Skip to main content
NCBI home page
Carcinogenesis logoLink to Carcinogenesis
. 2019 Mar 12;40(11):1376–1386. doi: 10.1093/carcin/bgz050

A novel mouse model of sporadic colon cancer induced by combination of conditional Apc genes and chemical carcinogen in the absence of Cre recombinase

Jeffrey S Souris 1,#,✉, Hannah J Zhang 1,#, Urszula Dougherty 2, Nai-Tzu Chen 3, Joseph V Waller 1, Leu-Wei Lo 1,4, John Hart 5, Chin-Tu Chen 1, Marc Bissonnette 2
PMCID: PMC6875902  PMID: 30859181

Abstract

Although valuable insights into colon cancer biology have been garnered from human colon cancer cell lines and primary colonic tissues, and animal studies using human colon cancer xenografts, immunocompetent mouse models of spontaneous or chemically induced colon cancer better phenocopy human disease. As most sporadic human colon tumors present adenomatous polyposis coli (APC) gene mutations, considerable effort has gone into developing mice that express mutant Apc alleles that mimic human colon cancer pathogenesis. A serious limitation of many of these Apc-mutant murine models, however, is that these mice develop numerous tumors in the small intestine but few, if any, in the colon. In this work, we examined three spontaneous mouse models of colon tumorigenesis based upon the widely used multiple intestinal neoplasia (Min) mouse: mice with either constitutive or conditional Apc mutations alone or in combination with caudal-related homeobox transcription factor CDX2P-Cre transgene — either with or without exposure to the potent colon carcinogen azoxymethane. Using the CDX2 promoter to drive Cre recombinase transgene expression effectively inactivated Apc in colonocytes, creating a model with earlier tumor onset and increased tumor incidence/burden, but without the Min mouse model’s small intestine tumorigenesis and susceptibility to intestinal perforation/ulceration/hemorrhage. Most significantly, azoxymethane-treated mice with conditional Apc expression, but absent the Cre recombinase gene, demonstrated nearly 50% tumor incidence with two or more large colon tumors per mouse of human-like histology, but no small intestine tumors — unlike the azoxymethane-resistant C57BL/6J-background Min mouse model. As such this model provides a robust platform for chemoprevention studies.

Conditional Apc gene expression and treatment with the carcinogen azoxymethane, in the absence of Cre recombinase, provides a novel mouse model of colon carcinogenesis suitable for tumor promotion and chemoprevention studies, with tumor development restricted to the colon.

Introduction

Colorectal cancer (CRC) is the third leading cause of cancer-related death in the USA and thought to arise from the acquisition of multiple mutations of genes controlling colon epithelial cell growth and differentiation that have accumulated over several decades (1–3). Among these genes, mutation of the adenomatous polyposis coli (APC) tumor suppressor gene has proven the most frequent mutation in familial and sporadic colon cancers, the latter of which is present in 75–80% of all CRCs (3–6). Germline carriers of APC mutations are especially susceptible to early CRC onset, with a nearly 100% risk of developing CRC by age 40 years (4,5). The APC transcript encodes a multidomain protein, and plays a number of critical roles in cell adhesion, migration, apoptosis and chromosomal segregation at mitosis, often through its association with other proteins (3). The primary tumor-suppressing function of APC, however, resides in its downstream suppression of canonical Wnt signals (7). This pathway induces nuclear translocation of transcriptional cofactor β-catenin and is tightly regulated to preclude excessive nuclear accumulation of β-catenin and its subsequent association with other gene activating/suppressing nuclear proteins (7,8). In healthy individuals, cytoplasmic APC — along with AXIN, GSK-3β, EB1 and other proteins — forms a complex that binds β-catenin to promote its phosphorylation and subsequent ubiquitin-mediated proteasomal degradation (9). Loss or mutation of both APC genes results in nuclear β-catenin accumulation and upregulation of numerous transcriptional targets including proto-oncogenes c-Myc, Cdk-4 and cyclin D1 (8).

To study the role of APC in CRC onset and tumorigenesis, a number of inactivating mutant alleles of the mouse Apc gene have been generated to date (10,11). A few germline mutant Apc alleles induce gastrointestinal tumorigenesis in the heterozygous state thereby mimicking, to a limited extent, the human tumor phenotype in that they are at least partially non-immunogenic and arise spontaneously in immunocompetent hosts (12). However, quite unlike human CRC, these heterozygous Apc-mutant mice generally develop significant numbers of adenomas in the small intestine and very few, if any, tumors in the large intestine (11). Tumor multiplicity of these strains is dependent upon the genomic location of the mutation, ranging from ~3 tumors per mouse for truncations at codon 1638 (Apc1638N) to >300 tumors per mouse for truncations at codon 716 (ApcΔ716), with almost none ever progressing to invasive adenocarcinomas, possibly a consequence of their host’s relatively short life span (13). Truncation of the Apc gene at codon 850 (Apcmin/+ or Min Mouse, for multiple intestinal neoplasia) leads to the development of up to 100 polyps in the small intestine in addition a few colon tumors (14–16). Use of the Apcmin/+ model greatly expanded with the discovery that administration of the potent colon-specific carcinogen azoxymethane (AOM), with or without the inflammation inducing agent dextran sodium sulfate (DSS), greatly enhanced colonic tumorigenesis (17,18). But the chemically augmented Apcmin/+ model’s dominant small bowel phenotype, short life span and severe colon inflammation in the case of DSS augmentation, severely compromises these models as mimics of sporadic human colon cancer (15,19).

To overcome the limitation of small bowel tumor phenotype, mice with conditional Apc mutations have been created and bred with transgenic mice that express Cre recombinase (Cre) driven by promoters of genes highly expressed in the colon, such as the caudal-related homeobox transcription factor gene CDX2 (20–25). As earlier studies have shown that Apcmin/+ mice given AOM develop additional colon tumors, we postulated that AOM treatment of conditional Apc+/LoxP mice might also promote colon tumors with minimal small intestine tumorigenesis — even in the absence of Cre-mediated recombination (15,18). To test this hypothesis, we decided to compare the time of onset, incidence, multiplicity and total burden of colon tumors of mice possessing germline Apc mutations with mice expressing conditional alleles of Apc. Tumorigenesis was induced by Apc mutation and/or AOM treatment. Quantitative real-time PCR (qRT-PCR) and immunohistochemical analyses for select tumor-associated glycoproteins were also performed on harvested murine tissues at various time-points, as an initial foray into correlating each model’s observed tumorigenic profile with clinically relevant, prognostic CRC biomarkers.

