬂask (Figure 15.2b) represents one of the simplest bioreactor models. It was ﬁrst designed

with the idea to use convection in order to maintain a well-mixed system. The scaffolds are threaded

into needles connected to the cover of the ﬂask, and submerged in the culture medium. Convection is

generated through the usage of a magnetic stir bar or a shaft that continuouslymixes the media surrounding

the scaffolds, providing a practically homogenous distribution of oxygen and nutrients [7,8]. The ﬂuid

dynamic environment at the external surface of the scaffolds is turbulent and characterized by the existence

of eddies that may enhance the transport of nutrients into the porosity, and locally expose cells residing at

the exterior of the construct to relatively high shear forces. The magnitude of the shear stresses can vary

signiﬁcantly between different locations; therefore, not all the cells are exposed to the same shear stresses.

The presence of convective forces external to the scaffolds may not suppress concentration gradients

appearing deep inside large three-dimensional constructs, where diffusion is the controlling mechanism

of nutrient transport [9,25].

15.2.2 Rotating-Wall Vessels

Initially designed by NASA as a microgravity environment for cell culture, the rotating-wall bioreactor

(Figure 15.2c) is now widely used in the formation of engineered bone, cartilage, and other tissues

[7,8,10–12]. This device consists of two concentric cylinders whose annular space contains the cell culture

medium [13]. The inner cylinder is static and permeable to allow gas exchange for oxygen supply. The

outer cylinder, on the other hand, is impermeable and horizontally rotates at a speed that causes centrifugal

forces that can balance, if tuned properly, the gravitational forces; thus, generating a pseudo microgravity

mikos: “9026_c015” — 2007/4/9 — 15:51 — page4—#4

15-4 Tissue Engineering

environment [7,13,14]. Unlike the spinner ﬂask, in the rotating-wall vessel the ﬂuid ﬂow is mostly laminar

and the range of shear forces experienced by the cells at the outer surface is relatively narrow, with the

existence of a stagnation zone at the upstream edge. As reported by Williams et al. [15], shear stresses

decrease in the direction of ﬂow and no signiﬁcant variations from scaffold to scaffold are observed.

Medium can be recirculated between the annular space and an external gas membrane. A modiﬁcation

of the original design, called rotating-wall perfused-vessel bioreactor, includes the rotation of the inner

cylinder. In this model, media is perfused from the vessel’s end cap to the pores of the inner cylinder [14].

The conditions of operation of a rotating-wall vessel must be carefully controlled. Large rotation speeds

of the outer wall will affect mass transport since most of the inlet ﬂuid bypasses the vessel volume, whereas

an increase on the differential rotation enhances the radial and axial distribution of the ﬂuid at low mean

shear stresses [14].

15.2.3 Perfusion Chambers and Flow Perfusion Systems

Flow perfusion bioreactors provide continuous ﬂow through chambers were the scaffolds are located. The

perfusion column (Figure 15.2e) was one of the ﬁrst designs of this kind of bioreactors. Culture medium

is continuously recirculated through the chamber, thus improving the transport of nutrients and oxygen

to the constructs [16,17,22]. Nevertheless, the ﬂow of medium in these chambers is distributed between

the inner network of the construct and its surroundings, minimizing convective ﬂow through the scaffold

[18]. To ensure that the ﬂow of medium occurs exclusively through the porosity of the material, new

designs of ﬂow perfusion bioreactors include the conﬁning of the construct in chambers (Figure 15.2d).

In this way, a more controllable ﬂow is achieved and nutrient transport limitations are virtually eliminated.

Internal ﬂow can also expose the cells inside the scaffold to ﬂuid shear forces that have been known to

be stimulatory for some cell types such as osteoblasts and endothelial cells [19,20]. A standard design

of this kind of reactors does not exist, but all of them are based on the same principle. A more detailed

description of a perfusion system is given later in this chapter.

15.3 Cell Seeding in Bioreactors

The ﬁrst step to culturing cells in a three-dimensional environment is the seeding of scaffolds [21].

