You are on page 1of 17

GASTROENTEROLOGY 2012;143:741–753

Obesity-Induced Increase in Tumor Necrosis Factor-␣ Leads to
Development of Colon Cancer in Mice
*Department of Internal Medicine, Faculty of Medical Sciences, ‡Department of Anatomy, Cell Biology, Physiology and Biophysics, State University of Campinas,
Campinas, São Paulo, Brazil

BACKGROUND & AIMS: Epidemiology studies have ing serine phosphorylation of insulin-receptor substrate
shown that obesity increases risk for colorectal cancer (IRS) proteins by kinases such as c-jun N terminal kinase
(CRC). We investigated the contribution of obesity-in- (JNK)11 and inhibitor of nuclear factor-␬B kinase (IKK).12
duced increases in levels of tumor necrosis factor (TNF)-␣ JNK and IKK activation induce inhibitory serine 307
and hyperinsulinemia to the development of CRC in mice. (Ser307) phosphorylation of IRS-1,11,13,14 which decreases
METHODS: Lean and obese mice (C57BL6/J and ob/ob) insulin-mediated phosphatidylinositol 3-kinase (PI3K)/
were given a combination of azoxymethane and dextran Akt/mammalian target of rapamycin (mTOR) pathway
sulfate sodium, which led to the development of CRC; activation.13
lean and obese severe combined immunodeficient mice TNF-␣, first identified as an antitumor agent, now also
were injected with HT-29 cells. We analyzed the roles of is recognized as a tumor-promoting cytokine that links
TNF-␣ and insulin in the development of obesity-medi- inflammation and cancer.15 The importance of TNF-␣
ated CRC using immunoblot, immunohistochemical, and and its intracellular mediators, such as JNK and IKK, as
apoptosis assays. RESULTS: Genetic- and diet-induced colonic tumorigenic promoters is strengthened by knock-
obesity increased the incidence and size of tumors that out and pharmacologic studies, using azoxymethane
developed after administration of azoxymethane and dex- (AOM) or AOM combined with dextran sulfate sodium
tran sulfate sodium, compared with lean mice. HT-29 (DSS) as inducers of colorectal carcinogenesis.16 –18 Fur-

xenograft tumors grew more rapidly in obese than lean thermore, the increased TNF-␣ levels associated with obe-

mice. Neutralization of TNF-␣ reduced activation of c-Jun sity are a potent liver tumor promoter in mice.19
N-terminal kinase, I␬B kinase, and the phosphatidylino- Although the evidence for the potential involvement of
sitol 3-kinase–Akt–mammalian target of rapamycin sig- inflammation and hyperinsulinemia in the development
naling pathway; it also reduced the growth and develop- of obesity-mediated cancer is quite extensive, a systematic
ment of tumors in obese mice. Reducing levels of insulin evaluation of the independent contribution of these fac-
levels had no effect on tumor growth in obese mice. tors to CRC development is lacking. Here, we examined
CONCLUSIONS: TNF-␣ contributes to colon tumor whether obesity modulates insulin signaling and inflam-
mation in the colon and CRC. We show that the obesity-
growth in obese mice. Reagents that inhibit TNF-␣
induced abnormal inflammatory response strongly pro-
might prevent the development or progression of CRC
motes CRC.
in obese individuals.
Keywords: JNK; Inflammation; Insulin Resistance; Mouse Materials and Methods
Model. Antibodies, Chemicals, and Buffers
All the reagents were from Sigma-Aldrich (St Louis,

C olorectal cancer (CRC) remains a major health bur-
den with more than 1 million cases worldwide and a
disease-specific mortality of approximately 33% in the
MO), unless otherwise specified. Octreotide was from Novartis
PharmaStein AG (Stein, Switzerland), pioglitazone was from
Takeda Pharmaceutical (Osaka, Japan), infliximab was from
Centocor (Horsham, PA), and DSS was from MP Biochemicals
developed world.1,2 The association between obesity and (Solon, OH). Anti–phospho-mTOR, anti-mTOR, anti–phospho-
the risk for CRC development is observed in both men
and women (relative risk, 1.2–2.0).3,4 In addition to its
Abbreviations used in this paper: AOM, azoxymethane; CRC, colorec-
association with obesity, inflammation and hyperinsulin-
tal cancer; DSS, dextran sulfate sodium; HFD, high-fat diet; IKK, inhib-
emia also primarily may contribute to the risk for devel- itor of nuclear factor-␬B kinase; IP, intraperitoneally; IRS, insulin recep-
opment of CRC.5– 8 tor substrate; ITT, insulin tolerance test; JNK, c-jun N terminal kinase;
Among the major mediators of the inflammatory re- mTOR, mammalian target of rapamycin; PBS, phosphate-buffered sa-
sponse is tumor necrosis factor (TNF)-␣, whose overex- line; PI3K, phosphatidylinositol 3-kinase; SCID, severe combined im-
mune deficiency; TNF, tumor necrosis factor.
pression in adipose tissue is a common feature in human © 2012 by the AGA Institute
and animal models of obesity.9,10 TNF-␣ contributes to 0016-5085/$36.00
the deregulation of the insulin-signaling pathway, includ-