Materials and methods

Animal models

All animal studies were conducted with the approval of the University of Chicago’s Institutional Animal Care and Use Committee (ACUP71350), in compliance with National Institutes of Health guidelines for the care and use of laboratory animals. Wild-type C57BL/6J (Stock No: 000664), Apcmin/+ (on C57BL/6J background; Stock No: 002020), and transgenic mice expressing Cre recombinase controlled by CDX2 promoter (CDX2P-Cre, on C67BL/6J background; Stock No: 009350), were obtained from Jackson Laboratories (Bar Harbor, ME). Transgenic mice with Apc LoxP-exon 14-LoxP (Apc-CKO) were obtained from the National Cancer Institute (strain 01XAA, on C57BL/6J background). All mice were bred and maintained in-house at 37°C and 65–70% relative humidity, on a regulated 12-h day/night cycle. Wild-type C57BL/6J mice served as controls for Apcmin/+ mice, whereas CDX2P-Cre mice and Apc-CKO mice were interbred to generate CDX2P-Cre-Apc+/LoxP mice. Genotypes were verified using PCR protocols recommended by Jackson Laboratory or NCI with wild-type littermates serving as controls (Apc+/LoxP). Animals were divided into multiple cohorts: one cohort of Apcmin/+ mice received intraperitoneal injections of AOM (7.5 mg/kg) at 10 and then 12 weeks of age, whereas select cohorts of CDX2P-Cre-Apc+/LoxP mice and their Apc+/LoxP counterparts received only one injection of AOM (7.5 mg/kg) at 10 weeks of age, because of their anticipated more aggressive tumor development. Two weeks after administration of AOM, Apcmin/+ mice were placed on a high 20% fat diet (modified Harlan Teklad Diet 06265, Madison, WI), to mimic Western dietary conditions that have been shown to increase tumor development in Apcmin/+ mice (26). All other mice were maintained on a standard 5% fat rodent diet (modified Harlan Teklad Diet 06266, Madison, WI).

Colonoscopy and tissue harvesting

Tumor onset and progression were monitored by white light colonoscopy (ColoView 64301AA; Karl Storz Endoskope, Goleta, GA). Prior to colonoscopy, animals were fasted overnight. Initial colonoscopies were performed at 8 weeks following the first AOM administration and then repeated at 4 week intervals. In operation, the 1.9-mm OD Hopkins telescope was held within a 2.3-mm OD exam sheath whose working channels permitted both air insufflation and saline/acetylcysteine irrigation, to enable optimal colon wall visualization and removal of residual feces/mucus, respectively. The entire left colon extending to the splenic flexure was thus visualized and imaged. When tumors of sufficient size were detected at colonoscopy, animals were euthanized and colons isolated. Harvested colons were then opened, longitudinally along the cephalon–caudal axis. The lumenal surface was inspected to record the size and distribution of polyps within the organ. Visible tumors were excised and remaining colon was either Swiss-rolled or sectioned in a bread loaf manner and flash frozen in liquid nitrogen or embedded in Optimal Cutting Temperature compound on dry ice. Samples were stored at −80°C for subsequent analyses.

Histology and immunohistochemistry

Hematoxylin and eosin (H&E) and immunostaining were performed on cryostat frozen sections prepared at the Human Tissue Resources Center at the University of Chicago. Tumors were graded by a clinical pathologist who was unaware of treatments or genotypes. For immunohistochemistry staining, slides were fixed in 4% paraformaldehyde followed by antigen retrieval at 92oC for 5 min. Endogenous peroxide activity was blocked by H2O2 treatment. Rabbit anti-β-catenin (1:1000) and anti-Ki67 [1:200 (Thermo Fisher, Waltham, MA)] as well as anti-MUC1 [1:500 (GeneTex, San Antonio, TX)] were used as primary antibodies followed by biotinylated donkey anti-rabbit IgG (Thermo Fisher) as secondary antibodies. Sav-HRP and DAB substrate (Abcam, Cambridge, UK) were used for color development. Slides were counterstained with hematoxylin. For negative controls, primary antibodies were replaced with normal serum. Negative control sections showed no specific staining.

Real-time PCR

For qRT-PCR analyses, total RNA was isolated from harvested tumors, as well as their visually normal adjacent and distal tissues, using a PureLink RNA Mini Kit (Invitrogen, Carlsbad, CA) and then reverse transcribed using a High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA), following the manufacturer’s instructions. qRT-PCR was performed in triplicate for each sample as described previously (27), using primers specific to mouse Mucin-1 (MUC1), Mucin-15 (MUC15) and epidermal growth factor-like repeats and discoidin I-like domains 3 (EDIL3) complementary DNA. The obtained 3 CT values were averaged and then normalized to β-actin which served as an endogenous control. Expression levels of messenger RNA (mRNA) were then expressed as folds-changed over corresponding controls using the comparative 2(−ΔΔCT) method. Primers were purchased from Integrated DNA Technologies (Coralville, IA) that comprised probes MUC1 (Mm.PT.58.31827570), MUC15 (Mm.PT.58.41389990), EDIL3 (Mm.PT.58.30149877) and BACT (Mn.PT.58.33257376.gs). PCR amplification was verified by melting curve.

Data analysis

Statistical analyses were performed using Prism software (GraphPad, La Jolla, CA). Continuous data were summarized as means ± standard error of the mean, with statistically significant differences between two groups being assessed using an unpaired Student’s t-test, and differences involving three or more groups determined by one-way analysis of variance using a post hoc test. Log-rank tests were conducted to determine the significance of tumor onset. P < 0.05 was considered statistically significant. Colon tumor incidence was defined as the percentage of mice bearing at least one tumor whereas tumor multiplicity was defined as the average number of tumors per mouse. Tumor multiplicities were compared among groups using analysis of variance with the Holm–Sidak post hoc test, with P-values <0.05 considered to be statistically significant. A total of seven cohorts of mice were studied in these experiments: C57BL/6J (n = 17), Apcmin/+ (n = 20), Apcmin/+ + AOM (n = 25), Apc+/LoxP (n = 17), CDX2P-Cre-Apc+/LoxP (n = 27), Apc+/LoxP + AOM (n = 20) and CDX2P-Cre-Apc+/LoxP + AOM (n = 17).

