Emaduddin M, Bicknell DC, Bodmer WF, Feller SM.. Cell growth, global phosphotyrosine elevation, and c-Met phosphorylation through Src family kinases in colorectal cancer cells. Proc Natl Acad Sci USA 105: 2358-2362
The heterogeneity of cancer cell signaling is a significant obstacle for the effective development and clinical use of molecularly targeted therapies. As a contribution to a better understanding of the diversity of signaling activities in colorectal cancers (CRCs), we have analyzed the activity of Src family kinases (SFKs), which are implicated in human cancer development, in 64 CRC cell lines. A striking diversity of SFK activity was observed within this panel. Importantly, all CRC lines tested depend on SFK activity for their growth. In addition, SFK activity levels strongly correlated with global levels of tyrosine-phosphorylated (pTyr) proteins in CRC lines. SFK inhibition substantially reduced these pTyr levels, suggesting that SFKs may function as signal integration points and master controllers for the pTyr protein status in CRC lines. The majority of analyzed CRC lines with high-SFK activity express activated c-Met (pYpY1234/1235), a receptor tyrosine kinase contributing to the regulation of cell proliferation, migration, and invasion. Inhibition of SFKs reduced c-Met phosphorylation in most cases, indicating a reversed signal flow from SFK to c-Met. We conclude that SFK activity is important for the growth of CRC lines, although only low activity levels are required. If this also is true for CRC patients, tumors with low-SFK activity may be particularly sensitive to SFK inhibitors, and such patients should be targeted in clinical trials testing SFK inhibitors. • colon • therapy • kinase inhibition • molecular heterogeneity
Cell growth, global phosphotyrosine elevation,
and c-Met phosphorylation through Src family
kinases in colorectal cancer cells
Muhammad Emaduddin*, David C. Bicknell
, Walter F. Bodmer
, and Stephan M. Feller*
*Cell Signalling Group and
Cancer and Immunogenetics Group, Weatherall Institute of Molecular Medicine, John Radcliffe Hospital, Headley Way,
Oxford OX3 9DS, United Kingdom
Contributed by Walter F. Bodmer, December 23, 2007 (sent for review December 12, 2007)
The heterogeneity of cancer cell signaling is a signiﬁcant obstacle
for the effective development and clinical use of molecularly
targeted therapies. As a contribution to a better understanding of
the diversity of signaling activities in colorectal cancers (CRCs), we
have analyzed the activity of Src family kinases (SFKs), which are
implicated in human cancer development, in 64 CRC cell lines. A
striking diversity of SFK activity was observed within this panel.
Importantly, all CRC lines tested depend on SFK activity for their
growth. In addition, SFK activity levels strongly correlated with
global levels of tyrosine-phosphorylated (pTyr) proteins in CRC
lines. SFK inhibition substantially reduced these pTyr levels, sug-
gesting that SFKs may function as signal integration points and
master controllers for the pTyr protein status in CRC lines. The
majority of analyzed CRC lines with high-SFK activity express
activated c-Met (pYpY1234/1235), a receptor tyrosine kinase con-
tributing to the regulation of cell proliferation, migration, and
invasion. Inhibition of SFKs reduced c-Met phosphorylation in most
cases, indicating a reversed signal ﬂow from SFK to c-Met. We
conclude that SFK activity is important for the growth of CRC lines,
although only low activity levels are required. If this also is true for
CRC patients, tumors with low-SFK activity may be particularly
sensitive to SFK inhibitors, and such patients should be targeted in
clinical trials testing SFK inhibitors.
colon 兩 therapy 兩 kinase inhibition 兩 molecular heterogeneity
enetic and epigenetic changes found in cancer cell genomes
lead to the deregulation of physiological-signaling path-
ways, as well as partial rewiring (signal spillover) and degrada-
tion (signal loss) of cellular-signaling networks. Moreover, in-
dividual tumors derived from the same cell type may have very
distinct sets of molecular lesions (1–5). This diversity makes it
desirable to develop molecularly targeted therapies that are
optimized with respect to the particular genetic and epigenetic
c onstitution of a patient’s cancer. Because cancers are hetero-
geneous in this respect, many individual cases need to be studied
to get an overall picture of the underlying pathological signals.
