Article Text

Download PDFPDF

Leptin acts as a growth factor for colorectal tumours at stages subsequent to tumour initiation in murine colon carcinogenesis
  1. Hiroki Endo1,
  2. Kunihiro Hosono1,
  3. Takashi Uchiyama1,
  4. Eiji Sakai1,
  5. Michiko Sugiyama1,
  6. Hirokazu Takahashi1,
  7. Noriko Nakajima2,
  8. Koichiro Wada3,
  9. Kiyoshi Takeda4,
  10. Hitoshi Nakagama5,
  11. Atsushi Nakajima1
  1. 1Division of Gastroenterology, Yokohama City University School of Medicine, Yokohama, Japan
  2. 2Department of Pathology, National Institute of Infectious Diseases, Tokyo, Japan
  3. 3Department of Pharmacology, Graduate School of Dentistry, Osaka University, Osaka, Japan
  4. 4Laboratory of Immune Regulation, Department of Microbiology and Immunology, Graduate School of Medicine, Osaka University, Osaka, Japan
  5. 5Biochemistry Division, National Cancer Center Research Institute, Tokyo, Japan
  1. Correspondence to Dr Atsushi Nakajima, 3-9 Fuku-ura, Kanazawa-ku, Yokohama 236-0004, Japan; nakajima-tky{at}umin.ac.jp

Abstract

Background and aims Obesity increases the risk of colorectal cancer (CRC). Serum leptin levels are markedly elevated in obese individuals, but the involvement of leptin in CRC growth remains unclear. We explored the hypothesis that leptin signalling regulates the growth of CRC, by examining the effects of leptin deficiency on murine colon tumour growth.

Methods We used genetic (leptin-deficient and leptin receptor-deficient) models of obesity and investigated carcinogen-induced colon polyp formation and cell proliferation in the colonic epithelium. Colonic tissues and cell lines were analysed by histopathology and molecular-biology methods.

Results A significant increase in the proliferative activity of normal colonic epithelial cells was observed in the obesity model; on the other hand, significant decrease of tumour cell proliferation was observed in leptin-deficient tumours, and tumour growth was dramatically inhibited in leptin-deficient and leptin-receptor-deficient mice despite the animals exhibiting severe obesity. Notably, a marked increase of the leptin receptor (ObR) expression levels was observed in colon tumours as compared to the normal epithelium. Nuclear β-catenin staining was pronounced in all tumours, irrespective of leptin deficiency, whereas altered cellular localisation of β-catenin was not observed in the normal colonic epithelial cells. In vitro, β-catenin knockdown decreased ObR expression, and stimulation of recombinant Wnt increased ObR expression. In addition, the proliferative and survival effects of leptin were found to be mediated by the ObR/signal transducer and activator of transcription 3 (STAT3) signalling in colon tumours.

Conclusions Our findings indicate that leptin is important for CRC growth in obesity, and acts as a growth factor for CRC at stages subsequent to tumour initiation in colorectal carcinogenesis. Thus, inhibition of leptin signalling may be an effective strategy for therapy and prevention of colonic adenoma and cancer, which show activation of Wnt signalling.

  • Leptin
  • colorectal cancer
  • obesity
  • tumour growth
  • STAT3
  • colon carcinogenesis
  • colorectal cancer
  • dietary - colon cancer
  • molecular carcinogenesis
  • obesity
View Full Text

Statistics from Altmetric.com

Significance of this study

What is already known about this subject?

  • Epidemiological studies have revealed that obesity raises the risk of colon adenoma and colorectal cancer (CRC), and the results of animal experiments suggest a link between obesity and CRC.

  • Obesity is strongly associated with adipose tissue dysfunction and altered serum levels of adipokines, including leptin.

  • Data concerning the effect of leptin on CRC development are still contradictory.

What are the new findings?

  • The proliferative activity of the normal colonic epithelial cells was significantly increased in the obese model, but tumour cell proliferation was significantly lower in leptin-deficient tumour, and tumour growth was dramatically inhibited in the leptin-deficient and leptin receptor-deficient mice despite their severe obesity.

  • Leptin receptor (ObR) expression levels were increased markedly in colon tumours as compared with the normal epithelium, and, in vitro, β-catenin knockdown decreased ObR expression and stimulation of recombinant Wnt increased ObR expression.

  • The ability of leptin to regulate CRC growth was mediated by colonic leptin signalling via the ObR/signal transducer and activator of transcription 3 (STAT3) pathway.

How might it impact on clinical practice in the foreseeable future?

