In his 1911 article “About cell fusion with
qualitatively abnormal chromosome dis-
tribution as cause for tumour formation”
Aichel first proposed the fusion theory of
tumour progression and exhorted future
scientists to “study chromosomes from all
angles” to investigate it further1. He pro-
posed that the source of aneuploidy could be
fusion of tumour-invading leukocytes with
cancer cells, suggesting that a combination
of extra chromosomes and the “qualitative
differences” in chromosomes from the
two cell types could lead to the metastatic
phenotype (reviewed in Refs 2–4). Decades
later, the same hypothesis — that metastasis
is caused by leukocyte–tumour cell fusion —
was proposed independently by Meckler5,6
and by Goldenberg7,8. Several laboratories
have now reported that hybrids produced by
fusion in vitro or in vivo were aneuploid and
of higher metastatic potential (reviewed in
Ref. 2–4). In 1984, LaGarde and Kerbel sum-
marized the emerging concepts9: “[Tumour
cell hybridization] can lead to major changes
in gene expression. These processes can
lead to the evolution of subpopulations of
tumour cells having major losses or gains
in their malignant aggressiveness and
therefore represents a large-scale genetic
mechanism capable of generating genotypic
and phenotypic diversification. If the normal
host cell happens to be a lymphoreticular-
haematopoietic cell, it could donate this phe-
notype to cell types which otherwise do not
normally express metastatic traits.” There is
now considerable evidence to support these
The pathways of invasion and metastasis
have been under intense scientific scrutiny
and much is now known about the steps
involved10,11. However, the actual genesis of
metastatic cells from within populations
of non-metastatic cells of the primary
tumour is not understood. What are the ini-
tiating mechanisms that cause a carcinoma
or melanoma cell in the epithelium to free
its adhesions to neighboring cells, adapt a
migratory phenotype, cross the basal lamina
into the dermis, intravasate into the blood
circulatory system or lymphatics, extrava-
sate, and form new tumours in lymph nodes
and distant tissues or organs? The long-
standing view is essentially Darwinian: the
unstable cancer genome combined with host
selective pressures generates metastatic cells
in the otherwise non-metastatic primary
tumour12,13. This view continues to provide
the best framework for envisioning tumour
progression. Yet it is difficult to imagine how
this might occur through successive, stepwise
mutations as generation of a metastatic
phenotype would require activation and
silencing of large numbers of genes in the
primary tumour cell10. One solution to this
problem lies in the activation of master regu-
latory genes that control multiple pathways
and initiate pro-metastatic cascades14. This
has been highlighted in reports that master
regulators of epithelial–mesenchymal transi-
tion (EMT) in development, such as SNAIL
(also known as SNAI1), SLUG (also known
as SNAI2), secreted protein, acidic, cysteine-
rich (SPARC) and TWIST (also known as
TWIST1), have analogous roles in invasion
and metastasis, in which they activate
mesoderm-associated pathways of cellular
adhesion and migration10,14. For example,
in breast cancer TWIST activates micro-
RNA-10b, which in turn causes increased
expression of the pro-metastatic gene RHOC,
with increased metastatic potential of the
affected cells14. However, the mechanisms
through which master regulators such
as TWIST are themselves upregulated in
cancer are not understood. We propose
that at least in some cases this could be
initiated by fusion of cancer cells with bone
marrow-derived cells (BMDCs). Although
a transition from epithelial to mesodermal
gene expression is indeed a characteristic of
invasion and metastasis, the expressed genes
are often remarkably similar to those associ-
ated with migratory BMDCs, such as macro-
phages and other myeloid-lineage cells3,4,15.
Fusion of migratory BMDCs and cancer cells
with co-expression of both fusion partner
genomes provides a potential explanation for
In our opinion the fusion theory comes
closer to a unifying explanation of tumour
progression than any yet proposed. Fusion
represents a non-mutational mechanism that
could explain the aberrant gene expression
patterns associated with malignant cells.
Studies of macrophage–tumour cell fusions
have demonstrated that genes from both
parental partners are expressed in hybrid
cells17. Gene expression in such cells reflects
combinations of myeloid lineage genes along
with those of the cancer cell lineage, all in a
background of deregulated cell division. In
fact, many molecules and traits associated
with tumour progression are expressed by
Fusion of tumour cells with bone
marrow-derived cells: a unifying
explanation for metastasis
John M. Pawelek and Ashok K. Chakraborty
Abstract | The causes of metastasis remain elusive despite vast information on
cancer cells. We posit that cancer cell fusion with macrophages or other
migratory bone marrow-derived cells (BMDCs) provides an explanation.
BMDC–tumour hybrids have been detected in numerous animal models and
recently in human cancer. Molecular studies indicate that gene expression in
such hybrids reflects a metastatic phenotype. Should BMDC–tumour fusion be
found to underlie invasion and metastasis in human cancer, new approaches
for therapy would surely follow.
NATURE REvIEWS | cancer?
vOLUME 8 | MAY 2008 | 377
© 2008 Nature Publishing Group
healthy myeloid lineage cells, for example,
angiogenesis, motility, chemotaxis and
tropism, immune signalling, matrix
degradation and remodelling, responses
to hypoxia, and multi-drug resistance to
chemotherapy3,4. Tumour fusion could also
account for aneuploidy and genetic rear-
rangements in metastatic cells2,18. It is further
possible that tumour–BMDC fusions are a
source of cancer stem cells19. From studies
in animal and human cancers there is little
doubt that tumour hybrids are generated
in vivo and that at least in animals they can
be a source of metastases2–4. This Perspective
reviews the molecular and cellular pathways
that are activated following fusion of tumour
cells with BMDCs, their expression in
macrophages and other BMDCs, and their
similarities to those governing tumour
progression in animal and human cancer.
Cell fusion mechanisms
Cell fusion is a widespread phenomenon
in biology20. The pathways vary between
different cell types, suggesting that they have
evolved separately in different systems21–23.
However, there are many mechanistic
similarities24 and it was recently shown that
myoblasts and macrophages use some of
the same molecular components in fusion25.
Fusion might occur following phagocytosis of
cancer cells or apoptotic bodies by tumour-
associated macrophages or other phagocytes2.
Horizontal transfer of oncogenes during
phagocytosis of cancer cells in vitro was
demonstrated26. Cancer cell fusion can be
induced by viruses18,27. Endometrial and
breast cancers fuse by means of the protein
syncytin (encoded by ERVWE1)28. Chronic
activation of protein kinase AKT2 leads to
multinucleation and cell fusion in human
epithelial kidney cells29. Cell–cell invasion
mechanisms of ‘cellocytosis’23 or ‘entosis’30
may also initiate fusion. A general requirement
is that the two fusing membranes be in close
contact. This is accomplished by receptor–
ligand interactions, as seen in virus–cell
fusions18,27 and in macrophage–macrophage
fusions in the formation of osteoclasts and
giant cells22,31. Regarding macrophages, several
genes are involved in fusion32. For osteoclast
formation, three receptor systems involved in
fusion are macrophage fusion receptor (MFR,
also known as signal-regulatory protein α
(SIRPA)), CD44 and dendritic cell-specific
transmembrane protein (DC-STAMP, also
known as TM7SF4)22. MFR and its ligand
CD47 (thrombospondin 2 receptor) belong
to the immunoglobulin superfamily33. MFR
is expressed by myeloid cells and neurons
whereas CD47 is expressed in many cell
types. CD44, for which the fusion ligand is
unknown, is also transiently expressed in
an early stage in fusion. The extracellular
domain of CD44 is cleaved by membrane
type I matrix metalloproteinases, possibly
bringing plasma membranes closer as
a prelude to fusion23,34. DC-STAMP is a
chemokine-like receptor that is essential
for macrophage fusion to form osteoclasts
and giant cells31. Although the DC-STAMP
ligand is as yet undetermined, a candidate is
the cytokine CCL2 (also known as monocyte
chemoattractant protein 1 (MCP1)), which
is an important component of osteoclast and
giant cell formation35,36.
Macrophages may thus fuse with cancer
cells through their inherent fusion capabili-
ties. Likewise, cancer cells may be prone
to fusion because of aberrant expression
of fusion-associated receptors or ligands.
For example, CD44 is widely expressed in
cancer, in which it is a cell surface receptor
for hyaluronan and associated with poor
outcome37,38. It is also a marker for putative
solid tissue cancer stem cells in several differ-
ent neoplasms (for an example see Ref. 39).
CD47 and CCL2 (and CCL2 receptors) are
each expressed by many different cancers40–42.
Close apposition of plasma cell membranes
between macrophages and melanoma cells is
readily observed in tumour biopsies, fulfilling
one of the requirements for fusion43.
Cancer cell fusion in vivo
Cancer cells fuse with many cell types in
vivo, including stromal cells44, epithelial
cells45 and endothelial cells46–48. There are
more than 30 reports of tumour cell fusion
with host cells and many of these implicate
macrophages or other BMDCs as host
fusion partners2–4,45,49–53. For example, when
the MDAY or A9 mouse sarcomas were
implanted in mice with allogeneic bone mar-
row transplants, hybrids between BMDCs
and tumour cells were generated51,54. Another
example was seen in the development of a
spontaneous melanoma metastasis to the
lungs in a Balb/c nude mouse52 (fIG. 1). Balb/c
mice are albino owing to a homozygous
mutation in tyrosinase (c/c), the rate-limiting
enzyme in melanogenesis. Although the
melanoma clone implanted into these mice
was genetically wild-type for tyrosinase
(C/C), the cells produced little or no melanin
in culture and formed amelanotic tumours in
mice. Metastases, though infrequent, were
generally small, amelanotic tumours in the
lung, and were well tolerated by the mice53.
However, in one experiment a mouse devel-
oped a melanin-producing in transit meta-
stasis near the site of implantation in the tail
dermis (fIG. 1a). Because of this the tail was
amputated and the mouse was followed to see
if distant metastases developed. After 5 weeks
the mouse became moribund with a massive,
highly pigmented pulmonary metastasis
(fIG. 1c). DNA analyses showed that cells from
the metastasis had a genotype of C/c, indicat-
ing they were hybrids formed from fusion of
the implanted tumour cells (C/C) with host
cells (c/c). Cells from the metastasis showed
an average 30–40% increase in DNA content,
increased chemotaxis in vitro, activation
of N-acetylglucosaminyltransferase v
(GnT-v, MGAT5, E.C.184.108.40.206), and pro-
duction of β1,6-branched oligosaccharides
(see below). They also produced ‘coarse
melanin’ — autophagosomes containing
melanosomes and other organelles (below).
