A novel mouse model of rectal cancer ... - Wiley Online Library

00392.fm Page 514 Friday, May 28, 2004 9:54 AM

A novel mouse model of rectal cancer established by orthotopic implantation of colon cancer cells Takuya Takahashi, Masami Morotomi and Koji Nomoto Yakult Central Institute for Microbiological Research, 1796 Yaho, Kunitachi, Tokyo 186-8650

A novel intraluminal colon tumor model was established in mice by intrarectal instillation of colon cancer cells followed by shortterm induction of colitis by an irritant agent. Male BALB/c mice were fed a diet containing 3% (w/w) dextran sulfate sodium (DSS) for 7 days to induce colitis, and colon 26 cells (1–2 × 10 6 cells/mouse) were infused intrarectally after the mice had been deprived of food for the last 18 h of DSS treatment. The tumor incidence (%) and size (mean volume ± SD, mm3) at the rectal mucosa were 35% (2 ± 3), 95% (96 ± 79), 95% (141 ± 137) and 94% (325 ± 270) at 1, 2, 3 and 4 weeks after instillation of tumor cells, respectively. Histopathological analyses revealed that a solid tumor was formed initially at the rectal mucosa at 1 week after instillation, then became invasive into the submucosal and muscular tissues at 3 weeks after implantation. Intrarectal instillation of human colon cancer cells, LS174T (1 × 10 7 cells/mouse), mixed with “Matrigel” (0.5 mg/mouse), an extracellular matrix solution, in SCID mice led to formation of rectal tumors at 4 weeks after instillation, and immunohistochemical analysis revealed that the tumor cells expressed human carcinoembryonic antigen, suggesting that the tumor nodule was derived from the instilled LS174T cells. Oral or intravenous administration of a camptothecin (CPT) derivative, CPT-11, resulted in a significant reduction in tumor incidence and tumor volume in the colon 26-intraluminal implantation system. In conclusion, it was suggested that the present intraluminal colon tumor model is useful for examination of chemotherapeutic agents and also intraluminal factors (dietary compounds, intestinal microflora, etc.) that might function to suppress or enhance the growth of colorectal cancer in situ. (Cancer Sci 2004; 95: 514–519)

C

olorectal cancer is one of the leading causes of cancer mortality in the Western world and is also increasing in Japan.1) A great deal of research is currently in progress, not only on treatment, but also on chemoprevention of colon cancer. In order to enable such activity to progress, suitable animal models are required. A heterotopic implantation model has been most widely used to evaluate antitumor agents because of its advantages in practical use, such as high reproducibility, relatively short period needed compared to chemically induced carcinogenic models, and easier technique than other models.2) However, the heterotopic models have problems in that they are only useful for examination of antitumor/chemopreventive agents with systemic efficacy,2) and are not suitable for examination of agents/factors that may act in situ. Several animal models produced by orthotopic implantation of colon tumor cells or tissues have been established and utilized to investigate the mechanisms of progression in situ and tumor metastasis, as well as the efficacy of antitumor drugs.3–6) In these rodent models, however, either single colon cancer cell suspensions were injected into the submucosal tissues of the cecum or rectum from the serosal aspect 3–5) or fragments of colon cancer tissue were sutured to the wall of the colon.6) Implantation by these procedures has been reported to induce colonic tumors in submucosal tissues or growth of tumors at the serosal aspect, but not at the mucosal surface. These animal models therefore appear to be inappropriate to examine the direct influence of the intraluminal environment (dietary compo514–519

