Upload
others
View
5
Download
0
Embed Size (px)
Citation preview
Vol. 4, 1251-1261, May 1998 Clinical Cancer Research 1251
Loss of �21WAF1/CiP1 Protein Expression Accompanies Progression of
Sporadic Colorectal Neoplasms but not Hereditary Nonpolyposis
Colorectal Cancers’
Frank A. Sinicrope,2 Gardiner Roddey,
Michael Lemoine, Sanbao Ruan,
L. Clifton Stephens, Marsha L. Frazier, Yu Shen,
and Wei ZhangDepartments of Gastrointestinal Medical Oncology & Digestive
Diseases [F. A. S., G. R., M. L., M. L. F.], Neuro-oncobogy [S. R..
W. Z.], Veterinary Medicine & Surgery [L. C. S.], and
Biomathematics [Y. S.], The University of Texas M. D. AndersonCancer Center, Houston, Texas 77030
ABSTRACTp2! (p21WAF1/CiP1), a cyclin-dependent kinase inhibitor,
induces G1 arrest and can inhibit the activity of the prolif-
erating cell nuclear antigen (PCNA). We analyzed p21 ex-
pression during colorectal tumorigenesis, its associationwith its transcriptional regulator p53, and its relationship torates of cell proliferation and apoptosis. p2! and p53 protein
expression were examined in sporadic tumors and heredi-tary nonpolyposis colorectal cancers (HNPCCs) by immu-
nohistochemistry (LHC) and immunoblotting. Apoptosis was
examined using a DNA nick end-labeling assay, and cellproliferation was examined by PCNA staining. In normal
colorectal epithelia, nuclear p2! staining was uniformly de-
� tected in crypt cells of the superficial compartment (upper
one-third) that stained negatively for PCNA. p21 and PCNAexpression were, therefore, mutually exclusive. In sporadic
cases, a decrease in the frequency of p2! expression accom-
panied adenoma development and progression to carci-
noma. Specifically, p21 was detected in 12 of 16 (75%)adenomas and 10 of 32 (31 %) carcinomas. In contrast tosporadic cases, HNPCCs with known mutations in DNA
mismatch repair genes expressed p21 in 12 of 15 (80%)
carcinomas. An inverse relationship between p2! and p53was observed wherein mutant p53 proteins were detected in4 of 15 (27%) HNPCCs versus 22 of 32 (69%) sporadiccarcinomas. Although p2! + carcinoma cells were generally
negative for p53, IHC revealed that some carcinoma cells
expressed both p2! and p53 proteins. Furthermore, p53-
Received 10/15/97; revised 1/6/98: accepted 2/19/98.The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to
indicate this fact.I This work was supported by a grant from the Glaxo Institute of
Digestive Health (to F. A. S.). F. A. S. is a recipient of a Career Devel-opment Award from the American Cancer Society.2 To whom requests for reprints should be addressed, at Box 78, 1515Holcombe Boulevard, Houston, TX 77030. Phone: (713) 792-2828;
Fax: (713)745-1163.
mutated SW480 colon carcinoma cells were found to coex-press p21 and p53, suggesting that p2! can also be activated
by a p53-independent mechanism. No association was foundbetween p2! or PCNA and apoptotic labeling indices in
adenomas or carcinomas. In conclusion, a decrease in p2!expression accompanies neoplastic progression in sporadiccases but not in HNPCCs. This finding appears related top53 status in that the frequency of p53 expression wassignificantly reduced in HNPCCs compared to sporadiccases, suggesting a difference in their molecular pathways of
tumorigenesis.
INTRODUCTIONThe p21 (p21WAFI/CIPI) gene is located on chromosome 6p
and encodes the p21 protein, which functions as a potent inhib-
itor of cdk activity ( 1 , 2). Members of a family of cdks function
as key positive regulators of the cell cycle that promote cell
proliferation. p21 induces G, arrest and blocks entry into S
phase by inhibiting cdks3 (1, 2) or by binding to the PCNA,
resulting in a blockade in DNA replication (3, 4). p21 is a
downstream target effector of the wt p53 protein, which tram-
scriptionaiiy activates p21 (5, 6). wt p53 proteins accumulate as
a consequence of DNA damage and can bind to specific sites on
the p21 promoter, resulting in induction of p21 expression (5).
Mutational inactivation of p53, a frequent event in coborectal
cancers (7), with resultant failure to tramsactivate p21, may
thereby produce uncontrolled cell proliferation.
Transfection of the p21 cDNA and overexpressiom of the
p21 protein have been shown to inhibit the growth of cultured
colon, brain, and lung cancer cells, as well as leukemia cells (5,
8, 9). Furthermore, p21 overexpression reduced the tumorige-
nicity of colon carcinoma cells when injected into nude mice
(9). Studies in colon cancer cells (6) and in p21-deficient murine
fibroblasts ( 10) have shown that p21 is necessary for p53-
mediated G1 arrest induced by DNA damage. The requirement
for p21 in p53-mediated growth arrest suggests that p21 plays a
critical role in p53-mediated tumor suppression (6, 1 1 ). Sup-
pression of tumor growth may occur secondary to cell cycle
arrest or, alternatively, by induction of apoptosis (12). Interest-
ingly, loss of p53 function has been associated with attenuated
apoptotic rates in vivo (12). Although p53 appears to be a key
regulator of p2 1 , abundant evidence indicates that p2 1 can also
be induced independently of p53 (13-16). In most normal tis-
3 The abbreviations used are: cdk, cyclin-dependent kinase; HNPCC,hereditary nonpolyposis coborectal cancer; PCNA, proliferating cellnuclear antigen: wt. wild-type; TUNEL. terminal deoxynucleotidyl
transferase-mediated nick end labeling; MMR. mismatch repair; IHC.
immunohistochemistry; RT, room temperature: MAb. monocbonal anti-body; DAB, diaminobenzidine: Al, apoptotic index; LI, labeling index.
Research. on October 24, 2020. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
Table 1 Analysis of HNPCC cases for mutations in hMLHI and
hMSH2 genes determined by DNA sequencing (23) or hMSH2
protein expression. p21 and p53 expression were also examined
by IHC
Protein expression
a Absence of staining (-) is consistent with germline or somatic
mutation (24).1� Percent tumor cell positivity (0, �5; I , 6-25; 2, 26-50: 3, 51-75;
4, �75).
‘ ND. not determined.
1252 p21w��c� in Coborectal Carcinomas and HNPCCs
sues of the adult mouse, including the intestinal epithelium, p21
expression is p53 independent (14, 15). p53-independent ex-
pression of p21 has also been reported in human leukemia (16),
non-small cell lung carcinoma ( 1 7), and ovarian carcinoma (18).
The p2 1 protein has been detected in a variety of normal
tissues, including colorectal epithelia, where its expression has
been associated with terminally differentiated or senescent cells
that have ceased to divide ( 14, 19). These observations indicate
that p21 is involved in regulating epithelial cell turnover. Po-
tentially. alterations in p2 1 expression may deregulate growth
and contribute to neoplastic development and/or progression. To
address this issue, we examined p21 and PCNA expression at
histological stages in the development of human colorectal
cancer. Given that wtp53 transcniptionally activates p21 (5), we
determined the relationship of p21 to p53 expression in these
tissues. The association between p2 1 and rates of spontaneous
apoptosis, as assessed by the TUNEL assay (20), was also
examined. p21 and p53 expression in sporadic coborectal carci-
nomas were compared to results in HNPCCs. HNPCC is a
dominantly inherited disorder characterized by multiple cases of
colorectal cancer within a family (21 ). HNPCC accounts for up
to 5% of all coborectal cancer cases and is caused by germ-line
mutations in at least four DNA MMR genes (2 1). Two MMR
genes, hMLH/ and hMSH2, account for the majority of MMR
mutations that result in genomic instability and early onset of
colorectal cancer. MMR-deficient tumors display microsateilite
instability as a consequence of an accumulation of mutations at
short tandem repeats of DNA, known as microsateilites, which
are distributed throughout the genome (22). Some of these
mutated genes contribute to tumonigenesis in HNPCC, and
studies are ongoing to identify these specific mutational targets.
