Caco-2 intestinal cell differentiation is associated with
G1 arrest and suppression of CDK2 and CDK4
QING-MING DING, TIEN C. KO, AND B. MARK EVERS
Department of Surgery, University of Texas Medical Branch, Galveston, Texas 77555
gut differentiation; cyclin-dependent kinases; cell cycle; cyclindependent kinase inhibitor
THE EPITHELIUM OF THE gastrointestinal tract is a complex and dynamic tissue composed of numerous cell
types with important cellular functions, including digestion, absorption, barrier and immune function, and
peptide secretion (35). The mammalian intestinal mucosa undergoes a process of continual renewal characterized by active proliferation of stem cells localized
near the base of the crypts, progression of these cells up
the crypt-villus axis with cessation of proliferation, and
subsequent differentiation into one of the four primary
cell types (i.e., absorptive enterocytes, goblet cells,
Paneth cells, and enteroendocrine cells) (7, 16, 36, 38,
The costs of publication of this article were defrayed in part by the
payment of page charges. The article must therefore be hereby
marked ‘‘advertisement’’ in accordance with 18 U.S.C. Section 1734
solely to indicate this fact.
43). The differentiated enterocytes, which make up the
majority of the cells of the gut mucosa, then undergo a
process of programmed cell death (i.e., apoptosis) and
extrusion into the gut lumen (22, 37). Remarkably, this
entire process of proliferation, differentiation, and apoptosis and extrusion occurs over a 3- to 5-day period,
depending on the species (7, 16, 36–38, 43). The cellular
mechanisms regulating this tightly regimented process
have not been clearly defined; however, this topic
represents an area of active investigation, since the
delineation of this process will lead to a better understanding of normal gut mucosal growth.
The mammalian cell cycle is regulated by the sequential activation and inactivation of a highly conserved
family of cyclin-dependent kinases (CDKs) (31, 46).
CDK activation requires the binding of a regulatory
protein (i.e., cyclin) (24) and is controlled by both
positive and negative phosphorylation (29). Cell cycle
progression is regulated at two key checkpoints, the
G1/S and the G2/M transition points (32, 33). Progression through early to mid-G1 is dependent on CDK4,
and possibly CDK6, which are activated by association
with one of the three D-type cyclins (D1, D2, and D3)
(40). Progression through late G1 and into the S phase
requires activation of CDK2, which is sequentially
regulated by cyclins E and A, respectively (11, 19). The
activities of the CDKs can be inhibited by the binding of
CDK inhibitory proteins (13, 41). Two families of CDK
inhibitory proteins have been identified. The first family consists of p15Ink4b, p16Ink4a, p18Ink4c, and p19Ink4d,
which appear to selectively bind and inhibit CDK4 and
CDK6 (17, 18, 21, 39), whereas members of the second
family, consisting of p21Waf1/Cip1, p27Kip1/Pic2, and p57Kip2,
are universal inhibitors of the cyclin/CDK complexes
(13, 41). In addition, important targets of CDK4 and
CDK2 include the retinoblastoma protein (pRb) and the
pRb-related proteins p107 and p130 (26, 47). In their
hypophosphorylated form, the pRb family of proteins
can sequester and inactivate the E2F family of transcription factors that regulate genes that participate in
S phase entry.
The human colon cancer cell line Caco-2 spontaneously differentiates to a small bowel-like phenotype, as
indicated by dome formation, presence of microvilli,
and expression of brush-border enzymes (i.e., sucrase
and alkaline phosphatase) after confluency. Caco-2 has
served as a useful in vitro model to further delineate
mechanisms triggering the terminal differentiation process in enterocytes (27, 34, 50). Previously, we showed
an early induction of the CDK inhibitor p21Waf1/Cip1
occurring in Caco-2 cells on day 3 postconfluency that
precedes induction of sucrase-isomaltase gene expression (14). Because p21Waf1/Cip1 is a universal inhibitor of
CDKs, its induction during Caco-2 differentiation may
0363-6143/98 $5.00 Copyright r 1998 the American Physiological Society
C1193
Downloaded from http://ajpcell.physiology.org/ by 10.220.32.247 on June 17, 2017
Ding, Qing-Ming, Tien C. Ko, and B. Mark Evers.
Caco-2 intestinal cell differentiation is associated with G1
arrest and suppression of CDK2 and CDK4. Am. J. Physiol.
275 (Cell Physiol. 44): C1193–C1200, 1998.—The cellular
mechanisms regulating intestinal proliferation and differentiation remain largely undefined. Previously, we showed an
early induction of the cyclin-dependent kinase (CDK) inhibitor p21Waf1/Cip1 in Caco-2 cells, a human colon cancer line that
spontaneously differentiates into a small bowel phenotype.
The purpose of our present study was to assess the timing of
cell cycle arrest in relation to differentiation in Caco-2 cells
and to examine the mechanisms responsible for CDK inactivation. Caco-2 cells undergo a relative G1/S block and cease to
proliferate at day 3 postconfluency; an increase in the activity
of terminally differentiated brush-border enzymes (sucrase
and alkaline phosphatase) was noted at day 6 postconfluency.
Cell cycle block was associated with suppression of both
CDK2 and CDK4 activities, which are important for G1/S
progression. Treatment of the CDK immune complexes with
the detergent deoxycholate (DOC) resulted in restoration of
CDK2, but not CDK4, activity at day 3 postconfluency,
suggesting the presence of inhibitory protein(s) binding to the
cyclin/CDK2 complex at this time point. An increased binding
of p21Waf1/Cip1 to CDK2 complexes at day 3 postconfluency was
noted, suggesting a potential role for p21Waf1/Cip1 in CDK2
inactivation; however, immunodepletion of p21Waf1/Cip1 from
Caco-2 protein extracts demonstrated that p21Waf1/Cip1 is only
partially responsible for CDK2 suppression at day 3 postconfluency. A decrease in the cyclin E/CDK2 complex appears to
contribute to the CDK2 inactivation noted at days 6 and 12
postconfluency. Taken together, our results suggest that
multiple mechanisms contribute to CDK suppression during
Caco-2 cell differentiation. Inhibition of CDK2 and CDK4
leads to G1 arrest and inhibition of proliferation that precede
Caco-2 cell differentiation.
