The cyclooxygenase (COX) pathway is important in colorectal carcinogenesis with the majority of cancers overexpressing COX-2; however, the role of COX-2 in the development of colorectal adenomas is less well defined. Accordingly, we analyzed 108 colorectal adenomas for COX-1 and COX-2 transcription in archival formalin-fixed, paraffin-embedded tissue using by real-time PCR and normalized to β-actin. Neither COX-1 nor COX-2 mRNA expression differed with regard to age or gender of the subject. COX-2 mRNA expression was significantly higher in distal adenomas (2.2 ± 1.9) compared with proximal (0.7 ± 0.5) adenomas (P < 0.0001) and in larger (≥7 mm) compared with smaller (<7 mm) adenomas (2.3 ± 2.2 and 1.7 ± 1.3, respectively, P = 0.04). COX-2 expression did not differ significantly in tubular compared with tubulovillous adenomas, although there appeared to be a trend toward higher COX-2 expression in tubulovillous adenomas with increasing villous content. Additionally, there was not a significant difference in either COX-1 or COX-2 based on the degree of dysplasia Therefore, if COX-2 inhibitors work through a COX-2 mechanism, these agents may have differential effects on colorectal adenomas that are distal and larger.

CRC3 is the second leading cause of cancer mortality in the United States (1). CRC is generally thought to evolve through a multistep carcinogenesis process that includes the CRA as an intermediate step in the process. Although removal of CRAs is the most effective method for reducing the risk of CRC (2), factors associated with the progression to CRC or with recurrence of CRAs are not fully understood, and it is important to determine the molecular pathways or events involved in this progress (3).

COX is the rate-limiting enzyme in the metabolism of arachidonic acid to eicosinoids, including PGs. COX-1 is constitutively expressed in many tissues, whereas the expression of COX-2 is inducible. Studies have demonstrated that COX-2 is elevated in CRAs and CRC as compared with normal mucosa with no apparent change in COX-1(4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14). PGs are important signal transduction molecules with many functions, and PGE2 is elevated in colonic tumors of carcinogen-treated rats in relation to the surrounding mucosa (15). The methodology used for measurement of COX expression in studies has varied tremendously with use of a number of different antibodies for immunohistochemistry as well as in situ hybridization (4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14); however, the availability of real-time PCR technology combined with newly developed methods for extraction of RNA from formalin-fixed, paraffin-embedded specimens now permits quantitative and accurate measurement of COX-1 and COX-2 gene expression in CRAs. In this study, we determined the feasibility of measuring COX-1 and COX-2 gene expression in paraffin-embedded archival CRAs using real-time PCR. We then determined the relationship between adenoma characteristics and COX-1 and COX-2 gene expression.

COX-1 and COX-2 analyses were performed on baseline sporadic CRAs from a completed Phase III WBF Chemoprevention Trial in 1429 participants (16). Informed consent and archival blocks were obtained as part of the WBF Trial with ∼80% of archival blocks retrieved. One hundred eighty-six CRAs were randomly selected by the Biometry Section, and 108 of these had tissue sufficient for analyses. A 5-μm section was stained with H&E and evaluated by a gastrointestinal pathologist to determine the area of adenoma to be microdissected. Adenomatous tissue from three additional 10-μm sections were pooled for RNA isolation according to a proprietary procedure of Response Genetics, Inc. (U.S. patent no. 6,248,535). After RNA isolation, cDNA was prepared from each sample as described previously (17).

