In vivo imaging using
bioluminescence: a tool for
probing graft-versus-host disease
Robert S. Negrin and Christopher H. Contag
Abstract | Immunological reactions have a key role in health and disease and are
complex events characterized by coordinated cell trafficking to specific locations
throughout the body. Clarification of these cell-trafficking events is crucial for
improving our understanding of how immune reactions are initiated, controlled
and recalled. As we discuss here, an emerging modality for revealing cell
trafficking is bioluminescence imaging, which harnesses the light-emitting
properties of enzymes such as luciferase for quantification of cells and uses low-
light imaging systems. This strategy could be useful for the study of a wide range
of biological processes, such as the pathophysiology of graft-versus-host and
Normal immune function is crucial for
maintaining health in a hostile environ-
ment containing many potential pathogens.
To protect against pathogens, immune
responses must be both rapid and sustained.
In addition, effector cells of the immune
system provide immune surveillance against
malignancy and promote tissue remodelling
and repair. The importance of a normal
immune response is further exemplified
by the observation that patients with many
diseases, including cancer and infectious and
auto immune diseases, have dysfunctional
immune responses. Therefore, a greater
understanding of normal and pathological
immune responses will not only provide
insights into basic biological mechanisms but
also will aid in the development of effective
treatments for a range of diseases.
In this Innovation article, we discuss
how bioluminescence imaging (BLI) can be
used to analyse aspects of complex immune
reactions in living animals. We use graft-versus-
leukaemia (GVL) reactions and graft-versus-host
disease (GVHD) as examples of what BLI can
teach us about clinically relevant immune
Nowhere in clinical medicine is the impor-
tance of effective immune responses clearer
than after allogeneic haematopoietic-cell
transplantation (HCT). Over the past several
decades, HCT has emerged as an effective
and often life-saving treatment for a broad
array of haematological malignancies, as well
as for genetic and acquired immune deficien-
cies1. HCT involves the transfer of the entire
haematopoietic and immune systems from a
donor to a recipient. The procedure involves
pretreatment of the patient with high doses
of chemotherapy, with or without irradiation,
to eliminate malignant or defective
haemato poietic cells. This is followed by the
transfer to the patient (recipient) of donor-
derived haematopoietic cells, which home
to the bone marrow and re-establish
haematopoiesis. After HCT, all of the haema-
topoietic cells in the recipient, including
professional antigen-presenting cells (APCs)
and immune effector cells, are of donor
origin. The functional consequences for the
recipient of the new donor-derived immune
system are dramatic and include the ability to
reject the underlying malignancy or replace
damaged haemato poietic-cell populations
with normal cells. For example, the damaged
red blood cells in patients with thalassaemia
or sickle-cell disease can be replaced with
healthy cells. The rejection of malignancy is
known as the GVL effect. The antigens on
the tumour cells that are recognized by the
donor leukocytes are largely unknown but
they include major and minor histocompatibility
antigens (depending on the genetic disparity
between the donor and the recipient) as well
as potential tumour-specific antigens.
Despite its success in promoting the rejec-
tion of malignancy, allogeneic HCT carries
the significant risk that the donor-derived
immune cells will recognize and respond to
recipient tissues and result in a syndrome
known as GVHD2,3. Severe GVHD limits the
overall effectiveness of HCT and precludes
the application of this life-saving therapy to
other clinical settings, such as for the treat-
ment of severe autoimmune disorders4 or for
the induction of tolerance to organ transplan-
tation. The risks of GVHD are substantial,
such that 20–60% of patients will develop
this complication following allogeneic HCT,
depending on disease-related factors, such as
the stage of disease, the age of the recipient
and the degree of genetic disparity between
the donor and recipient. The reasons why
some patients develop severe GVHD after
HCT, whereas others do not, remain unclear.
Given the substantial risk of GVHD, applica-
tion of current HCT procedures requires that
the donor is a fully matched histocompatible
sibling or unrelated donor. Unfortunately,
many patients that need HCT are unable to
secure a well-matched donor, resulting in
lethal consequences for the patient due to
disease progression. Therefore, the study of
GVHD and GVL reactions provides insight
into both normal and pathological immune
reactions and has significant implications for
the development of new and more effective
strategies for clinical management of disease.
Experimental models of HCT
HCT has been widely studied in both rodent
and canine models, and these studies have
been crucially important in developing the
theoretical basis of HCT. Initial studies in
mice established the concept of allogeneic
immune responses and GVHD. Mouse
studies have also been crucial for studying
allorecognition, and for exploring effector-cell
populations and mechanisms, owing to the
defined genetics and availability of strains
that lack key effector molecules. Early stud-
ies in dogs showed that in some litter-mates
long-term engraftment occurred, whereas
in others GVHD developed. The outbred
canine model has been useful in the develop-
ment of preparative regimens for transplan-
tation, which were translated to the clinic5,6,
and to study the use of drug prophylaxis
for both acute and chronic GVHD in large
animals. More recent studies have led to
the development of non-myeloablative HCT,
in which the intensity of the preparative
regimen of the recipient is reduced and
replaced with immuno suppressive
medications to prevent graft rejection. This
strategy of HCT is associated with reduced
transplant-related morbidity and mortality7.
Non-myeloablative HCT has been widely
applied in the clinic; however, GVHD
remains a serious complication of the
therapy with an associated mortality rate
484 | JUNE 2006 | VOLUME 6
© 2006 Nature Publishing Group
Transfer cells to
Inject luciferase reagent
of 10–20% (REF. 8). Recent studies in which
the immune environment of the recipient is
altered before transplantation using total
lymphoid irradiation (TLI) and depleting
antibodies that target T cells (anti-thymocyte
globulin) — an approach pioneered in
mouse model systems — have been trans-
lated to the clinic with an apparent reduced
risk of acute GVHD9.
