Diindolylmethane inhibits angiogenesis and the growth of transplantable human breast carcinoma in athymic mice

Abstract

Studies have linked the consumption of broccoli and other cruciferous vegetables to a reduced risk of breast cancer. The phytochemical indole-3-carbinol (I3C), present in cruciferous vegetables, and its major acid-catalyzed reaction product 3,3′-diindolylmethane (DIM) have bioactivities relevant to the inhibition of carcinogenesis. In this study, the effect of DIM on angiogenesis and tumorigenesis in a rodent model was investigated. We found that DIM produced a concentration-dependent decrease in proliferation, migration, invasion and capillary tube formation of cultured human umbilical vein endothelial cells (HUVECs). 

Consistent with its antiproliferative effect, which was significant at only 5 µM DIM, this indole caused a G ₁ cell cycle arrest in actively proliferating HUVECs. Furthermore, DIM downregulated the expression of cyclin-dependent kinases 2 and 6 (CDK2, CDK6), and upregulated the expression of CDK inhibitor, p27 ^Kip1 , in HUVECs. We observed further in a complementary in vivo Matrigel plug angiogenesis assay that, compared with vehicle control, neovascularization was inhibited up to 76% following the administration of 5 mg/kg DIM to female C57BL/6 mice. Finally, this dose of DIM also inhibited the growth of human MCF-7 cell tumor xenografts by up to 64% in female athymic (nu/nu) mice, compared with the vehicle control. 

This is the first study to show that DIM can strongly inhibit the development of human breast tumor in a xenograft model and to provide evidence for the antiangiogenic properties of this dietary indole.

aFGF, acidic fibroblast growth factor, CDK, cyclin-dependent kinase, CKI, cyclin-dependent kinase inhibitor, DIM, 3,3′-diindolylmethane, DMSO, dimethyl sulfoxide, HUVECs, human umbilical vein endothelial cells, I3C, indole-3-carbinol, s.c., subcutaneousTopic:

• angiogenesis
• endothelial cells
• diet
• angiogenesis inhibitors
• breast neoplasms
• indoles
• mice, inbred c57bl
• mice, nude
• neovascularization, pathologic
• rodentia
• transplantation, heterologous
• vegetables
• methanol
• mice
• neoplasms
• carcinogenesis
• cyclin-dependent kinase inhibitors
• cdk2 protein, human
• breast carcinoma
• tumorigenesis
• cell cycle arrest
• cdkn1b gene
• capillary tubes
• mcf-7 cells

Issue Section:

 Molecular Epidemiology and Cancer Prevention

Introduction

Angiogenesis, the development of new capillaries from an existing vascular network, is an important event in normal and pathological development. Angiogenesis can occur under normal physiological conditions, including embryonic development, female reproductive cycling and wound healing. However, it is also associated with a large number of diseases, including cancer, cardiovascular diseases, rheumatoid arthritis, psoriasis and diabetic retinopathy ( 1 , 2 ). Successful tumor establishment depends on the angiogenesis process to provide oxygen and nutrients to rapidly proliferating cells ( 2 , 3 ). 

Formation of new blood vessels during this process is critically dependent on the ability of the endothelial cells to proliferate, degrade extracellular matrix, migrate and differentiate ( 4 , 5 ). At the beginning of tumorigenesis, the tumor cell mass is not vascularized and it does not grow beyond a few cubic millimeters unless vascularization has occurred ( 3 , 6 ). In addition, the appearance of a vascular stage in the natural history of a tumor is associated with the development of metastases. Therefore, control of tumor angiogenesis is a major area of exploration for the development of preventive and therapeutic agents for cancer.

Results of epidemiologic studies and laboratory investigations with rodents and cultured tumor cells provide evidence that phytochemicals in broccoli and other cruciferous vegetables have anticarcinogenic properties ( 7 ). One of these phytochemicals, indole-3-carbinol (I3C), self-condenses at the low pH of stomach to multiple products ( 8 ). 3,3′-Diindolylmethane (DIM), a major acid-catalyzed product of I3C, has bioactivities relevant to the inhibition of carcinogenesis ( 9 ). Chen et al . ( 10 ) reported that the oral administration of DIM at a dose of 5 mg/kg on alternate days strongly inhibited the dimethylbenzanthracene-induced mammary tumor growth in female Sprague–Dawley rats. However, although results of studies with cultured cells suggest possible modes of DIM activity, little is known regarding the in vivo mechanism whereby DIM inhibits tumor growth. 

