Skip to main content
NCBI home page
Molecular Oncology logoLink to Molecular Oncology
. 2013 Nov 8;8(2):169–177. doi: 10.1016/j.molonc.2013.10.009

Targeted Cox2 gene deletion in intestinal epithelial cells decreases tumorigenesis in female, but not male, Apc Min/+ mice

Durga P Cherukuri 1, Tomo-o Ishikawa 1, Patrick Chun 1, Art Catapang 1, David Elashoff 2, Tristan R Grogan 2, James Bugni 3, Harvey R Herschman 1,✉
PMCID: PMC3963510  NIHMSID: NIHMS541065  PMID: 24268915

Abstract

Mice heterozygous for mutations in the adenomatous polyposis coli gene (Apc+/− mice) develop intestinal neoplasia. Apc+/− tumor formation is thought to be dependent on cyclooxygenase 2 (COX2) expression; both pharmacologic COX2 inhibition and global Cox2 gene deletion reduce the number of intestinal tumors in Apc+/− mice. COX2 expression is reported in epithelial cells, fibroblasts, macrophages and endothelial cells of Apc+/− mouse polyps. However, the cell type(s) in which COX2 expression is required for Apc+/− tumor induction is not known. To address this question, we developed ApcMin/+ mice in which the Cox2 gene is specifically deleted either in intestinal epithelial cells or in myeloid cells. There is no significant difference in intestinal polyp number between ApcMin/+ mice with a targeted Cox2 gene deletion in myeloid cells and their control littermate ApcMin/+ mice. In contrast, ApcMin/+ mice with a targeted Cox2 deletion in intestinal epithelial cells have reduced intestinal tumorigenesis when compared to their littermate control ApcMin/+ mice. However, two gender‐specific effects are notable. First, female ApcMin/+ mice developed more intestinal tumors than male ApcMin/+ mice. Second, targeted intestinal epithelial cell Cox2 deletion decreased tumorigenesis in female, but not in male, ApcMin/+ mice. Considered in the light of pharmacologic studies and studies with global Cox2 gene knockout mice, our data suggest that (i) intrinsic COX2 expression in intestinal epithelial cells plays a gender‐specific role in tumor development in ApcMin/+ mice, and (ii) COX2 expression in cell type(s) other than intestinal epithelial cells also modulates intestinal tumorigenesis in ApcMin/+ mice, by a paracrine process.

Keywords: APC, COX2, Cyclooxygenase 2, Mouse intestinal tumors, Gender‐specific effect

Highlights

  • Effect of tissue‐specific Cox2 deletion on polyposis in ApcMin/+ mice was examined.

  • Targeted myeloid cell Cox2 deletion has no effect on polyposis in ApcMin/+ mice.

  • Intestinal epithelial Cox2 deletion did not affect ApcMin/+ male mice polyposis.

  • Intestinal epithelial Cox2 deletion greatly reduced ApcMin/+ female mice polyposis.

  • In females, intestinal epithelial cell Cox2 deletion affected mainly large tumors.

1. Introduction

Heterozygous inactivating adenomatous polyposis coli (APC) gene mutations cause familial adenomatous polyposis (FAP) in patients. Most sporadic human colorectal cancers also exhibit APC gene mutations. Apc +/− mutant mice have been widely studied, to better understand mechanisms of Apc +/− intestinal tumorigenesis (McCart et al., 2008). One of the most commonly used Apc +/− mouse models, the multiple intestinal neoplasia (Min) or Apc Min/+ mouse (Moser et al., 1990), develops ∼30–100 tumors, predominantly in the small intestine (McCart et al., 2008).

Cyclooxygenase‐2 (COX2) is an inducible form of cyclooxygenase, the enzyme(s) that catalyze the first two steps in the conversion of arachidonic acid to prostaglandins. COX2 mRNA is often elevated in human colorectal adenomas and adenocarcinomas (Eberhart et al., 1994; Maekawa et al., 1998), suggesting that COX2 may play a role in the early stages of colorectal cancer. Markedly increased COX2 mRNA and protein levels are also observed in intestinal tumors of Apc Δ716/+ and Apc Min/+ mice (Oshima et al., 1996; Williams et al., 1996; Backlund et al., 2005). Cox2 gene disruption reduced intestinal tumorigenesis in both Apc Δ716/+ and Apc Min/+ Cox2 knockout mice (Oshima et al., 1996; Chulada et al., 2000); selective pharmacologic COX2 enzyme inhibition also decreased intestinal tumor frequency in Apc Δ716/+ and Apc Min/+ mice (Jacoby et al., 2000; Oshima et al., 2001). These studies with Apc +/− mice provide direct evidence for a role of COX2‐derived prostaglandins in Apc +/− intestinal tumorigenesis. Celecoxib (a COX2 selective inhibitor) treatment attenuated the number and size of polyps in FAP patients (Higuchi et al., 2003; Lynch et al., 2010) also suggesting a role for COX2 in hereditary APC‐dependent tumor formation.

Despite genetic and pharmacologic evidence, how COX2 promotes intestinal tumor development in Apc +/− mice and FAP patients is still unclear. Epithelial cell COX2 staining was observed in all adenocarcinomas (Brosens et al., 2008) and in epithelial and stromal cells (Jungck et al., 2004) from FAP patients. Oshima et al. (1996) observed COX2 expression in the interstitial cells of polyps from Apc Δ716/+ mice, but not in the dysplastic epithelium. Williams et al. (1996) concurrently reported the presence of COX2 in the epithelial cells of polyps from Apc Min/+ mice. Sonoshita et al. (2002) subsequently reported the presence of COX2 in fibroblasts and endothelial cells in polyps from both APC patients and Apc Δ716/+ mice. However, they did not observe COX2 expression in macrophages or lymphocytes. In contrast, Hull et al. (1999) reported the presence of COX2 immunostaining in the lamina propria and in macrophages of polyps from Apc Min/+ mice.