The functional consequences of genetic and epigenetic changes
are, however, sometimes not immediately apparent from DNA-
based or transcriptional studies but may be revealed by in-depth
analyses of signaling pathway activities and biological assays.
This study focuses on Src family kinases (SFKs) in a panel of
c olorectal cancer (CRC)-derived cell lines. CRC is a frequent
disease of ten diagnosed at a relatively late stage, when currently
available therapies are rarely curative (6). Molecularly targeted
therapies for CRC have recently entered into clinical use or are
currently in clinical trials (7). These therapies include kinase
inhibitors, such as dasatinib, that target SFKs.
Src, the first known and one of the most studied onc ogenes,
has been implicated in CRC development (8) and poor clinical
prognosis (9). Overexpression and hyperactivation of Src, not
apparently associated with mut ations in the Src gene, appears to
be predominant in CRC, although occasional mutations have
been reported (10, 11). In addition to Src, other SFK members,
such as Yes and Lck, also are activated or aberrantly expressed
in some CRC (12, 13).
Gene expression microarray data from 30 CRC lines have
provided evidence for the expression of mRNAs for the SFK
members Src, Yes, and Lyn in all cases and Fyn and Lck in some
CRC lines, but no evidence for the expression of Fgr, Hck, and
Blk (J. Wilding, D.C.B., and W.F.B., unpublished data; see also
ref. 14). Furthermore, some SFKs are known or suspected to
exist in several splice forms (15, 16), creating an additional layer
of c omplexity. This has led us to study the overall activity of SFK
proteins in CRC lines, rather than focus on a single SFK member
or isoform, as has been done in most prev ious studies. The
simult aneous analysis of all SFKs expressed in CRC is possible
because a key regulatory epitope of SFKs in the activation loop
of their cat alytic domains is well conserved (Fig. 1A), and
antibodies highly specific for this phosphoepitope [pY419 in
human c-Sc (UniProtKB/Swiss-Prot entry P12931) and pY416 in
chicken] have been generated. A genome-wide sequence search
of human DNA revealed that, apart from the tyrosine kinases
Btk (ca. 77 kDa) and Abl (ca. 145 kDa), which differ substantially
in size from all known SFK forms, no other proteins have high
homology to the sequence of the Y419 epitope.
Our results document not only a striking variability in the
activit y levels of SFKs in the CRC lines, but also show that some
SFK activity is required for the growth of all lines analyzed. The
results also indicate that SFKs play an important role in con-
trolling the overall t yrosine-phosphorylated (pTyr) levels in
CRC cells and that they may act as signal integration points for
multiple pathways. In addition, we show that SFK signals often
lead to the activation of c-Met, a multipotent receptor tyrosine
k inase implicated in a variety of cancers.
Tot al cell protein extracts from 64 CRC lines [see supporting
infor mation (SI) Table 1 for further details] were made 48 h after
the last medium change and were analyzed by Western blotting
for phosphorylation of the conserved key regulatory epitope of
human SFKs, designated pY419Src throughout this article (Fig.
1 A). The results shown in Fig. 1B reveal a striking heterogeneity
of SFK activit y not only in the absolute signal levels, but also in
the pattern of bands detected. The multiple bands observed in
some CRC lines could be due to differences in the SFKs
activated, but also may result from different splice variants or
phosphorylation states. Signals for pY419Src were seen in
Author contributions: M.E., W.F.B., and S.M.F. designed research; M.E., D.C.B., and S.M.F.
performed research; M.E., W.F.B., and S.M.F. analyzed data; and M.E., D.C.B., W.F.B., and
S.M.F. wrote the paper.
The authors declare no conﬂict of interest.
To whom correspondence may be addressed. E-mail: email@example.com or
This article contains supporting information online at www.pnas.org/cgi/content/full/
© 2008 by The National Academy of Sciences of the USA
February 19, 2008
no. 7 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0712176105
virtually all CRC lines after prolonged exposure (SI Fig. 6E),
raising the question as to whether all of the CRC lines depend,
at least to some extent, on basal SFK activity. To address this
issue, 16 CRC lines with either very high- or low-SFK activity
were selected and exposed to different doses of PP2, a well
characterized SFK inhibitor with a good selectivity profile (17).