Relative gene expression quantitation for COX-1 and COX-2 (with β-actin as an internal reference gene) was performed using a fluorescence-based, real-time detection method (ABI PRISM 7700 Sequence Detection system, TaqMan; Applied Biosystems, Foster City, CA), as described previously (18, 19). The primers and probe sequences used are given below. In each case, the first primer is the forward primer, the second is the reverse primer, and the third is the Taqman probe: COX-1; 5′-ATGATGGGCCTGCTGTTGA-3′; 5′-CTCACCATGCCAAACCAGAA-3′; 6FAM 5′-CTGGCCTCAGCACTCTGGAATGACAA-3′TAMRA, COX-2; 5′-GCTCAAACATGATGTTTGCATTC-3′; 5′-GCTGGCCCTCGCTTATGA-3′; 6FAM5′-TGCCCAGCACTTCACGCATCAGTT-3′TAMRA β-actin; 5′-TGAGCGCGGCTACAGCTT-3′; 5′-TCCTTAATGTCACGCACGATTT-3′; 6FAM5′-ACCACCACG-GCCGAGCG G-3′TAMRA. The PCR reaction mixture consisted of 600 nm of each primer, 200 nm probe, 2.5 units of AmpliTaq Gold Polymerase, 200 μm each of dATP, dCTP, and dGTP, 400 μm dUTP, 5.5 mm MgCl2, and 1 μl of TaqMan Buffer A with reference dye (all reagents from Applied Biosystems) and cDNA or calibrator material to a final volume of 20 μl (all reagents from Applied Biosystems). Cycling conditions were 50°C for 10 s and 95°C for 10 min, followed by 46 cycles at 95°C for 15 s and 60°C for 1 min. Significant contamination with genomic DNA was excluded by amplifying nonreverse-transcribed RNA. Colon, liver, and lung RNAs (all from Stratagene, La Jolla, CA) were used as control calibrators on each plate. All gene expression analyses were performed in a blinded fashion with the laboratory investigators unaware of the clinical data. The expression of individual COX-1 and COX-2 measurements was calculated relative to expression of β-actin using a modification of the method described by Lehmann et al.(17). After relative gene expression levels for COX-1 and COX-2 values are established for each sample, these values of COX-1 and COX-2 were used for statistical analyses.

Data on baseline adenoma characteristics (size, location, histology, and degree of dysplasia) were obtained from pathology reports and reviewed by the study pathologist (A. K. B.) as described previously (16). COX-1 and COX-2 levels were normalized relative to β-actin and log transformed to improve normality. Because of the presence of 0 values in the COX-1 data, a 1 was added to all COX-1 values before log transformation. Differences in log-transformed values among CRA characteristics were compared using t tests or ANOVAs. P < 0.05 was considered statistically significant. Stata 7.0 was used for all statistical analyses (Stata Corporation, College Station, TX).

Archival blocks from baseline CRAs were collected as part of the WBF Phase III Chemoprevention Trial (16). One hundred eight sporadic CRAs were analyzed for COX-1 and COX-2 mRNA levels in archival paraffin-embedded tissue by real-time PCR and normalized relative to β-actin. Forty-four (41%) females and 64 (59%) males with a median age of 70 years (range, 48–80 years) were included in this analysis.

Neither COX-1 nor COX-2 expression differed with regard to age or gender. Mean COX-2 was significantly higher in adenomas located in the distal colon (including the rectum and splenic flexure [2.2 ± 1.9]) compared with those proximal to the splenic flexure (0.7 ± 0.5, P < 0.0001). The difference in mean COX-1, by adenoma location, did not reach statistical significance (1.0 ± 0.6 and 0.7 ± 0.4 for distal and proximal adenomas, respectively, P = 0.07). Mean COX-1 and COX-2 values [±SE (SE)] by location within the colorectum are shown in Fig. 1. Furthermore, when rectal (23 of 88) adenomas were considered separately from other distal adenomas, COX-2 results were similar (2.2 ± 2.0 for distal CRAs, excluding rectal and 2.3 ± 1.4 for rectal adenomas).

COX-2 expression was significantly higher in larger adenomas when the median size was used as a cutoff (≥7 mm) versus smaller (<7 mm) adenomas (P = 0.04), whereas COX-1 was not significantly different. Fig. 2 demonstrates the relationship of mean COX-1 and COX-2 expression and adenoma size.

When COX-2 was evaluated by both location and size, there was a significant difference in smaller distal adenomas compared with larger (≥7 mm) distal adenomas (Table 1a). In contrast, there were no significant differences in COX-2 expression for proximal adenomas and size. Furthermore, COX-1 did not differ significantly when location and size were combined (Table 1b).