The diet composition is described in Supplementary Epididymal fat pads from mice were excised and minced Table 1. 3 p70S6K. Mice were provided from the Central Breeding Center of the State University of Campinas and Isolation of the Stromal Vascular Fraction and were divided randomly into 2 groups: control and high-fat diet Adipocyte of Adipose Tissue (HFD). Tumor counts were performed and every 5–10 minutes during the incubation. pernatants of these tissues were used. 1. anti–phospho-c-jun. mmol/L EDTA. MD). For bovine serum with the addition of antibiotics and fungicides. Mice were anesthetized with sodium amobarbital (15 MA). fixed in in complete RPMI 1640 for 30 minutes at 37°C in a shaking 4% paraformaldehyde. Serum TNF-␣ was measured used for protein analysis by immunoblotting. and anti–␤-tu.5% bovine serum albumin. Frederick. Brazil). and anti-TNF receptor 1 antibodies were from sodium pyrophosphate.02% colla- killed 10 days after the last cycle.52.22 a day for a total dosage of 0. anti-IKK␤. Indianapolis. The Animals extracts were centrifuged at 11. Infliximab was triphosphate nick-end labeling staining was performed using a given daily IP for a total dosage of 5 mg/kg body weight. and 0. anti– Aldrich]. Afterward. anti–phospho-Akt. flushed with phosphate-buffered saline (PBS). anti–phospho-JNK. 36 –50 kilodaltons) threitol. followed by 14 days minutes at 37°C and shaken vigorously.5-cm sections. 143. Collagenase type II was added at 1 mg/mL and incubated at Insulin Tolerance Test. pioglitazone. and homogenized in extraction buffer (1% Triton-X 100 [Sigma- anti-I␬B␣. 10 Santa Cruz Biotechnology (Santa Cruz. and paraffin-embedded.5 g for 5 minutes to separate floating adipocytes from the stromal IU insulin/kg body weight) as described previously. nas approved all experiments. Serum Insulin. anti-JNK. resis. MO). Anti–phospho-IR. anti–TNF-␣. IL). Tumor Induction and Analysis The intestines were opened longitudinally. anti–phospho-IRS-1 Tyr971. pH 7. the 2 fractions were assay kit (Linco. 10 mmol/L sodium vanadate. minced coarsely. anti–phospho-IKK␤. and heated in a boiling water bath. Sections (5 ␮m) water bath until complete digestion of the tissue. 100 mmol/L Tris. anti-IR. at 16 weeks of age. CA). phospho-IRS-1 Ser307. Tumors were removed. was purchased Whole-tissue extracts were homogenized in extraction from the Rio de Janeiro Cell Bank (Rio de Janeiro. washed in Hank’s Four-week-old male mice (C57BL6/J Unib and ob/ob) balanced salt solution. Charles.01 mg/animal/day. and the supernatant of regular water. and the final supernatant was passed Assessment of colitis disease scores was performed as previously through a 100-␮m filter. in the dorsal region with 1 ⫻ 106 HT-29 cells.4. using a mouse enzyme-linked immunosorbent assay kit (Thermo Scientific. . Tissue was were placed on standard chow or on a high-fat diet for 1 week placed in 10 mL fresh calcium/magnesium-free Hank’s balanced TRANSLATIONAL AT and then injected intraperitoneally (IP) with 12.1 mg of aprotinin/mL). salt solution containing 1 mmol/L EDTA and 1 mmol/L dithio- After 5 days. Octreotide was given subcutaneously twice Ki67 staining was performed as described previously. and cut into 0. total extracts. St. No. HT-29. 2 mmol/L phe- nylmethylsulfonyl fluoride. and buffer. the fraction was used for described.20 Plasma was vascular fraction pellet. Band intensities were quantified by optical densitom- Four-week-old male severe combined immune deficiency etry of developed autoradiographs using Scion Image software (SCID) mice (n ⫽ 10 per group) were inoculated subcutaneously (Scion Corporation. the tube was in a blinded fashion. anti-Akt. the tube was placed in a shaking water bath for 15 BASIC AND was given in the drinking water for 5 days. Samples were digested until adipocyte separated by centrifugation (1100 ⫻ g) for 15 minutes at 4°C fractions were free of adherent cells by these 2 quality control and stored at ⫺80°C until the assay. Tissue Extracts bulin antibodies were from Cell Signaling Technology (Beverly.21 protein analysis by immunoblotting. Colons were genase V and 0. anti– mg/kg body weight. Tumor volume (V) was calculated daily by measuring length (L) and width (W) of Immunohistochemistry and Terminal the tumor with calipers and using the following formula: V ⫽ Deoxynucleotidyl Transferase–Mediated {W ⫻ L ⫻ [(W ⫹ L)/2]} ⫻ 0. 742 FLORES ET AL GASTROENTEROLOGY Vol. in PBS with calcium chloride and 0. Serum insulin and leptin methods to ensure recovery of the majority of the stromal were measured by using a mouse enzyme-linked immunosorbent vascular fraction population. Deoxyuridine Triphosphate Nick-End Labeling Treatments with octreotide. or infliximab began Assay with cell inoculation. and the su- The Ethics Committee of the State University of Campi. The cell suspension was and TNF-␣ Quantification filtered through a 100-␮m filter and then centrifuged at 300 ⫻ The mice were given an insulin tolerance test (ITT. Pioglitazone was Terminal deoxynucleotidyl transferase–mediated deoxyuridine given orally for a total dosage of 50 mg/kg/day. Colonic Macrophage Isolation The colons of mice were removed and washed in PBS. Tumor sizes were measured with calipers. IN) removed. 2. 37°C for 20 minutes with shaking.5 mg/kg AOM.05% DNase (Roche Diagnostic. anti–IRS-1.5% DSS (molecular weight. At the start of were cut and stained with H&E. Her- Human Tumor Xenograft Models cules. Proteins were resolved in 8%–15% sodium dodecyl sulfate gels and blotted onto nitrocellulose membranes (Bio-Rad. Cell Culture Protein Analysis by Immunoblotting The human colon cancer cell line. The remaining tissue was digested with 0. anti-p70S6K.000 rpm and 4°C. This cycle was repeated twice and mice were separated. Leptin. IP). CA). Rockford. containing 100 mmol/L c-jun. 100 mmol/L sodium fluoride. similar-sized aliquots (50 ␮g protein) were sub- Cells were maintained at 37°C in a humid atmosphere and 5% jected to sodium dodecyl sulfate–polyacrylamide gel electropho- CO2. treated with Laemmli sample buffer containing 100 cells were cultured in McCoy’s medium containing 10% fetal mmol/L dithiothreitol. commercial apoptosis detection kit (Roche Diagnostics GmbH. Afterward. shaken vigorously.