Mouse studies have also been important in
the development of a conceptual framework
for GVHD reactions. As in humans, GVHD
in animal models shows an unusual tissue
tropism; it mainly affects the skin, gastrointes-
tinal tract and liver. Tissue injury in the recip-
ient owing to the preparative regimen results
in the release of pro-inflammatory cytokines
that fuel the GVHD reaction2,3. Alloreactive
donor-derived T cells can become activated in
the recipient, and can then infiltrate and sub-
sequently cause damage to tissues of the skin,
gut and liver, resulting in the pathophysiology
of GVHD. In mice, this alloreactivity trans-
lates as end-organ damage such as hair loss,
ruffled fur, weight loss, diarrhoea and eventu-
ally mortality. However, these experimental
end-points are reflective of end-stage disease
and provide little information about the
spatial and temporal events in the induction
of GVHD at time points when intervention
could affect the ultimate outcome. To explore
GVHD and GVL reactions in greater detail,
imaging strategies have been used to visualize
the spatial and temporal events in GVHD
Visualizing immune responses
Until recently, much of our understanding
of immune-cell trafficking and the factors
that control this process has been obtained
by using culture systems, in which the influ-
ence of intact organ structure, circulation,
endothelial barriers and tissue effects have
been removed. Insights into the specific
locations and timing of immune-cell migra-
tion and proliferation that can be gained
using imaging methods in living animals
hold promise for providing new informa-
tion on physiology and pathophysiology.
Various imaging modalities are emerg-
ing for the study of small animal models of
human biology and disease, and several
of these have been applied to the study of
immune-cell migration. There are preclinical
versions of clinical imaging systems, such as
magnetic resonance imaging (MRI), positron
emission tomography (PET) and single
photon emission computed tomography
(SPECT). For example, MRI has recently been
used to track cardiac-graft rejection in a rat
model10. The use of these tools in studies of
small animals allows ready translation of new
reagents and imaging approaches to the clinic.
However, the use of optical markers, such as
those that are bioluminescent or fluorescent,
to assess cell fate and function in whole
animals offers several advantages that can
be used to refine and accelerate the study of
mouse models of disease (BOX 1). The instru-
mentation for acquiring whole-body images
using optical reporters that are expressed in
small animals also has several advantages over
other approaches: it can usually accommodate
multiple animals in a single image at each
time point; it is user-friendly so dedicated
imaging technicians are not required; and
it is less expensive than the instrumenta-
tion required for SPECT, MRI and PET.
The use of two-photon intravital microscopy
offers high-resolution images of cell–cell
interactions in tissues11, but is constrained
by a limited field of view and can be severely
hampered by motion artefacts. As a result
of these limitations and constraints, only
certain tissues can be observed using this
In the field of optical imaging, there are
several approaches that have been used to
generate whole-body images of biological
processes in rodents. These include the use
of light scatter, absorbance, fluorescence
and bioluminescence. In this Perspective
article, we focus on the use of biolumin-
escence to generate images of cell migra-
tion and proliferation in vivo and on the
Box 1 | Molecular imaging in immunology: watching and waiting for an immune response
Imaging modalities. Two-photon intravital microscopy offers resolution of cells in vivo11, but it is
constrained by small fields of view and motion artefacts. Non-invasive measures of immunological
processes in vivo have been accomplished using positron emission tomography (PET)49–51, single
photon emission computed tomography (SPECT)52, magnetic resonance imaging (MRI)53,
bioluminescence imaging (BLI) and fluorescence imaging (FLI)49,54–56. Ultrasound and X-ray
computed tomography (CT) provide anatomical information, and when used in combination with
other modalities, this information improves localization of the signals obtained by PET, SPECT
or optical imaging57. PET, SPECT, ultrasound MRI and CT have potential clinical uses, and therefore
are useful in translational studies.
Optical methods. BLI and FLI can be used to refine and accelerate studies of animal models, but
they have limited clinical application. Imaging times for optical imaging methods are generally
short, which facilitates the analysis of greater numbers of animals. Optical methods also allow a
range of image resolutions from microscopic to macroscopic, produce images without the use of
ionizing radiation, offer the choice of many reporters and dyes, and benefit from user-friendly and
inexpensive instrumentation58. Signal-to-noise ratios (SNRs) for BLI are excellent59,60, enabling
detection of subtle changes non-invasively, thereby obviating the need to remove overlying tissue.
Reporter genes. The use of dyes and contrast agents allows visualization of the early events, but
they are diluted by cell division. To prevent loss of labels during cell division, genes that encode
reporter proteins can be integrated into the genome. Reporter genes are available for PET, SPECT,
MRI and optical imaging49,57, each with strengths and weaknesses. The radiotracers used for PET
and SPECT often produce signals from kidney, liver and bladder that can obscure the target tissue,
and MRI is generally less sensitive than imaging of reporter-gene expression with PET or SPECT.
In summary, optical imaging of reporter-gene expression in vivo offers the greatest versatility,
sensitivity and SNR of all of the modalities used for small animals.
Figure 1 | Schematic representation of a bio-
luminescence imaging strategy using cells
from a transgenic donor mouse. Cells express-
ing the transgene encoding a luciferase–GFP–
green fluorescent protein (GFP) fusion protein
are isolated from a transgenic donor animal and
selected by cell-sorting technologies, using the
GFP signal or fluorescent antibodies specific for
selected cell-surface markers. Luciferase–GFP-
positive cells are then transferred to recipient
syngeneic or allogeneic animals. Recipient ani-
mals are then injected with the luciferase sub-
strate to allow serial imaging of the biolumines-
cent signal in vivo. The tracking of effector cells
involved in graft-versus-host disease or graft-
versus-leukaemia reactions can be carried out in
recipient animals that have been prepared using
myelo ablative or non-myeloablative regimens.
