CD47 update: a
multifaceted actor in the
of potential therapeutic
E Sick1,2, A Jeanne1, C Schneider1, S Dedieu1, K Takeda2and L Martiny1
1CNRS, FRE 3481, Matrice Extracellulaire et Dynamique Cellulaire, Université de Reims
Champagne-Ardenne, Faculté des Sciences, Reims, France, and2CNRS, UMR 7213, Laboratoire
de Biophotonique et Pharmacologie, Université de Strasbourg, Faculté de Pharmacie Illkirch,
Dr Emilie Sick, CNRS, FRE 3481,
URCA, Moulin de la Housse, BP
1039, 51687 Reims Cedex 2,
21 July 2011
25 June 2012
28 June 2012
CD47 is a ubiquitous 50 kDa five-spanning membrane receptor that belongs to the immunoglobulin superfamily. This
receptor, also known as integrin-associated protein, mediates cell-to-cell communication by ligation to transmembrane
signal-regulatory proteins SIRPa and SIRPg and interacts with integrins. CD47 is also implicated in cell-extracellular matrix
interactions via ligation with thrombospondins. Furthermore, CD47 is involved in many and diverse cellular processes,
including apoptosis, proliferation, adhesion and migration. It also plays a key role in many immune and cardiovascular
responses. Thus, this multifaceted receptor might be a central actor in the tumour microenvironment. Solid tumours are
composed of not only cancer cells that actively proliferate but also other cell types including immune cells and fibroblasts that
make up the tumour microenvironment. Tumour cell proliferation is strongly sustained by continuous sprouting of new
vessels, which also represents a gate for metastasis. Moreover, infiltration of inflammatory cells is observed in most neoplasms.
Much evidence has accumulated indicating that infiltrating leukocytes promote cancer progression. Given its ubiquitous
expression on all the different cell types that compose the tumour microenvironment, targeting CD47 could represent an
original therapeutic strategy in the field of oncology. We present a current overview of the biological effects associated with
CD47 on cancer cells and stromal cells.
AML, acute myeloid leukaemia; CBD, cell-binding domain; DC, dendritic cell; Drp-1, dynamin-related protein-1; eNOS,
endothelial nitric oxide synthase; FAK, focal adhesion kinase; HIF, hypoxia-inducible factor; HSC, haematopoietic stem
cell; HSPGs, heparan sulfate proteoglycans; IAP, integrin-associated protein; ITIM, immunoreceptor tyrosine-based
inhibitory motif; LSC, leukaemia stem cell; mAb, monoclonal antibody; PCD, programmed cell death; PS,
phosphatidylserine; RBC, red blood cell; ROS, reactive oxygen species; SCID, severe combined immunodeficiency; SIRP,
signal-regulatory protein; Treg, T-regulatory cell; TSP, thrombospondin; VCAM-1, vascular cell adhesion molecule-1
It is now widely accepted that solid tumours are not solely
composed of proliferating aneuploïd cells particularly resist-
ant to cell death. Thus, solid tumours should be considered as
heterogeneous structures containing multiple distinct cell
types, including normal cells recruited into tissues surround-
ing tumoursthat activelysustaintumourigenesis by
British Journal of
British Journal of Pharmacology (2012) 167 1415–14301415
© 2012 The Authors
British Journal of Pharmacology © 2012 The British Pharmacological Society
cell-to-cell interactions and secretion of paracrine growth
factors. Together with the vascular network provided by ang-
iogenesis that ensures nutrition and oxygenation of tumours,
the ensemble of these actors defines the tumour microenvi-
ronment. In addition, cancer-associated fibroblasts have been
shown to enhance cancer cell proliferation, angiogenesis,
invasion and metastasis, especially in the context of chronic
inflammation (Tlsty and Coussens, 2006). Finally, many dif-
ferent types of immune cells are also found infiltrated in
tumour masses at early stages. Sustained inflammation is one
of the hallmarks of tumour promotion and correlates with
poor prognosis in many different types of cancer. These infil-
trating cells play key roles in both innate and acquired immu-
nity and include dendritic cells (DCs), macrophages, mast
cells, neutrophils, and T- and B-cells (Hanahan and Weinberg,
CD47 is an ubiquitously expressed membrane receptor
that has been implicated in many normal and pathophysi-
ological processes including apoptosis, proliferation, cell
adhesion, cardiovascular effects, inflammation and immu-
nity. Given this diversity of actions, CD47 appears as a
multifaceted receptor that might represent a key target
in the tumour microenvironment for development of
innovative therapeutic strategies against cancer. In this
review, we update the effects associated with activation or
inhibition of CD47 in both cancer cells and stromal cells,
paying particular attention to vascular cells and immune
CD47 structure and partners
CD47 is an ubiquitous 50 kDa membrane receptor that
belongs to the immunoglobulin superfamily (Frazier et al.,
2010). This receptor was identified through its association
with integrin avb3and was therefore initially called integrin-
associated protein (IAP; Brown et al., 1990). At that time, an
ovarian carcinoma antigen called OA3 was cloned (Campbell
et al., 1992) with a proposed structure having five putative
transmembrane domains, an extracellular N-terminal IgV
domain with five putative N-glycosylation sites and a
C-terminal intracellular tail characterized by alternative splic-
ing (Figure 1). A closely similar structure was concomitantly
proposed for IAP (Lindberg et al., 1993). The ovarian tumour
marker OA3 was subsequently shown to be the same protein
as IAP (Mawby et al., 1994). The OA3 marker was also found
not to be restricted to ovarian carcinoma, rather exhibiting a
larger distribution. CD47 was an orphan receptor until it was
reported that this protein was the receptor for the C-terminal
cell-binding domain (CBD) of thrombospondins (TSPs) (Gao
and Frazier, 1994). Later, structural analysis revealed that a
long-range disulfide bond between the extracellular and the
fourth membrane-spanning domains was required for CD47
ligand binding, signalling and localization into lipid rafts
(Rebres et al., 2001).
Four alternatively spliced forms of CD47 were identified
in both murine and human cells, with variants ranging in
length from 3 to 36 amino acids in the cytoplasmic tail and
Schematic representation of CD47 and its endogenous ligands TSP-1 and SIRP. CD47 is composed of five transmembrane domains, an extracellular
N-terminal IgV domain and an intracytoplasmic C-terminal alternatively spliced domain. TSP-1 comprises several different domains, including the
heparin-binding domain (HBD), the von Willebrand C domain (vWCD), three type 1 properdin repeats, three type 2 EGF-like repeats, seven type
3 calcium-binding repeats and the C-terminal CBD. Several domains of TSP-1 interact with different extracellular matrix components or membrane
receptors (arrows). Key amino acid sequences responsible for TSP-1 ligation to integrins (RGD) and CD47 (RFYVVMWK) are indicated in italics.
SIRPa and g are single transmembrane spanning proteins with 3 Ig-like motifs, including an extracellular IgV domain that interacts with the IgV
domain of CD47.
E Sick et al.
1416 British Journal of Pharmacology (2012) 167 1415–1430
having a very high interspecies conservation with respect to
both peptide sequence of the alternatively spliced regions
and gene structure (Reinhold et al., 1995). The second-
shortest form 2 is the most abundantly expressed isoform
while the second most abundant isoform, form 4 (the
longest), is predominantly expressed in the brain and in the
peripheral nervous system; only keratinocytes expressed sig-
nificant amounts of form 1 (Reinhold et al., 1995). Little is
known about the functional significance of this alternative
splicing. Indeed, only a single study has focused on the
expression of these alternatively spliced isoforms: during
memory consolidation in rats, isoforms 3 and 4 are thought
to be closely associated with memory consolidation, while
isoform 2 appears to be the principal signal transducer in
astrocytes (Lee et al., 2000). Using a yeast model, post-
translational modifications such as glycosylation were shown
to be necessaryfor membrane
(Parthasarathy et al., 2006). However, it is still unknown to
what extent glycosylation influences CD47 ligand-binding
properties and subsequent signalling. Recently, a high
molecular weight isoform of CD47 was identified in Jurkat
cells (Kaur et al., 2011). In these cells, very little CD47 at its
expected molecular weight (55 kDa) was found. Rather, most
immunoreactivity in Western blots migrated with an appar-
ent molecular weight >250 kDa, consistent with CD47 also
being expressed as a proteoglycan, because this isoform was
sensitive to chondroitinase and heparitinase (Kaur et al.,
2011). The glycosaminoglycan residues at Ser64are involved
in CD47-mediated CD69 induction in T-cells. This high
molecular weight isoform is also apparently widely expressed
on vascular cells (Kaur et al., 2011).
The first identified endogenous ligand for CD47 was TSP-1
(Gao and Frazier, 1994). Using synthetic overlapping peptides
from the C-terminal CBD of TSP-1, a consensus sequence
containing the motif VVM was proposed to be essential for
binding to CD47. This resulted in the development of several
CD47 peptide agonists (Gao et al., 1996a), among which pep-
tides 7N3 (1102FIRVVMYEGKK1112), 4N1 (1016RFYVVMWK1024)
and the extended 4N1K (K-1016RFYVVMWK1024-K; i.e. 4N1
flanked by two Lys residues) are still the most widely used
experimentally. Nevertheless, it should be noted that, in at
least two studies, the CD47-binding peptides 4N1 and 4N1K
were reported to induce aggregation of platelets and to
enhance Jurkat cell adhesion in CD47-deficient cells (Tulasne
et al., 2001; Barazi et al., 2002), raising doubt as to the spe-
cificity of these peptides. Thus, results obtained with these
peptides alone [i.e. without confirmation using native TSP-1,
function-blocking anti-CD47 monoclonal antibodies (mAbs)
or CD47-null cells] should be interpreted cautiously because
it cannot be ruled out that these peptides might also bind to
another yet unidentified receptor. It should also be noted that
such controls have in general not been done in published
work using 4N1 or 4N1K.
TSP-1 belongs to a family of multidomain calcium-
binding glycoproteins that associate as homotrimers (TSP-1
and -2) or homopentamers (TSPs 3–5) in mammals. Most
adult tissues co-express one or more TSPs and major altera-
tions occurduring pathophysiological
example, highly expressed levels of TSP-1 and -2 are com-
monly found in stromal fibroblasts and endothelial cells
within tumours (Streit et al., 1999; Hawighorst et al., 2001).
In addition to the C-terminal CD47-binding domain, TSP-1
has several other domains that bind to different cell mem-
brane receptors or extracellular matrix, thereby mediating
cell-cell and cell-extracellular matrix interactions (Figure 1).
The N-terminal domain contains a heparin-binding domain
that binds to heparan sulfate proteoglycans (HSPGs) and
integrin a4b1. TSP-1 also contains three types of repeats: type
1 properdin repeats that bind to CD36, collagen type V,
fibronectin and HSPGs; type 2 EGF-like repeats; and type 3
calcium-binding repeats that contain the arginine-glycine-
aspartate (RGD)-sequence involved in ligation of b3 integrin
subunits (Figure 1; Adams, 2001). To date, the X-ray structure
of the CD47/TSP-1 complex has not been resolved. However,
our molecular modelling study has contributed to under-
standing how the TSP-1 CBD interacts with CD47 (Floquet
et al., 2008). Observation of X-ray structure of the CBD of
TSP-1 shows that the 4N1 sequence is normally hidden
within a hydrophobic pocket, preventing any interaction. By
normal mode analysis and energy minimization, it was pro-
posed that the opening of the 4N1 hydrophobic cleft is
driven by large amplitude motions when in close proximity
to CD47 and the phospholipid bilayer. The main biological
effects induced by ligation of TSP-1 CBD to CD47 are sum-
marized in Table 1. Binding of TSP-1 to CD47 influences
several fundamental cellular functions including cell migra-
tion and adhesion, cell proliferation or apoptosis, and takes
part in the regulation of angiogenesis and inflammation.