  • Leptin acts as a growth factor for CRC at stages subsequent to tumour initiation in colon carcinogenesis.

  • Our findings suggest that leptin signalling is a direct pathway that is crucial for CRC growth, which is a reasonable explanation for the tendency of CRC to be more aggressive in obese individuals known to show elevated serum leptin levels.

  • Inhibition of leptin signalling may be efficacious for therapy and prevention of colonic adenoma and cancer with Wnt signalling activation.

Introduction

Obesity increases the risk of not only cardiovascular disease and type 2 diabetes mellitus,1 but also of various types of cancers.2 3 In particular, obesity has been shown to be associated with advanced progression of colorectal cancer (CRC).4 For a number of cancers, including CRC, the risk of the disease is also elevated in individuals with obesity.5 Epidemiological studies have revealed that obesity, especially visceral adipose tissue, raises the risk of colon adenoma6 and CRC,7 and the results of animal experiments suggest a link between obesity and CRC.7 Obesity is strongly associated with adipose tissue dysfunction and altered serum levels of adipokines, which might underlie the risk of CRC, but no definitive conclusions have been reached. Leptin, a 16-kDa product of the ob gene involved in energy balance and regulation of food intake,8 is secreted predominantly in adipose tissue and is correlated with the percentage of body fat.9 Serum leptin levels are markedly elevated in obese individuals,10 and thus we hypothesised an association between this adipokine and increased risk of CRC.

Data concerning the effect of leptin on CRC development are contradictory and difficult to interpret.11–20 In humans, several case–control studies have shown an elevated risk of CRC associated with high serum leptin level,11 12 although in some studies, no elevation of the serum leptin levels were found in patients with CRC.13 14 In experimental studies, although there has been general agreement that leptin acts as a growth factor for colon cancer cells in vitro,15–17 conflicting results have been reported from in vivo studies that have investigated the effects of leptin on rodent colonic epithelial cell proliferation15 18 and colon carcinogenesis.19 20 Overall, the role of leptin in CRC induction and growth remains unclear.

Here, we explored the hypothesis that leptin signalling might regulate the growth of CRC to account for the clinical observation that obesity correlates with increased progression of CRC. We confirmed that ablation of leptin or leptin receptor (ObR) markedly inhibited the growth of colon tumours. Furthermore, we found that the ability of leptin to regulate CRC growth was mediated by colonic leptin signalling via the ObR/signal transducer and activator of transcription 3 (STAT3) pathway. This suggests that leptin signalling is a direct pathway that is crucial for CRC growth, which is a reasonable explanation for the tendency of CRC to be more aggressive in obese individuals who are known to show elevated serum leptin levels.

Materials and methods

Animals and tumour induction

Six-week-old male C57BL/6J-ob/ob mice, C57BL/KsJ-db/db mice, and their respective control C57BL/6J and C57BL/KsJ mice (wild-type; WT) were obtained from the Jackson Laboratory (Bar Harbor, Maine, USA). The animals were fed either a normal diet (ND) or high-fat diet (HFD) until the end of the study (Supplementary figure 1). The compositions of the ND (MF; Oriental Yeast Co., Tokyo, Japan) and the HFD (High Fat Diet 32; CLEA Japan Inc., Tokyo, Japan) have been described previously.21

The protocols for azoxymethane (AOM)-induced aberrant crypt foci (ACF) or the tumour model were essentially as described previously.22 Briefly, mice were given 2- or 6-weekly intraperitoneal (i.p.) injections of 10 mg/kg AOM (Sigma, St. Louis, Missouri, USA) and were killed at 6 or 21 weeks following the initiation of AOM injection (Supplementary figure 1). Macroscopic tumours were counted and measured with a caliper. To facilitate the small tumour counting, the colons were stained with 0.2% methylene blue solution and were observed using stereomicroscopy. The number of ACF was counted as described previously.22 We repeated each experiment three times to confirm the reproducibility of our results.

Leptin treatment

Ob/ob mice were divided into two groups of eight mice each, injected with either leptin or vehicle. Leptin-treated mice received daily i.p. injections of 2 μg murine recombinant leptin protein (Peprotech, Rocky Hill, New Jersey, USA) per gram of body weight for 6 weeks. Vehicle-treated mice received a 0.9% saline endotoxin-free solution for 6 weeks, which was also used for leptin injection.