The lines also were ex posed to a control compound void of SFK
inhibitory activity (PP3) or to solvent alone (DMSO). Repre-
sent ative data for four CRC lines are shown in Fig. 2A, and the
values for all of the lines studied are given in Fig. 2B.All
of the tested CRC lines are sensitive to PP2, as measured by
growth inhibition. However, the fact that even at high concen-
trations of PP2 there was no apparent reduction in cell number
suggested that there was no cell-killing effect of the inhibitor.
This lack of cell killing was confirmed by light microscopic
examination of treated and untreated cells. Cells w ith low-SFK
activit y are on average significantly more sensitive to PP2 than
cells with high-SFK activity (mean IC
value ⫽ 4.64
M ⫾ 1.22
M ⫾ 2.12, respectively; P ⫽ 0.0017). These results were
not anticipated and would seem to indicate that a relatively low
threshold level of SFK activity is sufficient to maintain the
growth of the CRC lines. Some cell lines also were analyzed with
a different SFK inhibitor, SU6656, and again were found to be
growth-inhibited (SI Fig. 7).
It is well k nown that oncogenic SFKs such as v-Src in
fibroblasts cause hyperphosphorylation of tyrosyl residues on
multiple cellular proteins (18). To investigate whether high-SFK
activit y also impacts on the global steady-state pTyr levels in the
CRC cells, a subset of 16 lines was analyzed by Western blotting
tot al cell proteins with anti-pTyr mAb (4G10). The results (see
Fig. 3) show that this antibody detects substantially more Tyr
phosphorylation in CRC lines with high- compared with low-
SFK activ ity. Similar results also were obtained from further
CRC lines and with another pTyr mAb (SI Fig. 8). The patterns
of pTyr bands observed in the CRC lines with high SFK are very
variable probably because of the different genetic and epigenetic
changes in the cell lines, reflecting the heterogeneit y of CRCs.
However, in some cell lines, similarly sized pTyr bands are
detected. Whether these are indeed identical proteins remains to
be investigated. To determine whether the c onstitutively ele-
vated pTyr levels in CRC lines with high-SFK activ ity are a
c onsequence of these kinase activities, such cells were treated
with PP2 or a c ompound solvent (DMSO) and analyzed by
Western blot w ith anti-pTy r mAb (P-Ty r-100) and anti-
pY419Src. The results (see Fig. 4) document a substantial
reduction of pTyr protein levels in each case. Similar data were
obt ained with SU6656 (SI Fig. 9). This finding raises the
interesting possibility that SFKs function as signal integration
points and master regulators of pTyr levels in the CRC lines.
Activation of SFKs by a variety of signaling pathways and in a
wide range of different cell types has been described. For example,
receptor tyrosine kinases such as EGFR, IGF1R, and c-Met (11, 19,
20) have been reported as important SFK activity regulators when
ligand-activated, activated by mutations, overexpressed, or trans-
activated through interactions with other signaling proteins (refs.
21–23 and references therein). To determine whether these recep-
tors may be important for the maintenance of constitutively ele-
vated SFK activities, EGFR, IGFR, and c-Met were immunopre-
cipitated (IP) and analyzed with phospho-specific antibodies
rec ognizing important, activit y-regulating t yrosine residues,
namely, pYpY1135/1136 for IGF1R and pYpY1234/1235 for c-
Met, or with a pTyr mAb for EGFR (P-Tyr-100). c-Met was found
to be phosphorylated in five of the six high-SFK activity cell lines
analyzed (Fig. 5 Upper), whereas little phosphorylation was de-
tected on EGFR and IGF1R tyrosines (SI Fig. 10). C32 cells have
high-SFK activity, but apparently do not express c-Met. No clear
correlation between c-Met protein abundance and its pYpY1234/
1235 phosphorylation was observed, which is suggestive of multiple
deregulating mechanisms for this kinase, in addition to gene
expre ssion changes.