accompanied by Inflammation and Insulin Signaling increased IRS-1 Ser307 phosphorylation. We observed a slightly higher activation of insulin re. epididymal fat. matory state observed in the colon. Twenty amounts of macrophages and leukocytes associated with weeks after the AOM treatment. and a Colon sections from the obese mice showed higher single dose of AOM (12. Inflammatory Microenvironment Promotes formed as described previously. as described in tyrosine phosphorylation of IR␤ and IRS-1 (Figure 2E). and also increases ity in the obese animals. (Supplementary Table 2). as well as higher Impairment levels of phosphorylated Akt. to the same AOM⫹DSS protocol.22 Colon Cancer Mice kept on an HFD showed increased tumor Statistical Analysis incidence and size compared with control mice (Figure 2A Data are presented as means ⫾ standard error of the and B). inflammatory cytokines. and observed increased TNF-␣ expression. mentary Table 2). These alter- compared with control mice (Figure 1E). mentary Table 2). As shown in insulin. as higher level of histologic damage (Figure 3A). The alterations Inflammation and Insulin Signaling in colon tissue were accompanied by an increase in Ki67 Impairment staining and diminished apoptosis (Figure 1C) when com. we evaluated the BASIC AND owing to an obesity-related increase in circulating levels of adipose tissue. To verify observed previously. In agreement with these re. trols. During the protocol. including rectal bleeding and prolapse (Figure 3A). Statistical increased TNF-␣ expression. when compared with the con- Four-week-old male mice received a single dose of trol mice (Figure 2D). Insulin increased Akt phosphorylation in a time. Significance was established at the phorylated JNK and c-jun (Figure 2C). phos. mTOR. We examined the effects of obesity on several kinases . the mice were killed. accompanied by increased IRS-1 Ser307 phosphoryla. As the Materials and Methods section. As observed in Figure 1F. with maximum phosphorylation at We next examined proinflammatory protein expression 10 minutes (Supplementary Figure 1A). The HFD impaired insulin-induced AOM and 3 cycles of DSS (AOM⫹DSS). The Diet-Induced Obesity-Mediated mendations. Germany). serum flammatory proteins in the obese mice. Consistent with these signs. and p70S6K on the tumors in the obese mice. more clinical signs of inflammatory response than con- tion. As evidenced by the dosage of serum and tissue mice (Supplementary Figure 2B). is derived from the adipocytes and from the macrophages Genetic Obesity Promotes Colonic present in the adipose tissue (Figure 1B). measured with the ITT (Supple- JNK and c-jun phosphorylation in colonic tissues. and leptin when compared with controls (Supple- Figure 1D. leading to the colon was examined for the presence of tumors. We Results observed a slightly higher activation of IR␤ and IRS-1 Diet-Induced Obesity Engenders Colonic with increased tyrosine phosphorylation. HFD increases TNF-␣ expression. de- analysis was performed by using the analysis of variance test creased I␬B␣ expression. IKK␤ phosphorylation. mice were placed on an HFD or on a control diet. We submitted 4-week-old male ob/ob and WT mice pared with control animals. we examined the effects of obesity on several kinases involved in insulin signaling in carcinomas. dependent manner. resulting in in- pocytes and the stromal vascular fractions). with increased levels of phosphorylated JNK and c-jun in phorylated Akt levels are similar in the insulin-stimulated the ob/ob colon. with its two major components (the adi. IKK␤ phos- paired insulin-induced tyrosine phosphorylation of IR␤ phorylation. when compared with control mice (Figure and nonstimulated colons of obese (HFD-fed) mice. 3C). presented increased body weight. Similar to the colonic tissues. Analysis and documentation of results were per. obese mice presented a higher in- the source of the increased TNF-␣ leading to the inflam- TRANSLATIONAL AT flammatory state in the colon (Supplementary Figure 2A).5 mg/kg) was given IP. compared with the controls mors from HFD-fed obese mice. tion as well as higher levels of phosphorylated Akt. and increased levels of phos- with the Bonferroni post test. As expected. IKK␤ phos. We observed decreased insulin To rule out the effect of the inflammation caused by sensitivity in the obese animals. we observed a significantly mTOR. as measured with the ITT DSS in the colon. or the colonic creased tumor incidence and size compared with control tissue. and TNF-␣. Next. we performed an experiment in which (Supplementary Table 2). decreases I␬B␣ expression. Next. Mice exposed to HFD observed in Figure 2E.05 level. ob/ob mice showed significantly ceptor ␤ and IRS-1 with increased tyrosine phosphoryla. epididymal fat. and epithelial cell damage and adenoma formation. and p70S6K in the colons of the obese mice. associated and IRS-1 (Figure 1F). phosphorylated Akt levels are sim- presented increased body weight. and decreased I␬B␣ expression. such as TNF-␣. leptin.September 2012 OBESITY–MEDIATED COLON CANCER 743 Mannheim. The HFD im. colon tumors show mean (SEM) of at least 3 independent experiments. ob/ob mice sults. serum ilar in insulin-stimulated and nonstimulated colon tu- insulin. ations were accompanied by an increase in Ki67 staining ined the response of colonic tissue to insulin in control and diminished apoptosis (Figure 3B) when compared mice. P ⬍ . We observed decreased insulin sensitiv- phorylation. we exam. with lean animals. TNF-␣. As an inflammatory state in the colon (Figure 1A). we observed that the majority of the TNF-␣. we observed an increase in the activation of proin. according to the manufacturer’s recom.


mTOR. Data are the means ⫾ SEM. BASIC AND (B) Tumor incidence and size of colons removed at the end of protocol from CTL and HFD-fed mice. and Akt phos- phorylation in colon tumors. TRANSLATIONAL AT (A) Macroscopic changes in co- lonic tissues. Scale bars. IRS-1.05 vs HFD⫺. I␬B␣. and colonic expression of I␬B␣ TNF receptor 1 (TNFR1) and ␤-tubulin. 4™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™ Figure 1. . Western blots showing colonic tumor lysates from C57BL6/J control and HFD-fed mice. #P ⬍ . CTL. IRS-1. Data are the means ⫾ SEM. mTOR. mean ⫾ SEM.05 vs HFD⫺. 4 fields per tumor section. ‡P ⬍ . (D) Colonic tumor IR␤. §vs CTL⫹. and colonic SVF from CTL and HFD-fed mice.05 vs respective control group. IRS-1. (D) Western blots showing colonic tumor lysates from C57BL6 mice. #P ⬍ . Results are means ⫾ SEM (n ⫽ 8). white bars. §vs CTL ⫹. (E) Insulin-in- duced IR␤. IRS-1. black bars. JNK. IKK␤. *P ⬍ . TNFR1. adipose tissue stromal vascular fraction (SVF). and c-Jun expression and phosphorylation. and c- Jun expression and phosphory- lation.05 vs CTL⫺. Akt.05 vs respective lean group. JNK. HFD.September 2012 OBESITY–MEDIATED COLON CANCER 745 Figure 2. and p70S6K expression and phosphorylation. isolated adipocytes. (C) Colonic tumor TNF-␣. (E) Colonic IR␤. and inflammatory index of colons from control (CTL) and HFD-fed mice. (B) TNF-␣ dosage of serum. white bars. black bars. 5 mm. CTL. (A) Macroscopic changes in epithelial tissues. and Akt phosphorylation in the colon. Akt.05 vs CTL⫺. Diet-induced obesity promotes colonic inflammation and insulin signaling impairment. Colonic TNF-␣. (F) Insulin-induced IR␤. ‡P ⬍ . (C) Representative microphotographs and quantification of Ki67 and apoptotic (terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick-end labeling [TUNEL]) staining on colonic tissue sections of CTL and HFD-fed mice. IRS-1 Ser307. Results are means ⫾ SEM (n ⫽ 8). HFD. and p70S6K expression and phosphorylation. *P ⬍ . enterocytes. Diet-induced obesity promotes colon carcinogenesis. H&E staining of histologic sections. IKK␤. IRS-1 Ser307.

