Note: Descriptions are shown in the official language in which they were submitted.
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
1
VASCULARISED TISSUE GRAFT
This invention relates to the fields of tissue
engineering and transplantation, and particularly to the
generation of vascularised tissue.
BACKGROUND OF THE INVENTION
Tissue engineering utilising homologous starting
material offers the prospect of replacing missing or non-
functioning body parts with newly created, living tissue. it
has the potential to minimise loss of tissue and resultant pain
from the donor site experienced in conventional reconstructive
surgery. or to recreate specialised tissue for which there is no
donor site, while obviating the long-term immunosuppression
required for heterologous transplantation.
It combines the techniques of tissue culture, the
creation of bio-compatible materials and the manipulation of
angiogenesis in order to create new, vascularised tissue to
replace damaged tissue or tissue which is congenitally absent.
One of the major challenges faced in tissue engineering
is to create differentiated tissue of the appropriate size and
shape. Tissue created without a functional vasculature is
strictly limited in size by the constraints of oxygen
diffusion; if the tissue is too large it will become necrotic
before the host has time to create a new blood vessel supply.
Thus there are many advantages in creating new tissue
containing a functional vasculature. Additionally, as the new
tissue may need to be produced at a site on the body remote
from the defect, or on an immunosuppressed carrier animal or in
vitro with an extracorporeal circulation, the blood supply for
the new tissue must be defined, so that it can be brought with
the tissue intact to the site of reconstruction.
The creation of skin flaps, a living composite of skin
and its underlying fat, is a common technique used to repair
tissue defects in reconstructive surgery. Because these flaps
must retain their blood supply to remain viable after
transplantation, the origin of the flaps is limited to those
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
2
areas where there is an anatomically recognised blood vessel
source. In order to overcome this limitation, skin flaps can be
"pre-fabricated" by implanting short segments of blood vessels
into a desired site, and utilising the resultant angiogenesis
to vascularise a flap of the desired size and composition.
Subsequently this vascularised flap can be transferred by
microsurgery to the region of interest. This technique is,
however, limited by the availability of donor tissue, and the
disfigurement that results at the donor site.
In an extension to this technique, Erol and Spira
(1980) demonstrated that the creation of an anastomosed
arterio-venous (AV) loop beneath a skin graft could produce a
vascularised skin "flap".
However, while the generation of vascularised skin
using an AV loop has been demonstrated, the production of other
vascularised tissues suitable for grafting remains elusive.
Vascularised adipose tissue, for example, is often demanded in
reconstructive procedures; however, donor mature adipose tissue
is extremely fragile, and will rapidly become necrotic if not
immediately re-connected to a functional blood supply.
Furthermore, the use of conventional autologous transplantation
techniques involves "robbing Peter to pay Paul", producing
disfigurement at the donor site. The ability to produce new
tissue with a defined vasculature-would overcome this major
shortcoming.
Khouri et al. (1993) and Tanaka et al. (1996) have
demonstrated that an arteriovenous loop could intrinsically
generate new, vascularised tissue when it was lifted from the
body, sandwiched between sheets of collagenous matrix and
isolated from the surrounding tissue within a plastic chamber.
in the model described by Khouri et al., the generation of new
tissue relied on the addition of recombinant BB-homodimer of
Platelet-Derived Growth Factor (BB-PDGF), and even with this
supplement the tissue was labile, peaking in volume at 15 days
and subsiding by 30 days. Similarly, tissue growth in Tanaka's
model, where the chamber was supplemented with 0-Fibroblast
Growth Factor ((3-FGF or FGF-2), continued to increase in
volume, peaking at 2 weeks, but returned to the levels of the
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
3
unsupplemented control chambers after 4 weeks. This AV loop
model is not generally known in the field of tissue
engineering.
The classical notion that mature tissues do not contain
stem cells has changed considerably in recent years. Many
mature tissues which were previously regarded as largely non-
self renewing are now considered to harbour a stem cell
population. These stem cells possess the potential to change
their phenotype in response to their environment, and may be
able to provide a self-replenishing stem cell population
(Prockop, 1997). Micro-environmental cues are considered to
play a significant role in determining the behaviour of stem
cells, for example, in initiating stem cell division and
differentiation and/or maintaining stem cell quiescence. The
cues and mechanisms behind these processes are far from being
understood. However, it is clear that the ability to recruit,
stimulate, proliferate and differentiate stem cells is the crux
of tissue engineering. The behaviour of stem cells is largely
studied in vitro, although a small number of in vivo studies
have examined the behaviour of stem cells when injected either
under the capsule of mature organs or systemically. These
studies have a number of limitations in furthering the
knowledge of the use of stem cells for tissue engineering. in
particular, when the stem cells are injected into mature organs
they must interact with an established micro-environment and
derive a limited neovasculature from the host organ; when they
are systemically injected they become widely dispersed. In
order for stem cells to generate organs, it is expected that
they will require an expandable vascular supply to accommodate
and service de novo tissue generation. In order to assist in
directing stem cell expansion, development and differentiation,
an expandable microenvironment comprising an inert support
and/or extracellular matrix is also expected to be required. We
have now developed a model which satisfies these requirements,
and holds great promise for the study of stem cells. Its
application to tissue engineering is a significant advance in
the state of the art.
It will be clearly understood that, although a number
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
4
of prior art publications are referred to herein, this
reference does not constitute an admission that any of these
documents forms part of the common general knowledge in the
art, in Australia or in any other country.
We have now developed a system for producing
vascularised graft tissue, which is useful in transplant and
reconstructive surgery, and also provides a useful model
system.
SUMMARY OF THE INVENTION
In a first aspect, the invention provides a method of
producing donor vascularised tissue, suitable for
transplantation into a recipient animal in need of such
treatment, comprising the steps of:
a) creating a functional circulation on a vascular
pedicle in a donor subject;
b) partially or totally enclosing the vascular pedicle
within a fabricated chamber;
c) seeding the chamber with isolated cells or pieces of
tissue;
d) implanting the chamber containing the vascular
pedicle into a host animal at any site where such an anatomical
construct can be created; and
e) leaving the chamber in the implantation site for a
period sufficient to allow the growth of vascularised new
tissue.
In one preferred embodiment, the method comprises the
step after step (a) of surrounding the vascular pedicle with
added extracellular matrix and/or a mechanical support. in
another preferred embodiment, the method comprises a step after
step (b) of adding growth factors, drugs, antibodies,
inhibitors or other chemicals to the chamber.
Preferably in step (e) the chamber is left in the
implantation site for at least 4 weeks, more preferably at
least 6 weeks.
The vascularised tissue may be grown in vivo or in
vitro, or may be in situ in the host.
More preferably the chamber is implanted in the donor
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
body, beneath the skin, although it is not limited to
subcutaneous insertion. While externalization of the chamber
during tissue/organ growth is theoretically possible, the high
risk of infection makes this a rarely used alternative.
5 For the purposes of this specification, the term "donor
subject" is taken to mean an animal, especially a mammal and
most especially a human, in which the donor vascularised tissue
is created. For the purposes of this specification, the term
"recipient animal" is taken to mean an animal, especially a
mammal and most especially a human, that receives the donor
vascularised tissue graft. It would be appreciated by those
skilled in the art that as the generation of new vasculature,
angiogenesis, in all warm blooded animals is associated with
essentially the same physiological and pathological processes,
methods disclosed herein are directly applicable to all warm
blooded animals. The donor subject is preferably a mammal, and
may be a human or a non-human animal. Preferred mammals include
rodents, felines, canines, hoofed mammals such as horses, cows,
sheep and goats, pigs, and primates. in a particularly
preferred embodiment, the donor subject and recipient are
human.
The person skilled in the art will appreciate that a
"vascular pedicle" is an artificial or naturally occurring
arrangement of blood vessels or vessel replacements that
comprises an artery taking blood to the site of the construct
and a vein carrying it away. Preferably the vascular pedicle
comprises an arterio-venous (AV) loop or shunt. In an AV loop
or shunt the artery is either joined directly to the vein or
connected via a graft of a similar diameter so that there is no
impediment to blood flow (for example as illustrated in Figure
1). In one alternative arrangement, the artery and vein are
both ligated and'blood flow is via microscopic connections
between the two (for example as illustrated in Figure 3). in
another alternative the artery and vein are in a "flow through"
configuration with the blood vessels entering at one end of a
semi-closed chamber and exiting at the opposite side (for
example as illustrated in Figure 4).
It would be appreciated by those skilled in the art
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
6
that the term "functional circulation" as used herein describes
a circulation that has at least one of the following
properties: the vessels making up the circulation are patent,
the vessels are capable of sustaining blood or blood-substitute
flowing through them, the vessels are capable of supplying
nutrients and/or oxygen to nearby tissue and the vessels are
capable of forming new blood vessels by budding.
Optionally, the chamber may also be supplied with added
extracellular matrix, for example matrix deposited by cells in
situ, reconstituted basement membrane preparations such as
MatrigelTM or laminin (mouse origin), AmgelTM, HumatrixTM, or
laminin (all of human origin) with or without matrix
metalloproteinase inhibitors, polylactic-polyglycolic acid
variants (PLGA), fibrin or plasma glue (autologous or
heterologous) with or without fibrinolysis inhibitors, or
native collagen (autologous or heterologous) with or without
collagenase inhibitors.
in a preferred embodiment, extracellular matrix-like
polylactic-polyglycolic acid sponges, DexonTm sponges, or sea
sponges are added to the chamber. Combinations of matrices,
such as PLGA sponges coated with one or more other matrix-
forming components such as fibrin, laminin, fibronectin,
collagen, low molecular weight hyaluronan and vitronectin are
other preferred options. Freeze dried segments of tissues such
as muscle or organs such as liver may be used as sources of
matrix and growth factors. Preferably the segments of tissues
or organs are taken from the same species as the donor subject,
and most preferably taken from the donor individual.
in a particularly preferred embodiment of the
invention, the donor subject is the same individual as the
recipient animal, i.e. the graft is autologous. Alternatively
the donor subject may be an immunocompromised animal,'such as
an athymic mouse or pig, and the recipient may then be a
different individual, i.e. the graft is heterologous. Other
permutations and combinations of these procedures may include
the use of either autologous or immunocompromised blood
vessels, cells, tissue segments or growth factors implanted
back into either the original donor or a different recipient
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
7
individual. Whether or not the "maturity" of the graft confers
immunoprotection on a heterologous graft is another variant
that can be tested using routine techniques.
The tissue or cells used in the chamber may be
supplemented with additional growth factors selected from the
group consisting of "homing" factors to attract stem cells from
the circulation, exogenous growth factors such as a-Fibroblast
Growth Factor (aFGF or aFGF-1,), (3-Fibroblast Growth Factor ((3
FGF-l or (3FGF-2), Platelet-Derived Growth Factor (PDGF),
Vascular Endothelial Growth Factor (VEGF-A,B,C,D or E),
Angiopoietin-1 and -2, Insulin-like Growth Factor (IGF-1),
Bone Morphogenic Protein (BMP-2 and -7), Transforming Growth
Factor-a and -(3 (TGF-a, TGF-(3), Epidermal Growth Factor (EGF),
Connective Tissue Growth Factor (CTGF), Hepatocyte Growth
Factor (HGF), Human Growth Hormone (HGH), Keratinocyte Growth
Factor (KGF), Tumour Necrosis Factor-a (TNF-a), Leukemia
Inhibitory Factor (LIF), Nerve Growth Factor (NGF), Granulocyte
Macrophage Colony Stimulating Factor (GM-CSF) and other factors
such as 3-isobutyl-l-methylxanthine (IBMX), insulin,
indomethacin, dexamethasone, hyaluronan hexasaccharide, the
PPAR-y ligand Troglitazone, nitric oxide, prostaglandin El,
transferrin, selenium, parathyroid hormone (PTH), parathyroid
hormone related peptide (PTHrP), etc, many of which are
promoters of angiogenesis or vasculogenesis. Antibodies,
agonists or antagonists to some of these growth factors or
inhibitors of the chemical mediators can also be used to
influence the type of tissue formed and the rate of its
formation. The person skilled in the art will readily be able
to test which growth factor(s), anti-growth factor antibodies,
or inhibitors, or combination thereof, are most suitable for
any given situation.