To date, p2 1 expression has not been examined in HNPCCs.
MATERIALS AND METHODSTissue Specimens. Formalin-fixed, paraffin-embedded
tissue blocks were obtained from the Surgical Pathology Labo-
ratory at the University of Texas M. D. Anderson Cancer Center
and from referring institutions. Normal colorectal epithelia (ii =
I 1 ) were obtained from the resection margins of colorectal
cancer cases. Hyperplastic (n = 5) and adenomatous (n = 16)
polyps were obtained from both surgical and endoscopic biopsy
specimens. All adenomas had tubular or tubulovillous histology,
with the exception of one villous tumor. Three adenomas were
from patients with synchronous colorectal carcinomas, and an-
other three adenomas contained foci of severe dysplasia. All
nonneoplastic tissues were obtained from patients without din-
ical evidence of a familial colon cancer syndrome, i.e. , sporadic
cases. Primary, untreated colorectal adenocarcinomas were from
both sporadic [29 colon and 3 rectal (ii = 32)] and HNPCC
[colon (ii = 15)1 patients. Ten of 15 HNPCC cases had known
germ-line mutations in hMLHI or hMSH2 MMR genes as
previously determined by sequencing of genomic DNA (23).
The remaining five also came from families meeting the Am-
sterdam criteria (2 1 ), and four of these had loss of hMSH2
protein expression by IHC. Specifically, tumor cells lacked
nuclear hMSH2 staining in contrast to adjacent normal epithe-
hum. Absence of hMHS2 immunoreactivity is consistent with
either germ-line or somatic hMSH2 mutation (24). In the seven
no. MutationhMSH2” p21” p531’-
I hMSH2/hMLHI ND - 0 0
2 hMSH2 + - 1 0
3 hMHS2 + - 1 0
4 hMHS2 + + 4 1
5 hMLHI + 0 0
6 hMSH2 + - 1 0
7 hMSH2/hHLHJ ND - I 0
8 /MLHJ + 2 1
9 hMSH2//MLHI ND - 3 0
10 hMSH2/hMLHI ND - I 0
11 /MLHI + 0 1
12 hMHS2 + - 1 0
13 /iMSH2 + - I 0
14 hMSH2 + - 1 0
I 5 hMSH2//iMLHI ND + I 4
cases with hMSH2 mutations detected by DNA sequencing,
concordant results were found by IHC in all but one case. This
discordant case (patient 4; Table 1) had a missense mutation that
did not result in a loss of immunoreactivity. Tumor stage of
HNPCCs was not available. For sporadic cancers and using the
tumor-node-metastasis classification, there were 4 stage II, 22
stage III, and 6 stage IV tumors. Degree of tumor differentiation
for sporadic cases was as follows: 3 well, 24 moderately, and 5
poorly differentiated. Eight of 15 HNPCCs were poorly differ-
entiated, and 7 were moderately differentiated.
Consecutive 4-6-�i.m tissue sections were cut from each
tumor block for subsequent analysis. In additional to archival
materials, we also analyzed fresh tissue from sporadic colorectal
carcinomas (n = 5) and histologically normal tissue from the
resection margins (distance of >10 cm from primary tumor).
Tissue specimens were immediately frozen on dry ice and
ethanol and stored at -75#{176}Cuntil use. To ensure that tumor
tissue was being analyzed, H&E sections were prepared and
examined from the cut margin of the tumor and found to be
highly tumor enriched. Furthermore, formalin-fixed, paraffin-
embedded sections from these cases were examined for diagno-
sis, and additional sections were cut for a comparison of p2 1 and
p53 expression by IHC versus immunoblotting.
Cell Culture. The SW480, HT29, and RKO human co-
ion carcinoma cell lines were obtained from American Type
Culture Collection (Rockville, MD) and grown in DMEM:F-12
medium (Life Technologies, Inc.) supplemented with heat-mac-
tivated 10% FCS and l0� units/mi penicillin and streptomycin.
Cultures were maintained in a humidified environment at 37#{176}C
with 5% CO2. The LN-Z308 human gliobiastoma cell line
expresses low levels of p2 1 endogenously and lacks wt p53. As
previously reported (25), wt p53 transfected into LN-Z308 cells,
which activated p21 expression and inhibited cell growth. LN-
Research. on October 24, 2020. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
Clinical Cancer Research 1253
Z308 cells were maintained in DMEM:F-l2 medium supple-
mented with 10% FCS and maintained as described above.
Immunoblots. Frozen tissue samples and cultured cell
lines were processed per standard procedures and a cell lysate
was prepared (26). Protein extraction was performed on tissue
and cell lysates and the proteins were analyzed on a SDS-
polyacrylamide gel as described previously (26). Proteins were
then transferred to am Immobilon membrane (Millipore, Bed-
ford, MA) and incubated overnight with MAbs against p21, p53,
and actim (Oncogeme Science, Uniondale, NY). The level of
protein expression was measured using an enhanced chemilu-
minescence detection system (Amersham, Arlington Heights,
IL) per the manufacturer’s instructions. In an attempt to quan-
titate the level of p2 1 expression in immunoblots, autoradiog-
raphy films were scanned using a densitometer. Density values
for p21 were calculated in carcinomas relative to normal tissue,
in which mean density was designated as 1 .0. Actin levels were
analyzed as controls for protein loading.
IHC. An immunoperoxidase method using an avidin-
biotinylated horseradish peroxidase complex (Vectastain Elite
ABC) was used to stain all slides. p21 staining was performed
as follows. Slides were placed in xylene and then rehydrated
through graded alcohols. Endogenous peroxidase activity was
blocked by incubating slides for 30 mm in 3% H202 at RT,
followed by rinsing in deionized water. Sections were then
placed in 1 M sodium citrate buffer (pH 6) and microwaved on
high for 10 mm. After rinsing in PBS, sections were treated with
a nomserum protein block (Dako Corp., Carpinteria, CA) for 10
mm followed by normal horse serum for 20 mm. Slides were
then incubated overnight at RT with the WAFI murine IgG1
MAb (Ab-l ; Oncogene Science, Uniondale, NY) to p21 at a
dilution of 1 :75. Following washing in PBS for 5 mm, the
secondary antibody was applied for 30 mm at RT. Slides were
again rinsed in PBS, and them the ABC reagent was applied for
45 mm at RT. The chromogem DAB (Research Genetics, Hunts-
yule, AL) was subsequently applied to each section, and the
color reaction was observed using light microscopy. The reac-
tion was stopped by immersing slides in deionized water, and a
hematoxylim counterstain was applied. Slides were then cover-
slipped using a nonaqueous mounting medium.
The procedure to detect hMSH2 proteins was as outlined
for p21 above. Sections were incubated overnight with the
hMSH2 MAb (Ab-2; Oncogene Science) at a dilution of 1 :80 at
RT. For p53 (Ab-6; Oncogene Science), the procedure was
again as above except that tissue sections were placed in PBS
and microwaved on high for 4 mm, as described previously (25).
The staining method used for PCNA was identical to that
reported for p2 1 , excepting that (a) microwaving of slides was
not performed, and (b) slides were incubated with the primary
antibody at a dilution of 1 : 100 for 1 h at RT (27). The anti-
PCNA MAb (PC1O; Signet Laboratories, Dedham, MA) is a
murime antirat subclass IgG2�, known to recognize the DNA
polymerase � accessory protein in all vertebrate species.
For double immunostaining, anti-p21 was the first MAb
applied, and the procedure was performed as outlined for p21
above excepting that nickel chloride was added to DAB (0.5 mg
in 2 ml), and the color reaction was developed. The second
primary MAb, amti-p53, was then applied, and the remainder of
the procedure was identical to that described for p53.
