RECK—a newly discovered inhibitor of metastasis
with prognostic significance in multiple forms of cancer
Jonathan C. M. Clark & David M. Thomas &
Peter F. M. Choong & Crispin R. Dass
Published online: 9 September 2007
# Springer Science + Business Media, LLC 2007
Abstract The RECK (reversion-inducing cysteine rich
protein with Kazal motifs) protein was initially discovered
by its ability to induce reversion in ras-activated fibroblasts.
The key action of RECK is to inhibit matrix metal-
loproteinases (MMPs) involved in breakdown of the
extracellular matrix (ECM), and angiogenesis—namely
MMP-2, MMP-9 and MTP-1. To this effect, it plays
important physiological roles in embryogenesis and vascu-
logenesis. Additionally, it has a significant effect on
tumorigenesis by limiting angiogenesis and invasion of
tumours through the ECM. RECK has been studied in the
context of a number of human tumours including colorec-
tal, breast, pancreas, gastric, hepatocellular, prostate, and
non-small cell lung carcinoma. In many of these tumours,
RECK is down-regulated most likely as a result of
inhibition at the Sp1 promoter site. MMP-2 and MMP-9
generally show an inverse association with RECK expres-
sion, but there are exceptions to this rule. Likewise, a
reduction in tumour microvascular density (MVD) and
VEGF have also been correlated with increased RECK
levels, although more studies are required to define this
effect. The predominant finding across all human tumour
studies is a significantly improved prognosis (due to
decreased invasion and metastasis) in tumours with pre-
served RECK expression. Although further research is
required, RECK is a promising prognostic marker and
potential therapeutic agent in multiple cancers.
RECK is a recently characterised membrane bound protein
which, in the mouse, has been found to be important in
suppressing two key components in the metastatic cascade:
matrix metalloproteinases (MMPs), and angiogenesis . A
limited number of studies have also confirmed that RECK
levels are significantly down-regulated in common human
malignancies, compared with the normal surrounding
tissue. By contrast, in the minority of tumour samples
where RECK levels are normal or elevated, there is
generally a reduction in local invasion, metastasis and an
The protein RECK stands for ‘reversion-inducing cyste-
ine rich protein with Kazal motifs’. RECK was initially
discovered by screening a human fibroblast cDNA library
for genes giving rise to reversion-inducing clones when
transfected into v-Ki-ras transformed NIH 3T3 cells .
Hybridization cDNA screening in the mouse identified a
The aim of this review is to provide an overall
understanding of the mechanisms by which RECK inhibits
tumour invasion, the molecular pathways involved in RECK
Cancer Metastasis Rev (2007) 26:675–683
J. C. M. Clark:P. F. M. Choong:C. R. Dass (*)
Department of Orthopaedics, St. Vincent’s Hospital,
University of Melbourne,
D. M. Thomas
Sarcoma Genomics and Genetics Laboratory,
Peter MacCallum Cancer Centre,
P. F. M. Choong
Bone and Soft Tissue Sarcoma Service,
Peter MacCallum Cancer Centre,
down-regulation, and further direction for RECK manipula-
tion based on the patterns observed in some tumours.
2 Gene structure
The genomic structure for the RECK gene has been
identified on chromosome region 9p13→p12 , and as
noted by Kim et al. , the short arm of chromosome 9
contains a number of potential tumour suppressor genes ,
for example p16INK4a (CDKN2A) which has a well
established role in melanoma. These authors demonstrated
a loss of heterozygosity (LOH), an indicator of mutated
tumour suppressor genes, in greater than 50% of small cell
lung cancers at the 9p13 region, distinct from and
centromeric to the CDKN2A locus. The identity of putative
tumour suppressor gene(s) at the 9p13 locus remains
unknown, but RECK represents an interesting candidate.
The gene sequence was identified by Eisenberg et al. 
within two Bacterial artificial clones (BAC), using a
BLAST search. It includes 21 exons and 20 introns, with
13 SNPs (single nucleotide polymorphisms). Four of these
are found within the coding sequence which raises the
possibility of disease involvement—where polymorphisms
in the corresponding protein structure lead to abnormal
3 Protein structure
As originally described by Takahashi et al. , the protein
coded for by the RECK gene is 971 amino acid residues in
both humans and mice and there is 93% identity between
the two species. The overall structure (Fig. 1) consists of a
hydrophobic region at either end, a length of five cysteine
repeats near the NH2 end, two regions with epidermal
growth factor (EGF)-like activity in the middle of the
protein, and three regions with serine protease inhibitor
activity. The COOH end allows glycosylphosphatidylinosi-
tol (GPI) anchoring to the cell membrane .
The first third of the protein sequence (marked by five
cysteine repeats) appears to have particular functional
importance in that it also contains a number of glycosyl-
ation sites at asparagine (Asn) residues. These sites are
necessary for proper interaction with MMP-9 and MMP-2
as will be discussed later.
According to Takahashi et al. , the EGF-like regions
have weak homology to the EGF protein sequence. EGF-
like molecules are usually around 50–60 residues, they can
bind to the EGF receptor and stimulate mitosis. EGF-like
sequences within cell surface proteins, such as RECK, do
not always bind to EGF receptors, but instead can influence
cell development, adhesion, and protein interactions .
