Anatomic Pathology / KRAS Mutation in Colorectal Cancer
KRAS Gene Mutation in Colorectal Cancer Is Correlated
With Increased Proliferation and Spontaneous Apoptosis
Xiuli Liu, MD, PhD,1 Maureen Jakubowski, MT,1 and Jennifer L. Hunt, MD2
Key Words: Colorectal cancer; KRAS; Proliferation; Apoptosis; Microsatellite instability high; Polymerase chain reaction; Cycling sequencing
DOI: 10.1309/AJCP7FO2VAXIVSTP
Abstract
KRAS mutation occurs in 30% to 50% of
colorectal cancers (CRCs) and has been suggested
to be associated with proliferation and decreased
apoptosis. In this study, we analyzed KRAS in 198
CRCs and compared the clinicopathologic variables
between KRAS-mutated and wild-type CRCs. Also, a
subset of 90 and 66 CRCs from this cohort underwent
microsatellite instability testing and histomorphologic
evaluation, and the frequency of microsatellite
instability-high (MSI-H) and histomorphologic
variables were compared between KRAS-mutated and
wild-type CRCs. Clinicopathologic features (age, sex,
and tumor site, depth, size, grade, and metastasis) were
not different between KRAS-mutated and wild-type
CRCs. Compared with wild-type KRAS CRCs, KRAS-
mutated CRCs had a lower frequency of MSI-H (15%
vs 42%; P = .015), a higher chance of having brisk
mitosis (77% vs 43%, P = .022) and apoptosis (77%
vs 28%; P = .00012), and a greater mean of mitotic
figures (P = .0002) and apoptotic cells (P = .0008).
KRAS mutation was associated with higher tumor cell
turnover.
American Society for Clinical Pathology
Colorectal cancer (CRC) is the second leading cause of
cancer-related death in the United States. The development
of CRC is a multistep process characterized by accumulation
of genetic alterations that have long been considered to occur
in a stepwise process. Along the progression from normal
colonic epithelial cells, small adenoma, advanced adenoma,
and finally to carcinoma, the KRAS oncogene mutation has
a role in a significant proportion of CRCs. KRAS has been
reported to be mutated in about 30% of colorectal adenomas
and 30% to 50% of CRCs.1-4
The KRAS gene encodes a 21-kDa small protein that is
activated transiently as a response to extracellular stimuli or
signals such as growth factors, cytokines, and hormones via
cell surface receptors.5,6 On its activation, the KRAS protein
also is capable of turning off the signaling pathway by catalyz-
ing hydrolysis of guanosine triphosphates (GTP) to guanosine
diphosphates. The most common KRAS mutations in codons
12 and 13 are activation mutations, leading to continuous
activation of downstream pathways.5,6 The most frequently
observed types of mutations in KRAS in all human cancers are
G > A transition and G > T transversion. Although the precise
molecular and cellular mechanisms that constitute the oncogen-
ic effects of activating KRAS mutations remain incompletely
understood, in vitro and animal studies showed that KRASV12
regulated genes involve cytokine signaling, cell adhesion, cell
survival, proliferation, apoptosis, and colon development.5-8
Emerging data strongly suggest that KRAS mutations
are predictive of resistance to epidermal growth factor recep-
tor (EGFR) inhibitor treatment in CRCs.5,6 The mechanisms
by which KRAS-mutated CRCs are resistant to EGFR inhibi-
tors are not known, although some investigators have sug-
gested that activation mutation of KRAS would replace the
Am J Clin Pathol 2011;135:245-252 245
245 DOI: 10.1309/AJCP7FO2VAXIVSTP 245
Liu et al / KRAS Mutation in Colorectal Cancer
dependence of CRCs on increased signaling from upstream
EGFR. As an alternative, CRCs KRAS mutations may have
decreased potential to undergo apoptosis and imply a resis-
tance to drugs and agents with therapeutic effects through
apoptosis pathways.
A few studies have also reported that a KRAS mutation
was more commonly seen in well- to moderately differenti-
ated CRCs and in distally located tumors.9,10 Several reports
suggested that KRAS mutation was associated with poor
prognosis and advanced disease.1,3,10,11 The data regarding
the association of KRAS mutation with microsatellite sta-
tus have been mixed.9,12 Two small studies with relatively
detailed morphologic analysis of colorectal adenoma and
adenocarcinoma suggested that KRAS mutation is associated
with diffuse proliferation and decreased apoptosis.13,14 In
CRC, while KRAS mutation was frequently noted in mucinous
carcinoma,11 others reported KRAS mutation was more often
seen in nonmucinous carcinoma regardless of microsatellite
instability (MSI) status.15
In this study, we performed KRAS mutational analysis
on 217 CRCs using formalin-fixed, paraffin-embedded tissue
by polymerase chain reaction (PCR) and cycle sequencing
of the amplified PCR products. A subset of 90 CRCs also
had microsatellite stability testing. A detailed analysis of the
frequency of each mutation within codons 12 and 13 was
performed. Clinicopathologic parameters were obtained, and
a detailed histomorphologic analysis was performed. CRCs
with mutated KRAS and wild-type KRAS were compared
to identify any possible histologic features associated with
mutated KRAS in CRCs.
Materials and Methods
Study Group
We included all 217 cases of CRC with KRAS muta-
tion testing in our molecular diagnostic laboratory in the
Department of Anatomic Pathology, Cleveland Clinic
Foundation, Cleveland, OH, between May 2008 and January
2009. This study was approved by the institutional review
board. Since June 2008 in the Cleveland Clinic Foundation,
all clinical cases of CRC that are stage 3 or 4 have been tested
for KRAS mutations.
PCR Amplification and Sequencing of KRAS
PCR amplification and sequencing of KRAS exon 2,
targeting the mutation hotspots of codons 12 and 13, were
performed as follows. Briefly, slides from each case were
reviewed by a gastrointestinal pathologist (X.L.), and for-
malin-fixed paraffin-embedded tissue blocks containing the
highest density of malignant cells were chosen. For each case,
8 unstained sections were cut and mounted on uncharged
246 Am J Clin Pathol 2011;135:245-252
246 DOI: 10.1309/AJCP7FO2VAXIVSTP
glass slides, and the area with the highest density of malignant
cells was selected for microdissection. The microdissection
targets were confirmed by direct microscopic visualization.
Tumor DNA was extracted from the microdissected frag-
ments. The KRAS gene was examined by PCR for a 263-base-
pair product that includes the most common mutation sites
(codons 12 and 13) using the following primer pairs (forward,
5-GGTGAGTTTGTATTAAAAGGTACTGG; and reverse,
5-TCCTGCACCAGTAATATGCA). PCR amplification was
performed for 49 cycles, including a 15-second denaturation
at 95C, 15-second annealing at 55C, and 15-second exten-
sion at 72C following the initial 3-minute denaturation at
95C. Cycle sequencing was performed for the forward and
reverse strands, using the BigDye Terminator kit (Applied
Biosystems, Foster City, CA) and the same forward and
reverse primers, in separate reactions. The sequence was ana-
lyzed using the ABI 3730 (Applied Biosystems). The forward
and reverse sequences were visually inspected to identify the
presence of point mutations at codon 12 or 13 or other loca-
tions within the PCR product.
Microsatellite Stability Analysis
For the study, 90 CRCs were analyzed for high-level MSI
(MSI-H) according to the criteria for the determination of
MSI-H as previously described.16 Briefly, the microsatellite
stability test was performed using a fluorescent PCR-based
assay by amplifying 7 loci, including 5 mononucleotide
repeats (BAT 25, BAT 26, NR-21, NR-24, and MONO-
27) and 2 pentanucleotide repeats (PentaC and PentaD) on
genomic DNA extracted from CRC tissue and a paired normal
sample (MSI Analysis System, Promega, Madison, WI). The
fluorescently labeled, amplified PCR products from CRC and
normal samples were analyzed by capillary gel electropho-
resis on the ABI 3100 (Applied Biosystems). By comparing
the sizes of the PCR products from CRC and normal samples,
the presence of MSI was determined by the appearance of
new alleles in the CRC sample that were not present in the
corresponding normal sample. The 2 loci of pentanucleotide
repeats serve to confirm identity. CRCs containing 2 or more
loci with MSI were classified as MSI-H.
Histomorphologic Review and Quantitative Assessment
of Mitotic and Apoptotic Activity and Tumor-Infiltrating
Lymphocytes
Tumor-containing slides from a set of 66 randomly
selected, chemoradiation-naive, and surgically resected pri-
mary CRCs (including 40 CRCs with wild-type KRAS and 26
cases with mutated KRAS) used for KRAS mutation analysis
were further reviewed histologically by one of us (X.L.) to
assess the following morphologic features: tumor grade, the
presence of dirty necrosis, the presence of any mucinous dif-
ferentiation, the presence of prominent cribriform architecture
American Society for Clinical Pathology