The second identified endogenous ligand of CD47 was
SIRPa (Vernon-Wilson et al., 2000). This protein was alterna-
tively called BIT (Sano et al., 1997), p84 (Jiang et al., 1999),
SHPS-1 (Babic et al., 2000) or CD172a. As shown in Figure 1,
N-terminal and four tyrosine phosphorylation sites in the
tyrosine-based inhibitory motifs (ITIMs; Brooke et al., 1998).
Additionally, two other signal-regulatory protein (SIRP)
family members, SIRPb (also known as CD172b) and SIRPg
(also known as SIRPb2 and CD172g), have been characterized
in man (Kharitonenkov et al., 1997; Ichigotani et al., 2000).
The N-terminal IgV domain of SIRPa binds to the IgV domain
of CD47 (Vernon-Wilson et al., 2000) and the structure of the
CD47/SIRPa complex has been recently resolved by X-ray
crystallography (protein database code 2uv3; Hatherley et al.,
2008). In spite of the highly conserved structure for SIRP
family members, only SIRPa and SIRPg bind to CD47. Failure
of SIRPb to bind CD47 has been attributed at least in part to
the replacement of Val27 by Met in the IgV domain of SIRPa
(Hatherley et al., 2008). The structure of their cytoplasmic
regions is quite different. SIRPg has a very short cytoplasmic
tail like SIRPb and lacks a charged amino acid residue in its
transmembrane region preventing the association with the
adaptor protein DAP12. This molecular adaptor contains an
immunoreceptor tyrosine-based activation motif, thought to
be responsible for SIRPb-mediated intracellular signalling.
The ligand for SIRPb is still unknown. CD47 ligation of SIRPa
or SIRPg mediates bidirectional signalling responsible for dif-
ferent cell-to-cell effects including inhibition of phagocytosis,
stimulation of cell-cell fusion or T-cell activation (Table 1).
CD47 was initially characterized as associated with
avb3 (Brown et al., 1990) integrins. Since then, numerous
bears two immunoreceptor
CD47 in the tumour microenvironment
British Journal of Pharmacology (2012) 167 1415–14301417
Partners of CD47 and their associated actions
CD47 partners Main biological and cellular eventsReferences
Apoptosis, cell proliferation, cell survival, cell adhesion,
inhibition of angiogenesis, pro- and anti-inflammatory
effects, platelet activation and aggregation
Dorahy et al. (1997); Manna and Frazier (2004);
Rath et al. (2006a); Lamy et al. (2007); Isenberg
et al. (2009); Xing et al. (2009); Sick et al. (2011)
Serpin A1 (C-terminal domain)
Cell proliferationCongote and Temmel (2004)
Inhibition of phagocytosis, stimulation of cell-cell fusion,
T-cell activation, neutrophil transepithelial migration
Han et al. (2000); Oldenborg et al. (2000); Seiffert
et al. (2001); Liu et al. (2002)
Leukocyte transendothelial migration, T-cell proliferationPiccio et al. (2005); Stefanidakis et al. (2008)
Direct partners (*demonstrated experimentally by immunoprecipitation or FRET in the case of VEGFR-2)
a2* COX-2 expression and intestinal epithelial cell migration
a2b1* Migration and proliferation of smooth muscle cells
aIIbb3* Platelet activation
a4b1* Adhesion of sickle reticulocytes, B-cell migration
Glycosaminoglycan synthesis by chondrocytes
Fibrillar beta amyloid-mediated microglia activation and
Neutrophil transepithelial migration
avb3* C32 human melanoma cells spreading,
pro-inflammatory cytokine synthesis in human
monocytes, Coxiella burnetii phagocytosis,
promyelocytic leukaemia cell death
Broom et al. (2009)
Chung et al. (1997); Wang and Frazier (1998)
Chung et al. (1997)
Yoshida et al. (2000); Brittain et al. (2004)
Holledge et al. (2008)
Bamberger et al. (2003); Koenigsknecht and
Hofman et al. (2000)
Gao et al. (1996a); Capo et al. (1999); Hermann
et al. (1999); Saumet et al. (2005)
VEGFR-2* Inhibition of VEGFR-2 downstream signallingKaur et al. (2010)
CD36 Amyloid b-induced inhibition of NO signalling, inhibition
of NO-stimulated vascular cell responses and cGMP
Isenberg et al. (2006); Miller et al. (2010a)
Fas/CD95Stimulation of Fas-mediated apoptosis Manna et al. (2005)
Gi proteins* SykPlatelet activation (Lyn and FAK phosphorylation) Chung et al. (1997)
AC/PKAPlatelet aggregation (decrease in cAMP), smooth muscle
cell migration, T-cell death
Frazier et al. (1999); Wang et al. (1999); Manna and
ERKJurkat T-lymphoma cell adhesion (phosphorylation);
inhibition of smooth muscle cell migration
Wang et al. (1999); Wilson et al. (1999)
PI3K C32 melanoma cell spreading, T-cell migration,
astrocytoma cell proliferation
Gao et al. (1996b); Li et al. (2005); Sick et al.
PLIC-1*Jurkat T-cells spreading, regulation of heterotrimeric
Wu et al. (1999); N’Diaye and Brown (2003)
BNIP3*T-cell apoptosis Lamy et al. (2003)
RacNeurite and filopodium formation, neuronal
Miyashita et al. (2004); Murata et al. (2006)
Cdc42 B-cell migration, neurite and filopodium formation,
Miyashita et al. (2004); Yoshida et al. (2000);
Murata et al. (2006
Src and MEK kinasesEpithelial cell spreading and migrationShinohara et al. (2006)
Drp1Caspase-independent cell death of normal and
Bras et al. (2007)
Protein 4.2Rh complex integrity on red blood cells Bruce et al. (2002)
GCInhibition of NO signalling (decrease in cGMP)Isenberg et al. (2006)
E Sick et al.
1418British Journal of Pharmacology (2012) 167 1415–1430
interactions between CD47 and many different integrin
subunits have been reported (Table 1). For example, in asso-
ciation with a2b1 integrins, CD47 mediates migration and
proliferation of smooth muscle cells (Wang and Frazier,
1998). In cooperation with avb3, CD47 mediates diverse func-
tions such as cell spreading, cytokine synthesis, phagocytosis
and cell death (Gao et al., 1996b; Capo et al., 1999; Hermann
et al., 1999; Saumet et al., 2005). Interaction between avb3
integrin and subsequent intracellular signalling was depend-
ent on the IgV domain of CD47 but did not require the
transmembrane domain (Lindberg et al., 1996). However,
CD47-mediated cellular effects may occur independently of
integrins, especially in red blood cells (RBCs). This is why the
designation CD47 is now considered more appropriate than
the initial name IAP. CD47 has also been shown to cooperate
with the Fas pathway in mediating apoptosis (Manna et al.,
2005). Moreover, recent advances describing the role of CD47
in the field of cardiovascular diseases (Table 1) have under-
lined strong cooperation of CD47 with CD36 (Isenberg et al.,
2006) and VEGFR-2 (Kaur et al., 2010).
Many of the cellular effects of CD47 are mediated by Gi
proteins (Table 1). For example, apoptosis induced by the
peptide agonist 4N1K involves inhibition of AC and a subse-
quent decrease in intracellular cAMP levels (Manna and
Frazier, 2004). CD47-associated platelet activation requires Gi
proteins with activation of Syk kinases that induce phospho-
rylation of Lyn and focal adhesion kinase (FAK) (Chung et al.,
1997). Other kinases identified in the signalling pathway
downstream of Gi proteins include PI3K and MAPK. Taking
into account such G-protein-dependent signalling associated
with CD47 activation, Brown and Frazier (2001) proposed
that CD47 with five transmembrane domains associated to
an integrin dimer might mimic classical heptahelical GPCRs.
In agreement, CD47-dependent signalling involves not only
Gi proteins, but also direct activation of small G-proteins like
Rac and Cdc42, especially in neurones where CD47 plays an
important role in neurite formation and more generally in
neuronal development (Table 1). Furthermore, CD47 also
associates with protein linking integrin-associated protein to
cytoskeleton-1 protein (PLIC-1) known to regulate G-protein
signalling. Several studies have also identified cytoplasmic
Ca2+as an important mediator of CD47 signalling. TSP-1 was
reported to induce an increase in intracellular Ca2+in fibrob-
lasts via its RGD and C-terminal domain (Tsao and Mousa,
1995). CD47 has been shown to mediate fibronectin-induced
intracellular Ca2+rises in HUVECs (Schwartz et al., 1993).
CD47 cross-linking by mAb 1/1A4 resulted in a strong
increase in intracellular Ca2+in Jurkat cells (Waclavicek et al.,
1997). Peptide 4N1 also induced an increase in intracellular
Ca2+responsible for mast cell exocytosis (Sick et al., 2009). In
Jurkat T-cells, a recombinant TSP-1 C-terminal fragment
(E3CaG1) inhibited sGC via increased intracellular Ca2+(Ram-
anathan et al., 2011). In Jurkat cells, increases in Ca2+were
CD47 dependent but did not involve Gi protein activation,
contrary to what we found in mast cells (Sick et al., 2009). In
contrast, in HUVECs, TSP-1 via CD47 is thought to inhibit
acetylcholine-induced increases in intracellular Ca2+(Bauer
et al., 2010). Because decreased Ca2+concentrations inhibit
endothelial nitric oxide synthase (eNOS) Ca2+-dependent
activation, TSP-1 contributes to regulate blood pressure by
limiting eNOS activation and endothelial-dependent vasore-
laxation (Bauer et al., 2010). Several previous studies have
also confirmed that CD47 acts as a key regulator of the
NO/cGMP pathwayin vascular
endothelial cells and platelets (Isenberg et al., 2006; 2008a).
Taken together, these data indicate that CD47 signalling
may differ according to cell type. Given the ubiquitous
expression ofCD47, its intracellular
associated partners are likely to be crucial in defining a given
specific cellular response both in physiological and also
CD47 and tumour cells
Tumour cell apoptosis
Tumour growth is considered, at least in part, to result from
dysfunctions in the balance between apoptosis and cell pro-
liferation. CD47 actively takes part in apoptosis and therefore
plays a key role in maintaining tissue homeostasis. Exposure
to soluble anti-CD47 mAbs resulted in rapid (within 3 h)
apoptosis of Jurkat E6 cells and activated T-cells (Pettersen
et al., 1999). However, this first report underlines the com-
plexity and ambiguity concerning the functional effects of
anti-CD47 mAbs. Indeed, soluble anti-CD47 mAbs Ad22 and
1F7 induced Jurkat cell apoptosis whereas mAbs B6H12 and
2D3 were without effect. Ad22 was reported to react with the
IgV domain of CD47, in close proximity to epitopes defined
by B6H12 and 1F7, whereas 2D3 reacts with a distant region.