This functional description is certainly in line with the
current understanding of RECK function and its interaction
Located within the middle of the protein structure are
three serine-protease inhibitor-like domains (Fig. 1). SPIs
(serpins) are proteins which inhibit the action of proteases
by ‘trapping reactions’ and ‘reversible tight binding
interactions’. The first of these SPIs is identical to the
Kazal motif, which is a particular family of peptidase
inhibitors containing disulphide-rich proteins with small
alpha and beta folds . The other two SPIs in the RECK
protein are similar but not identical to the Kazal motif.
Given that MMPs are indeed proteases, these SPI-like
domains are likely to have a significant role in MMP
It appears that RECK is predominantly anchored to cell
membranes in vivo via the COOH-terminal hydrophobic
region and a GPI interaction. Serine proteases are known to
be anchored to cell membrane via GPI and are thought to
have some role in intracellular signal transductions . This
raises the possibility that RECK, being membrane-anchored
Fig. 1 The overall structure of
the RECK gene. Adapted from
Takahashi et al. 1998 
676 Cancer Metastasis Rev (2007) 26:675–683
in the same way, may facilitate signal transduction also.
RECK can be cleaved from the surface of cells in vitro
using phosphatidylinositol-specific phospholipase C (PI-
PLC) to make it available in soluble form .
Membrane anchoring of RECK is clearly of some
importance, and the mechanisms behind this may be
GPI linkage to the membrane are thought to be involved in
intracellular signal transduction. Noel and colleagues 
mention this as a possible mechanism for serine proteases.
Muller and coworkers  found that the phosphoinositol
component of GPI-anchored proteins interacts with protein
binding sites in glycolipid rafts and this results in signal
transduction to produce an insulin-mimetic response.
4 Physiological role of RECK
RECK is an important mediator of tissue remodeling. Its
main function is to inhibit MMP-2, MMP-9 and MT1-MMP
post-transcriptionally . These MMPs are active in break-
ing down the extra-cellular matrix (ECM) in both physio-
logical and pathological states, including neoplasia .
RECK, on the other hand, keeps these processes under
control. Researchers Oh et al.  found tissue disruption and
reduced collagen IV, laminin and fibronectin in RECK −/−
embryos when compared with tissues from the wild-type.
This indicates that in the absence of RECK, MMPs degrade
the matrix in an unchecked fashion.
RECK is required for normal embryonic development,
facilitating ordered tissue remodeling by maintaining ECM
around vessels and the neural tube . Adequate ECM
surrounding vessels promotes vessel stability and matura-
tion. Mouse embryos lacking both copies of the RECK
gene die slightly earlier than those lacking both RECK and
MMP-2 genes because removing RECK alone leaves the
proteolytic action of MMP-2 unchecked, resulting in
excessive destruction of collagen IV and laminin in the
basal lamina of vessel membranes . RECK −/− embryos
showed disrupted mesenchymal tissues and organogenesis.
They had smaller bodies and developed abdominal haem-
orrhages. The vessels studied in these mice were primitive
and underdeveloped, leaving the authors to conclude that
RECK was more vital to stabilising angiogenic sprouting
rather than later steps in maintaining vascular structure,
which requires recruitment of cells to an area of vascular
5 Pathophysiological role of RECK
RECK over-expression in a fibrosarcoma cell line HT1080
compared with a control showed significant reduction in
microvessel density (MVD) and branching . The
mechanism behind this, as mentioned above, is thought
to be inhibition of MMP-2 which leads to reduced tissue
remodeling and sprouting of vessels. Vessels in the
RECK-expressing tumours tend to develop in diameter,
but lack the ability to branch-off throughout the tumour
tissue. This results in death of tumour tissue in areas of
reduced MVD. The level of RECK expression did not
appear to affect tumour volume, but mice with RECK-
expressing tumours were noted to have a prolonged
survival compared with the control.
The findings of reduced MVD correlate with results
from the limited number of studies on human cancers where
MVD has been examined in relation to RECK expression.
There are three studies of note here, with somewhat
conflicting results. Takeuchi et al.  identified an inverse
relationship between RECK and both VEGF and MVD in
colorectal cancer. Takenaka et al.  also found the same
inverse relationship for MVD in non-small cell lung cancer
but this relationship was only present in tissues strongly
expressing VEGF, suggesting that the effects of RECK on
vascular structures is directly or indirectly dependent on
VEGF. A further study by Takenaka et al.  specifically
looking at RECK expression in stage IIA N2 non-small cell
lung cancer did not show significantly reduced MVD with
high RECK expression. The reason for this is largely
undetermined, but undoubtedly more studies will clarify the
true extent of such an association.
In addition to its effects on angiogenesis, RECK also
limits tumour invasion and metastasis by inhibiting matrix
metalloproteinases (MMPs). MMPs were previously men-
tioned in relation to tissue remodeling where they facilitate
ordered tissue break-down in concert with regulatory
factors such as RECK. In pathological states such as
neoplasia and rheumatoid arthritis, there is an imbalance
in remodeling factors towards favouring of MMPs. The
result is a destruction of the ECM allowing tumour cells to
extend into surrounding tissues, and the bloodstream.
