Carcinogenesis vol.31 no.9 pp.1531–1540, 2010
doi:10.1093/carcin/bgq133
Advance Access publication June 27, 2010
The PI3K–Akt mediates oncogenic Met-induced centrosome amplification and
chromosome instability
Hyun-Ja Nam1,2, Sunyoung Chae1,2, Seung-Hoon Jang1,2,
Hyeseong Cho1,2 and Jae-Ho Lee1,2,
1
Department of Biochemistry and Molecular Biology, Ajou University School
of Medicine and 2Department of Molecular Science and Technology,
Graduate School of Ajou University, Suwon 443-721, Korea
To whom correspondence should be addressed. Department of Biochemistry
and Molecular Biology, Ajou University School of Medicine, Suwon 443-721,
Korea. Tel: þ82 31 219 5053; Fax: þ82 31 219 5059;
Email: [email protected]
Correspondence may also be addressed to Hyeseong Cho.
Tel: þ82 31 219 5052; Fax: þ82 31 219 5059;
Email: [email protected]
The oncogenic ability of aberrant hepatocyte growth factor receptor (Met) signaling is thought to mainly rely on its mitogenic and
anti-apoptotic effects. Recently, however, cumulating evidences
suggest that genomic instability may be a crucial factor in tumorigenesis. Here, we address whether oncogenic Met receptor is
linked to the centrosome abnormality and genomic instability.
We showed that expression of the constitutive active Met (CAMet) induced supernumerary centrosomes probably due to
deregulated centrosome duplication, which was accompanied with
multipolar spindle formation and aneuploidy. Interestingly,
LY294002, a phosphoinositide 3-kinase (PI3K) inhibitor, significantly suppressed the appearance of supernumerary centrosomes.
Moreover, knockdown of Akt with small interfering RNAs and
overexpression of phosphatase and tensin homolog or dominantnegative Akt abrogated supernumerary centrosome formation,
evidencing the involvement of PI3K signaling. We further showed
that expression of CA-Met significantly increased aneuploidy in
p532/2 HCT116 cells, but not in p531/1 HCT116 cells, indicating
that the ability of CA-Met to induce chromosomal instability
(CIN) phenotype is related with p53 status. Together, our data
demonstrate that aberrant hepatocyte growth factor/Met signaling induces centrosome amplification and CIN via the PI3K–Akt
pathway, providing an example that oncogenic growth factor signals prevalent in a wide variety of cancers have cross talks to
centrosome abnormality and CIN.
Introduction
The Met tyrosine kinase is a high-affinity receptor for hepatocyte
growth factor (HGF) (1,2). Both Met and HGF are expressed in various tissues, and signaling via this receptor–ligand pair has been
shown to affect a variety of biological activities, including cell growth
(3,4), cellular motility (5) and angiogenesis (6). Aberrant activation of
HGF/Met signaling via paracrine or autocrine activation has been
implicated in the generation and spread of many types of human
cancers, including major carcinomas as well as sarcomas (7). Notably,
activating missense mutants of Met are found in human hereditary
papillary renal carcinomas (8) and these mutants have been shown
to possess constitutive kinase activity and transforming ability (9).
Activating mutations of Met are also found in ovarian cancer (10),
early-onset hepatocellular carcinoma (11) and gastric carcinoma (12),
Abbreviations: CA-Met, constitutive active Met; CIN, chromosomal
instability; Cdk2, cyclin-dependent kinase 2; DN-Akt, dominant-negative form
of Akt; ERK, extracellular signal-regulated kinase; HGF, hepatocyte growth
factor; JNK, c-jun N-terminal kinase; MAPK, mitogen-activated protein
kinase; PTEN, phosphatase and tensin homolog; PBS, phosphate-buffered
saline; PI3K, phosphoinositide 3-kinase; RTK, receptor tyrosine kinase; RT,
room temperature; siRNAs, small interfering RNAs.
indicating that the activation mutation of Met is a general mechanism
causing various human cancers.
Growth factors/growth factor receptors play a critical role in initiating signaling cascades that stimulate cell cycle progression at G1/S
phase. Oncogenic receptor tyrosine kinases (RTKs) derived
from mutant growth factor receptors allow cells to pass over the G1/S
checkpoint, triggering uncontrolled cell proliferation. The Ras–
extracellular signal-regulated kinase (ERK) and phosphoinositide
3-kinase (PI3K)–Akt are major signaling activators that are responsible
for uncontrolled cell proliferation as well as for anti-apoptotic signals
(13,14). So far, the ability of oncogenic RTKs to transform cells has
been thought to mainly rely on their mitogenic and anti-apoptotic effects (15). Recently, however, evidence is accumulating that genomic
instability is a crucial factor in tumorigenesis (16,17). Interestingly,
some of the key signaling molecules related to growth factors have
been reported to induce genomic instability. The activated Ras increased genomic instability of thyroid cells in which the Ras-mediated
cell transformation seemed to occur after acquisition of the impaired
DNA damage response (18). In addition, aberrant Wnt/b-catenin signaling, an important driving force of tumorigenesis in many cancers,
accompanied with chromosomal instability (CIN) in cells (19), suggesting that growth signaling molecules not only control cell cycle
progression at G1/S phase but may also extend to mitotic regulation.
CIN is characterized by losses or gains of chromosome number and
changes in chromosome structure (20) and can be caused by an increased rate of chromosomal missegregation during mitosis (21).
Abnormal number of centrosome, a microtubule-organizing center,
is often associated with the multipolar mitotic spindles that consequently cause chromosomal missegregation (22,23). Centrosome amplification is considered as a major contributing factor for CIN in
cancer development because a significant fraction of human cancers
commonly show centrosome defects (20,22,24,25). Cellular kinases
of the cyclin-dependent kinase 2 (Cdk2)/cyclin E, Aurora, Polo and
Nek families participate in the centrosome duplication cycle (26) and
amplification or mutations of those kinases increase the frequency of
centrosome abnormality (27). So far, whether oncogenic signaling
derived from aberrant growth factor/growth factor receptors are
linked to the centrosome abnormality remains largely unknown.