black bars. I␬B␣. Akt. IRS-1.05 vs HFD⫺. IRS-1. white bars. and Akt phosphorylation in the colon. Results are means ⫾ SEM (n ⫽ 10). No. (C) Colonic TNF-␣. and c-Jun expression and phosphorylation. H&E staining histologic sections. Western blots of 6 independent experiments showing colonic lysates from control or ob/ob mice. (B) Representative microphotographs and quantification of Ki67 and apoptotic (terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick-end labeling [TUNEL]) staining on colonic tissue sections of CTL and ob/ob mice. (A) Rectal prolapse presented in ob/ob mice during protocol treatment. §vs CTL⫹.05 vs CTL⫺. *P ⬍ .05 vs respective control group. IKK␤. TNFR1. JNK. Data are the means ⫾ SEM. ‡P ⬍ . 143. IRS-1 Ser307. mean ⫾ SEM. CTL. . and p70S6K expression and phosphorylation. Genetic obesity promotes colonic inflammation and insulin signaling impairment. and inflammatory index of colons from CTL and ob/ob mice. (D) Colonic IR␤. ob/ob. (E) Insulin-induced IR␤. #P ⬍ . 3 TRANSLATIONAL AT BASIC AND Figure 3. mTOR. 4 fields per tumor section. 746 FLORES ET AL GASTROENTEROLOGY Vol.

or decreasing Akt phosphoryla- The greater the degree of host obesity. mTOR. creased IRS-1 Ser307 phosphorylation. As shown in Figure 4D. treated HFD mice. in improving creased body weight. ylation. the faster the tion (Supplementary Figure 3D). and mice presented lower activation of IR␤ and IRS-1 after p70S6K (Figure 6D). or apoptosis when compared with untreated I␬B␣ expression. increased I␬B␣ expression. Next. accompanied by in. sulin resistance. interestingly. Akt phosphorylation. and p70S6K on the stimulation (Figure 5D). leptin. TNF-␣ expression. phosphorylated Akt levels are similar in insulin-stimu. caused by the increased proliferation of tumor cells. Next. which reached the largest size in (Supplementary Figure 3E and F) when compared with un- the HFD mice (Figure 5A). (Supplementary Figure 3C). we observed the discrete activation of IR␤ and Effects of Obesity on Tumor Formation IRS-1. Tumor Growth in SCID Mice The octreotide treatment was effective in reducing insulin The SCID mice submitted to HFD showed in. Thus. when compared with HFD High-Fat Diet Enhances HT-29 Xenograft mice (Figure 6E). c-jun. decreased I␬B␣ expression. As observed in Fig. IR␤ and IRS-1 lead colons of obese mice compared with the colons of control to impaired insulin signaling in the HFD mice. Carcinogenicity Associated With Inflammation the effect of obesity on HT-29 xenografts is characterized and Colonic Insulin Signaling Impairment by enhanced cell proliferation and decreased cell death. mTOR. we found that ob/ob mice also showed enhanced tumor incidence and Neutralization of TNF-␣ Reverses size compared with control mice (Figure 4A and B). we examined the response of tumor tissue to crease in IRS-1 Ser307 phosphorylation and. Thus. IKK␤ phosphorylation. and increased lev- the levels observed in lean animals (Figure 6A). effective in improving insulin sensitivity. In addition. we also examined the proliferation. As shown in Figure 4C. although it was not lonic tumors. increased tyrosine phosphorylation. with a lower activation of IR␤ and IRS-1 Akt also presents impaired activation after insulin stimu- after insulin stimulation (Figure 3E). The enhanced tumor growth in the HFD mice was stimulated and nonstimulated colons in ob/ob mice. IRS-1. infliximab treatment decreased the IKK␤ phosphor- lated and nonstimulated colon tumors from ob/ob mice. A slightly observed an increase in IRS-1 Ser307 phosphorylation (Fig- higher activation of IR␤ and IRS-1 was observed with ure 5C). insulin resistance in peripheral tissues (Supplementary Fig- and TNF-␣ levels as well as decreased insulin sensitivity ure 3B). Colons from ob/ob mice presented in. and p70S6K. ure 3E. The ob/ob mice also presented higher levels group. As observed in Figure 4E. but was not effective in altering tumor growth (Supplementary Table 2). ob/ob effect on tumor growth in HFD mice when compared mice present increased TNF-␣ expression and IKK␤ phos- with nontreated mice. When we analyzed the insulin signaling pathway of these TNF-␣ Neutralization Protects From the tumors. compared with the controls. represented by tyrosine phosphorylation and the We designed a pharmacologic approach to deter- higher activation of Akt. insulin in control and ob/ob mice. the xenografts also presented increased effective in altering tumor growth. Obesity-Induced Effects on Tumor Growth we examined the inflammatory pathway in tumor tissue Infliximab treatment exerted a stronger inhibitory of control and ob/ob mice. Tumors from ob/ob a decrease in the phosphorylation of Akt. Infliximab was effective in reducing insulin stimulation (Figure 4E). with higher phosphorylation of JNK and mice (Supplementary Figure 4).September 2012 OBESITY–MEDIATED COLON CANCER 747 involved in insulin signaling in the ob/ob colon. levels in HFD mice (Supplementary Figure 3A). phosphorylated Akt levels are similar in insulin. as well as higher and Akt of the control and obese SCID mice after insulin levels of phosphorylated Akt. serum insulin. The pioglitazone treatment also was As previously observed in the AOM⫹DSS–induced co. Next. when compared with control mice (Figure 5B). is owing to decreased than control mice. and diminished ap- Genetic Obesity Promotes Colon optosis (Figure 5F) relative to the control animals. proliferation. lus. despite increased steady-state Akt serine phosphorylation. and its ablation is capable of tyrosine phosphorylation and increased IRS-1 Ser307 phos- reducing tumor growth to levels observed in the control BASIC AND phorylation. ob/ob mice. Thus. we examined the phosphorylation of IR␤. or apoptosis HT-29 xenografts grew. whereas mice (Figure 3D). of phosphorylated Akt. when cate that ob/ob mice present a higher inflammatory state compared with the HFD mice. and reduced the phos- phorylation of JNK and c-jun. and decreased proliferation. mTOR. epididymal fat. As observed. the increased proinflammatory effect mediated by obesity. mTOR. shown by Ki67 staining (Figure 6B). and p70S6K when com- We observed that infliximab treatment induced a de- pared with control animals. as shown by Ki67 staining (Figure 5E). and insulin signaling pathway in the tumors from control and enhanced apoptosis of tumor cells (Figure 6C). ob/ob mice present a TNF-␣ plays a significant role in the proliferation of slightly higher activation of IR␤ and IRS-1 with increased TRANSLATIONAL AT tumor cells in obese mice. nearly restoring tumor growth to phorylation. These results indi- duced tumor growth of the infliximab-treated mice. The re- els of phosphorylated JNK and c-jun. we also mine the effects of TNF-␣ on tumor formation during . After the AOM⫹DSS protocol.