One possibility that has received little attention is the inhibition of angiogenesis. The objectives of this study were to investigate the antiangiogenic and antitumor properties of DIM in cultured human cells and in rodent models. Our results show that DIM strongly inhibited proliferation, migration, invasion and capillary tube formation in cultured HUVECs. Furthermore, we show that DIM strongly inhibited vascularization in a Matrigel plug assay and tumorigenesis in a human tumor cell xenograft assay in mice.

Materials and methods

Materials

Estradiol pellets were purchased from Innovative Research of America (Sarasota, Florida). Matrigel and cell culture inserts were obtained from BD-Discovery Labware (Bedford, MA). Anti-human cyclin-dependent kinases (CDKs) such as CDK2, CDK4, CDK6, p27 ^Kip1 and cyclin E were from Santa Cruz Biotechnology (Santa Cruz, CA), and anti-human p21 ^Waf1/Cip1 antibody was from Oncogene Research Products (Darmstadt, Germany). Chemiluminescence reagent was from PerkinElmer Life Sciences (Wellesley, MA). 

Prestained molecular weight marker was obtained from New England Biolabs (Beverly, MA). PVDF Immobilon-P transfer membranes were from Millipore (Bedford, MA). Acidic fibroblast growth factor (aFGF) was purchased from R&D Systems (Minneapolis, MN). Heparin and Drabkin Reagent kit 525 were obtained from Sigma (St Louis, MO). Model Z1 Coulter particle counter was from Coulter Corp. (Miami, FL). 

Microplate reader was from Molecular Devices Corporation (Sunnyvale, CA). DIM was prepared from I3C as described ( 8 , 9 , 11 ) and recrystallized in toluene. All cell culture plates and Petri dishes were ordered from Corning Incorporation (Corning, NY). Salts and other chemicals used were of the highest purity available, and purchased from Sigma Chemical Corp. (St Louis, MO).

Animals and cell lines

Female athymic (nu/nu) mice (7 weeks old) and female C57BL/6 mice (10 weeks old) were purchased from Simonsen Laboratories (Gilroy, CA). The animals were housed in microisolator cages under standard conditions (12:12 h light/dark cycle, 50% relative humidity at 21°C) and given free access to a semi-purified AIN-76A diet and water. MCF-7 cells were purchased from American type culture collection (Rockville, MD), grown in Dulbecco’s modified Eagle’s medium (DMEM; Gibco, Gaithersburg, MD) supplemented with 4.0 g/l glucose, 3.7 g/l sodium bicarbonate and 10% fetal bovine serum (FBS). Human umbilical vein endothelial cells (HUVECs) were purchased from Clonetics (San Diego, CA), and cultured in EGM-MV BulletKit medium containing 5% FBS, 12 µg/ml bovine brain extract, 10 ng/ml human epidermal growth factor, 1 µg/ml hydrocortisone and 1 µg/ml GA-1000 (Gentamicin, Amphotericin-B), according to the manufacturer’s instructions (Clonetics). Cells at 3–7 passages were used for the experiments. Cells were incubated in a humidified incubator at 37°C and 5% CO ₂ .

Cell proliferation

To determine the endothelial cell proliferation, 8 × 10 ⁴ HUVECs were seeded in 6-well plates and allowed to grow for 24 h. The cells were then treated with 0, 2, 5, 10 or 25 µM DIM at 24, 48 and 72 h. DIM was dissolved in dimethyl sulfoxide (DMSO) as 1000× stock solutions for each treatment. Cell numbers were assessed at various time points by trypsinization and counting with a Coulter Z1 cell counter.

Cell invasion assay

Invasion assays were carried out using the modified transwell Boyden chamber system. Chambers were assembled using 8 µm-pore BD Falcon transwell inserts as the upper chambers and the 12-well plates as the lower chambers. Cell culture inserts were coated with 10 µl/insert Matrigel. Coated inserts were left to dry overnight in a laminar flow fume hood and then rehydrated with DMEM supplemented with 0.1% bovine serum albumin (BSA) for 1 h at 37°C. Rehydrated Matrigel wells were washed with the same medium. 

HUVECs were harvested by trypsin/EDTA and resuspended in 0.1% BSA/DMEM. Medium containing 5% FBS was applied to the lower chamber as chemoattractant, cells were then seeded at 1.5 × 10 ⁵ cells/insert in the presence of 0, 5, 10 or 25 µM DIM and incubated for 2 h at 37°C and 5% CO ₂ . 