For p21, normal human epidermis was used as a positive
control (Fig. 1A). Appropriate positive and negative tissue con-
trols used for p53 and PCNA were described previously (27). In
negative control slides, the primary antibody was omitted, and
no immunostaiming was observed. In colorectal neoplasms, the
percentage of tumor cells expressing p2 1 was analyzed in each
section at light microscopy and arbitrarily categorized as
6-25%, or >25%. Scoring for p53 is shown in Table 1. Top-
ographical distribution of staining was also assessed. Tumors
with �5% immunoreactive cells were considered negative (25).
A PCNA LI was calculated as the ratio of the number of PCNA
positive cells to 500 total cells examined, expressed as a per-
centage (27).
Labeling of Apoptotic Cells by TUNEL Assay. A
method has been described that allows in situ detection of
apoptotic cells using terminal deoxynucleotidyl transferase in-
corporation of labeled nucleotides into the 3 ‘ end of DNA strand
breaks (20). Apoptotic cells are then identified using an immu-
noperoxidase detection system. After deparaffinization and re-
hydration, tissue sections were incubated in 20 p.g/ml of pro-
teimase K (Sigma Chemical Co., St. Louis, MO) for 15 mm at
RT. After rinsing in PBS for 5 mm, endogemous peroxidase
activity was blocked by incubating the sections in 3% H,02 for
30 mm at RT. Following rinsing in PBS. the slides were covered
with terminal deoxymucleotidyl transferase plus nucleotide mix-
ture (TUNEL reagent; In Situ Cell Death Detection kit; Boeh-
ringer Mannheim, Indianapolis, IN) at a 1 :35 dilution and incu-
bated for 60 mm in a humid chamber at 37#{176}C.The reaction was
terminated by immersing the slides in PBS containing 2% BSA
for 10 mm at RT. The sections were then covered with a
nomserum protein block (Dako Corp.) for 10 mm at RT. Fol-
lowing rinsing in PBS with 2% BSA for 5 mm, slides were
covered with am antifluorescein antibody conjugated with horse-
radish peroxidase for 15 mm at RT. Following rinsing in PBS
plus 2% BSA, DAB (Research Genetics) was applied, and the
reaction was observed under light microscopy. Positive control
slides consisted of a tissue section exposed to DNase I and am
irradiated murime lymphoma known to be radiation sensitive.
Exposure to DNase I produced staining of all nuclei in the
section. In the irradiated lymphoma, numerous TUNEL-stained
apoptotic cells and bodies were detected. For negative controls,
the TUNEL reagent was omitted, and no immunoreactive cells
were observed.
Quantitation of Apoptotic Cells in Histological Sections.Using light microscopy, TUNEL-stained cells were identified in
tissue sections by a brown reaction product confined to the
nucleus or nuclear fragments (Fig. 5; Ref. 27). In addition to
TUNEL staining, morphological features of apoptosis were also
required given that apoptosis can occur in the absence of DNA
strand breaks (28) and that TUNEL positivity may not always
indicate impending cell death (29). The number of apoptotic
nuclei per 100 cells in five high-power fields per slide was
determined. An Al was then calculated as follows:
number of apoptotic cellsAI= xl00
total number of cells
Statistics. The relationship between p21 and p53 protein
expression was analyzed using the nonparametric Kruskal-
Research. on October 24, 2020. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
p
1�1.
#{149}0
V
�- ��4:
;,C �
�I
. ; .‘�
‘. -.t 7�_:. �,. +,‘4 �
� �l(
.�:..r’
k �
� �t�i’
. s�q,\ � ‘�
� :� � � �� 4,
:r’�.r.- �
. �C#{149}
. .. . ,.. : � � :
.#{149}- � CII. .. e.
..
D
..�. .
“p �
,�1�#{149}
I#{149}� ij4
.� .‘,. ���i�’4 td
�‘ �. .‘ .:
�, �.
‘ .. F
1254 p2 1 WAFI/CipI in Coborectal Carcinomas and HNPCCs
,
.�. a �
. _1�_, �‘%p_� � #{149} � _; ..,. : - �
.... .� .-. .- ‘ .1
,. . . � ,_�,: .‘..� . - .. ‘ . , ..% .1. #{149}� #{149}#{149}�
‘�:#{149}�‘:. �i:.’:.#{149}�s.#{149}.%.:.;e.t�� �%
,�‘ � .�e. �
�, ‘I � le:jk.a�14%4I�&s,4r�& ,
.: ,. . . � - ‘ � : :. �. � :� � . .‘, � � �
, S. � � � <�/d$� � � “� �*1 � � .. � 4� #{149}�:‘ � � ::_.:� :� �‘
�‘ “� ..�4 � 4t:l� , . 0
� � 1, �
..
.�. *�:� #{149}4#{149}��#{149}� C�
.� .�I-. �
F VG”� �
�:#{149}�.‘ �
.;,. 4
Fig. I Immunohistochemical analysis of p2 1 expression in normal and neoplastic coborectal epithelia. Dark staining indicates a positive signal. A,
p2 1 staining is seen in epithelial cell nuclei in the suprabasal layer of normal epidermis (positive control). B, normal colorectal epithelia demonstrates
nuclear p21 staining in differentiated crypt cells of the superficial compartment. These p21-positive cells stained negatively for PCNA (see Fig. 2).
C, transition of normal and adenomatous coborectal epithelium. p21 staining in the adenoma localizes to the superficial region. D, adenoma displayingp21 staining in superficial crypt cells. E, p21 staining in tumor cell nuclei in a well-differentiated colon carcinoma. F. p21 staining in a poorly
differentiated colon carcinoma.
Research. on October 24, 2020. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
C �
. - ‘ .�‘‘.. .., ....... ‘oil
‘- #{149} I, . , #{149}�,. ...- 4.
. 4 _l �‘#{149}�: .� #{149}� # .
,l I � � �
. . � . , .I :‘� �
‘I � � �1
. . � .l�..’
� .� � L�1f��
�/H �
.,‘
.� .,
. :.‘ p�. t
‘I.)
..
i#{149}L� . . . .
i�/’� ..
�
,. C
; .l..4
‘l
;: �,�.:�‘‘::.#{149}�
#{176}1
Fig. 2 PCNA immunostaining in coborectal epi-thelia. A and B, PCNA-reactive cells are shown by
dark nuclear staining and are confined to theproliferative compartment of normal (A) and hy-
perplastic (B) mucosa. Localization of PCNA
staining in nonneoplastic epithelia is opposite of
that for p2l (Fig. 1). C, adenoma with diffuse
PCNA staining extending to superficial crypt
cells. D, PCNA-positive cells in malignant glandsof a colon carcinoma; overlying normal mucosa isnoted at the top.
Wallis statistic. Relationships between p21 and Al and PCNA
LI values in neopiasms were assessed by the Spearman’s cor-
relation test and the Kruskal-Waliis test. The x2 test was used to
assess the relationship or dependence of the categorical van-
ables between two groups. Statistical significance was defined
as a two-sided P of <0.05.
RESULTS
p21 and PCNA protein expression were examined in four
histological types of colorectal epithelia representing histologi-
cal stages in the evolution of colorectal cancer. p2 1-positive
cells were identified by dark brown nuclear staining. In normal
human epidermis that served as a positive control, p2 1 expres-
siom was detected in epithelial cells in the suprabasal layer (Fig.
1A). In normal and hyperplastic colorectal epithelia from spo-
Clinical Cancer Research 1255
radic cases, p21 was uniformly expressed in crypt cell nuclei in
the superficial one-third of the mucosa. including luminal epi-
thelial cells. p21 was therefore topographically restricted to
nomproliferating and differentiated crypt cells (Fig. 1B). In con-
trast to p2 1 , PCNA immunoreactive cells were confined to the
lower one-third of normal mucosa, corresponding to the prolif-
erative compartment (Fig. 2A). Thus, PCNA and p21 staining
were inversely distributed and nonoverlapping in their topogra-
phy (Figs. lB and 14). This finding is consistent with another
study in colonic epithelia. in which double staining for p21 and
Ki-67 proteins revealed that their expression was mutually cx-
clusive (19).