Western blots of 5 inde- pendent experiments showing colonic tumor lysates from con- trol and ob/ob mice. consistent with our pre- (Figure 7A). In addition. Consistent with these signs. Data are the means ⫾ SEM. reducing proliferation. we started treatment with infliximab mor growth in the ob/ob mice. TNFR1. 748 FLORES ET AL GASTROENTEROLOGY Vol.05 vs CTL⫺. we observed that vious studies. CTL. (D) Colonic tumor IR␤. 5 mm. I␬B␣.05 vs HFD⫺. ‡P ⬍ . (C) Colonic tumor TNF-␣. (E) Insulin-induced IR␤. #P ⬍ . and c-Jun expres- sion and phosphorylation. increased insulin sensitivity as measured by the and size in ob/ob mice (Figure 7B). ob/ ob. tolerance test9. Genetic obesity en- hances tumor formation. Results are means ⫾ SEM (n ⫽ 8). IRS-1. and p70S6K expression and phosphoryla- tion. 3 Figure 4. an effect that was not rate constant for glucose disappearance in the insulin observed with pioglitazone (data not shown). obesity. §vs CTL⫹. JNK. and observed that the infliximab treatment resulted in shown by Ki67 staining (Figure 7C). . 143. (B) BASIC AND Tumor incidence and size from colons removed at the end of the protocol from CTL and ob/ob mice. and Akt phosphorylation in the colon tumor. besides neutralizing infliximab treatment decreased colon tumor incidence TNF-␣. Furthermore.05 vs the respective control group. and enhancing apo- decreased levels of histologic damage in the colon sections ptosis (Figure 7D). IRS-1. mTOR. Akt. Scale bars. white bars.23 (Supplementary Table 3). IRS-1 Ser307. Concomitantly with the AOM⫹DSS protocol infliximab treatment exerted an inhibitory effect on tu- with the ob/ob mice. black bars. infliximab treatment. IKK␤. *P ⬍ . (A) TRANSLATIONAL AT Macroscopic changes in epithe- lial tissues. No.

(D) Insulin-induced IR␤. I␬B␣.September 2012 OBESITY–MEDIATED COLON CANCER 749 TRANSLATIONAL AT BASIC AND Figure 5. Western blots of 6 independent experiments showing xenografts lysates from control and high-fat diet–fed mice.05 vs the respective lean group. IRS-1 Ser307. Akt. black bars. and c-Jun expression and phosphorylation. TNFR1. (A) HT-29 tumor xenografts from normal chow and HFD-fed SCID mice and tumor growth were measured for 30 days. ‡P ⬍ . mean ⫾ SEM. and Akt phosphorylation in the xenograft tumor. CTL.05 vs HFD⫺. §vs CTL⫹. and p70S6K expression and phosphorylation. Representative microphotographs and quantification of (E) Ki67 and (F) apoptotic staining on HT-29 xenograft sections of control and HFD mice. IRS-1. *P ⬍ . JNK. IRS-1. 4 fields per tumor section. #P ⬍ .05 vs CTL⫺. The results are means ⫾ SD of 5– 8 mice per group. (C) HT-29 xenograft IR␤. white bars. (B) HT-29 xenograft TNF-␣.05 vs control. . IKK␤. Data are the means ⫾ SEM. mTOR. *P ⬍ . High-fat diet enhances HT-29 xenograft tumor growth in SCID mice. HFD.

The results are means ⫾ SD from 5– 8 mice per group. and p70S6K expression and phosphorylation. . (D) HT-29 xenograft IRS-1. (B and C) Representative microphotographs and quantification of (B) Ki67 and (C) apoptotic staining on HT-29 xenograft sections. I␬B␣. No. TNFR1. CTL. *P ⬍ . 4 fields per tumor section. (E) HT-29 xenograft TNF-␣.05 vs control. 750 FLORES ET AL GASTROENTEROLOGY Vol. mTOR. Akt. infliximab. JNK. Neutralization of TNF-␣ reverses obesity-induced effects on tumor growth. IKK␤. black bars. IRS-1 Ser307. (A) Tumor growth was measured for 30 days in SCID mice maintained on normal chow or HFD for 8 weeks before tumor inoculation with HT-29 cells (1 ⫻ 106 cells/mouse) and kept on normal chow or HFD for 30 days after inoculation and with the indicated treatment. 3 TRANSLATIONAL AT BASIC AND Figure 6. Western blots showing HT-29 xenograft lysates from tumor-bearing SCID mice. INFLIX. white bars. mean ⫾ SEM. HFD. and c-Jun expression and phosphorylation. 143.

reduced the magnitude decreased tumor growth or Akt phosphorylation with of carcinogenesis in HFD-fed and in ob/ob mice. and Akt The results of this study show that HFD induced phosphorylation in the colons of HFD-fed and ob/ob mice. we ob. that can activate the PI3K/Akt/mTOR pathway in the graft growth in HFD-fed SCID mice. (A) H&E stain- ing histologic section and inflam- matory index and (B) macro- scopic changes (scale bars. a TNF-␣ inhibitor. suggesting that TNF-␣ is a major modulator of the characteristics of carcinogenesis and reducing inflamma- effects of obesity-induced CRC. 5 mm). mean ⫾ SEM.September 2012 OBESITY–MEDIATED COLON CANCER 751 Figure 7. It is interesting to note dition. In ad.16 –18 Our data show that obesity induced a great also may influence CRC recurrence. also significant increase in host insulin levels and an impair. tumor incidence. decrease the tumor growth rate. Absence of TNF-␣ protects from obesity effects on tumor formation.05 vs control. increase in the colonic inflammatory index. as well as octreotide treatment. we did not observe of infliximab. inhibited the HT-29 xenograft growth in obese SCID Tumor-promoting inflammation is one of the enabling mice. decreases the synthesis of several intestinal peptides. 4 fields per tumor sec- tion. are critical linemia is an important mechanism by which obesity can in colon carcinogenesis induced by the treatment of confer an increased risk of colon cancer. accompanied . and insulin levels AOM. IRS-1. and genetic obesity markedly increased the development providing a biochemical correlation for decreased colonic of CRC. the HFD-fed and ob/ob mice also presented a that octreotide. beyond reducing insulin secretion. tion. octreotide and cose-dependent insulinotropic polypeptide is a hormone pioglitazone treatments did not influence HT-29 xeno. Our data indicate that obesity-induced colonic in vivo insulin sensitivity.27 TNF-␣ and its mediators. and size of colons removed at the end of the TRANSLATIONAL AT protocol from ob/ob mice treated or not with infliximab.6. Discussion served a blunted insulin-stimulated IR. tion during tumor promotion is pivotal for chemopreven- Epidemiologic studies have suggested that hyperinsu. nevertheless. inflammation increases JNK and IKK␤ expression and octreotide and pioglitazone treatments were unable to activity in the colon of AOM⫹DSS–treated mice. Moreover. such ment in insulin activation of the IR/IRS-1/Akt pathway in as glucose-dependent insulinotropic polypeptide.24 However. *P ⬍ . JNK and IKK. Consistent with these results. Representa- BASIC AND tive microphotographs and quan- tification of (C) Ki67 and (D) apo- ptotic staining of colonic tissue sections.25 Glu- the colon of these animals. The administration gastrointestinal tract26.