The chamber may be used with autologous or heterologous
cells, such as myoblasts transfected with Myo-D to promote
formation of the skeletal muscle phenotype, stem cells with
appropriate differentiation factors, keratinocytes seeded to
produce thin skin constructs for face and neck reconstruction,
etc. Optionally the chamber may also comprise isografted or
autologous cells selected from the group consisting. of
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
8
myoblasts, fibroblasts, pre-adipocytes and adipocytes,
cardiomyocytes, keratinocytes, endothelial cells, smooth muscle
cells, chondrocytes, pericytes, bone marrow-derived stromal
precursor cells, embryonic, mesenchymal or haematopoietic stem
cells, Schwann cells and other cells of the peripheral and
central nervous system, olfactory cells, hepatocytes and other
liver cells, mesangial and other kidney cells, pancreatic islet
(3-cells and ductal cells, thyroid cells and cells of other
endocrine organs.
Alternatively the chamber may be used with additional
autologous or isografted portions of skeletal or cardiac
muscle, pancreas, liver, epididymal and other subcutaneous fat,
nerves (peripheral, blood vessel-associated, etc), kidney,
bowel, ovary, uterus, testis, olfactory tissue or glandular
151. tissue from endocrine organs. For the purposes of the
specification the term "pieces of tissue" shall be taken to
encompass any aggregates of cells, with or without additional
extracellular material such as extracellular matrix, either
taken directly from an animal or produced as a result of
manipulation of cells in tissue culture, or a combination of
the two. In other variants such tissue segments may be rendered
ischaemic, cell-depleted or necrotic in order to provide cues
or signals to the surviving stem cells and other cells which
may influence tissue development.
Depending on the nature of the supplementation provided
to the cells, the vascularised tissue is enabled to
differentiate. in a particularly preferred embodiment, stem
cells,-together with appropriate extracellular matrix and
growth factor supplements, are supplied to the chamber in order
to produce vascularised, differentiated tissues or organs.
Suitable pluripotent stem cells can be derived from:
a) blood;
b) bone marrow;
c) specific organs or tissues, including mesenchymal
stem cells;
d) cultured cells, which may be transfected or
differentiated; or
e) placental stem cell banks.
CA 02419923 2003-02-19
WO 02/15914 PCT/AU01/01031
9
To date we have used sources such as bone marrow,
ischaemic skeletal muscle, and subcutaneous adipose tissue.
Other potential sources of pluripotent stem cells are blood,
especially from a fetus or newborn individual but also from an
adult, and human placenta. A number of stem cell banks such as
bone marrow or cord blood banks are already established. Human
embryos are a potential clinical source of stem cells, although
legal and ethical issues precludes their use at present in some
countries.
The type of differentiated cells produced depends on
the origin of the stem cells, the local environment, the
presence of tissue-specific growth or differentiation factors,
and other factors. For example, unexpectedly we have observed
that ischaemic skeletal muscle placed in the chamber with an AV
loop differentiates into predominantly adipose tissue after 4-6
weeks. Without wishing to be limited by any proposed mechanism,
we believe that in this case, mesenchymal stem cells in the
muscle, together with the stimulus of acidic ischaemic
metabolites, are potentially responsible for this
differentiation.. The chief advantage of using stem cells is
their huge proliferative capacity, so that relatively few cells
are required to generate a large colony for seeding the chamber
and the AV loop.
Preferably the vascular pedicle, such as an AV loop
comprises an artery joined to a venous graft, which is in turn
joined to a vein. Alternatively the AV loop comprises an artery
joined to a vein directly, or the AV loop comprises an artery
joined sequentially to a venous graft, an arterial graft, and a
vein. in another variant, which is useful where microsurgical
anastomosis of vessels is technically difficult or impossible,
a pedicle comprising the ligated stumps of an artery and vein
(eg. the femoral vein) placed side by side in the chamber can
be used as the blood vessel supply. in another preferred
embodiment of the invention, the AV loop vessels flow in and
out of the chamber from the same edge. In another variant the
artery and vein are neither divided nor formed into a shunt,
but instead flow in one side of the chamber and out the
opposite side (see, for example, Figure 4). In a third variant
CA 02419923 2003-02-19
WO 02/15914 PCT/AU01/01031
suitable for extremely small blood vessels, the artery and vein
are divided and placed side by side in the chamber, the vessels
both entering from the same edge; this is illustrated in Figure
3.
5 The graft portion of the AV loop may be derived from
the host or from a separate donor. Cold-stored or prefabricated
vessels may also be used.
in one preferred embodiment of the invention, an
additional step involves the incorporation of a nerve stump, so
10 that tissue in the chamber may become innervated. Skeletal
muscle, for example, requires proximity to a nerve for its
maintenance and maturity; otherwise it will,atrophy.
Preferably the chamber containing the vascular pedicle
has a defined internal dimension. The internal dimensions,
volume, and shape may be varied in order to influence the
volume and shape of the new tissue being produced. For example:
a) the internal volume of the chamber may be increased,
without altering the external size of the chamber, by providing
thinner walls;
b) the shape of the chamber may be constructed to
resemble that of the target organ or body part, such as an ear,
nose, breast, pancreas, liver, kidney, finger or other joint;
c) the degree of permeability of the walls of the
chamber may be varied; for example the chamber may include a
semi-permeable membrane component to allow selective perfusion
of molecules into and out of the chamber, or a plurality of
perforations may be placed in the walls of the chamber to allow,
an increased flow of metabolites and metabolic by-products,
growth factors and other factors that influence cell survival,
growth and differentiation between the inside and outside of
the chamber. The size, shape and number of the perforations may
be selected according to the size of the donor vascularised
tissue and the requirement to keep the contents of the chamber
isolated from direct contact with the implantation site.
Alternatively,
d) a semi-permeable component may be placed within the
chamber in order to isolate "feeder" cells from immune
reactions.
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
11
As an example of the latter, populations of fibroblasts
or other cells can be transfected, then used as a source of the
transfected gene product(s) within the chamber. This construct
is placed within a semi-permeable pocket out of contact with
the host's immune system. Drug delivery is used to switch the
transfected gene on or off. These cells will survive by
diffusion as long as they receive adequate nutrients, but will
eventually die.
The surface chemistry of the chamber walls may be
modified, in order to modify the interaction between the tissue
and the chamber wall, to provide a stimulus for differentiation
or to incorporate or be coated with a gel, such as alginate,
which mediates the slow release of a chemical or biological
agent to create a gradient.
The degree of internal support within the chamber may
be varied, eg there may be:
a) no support;
b) a solid support which directs, encourages or
inhibits the growth of the new tissue, or excludes new tissue,
or is incorporated into the new tissue;
c) a transient support based on resorbable materials;
d) a porous supporting material which supports cell and
vascular ingrowth, providing a skeleton over which the new
tissue can be generated, eg sponge-like materials such as blown
PTFE materials, PLGA sponges of variable composition and
porosity, etc;
e) a support formed from materials which direct tissue
differentiation, such as hydroxyapatite or demineralised,
granulated bone.
Preferably the exterior surface of the chamber bears a
means by which the chamber can be attached and/or immobilised
to the desired region of the body.
In a second aspect, the invention provides a
vascularised tissue graft, ie. the contents of the chamber,
comprising differentiated tissue or an organ with a mature
vascular supply.
Preferably the graft predominantly comprises tissue
CA 02419923 2003-02-19
WO 02/15914 PCT/AU01/01031
12
selected from the group consisting of adipose tissue,
cartilage, bone, skeletal muscle, cardiac muscle, loose
connective tissue, ligament, tendon, kidney, liver, neural
tissue, bowel, endocrine and glandular tissue. More preferably
the graft predominantly comprises vascularised adipose tissue,
skeletal muscle, cartilage or bone tissue or tissue comprising
pancreatic islet and/or ductal cells, kidney cells or liver
cells.
In a third aspect, the invention provides a method of
repairing a tissue deficit, comprising the step of implanting a
tissue chamber according to the invention into a patient in
need of such treatment, in which:
a) the tissue or "organ" graft is formed according to
the methods of the invention, and;
b) retained for sufficient time to mature ie. to
achieve the desired size, vascularity and degree of
differentiation, and;
c) transferred to the desired recipient site; and
d) the blood vessels of the graft are microsurgically
anastomosed to a local artery and vein.
For the purposes of the specification, the term "tissue
deficit" will be taken to comprise a shortfall in the normal
volume, structure or function of a tissue in the recipient.
Such a tissue may be selected from, but is not limited to
superficial tissues such as skin and/or underlying fat, muscle,
cartilage, bone or other structural or supporting elements of
the body, or all or part of an organ. The augmentation of
otherwise normal tissues for cosmetic purposes, such as forms
of breast augmentation, is also provided by the invention. A
person skilled in the art will readily recognise that such a
tissue deficit may be a result of trauma, surgical or other
therapeutic intervention, or may be congenitally acquired.
In a fourth aspect, the invention provides a method of
providing a subject with a gene product, comprising the steps
of:
a) constructing a tissue chamber according to the
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
13
invention to create vascularised tissue from a patient in need
of such therapy;
b) removing the chamber with its vascularised tissue
and culturing the chamber assembly in vitro;
c) transforming cells of the tissue in the chamber with
a desired gene; and
d) implanting the chamber or the contents minus its
chamber into the patient.
The timing of the genetic transformation of the tissue-
producing cells can be varied to suit the circumstances, for
example the cells may be transformed at the time of setting up
the chamber construct, during the incubation, or immediately
prior to transplantation.
The provision of gene products can take several forms.
One example is the transfection of myoblasts with the Myo-D
gene to create tissue with a normal skeletal muscle phenotype.
Such transfected cells may then be seeded into the desired
chamber, matrix and AV loop to generate vascularised skeletal
muscle. This may have implications for the treatment of
muscular dystrophy and other genetically inherited muscle
diseases. A second example is the transfection of pancreatic
islet cells with a "healthy" phenotype and their seeding into
the chamber. This approach may prove to be useful in the
treatment of diabetic patients. in a third example, cells are
transfected with a growth factor gene or an angiogenesis-
promoting gene, such as PDGF, bFGF or VEGF, prior to seeding
them into the chamber together with the AV loop and selected
matrix. This continuous production of growth factor is
designed to speed up the rate of development of, and the rate
of new blood vessel formation within, the new tissue/organ.
In a fifth aspect, the invention provides a model
system for vascularised tissue, comprising a tissue chamber
containing a vascular pedicle of the invention and optionally
an extracellular matrix, operably connected to an
extracorporeal circulation apparatus and renal dialysis filter.
The extracorporeal circulation apparatus and renal dialysis
filter may be of any suitable conventional type. The cells
CA 02419923 2003-02-19
WO 02/15914 PCT/AU01/01031
14
forming the tissue in the chamber are optionally transformed so
as to express a heterologous gene. This model system may be
used for culturing, recruiting, growing and studying the
behaviour of stem cells or tissue containing precursor cells,
either in vitro or in vivo. Because of the ability to alter
the environment of the chamber with added growth,
differentiation and chemical factors, it is possible to produce
a wide variety of tissues and organs by this process.
The ability to generate autologous vascularised tissue
of a defined composition and at any anatomical site in the body
where it is possible to create an arterio-venous loop or
suitable vascular pedicle has many other applications. At its
localised site the tissue in the chamber may, for example, be
manipulated by
a) gene transfection,
b) administration a local drug or other "factor", or
c) creating a site of circulatory stem cell homing.
Furthermore, the tissue and exudate in the chamber may
readily be harvested to monitor progress of tissue growth and
development. Above all, it is the ability to grow and
transplant new vascularised, differentiated tissues or
organoids that sets this invention apart from others.
For the purposes of this specification it will be
clearly understood that the word "comprising" means "including
but not limited to", and that the word "comprises" has a
corresponding meaning.
BRIEF DESCRIPTION OF THE FIGURES
Figure 1 illustrates how the femoral artery and vein
are anastomosed microsurgically to a vein graft of similar
diameter to form a loop (shunt). The AV loop is placed as shown
in a plastic chamber (made of polycarbonate or poly-L-lactic
acid, etc), the lid secured, and the chamber optionally filled
with an extracellular matrix with or without added cells or
growth factors. The chamber is anchored in position relative to
the surrounding tissue by means of stay sutures through
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
external holes.