COX-2 did not differ significantly in tubular compared with tubulovillous adenomas. No pure villous adenomas were found in this subset, but there appears to be a trend toward an increase of COX-2 in tubulovillous adenomas with increasing villous content (Table 2). COX-1 was significantly decreased in tubular compared with tubulovillous adenomas and with a higher villous content (P = 0.02, Table 2). When histology and size were combined, COX-2 was significantly increased in larger (≥7 mm) tubulovillous adenomas (P = 0.015, Table 1a), whereas no association was found for COX-1 (Table 1b).

Similarly, there was not a significant difference in either COX-1 or COX-2 based on the degree of dysplasia (mild to moderate versus high) as shown in Table 2. When a high degree of dysplasia and adenoma size are combined (Table 1a), there was a borderline significant increase in COX-2 in larger (≥7 mm) adenomas with high-grade dysplasia (P = 0.5). No relationship was seen for COX-1 (Table 1b).

We demonstrated that COX-2 mRNA expression was significantly higher in distal and in larger (≥7 mm) adenomas. In contrast, COX-1 mRNA expression was significantly reduced when villous histology was present. We also found that in distal, but not proximal adenomas, COX-2 was significantly higher in larger adenomas. When histology, degree of dysplasia, and percentage of villous content were evaluated by size, we found that COX-2 was significantly higher in adenomas with tubulovillous histology, high-grade dysplasia, or >25% villous content if adenomas were ≥7 mm. This relationship was not found for COX-1.

Most studies have determined that COX-2 expression (mRNA and/or protein) is increased in CRAs in comparison with normal mucosa (6, 13). There are little data that address COX-2 expression and adenoma location. Chapple et al.(20) found increased superficial interstitial cell COX-2 expression by immunohistochemistry in distal adenomas compared with proximal. Other studies have not demonstrated this relationship (13, 21). Dimburg et al.(10) found increased COX-2 in rectal cancers as compared with cancers in other locations.

A size-dependent increase in COX-2 has been described in CRCs (14) and in adenomas (4, 13), although Hao et al.(7) did not find adenoma size to be related to COX-2 expression. In a study by Yang et al.(8), PGs in familial adenomatous polyposis adenomas were increased when adenomas reached 6–7 mm. Similarly, Hasegawa et al.(6) found an increase in COX-2 in sporadic adenomas when adenomas reached 6 mm (as compared with normal mucosa).

In a study by Chapple et al.(9), although histology was an independent predictive factor for superficial interstitial cell COX-2 protein expression, COX-2 was reduced in villous adenomas as compared with tubular and tubulovillous adenomas. Until our present study, dysplasia has not been related to COX-2 expression (9, 13).

COX-1 expression (mRNA and/or protein) has been shown similar in matched normal colonic mucosa from adenomas and/or cancers with several studies reporting COX-2 as a ratio of COX-2/COX-1. Our findings and those of a small study by Chapple et al.(22) suggest that COX-1 in adenomas may be more variable and needs additional investigation.

Our study is the largest study published to date, with 108 sporadic CRAs and addresses multiple adenoma characteristics. The COX-1 and COX-2 mRNA real-time assays from paraffin-embedded archival CRAs facilitates the assessment of large numbers of archival samples and allows quantitation of mRNA. Many antibodies are currently available for COX-2 immunohistochemistry, and some controversy exists regarding the most appropriate antibody. There is also debate as to which cells within adenomas express COX-2, the epithelial or stromal cells, both of which may play a critical role in colorectal carcinogenesis.

It is essential to determine whether expression of the COX-2 gene occurs as a general phenomenon in CRAs or if COX-2 expression varies based on factors such as adenoma location within the colorectum, size, the presence of villous histology, and/or degree of dysplasia. Differences in COX-2 gene expression may result from multiple genetic/molecular pathways and/or varying environmental conditions within the colorectum.

It must be emphasized that a major limitation of this study is that the results are limited to RNA expression and that not every expressed RNA will be translated into active protein. It will be important for additional studies to determine the relationship of COX-1 and COX-2 to these adenoma characteristics, as well as other patient characteristics, including a history of nonsteroidal anti-inflammatory drugs use. With the extreme interest in use of COX-2-specific inhibitors in CRC chemoprevention, it is essential to develop a rational approach to their use. Our data and those of other investigators suggest that the COX-2 inhibitors may be most effective in the chemoprevention of the rectosigmoid colon and especially to inhibit progressive carcinogenesis in large adenomas (i.e., ≥7 mm). Whether these inhibitors are the treatment of choice for all subjects with CRAs remains to be determined.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported, in part, by Grant CA41108 from the National Cancer Institute, National Institutes of Health.