NATURE REVIEWS | IMMUNOLOGY
VOLUME 6 | JUNE 2006 | 485
© 2006 Nature Publishing Group
Number of GFP-positive cells
Number of cells per animal
Number of cells per well
Relative intensity of
bioluminescent signal (× 105)
0.1 1,000100 101 0.1
0.1 1,000100101 0.1
use of these data to guide sampling of the
appropriate tissues at times when biological
changes are occurring.
The use of in vivo bioluminescence was first
shown in tracking bacterial pathogens12.
In vivo BLI has since been applied to the
study of gene-expression patterns13, as a
measure of successful gene transfer14, to
indicate the extent of tumour growth and
response to therapy15, to assess the extent of
protein–protein interactions in vivo16 and to
determine the location and proliferation of
stem cells17. These examples show the versa-
tility of the approach and some of the basic
principles of the method (FIG. 1).
BLI is based on the expression of a
light-emitting enzyme (such as luciferase)
in target cells and tissues. In the presence of
its substrate (such as luciferin), an energy-
dependent reaction releases photons that
can pass through tissues and be detected
using sensitive detection systems. As with
all optical imaging approaches, BLI is sub-
ject to the optical properties of tissues18. For
example, in the visible region of the spec-
trum, haemoglobin is the main absorber in
the body, and this has a marked influence
on the transmission of light through tissues.
The influence of tissue on the detection of
bioluminescent reporters in vivo has been
studied using four luciferases in three animal
models19. The data generated were consist-
ent with those reported using external light
sources, and they showed that the longer
wavelengths of bioluminescence (>600 nm),
in the red and near-infrared regions of the
spectrum, are transmitted through mam-
malian tissues more efficiently than the
shorter wavelengths of light, in the blue and
green regions of the visible spectrum. This
indicates that using luciferases that have a
significant portion of their emission greater
than 600 nm, such as luciferase derived from
fireflies and click beetles (approximately
60% of the light emitted from these two
enzymes has wavelengths greater than 600
nm), will lead to more-sensitive detection
of the labelled cells in vivo. In each animal
model, the sensitivity of detection should
be measured by determining cell numbers
(FIG. 2), assessing luciferase activity in excised
tissues20 or using other imaging modalities21.
Cells can be counted after recovery from
animals (FIG. 2a–c) or before introduction
into animals (FIG. 2d,e). These measures
have indicated that, in mice, a minimum of
100–1,000 cells can be detected at specific
anatomical sites. Among the parameters
that affect the sensitivity of detection are
the wavelength of light emission, expression
levels of the enzyme in the target cells, the
location of the source of bioluminescence in
the animal, the efficiency of the collection
optics and the sensitivity of the detector.
Detection of internal biological sources
of light requires sensitive detection systems
with spectral sensitivity in the red region of
the spectrum. Charge-coupled device (CCD)
cameras (similar to a home camcorder) in
general, have a spectral range that is appropri-
ate for detecting biological light sources. But
most CCD detectors are not sensitive enough
to detect the light from inside the body that
is needed for the study of cell migration or
other biological processes. To increase their
sensitivity, these detectors can be cooled.
Alternatively, approaches can be used to
amplify or intensify the signal, but many
intensifiers used for amplifying optical signals
are not sensitive in the red region of the spec-
trum. So, CCD cameras, in which the CCD
chip is thinned, back illuminated, placed in
Figure 2 | Sensitivity of detection in bioluminescence imaging studies.
The detection of weak optical signals, such as those generated during bio-
luminescence imaging (BLI), from inside small animals is influenced by sev-
eral factors that include: the level of cell brightness (photon flux from
source), the depth at which the bioluminescent source is located in the tis-
sue, the wavelength of the emitted light, the quantum efficiency and noise
of the detector, the nature of the collection optics and the background
emission levels from the live animal. a | In this example, studies of the detec-
tion sensitivity of BLI were carried out in a mouse B-cell lymphoma model,
in which the tumour cells were labelled with luciferase and green fluor-
escent protein (GFP), by imaging whole animals at various times during the
disease course. One time point is shown for two mice. b | After recovery from
the animals, tumour cells from the liver and spleen of these animals can then
be quantified by flow cytometry using the GFP signal and fluorescent anti-
bodies specific for the B-cell marker CD19; the results for the two animals
in part a are shown. c | The quantity of GFP-positive cells detected by flow
cytometry correlated well with the bioluminescent signals detected in vivo,
showing that BLI is a sensitive and reliable measure of cell number in vivo.
d | Using the same detector, the detection sensitivity of BLI can also be ana-
lysed by measuring the bioluminescent signal emitted from known numbers
of cells in culture or following transfer in vivo (e). Images are adapted, with
permission, from REF. 24 © (2003) the American Society of Hematology.
486 | JUNE 2006 | VOLUME 6
© 2006 Nature Publishing Group
Days after transplantation of luciferase-positive splenocytes
Days after transplantation of luciferase-positive splenocytes
Figure 3 | Imaging of graft-versus-host disease. Bioluminescence imaging of luciferase-positive
splenocytes transplanted to either irradiated syngeneic (top panels) or allogeneic (bottom panels)
animals are shown. Serial images show markedly different patterns of lymphocyte trafficking, pro-
liferation and tissue infiltration. At defined time points, tissue sites of interest, as determined by
bioluminescence imaging, can then be further analysed. Images are reprinted, with permission, from
REF. 25 © (2005) the American Society of Hematology.
a vacuum and cooled to temperatures as low
as –105°C, are at present the most common
cameras for imaging weak biological sources
of light in the body.
For BLI, cells must be engineered to
express the reporter luciferase and the sub-
strate for the reaction must be injected into
the animal for light to be emitted. Luciferases
have been cloned from both marine (such
as Renilla luciferase) and terrestrial
(such as firefly and click-beetle luciferases)
eukaryotic organisms. The substrates that
these luciferases use seem to group with their
origins — marine bioluminescent organ-
isms use coelenterazine as a substrate and
terrestrial organisms use d-luciferin. The
biodistribution of these substrates has been
studied in animals and differs significantly.
d-luciferin has a longer circulation time than
coelenterazine and there is little catalysis
of d-luciferin by mammalian proteins22.