RECK appears to suppress MMPs secreted by tumour
cells via a glycosylation mechanism. In normal cells, four
separate asparagine residues in the RECK protein sequence
are N-glycosylated. Simizu et al.  developed mutant
RECK proteins where Asn residues were replaced with Glu
in certain positions, to observe the effect of absent
glycosylation on the expression of MMP-2 and MMP-9.
They found that glycosylation of the Asn297was required to
prevent MMP-9 secretion and Asn352glycosylation was
needed to inhibit MMP-2 activation. This finding was then
correlated with tumour cell invasion where HT1080-RECK
mutants (absent glycosylation at Asn297and Asn352)
showed increased tumour cell invasion in comparison to
HT1080 cells expressing wild-type RECK. Asn297is found
within one of the cysteine knot regions of RECK (CKM5).
Cancer Metastasis Rev (2007) 26:675–683677
RECK sequences with mutated CKM5 were likewise found
to be unable to suppress MMP-9.
6 Mechanisms by which RECK is down-regulated
The discovery of RECK and its ability to counteract MMPs
has major implications for treating cancer. Importantly
RECK inhibits both metastasis and angiogenesis, and it has
been confirmed that in many tumours, RECK is down-
regulated. The mechanism of this down-regulation appears
to be multifactorial and also tumour specific, with a
common target of inhibition being the Sp1 site on the
RECK promoter sequence (Fig. 2—summary of RECK
control). Early studies, Sasahara et al. [14, 15], speculated
that by activating the extracellular signal regulated kinase
(ERK) pathway, oncogenic Ras facilitates phosphorylation
or other modification of Sp1/Sp3 factors which increases
their affinity for the Sp1 site on the RECK promoter thus
reducing RECK expression. Sasahara et al.  also
hypothesised that histone deacetylation interaction with
Sp1 may contribute to transcriptional repression of RECK.
They used an inhibitor of HDAC, Trichostatin A (TSA), in
NIH-3T3 cells but found no specific correlation with RECK
levels. Transcription factor Sp1 has recently been shown to
be overexpressed in a number of human cancers and its
overexpression contributes to malignant transformation [16,
17]. Sp1 regulates the expression of a number of genes
participating in multiple aspects of tumorigenesis such as
angiogenesis, cell growth and resistance to apoptosis. Thus,
the intricate interplay between Ras, Sp1 and RECK may
dictate the degree of malignancy of a neoplastic cell.
Following on from the above study, Liu et al.  tested
a similar hypothesis in CL-1 human lung cancer cells, and
found increased cell surface RECK expression after adding
TSA. This is thought to occur because TSA inhibits the
interaction between HDAC and Sp1, and therefore reduces
binding to the Sp1 promoter site. Additionally, once RECK
was up-regulated, a corresponding inhibition of MMP-2
A further study published by the same investigators,
Chang et al. , demonstrates further evidence to support
RECK transcription inhibition via ras and histone deacet-
ylation. Interestingly, they found that Sp1 and Sp3 actually
increase RECK promoter activity rather than inhibit it.
Oncogenic ras activity, probably via the ERK phosphory-
lation pathway, resulted in increased Sp1 protein associated
with HDAC, and this is believed to increase the binding of
HDAC to the Sp1 site on the RECK promoter.
RECK is also inhibited by LMP-1, a product of the
Epstein Barr virus (EBV) LMP-1 acts by binding to the
Sp1 site in the promoter region of the RECK gene to
inhibit promoter function. LMP-1 was also found to
stimulate the ERK signaling pathway, but when this
pathway was inhibited by PD98059 (ERK pathway
inhibitor), there was diminished inhibition of RECK. This
suggests that an overactive ERK signaling pathway
(induced by LMP-1) may be responsible for RECK
Inhibits vessel branching
Via MMP-2 inhibition and
on RECK gene
Fig. 2 Summary of RECK
control. RECK expression is
inhibited via a number of dif-
ferent pathways acting on the
Sp1 promoter site of the RECK
gene. Specifically, oncogenic
Ras increases ERK pathway
activity leading to the combined
action of histone deacetylase
and the Sp1 protein in binding
and inhibiting the Sp1 promoter
site. NSAIDs and HDAC inhib-
itors act against this common
pathway, and show potential for
cancer therapy by reducing
RECK down-regulation. The
main function of RECK is to
inhibit MMP-2, MMP-9, and
MT1-MMP thereby reducing
MMP destruction of the ECM
and reducing metastasis. RECK
also inhibits angiogenesis
678Cancer Metastasis Rev (2007) 26:675–683
In a very similar mechanism to that used by LMP-1, the
HER-2/neu protein reduces RECK expression by increasing
the binding of Sp1 proteins to the Sp1 site. It does this by
inducing the ERK pathway to phosphorylate Sp1 proteins,
increasing their affinity for the RECK promoter, and thus
inhibiting RECK expression. Although this Sp1 protein
action may not directly down-regulate RECK (as per Chang
et al. ), HER-2/neu also recruits HDAC to Sp1 proteins
and this combination represses expression of the RECK
gene by binding to the promoter region .
Thus, it is evident from these three mechanisms that
binding and inhibiting the Sp1 region of the RECK
promoter provides the common pathway for down-regulating
RECK expression in tumours. This process of inhibition can
be achieved either through direct binding of the oncogenic
factor (that is, LMP-1) or indirectly via up-regulation of the
ERK signaling pathway which increases the affinity of
proteins binding to the Sp1 region specifically HDAC.
