Depleted Dopamine in Gastric Cancer Tissues: Dopamine Treatment Retards Growth of Gastric Cancer by Inhibiting Angiogenesis

Besides being a predominant neurotransmitter in the brain, the physiological significance of dopamine (DA) in the periphery has now been emphasized (1). In addition to its role in renal and cardiovascular systems (1), DA also has a wide spectrum of physiological actions in the stomach (2, 3, 4, 5). These actions of DA are mediated through DA receptors present in the stomach (1, 3, 4).

Tumor angiogenesis is essential for the growth and progression of solid tumors, including gastric cancer and vascular endothelial growth factor (VEGF) is thought to be the single most important angiogenic cytokine in malignancy (6, 7, 8, 9, 10). As substantial amounts of dopamine is produced by the stomach (3, 4, 5, 11, 12) and we (S. Basu and P. Dasgupta) have recently shown that VEGF-mediated angiogenesis is inhibited by DA (13), we therefore reasoned that endogenous dopamine present in stomach tissue might have a role on the growth and progression of gastric cancer in vivo. Accordingly, we investigated the status of DA in human gastric cancer tissues. As N-methyl-N′-nitro-N-nitrosoguanidine (MNNG)-induced gastric cancer in rats is an established in vivo model for human gastric adenocarcinoma (14), we also investigated the status of DA concentration in the gastric cancer tissues of these animals. We report here for the first time that unlike normal stomach tissues, there is depletion of DA in cancer tissues of the stomach in both humans and rats. Furthermore, treatment with a low nontoxic dose of DA inhibits the growth of MNNG-induced gastric cancer in rats and human gastric cancer xenografted in nude mice.

MNNG, domperidone, ethylene glycol tetra acetic acid, and reduced glutathione were obtained from Sigma (St. Louis, MO). DA was from TTK Healthcare (Chennai, India). OCT compound was from Miles Diagnostics (Elkhart, IN). Collagenase and DNase were from Roche Diagnostics Corporation (Indianapolis, IN). VEGFR-2, CD16/CD32, CD31 and CD34 monoclonal antibodies for flow cytometry analysis were from BD Biosciences PharMingen (San Diego, CA). CD31 monoclonal antibody for immunohistochemistry was from BD Biosciences PharMingen. DA D₂ receptor, VEGFR-2, and tyrosine hydroxylase antibodies for immunoblot were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Phosphotyrosine antibody was from Upstate Biotechnology (Lake Placid, NY), and the ABC kit was from Vector Lab (Burlingame, CA). Mouse VEGF quantikine ELISA kit and tumor TACS in situ apoptosis detection kit (TA 5411) were from R&D Systems (Minneapolis, MN). 5-Bromo-2′-deoxyurdine (BrdUrd) labeling and detection kit was from Roche Diagnostics Corporation. Hybond-enhanced chemiluminescence nitrocellulose membrane and enhanced chemiluminescence were from Amersham (Piscataway, NJ). DMEM and fetal bovine serum were purchased from the American Type Culture Collection (Manassas, VA).

Gastric cancer patients of both sexes with a median age of 45 years were selected. These patients did not receive any prior treatment, including medicines, and they were also free from disorders that could alter the status of DA in the stomach. Malignant gastric tumor tissue specimens were taken during surgery from different cancer patients. Similarly, benign control tissues were collected from the biopsy specimens of patients suffering from benign adenomatous polyps of the stomach, and normal tissues were collected from the histologically normal portions of the biopsy specimens not invaded by the tumors. Tissues were frozen at −80°C until assayed for DA concentration. The protocol to obtain tissue samples for this study was approved by the Institutional Review Board, and all samples were confirmed histologically.

Eight-week-old male Sprague Dawley rats were supplied from the institutional animal facility, and 4–6-week-old male nude mice were purchased from National Centre for Laboratory Animal Science (Hyderabad, India). All procedures in animals were approved by the Institutional Animal Care and Use Committee.