(defined as easily noticeable under 100 magnification), the
presence of brisk mitotic activity (defined as 3 mitotic fig-
ures/high-power field [HPF]), the presence of brisk apoptotic
activity (defined as apoptotic cells easily identifiable under
200 magnification), and the presence of tumor-infiltrating
lymphocytes (TILs; defined as 2 TILs/HPF).
Further detailed quantitative analysis of mitotic activity,
apoptosis, and TILs was performed in 26 cases of KRAS-
mutated CRCs and 40 cases of CRCs with wild-type KRAS.
More specifically, mitotic figures were counted in every 100
consecutive malignant cells in 10 areas with the most mitotic
activity for each tumor. Apoptotic cells were counted in every
100 consecutive malignant cells in 10 areas with the highest
apoptotic activity. When apoptotic cells were present on the
luminal surface or within the lumen of neoplastic glands, they
were excluded, as were areas of extensive necrosis. Only
typical apoptotic bodies were counted using the standard
morphologic criteria: overall shrinkage and homogeneously
dark basophilic nuclei, the presence of nuclear fragments, a
sharply delineated cell border surrounded by empty space,
and homogeneous eosinophilic cytoplasm. TILs were also
counted in every 100 consecutive malignant cells in 10 areas
with the most TILs. All assessments were performed on the
tumor sections used for KRAS mutation analysis. All sections
were examined by 1 observer to ensure internal consistency of
interpretation. The mitotic activity, apoptotic cells, and TILs
were expressed as mean counts per 100 consecutive malignant
cells obtained from the 10 areas.
Statistical Analysis
The 2 test or Fisher exact test was performed for cate-
gorical data. The Student t test was used to analyze continuous
data. The results for mitotic activity, apoptotic cells, and TILs
were expressed as mean SD. All P values were 2-sided, and
statistical significance was set at a P value of .05 or less.
Results
Clinicopathologic Features of CRCs in the Study
Population
In this study, 217 CRC cases were included, including
109 in-house cases and 108 reference cases from outside
Table 1
Characteristics of Colorectal Cancers With Wild-Type KRAS and Mutated KRAS
Total (n = 198)
Mean SD age at diagnosis (y) 59.8 14.0
No. (%) men 97 (49.0)
* Comparison of cases with at least 1 KRAS mutation with cases without a KRAS mutation.
American Society for Clinical Pathology
Anatomic Pathology / Original Article
facilities. Among them, 198 cases (91.2%) had successful
amplification of the KRAS gene and constituted the study
population. Of the 19 cases with failed amplification, testing
was repeated once in 16, and amplification failed again in 16
of 16; most of these were considered to be due to previous
chemoradiation therapy with resultant scarcity of tumor cells
and inappropriate fixation of the tissue. The mean age of the
patients was 59.8 years, and 49.0% were men Table 1. In
the 113 cases with more detailed clinicopathologic data, 23
tumors were in the rectum, 43 were in the right colon, 27 were
in the left colon, 12 were liver metastases, and 8 were lung or
thoracic metastases. In 95 CRCs for which the depth of inva-
sion was known, 6% (5/88), 6% (5/88), 69% (61/88), and 19%
(17/88) of them invaded submucosa (pT1), muscularis propria
(pT2), subserosa (pT3), and peritoneal or adjacent organ
(pT4), respectively. Approximately 63% (59/93) of CRCs and
25.0% (31/124) had lymph node metastases and distal metas-
tases, respectively, at the time of KRAS mutation testing. The
high proportion of lymph nodepositive and distant metastatic
cases in this study population reflects the practice pattern of
oncologists in which they routinely ordered KRAS mutation
analysis when EGFR inhibitor therapy was considered (stage
3 or 4 tumor).
Mutational Types of KRAS in the Study Population
In this study, among the 198 cases with analyzable KRAS
results, 65 tumors (32.8%) were found to harbor mutated
KRAS gene. As shown in Table 2, the KRAS mutations
were distributed between codon 12 (45/65 [69%]) and codon
13 (20/65 [31%]). The G > A transitions at nucleotides were
the most frequent mutations observed in codons 12 and 13
(61.7%). Among the 65 tumors with mutated KRAS genes,
3 had unique mutation patterns. More specifically, 1 CRC
showed only mutated nucleotide T at the first position in
codon 12 without an accompanying normal allele, 1 contained
only mutated nucleotide A at the second position in codon 13
without an accompanying normal allele, and 1 had mutated
nucleotide T at the first and second positions in codon 12. The
former 2 cases with mutated nucleotides without an accom-
panying normal allele may represent a homozygous mutation
(C12GGT > TGT with amino acid change from Gly to Cys
and C13GGC > GAC with amino acid change from Gly to
Asp) or a single mutation in one allele and loss of heterozygosity
Wild-Type KRAS (n = 133) Mutated KRAS (n = 65) P*
59.4 15.1 60.5 11.7 .61
63 (47.4) 34 (52) .547
Am J Clin Pathol 2011;135:245-252 247
247 DOI: 10.1309/AJCP7FO2VAXIVSTP 247
Liu et al / KRAS Mutation in Colorectal Cancer