Along with the characteristics of the material, this process plays a crucial role in the development of

efﬁcient constructs for tissue engineering. Seeding of scaffolds determines the initial number of cells in

the construct, as well as their spatial distribution throughout the matrix. Consequently, proliferation,

migration, and the speciﬁc phenotypic expression of the engineered tissue will be affected by the utilized

seeding technique [22]. In the case of tissues that require a ﬁbrous or porous material, static seeding has

been the most widely used method of cell seeding (Figure 15.2a). Burg et al. [26] compared different

seeding techniques using rat aortic cells in polyglycolide ﬁbrous meshes. Static seeding produced the

poorest cellular distribution. In addition to preventing a homogeneous spatial distribution of the cells,

static seeding also produces a low yield [23,26]. Holy et al. [23] reported a 25% efﬁciency of attachment

after seeding 0.5 to 10×10

cells on porous PLGA 75/25 scaffolds. A low yield diminishes the development

of speciﬁc functions related to cell–cell interactions and increases the required amount of cells; therefore,

the usage of new seeding techniques becomes imperative.

In order to address these issues, researchers have incorporated convection into the process of cell

seeding, suppressing some of the mass transfer limitations encountered in the static procedure. Spinner

ﬂask bioreactors (Figure 15.2b) have been implemented to create convection and, thereby, hydrodynamic

forces that could help increase mass transport. Poly(glycolic acid) (PGA) scaffolds were threaded onto

needles and chondrocytes suspensions with a total number of cells between 2 × 10

and 10 × 10

were

used. A yield of 60% was obtained after2hofseeding. A more uniform distribution of the cells in the

scaffold was seen (compared to the static seeding); nonetheless, the concentration of cells in the outer layer

of the construct was 60 to 70% higher than that in the bulk [24]. This behavior may be due to the poor

mikos: “9026_c015” — 2007/4/9 — 15:51 — page5—#5

Tissue Engineering Bioreactors 15-5

convection to the interior of the scaffold, making migration the only way for cells to reach the interior of

the scaffold.

It has been reported that, in the spinner ﬂask, the shear forces at the external surface of the scaf-

fold are highly nonuniform. Such variability may inﬂuence the homogeneity of the seeded cells even

when considering only the external surface area [25]. To avoid such problem, Mauney et al. rotated

the scaffold every2hfortheﬁrst6hofseeding so that each face of the construct could be exposed

to the ﬂow ﬁeld, reaching a homogenous distribution of the cells and a higher efﬁciency than that

of the static methodology. However, despite the high efﬁciency of seeding achieved with the spinner

ﬂask bioreactor, a more homogeneous distribution of the cells throughout the construct volume is still

desired.

One way to guarantee mass transfer to the interior of the scaffold and a better distribution of cells is

by applying perfusion [26,27]. In this technique, the construct is press ﬁtted into a chamber, and the cell

suspension is ﬂowed through it (Figure 15.2c). Li et al. used a depth ﬁltration system to seed poly (ethylene

terephthalate) matrices at a rate of 1 ml/min. The cell suspension was recycled to increase the yield. Cell

density increased along with the inoculation cell number, with an efﬁciency of about 65%, while with the

static seeding, the yield stayed constant and lower than that achieved with the perfusion [22]. Similarly,

Kim et al. seeded hepatocytes on polymeric matrices using a ﬂow perfusion system. A suspension of rat

hepatocytes at a density of 5 × 10

cells/ml was pumped through decellularized bone matrices at a ﬂow

rate of 1.5 ml/min for 4 h. A total of approximately 4.4 ×10

cells were attached to the matrix, which was

considered successful. Furthermore, scanning electron microscopy and histology conﬁrmed a uniform

distribution of hepatocytes throughout the scaffold.

Wendt et al. [27] monitored seeding efﬁciency and uniformity of static, spinner ﬂask, and perfusion

systems. Using the same inoculation concentrations, there was not statistical signiﬁcance among the

efﬁciencies of the static and perfused techniques, both producing a larger yield than the spinner ﬂask.