Human gastric carcinoma cell line Hs746T was purchased from the American Type Culture Collection and were maintained in culture as recommended by American Type Culture Collection. Briefly, the cells were cultured in DMEM with 4 mm l-glutamine adjusted to contain 1.5 g/liter sodium bicarbonate, 4.5 g/liter glucose, and 10% fetal bovine serum. The cultures were maintained at 37°C in an atmosphere of 5% CO₂/95% air. Cells were grown to ∼80% confluence before removal for xenografting. These cells were recently reported to secrete VEGF (15).

Eight-week-old male Sprague Dawley rats were divided into two groups: control and MNNG treatment groups. The MNNG treatment group was given MNNG dissolved in drinking water at a concentration of 80 mg/liter for the first 28 weeks (16). After 28 weeks, MNNG feeding was stopped, and these animals were given drinking water without MNNG for another 24 weeks (total of 52 weeks). The control group was given drinking water without MNNG for 52 weeks. At the 53^rd week of the experiment, all rats were sacrificed. As these tumors are flat, we therefore measured the tumor diameters by microcalipers to determine their growth, and this is the standard procedure for determining the growth of these tumors (16, 17). The stomach was then removed, and then the stomach specimens were examined grossly and their histology was evaluated.

To obtain tumor donor animals, three male nude mice received s.c. injections of a 0.2-ml cell suspension containing 10⁶ cells. The well developed tumors were dissected and minced, and 2-mm³ pieces of tumor tissue were transplanted s.c. to both flank areas of the experimental animals (18). When tumors became measurable, experimental groups were formed with about equal average tumor sizes, and the control and treatment groups were selected at random. Tumor growth was monitored weekly, and perpendicular tumor diameters were measured using Vernier scale calipers, and the tumor volume was calculated using the formula [(x² × y)/2] for an ellipsoid (19).

At the beginning of the 50^th week, rats with MNNG-induced gastric cancer verified to have tumor diameters of 1–2 mm in laparotomy were selected for the study. Then these tumor-bearing rats were randomly divided into three groups. The first group was injected i.p. with DA at a low nontoxic dose of 50 mg/kg, corresponding to ∼5% of the median lethal dose (LD₅₀) in rodents (20), beginning 50^th week after MNNG feeding, and was continued for 10 days. This i.p. dose of DA raises the level of DA in rats and mice to 2.1 ± 0.04 and 1.2 ± 0.01 nmol/ml, respectively, in plasma 1 min after DA injection (the normal plasma level of DA in rats and mice were 2.4 ± 0.01 and 1.5 ± 0.01 pmol/ml, respectively).⁵ A specific peripheral D₂ DA receptor antagonist, domperidone, was injected i.p. (10 mg/kg) to the second group of rats before DA treatment. The third group received the vehicle only. After completion of DA treatment, rats from all of the three groups were sacrificed.

The treatment of the mice bearing Hs746T tumors was started 30 days after transplantation, when the mean tumor volume was 100 mm³. These tumor-bearing mice were randomly divided into three groups. First group was injected with DA i.p at a dose of 50 mg/kg/day for 10 continuous days. A specific peripheral D₂ DA receptor antagonist domperidone was injected i.p. (10 mg/kg) to the second group of tumor-bearing mice before DA treatment. The third group received 0.9% NaCl, i.p., for 10 consecutive days. After the completion of the treatment, all of the mice were sacrificed, tumors were excised and processed for further experiment.

Samples were processed and assayed for DA by reverse-phase high-performance liquid chromatography with electrochemical detection. In short, tissue was weighed and then homogenized in 10.2 ml of ice-chilled 0.6 m perchloric acid containing 1.7 mg/ml ethylene glycol tetra acetic acid and 1.1 mg/ml reduced glutathione. After centrifugation for 15 min at 2500 × g at 0°C, 1.0 ml of the supernatant was adjusted to pH 8.6 with 6 m potassium hydroxide and then processed for determination of DA concentration after alumina-batch extraction. Twenty-five μl of 3 m potassium chloride were added to 200 μl of the 0.2 m perchloric acid eluate. After centrifugation, 50 μl of the supernatant were injected into the high-performance liquid chromatography system. The limit of sensitivity in the supernatant, defined as two times baseline noise, was 0.01 ng/ml for DA. If a concentration was lower than the limit of sensitivity, the concentration of the DA was assigned the value 0 (21).