In the present study, we first demonstrated that expression of oncogenic Met receptor in cells led to supernumerary centrosomes and
aneuploidy. We showed that inhibition of Akt activity in these cells by
using Akt small interfering RNAs (siRNAs) or overexpression of
dominant-negative form of Akt (DN-Akt) abrogated supernumerary
centrosome formation. Thus, oncogenic growth factor/growth factor
receptor signaling may have cross talk to centrosome abnormality and
genomic instability during tumorigenesis.
Materials and methods
Antibodies, plasmids and chemicals
Fetal bovine serum, trypsin–ethylenediaminetetraacetic acid, antibiotic–
antimycotic containing penicillin G sodium, streptomycin sulfate and amphotericin B were obtained from Gibco BRL (Carlsbad, CA). Protease inhibitor
cocktail was purchased from Roche Applied Science (Mannheim, Germany).
Phosphatase inhibitors, nocodazole, U0126 and LY294002 were from Sigma
Chemical Co. (St Louis, MO). Rabbit polyclonal antibodies against the following antigens were purchased from Cell Signaling Technology (Beverly,
MA): phopho-p44/42 mitogen-activated protein kinase (MAPK), p44/42
MAPK, phospho-p38 MAPK, p38 MAPK, phospho-stress-activated protein
kinase/c-jun N-terminal kinase (JNK), stress-activated protein kinase/JNK,
phospho-Akt and Akt. Mouse monoclonal c-tubulin antibody was purchased
from Sigma Chemical Co.. Antibody against the N-terminal half of Cep-170
was generously provided by Dr G.Guarguaglini (University of Rome, Rome,
Italy). CA-Met construct, which is wild-type murine Met expression vector,
was provided by Dr G.F.VandeWoude (Van Andel Research Institute). SRT
Ó The Author 2010. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]
1531
H.-J.Nam et al.
and SRT-tagged WT-Akt constructs were provided by Dr Myong-Joon Hanh
(Sungkyunkwan University School of Medicine).
Cell culture
HeLa cells purchased from American Type Culture Collection (Manassas, VA) and
HCT116 cells generously provided by Dr B.Vogelstein (Johns Hopkins University,
Baltimore, MD) were cultured in Dulbecco’s modified Eagle’s medium nutrient
mixture F-12 HAM (Sigma Chemical Co.) supplemented with 10% (vol/vol) fetal
bovine serum (Gibco BRL) in 37°C incubator with 5% CO2 in air. To generate
HeLa cells overexpressing CA-Met Met, cells were co-transfected with a plasmid
pSV2neo (28), which confers resistance to G418, and a plasmid encoding the
pMB1-CA-Met Met. MOCK cells were transfected with plasmid pSVneo and
pMB1 empty vector. Cells were selected for 2 weeks with 800 lg/ml G418 (Life
Technologies, Carlsbad, CA) and then grown as pools with 500 lg/ml G418.
Cell synchronization
For mitosis arrest, cells were treated with 100 ng/ml nocodazole for 12 h.
To arrest the cells at G1–S boundary, double thymidine block was performed.
Cells were treated with 1 lM thymidine for the first synchronization. Twenty
hours later, cells were washed with thymidine-free medium (first release) and
cultured in complete medium for 8 h. Then, cells were cultured again in
thymidine-containing medium for 14 h for the second round of synchronization. After then, cells were washed with thymidine-free medium and cultured
in complete medium (second or final release).
Cell cycle analysis by flow cytometry
Cells synchronized at G1–S boundary by double thymidine block were trypsinized, centrifuged and fixed with 70% cold ethanol for 30 min. After samples
were washed with phosphate-buffered saline (PBS) and then resuspended in
a solution containing RNase A for 5 min, propidium iodide was added. After
5 min of incubation, samples were subjected to FACS analysis using a FACS
Vantage flow cytometer (Becton Dickinson, Franklin Lakes, NJ).
Giemsa staining
MOCK and CA-Met cells were exposed to 100 ng/ml nocodazole for 12 h, and
mitotic cells were then collected, incubated at 37°C for 15 min in 0.075 M
hypotonic KCl, pelleted and fixed with methanol:acetic acid (3:1) solution.
After washing three times with PBS, samples were spreaded on slide glass,
dried at room temperature (RT) and then stained with 3% (wt/vol) Giemsa
staining solution.
Immunoblotting
Conventional immunoblotting was performed as previously described (29)
using corresponding antibodies. Briefly, cell lysates (50 lg) were resolved
by sodium dodecyl sulfate–polyacrylamide gel electrophoresis and then were
transferred to nitrocellulose membrane. After blocking for 1 h at RT with PBS
containing 0.05% (vol/vol) Tween-20 and 5% (wt/vol) nonfat milk, membranes were incubated with primary antibodies at 4°C, followed by washing
with PBS containing 0.05% Tween-20 and incubated with horseradish
peroxidase-conjugated anti-rabbit or anti-mouse antibody (Amersham
Biosciences, Piscataway, NJ) for 1 h at RT. Detection was carried out using
ECL reagents (Amersham Biosciences) and exposing them to x-ray film.
Immunofluoresence analysis
Cells were fixed in methanol:acetone (1:1) solution and permeabilized with
0.075% Triton X-100. Fixed cells were pre-incubated in blocking solution (1%
bovine serum albumin in PBS), followed by incubation overnight with primary
antibodies at 4°C. Cells were then washed three times with shaking and probed
with a rhodamine-conjugated anti-rabbit antibody (1:1000; Jackson ImmunoResearch Laboratories, West Grove, PA) or fluorescein isothiocyanateconjugated anti-mouse antibody for 1 h at RT. After washing, cells were mounted
in the mounting solution containing 4#,6-diamidino-2-phenylindole and examined by confocal microscope (LSM510; Carl Zeiss, Oberkochen, Germany).
Knockdown experiment
Pre-designed siRNA oligonucleotides targeting for Akt (Akt1 and Akt2) were
purchased from Life Technologies. HeLa cells (1 104) were seeded in
35 mm dishes. After 24 h, cells were transfected with 200 nmol each of specific
or non-silencing control siRNA oligonucleotides using Oligofectamine according to manufacturer’s instruction (Life Technologies).
Results
Constitutive activation of Met signaling induced supernumerary
centrosomes and multipolar spindle formation
In order to address whether oncogenic Met tyrosine kinase receptor
is linked to centrosome abnormality, we utilized the M1268T mu-
1532
tant, a strong constitutively activating form of Met that was originally found in human papillary renal cancer (8). Stable cell lines
were established by transfecting pMB1-M1268T complementary
DNA into HeLa cells, followed by selection with G418.