There are almost certainly other pathways by which
RECK is down-regulated in cancer. In their defining paper
on RECK, Takahashi et al.  found RECK down-
regulation in NIH3T3 cells transformed by a number of
other oncogenes apart from ras . These included v-fos, c-
myc, v-src, v-fms, v-fes, and v-mos. Needless to say, closer
attention to the molecular pathways intricately involved
with RECK downregulation in cancer will no doubt lead to
better ways of manipulating RECK expression via modu-
lation of regulators upstream or reinstatement of molecular
players downstream of RECK.
7 RECK down-regulation identified in many tumours
In studies over the last 6 years, a number of common
tumours have been linked to RECK down-regulation, and
in particular, down-regulation was associated with reduced
survival. At least 19 human tumour cell lines have already
demonstrated absent RECK mRNA . Exceptions to this
rule are being identified as RECK is analysed in a greater
spectrum of tumours, for example in hepatocellular carci-
noma, where RECK levels were actually higher than
normal liver in 40% of specimens . So far, tumours
studied in relation to prognosis and RECK expression
include; colorectal, breast, lung, gastric, hepatocellular and
pancreatic carcinomas. The key findings from these studies
will now be discussed briefly in turn (see Table 1 for
In a study by Takeuchi et al.  colorectal carcinoma
tissue samples obtained from 53 patients were analysed for
RECK and MMP-9 using immunohistochemistry, with
MMP-9 being found to be the predominant gelatinase on
gelatin zymography. RECK was confirmed to be down-
Table 1 Summary of RECK involvement in human cancer
No. 40.6% of tumours
had high RECK
No. 44.9% strong
No. 52% RECK
positive, 48% RECKnegative
No correlation with
Effect on level of
No effect on
Decreased MMP-9. No
effect on MMP-2 and
Decreased MMP-2. No
effect on MMP-9
van der Jagt et al.
Furumoto et al. 
Song et al. 
Takenaka et al. 
Masui et al. 
Kang et al. 
Cancer Metastasis Rev (2007) 26:675–683679
regulated in the majority of colon carcinoma specimens
compared with surrounding normal tissues, and in tumour
samples with high levels of expressed RECK there was no
poorly differentiated tissue. Statistically significant findings
included a higher frequency of lymph node metastases in
the low-RECK group, and an inverse relationship between
RECK levels and the Duke’s staging. There was no
relationship between RECK expression and MMP levels,
however there was a correlation between prognosis and the
RECK/MMP-9 ratio. This was based on the hypothesis that
tumours with the highest RECK and the lowest MMP-9
levels, should display the least angiogenesis, invasion and
found an inverse relationship between RECK and MMP-2
but not with MMP-9. This may be the result of preserved
Asn297glycosylation but absent glycosylation of Asn352(see
section on “Pathophysiological role of RECK” above).
These particular investigators believe that there is no direct
inhibition of MMP-9 by RECK in colorectal cancer.
Span et al.  found significantly reduced RECK
mRNA levels in invasive breast carcinoma specimens
compared with normal surrounding tissues. Consistent with
other studies, RECK expression had a significant bearing
on survival times in these patients. Measurements of
angiogenesis and MMPs were not performed in this study.
In a study by Riddick et al. , PCR was used to
examine expression of MMPs and their inhibitors, which
were either up or down-regulated in prostate cancer. Levels
of these various factors were compared between benign and
malignant prostate tissue. RECK and MMP2 were both
significantly down-regulated in malignant tissue when
compared with levels in the benign tumour tissue. The
down-regulation of MMP-2 was an unusual finding, given
that another study on prostate cancer shows increased
MMP-2, although immunohistochemistry was used to study
the MMP-2 protein rather than mRNA . Riddick et al.
 believe there are differences between the levels of
observed MMP-2 mRNA and MMP-2 protein. This may
relate to the fact that RECK is known to inhibit MMPs at
the post-transcriptional level . It is therefore more
beneficial to study MMP protein rather than mRNA levels
when looking at the effect of RECK. RECK down-
regulation was associated with an increased Gleason score,
which clearly demonstrates a correlation between low
RECK and a worse prognosis. The main MMP up-regulated
in malignant prostate cancer was MMP-26. So far, there
appears to be no documented relationship between this
MMP and RECK levels.
As previously noted, Furumoto et al.  found that in
26 cases of hepatocellular carcinoma, RECK expression
was actually higher than in the non-tumour control speci-
mens. Possible explanations for this will be addressed in the
discussion. The direct correlation of RECK with improved
prognosis was still evident in this study. With higher RECK
levels identified in tumours, patients had better survival and
tumours demonstrated reduced invasion.
When comparing NSCLC tumour groups with strong
and weak expression of RECK there was no significant
difference in the intra-tumour MVD . In addition, no
significant difference was found between strong and weak
RECK expression in the categories of age, sex, tumour
differentiation, clinical N factor, or number of N2 node
stations involved. There was also no difference in prolifer-
ative index, apoptotic index or p53 expression. The most
significant finding was that strong RECK expression corre-
lated with improved survival overall (42.9% 5 year survival
versus 23.1%), and in squamous cell carcinomas, weak
RECK expression was associated with a poor prognosis.