Figure 2 shows a configuration similar to Figure 1,
except that the lid of the chamber is dome-shaped and the edges
5 of the chamber are more rounded to minimise wound breakdown.
Figure 3 depicts an example of the thin-walled chamber
used for the pedicle model. in this case an artery and a vein
are ligated distally and placed adjacent to each other.
10 Microscopic connections between the artery and vein become
established, and form an AV loop in a similar manner to that
shown in Figures 1 and 2.
Figure 4 shows a model chamber similar to that in
15 Figure 3, but with exit holes for the blood vessels at either
end of the chamber. This allows an undivided, dissected length
of blood vessels, placed side by side, and in some variants
surrounded with extracellular matrix, to form new tissue.
Figure 5 shows the inner aspect of an AV loop-
containing chamber, 7 days after insertion. Fluorescence
microscopy shows labelled fibroblasts evenly distributed across
the chamber surface, magnification x 160 (see Example 2).
Figure 6 shows a reconstructed "breast" on a male
rabbit, constructed using a vascularised, tissue-engineered fat
and connective tissue flap created at a remote site (the groin
region) in the same rabbit (see Example 10).
DETAILED DESCRIPTION OF THE INVENTION
The invention will now be described in detail by way of
reference only to the following non-limiting examples and
drawings.
Experimental Procedures
Preparation of Tissue Chamber
A custom-made polycarbonate chamber was prepared. it
has a top and a bottom, and when the two halves are sealed
CA 02419923 2003-02-19
WO 02/15914 PCT/AU01/01031
16
together the internal volume is 0.45-0.50 ml. The general
construction of the chamber is illustrated in Figure 1.
The basic chamber for use in rats is made of
polycarbonate. in one variant the chamber is made of polylactic
acid or PLGA. The chamber is in the shape of a cylinder of
external dimensions 14 mm diameter and 4 mm high, with a saw
cut on one side to create an opening for the blood vessel entry
and exit. Another variant has cut openings on opposite sides of
the chamber to allow blood vessels to flow in one side and out
the other. The chamber has a base and a removable lid. The base
has holes to allow anchoring of the chamber to subcutaneous
tissue. The internal volume is approximately 0.45-0.50 ml. The
internal volume of this basic chamber can be varied,
maintaining the same external volume, by using thinner walls,
which may even be as thin as a standard plastic film used in
food storage. An alternative design is in the shape of a "dome"
with more rounded edges, as shown in Figure 2. Other variants
include an elongated, flattened cigar shape as shown in Figure
3 which fits readily into the subcutaneous space in the groin.
For the purposes of specific grafts, the shape of the chamber
may be designed to mimic the shape or contours of a particular
body part, for example a human finger joint or thumb, human
ear, human nose, human breast, etc.
The size of the chamber can be scaled up or down to
suit the size of the host. Hence the internal volume for a
chamber to be used in a mouse may be approximately 0.1-0.2 ml,
in a rabbit 10-12 ml, but in a human can be up to approximately
100-200 ml.
The chamber may optionally be sealed. In the standard
version the opening allows limited contact with the surrounding
tissue and total uninterrupted contact with the blood supply.
In a sealed variant, the opening is engineered to allow just
enough space for the ingoing artery and outflowing vein without
crushing the blood vessels. The vessel ports are sealed, for
example with fibrin glue, to avoid contact of the developing
graft with surrounding tissue.
The surface of the polycarbonate chamber can be left in
its native hydrophobic state, or can be rendered relatively
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
17
more hydrophilic by the use of polylactic acid or the pre-
treatment of polycarbonate with a thin film of poly-L-lysine.
in one useful configuration, the surface of the chamber
comprises,a plurality of perforations, allowing increased
contact with growth factors in the surrounding tissue. The size
and shape of the perforations may be tailored to optimise the
passage of the desired factors, while minimizing or preventing
the passage of cells.
if the chambers are made of glass or Pyrex they can be
coated with silicone.
The chamber design should ideally fit comfortably into
the recipient site, and should be of a rounded shape and of a
sufficiently small size to avoid wound break down.
The internal contents of the chamber are sufficiently
large to accommodate an osmotic pump (eg. an A1ZetTM osmotic
mini pump) to deliver drugs, growth factors, antibodies,
Inhibitors or other chemicals at a controlled rate. in one
alternative method of drug/factor delivery, the osmotic pump
may be placed subcutaneously outside the chamber with a plastic
tube leading from the pump placed inside the chamber, eg. at
the centre of the AV loop.
Creation of an Arteriovenous (AV) Shunt Loop inside the
Tissue Chamber
The basic model has been described by Tanaka et al
(1996). Briefly, male Sprague-Dawley rats (225-285 g) were
anaesthetised with intraperitoneal phenobarbitone (50 mg/kg;
2.5 ml of a 6 mg/ml solution). Under sterile conditions an
inferior-based flap was created in the right groin to expose
the femoral vessels from the inguinal ligament to the
superficial epigastric branch. A longitudinal incision was made
in the left groin to harvest the left femoral vein from
inguinal ligament to the superficial epigastric branch. This
vein graft (approximately 1.5 - 3 cm long; usually 2 cm) was
interposed between the recipient right femoral vein and artery
at the level of the superficial epigastric artery by
microsurgical techniques using 10-0 sutures. The shunt was
placed into the chamber, the lid closed and the construct
sutured to the groin musculature with the aid of small holes on
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
18
the base of the chamber. An adipose layer was placed over the
chamber and the wound closed with 4-0 silk sutures.
The growth chambers with the AV shunts were harvested
at either 2, 4 or 12 weeks post implantation.
Assessment of Vascularisation and Tissue Creation
At the specified time of exploration, the chamber was
opened, and the vessels cleaned and tested for patency. The
vessels were tied off with a 5-0 silk suture at the entrance of
the chamber and the flap harvested. in 2 of the 5 rats in each
group the flap was perfused, via the aorta, with India ink
prior to harvest (details below). The flaps were assessed for
volume and weight and placed in buffered 10% formal saline
(BFS) for histological examination. The animals were
sacrificed with an intracardiac dose of sodium pentabarbitone
(-3 ml of 250 mg/ml solution) at the completion of the
exploration.
Tissue mass and volume
The tissue in the chamber was removed and its wet
weight and volume recorded. The volume of the tissue was
assessed by a standard water displacement technique. The tissue
was suspended by a 5-0 silk suture in a container of normal
saline which had been zeroed previously on a digital balance.
Care was taken not to touch the container with the specimen.
The weight recorded was the volume of the tissue specimen (with
a density equal to that of normal saline, 1.00 g/ml). The mass
of the specimen was assessed at the same time on the same
digital scale by allowing the tissue to rest on the base of the
container, and recording the weight.
India ink perfusion
In order to perfuse the flaps with India ink, the
abdomen was opened via a midline incision. The intestines were
gently retracted to the periphery and the periaortic fat
stripped away. The proximal aorta and inferior vena cava were
ligated. The aorta was cannulated with a 22-gauge angiocatheter
which was secured with a distal suture around the angiocatheter
and aorta. A venotomy was carried out in the inferior vena
cava. The aorta was perfused with 10 ml of heparinised saline
to flush out the retained blood, the animal was sacrificed with
CA 02419923 2003-02-19
WO 02/15914 PCT/AU01/01031
19
intracardiac sodium pentabarbitone (3 ml of a 250 mg/ml
solution), the aorta infused with 3 ml buffered 10% formol
saline (BFS) and then with 5 ml India ink in 10% gelatin. The
flap vessels were then tied off. Tissue from the chamber was
removed, fixed in BFS, cleared in cedar wood oil and the
pattern of vessels visualised microscopically using transmitted
light and image analysis (Video Prom imaging).
Histology
Specimens were fixed in buffered formol saline and
embedded in paraffin. Sections (5 m) were cut and stained with
either haematoxylin & eosin (H & E) or Masson's Trichrome.
Example 1 - Creation of Vascularised Tissue in Chambers
with an AV loop
Three groups of five rats each were used. Each group
had an identical procedure performed as described above, and
the growth chambers with the AV shunts were harvested at either
2, 4 or 12 weeks post implantation.
The average mass of the AV shunt vessels prior to
insertion was 0.020 g (exsanguinated) and 0.039 g (when full of
blood). Two weeks after insertion the AV shunt and its
surrounding tissue weighed 0.18 0.03 g. The mass increased
progressively being 0.24 0.04,g at 4 weeks and 0.28 0.04 g
at 12 weeks. The volume of the new tissue closely paralleled
its weight. The increase in weight but not volume between 2 and
12 weeks was statistically significant (P<0.05, ANOVA/Dunnett's
test).
Two weeks after implantation the AV loop was surrounded
by a mass of coagulated exudate containing varying amounts of
clotted blood. At 4 weeks the mass of tissue around the loop
was larger and firmer, especially in its central part. By 12
weeks the newly formed tissue surrounding the loop-had
increased still further in volume and now filled approximately
two-thirds of the chamber. The surface coagulum was no longer
visible, and the whole mass had a uniformly firm consistency.
After 2 weeks of incubation the AV shunt was surrounded
by a cuff of newly-formed connective tissue composed of
fibroblasts, thin collagen fibres and vascular sprouts,
CA 02419923 2003-02-19
WO 02/15914 PCT/AU01/01031
arranged roughly vertical to the shunt. inflammatory cells,
both neutrophils and macrophages, were present in moderate
numbers in the outer part of the newly formed tissue and in the
surrounding mass of coagulated inflammatory exudate. In
5 occasional sections, branches of newly-formed blood vessels
arising from the venous lumen of the AV shunt could be
identified.
In the 4 weeks incubation group, the newly formed
tissue was more mature. The zone closest to the AVS contained a
10 dense plexus of newly formed vessels embedded in mature
collagenous stroma. Outside this layer was a less mature zone
similar to the newly formed tissue in the 2 weeks specimens.
Most of the surrounding coagulum was no longer visible, and
only small numbers of inflammatory cells were present in the
15 newly formed tissue. As at 2 weeks, communications between the
AV shunt and the newly formed vessels were visible in some
sections.
Twelve weeks after incubation, the newly formed tissue
had matured still further, and consisted of dense collagenous
20 connective tissue with fibroblasts aligned parallel to the
outer margin of the AV shunt. There was no apparent decrease in
vascularity and newly formed vessels formed a dense plexus
throughout the connective tissue. Few inflammatory cells were
visible.
At all three time points, the specimens which were
injected with India ink gave a clearer picture of the extent
and density of the newly formed vasculature. In most specimens
almost all vessels contained carbon in their lumen, indicating
that they communicated with the AV shunt.
Ideally, newly formed tissue must be stable and capable
of retaining its shape. The tissue formed around an AV loop has
both these characteristics. At 2 weeks the mass within the
chamber is soft and readily deformed. By 4 weeks it is firmer
and more rigid, and at 12 weeks it has the physical
characteristics of mature connective tissue. Surprisingly,
growth is continuous for at least 12 weeks after implantation,
with no indication of resorption or regression of the newly
formed tissue with increasing maturity.
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
21
Example 2 - Chambers with rat dermal fibroblasts
Culture of rat dermal fibroblasts
Rat skin was harvested in a 6 cm by 4 cm ellipse from
the groin area of an inbred Sprague-Dawley rat line (Monash
University Animal Services, Clayton, Victoria, Australia). The
inbred line comprised animals resulting from at least 20
generations of brother-sister matings.
The epidermis was trimmed off. Segments of dermis were
cut into 2mm by 2mm squares and 10 pieces were placed onto a
sterile Petri dish and attached to the base using rat plasma
"glue". This glue was made by the addition of 2 ml of rat
plasma, prepared from Sprague Dawley rats, to 0.3 ml of 2%
calcium chloride. The glue was allowed to set for 10 min at 37
C. Complete culture medium, comprising Dulbecco's Modification
of Eagle's Medium (DMEM), 10% fetal calf serum, penicillin and
streptomycin and glutamine, was added to the culture dish. The
skin segments were left undisturbed for 7 days, then the medium
was changed. There was considerable outgrowth of fibroblasts
by 10 days, at which time the skin segments were removed. The
fibroblasts were subcultured twice at weekly intervals, each
time growing the cells in 75 cm2 and 175 cm2 culture flasks
respectively.