C1194
CDK SUPPRESSION AND EPITHELIAL CELL DIFFERENTIATION
be sufficient to inhibit CDK activity and induce cell
cycle arrest. Therefore, the purpose of our present
study was to assess the timing of cell cycle arrest in
relation to differentiation of the Caco-2 cell line. In
addition, we evaluated the role of p21Waf1/Cip1 in inhibiting CDK2 activity during Caco-2 differentiation.
MATERIALS AND METHODS
RESULTS
Association of Caco-2 cell differentiation with G1 cell
cycle arrest and cessation of proliferation. Previously,
we showed that enterocytic differentiation of the Caco-2
cell line, which occurs after confluency, was preceded by
induction of the universal CDK inhibitor, p21Waf1/Cip1,
suggesting that Caco-2 cell differentiation was associated with cell cycle arrest (14). To further ascertain the
cell cycle distribution of preconfluent (day 22), confluent (day 0), and postconfluent Caco-2 cells, DNA flow
cytometry was performed (Fig. 1A). Our results indicate that 58% of preconfluent Caco-2 cells were in the
Downloaded from http://ajpcell.physiology.org/ by 10.220.32.247 on June 17, 2017
Materials. Radioactive compounds were purchased from
DuPont-NEN (Boston, MA). Immobilon P nylon membranes
for Western blots were purchased from Millipore (Bedford,
MA), and X-ray film was from Eastman Kodak (Rochester,
NY). The human glutathione S-transferase retinoblastoma
protein (GST-Rb) and all antibodies were from Santa Cruz
Biotechnology (Santa Cruz, CA). The enhanced chemiluminescence (ECL) system for Western immunoblot analysis and the
cell proliferation kit were purchased from Amersham (Arlington Heights, IL). Concentrated protein assay dye reagent was
purchased from Bio-Rad Laboratories (Hercules, CA). Tissue
culture media and reagents were obtained from GIBCO BRL
(Grand Island, NY). All other reagents were of molecular
biology grade and were purchased from either Sigma
(St. Louis, MO) or Amresco (Solon, OH).
Cell culture, cell counting, and 5-bromo-28deoxyuridine
immunohistochemistry. The human colon cancer cell line
Caco-2, obtained from the American Type Culture Collection
(Manassas, VA), was maintained in modified Eagle’s medium
supplemented with 15% (vol/vol) FCS. The cells were seeded
at 1 3 106 cells in 25-cm2 flasks and maintained in a
humidified atmosphere of 95% air-5% CO2 at 37°C. Studies
were performed on cells at various times either before confluency was reached (preconfluency) or postconfluency. For
assessment of cell proliferation, Caco-2 cells were counted
over a time course using a hemocytometer. 5-Bromo-28deoxyuridine (BrDU) immunohistochemistry was performed on
preconfluent and day 3 postconfluent cells using the cell
proliferation kit as described by the manufacturer, after a
60-min pulse with BrDU at 37°C.
Enzyme assays. Caco-2 cells were harvested with trypsin
and washed in PBS. The cell pellet was homogenized and
sonicated in Tris-mannitol buffer [2 mM Tris and 50 mM
mannitol (pH 7.1)] at 4°C. Sucrase (EC 3.3.1.48) activity was
measured according to the method of Messer and Dahlqvist
(28). Alkaline phosphatase (EC 3.1.3.1) activity was determined using an alkaline phosphatase determination kit from
Sigma. Values are expressed as milliunits per milligram
protein; one unit is defined as the activity that hydrolyzes 1
µmol substrate/min at 37°C. Proteins were assayed by the
method of Bradford (4).
Flow cytometry. Caco-2 cells were harvested with trypsin at
various time points, washed twice with PBS, and then
resuspended in PBS. Cells were then fixed in 80% ethanol for
30 min at room temperature and stored at 4°C. Before
processing, cells were collected by centrifugation and stained
by addition of 1 ml of propidium iodine solution (50 µg/ml).
RNase A (100 µg/ml) was added, and the sample was incubated for 15 min at room temperature. Cell cycle analysis was
performed using a FACScan flow cytometer (Becton Dickinson, San Jose, CA), and cell cycle distribution was analyzed
by the Modfit LT program (Verity, ME).
Protein preparation, Western immunoblot, and immunoprecipitation. Caco-2 cells were lysed with TNN buffer [in mM:
50 Tris · HCl (pH 7.5), 150 NaCl, 0.5 Nonidet P-40, 50 NaF, 1
sodium orthovanadate, 1 dithiothreitol (DTT), and 1 phenylmethylsulfonyl fluoride, and 25 µg/ml each of aprotinin,
leupeptin, and pepstatin A] at 4°C for 30 min. Lysates were
clarified by centrifugation (10,000 g for 30 min at 4°C), and
protein concentrations were determined using the method of
Bradford (4). Western immunoblot analyses were performed
as described previously (14). Briefly, protein samples (60 µg)
were resolved by SDS-PAGE and then electroblotted to
Immobilon P nylon membranes. Filters were incubated overnight at 4°C in blocking solution (Tris-buffered saline containing 5% nonfat dried milk and 0.05% Tween 20), followed by 3
h of incubation with the primary antibody. Filters were
incubated with a horseradish peroxidase-conjugated goat
anti-rabbit or anti-mouse IgG as a secondary antibody for 1 h.
After four final washes, the immune complexes were visualized using ECL detection. For determination of p21Waf1/Cip1
bound to CDK2, lysates were immunoprecipitated with antiCDK2 antibody, resolved by SDS-PAGE, and transferred to
nylon membranes. The membranes were then probed with an
antibody against p21Waf1/Cip1 (sc-397, Santa Cruz Biotechnology). Signals on the blots were visualized by autoradiography
and quantitated by densitometry using a Lynx 5000 digital
image analysis system.