Results

White light colonoscopies were performed at regular intervals on all cohorts, to detect the emergence and progression of colorectal polyps, and monitor tumor adjacent colonic tissue for early endoscopic signs of transformation (Figure 1A). CDX2 promoter-targeted Cre-mediated Apc inactivation resulted in both increased tumor incidence and earlier onset in the murine models studied. CDX2P-Cre-Apc+/LoxP animals treated with AOM, in particular, showed the earliest colon tumor onset (Figure 1B). On average, tumors did not appear until 120 days of age in CDX2P-Cre-Apc+/LoxP + AOM mice, compared to 139 days of age in Apcmin/+ mice (Figure 1C). A similar trend was observed in the median number of days until tumor onset between the three experimental groups: 138 days for Apcmin/+ + AOM mice, 129 days for CDX2P-Cre-Apc+/LoxP mice and 119.5 days for CDX2P-Cre-Apc+/LoxP + AOM mice (Figure 1D).

https://academic.

https://academic

Open in a new tab

CDX2 promoter-driven, Cre-targeted Apc inactivation leads to earlier tumor onset. (A) Representative colonscopic images of control and tumor-bearing murine colons in which C57BL/6J mice served as controls for Apcmin/+ animals, whereas wild-type littermates, determined by genotyping, served as the controls for the other respective groups. (B) Kaplan–Meier curve showing onset of detectable tumors of Apcmin/+ + AOM (red), CDX2P-Cre-Apc+/LoxP (green) and CDX2P-Cre-Apc+/LoxP + AOM (blue) mice. Log-rank analyses revealed that CDX2 promoter-driven, Cre-targeted Apc inactivation enhances early tumor development in colon, with additional acceleration afforded by AOM treatment (P < 0.05). (C) Mean animal age of tumor onset for CDX2P-Cre-Apc+/LoxP + AOM mice, CDX2P-Cre-Apc+/LoxP mice and Apcmin/+ + AOM mice showing earlier onset with conditional Apc inactivation, accelerated by AOM administration — though only differences in tumor onset age between CDX2P-Cre-Apc+/LoxP + AOM and Apcmin/+ + AOM age groups were statistically significant (*P < 0.05). (D) Similar trend mirrored in the median tumor onset ages of the three respective cohorts studied.

As shown by the representative images of Figure 2A, each harvested colon was longitudinally excised, reflected to display its luminal surface and the epithelium inspected and photographed at the time of harvest, to grossly characterize each animal’s tumorigenesis. Although small intestinal tumors were observed in Apcmin/+ + AOM mice (data not shown), these mice demonstrated a significant increase in colon tumor incidence relative to Apcmin/+ mice that did not receive AOM; with an average of 84.0% of Apcmin/+ + AOM mice possessing tumors at corresponding time-points versus 50.0% of Apcmin/+ controls, as shown in Figure 2B. CDX2P-Cre-Apc+/LoxP mice demonstrated the highest rate of tumor incidence of all models studied, with 96.3% of animals possessing colon tumors at corresponding time-points (Figures 2B). In comparison to Apcmin/+ controls, these CDX2P-Cre-Apc+/LoxP mice presented with higher colon tumor incidence (Figure 2C), but no visually detectable tumors within the small intestine (data not shown). AOM treatment of CDX2P-Cre-Apc+/LoxP animals resulted in a statistically significant increase in colon tumor incidence relative to their littermate wild-type controls (Apc+/LoxP + AOM), as shown in Figure 2B. More strikingly, AOM treatment of Apc+/LoxP mice resulted in ~45% of these mice developing colon tumors, but no tumor formation in Apc+/LoxP controls that did not receive AOM (Figure 2D).

Figure 2.

Figure 2.

Open in a new tab

CDX2 promoter-driven, Cre-targeted Apc inactivation leads to increased colon tumor incidence. (A) Harvested colons of experimental and control groups, longitudinally opened and reflected to reveal their lumenal surfaces. (B) Percentage of Apcmin/+ + AOM, CDX2P-Cre-Apc+/LoxP and CDX2P-Cre-Apc+/LoxP + AOM mice with visually detectable colon tumors revealing higher incidence rates than their respective controls (*P < 0.05, **P < 0.01, ****P < 0.0001). (C) Comparison of tumor incidence between mice with CDX2 promoter-driven, Cre-targeted Apc inactivation and Apcmin/+ controls revealed a significant increase in tumorigenesis in CDX2P-Cre-Apc+/LoxP mice (***P < 0.001). (D) Especially noteworthy is the effect that AOM had on conditional Apc+/LoxP animals in the absence of Cre recombinase, with no detectable tumors occurring in the absence of AOM (****P < 0.0001).

Tumor multiplicities (number of tumors per mouse) were also compared among Apcmin/+ + AOM, CDX2P-Cre-Apc+/LoxP and CDX2P-Cre-Apc+/LoxP + AOM groups. Although CDX2P-Cre-Apc+/LoxP + AOM mice showed the highest tumor multiplicity (4–5 tumors per mouse) whereas Apcmin/+ + AOM mice had the lowest tumor multiplicity (2–3 tumors per mouse), differences among the three groups were not statistically significant (Supplementary Figure 1, available at Carcinogenesis Online). Nevertheless, when compared to their respective controls, all three groups demonstrated significantly greater tumor multiplicities, as shown in Figure 3A, whereas Apcmin/+ mice not exposed to AOM averaged one tumor per mouse (data not shown). In addition, when animal age at the time of tissue harvest was considered, CDX2P-Cre-Apc+/LoxP animals older than 15 weeks presented with significantly greater numbers of tumors than their younger cohorts, suggesting that postnatal tumor initiation is extended over time. Surprisingly, however, this time dependence of tumor burden was not observed in the CDX2P-Cre-Apc+/LoxP mice treated with AOM. Instead, a trend of increasing tumor number (per animal) was observed for mice younger than 15 weeks in this group (Figure 3B). To assess overall tumor burden, we compared tumor multiplicities among groups based upon tumor volume: tumors <3 mm3, tumors 3–10 mm3 and tumors >10 mm3. Older animals (>20 weeks of age) presented with significantly greater numbers of large tumors (>10 mm3) in both CDX2P-Cre-Apc+/LoxP mice and CDX2P-Cre-Apc+/LoxP + AOM mice, as shown in Figure 3C. For the AOM-treated CDX2P-Cre-Apc+/LoxP group, older mice also tended to have tumors approximately twice the volume of those in younger cohorts, as shown in Figure 3D.