Assay for proliferation and apoptosis

The entire colon was removed, gently flushed with saline to remove any faecal contents, opened longitudinally, and fixed in 10% neutralised formalin. Paraffin sections were prepared at 3 μm thickness, and stained with H&E. We evaluated the 5-bromo-2-deoxyuridine (BrdU) (BD Biosciences, Franklin Lakes, New Jersey, USA) labelling index to determine the proliferative activity of the colonic epithelial cells as described previously.21 The apoptotic tumour cells were stained using a transferase deoxytidyl uridine end labelling (TUNEL) staining kit according to the manufacturer's instructions (Wako Pure Chemical, Osaka, Japan).

Immunohistochemistry, immunofluorescence and immunoblotting

Paraffin-embedded sections were deparaffinised and subjected to immunohistochemical staining with primary antibodies using a Histofine kit (Nichirei, Tokyo, Japan) in accordance with the manufacturer's instructions. Nuclear counterstaining was performed with haematoxylin. In the negative controls, the primary antibody was replaced by non-specific, non-immune immunoglobulin of the same isotype at an equivalent final concentration. For immunofluorescence of the cells, the cells grown on coverslips were paraformaldehyde-fixed and permeabilised with 100% ethanol at −20°C. Fixed cells were incubated with the primary antibodies and stained with Alexa Fluoro-conjugated secondary antibodies (Molecular Probes, Eugene, Oregon, USA). Nuclei were stained by 4′-diamidine-2′-phenylindole hydrochloride (DAPI; Molecular Probes). Confocal laser scanning microscopic images were then generated (Olympus, Tokyo, Japan).

Protein extracts were separated using SDS/PAGE, and the separated proteins were transferred to a polyvinylidene difluoride (PVDF) membrane (Amersham, London, UK). The membranes were probed with primary antibodies and glyceraldehyde-3-phosphate dehydrogenase (GAPDH; Trevigen, Gaithersburg, Maryland, USA). Horseradish-peroxidase-conjugated secondary antibodies and the enhanced chemiluminescence (ECL) detection kit (Amersham) were used for the detection of specific proteins.

Antibodies used were anti-p-ObR, anti-ObR (Santa Cruz Biotechnology, Santa Cruz, California, USA), anti-p-STAT3, anti-STAT3, anti-cleaved caspase-3 (Cell Signaling Technology, Danvers, Massachusetts, USA), and anti-β-catenin (BD PharMingen, San Diego, California, USA).

RT-PCR analysis

Total RNA was extracted from the colonic epithelium using the RNeasy Mini Kit (Qiagen, Hilden, Germany). For real-time reverse transcription polymerase chain reaction (RT-PCR), total RNA was reverse-transcribed into cDNA and amplified using real-time quantitative PCR using the ABI PRISM 7700 System (Applied Biosystems, Foster City, California, USA). Probes and primer pairs specific for ObRb and β-actin were purchased from Applied Biosystems. The concentrations of the target genes were determined using the delta-delta Ct method and the values were normalised to those of the internal control. Primer sequences are listed in the Supplementary Methods.

Cell culture and transfection

Colon cancer cell line SW480 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS), while human embryonic kidney cells HEK 293 cells were grown in DMEM. Transfection of siRNA was performed by using Lipofectamine 2000 (Invitrogen, Carlsbad, California, USA). The cells transfected with β-catenin siRNA (Invitrogen) were harvested at 48 h after transfection, and immunoblotting and RT-PCR analysis were performed. To confirm the Wnt3a requirement for ObR expression, the cells were grown under recombinant Wnt3a protein (Stem Cell Technologies, Vancouver, British Columbia, Canada) supplementation.23 The cells stimulated with 160 ng/ml of Wnt3a were harvested at 48 h, and RT-PCR and immunofluorescence analyses were performed.

Statistical analysis

Statistical analysis for comparisons of the number of ACF, the number and size of colon polyps, the BrdU labelling index, and the blood test results were conducted using the Mann–Whitney U test. Other statistical analyses were performed using the Student t test. Values of p<0.05 were regarded as denoting statistical significance.

Results

Leptin regulates colorectal tumour growth, but does not stimulate the formation of ACF