The fibroblasts were labelled with two fluorescent
labels, bisbenzamide (BB) and carboxyfluorescein diacetate
(CFDA). Three ml of 0.1% trypsin in phosphate buffered saline
(PBS) at pH 7.4 was added to a 175 cm2 cell culture flask
containing confluent fibroblasts for 5 min at 37 C. The
trypsin was neutralized by the addition of 17 ml of complete
DMEM media. The cell suspension was centrifuged at 2000 x g
for 10 min. The cell pellet was resuspended in 3 ml of media
and the suspension transferred in three 1 ml aliquots to
Eppendorf tubes. To each Eppendorf tube 13.5 l of a 10% CFDA
solution and 20 l of BB were added. The tubes were incubated
for 1 h at 37 C and shaken gently every 15 minutes. The cells
then were transferred into a 175 cm2 flask and recultured.
CFDA persists in the cytoplasm of cultured cells and survives
the division of cells into daughter cells. CFDA fluoresces
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
22
maximally at 513 run; BB fluoresces maximally at >430nm.
Labelled cells were protected from light, in an effort to
maintain maximal fluorescence.
Cell Counting
Prior to the addition of cells to the chambers, the
fibroblast culture flasks were trypsinized and the trypsin
neutralized. 10 l of suspended cells were counted using a
hemocytometer, and 0.05% Evan's blue dye in a 1:10 ratio. The
solution was centrifuged and the resulting cell pellet
suspended in an appropriate volume of bovine collagen solution
to yield a cell concentration of 1 million cells/ml.
Rat tail tendon collagen (RTTC)
The tendons from six rat tails were harvested and diced
into 2x2x2 mm cubes (yield approximately l0g). Four hundred ml
of cold 0.5 M acetic acid was added and the mixture homogenized
and left stirring at 4 C for 24 h.' The homgenate was
centrifuged (3000 rpm x 20 min) and the supernatant harvested.
This extraction procedure was repeated twice with further
additions of 300 ml of cold 0.5 M acetic acid. To the pooled
extracts a solution of 5 M NaCl was added slowly, with magnetic
stirring at 4 C, until the final concentration of salt was
approximately 0.7 M (100 ml of 5M NaCl added to every 600 ml of
extract). The solution was left for 1 h to allow full
precipitation of the native collagen. The precipitate was
collected by centrifugation (3000 rpm x 20 min at 4 C),
redissolved in 200 ml of 0.5 M acetic acid and dialysed twice
against 2 1 of cold 0.5 M acetic acid for 24 h, and twice
against sterile, cold distilled water, the final dialysis
solution containing a few drops of chloroform on the surface.
This results in a sterile stock solution of RTTC of
approximately 3 mg/ml, the concentration checked by a Bradford
protein assay (Bio Rad) with a Type I collagen standard.
Preparation of chambers
All procedures were carried out in a cell culture hood
using sterile technique. Chambers were coated internally with
RTTC by addition of 200 it of 2.5 mg/ml RTTC solution, pH 7.4,
to each half chamber. Chambers were incubated for 1 h at 37 C
to allow gel formation and dried for 24 h. After rinsing with
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
23
PBS to remove residual salt crystals, 0.25 x 106 of
fluorescently labelled fibroblasts in 150 l of complete DMEM
were added to each half chamber. After allowing 1 h for
adherence of the cells, chambers were immersed in complete DMEM
and incubated at 37 C under 5% CO2 in air for 24 h. The density
of labelled of cells was determined by counting the number of
cells in 7 randomly selected fields of each half chamber using
a x10 objective.
Insertion of chambers
Two groups of 6 inbred male Sprague-Dawley rats,
weighing between 230 - 280 g, were used. Two chambers were
inserted into the inguinal region of each rat, the chamber in
the right side containing an AV shunt (prepared as described
above) and that in the left side containing no shunt. In 6 rats
chambers were removed 2 days after implantation. The remaining
6 chambers were removed 7 days after implantation.
Examination of chambers after removal
The chamber was removed, the AV shunt examined for
patency and the flap removed. Ten l of 0.05% Evan's blue dye
was added to each half chamber and incubated for 5 min at 37 C.
The base of each half of the chamber was then examined, using a
x 10 ocular, to determine the number of Evan's blue-stained and
fluorescent cells in 7 randomly selected microscopic fields.
The number of labelled cells in 7 random fields on the surface
of the AV shunt was then determined.
Two days after insertion the shunt and surrounding
tissue covered approximately 20% of the surface of the chamber;
by 7 days this had increased to approximately 30%. On this
basis the overall density of cells in the chamber containing an
AV shunt was calculated by summation of the density of cells on
the surface of the chamber and 20% (2 days) or 30% (7 days) of
the labelled cells on the surface of the AV shunt.
Paired t-tests were used to compare number of cells per
grid in the control and experimental chambers and the
preoperative number of cells per grid using Microsoft Excel
and Graph Pad Prism software (San Diego, CA, USA).
After counting, the shunt and surrounding tissue was
fixed in 10% formol saline, embedded in methacrylate and thin
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
24
sections prepared and stained with either haematoxylin and
eosin or Masson's trichrome.
Comparison between labelling with bisbenzamide (BB) and
carboxyfluorescein diacetate (CFDA)
In both in vitro cultures and the in vivo chambers the
number and distribution of-labelled cells at the two
wavelengths examined (430 nm for BB; 573 nm for CFDA) was the
same. No cells were identified as being labelled with only one
fluorescent dye. Hence in the results which follow "fluorescent
cells" refers to cells labelled with both BB and CFDA.
Macroscopic findings
The AV loop was patent in every chamber.
Two days after insertion the AV shunt covered
approximately 20% of the surface of the chamber. By 7 days the
area covered by the AV shunt and new tissue arising from it had
increased to approximately 30%.
The 2 day mean weight of the shunt was 0.12 0.017 g
and the mean volume was 0.12 0.014 ml. By 7 days the mean
weight had risen to 0.23 0.018 g and the mean volume to 0.21
0.015 ml.
Density of the labelled cells
The density of the labelled cells in empty and AV shunt
containing chambers is shown in Table 1.
Table 1
Density of labelled cells,(mean number/grid) in empty
and AV shunt containing chambers, pre-operatively and 2 and 7
days after insertion.
Time after Pre- Empty AV Shunt-containing
Insertion operative Chamber chamber
In chamber Total*
2 days 8.6 1.74 4.0 0.94 4.8 0.59 5.7 0.62
7 days 10.2 1.7 4.8 1.3 11.7 1.4 15.5 1.14
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
* Total density calculated as number of cells / grid on
chamber surface plus 20% (2 days) or 30% (7 days) of labelled
cells in surface of tissue surrounding the AV shunt.
# Increase above density in 7 day empty chambers is
5 significant (p = 0.011).
It can be seen that in all chambers the cell density
decreased in the early stages after implantation, the values in
all 2 day chambers being less than their pre-insertion density.
10 Two days after insertion there was no significant difference in
the density of cells in empty and AV shunt-containing chambers.
At 7 days the density of the cells in empty chambers
did not differ significantly from the density 2 days after
insertion. In contrast, the cell density in AV shunt containing
15 chambers increased to almost 3 times its 2 day value, and both
the density of cells in the grid and the density (after
allowing for the number of labelled cells in the tissue
surrounding the shunt) were significantly greater than the
density in empty chambers (p= 0.013).
20 Evan's blue staining showed that in all chambers
examined virtually all labelled fibroblasts were viable, with
less than 1% of cells taking up the Evan's blue dye.
Histological Findings
After 2 days incubation the vessels of the AV shunt
25 were surrounded by blood clot and coagulated inflammatory
exudate. Small numbers of fibroblasts were visible migrating
from the vascular adventitia into coagulum.
By 7 days, many more fibroblasts were present within
the coagulum, and early vascular sprouts were visible arising
from the outer aspect of the AV shunt.
At both 2 and 7 days fluorescent studies showed
labelled fibroblasts on the surface of the coagulum surrounding
the AV shunt, but labelled cells were not seen within its
substance. The inner aspect of an AV shunt-containing chamber
removed 7 days after insertion is shown in Figure S.
Example 3 - Differentiation of stem cells in implanted
tissue chambers
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
26
Skeletal muscle, pancreas, fat, liver and kidney were
aseptically removed from four inbred Sprague-Dawley rats. They
were chopped into 1 mm cubes and placed in a tissue culture-
grade petri-dish (15-20 pieces each 7 cm2 of culture surface)
containing 1-2 ml of complete serum-free DMEM. They were then
incubated for a minimum of 24 h and up to 3 days. At the
appropriate time 4-6 pieces of tissue were adhered in a plasma
clot to each side of a chamber of the type described in Example
2. The chamber was then seeded with the AV loop and closed.
The proximal end of a femoral nerve was placed inside one half
of the chambers containing skeletal muscle explants. After 4-6
weeks the rats were sacrificed and the chambers examined.
After 4-6 weeks, the contents of chambers with tissue
explants differed from the contents of chambers without tissue
explants, in that they contained new and different cell
phenotypes. In all cases most of the necrotic tissue explants
had been replaced by clumps of new cells.
In the most dramatic of these experiments, 8 of the 11
chambers seeded with skeletal muscle explants contained up to
two thirds of their volume with mature, well-vascularised
adipose tissue together with mature skeletal muscle fibres,
surrounded by a thin capsule. The mature region of the new
tissue contained up to 90% vascularised adipose tissue. The
remaining chambers also had a lesser proportion of mature
adipose tissue and skeletal muscle fibres.
The chambers seeded with portions of pancreatic tissue
had a large population of well-demarcated large ovoid
eosinophilic cells, many giant cells and other smaller cells.
Without wishing to be limited by any proposed
mechanism, we believe that a "stem cell" population, either
attracted into the chamber from a circulating stem cell source
by the necrotic tissue explants, or contained within the tissue
explants, has given rise to the new tissue. in either case a
very small amount of explant tissue was used, in comparison to
the large amount required to isolate stem cells, and our
results indicate that this is a novel and efficient method to
obtain stem cells. The stem cells may have differed with
respect to their degree of commitment to a particular tissue
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
27
type, or else they may have responded to cues expressed by the
unique microenvironment of the different explants, to
proliferate and differentiate into the different cell types
observed.
The generation of encapsulated adipose tissue described
here is, to our knowledge the first time that such a neo-
organoid has been grown de novo on its own artery and vein.
A detailed study of the spatio-temporal and dynamic
changes in the chamber and the mechanism by which these events
give rise to the neo-organ may also have applications in
defining in vivo stem cell availability and behavior. The
chamber model is superior to any other in vivo model available
so far, since it enables a wide variety of manipulations of the
chamber contents and environment and stem cell sources.
Furthermore, it enables a study of stem cells in a naive
,environment without the influences of other nearby tissues, as
opposed to the growth of stem cells in an established tissue.
The finding that muscle explants can result in the
generation of a neo-organ, consisting almost entirely of mature
adipose tissue, indicates that:
a) a stem cell population can successfully seed the
chamber;
b) the chamber model supports the plasticity of stem
cells;
c) a satisfactory, appropriate and adequate
neovascularisation develops with, integrates and supports the
tissue construct;
d) the constructs are not overcome by fibroblastic in-
growth; and
e) the constructs are not overcome by inflammatory
cells.
These results demonstrate that application of the
chamber model to tissue engineering is feasible, and represent
a significant advance in the art of -tissue engineering".
Example 4 - Effect of Matrigel
A pilot study was devised to determine if there was any
initial loss of Matrigel during 20 minutes of contact with the
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
28
AV loop. Based on the results of the pilot study, time periods
of 2, 4 and 8 weeks were chosen. At the 4 week time period a
further comparison was done with growth factor-reduced
Matrigel. Six male Sprague-Dawley rats were used per group,
each weighing between 220 and 280 g. The arterio-venous loop
procedure was carried out as described in the Experimental
Procedures.
Matrigel (Collaborative Research Inc, Bedford, MA, USA)
was divided into in sterile 10 ml aliquots at an approximate
concentration of 12 mg/ml in DMEM containing 10 g/ml of
Gentamycin (Becton Dickinson). The Matrigel was stored at -20 C
and prior to use was thawed overnight at 4 C. Throughout the
preparation process the Matrigel was kept on ice and
manipulated using pre-cooled pipettes. Growth factor reduced
(GFR) Matrigel was prepared from Matrigel essentially as
described by Vikicevic et al (1992). This involved an
additional fractional ammonium sulphate step. The protein
concentration of the resultant GFR Matrigel was verified by
Bradford protein assay and by Coomassie blue staining after
SDS-PAGE to be consistent with that of normal growth factor-
replete Matrigel.
