The importance of large animal models in transplantation.

Abstract

Animal models have been extensively used in transplantation research. However, animal experimentation is contentious and subject to legal and ethical restrictions. Most experiments are carried out on rodents, but crucial prerequisites for the development of safe pre-clinical protocols in biomedical research are needed through suitable large animal models. In transplantation particularly, large animal models have developed dramatically. This article provides an overview of the large animal models commonly used to evaluate organ transplant experiments and analyzes the specificity of several models in various situations such as induction of allospecific tolerance and xenotransplantation. The key determination that remains be addressed is the most appropriate species and strains to model human immune and physiological systems. Because of their phylogenetic and physiologic similarities to man, non-human primates play an increasingly important role in pre-clinical testing. Nevertheless, a number of studies have shown the pig to be a reliable large animal model for transplantation research, and the availability of genetically defined or modified pigs establishes a stronger position for pigs as a large animal model.

Full text

Page 1

[Frontiers in Bioscience 12, 4864-4880, September 1, 2007]
4864
The importance of large animal models in transplantation

Jean-Paul Dehoux, Pierre Gianello

Universite catholique de Louvain, Faculty of Medicine, Experimental Surgery Unit, Avenue Hippocrate, 55, 1200 Brussels,
Belgium

TABLE OF CONTENTS

1. Abstract
2. Introduction
3. Ethical considerations
4. The importance of large animal models for transplantation research
5. Animal models
6. Factors controlling allograft or xenograft rejection
6.1. Allotransplantation
6.2. Xenotransplantation
7. Immunosuppressive Drugs
8. Immunosuppressive therapies in large animal models
8.1. Monoclonal antibody or fusion protein therapy
8.2. Depletion strategies
8.3. Mixed chimerism
8.4. Donor antigen infusion
9. Large Animals in Xenotransplantation
9.1. Concordant large animal models
9.2. Discordant large animal models
9.3. The prevention of discordant xenograft hyperacute rejection
9.3.1. Transplantation of transgenic pig organs
9.3.2. Alpha-1,3-galactosyltransferase knock-out pigs
9.4. Cell-mediated and chronic rejection
9.5. Accommodation
9.6. Induction of immunologic tolerance
10. Non-Immunologic hurdles
10.1. Anatomical and physiological obstacles
10.2. Zoonosis
11. Experimental cellular xenotransplantation
12. Perspectives
13. References

1. ABSTRACT

Animal models have been extensively used in
transplantation research. However, animal experimentation is
contentious and subject to legal and ethical restrictions. Most
experiments are carried out on rodents, but crucial
prerequisites for the development of safe pre-clinical protocols
in biomedical research are needed through suitable large
animal models. In transplantation particularly, large animal
models have developed dramatically. This article provides an
overview of the large animal models commonly used to
evaluate organ transplant experiments and analyzes the
specificity of several models in various situations such as
induction of allospecific tolerance and xenotransplantation.
The key determination that remains be addressed is the most
appropriate species and strains to model human immune and
physiological systems. Because of their phylogenetic and
physiologic similarities to man, non-human primates play an
increasingly important role in pre-clinical testing.
Nevertheless, a number of studies have shown the pig to be a
reliable large animal model for transplantation research, and
the availability of genetically defined or modified pigs
establishes a stronger position for pigs as a large animal model.

2. INTRODUCTION

Transplantation is today the treatment of choice
for several end-stage organ diseases. Experimental research
on large animal models has been critical for this
phenomenal development in human medicine. Murray and
co-workers (1,2) demonstrated in the early 1960s that a
treatment with anti-metabolite
significantly prolong the survival of renal homografts in
dogs: these experiments were crucial for the endorsement
of transplantation as an overwhelming therapeutic tool. The
remarkable results obtained in human transplantation,
however, have led to a major worldwide shortage of organs
and cells. This shortage is the impetus for research into
novel methods to directly or indirectly increase the number
of available organs and cells through tolerance induction,
use of living-related grafts or non-heart beating donors,
xenotransplantation, and cellular transplantation. Animal
models will again play a major role in these research topics.