A suspension of Hs746T-malignant human gastric tumors was made by passage of viable tissue through sieve and treatment with collagenase and DNase. The cells were washed, and the RBCs were lysed with PharM Lyse (BD PharMingen). The cell pellets were then resuspended in fluorescence-activated cell sorting buffer (1× PBS plus 1% BSA), preblocked with a Fc block (CD16/CD32), and then incubated with primary antibody on ice: phycoerythrin-conjugated anti-VEGF receptor-2 (anti-VEGFR-2), 1:100; CD31, 1:100; and CD34, 1:100. One million positive tumor endothelial cells were collected by fluorescence-activated cell sorting (22, 23).

Human and rat tissues were homogenized in 0.2% Triton in PBS (pH 7.5), assayed for protein content, and stored at −80°C. Either 20 μg of protein from human or rat stomach tissues or 2 μg of protein from rat adrenals was size fractionated by SDS-PAGE and then electroblotted onto Hybond-enhanced chemiluminescence nitrocellulose membrane at 30 V for 18 h in transfer buffer (20% methanol, 25 mm Tris, and 192 mm glycine). Blots were preblocked in 5% nonfat dry milk in Tris-buffered PBS and incubated with anti-tyrosine hydroxylase antibodies (1:500) for 1 h at 22°C. Blots were then incubated with horseradish peroxidase-conjugated mouse antigoat IgG (1:1000) for 1 h at 22°C and developed using enhanced chemiluminescence (5). Immunoprecipitation and Western blot analysis using antibodies against DA D₂ receptors (1:100) were also performed on Hs746T human gastric cancer cells and tumor endothelial cells isolated from Hs746T tumors. A similar experiment was performed on the same tumor endothelial cells using antibodies against VEGFR-2 (1:100) and phosphotyrosine (1:100).

The blood was collected from the rats in EDTA containing tubes on ice by cardiac puncture. The blood samples were centrifuged at 800 rpm for 30 min at 4°C, followed by 10 min at 3000 rpm. The plasma was collected and stored at −80°C until analysis. Plasma and tissue VEGF were then determined by ELISA following the protocol provided by the manufacturer (R&D Systems) that has been shown to recognize both rodent and murine VEGF (24, 25).

Tissue samples were fixed in 4% paraformaldehyde, rinsed in PBS, transferred to 30% sucrose in PBS at 4°C, and frozen in OCT compound. Immunohistochemistry was then performed on frozen tissue sections using antibody against CD31 (1:50). Microvessel density was quantitated by analyzing 10 random fields/section (26).

Tumor cell apoptosis was assessed by measuring DNA fragmentation in a standard terminal deoxynucleotidyl transferase-mediated nick end labeling assay according to the instruction of the manufacturer (R&D Systems). The number of apoptotic cells was expressed as an apoptotic index (8).

The cellular kinetics of the tumor specimens was determined by immunohistochemical staining of BrdUrd. One h before sacrificing the animals, 20 mg/kg body weight of BrdUrd were injected i.p. into each animal. Subsequently, the animals were sacrificed, and tumors were excised and fixed in buffered formalin and embedded in paraffin. Cross-sections of the tissues were cut 4-μm thick. Tissue sections were then immunostained by using BrdUrd labeling and detection kit. The percentage of stained cells was counted and expressed as labeling index (16).

Mean and SE were calculated. Differences among groups were evaluated by ANOVA and the unpaired Student’s test or Dunn’s multiple comparison tests.

MNNG-treated rats developed a single (Fig. 1,A) intestinal type of adenocarcinoma of the stomach similar to human gastric cancers (Fig. 1 B), and our results are in conformity with other studies (14, 27).

In normal stomach tissues of humans and rats substantial amounts of DA (Table 1) and tyrosine hydroxylase (Fig. 2, A and B), the rate-limiting enzyme required for DA synthesis (5) was observed. On the contrary, there was total absence of endogenous DA (Table 1) and tyrosine hydroxylase (Fig. 2, A and B) in malignant tumor tissues of the stomach collected from humans and rats.