Under sterile conditions, 0.5 ml of Matrigel was added
to each sterile chamber at room temperature where it gelled
rapidly (within 15 seconds). The chamber with matrigel was then
placed in position in the rat's right groin. The Matrigel is
gelatinous at room temperature, enabling immersion of the loop
within it. In the pilot study the AV loop was made and immersed
in the Matrigel for 20 minutes before implantation, to
determine whether there was any initial loss of Matrigel from
the chamber due to liquefaction of the matrix.
For the time course studies the new tissue flaps were
harvested at 2, 4 and 8 week periods. The flaps were harvested
at the above time periods, and assessed for weight, volume and
histology. Statistical analysis was carried out comparing the
2, 4 and 8 week groups with each other and the AV loop alone
(See Example 1). A further comparison was done at 4 weeks
between Matrigel, GFR Matrigel and the AV loop alone at 4
weeks.
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
29
In the pilot study Matrigel proved easy to manipulate
in vitro. There was minimal loss of Matrigel after 20 minutes
of contact with the AV loop.
in an AV loop alone (no added matrix), the average
weight of the new tissue flap formed after 4 weeks was 0.24
0.04 g, and the average volume was 0.23 0.03 ml. These
results acted as the control for this experiment and Example S.
At two weeks the average weight of flap in chamber
supplemented with Matrigel was 0.32 0.03 g and volume was
0.30 0.03 ml. This was significantly greater then the 4 week
loop alone flap (p=0.05). At four weeks the flaps were slightly
heavier than the 2 week flaps, with an average weight of 0.35
0.03 g and a volume of 0.33 0.03m1. A comparison of these two
groups showed no statistical significance. The weight (p=0.01)
and volume (p=0.01) were both significantly greater than the
control flaps produced by loop alone.
At 8 weeks the flaps had regressed, with an average
weight of 0.18 g 0.02 g and volume of 0.16 ml 0.02 ml.
Statistical analysis reveals that this is highly significant in
weight (p=0.002) and volume (p=0.001) when compared with both
the two week flaps and the four week flaps weight (p=0.0005)
and volume (p=0.0003). For this longer time course 8 rats were
operated on to compensate for infection or dehiscence. No such
problems were encountered, so all 8 have been included in the
analysis.
The GFR Matrigel flaps were smaller than the normal
Matrigel flaps at 4 weeks, weighing on average 0.27 0.02 g. A
comparison of weights showed no statistical significance. The
volume was 0.24 0.01 ml; this was significantly less than the
normal Matrigel (p=0.04). The GFR flaps were still larger than
the loop alone at the same time period (not statistically
significant). One of the chambers became infected, and had to
be removed. As a consequence there were 5 animals examined in
this group.
At 2 and 4 weeks a significant flap of tissue had
formed when compared to chambers containing the loop alone at
day 0. There was residual Matrigel in the chamber, and strands
of microvessels were visible running from the flap edge into
CA 02419923 2003-02-19
WO 02/15914 PCT/AU01/01031
the Matrigel. Microfil injection demonstrated good filling of
flap vessels, including the advancing microvessels. This
appearance was not apparent at 8 weeks, when the flaps were
smaller and with a more regular smooth surface. At 8 weeks
5 there was only residual fluid in the chamber, and no viscous
Matrigel was visible.
Histological examination showed that at 2 weeks there
were many immature vessels extending to the flap edge, with
haemorrhage within the peripheral tissue. There was early
10 collagen formation in the central portion and areas of
unincorporated Matrigel within the flap.
At 4 weeks the vessels had matured into arterioles and
venules, with larger branching vessels arising from the loop
and smaller branches at the periphery. There was still some
15 unincorporated Matrigel and small amounts of haemorrhage. The
unincorporated Matrigel contained sparse fibroblasts and the
occasional vessel. The general impression was of a maturing but
still growing flap with good vessel formation.
At 8 weeks the flap tissue appeared more mature, with
20 denser collagen and larger vessels nearer the loop. It was less
cellular with less vessels. A capsule had started to form
around the generated tissue, and there was residual Matrigel
remaining within the flap.
The GFR Matrigel flaps appeared to be more mature, with
25 larger vessels in the centre and less active angiogenesis at
the periphery. There was evidence of early capsule formation
and in some specimens more inflammatory cells were present.
At all time courses Microfil injection demonstrated
good vascular connection between the loop and the flap vessels.
Example 5 - Effect of Poly-L-lactic Polyglycolic Acid
(PLGA)
(a) PLGA prepared by the salt-leached method.
A PLGA insert for the tissue chamber was constructed
using a particulate leaching method as described by Patrick et
al (1999). In essence PLGA is dissolved in chloroform and mixed
with NaCl. After evaporation of the chloroform the resulting
scaffold is machined to the desired shape. The salt was then
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
31
leached from it leaving interconnected pores. The pore size is
a reflection of the size of the salt particle used. In this
experiment pores of 300-400 pim and a porosity of 84% were made.-
The PLGA was machined in two parts so.as to fit inside the
polycarbonate chamber. The lower part comprised a base plate
containing a groove for the loop and the upper part comprised a
flat disc to cover the loop and base plate. The PLGA discs
were 1.4 mm in diameter by 2.5 mm thick. The PLGA was
sterilised and pre-wetted by soaking in 100% alcohol for 30
minutes on a mechanical stirrer then subjecting them to three
30 minute washes in sterile saline washes, also on a mechanical
stirrer.
The arteriovenous loop was prepared as described above,
and placed into the base plate of PLGA sitting in the chamber.
The superior disc was placed on top and the chamber closed.
Each group of rats contained 6 male Sprague-Dawley rats, with
each rat weighing between 220 and 280 grams. The chambers were
harvested at either 2 or 4 weeks. Weight, volume and histology
were assessed at both time periods. Immunohistochemical
staining of flap sections for a-actin was carried out to detect
myofibroblasts. in each group, one chamber was excluded, one
due to infection and the other to dehiscence, leaving 5 rats in
each group.
At 2 weeks the vessels had almost entirely vascularised
the construct, with some uninvolved PLGA at the tip. The
capsule had begun to form proximally near the portal. At 4
weeks the construct was entirely encapsulated, and had shrunk
and retracted, withdrawing from the sides of the chamber.
Micro-fill injection demonstrated the extent of vessel
penetration.
The 2 week flap weight was 0.43 0.05g and the volume
0.38 0.04 ml. The 4 week flap weight was 0.33 g 0.04 g and
the volume 0.29 ml 0.04 ml. A comparison between the 2 and 4
week groups showed a reduction in flap size between 2 and 4
weeks. This result was not statistically significant. Further
comparison with other experiments was not possible due to the
presence of PLGA retained within the flap, which skewed the
results.
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
32
At both 2 and 4 weeks there was extensive vessel
outgrowth, with branching vessels found up to the edge of the
PLGA. Arterioles had formed, and healthy branching angiogenesis
was seen coming from the loop. The cellular infiltrate was
lying on the matrix and on,the surface of the structure. A
capsule had formed on the proximal part of the flap only at 2
weeks. a-Actin stain showed that this capsule contained
myofibroblasts. At 4 weeks the capsule was thicker proximally,
with more myofibroblasts and had extended to encompass the
whole flap.
(b) PLGA prepared by a fiber-spun method.
The vascular loop model described in Example 1 was used
in this experiment. The AV loop was placed within a round
polycarbonate chamber (0.5 ml volume) filled with a PLGA disc
(75% poly-L-lactic acid/25% polyglycolic acid) as the scaffold.
The PLGA scaffold was either manufactured by the salt leaching
method described above or a fiber spun technique. Each group
comprised five animals. After 4 weeks incubation and
immediately before harvest heparinised India Ink was infused
i.v. for 5 min. Tissue from the chamber was harvested, fixed in
buffered 10% formalin, paraffin embedded, cut into 5 m
sections and stained with haematoxylin & eosin (H & E) for
evaluation.
The salt-leached PLGA was less dense than the hard,
dense consistency of the fiber-spun PLGA. This was evidenced by
the subsequent cutting of the tissue/PLGA blocks for
histological evaluation. The salt-leached PLGA was brittle and
prone to crumbling. The fiber-spun PLGA was easy to section as
it had a solid consistency and did not crumble.
Histological examination showed a consistent pattern
for all specimens in their respective groups. in the salt-
leached PLGA group, considerable invasion into the PLGA by
microvasculature and new tissue was found throughout, with
numerous India Ink filled microvessels evident. The fiber-spun
PLGA'differed in character. The neovascularization and new
tissue formation developed predominantly in a two dimensional
plane. Initially, instead of invading the PLGA, tissue
preferentially surrounded the PLGA discs and migrated towards
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
33
the edge of the chamber. Tissue invaded the matrix at a much
slower rate. Once the edge of disc was reached further
thickening of new tissue grew around the disc but not
completely engorging it after 4 weeks.
Further modifications to the fibre-spun PLGA, such as
increasing the pore size and decreasing the density (and
therefore the hardness) may make this technique a viable
alternative to the salt-leached PLGA preparation.
Example 6 - Model system for vascularized tissue
The tissue chamber and graft system of the invention
may be used as a model to examine the behaviour of vascularised
tissue, through the use of an extracorporeal circulation
machine to maintain the developing tissue in vitro during its
generation. The chamber contents are established as specified
in Example 1. The host's blood or suitable transfused blood (at
least 90 ml) is taken and heparinised (up to 50 units/ml). The
blood vessel ends are connected to silicone tubing and the
blood is oxygenated via a renal dialysis filter. The oxygenated
blood is pumped through the tissue using conventional intensive
care unit instrumentation adapted for this purpose, and
maintained in vitro in this manner until the tissue/organ is
mature. During this phase blood samples are constantly
monitored to assess the degree of coagulation and the
maintenance of haemostasis. In a similar manner to the in vivo
studies, genetic modification of the tissue generating cells
can be applied to this model. Finally the tissue/organ
generated is microsurgically replaced into the appropriate site
in the host. A major advantage of this method is the ability to
produce tailor-made, off-the-shelf parts and organs.
The next step in testing our model is to add stem cells
to the system and see whether tissue is generated de novo. The
isolation, expansion and seeding of "stem cells" into the
chamber is a huge area for research in itself and is still in
its infancy. For various reasons, we have chosen an unorthodox
method of adding stem cells and environmental cues, with
unexpected results. We have investigated the behaviour of
injured/necrotic tissue explants placed in vivo in the chamber,
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
34
and have demonstrated conversion of muscle into fat (see
Example 3).
The hypothesis being tested in experiments such as
these is that these small tissue explants may harbour at least
a few stem cells, which perceive an injury to their parent
organs and respond by initiating tissue renewal. We have also
tested a number of tissues, including fat, liver and kidney,
and will shortly investigate neural, uterus, ovarian, thyroid
and glandular tissue. The results have been very promising,
because all of the tissues tested have "driven", by unknown
mechanisms, the generation of a cell phenotype not normally
present in the chamber. Mechanistically they have converted
the cellular/angiogenic response in the chamber from one
analogous to "inflammation and scar formation", involving the
de novo generation of tissue largely composed of fibroblasts,'
to one analogous to "tissue renewal and generation", also known
as "scarless" tissue repair in the fetus, comprising the
generation of vascularised tissue with a recognisable three
dimensional organisation and phenotype. Significantly, the new
tissue formed is free of fibroblastic in-growth and of
inflammatory cells.
Example 7 - Assessment of hypoxia within the tissue
growth chamber
For the study of hypoxia of the cells within the
chambers, AV shunt loops were created in anaesthetised male
rats as previously described in Example 1. Standard-sized
chambers (0.5 ml volume) were used. Chambers were filled with
Matrigel, as described in Example 5, and seeded with immortal
rat L6 myoblasts (1 x 106 cells / 0.5 ml Matrigel) distributed
over the entire surface area. Chambers were then positioned in
the groin of the rat.