Kinase assays. Caco-2 cells (preconfluent, confluent, and
postconfluent) were lysed with TNN buffer, and protein
samples (300 µg) were incubated with 1.5 µg of anti-CDK2 or
anti-CDK4 antibodies. Immune complexes were recovered
with protein A-Sepharose beads, washed twice with TNN
buffer and once with kinase buffer [in mM: 25 HEPES
(pH 7.4), 10 MgCl2, and 1 EGTA]. Pellets were resuspended in
40 µl of kinase buffer containing either 5 µg of histone H1
(Sigma; to measure CDK2-associated kinase activity) or
GST-Rb (to measure CDK4-associated kinase activity) at
30°C for 30 min. The kinase reaction was terminated by
addition of SDS sample loading buffer [50 mM Tris (pH 6.8),
100 mM DTT, 2% SDS, 0.1% bromphenol blue, and 10%
glycerol]. The samples were then heated to 95°C for 5 min and
resolved by 10% SDS-PAGE. The gels were dried, and the
phosphorylated proteins were visualized by autoradiography
and quantitated by densitometry.
DOC activation assays. Caco-2 cell extracts were immunoprecipitated with either anti-CDK2 or anti-CDK4 antibodies,
and the resultant immune complexes were incubated with
either 20 mM HEPES (pH 7.4) alone or 20 mM HEPES
containing 0.8% DOC on ice for 20 min. DOC was removed by
washing three times with TNN buffer and once with kinase
buffer. CDK2 and CDK4 activities were then analyzed as
described above.
Immunodepletion kinase assays. Protein extract (100 µg)
from day 3 postconfluent Caco-2 cells was heated for 5 min at
95°C and then immunoprecipitated with IgG or antip21Waf1/Cip1 antibody (2.0 µg). The immunodepleted supernatant was mixed with protein extract (100 µg) from confluent
(day 0) Caco-2 cells, incubated at 30°C for 30 min, and then
immunoprecipitated with anti-CDK2 antibody. The associated kinase activity was assayed as described above.
CDK SUPPRESSION AND EPITHELIAL CELL DIFFERENTIATION
C1195
G0/G1 phase and 28% of the cells were in the S phase of
the cell cycle. In contrast, 73% of Caco-2 cells at day 3
postconfluency were in G0/G1, with only 19% in S phase;
these percentages remained relatively stable up to day
12 postconfluency. Therefore, our results demonstrate a
32% decrease in the S phase population, consistent
with an induction of a relative G1/S cell cycle block
occurring in day 3 postconfluent Caco-2 cells.
We next determined the proliferative activity as well
as changes in brush-border enzyme activity (i.e., sucrase and alkaline phosphatase), which are markers of
enterocytic differentiation. Although other investigators have documented changes in cell proliferation and
enzyme activity in differentiating Caco-2 cells, variations can occur depending on culture conditions or
clonal differences from the parental cell lines (34, 44,
45, 50); therefore, we wanted to confirm when these
changes occurred in our own system and compare the
proliferative and enzymatic changes with alterations in
the cell cycle. Caco-2 cells (1 3 106 ) were plated in
25-cm2 flasks and then counted over a time course,
using a hemocytometer (Fig. 1B). Cells were 100%
confluent after 4 days in culture; a plateau growth was
achieved at day 3 postconfluency (i.e., after 7 days in
culture). As an additional proliferative index, preconfluent and day 3 postconfluent Caco-2 cells were assessed
by BrDU incorporation. Preconfluent Caco-2 cells were
actively proliferating, with the majority of cells staining positive for BrDU; in contrast, BrDU nuclear
staining was much less evident in day 3 postconfluent
Caco-2 cells (data not shown).
Finally, to determine the relationship between cell
proliferation and G1 arrest with changes in enterocytic
differentiation, the enzyme activities of sucrase and
alkaline phosphatase were measured. Increases in the
activities of both enzymes were first noted beginning on
day 6 postconfluency and continuing through day 12
postconfluency (Fig. 1C). Taken together, these findings
indicate that G1 cell cycle block and arrest of proliferation occur before differentiation of Caco-2 cells to an
enterocyte-like phenotype, as assessed by increases in
the activities of the terminally differentiated enzymes
sucrase and alkaline phosphatase.
Caco-2 cell differentiation is associated with decreases in CDK2 and CDK4 activities. To further examine the mechanisms responsible for G1 arrest at day 3
postconfluency and subsequent differentiation, we first
determined CDK2 and CDK4 kinase activities in preconfluent and postconfluent Caco-2 cells. These CDKs are
required for the progression from G1 to the S phase
(31, 46) and can be inhibited by p21Waf1/Cip1. Caco-2 cell
extracts from preconfluent, confluent (day 0), and
postconfluent cultures were immunoprecipitated with
either anti-CDK2 or anti-CDK4 antibodies, and the
immune complexes were assayed for kinase activity
using histone H1 (Fig. 2A) or a purified GST-Rb fusion
protein (Fig. 2B) as substrates, respectively. As demonstrated by the densitometric analyses (Fig. 2, A and B,
bottom), day 3 postconfluent cells had a 50% decrease
in both CDK2 and CDK4 activities compared with
preconfluent cells. A further decrease of CDK2 activity
to almost undetectable levels was noted at days 6 and
12 postconfluency. The suppression of both CDK2 and
CDK4 activities, occurring on day 3 postconfluency,
coincides with a decrease in the S phase population
(Fig. 1A). Potential mechanisms for the downregulation of CDK activities include the increased expression
of one or more cell cycle inhibitor proteins present in
Downloaded from http://ajpcell.physiology.org/ by 10.220.32.247 on June 17, 2017
Fig. 1. Cell cycle analysis, cell proliferation, and brush-border enzyme activities in differentiating Caco-2 cells. A:
cell cycle distribution was determined by DNA flow cytometric analysis as described in MATERIALS AND METHODS.
Samples from preconfluent (day 22), confluent (day 0), and postconfluent (days 3, 6, and 12) Caco-2 cells were
analyzed. Numbers indicate percentage of cells in G0/G1, S, and G2/M phases. Data are representative of 2 or 3
separate analyses. B: growth curve of Caco-2 cells seeded at 1 3 106 cells/25-cm2 flasks. Cell number was counted
using a hemocytometer. C: brush-border enzyme activities of differentiating Caco-2 cells; r, alkaline phosphatase
activity; j, sucrase activity.