At the end of the incubation, the cells in the upper chamber were removed with cotton swabs and cells that traversed the Matrigel to the lower surface of the insert were fixed with 10% formalin/PBS and stained with crystal violet in 10% formalin/PBS. Cells that migrated to the lower surface of the insert were counted under the light microscope at a magnification of ×40. 

Each treatment was in triplicate. The quantification of the invasive cells in the presence of DIM was expressed as the percentage of the quantity of invasive cells under control conditions (DMSO).

Wound migration assay

Migration was assessed using an in vitro wound assay. 1 × 10 ⁵ cells were seeded into two 12-well cell culture plates and cultured in EGM-MV BulletKit to confluency. A scrape was made in the center of the cell monolayers with a sterile pipette tip to create a gap of constant width. Cellular debris was removed by gently washing with PBS. 

The initial images of the wounds were captured under phase contrast microscopy and the wounded monolayers were incubated further for 18 h in fresh EGM-MV BulletKit medium in the presence of DMSO or DIM of various concentrations (5, 10 or 25 µM). Pictures were taken with a Nikon Coolpix990 digital camera connected to the Nikon Eclipse TS100 microscope at a ×100 magnification. 

To quantify the migration, photographs of the initial wounded monolayers were compared with the corresponding images of cells at the end of the incubation. Artificial lines fitting the cutting edges were drawn on pictures of the original wounds and overlaid on the images of cultures after incubation. Cells that migrated across the lines were counted. 

All quantification was done on full-size images with the weight of artificial lines negligible, when compared with the size of the cell body. At least five fields from each triplicate treatment were counted.

Tube formation

Endothelial cells plated onto a gel of basement membrane protein rapidly organize into multicellular tube-like structures ( 12 , 13 ). Matrigel-coated 12-well cell culture plates were ordered from BD-Discovery Labware and the assay was performed according to the manufacturer’s instruction. Briefly, the 12-well plates coated with Matrigel were allowed to solidify at 37°C for 1 h. 

Endothelial cells pre-treated with 0, 10 or 25 µM DIM for 24 h were trypsinized and 6 × 10 ⁴ cells were added per well in 1 ml of the medium. Cell viability was determined by Trypan blue exclusion stain before seeding. DIM treatments continued for the rest of the assay. Tube formation was observed periodically over time under a phase contrast microscope and representative pictures were taken at 3 and 24 h.

Flow cytometric analysis of DNA content

HUVECs were seeded in a 100 mm Corning tissue culture dish to grow to 70–80% confluency. Cells were then treated with 0, 5, 10 or 25 µM DIM for 24 h. Cells were hypotonically lysed in 1 ml of DNA staining solution (0.5 mg/ml propidium, 0.1% sodium citrate and 0.05% Triton X-100), and nuclear fluorescence (wavelength > 585 nm) was measured with a Coulter Elite instrument with laser output adjusted to deliver 15 mW at 488 nm. 

A total of 10 000 nuclei were analyzed from each sample. The percentage of cells in the G ₁ , S and G ₂ phases of the cell cycle were determined by an analysis of data with the Multicycle computer software provided by Phoenix Flow Systems in the Cancer Research Laboratory Microchemical Facility of the University of California at Berkeley.

Western blot analysis

After the indicated treatments, the cells were washed with PBS and harvested in RIPA buffer (150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS and 50 mM Tris) containing 10 µg/ml aprotinin, 5 µg/ml leupeptin and 1 mM phenylmethylsulfonyl fluoride. Protein concentrations were determined by Bradford–Coomassie dye binding assay. 

Equal amounts of protein were fractionated by electrophoresis on 4% polyacrylamide/0.1% SDS stacking gels and 15% polyacrylamide/0.1% SDS resolving gels. Prestained molecular weight marker was used as a standard. Proteins were electrotransferred to PVDF Immobilon-P transfer membranes using a transfer buffer (25 mM glycine, 25 mM ethanolamine and 20% methanol) and then blocked at room temperature for 1 h in 5% blocking buffer (1× TBS, 0.1% Tween-20 with 5% w/v non-fat dry milk). 

Blots were then incubated for 3 h at room temperature with primary mouse anti-human CDK2, CDK4, CDK6, p27 ^Kip1 , p21 ^Waf1/Cip1 and cyclin E antibody, and then washed with wash buffer (10 mM Tris–HCl, pH 9.5, 10 mM NaCl and 0.1% Tween-20). The membranes were incubated for another hour with secondary anti-mouse or anti-rabbit IgGs conjugated with horseradish peroxidase (1:1000 dilution). 

The proteins were visualized by chemiluminescence reagent and exposure to Kodak BioMax MR film. Images were scanned and quantified by the Molecular AnalystDensit software. Equal protein loading was confirmed by probing blots with antibody to β-actin.