Whereas p2 1 was expressed in all specimens of normal and
hyperplastic mucosa, p21 staining was found in 12 of 16 (75%)
sporadic adenomas (Table 2). Hence, a reduction in the fre-
Research. on October 24, 2020. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
Table 3 Percentage of p21 + tumor cells detected by IHC in
colorectal neoplasms
1256 p2l’�’�” in Coborectal Carcinomas and HNPCCs
Table 2 p2 1 and p53 protein expression during colorectal
tumorigenesis#{176}
HNPCC”
Carcinoma(n 15)
12(80%)
4 (27%)
Sporadic
Normal Hyperplastic Adenoma Carcinoma
(n 11) (n 5) (,i 16) (n 32)
p21+ 11 (l(X)%) 510()%) 12(75%) 10(31%)
p513+ 0 0 5(31%) 22(69%)
(4 Number and percentage of specimens that overexpress p2 1 or p53
as detected by IHC.
I, See Table I.
quency of p2 1 expression was associated with neoplastic devel-
opment in sporadic patients. p21 staining was found predomi-
nantly in crypt cells in superficial regions of adenomas and
tended to occur in clusters (Fig. 1, C and D). In the majority of
adenomas [10 of 16 (62%)], staining was found in �25% of the
total tumor cell populations (Table 3, Fig. 10). However, a
subset of adenomas with >25% p21-positive cells was found in
which p21 staining extended deep into the tubular glands (Table
3). Staining intensity was medium to strong in all adenomas and
did not differ appreciably from p21 staining intensity in non-
neoplastic epithelia (Fig. I). Whereas p21 staining was prefer-
entially expressed in superficial versus basal regions of ademo-
mas (Fig. I , C and D), no specific pattern of PCNA expression
was observed (Fig. 2C). Hence, the strict compartmentalization
of p21 and its inverse relationship with PCNA that were seen in
normal epithelia became disorganized in the adenomas.
A further reduction in the frequency of p21 expression
occurred during progression from adenoma to carcinoma in
sporadic patients. Specifically, 10 of 32 (31%) carcinomas cx-
pressed p21 proteins (Table 2). p21 nuclear staining was de-
tected in clusters of tumor cells, but no spatial distribution of
p21 staining was discernable, given the architectural disorgani-
zatiom of carcinomas. As with adenomas, a subset of carcinomas
(3 cases) expressed p21 in >25% oftumor cells, consistent with
a high level of expression (Table 3). Thus, the level of p21
expression in adenomas and carcinomas was heterogeneous. Of
all p21 + carcinomas, one-third had strong p21 staining, and the
remainder had staining that was medium in intensity. In contrast
to sporadic cases, 12 of 15 (80%) HNPCCs were found to
express p2l proteins (Table 2). At least 25% p21 + tumor cells
were detected in 3 (20%) cases (Table 3), and the pattern of p21
staining in HNPCCs was similar to that in sporadic cases.
Importantly, these results pertain to HNPCC cases with known
mutations in hMLHJ or hMSH2 genes (n = 14), as determined
by DNA sequencing or analysis of protein expression (Table 1;
see “Materials and Methods”). Absence of hMSH2 immumore-
activity is consistent with either germiime or somatic hMSH2
mutation (Fig. 3).
Nuclear p53 expression, consistent with the accumulation
of mutant p53 proteins (30), was detected in 5 of 16 (31%)
adenomas and 22 of 32 (69%) carcinomas from sporadic pa-
tients (Table 2; Ref. 25). p53 staining was focally expressed in
dysplastic crypt cells but was not detected in either normal or
hyperplastic colorectal epithelium. In HNPCCs, p53 expression
was detected in 4 of 15 (27%) carcinomas (Table 2), suggesting
a significant decrease in the frequency ofp53 mutation in these
% of p 21-positive cells
Total�5 6-25 >25
AdenomaCarcinoma
Sporadic
HNPCC
4(25%)
22 (69%)
3 (20%)
6(37%)
7 (22%)
9 (60%)
6(37%)
3 (9%)
3 (20%)
16(100%)
32 (100%)15 (100%)
tumors. Furthermore, 3 of 4 p53 + HNPCCs had < 25% p53 +
tumor cells (Table 3). This is in contrast to sporadic carcinomas,
in which 13 of 32 (41%) tumors had >75% p53+ tumor cells.
Given that wt p53 transcriptionaily activates p21, we hypothe-
sized that p53 expression would be inversely related to p21.
Therefore, we compared the frequency of p2! expression in
p53 + versus p53 - meopiasms. An inverse association between
p53 and p21 was found in HNPCCs. Specifically, p53-
HNPCCs expressed p21 in 9 of 15 (60%) cases, and p53+
tumors had detectable p21 in 3 of 15 (20%) cases. These data
suggest that the low rate of mutant p53 expression compared to
sporadic cases can explain the increased frequency of p21
expression. In sporadic cases, p2 1 staining was detected in 8 of
1 1 PS3- versus 4 of 5 p53+ adenomas (P = 0.75). In carci-
nomas, 5 of 1 1 p53 - versus 10 of 2 1 p53 + tumors expressed
p21 (P = 0.80).
Thirteen of 32 (41%) sporadic carcinomas had diffuse p53
staining and were found to express p21 in 6 (46%) cases,
suggesting that some tumor cells may coexpress p53 and p21
proteins. To address this issue, we performed double immuno-
staining and found that p53 and p21 were rarely detected withinthe same malignant gland. In some malignant glands, however,
tumor cell nuclei appeared to coexpress p53 and p21, and within
this subset, both proteins were clearly detected in adjacent tumor
cells. By immunoblottimg, frozen tissues from 3 of 5 sporadic
carcinomas were found to coexpress p53 and p2! proteins
(Table 4). Furthermore, SW480 colon carcinoma cells, known to
carry a p53 mutation (31, 32), overexpressed p53 and p21.
These findings suggest that p21 can be induced by a p.53-
independent mechanism. Overexpression of p53 was detected in
p53 mutated SW480 and HT29 colon carcinoma cells (Fig. 4;
Refs. 31 and 32) but not in wt p53-containing RKO cells (32).
Taken together, our data support p53-dependent and p53-inde-
pendent induction of p2 1 . Moreover, our results and those of
others (33) indicate that p53-dependent p21 induction is the
predominant mechanism in colorectal carcinomas.
We attempted to quantitate the level of p21 expression in
carcinomas relative to histologically normal mucosa from the
same patients. Frozen tissue from five sporadic colon cancers
was analyzed by immunobiotting, and autoradiographs were
subjected to demsitometry. The p53-deficient human glioblas-
toma cell line LN-Z308 was included as a positive control (34).
LN-Z308 cells were previously transfected with a wt p53 cx-
pression vector (34), and immunobiotting detected a protein
band of 21 kilodaltoms (Fig. 4). Colon carcinoma cell lines were
also examined, and a high level of p2! expression was detected
in SW480 and RKO cells (Fig. 4); however, p53 mutated HT-29
cells were observed to lack p2 1 expression. In carcinoma spec-
Research. on October 24, 2020. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
\‘#{149}\ :�;,
�: ‘ ‘:
.. :1, , � � �
�)
C � � �
..:
Table 4 p2 1 and p53 expression in sporadic coborectal carcinomas detected by IHC and immunobbot assays
p21 expression p53 expression
Patient IHC Immunobbot IHC
specimen (% positive cells)” (relative level of p21)” (% positive cells)”
.,3
4
5
6-25
6-25
6-25
8.0
6.8
2.3
0
>75
0
Immunobbot
(p53 status)’
+
+
+
Clinical Cancer Research 1257
Fig. 3 Colon carcinoma and adjacent normal co-
lonic epithelium from a patient with HNPCC ana-lyzed for hMSH2 protein expression. Immunoreac-
tivity is seen in nuclei of normal cobonic crypt cells
(left) but not in malignant glands (rig/It). Absence of
hMSH2 staining in the carcinoma is consistent with
/iMSH2 gene mutation (24).
