New Activity of Spironolactone
Inhibition of Angiogenesis In Vitro and In Vivo
Background The formation of new blood vessels (angiogenesis) is a critical component in a variety of pathological settings, including solid tumor growth, macular degeneration, and atherosclerosis.
Methods and Results We have found that orally administered spironolactone inhibited the area of angiogenesis induced by basic fibroblast growth factor (bFGF) in a rabbit corneal micropocket assay. Additionally, spironolactone inhibited bFGF- and vascular endothelial growth factor–stimulated capillary endothelial cell proliferation in vitro, inhibited bFGF-stimulated capillary endothelial cell chemotaxis in vitro, and caused avascular zones when placed on the chick chorioallantoic membrane. Experiments analyzing spironolactone metabolites revealed that the major human metabolites 6β-hydroxy-7α-thiomethyl spironolactone and canrenoic acid retained antiangiogenic activity. The antiangiogenic activity appears to be unrelated to the antiandrogenic and antimineralocorticoid effects of spironolactone.
Conclusions These experiments hold promise for the potential use of spironolactone as an orally administered drug for the treatment of many diverse diseases dependent on angiogenesis.
Neovascularization plays a critical role in the pathogenesis of a variety of diseases such as solid tumor growth,1 2 3 macular degeneration, hemangiomas, psoriasis,4 and atherosclerosis.5 6 7 Because the treatment of such diseases with antiangiogenic therapy is likely to be long-term, we have been searching for orally efficacious angiogenesis inhibitors.
SL (Fig 1⇓), a renal aldosterone antagonist, is used clinically for the treatment of hypertension and congestive heart failure. SL also has antiandrogenic activity, which makes it useful for the treatment of hirsutism,8 acne, and seborrhea.9 We hypothesized that this drug would inhibit angiogenesis on the basis of the observation that SL has an unexplained side effect of amenorrhea.10 11 We proposed that the cessation of menstruation associated with SL could be due to angiogenesis inhibition. In this study, we found that orally administered SL inhibits angiogenesis in a rabbit corneal neovascularization assay. Additionally, SL directly inhibits bFGF- and VEGF-stimulated capillary endothelial cell and smooth muscle cell proliferation in vitro, inhibits bFGF-stimulated capillary endothelial cell chemotaxis in vitro, and inhibits angiogenesis on the chick CAM.
CAM assays were performed as described,12 13 and the effects on the developing vasculature were recorded at 48 hours after implantation of a pellet containing 200 μg of various drugs in 10 μL of 0.5% carboxymethylcellulose. The dose of 200 μg was used because this was the maximal dose used13 when angiostatic steroids were tested on the CAM. Higher concentrations produced a mild angiogenic reaction on the CAM.13 The response was scored as effective when the avascular zone equaled or exceeded 180° around the pellet. The results were evaluated statistically with one-way ANOVA with 10 pairwise comparisons performed, and probability values were appropriately adjusted for multiple comparisons. SL, canrenoic acid, and heparin were purchased from Sigma Chemical Co. 6β-OH-7α-thiomethyl-SL (SC-26962); 6β,7β-epoxy-canrenone (6β,7β-epoxy-CAN, SC-17146); and cyproterone (E-00004) were a gift from Searle.
BCE cells were grown on gelatinized surfaces (1.5% gelatin in PBS) in DMEM containing 10% heat-inactivated BCS and supplemented with 2 mmol/L glutamine, 100 U/mL each of penicillin and streptomycin (GPS; Irvine Scientific), and 3 ng/mL human recombinant bFGF (Scios Nova). Media, PBS, and BCS were obtained from JRH Biosciences. Bovine aortic endothelial cells, Lewis lung carcinoma cells, human retinal pigmented epithelial cells, and melanoma B16 F-10 cells were grown in T-25 flasks in DMEM containing 10% heat-inactivated FCS (Hyclone Laboratories) and supplemented with GPS. NIH-3T3 fibroblast cells were grown in T-25 flasks in DMEM with high glucose (4.5 g/L) containing 10% heat-inactivated BCS (Colorado Serum Co) and GPS. Bovine aortic smooth muscle cells were grown in T-25 flasks in DMEM with 10% BCS and GPS.