Figure 3.

Figure 3.

Open in a new tab

CDX2 promoter-driven, Cre-targeted Apc inactivation enhances tumor multiplicity and total tumor burden (volume). (A) Average number of tumors was significantly higher in germline Apcmin/+ or conditional Apc+/LoxP mice treated with or without AOM, compared to their corresponding controls (*P < 0.05, ***P < 0.001). Apcmin/+ mice not given AOM presented with 1.25 tumors per mouse (data not shown). (B) Older animals (>15 weeks) developed significantly more tumors without AOM treatment whereas AOM treatment accelerated early tumor development (*P < 0.05). (C) Older animals (>20 weeks) also developed significantly more tumors that were >10 mm3, when compared to younger cohorts (*P < 0.05). (D) Average tumor volume was significantly greater in AOM treated older animals (>20 weeks) with CDX2 promoter-driven, Cre-targeted Apc inactivation than in younger counterparts (*P < 0.05).

To assess similarities and differences between the colon tumors that arose in these mouse models and those characteristic of human CRC, we conducted a series of histological and immunohistochemical analyses using murine colon tissues harvested immediately following colonoscopic verification pathology status. Representative hematoxylin and eosin–stained images of tumor sections are presented in Figure 4A. Colon tumor histologies ranged from well-differentiated adenomas to invasive adenocarcinomas. Histological inspection of specimens by a clinical pathologist revealed no readily apparent differences between comparably staged murine and human colon tumors for any of the animal groups studied. Mouse adenocarcinomas exhibited all of the most dominant features of human colorectal adenocarcinomas, with increased nuclear-to-cytoplasmic ratios, increased heterochromatin and back-to-back gland formation, and tumor invasion into the muscularis mucosa. No differences in Ki67 staining or β-catenin expression were observed between animal groups with regards to age or treatment conditions. Similarly, Ki67 and β-catenin expression levels were consistent between animal groups of comparable disease status and found to be closely correlated with tumor progression, as shown in Figure 4B and C. For Ki67, normal colon tissues showed low levels of expression, whereas moderate Ki67 staining was observed in adenomas and strong Ki67 staining noted in adenocarcinomas — a trend mirrored in the staining for β-catenin expression.

Figure 4.

Figure 4.

Open in a new tab

Histological and immunohistochemical analyses of observed colon carcinogenesis. (A) Representative images of hematoxylin and eosin (H&E)–stained tumors showing tumorigenesis (A: left to right) showing normal colon tissue, adenoma and adenocarcinoma, respectively. (B: left to right) Corresponding representative immunohistochemical stainings of the proliferation biomarker Ki67. (C: left to right) Corresponding representative immunohistochemical stainings of the tumor biomarker β-catenin. All images were acquired at ×200 magnification.

We also sought to assess similarities and differences in the expression of prognostic biomarkers of tumors that arose in these murine models relative to human CRC, specifically MUC1, MUC15 and EDIL3, as upregulation of these moieties has previously been shown to correlate with stage, metastatic potential and invasiveness in a variety of cancers, including CRC, and are under either clinical or preclinical investigation as targets for chemotherapy and/or immunotherapy (28–37). To this end, and for correlation with each model’s tumorigenesis profile, we measured the transcript expression levels for each of these moieties in harvested tissues. These studies revealed, somewhat surprisingly, that MUC15 and EDIL3 mRNA expression levels did not vary appreciably in colon tumors when compared to their genotype-matched controls (Supplementary Figure 2, available at Carcinogenesis Online). MUC1 mRNA expression, on the other hand, was significantly upregulated in colonic tumors. Comparison with genotype-matched controls revealed that the colon tumors of Apcmin/+ + AOM mice (Figure 5A), CDX2P-Cre-Apc+/LoxP mice (Figure 5B) and CDX2P-Cre-Apc+/LoxP + AOM mice (Figure 5C) all demonstrated significant upregulation of MUC1 mRNA expression. Interestingly, different patterns of MUC1 mRNA overexpression were observed in the grossly normal colon tissues collected both adjacent and distal to tumors, especially in mice administered AOM. MUC1 mRNA expression was significantly higher in these normal-appearing colonic tissues of Apcmin/+ mice treated with AOM compared to the normal colonic tissue of C57BL/6J controls, as shown in Figure 5A. In contrast, MUC1 transcript expression levels were similar between normal colon tissue of Apc+/LoxP controls and normal-appearing colon tissue in CDX2P-Cre-Apc+/LoxP mice (Figure 5B). AOM treatment of CDX2P-Cre-Apc+/LoxP mice also appeared to increase MUC1 mRNA expression in the normal-appearing tissues adjacent and distal to tumors when compared to similar tissues harvested from AOM-treated Apc+/LoxP controls (Figure 5C), albeit no statistical significance was achieved. Upregulation of MUC1 protein expression in colon tumors was confirmed by immunohistochemical analyses using antibodies specific for MUC1 protein (Figure 5D), with MUC1 being predominantly distributed on the surface of colonocytes. Immunohistochemical staining confirmed MUC1 protein upregulation, and AOM-induced potentiation of MUC1 protein upregulation, in colon tumors and normal-appearing mucosa adjacent and distal to tumors in both Apc+/LoxP and CDX2P-Cre-Apc+/LoxP mice, relative to untreated (AOM-free) Apc+/LoxP controls as shown in Figure 5D–F.

Figure 5.

Figure 5.

Open in a new tab

MUC1 is notably upregulated in colon tumors of all three murine models examined. (A–C) Compared to genotype-matched control mucosa, the colon tumors of Apcmin/+ + AOM mice (A), CDX2P-Cre-Apc+/LoxP mice (B) and CDX2P-Cre-Apc+/LoxP + AOM mice (C) each revealed substantial upregulation of MUC1 mRNA expression, as determined by qRT-PCR (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001). AOM substantially increased MUC1 mRNA expression in the normal-appearing mucosa in Apcmin/+ mice compared to normal colonic mucosa. Although MUC1 mRNA levels were similar between control mucosa in Apc+/LoxP mice and normal-appearing mucosa in CDX2P-Cre-Apc+/LoxP mice, AOM administration to CDX2P-Cre-Apc+/LoxP mice appeared to increase MUC1 mRNA expression in grossly normal tissues when compared to similar tissues harvested from AOM treated Apc+/LoxP controls, but differences were not statistically significant (C). (D–H) Immunohistochemical stainings confirmed MUC1 protein upregulation, and AOM-induced potentiation of MUC1 protein upregulation, in colon tumors and normal-appearing mucosa in both Apc+/LoxP and CDX2P-Cre-Apc+/LoxP mice. All images were taken at ×200 magnification.