|

Cancer Sci

| June 2004

|

vol. 95

|

no. 6

nents, intestinal microflora, bile acids, etc.) on the growth of colorectal cancers. The objective of this study was therefore to establish an experimental model in which colon tumors are formed at the luminal mucosal surface by instilling colon cancer cells intrarectally in mice. Materials and Methods Animals. Male specific-pathogen-free BALB/c mice (9 weeks old) and male C.B-17 SCID mice (scid/scid, 6 weeks old) were purchased from Japan SLC Inc. (Hamamatsu) and CLEA Japan, Inc. (Tokyo), respectively. The mice were acclimatized for a week until at the start of experiments, and were housed in plastic cages and maintained under the following standard conditions: 22±2°C, 45±10% relative humidity, 12 h light/12 h dark cycle each day. SCID mice were maintained in a laminar flow cabinet. The mice were given a standard diet (MF, Oriental Yeast Industry Co., Tokyo) and water ad libitum. All animal studies were conducted in accordance with the principles and procedures outlined in the Guide for the Care and Use of Laboratory Animals.7) Tumor cells. A mouse colonic cancer cell line, colon 26, was maintained in RPMI 1640 (Invitrogen Japan K.K., Tokyo) supplemented with 10% fetal bovine serum (FBS, Invitrogen Japan K.K.) and kanamycin monosulfate (KM, 100 µg/ml, Sigma-Aldrich Japan Co., Tokyo) at 37°C in a humidified atmosphere of 5% CO2 and air. The monolayer culture on plastic dishes (Asahi Techno Glass Co., Tokyo) was washed once with PBS and then harvested after incubation with 1 ml of Cell Dissociation Solution (Sigma-Aldrich Japan Co.). The single cell suspension was then washed and suspended at a density of 2.5–5×10 7 cells/ml in RPMI 1640 supplemented with KM (100 µg/ml) but without FBS. Human colon cancer cells, LS174T were maintained in Dulbecco’s modified Eagle’s medium (DMEM, Invitrogen Japan K.K.), supplemented with 10% FBS and 100 µg/ml KM at 37°C. Subconfluent cultures of LS174T cells were harvested by treatment with 0.25% trypsin and 1 mM EDTA-4Na in Hank’s balanced salt solution (HBSS, Invitrogen Japan K.K.), washed and suspended at a density of 2×10 8 cells/ml in DMEM. Induction of colitis. Colitis was induced by one of the following procedures (A)–(C). (A) Dextran sulfate sodium (DSS)8): Mice were fed a MF pellet containing 3% (w/w) DSS (Tokyo Kasei Kogyo Co., Ltd., Tokyo, MW 25,000) for 7 days. (B) 2,4,6-Trinitrobenzenesulfonic acid (TNBS)9, 10): Mice were anesthetized with an intraperitoneal (i.p.) injection of 50 mg/kg sodium pentobarbital after fasting for 18 h. A sterile polytetrafluoroethylene catheter (Angiocath, 0.7 mm outer diameter and 19 mm length, Becton Dickinson Infusion Therapy Systems, Inc., Sandy, UT) was inserted into the colon so that the tip of the catheter was 2 cm from the anus. Then, 1 mg of TNBS (Sigma-Aldrich Japan Co.) in 50 µl of 50% (v/v) ethanol was infused into the colonic lumen through the catheter attached to a 1 ml tuberculin syringe. The induction of colitis was carried out on day–7 and the day of instillation of cells. (C) E-mail: [email protected]

Takahashi et al.


Hydrochloric acid (HCl): Mice were anesthetized with sodium pentobarbital on the day of instillation of cells after fasting for 18 h. A sterile polytetrafluoroethylene catheter was inserted into the colonic lumen through the anus. The colorectal mucosa was traumatized by infusing 100 µl of 0.1 M HCl solution (Wako Pure Chemical Industries, Ltd., Tokyo) for 1 min, followed immediately by neutralizing infusion of 100 µl of 0.1 M KOH and subsequent flushing with 200 µl of HBSS through the catheter fitted to a tuberculin syringe. Intrarectal instillation of colon cancer cells and examination for tumor formation. Mice were fasted for 18 h (DSS) or 6.5 h (TNBS

or HCl) after colitis induction, and then anesthetized for 1 h with sodium pentobarbital. Colon 26 cells (1–2×10 6 cells/40 µl/mouse) were infused intrarectally with a micropipette (10– 100 µl, Eppendorff Co., Ltd., Tokyo) inserted 2 cm into the anus of the mice. The anus was compressed with a noncrushing microclamp (Natsume Seisakusho Co., Ltd., Tokyo) immediately after instillation of tumor cells for 30 min to prevent leakage. Mice were sacrificed at 1, 2, 3 and 4 weeks after the instillation of tumor cells, each colorectum was examined for the development of intraluminal tumors under a stereoscopic microscope, and the tumor volume was calculated according to the following formula: Tumor volume (mm3)=length (mm)×width2 (mm2)×1/2