The level of angiogenesis was examined by Takenaka et al.
, and it was found that MVD was reduced with high-
expression of RECK but only in tumours concurrently
expressing VEGF. This study covered NSCLC stages I–IIIA.
Similar to the pattern observed in hepatocellular carci-
noma, RECK was not significantly down-regulated in
pancreatic tumours studied by Masui et al. , with 52%
actually demonstrating strong expression. An improved
prognosis for RECK-expressing tumours was found in this
study with patients whose tumours were positive for RECK
having a 2 year survival of 42%, in comparison to 0% in
the RECK-negative group (P=0.0463). Additionally, an
inverse relationship was identified between RECK expres-
sion and MMP-2 activation ratio. On the other hand, RECK
had no significant influence over MMP-9 in these pancre-
atic tumour samples, and as yet there is no explanation for
this. Angiogenesis was not examined.
Similar to other cancers studied in relation to RECK, Song
et al.  have identified an inverse relationship between
RECK expression and gastric tumour invasion or metastasis.
Increased RECK expression was associated with a reduction
in MMP-9 levels, but not MMP-2 or MMP-7. This pattern is
essentially the opposite from that seen in pancreatic cancer,
where levels of MMP-9 were not affected by RECK.
Song et al.  also studied VEGF and found, not
surprisingly, that levels of VEGF 121 and 165 were
increased in most tumours, but these levels were not
significantly reduced in the presence of high RECK
expression. This finding was in contrast to colorectal and
non-small cell lung cancer where VEGF levels and MVD
were reduced with increased RECK. CD-34 immunohisto-
chemistry was used to assess lymphatic invasion but MVD
was not examined in this study, which may have shed more
light on a potential anti-angiogenic action of RECK in
Kang et al.  studied RECK in osteosarcoma cell lines
which included the established lines HOS, U-2OS, MG-63,
680Cancer Metastasis Rev (2007) 26:675–683
and SaOS-2, and 23 lines established inhouse from patients
prior to undergoing chemotherapy. They found that RECK
mRNA levels were reduced in the majority of cell lines
relative to expression in MG63 cells. Predictably, higher
RECK levels were inversely correlated with pro-MMP-2,
which was expressed across all cell lines. On the other
hand, pro-MMP-9 was only expressed in U-2OS cells. Cell
lines transfected with the RECK gene demonstrated that
pro-MMP-2 activation was significantly decreased but pro-
MMP-9 activation was not. Invasion assays were also
conducted in matrigel (specifically using the HOS line),
showing reduced cell invasion after RECK transfection.
Rheumatoid arthritis (RA) is a condition which exhibits
similarities to invasive cancer in that MMPs play a
significant role in ECM breakdown (particularly MT-
MMP and MMP-14). In a study by van Lent et al. ,
RECK expression was significantly lower in RA synovial
membranes comparedwithcontrols. RECK down-regulation
in the synovium may be related to cytokine signaling
proteins Sp1 and Sp3 which, as previously mentioned, are
able to regulate RECK expression by binding to its promoter
region. These researchers found no significant difference in
MMP mRNA levels between rheumatoid tissues and con-
trols. This does not rule out significant involvement of
MMPs since the key regulatory point appears to be post-
8 Overall characteristics of RECK
When studying the relationship of RECK and MMPs in
different tumours, there are a number of consistent
observations. Firstly, RECK appears to have prognostic
significance across the board. This finding alone may be
used clinically to enable a more specific estimation of the
tumour stage and grade at the initial biopsy. Secondly,
RECK is noted to be down-regulated in the majority of
tumours studied, with the exception of pancreatic and
hepatocellular cancer. While more studies have to confirm
this link, this currently provides a distinct avenue for further
studies to reverse RECK knockdown, effectively over-
expressing RECK in certain tumours and observing the
effects on local invasion, metastasis, MVD, and angiogenic
factors. Furthermore, while it may seem far-fetched at
present, it may be possible to administer recombinant
RECK (rRECK) to the patient either locally or systemically,
or apply gene transfer techniques in order to improve
There are still a number of features of RECK which need
clarification. More studies are required to define the ability
of RECK to inhibit angiogenesis and its interaction with
VEGF in different human tumours. Only three studies so
far, have examined this issue. The general pattern,
consistent with Oh et al. , is a reduction in MVD—most
likely through interaction with VEGF. However, as seen in
gastric cancer , VEGF levels can be unchanged with
RECK expression. These findings necessitate an explora-
tion of the mechanisms behind RECK interaction with
VEGF and the possible RECK-induced down-regulation of
other angiogenic factors.
Various glycosylation sites within the RECK protein,
and particularly the CKM5 region, are clearly important
structural features relating to RECK’s ability to inhibit
MMP-9 and MMP-2. As we have seen, there are at least
two tumours where RECK expression is not significantly
reduced (hepatocellular and pancreatic carcinomas). In
these cases, it may be that in some of the cells, RECK
protein is actually mutated at the important glycosylation
sites Asn297and Asn352rendering it unable to perform its
usual role of inhibiting MMP-9 and MMP-2, but leaving it
in large detectable quantities. More data is needed to
determine the amount of specific MMP expressed in
different human tumours, and to correlate this with the
presence or absence of mutated forms of RECK. This will
provide a clearer understanding of the mechanism of
RECK’s involvement in carcinogenesis.