Although much knowledge has been accumulated
about the immune systems of rodents and humans, the
evolutionary gap between them has often hampered the
drugs allowed to

Page 2

Large animal model
4865
direct applicability of the knowledge gained. Despite the
advantages of using small animal models in transplantation,
and the advantages conferred by congenic strains,
transgenic, and knock-out murine models notwithstanding,
recent data confirm that experiments in rodents are not
sufficiently relevant to predict human responsiveness to
potential treatment strategies. Extrapolating from rodent
studies to human applications is difficult for several reasons:

•
The young age of murine models certainly
has an important impact on the maturity of the
immune system and renders immunomodulation
easier.
•
The expression of major histocompatibility
complex (MHC) antigens is restricted to antigen-
presenting cells (3).
•
The general conditions of rodents differ from
those of large animals and humans. In fact,
rodents have a reverse activity cycle compared
with humans, so the chronobiology is different
(4).
•
The dosage/metabolism conversions between
the rodents and NHPs/humans are always
approximate (5).
•
Different pharmacokinetics parameters might
lead to incorrect
chronological times are used instead of allometric
ratios to determine the lifetime and the body
weight of the rodents (6).

In contrast, large animal models, although
complex and expensive, can be more relevant to clinical
transplantation and to understanding the immunological
barriers to transplantation. Large animal experimentation
has a long history in the field of transplantation research,
where it has been used from an anatomical and a surgical
point of view to plan the technical aspects of the surgery,
for research into physiological and immunological aspects
to overcome the rejection processes, to determine the
effectiveness of immunosuppressive agents, and to induce
donor-specific tolerance. In this way, great strides have
been made in recent years to develop reagents and tools to
study immunology of non-human primates (NHPs) and
pigs.

The aim of this article is to review the commonly
used large animal models (pigs and NHPs) in the field of
transplantation research, identify their advantages and
limits, and propose guidelines for the use of animal models.

3. ETHICAL CONSIDERATIONS

The ethics of large animal experimentation have
evolved to a level similar to that stated for humans: large
animal experiments should be based on prior in vitro and, if
possible, small animal experimentation. In each country of
the European Union, animal research is currently controlled
by national legislation. In Belgium, animal laboratory use is
controlled by the Animal Law of 1986 and the European
legislation—directive 86/609/EEC—for the protection of
animals used for experimental purposes. These regulations
aim to improve the controls on the use of laboratory
extrapolation since
animals, set minimum standards for housing and care,
standardize the training of staff who handle animals and
supervise the experiments, and reduce the numbers of
animals used for experiments. A research laboratory must
hold a personal license authorizing it to perform regulated
procedures on protected animals. This license authorizes
specified work for a well-defined purpose and must include
a detailed protocol for each regulated procedure. The
protocol must be approved by the local animal research
ethics committee of the institution maintaining the research
laboratories, and the research can be carried out only under
the supervision of an official animal care and welfare
officer and a named veterinary surgeon for a clearly
defined period of time.

Animal research in Belgium and in Europe is as
such governed by the three “R” rules: reduce (lower) the
number of animals used in experimentation, refine
(improve) their life conditions, and replace laboratory
animals if possible by in vitro or other experiments (7).

4. THE IMPORTANCE OF LARGE ANIMAL
MODELS FOR TRANSPLANTATION RESEARCH

No known animal model can provide data that are
completely comparable with humans, but most animal
data provide a reasonable prediction of efficiency and
safety. Animal models have played a critical role in
establishing basic paradigms of immunology and
physiology because they provide an in vivo milieu that
cannot be reproduced in vitro. Insights gleaned into the
mechanism of immune and physiological functions in
rodents form the basis of research; however, an
increasing knowledge of the complexity of the human
immune system reveals important hurdles that do not
appear in small animal models. Although numerous
strategies have been successfully applied to small rodent
models, extending most of these to large animals has
been difficult.