of Fisher’s exact probability test. Statistical analysis of tumor volume was performed with the Mann-Whitney U test. Results Orthotopic implantation of colon 26 cells. Tumor incidence, tumor volume and survival of mice were examined after instillation of colon 26 cells in BALB/c mice, which had been treated with DSS for a week until 1 day before instillation of cells. Tumor incidence at 1 week after implantation was only 35%, which, however, increased to more than 90% at 2 to 4 weeks after implantation (Fig. 1A). Tumor size increased gradually with time after implantation (Fig. 1B). In all the specimens tested, it was found that only a single tumor was formed at the rectal mucosa. No mice died until 3 weeks after implantation, but the survival rate gradually decreased thereafter to 20% by 6

Intrarectal implantation of human colon cancer cells, LS174T, in C.B-17 SCID mice. SCID mice were anesthetized after fasting for

18 h. Induction of colitis was performed by HCl infusion as described above with slight modifications. Briefly, the colorectal mucosa was traumatized by infusing 100 µl of 0.2 M HCl solution (Wako Pure Chemical Industries, Ltd., Tokyo) for 4 min, neutralized subsequently by infusion of 100 µl of 0.2 M KOH, and flushed with 200 µl of HBSS. Kanamycin sulfate at a dose of 30 mg/kg (Meiji Seika Kaisha, Ltd., Tokyo) was intramuscularly injected to protect mice from possible bacterial infections. At 6.5 h after traumatization, a suspension of LS174T cells (1×10 7 cells/100 µl/mouse) containing 0.5 mg of “Matrigel” basement membrane matrix (“Matrigel,” Becton Dickinson Bioscience, MA) was instilled into the colorectum. The occurrence of colorectal tumors was examined at 4 weeks after the instillation of tumor cells. Histopathological examination. For histopathological examination, excised colorectal specimens were embedded in paraffin and cut at 3–5 µm. The sections were stained with hematoxylin and eosin (H&E) and examined by pathologists who were blinded to the experimental groups. Human carcinoembryonic antigen (CEA) expression in the LS174T tumors was determined immunohistochemically using rabbit anti-human CEA polyclonal antibody (N1503, DAKO Cytomation Japan Co., Ltd., Kyoto) as the primary antibody and biotinylated F(ab′)2 fragments of swine anti-rabbit immunoglobulins (DAKO Cytomation Japan Co.) as the secondary antibody. Chemotherapeutic experiments. A camptothecin (CPT) derivative, CPT-11,11, 12) was used. Mice (25–26 mice/experiment) were randomly divided into two groups 2 days after intraluminal implantation of colon 26 cells. Administration of CPT-11 was performed according to the method described elsewhere13, 14) with slight modifications. Briefly, CPT-11 dissolved in saline was administered per os (p.o.) or intraveneously (i.v.) once after 3 days at a single dose of 800 mg/kg (p.o.) or 50 mg/kg (i.v.) in experiment I. CPT-11 was administered p.o. or i.v. once a day after 3, 7 and 11 days at a total dose of 800 mg/kg (p.o.) or 200 mg/kg (i.v.) in experiment II. The control group was treated with saline. CPT-11 group and the control group, which consisted of 10 mice and 15–16 mice, respectively, were sacrificed at 3 weeks after tumor implantation, and both the tumor incidence and the tumor volume were determined as described above. Statistical analysis. Tumor incidence was analyzed by means Takahashi et al.

Fig. 1. Tumor incidence, tumor volume and survival of mice after implantation of colon 26 cells. Mice were fed MF diet containing 3% (w/ w) DSS for 7 days. Colon 26 cells (1–2×10 6 cells/mouse) were then infused intrarectally as described in “Materials and Methods.” Mice were sacrificed at 1 (n=20), 2 (n=20), 3 (n=19) and 4 weeks (n=17) after cell instillation. A, Tumor incidence and B, tumor volume were examined as described in “Materials and Methods.” C, Survival of 20 mice was followed for 6 weeks after tumor implantation.