The pre-dominant MMPs inhibited by RECK, MMP-2
and MMP-9, do not display a consistent inhibition across
the different tumours studied. For example in colorectal
cancer , there was no significant effect of RECK on the
level of MMP-2 or MMP-9, even though tumours with
the highest RECK/MMP-9 ratio were confirmed to have the
best prognosis. In addition, the studies on gastric and
pancreatic cancer show an opposing pattern of RECK
interaction with MMP-2 and 9 (Table 1). Again, this may
the result of glycosylation-site specific mutations in the
RECK protein resulting in different MMP expression
9 RECK—potential for cancer therapy
While it is still early days. the consistent association of
RECK with improved prognosis in multiple cancers
suggests that this protein may have therapeutic potential.
Such evaluation, initially in cell culture, then in preclinical
(animal) models, and finally in clinical trials, will be
necessary to demonstrate this. With this foresight, a number
of therapeutic possibilities are now discussed.
Firstly, recombinant RECK could be used for therapy,
similar to insulin and erythropoietin administration. This
soluble RECK could then be administered either by
systemic or by local infusion. There are several methods
available for protein delivery, some even currently used
clinically [31, 32]. Another possibility would be gene
transfer, involving the delivery of viral or nonviral vectors
Cancer Metastasis Rev (2007) 26:675–683681
in vivo towards gene therapy. While this technique seems
less likely given the limited control over the actual amount
of RECK produced, studies with other proteins like PEDF
demonstrate its latent promise . With viral vectors
though, there is the potential for further neoplastic change
to occur when transfecting an already unstable genome
. Thus, use of non-viral vectors such as liposomes ,
microplexes [36, 37], cyclodextrins  and nanoparticles
, may be more appropriate for gene transfer. What is
needed now are studies outlining whether RECK is needed
for normal development, apart from what is now known
regarding its role in angiogenesis and vasculogenesis. The
effects of RECK overexpression or treatment with rRECK
on normal tissues should be elucidated in preclinical
studies. Furthermore, its effects on developing versus
established vasculature needs to be examined.
As previously outlined, the new evidence that HDAC
inhibitors (that is, trichostatin A) are able to increase RECK
levels by minimizing promoter inhibition, raises the
possibility for HDAC inhibitors to be used therapeutical-
ly. Furthermore, there is the use of non-steroidal anti-
inflammatories (NSAIDs) to increase RECK expression.
Liu et al.  hypothesise that this observation may be due
to NSAID inhibition of the ras/ERK/Sp1 pathway. They
also note that the mechanism behind RECK up-regulation is
independent of NSAID action on cyclooxygenase (COX)
since concurrent administration of prostaglandin E2(PGE2)
or overexpression of COX-2 by transfection in the lung
cancer cells, did not change the levels of RECK.
RECK offers significant promise as a future prognostic
indicator and may be even as potential cancer therapeutic.
By inhibiting MMP-2 and MMP-9, it inhibits the local
tissue invasion and metastasis of a number of common
human cancers, and also reduces angiogenesis. The
ultimate test of its effectiveness is an improvement in
prognosis, which so far, with the limited number of studies
hitherto, has been demonstrated in nearly every human
RECK is down-regulated in many cancers, and where
there are exceptions to this rule (that is, hepatocellular and
pancreatic carcinoma), high RECK levels clearly offer a
prognostic advantage. However, these cancers are not
associated with a favourable prognosis overall [41, 42].
Although late diagnosis is probably the most significant
factor in this contrasting observation, there may also be
variable mutations of RECK glycosylation sites within
tumour subclones such that one portion of cells cease to
produce RECK (perhaps as a consequence of oncogenic
ras), another produces effective unmutated RECK and still
another portion produce ineffective mutated RECK at the
The molecular mechanisms behind RECK transcriptional
regulation are clearly complex, and are still yet to be fully
defined. However, with the discovery of the role of HDAC,
via oncogenic ras, in reducing RECK transcription, the use
of HDAC inhibitors such as TSA offers another possible
means of manipulating RECK to provide cancer therapy.
This may be a promising alternative to the difficulties faced
by gene therapy. Additionally, new non-viral methods of
transfecting cells may provide an effective method of up-
regulating RECK production.
1. Oh, J., Takahashi, R., Kondo, S., Mizoguchi, A., Adachi, E., &
Sasahara, R. M., et al. (2001). The membrane-anchored MMP
inhibitor RECK is a key regulator of extracellular matrix integrity
and angiogenesis. Cell, 107, 789–800.
2. Takahashi, C., Sheng, Z., Horan, T. P., Kitayama, H., Maki, M., &
Hitomi, K., et al. (1998). Regulation of matrix metalloproteinase-9
and inhibition of tumor invasion by the membrane-anchored
glycoprotein RECK. Proceedings of the National Academy of
Sciences of the United States of America, 95, 13221–13226.
3. Eisenberg, I., Hochner, H., Sadeh, M., Argov, Z., & Mitrani-
Rosenbaum, S. (2002). Establishment of the genomic structure
and identification of thirteen single-nucleotide polymorphisms in
the human RECK gene. Cytogenetic and Genome Research, 97,
4. Kim, S. K., Ro, J. Y., Kemp, B. L., Lee, J. S., Kwon, T. J., &
Fong, K. M., et al. (1997). Identification of three distinct tumor
suppressor loci on the short arm of chromosome 9 in small cell
lung cancer. Cancer Research, 57, 400–403.