Uniformity, however, was optimized by the perfusion apparatus, while the static and the spinner ﬂask

generated cell-scaffold constructs with low spatial uniformity [27].

15.4 Bioreactor Applications in Functional Tissues

After being seeded with the speciﬁc type of cells needed for the application, scaffolds must be subjected

to longer periods of culture under the desired physical stimuli. The appropriate stimulation relies on the

mechanical, biological, and ultrastructural characteristics of the native tissue. Bioreactors for engineered

vascular grafts must mimic the natural ﬂuid dynamic conditions of blood ﬂow, including relatively high

ﬂow rates and pulsatile ﬂow [42]. Bone tissue has also been shown to be stimulated by ﬂuid ﬂow [19].

Engineered ligaments and tendons need mechanical loading to emulate the conditions that they normally

experience [87]. In this section we highlight some of the most important bioreactor applications in the

engineering of different tissues; different designs and their efﬁciency in the development of inductive

constructs are discussed.

15.4.1 Tissues of the Cardiovascular System

Several types of bioreactors have been used for the regeneration of tissues from the cardiovascular system.

Spinner ﬂasks and rotating-wall vessels have been employed in the regeneration of heart tissue. Cardiac

myocytes were cultured on PGA scaffolds in a spinner ﬂask, rotating-wall vessels and perfusion systems

[28,29]. After 14 days of culture, the medium in the rotating-wall vessels showed higher levels of oxygen.

The cell number in the spinner ﬂask and rotating-wall reactor was larger than that of the static culture.

Likewise, the uniformity of extracellular matrix throughout the scaffold was more homogeneous in the

rotating-wall vessels. The engineered constructs had similar structural characteristics to that of the native

tissue, and the cultured cells secreted proteins related to mature cardiac myocytes (myosin, troponin-T,

tropomyosin, etc.). However, the cell density was always lower in the interior of the constructs [29].

mikos: “9026_c015” — 2007/4/9 — 15:51 — page6—#6

15-6 Tissue Engineering

15.4.1.1 Vascular Grafts

The emulation of physiological conditions in tissue-engineered vascular grafts is extremely important.

Factors affecting the successful generation of a vascular graft include the selection of the appropriate

scaffolding material, the cell type (smooth muscle cells, endothelial cells, and ﬁbroblasts), and the culturing

conditions [30,35]. Evaluation of mechanical properties is necessary to determine the quality of the

engineered construct. Mechanical strength is directly related to the matrix structure, especially to the

alignment of smooth muscle cells and collagen ﬁbers [30,31].

Blood experiences rapid pulsating ﬂow in the body with a velocity of about 33 cm/sec in the aorta and

0.33 mm/sec in capillaries [32]. Moreover, the conditions of ﬂow will determine the differentiation path

and properties of endothelial cells [6,33]. The extracellular matrix organization and mechanical properties

of the graft, such as burst pressure and suture retention, are determinant factors of the graft quality [34].

Different systems with pulsatile ﬂow have been designed in order to achieve these goals [35–38].

Static culture of endothelial cells seeded on different synthetic or natural polymers resulted in grafts with

poor mechanical properties and morphological characteristics different from that of the native tissue [39].

In an attempt to mimic the natural environment of blood vessels, ﬂuid ﬂow has been employed during

in vitro culture of smooth muscle cells seeded on PGA scaffolds. Incorporation of pulsation improved the

mechanical properties of the constructs and their histological appearance resembled that of native arteries.

It is important to point out that plain pulsation generated grafts with inferior mechanical properties [40].

The implementation of biomimetic bioreactors that utilize pulsatile ﬂow in the physiological range appears

to stimulate the formation of an organized matrix similar to that of the native tissue [41,42].