The studies described above report, in Apc +/− mice and FAP patients, the presence of COX2 in intestinal tumor epithelial cells, stromal cells of the surrounding tumor microenvironment, and macrophages present in the tumors. In their Nature Reviews article discussing colorectal cancer and COX2, published over a decade ago, Gupta and DuBois (2001) comment “There is no consensus at present on what cell types within a tumor express COX2…” They pose the question “Does understanding the exact location of COX2 in colorectal tumors have any bearing on understanding how COX2 promotes colorectal cancer cell growth?” and end with the comment that “knowing the specific cell type that expresses COX2 would be essential for understanding how COX2 promotes colorectal cancer progression. The most direct experimental approach to address these issues could be through the use of mice with tissue specific genetic ablation of Cox2. In such experiments, it would be possible to find out whether removal of COX2 only from epithelial or macrophage lineages, for example, decreases polyp burden in Apc Min/+ mice (Gupta and DuBois, 2001).”

The ambiguities regarding the role(s) of cell‐type specific COX2 expression in driving tumor progression in Apc +/− tumors have not changed since Gupta and DuBois (Gupta and DuBois, 2001) wrote their review. To address these questions, we generated Apc Min/+ mice with specific Cox2 gene deletions in intestinal epithelial cells or in myeloid cells and examine here the effect of these targeted Cox2 deletions on COX2 expression, intestinal polyp number, and polyp size in Apc Min/+ mice.

2. Materials and methods

2.1. Mice

A conditional Cox2 gene knockout mouse, Cox2 flox/flox, in which exons 4 and 5 of the Cox2 gene are flanked by loxP sites was generated previously (Ishikawa and Herschman, 2006). Apc Min/+ mice (C57BL/6J‐Apc Min/J), LysMCre knock‐in mice (B6.129P2‐Lyz2 tm1(cre)Ifo/J) and VillinCre transgenic mice (B6.SJL‐Tg(Vil‐cre)997Gum/J) were purchased from the Jackson Laboratory (Bar Harbor, MA). All animal experiments were conducted with the approval of the Animal Research Committee, University of California, Los Angeles.

Apc Min/+ mice, LysM Cre/+ knock‐in mice, and VillinCre transgenic mice were crossed with Cox2 flox/+ or Cox2 flox/flox mice to generate Cox2 flox/flox Apc Min/+ mice, Cox2 flox/flox LysM Cre/+ mice, and Cox2 flox/flox VillinCre Tg mice. To generate Apc Min/+ mice with homozygous Cox2 deletions specifically in intestinal epithelial cells, genetic crosses between Cox2 flox/flox Apc Min/+ mice and Cox2 flox/flox VillinCre Tg mice were performed. These crosses produced Apc Min/+ mice homozygous for the Cox2 floxed alleles and positive for the VillinCre transgene (Cox2 flox/flox Apc Min/+ VillinCre Tg mice) or negative for the VillinCre Tg allele (Cox2 flox/flox Apc Min/+ mice). Similarly, Cox2 flox/flox Apc Min/+ mice were crossed with Cox2 flox/flox LysM Cre/+ mice to generate Apc Min/+ mice with myeloid cell‐specific homozygous Cox2 deletions (Cox2 flox/flox Apc Min/+ LysM Cre/+ mice) and their littermate controls (Cox2 flox/flox Apc Min/+ mice).

Cox2 flox/flox Apc Min/+ VillinCre Tg mice, in which the Cox2 gene is deleted in intestinal epithelial cells, will be referred to as Apc Min/+ Cox2 ΔE mice. Cox2 flox/flox Apc Min/+ LysM Cre/+ mice, in which the Cox2 gene is deleted in myeloid cells, will be referred to as Apc Min/+ Cox2 ΔM mice. Littermate control mice for both Apc Min/+ Cox2 ΔE and Apc Min/+ Cox2 ΔM mice will be referred to as Apc Min/+ mice. Cox2 gene deletion in myeloid cells and in intestinal epithelial cells, as a result of cell‐type specific Cre recombinase expression, was examined as previously described (Ishikawa et al., 2011).

2.2. Genotyping of mice

DNA was isolated from tail snips. Cre and Cox2 allele genotypes were determined as previously described (Ishikawa et al., 2011). A PCR‐based protocol from the Jackson Laboratory was adapted to genotype the Apc locus. PCR reactions for Apc wild type or Min alleles were carried out separately with appropriate positive, negative and no template controls. All PCR reactions were carried out using an MJ Research thermal cycler.

2.3. Tumor number and size measurements

Apc Min/+ Cox2 ΔM mice and their littermate Apc Min/+ controls were sacrificed at 150 days of age; Apc Min/+ Cox2 ΔE mice and their littermate Apc Min/+ controls were sacrificed at 180 days of age. The entire intestinal tract was removed, flushed with PBS several times, and the small intestine was cut into 4–6 pieces. Each section was opened longitudinally and kept flat between bibulous papers. The tissues were fixed in 10% formalin overnight, transferred to 70% alcohol and stored at 4 °C until use. Intestine sections were stained with 1% methylene blue, and tumors were counted using a dissecting microscope at 10× magnification. Maximum tumor diameter was measured with a calibrated eye piece reticle. Tumors were classified into the following size groups: <1 mm, 1–2 mm, 2–3 mm, 3–4 mm, 4–5 mm, and >5 mm (Oshima et al., 1996).