Among .50 clones, different clones (#1, #3 and #5) expressing
M1268T were selected and analyzed for the Met activation.
To determine the status of tyrosine phosphorylation of Met in these
clones, cell lysates were subjected to immunoprecipitation using
anti-murine Met antibody, followed by western blot analysis with
anti-phosphotyrosine antibody. All these clones exhibited significant increases in the levels of tyrosine phosphorylation of Met
(Figure 1A), verifying constitutively active states of Met (CAMet). Under microscope, all these clones also showed similar scattered growth pattern (data not shown), which is an indicator of cells
with Met activation (30,31).
We then determined the number of centrosomes in asynchronously
growing cells by staining for c-tubulin and counting .200 cells in
each experiment. We found that 2% of control cells transfected with
the pMB1 vector contained supernumerary centrosomes (n 3). In
contrast, 6–9% of cells expressing CA-Met showed supernumerary
centrosomes (Figure 1B), which was 3- to 4-fold higher than that of
control cells (P , 0.005). Since different CA-Met clones did not
yield differences in the activation of Met and in the formation of
supernumerary centrosomes, we rationalized that these CA-Met
clones can be pooled for further analysis.
At mitotic phase, centrosome duplication and segregation are completed and therefore, it is better to determine supernumerary centrosomes in mitotic cells. To count the number of centrosomes in mitotic
phase, cells were arrested at prometaphase by nocodazole treatment
and then released from the drug. At 60 min after release when most of
cells remained at mitosis, numbers of centrosomes in each cell were
determined by staining for c-tubulin (Figure 1C, upper panel).
Whereas 5% of the control cells contained .2 centrosomes (supernumerary centrosomes), 25% of cells expressing CA-Met contained
supernumerary centrosomes, which was significantly higher than that
of control cells (Figure 1C, lower panel). In addition, tripolar and
tetrapolar mitotic spindles with multidirectional chromosomal staining were significantly increased in CA-Met cells (15.8 ± 2.7% versus
2.4 ± 0.5%, P , 0.005) as shown in Figure 1D. Furthermore, chromosome segregation defects such as bridge chromosomes in anaphase
cells were often observed (Figure 1E, upper panel), and the frequency
was significantly higher than that of control cells (P , 0.005)
(Figure 1E, lower panel).
Supernumerary centrosomes can be generated either by multiple
cycles of centrosome duplication in a single cell cycle or by cytokinesis failure itself without overduplication of centrosomes (20). Cells
with supernumerary centosomes often fail to undergo cytokinesis due
to formation of aberrant spindles. In turn, the cytokinesis block plays
a secondary role in generation of supernumerary centrosomes. To
investigate the possibility of centrosome duplication, we analyzed
Cep170 protein, which localizes to only one mature centriole from
G1 to early G2 phases, whereas two Cep170-positive centrioles can be
shown through late G2 to mitotic phase as two separated centrosomes
(32). In cells showing four centrosomes due to S phase arrest by
hydroxyurea treatment, most contained one to three Cep170-positive
staining and no cells showed four Cep170-positive staining (supplementary Figure 1 is available at Carcinogenesis Online). In contrast,
cytokinesis failure by itself induced by cytochalasin D treatment
greatly increased cells with four Cep170-positive staining up
to .70%. Meanwhile, 80% of CA-Met cells contained one to three
Cep170-positive staining, whereas ,20% of them displayed four
Cep170-positive staining (supplementary Figure 1 is available at
Carcinogenesis Online). These results indicate that supernumerary
centrosomes in CA-Met cells are mainly driven by overduplication
of centrosomes during cell cycle and partly by division failure.
Together, this is the first indication that constitutive active Met signaling induces supernumerary centrosomes probably via deregulation
of centrosome duplication cycle, which is probably to contribute to
multipolar spindle formation and chromosomal bridge formation.
Met-mediated CIN via Akt activation
Fig. 1. Centrosome amplification and aberrant mitotic spindle formation accompanied with anaphase bridge formation in CA-Met-expressing cells.
(A) HeLa cells were transfected with empty vector or the constitutive active mutant Met construct M1268T-Met complementary DNA (CA-Met) and survived
clones were selected in the presence of G418. Cell lysates of control and different clones transfected with M1268T (#1, #3 and #5) were extracted using RIPA
buffer and subjected to immunoprecipitation using anti-murine Met antibody (Met), followed by western-blot analysis with either anti-phosphotyrosine antibody
(pTyr) or anti-murine Met antibody (lower panel). Note: anti-murine Met antibody does not cross-react with endogenous human Met protein. (B) Asynchronously
growing cells were seeded onto cover glasses and the numbers of centrosomes were determined by staining with anti-c-tubulin antibody. P , 0.005 by
Student’s t-test. (C) CA-Met clones were pooled and synchronized by treatment with 100 ng/ml nocodazole for 12 h. After removal of nocodazole, cells were
reseeded onto poly-L-lysine-coated cover glasses and allowed to progress through the cell cycle for 60 min. The centrosomes were visualized by staining with antic-tubulin antibody (upper panel). P , 0.0005 by Student’s t-test (lower panel). (D) Polar spindles and chromosomes were visualized by staining with antia-tubulin antibody and 4#,6-diamidino-2-phenylindole, revealing bi-, tri- and tetrapolar spindles (upper panel). P , 0.005 by Student’s t-test (lower panel).
(E) Increase of cells with anaphase bridge by CA-Met expression. Cells were synchronized at the G1/S boundary by double thymidine block. Cells at the late
M phase (9, 10 and 11 h after release) were immunostained with anti-a-tubulin antibody and propidium iodide (upper panel). P , 0.005 by Student’s t-test
(lower panel). All experiments from B to E were carried out by counting 300 cells and expressed as mean ± SD from three independent experiments.