5. Carpenter, G., & Cohen, S. (1990). Epidermal growth factor.
Journal of Biological Chemistry, 265(14), 7709–7712.
6. Rawlings, N. D., Tolle, D. P., & Barrett, A. J. (2004).
Evolutionary families of peptidase inhibitors. Biochemical Jour-
nal, 378, 705–716.
7. Noel, A., Maillard, C., Rocks, N., Jost, M., Chabottaux, V., &
Sounni, N. E., et al. (2004). Membrane associated proteases and
their inhibitors in tumour angiogenesis. Journal of Clinical
Pathology, 57(6), 577–584.
8. Müller, G., Jung, C., Frick, W., Bandlow, W., & Kramer, W.
(2002). Interaction of phosphatidylinositolglycan(-peptides) with
plasma membrane lipid rafts triggers insulin-mimetic signaling in
rat adipocytes. Archives of Biochemistry and Biophysics, 408(1),
9. Fisher, J. F., & Mobashery, S. (2006). Recent advances in MMP
inhibitor design. Cancer and Metastasis Reviews, 25(1), 115–136.
10. Takeuchi, T., Hisanaga, M., Nagao, M., Ikeda, N., Fujii, H., &
Koyama, F., et al. (2004). The membrane-anchored matrix
metalloproteinase (MMP) regulator RECK in combination with
MMP-9 serves as an informative prognostic indicator for
colorectal cancer. Clinical Cancer Research, 10(16), 5572–5579.
11. Takenaka, K., Ishikawa, S., Kawano, Y., Yanagihara, K.,
Miyahara, R., & Otake, Y., et al. (2004). Expression of a novel
matrix metalloproteinase regulator, RECK, and its clinical
significance in resected non-small cell lung cancer. European
Journal of Cancer, 40(10), 1617–1623.
12. Takenaka, K., Ishikawa, S., Yanagihara, K., Miyahara, R.,
Hasegawa, S., & Otake, Y., et al. (2005). Prognostic signifi-
682Cancer Metastasis Rev (2007) 26:675–683
cance of reversion-inducing cysteine-rich protein with Kazal Download full-text
motifs expression in resected pathologic stage IIIA N2 non-
small-cell lung cancer. Annals of Surgical Oncology, 12(10),
13. Simizu, S., Takagi, S., Tamura, Y., & Osada, H. (2005). RECK-
mediated suppression of tumor cell invasion is regulated by
glycosylation in human tumor cell lines. Cancer Research, 65(16),
14. Sasahara, R. M., Takahashi, C., & Noda, M. (1999). Involvement
of the Sp1 site in ras-mediated downregulation of the RECK
metastasis suppressor gene. Biochemical and Biophysical Re-
search Communications, 264(3), 668–675.
15. Sasahara, R. M., Brochado, S. M., Takahashi, C., Oh, J., Maria-
Engler, S. S., & Granjeiro, J. M., et al. (2002). Transcriptional
control of the RECK metastasis/angiogenesis suppressor gene.
Cancer Detection and Prevention, 26(6), 435–443.
16. Lou, Z., O'Reilly, S., Liang, H., Maher, V. M., Sleight, S. D., &
McCormick, J. J. (2005). Down-regulation of overexpressed sp1
protein in human fibrosarcoma cell lines inhibits tumor formation.
Cancer Research, 65(3), 1007–1017.
17. Kanai, M., Wei, D., Li, Q., Jia, Z., Ajani, J., & Le, X., et al.
(2006). Loss of Krüppel-like factor 4 expression contributes to
Sp1 overexpression and human gastric cancer development and
progression. Clinical Cancer Research, 12(21), 6395–6402.
18. Liu, L. T., Chang, H. C., Chiang, L. C., & Hung, W. C. (2003).
Histone deacetylase inhibitor up-regulates RECK to inhibit MMP-
2 activation and cancer cell invasion. Cancer Research, 63(12),
19. Chang, H. C., Liu, L. T., & Hung, W. C. (2004). Involvement of
histone deacetylation in ras-induced down-regulation of the
metastasis suppressor RECK. Cellular Signalling, 16(6), 675–679.
20. Liu, L. T., Peng, J. P., Chang, H. C., & Hung, W. C. (2003).
RECK is a target of Epstein–Barr virus latent membrane protein 1.
Oncogene, 22(51), 8263–8270.
21. Hsu, M. C., Chang, H. C., & Hung, W. C. (2006). HER-2/neu
represses the metastasis suppressor RECK via ERK and Sp
transcription factors to promote cell invasion. Journal of Biolog-
ical Chemistry, 281(8), 4718–4725.
& Nakao, T., et al. (2001). RECK gene expression in hepatocellular
carcinoma: Correlation with invasion-related clinicopathological
factors and its clinical significance. Reverse-inducing-cysteine-rich
protein with Kazal motifs. Hepatology, 33(1), 189–195.