Thus, CD47-induced apoptosis was thought to require liga-
tion of distinct epitopes on the IgV domain. Interestingly,
immobilized Ad22 and 2D3 had no influence on Jurkat cell
proliferation, suggesting that the mode of presentation of
anti-CD47 mAbs is critical with respect to functional
responses. Furthermore, the effects of anti-CD47 mAbs also
depend on the cell line tested and on the activation state of
the cell. For example, Ad22 induced cell death of Jurkat cells
and CD3-activated T-cells but not resting T-cells. In these
cells, CD47-induced apoptosis was thought to be independ-
ent of Fas and the tumour necrosis factor receptor-dependent
et al., 1999).
mediated apoptosis induced by TSP-1 or immobilized (B6H12
and BRIC126) but not soluble anti-CD47 mAbs was reported
in many leukaemic cells, including B-chronic lymphocytic
leukaemia (B-CLL) cells (Mateo et al., 1999), promyelocytic
leukaemia NB4 cells (Saumet et al., 2005) and Jurkat T-cells
(Roué et al., 2003). Activation of CD47 apparently leads to a
particular type of cell death called type-III programmed cell
death (PCD) in normal as well as leukaemia cell lines. Type-III
PCD is caspase independent and is characterized by cytoplas-
mic effects such as cell shrinkage, phosphatidylserine (PS)
exposure and mitochondrial matrix swelling, which result in
disruption of mitochondrial transmembrane potential (Dym)
and subsequent reactive oxygen species (ROS) production, as
described for B-cells cultured on immobilized B6H12 mAbs
(Mateo et al., 2002). Neither cytochrome c and apoptosis-
inducing factor nor release of other mitochondrial inter-
membrane proteins were required for CD47-dependent PCD
(Roué et al., 2003; Saumet et al., 2005). Mitochondrial trans-
membrane potential loss and PS externalization were pre-
vented by cytochalasin D, suggesting that actin cytoskeleton
CD47 in the tumour microenvironment
British Journal of Pharmacology (2012) 167 1415–1430 1419
rearrangement was necessary in CD47-dependent B-CLL cell
death (Mateo et al., 2002; Barbier et al., 2009). Additionally,
immobilized anti-CD47 mAbs induced morphological and
functional changes in B-cell leukaemia cell lines that involve
Rac1 and Cdc42 GTPases (Yoshida et al., 2000). Furthermore,
CD47 ligation failed to induce characteristic type-III PCD
cytoplasmic events in patients who do not express Wiskott–
Aldrich syndrome protein (WASP), indicating that both
Cdc42 and WASP are key components of the signalling
pathway responsible for actin polymerization in CD47-
induced caspase-independent apoptosis (Mateo et al., 2002).
Nevertheless, the molecular basis of CD47-induced type-
III PCD is still incompletely understood. Recently, dynamin-
related protein-1 (Drp-1) was identified as a key molecular
relay in activating PCD (Bras et al., 2007). Chymotrypsin-like
protease activation after CD47 ligation by immobilized
B6H12 mAbs or immobilized TSP-1 was shown to induce
Drp-1 translocation from cytosol to mitochondria, thus
leading to ROS production and ATP loss independently of its
GTPase activity in primary B-cells (Barbier et al., 2009). Of
note, in addition to induction of type-III PCD, immobilized
anti-CD47 1F7 mAbs have been reported to promote death of
Jurkat T-cells and normal murine T-cells by classical caspase-
dependent apoptosis. Indeed, the extracellular IgV domain of
CD47 might interact with Fas receptors to induce apoptosis
without mobilization of CD47-dependent signalling or locali-
zation of CD47 in rafts (Manna et al., 2005).
Considering the key role of TSP-1/CD47 interactions in
leukaemia, the putative therapeutic potential of anti-CD47
mAbs was investigated. A bivalent single-chain Ab fragment
(scFv) of a murine mAb was first shown to promote apoptosis
of hCD47-transfected L1210 murine leukaemic cells without
causing haemagglutination of RBCs (Kikuchi et al., 2004).
Bivalent single-chain Ab fragments were also effective against
the human myeloma cell line KPMM2 implanted in severe
combined immunodeficiency (SCID) mice (Kikuchi et al.,
2005). Antitumour activity of CD47 mAbs was reported in
xenograft models of acute lymphoblastic leukaemia and
B-cell chronic lymphocytic leukaemia. Interestingly, no apop-
totic side effects on CD34+ haematopoietic progenitor/stem
or human endothelial cells were observed in these models
(Uno et al., 2007). Recently, a disulfide-linked dimer of a
single-chain Ab fragment against CD47 induced prolonged
survival of SCID mice engrafted with JOK-1 leukaemic cells
(Sagawa et al., 2011). In vitro results indicate that this effect
genes BNIP3 and RTP801.
The role of CD47 in cancer development is not restricted
to only leukaemia, but has also been shown for vascularized
solid tumours. Triggering of CD47 by 4N1K induces caspase-
independent apoptosis in several human breast cancer cell
lines. Indeed, CD47 activation stimulates cell death of both
non-invasive (MCF-7 and AU-565) as well as invasive (MDA-
MB-231) cell lines through Gi signalling, leading to subse-
quent decrease in cAMP levels and inhibition of PKA (Manna
and Frazier, 2004).
Interestingly, CD47 has been reported to be involved in
resistance of cancer cells to taxol (Lih et al., 2006). The resist-
ance of human prostate cancer cells to taxanes was correlated
with overexpression of a recently identified protein Txr1,
reported to dramatically down-regulate TSP-1 expression.
Stimulation of CD47 with either TSP-1 or 4N1K in resistant
cells increased taxol-induced apoptosis (Lih et al., 2006).
Whether this mechanism is directly or indirectly mediated by
CD47 is unknown. In addition, primary vascular cells cul-
tured from TSP-1-null mice and CD47-null mice were shown
to be resistant to high-dose radiation injury (Isenberg et al.,
2008b). CD47 suppression in mice resulted in increased radia-
tion sensitivity of implanted melanoma and squamous cell
lung tumours, while conferring radioresistance in normal soft
tissues, bone marrow and tumour-associated leukocytes
(Maxhimer et al., 2009a). Radiation-induced cell death was
similar in HUVECs when treated either by mAbs directed
against TSP-1 (clone A6.1) or CD47 (clone B6H12) or by
antisense CD47 morpholinos, consistent with blockade of the
TSP-1/CD47 interaction being responsible for the decrease in
radiation sensitivity. Radioprotection in normal tissues may
be attributable at least in part to restoration of NO-mediated
cytoprotective activity (Isenberg et al., 2008b; Maxhimer
et al., 2009a).
Tumour cell survival and proliferation
Besides its well-described role in inducing apoptosis, CD47
has been paradoxically shown to modulate cell survival
and proliferation. Indeed, we demonstrated that activation
of CD47 by 4N1 results in protection of thyroid cells
from apoptosis induced by C2-ceramides by stimulating the
cAMP/PKA pathway (Rath et al., 2006a). Moreover, 4N1-
mediated activation of CD47 inhibited camptothecin- and
doxorubicin-induced apoptosis of human follicular thyroid
carcinoma FTC-133 cells (Rath et al., 2006b).
Furthermore, we showed that CD47 stimulated by the
agonist 4N1 induces the proliferation of astrocytoma cells but
not normal astrocytes, with downstream signalling subse-
quent to CD47 activation involving Gbg dimer-dependent
activation of the PI3K/Akt pathway in astrocytoma cells (Sick
et al., 2011). As well, a peptidic fragment resulting from
serpin A1 proteolytic cleavage exhibits CD47-dependent
mitogenic activity in both Hep G2 liver cancer cells and
MCF-7 breast cancer cells (Congote and Temmel, 2004).
The implication of CD47 in cell migration was first demon-
strated for neutrophils, with blocking Abs against CD47
inhibiting transmigration of neutrophils and monocytes
through endothelium (Cooper et al., 1995; de Vries et al.,
2002). Increased levels of CD47 expression were correlated to
improved rates of neutrophil migration during inflammatory
responses (Liu et al., 2001). It was also found that CD47 plays
a major role in smooth muscle cell proliferation and chemo-
taxis towards soluble collagen, in part via association with
a2b1 integrins, subsequent Gi-mediated decrease in cAMP
levels and inhibition of ERK activity (Wang and Frazier, 1998;
Wang et al., 1999). Recent data suggest that CD47 stimulates
intestinal epithelial cell migration on collagen-I through
COX-2 expression (Broom et al., 2009). Moreover, increased
expression of CD47 in fibroblasts promotes intercellular
adhesion and decreases the spontaneous migration of Jurkat
lymphocytes through a fibroblast monolayer. This effect was
dependent on Rho family GTPases and Gi proteins (Rebres
et al., 2005). These data suggest that CD47 may restrict the
E Sick et al.
1420British Journal of Pharmacology (2012) 167 1415–1430
infiltration of cancer cells in surrounding tissues and devel-
opment of metastasis. Paradoxically, CD47 was reported to
induce epithelial cell spreading and migration. Indeed, CD47
co-localizes with E-cadherin at cell-cell adhesion sites and
forced expression of CD47 induces epithelial cell spreading
and leads to a partial disruption of cell-cell adhesion (Shino-
hara et al., 2006).
Surprisingly, very few data are available regarding the role
of CD47 in the regulation of cancer cell migration. CD47 was
shown to stimulate migration of several lymphoma cell lines,
but not myeloma cell lines, involving the Cdc42 GTPase and
a tight molecular cooperation with a4b1 integrins (Yoshida
et al., 2000). In addition, CD47 was demonstrated to decrease
Melan-a cell migration but not that of B16F10 melanoma
cells. Aberrant N-glycosylation of SIRPa was observed in
B16F10 melanoma cells and results in impaired binding to
CD47. Binding of CD47 to SIRPa in Melan-a cells is supposed
to down-regulate SIRPa expression levels on adjacent cells by
stimulation of proteosomal degradation. Increased proteo-
somal degradation of SIRPa resulted in inactivation of SHP-2
phosphatase, inhibition of the dephosphorylation of FAK,
subsequent disassembly of focal adhesions and thereby to
inhibition of Melan-a cell motility (Ogura et al., 2004). Resist-
ance to this molecular mechanism in B16 melanoma is
thought to contribute, at least in part, to the more aggressive
behaviour of these cells (Ogura et al., 2004). Blockade of
CD47 by neutralizing Abs reduced migration and chemotaxis
in response to collagen IV of melanoma, prostate cancer and
ovarian cancer-derived cells. These effects were shown to be
dependent on avb3 integrins and intracellular calcium
(Shahan et al., 2000). Finally, CD47 was reported to stimulate
osteoclastogenesis and its disruption may protect against
metastatic tumours in bone (Uluçkan et al., 2009). Taken
together, these data are consistent with activation of CD47
contributing to cell migration, and it therefore may be that
CD47 antagonism represents a promising therapeutic strat-
egy to limit infiltration at tumour sites and to decrease cancer
cell dissemination and formation of metastases.
Cell adhesion and spreading
CD47 is required for post-adhesive events after neutrophil
transmigration across intestinal epithelial cells (Parkos et al.,
1996). C32 melanoma cell spreading onto vitronectin was
stimulated by CD47 through cooperation with avb3and avb5
integrins and involved activation of Gi proteins, FAK and
PI3K (Gao et al., 1996b). This effect apparently depends on
cholesterol, which was thought to maintain the structural
and functional integrity of the CD47/avb3/G-protein complex
in lipid rafts (Green et al., 1999). However, adhesion may also
be stimulated by CD47 independently of cholesterol content
and G-protein signalling. Indeed, CD47 improves avb3avidity
in human ovarian carcinoma cells after cholesterol depletion
in the cell membrane and similarly, a4b1avidity for vascular
cell adhesion molecule-1 (VCAM-1) and TSP-1 in sickle
reticulocytes (Brittain et al., 2004; McDonald et al., 2004).
