The first two authors contributed equally to this work.
IGFBP-2 overexpression reduces the appearance of dysplastic aberrant crypt foci and inhibits growth of adenomas in chemically induced colorectal carcinogenesis
Version of Record online: 26 NOV 2008
Copyright © 2009 Wiley-Liss, Inc.
International Journal of Cancer
Volume 124, Issue 9, pages 2220–2225, 1 May 2009
How to Cite
Diehl, D., Hessel, E., Oesterle, D., Renner-Müller, I., Elmlinger, M., Langhammer, M., Göttlicher, M., Wolf, E., Lahm, H. and Hoeflich, A. (2009), IGFBP-2 overexpression reduces the appearance of dysplastic aberrant crypt foci and inhibits growth of adenomas in chemically induced colorectal carcinogenesis. Int. J. Cancer, 124: 2220–2225. doi: 10.1002/ijc.24193
- Issue online: 24 FEB 2009
- Version of Record online: 26 NOV 2008
- Manuscript Accepted: 9 APR 2008
- Manuscript Received: 12 NOV 2007
- German Research Council. Grant Numbers: KFOR128, DFG HO 2003/4-2
- German National Genome Research Network. Grant Numbers: NGFN2, PMM-S31T06
- Hanns Seidel Foundation e.V
- growth inhibition;
- colon cancer;
Colon cancer patients frequently show increased levels of serum insulin-like growth factor-binding protein-2 (IGFBP-2), however, the pathogenetic relevance of this phenomenon for colorectal cancer is unclear. Therefore, we have used IGFBP-2 transgenic animals which overexpress IGFBP-2 systemically and locally in the intestine to study its role in chemically induced colorectal carcinogenesis. Mice received intraperitoneal injections of 1,2-dimethylhydrazine (DMH) (40 mg/kg body weight) once a week for 6 weeks to selectively induce aberrant crypt foci (ACF) and tumors in the colon. While tumor incidence was comparable in transgenic and control mice, the volume of adenomas in IGFBP-2 transgenic mice was reduced more than 2-fold. Furthermore, serum IGFBP-2 levels negatively correlated with tumor volume in the IGFBP-2 transgenic group. Histological examination showed that IGFBP-2 transgenic mice developed significantly less dysplastic ACF with a high potential to progress to advanced stages. The reduced tumor volume in IGFBP-2 transgenic animals was due to significantly reduced proliferative capacity, evidenced by a lower proportion of cells positive for Ki67. Our results demonstrate for the first time in an experimental model that IGFBP-2 overabundance prior to the onset and during colorectal carcinogenesis reduces tumor growth by inhibition of cell proliferation. © 2008 Wiley-Liss, Inc.
Abnormal expression of the insulin-like growth factors (IGFs), their receptors and binding proteins has been linked to several malignancies, including colorectal cancer.1 Because of their high affinities to IGFs, IGF binding proteins (IGFBPs) regulate the bioavailability of IGFs in the circulation and in many tissues.2 However, IGF-independent actions of selected IGFBPs have also been demonstrated.2
In vitro studies have shown that IGFBP-2 overabundance may be associated with inhibition or stimulation of proliferation depending on the cancer cell line tested.3 Increased serum concentrations of IGFBP-2 were found in patients suffering from different malignancies.3 In colorectal cancer patients, serum IGFBP-2 concentrations reflect the tumor load and the levels fall after curative resection.4 However, whether increased serum IGFBP-2 levels have any impact on the risk to develop colorectal cancer remains a controversial issue.5, 6
Most of the current knowledge about the IGF system is derived from in vitro experiments. To date no experimental studies have been published to investigate whether and during which stages IGFBP-2 overabundance might have an influence on colorectal carcinogenesis in vivo. To address this question we have induced colorectal carcinogenesis by injection of 1,2-dimethylhydrazine (DMH) leading to the development of aberrant crypt foci (ACF) and adenomas using a transgenic mouse line in which IGFBP-2 is strongly overexpressed.7 We have evaluated the effect of IGFBP-2 overexpression on the development and growth of adenomas in these animals. In addition, we have analyzed the impact of IGFBP-2 on the appearance of early hyperplastic and dysplastic ACF. Finally, we have explored potential mechanisms which could influence adenoma growth by determination of parameters of cell proliferation and apoptosis.