C1196
CDK SUPPRESSION AND EPITHELIAL CELL DIFFERENTIATION
Fig. 2. Cyclin-dependent kinase (CDK)
2 and 4 assays in differentiating Caco-2
cells. A and B, top: protein extracts
from preconfluent (Pre) or postconfluent (days 0, 3, 6, and 12) Caco-2 cells
were immunoprecipitated (IP) with
anti-CDK2 or anti-CDK4 antibodies.
Resultant immune complexes were analyzed for CDK2 activity using histone
H1 as substrate (A) or for CDK4 activity using glutathione S-transferase
retinoblastoma protein (GST-Rb) as substrate (B). A and B, bottom: densitometric analyses of kinase assays are expressed as relative densitometric units.
Fig. 3. Deoxycholate (DOC) activation
assays. A and B, top: protein extracts
from preconfluent or postconfluent (days
0, 3, 6, and 12) Caco-2 cells were immunoprecipitated with either anti-CDK2
or anti-CDK4 antibodies. Resultant immune complexes were incubated with
(1) or without (2) 0.8% DOC, and
kinase assays were performed as described in MATERIALS AND METHODS, using histone H1 (A) or GST-Rb (B) as
substrates for CDK2 and CDK4, respectively. A and B, bottom: densitometric
analyses of kinase assays are expressed
as relative densitometric units.
The activation of CDKs, which is necessary for cell
cycle progression, requires binding of their catalytic
partners, the cyclins (24). During G1, the D-type cyclins
and cyclin E accumulate and activate CDK4 and CDK2,
respectively (11, 40). Previously, we have shown a
decrease in steady-state cyclin E protein levels beginning on day 6 postconfluency in Caco-2 cells (14).
Therefore, we determined whether the decrease in
CDK2 activity might be the result of decreases in cyclin
E/CDK2 complex formation. Protein extracts from preconfluent and postconfluent Caco-2 cells were immunoprecipitated with anti-CDK2 antibody. The resultant
immune complexes were analyzed for cyclin E levels by
Western blot (Fig. 4A). Consistent with our previous
findings on cyclin E protein levels (14), the amount of
cyclin E bound to CDK2 remained relatively stable
until day 6 postconfluency, when the levels decreased
dramatically (Fig. 4A). We then determined the levels
of CDK2 protein in differentiating Caco-2 cells and
found no change in CDK2 protein expression at day 3
postconfluency compared with day 0 (confluent) cells
(Fig. 4B). Thus these results suggest that the suppression of CDK2 activity in day 3 postconfluent Caco-2
cells is not caused by a decrease in the amount of the
cyclin E/CDK2 complex. However, the decrease in
Downloaded from http://ajpcell.physiology.org/ by 10.220.32.247 on June 17, 2017
postconfluent (day 3) Caco-2 cells or, alternatively, a
decrease in cyclin/CDK complex formation.
Suppression of CDK2 activity in day 3 postconfluent
Caco-2 cells is associated with inhibitory protein(s) that
bind CDK2. To assess the possibility of inhibitory
protein(s) binding to the cyclin/CDK2 and/or cyclin/
CDK4 complexes in differentiating Caco-2 cells, we
used the detergent DOC, which has been shown to
preferentially dissociate certain protein-protein interactions and leave the cyclin/CDK complex intact (48). The
CDK2 immune complexes were treated with either
buffer alone or buffer containing DOC for 20 min and
then assayed for CDK2 activity (Fig. 3A). Treatment
with DOC, but not with buffer alone, restored the
kinase activity at day 3 postconfluency to preconfluent
levels. However, CDK2 activity was not altered by DOC
on days 6 and 12 postconfluency. This result indicates
that one or possibly more inhibitory proteins contribute
to the decrease of CDK2 activity in the day 3 postconfluent Caco-2 cells, but not in the days 6 and 12 postconfluent cells. In contrast, no increase of CDK4 activity was
noted after treating the CDK4 immune complexes with
DOC (Fig. 3B), which suggests that CDK4 inhibition in
differentiating Caco-2 cells is not the result of the
binding of inhibitory proteins.
CDK SUPPRESSION AND EPITHELIAL CELL DIFFERENTIATION
C1197
Fig. 4. Abundance of cyclin E/CDK2 complexes in differentiating Caco-2 cells. A and B, top: protein extracts
from preconfluent or postconfluent (days 0, 3, 6, and 12)
Caco-2 cells were immunoprecipitated with anti-CDK2
antibody. Immune complexes were resolved by SDSPAGE, transferred to a membrane, and probed with
either anti-cyclin E antibody (A) or anti-CDK2 antibody
(B). A and B, bottom: densitometric analyses of blots are
expressed as relative densitometric units.
tein extracts with day 3 heated extract immunodepleted with a nonspecific IgG. As expected, the control
mixed extracts (day 0 1 heated, IgG-depleted day 3)
produced a marked decrease in CDK2 activity compared with CDK2 activity in confluent (day 0) extracts
Fig. 5. Assessment of p21Waf1/Cip1 as a potential inhibitor of CDK2
activity at day 3 postconfluency. A: Western blot analysis of p21Waf1/Cip1
bound to CDK2. Protein lysates from preconfluent or postconfluent
(days 0, 3, 6, and 12) Caco-2 cells were immunoprecipitated with
anti-CDK2 antibody. Immune complexes were resolved by SDSPAGE, transferred to a membrane, and probed with anti-p21Waf1/Cip1
antibody. B: immunodepletion assay. Protein extract from confluent
(day 0) Caco-2 cells was mixed with heated day 3 postconfluent
Caco-2 extract that had been immunodepleted (ID) with IgG or
anti-p21Waf1/Cip1 antibody. Protein extract from confluent (day 0)
Caco-2 cells was used as a control (2). Each extract was then
immunoprecipitated with anti-CDK2 antibody, and kinase activities
were assessed using histone H1 (HH1) as substrate. C: Western blot
analysis of p21Waf1/Cip1 levels in protein extracts from preconfluent or
day 3 postconfluent Caco-2 cells that were either immunodepleted
with anti-p21Waf1/Cip1 antibody (1) or not depleted (2).