Table 2 of the other allele. The third case with 2 mutated nucleotides
Distribution of Various KRAS Mutations in 65 KRAS-Mutated in codon 12 may represent a double mutation (cis position;
Colorectal Cancers
GGT > TTT with amino acid change from Gly to Lys) or 2
Codon/Mutation No. (%) of Cases Amino Acid Change different mutations in two alleles (trans position; GGT > TGT
and GGT > GTT with amino acid changes from Gly to Cys
12 45 (69)
C12GAT 22 (33.8) Gly12Asp and from Gly to Val, respectively).
C12TGT 12 (18.5) Gly12Cys
C12GTT 8 (12.3) Gly12Val
C12CGT 1 (1.5) Gly12Arg
KRAS Mutation Was Not Associated With
C12GCT 1 (1.5) Gly12Ala Clinicopathologic Features of CRCs
Other* 1 (1.5)
13 20 (31) Clinicopathologic features of CRCs with mutated KRAS
C13GAC 18 (27.7) Gly13Asp and wild-type KRAS were compared in a subset of CRCs for
C13GCC 1 (1.5) Gly13Ala
Other 1 (1.5)
which clinicopathologic data were available. As shown in
Total 65 (100) Table 1 and Table 3, the KRAS mutation is not associated
*
with age, sex, tumor location, depth of tumor invasion, lymph
This case has mutated nucleotide T at the first and second positions in codon 12 and
there may be a double mutation in 1 allele (cis with amino acid change from Gly to node metastasis, distant metastasis, tumor grade, tumor size,
Lys) or 2 mutations in both alleles (trans with amino acid change from Gly to Cys
and Gly to Val).
or angiolymphatic invasion.
This case has mutated nucleotide A at the third position in codon 13 that represents