In fact, in vitro studies cannot mimic the most
basic in vivo immune responses because so many local
and systemic cellular and humoral actors to auto, allo-
or xenoantigens are involved. Animal models are thus
an obvious step needed to identify the most relevant
pathways involved in immune mechanisms that only can
be hypothesized in vitro. These models nevertheless
result from a stepwise progression that includes in vitro
experiments in well-controlled systems, followed by the
confirmation in vivo, in murine models. The need to
pursue research in large animal models depends on the
analysis of the data previously obtained. What exactly
does the survival of a transplanted organ for 100 days in
a young mouse mean, when it probably has a completely
immature or naive immune system? Is it indefinite
survival? Is it specific and indefinite tolerance? Thus,
large animal models must serve as a confirmation step
only when the data converge reasonably to suggest a
possible application in human medicine. In addition,
these data must have been reproduced in several
conditions and laboratories before extensive additional
experiments are considered.

Page 3

Large animal model
4866
Table 1. Success of tolerance induction protocols in small
versus large animal models and human
Method
Enhancement
Donor-Specific Transfusion
Peptides
Anti-MHC mAbs
Cyclosporin
Anti-Lymphocyte Serum
Anti-CD4 mAb
Anti-CD25 mAb
Total Lymphoid Irradiation
Anti-CD3 toxin
Co-stimulatory blockade
Chimerism
+: Tolerance induction success, -: Tolerance induction
failure

5. ANIMAL MODELS

The first immune characteristic of small animal
models is certainly the young age at the time of
experimentation, which correlates with an immature
immune system and a very active thymus. These young
animals are usually inbred and very much protected from
outside pathogens and immune challenges (viruses,
vaccines, germs, etc.). This “artificial” environment is
clearly different from an adult immune system, which has a
large repertoire of T-cell memory. Also, the expression of
class II antigens, which is in fact restricted to dendritic
cells, could play a major role in small animal
transplantation models: for instance, the lack of expression
of class II antigens on vascular endothelium is related to the
relative ease for inducing tolerance to primarily
vascularized allografts in rodents, even across full MHC
barriers (see Table 1) compared to NHPs or humans (8).

In contrast, large animal models are crucial,
especially in transplantation immunobiological research,
since class II expression of MHC antigens is constitutive on
the endothelium as it is in humans. The importance of class
II antigens has been extensively demonstrated in
transplantation pig models and in NHPs that have
undergone MHC antigen typing (9,10).

Among large animals dedicated to this research
field, three models—dogs, pigs, and NHPs—have been and
are still used frequently in transplantation research because
their physiology and
immunological properties approximate those of humans.

Large animal models were first used to develop
the technical and surgical aspects of allotransplantation. In
1961, a surgical technique of heart allograft in dogs, which
led to Barnard’s first human cardiac transplant in South
Africa in 1967, was developed (11,12,13,14); similarly,
liver allotransplantation was developed in pig and dog
models to eventually allow the first human liver transplant
(15,16,17,18,19). Calne and co-workers clearly initiated the
first step of allotransplantation in humans by showing
prolonged survival of renal allografts in beagles (1). Today,
large animal models still allow technical development such
as spleen allotransplantation to evaluate tolerance induction
(20).
Mouse
+
+
+
+
+
+
+
+
+
+
+
+
Primate/Human
-
-
-
-
-
-
-
-
+/-
+/-
+/-
+
general anatomical and
Dogs were abundantly used in early
transplantation research. Their use in renal, heart
(11,12,13,14), liver (18,19), lung (21), intestinal (22), bone
(23), bone marrow (24), pancreas (25) and pancreatic islet
transplantation (26) has been important for facilitating
surgical training. In recent decades, the dog model has
mainly been used in bone marrow transplantation, since the
dog MHC antigen has been well characterized in strains
such as beagles (dog leukocyte antigen [DLA]) (27). But
over time, dogs have been less used in response to societal
pressure. In addition, reagents for dog immunological
studies are very limited and do not allow the study of
immunological mechanisms. Most ethics committees
mandate the use of breeding colonies for dogs. This renders
their use more difficult and much more expensive. Dogs are
still used to assess drug
transplantation.