Downloaded from http://ajpcell.physiology.org/ by 10.220.32.247 on June 17, 2017
cyclin E (14) and cyclin E/CDK2 complex at days 6 and
12 postconfluency may contribute to the sustained
suppression of CDK2 activity at these time points.
Taken together, the findings presented in Figs. 3 and
4 suggest that the suppression in CDK2 activity, occurring by day 3 postconfluency, is most likely due to the
presence of inhibitory protein(s) binding to the cyclin/
CDK2 complex and not the result of decreased levels of
cyclin E or cyclin E/CDK2 complexes.
Potential role of p21Waf1/Cip1 in the suppression of
CDK2 activity. Previously, we showed that the CDK
inhibitor p21Waf1/Cip1 is induced in differentiating Caco-2
cells beginning on day 3 postconfluency (14). We next
assessed the potential role of this protein in blocking
CDK2 activity on day 3 postconfluency. p21Waf1/Cip1
inhibits the cyclin/CDK complex by binding the heterodimeric complex. To determine whether p21Waf1/Cip1 is
bound to the cyclin/CDK2 complexes during Caco-2
differentiation, immunoprecipitation of the Caco-2 protein extracts was performed using anti-CDK2 antibody,
and the immune complexes were assayed for the abundance of p21Waf1/Cip1, as shown in Fig. 5A. An increase in
the amount of p21Waf1/Cip1 bound to CDK2 was noted at
day 3 postconfluency and persisted on days 6 and 12
postconfluency.
The increase in p21Waf1/Cip1 bound to CDK2 on day 3
postconfluency may be responsible for CDK2 inactivation, and p21Waf1/Cip1 may be the inhibitory protein
identified by the DOC experiment (Fig. 3A). To determine the role of p21Waf1/Cip1 in CDK2 suppression on day
3 postconfluency, we examined the ability of protein
extract from this time point to inhibit CDK2 activity
after it has been immunodepleted of p21Waf1/Cip1. To
remove p21Waf1/Cip1, protein extracts from day 3 postconfluent cells were first boiled for 5 min, which releases
heat-stable CDK inhibitors such as p21Waf1/Cip1 from
cyclin/CDK complexes (20). Boiled extracts were then
immunodepleted of p21Waf1/Cip1 by incubation with antip21Waf1/Cip1 antibody. CDK2 activity was determined by
mixing an equal amount of protein extract from confluent (day 0) Caco-2 cells with heated, p21Waf1/Cip1depleted extract from day 3 postconfluent cells. Control
experiments were performed by incubating day 0 pro-
C1198
CDK SUPPRESSION AND EPITHELIAL CELL DIFFERENTIATION
DISCUSSION
As intestinal epithelial cells progress up the cryptvillus unit, they enter a terminal differentiation program that in the absorptive enterocyte involves cessation of proliferation and expression of the terminally
differentiated enzymes sucrase and alkaline phosphatase (7, 16, 36, 38, 43). In our present study, we
demonstrate growth inhibition and a relative G1/S
block of the gut-derived Caco-2 cell line at day 3
postconfluency that precede the increases of both sucrase and alkaline phosphatase activity. This is, to our
knowledge, the first demonstration of G1/S block preceding Caco-2 cell differentiation. Others have also demonstrated the importance of G1 arrest in the subsequent
differentiation of nonintestine-derived cell types. For
example, Decker (8) demonstrated that nerve growth
factor (NGF)-mediated differentiation of the rat pheochromocytoma cell line PC-12 to a neuronal phenotype
is associated with G1 arrest and p21Waf1/Cip1 induction.
Furthermore, cell cycle arrest was also noted during
myoblast differentiation, interleukin-6-induced B cell
differentiation, and differentiation of the NB4 promyelocytic cell line mediated by retinoic acid (2, 3, 30). Taken
together, these results suggest that cell cycle arrest
may be a prerequisite for terminal differentiation of
various cell types.
Important regulators of the mammalian G1/S transition include the CDK2 and CDK4 proteins (31, 46).
Suppression of CDK2 and CDK4 activities induces
neuronal differentiation of mouse neuroblastoma cells
(25); inhibition of CDK2 and CDK4 is associated with
glial cell differentiation of central glia-4 cells and
hexamethylene bisacetamide-induced differentiation of
murine erythroleukemia cells, respectively (23, 42).
Conversely, the overexpression of CDK2 can block
NGF-induced differentiation of PC-12 cells (10). Concomitant with the increase in the G1 phase population
of Caco-2 cells noted at day 3 postconfluency, we found a
decrease in the kinase activities of both CDK2 and
CDK4, suggesting that suppression of CDKs may be
important in subsequent intestinal cell differentiation.
This hypothesis is further supported by the recent in
vivo findings of Chandrasekaran et al. (6) demonstrating a progressive decrease of CDK2 protein levels as
cells progress up the crypt-villus axis, with undetectable levels noted in the villus.
CDK activity can be inhibited by multiple mechanisms, including decreased expression of cyclin or CDK
proteins, decreased formation of cyclin/CDK complexes, negative phosphorylation, or increased binding
of inhibitory protein(s). Several of these mechanisms
appear to be responsible for suppression of CDK2 and
CDK4 in differentiating Caco-2 cells. Treatment of the
CDK2 immune complexes with the detergent DOC
reversed the CDK2 suppression at day 3 postconfluency, but not at days 6 and 12, suggesting that CDK2
inactivation on day 3 is the result of binding of inhibitory protein(s). These findings were further supported
by the fact that levels of cyclin E/CDK2 complexes were
not decreased on day 3 postconfluency. Other mechanisms likely contribute to the sustained CDK2 inhibition noted on days 6 and 12 postconfluency. Previously,
we showed a decrease in cyclin E and CDK2 expression
in days 6 and 12 postconfluent cells. In the present
study, we also noted a dramatic decrease in cyclin
E/CDK2 complexes at these time points; therefore, the
decreases noted in the cyclin E/CDK2 complexes may
contribute to CDK2 suppression at days 6 and 12. In
contrast to CDK2, CDK4 activity remains suppressed
even after the addition of DOC. These results, in
conjunction with our previous findings of decreased
D-type cyclin expression (14), suggest that suppression
of CDK4 is secondary to a decrease in D-type cyclin
protein levels. Interestingly, levels of cyclin D1 fall
rapidly in vivo in association with enterocytic differentiation (6), which further supports a role for the D-type
cyclins in intestinal cell differentiation. Therefore, distinct mechanisms may be responsible for the suppression of the CDKs, depending on the specific CDK and
the time point during the differentiation process.