To investigate whether c-Met activates SFKs or whether SFKs
af fect c-Met through a reversed signal flow, cells with detectable
c-Met pYpY1234/1235 were treated with PP2 to inhibit SFK, or
with a c-Met inhibitor, SU11274, and then analyzed for the
ef fects on c-Met or SFK activity, respectively. Treatment with
PP2 led to a significant reduction of c-Met pYpY1234/1235 in
four of five CRC lines analyzed (Fig. 5 Lower). COLO 320DM
cells, however, showed no reduction of c-Met phosphorylation
Fig. 1. Heterogeneity of steady-state SFK activity in a panel of 64 human CRC
cell lines. (A)(Upper) Schematic representation of the human c-Src domain
structure. Phosphorylation of Tyr-419 in the activation loop of the catalytic
domain (corresponds to Tyr-416 in chicken c-Src) is crucial for tyrosine kinase
activity. (Lower) Sequence alignment of all human SFK regions corresponding
to the Tyr-419 epitope in human c-Src. Names of SFKs with detectable mRNA
expression in CRC lines are in bold. Divergent amino acids in this conserved
epitope are boxed. Note the high degree of sequence identity in this key
regulatory epitope for all SFKs expressed in the CRC lines. (B) Western blots of
total cell protein extracts from 64 human CRC lines grown as described in
Methods were separated by SDS/PAGE and immunoblotted with the anti-
phosphoTyr419Src antibody that recognizes the SFK members expressed in
the CRC lines. Membranes were exposed to x-ray ﬁlm for 1–2 min. Standard-
ization of signal intensities between the four panels was done by using a single
batch of lysate from K562 CML cells, which are known to contain activated
SFKs. There is a striking degree of diversity in signal intensity and detectable
band sizes. For longer exposures of the same blots, see SI Fig. 6E.
Emaduddin et al. PNAS
February 19, 2008
on this epitope. The reason for this insensitivity is currently
unclear, but one of several possible ex planations would be the
presence of an activating c-Met mutation. No detectable effect
of SU11274 on global pTyr levels or SFK activity in any of the
five high-SFK activity lines with c-Met expression lines was
observed (SI Fig. 11). These results led us to conclude that
retrograde signaling occurs from SFK to c-Met, identifying
c-Met as a frequent effector protein in CRC lines with high
steady-st ate SFK activity.
The main hope for improved therapies for patients with more
advanced disease must come from novel molecularly targeted
drugs, such as those currently undergoing clinical evaluation. In
addition to the SFK inhibitor, dasatinib, which is already approved
for some advanced forms of chronic myelogenous leukemia (CML)
(24), other SFK inhibitors with activity in preclinical animal models
are currently being tested in patients. Therefore, the routine clinical
use of SFK inhibitors for several solid cancers may be only a few
years away. However, it is probable that most of these drugs will not
be highly specific for one particular SFK, but will affect multiple, if
not all, SFK family members. To make optimal use of SFK
inhibitors in the clinic, it will thus be important to obtain a better
understanding of the role of SFK activities in human cancer
development. The large CRC panel accumulated in the Cancer and
Immunogenetics Laboratory (Weatherall Institute of Molecular
Medicine) over the last two decades (25–27) is a valuable resource
for such studies and allows detailed investigation of the molecular
heterogeneity in CRC signaling protein activities. Based on our
results, we suggest that studying the overall SFK activity may be
more productive than focusing on a single SFK member. In support
of this suggestion, we have found that overall SFK activity levels do
not correlate well with c-Src activity (SI Fig. 12). Our results show
that many, if not all, CRC lines require SFK activity, although even
quite low levels seem to be sufficient for cell growth.
The dependence of the CRC lines on basal activity levels of
SFK may have sign ificant clinical implications. Thus, if low-SFK
activit y is sufficient for the proliferation of tumor cells in
patients, individuals who have low tumor SFK activity could be
Fig. 2. SFK activity is required for the growth of CRC lines: Enhanced sensitivity of low-SFK activity CRC lines to an SFK inhibitor. For growth assays, CRC lines
were incubated with 0.3% DMSO or the indicated concentrations of the SFK inhibitor PP2 (
M, micromolar) or the control compound PP3. Cell growth at the
indicated times was measured by crystal violet staining (shown as arbitrary units with cells at day zero as 100%). (A) Representative results of growth assays from
four CRC lines, two (C10 and CC20) with high-SFK activity and two (C75 and LoVo) with low-SFK activity are shown. Error bars (SE) are indicated but are usually
smaller than symbol size (n ⱖ 6). (B) The IC
values of PP2 in growth assays for 16 CRC lines on day 5 of incubation with compound. High-SFK activity is indicated
by a plus symbol; 11 of these lines were selected to represent different patterns observed with the pY419Src mAb (also see Fig. 1B). Statistical analysis of the
values indicates that CRC with low-SFK activity (mean value ⫽ 4.64
M ⫾ 1.22) is, on average, signiﬁcantly more sensitive to PP2 than CRC with
high-SFK activity (mean value ⫽ 17.38
M ⫾ 2.12; P ⫽ 0.0017).