2.4. Immunohistochemistry

COX2 immunohistochemical detection on formalin‐fixed, paraffin‐embedded tissues was performed as described previously (Ishikawa et al., 2011). Briefly, 4 μm thickness sections were used for immunostaining, and COX2 protein was detected with a polyclonal anti‐COX2 antibody (Thermo Scientific).

2.5. Statistics

Variation in tumor counts between mice was investigated utilizing two‐way ANOVA models. These models contained the main effects of gender (m/f), Cre (+/−), and the interaction effect of gender and Cre. This model was run separately for both cell types (LysM Cre and VillinCre Tg). Average tumor counts are reported as mean ± 1 standard error. Residual analysis yielded no evidence that the normal model was an inadequate fit. The significance threshold was set at 0.05. Statistical analyses were performed using SAS 9.3 (SAS Institute Inc., Cary, NC).

3. Results

3.1. Myeloid cell‐specific Cox2 gene deletion in Apc Min/+ mice does not affect polyp formation

Macrophages are among the several cell types in which COX2 is reported to be expressed in Apc Min/+ intestinal tumors (Hull et al., 1999). To investigate whether COX2 expression in myeloid cells, including macrophages, contributes to tumor burden in Apc Min/+ mice, we created Apc Min/+ mice with a targeted deletion in the myeloid cell Cox2 gene. Cox2 flox/flox Apc Min/+ mice were crossed with Cox2 flox/flox LysM Cre/+ mice to create both Apc Min/+ Cox2 ΔM mice with a myeloid‐specific Cox2 gene deletion and their littermate Apc Min/+ controls. Apc Min/+ Cox2 ΔM mice (n = 8) developed 60 ± 7.9 polyps at 150 days of age (Figure 1A). Littermate control Apc Min/+ mice (n = 9) developed 54 ± 4.6 polyps. There was no significant difference (p = 0.35, group effect from two‐way ANOVA model) in intestinal polyp frequency in Apc Min/+ Cox2 ΔM mice and littermate control Apc Min/+ mice. In addition, there was no difference in size distribution in the polyps found in Apc Min/+ mice and Apc Min/+ Cox2 ΔM mice (data not shown). Thus we did not detect any effect on intestinal tumorigenesis in Apc Min/+ mice unable to express COX‐2 in myeloid cells.

Figure 1.

Figure 1

Open in a new tab

Intestinal tumor formation in ApcMin/+ mice with cell‐type specific Cox2 gene deletions. (Panel A) The Cox2 gene was deleted in myeloid cells of ApcMin/+Cox2ΔM mice. Intestinal tumors in ApcMin/+Cox2ΔM mice and littermate control ApcMin/+ mice were counted at 150 days of age. (Panel B) The Cox2 gene was deleted in intestinal epithelial cells of ApcMin/+Cox2ΔE mice. Intestinal tumors in ApcMin/+Cox2ΔE mice and control littermate ApcMin/+ mice were counted at 180 days of age. p‐values for the Cre effect were computed from the two‐way ANOVA models. Means ± S.E. are shown.

3.2. Intestinal epithelial cell‐specific Cox2 gene deletion in Apc Min/+ mice reduces polyp formation

COX2 expression is observed in dysplastic epithelial cells of the intestinal tumors of Apc Min/+ mice (Williams et al., 1996), and in epithelial cells of polyps from FAP patients (Brosens et al., 2008; Jungck et al., 2004). To determine whether intrinsic intestinal epithelial cell COX2 expression modulates polyp formation in Apc Min/+ mice, we created Apc Min/+ mice with a targeted Cox2 deletion in their intestinal epithelial cells. Cox2 flox/flox Apc Min/+ mice were crossed with Cox2 flox/flox VillinCre Tg mice to generate both Apc Min/+ Cox2 ΔE mice with targeted intestinal epithelial cell Cox2 deletions and their littermate control Apc Min/+ mice. The control Apc Min/+ mice (n = 15) developed 40 ± 6.3 tumors at 180 days of age (Figure 1B). In contrast, Apc Min/+ Cox2 ΔE mice (n = 17), in which the Cox2 gene is deleted in intestinal epithelial cells, developed 27 ± 3.0 polyps. The 31% reduction in polyp formation as a consequence of deleting the intestinal epithelial Cox2 gene is – in contrast to the targeted deletion of the Cox2 gene in myeloid cells – highly significant (p = 0.004, group effect from two‐way ANOVA).

In addition to reducing the number of polyps that develop in Apc Min/+ mice, targeted Cox2 gene deletion in intestinal epithelial cells also altered the polyp size distribution (Figure 2). ”Small” polyps, i.e., polyps less than 2 mm in diameter (Oshima et al., 1996), were similar in number in Apc Min/+ mice and Apc Min/+ Cox2 ΔE mice. However, there was a substantial reduction in “large” polyps; e.g., polyps 2 mm or greater in diameter (Oshima et al., 1996). These data on polyp frequency and size distribution suggest that intrinsic intestinal epithelial cell COX2 expression is a strong driving force in polyp promotion in Apc Min/+ mice.

Figure 2.

Figure 2

Open in a new tab

Sizes of intestinal tumors in ApcMin/+Cox2ΔE mice (n = 17), in which the Cox2 gene is specifically deleted in intestinal epithelial cells, and of littermate ApcMin/+ mice (n = 15). Data are presented as average number of tumors per mouse in each group. Means ± S.D. are shown.