An example of a pulsatile-ﬂow bioreactor for vascular grafts is shown in Figure 15.3. This design, used

in the Laboratory of Cardiovascular Tissue Engineering at the University of Oklahoma, resembles the

shell and tube concept of some heat exchangers. Flow circuits are separated into shell-side and tube-

side to monitor (and control) system pressure and ﬂow rates independently. Control valves downstream

Waste

Media

(Shell-side)

Pump

Compliance chamber

Tissue-engineered construct

within bioreactor

Pressure

transducers

Media

(Tube-side)

CPU

FIGURE 15.3 Shell–tube bioreactor for vascular grafts. (Courtesy of Dr. P. McFetridge from the School of Chemical

Engineering and Materials Science, University of Oklahoma Bioengineering.)

mikos: “9026_c015” — 2007/4/9 — 15:51 — page7—#7

Tissue Engineering Bioreactors 15-7

of the bioreactor can be adjusted automatically in response to pressure variation. Typically, vascular

bioreactors require different cell types seeded onto each side of the construct; as such, the dual-circuit-

process ﬂow allows different media types (endothelial and smooth muscle cell) to be run independently.

Media is recirculated until nutrients, antibiotics, and pH require correction to remain within physiological

parameters.

In another model designed by Thompson et al. [35], pressure is created by cyclical air inﬂow and the

amount of pressure introduced in the system is controlled by a check valve. The air comes into contact

with the culture medium in the so-called driving shaft. Tubing with medium is connected to the unit

where the constructs are placed; this unit is called the manifold. Constructs can be accommodated in

series or parallel, with a capacity of six scaffolds. Real-time ﬂow and pressure can be monitored using an

in-line ﬂow meter and pressure transducer [35]. Hoerstrup et al. [43] designed a similar system in which

the pulsatile ﬂow is generated by the inﬂation/deﬂation of an elastic membrane.

Pulsatile ﬂow, however, is not the only stimulation that has been used in the formation of tissue-

engineered blood vessels. Selitkar et al. [44] designed a bioreactor to apply cyclic strain to cell-seeded

scaffolds. After seeding aortic rat SMCs and culturing them for 8 days, circumferential orientation and a

homogeneous distribution was observed in the scaffold. Collagen ﬁbers were also produced in an organ-

ized fashion, forming bound assemblies. Mechanical properties were improved compared to unstrained

constructs, achieving an ultimate stress of 58 kPa and modulus of 142 kPa, which are much greater than

those achieved without any stimulation (16 and 68 kPa, respectively) [44].

Different stimuli have also been combined in order to enhance the formation of the extracellular matrix

and mechanical properties. Peng et al. [45] used both cyclic stretching and pulsatile ﬂow with perfusion

to culture vascular cells. Endothelial cells were grown in distensible silastic tubes and subjected to pulse

pressures and shear stresses comparable to the physiological values (up to 150 mmHg and 15 dyn/cm

Cells aligned in the direction of higher pulsation, almost parallel to the ﬂow. Actin ﬁbers showed a

peripheral longitudinal orientation at greater pulsatilities.

15.4.1.2 Heart Valves

The environment in which heart valves perform is mechanically complex. Heart valves must operate at

a frequency of approximately 1Hz, undergoing shear and bending stresses caused by blood ﬂow [46,47].

Hoerstupet al. designed a pulsatile-ﬂow system for the engineering of heart valves. This bioreactor basically

consists of two chambers with a silicone diaphragm between them. The lower chamber maintains air, while

the upper one is the culture medium chamber. Air is pumped, using a ventilator, into the lower chamber

to displace the diaphragm and create the pulsation. Pulsatile ﬂows from 50 to 2000 ml/min and systemic

pressures from 10 to 240 mmHg can be achieved using this system [38]. Dumont et al. [37] designed a

similar bioreactor for the formation of engineered aortic heart valves. This apparatus consists of a left

ventricle (LV) made out of a silicone and an after-load with a compliance chamber and a resistance. The

stroke and frequency of the machine is controlled by a piston that compresses and decompresses the LV.

In this design, air is also incorporated in a compliance chamber, being pumped under conditions that

resemble the standard systolic and diastolic pressures. The tissue-engineered aortic heart valve is placed

between the left ventricle and the compliance chamber [37].