Downloaded from http://ajpcell.physiology.org/ by 10.220.32.247 on June 17, 2017
alone (Fig. 5B), which indicates the presence of heatstable CDK2 inhibitory molecules. Surprisingly, depletion of day 3 extract with anti-p21Waf1/Cip1 antibody also
resulted in significant inhibition of CDK2 activity in
day 0 protein extracts compared with control (Fig. 5B).
However, the level of CDK2 activity in the presence of
extract depleted of p21Waf1/Cip1 was greater than that of
extract depleted with IgG, indicating some inhibitory
role of p21Waf1/Cip1. To confirm that p21Waf1/Cip1 protein
was indeed depleted by our protocol, heated p21Waf1/Cip1depleted protein extract from day 3 postconfluency was
analyzed for p21Waf1/Cip1 expression by Western blotting
(Fig. 5C). Immunodepletion of extracts using antip21Waf1/Cip1 antibody resulted in no apparent p21Waf1/Cip1
protein in the extract compared with the nondepleted
day 3 extract. Furthermore, the finding of increased
p21Waf1/Cip1 in day 3 postconfluent Caco-2 cells compared
with preconfluent cells confirms our previous findings
of p21Waf1/Cip1 induction occurring by day 3 postconfluency (14). These findings suggest that, even though
p21Waf1/Cip1 levels are increased in day 3 postconfluent
Caco-2 cells, this protein is only partially responsible
for the inhibition of CDK2 at this time point. To
determine whether other members of the Cip/Kip family of CDK inhibitors (i.e., p27Kip1/Pic2 and p57Kip2 ) might
play a role during Caco-2 differentiation, we analyzed
binding of these inhibitors to CDK2 in day 3 postconfluent Caco-2 cells. We were unable to detect binding of
either p27Kip1/Pic2 or p57Kip2 to CDK2 on day 3 postconfluency (data not shown).
Taken together, our results suggest that multiple
mechanisms contribute to the CDK suppression and
the cell cycle block noted in differentiating Caco-2 cells,
which may include the binding of p21Waf1/Cip1 and possibly other unidentified CDK inhibitors. The continued
suppression of CDK2 and CDK4 activities in Caco-2
cells may result from the decreased levels of their
partner cyclins, such as cyclin E and the D-type cyclins
(14), as well as a decrease in cyclin/CDK complexes.
CDK SUPPRESSION AND EPITHELIAL CELL DIFFERENTIATION
We thank Drs. Mark R. Hellmich, Zizheng Dong, and E. Aubrey
Thompson for helpful advice and discussion. In addition, we thank
Eileen Figueroa and Karen Martin for manuscript preparation.
This work was supported by National Institutes of Health Grants
RO1-DK-48498 and RO1-AG-10885 (to B. M. Evers), KO8-CA-64191
(to T. C. Ko), and PO1-DK-35608 and by the James E. Thompson
Memorial Foundation.
Q.-M. Ding is a visiting scientist from the Institute of Radiation
Medicine, Beijing, China.
Address for reprint requests: B. M. Evers, Dept. of Surgery,
University of Texas Medical Branch, 301 University Blvd., Galveston, TX 77555.
Received 28 April 1998; accepted in final form 22 July 1998.
REFERENCES
1. Abraham, C., B. Scaglione-Sewell, S. F. Skarosi, W. Qin, M.
Bissonnette, and T. A. Brasitus. Protein kinase C a modulates
growth and differentiation in Caco-2 cells. Gastroenterology 114:
503–509, 1998.
2. Andres, V., and K. Walsh. Myogenin expression, cell cycle
withdrawal, and phenotypic differentiation are temporally separable events that precede cell fusion upon myogenesis. J. Cell
Biol. 132: 657–666, 1996.
3. Bocchia, M., Q. Xu, U. Wesley, Y. Xu, T. Korontsvit, F.
Loganzo, A. P. Albino, and D. A. Scheinberg. Modulation of
p53, WAF1/p21 and BCL-2 expression during retinoic acidinduced differentiation of NB4 promyelocytic cells. Leuk. Res. 21:
439–447, 1997.
4. Bradford, M. M. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of
protein-dye binding. Anal. Biochem. 72: 248–254, 1976.
5. Brugarolas, J., C. Chadrasekaran, J. Gordon, D. Beach, T.
Jacks, and G. Hannon. Radiation-induced cell cycle arrest
compromised by p21 deficiency. Nature 377: 552–557, 1995.
6. Chandrasekaran, C., C. M. Coopersmith, and J. I. Gordon.
Use of normal and transgenic mice to examine the relationship
between terminal differentiation of intestinal epithelial cells and
accumulation of their cell cycle regulators. J. Biol. Chem. 271:
28414–28421, 1996.
7. Cheng, H., and C. P. Leblond. Origin, differentiation and
renewal of the four main epithelial cell types in the mouse small
intestine. V. Unitarian theory of the origin of the four epithelial
cell types. Am. J. Anat. 141: 537–561, 1974.
8. Decker, S. J. Nerve growth factor-induced growth arrest and
induction of p21Cip1/WAF1 in NIH-3T3 cells expressing TrkA. J.
Biol. Chem. 270: 30841–30844, 1995.
9. Deng, C., P. Zhang, J. W. Harper, S. J. Elledge, and P. Leder.
Mice lacking p21CIP1/WAF1 undergo normal development, but
are defective in G1 checkpoint control. Cell 82: 675–684, 1995.