Immunostaining in the intestinal tumors of control Apc Min/+ mice showed COX2 expression in many, but not all, intestinal epithelial cells and in cells of the lamina propria (Figure 3A and B). In the normal intestinal mucosa adjacent to the tumors, COX2 expression was present in lamina propria cells, but not in the epithelium (Figure 3C). In contrast to the results for control littermate Apc Min/+ mice, few if any dysplastic epithelial cells in Apc Min/+ Cox2 ΔE mice intestinal tumors stain positive for COX2 (Figure 3D and E). However, COX2 expression was detected in the cells of the tumor lamina propria. As in the normal intestinal mucosa of Apc Min/+ tumors (Figure 3C), cells of the lamina propria, but not epithelial cells, were positive for COX2 expression in the normal mucosa adjacent to tumors of Apc Min/+ Cox2 ΔE mice (Figure 3F).

Figure 3.

Figure 3

Open in a new tab

COX2 protein expression in intestinal tumors of ApcMin/+ and ApcMin/+Cox2ΔE mice. (A) Control littermate ApcMin/+ mouse intestinal tumor, with extensive COX2 expression in tumor epithelial cells (100×). (B) Higher magnification of control littermate ApcMin/+ mouse intestinal tumor (200×). Arrows indicate positive COX2 protein staining in epithelial cells. (C) Tumor‐adjacent normal mucosa of control littermate ApcMin/+ mouse (100×). Arrows indicate positive COX2 staining in lamina propria cells. (D) ApcMin/+Cox2ΔE mouse intestinal tumor, with COX2 expression present predominantly in non‐epithelial cells (100×). (E) Higher magnification of ApcMin/+Cox2ΔE mouse intestinal tumor (200×). Arrows indicate positive COX2 protein staining in stromal, non‐epithelial cells. (F) Tumor‐adjacent normal mucosa of ApcMin/+Cox2ΔE mouse (100×). Arrows indicate positive COX2 staining in lamina propria cells.

3.3. Gender‐specific differences in polyp formation in control Apc Min/+ mice

We observed an unexpected (to us) gender difference in polyp formation in the control Apc Min/+ mice. For both control Apc Min/+ mice in the myeloid Cox2 gene deletion study (Figure 4A) and the control Apc Min/+ mice in the intestinal epithelial cell Cox2 gene deletion study (Figure 4B), we found significantly more polyps in the female Apc Min/+ mice than in the control male Apc Min/+ mice. Although there are hundreds of publications studying characteristics of Apc +/− mice, we found only two other reports of gender‐based differences in polyp frequency (McAlpine et al., 2006; Yoo et al., 2008).

Figure 4.

Figure 4

Open in a new tab

Gender effects on the frequency of intestinal tumors in control ApcMin/+ mice for targeted deletion studies with both ApcMin/+Cox2ΔM mice and ApcMin/+Cox2ΔE mice. (Panel A) intestinal tumors in male versus female control littermate ApcMin/+ mice for the experiment in which the Cox2 gene was deleted in myeloid cells. (Panel B) intestinal tumors in male versus female control littermate ApcMin/+ mice for the experiment in which the Cox2 gene was deleted in intestinal epithelial cells. p‐values were computed from the pairwise differences estimated by the two‐way ANOVA models. Means ± S.E. are shown.

3.4. Intestinal epithelial cell‐specific Cox2 gene deletion reduces polyp formation in female Apc Min/+ mice, but not in male Apc Min/+ mice

Once we observed the gender‐based differences in polyp formation in control Apc Min/+ mice (Figure 4), we revisited the effect of myeloid cell‐ and intestinal epithelial cell‐targeted Cox2 gene deletion, to determine whether there might be gender‐based differences in the roles of myeloid or intestinal epithelial cell COX2‐driven polyp formation in Apc Min/+ mice. Deletion of myeloid cell Cox2 in Apc Min/+ Cox2 ΔM mice had no effect on intestinal tumorigenesis, in either male or female mice (data not shown).

Deletion of the Cox2 gene in intestinal epithelial cells of male Apc Min/+ Cox2 ΔE mice had no effect on the frequency of polyps, when compared to their control Apc Min/+ littermates (Figure 5). In contrast, Cox2 gene deletion in intestinal epithelial cells of female Apc Min/+ Cox2 ΔE mice caused a greater than two fold, highly significant (p < 0.001, pairwise difference in two‐way ANOVA model) decrease in polyp formation when compared to their female Apc Min/+ littermates (Figure 5). Perhaps coincidentally, the observed polyp frequency in female Apc Min/+ Cox2 ΔE mice was reduced to the same frequency observed in male Apc Min/+ and Apc Min/+ Cox2 ΔE mice.

Figure 5.

Figure 5

Open in a new tab

Intestinal epithelial Cox2 deletion as a result of targeted Cre recombinase expression has no effect on male ApcMin/+Cox2ΔE mice, but significantly decreased tumor formation in female ApcMin/+Cox2ΔE mice. The p‐value was computed from the pairwise difference estimated by the two‐way ANOVA model. Means ± S.E. are shown.

We analyzed polyp size distribution in male and female Apc Min/+ Cox2 ΔE mice and their littermate male and female Apc Min/+ control mice in two ways. Deletion of the Cox2 gene in intestinal epithelial cells of female Apc Min/+ mice affects the size of polyps (Figure 6A), as well as the total number of polyps (Figure 5). Comparing the numbers of polyps of different sizes (Figure 6A) demonstrates that there is a striking (greater than 75%) reduction in the numbers of polyps of diameter 2 mm or greater in female Apc Min/+ Cox2 ΔE mice when compared to littermate female Apc Min/+ mice. Figure 6B shows the percentage (rather than the numbers, as in Figure 6A) of polyps of differing diameters in female Apc Min/+ and Apc Min/+ Cox2 ΔE mice. Many authors classify polyps of Apc Min/+ mice as either “small” (less than 2 mm in diameter), or “large” (2 mm or greater in diameter). By these criteria, not only is the number of polyps reduced in female Apc Min/+ mice by targeted deletion of the Cox2 gene in intestinal epithelial cells (Figure 5); in addition, the polyps that do form in female Apc Min/+ Cox2 ΔE mice are skewed heavily toward “small” polyps, in contrast to the predominantly “large” polyps found in control littermate female Apc Min/+ mice.