The pig model is thus slowly replacing the dog
model in many experimental protocols because this animal
is accepted for human consumption. Thus, there is little
public resentment of public research and its use is less
expensive. Regarding the European legislation and in
contrast to classical laboratory animals and domestic
animals, outbred pigs for laboratory research may come
from farms, and colony breeding is not mandatory. The pig
has distinct advantages:

•
It breeds very easily.
•
Its anatomy, physiology, and immunology are
well known and quite comparable to those of humans.
•
It is reared and adapted easily to experimental
conditions and has few disease problems.
•
The ease of pig genetic engineering is a
major advantage since
biotechnology allow their transgenic breeding and cloning
(28,29).

The pig is thus very important as an animal
model and, as discussed later, is also a putative important
source of cells and organs for humans. The use of pigs as a
model has been paralleled by marked improvements in pig
immunological markers such as monoclonal antibodies and
pig cytokine detection (30).

The use of outbred pigs for transplantation
immunology research is common, but the lack of MHC
antigen control does not allow precise studies and
consequently well-defined result analysis. For this reason
mainly, Sachs D.H. developed his own inbred miniature pig
model that allows researchers to control all the MHC
combinations in several transplant models such as renal,
cardiac, thymic, and bone marrow. In fact, inbred pigs with
well-identified swine leukocyte antigen (SLA) (31) have
been used, particularly in allotransplant research (10). This
miniature pig model mimics all human MHC combinations
(10) and has a slow growing curve, but has one major
disadvantage related to breeding difficulties: small litters
caused by significant consanguinity. Certain minipig breeds
(Gottingen minipigs, Yucatan pigs) are also used,
pharmacokinetics in
the recent advances in

Page 4

Large animal model
4867
particularly for biomedical research because of their
reasonable size (adult weight = 30 kg).
Pigs need to be used at a reasonable weight to
perform transplantation surgery in laboratories with sterile
conditions, so the age of the animals is mostly approximate
to very young adults rather than to adults. As in rodents, the
immune systems of young pigs are still immature, and the
memory phenotype is still developing and the peripheral T
cell repertoire is predominantly naive. In addition, these
young animals have still a very active thymus at 3 to 6
months of age, which must be taken into consideration for
transplantation research. The data with experiments in adult
pigs should therefore still be confirmed when results are
found in young pigs. For this reason, the use of miniature
pigs from inbred colonies is a considerable advantage since
adult pigs can still be used at 150 kg, whereas outbred pigs
of similar age have often a weight of 250 to 300 kg, which
precludes their use in the laboratory.

NHPs have several important properties that
allow them to serve as a bridge between the basic insights
obtained in other animal models and their application to
humans (32). According to the Convention on International
Trade in Endangered Species of wild fauna and flora
(www.cites.org/eng/disc/text.shtml), experimental research
with Hominidae (the great apes) is prohibited except for
research in virology (AIDS and hepatitis). The two
Cercopithecidae families (baboons and macaques (rhesus
and cynomolgus) from the Old World provide most of the
monkeys used for biomedical research, especially in
hematology, immunology, and virology (33).

The primate immune system has important
similarities to humans:

•
A high level of homology has been reported
between humans and baboons for MHC class I molecules
(90%) (34).
•
Nucleotide and protein sequences of several
cytokines are almost homologous in humans and macaques
(93% to 99%) (35).
•
The similarities between the sequences of
macaque and human antibodies are equivalent to those
between human antibodies that originate from different
individuals (36).
•
The ABO system, hemoglobin, coagulation
and fibrinolysis physiology are comparable between
humans and baboons (37,38).

Although the immune system of NHPs has
important homology to the human immune system, there
are some notable exceptions such as antibodies for
Macacus CD3, which do not cross-react with human CD3
molecules and vice versa. In contrast, many CDs are still
unknown in Macacus in comparison to human molecules
and the determination of immunoglobulin subtypes is still
imprecise in baboons (39).