CA-Met enhanced CIN
It has been shown that cells with severe aberrant chromosomes do not
survive through cell cycle progression, whereas cells with relatively
mild alterations in chromosome survive, increasing the chance
for accumulation of CIN (33). To address whether the cells with
supernumerary centrosomes and multipolar spindle survived and revealed accumulation of CIN in interphase cells, we measured the
frequency of multinuclei (Figure 2A, a’) and micronuclei formation
(Figure 2A, b’) that would be derived from aberrant chromosomal
segregation during mitosis. As shown in the right panel of Figure
2A, numbers of multinucleated cells and micronuclei-containing cells
were significantly increased in CA-Met cells compared with control
cells (P , 0.05). The same phenomenon was also observed in cells
transiently transfected with M1268T expression vector (Figure 2B),
showing supernumerary centrosomes and subsequent multinuclei/
micronuclei formation. Moreover, we observed that repetitive expo-
sure to HGF induces supernumerary centrosomes in WI-38 human
fibroblast cells (supplementary Figure 2 is available at Carcinogenesis
Online), evidencing that prolonged activation of HGF/Met signaling
contributes to accumulation of CIN.
Increasing extent of aneuploidy is one of the most representative
indicator of CIN (34). Using metaphase spread, we first determined
numbers of chromosomes from vector control cells (n 5 300) and
then divided them into three categories—N1, N2 and N3
(Figure 2C). The average number of chromosomes per one cell
was determined as 59.2 in parental HeLa cells. To quantitatively
express the extent of newly generated aneuploidy, we formulated,
so called, an ‘aneuploidy index’. The ‘N2’ window was defined as
the window around the peak that 95% of cell population from parental HeLa cells belonged to and the aneuploidy index was defined
as [(N1 þ N3)/(N1 þ N2 þ N3)] 100 (%). Thus, the aneuploidy
index from parental HeLa cells should be 5% and the same window
1533
H.-J.Nam et al.
Fig. 2. Induction of CIN by CA-Met expression. (A) Representative images showing multinucleated cells and micronuclei formation under conditions
indicated. Immunocytochemistry was performed using a-tubulin antibody and 4#,6-diamidino-2-phenylindole staining to observe the nuclei (upper panel). Boxes
on the right are magnifications of those on the left, clearly showing the micronuclei indicated by arrows (a’ and b’). The numbers of multinucleated cells and
cells with micronuclei in vector cells and CA-Met cells were counted and shown at the right panel. Mean ± SD from three independent experiments. P , 0.05 by
Student’s t-test. (B) HeLa cells with transient expression of CA-Met complementary DNA were also analyzed for supernumerary centrosomes (left panel),
multinucleated cells and micronuclei-containing cells (right panel). Mean ± SD from three independent experiments. P , 0.05 by Student’s t-test. (C) Metaphase
spread of cells with either vector (upper panel) or CA-Met expression (lower panel) was prepared by treatment with nocodazole for 12 h, and Giemsa
staining was performed to count the numbers of chromosomes in 300 cells. The figure shows the numbers of cells with designated chromosome number on the
x-axis. Cells were divided into three groups (N1, N2 and N3), and aneuploidy indices were calculated as described in the Materials and Methods. Representative
photomicrographs are presented (insets).
was applied to either vector control cells or CA-Met cells for the
comparison. We observed that the average number of chromosomes
in CA-Met cells (average 5 62.6) was similar to that in vector cells
but the CA-Met cells displayed a broader range of chromosome
numbers (Figure 2C, lower panel) as expected by their large standard deviation (SD 5 16.7). Notably, the aneuploidy index of CAMet cells was determined as 19.7% that was significantly
higher than that of vector control cells (7.44%). Collectively, CAMet enhanced the extent of CIN probably through the formation of
supernumerary centrosomes.
Activation of the PI3K–Akt pathway was necessary for centrosome
amplification
Cdk2/cyclin E kinase is a major player initiating centrosome duplication at G1/S phase and constitutive activation of Cdk2/cyclin E
1534
often leads to centrosome amplification (35,36). We first determined
whether Cdk2/cyclin E kinase activity remained elevated in CA-Met
cells using in vitro kinase assay using S phase cell lysates by releasing
cells from the second thymidine block. We found that Cdk2/cyclin E
kinase activity was not different between control and CA-Met cells
(supplementary Figure 3 is available at Carcinogenesis Online), indicating that Cdk2/cyclin E was not a causative factor for centrosome
amplification in CA-Met cells. We next assumed that cellular signaling kinases downstream of CA-Met might be involved in centrosome
amplification. As shown in Figure 3A, western blot analysis revealed
increased levels of phospho-ERK and phospho-Akt but not of
phospho-p38 or phospho-JNK in CA-Met cells. Densitometry quantification revealed 8-fold increase of phospho-ERK and 2-fold increase of phospho-Akt (Figure 3A, right panel) compared with control
cells. To address which signaling pathway is responsible, we next
Met-mediated CIN via Akt activation
Fig. 3. Abrogation of the increase of cells with multicentrosomes and multinuclei by treatment with LY294002, but not with U0126. (A) Cell lysates from
vector and CA-Met transfectants were (50 lg protein) subjected to western blot analysis using anti-phospho-ERK1/2, phospho-JNK, phospho-Akt and phosphop38 antibodies. Representative data from three independent experiments are presented. At the right panel, the band intensities were determined by densitometry
and the phosphorylation status was shown as fold changes to control. (B) CA-Met transfectants were treated with U0126 (10 lM) or LY294002 (30 lM) for 12 h
and further incubated in the presence of 100 ng/ml nocodazole for 12 h. After removal of nocodazole, cells were reseeded onto cover glasses and allowed to
progress for 60 min. The centrosomes were visualized by staining with anti-c-tubulin antibody. Cell number 5 300, mean ± SD from three independent
experiments. P , 0.05 by Student’s t-test. (C) Vector and CA-Met transfectants were treated with U0126 (10 lM) or LY294002 (30 lM) for 12 h and then cells
were subjected to 4#,6-diamidino-2-phenylindole staining for nucleus and immunostaining for a-tubulin. Cell number 5 300, mean ± SD from three
independent experiments. P , 0.005 by Student’s t-test. (D) Cell lysates from indicated samples (each containing 50 lg protein) were subjected to western blot
analysis using anti-phospho-Akt, phospho-ERK1/2, anti-Akt and anti-ERK1/2 antibodies.
applied chemical inhibitors against the ERK or PI3K to CA-Met cells to
investigate their effects on the centrosome amplification. Synchronized
cells were treated with each signaling inhibitor (U0126 for MEK1
and LY294002 for PI3K) and subjected to centrosome analysis. We
analyzed only the cells in the mitotic phase (mainly metaphase) because these inhibitors may also interfere with cell cycle progression
(37). Interestingly, pretreatment with LY294002, but not U0126, significantly reduced the number of CA-Met cells with multiple centrosomes (3) (Figure 3B, P , 0.05). Moreover, inhibition of the
PI3K pathway by LY294002 abrogated multinucleated cell formation
in CA-Met cells (Figure 3C, P , 0.005). At the concentration used for
the experiment, both signaling pathways were effectively blocked as
shown in Figure 3D.