Material and methods
Treatment of animals
Female C57BL/6 mice, which overexpress mouse IGFBP-2 under the control of the CMV-promoter in several tissues including colon7 and their nontransgenic littermates received intraperitoneal injections of DMH (40 mg/kg body weight) once a week for 6 weeks starting at the age of 4 weeks. All experiments were carried out according to the German Animal Protection Law. Mice were killed 10 weeks (13 IGFBP-2 transgenic and 10 wild-type mice) or 34 weeks (20 IGFBP-2 transgenic and 19 wild-type mice) after the first carcinogen injection.
Identification of CMV-IGFBP-2 transgenic animals
IGFBP-2 transgenic mice were distinguished from wild-type mice by PCR as described previously.7
Detection of serum IGF and IGFBP-2 levels
Serum levels of IGFBP-2, IGF-I and IGF-II were determined by RIA as described before.8
Identification and characterization of ACF and tumors
Recovery of colons and detection of ACF by methylene blue staining and tumor evaluation were performed in both age groups (14 and 38 weeks) as previously described.8 After dissection colons were opened longitudinally, fixed flat between filter paper and a microscopic slide in 4% neutral buffered paraformaldehyde and stained with 0.2% methylene blue. Thereby ACF can be differentiated from unaltered squamous tissue by their bigger size, round to elliptic lumina of variable width and thickened epithelia.9 Thirty-four weeks after DMH injection visible tumors were numbered and their location was recorded. Colonic mucosa and visible tumors were routinely processed and embedded in paraffin. Tumors were cut in the middle and embedded with the section flat down. Colons were embedded with the mucosal side flat down which was described as an “en face” preparation.10 Serial sections (3–4 μm) from colon tissue and tumors were mounted on glass slides. Hematoxylin and eosin (H&E)–stained serial sections of ACF and tumors were used for histological grading. Since dysplastic ACF have a higher potential of malignant progression, we have performed a histopathological study in selected animals 10 weeks after the first DMH-treatment (wt: n = 4; tg: n = 5) by systematic analysis of serial sections over the complete length of the colon (at least every forth section). The histopathological analysis included hyperplastic and dysplastic ACF. The grading of ACF and tumors was based on criteria described elsewhere.8, 10, 11 The degree of dysplasia was evaluated according to a number of criteria: (i) number of nuclei per crypt, (ii) morphology of nuclei, (iii) differentiation of cells (loss of goblet cells) and (iv) structure of crypts.
Immunohistochemistry (IHC) of formalin-fixed, paraffin-embedded sections of ACF and adenomas was performed using the peroxidase-conjugated avidin-biotin method as previously described.8 The proliferative activity in adenomas was determined by incubating the tissue sections with rat monoclonal anti–Ki-67-TEC3 (DakoCytomation, Glostrup, Denmark) at a dilution of 1:200. The rabbit polyclonal cleaved caspase-3 (Asp175) antibody (Cell Signaling, Frankfurt, Germany) served for detection of apoptotic cell nuclei at a dilution of 1:200. Biotinylated anti-rat or anti-rabbit IgG (dilution 1:200) (Vector Laboratories, Wertheim, Germany) were used as secondary antibodies. Negative controls for all stainings were performed by omitting primary antibodies. Proliferation and apoptosis indices in adenomas were determined as previously described.8
Total RNA was isolated from colon and tumor tissues with the Trizol reagent (Invitrogen, Karlsruhe, Germany) according to the manufacturer's recommendations. Total RNA preparations were incubated for 30 min at 37°C with RNase-free DNase I (Roche Diagnostics, Mannheim, Germany) to digest residual genomic DNA. The reaction was stopped by incubation for 10 min at 75°C. Reverse transcription was performed for 2 hr at 42°C in RT buffer (50 mM Tris-HCl, pH 8.3, 75 mM KCl, 1.5 mM MgCl2), 10 mM DTT, dNTPs (1 mM each), random hexamer primers (250 ng) and 50 U Superscript II (Invitrogen). The reaction was terminated by incubation for 10 min at 70°C. Subsequent quantitative PCR analyses were performed on an ABI Prism 7700 Sequence Detector (Applied Biosystems, Darmstadt, Germany) in a final volume of 25 μL containing cDNA, forward and reverse primers (200 nM each), the appropriate fluorescence-labeled probe (200 nM) from the Universal Probe Library (Roche Diagnostics) and Fast Start TaqMan Probe Master solution (Roche Diagnostics). The following primers and universal probes were selected using the web-based ProbeFinder Software: IGFBP-2 forward 5′ GCGGGTACCTGTGAAAAGAG 3′, IGFBP-2 reverse 5′ CCTCAGAGTGGTCGT CATCA 3′, IGFBP-2 universal probe #62. Taq polymerase was activated for 10 min at 95°C followed by 40 cycles of 15 sec at 95°C (denaturation) and 60 sec at 60°C (annealing and extension). The relative expression levels of genes were calculated in relation to the expression of housekeeping genes.