a silent mutation.
KRAS Mutation Was Less Frequently Seen in MSI-H
CRCs
Table 3
Correlation Between KRAS Mutation and Clinicopathologic Of the 198 CRC cases, 90 tumors (45.5%) were tested
Factors of Colorectal Carcinomas* for microsatellite stability as well. Among these 90 CRCs
tested for microsatellite status, 59 were microsatellite stable
Total No. Wild-Type KRAS
Factor of Cases KRAS Mutated P (66%) and 31 CRCs were MSI-H (34%). As shown in Table
4, only 13% (4/31) CRCs with mutated KRAS were MSI-H,
Tumor site 113 80 33 .201 in contrast with 37% (22/59) CRCs with wild-type KRAS (P
Rectum 23 20 (87) 3 (13)
Left 27 18 (67) 9 (33) = .015).
Right 43 27 (63) 16 (37)
Others 20 15 (75) 5 (25)
Depth of invasion 88 63 25 .495
More KRAS-Mutated CRCs Demonstrated Brisk Mitotic
T1 5 4 (80) 1 (20) and Apoptotic Activity
T2 5 5 (100) 0 (0)
T3 61 42 (69) 19 (31) Although KRAS activation mutation has been shown to
T4 17 12 (71) 5 (29) be associated with increased proliferation in cancer cell lines,
Lymph node involvement 83 58 25 .517
No 24 18 (75) 6 (25)
its exact effect in human cancer is relatively unclear. In a few
Yes 59 40 (68) 19 (32) studies, KRAS mutation has been shown to be associated with
Distant metastasis 124 84 40 .375 a diffuse proliferation pattern, polypoid growth, and high
Yes 31 19 (61) 12 (39)
No 93 65 (70) 28 (30) cytologic grade in colorectal adenomas and early cancers.14 In
Tumor grade 93 68 25 .302 a small study, a KRAS mutation was suggested to be associ-
Well-moderate 63 44 (70) 19 (30)
Poor 30 24 (80) 6 (20) ated with decreased apoptosis.13
Tumor size (cm) 85 62 23 .963 Because there is lack of detailed histomorphologic analy-
>5 41 30 (73) 11 (27)
<5 44 32 (73) 12 (27) sis of CRCs with mutated KRAS in the literature, in this study,
Angiolymphatic invasion 89 64 25 .827 40 CRCs with wild-type KRAS and 26 tumors with mutated
Present 55 40 (73) 15 (27)
Absent 34 24 (71) 10 (29)
KRAS were randomly chosen for detailed histomorpho-
logic review by a gastrointestinal pathologist (X.L.) who was
* Data are given as number (percentage).
blinded to the KRAS mutation results. These 66 tumors were
chemoradiation-naive, surgically resected, primary CRCs.
Table 4 Tumors were given a single grade of differentiation (well,
Correlation of KRAS Mutation With Microsatellite Stability in moderate, or poor) based on the criteria of Jass et al17 and
90 Colorectal Carcinomas*
Greenson et al18 with minor modification. The worst grade
Total No. of tumor seen was used for the overall grade, unless the
of Cases Wild-Type Mutated
(n = 90) KRAS KRAS P worst area was smaller than 10% of the tumor volume and
at the advancing margin of the tumor. In addition, morpho-
Microsatellite stable 59 37 (63) 22 (37)
Microsatellite instability-high 31 27 (87) 4 (13) .015
logic features assessed included the presence of dirty necrosis
(>10% of tumor volume, infarcted tumor areas not included),

248 Am J Clin Pathol 2011;135:245-252 American Society for Clinical Pathology

248 DOI: 10.1309/AJCP7FO2VAXIVSTP
 Anatomic Pathology / Original Article

mucinous differentiation (any), cribriform architecture of Table 5

glands (easily noticeable at 100 magnification), the pres- Histomorphologic Features of KRAS-Mutated and Wild-Type
ence of brisk mitotic activity (>3 mitotic figures/HPF), and Colorectal Carcinomas*
the presence of brisk apoptotic activity (easily identifiable at Wild-Type KRAS
200 magnification). Tumors were also evaluated for TILs as KRAS Mutated
Feature (n = 40) (n = 26) P
previously described,18 and the tumor was considered to be
positive for the presence of TILs if more than 2 lymphocytes Presence of dirty necrosis .364
Yes 11 (28) 9 (35)
per HPF were identified. No 29 (73) 17 (65)
As shown in Table 5, the frequency of the presence Tumor differentiation .252
Well-moderate 29 (73) 16 (62)
of dirty necrosis, the presence of poor differentiation, the Poor 11 (28) 10 (38)
presence of mucinous differentiation, the presence of TILs, Any mucinous differentiation .281
Yes 20 (50) 9 (35)
and the presence of cribriform architecture were not differ- No 20 (50) 17 (65)
ent between the 40 wild-type KRAS-containing and the 26 Prominent cribriform architecture .407
Yes 30 (75) 21 (81)
KRAS-mutated CRCs (presence of dirty necrosis, 11 [28%] No 10 (25) 5 (19)
vs 9 [35%], P = .364; poor tumor differentiation, 11 [28%] Presence of brisk mitotic activity .006
vs 10 [38%], P = .252; mucinous differentiation, 20 [50%] Yes 17 (43) 20 (77)
No 23 (57) 6 (23)
vs 9 [35%], P = .281; cribriform architecture, 30 [75%] vs Presence of brisk apoptosis .00012
21 [81%], P = .407; and presence of TILs, 10 [25%] vs 8 Yes 11 (28) 20 (77)
No 29 (73) 6 (23)
[31%], P = .122). However, 77% (20/26) of KRAS-mutated Presence of TILs .61
CRCs had brisk mitotic activity compared with 43% (17/40) Yes 10 (25) 8 (31)
No 30 (75) 18 (69)
of CRCs with wild-type KRAS (P = .006). In addition, 77%
of the KRAS-mutated CRCs (20/26) demonstrated brisk apop- TILs, tumor-infiltrating lymphocytes.
* Data are given as number (percentage).
totic activity compared with only 28% of CRCs (11/40) with
wild-type KRAS (P = .00012).