BCECs were plated at 24 000 cells/mL into gelatinized 24-well dishes in the absence of added growth factors. All other cell types were plated at 20 000 cells/mL in 24-well dishes. After the cells were allowed to attach overnight, the appropriate fresh media were applied, containing differing concentrations of SL or aldosterone (Sigma). Drugs were made soluble in DMSO (Fisher Scientific) or 100% ethanol, and control wells received equal volumes (0.1%) of vehicle alone. All reagents were added to the wells in a volume of 500 μL DMEM. The media for BCECs were supplemented with either bFGF (1 ng/mL, bFGF experiments) or VEGF (R&D Systems, 5 ng/mL, VEGF experiments). BCE, NIH-3T3 fibroblast, and bovine aortic smooth muscle cells were assayed in 5% BCS, whereas the other cell types were assayed in 2.5% FCS because of their more rapid growth curves. The cells were incubated for 3 days at 37°C under 10% CO2 in air. Cells were then washed with PBS, detached by trypsinization (0.05% trypsin, 0.53 mmol/L EDTA, from Life Technologies), resuspended in azide-free isotonic diluent (Hematall, Fisher Scientific), and counted by Coulter counter. Each condition was prepared in triplicate, and the experiments were carried out two or three times. The IC50 values were obtained by linear regression and are reported as mean±SEM.
Migration was measured with a multiwell chamber assay.14 Polycarbonate filters (Nucleopore Corp) were coated with collagen type I, 100 μg/mL (Collaborative Biomedical Products). The migration of 7100 cells placed in the upper well toward bFGF (10 ng/mL) in the lower well was assayed after a 4-hour incubation at 37°C. At the end of this period, the chambers were disassembled, and the side of the filter to which the cells were added was scraped. The migrated cells on the distal side of the filter were fixed in formalin, washed in PBS, and stained with Gill's triple-strength hematoxylin (Polysciences). All cells within an area were counted visually.
Rabbit Corneal Micropocket Assay
Forty female New Zealand Albino rabbits were anesthetized with ketamine 30 mg/kg IM and xylazine 2.2 mg/kg IM. Corneal neovascularization was induced by pellets containing bFGF and sucrose aluminum sulfate (Sucralfate, Bukh Meditec) implanted into micropockets of both corneas of each rabbit as described15 with the following modifications: Pellets were made by mixing 30 μg bFGF with 120 mg sucrose aluminum sulfate. To this suspension, 336 μL of 12% (wt/vol) of poly(hydroxyethyl methacrylate) (Hydron; Interferon Sciences) in ethanol and 420 μL saline was added. Aliquots (10 μL) of this mixture were then pipetted onto 75 Teflon pegs and allowed to dry. The final amount of bFGF per pellet was 400 ng. The animals treated with SL, canrenoic acid, or vehicle alone were fed daily from 1 day after implantation by gastric lavage. SL (100 mg/kg) was suspended in filtered peanut oil, canrenoic acid (100 mg/kg) was suspended in 0.9% normal saline, and controls received equal volumes of peanut oil. Administration of drug or vehicle was followed by 10 mL of 0.9% normal saline through the lavage tube. Amiloride (10 mg/kg) (Sigma) was suspended in 0.9% normal saline and injected intraperitoneally daily from 1 day after implantation.
The animals were examined with a slit lamp every other day from day 6 through day 10 by the same examiner, who was blinded to the experiment. The area of corneal neovascularization was determined by measurement with a reticule of the vessel length (L) from the limbal vessels and the number of clock hours (C) of the limbus involved. A formula was used to determine the area of a circular band segment as previously described15 : C/12×3.1416[r2−(r−L)2], where r=6 mm, the measured radius of the rabbit cornea. Statistical analysis was performed with one-way ANOVA on ranked data (this yields a nonparametric analysis). When only one drug was tested in an experiment (eg, SL), comparisons between control and SL-treated rabbits were made with a Wilcoxon two-sample test. When multiple drugs were tested in an experiment, comparisons were made between control and SL-, canrenoic acid–, and amiloride-treated rabbits and between canrenoic acid– and SL-treated rabbits with a pairwise test. The probability values were appropriately adjusted for multiple comparisons. Rabbit weights were monitored throughout the study. Serum electrolytes, blood urea nitrogen, and creatinine in the control and SL-treated animals on day 10 were measured by Corning Bioran. Animal studies were reviewed and approved by the Animal Care and Use Committee of Children's Hospital and are in accordance with the guidelines of the Department of Health and Human Services.
UPA Activity Assay
The thiobenzyloxycarbonyl-l-lysinate assay for trypsin-like enzymes was performed as described,16 with a substrate concentration of 7.3×10−5 mol/L. Human urokinase (Mr, 33 000, 0.3 μg/mL) was purchased from American Diagnostica Inc. N-CBZ-l-lysinethiobenzylester and 5,5′-dithio-bis-(2-nitrobenzoic acid) were obtained from Sigma. Experiments were performed three times, and data are presented as mean±SEM.