Fig. 3. CRC lines with high-SFK activity show elevated global levels of cellular
tyrosine phosphorylation. (Upper) Total cell lysates from eight CRC lines with
either high- or low-SFK activity, as indicated, were immunoblotted, after
normalization for protein concentration, with anti-pTyr mAb (clone 4G10).
MWM, molecular weight marker. (Lower) The membrane was then stripped
and reprobed for actin (see SI Fig. 8 for a similar experiment).
www.pnas.org兾cgi兾doi兾10.1073兾pnas.0712176105 Emaduddin et al.
more sensitive to SFK inhibitors and so have a better therapeutic
window for the dr ug. It may therefore be quite inappropriate to
include only patients with elevated SFK activity in clinical trials
of SFK inhibitors.
The identification of a reversed signal flow from SFK to c-Met
is not entirely surprising because c-Met has long been known to
for m a complex with c-Src (28). However, other SFK members
are not well studied in this respect. c-Met is a multifunctional
receptor tyrosine k inase that can drive various biological pro-
cesses, so the actual c onsequences of c-Met activation by SFKs
in different CRC lines still need to be elucidated. Many questions
remain to be answered. For example, what are the underlying
genetic, epigenetic, and signaling protein activit y changes that
are responsible for the constitutively elevated SFK activities in
a substantial proportion of the CRC lines? It could be that other
receptor or nonreceptor tyrosine k inases, which we have not
analyzed so far, are responsible. Suppression of certain phos-
phat ases, either directly or indirectly by mutation or epigenetic
changes, also may lead to steady-state activation of SFKs (29).
Another unexpected finding that warrants further studies is the
predominant link between SFK and c-Met. Both EGFR and IGF1R
are known to physically and functionally interact with SFK in
various contexts. Furthermore, although c-Met is clearly affected by
SFK activity, it is by no means certain that this receptor kinase is
a key player in driving SFK-mediated CRC cell proliferation. For
example, Colo 320DM cells are sensitive to SFK inhibition, but this
sensitivity does not appear to affect the c-Met pYpY1234/1235
phosphorylation. However, because c-Met is a multipotent recep-
tor, other functional consequences of the SFK–c-Met interaction
remain to be explored. Whether c-Met is a direct target of SFK also
is not certain. Although Src and c-Met are known to form a complex
in different cells, thereby leading to a high local concentration of
both and potentially driving a direct phosphorylation, an involve-
ment of other tyrosine kinases cannot be formally excluded at this
point. Many of the proteins hyperphosphorylated in CRC lines with
high-SFK activity remain unidentified, and corresponding mass
spectrometric analyses, which may unravel important new SFK
targets, are urgently needed.
In this study, we have solely relied on inhibitor compounds to
analyze the roles of SFK in part because of the observation that
CRC cells are difficult to transfect with oligonucleotide siRNAs,
but also because a simultaneous k nockdown of multiple SFK
members, if technically feasible (for example with multicistronic
viral vectors), is likely to elicit some off-target effects that are not
dissimilar to those seen with many small-molecule inhibitors.
Hence, additional investigations are warranted into some of the
intriguing findings made in the course of this study. Further
improved understanding of the SFK-regulated signaling path-
Fig. 4. Elevated global phosphotyrosine levels in CRC lines with high-SFK
activity are greatly reduced by SFK inhibition. Six CRC lines with high-SFK
activity were grown for 24 h after splitting and then treated with either 0.3%
DMSO (D) or 30
M SFK inhibitor PP2 (P), which were added into the medium
for a further 24 h. Total cell lysates were separated by SDS/PAGE and then
immunoblotted with anti-pTyr mAb (P-Tyr-100), anti-pY419Src, or anti-actin
as indicated. The signals were quantiﬁed as described in Methods, and ob-
served signal reductions upon PP2 treatment of cells are indicated in percent-
age below the corresponding lanes. A similar experiment conducted with the
SFK inhibitor SU6656 is shown in SI Fig. 9.