Chambers were harvested at 3 days, 7 days, and 2 and 4
weeks incubation. At the time of exploration the animals were
again anesthetised with sodium phenobaritone (30 mg/ml) and an
assessment of anoxia was made by injection of nitroimidazole
(60 mg/kg, i.p.) 2 hours before the time of chamber harvest:
Rats were sacrificed with a lethal dose of pentobarbitone
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
sodium (3 ml of a 325 mg/ml solution) after harvesting the
chambers. Specimens within the chambers were processed for
histology and immunostaining with nitroimidazole antibody.
Under these circumstances, the only cells which label are those
5 which are hypoxic (<10 mm Hg) and which are proliferating.
An assessment of the degree of oxygenation of tissue at
days 3 and 7 showed proliferating, hypoxic cells in the
immediate vicinity of the vascular loop at both time points.
After 2 weeks the only labelled cells were at the periphery of
10 the growing mass of new tissue. By 4 weeks, no cells were
labelled with nitroimidazole.
The results from this study indicate that a state of
hypoxia and active biosynthesis exists in cells close to the
blood vessel loop. This strongly suggests that hypoxia is a
15 driving force of angiogenesis in the polycarbonate chamber
particularly in the first week. Those cells remote from the AV
loop were undoubtedly hypoxic but were not proliferating.
During week 2 the hypoxic, proliferating cells were located in
the advancing edge of the new tissue, but by the end of week 4
20 the chamber was well oxygenated throughout and new tissue
formation had slowed considerably. Studies such as this enable
the researcher to invetigate how hypoxia can influence the
growth of new tissue within the chamber.
25 Example 8 - Isolated, cultured cells added to chambers
in the rat AV loop model
(a) Addition of myoblasts to chambers
Skeletal muscles from various parts of the body (eg.
gastrocnemius, rectus femoris, latissimus dorsi, etc) were
30 harvested from neonatal rats 5 days after they were weaned.
Myoblasts were generated from this harvested tissue by
collagenase digestion and culturing in Ham's F10 culture medium
containing 20% fetal calf serum with 2 ng/ml of bFGF. Myoblasts
were identified by desmin immunostaining. Fibroblasts were
35 removed by serial subculturing, taking advantage of the fact
that they adhere to plastic within half an hour whereas
myoblasts adhere after that time. Enriched myoblasts (2-4 x 106
cells) were inserted into either (1) Matrigel alone
CA 02419923 2003-02-19
WO 02/15914 PCT/AU01/01031
36
(approximately 0.5 ml) or (2) Matrigel (approximately 0.15 ml)
with PLGA making up the balance of the volume. These matrices
were placed around an AV loop within a standard 0.5 ml chamber,
as previously described. These constructs were incubated
subcutaneously for either 2, 4, 6, 12 or 16 weeks. At the time
of exploration, the rats were placed under general anaesthesia,
and the tissue formed within the chamber (also known as the
"flap") was removed. Approximately half of the tissue was
frozen in isopentane and the other half fixed in formalin, and
sectioned, prior to morphological, histological and
immunohistochemical staining.
1. Matrigel only group
Group A - 2 weeks (n=6)
The chambers from six rats were examined at 2 weeks.
There was a large amount of muscle in four of these; and of
these, 3 contained identifiable desmin-positive myoblasts and
evidence of myotube formation. The other two contained no
desmin-positive tissue.
Group B - 6 weeks (n=9)
Of the 9 rats in this group, 2 constructs contained
muscle and myotubes, 4 flaps contained no identifiable muscle,
and 3 rats died prematurely.
Group C- 12 weeks (n=11)
Of the 11 rats in this group, no constructs contained
muscle, 5 flaps contained no muscle but did contain some (as
yet identified) tissue, 2 chambers contained no flap (possibly
because it slipped out of the chamber) and 3 rats died
prematurely.
2. PLGA/Matrigel group
Group A - 2 weeks (n=3)
No results for this group.
Group B - 6 weeks (n=6)
Of the 6 rats in this group, 2 constructs contained
muscle and myotubes, and 4 flaps contained no muscle. In one
chamber in which the myoblasts were fluorescently labelled with
CFDA prior to being seeded into the chamber, there was evidence
of myoblasts still surviving after 4 weeks' incubation in vivo.
Group C - 12 weeks (n=7)
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
37
Of the 7 rats in this group, 2 constructs contained
desmin-stained myoblasts, 5 flaps contained unidentified tissue
but no muscle, and 1 rat died prematurely.
Group D - 16 weeks (n=5)
No results for this group.
In H&E stained sections of flaps after 2 weeks
incubation, myoblasts were evident in some tissue specimens,
with their presence confirmed by immunostaining for desmin.
Within 2 weeks, groups of myoblast nuclei had aligned and
formed into myotubes which stained positively for dystrophin
and formed mature striated skeletal muscle. By 6 weeks,
myotubes and mature muscle were present in some specimens but
connective tissue formed in others. At both 2, 4 and 6 weeks
mononuclear leukocyte infiltrate was present, probably due to
the use of Matrigel, which originates from mouse cells.
However, by 12 weeks, much of the flap tissue was resorbed.
Interestingly, in some of the early experiments with "less
pure" myoblasts seeded, isolated pockets of osteoid (bone
tissue) and adipose tissue (fat) were also observed after 2 and
4 weeks in the Matrigel only experiments.
In preliminary experiments, a femoral nerve severed
distally was incorporated into Matrigel matrix, adjacent to the
loop and surrounded by the seeded myoblasts (n=6, 2 weeks
incubation). There appeared to be a trend towards reduced
desmin-positive muscle cells (compared with the nerve-free
controls, Group 1A) but there was positive immunostaining for
S100, a Schwann cell marker, in most of the newly generated
tissue.
We know from previous work that this model provides a
good angiogenic stimulus, and we have now shown that this model
can sustain the survival, expansion and differentiation of
myoblasts. The vascularised chamber can also support this cell
line and provide an optimal environment in which the chosen
cell can differentiate in a normal and expected fashion.
Histological evidence demonstrates that the seeded myoblasts
both survive and differentiate to form myotubes, which in turn
coalesce to form mature skeletal muscle in this model, over a
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
38
period as short as 2 weeks.
(b) Stem Cell addition
Using the same AV loop model, we have investigated the
fate of green fluorescent protein (GFP) labelled and non-
labelled rat bone marrow-derived stem cells into these
chambers.
Bone marrow-derived stromal cells were harvested from
rat femurs by flushing them with normal saline. These cells
were then labelled and sorted on a FACS machine. The stromal
cell subpopulation was expanded by culturing in a-MEM medium
containing 20% fetal calf serum. The expanded cells were
retrovirally transfected with Green Fluorescent Protein (GFP)
and a neomycin plasmid to enable them to be tracked within our
flap. When sufficient cells were available we placed them at a
concentration of 2 x 106 per 0.5 ml Matrigel into our AV loop
chamber model.
Nine AV loops in chambers containing these stem cells
were constructed using either Matrigel alone (n=8) or
Matrigel/PLGA (n=1) and the matrix. Rats have been examined at
2 weeks (n=4) or 4 weeks (n=4). In frozen sections some
fluorescence is seen in these specimens, although it is not
clear whether this is genuine GFP fluorescence or
autofluoresence. In subsequent experiments the resultant tissue
from our GFP-labelled flaps has been cultured in the presence
of neomycin-rich media. Surviving GFP-labelled cells have been
detected under such conditions after 2 and 4 weeks in the
chamber, whereas non-GFP-labelled cells failed to survive under
these conditions. However, to date we have found no evidence of
specific tissue phenotype or clone formation in new tissue
arising from these seeded cells.
Example 9 - Pancreatic cells added to chambers in the
rat AV loop model to form a transplantable pancreatic organoid-
All experiments were performed using inbred Sprague-
Dawley rats. The experimental model used of an arteriovenous
(AV) fistula created with a vein graft in the right groin and
placed within a 0.5 ml internal volume polycarbonate chamber,
was consistent throughout all experimental groups.
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
39
Rats were anaesthetised with pentobarbitone prior to
surgery as described in previous examples. Pancreatic tissue
for transplantation was prepared by various methods:
(a) "Ficoll islets": Using adult donor rats, the
isolated pancreas was digested with collagenase P (Boehringer
Mannheim, Germany) in vitro, and the islets purified by
centrifugation on a Ficoll density gradient.
(b) "Histopaque islets": Using adult donor rats, the
vasculature of the pancreas was perfused in vivo with 7 ml of
collagenase (Worthington Biochemicals, USA) at 1.3 U/ml. The
resultant islets were isolated and purified using Histopaque
[Liu and Shapiro, 1995].
(c) "Digested pancreas": Using adult donor rats, the
isolated pancreas was digested with collagenase P (Boehringer
Mannheim, Germany) in vitro, but the preparation was not
subjected to any further purification step.
(d) "Filtered pancreas": Using adult donor rats, the
isolated pancreases were not enzymically digested but simply
homogenised and the crude extract sieved through a range of
different sized filters. The fraction which passed through the
450 pm filter but was retained by the 100 pm filter was used in
further experiments.
The extracellular matrix used as a support for seeding
the islet preparations were used in one of the following
configurations:
(i) The chamber was filled with Matrigel, and the
islets were dispersed throughout.
(ii) The chamber was filled with Matrigel and the
islets / pancreatic tissue was placed in centre of chamber / AV
loop.
(iii) 150 l of Matrigel containing the islets /
pancreatic tissue was placed in centre of chamber in close
proximity to the AV loop.
(iv) 150 l of rat plasma clot containing the islets /
pancreatic tissue was placed in centre of chamber in close
proximity to the AV loop.
The experimental groups were devised as follows:
Group 1. Old (400-500 g) inbred Sprague Dawley rats
CA 02419923 2003-02-19
WO 02/15914 PCT/AU01/01031
were used. "Ficoll islets" were placed in Matrigel. There were
3 recipient rats. We used a 2.5:1 (donor:recipient) ratio, and
10-17 days incubation.
Group 2. Old (400-500 g) inbred rats were used.
5 "Digested pancreas" were placed in Matrigel. There were 3
recipient rats. We used a 1:1 (donor:recipient) ratio, and 11
days incubation.
Group 3. Adult (230-260 g) inbred rats were used.
"Digested pancreas" was placed in Matrigel. There were 6
10 recipient rats. We used a 1:1 (donor:recipient) ratio, and 7-14
days incubation.
Group 4. Adult (230-260 g) inbred rats were used.
"Histopaque islets" were placed in Matrigel. There were 8
recipient rats. We used both 1:1 and 4:1 (donor:recipient)
15 ratios, and 6-21 days incubation.
Group S. Adult (230-260 g) inbred rats were used.
"Filtered pancreas" was placed in a plasma clot. There were 8
recipient rats. We used a 1:2 (donor:recipient) ratio, and 8-24
days incubation.
20 In vitro experiments
Islets were kept in culture in Matrigel, with DMEM
media changes twice weekly, in parallel with the above in vivo
experiments to test the longevity of islets in culture.
Insulin immunostaining was performed on several such cultures
25 at one and two months with positive staining results.
Serum insulin level measurements
At the time of chamber harvest, blood samples (100 l)
were taken from the loop artery and vein and systemic venous
circulation, for measurement of insulin levels by
30 radioimmunoassay for the rat isoform.
Chamber harvest and flap manipulation
Chambers were harvested at the above time points, and
tissues were preserved in Buffered Formal Saline and routine
histological preparation, followed by paraffin embedding.
35 Histological sections were subjected to routine (H&E) and
immunostaining (for insulin and glucagon).
In 'vitro culture
Survival of islets was demonstrated to 4 and 8 weeks in
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
41
culture. H&E and insulin staining showed functional survival
at these time points. The islet clusters had begun to
dissociate into individual cells and clumps of cells between 4
and 8 weeks.
Serum insulin levels
Serum insulin levels in were tested in experimental
groups 3 and 4 described above. Venous (outflow) blood
exhibited serum insulin levels that were 30-50% lower than
those in the arterial (inflow) serum in most animals. in two
animals, levels were 40% and 100% higher in the venous system.
Analysis of the Chambers
Tissue in the chambers was divided into four parts and
serial sections-made. Large amounts of angiogenesis and
collagen deposition were confirmed, in keeping with the
original model. H&E staining demonstrated occasional islet
persistence in all groups, but not in all flaps. inflammatory
infiltrates were present in most flaps, consisting mainly of
lymphocytes. Ductal elements were observed in the Group 5
"filtered pancreas" chambers, although no confirmatory
immunohistochemistry was performed. Insulin and glucagon
immunohistochemistry demonstrated occasional positive staining,
particularly for glucagon.