Figure 6.

Figure 6

Open in a new tab

Sizes of intestinal tumors in male and female ApcMin/+Cox2ΔE mice, in which the Cox2 gene is specifically deleted in intestinal epithelial cells, and in littermate ApcMin/+ mice. (Panel A) Numbers of intestinal tumors per mouse in female ApcMin/+Cox2ΔE mice (n = 7) and their female littermate ApcMin/+ mice (n = 6). Data are shown as number of tumors per mouse in each group. (Panel B) Percent of tumors in each size class for female ApcMin/+Cox2ΔE mice and their female littermate ApcMin/+ mice. Small polyps are defined as less than 2 mm in size; large polyps are 2 mm or greater in size. (Panel C) Numbers of intestinal tumors per mouse in male ApcMin/+Cox2ΔE mice (n = 10) and their male littermate ApcMin/+ mice (n = 9). (Panel D) Percent of tumors in each size class for male ApcMin/+Cox2ΔE mice and their male littermate ApcMin/+ mice. Means ± S.D. are shown.

In contrast, to the results for female Apc min/+ mice, neither the number (Figure 5) nor the size distribution (Figure 6C and D) of polyps in male Apc Min/+ mice was affected by targeted deletion of the Cox2 gene in intestinal epithelial cells. In summary, our results demonstrate that intestinal epithelial COX2 expression plays a strong modulatory role in tumor development in female Apc Min/+ mice, but not in male Apc Min/+ mice.

4. Discussion

Increased COX2 expression has been reported in human intestinal tumors of multiple origins, and in intestinal tumors that occur in a variety of animal models. A causal role for COX2 expression in the progression of intestinal cancer has been most clearly defined in hereditary familial adenomatous polyposis (FAP) patients, in whom the APC gene is mutated, and in Apc +/− mice. Treatment of FAP patients with sulindac (Giardiello et al., 1993), celecoxib (Steinbach et al., 2000) or rofecoxib (Higuchi et al., 2003), all selective COX2 inhibitors, reduces polyp frequency and size, strongly suggesting a causal role for COX2 expression as a driver for FAP. Similarly, both celecoxib (Jacoby et al., 2000) and rofecoxib (Oshima et al., 2001) inhibition of Apc Min/+ mouse intestinal polyp induction reinforce the suggestion for a COX2 causal role in Apc +/− intestinal cancer. The reduction in intestinal tumor number and size in Apc Δ716/+ (Oshima et al., 1996) and Apc Min/+ (Chulada et al., 2000) mice with global Cox2 gene deletions provided conclusive proof that COX2 expression, in some cell type(s), plays a major role in Apc +/− tumor progression.

Dysplastic intestinal epithelial cells, stromal fibroblasts, endothelial cells and macrophages are all reported to express COX2 (Oshima et al., 1996; Williams et al., 1996; Sonoshita et al., 2002; Hull et al., 1999) in intestinal tumors of Apc +/− mice. Each cell type has been suggested as a (and/or the) cell type in which COX2 must be expressed to drive tumor progression. However, neither pharmacologic studies with COX2 inhibitors nor studies of mice with global Cox2 gene deletions can conclusively identify the cell type(s) in which COX2 expression is required for Apc +/− tumorigenicity, since both interventions eliminate COX2 function in all cells.

Hull et al. (1999) demonstrated COX2 expression in macrophage of the lamina propria of Apc Min/+ mice, and suggested a paracrine effect of macrophage COX2 function on epithelial cells in Apc Min/+ mouse adenomas. However, myeloid cell‐specific Cox2 gene deletion has no effect on the number or size of Apc Min/+ intestinal tumors (Figure 1A); myeloid/macrophage derived, COX‐2 dependent prostanoids do not modulate tumor formation in Apc Min/+ mice. In contrast, Cox2 intestinal epithelial cell‐specific deletion significantly reduced the frequency of Apc Min/+ tumors in our cohort of Apc Min/+ Cox2 ΔE mice (Figure 1B), initially suggesting a required role in these cells for COX2 function in Apc +/− tumor progression.

When we considered gender as a variable in Apc Min/+ intestinal tumor induction we found that control Apc Min/+ female mice, both for the myeloid‐specific Cox2 knockout (Apc Min/+ Cox2 ΔM) mice and for the intestinal epithelial cell‐specific Cox2 knockout (Apc Min/+ Cox2 ΔE) mice, developed more tumors than did their male Apc Min/+ counterparts (Figure 4). Although there are hundreds of papers reporting studies with Apc +/− mice, we found only two other studies that report on the role of gender in tumor formation. McAlpine et al. (2006) and Yoo et al. (2008) also observed greater intestinal tumor frequency in female Apc +/− mice than in male Apc +/− mice. However, it is clear that this gender distinction is not widely appreciated.

When we analyzed gender‐specific effects in our Cox2 targeted deletion studies, we found that intrinsic COX‐2 expression in intestinal epithelial cells of male Apc Min/+ mice plays no role, in either tumor frequency or tumor size, in intestinal polyp development; these characteristics were the same in male Apc Min/+ Cox2 ΔE mice and their littermate Apc Min/+ control mice. In contrast, both polyp frequency and size were reduced in female Apc Min/+ Cox2 ΔE mice when compared to their littermate control Apc Min/+ mice. Despite the reduction in polyps in the female Apc Min/+ Cox2 ΔE mice, there remained a cohort of intestinal polyps that formed in these animals; i.e., in female mice unable to make COX2‐dependent prostanoids in their intestinal epithelial cells.