1.5
50-756.2 50-75
(4 Percentage of positive cells was determined by IHC and assigned to predetermined categories (see �‘Materials and Methods”).
I, Relative level of p2 1 proteins was determined by scanning autoradiography films using a densitometer. Values for p2 1 refer to density relative
to normal cobonic tissue (mean density designated as 1 .0). Immunoblots were standardized by the level of actin to control for protein loading.
4 See Fig. 4.
imens, p2 1 expression was detected in 5 of S tumors, and a
1.5-8-fold increase in the level of p21 was found relative to
normal mucosa (Fig. 4, Table 4). In paraffin sections from these
same tumors, the percentage of p21 + cells did not predict the
relative level of p21 (Table 4). However, normal mucosa was
not present on these sections, and therefore, the absence of a
denominator precluded a direct comparison.
As shown in this study and in others (19, 33), p21 and
PCNA expression in normal colorectal epithelia were inversely
distributed and mutually exclusive in their topography. This
inverse relationship was maintained in hyperpiastic polyps. In
sporadic adenomas and carcinomas, significantly fewer p2 1 -
reactive cells were seen compared to PCNA (Figs. lB and 14).
Furthermore, the percentage of PCNA-positive tumor cells was
greater in neoplastic than in nonneoplastic mucosa (Fig. 2). We
then examined the PCNA LI stratified by p21 staining. The
mean PCNA LI of p2 1-positive versus p2 1-negative adenomas
was 43.3 and 32.4, respectively (P = 0. 1 1 ; Table 5). p2 1-
positive carcinomas had a mean PCNA LI of 44.3 versus 43.2
for p21-negative cancers (n = 21; P = 0.85). Thus, p21 expres-
sion was not significantly related to overall proliferative activity
in colorectal neoplasms. To determine whether an association
existed between p2 1 and rates of spontaneous apoptosis, we
quantified apoptotic cells in paraffin sections using the TUNEL
assay in addition to morphological criteria (Fig. 5), as described
previously (27). We found that the mean Al in p21-positive
versus p21-negative adenomas was 2.7 and 1 .9 (P 0. 18) and
for carcinomas was 2.5 and 2.3, respectively (P = 0.66; Table
5). The above results did not change significantly when the p53
status of these neoplasms was considered. Thus, no relationship
between p2 1 expression and in situ apoptotic rates was ob-
served.
DISCUSSION
Our results demonstrate a significant difference in p21
expression in sporadic colorectal carcinomas versus HNPCCs
with known mutations in MMR genes. In sporadic patients, we
found a reduction in the frequency of p21 expression during
adenoma development and progression to carcinoma. Complete
Research. on October 24, 2020. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
p53 -3’
actrn -3’
p21 -3’
-
1258 �21WAFh�1 in Colorectal Carcinomas and HNPCCs
LN-Z308
>Lt�
u� TNTNTN
1 2 3 4 5
_____ __ __ TNTN ____
� � :- � �,
Fig.4 Expression of p2 1 and p53 proteins was analyzed by immunobbotting in colon carcinoma cell lines as well as colorectal carcinomas (tumorenriched) and normal mucosa from the same patients. Total protein was extracted from cell lines, and tissues and blots were probed for protein
expression using specific antibodies. Lane 1. p53-null LN-Z308 glioblastoma cells transfected with the C’MV.neo vector were used as negative control
for p53 and p2l proteins, respectively. Lane 2, wt p.53 transfection of LN-Z308 cells activates p21 expression as shown. Patient specimens 1-5
included tumor (T) and normal (N) tissues. In tumors 1, 2. 4, and 5, overexpression ofp2l and p53 was detected. Coexpression ofp2l and p53 proteins
in SW480 cells is consistent with p53-independent activation of p21 . The levels of p2 1 proteins in tumors relative to normal tissue were determinedby densitometry (see Table 4). IHC results for p21 and p53 in these same tissue specimens are shown in Table 4.
Table 5 Relationship of p2! to PCNA expression and apoptotic
indices in sporadic coborectal neoplasms
PCNA LI (%)“ Al (mean)”
Adenoma (ii = 16)p21- 32.4 1.9
P=0.ll P=0.l8
p21+ 43.3 2.7
Carcinoma (n = 32)
p21- 43.2 2.3
P=0.85 P=0.66
p21+ 44.3 2.5
4, PCNA labeling index (27).
I, Al, apoptotic index (see �‘Materials and Methods”).
loss of p2 1 expression occurred in 25% of ademomas and 69%
of carcinomas. Given that transfectiom and overexpression of
p2 1 in colon carcinoma cells has been shown to suppress their
proliferation, anchorage-independent growth, and tumorigenic-
ity when injected into nude mice (5, 6, 9), loss of p21 may
contribute to malignant transformation and progression. Al-
though p2 1 acts as a tumor suppressor in vitro and in vivo, mice
lacking a functional p21 gene do not develop tumors, indicating
that additional genetic alterations are required for tumonigenesis
( I 0). In contrast to sporadic cancers, loss of p2 1 was observed
in only 20% of HNPCCs. The disparity in the frequency of p21
expression may be related to the inverse relationship between
p21 and p53 expression found in HNPCCs, as discussed below.
p53 expression, consistent with the accumulation of mutant
p53 proteins, was found infrequently in HNPCCs. Specifically,
mutant p53 proteins were detected in 27% of HNPCCs versus
69% of sporadic carcinomas. This finding suggests that muta-
tional inactivation ofp53 in sporadic cases, with resultant failure
to transactivate p21, accounts for loss of p21 expression. The
low frequency of p53 expression found in HNPCCs is consistent
with some studies (35), but not all (36). In our study, however,
only HNPCCs with known MMR mutations were analyzed.
Additionally, Cottu et a!. (37) found that 4 MMR-deficient
human colon cancer cell lines had wt p53, whereas 15 of 17
MMR-proficient lines contained p53 mutations. Although pa-
tients with HNPCC show early onset of carcinoma, intact p53
function and overexpression of p2 1 may contribute to the better
prognosis seen in HNPCCs versus sporadic colorectal cancers
(21). Taken together, these findings suggest that HNPCCs,
unlike sporadic cases, frequently develop through a pathway
that is independent of p53 mutation.
Our results suggest that loss of p2 1 expression in sporadic
colorectal cancers may be a baseline event, and those tumors
retaining p21 expression can be considered a high p21 group. In
this regard, we found that the level of p21 in sporadic carcino-
mas was increased 1 .5- 8-fold relative to normal mucosa. Fur-
thermore, HNPCCs were found to overexpress p21 in 80% of
cases, reflecting am increased level of expression. These results
are consistent with data in human gliomas, in which the level of
p2 1 expression was markedly elevated compared to normal
brain tissue (26). Similarly, p21 RNA and protein were cx-
pressed at higher levels in non-small cell lung carcinomas than
in normal respiratory epithelia (17). Increased p21 expression
has also been reported in breast (38) and ovarian ( 18) carcino-
mas, as well as in acute myeiogenous leukemia (39). The
heterogeneity of p21 expression found in colorectal and brain
tumors is unlikely to be related to p21 mutations because so-
matic mutations in the p21 gene have not been found in either
tumor type, although polymorphisms were detected (40, 41). At
present, the biological and clinical implications of p21 overex-
pression in human cancers are poorly understood. However, a
recent study found that a high level of p2 1 in acute myelogenous
leukemia cells was associated with resistance to chemotherapy
(39). Additionally, data in breast carcinomas indicate that p21
overexpression in untreated tumors is associated with adverse
prognostic features and shorter disease-free survival (38). These
findings suggest that cell cycle inhibition by p21 may adversely
effect treatment response and/or clinical outcome.