To investigate the impact of leptin on obesity-related colorectal carcinogenesis and to determine whether it might act as a tumour promoter, we examined the formation of chemically induced ACF, as a marker of experimental colorectal carcinogenesis,24 and of polyps in the colon specimens. The experimental protocol based on AOM treatment is shown in Supplementary figure 1. We used both dietary (HFD) and genetic (leptin-deficient; ob/ob) models of obesity for comparison with lean controls. To avoid the possibility that the differences in tumourigenecity of AOM could be due to the effects of dietary alterations on AOM metabolism, we examined the ACF model by alternating the diet 1 week after the last injection of AOM (Supplementary figure 2). The body weights and visceral fat were much higher in the ob/ob and WT mice fed a HFD than in the WT mice fed a ND (Supplementary figure 3). As expected, HFD exposure increased the serum leptin levels in WT mice; meanwhile, the levels of insulin and cholesterol were significantly higher in ob/ob mice than in WT mice, and there was no significant difference in the serum adiponectin level between the WT and ob/ob mice (figure 1A). We found that BrdU labelling index of the normal mucosa was significantly higher in the obese than in lean WT mice (figure 1B,C). The number of ACF in the obesity model was also significantly higher than in lean controls (figure 1D,E, Supplementary figure 3). These results suggest that obesity enhanced the development of early-stage colorectal carcinogenesis irrespective of leptin signalling.

Figure 1

Leptin is not involved in the early-stage of colorectal carcinogenesis in obesity. (A) AOM-treated wild-type (WT) and ob/ob mice were bled and the levels of metabolic factors in serum were determined by ELISA. Results are averages ± SEM. (n=8). ND, normal diet; HFD, high-fat diet. (B) BrdU incorporation in normal colonic epithelial crypts of WT and ob/ob mice. Scale bars=100 μm. (C) BrdU labelling indices of colonic normal mucosa of AOM-treated WT and ob/ob mice. Results are averages ± SEM. (n>10). *p<0.01. **p<0.005. (D) Stereoscopic observations of ACF (arrows) in colonic tissues of WT and ob/ob mice. The samples were stained with 0.2% methylene blue. Scale bars=100 μm. (E) ACF multiplicity. Results are averages ±SEM. (n=10). *p<0.05. **p<0.001. ACF, aberrant crypt foci; AOM, azoxymethane; BrdU, bromodeoxyuridine; WT, wild type.

Therefore, we focused on the later stages of cancer progression and observed that leptin deficiency dramatically decreased the tumour sizes despite ob/ob mice developing overt obesity (figure 2A–C,E). These findings were closely correlated with the serum levels of leptin, but were not associated with the dietary conditions. It was noteworthy that the absence of leptin had a stronger effect on colonic tumour growth than either HFD exposure or hyperinsulinaemia, which have also been reported to increase the risk of CRC.25 26 Tumour multiplicity was also reduced more in ob/ob than in WT mice, but was comparable under HDF conditions (figure 2A,D). Supplementary table 1 summarises the histological findings of the tumours in WT and ob/ob mice.

Figure 2

Leptin regulates AOM-induced colon tumour growth. (A) Macroscopic findings of colon tumours. Arrows indicate large tumours. Scale bars=1 cm. ND, normal diet; HFD, high-fat diet. (B) Macroscopic (left panel) and stereomicroscopic (right panel) findings of small colon tumours. Scale bars=1 mm (left panel) and 200 μm (right panel). (C) Tumour size. Results are averages ±SEM. (n=10). *p<0.05. **p<0.005. (D) Tumour multiplicity in WT and ob/ob mice fed ND or HFD. Results are averages ±SEM. (n=10). *p<0.05. (E) Histogram showing size distribution of tumours. (F) BrdU incorporation in colon tumour of WT and ob/ob mice. Scale bars=100 μm. (G) BrdU labelling indices in colon tumours of the AOM-treated WT and ob/ob mice. Results are averages ±SEM. (n>10). *p<0.05. **p<0.005. AOM, azoxymethane; BrdU, bromodeoxyuridine; WT, wild type.

Next, we analysed cell proliferation and apoptosis in WT and leptin-deficient tumours to explain the differences in tumour growth. We found that BrdU incorporation was significantly lower in tumours of ob/ob mice than WT mice (figure 2F,G), which was consistent with decreased tumour growth in the absence of leptin. Interestingly, TUNEL and cleaved caspase-3 revealed a reciprocal increase in the apoptotic response of the colon tumours between WT and ob/ob mice (Supplementary figure 4), which suggests that tumour cell survival also relies on leptin signalling. Taken together, these data indicate that leptin enhances tumour proliferation, but it might not exert the same effects on normal mucosa and premalignant lesions.

Leptin receptors are required for colorectal tumour growth

To clarify why the effects of leptin are limited to tumour cells, we investigated the roles of ObR in colon. We found strong ObR expression in tumour cells, but little expression in normal epithelial cells of the colonic mucosa (figure 3A,B). Expression of the long form of ObR (ObRb) mRNA was found to be significantly higher in colon tumours than in normal mucosa (figure 3C). Therefore, differences in cell proliferation dependence on leptin between tumours and normal mucosa might be explained by altered expression of ObRb.