KRAS-Mutated CRCs Were Mitotically and

Apoptotically Active by Quantitative Assessment
To confirm our qualitative results, a quantitative assess-
ment of mitotic activity, apoptotic cells, and TILs were
performed in that subset of CRCs (40 CRCs with wild-type
KRAS and 26 cases with mutated KRAS) using the methods as
described in the Materials and Methods section. As shown
in Figure 1, the quantitative assessment revealed results sim-
ilar to those obtained by the overall assessment method. The
mean SD numbers of mitotic figures (in 100 consecutive
malignant cells) were 3.15 1.05 (n = 26) in KRAS-mutated
CRCs compared with 1.83 1.47 (n = 40) in CRCs with wild-
type KRAS (P = .0002). The mean SD number of apoptotic
cells (in 100 consecutive malignant cells) were 10.9 4.05 (n
= 26) in KRAS-mutated CRCs compared with 7.15 4.34 (n
= 40) in CRCs with wild-type KRAS (P = .0008). The number
of TILs (in 100 consecutive malignant cells) in KRAS-mutated
and KRAS wild-type CRCs were not significantly different
(2.03 2.95 vs 3.61 6.01, n = 26 and n = 40, respectively,
for KRAS-mutated and wild-type CRCs, P = .22).
Discussion
KRAS is an oncogene that encodes a 21-kDa small pro-
tein with GTPase activity.5,6 Oncogenic KRAS alleles that
harbor activation mutations produce a protein product with
enzymatically impaired GTPase function or that is refractory
to GTPase activating proteins. This leads to elevated steady-
state levels of RAS-GTP and causes prolonged effector path-
way stimulation.5,6 Previous intensive exploration in the field
has demonstrated that the RAS-mediated pathway regulates
different cellular responses such as proliferation, transfor-
mation, differentiation, senescence, and apoptosis.5,6,19,20
Although KRAS mutation has been described in a significant
number of human CRCs, detection of the mutation was not
known to have clinical relevance and was largely performed
in research and clinical trial settings. Emerging data strongly
suggest, however, that KRAS mutations in CRCs have a
predictive role in determining resistance to EGFR inhibitor
(cetuximab and panitumumab) treatment in patients in whom
previous conventional chemotherapy regimens failed.21,22
After the initial announcements of these results, KRAS muta-
tional analysis quickly became an important molecular test
for patients with CRC who were being considered for EGFR
inhibitor treatment. This dramatic shift in practice can be seen
in our study in which 217 requests for KRAS mutation analy-
ses have been requested for patients with CRC during the past
9-month period. Currently, only a handful of molecular diag-
nostic laboratories are offering KRAS mutational analysis, but
more laboratories are expected to validate this assay.
In our study population, the mutational rate for KRAS
at codons 12 and 13 was 32.8%, which was similar to the

American Society for Clinical Pathology Am J Clin Pathol 2011;135:245-252 249

249 DOI: 10.1309/AJCP7FO2VAXIVSTP 249
Liu et al / KRAS Mutation in Colorectal Cancer

A B

Consecutive Malignant Cells

12.0 10.9

Apoptotic Cells in 100

Mitotic Figures in 100
Consecutive Cells

3.15 10.0
3.5
3.0 8.0 7.15
2.5 1.83
2.0 6.0
1.5
1.0 4.0
0.5
0 2.0
Mutated KRAS Wild-Type KRAS
0
Mutated KRAS Wild-Type KRAS

C
Figure 1 Correlation of KRAS mutation with mitotic activity,
spontaneous apoptosis, and tumor-infiltrating lymphocytes
(TILs) in colorectal carcinomas (CRCs). A, Mitotic figures in
Consecutive Malignant Cells

KRAS-mutated and KRAS wild-type CRCs. B, Apoptotic cells

4.0 3.61 in KRAS mutated and KRAS wild-type CRCs. C, TILs in KRAS-
3.5
TILs in 100