These experiments demonstrate that the AV loop chamber
model creates a suitable environment to support the survival of
islets in a significant number of the constructs for periods up
to 24 days. Insulin and glucagon production was identified by
immunostaining in histological sections of tissue during this
same period. However, the long term viability of this new
"organoid" and its continued insulin production remains to be
evaluated.
Example 10 - Increasing the amount of tissue in the rat
model through the use of larger chambers.
(a) Rat Experiments
The amount of tissue produced in the rat using the
standard chamber model (-0.3 ml) is quite substantial in
comparison with the animal's body size, and corresponds to a
small "breast" or small "organ" within the body. In order to be
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
42
able to reproduce this finding in the humans it is essential to
test the limits of tissue production. This can be done firstly
in the rat, through the use of larger volume chambers.
Therefore, the aim of this study was to assess whether larger
amounts of tissue could be grown over a longer period of time
(4-8 weeks) inside larger chambers. in this fashion it is
proposed that this method can be used to produce clinically
useful amounts of new tissue which, if necessary, could be
transferred on its own vascular pedicle to another part of the
same individual.
The basic model of the arteriovenous (AV) shunt loop in
an enclosed growth chamber has been described in detail in
Example 1. The AV shunt was placed within a dome-shaped chamber
(Figure 2). The chamber was made of polycarbonate, had a
proximal opening for the pedicle and consisted of a base plate
and a lid. It had a base diameter of 17 mm, a centre-of-base to
top-of-dome distance of 1.3 mm and an internal volume of 1.9
ml. In contrast, the standard chamber described in previous
studies (for instance Examples 1 and 2) had a volume of 0.5 ml.
The AV shunt was sandwiched between two custom-made disks of
PLGA which was used as a matrix to fill the chamber. The PLGA
was prepared according to the salt leaching method described by
Patrick et al.(1999). Pore sizes between 300-420 nm and a
porosity of 80-90% was achieved. The disks were sterilised by
four cycles of mechanical stirring for 30 minutes in 100%
ethanol, then three times sterile, phosphate buffered saline,
before use.
After positioning the lop in the chamber the lid of the
chamber was closed and the chamber embedded beneath the
inguinal skin and secured with three 6-0 prolene holding
sutures. The wound was closed with 4-0 silk sutures. Chambers
were harvested from rats under general anaesthesia at 2, 4, 6,
and 8 weeks incubation for further analysis (n=6 per group).
The animal was finally killed by an overdose of Lethobarb (3
ml) administered by intracardiac injection.
whole mount specimens were fixed in buffered formal
saline (BFS) and cut into 1 mm slices. Half of these slices, in
alternating order, were embedded in paraffin and stained with
CA 02419923 2003-02-19
WO 02/15914 PCT/AU01/01031
43
H&E for histological comparison of the maturity of the newly
formed tissue and its vasculature. The other half of these
slices were stored in 100% ethanol and used for point counting
on a grid to assess the percentage of the newly formed tissue,
the remaining PLGA, and the AV loop in the specimen. Every
fifth field of 100 points was counted on the front and back of
each tissue slice. For this purpose, the slices were dipped in
haematoxylin briefly before counting. This enabled newly formed
tissue to be readily distinguished from PLGA. The results of
point counting on the grid enabled calculation of the
percentages of newly formed tissue, remaining PLGA, and AV loop
and comparisons of those values at 2, 4, 6, and 8 weeks.
Statistical differences between newly formed tissue weight and
PLGA weight in time were calculated using Student's t-test with
p<0.05 being statistical significant.
Weight and volume measurements: All specimens harvested
from the chambers were assessed for volume and weight. The
volumes of the specimens, as measured by fluid displacement,
was not statistically significant different from the measured
weights. The total average weight (equivalent to volume) of the
specimens decreased progressively in time. The total average
weight t standard deviation (SD) of each group of specimens was
1.07 t 0.06, 1.03 t 0.06, 0.96 t 0.06, and 0.81 t 0.18 grams,.
at 2, 4, 6 and 8 weeks, respectively. This resulted in a
statistically significant decrease of specimen weight between
time points apart 4 weeks or longer, which may be accounted for
by the progressive gradual resorption of PLGA matrix.
The amount of PLGA and tissue in the specimen was
studied to assess their involvement in the overall decrease in
weight of the specimens. All specimens were point counted
microscopically with the aid of a grid to determine the
percentage of specimen taken up by PLGA or tissue. The decrease
in specimen wet weight was attributed to resorption of PLGA.
The total average weight of PLGA t SD at 2, 4, 6, and 8 weeks,
respectively, was 0.89 t 0.07, 0.56 10.14, 0.34 t 0.07, and
0.20 t 0.09 g. On the other hand the newly formed tissue
component of the specimen showed a progressive increase of
weight in time. The total average weight of tissue t SD at 2,
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
44
4, 6, and 8 weeks, respectively, was 0.13 0.04, 0.42 t 0.09,
0.57 t 0.06, and 0.58 t 0.10 g. The increase in tissue weight
was statistically significant over all consecutive time
periods, except for the period between 6 and 8 weeks (P<0.05).
Over this 6-8 week period, tissue growth reached a plateau,
although it also did not decrease as noticed in previous
experiments in smaller sized chambers filled with PLGA (Example
51.
Macroscopic findings: After India-ink injection,
neovascularisation could be readily identified during
processing of the tissue. New vasculature.did not reach the
outer edge of the PLGA scaffold at any time point. However, in
one serendipitous finding, a chamber was inadvertently left
incubating in a rat for 10 months. When harvested, the chamber
was totally full of soft connective tissue, which was well
vascularised and had patent blood vessels supplying nutrition
to the tissue "flap".
(b) Rabbit Pilot Experiment
Preliminary results from the rat experiments indicated
that the larger chambers were able to grow more tissue and for
a longer period than the standard chambers. Where the walls of
the large chambers were perforated with numerous holes, a
further improvement in the rate of new tissue growth, the
amount of tissue produced and growth to the edges of the
chamber were found [Tanaka Y,, 2000, unpublished findings].
These latter conditions approach the optimal conditions-for
.tissue growth in this model. The major aim of this pilot study
was to assess whether tissue production could be scaled up in
an animal which is 8-10 times the size of a rat, and whether
the tissue would maintain its size and shape.
The experimental model used was the basic AV shunt loop
in an enclosed growth chamber, however the experimental animal
was the New Zealand White rabbit.
Pre-operative analgesia was given in the form of
carprofen (1.5 mg/kg, s.c.). New Zealand White rabbits (2.0 to
2.8 kg) were anaesthetised with i.v. pentobarbitone (30 mg/kg)
and maintained in a face mask with halothane and oxygen (2.0
L/min). Under sterile conditions a graft of 4-6cm (rabbits)
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
respectively was harvested from the left femoral vein, and used
to create an AV shunt between the proximal ends of the divided
right femoral artery and vein. The AV shunt was placed within a
dome-shaped chamber, in this case made of polyurethane, with
5 the approximate dimensions 3.0 cm diameter, 2.0 cm high, with
an opening for the vessel entry and egress (Figure 2). In some
instances the anatomy of the rabbit permitted the use of an AV
pedicle rather than an AV loop, because the small connecting
vessels in the surrounding tissue of the pedicle made it a
10 naturally occurring flow-through loop. in this latter example
the effect of the AV blood flow was comparable but the
operating time and postoperative pain was less. in the usual
configuration this chamber had a plurality of small
perforations in the chamber walls. Subcutaneous fat in the
15 groin region was used as a source -of adipocytes and adipogenic
precursor cells.(Zuk et al, 2001). The fat tissue was formed
into a crude slurry by injection through an 18 gauge needle.
These cells were donated by and implanted into the same rabbit.
The AV shunt loop or pedicle was placed within the
20 chamber, which was filled with a 3-dimensional matrix made of a
combination of PLGA which was machined to fit the chamber,
Matrigel, Type 1 porcine skin collagen or a similar suitable
composition, and the preadipocyte-rich fat tissue slurry. The
Matrigel was then allowed to gel. The lid was closed and the
25 chamber embedded beneath the inguinal skin. The wound closed
with 4-0 nylon sutures.
Approximately 6-8 weeks later, with the animal again
under general anaesthesia, the chamber with its associated
blood vessels was removed from the groin and the chamber. Two
30 flaps have been analysed to date.
The tissue in the chamber was removed and its wet
weight recorded. The tissue was also be suspended by a fine
cotton suture thread and wholly immersed in a beaker of water
on a balance. The mass, assuming a density of 1.00 g/ml, is the
35 tissue volume. Specimens were fixed in buffered formol saline
(BFS), embedded in paraffin and stained with either H&E or
Masson's Trichrome (a connective tissue stain).
The volume of new tissue generated after 8 weeks growth
CA 02419923 2003-02-19
WO 02/15914 PCT/AU01/01031
46
was 10-11 ml (compared with a total volume of the chamber
estimated to be 12 ml). The composition of the flap was
adjudged to be a mixture of adipose and other connective
tissue. The shape was preserved when transferred under the
nipple of the same male rabbit and the volume sufficient to
enable the construction of a medium-sized breast on this animal
(see Figure 6).
We have achieved the production of clinically useful
amounts of tissue in the rat and rabbit. The tissue thus
produced was of a size and shape potentially suitable for
breast reconstruction and similar applications. Flaps such as
these with their associated patent blood vessels have the
potential to be transferred to another part of the body for
reconstructive purposes.
Example 11 - A new model of vascularised tissue
engineering in the mouse.
In order to investigate the fundamental processes of
tissue engineering it is desirable to develop a suitable tissue
engineering model in the mouse for the following reasons:
Genetic Technology: Transgenic and gene knock-out
technology is much further advanced in mice, allowing us to
probe the influence of a number of factors involved in tissue
engineering such as growth promoters and inhibitors.
Stem Cell Biology: Stem cells are pluripotent cells
that give rise-to all tissues;, they are highly durable and can
therefore theoretically resist the initially hostile ischaemic
environment of the chamber. This.makes them attractive cells to
seed in the chamber. Stem cell biologists have cloned a wide
variety of stem-cell sub-types in mice that can be seeded into
the mouse model in order to attempt to generate specific tissue
types.
Cost: There are also significant cost benefits in using
mice. Purchase, housing and caring for mice is less expensive
than for larger animals. Also there will be a reduction in the
use of expensive laboratory consumables such as growth factors.
We investigated two different types of vascular
configurations that have been shown to be angiogenic in
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
47
previous work, in order to determine the best technique to use
in the mouse. The first was a tied off arteriovenous pedicle
(AVP) of the femoral artery and vein (Khouri et al, 1993;
Figure 3) and the second was a "flow through loop" pedicle
(FTLP) configuration (Morrison et al, 1990; Figure 4).
The polycarbonate chamber, when used in the rat model,
did not adversely affect the patency rate of the high-flow
microsurgical arteriovenous loop. It was also tolerated well by
these animals. However this material is hard and has sharp
edges which was felt might affect the patency rate in the mouse
due to the lower flow rate of the proposed vascular
configurations and smaller diameter vessels in this animal.
Therefore polycarbonate chambers were compared with softer
silicone chambers in order to determine the most suitable
material to use in the construct of the chamber. We also
compared the two main extracellular matrices used in the rat
model (Matrigel and PLGA) in the mouse to judge which was best
with regard to angiogenesis and tissue growth. A total of 88
C57BL/6 wild-type mice (male and female; 18-24g body weight)
were used for this set of experiments.
Initially two vascular configurations were examined
using specially constructed polycarbonate chambers. The first
was a tied off AVP of the mouse femoral artery and vein as
described by Khouri et al [1993] in the rat (Figure 3). The
second was a FTL pedicle comprising the superficial epigastric
vessels encapsulated within a modified version of the
polycarbonate chamber as described by Morrison et al [1990]
(Figure 4). There were 3 groups of 6 animals for both vascular
configurations. Each configuration was examined at the 2, 4 and
6 weeks. The experiment was to be performed using both
Matrigel and PLGA as extracellular matrices (n=2x2x3x6=72).