Figure 3

Leptin receptors are required for colorectal tumour growth. (A) and (B) Section of colon from an AOM-treated mouse showing a representative tumour protruding into the colonic lumen was stained with antibodies to ObR. Magnified views of the boxed area confirms the presence of ObR-positive cells in the cytoplasm of tumour cells (upper panel), but scarce expression of ObR in normal epithelial cells (lower panel) of the colonic mucosa. Scale bars=400 μm (left panel), 10 μm (right, upper panel), and 100 μm (right, lower panel). (C) Relative expression level of ObRb mRNA in normal colonic mucosa and tumours of AOM-treated mice, analysed by real-time PCR. Results are averages ±SEM. (n=5). *p<0.005. (D) Expression of ObRb in the isolated colonic mucosa from WT and leptin-receptor-deficient db/db mice. Results are averages ±SEM. (n=5). *p<0.005. (E) Macroscopic findings of the colon tumours. Arrows indicate large tumours. Scale bars=1 cm. (F) Tumour multiplicity in WT and db/db mice subjected to induction of colon tumours. Results are averages ±SEM. (n=10). *p<0.05. (G) Tumour size in WT and db/db mice. Results are averages ±SEM. (n=10). *p<0.0001. AOM, azoxymethane; WT, wild type.

Furthermore, to elucidate the contribution of ObRb to colonic tumourigenesis, we used mice with ObRb deletion (db/db mice).27 As expected, db/db mice, which exhibited the same obese phenotype as the ob/ob mice (Supplementary figure 5), were devoid of ObRb mRNA in colonic mucosa, whereas WT mice expressed ObRb (figure 3D). In the tumour experimental protocol (Supplementary figure 1B), we observed a significant increase in the frequency and size of tumours in WT mice as compared to db/db mice (figure 3E–G); meanwhile, there was no significant difference in tumour size and multiplicity between the db/db mice fed ND and those fed a HFD (Supplementary figure 6). Supplementary table 2 summarises histological findings of tumours in WT and db/db mice. On the other hand, the number of ACF in db/db mice was significantly higher than in WT mice (Supplementary figure 7). These results suggest that epithelial ObRb is required for transduction of tumour-promoting signals from leptin.

Wnt signalling stimulates expression of ObRb

We explored the mechanism of ObRb expression in tumours. Expression of ObRb was strong in the tumour epithelium where Wnt signalling was activated. Frequent gene mutations of β-catenin and altered cellular localisation of the protein are features of AOM-induced colon tumours in mice.28 29 Using immunohistochemical analysis, we examined the expression of β-catenin in colon tumours induced by AOM in comparison with that in the adjacent normal mucosa. Cytoplasmic and nuclear β-catenin staining was pronounced in all tumour tissues of WT and ob/ob mice, whereas antibody binding was limited to the membranes at the intercellular borders in normal epithelial cells (Supplementary figure 8). Importantly, the stabilised β-catenin in the nuclei was observed in tumour, irrespective of leptin deficiency. To elucidate the roles of Wnt signalling activation in the regulation of leptin/ObRb signalling, we examined the effects of β-catenin knockdown on leptin/ObRb pathway in the human SW480 colon cancer cell line. Transfection of siRNA for the β-catenin gene markedly reduced the protein expression level (figure 4A). Notably, Wnt signalling inhibition by β-catenin knockdown markedly reduced ObRb expression (figure 4A). We confirmed that the ObRb mRNA level was significantly reduced in the β-catenin siRNA-transfected SW480 cells (figure 4B). Reductions of ObRb protein expression by transfection of β-catenin siRNA was also observed in other colon cancer cell lines (Supplementary figure 9). It has been shown that the Wnt/β-catenin pathway can be stimulated in HEK293 cells by addition of Wnt3a.30 We tested ObRb expression level following Wnt3a stimulation in HEK293 cells that normally contain trace amounts of nuclear β-catenin. Wnt3a-stimulated HEK293 cells showed a marked increase in ObR expression (figure 4C,D). These results are consistent with the increased expression levels of ObR in tumours as compared with those in normal mucosa (figure 3A–C), and indicate that the Wnt signalling activates ObRb expression in colonic epithelium.