These two differences determine the use
of the enzyme–substrate pair for in vivo
BLI in which luciferase–luciferin reactions
provide a longer duration of signal at a longer
wavelength that is less influenced by tissue
absorption than enzyme–substrate reactions
that emit blue light (for example, that are pro-
duced by luciferases from marine organisms
such as Renilla spp.). Coelenterazine-using
enzymes can provide a short-lived signal
that is useful when combining the assays
with luciferin-using enzymes22,23. However,
all enzymes characterized so far that use
co elenterazine emit blue light, which is highly
absorbed by mammalian tissues. As more
enzyme–substrate pairs are characterized and
existing reactions are optimized for in vivo
applications, more reagents will become avail-
able for accelerating and refining animal stud-
ies using BLI. At present, the luciferin-using
enzymes from fireflies and click beetles that
emit light greater than 600 nm offer the great-
est sensitivity, and the coelenterazine-using
enzymes can be used as secondary markers,
despite their severe limitations as convenient
and sensitive markers.
Using BLI to explore GVHD
BLI is an effective means of evaluating com-
plex biological processes such as stem-cell
engraftment, GVHD and GVL reactions17,24,25.
BLI is remarkably sensitive; it can detect
as few as 10 cells in vitro and 100–1,000
cells in vivo (FIG. 2). The ability to track cell
populations serially and non-invasively, so
as to define key time points and locations for
further analysis, has provided important new
insights. Whole-body imaging of cell migra-
tion guides investigators to specific times and
organs for more labour-intensive assays.
A central limitation of all reporter-gene
strategies is the need to introduce the reporter
construct into the cell populations of interest.
Gene transfer using viral vectors or non-viral
strategies have been useful; however, they
suffer from variable efficiency, especially
when attempting to introduce genes into
certain primary cells. These techniques also
often require cell activation and culturing
for variable periods of time, and this might
alter the biological activity of the cells.
Transduction of haematopoietic stem cells
(HSCs) has been carried out at high frequen-
cies using viral vectors; these transduced cells
can then be used directly in vivo to visualize
engraftment26. In addition, transduced HSCs
can be transplanted into immunodeficient
mice, such as recombination-activating
gene 2 (Rag2)–/– animals (which lack B and
T cells), allowed to engraft and then effector
cells can be isolated for secondary transfer27.
This approach can be particularly useful for
studying the migration of cells derived from
particular strains of mice that lack key effec-
tor molecules, as backcrossing these animals
to reporter-gene-expressing mice is expensive
and time consuming. However, if single HSCs
or small numbers of HSCs are to be studied,
then uniform integration of the reporter gene
into a given genomic site17 is preferred to
retroviral transduction, which results in each
cell having the reporter gene integrated into
An alternative approach to gene transfer
has been the generation of transgenic mice
that express a luciferase–green fluorescent
protein (GFP) fusion protein under the
control of the chicken β-actin promoter and
the cytomegalovirus enhancer in all haemato-
poietic cells. Cells from these transgenic mice
(known as L2G85 mice) can then provide
a source of luciferase-positive donor cells
for transplantation studies17. The presence
of GFP facilitates the isolation of these cells
from the donor animals and can also be used
as a marker in fluorescence microscopy or
flow-cytometry studies of transplanted cells
in tissues from recipient animals. Although
transgenic animals expressing GFP alone
have been used to provide donor cells in
studies of GVHD that show widespread infil-
tration of tissues by donor-derived cells28, it is
possible that the cells of interest lose expres-
sion of the transgene or are recognized by
the immune system and deleted29. Therefore,
control experiments with wild-type cells are
important to verify these results.
In studies with cells from L2G85 mice,
syngeneic (genetically identical; FVB; H2q)
or allogeneic (genetically different; BALB/c;
H2d) recipient animals first received lethal
irradiation (to delete the existing haematopo-
ietic cells and simulate the clinical setting of
HCT) followed by injection of T-cell-depleted
bone-marrow cells from wild-type donor
animals (to re-establish haematopoiesis) and
NATURE REVIEWS | IMMUNOLOGY
VOLUME 6 | JUNE 2006 | 487
© 2006 Nature Publishing Group
T cells and
splenocytes from L2G85 animals (to induce
GVHD)25. To visualize the donor cells, luci-
ferin, the substrate for luciferase, was injected
into the recipient just before imaging. Serial
imaging showed striking differences between
syngeneic and allogeneic recipients (FIG. 3).
In the syngeneic animals, a waxing and
waning BLI signal from the transplanted
luciferase-positive cells was observed, which
ultimately resulted in bone-marrow engraft-
ment, probably from residual stem cells in
the splenocyte preparations25. By marked
contrast, the transplanted cells in allogeneic
recipients showed early (in the first 24–48
hours) infiltration of cervical lymph nodes
and structures in the gut. At 2–4 days after
transplantation, marked proliferation of the
donor cells was observed at these lymph node
and gut sites, indicated by the increase in
BLI signal. By day 6, infiltration of the skin
(most obviously in the ears and tail) was
readily apparent25. This study identified the
key target structures and organs involved in
the induction of GVHD, so further analyses,
using ex vivo BLI-, immunofluorescence- and
flow-cytometry-based approaches, can focus
on these tissues.
Visualizing early migration of donor T cells.
Before donor-cell infiltration of GVHD target
organs, such as the skin, gut and liver, infiltra-
tion of secondary lymphoid structures was
seen. Histological analysis showed that, in the
gut, these early sites of donor-cell prolifera-
tion were the Peyer’s patches and mesenteric
lymph nodes. These structures are extremely
difficult to detect in irradiated animals but
became easily visible by BLI following HCT25.