The use of NHPs however, faces considerable
hurdles, including social pressures, pathogen-free animals
(viruses), major expense, complexity of care, and low rate
of fecundity, but they represent crucial models for
transplantation research, especially in the field of tolerance
induction. Their size could present a problem because of
practical surgical aspects, but allows easier drug dosing.
Adult baboons weigh 20 to 25 kg, which allows repetitive
blood sampling and represents a good model compared
with pig organs in a xenotransplantation context. But
because there are no breeding colonies, the availability of
this species is limited. The most used NHP models in allo-
and xenotransplantation are today the cynomolgus and
rhesus macaques, as colonies live in various countries and
delivery is possible by following the rules in force.

The importance of using NHPs for biomedical
research has been emphasized by the European Union. In
the absence of in vitro models that can take into account the
complexity of humans, the use of NHPs remains the gold
standard before clinical trials. As will be described later,
the NHP model is the only real model to allow
comprehensive studies in xenotransplantation.

6. FACTORS CONTROLLING ALLOGRAFT OR
XENOGRAFT REJECTION

6.1. Allotransplantation
The rejection of human allografts may be
mimicked by large animal models, and immune responses
are mostly dependent on similar immune T-cells that
respond to the major histocompatibility antigens. However,
minor antigens, blood group and recipient status are also
involved:

•
MHC antigens: these are of overwhelming
importance in determining
allotransplantation in large animal models and in humans.
The matching of class II antigens has been particularly
emphasized in human and pig models (10). In hyperacute
and accelerated allograft rejection, sensitized recipients
may reject their allografts within minutes or hours, through
pre-existing anti-MHC antibodies against the donor.
Models for this type of rejection include pig models in
which the recipient is previously sensitized by skin grafts
from the future donor. This hyperimmunized pig model is
of overwhelming importance to study the control of anti-
MHC antibodies elicited in humans caused by previous
grafts, blood transfusions or pregnancies (40).
•
Non-MHC linked immune response and minor
antigens may play a part in allograft rejection in several
animal models.
•
Blood group antigens may also produce
hyperacute rejection (HAR) and need to be controlled in all
large animal models. Although blood group is easily
controlled in inbred pig models, outbred herds show ABO
incompatibilities and 16 porcine blood groups are
recognized (41). These ABO antigens must also be
matched to avoid unexpected conclusions; e.g., humoral
rejection, on data generated by such models. The same has
to be done in NHP models, at least in allotransplantation.

6.2. Xenotransplantation
In xenografts, less is known about cellular
rejection in discordant xenogeneic grafts, but enough data
seem to evidence the similarity of the humoral immune
the outcome of

Page 5

Large animal model
4868
response to xenoantigens to consider the pig-to-primate
model as the only model that is appropriate for predicting
data in humans.

Xenotransplantation is divided into two types,
concordant and discordant, according to the discrepancy in
species. Concordant xenotransplantation is a transplantation
across species in which there are no serum preformed
antibodies; e.g., chimpanzee-to-human or chimpanzee-to-
baboon. Discordant xenotransplantation is a xenograft across
more distant species in which the recipient possesses
circulating preformed antibodies against the donor antigens,
e.g. pig to primate or pig to human. Humans, apes and old-
world monkeys (baboon, Macacus) have antibodies against
Gal-α(1,3)-Galactosyl determinants, which are widely
expressed on pig vascular endothelial cells. These xeno-
antibodies, which are comparable to ABO isoagglutinins in
humans, probably result from sensitization to enterobacteria
that rapidly colonize the small bowel after birth (42). These
pre-existing (or xenoreactive) natural antibodies (XNAs),
mainly of IgM isotypes, bind to α-Gal epitopes, activate the
classical complement cascade, attract platelets, induce
thrombus formation and activate the coagulation pathways that
lead to rapid interruption of the graft function within seconds
or minutes (43). This phenomenon is called hyperacute
rejection. After overcoming this rejection process by either
eliminating the XNA reaction or by blocking the complement
activation, the organ graft can undergo acute vascular rejection
that is characterized by antibodies elicited against Gal residues,
infiltration of monocytes and natural killer (NK) cells,
interstitial hemorrhage and intravascular coagulation. Even if
these two types of vascular rejection are controlled, the xeno-
organ can still be rejected by cellular mediated rejection in
which NK cells and T-lymphocytes are involved (44,45,46).
So far, chronic rejection in xenotransplantation has not been
really studied in large animal models because no model
survived enough long to evaluate this type of immune
response.