Figure 4

Wnt signalling increases the expression levels of ObRb. Western blot (A) and RT-PCR (B) analyses of ObRb expression in SW480 cells transfected with β-catenin siRNA (50 nM). Samples were prepared 48 h after transfection. ObRb mRNA and protein expression was decreased by β-catenin siRNA. GAPDH and β-actin are shown as the loading controls. (C) RT-PCR analysis of ObRb expression in HEK293 cells with and without Wnt3a stimulation. ObRb mRNA expression levels were increased by Wnt3a stimulation. β-actin is shown as a loading control. (D) Immunofluorescence of β-catenin (red) and ObR (green) in HEK293 cells treated with recombinant Wnt3a. Wnt3a induced translocation of β-catenin from the cytoplasm to the nuclei and induced the expression of ObR. DAPI, 4′-diamidine-2′-phenylindole hydrochloride; GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Leptin activates STAT3 signalling to promote colorectal tumour growth

Phosphorylation of Tyr1138 in ObRb induces STAT3 activation.31 Increased amounts of phosphorylated Tyr1138-ObRb and STAT3 (p-STAT3) were observed in tumours as compared with those in normal mucosa (figure 5A). These data suggest that leptin exerts a stimulatory action on colon tumours through the ObRb/STAT3 pathway.

Figure 5

Leptin increases STAT3 phosphorylation in colon tumours. (A) Immunoblot analysis for Tyr1138-phosphorylated ObRb and phosphorylated STAT3 (p-STAT3) in normal colonic mucosa and tumours of AOM-treated mice. (B) Paraffin-embedded sections of tumour-containing colons from WT and ob/ob mice were stained with H&E (upper panel) and with anti-p-STAT3 (lower panel). Insets in lower panels demonstrate nuclear localisation of p-STAT3 in WT mice, which was absent from tumours in ob/ob mice. Scale bars=100 μm. ND, normal diet; HFD, high-fat diet. (C) Normal colon mucosa from AOM-treated WT and ob/ob mice was stained with anti-p-STAT3. Scale bars=100 μm. (D) Expression of cell-cycle and apoptosis regulators in isolated colonic tumours from WT and ob/ob mice. Relative mRNA expression levels were determined by real-time PCR. Results are averages ±SEM. (n=6). *p<0.05. **p<0.01. AOM, azoxymethane; WT, wild type.

To determine the contribution of leptin to changes in tumour cell proliferation and survival, we analysed colon tumours of WT and ob/ob mice for activation of STAT3 and the expression of its target genes. Immunohistochemical analysis revealed tumour cell nuclear localisation of p-STAT3 in colon tumour cells in WT mice, while this signal was almost completely absent from similar tumours in ob/ob mice (figure 5C). The frequency of p-STAT3-positive cells was significantly higher in tumours of WT mice fed a HFD than in those of the mice fed a ND (figure 5C, Supplementary figure 10), closely matching the increase in serum leptin levels (figure 1A). Meanwhile, p-STAT3-positive cells were almost undetectable in normal mucosa of WT and ob/ob mice (figure 5B). Importantly, the lack of STAT3 activation in colonic mucosa coincided with the lack of colonic leptin signalling, namely, lack of functional leptin (figure 1A) or lack of colonic ObR expression (figure 3A–C). Therefore, our data indicate that leptin is a crucial STAT3 activator in colonic epithelium during tumour growth. Next, we analysed the STAT3-mediated proliferative response in WT and leptin-deficient tumours. To do that, we investigated the expression of cell-cycle genes and of the cyclin-dependent kinase (Cdk) inhibitor p21cip in tumours. We found that mRNA expressions of cyclin D1, c-Myc, cyclin B1, cyclin E and cdc2 were increased to a greater degree in WT mice than in ob/ob mice (figure 5D). This suggested a stimulatory effect of STAT3 on the cell cycle, and this observation was consistent with the downregulation of the Cdk inhibitor p21cip in tumours of WT mice. Furthermore, we observed elevated expression levels of Bcl-XL and survivin in WT mice as compared with those in ob/ob mice (figure 5D). These results suggest that impaired induction of Bcl-XL and survivin protein expression may account for the increased rate of apoptosis observed in leptin-deficient ob/ob mice. Collectively, these results strongly support the notion that the STAT3-associated proliferative and antiapoptotic effects are important for tumour epithelia.

Exogenous leptin compensates for suppressed tumour growth in leptin-deficient mice

We found that continuous treatment with recombinant leptin during the late-stage of CRC (figure 6A) resulted in an increase of tumour sizes (figure 6B–D), whereas tumour multiplicity was not affected (figure 6E). Supplementary table 3 summarises the histological findings of tumours. As expected, treatment with recombinant leptin resulted in elevated serum levels of leptin (figure 6F). Importantly, leptin supplementation enhanced STAT3 phosphorylation in colonic tumours (figure 6G). Thus, leptin signalling can increase tumour size without affecting tumour multiplicity, which has an impact on tumour growth.