All operations were performed under general anaesthesia
(chloral hydrate, 4 mg/g body weight, i.p.). The right groin
and upper leg were rendered hair free using a combination of
clipping and a depilatory cream. The skin was decontaminated
using an alcohol preparation. The tied off pedicle technique
required a vertical incision extending from the groin crease to
the knee just offset from the saphenous vessels which are
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
48
visible through the-skin. The saphenous vessels were tied off
at the knee and then dissected free from their accompanying
nerve back to the origin of the femoral artery at the inguinal
ligament. The flow -through model was performed using a
transverse groin incision sited just above the groin fat pad.
The superficial epigastric (SE) vessels were dissected free of
the. surrounding tissue from their origin at the femoral vessels
for a distance of approximately 1 cm to their entry into the
groin fat pad. Here the vessels course through the fat pad
sending nutritional branches to the fat and glandular tissue
around them.-They then anastomose directly with an ilio-
inguinal vessel (a direct branch of the infra-renal aorta) that
pierces the abdominal wall at the lateral aspect of the
inguinal ligament to enter the fat pad from the lateral side.
The entire fat pad is mobilised free of the skin and underlying
muscle thus creating a space into which the chamber will alter
be introduced. Thus the SE vessels have an arterial input and
venous drainage from both sides which we felt would augment the
long term patency rate in this model. To our knowledge this is
the first time that this vascular arrangement has been
described in the mouse. The first cm of the SE vessels (where
they are free of the fat pad) is then encapsulated in a
modified polycarbonate chamber that is split down one side and
the appropriate extracellular matrix (Matrigel or PLGA) is
inserted into the chamber. The chamber is then sealed at the
proximal end and along the lateral. split using melted bone wax
(Ethicon bone waxTM) taking care not to apply the heated wax
directly to the vessels. The seal is augmented by two 10/0
nylon microsutures placed at either end of the lateral split
and the whole chamber is anchored to the underlying muscle near
the origin of the SE vessels in order to prevent the pedicle
from being dislodged during post-operative mobilisation. A
small amount of fatty tissue surrounding the vessels as they
enter the fat pad is allowed to "plug" the distal end of the
chamber. This plug is then augmented with wax sealant and the
whole construct is carefully placed in the groin so that it
lies in the dissected space lateral to the femoral vessels. The
wounds were closed using a combination of buried interrupted
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
49
horizontal mattress sutures and a running suture (both 6/0
silk) as these animals tend to gnaw at their wounds.
Following early analysis of the results in each group
the tied off AVP group of the experiment was discontinued. This
was because the thrombosis rate was unacceptably high (11/14
animals) and pursuit of this line of investigation seemed
futile and wasteful of animals. This observation contrasts with
Khouri's work in the rat and our own experience in the rat and
the rabbit where the tied-off AVP remains patent in the
majority of cases. The most likely reason for the high
thrombosis rate in our study is that the mouse vessels are
extremely small (internal diameter approximately 0.2 mm) and
very sensitive to dissection. Flow rates in vessels of this
size are also very low. The thrombosis rate in the FTLP group
was better (3/11) but still seemed excessively high.
We postulated that the polycarbonate material we were
using was too hard and sharp for the delicate vessels of the
mouse. Our experimental plan was modified to include a cohort
of animals with chambers made from medical grade silicone
(Animal Ethics committee approval was obtained for this
modification). Two cohorts (1 with PLGA and 1 with Matrigel )
of 3 groups of 4 animals were used for this modified aspect of
the experiment (n=2x3x4=24) using only the flow through
vascular configuration. Accurate estimation of the volume and
weight of the specimens proved impossible. The volume of the
chamber is approximately 80-100ul. This varies for several
reasons such as the amount of wax or fat that encroaches on the
entry points of the chamber. Also it is difficult to measure
the exact volume of extracellular matrix that is used in each
chamber. Matrigel is usually added as a liquid and allowed to
gel in vivo. Some spillage may occur during infusion or during
manipulation of the chamber. We also noted that the volume of
the Matrigel declined by at least 50% over the first two weeks
such that the specimen that was removed was actually smaller
than that inserted.
The weight of the PLGA used in each chamber could be
accurately measured but the volume was impossible to ascertain
as the structure was porous and had to be broken up into crumbs
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
in order to easily fit it into the small chamber. Given these
inaccuracies we did not attempt to evaluate quantitatively the
chamber tissue and looked at the more qualitative aspects of
the device such as morphology of the newly formed tissue.
5 Patency of the vessels was determined at microsurgical
exploration and via India ink perfusion studies. If the vessel
was extensively thrombosed within the chamber it was usually
possible to see this under the operating microscope. However
ascertaining definitive patency was not always possible.
10 Therefore India ink perfusion studies were performed under
general anaesthesia on each animal prior to sacrifice. Under
the operating microscope the groin incision was reopened and
the chamber exposed taking care not to damage the pedicle. A
laparotomy was then performed and the abdominal aorta was
15 dissected free of the vena cava and cannulated just below the
renal vessels using a fine (30G) bore silicone tube. This was
then flushed using heparinised saline to ensure that the
cannula was in the correct position. Next a solution of neat
commercial India ink containing 10 i.u. heparin per ml was
20 infused under gentle hand pressure using in a pulsatile fashion
until the animal's liver had turned completely black. Previous
descriptions also advise the use of gelatin in this solution
but in our preliminary trials of the technique we found that
the gelatin formed clumps that obstructed the fine bore tube
25 and resulted failure of the procedure (this occurred even if
the gelatin was warmed to body temperature prior to infusion).
Patency.could be confirmed under direct visualization of the
transparent chamber as the India ink could clearly be seen
tracking into the chamber along the vascular pedicle. Following
30 this the animals were sacrificed via a lethal overdose of
phenobarbitone and the chambers were carefully removed cutting
the pedicle(s) flush with their entry into the device. ,
The specimens were fixed in formalin and taken through
graded alcohol solutions to absolute alcohol. They were then
35 immersed in methyl salicylate and allowed to clear over 72
hours. This allows direct visualization of the vascular tree
which has been perfused with India ink. All specimens were then
examined as whole-mount preparations under microcater and
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
51
vessel counts were performed. After this the specimens were
processed for histological examination and embedded in wax. The
wax blocks were then sectioned at 5pm and stained with
haematoxylin and eosin in a standard fashion. Vessel area
density was estimated on all cleared specimens using a
microcater which allowed visualization throughout the depth of
these small tissue specimens. Three fields were randomly
selected at 3 depth intervals of 500 pm and the vessel density
was assessed with the aid of a stereometric grid. Following
completion of this process the specimens were committed to
histological processing. The stained sections were
morphologically assessed in terms of angiogenesis and the
cellular characteristics of the newly generated tissue.
Univariate analysis of the patency rates and vessel density was
performed using the Student t-test. The patency rate was
assessed for the two vascular configurations and for the
different materials used in the make-up of the chamber.
The patency rate for the tied off arteriovenous pedicle
was 21% versus 88% for the flow-through pedicle. The patency
rate in the polycarbonate chambers (excluding the tied off AV
pedicle group) was 88% versus 97% in the silicone chambers. The
new vessels in the tied off AVP group were seen to be arising
from outside the chamber and growing in along the thrombosed
pedicle. The vessel densities in the flow-through chambers were
similar at 2, 4 and 6 weeks. Similarly there was no difference
in vessel density between PLGA and Matrigel. Morphologically
there was good angiogenesis in Matrigel and PLGA but
qualitatively it was better in the Matrigel . The new vessels
seemed to be more numerous and occurred throughout the
construct in the Matrigel . The angiogenesis in the PLGA was
more to the periphery of the construct with fewer vessels in
the central aspect probably due to the solid nature of this
ECM.
In terms of cellular morphology the PLGA seemed to
promote a predominately fibrous foreign-body type reaction.
Fibroblasts are the predominant cell seen both peripherally
where the matrix lay against the chamber wall and centrally
within the substance of the matrix. The Matrigel group also
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
52
showed a fibroblastic response at the ECM-chamber interface. On
the other hand the central aspect of the Matrigel shows the
presence of fat in the chamber that has clearly migrated
through the matrix and survived, presumably nourished by the
newly generated vascular tree. This phenomenon has been
reported before in non-encapsulated Matrigel in mice using
growth factors and pre-adipocytes. The presence of mature
viable fat in the chamber suggests this model is capable of
supporting the migration, maturation and possibly the
reproduction of fat cells and their precursors. in female
animals the fat pad contains some mammary tissue and associated
ducts. which are occasionally found in the distal part of the
chamber where this tissue is used as a "plug- to seal the
distal aperture. In the Matrigel group we observed that in
some of these animals the breast ductal/acinar tissue seemed to
be growing into the Matrigel and in others there is clear
morphological evidence of newly forming ductal/acinar tissue.
This suggests that the chamber is capable of supporting the
development of glandular tissue as well as fat. To our
knowledge this has not been reported before.
We have seeded the chamber with clones of mouse
mesenchymal stem cells that were cultured from a C57 Immorto
mouse and also with a mouse mammary tumour cell line. Both were
labelled with flourescent markers (GFP or CFDA) and we were
able to demonstrate that the implanted cells were alive at 48
hours, 4 days, 2 weeks, 4 weeks and 8 weeks. The mammary tumour
line has been seen histologically at 4 weeks demonstrating that
the chamber is capable of supporting cell lines in the longer
term. We have also successfully grown foetal pancreas, liver,
heart, bowel and limb bud (composite skin, bone , cartilage,
muscle, vessel and nerve) in immunodeficient SCID mice. As well
as this we have been successful in getting cultured adult
pancreatic islets to survive and produce hormones at 2 and 10
weeks in wild type mice (C57BL6). This effectively means that
we have successfully grown functioning islet allograft in these
animals which has not been achieved in other models of
pancreatic transplantation. This means that the chamber may
confer some immuno-privileged status to the cells that grow
CA 02419923 2010-03-31
WO 02/15914 PCT/AU01/01031
53
within it. This has therapeutic implications in that it may be
possible to use unmatched allograft or even xenograft in the
chamber with or without local inunu osuppression or Sertoli cell
co-culture as a treatment of Diabetes Mellitus.
It will be apparent to the person skilled in the art
that while the invention has been described in some detail for
the purposes of clarity and understanding, various
modifications and alterations to the embodiments and methods
described herein may be made without departing from the scope
of the inventive concept disclosed in this specification.
References cited herein are listed on the following
pages.
CA 02419923 2003-02-19
WO 02/15914 PCT/AUO1/01031
54
REFERENCES
Erol, O.O., and Spira, M. Surgery (1980) 66: 109-115.
Khouri, R.K., Koudsi, B., Deune, E.G., Hong, S.P., Ozbek,
M.R., Serdar, C.M., Song, S-Z and Pierce, G.F.
Surgery (1993) 114:374-380.
Knight, K.R., Mian, R., Tanaka, Y., Penington, A.J.,
Hurley, J.V., Cassell, 0., Romeo, R., and
Morrison, W.A. 7th Annual Meeting Australian
Vascular Biology Society (1999).
Liu, M., and Shapiro,=M.E. Transplant. Proc. (1995) 27:
3205-3207.
Morrison, W.A., Dvir, E., Doi, K., Hurley, J.V., Hickey,
M.J., and O'Brien, B.M. Br. J. Plast. Surg.
(1990) 43: 645-654.
Patrick, C.W.Jr., Chauvin, P.B., Hobley, J., and Reece,
G.P. Tissue Eng. (1999) 5: 139-151.
Prockop, D.J. Science (1997) 276: 71-74.
Tanaka, Y., Tajima, S., Tsutsami, A., Akamatsu, J., and
Ohba, S. J. Jpn. P.R.S. (1996) 16: 679-686.
Tanaka, Y., =Tsutsumi, A., Crowe, D.M., Tajima, S., and
Morrison, W.A. Br. J. Plast. Surg. (2000) 53:
51-57.
Vukicevic, S., Kleinman, H.K., Luyten, F.P., Roberts,
A.B., Roche, N.S. and Reddi, A.H. Exp. Cell Res.
(1992) 202: 1-8.
Zuk, P.A., Zhu, M., Mizuno, H., Huang, J., Futrell, W.,
Katz, A.J., Benhaim, P., Lorenz, H.P., amd
Hedrick, M.H. Tissue Eng. (2001) 7: 211-228.