Small numbers of highly melanized, coarse
melanin-producing cells were found within
the original implanted tumour (fIG. 1b). These
were not present in the cultured parental
melanoma cells and were thus generated
in vivo52. Morphologically identical cells were
cultured from the metastasis and determined
to be C/c hybrids with host cells, indicating
that fusion and hybridization had occurred
in the original implant. Histopathology stud-
ies of the original implant revealed that it was
infiltrated with macrophages, supporting the
possibility that macrophage–tumour fusion
had occurred there.
BMDCs in human cancer and stem cell-
like distribution patterns. The first and, as
yet, sole confirmation of BMDC–tumour
cell fusion in humans has been reported.
Transcriptionally active malignant nuclei and
normal nuclei were observed in tumour-
associated osteoclasts from myeloma patients.
In the osteoclast population, 30% of the nuclei
were of malignant-cell origin, indicating a
remarkably high incidence of osteoclast–
tumour cell fusion55. The potential relevance
of this finding to myeloma pathobiology is
not yet known. Other studies have demon-
strated the presence of donor genes in carci-
noma cells of secondary malignancies arising
after allogeneic haematopoietic stem cell
(HSC) transplant; however, for largely techni-
cal reasons, definitive proof for or against
donor–host fusion was lacking in each. In
the first reported case, a renal cell carcinoma
(RCC) developed in a child following an HSC
transplant from his cancer-free brother56. A
lymph node metastasis of this tumour (the
only tissue available) was analysed by laser
capture microscopy of tumour cells and
PCR-based analyses for donor genes (fIG. 2).
Carcinoma cells throughout the tumour con-
tained the donor-specific A allele of the ABO
378 | MAY 2008 | vOLUME 8
© 2008 Nature Publishing Group
Nature Reviews | Cancer
blood group, indicating that HSCs had in
some manner become incorporated into the
tumour. The patient history of radiation and
immunosuppression before HSC transplant
increased the likelihood that the tumour
arose de novo in the patient and that donor
BMDCs became incorporated through fusion
with pre-existing tumour cells. However,
because a suitable patient-specific DNA
sequence was unavailable, evidence for donor
and patient genes in the same cells was lack-
ing56. Nonetheless, carcinoma cells through-
out the tumour produced β1,6-branched
oligosaccharides (fIG. 2), a risk factor in
several cancers and a characteristic of other
BMDC–tumour cell hybrids, as discussed
below. In the second case57, tumour cells from
a primary papillary RCC (PRCC) arising after
a male-to-female HSC transplant were found
to exhibit a trisomy 17, a common abnormal-
ity in PRCC and other cancers58 (fIG. 3). About
1% of the trisomy 17-containing tumour cells
also contained the donor Y chromosome in
the same nucleus57. As above, this combined
with the patient history suggested that fusion
had occurred between tumour cells and
donor HSC cells after development of the
tumour57. However, the possibility that the
tumour was derived solely from a donor
HSC, without fusion, followed by growth
and widespread loss of the Y was not ruled
out59. Nonetheless, it is of note that the Y- and
trisomy 17-containing carcinoma cells were
distributed within the tumour in pairs and
clusters resembling post-mitotic daughter
cells — a pattern that would be predicted for
cancer stem cells19,60. Also, Y-containing car-
cinoma cells were localized to a region cover-
ing only about 10% of the tumour, suggesting
a clonal emergence of these cells. Supporting
this, Y-containing carcinoma cells differed
from the majority of carcinoma cells in this
tumour through their high expression of
β1,6-branched oligosaccharides (fIG. 3). In
other reports, Y-containing cancer cells were
found in two cases of intestinal adenoma and
one case of lung cancer in females who had
previously received male HSC transplants61.
XY fluorescence in situ hybridization of a
limited number of these cells revealed no
evidence of the XXY or XXXY cells that
could have supported (but not proven) the
presence of BMDC–tumour hybrids. The
authors proposed that some BMDCs come
to resemble cancer cells through “develop-
ment mimicry” rather than being “direct
seeds of the cancer”61. However, in the case
above57, as the donor Y chromosome was
present in the same cells with a trisomy 17 it
seems unlikely that HSC donor cells could
have acquired this aneuploid karyotype
Figure 1 |?Spontaneous?in vivo?fusion?in?melanoma52.?Cells from a clone of the Cloudman S91
mouse melanoma were implanted subcutaneously in the tail of a Balb/c nu/nu mouse. The mice
were albino due to a homozygous mutation in tyrosinase (c/c), the rate-limiting enzyme in
melanogenesis. Although the melanoma clone was genetically wild-type for tyrosinase (C/C),
the cells produced little or no melanin in culture and formed amelanotic tumours in mice.
Metastases, though infrequent, were generally small, amelanotic tumours in the lung, and were
well tolerated by the mice53. in one experiment (designed for other purposes), what appeared
to be a melanin-producing in transit metastasis developed (a, arrow) near the site of implant
(bracket). The tail was amputated and the implanted tumour was formalin-fixed, embedded in
paraffin and sectioned serially. Small numbers of highly melanized, coarse melanin-producing
cells were found within the implanted tumour that were not seen in cultures of the parental
melanoma cells and had thus been generated in vivo (b, arrows). Five weeks after removal of the
tail the mouse became moribund with a massive, highly pigmented pulmonary metastasis (c,
asterisk). Cells from the metastasis were cloned in soft agar. DNA analyses revealed that 12 of
12 randomly picked clones had a genotype of C/c, indicating they were hybrids formed from
fusion of the implanted tumour cells (C/C) with host cells (c/c). Cells from the metastasis showed
an average 30–40% increase in DNA content, increased chemotaxis in vitro, activation of the
glycosyltransferase GnT-v, and production of its enzymatic product, β1,6-branched oligo-
saccharides52. Like the pigmented cells found in the primary implant (b), they also produced
‘coarse melanin’ — autophagosomes containing melanosomes and other organelles52. Similar
cells were cultured from the metastasis and were also seen in histolopathology sections of the
pulmonary tumour. This indicated that the coarse melanin-containing cells originated in the
primary implant through host–tumour cell fusion(s). Coarse melanin was also observed in
another in vivo melanoma hybrid ‘PADA’84 and in experimentally fused macrophage–melanoma
hybrids and is a common characteristic of human melanomas159. reproduced, with permission,
from Ref. 52 American Association for Cancer research (2000).
NATURE REvIEWS | cancer?
vOLUME 8 | MAY 2008 | 379
© 2008 Nature Publishing Group
Nature Reviews | Cancer
simply through mimicry of carcinoma cells,
as mimicry would presumably not include
genetic aberrations61. In another study of
secondary solid tumours following female-
to-male HSC transplants, tumour cells were
found with two X chromosomes but no Y,
suggesting they originated at least in part
from the female donor BMDCs62. However,
this study did not report on the potential
presence of XXXY (tetraploid) or XXY
(aneuploid) cells that might have been indica-
tors of BMDC–tumour cell fusion50,61,62, nor
did it rule out the widespread loss of the Y
chromosome that occurs in many cancers as
an explanation for the XX karyotype of some
carcinoma cells50,59. Nonetheless, as above57
the XX tumour cells tended to be in clusters,
suggesting a stem cell-like pattern within the
In summary, although host cell–cancer
cell fusion has been demonstrated in ani-
mals, there is as yet far less information in
human cancer. HSCs have been shown to
incorporate into human cancers; however,
the mechanisms of incorporation — fusion
versus direct transformation — remain to
be elucidated. In the limited number of
cases so far, some of the HSCs incorporated
into human solid tumours showed a clonal
distribution pattern that might be expected
for cancer stem cells, consistent with a recent
proposal that BMDC–tumour cell fusion is a
potential source of cancer stem cells19.
Cancer cell fusion and the hybrid pheno-
type. Fusion-induced enhancement of
metastasis and a differentiated trait such
as melanin production is in contrast to
previous studies in which hybrids that
were formed in vitro between normal epi-
thelial cells or fibroblasts and tumorigenic
cancer cells were generally suppressed in
tumorigenicity compared with the parental
cancer cells63–69, with some exceptions70,71.
These important observations led to the
concept of, and subsequent identification
of, a number of different tumour suppres-
sor genes that have been largely involved
in control of progression through the cell
cycle69. Differentiated traits were also sup-
pressed in such hybrids. For example, poly-
ethylene glycol- and Sendai virus-induced
hybrids between fibroblasts and pigmented,
tumorigenic melanoma cells were non-
pigmented and non-tumorigenic72–76. The
tendency of hybrids to lose chromosomes
with successive cell divisions was exploited
for chromosomal mapping of suppressor
genes. However, when healthy leukocytes
were used as fusion partners with cancer
cells, co-activation of differentiated func-
tions between parental genomes was seen,
for example, in leukocyte–hepatoma
hybrids77,78, leukocyte–myeloma hybrids79,
and macrophage–melanoma hybrids
discussed herein. Thus, unlike tumour-
suppressive fibroblasts and epithelial cells,
haematopoietic cells enhanced malignancy
and differentiation when hybridized with
transformed cells. Expression of genes from
both parental lineages in cancer cell hybrids
could explain many properties of metastatic
cells3,4. For example, tropism to lymph
nodes and organs and tissues such as bone
marrow, brain, lung and liver is a common
trait of macrophages and metastatic cells
alike. Likewise, the notorious multidrug
resistance of malignant cells to chemother-
apy owing to high levels of p-glycoprotein81
could reflect the fact that macrophages also
express this phenotype82.
Tumour–BMDC fusions might explain
how common gene expression patterns
emerge for different tumour types.