To verify the importance of Akt in centrosome amplification of CAMet cells, we suppressed Akt activity by knockdown using siRNA or
by overexpressing the DN-Akt or phosphatase and tensin homolog
(PTEN), the phosphatase for PI3K products. Two different siRNAs
specifically targeting Akt messenger RNA (Life Technologies) was
applied to CA-Met cells 48 h before nocodazole treatment and re-
sulted in a clear abrogation of centrosome amplification (Figure 4A,
P , 0.05) as well as multinucleated cell formation (Figure 4B,
P , 0.05). Consistently, transfection of DN-Akt or PTEN into CAMet cells also significantly suppressed centrosome amplification as
well as multinucleated cell formation (Figure 4C and D), verifying the
crucial role for the PI3K–Akt in centrosome amplification of CA-Met
cells. Western blot analysis showed that 40–60% of Akt protein levels
were diminished by Akt siRNA (Figure 4E) and cell cycle profiles
were not changed by overexpression of PTEN (Figure 4F). Together,
our data for the first time provide the evidence that the PI3K–Akt
pathway is involved in deregulation of centrosome duplication in CAMet cells.
Increase of aneuploidy by CA-Met was observed in p53/ cells, but
not in p53þ/þ cells
Loss of p53 tumor suppressor function in cells leads to severe impacts
on centrosome duplication cycle (38) as well as maintenance of
genomic integrity (39). The p53 level in HeLa cells is very low, if
any, because E6 and E6AP actively ubiquitinates and induces
1535
H.-J.Nam et al.
Fig. 4. Reduction of the cells with multicentrosomes and multinuclei by suppression of PI3K pathway. (A and C) Vector and CA-Met transfectants were additionally
transfected with one of the Akt-specific siRNAs (Akt1 and Akt2) or control scrambled siRNA using oligofectamine reagent (A) or with dominant-negative form of
Akt (DN-Akt) or PTEN using polyethylenimine reagent (C) and cultured for 48 h before nocodazole treatment. After 12 h, cells released from nocodazole treatment
were analyzed for multicentrosomes by staining with anti-c-tubulin antibody. Cell number 5 300, mean ± SD from three independent experiments. P , 0.05,
P , 0.005, by Student’s t-test. (B and D) Vector and CA-Met transfectants were transfected with either siRNAs (B) or indicated plasmids (D) in the same manner
with (A) or (C), respectively. After 48 h, multinucleated cells were analyzed by 4#,6-diamidino-2-phenylindole staining for nuclei and immunostaining for a-tubulin.
Cell number 5 300, mean ± SD from three independent experiments. P,0.05, P , 0.005, by Student’s t-test. (E) Cell lysates from the samples harvested as
described above were subjected to western blot analysis using anti-phospho-Akt antibody. Actin was used as a loading control. (F) Vector and CA-Met transfectants
were additionally transfected with PTEN expression vector and cell cycle profiles were determined by flow cytometry at 2 days after transfection.
degradation of the p53 protein (40). In order to clarify the role for p53
in CA-Met-mediated CIN, we employed HCT116 human colon cancer
cells with the background of p53þ/þ or p53/ status. HCT116 p53þ/þ
and HCT116 p53/ cells were transfected with CA-Met expression
vector and treated with G418 (500 lg/ml). The resultant stable clones
(100) were pooled and further cultivated for 2 months. First, we
compared the number of centrosomes in these cells. Notably, overexpression of CA-Met in HCT116 p53þ/þ cells did not increase supernumerary centrosomes (Figure 5A). In contrast, cells with supernumerary
centrosomes were already abundant in HCT116 p53/ cells, reaching
to 30% and overexpression of CA-Met did not cause a significant
increase in these cells (Figure 5A). Next, we examined for progression
of karyotypic profiles upon cell passages. The cells at the time of antibiotics selection were indicated as 0 month and further collected after 1
and 2 months through continuous passages. The passage numbers of
each cell were essentially the same. In each analysis, three hundred
metaphase cells were analyzed for karyotypic profiles. Notably, we
observed that expression of CA-Met did not increase aneuploidy in
HCT116 p53þ/þ cells (Figure 5B) that is different from what we have
observed in HeLa cells (Figure 2B), suggesting that lack of p53 activity
1536
in HeLa cells probably contributed to aneuploidy formation. Indeed,
the aneuploidy indice of HCT116 p53/ cells were significantly higher than those of HCT116 p53þ/þ cells through cell passages (Figure 5B
and C). Moreover, expression of CA-Met significantly increased aneuploidy in p53/ HCT116 cells (Figure 5C). CA-Met expression in
HCT116 p53/ cells resulted in almost 3-fold higher aneuploidy index
than that in vector control cells at 1 month (Figure 5C, middle panel).
The increased aneuploidy index in HCT116 p53/ cells also remained
high after 2 months. Increase in the aneuploidy index was even observed at 0 month (29.1% for CA-Met versus 17.0% for control) because several rounds of cell proliferation had already occurred during
the selection process. However, the average numbers of chromosomes
were not significantly different among any of the cells tested (data not
shown), suggesting that these cells might undergo relatively equivalent
gain and loss of chromosomes. Together, our data indicate that the
increase of aneuploidy by CA-Met is related with the p53 status of
the cell.
Given that the Akt activation is important for centrosome amplification and multinucleated cell formation in CA-Met cells, inhibition
of Akt activation may reduce aneuploidy formation. To address that,
Met-mediated CIN via Akt activation
Fig. 5. Aneuploidy induction by CA-Met is related with p53 status. (A) HCT 116 p53þ/þ and p53/ cells were transfected with vector or CA-Met and selected
in the presence of G418 (500 lg/ml) to establish stable cell lines. The numbers of centrosomes were determined by staining with anti-c-tubulin antibody.