Figure 6

Leptin signalling stimulates tumour growth. (A) Scheme of treatment with recombinant leptin during the late stage of CRC growth. Mice were injected i.p. with 2 μg recombinant leptin, or control saline every day from 15 weeks after initial AOM injection. (B) Macroscopic findings of colon tumours. Arrows indicate large tumours. Scale bars=1 cm. (C) Tumour size. Results are averages ±SEM. (n=6). *p<0.05. (D) Histogram showing size distribution of colon tumours. (E) Tumour multiplicity. Results are averages ±SEM. (n=6). (F) Vehicle- or leptin-treated ob/ob mice were bled and the levels of leptin (left panel) and insulin (right panel) in serum were determined by ELISA. Results are averages ±SEM. (n=6). *p<0.001. (G) Paraffin-embedded sections of tumour-containing colons of vehicle- and leptin-treated ob/ob mice stained with anti-p-STAT3. Scale bars=100 μm. AOM, azoxymethane; CRC, colorectal cancer.

Discussion

The existence of a relationship between obesity-related factors and CRC has been speculated upon in recent years, but no definitive conclusions have been reached. The present investigation to elucidate the precise mechanisms involved was necessary because of the major clinical implications. We identified a novel mechanism to explain how leptin deficiency might suppress colon tumour growth, even in the presence of marked increase in the levels of other obesity-related factors. Our finding suggests that leptin is a crucial factor for colon tumour growth among the various obesity-related factors. We demonstrated an increase in the proliferative activity of the normal colonic epithelial cells and ACF formation in the obesity model but, unexpectedly, tumour growth was inhibited dramatically in the leptin-deficient obesity model, indicating the importance of leptin signalling for colon tumour growth. Taken in combination, our data indicate that leptin acts as growth factor for CRC at stages subsequent to cancer initiation.

Previous studies have provided much evidence of an association between metabolic factors and increased risk of colorectal carcinogenesis.7 Therefore, we hypothesised at first that ob/ob mice, which have obese metabolic phenotypes with elevated levels of insulin, glucose and lipid, would show increased susceptibility to CRC development as compared to their lean littermates. However, to our surprise, we found that ob/ob mice developed far smaller tumours than the corresponding WT mice, despite the animals exhibiting severe obesity. In contrast, administration of a HFD to WT mice resulted in increased tumour sizes, despite the finding that levels of various obesity-related metabolic factors, with the exception of leptin, in these mice were not as high as those in ob/ob mice. These results strongly indicate that, in vivo, leptin is important for the regulation of colon tumour growth, irrespective of obesity. Furthermore, these results also explain that CRC does not grow under leptin-deficient conditions, regardless of the serum insulin levels. Leptin-deficient mice exhibited few and small tumours despite a high intake of dietary fat. These findings suggest that leptin is a crucial factor for CRC development, regardless of dietary composition. Adiponectin has also been reported to influence colorectal carcinogenesis. Recently, we have demonstrated that adiponectin deficiency might promote the development of CRC only under HFD conditions, using adiponectin-knockout mice.21 Furthermore, a human epidemiological study has shown that decreased levels of plasma adiponectin are associated with increased risk of CRC.32 In a cell model study, adiponectin has been shown to block leptin-induced colon epithelial cell proliferation.33 However, there was no significant difference in the serum adiponectin level between the WT and ob/ob mice in the present study. We speculate that there are many factors that influence colon carcinogenesis in an obesity background, and adiponectin may be one of these factors. Further studies in animal CRC models are necessary to address the interaction between adiponectin and leptin.