Determination of food consumption
Food consumption was measured in adult control and age-matched IGFBP-2 transgenic male mice over a period of 11 days. Fresh tap water and fixed formula food for laboratory mice was supplied ad libitum (Altromin 1314: protein 22.5%, fat 5%, raw fiber 4.5%, ash 6.5%, germ reduced by heat during the production process, Altromin GmbH, Lage, Germany). The animals were housed in Makrolon-cages Type II (Ebeco, Castrop-Rauxel, Germany), with wood shavings for bedding, in a barrier system under controlled environmental conditions with a 12 hr light:dark cycle (room temperature 22°C; humidity 50–60%). Recordings were performed on days 4, 8 and 11 both for body and food weight. For the duration of the analysis all mice were caged individually, and dispersed food in the cages was collected and included in the weight recordings. In order to normalize for body weight, we have used metabolic body weight (body weight0.75) for the individual animals.
Data were expressed as the mean ± SEM and were analyzed using Graph Pad Prism Version 3.0 (Prism; Graph Pad, San Diego, CA). Data were analyzed using the Student's t-test, the Mann-Whitney U test and the Fisher's exact test. For the analysis of food consumption the comparison of means was performed using the Student's t-test. The significance level was set at p < 0.05.
Elevated serum and tissue IGFBP-2 levels in IGFBP-2 transgenic mice are associated with reduced colonic tumor growth
We have established transgenic mouse lines in which the expression of a murine IGFBP-2 cDNA is controlled by the CMV-promoter. These mice overexpress IGFBP-2 in several organs including heart, stomach, spleen and pancreas on the mRNA and protein levels and IGFBP-2 was biologically active.7 IGFBP-2 mRNA is also strongly overexpressed in colonic tissues of transgenic animals (Fig. 1a) and the synthesized protein is biologically active.7 Likewise, serum IGFBP-2 levels were significantly increased in IGFBP-2 transgenic mice. In contrast, serum concentrations of IGF-I and IGF-II were not different between both genotypes (Fig. 1b). To elucidate the question whether elevated IGFBP-2 levels may possibly influence the growth of neoplastic colonic cells we have selectively induced colorectal carcinogenesis by injection of DMH. Using this protocol, all animals which received carcinogen injections developed ACF.
Thirty-four weeks post-initiation with DMH 55% (11/20) of IGFBP-2 transgenic and 53% (10/19) of wild-type mice had developed macroscopically visible tumors in the colon (Fig. 1c). In most cases, only one visible tumor was seen. Two visible tumors were recorded in 3 wild-type and 5 transgenic animals, respectively. One animal in each experimental group developed 3 tumors (Table I). These tumors were histologically classified as adenomas in both genotypes, with the exception of one carcinoma in an IGFBP-2 transgenic animal. While the tumor incidence in both genetic groups was similar, the tumor volume in IGFBP-2 transgenic animals (5.1 ± 1.6 mm3) was smaller than in wild-type animals (11.7 ± 3.5 mm3; p < 0.05) (Fig. 1c). Notably, serum IGFBP-2 concentrations correlated negatively with the volume of adenomas in IGFBP-2 transgenic animals (p < 0.01) (Fig. 1d). Thus, the permanent presence of high IGFBP-2 levels prior to initiation and during the development of colorectal carcinogenesis obviously inhibited tumor growth.