(B and C) The same cell lines were analyzed for the aneuploidy index. The cells at the time of selection were designated as cells at 0 month, and cell samples
were collected 1 and 2 months later. Metaphase spread of indicated cells was prepared by treatment with nocodazole for 12 h, and Giemsa staining was
performed to count the chromosome numbers in 100 cells. The numbers of cells with the designated chromosome number on the x-axis are presented. Western blot
analysis was performed to verify the presence or absence of p53 and/or CA-Met from the indicated cell lysates (inset).
we transfected M1268T complementary DNA along with PTEN into
HCT116 cells lacking p53 gene (HCT116 p53/ cells) and survival
clones were selected in the presence of G418 (500 lg/ml). Western
blotting showed that CA-Met and PTEN proteins were properly expressed in these cells (Figure 6A). Notably, the aneuploidy index in
HCT116 p53/ cells was relatively high (Figure 6B), which is almost
twice of HeLa cells (Figure 2C), suggesting that loss of p53 contributed to increased aneuploidy in these cells. Expression of CA-Met
increased the aneuploidy index from 15 to 26%, whereas co-expression
of PTEN partly suppressed the CA-Met-induced aneuploidy
formation (Figure 6B). Collectively, these data suggest that CA-Metinduced CIN is at least partly mediated by the PI3K–Akt pathway.
1537
H.-J.Nam et al.
Fig. 6. Reduced aneuploidy formation by overexpression of PTEN in CAMet cells. (A) HCT 116 p53/ cells were transfected with CA-Met along
with PTEN followed by selection in the presence of G418 (500 lg/ml).
Expression of Met and PTEN were determined by western blotting.
(B) Metaphase spread of indicated cells was prepared by treatment with
nocodazole for 12 h, and Giemsa staining was performed to count the
chromosome numbers in 100 cells. The numbers of cells with the designated
chromosome number on the x-axis are presented. Aneuploidy indices were
calculated as described in the Materials and Methods.
Discussion
Oncogenesis is a complicated process that mediates many different
cellular and physiological events. Met is overexpressed and mutated
in a variety of human malignancies (15) and its pleiotropic actions on
angiogenesis (41), invasion, cellular motility (5), cell growth (3,4) and
cell death (42) appear to be deeply related with its oncogenic property.
Here, we demonstrated an additional function of oncogenic Met promoting centrosome amplification and CIN. Recently, centrosomes
have attracted much attention due to their critical roles in the proper
segregation of chromosomes. Faulty centrosomes have been strongly
implicated as a primary cause of CIN and aneuploidy in cancer development (22,24,25). Interestingly, in hereditary papillary renal carcinoma chromosome harboring germ line mutation of Met was found
to be selectively amplified (43). Although the mechanism involved in
selective amplification of chromosome harboring Met mutant allele
has not been suggested, amplification of the activating Met mutation
was found to play key roles in promoting tumorigenesis in vivo (44).
Furthermore, 4 of 10 tumor explants induced by injection of tetraploid
1538
p53/ mammary epithelial cells showed amplification of chromosome region harboring Met (16), suggesting the importance of aberrant activation of Met induced by CIN. Taken together with our results
showing the induction of CIN by activating mutant of Met, we assume
that CIN and activation of Met might act as a positive feedback in the
process of carcinogenesis.
RTKs are key initiators of signaling cascades and therefore, mutations on them have multiple effects on various cellular proteins.
Oncogenic RTKs often utilize the Ras–Raf–MAPK pathway to stimulate cellular proliferation (45). The P13K cascade is another important contributor in tumorigenesis. P13K can initiate changes in gene
transcription, cell motility and anti-apoptotic signaling (46,47). In the
present study, we showed that the PI3K–Akt pathway also mediates
centrosome amplification. Knock down of Akt expression by siRNA
or overexpression of the DN-Akt resulted in a clear abrogation of
centrosome amplification in CA-Met cells (Figure 4). As far as we
know, this is the first observation illustrating Akt activity in deregulated centrosome biogenesis. Intriguingly, Akt as well as other proteins in the PI3K–mTOR signaling pathway can localize to the
centrosome (48–50). TSC1, a known substrate protein of Akt, localizes to the centrosome and plays a role in centrosome duplication via
the mTOR pathway (50). It has also been shown that glycogen synthase kinase 3, one of the target molecules of Akt, is present along the
length of spindle microtubules and that a fraction of glycogen synthase kinase 3 is phosphorylated during mitosis particularly at the
spindle poles with concomitant appearance of phosphorylated Akt
at the centrosome (48). We investigated the possibility of centrosome
localization of phosphorylated Akt and glycogen synthase kinase 3,
but failed (data not shown). Although our preliminary test revealed
supernumerary centrosome formation by overexpression of wild-type
Akt (H.-J.Nam unpublished results), the underlying mechanisms involved in Akt-mediated centrosome amplification should be resolved
in the future study.
Centrosome biogenesis and maintenance of genomic integrity have
been closely correlated with p53 status. We observed that the CA-Met
induced centrosome amplification in HeLa cells (Figure 1C), which
have low or no p53 activity, and the ability of CA-Met to induce
aneuploidy was influenced by p53 status (Figure 5). p53 is shown to
participate in the regulation of centrosome duplication by direct binding
to centrosome, independent of its transactivation function (51). In addition, centrosome reduplication can be triggered by high Cdk2/cyclin
E activity in the absence of p53 (38,52), which was not the case in our
system (supplementary Figure 3 is available at Carcinogenesis Online).
It is also possible that the hyperactive Met–PI3K–Akt axis accelerates
centrosome duplication cycle and p53 behaves as a safeguard to suppress formation of supernumerary centrosomes. It is noteworthy that
frequency of cells with multicentrosome showed almost no difference
between CA-Met cells and vector control cells in p53/ background.