3.0 mutated CRCs and KRAS wild-type CRCs. Mitotic figures,

2.5 2.03 apoptotic cells, and TILs were counted in 100 consecutive
2.0
1.5 malignant cells in 10 fields as described in the Materials and
1.0 Methods section in 26 cases of KRAS-mutated CRCs and 40
0.5
0
cases of KRAS wild-type CRCs. The results are expressed as
Mutated KRAS Wild-Type KRAS means, as indicated by the numbers on top of the bars.

previously reported mutational rates of 30% and 37% in 3
large studies.1,3,23 The mutation at codon 12 accounted for
approximately 69% of cases. A G > A transition was the most
frequent mutation (61.7% of the total mutations detected).
These findings are in line with previously published results.10
In our study, we observed a total of 20 mutations at codon 13,
and 1 of them was a silent mutation (1 case). It is interesting
that 2 CRCs (3%) in our study showed a single mutant peak
on the sequencing electrophoretogram, with no correspond-
ing normal allele. One case showed the mutation at the first
position in codon 12, and the other was in the second posi-
tion in codon 13. The possible explanation is a homozygous
mutation with both alleles harboring identical mutations or a
single mutant allele with corresponding loss of heterozygos-
ity for the second allele. A third case had mutated nucleotides
T at the first and second positions of codon 12, which could
represent a double mutation (cis, GGT > TTT in 1 allele with
amino acid Gly > Lys) or 2 independent mutations, 1 on
each allele (trans, GGT > TGT and GGT > GTT with amino
acids Gly > Cys and Gly > Val). Owing to the small num-
ber of cases, the biologic significance of these homozygous
mutations, double mutation, or single mutation with loss of
heterozygosity remains unclear. However, all 3 cases showed
poorly differentiated morphologic features (data not shown),
and histomorphologic analysis of the tumor with double muta-
tion (cis) or 2 independent mutations (trans) revealed that this
tumor was mitotically and apoptotically active.
Our study did not identify any association of KRAS muta-
tion with the following clinicopathologic variables: age, sex,
tumor site, depth of tumor invasion, histologic grade, distant
metastasis, vascular invasion, and lymphocytic response.
Although owing to the relative lack of earlier stage cases the
study population may be biased to allow meaningful cor-
relation of KRAS mutation with clinicopathologic variables
including nodal status and distant metastases, our current find-
ings were in line with a previous large study of 3,439 cases of
CRC collected in multiple centers and many countries.1
We performed a detailed histomorphologic analysis for
KRAS-mutated CRCs to identify possible histologic features
associated with mutated KRAS in CRCs. For mitotic activ-
ity and apoptosis, we did not perform ancillary tests such as
immunohistochemical analysis for Ki-67 and the terminal
deoxynucleotidyl transferase (TdT)-mediated deoxyuridine
triphosphate (dUTP)-biotin end-labeling assay because we
wanted to use a simple method that would be applicable in
routine surgical pathology practice examining H&E-stained
sections. Our results showed that KRAS-mutated CRCs had
increased mitotic and apoptotic activity compared with wild-
type KRAS tumors. The higher mitotic activity in mutated
KRAS CRCs observed in our study is consistent with results
previously reported.8,11,14 Brisk apoptotic activity in mutated
KRAS CRCs was also observed in our study, consistent with
a previous report of an association of KRAS G > A transi-
tions with increased apoptosis,24 but contrasting with the