Discussion

Although great insights into the pathways driving CRC have been garnered from in vitro studies of human CRC cell lines and primary human colon cancer tissues, and in vivo animal studies of human colon cancer xenografts, immune-intact mouse models of spontaneous or chemically induced CRC better phenocopy human disease and offer more relevant guidance for deciphering the molecular mechanisms of CRC biology and for studying the chemopreventive efficacy of agents (38). As the overwhelming majority of sporadic human colorectal tumors arise from mutations in the APC gene, considerable effort has gone into developing mice bearing mutant Apc alleles that mimic human pathogenesis, such as the multiple intestinal neoplasia or Min Mouse (Apcmin/+). A serious limitation of this widely used murine model, however, is that heterozygous mice carrying the germline Apc mutation develop numerous tumors in the small intestine but few, if any, in the colon, whereas germline Apc mutant homozygotes are embryonically lethal.

In this study, we examined three spontaneous mouse models of colon tumorigenesis based upon mutation of the Apc gene: mice with either germline Apc mutation alone or in combination with CDX2P-Cre transgene derived Apc mutation, with or without AOM treatment. Depending upon the model studied we observed several histological types of malignant transformation — ranging from flat, poorly differentiated to well-differentiated adenomas to fully invasive adenocarcinomas — that reflect the spectrum of malignant transformation in human colonic epithelium. AOM treatment of Apcmin/+ mice significantly increased the percentage of animals developing colon tumors relative to their C57BL/6J and Apcmin/+ controls: with only 24% of C57BL/6J and 50% of Apcmin/+ mice bearing colonic tumors compared to 84% of Apcmin/+ + AOM mice. However, even with the induction of colon tumors via administration of a potent colon-specific carcinogen, AOM-augmented Apcmin/+ mice retained considerable small intestinal tumor burden that resulted in shortened life spans, as well as disease that failed to recapitulate many of the characteristics of human CRC pathogenesis.

In contrast to mice with germline Apcmin/+ mutations that develop numerous small intestinal tumors, mice with conditional mutation (inactivation) of Apc did not develop any detectable small intestine tumors, even in the distal small intestine. CDX2 is a hind gut homeobox gene highly expressed in the nuclei of colonic epithelial cells (39–43). Using the CDX2 promoter-driven Cre recombinase transgene system we effectively inactivated Apc in colonocytes, creating a colon cancer model without the complications that occur in Apcmin/+ mice; from relatively rare instances of sepsis arising from intestinal wall perforation/ulceration to very frequent early deaths arising from progressive anemia. AOM treatment of mice bearing CDX2P-Cre-mediated Apc inactivation proved an especially advantageous model with earlier tumor onset, higher tumor incidence and greater total tumor burden compared to the more widely used germline mutant Apcmin/+ model. Although tumor incidence showed an age dependence in CDX2P-Cre-Apc+/LoxP mice — with older mice possessing higher incidence rates — no such relationship between animal age and tumor incidence was found in CDX2P-Cre-Apc+/LoxP + AOM mice. Instead, a trend for increased tumor multiplicity (number of tumors per mouse) was observed for mice younger than 15 weeks of age, reflecting AOM’s acceleration of tumor development. As all mice received AOM at approximately the same age (either as a single injection at 10 weeks or 2 injections at 10 and 12 weeks) on a weight adjusted basis (7.5 mg/kg), we infer that the age dependencies observed simply reflect tumor growth curve dynamics and not an age-dependent susceptibility to AOM exposure.

Previous studies have shown the effectiveness of AOM in colorectal tumor initiation and promotion to be highly dependent upon genetics (strain background), age, dose, route of administration and diet. In one study, 3–5-month-old male and female mice of various strains (A/J., SWR/J, AKR/J, C57BL/6J, DBA/2J) were injected either intraperitoneally or subcutaneously with 5, 10 or 20 mg AOM per kg body weight for 2, 4 or 8 weeks (44). Although no sex-dependent differences were observed, increasing the number and thus total dose of AOM revealed a significant strain-dependent effect on tumor promotion (independent of susceptibility to tumor initiation), as did small dietary modifications of fat and protein intake. In another study of young (4–5 months) versus old (21–22 months) C57BL/6J mice, old mice had significantly more aberrant crypt foci (ACF) than young mice when given the same (12 or 15) mg/kg body weight dose of AOM. But when the dose was fixed (i.e. not body weight dependent), young mice demonstrated significantly higher ACF frequency than older mice at higher total AOM dose (1.8 and 2.2 mg), but similar ACF frequency at lower total AOM dose (1.5 mg) (45). And studies of Apcmin/+ pups and young adults (4–5 week) revealed 17-fold and 10-fold increases in the number of dysplastic ACF when compared to their vehicle-treated controls, respectively (15). Ongoing efforts in our laboratory are aimed at determining the significance these various endogenous/exogenous factors play in the AOM induction of tumorigenesis in the murine models presented in the current work.

In CRC a growing body of evidence suggests that the overexpression and aberrant glycosylation of normally occurring glycoproteins, and the expression of structurally altered glycoproteins, play significant roles in the malignant epithelial cell transformation of normal colorectal mucosa epithelia (37). Altered glycosylation and expression of membrane-bound glycoproteins like mucins have been found to modulate many cancer cellular properties, such as proliferation, adhesion, migration and invasion, and led to the development of novel transgenic mouse models and immunotherapies, such as MUC1 antigen-targeted peptide and non-peptide vaccines (46). As an initial foray into assessing the correlation between altered glycoprotein expression and the tumorigenesis profiles of our murine models, we conducted a series of qRT-PCR and immunohistochemical analyses of harvested tissues to characterize the expression of three very differently acting glycoproteins: MUC1, MUC15 and EDIL3 (28,37,47).