Matrigel plug angiogenesis assay

In vivo angiogenesis assays were performed on female C57BL/6 mice as described before ( 14 ). The mice were acclimated to semi-purified phytoestrogen-free AIN-76 ad libitum for 7 days prior to the study and randomly grouped (5/group). DMSO vehicle or DIM at 5 mg/kg was injected subcutaneously 5 days before the Matrigel inoculation. Heparin (64 U/ml) and aFGF (100 ng/ml) were gently mixed with cold liquid Matrigel at 4°C. 

The Matrigel solution (0.3 ml) was injected subcutaneously in the bilateral flanks of the mice. Animals were treated with DMSO control or DIM for 2 more weeks and killed by CO ₂ inhalation. The gels were surgically removed and the vascularization of Matrigel plugs was quantified by measuring the hemoglobin content using Drabkin Reagent kit 525.

MCF-7 human breast carcinoma xenograft study

A colony of 20 female athymic (nu/nu) mice were acclimated to semi-purified phytoestrogen-free AIN-76 ad libitum for 7 days prior to the study. They were implanted with a 60-day release 0.72 mg estradiol pellet in the subscapular region. 

The mice were then given a subcutaneous inoculation in the bilateral flanks with 0.1 ml of Matrigel containing 3 × 10 ⁶ MCF-7 human breast cancer cells and randomly grouped (10/group) to receive s.c. injections of either 5 mg/kg DIM or DMSO in a PBS vehicle, five times weekly. Feed intake and body weight were measured weekly and palpable tumor diameters were measured twice per week. 

Tumor volumes were calculated as: (π/6) × [length (mm) × width ² (mm ² )] ( 15 ). The experiment was terminated at 34 days. All the animals were killed by CO ₂ asphyxiation.

Statistics

For statistical analyses, means were compared by ANOVA followed by Tukey’s test. SigmaStat 2.03 software was used for all the statistics. Differences were considered significant at P < 0.05. The results are expressed as mean ± SD, unless otherwise stated.

Results

DIM inhibits proliferation, invasion, migration and tube formation of HUVECs

We conducted a series of assays in primary cell culture to determine whether DIM exhibits an antiangiogenic potential. First, the effect of DIM on proliferation of HUVECs was examined. HUVECs were treated with 0, 2, 5, 10 or 25 µM DIM for 24–72 h. DIM inhibited the proliferation of HUVECs in a concentration- and time-dependent manner, with up to 70% inhibition of proliferation at 72 h treatment with 25 µM DIM ( Figure 1 ). Significant inhibition of proliferation occurred in cells treated with only 5 µM DIM at 72 h. 

Cell survival and viability, as determined by trypan blue exclusion, were not obviously affected by the conditions of this assay. These results suggest that DIM might act as an angiogenesis inhibitor by directly reducing the proliferation of vascular endothelial cells.

Fig. 1.

Open in new tabDownload slide

The effect of DIM on HUVECs proliferation. HUVECs were seeded in 6-well plates at 8 × 10 ⁴ and allowed to grow for 24 h. Cells were then treated by 0, 2, 5, 10 or 25 µM DIM for 24, 48 and 72 h. Cell numbers were assessed at various time points by trypsinization and counting with a Coulter Z1 cell counter. Values were expressed as mean ± SD. 

Results are representative of at least three independent experiments. ^* indicates significant difference from control at a level of P ≤ 0.05. ^** indicates significant difference at a level of P ≤ 0.01.

We next conducted a cell invasion assay to examine the effect of DIM on HUVEC movement through a simulated extracellular matrix. In a modified Boyden chamber assay, the transwell inserts were coated with a thin layer of Matrigel and inserted into wells containing 5% FBS medium as a chemoattractant. As a negative control, serum-free medium containing 0.1% BSA was used in the lower chamber. Cells that transversed the Matrigel and migrated to the lower part of the insert were photographed and quantified. 

The results showed that no cell invasion was observed with serum-free medium in the lower chamber ( Figure 2A , negative control). In response to 5% FBS, however, HUVECs traversed the Matrigel and migrated to the lower chamber ( Figure 2A , DMSO). The invasion of HUVECs was decreased by DIM in a concentration-dependent manner, in which DIM at 25 µM decreased the cell invasion to 52% of the control ( Figure 2A and B ). The trend of decreased cell invasion was seen at 5 µM DIM and a statistically significant effect was seen at 10 µM exposure.

Fig. 2.