363:1346 –1353. decrease in tumor growth in the HT-29 xenograft and in 2. et al.17:3993– 4005. J Physiol 21.30 More. adults.9: colitis-associated cancer development. Nat Rev Cancer 2009. Dias MM. which is known to be an important 2010. et al. Rocha GZ.35. Insulin. De Souza CT. References induced and genetic obesity. plasma TNF-␣ levels and created fertile soil in the colonic obesity.118:560 –570.31:925–943.577: 997–1007. Cell However. visit the online version of 22. 4. IKKbeta links inflamma- serves further investigation. et al. the effect of TNF-␣ neutraliza- Chem 2002.140:197–208. production. J Biol Chem 2002. Giovannucci EL.43– 45 Along this line. Blocking TNF-alpha in mediated CRC.18. Pauli JR. Gastroen- work of inflammatory mediators. immunosuppressors. Phosphorylation of Ser307 in These different results may be related to the environment insulin receptor substrate-1 blocks interactions with the insulin that the animals were bred in or to the differences in the receptor and inhibits insulin action. 9. showed a consistent colorectal cancer.277:1531– inflammatory levels obtained with the different protocols. Renehan AG. Eckmann L. Karin M. and mortality from cancer in a prospectively studied co- epithelium that increased DSS and AOM action.40 – 42 13. Swanton C.94:972– ulcerative colitis31 and induces proinflammatory cytokine 980. No. and at http:// motherapy-induced AMPK activation and antitumoral growth. Altogether. et al.1053/j. of PTP1B and IRS-1 serine phosphorylation.gastro.148:5991–5997. sensitivity in human beings still is controversial and de. ment of complicated Crohn’s disease and was approved 10. 19.109(Suppl):S81–S96. Endo H. Sandhu MS. Reversal of diet-induced TNF-␣ blockers as attractive players in colon cancer che. et al. Overweight. et al.and to the standard therapies. Gao Z.doi. Cancer Res 2011. et al.277:48115– 48121. Zwahlen M. 16. Wu Y. et al. 12. TNF-␣ neutralization in the DSS-induced coli- BASIC AND 2002. homeostasis in an animal model of diet-induced obesity and dia- betes. et al. Ropelle ER.29. cancer growth. 752 FLORES ET AL GASTROENTEROLOGY Vol.59:366 –378. Fujisawa T. the low- 489 – 499.42 The results presented herein extend 2009. Insulin-like growth factor (IGF)-I. Gut 2009. Tumour necrosis factor and cancer. Kitamura K.34 after infliximab treatment 8. N Engl J Med 2003. Werner ED. et al. Minder C. these data. Sinicrope FA.9:361–371. such as cytokines and terol Clin North Am 2002.S. Greten FR.15 This network then 5. chemo. Calle EE.278:2896 –2902. infliximab is recommended for the treat. Walker-Thurmond K. Hans W. Popivanova BK. Smith RA. IRS-1-mediated inhi- for the treatment of ulcerative colitis that fails to respond bition of insulin receptor tyrosine kinase activity in TNF-alpha. 143.32 Our findings showing a reduction in the 7.33. Greten TF. J Clin Invest 2008. Clin Cancer Res Akt/mTOR pathway. as well as the treatment 1.271:665– 668.05.58:1637–1643. et al. Sargent DJ. Kojouharoff G. Epidemiological and of obese mice bearing colon cancer xenografts. Hosono K. Ueno M. regulator of CRC growth. showing that infliximab also prevents obesity. Lee YH. obesity-mediated colon carcinogenesis and point to 20. an epithelial damage model. Araujo EP. Peraldi P. Genetic prognostic and predictive markers in colorectal cancer. Obesity is an indepen- dent prognostic variable in colon cancer survivors. insulin-like PI3K/Akt pathway can be activated by TNF-␣.16. cancer. et al. Metformin amplifies che- Gastroenterology at www. the PI3K/Akt pathway is activated in patients with tions. and colorectal cancer. et al. our findings establish a role for TNF-␣ in liver inflammation and tumorigenesis by enhancing IL-6 and TNF expression. J Natl Cancer Inst 2002. growth factor-I (IGF-I).gastrojournal. Davis RJ. CA Cancer J Clin 2009. Lancet 2004. Clin Note: To access the supplementary material Exp Immunol 1997. Involvement of JNK pathway resistance and colon cancer risk. Ghosh S. c-Jun N-terminal kinase (JNK) neutralization of TNF-␣ is also an effective therapy in mediates feedback inhibition of the insulin signaling cascade. Jemal A. Serine phosphorylation of insulin receptor substrate 1 by inhibitor kappa B kinase complex. It is interesting to note that despite the contradictory 14. and cancer risk: systematic review acts by modulating the cancer microenvironment and and meta-regression analysis. IGF binding protein-3. Park EJ. Nat Rev Cancer in all studies.118:285–296. Aguirre V. Rocha GZ. J Biol results related to colitis.35–37 In accordance with our study. grade inflammatory state observed in obesity increased 3. Endocrinology 2007. et al.2012. 18.045.348:1625–1638. Arq major role of this pathway in obesity-mediated colon Bras Endocrinol Metabol 2009. Center MM. hort of U. Budavari A. TNF-␣ is an important component of the cancer net. Bataille F. Science 1996. Cell 2004. vascular endothelial growth factor. accompanying this article. Giraud J. Osorio-Costa F. tralization should be evaluated. .39 J Biol Chem 2003. TRANSLATIONAL AT various animal models of T-cell–mediated colitis. Worldwide variations in using the TNF-␣ blocker. Clin dx. Hotamisligil GS. et al.107:353–358. suggests a molecular mechanisms aspects linking obesity and cancer. et al.38. 17. Dunger DB. infliximab. although the possible in the promotion of the early stage of colorectal carcinogenesis contribution of anti–TNF-␣ therapy to improve insulin under high-fat dietary conditions. tis. Missing pieces in the NF-kappaB puzzle. tion and tumorigenesis in a mouse model of colitis-associated prevention of CRC in obese individuals with TNF-␣ neu. Yu GY. 3 by an increase in IKK and JNK phosphorylation in diet. Lee JH. et al. Dias MM. Giraud J. Modifiable risk factors for colon cancer. allowing cancer growth. Prada PO. Walther IGF binding proteins. et al. tion or inhibition of IKK␤ was effective in reducing CRC 15. Foster NR. Neutralization of Supplementary Material tumour necrosis factor (TNF) but not of IL-1 reduces inflammation in chronic dextran sulphate sodium-induced colitis in mice. Cell 2010. Giovannucci E. Balkwill F. such as corticosteroids and obesity-induced insulin resistance. Ropelle ER. is still controversial. Thus. these results suggest that the mice reduces colorectal carcinogenesis associated with chronic use of anti–TNF-␣ therapy could decrease both insulin colitis. 1537. insulin resistance with a single bout of exercise in the rat: the role moprevention for obese individuals. Dietary and genetic obesity promote In summary. Infliximab restores glucose Currently. Hwang D. Johnstone E. their biologic interac- over.28 It also has been shown that the 6.16:1884 –1893. Obermeier F. 11. Rodriguez C.53:213–226.