In all of the lymphoid tissues analysed, CD4+
T cells infiltrated first (as early as 24 hours
after transfer), followed (several days later)
by CD8+ T cells.
Previous studies have indicated a key role
for Peyer’s patches in the induction of GVHD
as, under certain experimental conditions,
animals that lack such structures do not
develop disease30. Using a myeloablative-
conditioning regimen, we have found that
alloreactive T-cell activation and proliferation
occurs in several sites, including Peyer’s
patches, mesenteric lymph nodes and other
nodal sites as well as the spleen (A. Beilhack,
S. Schulz and R.S.N., unpublished observa-
tions). This is consistent with results reported
by others indicating that animals that lack
Peyer’s patches still develop GVHD31. These
studies indicate that priming (activation) of
alloreactive T cells seems to occur in second-
ary lymph nodes and the spleen. Following
activation, alloreactive cells upregulate
the expression of key molecules, such as
α4β7-integrin, that are required for entry to
GVHD target organs. A central question
is whether cells are imprinted for entry to
specific sites at their site of priming — for
example, are cells that are activated in Peyer’s
patches and mesenteric lymph nodes destined
to enter the gut? And similarly, are other sites
of activation required for cell infiltration of
the skin? In vitro studies support this hypo-
thesis and have indicated that APCs from
particular sites, such as the Peyer’s patches,
activate cells that can infiltrate the gut but not
BLI is therefore extremely useful to
define the time points and sites of donor-cell
infiltration, and to direct further analyses
using pheno typic and functional assays.
Validation of these concepts in vivo with
re-transplantation of luciferase-positive cells
activated at specific sites and analysis by BLI
will be invaluable for directly answering the
Using BLI to evaluate strategies to reduce
GVHD. A major goal of these studies is to
develop strategies that can reduce the risk of
GVHD but do not interfere with GVL reac-
tions. The studies described earlier indicate
that there are several possible approaches to
control GVHD, which include blocking or
limiting access of T cells to priming sites,
controlling alloreactive T-cell proliferation
and blocking entry to GVHD target organs.
Because access to secondary lymphoid
structures was key to activation and prolif-
eration of alloreactive T cells, several studies
have indicated that different populations
of T cells might have variable access to
such sites and, therefore, differential abil-
ity to induce GVHD25,33,34. For example,
effector memory T cells (defined by a
CD4+CD44hiCD62Llow phenotype) can
provide increased immune reconstitution
and GVL effects but have limited capacity to
induce GVHD33,34. By contrast, naive T cells
(defined by a CD4+CD44lowCD62Lhi pheno-
type) readily induce GVHD. Using BLI to
follow the trafficking and survival of these
T-cell populations, it was shown that, unlike
naive T cells, effector memory T cells did not
infiltrate and proliferate in secondary lymph
nodes25. Other studies have highlighted the
importance of expression of the leukocyte-
adhesion molecule CD62L by T cells in
GVHD induction, also indicating that there
is a need to access secondary lymphoid
structures to initiate GVHD35.
Other populations of T cells and natural
killer (NK) cells with cytolytic activity have
also been studied in models of GVHD
and GVL. The generation of T cells with
defined reactivity — for example, against
minor histo compatibility antigens that are
expressed exclusively by malignant cells or
against viral antigens in Epstein–Barr-virus-
associated diseases — is currently being
explored and holds significant promise
for inducing GVL reactions36. Another
T-cell population that can be expanded
ex vivo, known as cytokine-induced killer
(CIK) cells, has a limited capacity to induce
GVHD in mouse models due, at least in
part, to the production by CIK cells of
interferon-γ (IFNγ)37. BLI studies of CIK
cells have indicated that they might have
reduced proliferative capacity compared
with naive T cells, which is consistent with
their attenuated capacity to induce GVHD.
NK cells have also been observed to lack
the capacity to induce GVHD but can have
GVL effects38,39. Accordingly, infusion of
NK cells from donors to HLA-mismatched
recipients with relapsed malignancies results
in very limited GVHD, but in some recipi-
ents, especially those with acute myeloid
leukaemia, beneficial GVL responses were
observed40. The precise mechanisms under-
lying the inability of NK cells to induce
GVHD are unclear. BLI studies of NK cells
Figure 4 | Effect of transfer of conventional
CD4+ and CD8+ T cells with and without CD4+
CD25+ regulatory T cells on tumour progres-
sion. Leukaemia cells expressing a transgene
encoding the fusion protein luciferase–yellow
fluorescent protein were injected into recipient
mice and, using bioluminescence imaging, can be
observed infiltrating the bone marrow. Recipient
mice that received irradiation and T-cell-depleted
bone marrow only have progressive tumour
growth at day 5 and 15 (left panels). Animals that
received T-cell-depleted bone marrow and con-
ventional T cells die rapidly due to acute lethal
graft-versus-host disease (GVHD). By contrast,
recipient mice that received both conventional
T cells and CD4+CD25+ regulatory T cells
(TReg cells) in equal proportions retain the ability
to reject the tumour without significant GVHD.
Images are reproduced, with permission, from
Nature Medicine REF. 27 © (2003) Macmillan
488 | JUNE 2006 | VOLUME 6
© 2006 Nature Publishing Group
and have resulted in reduced levels of acute
GVHD with retention of GVL activity9.
Another well-characterized population
of regulatory T cells is the subset of naturally
occurring CD4+ T cells that express the IL-2
receptor α-chain (also known as CD25)43
and the transcription factor forkhead box
P3 (FOXP3): that is, CD4+CD25+ regula-
tory T cells (TReg cells). Several groups have
shown that the infusion of grafts containing
an equal mix of TReg cells and conventional
T cells results in the control of GVHD44–46;
crucially, GVL reactions are retained in ani-
mals treated in this way27,47. Using BLI, active
rejection of the leukaemia could be observed
following the infusion of equal numbers
of luciferase-positive conventional T cells
and TReg cells27 (FIG. 4). The mechanism of
control of GVHD and retention of GVL
seems to be due to the ability of TReg cells to
suppress conventional T-cell proliferation, as
shown using BLI to evaluate donor-derived
conventional T-cell trafficking and numbers
in the presence and absence of TReg cells27.