Schenke-Layland et al. [48] cultured endothelial cells and myoﬁbroblasts on decellularized porcine

heart valves under static and pulsatile conditions. Increased cellularity, collagen and elastin content, and

mechanical strength were observed when the heart valves were cultured under pulsatile ﬂow. The level

of pulsation is also a critical parameter in the formation of functional heart valves, as demonstrated by

Mol et al. [49].

15.4.2 Bone

Bone is a hard connective tissue that provides mechanical support to the human body and is a frame for

locomotion. Bone grafts have been generated under a wide variety of culturing conditions, including static

and dynamic systems. Among the most popular dynamic systems are spinner ﬂasks, rotating-wall vessels,

mikos: “9026_c015” — 2007/4/9 — 15:51 — page8—#8

15-8 Tissue Engineering

and perfusion systems [20,24,50,51]. As for every tissue, before deciding upon the kind of bioreactor

to be used, considerations concerning the carrier matrix, cells (osteoblasts, mesenchymal stem cells,

etc.), and growth factors must be taken into account. In the case of bone, the matrix to be used must be

osteoconductive, provide mechanical support, deliver cells and allow their attachment, growth, migration,

and osteoblastic differentiation [52]. Synthetic and natural polymers have been implemented. Among the

synthetic polymers poly-α-hydroxy esters, poly(ε-caprolactone), poly(propylene fumarate), poly(sebacic

acid), and their copolymers have been widely used. Materials such as ceramics and titanium have also been

used for bone replacement [53,54]. Cell number and calcium deposition are good markers to evaluate

the evolution of bone matrix. Furthermore, alkaline phosphatase activity (ALP) is used to assess early

differentiation activity of osteoblastic cells. Production of extracellular matrix proteins such as osteocalcin,

osteopontin, and bone sialoprotein is also taken into consideration [55].

As mentioned before, static culture was one of the ﬁrst attempts to produce bone matrix. Ishaug et al.

[56] seeded marrowstromal cells on top of poly (dl-lactic-co-glycolic acid) (PLGA) foams of different pore

sizes at different densities. Cell proliferation was supported by the scaffold, and high level of ALP activity

and calcium matrix deposition were observed. It was found that the depth of mineralized tissue increased

over time, but the maximum penetration was only around 240 µm, resulting in a nonhomogeneous cell

and matrix distribution [56].

Improvement in the development of bone matrix in vitro has been achieved with the addition of

convection in the in vitro culture stage, which ultimately translates in a better transport of nutrients and

gases. After statically seeding 1 ×10

marrow stromal cells on 75 : 25 PLGA scaffolds, Sikavitsas et al. [7]

cultured these constructs under three different conditions: statically, in a spinner ﬂask and in a rotating-

wall vessel. The culture was carried out for 21 days, and samples were analyzed at 7, 14, and 21 days.

Scaffolds cultured in the spinner ﬂask bioreactor showed the largest number of cells at all time points,

followed by the static culture. At the end of the culture period, constructs in the spinner ﬂask presented

higher calcium contents than those encountered in the static and rotating-wall vessel [7].

Shea et al. [57] also utilized a spinner ﬂask to culture poly(lactic acid) foams seeded with MC3T3-E1

preosteoblasts and evaluated their differentiation. Cells were seeded statically and cultured for 12 weeks.

Proliferation was observed over time; however, their distribution throughout the scaffold lacked homo-

geneity. Cells were densely located only at a thin layer of 200 µm near the scaffold’s surface. The density

dramatically decreased deeper into the construct. The same behavior was seen for the formation of

extracellular matrix and calcium deposition.

It has been shown that mechanical stimulation augments the production of alkaline phosphatase, osteo-

blast proliferation, and mineral deposition in osteoblastic cells seeded on different scaffolding materials

[58]. Osteoblastic cells have been shown to be responsive to shear stress induced by ﬂuid ﬂow. The stim-

ulatory effect of shear stresses has shown to induce an increase in the release of important regulatory

factors such as nitric oxide and prostaglandin E2 [59–61]. Interestingly, osteoblasts have been found to

be more responsive to ﬂuid shear forces than mechanical strain [62]. A question arises then, what is the

physiological relevance of the stimulatory effect of ﬂuid ﬂow on bone cells? It has been hypothesized that

mechanical strains on bone tissue cause ﬂuid ﬂow in the lacunar–canalicular porosity of bone [63–65].