Cancer Sci | June 2004

|

vol. 95 |

no. 6 | 515


A

B

weeks after implantation (Fig. 1C), and dissection of the dead mice revealed that all had a tumor in the rectum, but not elsewhere. As shown in Fig. 2A, the site of tumor formation was found to be limited to the rectum. No tumors were formed in either the group treated only with DSS or in that infused with cells but not treated with DSS (Fig. 2, B and C). Histological analysis revealed that severe inflammation of the colonic mucosa was induced by DSS treatment for a week (Fig. 3A). In the group implanted with colon 26 after DSS treatment, the tumors at the rectal mucosa were solid at 1 week after instillation (Fig. 3B) and had become invasive into the submucosal and muscular tissues at 3 weeks after implantation (Fig. 3C). A tumor in nodular form growing at the intraluminal aspect of the colorectal mucosa was clearly seen (Fig. 3, B and C). Either induction of colitis with DSS or instillation of tumor cells alone was found to lead to no tumor formation even at 4 weeks after treatment (data not shown). Then, in the next series of experiments, the effect of three different methods for induction of colitis on the implantation efficiency was examined. The incidences of intraluminal rectal tumors observed at the rectal mucosa in the DSS-, TNBS- and HCl-treated groups were 90%, 80% and 70%, respectively. The mean tumor volume in the DSS-treated group was largest among the three groups (Table 1). Treatment with TNBS, HCl or DSS alone gave rise to no tumors (Table 1). Colonic tumor formation induced by an intrarectal instillation of LS174T cells in SCID mice. Instillation of LS174T colon tumor cell

suspension in a mixture with “Matrigel” solution resulted in tumor formation at the rectal mucosal surface (Fig. 4A). Immunohistochemical analysis using anti-human CEA antibody revealed that tumor cells expressed human CEA, suggesting that the tumor nodule was derived from the instilled LS174T cells, and that the tumor cells invaded the submucosal tissues through the lamina muscularis mucosa (Fig. 4B). The combined results of three repeated experiments showed that the tumor incidence was 27.8% (5/18). Effect of CPT-11 in BALB/c-colon 26 implantation model. Oral administration of CPT-11 as a single shot on day 3 or multiple treatments on days 3, 7 and 11 at a total dose of 800 mg/kg after implantation of colon 26 resulted in a significant reduction 516

Fig. 2. Macroscopic features of tumors formed after intrarectal instillation of colon 26 cells. A, Mice were fed MF diet containing 3% (w/w) DSS for 7 days, and colon 26 cells (1×10 6 cells/mouse) were instilled intrarectally. B, DSStreated control (no cells instilled). C, Colon 26 cells-instilled control (no DSS treatment). At 3 weeks after instillation of colon 26 cells, the colorectum was cut open longitudinally, washed once with saline, and then fixed in 10% formalin in phosphate-buffered saline. Scale indicates centimeters (cm).

C

of both the tumor incidence and tumor volume. Administration of CPT-11 by the i.v. route with the same treatment schedule as that for the oral route significantly inhibited tumor formation and growth (Table 2). Discussion