By contrast, GVL reactions mainly required
activation of CD8+ T cells, and this occurred
even in the presence of TReg cells. However,
important questions remain: where, how
and for how long do TReg cells exert their
immunological control? Clues have come
from studies of TReg cells that are divided into
subsets on the basis of CD62L expression.
Both CD62L+ and CD62L– TReg cells sup-
press cell proliferation in an in vitro T-cell
proliferation assay and express FOXP3;
however, only CD62L+ TReg cells can suppress
GVHD in vivo46,48, indicating that CD62L-
mediated homing of TReg cells is required
to control GVHD. The clinical application
of TReg cells is under active development by
BLI has provided new insights into complex
biological processes that, until now, could
not be evaluated. Future goals include
improvement of the techniques that allow
visualization of more than one population
of cells simultaneously and improvement
in quantification of cell numbers. Improved
quantification might be possible through
spectral imaging, in which the differential
transmission of blue and green components
of the luciferase emission spectra relative
to the red components can be used to
determine the depth of the signal in the
body. Reconstruction of three-dimensional
images from multiple views will also
improve the quantification of biolumines-
cence signals. Although the ability to carry
out imaging studies using luciferase in
humans will be limited to very superficial
sites in which expression of a foreign gene
is not problematic, lessons learned from
using BLI in animal models will greatly
affect the design of clinical trials for cellular
transplantation. Perhaps the greatest benefit
of BLI to clinical medicine will be through
accelerating and refining preclinical mod-
els. Evaluating the fate of transferred cell
populations is likely to prove crucial to the
evaluation of HCT and other cell-based
Robert S. Negrin and Christopher H. Contag are at the
Departments of Medicine, Center for Clinical Research
Building, 269 West Campus Drive and Pediatrics, Clark
Center, 318 Campus Drive, Stanford University,
Stanford, California 94305, USA.
Correspondence to R.S.N.
Negrin, R. S. & Blume, K. in Williams Hematology
7th edn Ch. 22 (eds Lichtman, M. A. et al.) 209–247
(McGraw-Hill Professional, 2005).
Ferrara, J. L. & Deeg, H. J. Graft-versus-host disease.
N. Engl. J. Med. 324, 667–674 (1991).
Kaitin, K. I. Graft-versus-host disease. N. Engl. J. Med.
325, 357–358 (1991).
Sykes, M. & Nikolic, B. Treatment of severe
autoimmune disease by stem-cell transplantation.
Nature 435, 620–627 (2005).
Thomas, E. D., Storb, R. & Clift, R. A. Bone-marrow
transplantation. N. Engl. J. Med. 292, 832–843
Thomas, E. D., Storb, R. & Clift, R. A. Bone-marrow
transplantation. N. Engl. J. Med. 292, 895–902
Storb, R. et al. Stable mixed hematopoietic chimerism
in dogs given donor antigen, CTLA4Ig, and 100 cGy
total body irradiation before and pharmacologic
immunosuppression after marrow transplant. Blood
94, 2523–2529 (1999).
in mouse models are ongoing, and it is
hoped that they will provide further clues
to why this cell population has such limited
capacity for GVHD induction.
Monitoring regulatory T-cell function. An
alternative approach to controlling GVHD
has been to harness regulatory mechanisms
in an effort to control alloimmune responses.
Several populations of regulatory T cells
that can control immune reactions, such as
a mixed lymphocyte reactions in vitro and
autoimmune diseases in vivo, have been
described. One such population of regulatory
T cells, known as natural killer T (NKT) cells,
co-expresses both NK-cell and T-cell markers,
recognizes CD1 through an invariant T-cell
receptor and produces large amounts of
interleukin-4 (IL-4) after activation41. Mouse
recipients treated with TLI and anti-
thymocyte serum (ATS) to delete the existing
T cells have reduced numbers of conven-
tional CD4+ and CD8+ T cells but increased
numbers of NKT cells42. Interestingly, these
animals are resistant to GVHD induction
such that 1,000-fold more allogeneic donor
cells can be transferred without GVHD
developing42. The use of BLI to follow the fate
of donor cells transplanted to animals treated
with TLI and ATS compared with animals
receiving total body irradiation is under
active investigation. Recipient preparation
strategies involving TLI and T-cell depletion
before HCT have been translated to the clinic
Allorecognition occurs when the host immune system
detects same-species, non-self antigens and triggers
allograft rejection. It can occur by direct or indirect
pathways: the direct pathway involves recognition of
foreign MHC molecules on donor cells, and the indirect
pathway involves processing and presentation of donor-
derived MHC molecules by host antigen-presenting cells.
(GVHD). Tissue damage in a recipient of allogeneic tissue
(usually a bone-marrow transplant) that results from the
activity of donor cytotoxic T cells recognizing the tissues
of the recipient as foreign. GVHD varies markedly in extent,
but it can be life threatening in severe cases. Damage to
the liver, skin and gut mucosae are common clinical
Graft versus leukaemia
Hosts with leukaemia who receive an allogeneic bone-
marrow transplant have far fewer disease relapses than
individuals who obtain autologous bone-marrow
transplants. This results from the transplanted T cells
recognizing alloantigens expressed by the leukaemia.
The commitment and differentiation processes that lead
from a haematopoietic stem cell to the production of
mature cells of all lineages: erythrocytes, myeloid cells
(such as macrophages, mast cells, neutrophils and
eosinophils), B and T cells, and natural killer cells.