Exp Immunol 1997.F.46:920 –926. Effects of etanercept in 338 –348.1:553–562. Zoncu R.B. Rutgeerts P. Ann Rheum Dis 2005. PI3K/Akt signaling pathway is 85:1316 –1319. mTOR: from growth signal inte- gration to cancer. Paquot N. Li B. Brazil. Sabatini DM. Phosphatidylinositol 3-kinase as a mediator of TNF-induced NF-kappa B activation. Fuss I. Stathopoulos GT. J Immunol 1997. tion and maintenance therapy for ulcerative colitis.96:546 –553. 31:262–266. J Gastroenterol Hepatol 2003. 33. São Paulo. 2005. and G.12:21–35. Desenvolvimento Científico e Tecnológico. Inhibition of insulin. protects synovial cells from nitric oxide induced apoptosis through 44. J Immunol 2000. Ranganath LR. 42. Yang X. Pasparakis M. et al. 2012. Received August 18. N Engl J Med M. Chung JW. J Rheumatol treatment on insulin resistance in patients with rheumatoid arthri- 2006. in chronic dextran sulphate sodium-induced colitis in mice. Kim YJ.67:9825–9834. 41. Powrie F. Cancer Res 2009. 2011. Infliximab reverses 38. Neutralization of cose-dependent insulinotropic polypeptide (GIP) and glucagon-like tumour necrosis factor (TNF) but not of IL-1 reduces inflammation peptide-1 (GLP-1) secretion by octreotide has no effect on post. Rheuma. Vasakos S. Kiortsis DN. Castillo MJ. 34. No increased insulin inhibition of NF-kappaB and PI3 kinase/Akt signal pathway in sensitivity after a single intravenous administration of a recombi- human rheumatoid arthritis fibroblast-like synoviocytes.18:560 –569. Prevention of colitis-associated This study was supported by grants from the Fundação de Amparo carcinogenesis with infliximab. disease by using a caspase-dependent pathway. Inflamm Res 2011. Obermeier F. Biology of incretins: GLP-1 and GIP. Takagi T. Nagaishi T. 24. MD. A prospective study of 39. Kollintza A. 25. Inhibition of Th1 responses plasma C-peptide and colorectal cancer risk in men. diabetes and ageing. Anti-malarial agent artesunate inhibits 766. Giovannucci E. Hong KS.S. 40. J Endocrinol 2007. Arch Intern Med 2006.69: 43. et al.296:G850 –G859. et al. Mauze S. 32. Low levels of tumor necrosis factor involved in colitis-associated carcinogenesis. Hallmarks of cancer: the next genera. et al. nant human tumor necrosis factor receptor: Fc fusion protein in tology (Oxford) 2007. Mavridis AK. Berry J. 29. Kim S. Weinberg RA. Schmidt M. Huang JH. Cancer Prev Res (Phila) 2010. The authors thank Dr Nicola Conran and Sherian Bowyer for 35. Baggio LL. Accepted May 24.27:1743–1750. Gastroenterol- ogy 2001. et al. Lugering A. e-mail: carvalheirajbc@uol. et al. of monocytes recruited to the tumor site. contributed equally to this work. 13083-887 Campinas. J Clin Endocrinol Metab 2000. Handa O. Zhang XH.60:727–734. Pattacini L. Lugering N. Funding 31. Xu J. et al. tion induced by dextran sulfate sodium in tumor necrosis factor- troenterology 2007.3: à Pesquisa do Estado de São Paulo and the Conselho Nacional de 1314 –1333. Neurath MF.194:539 –550. Vincent A. glu. Huang XL. Signaling pathway via tion. Immunity 1994. et al. TNF-alpha-induced production of proinflammatory cytokines via 45. et al. Pollak M.132:2131–2157. Leach MW.R. with CD45RBhi CD4⫹ T cells. He Y. et al. et al. Kojouharoff G. fax: (55) 19-3521-8950.121:1145–1157. et al. Ma J. Infliximab for induc. Bertolo MB. Kanai T. Cell 2011. Moschos C.107:353–358. 36. Milanski M. Predominant pathogenic steatosis and improves insulin signal transduction in liver of rats role of tumor necrosis factor in experimental colitis in mice.144:646 – 674. Morgan LM. Naito Y. Horm Metab Res 1999. C. Feagan BG. Efeyan A. Lefebvre PJ. Chen Q.33:1061–1068. Carvalheira. 2007. et al. TRANSLATIONAL AT Address requests for reprints to: José B. English language editing. obese insulin-resistant patients. Casali B. Enhanced intestinal inflamma- 26. TNF-alpha/NF-kappaB in intestinal epithelial cells may be directly 28.Z. Liao WS. 27. Am J Physiol Gas- alpha increase tumor growth by inducing an endothelial phenotype trointest Liver Physiol 2009. Gas. Hans W. Cancer Res University of Campinas. Reprint requests 164:1355–1363. et al. Clin heparin plasma lipoprotein lipase activity. Tumor necrosis factor-alpha 166:902–908. Onizawa M. et al. Drucker DJ. J Natl Cancer prevents inflammatory bowel disease in scid mice reconstituted Inst 2004. .64:765– 30. Hanahan D. tis and ankylosing spondylitis. Faculty of Medical Sciences-State BASIC AND factor-alpha promotes malignant pleural effusion. alpha deficient Reddy SA. Effects of infliximab phosphoinositide 3-kinase Akt signal transduction. Nat Rev Mol Cell Biol Acknowledgments 2011. Barbuio R.September 2012 OBESITY–MEDIATED COLON CANCER 753 23. patients with the metabolic syndrome. Eur fed a high-fat diet. Xu H. involved in the pathogenesis of ulcerative colitis. Cates J. et al. et al. Tumor necrosis Department of Internal Medicine.353:2462–2476. Infliximab induces Conflicts of interest apoptosis in monocytes from patients with chronic active Crohn’s The authors disclose no conflicts. Bernstein LE. et al. Sandborn WJ. Beety JM.