The characteristics of the present orthotopic implantation model are: 1) induction of colitis is necessary for the successful orthotopic implantation of colon cancer cells at the rectal mucosa, 2) irritant agents with different mechanisms are able to induce rectal tumors, 3) the tumor incidence is over 90% 2 weeks after instillation of tumor cells, 4) the tumor is formed intraluminally at the mucosal tissue, but never on the serosal surface, 5) the site of tumor formation is limited to the rectum, and 6) the implantation procedure is applicable to human colon cancer cells such as LS174T cells in immunodeficient SCID mice, although the incidence was lower than that in the colon 26-BALB/c system. The present experimental model appears to mimic human rectal cancer, because the site of tumor formation was found to be the rectum (Fig. 2A). The tissue tropism of tumor formation appeared to be due to shutting tumor cells in the colorectum for 30 min by clamping the anus immediately after instillation of tumor cell suspension, which might then allow the tumor cells to pool at the rectal mucosa during anesthesia. It has been reported that patients with inflammatory bowel disease such as ulcerative colitis have a high risk for colorectal cancer.15) In animal experiments, irritants such as DSS and TNBS have been used for induction of chronic intestinal inflammation. The mechanisms by which the two agents induce inflammation are different: DSS treatment has been reported to cause accumulation of macrophages phagocytosing DSS in the mucosal lesion, and subsequent suppression of bacterial phagocytosis by macrophages results in the breakdown of the mucosal defense system against intestinal bacteria.8, 16) In the TNBSinduced colitis model, which resembles human Crohn’s disease, it has been reported that CD45RBlowCD4 + T cells in the lamina propria contribute to the development of colitis and that immune responses dominated by type 1 helper T lymphocytes are induced by TNBS treatment.9, 10, 17) Not only chronic inflammatory agents such as DSS and TNBS, but also acute inflammaTakahashi et al.


Table 1. Tumorgenicity of colon 26 cells instilled into rectal lumen of BALB/c mice after induction of colitis

Fig. 3. Histopathology of colon 26 tumors at rectum of BALB/c mice. Histopathological examination (H&E staining) of the rectal mucosa was performed immediately before implantation (A, ×40), and of the excised colon 26 tumors at 1 week (B, ×40) and 3 weeks (C, ×20) after implantation.

tion induced by intrarectal instillation of HCl followed by neutralization with KOH made it possible to induce rectal tumors at a high incidence, suggesting that inflammation due to different mechanisms enhances rectal tumor formation in the mucosa. The tumor growth in the DSS-pretreated group appeared to be more progressive compared with that in the groups pretreated with TNBS or HCl (Table 1), which might be due to long-term immunosuppression by DSS.7, 15) Intrarectal instillation of LS174T, as a mixture with “Matrigel” solution, in SCID mice led to the formation of a rectal tuTakahashi et al.

Induction of colitis by:

Instillation of colon 26 cells

Tumor incidence (%)

Tumor volume (mm3/mouse, mean±SD, range)

None

− +

0/3 (0) 0/6 (0)

— —

DSS

− +

0/5 (0) 17/19 (90)

— 141±137 (0–536)

TNBS

− +

0/5 (0) 16/20 (80)

— 39±59 (0–238)

HCl

− +

0/5 (0) 14/20 (70)

— 15±26 (0–89)

Fig. 4. Histopathology of LS174T tumor formed at rectum of C.B-17 SCID mouse. A LS174T tumor formed at the rectum was examined histopathologically 4 weeks after instillation of 1×10 7 LS174T cells (H&E staining, A, ×40). Immunohistochemical analysis of human CEA was performed as described in “Materials and Methods” (Immunoperoxidase staining, B, ×200).

mor, while instillation of tumor cells by themselves did not yield any tumors in SCID mice (data not shown). “Matrigel” contains extracellular matrix components such as laminin and collagen IV18) and tumor growth factors such as TGF-β, fibroblast growth factor and tissue plasminogen activator.19) “Matrigel” has frequently been used for implantation of several types of human cancer cells in immunodeficient mice, and enhanced the growth of tumor cells.4, 20) Tsutsumi et al. have reported that Cancer Sci | June 2004

|

vol. 95 |

no. 6 | 517


Table 2. Antitumor effects of CPT-11 in rectal colon 26 tumor model Treatment schedule (day)

Route

Total dose (mg/kg)

Tumor incidence (%)

Tumor volume (mm3 /mouse, mean±SD, range)

Experiment I p.o. 3

i.v.

Saline CPT-11 Saline CPT-11

— 800 — 50

14/16 2/101) 13/15 6/10

(88) (20) (87) (60)

109±181 4±116) 145±155 30±724)

(0–495) (0–36) (0–468) (0–230)

Saline CPT-11 Saline CPT-11

— 800 — 200

16/16 3/103) 14/16 3/102)

(100) (30) (88) (30)

156±142 21±656) 91±92 10±195)

(12–608) (0–207) (0–302) (0–51)

Experiment II p.o. 3, 7, 11

1), 2), 3) 3) P