MNNG induced gastric cancer tissues showed striking angiogenic response as evident from CD31 expression (Fig. 3,A). In contrast, we found CD31 expression in normal rat gastric tissues to be very low or insignificant (20–30/10 randomly chosen high power fields), and our results are in conformity with a recent study that indicated similar results (28). Furthermore, these malignant tumor tissues also showed a significantly higher concentration of VEGF in comparison to normal controls (Table 2; 6.7 ± 0.8 versus 0.7 ± 0.02 ng/g tissue, mean ± SE, P < 0.05). The plasma VEGF levels were also significantly higher in the malignant tumor-bearing rats when compared with normal controls (Table 2; 25.7 ± 1.6 versus 8.8 ± 1.0 pg/ml, mean ± SE, P < 0.05). These results corroborate the striking angiogenesis observed in these tumors and, to our knowledge, is the first article that indicates that MNNG-induced malignant gastric tumors in rats also secrete VEGF.

DA when administered to gastric adenocarcinoma-bearing rats showed 3-fold decrease (P < 0.5) in tumor diameter than control animals (Table 3). This was also associated with significant inhibition of tumor cell proliferation as evident from reduction of BrdUrd incorporation in malignant stomach tissues after DA treatment in both rats and mice (Table 4). These DA-treated animals with decreased tumor diameter also showed significant inhibition of tumor angiogenic response than untreated controls as evidenced from decreased frequency of CD31⁺ cells in DA-treated tumor tissues (Fig. 3, A–D). Decreased angiogenic response in tumor tissues after DA treatment was also associated with increased apoptosis of tumor cells when compared with controls (Fig. 4, A–D). The increase in apoptosis of tumor cells after DA treatment are in conformity with other studies that indicate that inhibition of angiogenesis results in apoptosis of tumor cells (8, 26). In contrast, DA did not inhibit tumor growth (Table 3) in rats that received prior treatment of domperidone (10 mg/kg i.p.), a specific peripheral D₂ DA receptor antagonist, thereby confirming that the action of DA is through D₂ DA receptor. The other two catecholamines, epinephrine and norepinephrine, did not inhibit either angiogenesis or tumor growth (data not shown).

Similarly DA when administered to Hs746T-bearing mice also showed 3-fold decrease (P < 0.5) in tumor volume than control animals (Table 5). This was also associated with significant inhibition of tumor cell proliferation as evident from reduction of BrdUrd incorporation in malignant stomach tissues after DA treatment in mice (Table 4). Decreased angiogenic response in tumor tissues after DA treatment was also associated with increased apoptosis of tumor cells when compared with controls (Fig. 5). The increase in apoptosis of tumor cells after DA treatment are in conformity with other studies that indicate that inhibition of angiogenesis results in apoptosis of tumor cells (8, 19). These DA-treated animals with decreased tumor volume also showed significant inhibition of tumor angiogenic response than untreated controls as evidenced from decreased frequency of CD31⁺ cells in DA-treated tumor tissues (Fig. 3, E–H). In contrast, DA did not inhibit tumor growth (Table 5) in mice that received prior treatment of domperidone (10 mg/kg i.p.), a specific peripheral D₂ DA receptor antagonist, thereby confirming that the action of DA is through D₂ DA receptor. This result was additionally strengthened as we found D₂ DA receptors in tumor endothelial cells isolated from these tumor-bearing mice (Fig. 6,A). The D₂ DA receptors were absent in Hs746T gastric cancer cells (Fig. 6,A). Furthermore, VEGF is thought to induce angiogenesis by phosphorylating VEGFR-2, and our recent article had indicated that DA inhibits VEGFR-2 phosphorylation in vitro (13). In the present investigation, we found that DA treatment strongly inhibited VEGF-induced phosphorylation of VEGFR-2 in tumor endothelial cells isolated from Hs746T-malignant gastric tumors (Fig. 6,B) and pretreatment with domperidone, a specific peripheral D₂ DA receptor antagonist, significantly abrogated this effect (Fig. 6 B), indicating the specificity of DA action. The other two catecholamines, epinephrine and norepinephrine, did not inhibit either angiogenesis or tumor growth (data not shown).