250 Am J Clin Pathol 2011;135:245-252 American Society for Clinical Pathology

250 DOI: 10.1309/AJCP7FO2VAXIVSTP

findings reported by Ward et al,13 who found no association.
The discrepancy may be due to several factors: The study by
Ward et al13 was relatively small, used a different method to
measure apoptotic activity (TdT-mediated dUTP-biotin end-
labeling), and included more tumors in stages 1 and 2, and the
KRAS mutation analysis only detected codon 12 mutations.
However, the clinicians KRAS test-ordering pattern in which
KRAS tests were ordered only for patients with stage 3 or 4
CRCs during the study period in our study resulted in a rela-
tive lack of earlier stage cases in this retrospective study and
could have led to bias.
The exact mechanism by which an activating KRAS
mutation leads to relatively brisk spontaneous apoptosis in
CRCs remains unclear. It was initially thought that there was
a progressive inhibition of apoptosis during colorectal tum-
origenesis,25 and KRAS mutation appeared to further inhibit
apoptosis in CRCs because CRCs from compound KRASV12G/
APC+/1638N mice showed reduced apoptosis relative to tumors
arising from APC+/1638N transgenic mice.26 However, emerg-
ing data have suggested that KRAS mutation (KRAS13D)
promotes the induction of apoptosis in ReovirusT3D-infected
or oxaliplatin-treated tumor cells by sensitizing these cells to
activation of the intrinsic apoptosis cascade27 and increasing
sensitivity of colon cancer cells to 5-fluorouracil-induced
apoptosis.28 Additional reports have also showed that onco-
genic KRAS mutations mediated apoptosis through NORE1,
RASSF1, and other RASSF proteins5,29 and that KRAS
mutations induced apoptosis via a p53-independent pathway
when NF-B activation was inhibited.30 Further in vitro
experiments in a lung cell line showed that mutant KRASV12
increased COX-2, peroxides, and DNA damage, which led
to activation of an intrinsic apoptotic cascade.31 However,
the intermediate steps between oncogenic KRAS mutation,
COX-2 activation/up-regulation, DNA damage, NORE1/
RASSF1 and other RASSF proteins, and final apoptosis of
cancer cells are largely unexplored in CRCs. Further studies
in these fields will yield more mechanistic insights into KRAS
mutation-mediated spontaneous apoptosis in CRC.
We also assessed KRAS in relation to MSI and found
that KRAS mutations were less frequently seen in CRCs with
MSI-H. This finding is consistent with previously reported
results.12 Most of the MSI-H CRCs in our study had loss
of hMLH1 protein (but no other mismatch repair [MMR]
proteins) shown by immunohistochemical analysis and, there-
fore, were most likely sporadic tumors (data not shown). Our
results, along with results reported by Oliveira et al12 strongly
suggest that sporadic MSI-H CRCs have a lower frequency of
KRAS mutation compared with sporadic microsatellite-stable
CRCs and MSI-H CRCs arising in the setting of hereditary
nonpolyposis CRC. Furthermore, in a small study, Zhao et
al32 reported an association of sequence alterations in hMLH1
with sporadic KRAS mutations in sporadic MSI-H CRCs.
American Society for Clinical Pathology
Anatomic Pathology / Original Article
Therefore, the emerging data appear to support that muta-
tions in MMR genes are associated with KRAS mutation in
hereditary nonpolyposis CRC and sporadic MSI-H CRCs. It is
interesting that of 31 MSI-H CRCs in the current study, 4 had
KRAS mutation and 3 of them (75%) were C13GGC > GAC
with amino acid change from Gly to Asp, therefore suggest-
ing a possible association of C13GGC > GAC with MSI-H
in CRCs because the frequency of C13GGC > GAC in the
current entire cohort was only 28% (18/65). However, large
studies are needed to confirm this possible association. The
interplay of promoter hypermethylation of hMLH1 and other
genes, MMR gene mutation, and KRAS mutation (includ-
ing mutational location and type) is complex, and detailed
analysis of larger studies will be essential to enhance our
understanding of early predominant genetic alterations during
tumorigenesis of sporadic CRCs.
Our study showed an excellent success rate of KRAS
mutation analysis in routinely processed clinical surgical
pathology material using the traditional cycle sequencing
reaction. Approximately 30% of CRCs had KRAS mutation
at codons 12 and 13. KRAS mutation was less frequently seen
in MSI-H CRCs. Furthermore, our study showed that KRAS
mutation in CRCs was associated with higher mitotic activity
and brisker spontaneous apoptosis. Additional work in larger
series is needed to determine whether these histologic param-
eters may be a screening tool for triaging cases for up-front
KRAS mutational testing. It will also be important to deter-
mine whether these parameters have biologic significance in
the background of KRAS mutations and MSI in larger series
including early- and late-stage CRCs.
From the Departments of 1Anatomic Pathology, Cleveland Clinic
Foundation, Cleveland, OH; and 2Pathology, Massachusetts
General Hospital, Boston.
Address reprint requests to Dr Hunt: Dept of Pathology,
Massachusetts General Hospital, Warren 202, 55 Fruit St, Boston,
MA 02114.
References
1. Russo A, Bazan V, Agnese V, et al. Prognostic and
predictive factors in colorectal cancer: Kirsten Ras in CRC
(RASCAL) and TP53CRC collaborative studies. Ann Oncol.
2005;16(suppl 4):iv44-iv49.
2. Nosho K, Irahara N, Shima K, et al. Comprehensive
biostatistical analysis of CpG island methylator phenotype in
colorectal cancer using a large population-based sample. PLoS