Consequently, the incorporation of ﬂuid ﬂow through the porous network is desired in order to stimulate

a faster and more efﬁcient formation of bone matrix. This goal has been reached with the implementation

of ﬂow perfusion bioreactors [20,66,67].

Bancroft et al. [8] developed a perfusion system (Figure 15.4) where medium is pumped through the

scaffold, thereby maintaining mechanical stimulation and transport of nutrients through the pores. The

scaffolds are tightly ﬁt into cassettes in order to ensure ﬂuid ﬂow exclusively through the porous network.

Constructs are later placed in ﬂow chambers that are capped and secured with o-rings to restrict the ﬂow

around them (Figure 15.4a). The medium is pumped from a ﬂask to the top of the chamber and sent to

another reservoir from the bottom. This direction of ﬂow helps avoiding the entrance of air bubbles into

the ﬂow chamber. Both ﬂasks are connected so that the medium is in continuous recirculation. The main

body of the reactor consists of a total of six chambers and is made out of Plexiglas to allow the visualization

and monitoring of the ﬂow inside the chambers. Each chamber corresponds to an independent circuit

mikos: “9026_c015” — 2007/4/9 — 15:51 — page9—#9

Tissue Engineering Bioreactors 15-9

To pump

(b)

(a)

From pump

FIGURE 15.4 Schematics of a ﬂow perfusion bioreactor. (a) Close up of the perfusion chamber where the scaffold is

press ﬁtted. (b) Lateral view of the main body of the bioreactor.

using one of the heads of a peristaltic pump that produces ﬂow rates from 0.1 to 10 ml/min (Figure 15.4b).

The tubing permits the exchange of carbon dioxide and oxygen with the atmosphere in the incubator.

A complete change of medium can be done due to the two-reservoir set up [18].

To study how the shear rate affects the growth of bone matrix in vitro, Bancroft et al. [19] varied

the ﬂow rate when culturing titanium ﬁber meshes seeded with rat marrow stromal cells for 16 days.

Controls have been cultured under static conditions, and the ﬂow rates used in the perfusion culture were

0.3, 1.0, and 3.0 ml/min. It was found that the deposition of calcium was greatly increased in the ﬂow

perfusion culture as compared with the static conditions. It was also observed that increased medium ﬂow

improves the distribution of extracellular matrix throughout the construct volume [19]. The increased

calcium deposition could have been due to the increased shear forces or increased chemotransport in the

porosity of the scaffolds when higher ﬂow rates were employed. To isolate the effects of shear forces from

the mass transport effects, the shear forces were changed by varying the viscosity of the culture medium

under constant ﬂow rate [68]. An increase in viscosity, which translates into greater shear forces, was

found to enhance the deposition of mineral matrix and the ECM distribution throughout the construct,

demonstrating the importance of ﬂuid-ﬂow induced shear forces on the creation of bone tissue-engineered

grafts.

Meinel et al. [67] cultured human mesenchymal stem cells on silk scaffolds for 5 weeks in a ﬂow

perfusion chamber (at 0.2 ml/min) and a spinner ﬂask. Scaffolds cultured under ﬂow perfusion showed

a more homogenous distribution of the mineralized matrix throughout the construct although those

cultured in the spinner ﬂask produced a greater amount of deposited calcium.

Other kinds of stimulation include mechanical strain and electrical current. A bioreactor was developed

that allowed the continuous exposure of cells to continuous cyclic stretching [69]. Primary osteoblast-like

cells were seeded on silicone rubbers and subjected to 1000 microstrains at 1 Hz either continuously or in

periods of 60 min. Cellularity and calcium deposition were enhanced under the presence of mechanical

strain and the intermittent procedure was the most efﬁcient of them [70]. Another bioreactor that can

expose cells to mechanical load has been developed by Shimko et al. [71]. Unlike the previous study,

mechanical loading had a negative impact in the mineral deposition.