Minor histocompatibility antigens
Polymorphic peptides derived from normal cellular
proteins that can be recognized in the context of
MHC molecules. Immune responses to these
polymorphic antigens can result in graft-versus-host
reactions, graft rejection or beneficial antitumour
An allogeneic haematopoietic-cell transplantation in a
recipient who has received a conditioning regimen to
achieve immunosuppression and prevent graft rejection
without the complete ablation of host haematopoiesis.
The recipient might develop (transient) mixed chimerism,
owing to haematopoietic recovery of the host and
engraftment of donor haematopoietic cells.
Two-photon intravital microscopy
Laser-scanning microscopy that uses pulsed infrared laser
light for the excitation of conventional fluorophores or
fluorescent proteins. The main advantage is deep tissue
penetration of the infrared light, owing to the low level of
light scattering in the tissue.
NATURE REVIEWS | IMMUNOLOGY
VOLUME 6 | JUNE 2006 | 489
© 2006 Nature Publishing Group
8. Download full-text
McSweeney, P. A. et al. Hematopoietic cell
transplantation in older patients with hematologic
malignancies: replacing high-dose cytotoxic
therapy with graft-versus-tumor effects. Blood
97, 3390–3400 (2001).
Lowsky, R. et al. Protective conditioning for
acute graft-versus-host disease. N. Engl. J. Med.
353, 1321–1331 (2005).
10. Wu, Y. L. et al. In situ labeling of immune cells with
iron oxide particles: an approach to detect organ
rejection by cellular MRI. Proc. Natl Acad. Sci. USA
11. Sumen, C., Mempel, T. R., Mazo, I. B. &
von Andrian, U. H. Intravital microscopy: visualizing
immunity in context. Immunity 21, 315–329
12. Contag, C. H. et al. Photonic detection of
bacterial pathogens in living hosts. Mol. Microbiol.
18, 593–603 (1995).
13. Contag, C. et al. Visualizing gene expression in
living mammals using a bioluminescent reporter.
Photochem. Photobiol. 66, 523–531 (1997).
14. Lipshutz, G. S., Flebbe-Rehwaldt, L. & Gaensler, K. M.
Reexpression following readministration of an
adenoviral vector in adult mice after initial in utero
adenoviral administration. Mol. Ther. 2, 374–380
15. Sweeney, T. J. et al. Visualizing the kinetics of tumor cell
clearance in living animals. Proc. Natl Acad. Sci. USA
96, 12044–12049 (1999).
16. Massoud, T. F., Paulmurugan, R. & Gambhir, S. S.
Molecular imaging of homodimeric protein–protein
interactions in living subjects. FASEB J. 18,
17. Cao, Y. A. et al. Shifting foci of hematopoiesis during
reconstitution from single stem cells. Proc. Natl Acad.
Sci. USA 101, 221–226 (2004).
18. Jobsis, F. F. Noninvasive, infrared monitoring of cerebral
and myocardial oxygen sufficiency and circulatory
parameters. Science 198, 1264–1267 (1977).
19. Zhao, H. et al. Emission spectra of bioluminescent
reporters and interaction with mammalian tissue
determine the sensitivity of detection in vivo.
J. Biomed. Opt. 10, 41210 (2005).
20. Lipshutz, G. S. et al. In utero delivery of adeno-
associated viral vectors: intraperitoneal gene
transfer produces long-term expression. Mol. Ther.
3, 284–292 (2001).
21. Rehemtulla, A. et al. Rapid and quantitative
assessment of cancer treatment response using in vivo
bioluminescence imaging. Neoplasia 2, 491–495
22. Zhao, H. et al. Characterization of coelenterazine
analogs for measurements of Renilla luciferase activity
in live cells and living animals. Mol. Imaging 3, 43–54
23. Bhaumik, S. & Gambhir, S. S. Optical imaging of
Renilla luciferase reporter gene expression in living
mice. Proc. Natl Acad. Sci. USA 99, 377–382
24. Edinger, M. et al. Revealing lymphoma growth and the
efficacy of immune cell therapies using in vivo
bioluminescence. Blood 101, 640–648 (2003).
25. Beilhack, A. et al. In vivo analyses of early events in
acute graft-versus-host disease reveal sequential
infiltration of T-cell subsets. Blood 106, 1113–1122
26. Wang, X. et al. Dynamic tracking of human
hematopoietic stem cell engraftment using in vivo
bioluminescence imaging. Blood 102, 3478–3482
27. Edinger, M. et al. CD4+CD25+ regulatory T cells
preserve graft-versus-tumor activity while inhibiting
graft-versus-host disease after bone marrow
transplantation. Nature Med. 9, 1144–1150
28. Panoskaltsis-Mortari, A. et al. In vivo imaging of graft-
versus-host-disease in mice. Blood 103, 3590–3598
29. Murakami, T. et al. Immune evasion by murine
melanoma mediated through CC chemokine
receptor-10. J. Exp. Med. 198, 1337–1347
30. Murai, M. et al. Peyer’s patch is the essential site in
initiating murine acute and lethal graft-versus-host
reaction. Nature Immunol. 4, 154–160 (2003).
31. Welniak, L. A. et al. Peyer patches are not required for
acute graft-versus-host disease after myeloablative
conditioning and murine allogeneic bone marrow
transplantation. Blood 107, 410–412 (2005).
32. Mora, J. R. et al. Selective imprinting of gut-homing
T cells by Peyer’s patch dendritic cells. Nature
424, 88–93 (2003).
33. Anderson, B. E. et al. Memory CD4+ T cells do
not induce graft-versus-host disease. J. Clin. Invest.
112, 101–108 (2003).