The problems with vascularized organs need
more research in large animal models, but xenogeneic cell
transplantation might be less invasive and cause less severe
immune reaction, as will be discussed later.

7. IMMUNOSUPPRESSIVE DRUGS

Humans and large animal models show generally
similar physiological responses to most drugs tested if their
delivery is adequately controlled. The main differences
between large animals and humans are, however, the
consequence of drug metabolism or absorption and lead to a
drug response that is difficult to predict. In general, the only
way to use drugs in large animals is to refer to whole blood or
plasma levels and not to the dose given. For instance, to reach
a therapeutic level in pigs or rodents, calcineurin inhibitors
must be used at significantly higher doses than in humans
(47,48), but this higher dosage does not preclude lower level
use in humans if the circulating levels are comparable. In pigs,
an increased volume of distribution and clearance, as well as a
diminished systemic availability, explain lower plasma
concentrations of cyclosporin A and prednisolone compared to
humans. Thus, to obtain the same target concentrations of CyA
and prednisolone in both species, pigs require about 2 to 4
times higher I.V. or oral doses of CyA and 10 to 30 times
higher I.V. or oral doses of steroids (49).

As a consequence, all the experiments done in pigs
with a high dosage of calcineurin inhibitors should not be
discarded as nonapplicable to humans, but should be evaluated
carefully. In the early era of cyclosporin A, the dosages used in
humans were much too high. Unfortunately, some renal
allografts were lost because of the high dosages and
consequent nephrotoxicity. Nevertheless, each situation must
be clearly analyzed because stable tolerance has been
extensively demonstrated in miniature swine with a regimen of
12 days of CyA at 10mg/kg/day. These results could be
categorized as inapplicable to humans because the dosage is
too high, but adaptation to the pharmacology in pigs would
allow researchers to compare the effect with a two- to fourfold
reduction in other mammalian species such as humans. The
truth is that very few physicians would consider using only one
simple drug to induce tolerance in humans, but several clinical
trials recently demonstrated very good results in liver
transplantation with a regimen that included only
tacrolimus (50). In addition, true tolerance has been
induced in inbred swine for completely mismatched renal
allografts with clinically applicable low doses of tacrolimus
(0.1 mg/kg/day: 7 to 20 ng/ml level in the serum) as well as
liver tolerance with 0.1mg/kg/d for 12 days in semi-identical
pigs (51,52).

In total, cyclosporin A and tacrolimus have been
evaluated in dogs, pigs and NHPs with reproducible
prolongation of graft survival, or even tolerance. Azathioprine
and mycophenolate mofetil have been extensively tested in
dogs, whereas sirolimus was not well tolerated by dogs or
NHPs. In fact, the poor absorption of sirolimus by NHPs has
led to the use of increasing quantities of the drug and
consequent bowel lesions. Finally, sirolimus has mainly been
evaluated in monkeys at low doses and in the context of
multiple drug therapies (53,54).

8. IMMUNOSUPPRESSIVE THERAPIES IN LARGE
ANIMAL MODELS

8.1. Monoclonal antibody or fusion protein therapy
The main problem with monoclonal antibody
therapy in large animals is obviously the specificity.
Therefore, many monoclonal antibodies have been tested in
NHPs but very few in other large animal models because
there is little or no cross-reaction (55,56). New monoclonal
antibodies have emerged that may be viable candidates for
use in innovative
immunosuppressive combinations. With the high degree of
human specificity, many mAbs used in animal models have
been shown to be relevant in clinical human transplantation
and identified to target immune activation, except CD3
whose human-CD3-specific agents generally do not cross-
react with rhesus macaques (57).

Blocking the two main co-stimulatory targets
(B7-CD28 and CD40-CD154) has been the main objective
and blocking these two receptors synergistically induced
tolerance in mice (58,59). In NHPs, monoclonal antibodies
low-dose drug-sparing