Open in new tabDownload slide

The effect of DIM on HUVECs invasion. ( A ) Invasion assays were carried out using a modified Boyden transwell chamber system. Cell culture inserts for 12-well plate were coated with a thin layer Matrigel. Cells were seeded in 0.1% BSA/DMEM in the presence of 0, 5, 10 or 25 µM DIM and incubated for 2 h at 37°C and 5% CO ₂ . Cells that traversed the Matrigel and attached to the lower surface of the insert were fixed with 10% formalin and stained with crystal violet. Each treatment was done in triplicate. 

Representative pictures are shown. ( B ) Cells on the lower surface of inserts were counted under the microscope. Mean ± SD ( n = 3) was calculated and the values reported as the percentage of the control. Results are representative of at least three independent experiments. ^* indicates significant difference compared from control at a level of P ≤ 0.05. ^** indicates significant difference at a level of P ≤ 0.01.

A complementary wound migration assay was also conducted as described in Materials and Methods to determine whether DIM directly affects cell migration. For this assay, confluent cultures of HUVECs were wounded and then incubated for 18 h in fresh complete medium. To quantify the migration, artificial lines fitting the cutting edges were drawn on pictures of the original wounded cells and overlaid on the images of cells after incubation. Cells that migrated across the lines were counted.

 Figure 3A shows the representative pictures of the migration assay, including an example of the artificial lines drawn for quantification. The wound was largely closed by cells migrating from both edges of the wound after 18 h in the DMSO control ( Figure 3A ). However, DIM inhibited this process in a concentration-dependent manner. HUVECs treated with DIM showed a dramatic decrease in the migration of cells across the wound, with 25 µM DIM decreasing the number to only 40% of the control ( Figure 3B ).

Fig. 3.

Open in new tabDownload slide

The effect of DIM on HUVECs migration. (A) Cells were seeded into cell culture plates (12-well) and cultured in EGM-MV BulletKit to near confluency. The cell monolayers were wounded using a sterile pipette tip and the cells were incubated for 18 h in the presence of DMSO or DIM of indicated concentrations. Representative pictures were taken which show the wound areas before and after treatment. 

( B ) The total number of cells that migrated across the lines were counted in five random fields for each treatment. Values were calculated as mean ±SD ( n = 4) and reported as the percentage of the control. Results are representative of at least three independent experiments. ^* indicates significant difference compared from control at a level of P ≤ 0.05. ^** indicates significant difference at a level of P ≤ 0.01.

In a further assay relevant to angiogenesis, we examined the effect of DIM on HUVEC tube formation in culture. Capillary formation on Matrigel is a process that requires a cell–matrix interaction. For the assay, HUVECs were pre-treated with or without 10 or 25 µM of DIM for 24 h and then seeded on a thin layer of Matrigel. Capillary tube structure formation was obvious at 3 h of incubation and was almost completed at 24 h in the control medium ( Figure 4 ). 

However, DIM-treated HUVECs showed a decreased ability to extend and differentiate into tube-like structures with effects obvious following a 3-h incubation with 10 µM DIM.

Fig. 4.

Open in new tabDownload slide

The effect of DIM on tube formation of HUVECs. Endothelial cells pre-treated with 0, 10 or 25 µM DIM for 24 h were trypsinized and 6 × 10 ⁴ cells were seeded per well in 1 ml medium in 12-well plate coated with Matrigel. 

DIM treatments continued for the rest of the assay. Cell viability in the presence of DIM was determined by Trypan blue exclusivity stain. Tube formation was observed periodically over time under a phase contrast microscope and representative pictures were taken at 3 and 24 h. Results are representative of at least three independent experiments.

Taken together, these results indicate that DIM acted directly on cultured endothelial cells to inhibit the processes of proliferation, migration and tube formation, which are important markers of potential antiangiogenic activity, in vivo .

DIM induces a G ₁ cell cycle arrest in HUVECs

To further characterize the antiproliferative activity of DIM in HUVECs, we examined the effects of this indole on cell cycle regulation. For the cell cycle studies, HUVECs treated with vehicle control or 25 µM DIM for 24 h were analyzed by flow cytometry as described in Materials and Methods. Figure 5A shows representative histograms of cell cycle distribution in the control and 25 µM DIM-treated HUVECs. DIM treatment increased the cell population in G ₁ phase in a concentration-dependent manner, which became evident following a 24-h treatment of an asynchronous growing cell population with 5 µM DIM ( Figure 5B ). Conversely, the cell population in S phase significantly decreased. For example, the 25 µM DIM treatment increased the proportion of cells in the G ₁ phase from 71.8 ± 2.5% to 85.7 ± 2.9%, and decreased the proportion of S phase cells from 15.2 ± 1.3 to 2.4 ± 0.9% ( P < 0.01), clearly indicating a G ₁ block in cell cycle progression ( Figure 5B ).