One. 2008;3:e3698. doi:10.1371/journal.pone.0003698.
3. Suehiro Y, Wong CW, Chirieac LR, et al. Epigenetic-genetic
interactions in the APC/WNT, RAS/RAF, and P53 pathways
in colorectal carcinoma. Clin Cancer Res. 2008;14:2560-2569.
4. Castagnola P, Giaretti W. Mutant KRAS, chromosomal
instability and prognosis in colorectal cancer. Biochim Biophys
Acta. 2005;1756:115-125.
Am J Clin Pathol 2011;135:245-252 251
251 DOI: 10.1309/AJCP7FO2VAXIVSTP 251
Liu et al / KRAS Mutation in Colorectal Cancer
5. Malumbres M, Barbacid M. RAS oncogenes: the first 30 years.
Nat Rev. 2003;3:7-13.
6. Downward J. Targeting RAS signalling pathways in cancer
therapy. Nat Rev. 2003;3:11-22.
7. Roberts ML, Drosopoulos KG, Vasileiou I, et al. Microarray
analysis of the differential transformation mediated by
Kirsten and Harvey Ras oncogenes in a human colorectal
adenocarcinoma cell line. Int J Cancer. 2006;118:616-627.
8. Haigis KM, Kendall KR, Wang Y, et al. Differential
effects of oncogenic K-Ras and N-Ras on proliferation,
differentiation and tumor progression in the colon. Nat
Genet. 2008;40:600-608.
9. Kawabata Y, Tomita N, Monden T, et al. Molecular
characteristics of poorly differentiated adenocarcinoma
and signet-ring-cell carcinoma of colorectum. Int J Cancer.
1999;84:33-38.
10. Poehlmann A, Kuester D, Meyer F, et al. K-ras mutation
detection in colorectal cancer using the pyrosequencing
technique. Pathol Res Pract. 2007;203:489-497.
11. Bazan V, Migliavacca M, Zanna I, et al. Specific codon
13 K-ras mutations are predictive of clinical outcome in
colorectal cancer patients, whereas codon 12 K-ras mutations
are associated with mucinous histotype. Ann Oncol.
2002;13:1438-1446.
12. Oliveira C, Westra JL, Arango D, et al. Distinct patterns
of KRAS mutations in colorectal carcinomas according to
germline mismatch repair defects and hMLH1 methylation
status. Hum Mol Genet. 2004;13:2303-2311.
13. Ward RL, Todd AV, Santiago F, et al. Activation of the
K-ras oncogene in colorectal neoplasms is associated with
decreased apoptosis. Cancer. 1997;79:1106-1113.
14. Kobayashi M, Watanabe H, Ajioka Y, et al. Effect of K-ras
mutation on morphogenesis of colorectal adenomas and early
cancers: relationship to distribution of proliferating cells.
Hum Pathol. 1996;27:1042-1049.
15. Song GA, Deng G, Bell I, et al. Mucinous carcinomas of the
colorectum have distinct molecular genetic characteristics.
Int J Oncol. 2005;26:745-750.
16. Bacher JW, Flanagan LA, Smalley RL, et al. Development
of a fluorescent multiplex assay for detection of MSI-high
tumors. Dis Markers. 2004;20:237-250.
17. Jass JR, Atkin WS, Cuzick J, et al. The grading of rectal
cancer: historical perspectives and a multivariate analysis
of 447 cases. Histopathology. 1986;10:437-459.
18. Greenson JK, Bonner JD, Ben-Yzhak O, et al. Phenotype
of microsatellite unstable colorectal carcinoma: well-
differentiated and focally mucinous tumors and the absence
of dirty necrosis correlate with microsatellite instability. Am J
Surg Pathol. 2003;27:563-570.
19. Bourne HR, Sanders DA, McCormick F. The GTPase
superfamily: a conserved switch for diverse cell functions.
Nature. 1990;348:125-132.
252 Am J Clin Pathol 2011;135:245-252
252 DOI: 10.1309/AJCP7FO2VAXIVSTP
20. Campbell SL, Khosravi-Far R, Rossman KL, et al. Increasing
complexity of Ras signaling. Oncogene. 1998;17:1395-1413.
21. De Roock W, Piessevaux H, De Schutter J, et al. KRAS
wild-type state predicts survival and is associated to early
radiological response in metastatic colorectal cancer treated
with cetuximab. Ann Oncol. 2007;19:508-515.
22. Livre A, Bachet J-B, Boige V, et al. KRAS mutations as an
independent prognostic factor in patients with advanced
colorectal cancer treated with cetuximab. J Clin Oncol.
2008;26:374-379.
23. Brink M, de Goeij AFPM, Weijenberg MP, et al. K-ras
oncogene mutations in sporadic colorectal cancer in the
Netherlands Cohort Study. Carcinogenesis. 2003;24:703-710.
24. Risio M, Malacarne D, Giaretti W. KRAS transitions
and villous growth in colorectal adenomas. Cell Oncol.
2005;27:363-366.
25. Bedi A, Pasricha PJ, Akhtar AJ, et al. Inhibition of apoptosis
during development of colorectal cancer. Cancer Res.
1995;55:1811-1816.
26. Janssen K-P, Alberici P, Fsihi H, et al. APC and oncogenic
KRAS are synergistic in enhancing Wnt signaling in
intestinal tumor formation and progression. Gastroenterology.
2006;131:1096-1109.
27. Smakman N, van den Wollenberg DJM, Elias SG, et al.
KRASD13 promotes apoptosis of human colorectal tumor
cells by ReovirusT3D and oxaliplatin but not by tumor
necrosis factorrelated apoptosis-inducing ligand. Cancer Res.
2006;66:5403-5408.
28. Klampfer L, Swaby L-A, Huang J, et al. Oncogenic Ras
increases sensitivity of colon cancer cells to 5-FU-induced
apoptosis. Oncogene. 2005;24:3932-3941.
29. Ikeda M, Hirabayashi S, Fujiwara N, et al. Ras-association
domain family protein 6 induces apoptosis via both caspase-
dependent and caspase-independent pathways. Exp Cell Res.
2007;313:1484-1495.
30. Mayo MW, Wang C-Y, Cogswell PC, et al. Requirement of
NF-kappaB activation to suppress p53-independent apoptosis
induced by oncogenic Ras. Science. 1997;278:1812-1815.
31. Maciag A, Sithanandam G, Anderson LM. Mutant K-rasV12
increases COX-2, peroxides and DNA damage in lung cells.
Carcinogenesis. 2004;25:2231-2237.
32. Zhao Y, Miyashita K, Ando T, et al. Exclusive KRAS
mutation in microsatellite-unstable human colorectal
carcinomas with sequence alterations in the DNA mismatch
repair gene, MLH1. Gene. 2008;423:188-193.
American Society for Clinical Pathology