34. Foster, A. E. et al. Human CD62L– memory T cells
are less responsive to alloantigen stimulation
than CD62L+ naive T cells: potential for adoptive
immunotherapy and allodepletion. Blood 104,
35. Dutt, S. et al. L-selectin and β7 integrin on donor
CD4 T cells are required for the early migration to
host mesenteric lymph nodes and acute colitis of
graft versus host disease. Blood 106, 4009–4015
36. Rooney, C. M. et al. Infusion of cytotoxic T cells
for the prevention and treatment of Epstein-Barr virus-
induced lymphoma in allogeneic transplant recipients.
Blood 92, 1549–1555 (1998).
37. Baker, J., Verneris, M. R., Ito, M., Shizuru, J. A. &
Negrin, R. S. Expansion of cytolytic CD8+ natural
killer T cells with limited capacity for graft-versus-host
disease induction due to interferon γ production.
Blood 97, 2923–2931 (2001).
38. Asai, O. et al. Suppression of graft-versus-host
disease and amplification of graft-versus-tumor effects
by activated natural killer cells after allogeneic bone
marrow transplantation. J. Clin. Invest. 101,
39. Ruggeri, L. et al. Effectiveness of donor natural
killer cell alloreactivity in mismatched hematopoietic
transplants. Science 295, 2097–2100 (2002).
40. Miller, J. S. et al. Successful adoptive transfer and
in vivo expansion of human haploidentical NK cells
in patients with cancer. Blood 105, 3051–3057
41. Bendelac, A. et al. CD1 recognition by mouse NK1+
T lymphocytes. Science 268, 863–865 (1995).
42. Lan, F., Zeng, D., Higuchi, M., Higgins, J. P. &
Strober, S. Host conditioning with total lymphoid
irradiation and antithymocyte globulin prevents
graft-versus-host disease: the role of CD1-reactive
natural killer T cells. Biol. Blood Marrow Transplant.
9, 355–363 (2003).
43. Sakaguchi, S., Sakaguchi, N., Asano, M., Itoh, M. &
Toda, M. Immunologic self-tolerance maintained by
activated T cells expressing IL-2 receptor α-chains
(CD25). Breakdown of a single mechanism of self-
tolerance causes various autoimmune diseases.
J. Immunol. 155, 1151–1164 (1995).
44. Hoffmann, R., Ermann, J., Edinger, M., Fathman, C. G.
& Strober, S. Donor type CD4+CD25+ regulatory
T cells suppress lethal acute graft-versus-host
disease after allogeneic bone marrow transplantation.
J. Exp. Med. 196, 389–399 (2002).
45. Taylor, P. A., Lees, C. J. & Blazar, B. R. The infusion of
ex vivo activated and expanded CD4+CD25+ immune
regulatory cells inhibits graft-versus-host disease
lethality. Blood 99, 3493–3499 (2002).
46. Cohen, J. L., Trenado, A., Vasey, D., Klatzmann, D. &
Salomon, B. L. CD4+CD25+ immunoregulatory
T cells: new therapeutics for graft-versus-host disease.
J. Exp. Med. 196, 401–406 (2002).
47. Trenado, A. et al. Recipient-type specific CD4+CD25+
regulatory T cells favor immune reconstitution and
control graft-versus-host disease while maintaining
graft-versus-leukemia. J. Clin. Invest. 112,
48. Ermann, J. et al. Only the CD62L+ subpopulation
of CD4+CD25+ regulatory T cells protects from lethal
acute GVHD. Blood 105, 2220–2226 (2005).
49. Herschman, H. R. Molecular imaging: looking at
problems, seeing solutions. Science 302, 605–608
50. Ponomarev, V. et al. Imaging TCR-dependent NFAT-
mediated T-cell activation with positron emission
tomography in vivo. Neoplasia 3, 480–488 (2001).
51. Dubey, P. et al. Quantitative imaging of the T cell
antitumor response by positron-emission tomography.
Proc. Natl Acad. Sci. USA 100, 1232–1237 (2003).
52. van Montfrans, C. et al. In vivo evaluation of 111In-
labeled T-lymphocyte homing in experimental colitis.
J. Nucl. Med. 45, 1759–1765 (2004).
53. Bulte, J. W. & Kraitchman, D. L. Iron oxide MR
contrast agents for molecular and cellular imaging.
NMR Biomed. 17, 484–499 (2004).
54. McCaffrey, A., Kay, M. A. & Contag, C. H. Advancing
molecular therapies through in vivo bioluminescent
imaging. Mol. Imaging 2, 75–86 (2003).
55. Contag, C. H. & Ross, B. D. It’s not just about
anatomy: in vivo bioluminescence imaging as an
eyepiece into biology. J. Magn. Reson. Imaging
16, 378–387 (2002).
56. Ntziachristos, V., Ripoll, J., Wang, L. V. & Weissleder, R.
Looking and listening to light: the evolution of
whole-body photonic imaging. Nature Biotechnol.
23, 313–320 (2005).
57. Hildebrandt, I. J. & Gambhir, S. S. Molecular
imaging applications for immunology. Clin. Immunol.
111, 210–224 (2004).
58. Contag, C. H. & Bachmann, M. H. Advances in in vivo
bioluminescence imaging of gene expression. Annu.
Rev. Biomed. Eng. 4, 235–260 (2002).
59. Rice, B. W., Cable, M. D. & Nelson, M. B. In vivo
imaging of light-emitting probes. J. Biomed. Opt.
6, 432–440 (2001).
60. Troy, T., Jekic-McMullen, D., Sambucetti, L. & Rice, B.
Quantitative comparison of the sensitivity of detection
of fluorescent and bioluminescent reporters in animal
models. Mol. Imaging 3, 9–23 (2004).
Competing interests statement
The authors declare competing financial interests: see web
version for details.
The following terms in this article are linked online to:
CD4 | CD8 | CD25 | CD44 | CD62L | FOXP3 | interferon-γ | IL-4
Christopher H. Contag’s laboratory:
Robert S. Negrin’s homepage:
Access to this links box is available online.
490 | JUNE 2006 | VOLUME 6
© 2006 Nature Publishing Group