Overexpression of MUC1 has been shown to inhibit apoptosis and stimulate cell proliferation via growth factor, β-catenin and extracellular signal-regulated kinase signaling, whereas MUC1 glycosylation status appears to be correlated with immune recognition level and invasion/metastatic potential (31,48,49). Recent studies comparing mucin expression in colorectal mucosa, polyps and carcinomas have revealed MUC1 expression levels that correlate stepwise within polyp-adenoma-carcinoma progression; with no expression in normal mucosa, little expression in hyperplastic polyps, and >89% MUC1 positive cells in intramucosal carcinomas (50–53). Immunotherapeutic targeting of MUC1 expression in CRC has included the development of peptide and non-peptide based vaccines, as well as the derivation of bigenic murine models from crossing Apcmin/+ mice with human MUC1 transgenic mice (46).

Unlike most other mucins (including MUC1), the extracellular domain of MUC15 has no epidermal growth factor-like domains and tandem repeats. MUC15 is frequently overexpressed in colorectal adenocarcinomas, especially those undergoing rapid growth, and has been shown to enhance cell–extracellular matrix adhesion, colony formation and invasion (via extracellular signal-regulated kinase–mitogen-activated protein kinase signaling pathway activation) (28,29). By comparison, the tumor-associated glycoprotein EDIL3 differs from both MUC1 and MUC15 in that it contains three epidermal growth factor-like repeats and binds with αvβ3 integrin, potently affecting angiogenesis and vascular morphogenesis. High expression of EDIL3 has been correlated with especially poor prognoses in hepatocellular carcinoma and is overexpressed in breast, colon, pancreatic and bladder cancer (36,54–56). EDIL3 binding leads to focal contact formation, clustering of integrin receptors and the phosphorylation of signaling molecules such as p125FAK and mitogen-activated protein kinase (57).

In the current work, qRT-PCR studies revealed that MUC1 mRNA expression is significantly upregulated in colon tumors — especially those arising in Apcmin/+ mice treated with AOM, CDX2P-Cre-Apc+/LoxP mice and CDX2P-Cre-Apc+/LoxP mice treated with AOM — relative to genotype-matched controls. Indeed, MUC1 mRNA was found to be upregulated in all the colonic tumors examined and potentiated by AOM exposure, as well as in normal tissues of animals treated with AOM, albeit the latter to a much lesser degree; findings that were confirmed at the protein level by immunohistochemical analyses of sections from the same specimens and reflecting the field effect known to occur with AOM exposure. Somewhat surprisingly, however, neither MUC15 nor EDIL3 mRNA levels were found to vary appreciably between colon tumors and genotype-matched controls in any of the models tested, regardless of animal age or dysplasia stage — with levels of expression 2 or more orders of magnitude lower than MUC1. Immunohistochemical staining of these tissues for MUC15 and EDIL3 protein expression confirmed qRT-PCR measurements. Studies of EDIL3 autoantibodies in AOM- or DSS-only (dextran sodium sulfate) treated FVB/N mice have shown correlation with the presence of distal colon tumors, with animals exposed to the inflammatory agent DSS developing significantly higher levels of EDIL transcripts (35).

Perhaps the most interesting finding in the current studies was that of the almost binary-like susceptibility of tumorigenesis of Apc+/LoxP mice to AOM modulation. As anticipated, mice with conditional Apc expression, in the absence of a Cre recombinase, did not develop tumors and were thus used as negative controls. Surprisingly, however, after a single injection of AOM (7.5 mg/kg), 45% of Apc+/LoxP mice developed colonic tumors. We speculate that the conditional Apc+/LoxP construct predisposes animals to AOM-induced loss of function of the conditional Apc allele in the colon, and the remaining Apc+ wild-type allele is likely either silenced or mutated (e.g. via a mechanism involving loss of heterozygosity) allowing for the emergence of colonic adenomas. However, the exact nature of this predisposition remains under investigation as the presence of the LoxP sequence should not interfere with gene expression (i.e. Apc+/LoxP mice are essentially Apc wild-type mice until Cre recombination occurs).

It is important to point out that tumors generated in these mice possess the same histological features as those observed in other mouse models of human CRC. As such, this novel discovery provides an alternative colonic tumor model to those commonly used. As the tumor incidence in the AOM treated Apc+/LoxP mouse is close to 50%, with average multiplicity of 2 or more tumors per mouse, this model provides a robust platform to reliably detect tumor-promoting and chemopreventive effects of agents. The model also simplifies breeding strategies as no Cre recombinase gene is needed and only one conditional Apc allele is required. The AOM treatment protocol is similar to that used in AOM/DSS models of CRC tumorigenesis, but without the DSS-associated inflammation/colitis that does not mimic sporadic human colon cancer. Although AOM induces β-catenin mutations, AOM treatment of Apcmin/+ mouse has been shown to result in Apc loss of heterozygosity that phenocopies 85% of sporadic human colon cancers. Taken together, this conditional Apc model could have substantial and broad applications in the exploration of chemopreventive agents and screening for tumor promoting effects such as Western diet implicated in human CRC.

This model could also potentially be used to investigate the colon tumor phenotype of genes expressed by cells other than colonocytes (e.g. tumor stroma T/B-cell lymphocytes, macrophages, fibroblasts and dendritic cells); a challenging endeavor as conventional colonocyte-specific Cre would be required for Apc mutation/truncation, whereas a different tissue-specific Cre would be required for the targeted gene’s mutation/truncation in the non-colonocyte cell type (58). By crossing mice that express a conditional non-Apc gene (e.g., LoxP-NFκB, Stat3, Cox-2) with mice carrying a Cre gene expressed specifically in one of these non-colonocyte cell types, one could theoretically allow assembly of Cre-targeting non-colonocyte cells, conditional targeted genes of interest and Apc+/LoxP — the latter of which could be inactivated/mutated by colon-specific AOM (via induced genomic instability in the colonocytes), even on the frequently used C57BL/6J background, but without DSS-induced inflammation — though additional studies would be required to assure the fidelity of the single-Cre process.

Funding

National Institutes of Health (R01-CA171785 to J.S.S.); (R01-CA164124 to M.B.).

Conflict of Interest Statement: None declared.