10. Dobashi, Y., T. Kudoh, A. Matsumine, K. Toyoshima, and T.
Akiyama. Constitutive overexpression of CDK2 inhibits neuronal differentiation of rat pheochromocytoma PC12 cells. J. Biol.
Chem. 270: 23031–23037, 1995.
11. Dou, Q. P., A. H. Levin, S. Zhao, and A. B. Pardee. Cyclin E
and cyclin A as candidates for the restriction point protein.
Cancer Res. 53: 1493–1497, 1993.
12. Durand, B., F. B. Gao, and M. Raff. Accumulation of the
cyclin-dependent kinase inhibitor p27/Kip1 and the timing of
oligodendrocyte differentiation. EMBO J. 16: 306–317, 1997.
13. Elledge, S. J., and J. W. Harper. Cdk inhibitors: on the
threshold of checkpoints and development. Curr. Opin. Cell Biol.
6: 847–852, 1994.
14. Evers, B. M., T. C. Ko, J. Li, and E. A. Thompson. Cell cycle
protein suppression and p21 induction in differentiating Caco-2
cells. Am. J. Physiol. 271 (Gastrointest. Liver Physiol. 34):
G722–G727, 1996.
15. Gartel, A. L., M. S. Serfas, M. Gartel, E. Goufman, G. S. Wu,
W. S. El-Deiry, and A. L. Tyner. p21 (WAF1/CIP1) expression is
induced in newly nondividing cells in diverse epithelia and
during differentiation of the Caco-2 intestinal cell line. Exp. Cell
Res. 227: 171–181, 1996.
16. Gordon, J. I., G. H. Schmidt, and K. A. Roth. Studies of
intestinal stem cells using normal, chimeric, and transgenic
mice. FASEB J. 6: 3039–3050, 1992.
17. Guan, K. L., C. W. Jenkins, Y. Li, C. L. O’Keefe, S. Noh, X.
Wu, M. Zariwala, A. G. Matera, and Y. Xiong. Isolation and
characterization of p19INK4d, a p16-related inhibitor specific to
CDK6 and CDK4. Mol. Biol. Cell 7: 57–70, 1996.
18. Hannon, G. J., and D. Beach. p15INK4B is a potential effector
of TGF-beta-induced cell cycle arrest. Nature 371: 257–261,
1994.
Downloaded from http://ajpcell.physiology.org/ by 10.220.32.247 on June 17, 2017
Because the suppression of CDKs at day 3 postconfluency appears to be an important event in subsequent
Caco-2 cell differentiation, we wanted to determine the
potential inhibitory protein(s) responsible for CDK2
inactivation. Previously, we demonstrated an induction
of p21Waf1/Cip1 protein in day 3 postconfluent Caco-2 cells
(14); this early induction of p21Waf1/Cip1 in Caco-2 cells
has been confirmed by Gartel et al. (15) and Abraham
et al. (1). Therefore, we tested the hypothesis that
p21Waf1/Cip1 is responsible for the inhibition of CDK2
activity noted on day 3 postconfluency. Levels of p21Waf1/Cip1
bound to CDK2 were increased by day 3 postconfluency; however, heated protein extracts from day 3
postconfluent Caco-2 cells depleted of p21Waf1/Cip1 were
still able to partially inhibit CDK2 activity, although
not as effectively as heated extracts not depleted
of p21Waf1/Cip1. Our findings suggest that although
p21Waf1/Cip1 appears to contribute to CDK2 inactivation
on day 3 postconfluency, other inhibitory proteins are
also likely to be involved in this process. Moreover,
these results also suggest that additional mechanisms,
other than p21Waf1/Cip1 induction, may contribute to the
cell cycle arrest and subsequent differentiation of the
Caco-2 intestinal cell line. This supposition is further
supported by the in vivo findings of Brugarolas et al. (5)
and Deng et al. (9), who independently demonstrated
normal intestinal development in mice lacking
p21Waf1/Cip1, suggesting either that p21Waf1/Cip1 is not
essential for gut differentiation or, alternatively, that
other mechanisms can compensate for its loss. In
addition to p21Waf1/Cip1, the CDK inhibitors p27Kip1/Pic2
and p57Kip2 have been postulated to play a role in
cellular differentiation (12, 13, 24, 49); however, we did
not detect binding of either of these proteins to CDK2 in
the day 3 postconfluent Caco-2 cells. Future studies
will attempt to identify other CDK2 inhibitory protein(s) expressed during Caco-2 cell differentiation.
In conclusion, we have demonstrated that a relative
G1/S cell cycle block and suppression of the CDK2 and
CDK4 activities precede the increased differentiation of
Caco-2 cells. We have also shown that multiple mechanisms are responsible for suppression of CDK activities
observed during Caco-2 cell differentiation. These potential mechanisms include the downregulation of cyclin
and CDK protein expression, inhibition of cyclin/CDK
complex formation, the induction of the CDK inhibitor
p21Waf1/Cip1, and the potential involvement of other
inhibitory proteins. Identification of the cell cycle mechanisms regulating Caco-2 cell differentiation will provide important insights into the regulation of mammalian intestinal cell proliferation, differentiation, and
subsequent apoptosis.
C1199
C1200
CDK SUPPRESSION AND EPITHELIAL CELL DIFFERENTIATION
35. Podolsky, D. K., and M. W. Babyatsky. Growth and development of the gastrointestinal tract. In: Textbook of Gastroenterology, edited by T. Yamada. Philadelphia, PA: Lippincott, 1995, p.
546–577.
36. Ponder, B. A., G. H. Schmidt, M. M. Wilkinson, M. J. Wood,
M. Monk, and A. Reid. Derivation of mouse intestinal crypts
from single progenitor cells. Nature 313: 689–691, 1985.
37. Potten, C. S. Epithelial cell growth and differentiation. II.
Intestinal apoptosis. Am. J. Physiol. 273 (Gastrointest. Liver
Physiol. 36): G253–G257, 1997.
38. Potten, C. S., and M. Loeffler. Stem cells: attributes, cycles,
spirals, pitfalls and uncertainties. Lessons from the crypt. Development 110: 1001–1020, 1990.