3

The abbreviations used are: CRC, colorectal cancer; CRA, colorectal adenoma; COX, cyclooxygenase; PG, prostaglandin; WBF, Wheat Bran Fiber.

Fig. 1.

Mean COX-1 and COX-2 (normalized to β-actin) with SE bars in distal () adenomas compared with proximal (▪) adenomas. COX-2 was significantly different in distal versus proximal adenomas (∗, P < 0.0001).

Fig. 1.

Mean COX-1 and COX-2 (normalized to β-actin) with SE bars in distal () adenomas compared with proximal (▪) adenomas. COX-2 was significantly different in distal versus proximal adenomas (∗, P < 0.0001).

Close modal
Fig. 2.

Mean COX-1 and COX-2 (normalized to β-actin) with SE bars in adenomas ≥7 mm () compared with adenomas <7 mm (▪). COX-2 was significantly different in larger versus smaller adenomas (∗, P < 0.04).

Fig. 2.

Mean COX-1 and COX-2 (normalized to β-actin) with SE bars in adenomas ≥7 mm () compared with adenomas <7 mm (▪). COX-2 was significantly different in larger versus smaller adenomas (∗, P < 0.04).

Close modal
Table 1

Expression by adenoma size, location, histology, and degree of dysplasiaa

a. COX-2
Adenoma size
<7 mm (n)≥7 mm (n)P              b
Location    
 Distal 1.8 ± 1.3 (39) 2.5 ± 2.3* (49) 0.02 
 Proximal 1.0 ± 0.6 (11) 0.8 ± 0.3 (9) 0.46 
Histology    
 Tubular 1.7 ± 1.3 (41) 1.9 ± 1.4 (81) 0.59 
 Tubulovillous 1.3 ± 0.9 (9) 2.8 ± 2.9* (24) 0.02 
Degree of dysplasia    
 Mild-Moderate 1.7 ± 1.4 (35) 1.6 ± 1.0 (14) 0.80 
 High 1.4 ± 0.9 (15) 2.5 ± 2.4* (44) 0.05 
a. COX-2
Adenoma size
<7 mm (n)≥7 mm (n)P              b
Location    
 Distal 1.8 ± 1.3 (39) 2.5 ± 2.3* (49) 0.02 
 Proximal 1.0 ± 0.6 (11) 0.8 ± 0.3 (9) 0.46 
Histology    
 Tubular 1.7 ± 1.3 (41) 1.9 ± 1.4 (81) 0.59 
 Tubulovillous 1.3 ± 0.9 (9) 2.8 ± 2.9* (24) 0.02 
Degree of dysplasia    
 Mild-Moderate 1.7 ± 1.4 (35) 1.6 ± 1.0 (14) 0.80 
 High 1.4 ± 0.9 (15) 2.5 ± 2.4* (44) 0.05 
b. COX-1
Adenoma size
<7 mm (n)≥7 mm (n)P              b
Location    
 Distal 1.0 ± 0.7 (39) 1.0 ± 0.6 (49) 0.75 
 Proximal 0.7 ± 0.4 (11) 0.7 ± 0.5 (9) 0.93 
Histology    
 Tubular 1.0 ± 0.6 (41) 1.1 ± 0.7 (34) 0.39 
 Tubulovillous 0.7 ± 0.4 (9) 0.7 ± 0.4 (24) 0.68 
Degree of dysplasia    
 Mild-Moderate 1.0 ± 0.5 (35) 1.0 ± 0.8 (14) 1.00 
 High 1.0 ± 0.8 (15) 1.0 ± 0.6 (44) 0.75 
b. COX-1
Adenoma size
<7 mm (n)≥7 mm (n)P              b
Location    
 Distal 1.0 ± 0.7 (39) 1.0 ± 0.6 (49) 0.75 
 Proximal 0.7 ± 0.4 (11) 0.7 ± 0.5 (9) 0.93 
Histology    
 Tubular 1.0 ± 0.6 (41) 1.1 ± 0.7 (34) 0.39 
 Tubulovillous 0.7 ± 0.4 (9) 0.7 ± 0.4 (24) 0.68 
Degree of dysplasia    
 Mild-Moderate 1.0 ± 0.5 (35) 1.0 ± 0.8 (14) 1.00 
 High 1.0 ± 0.8 (15) 1.0 ± 0.6 (44) 0.75 
a

COX-1 and COX-2 were normalized to β-actin and values are means ± SD.

b

t test on log transformed COX values.