TPA Activity Assay
BCECs were plated at 24 000 cells/mL into gelatinized 24-well dishes in DMEM with 10% BCS, GPS, and 3 ng/mL bFGF. At confluence, the cells were incubated for 24 to 30 hours in DMEM with 5% BCS, GPS, and 3 ng/mL bFGF in the presence of SL (10 μmol/L), hydrocortisone 21-phosphate disodium salt (10 μmol/L, Sigma), or 0.1% DMSO alone. The conditioned medium was assayed for TPA activity with the Spectrolyse TPA Activity and Inhibitor Assay Kit (product 452, American Diagnostica Inc) according to the manufacturer's instructions, with the following modifications: assays were performed in 96-well flat-bottom plates with 100 μL tissue activator reagent, 4 μL sample, 4 μL Desafib, and 120 μL Stop solution. Absorption was read in a Dynatech Laboratories Inc MR 700 ELISA reader. The BCECs were washed with PBS, detached by trypsinization, resuspended in Hematall, and counted by Coulter counter. A TPA standard curve was run with each assay, and TPA activities of the samples in IU/mL were determined from comparison to the standard curve. TPA activity (IU/mL) was divided by the number of cells per well. Each condition was prepared in triplicate; the experiments were carried out three times, and data are presented as mean±SEM.
Our initial investigations were performed on the CAM. Antiangiogenic activity of the developing vasculature was seen when SL (200 μg) was placed on the CAM (35% avascular zones, P≤.001, Table 1⇓). When heparin (50 μg) was added to SL, there were 72% avascular zones (P≤.001). This was statistically different from SL alone (P≤.001).
SL is rapidly and extensively metabolized in humans. Sulfur-containing products, including 6β-OH-7α-thiomethyl-SL (Fig 1⇑), are the predominant metabolites.17 Canrenone, which exists in steady-state equilibrium with canrenoic acid (Fig 1⇑), is another active metabolite of SL and is further metabolized to 6β,7β-epoxy-CAN (Fig 1⇑) when incubated with liver microsomes in vitro.18 Because SL is extensively metabolized in humans, we wanted to see whether the major metabolites retained antiangiogenic activity. When pellets containing 6β,7β-epoxy-CAN (200 μg) were placed on the CAM, 31% had avascular zones (P≤.02). Similarly, pellets containing 6β-OH-7α-thiomethyl-SL (200 μg) produced 35% avascular zones (P≤.004, Table 1⇑). There was no statistical difference between the effects of SL and those of 6β,7β-epoxy-CAN; those of SL and of 6β-OH-7α-thiomethyl-SL; or those of 6β,7β-epoxy-CAN and of 6β-OH-7α-thiomethyl-SL. Pellets containing canrenoic acid (200 μg) induced scar formation around the pellet, making the results uninterpretable (data not shown).
SL is a testosterone-receptor antagonist that competitively inhibits the binding of dihydrotestosterone to cytoplasmic receptors.19 To see whether the inhibition of angiogenesis by SL correlated with antiandrogenic activity, we tested cyproterone (Fig 1⇑), another antiandrogen with a steroidal structure that inhibits dihydrotestosterone binding to cytoplasmic receptors.20 When pellets containing cyproterone (200 μg) were placed on the CAM, only 4% had avascular zones, which was not statistically significant (Table 1⇑).
We next tested the effect of SL on the inhibition of vascular endothelial cell proliferation. SL inhibited the proliferation of both bFGF- and VEGF-stimulated BCECs in a concentration-dependent manner. The IC50 for bFGF-stimulated BCECs was 5±1 μmol/L. VEGF-stimulated BCECs and smooth muscle cells were as sensitive as bFGF-stimulated BCECs to SL (Fig 2⇓). Other cell types were sensitive to SL but only at higher concentrations (Table 2⇓). Aldosterone (10 μmol/L) had no effect on BCEC proliferation in the presence or absence of bFGF (data not shown).
SL inhibited BCEC migration toward bFGF (10 ng/mL) in a dose-dependent fashion (Fig 3⇓). Chemotaxis inhibition was almost complete with 10 μmol/L and 20 μmol/L SL. DMSO, the vehicle used to dissolve the SL, had no significant effect on the migration of endothelial cells. These results suggest that SL may inhibit angiogenesis by preventing the migration of endothelial cells in response to angiogenic factors such as bFGF.
Rabbit Corneal Micropocket Assay
We subsequently evaluated the in vivo effect of SL on angiogenesis induced by bFGF in the rabbit corneal micropocket model. In the first experiment, treatment with 100 mg/kg PO daily of SL resulted in an inhibition of the area of vascularized cornea on day 10 of 39%. The median area for the control group was 52 mm2 (n=15 eyes) and for the SL-treated group, 32 mm2 (n=16 eyes, P=.0007). In our second experiment, we compared the effects of SL (n=10 eyes), canrenoic acid (n=10 eyes), and amiloride (n=10 eyes) with those in control (n=10 eyes) rabbits. SL was inhibitory at day 6 (P=.02), day 8 (P=.004), and day 10 (P=.05) (Fig 4⇓). Canrenoic acid was also found to be inhibitory at days 6 (P=.003) and 8 (P=.004) (Fig 4⇓). The effect of canrenoic acid was not statistically different from that of SL at all time points. Amiloride, like SL, is a potassium-conserving diuretic, although it is chemically unrelated to SL (Fig 1⇑). The area of neovascularization in amiloride-treated rabbits was not statistically different from control at day 6, 8, or 10 (Fig 4⇓).