|Overall tumor prevalence||Animals with 1 tumor||Animals with 2 tumors||Animals with 3 tumors|
|Wild-type||53% (10/19)||60% (6/10)||30% (3/10)||10% (1/10)|
|IGFBP-2 transgenic||55% (11/20)||45.5% (5/11)||45.5% (5/11)||9% (1/11)|
IGFBP-2 overexpression reduces the appearance of dysplastic foci
In order to identify potential mechanisms involved in a reduction of tumor growth we first performed a histological analysis of colonic tissue 10 weeks after the first injection of DMH. At this time, single aberrant crypts (AC) and ACF were observed in transgenic and wild-type mice after treatment with DMH (Fig. 2a) but never with vehicle (data not shown). The total number of ACF in the colon of IGFBP-2 transgenic mice was higher when compared with nontransgenic littermates (Fig. 2b). In both genotypes, the total number of ACF was reduced during the next 24 weeks, with a modest reduction of 7% in wild-type mice but of 38% (p < 0.05) in IGFBP-2 transgenic mice (Fig. 2b). In order to examine potential differences in the distribution of hyperplastic and dysplastic ACF we performed histological analyses (Fig 2c). The numbers of hyperplastic ACF per mouse were similar in both experimental groups (17.0 in wild-type vs. 20.0 in transgenic mice). In contrast, in the colons of IGFBP-2 transgenic mice the fraction of dysplastic ACF was significantly smaller than in the colons of wild-type animals (p < 0.05) (Fig. 2d). The number of ACF with mild dysplasia was lower in IGFBP-2 transgenic mice (16.5 in wild-type vs. 7.8 in transgenic mice). The effect was even more pronounced with ACF demonstrating features of moderate (wild-type: 6.0 vs. transgenic: 1.6) and severe (wild-type: 1.7 vs. transgenic: 0.4) dysplasia. This finding is important in view of the fact that dysplastic ACF contain a higher potential to progress into adenomas.9
Inhibition of tumor growth by IGFBP-2 is due to a reduced proliferation rate
The reduced tumor volume in IGFBP-2 overexpressing mice might be a consequence of an inhibition of cellular proliferation or an increase of the apoptosis rate. Therefore, serial sections of 6 adenomas per genotype were evaluated immunohistochemically for the presence of Ki-67 and cleaved caspase-3 as markers for proliferation and apoptosis, respectively. In adenomas of IGFBP-2 transgenic animals, the staining intensity of Ki-67 was clearly diminished when compared with adenomas derived from wild-type animals (Fig. 3a). The proportion of proliferating cells was significantly (p < 0.05) lower than in wild-type mice (Ki-67 labeling index: 33% vs. 48%) (Fig. 3c). In contrast, the proportion of cleaved caspase-3 positive cells in adenomas of IGFBP-2 transgenic mice was similar to that of controls (2.6% vs. 2.3%) (Fig. 3b). Thus, the reduced tumor growth in IGFBP-2 transgenic mice apparently is due to a decreased proliferation rate.
Body weight and food uptake in IGFBP-2 transgenic mice
IGFBP-2 transgenic mice are characterized by lower body weights when compared with their nontransgenic littermates. At ages of 10 and 34 weeks nontransgenic female mice have a body weight of 19.8 ± 0.37 g and 27.28 ± 0.98 g, while the weights of age matched female IGFBP-2 transgenic littermates correspond to 17.88 ± 0.36 g and 22.92 ± 0.98 g, respectively. In order to exclude effects of differences in food intake on tumor growth we have determined food intake in adult nontransgenic mice (n = 16) and age-matched transgenic littermates (n = 7) over a period of 11 days with repeated determinations of food consumption (Table II). In both genotypes, identical amounts of food were consumed after correction for the metabolic body weight, respectively.
|Day 4||Day 8||Day 11|
|Wild-type (g)||0.469 ± 0.048||0.446 ± 0.032||0.424 ± 0.024|
|IGFBP-2 transgenic (g)||0.463 ± 0.036||0.426 ± 0.019||0.415 ± 0.0018|
In several malignancies, including colon cancer, enhanced levels of IGFBP-2 in serum and tissues correlate with the presence of pathological and neoplastic alterations.3 In ovarian and endometrial cancer a potential correlation between serum IGFBP-2 levels and the risk to develop cancer has been suggested.12, 13 Renehan et al.4 described a positive correlation between serum IGFBP-2 levels and concentrations of the tumor marker carcinoembryonic antigen in patients suffering from colon cancer. Accordingly, IGFBP-2 levels have been suggested as a diagnostic parameter in the surveillance of patients with colorectal cancer.4 However, it remains controversial whether serum IGFBP-2 levels correlate with colon cancer risk. One study linked high IGFBP-2 levels with a significantly decreased risk to develop colorectal cancer,6 while a second one could not find a clear association.5