*

Significantly different (P < 0.05).

Table 2

COX-1 and COX-2 expression in adenomas by histology, degree of dysplasia, and percentage of villous contenta

nCOX-1COX-2
Histology    
 Tubular 75 1.0 ± 0.7b 1.8 ± 1.3 
 Tubulovillous 33 0.8 ± 0.5b 2.4 ± 2.6 
Degree of dysplasia    
 Mild-Moderate 49 0.9 ± 0.6 1.7 ± 1.2 
 High 59 0.9 ± 0.6 2.2 ± 2.2 
Percentage villous component    
 0 75 1.0 ± 0.7c 1.8 ± 1.4 
 1–25 12 0.9 ± 0.5c 1.8 ± 1.4 
 ≥26 21 0.6 ± 0.4c 2.6 ± 3.1 
nCOX-1COX-2
Histology    
 Tubular 75 1.0 ± 0.7b 1.8 ± 1.3 
 Tubulovillous 33 0.8 ± 0.5b 2.4 ± 2.6 
Degree of dysplasia    
 Mild-Moderate 49 0.9 ± 0.6 1.7 ± 1.2 
 High 59 0.9 ± 0.6 2.2 ± 2.2 
Percentage villous component    
 0 75 1.0 ± 0.7c 1.8 ± 1.4 
 1–25 12 0.9 ± 0.5c 1.8 ± 1.4 
 ≥26 21 0.6 ± 0.4c 2.6 ± 3.1 
a

COX-1 and COX-2 were normalized to β-actin, and values are given as means ± SDs.

b

t test P = 0.02.

c

ANOVA, P = 0.02.