We, and others, have found that when
BMDC–tumour cell hybrids were isolated
in vitro with no selective pressure other
than for growth in drug-containing media,
remarkably high numbers of them exhibited
a metastatic phenotype in mice. Of 75 clones
of polyethylene glycol-fused macrophage–
melanoma hybrids isolated in vitro, about
half showed increased chemotaxis in vitro
and metastasis in mice53,83,84. Similar results
were obtained in T-cell hybridomas from
the fusion of healthy T lymphocytes with
T lymphoma cells85, and in hybrids between
mouse T-cell lymphoma cells and bone
marrow-derived macrophages or spleen
lymphocytes86,87. High-frequency emergence
of a common metastatic phenotype in vitro
without host-selective pressure was surpris-
ing, particularly in view of the apparently
chaotic nature of aneuploidy. In fact, little is
Figure 2 |?a?renal?cell?carcinoma?arising?after?allogeneic?stem?cell?transplant.?These samples
were taken from a lymph node metastasis arising in a boy after receiving a haematopoietic stem cell
(HSC) transplant from his cancer-free brother56. Tumour cells throughout the metastasis contained the
ABO blood group A allele of the HSC donor, suggesting that the cells were hybrids between donor
HSC(s) and patient tumour cell(s), although direct transformation of bone marrow-derived cells into
tumour cells was not ruled out. The sections were stained by lectin histochemistry for β1,6-branched
oligosaccharides with the plant lectin leukocytic phytohaemagglutinin (LPHA), which exhibits high
specificity for β1,6-branching on N-glycoproteins. β1,6-branched oligosaccharides were present in
cells throughout the tumour, consistent with the wide distribution of the donor A allele56. a | Adjacent
lymphocytes in the same sections were negative for LPHA staining. b | Higher power revealed that
LPHA stained in a coarse vesicular, autophagosome-like pattern similar to that seen with coarse
melanin in macrophage-melanoma hybrids (fIG. 1). c,d | Low-power fields demonstrating homogene-
ous staining of tumour cells for β1,6-branched oligosaccharides. A similar coarse vesicular,
autophagosome-like staining pattern for β1,6-branched oligosaccharides is widespread in human
cancer159. reproduced, with permission, from Ref. 56 Nature Publishing Group (2004).
380 | MAY 2008 | vOLUME 8
© 2008 Nature Publishing Group
Nature Reviews | Cancer
known of the regulation of gene expression
in hybrids at the molecular level. Evidence
that BMDC–tumour hybrids express many
of the same genes associated with invasive
and metastatic cancers and that these genes
are also expressed by macrophages and other
migratory BMDCs is summarized below.
SPARC. SPARC (also known as osteonectin
and BM40) is a modulator of cell–matrix
interactions during development and is a
key component of wound healing, tissue
repair and hard-tissue formation88,89. SPARC
modulates cellular shape and as such is a
counter-adhesive factor89. SPARC binds to
several proteins of the extracellular matrix
and is also a chaperone aiding proper folding
of collagen in the endoplasmic reticulum90.
In development, SPARC is expressed in
late gastrulation during differentiation of
invaginated epithelial cells into mesoderm91.
Interestingly, SPARC is important in osteo-
clast formation92,93. In tissue macrophages
SPARC is expressed in regions of neovascu-
larization, for example, in wound repair94 and
degenerative aortic stenosis95. High SPARC
expression is associated with tumour
progression and poor outcome in melanoma
and a number of carcinomas including breast,
colorectal, ovarian and lung96. SPARC acts as
a regulator of melanoma EMT by downregu-
lating melanoma E-cadherin (also known as
CDH1) with loss of homotypical adhesion,
and stimulates motility and increases expres-
sion of mesenchymal markers such as matrix
metalloproteinase MMP9 (Ref. 97). The actions
of SPARC are mediated through SNAIL, a
transcription factor in the initiation of EMT
during normal development and cancer98.
The SPARC gene provides an example
of gene regulation in BMDC–tumour
fusion. In fusions between mouse macro-
phages or human blood monocytes and
weakly metastatic mouse Cloudman S91
melanoma cells, unfused melanoma cells,
macrophages and monocytes all expressed
SPARC mRNA; however, the levels were
3–4-fold higher per µg total RNA in
hybrids17,99. SPARC mRNA levels were high-
est in hybrids of high metastatic potential
and lowest in weakly metastatic hybrids and
parental melanoma cells. Moreover, hybrids
between human monocytes and mouse
melanoma cells expressed both human and
mouse SPARC mRNA17. This indicated
that genomes from cells of the two dif-
ferent developmental lineages were both
activated. Thus, for SPARC, gene expression
was enhanced by hybridization of tumour
cells with macrophages, high expres-
sion was correlated with high metastatic
potential, and SPARC mRNA was produced
in hybrids from the genomes of both
parental fusion partners. That increased
SPARC expression was a characteristic of
macrophage–melanoma hybrids provides a
possible explanation for increased SPARC
and SPARC-mediated pathways in human
melanoma and other cancers. It is not
known whether other regulators of EMT
and development in addition to SPARC
were expressed in macrophage–tumour
cell fusion hybrids (transcription factors
TWIST, SNAIL and others)10,14. However
at least one, TWIST, is activated in
macrophages and regulates inflammatory
cytokine production100,101. By analogy to
SPARC, this suggests that TWIST expres-
sion in some invasive carcinomas reflects
expression of macrophage-lineage genes
following macrophage–tumour cell fusion.
Figure 3 |?Tumour?β1,6-branched?oligosaccharides?after?allogeneic?stem?
cell?transplant.?The papillary renal cell carcinoma (PrCC) arose in the kid-
ney of a female 2 years after she had received a male haematopoietic stem
cell transplant from her cancer-free 15 year-old son57. Karyotypes revealed
that some of the carcinoma cells contained a trisomy 17, a common abnor-
mality for PrCC. Cells containing both the Y and ≥3 copies of chromosome
17 were localized to a small region covering about 10% of the section area
where they comprised about 1% of the tumour cells. These cells were puta-
tive fusion hybrids between bone marrow-derived cells (BMDCs) and carci-
noma cells, although direct transformation of BMDCs in carcinoma cells
without fusion was not definitively ruled out57. Sections were stained with
leukocytic phytohaemagglutinin (LPHA), a selective marker for β1,6-
branched oligosaccharides. LPHA-positive cells were photographed, and the
sections were processed by fluorescence in situ hybridization (FiSH) for the
Y (red) and 17 (green) chromosomes. a–e | Left: LPHA-positive carcinoma
cells. right: FiSH analyses of the same section for the Y and 17. Arrows
show cells containing both the Y and trisomy 17, demonstrating the pres-
ence of donor genes in the carcinoma nuclei. Asterisks denote Y-containing
carcinoma cells in both left and right panels. f | A region that was devoid of
LPHA-positive cells. Left: LPHA-negative carcinoma cells. right: a FiSH-
labelled sequential section of the same region displaying only chromosome
17 and not the Y. Of the 70 LPHA-positive cells studied in this manner, 46
nuclei gave positive FiSH signals, and of these 37 (80%) contained a
Y chromosome. The majority of tumour cells were LPHA-negative and dis-
played 17 but not the Y 3,4,57. Thus, tumour-incorporated BMDCs were the
main source of tumour cell-associated β1,6-branched oligosaccharides for
this tumour. As with other such cases discussed herein and for a wide
number of human cancers159, staining for β1,6-branched oligosaccharides
revealed a coarse vesicular phenotype (for example, e, left). reproduced,
with permission, from Ref. 57 Nature Publishing Group (2005).
NATURE REvIEWS | cancer?
vOLUME 8 | MAY 2008 | 381
© 2008 Nature Publishing Group
MCR1 and MET. The melanocortin 1
(MC1, melanocyte-stimulating hormone)
receptor (MC1R) is activated by MC1 in
healthy melanocytes and melanoma cells
in which, through cyclic AMP-dependent
mechanisms, it activates melanogenesis and
regulates proliferation along with several
other actions102,103. MC1R appears to have
a role in melanoma progression, at least
in part through its activation of the proto-
oncogene MET, whose signalling pathway
is a key regulator of metastasis in melanoma
and many other cancers104–106.
As with SPARC, gene expression for
both MC1R and MET was increased in
highly metastatic macrophage–melanoma
hybrids107,108. Moreover, each was involved
in the induction of chemotactic motility in
hybrids83,107. Upregulated MC1R mRNA
expression in hybrids was associated with
increased cellular binding of its ligand MC1,
and amplified responsiveness to MC1, as
shown by increased chemotactic motility,
dendricity and melanization83,84. Exposure
of hybrids to MC1 also increased both the
production of MET mRNA and responsive-
ness to hepatocyte growth factor (HGF) as
a chemoattractant108. Thus the MC1–MC1R
and HGF–MET pathways appeared to act
together in a positive autocrine loop to
control chemotaxis and other functions in
hybrid cells. This same relationship appears
to be operative in malignant melanoma105.
In melanoma, MET and MC1R are each
regulated through the master transcription
factor microphthalmia-associated transcrip-
tion factor (MITF)104, which itself is associ-
ated with tumour progression109. Although
it was not determined whether MITF was
upregulated in experimental macrophage–
melanoma hybrids, this appears to have been
the case, as levels of the mRNAs for both
MET and MC1R were increased, an expected
consequence of increased MITF104,109,110.
High expression of MITF111, MET112,113
and MC1R104,114–117 are all characteristics of
monocytes, macrophages and other BMDCs.
GnT-V and β1,6-branched oligosaccharides.
GnT-v is a Golgi complex enzyme that is
highly expressed in myeloid cells and meta-
static cancer cells. GnT-v and its enzymatic
products, β1,6-branched oligosaccharides
conjugated to N-glycoproteins, are associ-
ated with poor outcome in melanoma43,118
and carcinomas of the breast119,120, colon121,122
lung123 and endometrium124. β1,6-branched
oligosaccharides were first purified from
granulocytes125. From structural analyses
they are composed of poly-N-acetyllactose
amines that are carriers of sialyl lewisx
antigen (sialyl lex) and thereby used by both
leukocytes and metastatic cancer cells for
binding to E-selectin (SELE) and/or galectin 3
(also known as lectin, galactoside-binding,
soluble 3 (LGALS3)) on endothelial cells
during systemic migration125,126.