Fig. 5.

Open in new tabDownload slide

DIM causes G ₁ cell cycle arrest in HUVECs. ( A ) Histograms of distribution of DNA content in HUVECs with or without 25 µM DIM treatment for 24 h by flow cytometry. Cells were hypotonically lysed in 1 ml of propidium iodide solution. Nuclear fluorescence was measured and analyzed. The percentages of cells in G ₁ , S and G ₂ phase of the cell cycle were determined by analysis with the Multicycle computer software. ( B ) Distribution of cell population in G ₁ , S and G ₂ phases in control and 5, 10 and 25 µM DIM-treated HUVECs. Values were expressed as mean ± SD ( n = 3). Results are representative of at least three independent experiments. ^* indicates significant difference compared from control at a level of P ≤ 0.05. ^** indicates significant difference at a level of P ≤ 0.01.

DIM downregulates CDK2 and CDK6, and upregulates CDK inhibitor p27 ^Kip1 in HUVECs

Our observation that DIM induces a G ₁ block in cell cycle progression suggested that DIM might selectively regulate the activities of G ₁ cell cycle regulating components. To examine this possibility, the expressions of G ₁ cell cycle components were investigated by a western blot analysis. Results shown in Figure 6A indicate that DIM treatment reduced the expression of CDK2 and CDK6 proteins and strongly increased the expression of a cell cycle inhibitor (CKI) p27 ^Kip1 . Levels of CDK4, cyclin E and p21 ^Waf1/Cip1 , however, were not significantly affected by DIM treatment. The 25 µM DIM treatment significantly decreased the CDK2 expression to 35% and CDK6 to 40% of the control ( Figure 6B ). In contrast, 25 μM DIM administration significantly increased the CKI, p27 ^Kip1 to nearly 3.5-fold of the control ( Figure 6B ). These data indicate that DIM induces a G ₁ cell cycle arrest in HUVECs, which is accompanied by a downregulation of the expression of CDK2 and CDK6 G ₁ -related kinases and an upregulation of the expression of the CKI, p27 ^Kip1 .

Fig. 6.

Open in new tabDownload slide

The effect of DIM on cell cycle components of HUVECs. ( A ) HUVECs were treated with DMSO or DIM of indicated concentrations for 24 h. Cell lysates were subjected to SDS–PAGE and transferred onto PVDF membrane. Western blots were performed with the indicated antibodies as described in Materials and methods. ( B ) Western blots were quantified using densitometry. The value of the control was set as 100% and the treatments were expressed as the ratio to control for each individual blot. Three blots from independent experiments were quantified and values were reported as mean ± SD (%). ^* indicates significant difference from control at a level of P ≤ 0.05. ^** indicates significant difference at a level of P ≤ 0.01.

DIM reduces in vivo angiogenesis

Using a rodent Matrigel plug angiogenesis assay, we next investigated whether DIM affects the neovascularization in vivo . Matrigel is a urea extract of Engelbreth–Holm–Swarm tumor, which contains laminin, collagen IV, heparan sulfate proteoglycan and several growth factors, all of which are present in the authentic basement membrane of solid tumors. When injected subcutaneously into rodents, Matrigel forms a solid ‘plug’ beneath the skin. 

The hemoglobin content in the Matrigel parallels the blood vessel development in the gel, thereby allowing a quantitation ( 14 ). Matrigel (0.3 ml) containing 64 U/ml heparin and 100 ng/ml aFGF were injected subcutaneously into the bilateral flanks of C57/BL mice. DMSO vehicle or 5 mg/kg of DIM was injected subcutaneously 5 days before the Matrigel inoculation and continued for the remaining 2 weeks of the experiment. The result showed that DIM treatment reduced neovascularization up to 76% compared with controls, as indicated by the significantly lower hemoglobin content shown in Figure 7 . Thus, DIM exhibits direct inhibitory activity on angiogenesis.

Fig. 7.

Open in new tabDownload slide

The effect of DIM on Matrigel plug in vivo angiogenesis. Neovascularization in Matrigel plugs was quantified by hemoglobin (Hb) content after injecting female C57BL/6 mice with 0.3 ml of Matrigel mixed with heparin and aFGF in the bilateral flanks. Mice were pre-treated with s.c. injection of 5 mg/kg DIM before Matrigel implant and the treatment continued for the rest of the assay. Matrigel plugs were removed at day 14. Hb content was expressed as mean ± SE of n = 10. Values are reported as the percentage of the control. Results are representative of three independent experiments. ^* indicates significant difference from control at a level of P ≤ 0.05.