Thus, CD47 can also transmit inside-out signals contributing
to the regulation of integrin activation and clustering. Differ-
ent pools of integrins appear to be differentially regulated by
CD47 and this could be related to the localization of CD47 in
specific microdomains at the plasma membrane. CD47 can
also control cell-to-cell adhesion by interaction with SIRPg,
for example, in T-cell adhesion to antigen-presenting cells
(Piccio et al., 2005). Moreover, in CD47-deficient fibroblasts,
CD47 expression induces intercellular adhesion, resulting in
cell aggregation even in the absence of active integrins,
SIRPa1 or TSP-1 binding (Rebres et al., 2005). This effect is
proposed to involve CD47 homophilic interaction, but unlike
endothelial cell adhesion molecule-1 or cadherins), expres-
sion of CD47 is not sufficient to induce aggregation and may
require activation by serum (concentration above 4%) or by
4N1K. It is also interesting to note that anti-CD47 2D3 mAbs,
which inhibit CD47-dependent cell aggregation, enhanced
SIRPa1 binding. This effect was also mediated by Fab frag-
ments, and suggests that even monovalent binding causes
CD47 to undergo conformational or clustering changes that
are relevant for SIRPa binding (Rebres et al., 2005). Even if the
homotypic interaction of CD47 with itself was not further
explored, this study underlines two important features that
may explain some of the reported paradoxical results, namely
the influence of serum concentration on the aggregation state
of cells and differential SIRP binding in the presence of CD47
mAbs. CD47 also influences cell adhesion indirectly. Indeed,
TSP-1 induced the up-regulation of the adhesion molecules
intercellular adhesion molecule-1 and VCAM-1 in human
brain microvascular endothelial cells (Xing et al., 2009). In
view of these results, CD47 modulation by blocking strategies
may have detrimental effects resulting in increased cell
migration depending on the cell type considered. Such effects
have to be taken into account in developing anti-metastatic
strategies where CD47 is the target.
CD47 and stromal cells
Angiogenesis represents a key component of solid tumour
development, allowing both nutrition and oxygenation of
tumour cells, thus subsequently contributing to dissemina-
tion and metastasis. Strategies targeting angiogenesis have
emerged over the past decade based on a better understand-
ing of molecular mechanisms underlying angiogenesis. This
has led to the development of anti-angiogenic approaches for
the treatment of cancer, with notably several vascular
endothelial growth factor (VEGF)-neutralizing mAbs being
approved for clinical use in cancer (for review, see Chames
et al., 2009). However, even if effective in certain cancers,
significant problems have appeared. For example, many
patients with metastatic diseases are refractory or acquire
resistance towards VEGF inhibitors (Bergers and Hanahan,
2008). Tumour vessel abnormalities resulting in heterogene-
ous and tortuous vessel networks are also believed to decrease
the efficiency of anti-VEGF strategies (Carmeliet and Jain,
2011). To date, even if current anti-VEGF agents have proved
to be less than ideally efficacious, targeting angiogenesis still
represents a most promising therapeutic strategy.
The first evidence for CD47 involvement in the vascular
system was described in platelets where 4N1 was shown to
stimulate platelet activation and aggregation (Dorahy et al.,
1997). The signalling pathway responsible for platelet activa-
tion involves a Gi protein-mediated decrease in cAMP
CD47 in the tumour microenvironment
British Journal of Pharmacology (2012) 167 1415–14301421
production (Frazier et al., 1999). More recently, several groups
have focused on the putative role of CD47 in angiogenesis. In
vitro, TSP-1 and 4N1K were reported to induce cytotoxicity of
brain endothelial cells and inhibit endothelial cell migration
and tube formation (Xing et al., 2009). These effects were
confirmed in vivo. Using recombinant fragments, the anti-
angiogenic properties of TSP-1 were demonstrated to be
mediated by the interaction between its C-terminal domain
and CD47 (Isenberg et al., 2006), with this interaction being
responsible for the inhibition of NO-induced GC activation.
Therefore, CD47 can modulate blood flow and tissue per-
fusion by interfering with the NO-dependent balance of the
relative state of constriction and relaxation of arteries.
Although CD36 ligation by TSP-1 inhibits NO-stimulated
vascular cell responses and cGMP signalling, the anti-
angiogenic effects of TSP-1 at picomolar physiological con-
centrations are driven solely by CD47 ligation. Moreover,
angiogenic responses following CD36 ligation are also
dependent on CD47. The NO-driven delay in thrombin-
induced platelet activation is blocked by TSP-1 ligation to
CD47 and/or CD36 as shown using, respectively, TSP-1
C-terminal domain and type 1 repeat recombinant fragments
(Isenberg et al., 2008a).
Restoration of NO signalling consecutive to CD47 antago-
nism that results in a rapid restoration of NO-driven blood
flow would be essential for graft survival. Consistently, in a
murine skin graft model, graft survival was dramatically
increased by both morpholino suppression of CD47 expres-
sion and CD47 blockade by miap 301 mAbs (Isenberg et al.,
2008c). In a comparable murine ischaemic model of full
thickness flaps, perfusion, survival and angiogenesis were
optimized using combined treatment with nitrite and TSP-1/
CD47 blockade via anti-TSP-1 A6.1 mAbs or gene silencing
(Isenberg et al., 2009). In this study, a survival rate of almost
100% was attained, indicating that optimal results were
obtained by combined treatment. The efficiency of such an
approach is especially promising for the treatment of burns
and diverse pathologies requiring skin grafts. In support, an
increase in tissue survival following myocutaneous flap
surgery was observed when using CD47-blocking mAbs
(clone miap 301), antisense morpholino or CD47 silencing in
mice (Isenberg et al., 2007). Moreover, increased survival con-
secutive to CD47-blocking mAb treatment was described in
CD36-null mice, suggesting that this response is CD36 inde-
pendent. Flap survival was also increased by CD47-blocking
OX101 mAbs in a rat model (Maxhimer et al., 2009b). In
addition to flap reperfusion enhancement, CD47 blockade
contributed to a decrease in the inflammatory response in
flap tissues after ischaemia-reperfusion injury. Random cuta-
neous flap survival, blood vessel permeability and blood flow
were also drastically increased by antisense CD47 mor-
pholino and anti-TSP mAbs (A6.1) in a porcine model (Isen-
berg et al., 2008d). Considering the high level of homology
between porcine and human tissues, the potential therapeu-
tic targeting of CD47 is supported by these results. Finally,
TSP-1 and CD47 suppression or pretreatment with a CD47-
antagonist mAb miap 301 was shown to enhance tissue sur-
vival and perfusion after ischaemia-reperfusion injury in a
murine liver model, suggesting the inhibition of CD47 could
not only enhance skin graft but also organ survival after
transplantation surgery (Isenberg et al., 2008e).
CD47 ligation by TSP-1 was reported to inhibit the
NO-dependent and NO-independent activation of sGC in
platelets as well as in smooth muscle cells (Miller et al.,
2010b). Furthermore, CD47 activation by TSP-1 controls
intracellular cAMP levels in smooth muscle, directly via
Gi-dependent inhibition of adenylate cyclase and indirectly
through PDE3 inhibition in a cGMP-dependent manner (Yao
et al., 2011). In addition to the modulation of cyclic nucle-
otide levels, CD47 ligation by native TSP-1 or CBD recom-
binant domain disrupted the constitutive association of
VEGFR-2 and CD47 in endothelial cells. This in turn inhibits
VEGFR-2 phosphorylation and its downstream signalling,
without affecting VEGF binding (Kaur et al., 2010). Finally,
CD47 silencing was recently described to result in enhance-
ment of the pro-angiogenic properties of endothelial colony
forming cells by interfering with the SDF-1 chemokine
pathway (Smadja et al., 2011).
The role of NO is quite ambiguous in tumours. Indeed,
NO is reported to increase tumour perfusion via induction of
local vasodilatation. On the contrary, NO is also thought to
decrease tumour perfusion by preferential relaxation of the
normal vascular network outside the tumour, resulting in
blood redistribution away from the tumour (Isenberg et al.,
2008f). Even if the capacity of tumour vasculature to respond
to vasoactive agents is quite limited, recovery of basal per-
fusion after a vasoactive treatment (vasoconstrictor or vasodi-
latator) is modulated by TSP-1 through CD47 interaction.
Indeed, overexpression of truncated TSP-1 lacking parts of its
CBD domain failed to reproduce this activity (Isenberg et al.,
2008f). These data indicate that TSP-1/CD47 interaction may
regulate long-term vascular responses in the tumour context.
The concept that inflammation sustains cancer development
is not new. Indeed, in 1863, Rudolf Virchow first hypoth-
esized that the origin of cancer was at sites of chronic inflam-
mation (Balkwill and Mantovani, 2001). After a long eclipse,
this hypothesis has been more recently supported by clinical
and experimental studies demonstrating that infiltration of
leukocytes into the tumour microenvironment is a key event
in promoting cancer growth (for review, see Tlsty and Cous-
sens, 2006). In this context, cytokines and chemokines as
well as activation of transcription factors such as NF-kB, Stat3
or HIF play major roles in tumour progression. Moreover,
tumour cells are thought to have altered self-antigens that
impair recognition by the immune system, leading to tumour
escape from immune surveillance (for review, see Dunn et al.,
2004). Accordingly, inflammation and escape of tumoural
cells from surveillance by the immune system are fundamen-
tal steps in tumour promotion and, therefore, targeting
immune cells may represent a promising therapeutic strategy
Among tumour infiltrating leukocytes, macrophages rep-
resent the main actors of tumour promotion (for review, see
Mantovani et al., 2008). Several studies have underlined a
role for CD47 as a marker of ‘self’, initially in RBCs. Thus,
RBCs from CD47-/-mice were rapidly cleared by macrophages
when transfused into wild-type mice, whereas they circulated
normally in CD47-/-mice (Oldenborg et al., 2000). This
occurred independently of complement and antibodies and
was mediated by the interaction between SIRPa and CD47.
E Sick et al.
1422British Journal of Pharmacology (2012) 167 1415–1430
Such a molecular interaction is also responsible for maintain-
ing platelet homeostasis (Olsson et al., 2005). The ITIM
domains of SIRPa become phosphorylated upon ligation with
CD47 and recruit SHP phosphatases, resulting in deactivation
of myosin-II without affecting F-actin (Tsai and Discher,
2008). The absence of self-recognition at the cell surface thus
allows contractile engulfment. Several paired receptors
engaged at the immunological synapse trigger phagocytosis
but CD47/SIRPa is the only identified negative regulator.
Interestingly, the expression level of CD47 is likely to vary
under pathophysiological situations. Indeed, CD47 expres-
sion is up-regulated on normal haematopoietic stem cells
(HSCs) when they are recruited to the periphery upon expo-
sure to pro-inflammatory stimuli like granulocyte colony
stimulating factor or LPS (Jaiswal et al., 2009). Increased
CD47 at the cell surface is therefore thought to protect hae-
matopoietic cells from phagocytosis during their trafficking
from bone marrow to the periphery. On the contrary,
decreased expression of CD47 was observed in apoptotic neu-
trophils and fibroblasts and correlated with the degree of
apoptosis. Apoptotic cells lose their ability to activate SIRPa
and are engulfed, unlike viable cells expressing normal
amounts of CD47. Moreover, on apoptotic cells, CD47 is
redistributed on the plasma membrane to form patches. This
is thought to decrease the ability of CD47 to interact with
SIRPa. Additionally, CD47 becomes segregated away from
calreticulin and PSs, which are ‘eat me’ stimuli triggering
low-density lipoprotein receptor-related protein 1-dependent
efferocytosis (Gardai et al., 2005). Interestingly, CD47 expres-
sion levels were shown to increase in several leukaemic cell
lines (Jaiswal et al., 2009), as well as in bladder tumour-
initiating cells (Chan et al., 2009), non-Hodgkin lymphoma
(Chao et al., 2010a) and acute lymphoblastic leukaemia
(Chao et al., 2011). CD47 is thus considered as an adverse
prognostic factor. Minimally expressed on most normal cells,
calreticulin levels are increased at the cell surface of several
cancer cells and are commonly correlated with enhanced
CD47 expression. This is thought to protect tumour cells
2010a,b). Increased levels of CD47 expression observed in
leukaemic mice were confirmed for dysplastic human hae-
matopoietic cells as compared with normal umbilical cord
blood and peripheral blood stem cells. Indeed, granulocyte-
myeloproliferative disorders such as atypical chronic myeloid
leukaemia, proliferative phase chronic myelomonocytic leu-
kaemia and acute myeloid leukaemia (AML) (Jaiswal et al.,
2009). MOLM-13 cells having low CD47 levels derived from a
patient with AML failed to engraft when implanted in immu-
nocompromised mice. In contrast, implanted MOLM-13 cells
overexpressing CD47 lead to development of tumours, con-
sistent with CD47 overexpression enabling these cells to
escape from phagocytosis by macrophages. Moreover, macro-
phages were shown to exert a selective pressure for high-
expressing CD47 clones in vivo. Thus, leukaemic progenitors
are thought to hijack the protective effect of overexpressing
CD47 used by normal HSCs to escape from macrophage-
dependent killing observed during inflammatory states. Con-
sistently, a blocking antibody directed against CD47 was
reported to trigger phagocytosis of human bladder cancer
cells in vitro (Chan et al., 2009). Finally, anti-CD47-blocking
mAbs (clones BRIC126 and B6H12) were shown to promote
macrophage-mediated phagocytosis of acute lymphoblastic
leukaemia cells and leukaemia stem cells (LSCs) and to inhibit
tumour engraftment in vivo (Majeti et al., 2009; Chao et al.,
The interaction of CD47 with SIRPa contributes to fusion
of rat alveolar macrophages and the generation of multinu-
cleated cells (Han et al., 2000) that can differentiate into
osteoclasts in bone. Formation of such giant multinucleated
cells was strongly inhibited both in vivo and in vitro in the
absence of SIRPa/CD47 interaction (Lundberg et al., 2007).