GnT-v mRNA, protein and/or enzymatic
activity were increased in highly metastatic
macrophage–melanoma hybrids in vitro127,
and in host–tumour fusions in both lympho-
mas and melanomas growing in mice51,128,129.
In human cancer, β1,6-branched oligosac-
charide production was a characteristic of
putative BMDC–tumour hybrids in the
two RCCs discussed above that developed
after allogeneic HSC transplant56,57 (fIGs 2,3).
Moreover, multiple pathways in invasion
and metastasis that are regulated by GnT-v
were increased in macrophage–melanoma
hybrids, such as motility-associated
integrin subunits, cell surface expression of
lysosomal-associated membrane protein 1
(LAMP1) and autophagy.
Motility-associated integrins. The integrin
subunits α2, α3, α5, α6, αv, β1 and β3 are all
involved with migration of leukocytes and
cancer cells. These same integrin subunits
were significantly upregulated at the protein
level in metastatic macrophage–melanoma
hybrids compared with weakly metastatic
hybrids and parental melanoma cells127
(J.M.P. and A.K.C., unpublished data).
Following stimulation with MC1, protein
levels were further increased in highly
metastatic hybrids. These results correlated
with findings that metastatic hybrids had
acquired an MC1-inducible chemotactic
phenotype that was directed toward
fibronectin (FN1) through the action of
integrin α5β1 (Ref. 83). Of great interest, all
the above subunits have been identified as
substrates for GnT-v and their actions are
strongly affected by their glycosylation status
with β1,6-branched oligosaccharides129–139.
For example, in human fibrosarcoma cells
addition of β1,6 branched oligosaccharides
onto the β1 integrin subunit by GnT-v
reduced α5β1 integrin clustering and stim-
ulated cell migration139. Further, the above
integrin subunits are each involved in
metastasis. Levels of the α3β1 integrin are
increased and associated with increased
migration and invasion in several types
of metastatic cancers140. α5β1 is a well-
characterized receptor for fibronectin
that is overexpressed in metastasis141–143.
Upregulation of αvβ3, a vitronectin
(vTN) receptor, was described in various
cancers including malignant melanoma
Expression of the β1-integrin subunit is a
key component of melanoma metastasis147.
The above integrins and integrin subunits
are also highly expressed in macrophages,
in which they are involved with many func-
tions, including cell adhesion and migration,
signal transduction, cell–cell recognition
Cell surface expression of LAMP1. LAMP1
is a preferred substrate for GnT-v and
therefore a major carrier of sialyl lex and
poly-N-acetyllactose amines that bind to
E-selectins and galectins130. Cell surface
LAMP1 thus mediates binding to endothe-
lial cells by both leukocytes and cancer
cells153–155. Macrophage–melanoma hybrids
showed increased expression of cell surface
LAMP1 (Ref. 127). This was seen in highly
metastatic macrophage–melanoma hybrids
as well as peritoneal macrophages compared
with that in parental melanoma cells and less
Autophagy and course melanin. As
mentioned, the spontaneous mouse
melanoma–host hybrid described above
showed a high level of autophagy and coarse
melanin52 (fIG. 1). This was also a characteris-
tic of another spontaneous melanoma–host
hybrid described previously (‘PADA’)84 and
of macrophage–melanoma hybrids fused
in vitro52,53,84. Electron microscope studies
revealed that melanin was localized largely to
heavily melanized melanosomes packaged in
autophagosomes. Autophagosomes were ver-
ified by the presence of double limiting mem-
branes and heterogeneous morphologies.
They were also strongly positive for β1,6-
branched oligosaccharides, implicating a role
for GnT-v in their formation156–158. These
were surprising findings because healthy
melanocytes do not appear to use GnT-v in
melanogenesis and the melanosomes are not
packaged in autophagosomes but exist singly
in the cytoplasm. That several independently
isolated melanoma hybrids all showed high
levels of autophagy and coarse melanin
raised the question as to whether this trait
might be a signature of BMDC–melanoma
fusion in human melanoma. Although coarse
melanin in melanoma had been known to
pathologists for more than a century and was
shown to be due to autophagy (reviewed in
Ref. 159), its frequency in human cancers
had not been evaluated160. Analyses of several
hundred cases have revealed that it is a
common trait, expressed by 85% or more of
melanomas43,159. It was further determined
that coarse melanin-producing melanoma
cells and melanophages (macrophages with
382 | MAY 2008 | vOLUME 8
© 2008 Nature Publishing Group
Nature Reviews | Cancer
Membrane apposition and fusion
Deregulated cell cycle
autophagolysosomal vesicles containing
undigested melanin) account for the well-
known hypermelanotic regions of cutane-
ous malignant melanoma used in clinical
diagnosis43,159. As in macrophage–melanoma
hybrids, coarse melanin vesicles in human
melanomas contained β1,6-branched oligo-
saccharides43,159. In cutaneous malignant
melanoma, β1,6-branched oligosaccharide-
positive, coarse melanin-producing
melanoma cells emerge clonally as ‘nests’
within the in situ tumour and have the capac-
ity for invasion into the dermis43, 159. This is
consistent with BMDC–tumour cell fusion
as an explanation for the appearance of
these cells (fIG. 1b). Moreover, β1,6-branched
oligosaccharide-positive coarse vesicles with-
out melanin were common in all 22 types of
human cancers studied and predicted worse
outcome in primary breast carcinomas119,159.
Although it is not certain that the coarse
vesicular structures seen in other neoplasms
were always due to autophagy, it nonetheless
suggested that high levels of autophagy might
be widespread, if not universal, in cancer.
This was supported by separate molecular
genetic studies also indicating that high
levels of autophagy are common in cancer,
in which they are associated with tumour
survival and progression161–166. This could
seem counterintuitive as autophagy has long
been thought to be a catabolic event associ-
ated with cell death. However, more recent
evidence indicates that autophagy can act as
a pro-survival factor by producing a useable
energy source for cancer cells deprived of
an adequate blood supply. Thus autophagy
might help drive metastatic progression
where cells can produce nutrients distant
from the primary tumour and its nutrient
Whether through BMDC–melanoma
cell fusion or some other mechanism, the
generation of β1,6-branched oligosaccharide-
positive coarse melanin appears to account in
part for the well-known immunogenicity of
malignant melanoma. These highly melan-
ized melanoma cells are immunogenic and
attractive to macrophages43. One immune
escape mechanism appears to involve the
generation of variant tumour cells that no
longer attract macrophages, for example
through loss of melanin production and
generation of amelanotic variants176. In cuta-
neous malignant melanoma, dermal nests
of melanophage-free melanoma cells with
reduced or absent melanin were nonetheless
positive for β1,6-branched oligosaccharides
and associated with worse patient outcome43.
Could autophagy in human cancer
result from fusions between cancer cells
and macrophages or other phagocytes? In
fact, macrophages express active autophagy
as a part of the pathway for digestion of
phagocytosed microorganisms and cells167,168.
Autophagy in macrophages is linked to
phagocytosis, interestingly, another charac-
teristic of metastatic cancers169–173. Moreover,
macrophage vesicles, like those in experi-
mental macrophage–melanoma hybrids and
cancer cells, are positive for β1,6-branched
oligosaccharides118,119,159. Therefore, activa-
tion of phagocytic and autophagic pathways
in human cancers could reflect expression
of imprinted genes of myeloid lineage in
macrophage–tumour cell fusion hybrids. We
suggest that, should cancer cell autophagy
be linked to phagocytosis as it is in macro-
phages, nutrients could be continuously
phagocytosed from external sources and
digested through autophagy, rendering
metastatic cells constitutively independent of
a direct blood supply.
In summary, metastatic macrophage–
melanoma hybrids show high expression of
SPARC, MET, MC1R, integrin subunits α3,
α5, α6, αv, β1, β3, cell-surface LAMP1 and
GnT-v and high levels of autophagy. This
is paralleled in melanoma, and in a number
of other cancers in which these molecules
are associated with a migratory phenotype,
enhanced survival, metastasis and poor
outcome. Central to the metastatic pheno-
type is GnT-v which, through addition of
β1,6-branched oligosaccharides to several of
the above proteins, causes many phenotypic
changes, including increased chemotaxis,
melanogenesis and possibly autophagy.
Expression of MC1R, MITF, MET, motil-
ity-related integrins, cell-surface LAMP1
and GnT-v, and high levels of autophagy
are also characteristic of monocytes and
macrophages and other BMDCs. Thus,
expression of these molecules in cancer
could be a result of fusion of cancer cells
with migratory BMDCs and co-expression
of imprinted genes from both parental fusion
partners. Although these molecules and traits
are of course not the only factors involved
in tumour progression, their high expres-
sion in BMDC–tumour hybrids provides a
framework for understanding how fusion can
explain metastasis (fIG. 4).
Problems and pitfalls
To prove fusion and genomic hybridiza-
tion requires identification of genes or
chromosomes from both of the putative
fusion partners in the same cell or cells.
Hence, fusion has been well-documented
in tumour xenografts in animals where
hybrids were identified by the presence of
both tumour and host genes. Little is yet
known of the extent of cancer cell fusion in
humans. Although a few human cases have
recently been reported55–57,61,62, only one
of these, involving macrophage–myeloma
Figure 4 | a?model?for?generation?of?a?meta-
noma?cell?with?a?macrophage.?a | A macrophage
is attracted to a non-migratory melanoma cell
in situ. The epigenomes of the two cells reflect
their myeloid and melanocytic lineages respec-
tively. The melanoma cell produces ‘fine’ or ‘dusty’
melanin — individual melanosomes in the cyto-
plasm, generally with a golden-brown colour.
Melanoma-associated macrophages are known
as melanophages because they are laden with
autophagolysosomal vesicles containing residual
melanin from engulfed and digested melanoma
cells, and thus at times are difficult to distinguish
from melanoma cells at the light microscope
level43,159,160. b | The macrophage and melanoma
plasma membranes form close appositional con-
tacts, normally as a prelude to ingestion and
destruction of the melanoma cell43. However in
some cases the two cells fuse. c | Following fusion
a heterokaryon is formed with the two nuclei
separate in the cytoplasm. d | Genomic hybridiza-
tion occurs and a mononuclear macrophage–
melanoma hybrid emerges. From studies of
macrophage–melanoma hybrids generated
experimentally in vitro and of melanoma–host
hybrids generated spontaneously in mice, such
hybrids have a deregulated cell cycle, are aneu-
ploid and exhibit epigenomes of both parental
lineages. Some exhibit the myeloid capability for
chemotaxis in vitro and tropism in vivo, common
characteristics of metastatic cells.