Using genetic models, we demonstrated that leptin is an important regulator of CRC development. However, leptin signalling did not have a significant effect on promotion of premalignant lesions in the CRC model, because its absence did not alter the number of ACF. These data indicate that leptin does not act as a growth-promoting agent at an early stage of colon carcinogenesis. We observed that the number of ACF was significantly greater in ob/ob and db/db mice than in WT mice, which suggests that metabolic factors other than leptin act as promoters of early-stage colon carcinogenesis. Furthermore, leptin signalling did not promote normal colonic epithelial cell proliferation either. Why did leptin enhance tumour cell proliferation, but not induce formation of ACF or proliferation of normal colonic mucosa? Here, we noted a difference in ObR between tumours and normal mucosa. A marked increase in ObR expression level was observed in tumours as compared with that in the normal mucosa. Carcinogen-induced tumours frequently show mutation of β-catenin that leads to stabilisation and nuclear translocation of β-catenin, thereby activating the Wnt pathway. On the other hand, mutation and altered cellular localisation of β-catenin are not observed in normal mucosa or ACF.34 Based on this evidence, we hypothesised that activation of the Wnt pathway not only triggers the formation of colon tumours, but also induces the expression of ObR in colon tumours. To elucidate the roles of Wnt signalling activation on regulation of ObR expression, we examined the effects of β-catenin knockdown on ObR expression in colon cancer cell lines, and confirmed decrease in ObR mRNA and protein levels. Furthermore, we also confirmed increased ObR expression levels in exogenous Wnt-stimulated HEK293 cells. Thus, we propose that Wnt signalling contributes to the upregulation of ObR in colonic epithelium. Based on these results, we conclude that leptin stimulates the proliferation of tumour cells that carry activating alterations in the canonical Wnt pathway (Supplementary figure 11). Furthermore, these data also suggest that leptin is not involved in early-stage colorectal carcinogenesis. Collectively, our observations provided a novel finding that leptin acts as growth factor for CRC only after the tumour initiation stage during the process of colorectal carcinogenesis (Supplementary figure 12).

Our data define a novel role for leptin signalling in the control of tumour growth in addition to its essential role in food intake and energy regulation. The role of leptin signalling is evident from the finding of increased ObR expression in colon tumours, and of such increased expression coinciding with the activation of STAT3. Furthermore, absence of leptin signalling prevented tumour growth, and suppressed STAT3 activation in these tumours. These findings demonstrated that activation of STAT3 in tumours is crucially dependent on leptin signal transduction. Finally, the leptin signalling mechanism of action was revealed operationally by the finding that treating mice with recombinant leptin increased tumour growth. Taken together, these data provide strong evidence to indicate that leptin signalling controls tumour growth in vivo.

It has been shown previously that recombinant leptin does not stimulate cell proliferation and carcinogenesis in vivo.16 18 20 While continuous treatment with recombinant leptin enhanced tumour growth in AOM-treated mice, the effect of exogenous leptin was not as strong as we had expected. On the other hand, there is general agreement that leptin acts as a growth factor for colon cancer cells in vitro.15–17 These discrepancies between in vivo and in vitro studies could be explained by the complicated interaction between various hormones and cytokines. The effects of leptin in vivo are not as simple as those in vitro. Leptin is known to regulate the secretion of several hormones. Importantly, the actions of leptin involve amelioration of hyperinsulinaemia.35 36 We observed such actions of leptin on insulin levels in mice treated with recombinant leptin. Insulin has the effect of promoting the development of chemically induced tumours in the colon.37 Therefore, in vivo, the effects of exogenous leptin on promotion of colonic tumourigenesis might be suppressed through a decrease in insulinaemia.

In conclusion, we clearly demonstrated a relationship between leptin signalling and growth of colon tumours, using leptin-deficient or leptin-receptor-deficient mice. The dramatic suppression of colon tumour growth resulting from inhibition of leptin signalling indicates that leptin is an important growth factor for colon cancer progression. We speculate that dietary intake of excessive fat and calories might result in energy storage in the visceral and subcutaneous adipose tissue compartments, and that any surplus energy might be used for growth of CRC through leptin signalling. On the basis of the current results, it is reasonable to conceive that colon tumours might have a tendency to develop in obese individuals who over-eat and who show elevated serum leptin levels. Future study is warranted to address the importance of leptin signalling in the metastatic spread of CRC. Our data provide novel insights into leptin signalling in CRC and suggest novel therapeutic and preventive targets against colon polyps and cancers based on inhibition of leptin-dependent STAT3 signalling.

Acknowledgments

We thank Machiko Hiraga and Yuko Sato for their technical assistance.

References

View Abstract

Supplementary materials

Footnotes

  • Funding This work was supported in part by a Grant-in-Aid for research on the Third Term Comprehensive Control Research for Cancer from the Ministry on Health, Labor and Welfare, Japan, to AN, a grant from the National Institute of Biomedical Innovation (NBIO) to AN, a grant from the Ministry of Education, Culture, Sports, Science and Technology, Japan (KIBAN-B), to AN, and a grant program ‘Collaborative Development of Innovative Seed’ from the Japan Science and Technology Agency (JST) to AN.

  • Competing interests None.

  • Ethics approval All animal experiments were conducted with the approval of the institutional Animal Care and Use Committee of Yokohama City University School of Medicine.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • Digest
    Emad El-Omar Bill Grady Alexander L Gerbes