The distance between two cells interacting via CD47 and
SIRPa is supposed to be around 14 nm, a distance similar to
that observed during the establishment of the immunological
synapse (Hatherley et al., 2009). However, this distance may
be reduced to 5–10 nm in macrophages by interaction of
CD47 with a short form of SIRPa that lacks the C1 domain,
thereby leading to fusion (Han et al., 2000). To date, even if
the mechanism is poorly documented, several receptors
including CD44 (Sterling et al., 1998), DC-STAMP (Yagi et al.,
2006), RANK/RANKL and tetraspanin CD9 (Ishii et al., 2006)
have been reported to take part in the macrophage fusion
process leading to formation of osteoclasts or giant cells. The
fusion of cancer cell with macrophages is believed to repre-
sent a putative mechanism for dissemination and develop-
ment of metastasis, as described for melanoma cells fused
with macrophages (Rachkovsky et al., 1998; Chakraborty
et al., 2000). Nevertheless, to date, this proposed mechanism
remains controversial because of the absence of proof of
concept in vivo. However, fusion between intestinal epithelial
cells and macrophages was very recently shown to occur in
vivo in a cancer context and was related to nuclear reprogram-
ming resulting in a transcriptomic identity similar to the
parental cells but also with a unique subset of transcripts
(Powell et al., 2011). Thus, cancer cell fusion with circulating
bone marrow-derived cells contributes to masking the cancer
phenotype, thereby permitting cancer cells to escape from
immune surveillance and to infiltrate in a distant niche. In
this context, even if poorly characterized at present, thera-
peutic inhibition of CD47/SIRPa interactions might provide
not only enhanced clearance and degradation of cancer cells
but also limited metastatic potential.
DCs represent another key class of cells in the tumour
presenting cells of the immune system and thus playing a
critical role in establishing innate and adaptive immune
responses. Tumour-bearing mice as well as patients with
breast cancer, prostate cancer and glioma present a decreased
number of DCs in lymph nodes spleen and skin (Rabinovich
et al., 2007). Indeed, cancer patients have significantly
decreased numbers of circulating and peripheral DCs whereas
immature DCs accumulate. Immature DCs have a reduced
capacity for capturing antigens and a reduced capacity to
stimulate naïve T-cells (Petersen et al., 2010). Interestingly,
stimulation of CD47 by 4N1K inhibits secretion of pro-
inflammatory cytokines like IL-12, TNFa, IL-6 and GM-CSF
by maturing DCs (Demeure et al., 2000). Furthermore, CD47
activation inhibits maturation (i.e. up-regulation of major
histocompatibility complex class II antigens, co-stimulatory
molecules, acquisition of a potent T-cell stimulatory activity)
of immature DCs. CD47 stimulation by 4N1K also directly
most potent antigen-
CD47 in the tumour microenvironment
British Journal of Pharmacology (2012) 167 1415–14301423
affects DC populations by inducing apoptosis-like cell death
(Johansson and Londei, 2004). Thus, CD47 may participate
in reducing the efficiency of immune responses against
tumour by inducing T-cell tolerance towards tumours.
Various and sometimes contradictory functional effects
have been reported when CD47 is activated on T-cells. It was
reported that CD47-deficient animals exhibit sustained
inflammation related to decreased local apoptosis of T-cells
(Lamy et al., 2007). SIRPg expressed on T-cells which interacts
with CD47 on endothelial cells was critical for T-cell transen-
dothelial migration and in inducing apoptosis of Jurkat cells,
similar to what has been described for SIRPa ligation (Brooke
et al., 2004; Stefanidakis et al., 2008). Furthermore, TSP-1 may
be a potent inhibitor of T-cell receptor (TCR)-mediated T-cell
activation. Indeed, TSP-1 signalling through CD47 and
HSPGs inhibits TCR signalling including suppression of the
autocrine growth factor IL-2 and early activation markers
such as CD69, Egr-1 and PAC-1 (Li et al., 2001). T-regulatory
cells (Tregs) are critical in the limitation of damage induced
by exacerbated responses to ‘self’ or ‘foreign’ antigens, but in
cancer, they have been shown to proliferate. Because tumour
antigens are derived from the host, tumours are more likely
regarded as self by Tregs that actively promote their tolerance.
Increased infiltration of Tregs in tumours is thus correlated
with poor prognosis in several types of cancers (Elkord et al.,
2010). CD47 stimulation by TSP-1 or 4N1K was demonstrated
to trigger the conversion of naïve or memory CD4+ CD25-
T-cells into Tregs (Grimbert et al., 2006). Furthermore, CD47
cross-linking was shown to induce apoptosis of normal T- and
B-cells when cultured on immobilized B6H12 mAbs (Mateo
et al., 2002) or in the presence of soluble anti-CD47 Ad22
mAbs (Lamy et al., 2003). Interestingly, haematopoietic pro-
genitors CD34+ and immature DCs were described to be
resistant to CD47-induced apoptosis (Mateo et al., 2002).
CD47 was also reported to induce the development of
hyporesponsive or anergic T-cells (Avice et al., 2001).
Finally, levels of CD47 expression in head-and-neck squa-
mous cell carcinoma apparently influence the cytotoxicity of
NK cells, with increased expression of CD47 being related to
a decrease in NK cytotoxicity (Kim et al., 2008). The mecha-
nism remains unknown but may represent a possible supple-
mentary pathway for immune escape in cancer.
Taken together, these data indicate that blockade of CD47
receptors apparently triggers immune system responses
against tumours at many different levels. This may thus rep-
resent a promising therapeutic strategy resulting in increased
tumour cell phagocytosis, enhanced DC maturation and sub-
sequently augmented T-cell-mediated cytotoxicity.
This overview of CD47 indicates that CD47 very likely plays
a central role in the tumour microenvironment. Considering
its ubiquitous expression, its ability to modulate tumour cell
growth and migration and to limit angiogenesis, and the
development of immune responses against tumours, CD47
therefore appears to represent an exciting target candidate for
new therapeutic approaches against cancer (Figure 2).
A number of recent studies using function-blocking anti-
CD47 mAbs are consistent with an eventual therapeutic
application in the field of leukaemia. For example, single-
chain Ab fragments against CD47 were successfully used to
induce apoptosis of malignant lymphoid cells isolated from
patient with B-CLL and to significantly increase survival rate
of mice implanted with human lymphocytic leukaemia cell
lines (Sagawa et al., 2011). In addition to directly inducing
apoptosis, other mechanisms of action for anti-CD47 mAbs
includethe stimulation of
dependent cytotoxicity and the induction of leukaemic stem
cell phagocytosis by disruption of CD47/SIRPa interaction. A
demonstration of this latter effect was the phagocytosis of
human AML LSCs treated with B6H12 mAbs by mouse mac-
rophages in vivo (Majeti et al., 2009). Of note, the mouse
anti-CD47-blocking mAb miap 301 promoted phagocytosis
of mouse AML while not depleting normal HSCs in vivo
(Majeti et al., 2009), consistent with the feasibility of thera-
peutic targeting of CD47. A synergic effect of anti-CD47
mAbs in combination with rituximab, an anti-CD20 mAb, in
promoting phagocytosis was described for the eradication of
non-Hodgkin lymphoma (Chao et al., 2010a). To date, most
of the studies reporting successful eradication of leukaemia
cells by anti-CD47 mAbs were done using human tumour
xenograft models in immunodeficient mice. Because only
human leukaemic cells were targeted by the human anti-
CD47 mAbs used in such experimental models, interpreta-
tion of the observed depletion of leukaemic cells is somewhat
hampered. Notably, characterization of possible side effects
that could occur in immunocompetent mice or in humans is
not accessible with these models.
To our knowledge, very few studies have addressed the
role of CD47 in solid tumours, which are highly vascular-
ized, and for which little is known about pathological
cell-to-cell interactions in tumour stroma. An accurate
description of CD47 function during vascularization at the
site of tumours is still lacking and available data sometimes
appear contradictory or paradoxical. TSP-1, a well-known
inhibitor of angiogenesis, acts physiologically via CD47
through inhibition of NO signalling. However, little is
known about TSP-1 responsiveness in tumour vasculature
where NO could act as a friend or a foe, either by dilatation
of blood vessels in the tumour or dilatation of peripheral
blood vessels resulting in redistribution of blood away from
the tumour. Thus, in this context, it is currently very diffi-
cult to predict outcomes if systemic blocking strategies
against CD47 were to be tested.
Furthermore, as described beforehand, depending on the
cell type considered, CD47 activation is reported either to
induce apoptosis or in contrast to enhance proliferation and
survival. On the other hand, cell migration appears to be
quite universally stimulated by CD47 ligation and activation
(Figure 2). Thus, antagonism of CD47 should reduce cell
migration, thereby representing a plausible strategy to fight
against cancer cell dissemination. Blockade of SIRPa/CD47
interaction that subsequently promotes immune system
responses against tumours (Figure 2), especially in cases
where CD47 is highly overexpressed, likewise appears to be a
Given the ubiquitous expression of CD47, specific target-
ing will be a critical issue to resolve. This might be addressed
using mAbs directed against both CD47 and a specific tumour
cell marker. For example, because high levels of CD47 are
E Sick et al.
1424 British Journal of Pharmacology (2012) 167 1415–1430
found in prostate cancer (Vallbo and Damber, 2005), devel-
opment of blocking CD47 mAbs that also target the prostate-
specific antigen could be of particular therapeutic interest. It
should nevertheless be pointed out that strategies using anti-
CD47 mAbs may also mask binding sites for different endog-
enous ligands or partners, which might prevent some of their
beneficial effects against tumours. Therefore, the develop-
ment of alternative strategies using synthetic competitive
antagonists of CD47 is likely required. In this context,
detailed X-ray analysis and molecular modelling studies of
the TSP-1/CD47 complex would be helpful. Indeed, this
might lead to the characterization of non-evident binding
sites or other sites of interaction, thereby perhaps contribut-
ing to the development of new highly specific and efficacious
small molecule antagonists of CD47.
In conclusion, it thus may be that CD47 represents an
interesting therapeutic target for the treatment of cancer.
However, evaluation is clearly hampered, given the current
lack of candidate molecules and accordingly, the absence of
E. S. was a recipient of a Fellowship from the Région
Champagne-Ardenne. A. J. was a recipient of a Fellowship
from the Ministère de la Recherche.
Conflict of interest
Adams JC (2001). Thrombospondins: multifunctional regulators of
cell interactions. Annu Rev Cell Dev Biol 17: 25–51.