In studies with Apc Δ716/+ Cox2 −/− mice and Apc Min/+ Cox2 −/− that had homozygous global Cox2 gene deletions, intestinal polyp formation was reduced by 84–86% (Oshima et al., 1996; Chulada et al., 2000). Although gender differences were not considered in these studies, the data suggest that COX2‐derived prostanoid product(s) play a role in the development of a great majority of the polyps that occur in Apc +/− mice. The most likely interpretation of the frequency differences in tumor formation in Apc Min/+ mice in which there is a global Cox2 knockout (Apc Min/+; Cox2 −/− mice) and tumor formation in Apc Min/+ mice in which there is a targeted epithelial cell‐specific Cox2 knockout (Apc Min/+; Cox2 ΔE mice) is that COX2 expression in cells of the tumor microenvironment, and thus a paracrine effect(s) of COX2‐derived prostaglandins from these cells, is also required for tumor development in Apc +/− mice. A less likely possibility is that a subset of intestinal tumors that develop in these animals does not require COX2 function in any cell type. In light of our results, unequivocal clarification of these alternative possibilities would require re‐examination of the occurrence of tumors in Apc +/−; Cox2 −/− mice. If intestinal tumors do, in fact, occur in Apc +/−; Cox2 −/− mice, the result would suggest a COX2‐independent pathway to (at least some) intestinal tumors in Apc +/− mice. If tumor frequency is totally (or even largely) eliminated in Apc +/− Cox2 −/− mice, identification of a cell type(s) in which COX2 expression is causally needed in a paracrine fashion for intestinal tumorigenesis would require analysis of Apc +/− mice with targeted Cox2 gene deletions in alternative candidate cell types (e.g., fibroblasts, myofibroblasts, and endothelial cells).

In contrast to the results for male Apc Min/+ mice, in which tumor formation appears to be independent of COX‐2 expression in intestinal epithelial cells (Figure 5), tumor formation is reduced by nearly two‐thirds in female Apc Min/+ Cox2 ΔE mice, in which the Cox2 gene is specifically deleted in intestinal epithelial cells (Figure 5). Tumor frequency in female Apc Min/+ Cox2 ΔE mice is similar to tumor frequency in male Apc Min/+ mice; intestinal epithelial cell COX2 expression is largely, if not completely, responsible for the greater intestinal tumor frequency in female Apc Min/+ mice.

The frequency (Figure 6A) and percentage (Figure 6B) of “small” tumors, less than 2 mm in diameter, do not differ in female Apc Min/+ mice and Apc Min/+; Cox2 ΔE mice. However, “large” tumors are less frequent in female mice with an intestinal epithelial cell‐targeted Cox2 gene deletion (Apc Min/+; Cox2 ΔE mice) than in their female Apc Min/+ littermates (Figure 6A); the large tumors are a greater percentage of the intestinal epithelial tumors that form in female Apc Min/+; Cox2 ΔE mice (Figure 6B). In contrast, the distribution of tumor sizes in male Apc Min/+; Cox2 ΔE mice and their littermate Apc Min/+ control mice are similar (Figure 6C and D). Oshima et al. (1996) report that COX2 expression could not be detected in “small” tumors of Apc Δ716/+ mice, but was easily detectable in larger tumors. Collectively, these data support the suggestions (i) that tumor frequency and size are driven, to a much greater extent in female Apc +/− mice than in male Apc +/− mice, by an intrinsic intestinal epithelial cell COX2 product and (ii) that tumor formation in Apc +/− mice of both genders is also dependent on the paracrine action of a COX2 product of a cell type in the microenvironment. Seno et al. (2002) suggest that stromal cell COX2 expression results in PGE2‐driven angiogenesis and consequent tumor promotion in Apc Δ716/+ mice.

In addition to genetic studies of the Apc +/− mouse model of hereditary human APC colon cancer, the role of COX‐2 has been examined genetically in other mouse colon cancer models. Ulcerative colitis in patients is a frequent precursor to colon cancer. Colitis‐dependent colon cancer has been modeled by exposing mice to a single injection of the carcinogen azoxymethane (AOM), followed by colon tumor promotion with dextran sodium sulfate (DSS) oral administration. Colon tumors induced in mice by AOM/DSS administration express increased COX‐2 levels relative to surrounding colon tissue (Ishikawa and Herschman, 2010). However, despite the presence of elevated COX‐2 protein in the AOM/DSS‐induced tumors, and in contrast to results with Apc +/− mice, global Cox2 gene deletion had no effect on AOM/DSS colon cancer induction; COX2‐derived prostanoids are not required for tumor induction in this mouse model of colitis‐associated colon cancer (Ishikawa and Herschman, 2010). Clearly, one cannot generalize from one type of colon cancer to another with respect to roles for COX2 modulation of colon tumor progression.

We have not found a previous clinical study in patients or preclinical study in animal models in which gender‐specific differences in COX2 expression have been correlated with intestinal cancer detection, prognosis, survival or therapeutic outcome, nor have we found another report in any animal model in which differential COX2 expression mediates gender differences in tumor development. However, there are a few clear examples of sexual dimorphism in cyclooxygenase‐modulated biological processes. Estrogen‐driven COX2 expression in female mice leads to prostacyclin‐dependent atheroprotection in female mice (Egan et al., 2004). In chronic Freud's adjuvant‐induced arthritis and inflammatory pain, reduced edema and joint destruction were observed in female Cox2 −/− mice relative to appropriate controls, but not in males (Chillingworth et al., 2006). Naor et al. (2009) report that indomethacin was more effective in blocking LPS‐promoted lung colonization in male rats than in female rats. Elucidation of gender‐mediated, COX2‐dependent tumor number and progression in Apc +/− mice will require both endocrine and genetic studies.