NATURE REvIEWS | cancer?
vOLUME 8 | MAY 2008 | 383
© 2008 Nature Publishing Group
fusion in osteoclast formation, definitively
proved fusion55. The use of myeloma clone-
specific immunoglobulin rearrangements
as parental markers of myeloma cells can
thus be used to further investigate questions
of fusion in myeloma55. Other studies have
suggested that incorporation of BMDCs
into tumour cells can occur through dif-
ferentiation or neoplastic transformation
without fusion61,62,174. It is possible that both
mechanisms are operative in cancer as well
as in healthy tissue regeneration and repair,
and this remains to be resolved. The use
of allogeneic HSC transplants in medicine
followed by the unfortunate development
of secondary malignancies provides a
potential source of pathology material
for study56,57. However, such cases are in
limited supply and it will take some time
to determine the extent of fusion in human
cancer by this technique alone. Another
problem is that the frequency of cancer cell
fusion may be low, as it is in culture (~1 in
105–107 non-fused cells), making fusion
events difficult if not impossible to follow
in vivo53. Also, depending on the time when
a particular tumour is analysed, the number
of hybrid cells could range from none,
should hybridization not have occurred,
to 100% if hybrids had overgrown a pre-
existing tumour or initiated a new tumour,
for example, a metastasis. Further, hybrid
cells in a tumour could result from a single
progenitor hybrid or from multiple hybrids
formed from separate fusions. It is thus dif-
ficult to study the molecular mechanisms
of cancer cell fusion in vivo, or to estimate
its frequency. Until more progress is made
in these and other areas, the effect of
BMDC incorporation into human tumours,
whether by fusion or other mechanisms,
remains to be determined.
Tumour cell–BMDC fusion as a source
of metastatic cells would imply that
prevention of fusion or of early, rate-
limiting post-fusion events might prevent
metastasis (for example see Ref. 175). With
better understanding should come better
strategies for targeting vulnerable steps
in fusion and the generation of hybrids.
Post-fusion events and hybrid formation
could present other fruitful areas of focus,
for example, molecular steps governing
the integration of parental fusion partner
genes into hybrid genomes, or those
involved with activation of master regula-
tory genes that are rate-limiting in the
development of a migratory phenotype.
Early post-fusion cells are also likely to
express unique antigenic profiles, making
them susceptible to immunotherapy.
The cancer cell–BMDC fusion theory
presents a unifying explanation for tumour
progression. It seems that this theory is
not only possible but likely to be correct to
at least some degree, with the remaining
question being how extensively does it con-
tribute to progression of human cancers?
In our opinion the theory deserves far
more attention from the cancer research
community than it currently receives.
Should cancer cell–BMDC fusion be
determined to drive tumour progression in
humans, surely new therapeutic strategies
John M. Pawelek and Ashok K. Chakraborty are at the
Department of Dermatology and the Yale Cancer
Center, Yale University School of Medicine, 333 Cedar
Street, New Haven, Connecticut 06520–08059, USA.
Correspondence to J.M.P.
Published online 3 April 2008
Aichel, O. in Vorträge und Aufsätze über
Entvickelungsmechanik Der Organismen Chapter XIII
(ed Roux, W.) 92–111 (Wilhelm Engelmann, Leipzig,
Pawelek, J. M. Tumour cell hybridization and metastasis
revisited. Melanoma Res. 10, 507–514 (2000).
Pawelek, J. M. Tumour-cell fusion as a source of
myeloid traits in cancer. Lancet Oncol. 6, 988–993
Pawelek, J. et al. Co-opting macrophage traits in
cancer progression: a consequence of tumor cell
fusion? Contrib. Microbiol.13, 138–155 (2006).
Mekler, L. B. [A general theory of oncogenesis.]
Materials of Symposia on General Immunology. The
Club of Immunologists of NF Gamaleya Inst of
Epidemiology and Microbiology 3, 91–100 (1968)
Mekler, L. B. [Hybridization of transformed cells with
lymphocytes as 1 of the probable causes of the
progression leading to the development of metastatic
malignant cells.] Vestn Acad. Med. Nauk. SSR (Bulletin
of the USSR Acad Med Sci) 26, 80–89 (1971)
Goldenberg,.DM. [On the progression of malignity: a
hypothesis.] Klin. Wschr. 46, 898 (1968) (in German).
Goldenberg, D. M. & Gotz, H. On the ‘human’
nature of highly malignant heterotransplantable
tumors of human origin. Europ. J. Cancer 4, 547–548
Lagarde, A. E. & Kerbel, R. S. Somatic cell
hybridization in vivo and in vitro in relation to the
metastatic phenotype. Biochim. Biophys. Acta 823,
10. Gupta, P. B., Mani, S., Yang, J., Hartwell, K. &
Weinberg, R. A. The evolving portrait of cancer
metastasis Cold Spring Harb. Symp. Quant. Biol. 6,
11. Chambers, A. F., Groom, A. C. & MacDonald, I, C.
Dissemination and growth of cancer cells in metastatic
sites. Nature Rev. Cancer 2, 563–572 (2002).
12. Nowell, P. C. The clonal evolution of tumor cell
populations. Science 194, 23–28 (1976).
13. Fidler, I. J. & Kripke, M. L. Metastasis results from
preexisting variant cells within a malignant tumor.
Science 197, 893–895 (1977).
14. Ma, L., Teruya-Feldstein, J. & Weinberg, R. A. Tumour
invasion and metastasis initiated by microRNA-10b in
breast cancer. Nature 449, 682–688 (2007).
15. Chakraborty, A. K. & Pawelek, J. M. GnT-V,
macrophages, and cancer metastasis: A common link.
Clin. Exp. Metastasis 20, 365–373 (2003).
16. Munzarova, M., Lauerova, L. & Capkova, J. Are
advanced malignant melanoma cells hybrids between
melanocytes and macrophages? Melanoma Res. 2,
17. Chakraborty. A. K., de Freitas Sousa, J., Espreafico,
E. M. & Pawelek, J. M. Human monocyte × mouse
melanoma fusion hybrids express human gene. Gene
275, 103–106 (2001).
18. Duelli, D. & Lazebnik, Y. Cell-to-cell fusion as a link
between viruses and cancer. Nature Rev. Cancer 7,
19. Bjerkvig, R., Tysnes, B. B., Aboody, K. S., Najbauer, J.
& Terzis, A. J. Opinion: the origin of the cancer stem
cell: current controversies and new insights. Nature
Rev. Cancer 5, 899–904. Erratum in Nature Rev.
Cancer 5, 995 (2005).
20. Sapir, A., Avinoam, O., Podbilewicz, B. &
Chernomordik, L. V. Viral and developmental cell
fusion mechanisms: conservation and divergence. Dev
Cell. 14, 11–21 (2008).
21. Chen, E. H. & Olson, E. N. Unveiling the mechanisms
of cell–cell fusion. Science 308, 369–373 (2005).
22. Vignery, A. Macrophage fusion: the making of
osteoclasts and giant cells. J. Exp. Med. 202,
23. Vignery, A. Macrophage fusion: are somatic and
cancer cells possible partners? Trends Cell Biol. 4,
24. Chen, E. H., Grote, E., Mohler. W. & Vignery, A. Cell–cell
fusion. FEBS Lett. 581, 2181–2193 (2007).
25. Pajcini, K. V., Pomerantz, J. H., Alkan, O., Doyonnas,
R. & Blau, H. M. Myoblasts and macrophages share
molecular components that contribute to cell–cell
fusion. J. Cell Biol. 180, 1005–1019 (2008).
26. Holmgren, L., Bergsmedh, A. & Spetz, A. L. Horizontal
transfer of DNA by the uptake of apoptotic bodies.
Vox Sang. 83 (Suppl. 1), 305–306 (2002).
27. Duelli, D. M. et al. A virus causes cancer by inducing
massive chromosomal instability through cell fusion.
Curr. Biol. 17, 431–437 (2007).
28. Larsson, L. I., Bjerregaard, B., Wulf-Andersen, L. &
Talts, J. F. Syncytin and cancer cell1 fusions.
Sci. World J. 7, 1193–1197 (2007).
29. Jin, J. & Woodgett, J. R. Chronic activation of protein
kinase Bβ/Akt2 leads to multinucleation and cell
fusion in human epithelial kidney cells: events
associated with tumorigenesis. Oncogene 24,
30. Overholtzer, M. et al. A nonapoptotic cell death
process, entosis, that occurs by cell-in-cell invasion.
Cell 131, 966–979 (2007).
31. Yagi, M., Miyamoto, T., Toyama, Y. & Suda, T. Role of
DC-STAMP in cellular fusion of osteoclasts and
macrophage giant cells. J. Bone Miner. Metab. 24,
32. Teitelbaum, S. L. & Ross, F. P. Genetic regulation of
osteoclast development and function. Nature Rev.
Genet. 4, 638–649 (2003).
33. Han, X. et al. CD47, a ligand for the macrophage fusion
receptor, participates in macrophage multinucleation.
J. Biol. Chem. 275, 37984–37992 (2000).
34. Kajita, M. et al. Membrane-type 1 matrix
metalloproteinase cleaves CD44 and promotes cell
migration. J. Cell Biol. 153, 893–904 (2001).
35. Kyriakides, T. R., et al. The CC chemokine ligand,
CCL2/MCP1, participates in macrophage fusion and
foreign body giant cell formation. Am. J. Pathol. 165,
36. Kim, M. S., Magno, C. L., Day, C. J. & Morrison, N. A.
Induction of chemokines and chemokine receptors
CCR2b and CCR4 in authentic human osteoclasts
differentiated with RANKL and osteoclast like cells
differentiated by MCP-1 and RANTES. J. Cell.
Biochem. 97, 512–518 (2006).