It is possible that basal machineries yet uncharacterized leading to
supernumerary centrosome are more active in HCT116 cells than in
HeLa cells. Without p53, these unknown machineries may not be
checked, leading high frequency of cells with multicentrosome
(25%) as shown in Figure 5A. Therefore, it is probably that it
would be hard to see additional effect of CA-Met on supernumerary
centrosomes in HCT116 p53/ cells. The ability of CA-Met inducing higher aneuploidy index than vector control cells in spite
of almost the same level of supernumerary centrosome in HCT116
p53/ cells also suggest the possibility that CA-Met induces CIN
through unidentified mechanisms along with supernumerary centrosome formation in these cells. Thus, the principal mechanism for
CIN induction by CA-Met might be cell line context-dependent.
Anyhow, the oncogenic property of CA-Met mediated through centrosome amplification and aneuploidy largely depends on p53 status.
This may not be greatly surprising because cellular p53 induces
apoptosis, cell cycle arrest and senescence when cells encounter
oncogenic stresses (53). Recently, Kim et al. (54) also reported that
activation of PI3K signaling by mutations in PTEN or overexpression of PIK3CA, a constitutively active form of PI3K, could lead to
activation of p53-mediated senescence-like growth arrest in human
Met-mediated CIN via Akt activation
cells. Therefore, it is probable that the critical role of p53 in prevention of CIN in our system is due to its dual role, namely prevention of centrosome amplification and removal of cells with CIN.
In summary, our data show that HGF/Met signaling induces centrosome amplification and CIN via the PI3K–Akt pathway. Thus, it
seems that oncogenic growth factor/growth factor receptor not only
control cell cycle progression but also clearly have cross talk to regulate mitosis and that the widespread appearance of CIN in human
cancers might be due to the widespread abnormalities in growth factor
signaling in human cancers.
Supplementary material
Supplementary Figures 1–3 can be found at http://carcin
.oxfordjournals.org/
Funding
Korea Science and Engineering Foundation through Chronic Inflammatory Disease Research Center (R13-2003-019).
Acknowledgements
We thank Dr Woon Ki Paik (Korea University, Seoul, Korea) for his critical
review of this manuscript, Dr G.F.Vande Woude (Van Andel Research Institute,
MI 49503, USA) for CA-Met cDNA, Dr Myong-Joon Hanh (Sungkyunkwan
University School of Medicine, Suwon, Korea) for SRT antibody and SRTtagged WT-Akt constructs, Dr G.Guarguaglini (University of Rome, Rome,
Italy) for anti-Cep170 antibody and Dr B.Vogelstein (Johns Hopkins Oncology,
Baltimore, USA) for HCT116 cells.
Conflict of Interest Statement: None declared.
References
1. Naldini,L. et al. (1991) The tyrosine kinase encoded by the MET protooncogene is activated by autophosphorylation. Mol. Cell. Biol., 11, 1793–
1803.
2. Bottaro,D.P. et al. (1991) Identification of the hepatocyte growth factor
receptor as the c-met proto-oncogene product. Science, 251, 802–804.
3. Nakamura,T. et al. (1984) Partial purification and characterization of hepatocyte growth factor from serum of hepatectomized rats. Biochem. Biophys. Res. Commun., 122, 1450–1459.
4. Higashio,K. et al. (1993) Tumor cytotoxic activity of HGF-SF. EXS, 65,
351–368.
5. Stoker,M. et al. (1987) Scatter factor is a fibroblast-derived modulator of
epithelial cell mobility. Nature, 327, 239–242.
6. Bussolino,F. et al. (1992) Hepatocyte growth factor is a potent angiogenic
factor which stimulates endothelial cell motility and growth. J. Cell Biol.,
119, 629–641.
7. Jeffers,M. et al. (1996) Enhanced tumorigenicity and invasion-metastasis
by hepatocyte growth factor/scatter factor-met signalling in human cells
concomitant with induction of the urokinase proteolysis network. Mol.
Cell. Biol., 16, 1115–1125.
8. Schmidt,L. et al. (1997) Germline and somatic mutations in the tyrosine
kinase domain of the MET proto-oncogene in papillary renal carcinomas.
Nat. Genet., 16, 68–73.
9. Jeffers,M. et al. (1997) Activating mutations for the met tyrosine kinase
receptor in human cancer. Proc. Natl Acad. Sci. USA, 94, 11445–11450.
10. Tanyi,J. et al. (1999) Evaluation of the tyrosine kinase domain of the Met
proto-oncogene in sporadic ovarian carcinomas. Pathol. Oncol. Res., 5,
187–191.
11. Park,W.S. et al. (1999) Somatic mutations in the kinase domain of the Met/
hepatocyte growth factor receptor gene in childhood hepatocellular carcinomas. Cancer Res., 59, 307–310.
12. Lee,J.H. et al. (2000) A novel germ line juxtamembrane Met mutation in
human gastric cancer. Oncogene, 19, 4947–4953.
13. Roberts,P.J. et al. (2007) Targeting the Raf-MEK-ERK mitogen-activated
protein kinase cascade for the treatment of cancer. Oncogene, 26, 3291–3310.
14. Vivanco,I. et al. (2002) The phosphatidylinositol 3-Kinase AKT pathway in
human cancer. Nat. Rev., 2, 489–501.
15. Comoglio,P.M. (2001) Pathway specificity for Met signalling. Nat. Cell
Biol., 3, E161–E162.
16. Fujiwara,T. et al. (2005) Cytokinesis failure generating tetraploids
promotes tumorigenesis in p53-null cells. Nature, 437, 1043–1047.
17. Weaver,B.A. et al. (2007) Aneuploidy acts both oncogenically and as
a tumor suppressor. Cancer Cell, 11, 25–36.
18. Abulaiti,A. et al. (2006) Ras induces chromosome instability and abrogation of the DNA damage response. Cancer Res., 66, 10505–10512.
19. Hadjihannas,M.V. et al. (2006) CIN By WNT: growth pathways, mitotic
control and chromosomal instability in cancer. Cell Cycle (Georgetown,
Tex.), 5, 2077–2081.
20. Fukasawa,K. (2005) Centrosome amplification, chromosome instability
and cancer development. Cancer Lett., 230, 6–19.
21. Gollin,S.M. (2004) Chromosomal instability. Curr. Opin. Oncol., 16, 25–
31.
22. Lingle,W.L. et al. (2002) Centrosome amplification drives chromosomal
instability in breast tumor development. Proc. Natl Acad. Sci. USA, 99,
1978–1983.