39. Serrano, M., G. J. Hannon, and D. Beach. A new regulatory
motif in cell-cycle control causing specific inhibition of cyclin
D/CDK4. Nature 366: 704–707, 1993.
40. Sherr, C. J. D-type cyclins. Trends Biol. Sci. 20: 1887–1890,
1995.
41. Sherr, C. J., and J. M. Roberts. Inhibitors of mammalian G1
cyclin-dependent kinases. Genes Dev. 9: 1149–1163, 1995.
42. Tikoo, R., P. Casaccia-Bonnefil, M. V. Chao, and A. Koff.
Changes in cyclin-dependent kinase 2 and p27kip1 accompany
glial cell differentiation of central glia-4 cells. J. Biol. Chem. 272:
442–447, 1997.
43. Traber, P. G. Differentiation of intestinal epithelial cells: lessons
from the study of intestine-specific gene expression. J. Lab. Clin.
Med. 123: 467–477, 1994.
44. Vachon, P. H., and J.-F. Beaulieu. Transient mosaic patterns
of morphological and functional differentiation in the Caco-2 cell
line. Gastroenterology 103: 414–423, 1992.
45. Vachon, P. H., N. Perreault, P. Magny, and J.-F. Beaulieu.
Uncoordinated, transient mosaic patterns of intestinal hydrolase
expression in differentiating human enterocytes. J. Cell. Physiol.
166: 198–207, 1996.
46. Van den Heuvel, S., and E. Harlow. Distinct roles for cyclindependent kinases in cell cycle control. Science 262: 2050–2054,
1993.
47. Weinberg, R. A. The retinoblastoma protein and cell cycle
control. Cell 81: 323–330, 1995.
48. Woo, M. S., I. Sanchez, and B. D. Dynlacht. p130 and p107
use a conserved domain to inhibit cellular cyclin-dependent
kinase activity. Mol. Cell. Biol. 17: 3566–3579, 1997.
49. Yan, Y., J. Frisen, M. H. Lee, J. Massague, and M. Barbacid.
Ablation of the CDK inhibitor p57Kip2 results in increased
apoptosis and delayed differentiation during mouse development. Genes Dev. 11: 973–983, 1997.
50. Zweibaum, A., and I. Chantret. Human colon carcinoma cell
lines as in vitro models for the study of intestinal cell differentiation. In: Adaptation and Development of Gastrointestinal Function, edited by M. W. Smith and F. V. Sepulveda. Manchester, UK:
Manchester University Press, 1989, p. 103–112.
Downloaded from http://ajpcell.physiology.org/ by 10.220.32.247 on June 17, 2017
19. Hatakeyama, M., J. A. Brill, G. R. Fink. Collaboration of G1
cyclins in the functional inactivation of the retinoblastoma
protein. Genes Dev. 8: 1759–1771, 1994.
20. Hengst, L., V. Dulic, J. M. Slingerland, E. Lees, and S. I.
Reed. A cell cycle-regulated inhibitor of cyclin-dependent kinases. Proc. Natl. Acad. Sci. USA 91: 5291–5295, 1994.
21. Hirai, H., M. F. Roussel, J. Y. Kato, R. A. Aashmun, and C. J.
Sherr. Novel INK4 proteins, p19 and p18, are specific inhibitors
of the cyclin D-dependent kinases CDK4 and CDK6. Mol. Cell.
Biol. 15: 2672–2681, 1995.
22. Jones, B. A., and G. J. Gores. Physiology and pathophysiology
of apoptosis in epithelial cells of the liver, pancreas, and intestine. Am. J. Physiol. 273 (Gastrointest. Liver Physiol. 36): G1174–
G1188, 1997.
23. Kiyokawa, H., V. M. Richon, R. A. Rifkind, and P. A. Marks.
Suppression of cyclin-dependent kinase 4 during induced differentiation of erythroleukemia cells. Mol. Cell. Biol. 14: 7195–
7203, 1994.
24. Ko, T. C., W. A. Bresnahan, and E. A. Thompson. Intestinal
cell cycle regulation. Prog. Cell Cycle Res. 3: 43–52, 1997.
25. Kranenburg, O., V. Scharnhorst, A. J. van der Eb, and A.
Zantema. Inhibition of cyclin-dependent kinase activity triggers
neuronal differentiation of mouse neuroblastoma cells. J. Cell
Biol. 131: 227–234, 1995.
26. Kranenburg, O., A. J. van der Eb, and A. Zantema. Cyclindependent kinases and pRb: regulators of the proliferationdifferentiation switch. FEBS Lett. 367: 103–106, 1995.
27. Markowitz, A. J., G. D. Wu, A. Bader, Z. Cui, L. Chen, and
P. G. Traber. Regulation of lineage-specific transcription of the
sucrase-isomaltase gene in transgenic mice and cell lines. Am. J.
Physiol. 269 (Gastrointest. Liver Physiol. 32): G925–G939, 1995.
28. Messer, M., and A. Dahlqvist. A one step ultramicro method for
the assay of intestinal disaccharidases. Anal. Biochem. 14:
376–392, 1966.
29. Morgan, D. O. Principles of CDK regulation. Nature 374:
131–134, 1995.
30. Morse, L., D. Chen, D. Franklin, Y. Xiong, and S. ChenKiang. Induction of cell cycle arrest and B cell terminal differentiation by CDK inhibitor p18(INK4c) and IL-6. Immunity 6:
47–56, 1997.
31. Nigg, E. A. Cyclin-dependent protein kinases: key regulators of
the eukaryotic cell cycle. Bioessays 17: 471–480, 1995.
32. Nurse, P. Ordering S phase and M phase in the cell cycle. Cell 79:
547–550, 1994.
33. Pardee, A. B. G1 events and regulation of cell proliferation.
Science 246: 603–608, 1989.
34. Pinto, M., S. Robine-Léon, M. D. Appay, M. Kedinger, N.
Triadou, E. Dussaulx, B. Lacroix, P. Simon-Assmann, K.
Haffen, J. Fogh, and A. Zweibaum. Enterocyte-like differentiation and polarization of the human colon carcinoma cell line
Caco-2 in culture. Biol. Cell 47: 323–330, 1983.