DIM inhibits the growth of transplanted MCF-7 human breast carcinoma cells

Since we observed an inhibition of the in vivo angiogenesis by DIM in the Matrigel plug assay, we determined whether DIM could inhibit the growth of transplanted breast carcinoma cells in female athymic (nu/nu) mice. Mice were injected subcutaneously with MCF-7 human breast cancer cells in the bilateral flanks. Mice were randomly assigned (10 mice/group) to receive subcutaneous injections of either DIM at 5 mg/kg dose or vehicle control DMSO in PBS, five times weekly. 

Feed intake, body weight and palpable tumor diameter were measured twice per week. Feed intake (22.2 ± 1.3 g/wk versus 22.9 ± 1.0 g/wk) and body weight gain (2.3 ± 0.2 versus 2.4 ± 0.2 g) were not altered by DIM administration. Furthermore, relative organ weights were not significantly affected by DIM treatment (data not shown), indicating that the dose of DIM used in this study did not cause overt toxicity. Tumor volumes were calculated as: (π/6) × [length (mm) × width ² (mm ² )]. DIM treatment reduced tumor growth by 40% ( P = 0.22) after 3 weeks and by 64% ( P < 0.05) at termination of the study on day 34 ( Figure 8 ). 

Final average tumor volume was significantly lower in the DIM group (1125 ± 434 mm ³ ) compared with the control group (3121 ± 554.5 mm ³ ). These results show that DIM can strongly decrease the development of estrogen-dependant human breast cell tumors in a rodent xenograft model.

Fig. 8.

Open in new tabDownload slide

The effect of DIM on growth of transplantable human breast carcinoma in athymic mice. The tumor growth curves of transplantable MCF-7 human breast carcinoma in female athymic (nu/nu) mice. Mice were inoculated subcutaneously in the bilateral flanks with 0.1 ml Matrigel containing 3 × 10 ⁶ human breast cancer cells MCF-7. DMSO or 5 mg/kg DIM were injected subcutaneously five times weekly. Tumor sizes were measured twice per week using a caliper and calculated as (π/6) × [length (mm) × width ² (mm ² )]. The experiment was terminated on day 34. Values are mean ± SE, n = 10. ^* indicates significant difference from control at a level of P ≤ 0.05.

Discussion

We report here for the first time that HUVECs are highly sensitive to DIM, which strongly inhibits their proliferation, migration, invasion and tube formation in culture. DIM also inhibits angiogenesis and the development of human breast tumor cell xenografts in rodent models.

The growth of vascular endothelial cells is of key importance in angiogenesis. Our results show that DIM strongly retarded the proliferation of endothelial cells at a concentration of only 5 µM. Along with retarding the proliferation of cultured HUVECs, we observed that DIM caused a rapid induction of G ₁ cell cycle arrest in these cells. The eukaryotic cell cycle progression is controlled by CDKs that are activated by cyclin binding ( 16 , 17 ) and inhibited by the CKIs at several key checkpoints ( 18 , 19 ). In mammalian cells, control of proliferation is primarily accomplished in G ₁ phase ( 20 , 21 ). Passage from G ₁ into S phase is mainly regulated by the activities of cyclin E–CDK2 complex and cyclin D–CDK4/CDK6 complexes ( 22 , 23 ). p21 ^Waf1/Cip1 and p27 ^Kip1 belong to the members of the kinase inhibitor protein family, which can inhibit the cyclin E–CDK2 complex ( 24 – 26 ). In this study, we observed that the expressions of CDK2 and CDK6 were downregulated, while CKI p27 ^Kip1 was strongly upregulated by DIM treatment. Thus, DIM is an effective growth inhibitor of actively proliferating vascular endothelial cells by retarding the cell cycle progression.

Our results also show that the primary human vascular cells were highly sensitive, compared with tumor cell lines, to the effects of DIM. This indole significantly induced a G ₁ cell cycle arrest and inhibited migration, invasion and capillary formation at 5–10 µM in the medium of cultured cells. In contrast, we reported previously that the antiproliferative effects of DIM in the MCF-7 human breast tumor cell line required concentrations of the indole of 50 µM or higher ( 27 ). It is also of interest to note that although DIM produced a G ₁ cell cycle arrest in both types of cells, the antiproliferative effects in HUVECs occurred with the increased expression of p27 ^Kip1 , and a decreased expression of CDK2 and CDK6, whereas the cytostatic effects in MCF-7 cells occurred with a strong increase in the expression of p21 ^Waf1/Cip1 , and no change in the expression of the G ₁ acting CDKs ( 27 ). Thus, our results establish the relatively high sensitivity of HUVECs to DIM and show that the antiproliferative modes of action of DIM are cell specific.