Avice MN, Rubio M, Sergerie M, Delespesse G, Sarfati M (2001).
Role of CD47 in the induction of human naïve T cell anergy.
J Immunol 167: 2459–2468.
Babic I, Schallhorn A, Lindberg FP, Jirik FR (2000). SHPS-1 induces
aggregation of Ba/F3 pro-B cells via interaction with CD47.
J Immunol 164: 3652–3658.
Balkwill F, Mantovani A (2001). Inflammation and cancer: back to
Virchow? Lancet 357: 539–545.
Bamberger ME, Harris ME, McDonald DR, Husemann J,
Landreth GE (2003). A cell surface receptor complex for fibrillar
b-amyloid mediates microglial activation. J Neurosci 23: 2665–2674.
Barazi HO, Li Z, Cashel JA, Krutzsch HC, Annis DS, Mosher DF et al.
(2002). Regulation of integrin function by CD47 ligands. J Biol
Chem 277: 42859–42866.
Barbier S, Chatre L, Bras M, Sancho P, Roué G, Virely C et al.
(2009). Caspase-independent type III programmed cell death in
chronic lymphocytic leukemia: the key role of the F-actin
cytoskeleton. Haematologica 94: 507–517.
Potential consequences of therapeutic blockade of CD47 in the tumour microenvironment. (a) Modulation of apoptosis of tumour cells.
(b) Modulation of tumour cell proliferation or survival. (c) Inhibition of tumour cell migration and transmigration across the endothelium.
(d) Stimulation of tumour cell phagocytosis by disruption of CD47/SIRPa interaction. (e) Modulation of angiogenesis. (f) Inhibition of endothelial
cell apoptosis. (g) Prevention of apoptosis of normal cells after irradiation. The question marks indicate effects that might also be influenced by
the cellular environment.
CD47 in the tumour microenvironment
British Journal of Pharmacology (2012) 167 1415–1430 1425
Bauer EM, Qin Y, Miller TW, Bandle RW, Csanyi G, Pagano PJ et al.
(2010). Thrombospondin-1 supports blood pressure by limiting
eNOS activation and endothelial-dependent vasorelaxation.
Cardiovasc Res 88: 471–481.
Bergers G, Hanahan D (2008). Modes of resistance to
anti-angiogenic therapy. Nature Rev Cancer 8: 592–603.
Bras M, Yuste VJ, Roué G, Barbier S, Sancho P, Virely C et al. (2007).
Drp1 mediates caspase-independent type III cell death in normal
and leukemic cells. Mol Cell Biol 27: 7073–7088.
Brittain JE, Han J, Ataga KI, Orringer EP, Parise LV (2004).
Mechanism of CD47-induced a4b1 integrin activation and
adhesion in sickle reticulocytes. J Biol Chem 279: 42393–42402.
Brooke G, Holbrook JD, Brown MH, Barclay AN (2004). Human
lymphocytes interact directly with CD47 through a novel member
of the signal regulatory protein (SIRP) family. J Immunol 173:
Brooke GP, Parsons KR, Howard CJ (1998). Cloning of two
members of the SIRPa family of protein tyrosine phosphatase
binding proteins in cattle that are expressed on monocytes and a
subpopulation of dendritic cells and which mediate binding to CD4
T cells. Eur J Immunol 28: 1–11.
Broom OJ, Zhang Y, Oldenborg PA, Massoumi R, Sjölander A
(2009). CD47 regulates collagen I-induced cyclooxygenase-2
expression and intestinal epithelial cell migration. PLoS ONE 4:
Brown E, Hooper L, Ho T, Gresham H (1990). Integrin-associated
protein: a 50-kD plasma membrane antigen physically and
functionally associated with integrins. J Cell Biol 111: 2785–2794.
Brown EJ, Frazier WA (2001). Integrin-associated protein (CD47)
and its ligands. Trends Cell Biol 11: 130–135.
Bruce LJ, Ghosh S, King MJ, Layton DM, Mawby WJ, Stewart GW
et al. (2002). Absence of CD47 in protein 4.2-deficient hereditary
spherocytosis in man: an interaction between the Rh complex and
the band 3 complex. Blood 100: 1878–1885.
Campbell IG, Freemont PS, Foulkes W, Trowsdale J (1992). An
ovarian tumour marker with homology to vaccinia virus contains
an IgV-like region and multiple transmembrane domains. Cancer
Res 52: 5416–5420.
Capo C, Lindberg FP, Meconi S, Zaffran Y, Tardei G, Brown EJ et al.
(1999). Subversion of monocyte functions by Coxiella burnetii:
impairment of the cross-talk between avb3 integrin and CR3.
J Immunol 163: 6078–6085.
Carmeliet P, Jain RK (2011). Molecular mechanisms and clinical
applications of angiogenesis. Nature 473: 298–307.
Chakraborty AK, Sodi S, Rachkovsky M, Kolesnikova N, Platt JT,
Bolognia JL et al. (2000). A spontaneous murine melanoma lung
metastasis comprised of host x tumour hybrids. Cancer Res 60:
Chames P, Van Regenmortel M, Weiss E, Baty D (2009). Therapeutic
antibodies: successes, limitations and hopes for the future. Br J
Pharmacol 157: 220–233.
Chan KS, Espinosa I, Chao M, Wong D, Ailles L, Diehn M et al.
(2009). Identification, molecular characterization, clinical
prognosis, and therapeutic targeting of human bladder
tumour-initiating cells. Proc Natl Acad Sci USA 106: 14016–14021.
Chao MP, Alizadeh AA, Tang C, Myklebust JH, Varghese B, Gill S
et al. (2010a). Anti-CD47 antibody synergizes with rituximab to
promote phagocytosis and eradicate non-Hodgkin lymphoma. Cell
Chao MP, Jaiswal S, Weissman-Tsukamoto R, Alizadeh AA,
Gentles AJ, Volkmer J et al. (2010b). Calreticulin is the dominant
pro-phagocytic signal on multiple human cancers and is
counterbalanced by CD47. Sci Transl Med 2: 63ra94.
Chao MP, Alizadeh AA, Tang C, Jan M, Weissman-Tsukamoto R,
Zhao F et al. (2011). Therapeutic antibody targeting of CD47
eliminates human acute lymphoblastic leukaemia. Cancer Res 71:
Chung J, Gao AG, Frazier WA (1997). Thrombospondin acts via
integrin-associated protein to activate the platelet integrin aIIb b3.
J Biol Chem 272: 14740–14746.
Congote LF, Temmel N (2004). The C-terminal 26-residue peptide
of serpin A1 stimulates proliferation of breast and liver cancer cells:
role of protein kinase C and CD47. FEBS Lett 576: 343–347.
Cooper D, Lindberg FP, Gamble JR, Brown EJ, Vadas MA (1995).
Transendothelial migration of neutrophils involves
integrin-associated protein (CD47). Proc Natl Acad Sci USA 92:
Demeure CE, Tanaka H, Mateo V, Rubio M, Delespesse G, Sarfati M
(2000). CD47 engagement inhibits cytokine production and
maturation of human dendritic cells. J Immunol 164: 2193–2199.
de Vries HE, Hendriks JJ, Honing H, De Lavalette CR,
van der Pol SM, Hooijberg E et al. (2002). Signal-regulatory protein
a-CD47 interactions are required for the transmigration of
monocytes across cerebral endothelium. J Immunol 168:
Dorahy DJ, Thorne RF, Fecondo JV, Burns GF (1997). Stimulation of
platelet activation and aggregation by carboxyl-terminal peptide
from thrombospondin binding to the integrin-associated protein
receptor. J Biol Chem 272: 1323–1330.
Dunn GP, Old LJ, Schreiber RD (2004). The three Es of cancer
immunoediting. Annu Rev Immunol 22: 329–360.
Elkord E, Alcantar-Orozco EM, Dovedi SJ, Tran DQ, Hawkins RE,
Gilham DE (2010). T regulatory cells in cancer: recent advances and
therapeutic potential. Expert Opin Biol Ther 10: 1573–1586.
Floquet N, Dedieu S, Martiny L, Dauchez M, Perahia D (2008).
Human thrombospondin’s (TSP-1) C-terminal domain opens to
interact with the CD47 receptor: a molecular modeling study. Arch
Biochem Biophys 478: 103–109.
Frazier WA, Gao AG, Dimitry J, Chung J, Brown EJ, Lindberg FP
et al. (1999). The thrombospondin receptor integrin-associated
protein (CD47) functionally couples to heterotrimeric Gi. J Biol
Chem 274: 8554–8560.
Frazier WA, Isenberg JS, Kaur S, Roberts DD (2010). CD47. Nature
Signalling Gateway. doi: 10.1038/mp.a002870.01 Available at:
A002870 (accessed 19 September 2012).
Gao AG, Frazier WA (1994). Identification of a receptor candidate
for the carboxyl-terminal cell binding domain of thrombospondins.
J Biol Chem 269: 29650–29657.
Gao AG, Lindberg FP, Finn MB, Blystone SD, Brown EJ, Frazier WA
(1996a). Integrin-associated protein is a receptor for the C-terminal
domain of thrombospondin. J Biol Chem 271: 21–24.
Gao AG, Lindberg FP, Dimitry JM, Brown EJ, Frazier WA (1996b).
Thrombospondin modulates avb3 function through
integrin-associated protein. J Cell Biol 135: 533–544.
Gardai SJ, McPhillips KA, Frasch SC, Janssen WJ, Starefeld A,
Murphy-Ullrich JE et al. (2005). Cell-surface calreticulin initiates
clearance of viable or apoptotic cells through trans-activation of
LRP on the phagocyte. Cell 123: 321–334.
E Sick et al.
1426 British Journal of Pharmacology (2012) 167 1415–1430
Green JM, Zhelesnyak A, Chung J, Lindberg FP, Sarfati M,
Frazier WA et al. (1999). Role of cholesterol in formation and
function of a signalling complex involving avb3, integrin-
associated protein (CD47), and heterotrimeric G proteins. J Cell
Biol 146: 673–682.
Grimbert P, Bouguermouh S, Baba N, Nakajima T, Allakhverdi Z,
Braun D et al. (2006). Thrombospondin/CD47 interaction: a
pathway to generate regulatory T cells from human CD4+ CD25- T
cells in response to inflammation. J Immunol 177: 3534–3541.
Han X, Sterling H, Chen Y, Saginario C, Brown EJ, Frazier WA et al.
(2000). CD47, a ligand for the macrophage fusion receptor,
participates in macrophage multinucleation. J Biol Chem 275:
Hanahan D, Weinberg RA (2011). Hallmarks of cancer: the next
generation. Cell 144: 646–674.
Hatherley D, Graham SC, Turner J, Harlos K, Stuart DI, Barclay AN
(2008). Paired receptor specificity explained by structures of signal
regulatory proteins alone and complexed with CD47. Mol Cell 31:
Hatherley D, Graham SC, Harlos K, Stuart DI, Barclay AN (2009).
Structure of signal-regulatory protein a: a link to antigen receptor
evolution. J Biol Chem 284: 26613–26619.
Hawighorst T, Velasco P, Streit M, Hong YK, Kyriakides TR,
Brown LF et al. (2001). Thrombospondin-2 plays a protective role in
multistep carcinogenesis: a novel host anti-tumour defence
mechanism. EMBO J 20: 2631–2640.
Hermann P, Armant M, Brown E, Rubio M, Ishihara H, Ulrich D
et al. (1999). The vitronectin receptor and its associated CD47
molecule mediates proinflammatory cytokine synthesis in human
monocytes by interaction with soluble CD23. J Cell Biol 144:
Hofman P, Piche M, Far DF, Le Negrate G, Selva E, Landraud L et al.
(2000). Increased Escherichia coli phagocytosis in neutrophils that
have transmigrated across a cultured intestinal epithelium. Infect
Immun 68: 449–455.