37. Marhaba, R. & Zöller, M. CD44 in cancer progression:
adhesion, migration and growth regulation. J. Mol.
Histol 35, 211–231 (2004).
38. Götte, M. & Yip, G. W. Heparanase, hyaluronan, and
CD44 in cancers: a breast carcinoma perspective.
Cancer Res 66, 10233–10237(2006).
39. Dalerba, P. et al. Phenotypic characterization of
human colorectal cancer stem cells. Proc. Natl Acad.
Sci. USA 104, 10158–10163 (2007).
40. Zijlmans, H. J. et al. The absence of CCL2 expression
in cervical carcinoma is associated with increased
survival and loss of heterozygosity at 17q11.2.
J. Pathol. 208, 507–517 (2006.
41. Baier, P. K., Eggstein, S., Wolff-Vorbeck, G.,
Baumgartner, U. & Hopt, U. T. Chemokines in human
colorectal carcinoma. Anticancer Res. 25, 3581–3584
42. Rendlew-Danielsen, J. M., et al. Dysregulation of
CD47 and the ligands thrombospondin 1 and 2 in
multiple myeloma. Br. J. Haematol. 138, 756–760
384 | MAY 2008 | vOLUME 8
© 2008 Nature Publishing Group
43. Handerson, T. et al. Melanophages reside in
hypermelanotic, aberrantly glycosylated tumor areas
and predict improved outcome in primary cutaneous
malignant melanoma. J. Cutaneous Pathol. 34,
44. Jacobsen, B. M., et al. Spontaneous fusion with, and
transformation of mouse stroma by, malignant human
breast cancer epithelium. Cancer Res. 66, 8274–8279
45. Rizvi, A. Z., et al. Bone marrow-derived cells fuse with
normal and transformed intestinal stem cells. Proc.
Natl Acad. Sci. USA 103, 6321–6325 (2006).
46. Mortensen, K., Lichtenberg, J., Thomsen, P. D. &
Larsson, L. I. Spontaneous fusion between cancer
cells and endothelial cells. Cell. Mol. Life Sci. 61,
47. Bjerregaard, B., Holck, S., Christensen, I. J. &
Larsson, L. I. Syncytin is involved in breast cancer-
endothelial cell fusions. Cell. Mol. Life Sci. 63,
48. Streubel, B. et al. Lymphoma-specific genetic
aberrations in microvascular endothelial cells in B-cell
lymphomas. N. Engl. J. Med. 351, 250–259 (2004).
49. Alison, M. R., Lovell, M. J., Direkze, N. C., Wright,
N. A. & Poulsom, R. Stem cell plasticity and tumour
formation. Eur. J. Cancer 42, 1247–1256 (2006).
50. Herzog, E. L., et al. Lung-specific nuclear
reprogramming is accompanied by heterokaryon
formation and Y chromosome loss following bone
marrow transplantation and secondary inflammation.
FASEB J. 21, 2592–12601 (2007).
51. Kerbel, R. S., Lagarde, A. E., Dennis, J. W. &
Donaghue, T. P. Spontaneous fusion in vivo between
normal host and tumor cells: possible contribution to
tumor progression and metastasis studied with a
lectin-resistant mutant tumor. Mol. Cell. Biol. 3,
52. Chakraborty, A. K. et al. A spontaneous murine
melanoma lung metastasis comprised of host × tumor
hybrids. Cancer Res. 60, 2512–2519 (2000).
53. Rachkovsky, M. S. et al. Melanoma × macrophage
hybrids with enhanced metastatic potential. Clin. Exp.
Metastasis 16, 299–312 (1998).
54. Wiener, F., Fenyö, E. M. & Klein, G. Tumor-host cell
hybrids in radiochimeras. Proc. Natl Acad. Sci. USA 7,
55. Andersen, T. L. et al. Osteoclast nuclei of myeloma
patients show chromosome translocations specific for
the myeloma cell clone: a new type of cancer-host
partnership? J. Pathol. 211, 10–17 (2007).
56. Chakraborty, A. et al. Donor DNA in a renal cell
carcinoma metastasis from a bone marrow transplant
recipient. Bone Marrow Transplant. 34, 183–186
57. Yilmaz, Y., Lazova, R., Qumsiyeh, M., Cooper, D. &
Pawelek, J. Donor Y chromosome in renal carcinoma
cells of a female BMT recipient: visualization of
putative BMT-tumor hybrids by FISH. Bone Marrow
Transplant. 35, 1021–1024 (2005).
58. Salama, M. E., Worsham, M. J. & DePeralta-Venturina,
M. Malignant papillary renal tumors with extensive
clear cell change: a molecular analysis by microsatellite
analysis and fluorescence in situ hybridization. Arch.
Pathol. Lab. Med.127, 1176–1181 (2003).
59. Lau, L. C., Tan, P. H., Chong, T. W., Foo, K. T. & Yip, S.
Cytogenetic alterations in renal tumors: a study of 38
Southeast Asian patients. Cancer Genet.
Cytogenet.175, 1–7 (2007).
60. Guo, W., Lasky, J. L. 3rd & Wu, H. Cancer stem cells.
Pediatr. Res. 59, 59R–64R (2006).
61. Cogle, C. R. et al. Bone marrow contributes to epithelial
cancers in mice and humans as developmental mimicry.
Stem Cells 25, 1881–1887 (2007).
62. Avital, I. et al. Donor-derived human bone marrow
cells contribute to solid organ cancers developing
after bone marrow transplantation. Stem Cells 25,
63. Wiener, F., Klein, G. & Harris, H. The analysis of
malignancy by cell fusion. J. Cell Sci. 15, 177–183
64. Stanbridge, E. J. Suppression of malignancy in human
cells. Nature 260, 17–20 (1976).
65. Sidebottom, E., The analysis of malignancy by cell
fusion. In Vitro 16, 77–86 (1980).
66. Ramshaw, I. A., Carlsen, S., Wang, H. &
Badenoch-Jones, P. The use of cell fusion to analyse
factors involved in tumour cell metastasis.
Int. J. Cancer 32, 471–478 (1983).
67. Harris, H. The analysis of malignancy by cell fusion:
the position in 1988. Cancer Res., 48, 3302–3306
68. Weinberg, A. S. Tumor suppressor genes. Science
254, 1138–1146 (1991).
69. Levine, A. J. In The Molecular Basis of Cancer.
(eds Mendelsohn, J., Howley, P. M., Israel, M. A. &
Liotta, L. A) 86–104 (WB Saunders, Philadelphia,
70. Scaletta, L. J. & Ephrussi, B. Hybridization of normal
and neoplastic cells in vitro. Nature 205, 1169
71. Defendi, V., Ephrussi, B., Koprowski, H. &
Yoshida, M. C. Properties of hybrids between
polyoma-transformed and normal mouse cells. Proc.
Natl Acad. Sci. USA 57, 299–305 (1967).
72. Jonasson, J., Povey, S. & Harris, H. The analysis of
malignancy by cell fusion. VII. Cytogenetic analysis of
hybrids between malignant diploid cells and of
tumours derived from them. J. Cell Sci. 24, 217–254
73. Davidson, R. L., Ephrussi, R. L. B. & Yamamoto, K.
Regulation of pigment synthesis in mammalian cells,
as studied by somatic hybridization. Proc. Natl Acad.
Sci. USA 56, 1437–1440 (1966).
74. Powers, T. P. & Davidson, R. L. Coordinate extinction
of melanocyte-specific gene expression in hybrid cells.
Som. Cell Mol. Gen. 22, 41–56 (1996).
75. Gourdeau, H. & Fournier, R. E.K, Genetic analysis of
mammalian cell differentiation. Ann. Rev. Cell Biol. 6,
76. Powers, T. P., Shows, T. B. & Davidson, R. L.
Pigment-cell-specific genes from fibroblasts are
transactivated after chromosomal transfer into
melanoma cells. Mol. Cell Biol. 14, 1179–1190
77. Darlington, G. J., Bernhard, H. P. & Ruddle, F. H.
Human serum albumin phenotype activation in mouse
hepatoma — human leukocyte cell hybrids. Science
185, 859–862 (1974).
78. Malawista, S. E. & Weiss, M. C. Expression of
differentiated function in hepatoma cell hybrids: high
frequency of induction of mouse albumin production in
rat hepatoma-mouse lymphoblast hybrids. Proc. Natl
Acad. Sci. USA 71, 927–931 (1974).
79. Giacomoni, D. Tumorigenicity and intracisternal
A-particle expression of hybrids between murine
myeloma and lymphocytes. Cancer Res. 39,
80. Kohler, G. & Milstein, C. Continuous cultures of fused
cells secreting antibody of predefined specificity.
Nature 256, 495–497 (1975).
81. Gottesman, M. M. & Ling, V. The molecular basis of
multidrug resistance in cancer: the early years of
P-glycoprotein research. FEBS Lett. 580, 998–1009
82. Lemaire, S., Van Bambeke, F., Mingeot-Leclercq, M. P.
& Tulkens, P. M. Modulation of the cellular
accumulation and intracellular activity of daptomycin
towards phagocytized Staphylococcus aureus by the
P-glycoprotein (MDR1) efflux transporter in human
THP-1 macrophages and madin-darby canine kidney
cells. Antimicrob. Agents Chemother. 51, 2748–2757
83. Rachkovsky, M. & Pawelek, J. Acquired melanocyte
stimulating hormone-inducible chemotaxis following
macrophage fusion with Cloudman S91 melanoma
cells. Cell Growth Differ. 10, 515–524 (1999).
84. Pawelek, J. et al. Altered N-glycosylation in
macrophage x melanoma fusion hybrids. Cell. Mol.
Biol. 45, 1011–1027 (2000).
85. Roos, E., La Riviere, G., Collard, J. G., Stukart, M. J. &
De Baetselier, P. Invasiveness of T-cell hybridomas
in vitro and their metastatic potential in vivo. Cancer
Res. 45, 6238–6243 (1985).
86. Larizza, L., Schirrmacher, V., Stöhr, M., Pflüger, E. &
Dzarlieva, R. Inheritance of immunogenicity and
metastatic potential in murine cell hybrids from the
T-lymphoma ESb08 and normal spleen lymphocytes.