Fig. 5. High-SFK activity increases c-Met receptor tyrosine kinase phosphory-
lation at the key regulatory site Y1234/Y1235. Six cell lines each with either high-
or low-SFK activity were analyzed for the level of protein expression and persist-
ing phosphorylation of the c-Met receptor 48 h after their last feed with fresh
medium. For phospho-status analysis at the c-Met key regulatory site, Y1234/
Y1235, c-Met was immunoprecipitated from 1 mg of total cell lysate and blotted
with the phosphoepitope-speciﬁc antibody (uppermost) or anti-c-Met to control
for total c-Met in the IPs (second from top). Protein expression levels of c-Met in
total cell lysates (TCL) are shown (third from top). The asterisk indicates partially
processed c-Met protein in LoVo cells as described previously (32). Total cell lysates
also were blotted for actin as a loading control. The ﬁve CRC lines with detectable
c-Met phosphorylation on the Y1234/Y1235 epitope were subsequently treated
with PP2 inhibitor (P) or solvent control (DMSO; D). Four of the ﬁve lines showed
a substantial reduction of c-Met phosphorylation at the Y1234/Y1235 epitope
after incubation with PP2 (bottom). No changes were detected in COLO 320DM
cells. Actin was also analyzed in TCL (second from bottom).
Emaduddin et al. PNAS
February 19, 2008
ways in CRC will undoubtedly increase the range of novel
molecular strategies and targets for the development of im-
proved therapies that have greater specificity for a given cancer
and fewer side effects.
CRC Lines and Cell Lysates. The 64 lines in the panel have been previously
HLA-typed and characterized for other genetic changes to determine whether
they are derived from cancers of different donors. They are derived from 63
different patients; LS 174T and LS 180 originate from the same individual (30).
Information about line origins is given in SI Table 1. OXCO-1 and OXCO-3 lines
are a gift from Khoon Lin Ling and Vincenzo Cerundolo (Weatherall Institute
of Molecular Medicine). CRC lines were grown in DMEM supplemented with
L-Glu, antibiotics, and 10% FBS at 37°C in humidiﬁed atmosphere with
. Cells were cultured for 48 h from the last feeding before lysis,
washed three times with chilled PBS, and lysed in a RIPA-type buffer [20 mM
Tris䡠HCl (pH 7.5), 100 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.5% deoxy-
cholic acid, and 0.1% SDS] supplemented with 2x complete protease inhibitor
mix (11697498001; Roche Diagnostics) and phosphatase inhibitor cocktails 1
and 2 (P2850 and P5726; Sigma–Aldrich) at 4°C for 30 min on a nutator. Lysates
were then clariﬁed by centrifugation at 22,000 ⫻ g for 30 min at 4°C. Protein
concentrations were determined by the Bradford method (31), and Coomassie
blue staining of SDS/PAGE-separated proteins was used to further ensure
approximately equal loading of samples (SI Fig. 6 A–D).
Western Blot Analyses and IPs. Total lysates or IPs were separated by SDS/PAGE
and transferred onto Hybond-P (Amersham). Membranes were blocked over-
night at 4°C with BSA or nonfat dry milk in TBST [20 mM Tris䡠HCl (pH 7.5), 100
mM NaCl, and 0.1% Tween 20] as recommended by the antibody manufac-
turer, followed by incubation with primary antibody for 4 – 8 h at room
temperature. After washing extensively with TBST, bound antibodies were
detected by using peroxidase-conjugated anti-mouse or anti-rabbit IgG sec-
ondary antibody and ECL.
The pY419Src (pAb, 2101; Cell Signaling Technology) recognizes most
members of the SFK family (see Fig. 1 for details). Two different anti-pTyr
mAbs, 4G10 and P-Tyr-100 (9411; Cell Signaling Technology), were used as
indicated. Actin expression was analyzed by using a mAb (clone AC-40; Sigma–
Aldrich). IPs were carried out with 1 mg of total cell lysate by using anti-c-Met
mAb (clone DO-24; Upstate), anti-IGF1R pAb (3027; Cell Signaling Technol-
ogy), or anti-EGFR mAb (clone 528; Calbiochem). Western blots of IPs were
performed with anti-c-Met mAb (clone DO-21; Upstate), anti-phospho-c-Met
(pYpY1234/1235), pAb (3126; Cell Signaling Technology), anti-phospho-IGF1R
(pYpY1135/1136), pAb (3024; Cell Signaling Technology), anti-IGF1R pAb
(3027; Cell Signaling Technology), anti-EGFR pAb (2232; Cell Signaling Tech-
nology), or P-Tyr-100 as indicated.