Supplementary Material

bgz050_Suppl_Supplementary_Figures-1
Click here for additional data file. (160KB, png)
bgz050_Suppl_Supplementary_Figures-2
Click here for additional data file. (494.1KB, png)

Glossary

Abbreviations

AOM

azoxymethane

APC

adenomatous polyposis coli

CRC

colorectal cancer

DSS

dextran sodium sulfate

EDIL3

Epidermal growth factor-like repeats and discoidin I-like domains 3

mRNA

messenger RNA

MUC1

Mucin-1

MUC15

Mucin-15

qRT-PCR

quantitative real-time PCR

References

  • 1. American Cancer Society. (2017) Colorectal Cancer Facts & Figures 2017–2019. American Cancer Society, Atlanta, GA. [Google Scholar]
  • 2. Siegel R.L., et al. (2017) Colorectal cancer statistics, 2017. CA. Cancer J. Clin., 67, 177–193. [DOI] [PubMed] [Google Scholar]
  • 3. Markowitz S.D., et al. (2009) Molecular origins of cancer: molecular basis of colorectal cancer. N. Engl. J. Med., 361, 2449–2460. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. Fodde R., et al. (2001) APC, signal transduction and genetic instability in colorectal cancer. Nat. Rev. Cancer, 1, 55–67. [DOI] [PubMed] [Google Scholar]
  • 5. Kwong L.N., et al. (2009) APC and its modifiers in colon cancer. Adv. Exp. Med. Biol., 656, 85–106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6. Zhang B., et al. (2014) Proteogenomic characterization of human colon and rectal cancer. Nature, 513, 382–387. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7. Polakis P. (2000) Wnt signaling and cancer. Genes Dev., 14, 1837–1851. [PubMed] [Google Scholar]
  • 8. Sansom O.J., et al. (2004) Loss of Apc in vivo immediately perturbs Wnt signaling, differentiation, and migration. Genes Dev., 18, 1385–1390. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. Shao L., et al. (2013) A20 restricts wnt signaling in intestinal epithelial cells and suppresses colon carcinogenesis. PLoS One, 8, e62223. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Karim B.O., et al. (2013) Mouse models for colorectal cancer. Am. J. Cancer Res., 3, 240–250. [PMC free article] [PubMed] [Google Scholar]
  • 11. Johnson R.L., et al. (2013) Animal models of colorectal cancer. Cancer Metastasis Rev., 32, 39–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Rosenberg D.W., et al. (2009) Mouse models for the study of colon carcinogenesis. Carcinogenesis, 30, 183–196. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13. Aoki K., et al. (2003) Colonic polyposis caused by mTOR-mediated chromosomal instability in Apc+/Δ716 Cdx2+/- compound mutant mice. Nat. Genet., 35, 323–330. [DOI] [PubMed] [Google Scholar]
  • 14. Moser A.R., et al. (1990) A dominant mutation that predisposes to multiple intestinal neoplasia in the mouse. Science, 247, 322–324. [DOI] [PubMed] [Google Scholar]
  • 15. Paulsen J.E., et al. (2003) Age-dependent susceptibility to azoxymethane-induced and spontaneous tumorigenesis in the Min/+ mouse. Anticancer Res., 23, 259–265. [PubMed] [Google Scholar]
  • 16. Shoemaker A.R., et al. (1997) Studies of neoplasia in the Min mouse. Biochim. Biophys. Acta, 1332, F25–F48. [DOI] [PubMed] [Google Scholar]
  • 17. Kanneganti M., et al. (2011) Animal models of colitis-associated carcinogenesis. J. Biomed. Biotechnol., 2011, 342637. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18. Suzui M., et al. (2002) Enhanced colon carcinogenesis induced by azoxymethane in min mice occurs via a mechanism independent of beta-catenin mutation. Cancer Lett., 183, 31–41. [DOI] [PubMed] [Google Scholar]
  • 19. De Robertis M., et al. (2011) The AOM/DSS murine model for the study of colon carcinogenesis: from pathways to diagnosis and therapy studies. J. Carcinog., 10, 9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20. Robanus-Maandag E.C., et al. (2010) A new conditional Apc-mutant mouse model for colorectal cancer. Carcinogenesis, 31, 946–952. [DOI] [PubMed] [Google Scholar]
  • 21. Hinoi T., et al. (2007) Mouse model of colonic adenoma-carcinoma progression based on somatic Apc inactivation. Cancer Res., 67, 9721–9730. [DOI] [PubMed] [Google Scholar]
  • 22. Shibata H., et al. (1997) Rapid colorectal adenoma formation initiated by conditional targeting of the Apc gene. Science, 278, 120–123. [DOI] [PubMed] [Google Scholar]
  • 23. Gu H., et al. (1993) Independent control of immunoglobulin switch recombination at individual switch regions evidenced through Cre-loxP-mediated gene targeting. Cell, 73, 1155–1164. [DOI] [PubMed] [Google Scholar]
  • 24. Calon A., et al. (2007) Different effects of the Cdx1 and Cdx2 homeobox genes in a murine model of intestinal inflammation. Gut, 56, 1688–1695. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25. Chawengsaksophak K., et al. (1997) Homeosis and intestinal tumours in Cdx2 mutant mice. Nature, 386, 84–87. [DOI] [PubMed] [Google Scholar]
  • 26. Newmark H.L., et al. (2001) A Western-style diet induces benign and malignant neoplasms in the colon of normal C57Bl/6 mice. Carcinogenesis, 22, 1871–1875. [DOI] [PubMed] [Google Scholar]
  • 27. Dougherty U., et al. (2009) Epidermal growth factor receptor is required for colonic tumor promotion by dietary fat in the azoxymethane/dextran sulfate sodium model: roles of transforming growth factor-{alpha} and PTGS2. Clin. Cancer Res., 15, 6780–6789. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28. Byrd J.C., et al. (2004) Mucins and mucin binding proteins in colorectal cancer. Cancer Metastasis Rev., 23, 77–99. [DOI] [PubMed] [Google Scholar]
  • 29. Huang J., et al. (2009) Overexpression of MUC15 activates extracellular signal-regulated kinase 1 and kinase 2 and promotes the oncogenic potential of human colon cancer cells. Carcinogenesis, 30, 1452–1458. [DOI] [PubMed] [Google Scholar]