J. Natl Cancer Inst. 72, 1371–1381 (1984).
87. Larizza, L. et al. Suggestive evidence that the highly
metastatic variant ESb of the T-cell lymphoma Eb is
derived from spontaneous fusion with a host
macrophage. Int. J. Cancer 34, 699–707 (1984).
88. Lane, T. F. & Sage, E. H. The biology of SPARC, a
protein that modulates cell-matrix interactions.
FASEB J. 8, 163–173 (1994).
89. Bradshaw, A. D. & Sage, E. H. SPARC, a matricellular
protein that functions in cellular differentiation and
tissue response to injury. J. Clin. Invest. 107,
90. Martinek, N., Shahab, J., Sodek, J. & Ringuette, M.
Is SPARC an evolutionarily conserved collagen
chaperone? J. Dent. Res. 86, 296–305 (2007).
91. Damjanovski, S., Huynh, M. H., Motamed, K., Sage,
E.H & Ringuette, M. Regulation of SPARC expression
during early Xenopus development: evolutionary
divergence and conservation of DNA regulatory
elements between amphibians and mammals. Dev.
Genes Evol. 207, 453–461 (1998).
92. Fujita, T. et al. SPARC stimulates the synthesis of
OPG/OCIF, MMP-2 and DNA in human periodontal
ligament cells. J. Oral Pathol. Med. 31, 345–352
(2002). Erratum in J. Oral Pathol. Med. 31, 504
93. Mansergh, F. C. et al. Osteopenia in Sparc
(osteonectin)-deficient mice: characterization of
phenotypic determinants of femoral strength and
changes in gene expression. Physiol. Genomics. 32,
94. Reed, M. J. et al. Differential expression of SPARC
and thrombospondin 1 in wound repair:
immunolocalization and in situ hybridization.
J. Histochem. Cytochem. 41, 1467–1477 (1993).
95. Charest, A. et al. Distribution of SPARC during
neovascularisation of degenerative aortic stenosis.
Heart 92, 1844–1849 (2006).
96. Robert, G. et al. SPARC represses E-cadherin and
induces mesenchymal transition during melanoma
development. Cancer Res. 66, 7516–7523 (2006).
97. Alonso, S. R. et al. A high-throughput study in
melanoma identifies epithelial–mesenchymal
transition as a major determinant of metastasis.
Cancer Res. 67, 3450–3460 (2007).
98. Barrallo-Gimeno, A. & Nieto, M. A. The Snail genes as
inducers of cell movement and survival: implications
in development and cancer. Development 132,
99. Chakraborty, A. K. & Yamaga, S. Differential gene
expression in genetically matched mouse melanoma
cells with different metastatic potential. Gene 315,
100. Sharif, M. N. et al. Twist mediates suppression of
inflammation by type I IFNs and Axl. J. Exp. Med.
203, 1891–1901 (2006).
101. Sosi, D., Richardson, J. A., Yu, K., Ornitz, D. M. &
Olson, E. N. Twist regulates cytokine gene expression
through a negative feedback loop that represses
NF-κB activity. Cell 112, 169–180 (2003).
102. Carlson, J. A., Linette, G. P., Aplin, A., Ng, B. &
Slominski, A. Melanocyte receptors: clinical
implications and therapeutic relevance. Dermatol.
Clin. 25, 541–557, viii–ix (2007).
103. Kanetsky, P. A. et al. Population-based study of
natural variation in the melanocortin-1 receptor gene
and melanoma. Cancer Res. 66, 9330–9337 (2006).
104. McGill, G. G., Haq, R., Nishimura, E. K. & Fisher, D. E.
c-Met expression is regulated by Mitf in the melanocyte
lineage. J. Biol. Chem. 281, 10365–10373 (2006).
105. Beuret, L. et al. Up-regulation of MET expression by
α-melanocyte-stimulating hormone and MITF allows
hepatocyte growth factor to protect melanocytes and
melanoma cells from apoptosis. J. Biol. Chem. 282,
106. Boccaccio, C. & Comoglio, P. M. Invasive growth:
a MET-driven genetic programme for cancer and stem
cells. Nature Rev. Cancer. 6, 637–645 (2006).
107. Chakraborty, A. K. et al. Upregulation of mRNA for
the melanocortin-1 receptor but not for melanogenic
proteins in macrophage x melanoma fusion hybrids
exhibiting increased melanogenic and metastatic
potential. Pig. Cell Res. 12, 355–366 (1999).
108. Chakraborty, A. K. et al. Expression of c-Met proto-
oncogene in metastatic macrophage × melanoma
fusion hybrids: implication of its possible role in MSH-
induced motility. Oncol. Res. 14, 163–174 (2003).
109. Levy, C., Khaled, M. & Fisher, D. E. MITF: master
regulator of melanocyte development and melanoma
oncogene. Trends Mol. Med. 12, 406–414 (2006).
110. Garraway, L. A. et al. Integrative genomic analyses
identify MITF as a lineage survival oncogene amplified
in malignant melanoma. Nature 436, 117–122
111. Bronisz, A. et al. Microphthalmia-associated
transcription factor interactions with 14–3–3
modulate differentiation of committed myeloid
precursors. Mol. Biol. Cell 17, 3897–3906 (2006).
112. Beilmann, M. et al. Neoexpression of the c-met/
hepatocyte growth factor-scatter factor receptor gene
in activated monocytes. Blood 90, 4450–4458
113. Gaasch, J. A., Bolwahnn, A. B. & Lindsey, J. S.
Hepatocyte growth factor-regulated genes in
differentiated RAW 264.7 osteoclast and
undifferentiated cells. Gene 369, 142–152 (2006).
NATURE REvIEWS | cancer?
vOLUME 8 | MAY 2008 | 385
© 2008 Nature Publishing Group
114. Lam, C. W., Getting, S. J. & Perretti, M. In vitro and Download full-text
in vivo induction of heme oxygenase 1 in mouse
macrophages following melanocortin receptor
activation. J. Immunol. 174, 2297–2304 (2005).
115. Lam, C. W., Perretti, M. & Getting, S. J. Melanocortin
receptor signaling in RAW264.7 macrophage cell line.
Peptides 27, 404–412 (2006).
116. Manna, S. K., Sarkar, A. & Sreenivasan, Y.
α-Melanocyte-stimulating hormone down-regulates
CXC receptors through activation of neutrophil
elastase. Eur. J. Immunol. 36, 754–769 (2006).
117. Taylor, A. W. The immunomodulating neuropeptide
α-melanocyte-stimulating hormone (α-MSH)
suppresses LPS-stimulated TLR4 with IRAK-M in
macrophages. J. Neuroimmunol. 162, 43–50 (2005).
118. Fernandes, B., Sagman, U., Auger, M., Demetrio, M. &
Dennis, J. W. β1,6-branched oligosaccharides as a
marker of tumor progression in human breast and
colon neoplasia. Cancer Res. 51, 718–723 (1991).
119. Handerson, T., Camp, R., Harigopal, M., Rimm, D. &
Pawelek, J. β1,6-Branched oligosaccharides are
associated with metastasis and predict poor outcome in
breast carcinoma. Clin. Cancer Res. 11, 2969–2973
120. Seelentag, W. K. et al. Pronostic value of β1,6-
branched oligosaccharides in human colorectal
carcinoma. Cancer Res. 58, 5559–5564 (1998).
121. Murata, K. et al. Expression of
N-acetylglucosaminyltransferase V in colorectal
cancer correlates with metastasis and poor prognosis.
Clin. Cancer Res. 6, 1772–1777 (2000).
122. Dosaka-Akita, H. et al. Expression of
N-acetylglucosaminyltransferase V is associated with
prognosis and histology in non-small cell lung cancers.
Clin. Cancer Res. 10, 1773–1779 (2004).
123. Fukuda, M., Spooncer, E., Oates, J. E., Dell, A. &
Klock, J. C. Structure of sialylated fucosyl
lactosaminoglycan isolated from human granulocytes.
J. Biol. Chem. 25 10925–10935 (1984).
124. Yamamoto, E. et al. Expression of N-acetyl-
glucosaminyltransferase V in endometrial cancer
correlates with poor prognosis. Br. J. Cancer 97,
125. Fukuda, M., Spooncer, E., Oates, J. E., Dell, A. &
Klock, J. C. Structure of sialylated fucosyl
lactosaminoglycan isolated from human granulocytes.
J. Biol. Chem. 259, 10925–10935 (1984).
126. Mizoguchi, A., Takasaki, S., Maeda, S. & Kobata, A.
Changes in asparagine-linked sugar chains of human
promyelocytic leukemic cells (HL-60) during
monocytoid differentiation and myeloid differentiation.
Decrease of high-molecular-weight oligosaccharides in
acidic fraction. J. Biol. Chem. 259, 11949–11957
127. Chakraborty, A. K. et al. Fusion hybrids with
macrophage and melanoma cells up-regulate
N-acetylglucosaminyltransferase V, β1–6 branching,
and metastasis. Cell Growth Differentiation 12,
128. Dennis, J., Waller, C. A. & Schirrmacher, V.
Identification of asparagine-linked oligosaccharides
involved in tumor cell adhesion to laminin and type IV
collagen. J. Cell Biol. 99, 1034–1044 (1984).
129. Demetriou, M., Nabi, I. R., Coppolino, M., Dedhar, S.
& Dennis, J. W. Reduced contact-inhibition and
substratum adhesion in epithelial cells expressing
GlcNAc-transferase, V. J. Cell Biol. 130, 383–392
130. Saitoh, O., Wang, W. C., Lotan, R. & Fukuda, M.
Differential glycosylation and cell surface expression
of lysosomal membrane glycoproteins in sublines of a
human colon cancer exhibiting distinct metastatic
potentials. J. Biol. Chem. 267, 5700–5711 (1992).
131. Chammas, R., Veiga, S. S., Travassos, L. R. &
Brentani, R. R. Functionally distinct roles for
glycosylation of α and β integrin chains in cell
matrix interactions. Proc. Natl Acad. Sci. USA 90,
132. Zheng, M., Fang, H. & Hakomori, S. Functional role
of N-glycosylation in α5β1 integrin receptor.