Cell Growth Assays. Brieﬂy, 1,000–2,000 cells were seeded into 96-well plates,
allowed to adhere for 1 day, and either left in medium alone or incubated in
medium with DMSO (inhibitor solvent; D2650; Sigma) or different concentra-
tions of SFK inhibitor PP2 (529573; Calbiochem) or the structural analogue
PP3, which lacks SFK inhibitory activity (529574; Calbiochem). Cell growth at
the indicated times was assessed by a colorimetric assay using crystal violet
(C3886; Sigma–Aldrich). Cells were stained and ﬁxed in 0.1% crystal violet
dissolved in 50% methanol for 15 min at room temperature before being
washed three times with PBS and once with H
O and then air dried. Bound dye
was solubilized in 10% acetic acid for 30 min at room temperature, and
absorbance was measured at 560 nm by using a
-Quant microplate reader
(Bio-Tek). Cell growth was expressed as a percentage of control cells (in
medium alone at day 0). Similar experiments were conducted with the SFK
inhibitor SU6656 (572635; Calbiochem) for some cell lines (SI Fig. 7).
Inhibition of SFK and c-Met for Phospho-Protein Analyses. CRC lines were
incubated as indicated with PP2 or SU11274 (448101; 1
M for 24 h; Calbiochem)
for c-Met inhibition, lysed, and analyzed by Western blotting as described in
Western Blot Analyses and IPs. Note that DMSO concentrations of ⬎0.06%
affected the phospho-status of c-Met at least in some CRC lines. Therefore, in all
experiments analyzing c-Met phosphorylation, DMSO concentrations were kept
Quantification of Inhibitor Effects on Cell Growth or Protein Phosphorylations.
values were calculated by using Prism software (GraphPad) by comparing
PP2-treated cells to DMSO-treated controls. Levels of protein phosphorylation
were determined by using Alpha Scan (Alpha Innotech), and data were
analyzed with ImageQuant (Molecular Dynamics). The P value for the differ-
ence between the mean IC
values for the lines with high- versus low-SFK
activity was determined by Student’s t test.
ACKNOWLEDGMENTS. We thank our colleagues who provided cell lines to
make this study possible, especially Khoon Lin Ling and Enzo Cerundolo for
providing the newly established OXCO lines before publication, and Jenny
Wilding and Val Macaulay for useful discussions and sharing unpublished
information. This work was supported by a Cancer Research U.K. program
grant (to W.F.B.) and a Heads Up (U.K. cancer charity) core grant (to S.M.F.).
1. Smith G, et al. (2002) Mutations in apc, kirsten-ras, and p53-alternative genetic
pathways to colorectal cancer. Proc Natl Acad Sci USA 99:9433–9438.
2. Wood LD, et al. (2007) The genomic landscapes of human breast and colorectal cancers.
Science 318:1108 –1113.
3. Fearon ER, Vogelstein B (1990) A genetic model for colorectal tumorigenesis. Cell
4. Bodmer W (1997) The somatic evolution of cancer. The harveian oration of 1996. JR
Coll Physicians Lond 31:82–89.
5. Bodmer WF (2006) Cancer genetics: Colorectal cancer as a model. J Hum Genet 51:391–396.
6. Venook A (2005) Critical evaluation of current treatments in metastatic colorectal
cancer. Oncologist 10:250 –261.
7. Hwang J, Marshall JL (2006) Targeted therapy for colorectal cancer. Curr Opin Investig
8. Summy JM, Gallick GE (2003) Src family kinases in tumor progression and metastasis.
Cancer Metastasis Rev 22:337–358.
9. Aligayer H, et al. (2002) Activation of src kinase in primary colorectal carcinoma: An
indicator of poor clinical prognosis. Cancer 94:344 –351.
10. Irby RB, et al. (1999) Activating src mutation in a subset of advanced human colon
cancers. Nat Genet 21:187–190.