8 33.4 ⫾ 0.5a 3.05 1. rate constant for glucose disappearance in the insulin tolerance test.2 7.3 ⫾ 2.2 ⫾ 4.15 ⫾ 0.4 3.4 ND 0.2 3.8 ⫾ 0.4a TNF-␣. g High-fat diet.5 AIN.5 ⫾ 0. Supplementary Figure 1. control (chow-fed) mice. Kitt.1 ⫾ 0.0 ⫾ 0.3 ⫾ 0.5a Epididymal fat.6 ⫾ 9. Supplementary Figure 2.6a 63. (A) Macroscopic changes in colonic tissues.5 ⫾ 2.0 ⫾ 0. aP ⬍ . Characteristics of Chow-Fed and HFD-Fed C57BL6/J and SCID Mice and Chow-Fed C57BL6/J and ob/ob Mice SCID Characteristics CTL HFD C57 ob/ob CTL HFD Body weight.02a Insulin.7 ⫾ 0.e1 FLORES ET AL GASTROENTEROLOGY Vol. Values are presented as mean ⫾ standard deviation. g 4. n ⫽ 8 in each group.46 ⫾ 0. g Casein 202 200 Sucrose 100 100 Cornstarch 397 115.57 ⫾ 1.6 169.1 12.1a 6. American Institute of Nutrition.8 ⫾ 0. (B) Tumor incidence and size of colons removed at the end of the protocol from CTL and HFD-fed mice.1a NOTE.5a 0. Diet-induced obesity promotes colon carcinogenesis in mice treated with AOM alone.0 ⫾ 0. Supplementary Table 2.0 ⫾ 0.5a 19.1 11. 3 Supplementary Table 1.0 ⫾ 1.1 2.1 3.05 ⫾ 0.1a 22. pg/mL 53. ng/mL 3.6 ⫾ 1. CTL. ng/mL 2.5a 0.4a Kitt.4 ⫾ 0.05 vs CTL.53 ⫾ 0.5 132 Lard — 312 Soybean oil 70 40 Cellulose 50 50 Mineral mix AIN-93 35 35 Vitamin mix AIN-93 10 10 L-Cystine 3 3 Choline 2.4a 4.2 55.1 10.3 ⫾ 2. Components of Standard Chow and High-Fat Diet Ingredients Standard chow.2 ⫾ 0.8 ⫾ 5. *P ⬍ .01 0. %/min 5.6 ⫾ 0.0 ⫾ 0.5a 0.8 104 ⫾ 9. Western blot of colonic time-course of insulin-induced Akt phosphorylation in lean mice. 5 mm.5 Dextrinated starch 130.01 1.5 ⫾ 1.1 ⫾ 0.753. g 24.15a Leptin.5a 1. No. . Results are means ⫾ SEM (n ⫽ 8).4 ⫾ 0.2 ⫾ 3.2 ⫾ 7.5 2.01 vs respective control group.2 ⫾ 0. 143. Scale bars.1a 2.1 8.6 28.

05 vs control. (E and F) Representative microphotographs and quantification of (E) Ki67 and (F) apoptotic staining on HT-29 xenografts sections. (A) Dosage of insulin from the serum of CTL. immunoblotted with pAkt and Akt.September 2012 OBESITY–MEDIATED COLON CANCER 753. Octreotide reduces insulin levels with no effects on tumor growth. and apoptosis. Akt. (D) Western blots showing HT-29 xenografts lysates from CTL. IRS-1. treated or not with OCT. (B) Insulin-induced IR␤. and ␤-tubulin phosphorylation and expression in muscle lysates from HFD-fed mice. .e2 Supplementary Figure 3. HFD-fed mice. mean ⫾ SEM. #P ⬍ . 4 fields per tumor section.05 vs HFD. treated or not with OCT. HFD. *P ⬍ . (C) Tumor growth was measured for 30 days in SCID mice maintained on normal chow or HFD for 8 weeks before tumor inoculation with HT-29 cells (1 ⫻ 106 cells/mouse) and kept on normal chow or HFD for 30 days after inoculation and with the indicated treatment. and HFD and octreotide-treated (OCT) mice. proliferation.

or apoptosis. 143. (B) Insulin-induced IR␤. *P ⬍ .05 vs control. Pioglitazone reduces insulin levels with no effects on tumor growth. . proliferation. 4 fields per tumor section. HFD. #P ⬍ . 3 Supplementary Figure 4. Akt. No.753. and HFD and pioglitazone-treated (PIO) mice. IRS-1.05 vs HFD. (C) Tumor growth was measured for 30 days in SCID mice maintained on normal chow or HFD for 8 weeks before tumor inoculation with HT-29 cells (1 ⫻ 106 cells/mouse) and kept on normal chow or HFD for 30 days after inoculation and with the indicated treatment. treated or not with PIO. mean ⫾ SEM. HFD-fed mice. treated or not with PIO. (A) Dosage of insulin from the serum of CTL. (E and F) Representative microphotographs and quantification of (E) Ki67 and (F) apoptotic staining on HT-29 xenograft sections.e3 FLORES ET AL GASTROENTEROLOGY Vol. The results are means ⫾ SD from 5– 8 mice per group. immunoblotted with pAkt and Akt. (D) Western blots showing HT-29 xenograft lysates from CTL. and ␤-tubulin phosphorylation and expression in muscle lysates from HFD-fed mice.

7a NOTE. aP ⬍ .e4 Supplementary Table 3.1 ⫾ 0.9 ⫾ 0.67 ⫾ 0.18 1. n ⫽ 4 –5 in each group. rate constant for glucose disappearance in the insulin tolerance test.05 vs vehicle-treated group.6 5. Characteristics of HFD-Fed Mice Treated With Vehicle or Infliximab HFD ⫹ vehicle HFD ⫹ infliximab TNF-␣. %/min 1.4 ⫾ 31.September 2012 OBESITY–MEDIATED COLON CANCER 753.43 Kitt. Values are presented as mean ⫾ standard deviation.4a Leptin. pg/mL 133.5 ⫾ 23. ng/mL 1.9 38. Kitt.95 ⫾ 0. .