Consistent with the inhibitory effect of DIM on endothelial cell proliferation and function in vitro , this indole also inhibited Matrigel plug vascularization and human tumor xenograft growth in vivo . The potencies of DIM in our rodent assays of tumorigenesis were similar to published findings in which DIM dose of 5 mg/kg body weight was reported to inhibit a carcinogen-induced mammary carcinogenesis in rodents ( 10 ). Because of the high sensitivity of vascular endothelial cells relative to tumor cells, to the growth inhibitory effects of DIM, our results suggest that the retardation in neovascularization may be a key mechanism in DIM-altered tumor growth.

There are several naturally occurring, as well as synthetic antiangiogenic agents that inhibit tumor neovascularization mainly by inhibiting the growth of vascular endothelial cells. Some examples are endostatin ( 28 ), squalamine (MSI-1246) ( 29 ), platelet factor-4 (PF-4) ( 30 ), and TNP-470 ( 31 ). Interestingly, PF-4, a 28-kDa-protein originally purified from platelet α granules, was found to directly interfere with the cell cycle machinery by an impaired downregulation of p21 ^Waf1/Cip1 ( 32 ). TNP-470, a low molecular weight synthetic analog of the fungal product fumagillin, is used in clinical trials as an anticancer agent. This drug was also found to inhibit neovascularization by a mechanism that involves the alteration of endothelial cell cycle progression ( 33 , 34 ). TNP-470 administrated subcutaneously at 30 mg/kg, effectively suppresses the development of a wide spectrum of tumor xenografts in mice. However, it causes a severe weight loss in the animals and neurotoxicity in patients ( 35 – 37 ). In our animal study, DIM effectively inhibited angiogenesis and tumor growth at a dose as low as 5 mg/kg, and we did not observe any weight loss or toxicity to major organs of the animals. Several other compounds extracted from a plant-based diet can prevent neovascularization and suppress the growth of malignant tumor xenografts. For example, the isoflavonoid, genistein, is a potent inhibitor of in vitro markers of angiogenesis ( 38 ). Genistein can block the proliferation of vascular endothelial cells in vitro , and reduce the growth of transplantable human prostate carcinomas and tumor vascularization in mice ( 39 ). Green tea catechins inhibit angiogenesis through the inhibition of matrix metalloproteinase ( 40 ) and the suppression of interleukin-8 production ( 41 ). Resveratrol, a natural compound in grapes, could suppress angiogenesis and tumor growth by inhibiting endothelial cell growth and vascular endothelial growth factor-induced neovascularization ( 42 ).

The cytostatic effect of DIM on actively proliferating vascular endothelial cells, which leads to reduced angiogenesis, has important implications for the use of DIM as a potential anticancer agent. Although neoplastic cells readily acquire resistance to cytotoxic chemotherapy due to a high mutation rate ( 43 ), vascular endothelial cells are genetically stable and the possibility of developing drug resistance through a genetic alteration is low ( 44 , 45 ). Furthermore, because normal vascular endothelial cells turn over slowly, it is suggested that an antiangiogenic approach targeting fast proliferating vascular endothelial cells should offer an improved efficacy and reduced toxicity for cancer therapy ( 46 ). However, in order to arrest tumor development in a dormant stage, long-term delivery of the antiangiogenic drugs is necessary ( 1 , 47 ). Since the sources of DIM are cruciferous vegetables, and the antiproliferative effects of DIM on active endothelial cells were observed at physiologically relevant concentrations ( 48 ), it is anticipated that DIM will be effective with low chronic toxicity. Although there have been no reports of interference of normal vasculogenesis or physiological angiogenesis processes through the consumption of cruciferous vegetables or DIM, this possibility should be examined more fully. Further studies of the modes of action and usefulness of DIM in the control of tumorigenic angiogenesis are in progress.

This work was supported by the Department of Defense, Army Breast Cancer Research Program Grant DAMDI 7-96-1-6159 and the NIH grant CA69056.

Author notes

1Department of Nutritional Sciences and Toxicology and 2Department of Molecular and Cell Biology, University of California, Berkeley, CA 94720, USA