Research. on October 24, 2020. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
.-�‘ �Fig. 5 In situ detection of apoptotic cells in a
colonic adenoma (A) and carcinoma (B) using
the TUNEL assay. Immunopositive apoptoticcells (indicated by arrows) display condensed
chromatin and nuclear fragmentation. Cob-lapsed nuclei are surrounded by clear halos.
‘1
I
#{149}.�;?
a...#{149})
p.
-�
‘I.
‘a’
�‘ a #{149}
#{149}�:‘. �
. (.S,� , � �
- . _t�� �;�;
Clinical Cancer Research 1259
t- .
‘�‘#{149}�“;�
.�‘.. ,:� “�
1� f’4;.�5’,
A comparison of p21 expression in normal and p53-null
mice revealed a similar pattern of expression in intestinal epi-
thelia (14, 19). Thus, p21 is constitutively expressed in normal
colorectal epithelia and is therefore p53 independent. In a subset
of sporadic colorectal carcinomas, some tumor cells were found
to coexpress p21 and p53 proteins, suggesting p53-independent
activation of p2 1 . Furthermore, we detected coexpression of p21
and p53 in p53-mutated SW480 colon cancer cells. Similarly,
studies in Caco-2 colon cancer cells have shown that p21
increases during spontaneous differentiation in the absence of
functional wt p53 protein (42). Our findings, as shown by IHC
and immunoblotting, are in agreement with in vitro data (13-
16), as well as data in human brain tumors (26), non small-cell
lung carcinomas (17), and ovarian carcinomas (18), in which
p53-independent expression of p21 has been demonstrated. Al-
though it is theoretically possible that some p.53 mutations may
retain transcriptional activity allowing induction of p21, Matsui
et a!. (3 1 ) showed that mutated p53 in SW480 cells failed to
activate p21. Additionally, p53 expression may not always
indicate p53 mutation, although expression by IHC is strongly
correlated with accumulation of mutant p53 proteins (30). Of
note, we found complete concordance between results for p53
using IHC and immunoblotting. Our data also support p53-
dependent activation of p21, because p53 and p21 staining were
rarely detected within the same tumor cells or malignant glands.
Furthermore, RKO colon cancer cells containing wt p53 (31)
expressed p2 1 . These data are consistent with the findings of
Doglioni et a!. (33), in which an inverse relationship between
p21 and p53 staining was generally observed in colorectal
carcinomas. Furthermore, this inverse relationship was clearly
seen in our HNPCC cases. Our results and those of others (33)
suggest that p53-dependent expression of p21 is the predomi-
nant mechanism of p2 1 induction in colorectal cancers and may,
therefore, contribute to the heterogeneity of p2 1 expression
found in these tumors.
Treatment of colon carcinoma cells by TGF-�3 has been
shown to transcriptionally induce p21 through a p53-independ-
ent mechanism (43). The increase in p21 mRNA is due to
transcriptional activation of the p21 promotor by TGF-�3. Induc-
tiom of p2 1 , may therefore, mediate the growth inhibitory effects
of TGF-� against colon cancer cells. The growth inhibitory
signal of TGF-�3 is transduced by a tramsmembrane type II
receptor. Interestingly, studies indicate that the TGF-�3 type II
receptor gene is frequently mutated in HNPCCs (44). Inactiva-
tion of the type II receptor leads to resistance to growth inhibi-
tion by TGF-�3 and may contribute to accelerated tumorigenesis
in HNPCC. Our finding of frequent expression of p2 1 in
HNPCCs, in contrast to sporadic colorectal cancers, suggests
that p2 1-mediated cell cycle inhibition may represent an alter-
native mechanism of growth control in the presence of TGF-�3
resistance.
In addition to inhibition of cdk activity, p2 1 binds PCNA,
resulting in a direct interference with DNA replication and,
hence, cell proliferation ( 1 , 3, 4). In normal epithelia, p21
expression was restricted to crypt cells in the superficial com-
partment, including terminally differentiated cells at the luminal
surface that stained negatively for PCNA. This finding suggests
that p2 1 activation accompanies migration and differentiation of
epithelial cells from the proliferative (PCNA+) to the nonpro-
liferative (PCNA-) compartment and is responsible for their
cell cycle arrest (14, 42, 45). Thus, p21 appears to be involved
in regulating growth in normal mucosa. This inverse relation-
ship between p21 and PCNA was maintained in hyperplastic
epithelia. In sporadic adenomas and carcinomas. however. the
PCNA LI did not differ significantly between p21 - and p2 1 +
neoplasms. Thus, adenoma development and progression to
carcinoma are associated with a loss of p21-mediated growth
suppression. These results are consistent with studies in human
gliomas (26) and ovarian carcinomas (46), in which no relation-
ship was found between p21 and proliferative activity. There-
Research. on October 24, 2020. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
1260 �11WAII/(,pI in Coborectal Carcinomas and HNPCC5
fore, p2 1 expression was not sufficient to suppress overall
proliferative activity. It is reported that the growth inhibitory
effects of p2 1 can be overcome by activated oncogenes, includ-
img c-mvc (47) or B-mvb (48). Furthermore, p53 mutations may
potentially counteract p2 1 function by elevation of positive cell
cycle regulators (34).
The correlation between p2 1 and the Al in untreated,
sporadic colorectal carcinomas was examined. Importantly, the
spontaneous Al in experimental murine tumors has been shown
to predict susceptibility to chemotherapy-induced (49) and ra-
diation-induced (SO) apoptosis. Controlling for p53 status, no
association between p21 and AIs was found. This result is
consistent with data indicating that p21 induction alone is in-
sufficient to promote apoptosis (5 1 ). However, p2 1 has been
shown to influence susceptibility to apoptosis induction in colon
carcinoma cells (52). Specifically. p2/-deleted HCT 1 16 colon
carcinoma cells were shown to undergo apoptosis in response to
radiation or chemotherapeutic drugs, whereas parental cells
expressing p21 underwent growth arrest (52). HCT I 16 cells
contain a mutated hMLHI gene and are MMR deficient (53).
p21 deficiency in HCT I 16 cells was associated with defective
DNA repair, which may account for their increased sensitivity to
DNA damage (54). These observations were confirmed in vito
in xenografts established from HCT 1 16 cell lines differing only
in their p21 status (55). Following irradiation, a cure fraction
was seen in p21-deficient tumors, in contrast to tumors express-
ing p2 1 . Extrapolation of these findings to our results would
suggest that HNPCCs, and potentially the subset of colorectal
cancers overexpressing p2 1 , may exhibit reduced sensitivity to
anticancer treatments. Studies are warranted to further address
the role of p2 1 in treatment resistance in colorectal cancer.
ACKNOWLEDGMENTSWe thank Hong Li. Wan Liang Dong, and David Voehringer for
their technical assistance and Dr. Raymond Meyn for helpful discus-
sions.
REFERENCES
I. Harper, J. W., Adami, G. R.. Wei. N.. Keyomarsi. K., and Elledge.
S. J. The p2l cdk-interacting protein CipI is a potent inhibitor of G1
cyclin-dependent kinases. Cell, 75: 805-816. 1993.
2. Xiong. Y.. Hannon, G. J., Zhang. H.. Casso, D.. Kobayashi. R.. and
Beach. D. p21 is a universal inhibitor of cyclin kinases. Nature (Lond.).
366: 70l-7()4. 1993.
3. Waga. S.. Hannon. G. J.. Beach, D., and Stillman, B. The p21
inhibitor of cyclin-dependent kinases controls DNA replication by in-
teraction with PCNA. Nature (Lond.). 369: 574-578, 1994.
4. Luo. Y.. Hurwitz. J.. and Massague. J. Cell-cycle inhibition by
independent CDK and PCNA binding domains in p2lCipl. Nature
(Lond.). 375: 159-161, 1995.
5. El-Deiry. W. S., Tokino. T., Velculescu, V. E.. Levy, D. B., Parsons,R.. Trent. J. M.. Lin. D.. Mercer. W. E., Kinzler, K. W.. and Vogebstein,
B. WAFI. a potential mediator of p53 tumor suppression. Cell. 75:
817-825. 1993.