Endogenous DA produced in substantial amounts in normal stomach tissues regulates several physiological functions of the stomach, including reduction of gastric acid secretion, stimulation of bicarbonate, and mucus secretion (2, 3, 4). Here, we show for the first time that DA is depleted in both human gastric cancer and MNNG-induced adenocarcinoma of the stomach in rats. Furthermore, this finding is also significant because our results indicate that there is a significant retardation in the growth of gastric cancer after treatment with exogenous DA at a low nontoxic dose.

VEGF is thought to be the single most important cytokine essential for tumor angiogenesis (6, 7, 8, 9, 10), and we had recently demonstrated that DA specifically and significantly inhibits VEGF-mediated angiogenesis (13). DA acted through D₂ DA receptors to induce endocytosis of VEGFR-2, which is critical for promoting angiogenesis, thereby preventing VEGF binding, receptor phosphorylation, and subsequent signaling steps (13). Malignant human gastric tumors secrete VEGF and therefore are highly angiogenic tumors (8). In the present experiment, MNNG-induced malignant gastric tumor-bearing rats showed significant elevation of plasma VEGF, and this correlated well with their angiogenic nature. There is also a recent article that indicates that the human gastric cancer cell line Hs746T selected for our experiment also secretes considerable amount of VEGF (15). Therefore, one of the important mechanisms by which DA may retard gastric cancer growth is through inhibition of VEGF-mediated angiogenesis. Our present results confirm that DA retards the growth of gastric cancer by inhibiting VEGF-induced angiogenesis. The underlying mechanism of this action has been shown to be mediated by inhibition of VEGFR-2 phosphorylation in tumor endothelial cells involving DA D₂ receptors present on these cells as treatment with domperidone, a specific peripheral DA D₂ receptor antagonist, showed significant abrogation of this inhibitory effect of DA. However, there may be other mechanisms also by which DA may inhibit tumor growth (29). It would also be noteworthy to mention here that there are few contradictory studies that have shown that administration of a DA agonist stimulated MNNG-induced carcinogenesis and a DA antagonist inhibited MNNG-induced carcinogenesis (30, 31, 32). However, in those articles, the authors had only shown the effects of DA agonist and antagonist on the process of MNNG-induced carcinogenesis (incidence of malignant gastric tumors), and nothing was mentioned regarding the effects of those agonist and antagonist on the growth and progression of a developed gastric cancer. In addition, there is also an article that indicates that DA inhibits the process of MNNG-induced glandular stomach carcinogenesis in rats (33). However, that study was also limited to the incidence of gastric cancer in the rats, and no effect of DA on the growth and progression of malignant gastric tumor was mentioned (33). Therefore, in brief, the present results are of both physiological and clinical significance because it has been shown for the first time that there is an inverse correlation between endogenous stomach DA and gastric cancer. These results also indicate a direct gastric cancer growth inhibitory action of nontoxic dose of DA in vivo.

Finally, gastric cancer is one of the leading causes of cancer death worldwide (34). Radiation therapy or chemotherapy does not affect the length and quality of life of patients with advanced gastric cancer (8). Thus, novel therapies are needed to target the progression of these tumors. DA is already in clinical use with an established safety record (35). Therefore, DA can be tested without the usual delays that would be required to assess the safety of a new drug in patients with gastric cancer in which VEGF-mediated angiogenesis has an important pathogenic role (8). In addition, information concerning the potent antitumor angiogenic activity of DA in gastric cancer needs to be disseminated widely and rapidly so that clinical trials can be started as soon as possible to combat this very common cancer, which, as of yet, has no significant treatment.

We thank Uttio Roy Chowdhury of Signal Transduction and Biogenic Amines laboratory for technical help, Drs. Jahar Majumder and Jaydip Biswas, Department of Surgical Oncology, Chittaranjan National Cancer Institute, and Dr. Partha Bhattacharjee of Reliance Gastrointestinal Endoscopy Clinic (Calcutta, India) for providing human gastric tissue samples.