Conflict of interest statement

None declared.

Acknowledgments

We thank Dr. Gregory Lawson for extensive pathology consultation, and both the members of the Herschman laboratory and Dr. Sotiris Tetradis for helpful discussions. These studies were supported by National Cancer Institute Award P50 (grant number CA 86306 to HRH). Statistical analyses were supported by National Institutes of Health/National Center for Advancing Translational Science (NCATS) UCLA CTSI (Grant Number UL1TR000124).

Cherukuri Durga P., Ishikawa Tomo-o, Chun Patrick, Catapang Art, Elashoff David, Grogan Tristan R., Bugni James and Herschman Harvey R., (2014), Targeted Cox2 gene deletion in intestinal epithelial cells decreases tumorigenesis in female, but not male, Apc Min/+ mice, Molecular Oncology, 8, doi: 10.1016/j.molonc.2013.10.009.

References

  1. Backlund, M.G. , Mann, J.R. , Holla, V.R. , Buchanan, F.G. , Tai, H.H. , Musiek, E.S. , Milne, G.L. , Katkuri, S. , DuBois, R.N. , 2005. 15-hydroxyprostaglandin dehydrogenase is down-regulated in colorectal cancer. J. Biol. Chem.. 280, 3217–3223. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Brosens, L.A. , Keller, J.J. , Pohjola, L. , Haglund, C. , Morsink, F.H. , Lacobuzio-Donahue, C. , Goggins, M. , Giardiello, F.M. , Ristimaki, A. , Offerhaus, G.J. , 2008. Increased expression of cytoplasmic HuR in familial adenomatous polyposis. Cancer Biol. Ther.. 7, 424–427. [DOI] [PubMed] [Google Scholar]
  3. Chillingworth, N.L. , Morham, S.G. , Donaldson, L.F. , 2006. Sex differences in inflammation and inflammatory pain in cyclooxygenase-deficient mice. Am. J. Physiol. Regul. Integr. Comp. Physiol.. 291, R327–R334. [DOI] [PubMed] [Google Scholar]
  4. Chulada, P.C. , Thompson, M.B. , Mahler, J.F. , Doyle, C.M. , Gaul, B.W. , Lee, C. , Tiano, H.F. , Morham, S.G. , Smithies, O. , Langenbach, R. , 2000. Genetic disruption of Ptgs-1, as well as Ptgs-2, reduces intestinal tumorigenesis in Min mice. Cancer Res.. 60, 4705–4708. [PubMed] [Google Scholar]
  5. Eberhart, C.E. , Coffey, R.J. , Radhika, A. , Giardiello, F.M. , Ferrenbach, S. , DuBois, R.N. , 1994. Up-regulation of cyclooxygenase-2 expression in human colorectal adenomas and adenocarcinomas. Gastroenterology. 107, 1183–1188. [DOI] [PubMed] [Google Scholar]
  6. Egan, K.M. , Lawson, J.A. , Fries, S. , Koller, B. , Rader, D.J. , Smyth, E.M. , Fitzgerald, G.A. , 2004. COX2-derived prostacyclin confers atheroprotection on female mice. Science. 306, 1954–1957. [DOI] [PubMed] [Google Scholar]
  7. Giardiello, F.M. , Hamilton, S.R. , Krush, A.J. , Piantadosi, S. , Hylind, L.M. , Celano, P. , Booker, S.V. , Robinson, C.R. , Offerhaus, G.J. , 1993. Treatment of colonic and rectal adenomas with sulindac in familial adenomatous polyposis. N. Engl. J. Med.. 328, 1313–1316. [DOI] [PubMed] [Google Scholar]
  8. Gupta, R.A. , DuBois, R.N. , 2001. Colorectal cancer prevention and treatment by inhibition of cyclooxygenase-2. Nat. Rev.. 1, 11–21. [DOI] [PubMed] [Google Scholar]
  9. Higuchi, T. , Iwama, T. , Yoshinaga, K. , Toyooka, M. , Taketo, M.M. , Sugihara, K.A. , 2003. Randomized, double-blind, placebo-controlled trial of the effects of rofecoxib, a selective cyclooxygenase-2 inhibitor, on rectal polyps in familial adenomatous polyposis patients. Clin. Cancer Res.. 9, 4756–4760. [PubMed] [Google Scholar]
  10. Hull, M.A. , Booth, J.K. , Tisbury, A. , Scott, N. , Bonifer, C. , Markham, A.F. , Coletta, P.L. , 1999. Cyclooxygenase 2 is up-regulated and localized to macrophages in the intestine of Min mice. Br. J. Cancer. 79, 1399–1405. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Ishikawa, T.O. , Herschman, H.R. , 2006. Conditional knockout mouse for tissue-specific disruption of the cyclooxygenase-2 (Cox-2) gene. Genesis. 44, 143–149. [DOI] [PubMed] [Google Scholar]
  12. Ishikawa, T.O. , Herschman, H.R. , 2010. Tumor formation in a mouse model of colitis-associated colon cancer does not require COX-1 or COX-2 expression. Carcinogenesis. 34, 729–736. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Ishikawa, T.O. , Oshima, M. , Herschman, H.R. , 2011. Cox2 deletion in myeloid and endothelial cells, but not in epithelial cells, exacerbates murine colitis. Carcinogenesis. 32, 417–426. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Jacoby, R.F. , Seibert, K. , Cole, C.E. , Kelloff, G. , Lubet, R.A. , 2000. The cyclooxygenase-2 inhibitor celecoxib is a potent preventive and therapeutic agent in the Min mouse model of adenomatous polyposis. Cancer Res.. 60, 5040–5044. [PubMed] [Google Scholar]