6. Waldman, T., Kinzler, K. W., and Vogelstein, B. p21 is necessary for
the p53-mediated GI arrest in human cancer cells. Cancer Res., 55.
5187-5190. 1995.
7. Baker, S. J., Preisinger. A. C.. Jessup. J. M.. Paraskeva. C.. Markow-itz, S., Willson, J. K. V.. Hamilton. S.. and Vogebstein. B. p.53 gene
mutations occur in combination with l’7p allelic deletions as late events
in colorectal tumorigenesis. Cancer Ret., 50: 7717-7722, 1990.
8. Yang. Z. Y.. Perkins. N. D.. Ohno. T., Nabel. E. G.. and Nabel, G. J.
The p2 1 cyclin-dependent kinase inhibitor suppresses tumorigenicity in
)‘iVO. Nat. Mcd, 1: 1052-1056, 1995.
9. Chen. Y. Q.. Cipriano. S. C., Arenkiel. J. M., and Miller, F. R. Tumor
suppression by �21wAF) Cancer Res., 55: 4536-4539. 1995.
10. Deng, C.. Zhang, P.. Harper. J. W.. Elledge. S. J., and Leder. P.Mice lacking p2ld1��fl undergo normal development but are defec-tive in G1 checkpoint control. Cell. 82: 675-784, 1995.
1 1. El-Deiry, W. S., Harper, J. W., O’Connor. P. M., Velculescu, V. E.,
Canman, C. E.. Jackman, J., Pietenpol, J. A.. Burrell. M., Hill, D. E..Wang, Y., Wiman, K. G., Mercer, W. E., Kastan, M. B., Kohn, K. W.,Elledge, S. J., Kinzler, K. W., and Vogelstein, B. WAFI/Cipl is induced
in p53-mediated G1 arrest and apoptosis. Cancer Res.. 54: 1169-I 174.1994.
12. Symonds. H.. Krall, L., Remington, L., Saenz-Robles, M., Lowe, S.,Jacks, T., and Van Dyke. T. p53-dependent apoptosis suppresses tumor
growth and progression in vivo. Cell, 78: 703-71 1, 1994.
13. Michieli, P., Chedid, M.. Lin, D.. Pierce, J. H., Mercer, W. E.. and
Givol, D. Induction of WAF1/Cipl by a p53-independent pathway.
Cancer Res., 54: 3391-3395, 1994.
14. Parker. S. B.. Eichele, G.. Zhang. P.. Rawls, A., Sands, A. T.,Bradley, A., Olson, E. N., Harper, J. W., and Elledge, S. J. p53-
independent expression of p21WAFI/Cipl in muscle and other termi-
nally differentiating cells. Science (Washington DC). 267: 1024-1027,
1995.
15. Macleod, K. F., Sherry, N.. Hannon, G., Beach, D., Tokino, T.,
Kinzler, K., Vogelstein. B., and Jacks, T. p53-dependent and independ-
ent expression of p21 during cell growth differentiation, and DNA
damage. Genes Dcv., 9: 935-944. 1995.
16. Zhang. W.. Grasso, L.. McClain, C. D., Gambel, A. M.. Cha, Y.,Travali, S.. Deisseroth, A. B., and Mercer, W. E. p53-independent
induction of WAF1/Cipl in human leukemia cells is correlated with
growth arrest accompanying monocyte/macrophage differentiation.
Cancer Res., 55: 668-675, 1995.
17. Marchetti, A., Doglioni, C.. Barbareschi, M., Buttitta, F., Pellegrini,
S., Bertacca, G., Chella, A.. Merbo. G.. Alberto, Angeletti, C., Dalla
Palma, P., and Bevilacqua. G. p21 RNA and protein expression in
non-small cell lung carcinomas: evidence of p53-independent expres-
sion and association with tumoral differentiation. Oncogene. 12: 13 19-
1324, 1996.
18. Elbendary, A. A.. Cirisano, F. D., Evans, A. C.. Jr., Davis, P. L.,
Iglehart, J. D., Marks, J. R., and Berchuck, A. Relationship between p21
expression and mutation of the p53 tumor suppressor gene in normal and
malignant ovarian epithelial cells. Clin. Cancer Res.. 2: 1571-1575,
1996.
19. El-Deiry, W. S., Tokino, T.. Waldman, T., Oliner, J. D.. Velculescu,V. E., Burrell, M., Hill, D. E., Healy. E., Rees, J. L., Hamilton, S. R.,Kinzler, K. W., and Vogelstein, B. Topological control of �21WAFh/C��I
expression in normal and neoplastic tissues. Cancer Res., 55: 2910-
2919, 1995.
20. Gavrieli. Y.. Sherman, Y.. and Ben-Sasson, S. A. Identification ofprogrammed cell death in situ via specific labeling of nuclear DNA
fragmentation. J. Cell Biol., 119. 493-501, 1992.
21. Lynch, H. T., and Smyrk, T. Hereditary nonpolyposis coborectal
cancer (Lynch syndrome). An updated review. Cancer (Phila.). 78:
1149-1167. 1996.
22. Dietmaier, W., Wallinger, S.. and Bocker, T. Diagnostic microsat-
ellite instability: definition and correlation with mismatch repair protein
expression. Cancer Res. 57: 4749-4756. 1997.
23. Jeon, H. M., Lynch, P. M., Howard, L., Ajani, J., Levin, B., and
Frazier, M. L. Mutation of the hMSH2 gene in two families with
hereditary nonpolyposis coborectal cancer. Hum. Mutat., 7. 327-333,
1996.
24. Thibodeau, S. N.. French, A. J., Roche, P. C.. Cunningham. J. M.,
Tester, D. J., Lindor, N. M.. Moslem, G., Baker, S. M., Liskay, R. M.,Burgart. L. J., Honchel, R., and Hailing, K. C. Altered expression of
hMSH2 and hMLH I in tumors with microsatellite instability and ge-
Research. on October 24, 2020. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
Clinical Cancer Research 1261
netic alterations in mismatch repair genes. Cancer Res, 56: 4835-4840,
1996.
25. Sinicrope, F. A., Ruan, S., Cleary, K. R., Stephens, L. C., Lee, J. J.,
and Levin, B. Bcl-2 and p53 oncoprotein expression during colorectal
tumorigenesis. Cancer Res., 55: 237-241, 1995.
26. Jung, J. M., Bruner, J. B., Ruan, S., Langford, L. A., Athanassios,
P. K., Kobayashi, T., Levin, V., and Zhang, W. Increased levels of
p21W��1�1P1 in human brain tumors. Oncogene, 11: 2021-2028, 1995.
27. Sinicrope, F. A., Roddey, G., McDonnell, T. J., Shen, Y., and
Stephens, L. C. Increased apoptosis accompanies neoplastic develop-
ment in the human coborectum. Clin Cancer Res., 2: 1999-2006, 1996.
28. Schulze-Osthoff, K., Wakzcak, H., Droge. W., and Krammer, P. H.Cell nucleus and DNA fragmentation are not required for apoptosis.
J Cell Biol, 127: 15-20, 1994.
29. Ansani, B., Coates, P. J., Greenstein, B. D., and Hall, P. A. In situ
end-labeling detects DNA strand breaks in apoptosis and other physio-logical and pathological states. J. Pathol., 170: 1-8, 1993.
30. Baas, I. 0., Mulder, J. R., Offerhaus, G. J. A., Vogeistein, B., andHamilton, S. J. An evaluation of six antibodies for immunohistochem-istry of mutant p53 gene product in archival colorectal neoplasms.
J. Pathol., 172: 5-12, 1994.
31. Matsui, S. I., Arrendondo, M. A., Wrzosek, C., and Rustum, Y. M.DNA damage and p53 induction do not cause ZD1694-induced cellcycle arrest in human colon carcinoma cells. Cancer Res, 56: 4715-
4723, 1996.