11. Ishizawar R, Parsons SJ (2004) C-src and cooperating partners in human cancer. Cancer
Cell 6:209 –214.
12. Veillette A, Foss FM, Sausville EA, Bolen JB, Rosen N (1987) Expression of the lck tyrosine
kinase gene in human colon carcinoma and other non-lymphoid human tumor cell
lines. Oncogene Res 1:357–374.
13. Park J, Meisler AI, Cartwright CA (1993) C-yes tyrosine kinase activity in human colon
carcinoma. Oncogene 8:2627–2635.
14. Hirsch CL, Smith-Windsor EL, Bonham K (2006) Src family kinase members have a
common response to histone deacetylase inhibitors in human colon cancer cells. Int J
15. Davidson D, Chow LM, Fournel M, Veillette A (1992) Differential regulation of t cell
antigen responsiveness by isoforms of the src-related tyrosine protein kinase p59fyn.
J Exp Med 175:1483–1492.
16. Yi TL, Bolen JB, Ihle JN (1991) Hematopoietic cells express two forms of lyn kinase
differing by 21 amino acids in the amino terminus. Mol Cell Biol 11:2391–2398.
17. Bain J, McLauchlan H, Elliott M, Cohen P (2003) The speciﬁcities of protein kinase
inhibitors: An update. Biochem J 371:199 –204.
18. Martin GS (2001) The hunting of the src. Nat Rev Mol Cell Biol 2:467– 475.
19. Hynes NE (2000) Tyrosine kinase signalling in breast cancer. Breast Cancer Res 2:154–157.
20. Birchmeier C, Birchmeier W, Gherardi E, Vande Woude GF (2003) Met, metastasis,
motility and more. Nat Rev Mol Cell Biol 4:915–925.
21. Buchanan FG, et al. (2006) Role of beta-arrestin 1 in the metastatic progression of
colorectal cancer. Proc Natl Acad Sci USA 103:1492–1497.
22. Sekharam M, Nasir A, Kaiser HE, Coppola D (2003) Insulin-like growth factor 1 receptor
activates c-src and modiﬁes transformation and motility of colon cancer in vitro.
Anticancer Res 23:1517–1524.
23. Herynk MH, Zhang J, Parikh NU, Gallick GE (2007) Activation of src by c-met overexpression
mediates metastatic properties of colorectal carcinoma cells. J Exp Ther Oncol 6:205–217.
24. Jabbour E, Cortes JE, Giles FJ, O’Brien S, Kantarjian HM (2007) Current and emerging
treatment options in chronic myeloid leukemia. Cancer 109:2171–2181.
25. Browning MJ, et al. (1993) Tissue typing the HLA-A locus from genomic DNA by
sequence-speciﬁc pcr: Comparison of HLA genotype and surface expression on colo-
rectal tumor cell lines. Proc Natl Acad Sci USA 90:2842–2845.
26. Wheeler JM, et al. (1999) Mechanisms of inactivation of mismatch repair genes in
human colorectal cancer cell lines: The predominant role of hmlh1. Proc Natl Acad Sci
USA 96:10296 –10301.
27. Liu Y, Bodmer WF (2006) Analysis of p53 mutations and their expression in 56 colorectal
cancer cell lines. Proc Natl Acad Sci USA 103:976 –981.
28. Ponzetto C, et al. (1994) A multifunctional docking site mediates signaling and transfor-
mation by the hepatocyte growth factor/scatter factor receptor family. Cell 77:261–271.
29. Wang Z, et al. (2004) Mutational analysis of the tyrosine phosphatome in colorectal
cancers. Science 304:1164 –1166.
30. Tom BH, et al. (1976) Human colonic adenocarcinoma cells. I. Establishment and
description of a new line. In Vitro 12:180 –191.
31. Bradford MM (1976) A rapid and sensitive method for the quantitation of microgram
quantities of protein utilizing the principle of protein-dye binding. Anal Biochem
32. Mondino A, Giordano S, Comoglio PM (1991) Defective posttranslational processing
activates the tyrosine kinase encoded by the met proto-oncogene (hepatocyte growth
factor receptor). Mol Cell Biol 11:6084 – 6092.
www.pnas.org兾cgi兾doi兾10.1073兾pnas.0712176105 Emaduddin et al.