Holledge MM, Millward-Sadler SJ, Nuki G, Salter DM (2008).
Mechanical regulation of proteoglycan synthesis in normal and
osteoarthritic human articular chondrocytes: roles for a5 and avb5
integrins. Biorheology 45: 275–288.
Ichigotani Y, Matsuda S, Machida K, Oshima K, Iwamoto T,
Yamaki K et al. (2000). Molecular cloning of a novel human gene
(SIRP-B2) which encodes a new member of the SIRP/SHPS-1 protein
family. J Hum Genet 45: 378–382.
Isenberg JS, Ridnour LA, Dimitry J, Frazier WA, Wink DA,
Roberts DD (2006). CD47 is necessary for inhibition of nitric
oxide-stimulated vascular cell responses by thrombospondin-1.
J Biol Chem 281: 26069–26080.
Isenberg JS, Romeo MJ, Abu-Asab M, Tsokos M, Oldenborg A,
Pappan L et al. (2007). Increasing survival of ischemic tissue by
targeting CD47. Circ Res 100: 712–720.
Isenberg JS, Romeo MJ, Yu C, Yu CK, Nghiem K, Monsale J et al.
(2008a). Thrombospondin-1 stimulates platelet aggregation by
blocking the antithrombotic activity of nitric oxide/cGMP
signalling. Blood 111: 613–623.
Isenberg JS, Maxhimer JB, Hyodo F, Pendrak ML, Ridnour LA,
DeGraff WG et al. (2008b). Thrombospondin-1 and CD47 limit cell
and tissue survival of radiation injury. Am J Pathol 173: 1100–1112.
Isenberg JS, Pappan LK, Romeo MJ, Abu-Asab M, Tsokos M,
Wink DA et al. (2008c). Blockade of thrombospondin-1-CD47
interactions prevents necrosis of full thickness skin grafts. Ann Surg
Isenberg JS, Romeo MJ, Maxhimer JB, Smedley J, Frazier WA,
Roberts DD (2008d). Gene silencing of CD47 and antibody ligation
of thrombospondin-1 enhance ischemic tissue survival in a porcine
model: implications for human disease. Ann Surg 247: 860–868.
Isenberg JS, Maxhimer JB, Pwers P, Tsokos M, Frazier WA,
Roberts DD (2008e). Treatment of liver ischemia-reperfusion injury
by limiting thrombospondin-1/CD47 signalling. Surgery 144:
Isenberg JS, Hyodo F, Ridnour LA, Shannon CS, Wink DA,
Krishna MC et al. (2008f). Thrombospondin-1 and vasoactive agents
indirectly alter tumour blood flow. Neoplasia 10: 886–896.
Isenberg JS, Shiva S, Gladwin M (2009). Thrombospondin-1-CD47
blockade and exogenous nitrite enhance ischemic tissue survival,
blood flow and angiogenesis via coupled NO-cGMP pathway
activation. Nitric Oxide 21: 52–62.
Ishii M, Iwai K, Koike M, Ohshima S, Kudo-Tanaka E, Ishii T et al.
(2006). RANKL-induced expression of tetraspanin CD9 in lipid raft
membrane microdomain is essential for cell fusion during
osteoclastogenesis. J Bone Miner Res 21: 965–976.
Jaiswal S, Jamieson CH, Pang WW, Park CY, Chao MP, Majeti R
et al. (2009). CD47 is upregulated on circulating hematopoietic
stem cells and leukaemia cells to avoid phagocytosis. Cell 138:
Jiang P, Lagenaur CF, Narayanan V (1999). Integrin-associated
protein is a ligand for the P84 neural adhesion molecule. J Biol
Chem 274: 559–562.
Johansson U, Londei M (2004). Ligation of CD47 during monocyte
differentiation into dendritic cells results in reduced capacity for
interleukin-12 production. Scand J Immunol 59: 50–57.
Kaur S, Martin-Manso G, Pendrak ML, Garfield SH, Isenberg JS,
Roberts DD (2010). Thrombospondin-1 inhibits VEFG receptor-2
signalling by disrupting its association with CD47. J Biol Chem
Kaur S, Kuznetsova SA, Pendrak ML, Sipes JM, Romeo MJ, Li Z et al.
(2011). Heparan sulfate modification of the transmembrane
receptor CD47 is necessary for inhibition of T cell receptor
signalling by thrombospondin-1. J Biol Chem 286: 14991–15002.
Kharitonenkov A, Chen Z, Sures I, Wang H, Schilling J, Ullrich A
(1997). A family of proteins that inhibit signalling through tyrosine
kinase receptors. Nature 386: 181–186.
Kikuchi Y, Uno S, Yoshimura Y, Otabe K, Iida S, Oheda M et al.
(2004). A bivalent single-chain Fv fragment against CD47 induces
apoptosis for leukemic cells. Biochem Biophys Res Commun 315:
Kikuchi Y, Uno S, Kinoshita Y, Yoshimura Y, Iida S, Wakahara Y
et al. (2005). Apoptosis inducing bivalent single-chain antibody
fragments against CD47 showed antitumour potency for multiple
myeloma. Leuk Res 29: 445–450.
Kim MJ, Lee JC, Lee JJ, Kim S, Lee SG, Park SW et al. (2008).
Association of CD47 with natural killer cell-mediated cytotoxicity
of head-and-neck squamous cell carcinoma lines. Tumour Biol 29:
Koenigsknecht J, Landreth G (2004). Microglial phagocytosis of
fibrillar b-amyloid through a b1 integrin-dependent mechanism.
J Neurosci 24: 9838–9846.
Lamy L, Ticchioni M, Rouquette-Jazdanian AK, Samson M,
Deckert M, Greenberg AH et al. (2003). CD47 and the 19 kDa
interacting protein-3 (BNIP3) in T cell apoptosis. J Biol Chem 278:
CD47 in the tumour microenvironment
British Journal of Pharmacology (2012) 167 1415–14301427
Lamy L, Foussat A, Brown EJ, Bornstein P, Ticchioni M, Bernard A
(2007). Interactions between CD47 and thrombospondin reduce
inflammation. J Immunol 178: 5930–5939.
Lee EH, Hsieh YP, Yang CL, Tsai KJ, Liu CH (2000). Induction of
integrin-associated protein (IAP) mRNA expression during memory
consolidation in rat hippocampus. Eur J Neurosci 12: 1105–1112.
Li SS, Forslöw A, Sundqvist KG (2005). Autocrine regulation of
T cell motility by calreticulin-thrombospondin-1 interaction.
J Immunol 174: 654–661.
Li Z, He L, Wilson KE, Roberts DD (2001). Thrombospondin-1
inhibits TCR-mediated T lymphocyte early activation. J Immunol
Lih CJ, Wei W, Cohen SN (2006). Txr1: a transcriptional regulator
of thrombospondin-1 that modulates cellular sensitivity to taxanes.
Genes Dev 20: 2082–2095.
Lindberg FP, Gresham HD, Schwarz E, Brown EJ (1993). Molecular
cloning of integrin-associated protein: an immunoglobulin family
member with multiple membrane-spanning domains implicated in
avb3-dependent ligand binding. J Cell Biol 123: 485–496.
Lindberg FP, Gresham HD, Reinhold MI, Brown EJ (1996).
Integrin-associated protein immunoglobulin domain is necessary
for efficient vitronectin bead binding. J Cell Biol 134: 1313–1322.
Liu Y, Merlin D, Burst SL, Pochet M, Madara JL, Parkos CA (2001).
The role of CD47 in neutrophil transmigration. Increased rate of
migration correlates with increased cell surface expression of CD47.
J Biol Chem 276: 40156–40166.
Liu Y, Bühring HJ, Zen K, Burst SL, Schnell FJ, Williams IR et al.
(2002). Signal regulatory protein (SIRPa), a cellular ligand for CD47,
regulates neutrophil transmigration. J Biol Chem 277:
Lundberg P, Koskinen C, Baldock PA, Löthgren H, Stenberg A,
Lerner UH et al. (2007). Osteoclast formation is strongly reduced
both in vivo and in vitro in the absence of CD47/SIRPa-interaction.
Biochem Biophys Res Commun 352: 444–448.
McDonald JF, Zheleznyak A, Frazier WA (2004).
Cholesterol-independent interactions with CD47 enhance avb3
avidity. J Biol Chem 279: 17301–17311.
Majeti R, Chao MP, Alizadeh AA, Pang WW, Jaiswal S, Gibbs KD Jr
et al. (2009). CD47 is an adverse prognostic factor and therapeutic
antibody target on human acute myeloid leukaemia stem cells. Cell
Manna PP, Frazier WA (2003). The mechanism of CD47-dependent
killing of T cells: heterotrimeric Gi-dependent inhibition of protein
kinase A. J Immunol 170: 3544–3553.
Manna PP, Frazier WA (2004). CD47 mediates killing of breast
tumour cells via Gi-dependent inhibition of protein kinase A.
Cancer Res 64: 1026–1036.
Manna PP, Dimitry J, Oldenborg PA, Frazier WA (2005). CD47
augments Fas/CD95-mediated apoptosis. J Biol Chem 280:
Mantovani A, Marchesi F, Portal C, Allavena P, Sica A (2008).
Linking inflammation reactions to cancer: novel targets for
therapeutic strategies. Adv Exp Med Biol 610: 112–127.
Mateo V, Lagneaux L, Bron D, Biron G, Armant M, Delespesse G
et al. (1999). CD47 ligation induces caspase-independent cell death
in chronic lymphocytic leukemia. Nat Med 5: 1277–1284.
Mateo V, Brown EJ, Biron G, Rubio M, Fisher A, Deist FL et al.
(2002). Mechanisms of CD47-induced caspase-independent cell
death in normal and leukemic cells: link between
phosphatidylserine exposure and cytoskeleton organization.
Blood 100: 2882–2890.
Mawby WJ, Holmes CH, Anstee DJ, Spring FA, Tanner MJ (1994).
Isolation and characterization of CD47 glycoprotein: a
multispanning membrane protein which is the same as
integrin-associated protein (IAP) and the ovarian tumour marker
OA3. Biochem J 304: 525–530.
Maxhimer JB, Soto-Pantoja DR, Ridnour LA, Shih HB, Degraff WG,
Tsokos M et al. (2009a). Radioprotection in normal tissue and
delayed tumour growth by blockade of CD47 signalling. Sci Transl
Med 1: 3ra7.
Maxhimer JB, Shih HB, Isenberg JS, Miller TW, Roberts DD (2009b).
Thrombospondin-1/CD47 blockade following ischemia-reperfusion
injury is tissue protective. Plast Reconstr Surg 124: 1880–1889.
Miller TW, Isenberg JS, Shih HB, Wang Y, Roberts DD (2010a).
Amyloid-b inhibits NO-cGMP signalling in a CD36- and
CD47-dependent manner. PLoS ONE 5: e15686.
Miller TW, Isenberg JS, Roberts DD (2010b). Thrombospondin-1 is
an inhibitor of pharmacological activation of soluble guanylate
cyclase. Br J Pharmacol 159: 1542–1547.
Miyashita M, Ohnishi H, Okazawa H, Tomonaga H, Hayashi A,
Fujimoto TT et al. (2004). Promotion of neurite and filopodium
formation by CD47: roles of integrins, Rac, and Cdc42. Mol Biol
Cell 15: 3950–3963.
Murata T, Ohnishi H, Okazawa H, Murata Y, Kusakari S, Hayashi Y
et al. (2006). CD47 promotes neuronal development through Src-
and FRG/Vav2-mediated activation of Rac and Cdc42. J Neurosci
N’Diaye EN, Brown EJ (2003). The ubiquitin-related protein PLIC-1
regulates heterotrimeric G protein function through association
with Gbg. J Cell Biol 163: 1157–1165.