23. van der Jagt, M. F., Sweep, F. C., Waas, E. T., Hendriks, T.,
Ruers, T. J., & Merry, A. H., et al. (2006). Correlation of
reversion-inducing cysteine-rich protein with kazal motifs
(RECK) and extracellular matrix metalloproteinase inducer
(EMMPRIN), with MMP-2, MMP-9, and survival in colorectal
cancer. Cancer Letters, 237(2), 289–297.
24. Span, P. N., Sweep, C. G., Manders, P., Beex, L. V., Leppert, D.,
& Lindberg, R. L. (2003). Matrix metalloproteinase inhibitor
reversion-inducing cysteine-rich protein with Kazal motifs: A
prognostic marker for good clinical outcome in human breast
carcinoma. Cancer, 97(11), 2710–2715.
25. Riddick, A. C., Shukla, C. J., Pennington, C. J., Bass, R.,
Nuttall, R. K., & Hogan, A., et al. (2005). Identification of
degradome components associated with prostate cancer progres-
sion by expression analysis of human prostatic tissues. British
Journal of Cancer, 92(12), 2171–2180.
26. Brehmer, B., Biesterfeld, S., & Jakse, G. (2003). Expression of
matrix metalloproteinases (MMP-2 and -9) and their inhibitors
(TIMP-1 and -2) in prostate cancer tissue. Prostate Cancer and
Prostatic Diseases, 6(3), 217–222.
27. Masui, T., Doi, R., Koshiba, T., Fujimoto, K., Tsuji, S., &
Nakajima, S., et al. (2003). RECK expression in pancreatic
cancer: Its correlation with lower invasiveness and better
prognosis. Clinical Cancer Research, 9(5), 1779–1784.
28. Song, S. Y., Son, H. J., Nam, E., Rhee, J. C., & Park, C. (2006).
Expression of reversion-inducing-cysteine-rich protein with Kazal
motifs (RECK) as a prognostic indicator in gastric cancer.
European Journal of Cancer, 42(1), 101–108.
29. Kang, H. G., Kim, H. S., Kim, K. J., Oh, J. H., Lee, M. R., &
Seol, S. M., et al. (2007). RECK expression in osteosarcoma:
Correlation with matrix metalloproteinases activation and tumor
invasiveness. Journal of Orthopaedic Research, 25, 696–702.
30. van Lent, P. L., Span, P. N., Sloetjes, A. W., Radstake, T. R.,
van Lieshout, A. W., & Heuvel, J. J., et al. (2005). Expression and
localisation of the new metalloproteinase inhibitor RECK (rever-
sion inducing cysteine-rich protein with Kazal motifs) in inflamed
synovial membranes of patients with rheumatoid arthritis. Annals
of the Rheumatic Diseases, 64(3), 368–374.
31. Dass, C. R., & Choong, P. F. (2006). Biophysical delivery of
peptides: Applicability for cancer therapy. Peptides, 27(12),
32. Dass, C. R., & Choong, P. F. (2006). Carrier-mediated delivery of
peptidic drugs for cancer therapy. Peptides, 27(11), 3020–3028.
33. Ek, E. T., Dass, C. R., & Choong, P. F. (2006). Pigment
epithelium-derived factor: A multimodal tumor inhibitor. Molec-
ular Cancer Therapeutics, 5(7), 1641–1646.
34. Williams, D. A., & Baum, C. (2003). Medicine. Gene therapy—
New challenges ahead. Science, 302(5644), 400–401.
35. Dass, C. R., Walker, T. L., & Burton, M. A. (2002). Liposomes
containing cationic dimethyl dioctadecyl ammonium bromide:
Formulation, quality control, and lipofection efficiency. Drug
Delivery, 9(1), 11–18.
36. Dass, C. R., Walker, T. L., Kalle, W. H., & Burton, M. A. (2000).
A microsphere–liposome (microplex) vector for targeted gene
therapy of cancer. II. In vivo biodistribution study in a solid tumor
model. Drug Delivery, 7(1), 15–19.
37. Dass, C. R., & Burton, M. A. (2003). Modified microplex vector
enhances transfection of cells in culture while maintaining
tumour-selective gene delivery in-vivo. Journal of Pharmacy
and Pharmacology, 55(1), 19–25.
38. Dass, C. R. (2004). Cyclodextrins and oligonucleotide delivery to
solid tumours. Journal of Drug Targeting, 12(1), 1–9.
39. Dass, C. R., Contreras, K. G., Dunstan, D. E., & Choong, P. F.
(2007). Chitosan microparticles encapsulating PEDF plasmid
demonstrate efficacy in an orthotopic metastatic model of
osteosarcoma. Biomaterials, 28, 3026–3033.
40. Liu, L. T., Chang, H. C., Chiang, L. C., & Hung, W. C. (2002).
Induction of RECK by nonsteroidal anti-inflammatory drugs in
lung cancer cells. Oncogene, 21(54), 8347–8350.
41. Llovet, J. M., Burroughs, A., & Bruix, J. (2003). Hepatocellular
carcinoma. Lancet, 362(9399), 1907–1917.
42. Stojadinovic, A., Brooks, A., Hoos, A., Jaques, D. P., Conlon,
K. C., & Brennan, M. F. (2003). An evidence-based approach to
the surgical management of resectable pancreatic adenocarcinoma.
Journal of the American College of Surgeons, 196(6), 954–964.
Cancer Metastasis Rev (2007) 26:675–683683