  15. Jungck, M. , Grunhage, F. , Spengler, U. , Dernac, A. , Mathiak, M. , Caspari, R. , Friedl, W. , Sauerbruch, T. , 2004. E-cadherin expression is homogeneously reduced in adenoma from patients with familial adenomatous polyposis: an immunohistochemical study of E-cadherin, beta-catenin and cyclooxygenase-2 expression. Int. J. Colorectal Dis.. 19, 438–445. [DOI] [PubMed] [Google Scholar]
  16. Lynch, P.M. , Ayers, G.D. , Hawk, E. , Richmond, E. , Eagle, C. , Woloj, M. , Church, J. , Hasson, H. , Patterson, S. , Half, E. , Burke, C.A. , 2010. The safety and efficacy of celecoxib in children with familial adenomatous polyposis. Am. J. Gastroenterol.. 105, 1437–1443. [DOI] [PubMed] [Google Scholar]
  17. Maekawa, M. , Sugano, K. , Sano, H. , Miyazaki, S. , Ushiama, M. , Fujita, S. , Gotoda, T. , Yokota, T. , Ohkura, H. , Kakizoe, T. , Sekiya, T. , 1998. Increased expression of cyclooxygenase-2 to -1 in human colorectal cancers and adenomas, but not in hyperplastic polyps. Jpn. J. Clin. Oncol.. 28, 421–426. [DOI] [PubMed] [Google Scholar]
  18. McAlpine, C.A. , Barak, Y. , Matise, I. , Cormier, R.T. , 2006. Intestinal-specific PPARγ deficiency enhances tumorigenesis in Apc Min/+ mice. Int. J. Cancer. 119, 2339–2346. [DOI] [PubMed] [Google Scholar]
  19. McCart, A.E. , Vickaryous, N.K. , Silver, A. , 2008. Apc mice: models, modifiers and mutants. Pathol. Res. Pract.. 204, 479–490. [DOI] [PubMed] [Google Scholar]
  20. Moser, A.R. , Pitot, H.C. , Dove, W.F. , 1990. A dominant mutation that predisposes to multiple intestinal neoplasia in the mouse. Science. 247, 322–324. [DOI] [PubMed] [Google Scholar]
  21. Naor, R. , Domankevich, V. , Shemer, S. , Sominsky, L. , Rosenne, E. , Levi, B. , Ben-Eliyahu, S. , 2009. Metastatic-promoting effects of LPS: sexual dimorphism and mediation by catecholamines and prostaglandins. Brain Behav. Immun.. 23, 611–621. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Oshima, M. , Dinchuk, J.E. , Kargman, S.L. , Oshima, H. , Hancock, B. , Kwong, E. , Trzaskos, J.M. , Evans, J.F. , Taketo, M.M. , 1996. Suppression of intestinal polyposis in Apc Δ716 knockout mice by inhibition of cyclooxygenase 2 (COX2). Cell. 87, 803–809. [DOI] [PubMed] [Google Scholar]
  23. Oshima, M. , Murai, N. , Kargman, S. , Arguello, M. , Luk, P. , Kwong, E. , Taketo, M.M. , Evans, J.F. , 2001. Chemoprevention of intestinal polyposis in the Apcdelta 716 mouse by rofecoxib, a selective cyclooxygenase-2 inhibitor. Cancer Res.. 61, 1733–1740. [PubMed] [Google Scholar]
  24. Seno, H. , Oshima, M. , Ishikawa, T.O. , Oshima, H. , Takaku, K. , Chiba, T. , Narumiya, S. , Taketo, M.M. , 2002. Cyclooxygenase 2- and prostaglandin E(2) receptor EP(2)-dependent angiogenesis in Apc(Delta 716) mouse intestinal polyps. Cancer Res.. 62, 506–511. [PubMed] [Google Scholar]
  25. Sonoshita, M. , Takaku, K. , Oshima, M. , Sugihara, K.I. , Taketo, M.M. , 2002. Cyclooxygenase-2 expression in fibroblasts and endothelial cells of intestinal polyps. Cancer Res.. 62, 6846–6849. [PubMed] [Google Scholar]
  26. Steinbach, G. , Lynch, P.M. , Phillips, R.K. , Wallace, M.H. , Hawk, E. , Gordon, G.B. , Wakabayashi, N. , Saunders, B. , Shen, Y. , Fujimura, T. , Su, L.K. , Levin, B. , Godio, L. , Patterson, S. , Rodriguez-Bigas, M.A. , Jester, S.L. , King, K.L. , Schumacher, M. , Abbruzzese, J. , DuBois, R.N. , Hittelman, W.N. , Zimmerman, S. , Sherman, J.W. , Kelloff, G. , 2000. The effect of celecoxib, a cyclooxygenase-2 inhibitor, in familial adenomatous polyposis. N. Engl. J. Med.. 342, 1946–1952. [DOI] [PubMed] [Google Scholar]
  27. Williams, C.S. , Luongo, C. , Radhika, A. , Zhang, T. , Lamps, L.W. , Nanney, L.B. , Beauchamp, R.D. , DuBois, R.N. , 1996. Elevated cyclooxygenase-2 levels in Min mouse adenomas. Gastroenterology. 111, 1134–1140. [DOI] [PubMed] [Google Scholar]
  28. Yoo, C.B. , Chuang, J.C. , Bun, H.M. , Egger, G. , Yang, A.S. , Dubeau, L. , Long, T. , Laird, P.W. , Marquez, V.E. , Jones, P.A. , 2008. Long-term epigenetic therapy with oral zebularine has minimal side effects and prevents intestinal tumors in mice. Cancer Prev. Res.. 1, 233–240. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Molecular Oncology are provided here courtesy of Wiley

RESOURCES