De-N-glycosylation induces dissociation or altered
association of α5 and β1 subunits and concomitant
loss of fibronectin binding activity. J. Biol. Chem.
269, 12325–12331, (1994).
133. Leppa, S., Heino, J. & Jalkanen, M. Increased
glycosylation of β1 integrin affects the interaction of
transformed s115 mammary epithelial cells with
laminin-1. Cell Growth Differ. 6, 853–861, (1995).
134. Dennis, J. W., Granovsky, M. & Warren, C. E.
Glycoprotein glycosylation and cancer progression.
Biochim. Biophys. Acta 1473, 21–34 (1999).
135. Yamamoto, H. et al. β1,6 N-acetyl-glucosamine
bearing N-glycans in human gliomas; implications for
role in regulating invasivity. Cancer Res. 60, 134–142
136. Ochwat, D., Hoja-Lukowicz, D. & Litynska, A.
N-glycoproteins bearing β1,6-branched
oligosaccharides from the A375 human melanoma
cell line analysed by tandem mass spectrometry.
Melanoma Res. 14, 479–485 (2004).
137. Guo, H.-B., Lee, I., Kamar, M., Akiyama, S. K. &
Pierce, M. Aberrant N-glycosylation of β1 integrin
causes reduced α5β1 integrin clustering and
stimulates cell migration. Cancer Res. 62,
138. Poche, E., Litysk, A., Amoresano, A. & Casbarra, A.
Glycosylation profile of integrin α3β1 changes with
melanoma progression. Biochim. Biophys. Acta Mol.
Cell Res. 1643, 113–123 (2003).
139. Jasiulionis, M. G., Chammas, R., Ventura, A. M.,
Travassos, L. R. & Brentani, R. R. α6β1-Integrin, a
major cell surface carrier of β1-6-branched
oligosaccharides, mediates migration of
EJ-ras-transformed fibroblasts on laminin-1
independently of its glycosylation state. Cancer Res.
56, 1682–1689 (1996).
140. Giannelli, G. et al. Role of the α3β1 and α6β4
integrins in tumor invasion. Clin. Exp. Metastasis 19,
141. Danen, E. H. J. et al. Emergence of α5β1 fibronectin-
and αvβ3 vitronectin-receptor expression in
melanocytic tumour progression. Histopathology 24,
142. Natali, P. G., Nicotra, M. R., Di Filippo, F. & Bigotti, A.
Expression of fibronectin, fibronectin isoforms and
integrin receptors in melanocytic lesions. Br. J. Cancer
71, 1243–1247 (1995).
143. Galbraith, C. G., Yamada, K. M. & Galbraith, J. A.
Polymerizing actin fibers position integrins primed
to probe for adhesion sites. Science 315, 992–995
144. Gladson, C. L. & Cheresh, D. A. Glioblastoma
expression of vitronectin and the αvβ3 integrin.
Adhesion mechanism for transformed glial cells.
J. Clin. Invest. 88, 1924–1932 (1991).
145. Natali, P. G., et al. Clinical significance of αvβ3
integrin and intercellular adhesion molecule-1
expression in cutaneous malignant melanoma lesions.
Cancer Res. 57, 1554–1560 (1997).
146. Wong, N. C. et al. αv integrins mediate adhesion and
migration of breast carcinoma cell lines. Clin. Exp.
Metastasis 16, 50–61 (1998).
147. Juliano, R. L. The role of β1 integrins in tumors.
Semin. Cancer Biol. 4, 277–283 (1993).
148. Ammon, C. et al. Comparative analysis of integrin
expression on monocyte-derived macrophages and
monocyte-derived dendritic cells. Immunology 100,
149. Aplin, A. E., Howe, A., Alahari, S. K. & Juliano, R. L.
Signal transduction and signal modulation by cell
adhesion receptors: the role of integrins, cadherins,
immunoglobulin-cell adhesion molecules, and
selectins. Pharmacol. Rev. 50, 197–263 (1998).
150. Elsegood, C. L. et al. M-CSF induces the stable
interaction of cFms with αVβ3 integrin in
osteoclasts. Int. J. Biochem. Cell Biol. 38,
151. Shinji, H. et al. Expression and distribution of very
late antigen-5 in mouse peritoneal macrophages upon
ingestion of fibronectin-bound Staphylococcus aureus.
Microbiol. Immunol. 51, 63–171 (2007).
152. Kurita-Taniguchi, M. et al. Molecular assembly of
CD46 with CD9, α3-β1 integrin and protein tyrosine
phosphatase SHP-1 in human macrophages through
differentiation by GM-CSF. Mol. Immunol. 38,
153. Chang, M. H. et al. Transthyretin interacts with the
lysosome-associated membrane protein (LAMP-1) in
circulation. Biochem. J. 382, 481–489 (2004).
154. Sawada, R., Lowe, J. B. & Fukuda, M.
E-selectin-dependent adhesion efficiency of colonic
carcinoma cells is increased by genetic manipulation
of their cell surface lysosomal membrane
glycoprotein-1 expression levels. J. Biol. Chem. 268,
155. Sarafian, V. et al. Expression of Lamp-1 and Lamp-2
and their interactions with galectin-3 in human tumor
cells. Int. J. Cancer 75, 105–111 (1998).
156. Chakraborty, A. K. & Pawelek, J. M. β1,6-branched
oligosaccharides regulate melanin content and
motility in macrophage-melanoma fusion hybrids.
Melanoma Res. 17, 9–16 (2007).
157. Rupani, R., Handerson, T. & Pawelek, J. Co-localization
of β1,6-branched oligosaccharides and coarse melanin
in macrophage-melanoma fusion hybrids and human
melanoma cells in vitro. Pig. Cell Res. 17, 281–288
158. Hariri, M. et al. Biogenesis of multilamellar bodies via
autophagy. Mol. Biol. Cell 11, 255–268 (2000).
159. Handerson, T. & Pawelek, J. β1,6-branched
oligosaccharides and coarse vesicles: A common and
pervasive phenotype in melanoma and other human
cancers. Cancer Res. 63, 5363–5369 (2003).
160. Clark, W. H. et al. Current concepts of the biology of
human cutaneous malignant melanoma. Adv. Cancer
Res. 24, 267–338 (1977).
161. Hait, W. N., Jin, S. & Yang, J.-M. A matter of life or
death (or both): understanding autophagy in cancer.
Clin. Cancer Res. 12, 1961–1965 (2006).
162. Hait, W. N., Wu, H., Jin, S. & Yang, J. M. Elongation
factor-2 kinase: its role in protein synthesis and
autophagy. Autophagy 2, 294–296 (2006).
163. Hait, W. N., Jin, S. & Yang, J. M. Elongation factor-2
kinase regulates autophagy in human glioblastoma
cells. Clin. Cancer Res. 12, 1961–1965 (2006).
164. Mathew, R., Karantza-Wadsworth, V. & White, E.
Role of autophagy in cancer. Nature Rev. Cancer 7,
165. Degenhardt, K. et al. Autophagy promotes tumor cell
survival and restricts necrosis, inflammation, and
tumorigenesis. Cancer Cell 10, 51–64 (2006).
166. Levine, B. Cell biology: autophagy and cancer. Nature
446, 745–747 (2007).
167. Amer, A. O. & Swanson, M. S. Autophagy is an
immediate macrophage response to Legionella
pneumophila. Cell Microbiol. 7, 765–778 (2005).
168. Amer, A. O., Byrne, B. G. & Swanson, M. S.
Macrophages rapidly transfer pathogens from lipid raft
vacuoles to autophagosomes. Autophagy 1, 53–58
169. Lugini, L. et al. Cannibalism of live lymphocytes by
human metastatic but not primary melanoma cells.
Cancer Res. 66, 3629–3638 (2006).
170. Lugini, L. et al. Potent phagocytic activity
discriminates metastatic and primary human
malignant melanomas: a key role of ezrin. Lab. Invest.
83, 1555–1567 (2003).
171. Damiani, M. T. & Colombo, M. I. Microfilaments and
microtubules regulate recycling from phagosomes.
Exp. Cell Res. 289, 152–161 (2003).
172. Coopman, P. J., Do, M. T., Thompson, E. W. & Mueller,
S. C. Phagocytosis of cross-linked gelatin matrix by
human breast carcinoma cells correlates with their
invasive capacity. Clin. Cancer Res. 4, 507–515
173. Montcourrier, P. et al. Characterization of very acidic
phagosomes in breast cancer cells and their
association with invasion. J. Cell Sci. 107, 2381–2391
174. Houghton, J. et al. Gastric cancer originating from
bone marrow-derived cells. Science 306, 1568–1571
175. Parris, G. E. 2-Deoxy-d-glucose as a potential drug
against fusogenic viruses including HIV. Med.
Hypotheses 70, 776–782 (2008).
176. Halaban, R. et al. Aberrant retention of tyrosinase in
the endoplasmic reticulum mediates accelerated
degradation of the enzyme and contributes to the
dedifferentiated phenotype of amelanotic melanoma
cells. Proc. Natl Acad. Sci. USA 94, 6210–6215
We gratefully acknowledge the many and invaluable contri-
butions of D. Bermudes, J. Bolognia, D. Brash, D. Cooper,
T. Henderson, R. Lazova, L. Margulis, J. Pawelek, J. Platt,
M. Rachkovsky, S. Sodi and Y. Yilmaz. We thank
R. Sorensen, D. Schafer and L. Hummel for their critical
readings of the manuscript. Supported in part by a gift from
Vion Pharmaceuticals (J.M.P.), and a grant from Avon
entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query.
AKT2 | CCL2 | CD44 | CD47 | CDH1 | ERVWE1 | FN1 | HGF |
LAMP1 | LGALS3 | MC1r | MET | MiTF | MMP9 | RHOC | SeLe |
SirPA | SNAi1 | SNAi2 | SPArC | TM7SF4 | TWiST1 | vTN
National cancer institute: http://www.cancer.gov/
breast cancer | colorectal carcinoma | endometrial cancer |
lung carcinoma | melanoma | ovarian carcinoma
386 | MAY 2008 | vOLUME 8
© 2008 Nature Publishing Group