Ogura T, Noguchi T, Murai-Takebe R, Hosooka T, Honma N,
Kasuga M (2004). Resistance of B16 melanoma cells to
CD47-induced regulation of motility as a result of aberrant
N-glycosylation of SHPS-1. J Biol Chem 279: 13711–13720.
Oldenborg PA, Zheleznyak A, Fang YF, Lagenaur CF, Gresham HD,
Lindberg FP (2000). Role of CD47 as a marker of self on red blood
cells. Science 288: 2051–2054.
Olsson M, Bruhns P, Frazier WA, Ravetch JV, Oldenborg PA (2005).
Platelet homeostasis is regulated by platelet expression of CD47
under normal conditions and in passive immune
thrombocytopenia. Blood 105: 3577–3582.
Parkos CA, Colgan SP, Liang TW, Nusrat A, Bacarra AE, Carnes DK
et al. (1996). CD47 mediates post-adhesive events required for
neutrophil migration across polarized intestinal epithelia. J Cell
Biol 132: 437–450.
Parthasarathy R, Subramanian S, Boder ET, Discher DE (2006).
Post-translational regulation of expression and conformation of an
immunoglobulin domain in yeast surface display. Biotechnol
Bioeng 93: 159–168.
Petersen TR, Dickgreber N, Hermans IF (2010). Tumour antigen
presentation by dendritic cells. Crit Rev Immunol 30: 345–386.
Pettersen RD, Hestdal K, Olafsen MK, Lie SO, Lindberg FP (1999).
CD47 signals T cell death. J Immunol 162: 7031–7040.
Piccio L, Vermi W, Boles KS, Fuchs A, Strader CA, Facchetti F et al.
(2005). Adhesion of human T cells to antigen-presenting cells
through SIRPb2-CD47 interaction costimulates T-cell proliferation.
Blood 105: 2421–2427.
Powell AE, Anderson EC, Davies PS, Silk AD, Pelz C, Impey S et al.
(2011). Fusion between intestinal epithelial cells and macrophages
in a cancer context results in nuclear reprogramming. Cancer Res
E Sick et al.
1428 British Journal of Pharmacology (2012) 167 1415–1430
Rabinovich GA, Gabrilovich D, Sotomayor EM (2007).
Immunosuppressive strategies that are mediated by tumour cells.
Annu Rev Immunol 25: 267–296.
Rachkovsky M, Sodi S, Chakraborty A, Avissar Y, Bolognia J,
McNiff JM et al. (1998). Melanoma x macrophage hybrids with
enhanced metastatic potential. Clin Exp Metastasis 16: 299–312.
Ramanathan S, Mazzalupo S, Boitano S, Montfort WR (2011).
Thrombospondin-1 and angiotensin II inhibit soluble guanylyl
cyclase through an increase in intracellular calcium concentration.
Biochemistry 50: 7787–7799.
Rath GM, Schneider C, Dedieu S, Sartelet H, Morjani H, Martiny L
et al. (2006a). Thrombospondin-1 C-terminal-derived peptide
protects thyroid cells from ceramide-induced apoptosis through the
adenylyl cyclase pathway. Int J Biochem Cell Biol 38: 2219–2228.
Rath GM, Schneider C, Dedieu S, Rothhut B, Soula-Rothhut M,
Ghoneim C et al. (2006b). The C-terminal CD47/IAP-binding
domain of thrombospondin-1 prevents camptothecin- and
doxorubicin-induced apoptosis in human thyroid carcinoma cells.
Biochim Biophys Acta 1763: 1125–1134.
Rebres RA, Vaz LE, Green JM, Brown EJ (2001). Normal ligand
binding and signalling by CD47 (integrin-associated protein)
requires a long range disulfide bond between the extracellular and
membrane-spanning domains. J Biol Chem 276: 34607–34616.
Rebres RA, Kajihara K, Brown EJ (2005). Novel CD47-dependent
intercellular adhesion modulates cell migration. J Cell Physiol 205:
Reinhold MI, Lindberg FP, Plas D, Reynolds S, Peters MG, Brown EJ
(1995). In vivo expression of alternatively spliced forms of
integrin-associated protein (CD47). J Cell Sci 108: 3419–3425.
Roué G, Bitton N, Yuste VJ, Montange T, Rubio M, Dessauge F et al.
(2003). Mitochondrial dysfunction in CD47-mediated
caspase-independent cell death: ROS production in the absence of
cytochrome C and AIF release. Biochimie 85: 741–746.
Sagawa M, Shimizu T, Fukushima N, Kinoshita Y, Ohizumi I, Uno S
et al. (2011). A new disulfide-linked dimer of a single-chain
antibody fragment against human CD47 induces apoptosis in
lymphoid malignant cells via the hypoxia inducible factor-1a
pathway. Cancer Sci 102: 1208–1215.
Sano S, Ohnishi H, Omori A, Hasegawa J, Kubota M (1997). BIT, an
immune antigen receptor-like molecule in the brain. FEBS Lett 411:
Saumet A, Slimane MB, Lanotte M, Lawler J, Dubernard V (2005).
Type 3 repeat/C-terminal domain of thrombospondin-1 triggers
caspase-independent cell death through CD47/avb3 in
promyelocytic leukemia NB4 cells. Blood 106: 658–667.
Schwartz MA, Brown EJ, Fazeli B (1993). A 50-kDa
integrin-associated protein is required for integrin-regulated calcium
entry in endothelial cells. J Biol Chem 268: 19931–19934.
Seiffert M, Brossart P, Cant C, Cella M, Colonna M, Brugger W et al.
(2001). Signal-regulatory protein alpha (SIRPa) but not SIRPb is
involved in T-cell activation, binds to CD47 with high affinity, and
is expressed on immature CD34(+)CD38(-) hematopoietic cells.
Blood 97: 2741–2749.
Shahan TA, Fawzi A, Bellon G, Monboisse JC, Kefalides NA (2000).
Regulation of tumour cell chemotaxis by type IV collagen is
mediated by a Ca2+-dependent mechanism requiring CD47 and the
integrin avb3. J Biol Chem 275: 4796–4802.
Shinohara M, Ohyama N, Murata Y, Okazawa H, Ohnishi H,
Ishikawa O et al. (2006). CD47 regulation of epithelial cell
spreading and migration, and its signal transduction. Cancer Sci 97:
Sick E, Niederhoffer N, Takeda K, Landry Y, Gies JP (2009).
Activation of CD47 receptors causes histamine secretion from mast
cells. Cell Mol Life Sci 66: 1271–1282.
Sick E, Boukhari A, Deramaudt T, Rondé P, Bucher B, André P et al.
(2011). Activation of CD47 receptors causes proliferation of human
astrocytoma but not normal astrocytes via an Akt-dependent
pathway. Glia 59: 308–319.
Smadja DM, d’Audignier C, Bièche I, Evrard S, Mauge L, Dias JV
et al. (2011). Thrombospondin-1 is a plasmatic marker of peripheral
arterial disease that modulates endothelial progenitor cell
angiogenic properties. Arterioscler Thromb Vasc Biol 31: 551–559.
Stefanidakis M, Newton G, Lee WY, Parkos CA, Luscinskas FW
(2008). Endothelial CD47 interaction with SITPg is required for
human T-cell transendothelial migration under shear flow
conditions in vitro. Blood 112: 1280–1289.
Sterling H, Saginario C, Vignery A (1998). CD44 occupancy
prevents macrophage multinucleation. J Cell Biol 143: 837–847.
Streit M, Velasco P, Brown LF, Skobe M, Richard L, Riccardi L et al.
(1999). Overexpression of thrombospondin-1 decreases angiogenesis
and inhibits the growth of human cutaneous squamous cell
carcinomas. Am J Pathol 155: 441–452.
Tlsty TD, Coussens LM (2006). Tumor stroma and regulation of
cancer development. Annu Rev Pathol 1: 119–150.
Tsai RK, Discher DE (2008). Inhibition of ‘self’ engulfment through
deactivation of myosin-II at the phagocytic synapse between
human cells. J Cell Biol 180: 989–1003.
Tsao PW, Mousa SA (1995). Thrombospondin mediates calcium
mobilization in fibroblasts via its Arg-Gly-Asp and
carboxyl-terminal domains. J Biol Chem 270: 23747–23753.
Tulasne D, Judd BA, Johansen M, Asazuma N, Best D, Brown EJ
et al. (2001). C-terminal peptide of thrombospondin-1 induces
platelet aggregation through the Fc receptor g-chain-associated
signaling pathway and by agglutination. Blood 98: 3346–3352.
Uluçkan O, Becker SN, Deng H, Zou W, Prior JL, Piwnica-Worms D
et al. (2009). CD47 regulates bone mass and tumour metastasis to
bone. Cancer Res 69: 3196–3204.
Uno S, Kinoshita Y, Azuma Y, Tsunenari T, Yoshimura Y, Iida S
et al. (2007). Antitumour activity of a monoclonal antibody against
CD47 in xenograft models of human leukemia. Oncol Rep 17:
Vallbo C, Damber JE (2005). Thrombospondins, metalloproteases
and thrombospondin receptors messenger RNA and protein
expression in different tumour sublines of the Dunning prostate
cancer model. Acta Oncol 44: 293–298.
Vernon-Wilson EF, Kee WJ, Willis AC, Barclay AN, Simmons DL,
Brown MH (2000). CD47 is a ligand for rat macrophage membrane
signal regulatory protein SIRP (OX41) and human SIRPa 1. Eur J
Immunol 30: 2130–2137.
Waclavicek M, Majdic O, Stulnig T, Berger M, Baumruker T,
Knapp W et al. (1997). T cell stimulation via CD47: agonistic and
antagonistic effects of monoclonal antibody 1/1A4. J Immunol 159:
Wang XQ, Frazier WA (1998). The thrombospondin receptor CD47
(IAP) modulates and associates with a2b1 integrin in vascular
smooth muscle cells. Mol Biol Cell 9: 865–974.
Wang XQ, Lindberg FP, Frazier WA (1999). Integrin-associated
protein stimulates a2b1-dependent chemotaxis via Gi-mediated
CD47 in the tumour microenvironment
British Journal of Pharmacology (2012) 167 1415–14301429
inhibition of adenylate cyclase and extracellular-regulated kinases. Download full-text
J Cell Biol 147: 389–400.
Wilson KE, Li Z, Kara M, Gardner KL, Roberts DD (1999). b1
integrin- and proteoglycan-mediated stimulation of T lymphoma
cell adhesion and mitogen-activated protein kinase signalling by
thrombospondin-1 and thrombospondin-1 peptides. J Immunol
Wu AL, Wang J, Zheleznyak A, Brown EJ (1999). Ubiquitin-related
proteins regulate interaction of vimentin intermediate filaments
with the plasma membrane. Mol Cell 4: 619–625.
Xing C, Lee S, Kim WJ, Wang H, Yang YG, Ning M et al. (2009).
Neurovascular effects of CD47 signalling: promotion of cell death,
inflammation, and suppression of angiogenesis in brain endothelial
cells in vitro. J Neurosci Res 87: 2571–2577.
Yagi M, Miyamoto T, Toyama Y, Suda T (2006). Role of DC-STAMP
in cellular fusion of osteoclasts and macrophage giant cells. J Bone
Miner Metab 24: 355–358.
Yao M, Roberts DD, Isenberg JS (2011). Thrombospondin-1
inhibition of vascular smooth muscle cell responses occurs via
modulation of both cAMP and cGMP. Pharmacol Res 63: 13–22.
Yoshida H, Tomiyama Y, Ishikawa J, Oritani K, Matsumura I,
Shiraga M et al. (2000). Integrin-associated protein/CD47 regulates
motile activity in human B-cell lines through CDC42. Blood 96:
E Sick et al.
1430 British Journal of Pharmacology (2012) 167 1415–1430