23. Pihan,G.A. et al. (2003) Centrosome abnormalities and chromosome instability occur together in pre-invasive carcinomas. Cancer Res., 63, 1398–
1404.
24. Brinkley,B.R. (2001) Managing the centrosome numbers game: from chaos
to stability in cancer cell division. Trends Cell Biol., 11, 18–21.
25. Pihan,G.A. et al. (1998) Centrosome defects and genetic instability in
malignant tumors. Cancer Res., 58, 3974–3985.
26. Malumbres,M. et al. (2007) Cell cycle kinases in cancer. Curr. Opin. Genet.
Dev., 17, 60–65.
27. Li,J.J. et al. (2006) Mitotic kinases: the key to duplication, segregation, and
cytokinesis errors, chromosomal instability, and oncogenesis. Pharmacol.
Ther., 111, 974–984.
28. Southern,P.J. et al. (1982) Transformation of mammalian cells to antibiotic
resistance with a bacterial gene under control of the SV40 early region
promoter. J. Mol. Appl. Genet., 1, 327–341.
29. Park,M. et al. (2005) Presence of autocrine hepatocyte growth factor-Met
signaling and its role in proliferation and migration of SNU-484 gastric
cancer cell line. Exp. Mol. Med., 37, 213–219.
30. Martel,C. et al. (1997) Inactivation of retinoblastoma family proteins by
SV40 T antigen results in creation of a hepatocyte growth factor/scatter
factor autocrine loop associated with an epithelial-fibroblastoid conversion
and invasiveness. Cell Growth Differ., 8, 165–178.
31. Grotegut,S. et al. (2006) Hepatocyte growth factor induces cell scattering
through MAPK/Egr-1-mediated upregulation of Snail. EMBO J., 25, 3534–
3545.
32. Guarguaglini,G. et al. (2005) The forkhead-associated domain protein
Cep170 interacts with Polo-like kinase 1 and serves as a marker for mature
centrioles. Mol. Biol. Cell., 16, 1095–1107.
33. Wang,Q. et al. (2004) The centrosome in normal and transformed cells.
DNA Cell Biol., 23, 475–489.
34. Jefford,C.E. et al. (2006) Mechanisms of chromosome instability in cancers. Crit. Rev. Oncol. Hematol., 59, 1–14.
35. Mussman,J.G. et al. (2000) Synergistic induction of centrosome hyperamplification by loss of p53 and cyclin E overexpression. Oncogene, 19,
1635–1646.
36. Kawamura,K. et al. (2004) Induction of centrosome amplification and
chromosome instability in human bladder cancer cells by p53 mutation
and cyclin E overexpression. Cancer Res., 64, 4800–4809.
37. Roberts,E.C. et al. (2002) Distinct cell cycle timing requirements for extracellular signal-regulated kinase and phosphoinositide 3-kinase signaling
pathways in somatic cell mitosis. Mol. Cell. Biol., 22, 7226–7241.
38. Tarapore,P. et al. (2002) Loss of p53 and centrosome hyperamplification.
Oncogene, 21, 6234–6240.
39. Gomez-Lazaro,M. et al. (2004) p53: twenty five years understanding the
mechanism of genome protection. J. Physiol. Biochem., 60, 287–307.
40. Talis,A.L. et al. (1998) The role of E6AP in the regulation of p53 protein
levels in human papillomavirus (HPV)-positive and HPV-negative cells.
J. Biol. Chem., 273, 6439–6445.
41. Bussolino,F. et al. (1992) Synthesis of platelet-activating factor by polymorphonuclear neutrophils stimulated with interleukin-8. J. Biol. Chem.,
267, 14598–14603.
42. Fan,S. et al. (2001) The multisubstrate adapter Gab1 regulates hepatocyte
growth factor (scatter factor)-c-Met signaling for cell survival and DNA
repair. Mol. Cell. Biol., 21, 4968–4984.
43. Zhuang,Z. et al. (1998) Trisomy 7-harbouring non-random duplication of
the mutant MET allele in hereditary papillary renal carcinomas. Nat.
Genet., 20, 66–69.
44. Graveel,C. et al. (2004) Activating Met mutations produce unique tumor
profiles in mice with selective duplication of the mutant allele. Proc. Natl
Acad. Sci. USA, 101, 17198–17203.
1539
H.-J.Nam et al.
45. Porter,A.C. et al. (1998) Tyrosine kinase receptor-activated signal transduction pathways which lead to oncogenesis. Oncogene, 17, 1343–1352.
46. Xiao,G.H. et al. (2001) Anti-apoptotic signaling by hepatocyte growth
factor/Met via the phosphatidylinositol 3-kinase/Akt and mitogen-activated
protein kinase pathways. Proc. Natl Acad. Sci. USA, 98, 247–252.
47. Rosario,M. et al. (2003) How to make tubes: signaling by the Met receptor
tyrosine kinase. Trends Cell Biol., 13, 328–335.
48. Wakefield,J.G. et al. (2003) A role for glycogen synthase kinase-3 in mitotic
spindle dynamics and chromosome alignment. J. Cell Sci., 116, 637–646.
49. Kalous,J. et al. (2006) PKB/AKT is involved in resumption of meiosis in
mouse oocytes. Biol. Cell., 98, 111–123.
50. Astrinidis,A. et al. (2006) Hamartin, the tuberous sclerosis complex 1 gene
product, interacts with polo-like kinase 1 in a phosphorylation-dependent
manner. Hum. Mol. Genet., 15, 287–297.
1540
51. Shinmura,K. et al. (2007) Direct evidence for the role of centrosomally
localized p53 in the regulation of centrosome duplication. Oncogene, 26,
2939–2944.
52. Tarapore,P. et al. (2002) A mammalian in vitro centriole duplication
system: evidence for involvement of CDK2/cyclin E and nucleophosmin/
B23 in centrosome duplication. Cell Cycle (Georgetown, Tex.), 1,
75–81.
53. Rodier,F. et al. (2007) Two faces of p53: aging and tumor suppression.
Nucleic Acids Res., 35, 7475–7484.
54. Kim,J.S. et al. (2007) Activation of p53-dependent growth suppression in
human cells by mutations in PTEN or PIK3CA. Mol. Cell. Biol., 27, 662–
677.
Received July 10, 2009; revised June 10, 2010; accepted June 13, 2010