1
Jemal A., Murray T., Samuels A., Ghafoor A., Ward E., Thun M. J. Cancer statistics, 2003.
CA - Cancer J. Clin.
,
53
:
5
-26,  
2003
.
2
Winawer S. J., Zauber A. G., Ho M. N., O’Brien M. J., Gottlieb L. S., Sternberg S. S., Waye J. D., Schapiro M., Bond J. H., Panish J. F. Prevention of colorectal cancer by colonoscopic polypectomy. The National Polyp Study Workgroup.
N. Engl. J. Med.
,
329
:
1977
-1981,  
1993
.
3
van Stolk R. U., Beck G. J., Baron J. A., Haile R., Summers R. Adenoma characteristics at first colonoscopy as predictors of adenoma recurrence and characteristics at follow-up. The Polyp Prevention Study Group.
Gastroenterology
,
115
:
13
-18,  
1998
.
4
Elder D. J., Baker J. A., Banu N. A., Moorghen M., Paraskeva C. Human colorectal adenomas demonstrate a size-dependent increase in epithelial cyclooxygenase-2 expression.
J. Pathol.
,
198
:
428
-434,  
2002
.
5
Khan K. N., Masferrer J. L., Woerner B. M., Soslow R., Koki A. T. Enhanced cyclooxygenase-2 expression in sporadic and familial adenomatous polyposis of the human colon.
Scand. J. Gastroenterol.
,
36
:
865
-869,  
2001
.
6
Hasegawa K., Ichikawa W., Fujita T., Ohno R., Okusa T., Yoshinaga K., Sugihara K. Expression of cyclooxygenase-2 (COX-2) mRNA in human colorectal adenomas.
Eur. J. Cancer
,
37
:
1469
-1474,  
2001
.
7
Hao X., Bishop A. E., Wallace M., Wang H., Willcocks T. C., Maclouf J., Polak J. M., Knight S., Talbot I. C. Early expression of cyclo-oxygenase-2 during sporadic colorectal carcinogenesis.
J. Pathol.
,
187
:
295
-301,  
1999
.
8
Yang V. W., Shields J. M., Hamilton S. R., Spannhake E. W., Hubbard W. C., Hylind L. M., Robinson C. R., Giardiello F. M. Size-dependent increase in prostanoid levels in adenomas of patients with familial adenomatous polyposis.
Cancer Res.
,
58
:
1750
-1753,  
1998
.
9
Chapple K. S., Cartwright E. J., Hawcroft G., Tisbury A., Bonifer C., Scott N., Windsor A. C., Guillou P. J., Markham A. F., Coletta P. L., Hull M. A. Localization of cyclooxygenase-2 in human sporadic colorectal adenomas.
Am. J. Pathol.
,
156
:
545
-553,  
2000
.
10
Dimberg J., Samuelsson A., Hugander A., Soderkvist P. Differential expression of cyclooxygenase 2 in human colorectal cancer.
Gut
,
45
:
730
-732,  
1999
.
11
DuBois R. N., Radhika A., Reddy B. S., Entingh A. J. Increased cyclooxygenase-2 levels in carcinogen-induced rat colonic tumors.
Gastroenterology
,
110
:
1259
-1262,  
1996
.
12
Kargman S. L., O’Neill G. P., Vickers P. J., Evans J. F., Mancini J. A., Jothy S. Expression of prostaglandin G/H synthase-1 and -2 protein in human colon cancer.
Cancer Res.
,
55
:
2556
-2559,  
1995
.
13
Eberhart C. E., Coffey R. J., Radhika A., Giardiello F. M., Ferrenbach S., DuBois R. N. Up-regulation of cyclooxygenase 2 gene expression in human colorectal adenomas and adenocarcinomas.
Gastroenterology
,
107
:
1183
-1188,  
1994
.
14
Fujita T., Matsui M., Takaku K., Uetake H., Ichikawa W., Taketo M. M., Sugihara K. Size- and invasion-dependent increase in cyclooxygenase 2 levels in human colorectal carcinomas.
Cancer Res.
,
58
:
4823
-4826,  
1998
.
15
Yamaguchi A., Ishida T., Nishimura G., Katoh M., Miyazaki I. Investigation of colonic prostaglandins in carcinogenesis in the rat colon.
Dis. Colon Rectum
,
34
:
572
-576,  
1991
.
16
Alberts D. S., Martinez M. E., Roe D. J., Guillen-Rodriguez J. M., Marshall J. R., van Leeuwen J. B., Reid M. E., Ritenbaugh C., Vargas P. A., Bhattacharyya A. B., Earnest D. L., Sampliner R. E. Lack of effect of a high-fiber cereal supplement on the recurrence of colorectal adenomas. Phoenix Colon Cancer Prevention Physicians’ Network.
N. Engl. J. Med.
,
342
:
1156
-1162,  
2000
.
17
Lehmann U., Kreipe H. Real-time PCR analysis of DNA and RNA extracted from formalin-fixed and paraffin-embedded biopsies.
Methods
,
25
:
409
-418,  
2001
.
18
Gibson U. E., Heid C. A., Williams P. M. A novel method for real time quantitative RT-PCR.
Genome Res.
,
6
:
995
-1001,  
1996
.
19
Heid C. A., Stevens J., Livak K. J., Williams P. M. Real time quantitative PCR.
Genome Res.
,
6
:
986
-994,  
1996
.
20
Chapple K. S., Scott N., Guillou P. J., Coletta P. L., Hull M. A. Interstitial cell cyclooxygenase-2 expression is associated with increased angiogenesis in human sporadic colorectal adenomas.
J. Pathol.
,
198
:
435
-441,  
2002
.
21
Sheehan K. M., Sheahan K., O’Donoghue D. P., MacSweeney F., Conroy R. N., Fitzgerald D. J., Murray F. E. The relationship between cyclooxygenase-2 expression and colorectal cancer.
J. Am. Med. Assoc.
,
282
:
1254
-1257,  
1999
.
22
Chapple K. S., Scott N., Guillou P. J., Coletta P. L., Hull M. A. Analysis of cyclooxygenase expression in human colorectal adenomas.
Dis. Colon Rectum
,
45
:
1316
-1324,  
2002
.