We have previously reported a stimulating effect of IGF-II on growth of adenomas induced by DMH.8 Furthermore, IGFBP-2 potentiated the mitogenic effect of IGF-II in osteoblasts14 and conferred enhanced growth potential to adrenocortical tumor cells.15 One might thus have speculated that IGFBP-2 promoted the growth of DMH-induced adenomas through a positive interaction with IGF-II. However, the inhibitory effect of IGFBP-2 on adenoma growth might rather be the consequence of an interaction with IGF-I and/or because of the delivery of IGF-independent inhibitory signals. Although it must be mentioned that IGF-II expression in rodents is shut down after birth and we have found no evidence of an IGF-II re-expression in adenomas (data not shown), an inhibitory effect of IGFBP-2 on the development of dysplastic ACF or adenoma volume could not be predicted beforehand. Our results showed a clear and more than 2-fold reduction of the volume of adenomas, which developed in the colon of IGFBP-2 transgenic animals and serum IGFBP-2 levels correlated negatively with the volume of adenomas. Caloric restriction resulted in reduced numbers of ACF in rats16 and several studies demonstrated that food consumption can affect colon cancer.17 Since IGFBP-2 transgenic mice are characterized by reduced body weight we investigated food consumption in our model. Food consumption was a function of body weight but not of genotype and was identical in transgenic and nontransgenic mice when corrected for the metabolic body weight, excluding caloric restriction as cause for the reduced body weight or the negative effects on dysplastic ACF and adenoma growth. Interestingly, in addition to adiposity also body height was positively associated with colon cancer in humans indicating that both adiposity and elevated growth in fact are responsible for an increased colon cancer prevalence.18, 19 Therefore, the inhibitory effects of IGFBP-2 on somatic growth might be related to the negative effect on ACF development or adenoma growth. In mammals, after birth somatic growth is mainly depending on growth hormone and IGF-I and both hormones have been related to increased risk or prevalence of colon cancer.20–22 Since IGF-I transgenic mice23 displayed an almost exactly inverse phenotype of tissue specific growth effects when compared with IGFBP-2 transgenic mice and we were unable to identify specific effects of IGFBP-2 on food consumption it may well be that IGFBP-2 affects colon development by interfering with IGF-dependent effects.
Our results are in favor of a potential tumor-inhibiting effect of IGFBP-2 for colorectal cancer, which is supported by further experimental data. Ten weeks after the initial treatment with DMH ACF were detectable which represent early preinvasive lesions. Hyperplastic ACF have a low tendency to progress towards a tumor. This is supported by the significant decline in the total numbers of ACF at a later stage of azoxymethane-induced colorectal carcinogenesis.24 Our results are in agreement with these data as we also see a decline of the total number of ACF at the age of 38 weeks. In contrast and most interestingly, crypts with dysplastic morphology which possess a higher potential for malignant progression were only present at low abundance in IGFBP-2 transgenic animals. This finding is also compatible with a tumor inhibitory activity of IGFBP-2.
The strong reduction of adenoma volume in IGFBP-2 transgenic mice could be due to an alteration in proliferation or apoptosis. Our investigations have shown that the apoptotic activity in adenomas of wild-type and IGFBP-2 transgenic mice was similar. However, we have provided good evidence that the proliferation rate in adenomas of IGFBP-2 transgenic animals was decreased. The expression of the nuclear antigen Ki-67 correlates with other measurements of proliferation, e.g., S-phase or bromodeoxyuridine uptake. Comparing the fractions of cells positive for Ki-67 staining we observed a massive reduction of proliferating cells by approximately one-third in adenomas derived from IGFBP-2 transgenic animals. Thus, our results conclusively suggest that IGFBP-2 inhibits tumor growth by a negative effect on the proliferation rate of colonic cells.
In conclusion, we have shown for the first time in an animal model of chemical carcinogenesis that IGFBP-2 reduces the appearance of dysplastic ACF which are known to have a high potential to progress to advanced stages of colon cancer. In addition, IGFBP-2 strongly inhibited the growth of developing adenomas by triggering a pronounced reduction of the proliferation rate. As the localization and histology of the induced tumors are similar to human colon cancers, our findings are also of relevance for the biology of clinical cancers.
The authors thank Ms. M. Berauer and Ms. C. Spiller for excellent technical assistance, Ms. K. Weber for preparing the serum IGFBP-2 RIA, Ms. P. Renner, Ms. T. Mittmann and Ms. K. Zorn for expert care of the animals, Ms. E. Wytrwat for her help with statistical analysis and the pathologist Dr. L. Quintanilla-Martinez for her expert diagnostic support in the evaluation of the histology of the tumors. A.H., H.L. and D.D. were supported by the German Research Council and—in part—by the German National Genome Research Network and E.H. was supported by a grant of the Hanns Seidel Foundation e.V.
- 19Co-twin control and cohort analyses of body mass index and height in relation to breast, prostate, ovarian, corpus uteri, colon and rectal cancer among Swedish and Finnish twins. Int J Cancer 2007; 121: 810–8., , , , , , .