  • 30. Jiang S.H., et al. (2016) Overexpressed EDIL3 predicts poor prognosis and promotes anchorage-independent tumor growth in human pancreatic cancer. Oncotarget, 7, 4226–4240. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31. Khanh d.o.T., et al. (2013) Transmembrane mucin MUC1 overexpression and its association with CD10⁺ myeloid cells, transforming growth factor-β1 expression, and tumor budding grade in colorectal cancer. Cancer Sci., 104, 958–964. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32. Krishn S.R., et al. (2016) Mucins and associated glycan signatures in colon adenoma-carcinoma sequence: prospective pathological implication(s) for early diagnosis of colon cancer. Cancer Lett., 374, 304–314. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33. Zeng Y., et al. (2015) MUC1 predicts colorectal cancer metastasis: a systematic review and meta-analysis of case controlled studies. PLoS One, 10, e0138049. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34. Niv Y., et al. (2018) Mucin expression in colorectal cancer (CRC): systematic review and meta-analysis. J Clin Gastroenterol., 52,s437–444. [DOI] [PubMed] [Google Scholar]
  • 35. Barderas R., et al. (2013) Sporadic colon cancer murine models demonstrate the value of autoantibody detection for preclinical cancer diagnosis. Sci. Rep., 3, 2938. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36. Zou X., et al. (2009) Downregulation of developmentally regulated endothelial cell locus-1 inhibits the growth of colon cancer. J. Biomed. Sci., 16, 33. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37. Anandkumar A., et al. (2013) Tumour immunomodulation: mucins in resistance to initiation and maturation of immune response against tumours. Scand. J. Immunol., 78, 1–7. [DOI] [PubMed] [Google Scholar]
  • 38. Sebolt-Leopold J.S. (2018) Development of preclinical models to understand and treat colorectal cancer. Clin. Colon Rectal Surg., 31, 199–204. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39. da Costa L.T., et al. (1999) CDX2 is mutated in a colorectal cancer with normal APC/beta-catenin signaling. Oncogene, 18, 5010–5014. [DOI] [PubMed] [Google Scholar]
  • 40. Hryniuk A., et al. (2014) Cdx1 and Cdx2 function as tumor suppressors. J. Biol. Chem., 289, 33343–33354. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41. Olsen J., et al. (2014) The clinical perspectives of CDX2 expression in colorectal cancer: a qualitative systematic review. Surg. Oncol., 23, 167–176. [DOI] [PubMed] [Google Scholar]
  • 42. Bonhomme C., et al. (2003) The Cdx2 homeobox gene has a tumour suppressor function in the distal colon in addition to a homeotic role during gut development. Gut, 52, 1465–1471. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43. Saandi T., et al. (2013) Regulation of the tumor suppressor homeogene Cdx2 by HNF4α in intestinal cancer. Oncogene, 32, 3782–3788. [DOI] [PubMed] [Google Scholar]
  • 44. Bissahoyo A., et al. (2005) Azoxymethane is a genetic background-dependent colorectal tumor initiator and promoter in mice: effects of dose, route, and diet. Toxicol. Sci., 88, 340–345. [DOI] [PubMed] [Google Scholar]
  • 45. Chung H., et al. (2003) Effect of age on susceptibility to azoxymethane-induced colonic aberrant crypt foci formation in C57BL/6JNIA mice. J. Gerontol. A. Biol. Sci. Med. Sci., 58, B400–B405. [DOI] [PubMed] [Google Scholar]
  • 46. Akporiaye E.T., et al. (2007) Characterization of the MUC1.Tg/MIN transgenic mouse as a model for studying antigen-specific immunotherapy of adenomas. Vaccine, 25, 6965–6974. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47. Bhatia R., et al. (2019) Cancer-associated mucins: role in immune modulation and metastasis. Cancer Metastasis Rev. doi:10.1007/s10555-018-09775-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48. Niv Y. (2008) MUC1 and colorectal cancer pathophysiology considerations. World J. Gastroenterol., 14, 2139–2141. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49. Tian E., et al. (2009) Recent insights into the biological roles of mucin-type O-glycosylation. Glycoconj. J., 26, 325–334. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50. Biemer-Hüttmann A.E., et al. (1999) Immunohistochemical staining patterns of MUC1, MUC2, MUC4, and MUC5AC mucins in hyperplastic polyps, serrated adenomas, and traditional adenomas of the colorectum. J. Histochem. Cytochem., 47, 1039–1048. [DOI] [PubMed] [Google Scholar]
  • 51. Perçinel S., et al. (2007) Mucins in the colorectal neoplastic spectrum with reference to conventional and serrated adenomas. Turk. J. Gastroenterol., 18, 230–238. [PubMed] [Google Scholar]
  • 52. Matsuda K., et al. (2000) Clinical significance of MUC1 and MUC2 mucin and p53 protein expression in colorectal carcinoma. Jpn. J. Clin. Oncol., 30, 89–94. [DOI] [PubMed] [Google Scholar]
  • 53. Molaei M., et al. (2010) Mucins in neoplastic spectrum of colorectal polyps: can they provide predictions? BMC Cancer, 10, 537. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54. Feng M.X., et al. (2014) Elevated autocrine EDIL3 protects hepatocellular carcinoma from anoikis through RGD-mediated integrin activation. Mol. Cancer, 13, 226. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55. Beckham C.J., et al. (2014) Bladder cancer exosomes contain EDIL-3/Del1 and facilitate cancer progression. J. Urol., 192, 583–592. [DOI] [PubMed] [Google Scholar]
  • 56. Lee J.E., et al. (2016) Identification of EDIL3 on extracellular vesicles involved in breast cancer cell invasion. J. Proteomics, 131, 17–28. [DOI] [PubMed] [Google Scholar]
  • 57. Penta K., et al. (1999) Del1 induces integrin signaling and angiogenesis by ligation of alphaVbeta3. J. Biol. Chem., 274, 11101–11109. [DOI] [PubMed] [Google Scholar]
  • 58. Abram C.L., et al. (2014) Comparative analysis of the efficiency and specificity of myeloid-Cre deleting strains using ROSA-EYFP reporter mice. J. Immunol. Methods, 408, 89–100. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

bgz050_Suppl_Supplementary_Figures-1
Click here for additional data file. (160KB, png)
bgz050_Suppl_Supplementary_Figures-2
Click here for additional data file. (494.1KB, png)

Articles from Carcinogenesis are provided here courtesy of Oxford University Press

RESOURCES