32. Nagasawa, H., Li, C. Y., Maid, C. G., Imnich, A. C., and Little, J. B.Relationship between radiation-induced G1 phase arrest and p53 func-
tion in human tumor cells. Cancer Res., 55: 1842-1846, 1995.
33. Doglioni, C., Pelosio, P., Laurino, L., Macri, E., Meggiolaro, E.,
Favretti, F., and Barbareschi, M. p2! WAF1/Cipl expression in normalmucosa and in adenomas and adenocarcinomas of the colon: its rela-tionship with differentiation. J. Pathol., 179: 248-253, 1996.
34. Jung, J. M., Li, H., Kobayashi, T., Kyritsis, A. P., Langford. L. A.,Bruner, J. M., Levin, V., and Zhang, W. Inhibition of human glioblas-
toma cell growth by WAF1/Cipl can be attenuated by mutant p53. CellGrowth Differ., 6: 909-91 3, 1995.
35. Konishi, M., Kikuchi-Yanoshita, R., Tanaka, K., Muraoka, M.,
Onda, A., Okumura, Y., Kishi, N., Iwama, T., Mori, T., Koike, M.,Ushio, K., Chiba, M., Nomizu, S., Konishi, F., Utsunomiya, J., andMiyaki, M. Molecular nature of colon tumors in hereditary non-polyp-osis colon cancer, familial polyposis, and sporadic colon cancer. Gas-
troenterology, 111: 307-317, 1996.
36. Losi, L., Fante, R., Di Gregorio, C., Aisoni, M. L., Lanza, G.,Maestni, I., Roncucci, L., Pedroni, M., and Ponz de Leon, M. Biologic
characterization of hereditary non-polyposis colorectal cancer. Nuclearpboidy, AgNOR count, microvessel distribution, oncogene expression,and grade-related parameters. Am. J. Clin. Pathol., 103: 265-270, 1995.
37. Cottu, P. H., Muzeau, F., Estreicher, A., Flejou, J. F., Iggo, R.,Thomas, G., and Hamelin, R. Inverse correlation between RER+ status
and p53 mutation in colorectal cancer cell lines. Oncogene, 13: 2727-
2730, 1996.
38. Caffo, 0., Doglioni, C., Veronese, S., Bozanini, M., Marchetti, A.,
Buttitta, F., Fina, P., Leek, R., Morelli, L., Dalla Palma, P., Harris, A. L.,and Barbareschi, M. Prognostic value of �21WAFI and p53 expression in
breast carcinoma: an immunohistochemical study in 261 patients withlong-term follow-up. Clin. Cancer Res., 2: 1591-1599, 1996.
39. Zhang, W., Kornblau, S. M., Kobayashi, T., Gambel, A., Claxton,D., and Deisseroth, A. B. High level ofconstitutive WAFI/CipI protein
are associated with chemoresistance in acute myebogenous leukemia.
Clin. Cancer Res., 1: 1051-1057, 1995.
40. Li, Y. J., Laurent-Puig, P., Salmon, R. J.. Thomas, G., and Hamelin,R. Polymorphisms and probable lack of mutation in the WAFJ-Cipi
gene colorectal cancer. Oncogene, 10: 599-601, 1995.
41. Koopman, J., Maintz, D., Schild, S., Schramm, J., Lousi, D. N.,Wiestler, 0. D., and von Deimling, A. Multiple polymorphisms. but nomutations, in the WAFI/CIPJ gene in human brain tumors. Br. J.Cancer, 72: 1230-1233, 1995.
42. Gartel, A. L., Serfas, M. S., and Tyner, A. L. p21-negative regulatorof the cell cycle. Proc. Soc. Exp. Biol. Med., 213: 138-149, 1996.
43. Li, C. Y., Suardet, L., and Little, J. B. Potential role of WAFI/
Cipl/p21 as a mediator of TGF-�3 cytoinhibitory effect. J. Biol. Chem.,270: 4971-4974, 1995.
44. Markowitz, S., Wang, J., Myeroff, L., Parsons, R., Sun, L.. Lutter-baugh, J., Fan, R. S., Zborowska, E., Kinzler, K. W., Vogebstein, B.,Brattain, M., and Wilison, J. K. V. Inactivation of the type II TGF-betareceptor in colon cancer cells with microsatellite instability. Science(Washington DC), 268: 1336-1338, 1995.
45. Steinman, R. A. Hoffman, B., ho, A., Guillouf, C., Liebermann,D. A., and El-Houseini, M. E. Induction of p21 (WAF1/Cipl) during
differentiation. Oncogene, 9: 3389-3396, 1994.
46. Barboule, N., Mazars, P., Baldin, V., Vidal, S., Jozan, S., Mantel, P.,and Valette, A. Expression of p21WAF1/Cipl in heterogeneous andunrelated to proliferation index in human ovarian carcinoma. Int. J.
Cancer, 63: 611-615, 1995.
47. Hermking. H., Funk, J. 0., Reichert, M., Ellwart, J. W., and Eick,D.Abrogation of p53-induced cell cycle arrest by c-Myc: evidence foran inhibitor ofp2l WAF1/Cipi/SDI1. Oncogene, 11: 1409-1415, 1995.
48. Lin, D., Fiscella, M., O’Connor, P. M., Jackman, J., Chen, M., Luo,
L. L., Sala, A., Travali, S., Apella, E., and Mercer, W. E. Constitutiveexpression of B-myb can bypass p53-induced WAF1/Cipl-mediated G1arrest. Proc. Natl. Acad. Sci. USA, 91: 10079-10083, 1994.
49. Milross, C. G., Mason, K. A., Hunter, N. R., Chung, W. K., Peters,L. J., and Milas. L. Relationship of mitotic arrest and apoptosis to
antitumor effect of paclitaxel. J. Natl. Cancer Inst., 88: 1308-1314,1996.
50. Meyn, R. E., Stephens, L. C., and Milas, L. Programmed cell death
and radioresistance. Cancer Metastasis Rev., 15: 1 19-131, 1996.
51. Kobayashi, T., Consoli, U., Andreeff, M., Shiku, H., Deisseroth,A. B., and Zhang, W. Activation of p21WAF1/Cipl expression by atemperature-sensitive mutant of human p53 does not lead to apoptosis.
Oncogene, 11: 2311-2316, 1995.
52. Waldman, T., Lengauer. C., Kinzler, K. W., and Vogelstein, B.Uncoupling of S-phase and mitosis induced by anti-cancer agents incells lacking p21. Nature (Lond.), 381: 713-716, 1996.
53. Malkhosyan, S., McCarty, A., Sawai, H., and Perucho, M. Differ-
ences in the spectrum of spontaneous mutations in the hprt gene be-tween tumor cells of the microsatellite mutator phenotype. Mutat. Res.316: 249-259, 1996.
54. McDonald, E. R., Wu, G. S., Waldman, T., and El-Deiry, W. S.
Repair defect in p21WAF1/CIPI -/- human cancer cells. Cancer Res..
56: 2250-2255, 1996.
55. Waldman, T., Zhang. Y., Dillehay, L.. Yu. J., Kinzler, K.,Vogelstein, B., and Williams, J. Cell-cycle arrest versus cell death incancer therapy. Nat. Med.. 3: 1034-1036. 1997.
Research. on October 24, 2020. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from
1998;4:1251-1261. Clin Cancer Res F A Sinicrope, G Roddey, M Lemoine, et al. hereditary nonpolyposis colorectal cancers.progression of sporadic colorectal neoplasms but not Loss of p21WAF1/Cip1 protein expression accompanies
Updated version
http://clincancerres.aacrjournals.org/content/4/5/1251
Access the most recent version of this article at:
E-mail alerts related to this article or journal.Sign up to receive free email-alerts
Subscriptions
Reprints and
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Permissions
Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)
.http://clincancerres.aacrjournals.org/content/4/5/1251To request permission to re-use all or part of this article, use this link
Research. on October 24, 2020. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from