None of the rabbits exhibited signs of toxicity (ie, lethargy, diarrhea) during the study. There was no evidence of hyponatremia, hyperkalemia, hypochloremia, or azotemia (Table 3⇓). Weight loss between day 0 and day 10 in the SL-treated rabbits was only 3.0%, which was not statistically significant and would not account for the effect on angiogenesis. Weight loss in the canrenoic acid–treated animals was only 0.7%, which was not statistically significant.
Assay for UPA Activity
Amiloride is an inhibitor of UPA activity.21 We compared the effect of SL on the catalytic activity of UPA with that of amiloride by use of a colorimetric assay based on the cleavage of a general substrate for trypsin-like proteases. Consistent with a previous report,21 amiloride inhibited UPA activity. The effect on UPA activity as percentage of control for each concentration of amiloride was as follows: 1 μmol/L, 126±7.5%; 10 μmol/L, 74.7±23%; and 100 μmol/L, 48.8±11.3%. SL did not inhibit the UPA-catalyzed hydrolysis of this compound. The effect on UPA activity as percentage of control for each concentration of SL was as follows: 1 μmol/L, 114±12%; 10 μmol/L, 99±14.7%; and 100 μmol/L, 211±64.6%.
Assay for TPA Activity
A class of angiostatic steroids that inhibit angiogenesis when administered with heparin has been described.13 The mechanism of action of angiostatic steroids is through suppression of TPA activity via stimulation of plasminogen activator inhibitor synthesis.22 Since SL is a steroid analogue, we compared the effect of SL with that of hydrocortisone on TPA activity in the conditioned medium of BCECs stimulated with bFGF. Hydrocortisone (10 μmol/L) served as a positive control for TPA activity inhibition. The TPA activity (percentage of control) was 39.7±10% for hydrocortisone. Neither SL (10 μmol/L) nor canrenoic acid (10 μmol/L) inhibited TPA activity in three experiments. The TPA activity (percentage of control) was 231.8±51% and 103.5±26% for SL and canrenoic acid, respectively.
This study examined the antiangiogenic activity of SL both in vitro and in vivo. Orally administered SL inhibited angiogenesis induced by bFGF in the rabbit corneal micropocket assay. The dose used produced no signs of toxicity in the rabbits, as demonstrated by normal weights, behavior, serum electrolytes, and renal function tests. Topical administration of SL induced avascular zones on the chick CAM. SL also directly inhibited both bFGF- and VEGF-stimulated endothelial cell proliferation in vitro and inhibited capillary endothelial cell chemotaxis toward bFGF in vitro. These results suggest that SL may inhibit angiogenesis by preventing proliferation and migration of endothelial cells in response to angiogenic factors.
SL is rapidly and extensively metabolized in humans, and the major breakdown product, 7α-thiomethyl-SL,17 produced avascular zones on the CAM. In addition, orally administered canrenoic acid, another metabolite formed in humans,18 also inhibited angiogenesis induced by bFGF in the rabbit corneal micropocket assay to the same extent as SL. This finding suggests that if SL were used for treating human angiogenesis-dependent diseases, the antiangiogenic activity of SL would not be lost by metabolism. Furthermore, we propose that the unexplained association between SL administration and amenorrhea that has been reported in the literature10 11 may be due to the antiangiogenic actions of SL and its metabolites. Ovarian and uterine tissues have been shown to contain and produce angiogenic factors23 ; thus, amenorrhea would be an expected side effect of an angiogenesis inhibitor.
Angiostatic steroids have been shown to inhibit angiogenesis on the CAM when administered with heparin.13 SL is unique in this class of steroids in that it is antiangiogenic in the absence of heparin. The effects of angiostatic steroids are mediated through dissolution of basement membrane24 via inhibition of TPA activity that results from increased plasminogen activator inhibitor activity.22 We found that SL had no effect on TPA activity, thus implicating a mechanism different from that of angiostatic steroids.
SL is a testosterone antagonist.19 Since cyproterone, another testosterone antagonist,20 was not effective at inhibiting angiogenesis on the CAM, it is unlikely that SL is inhibiting angiogenesis through its antiandrogenic effects.
SL is also an aldosterone antagonist. No aldosterone was present in our culture media used for our cell proliferation or migration assays, and adding aldosterone to the media had no effect on BCEC proliferation. It is therefore unlikely that SL is inhibiting angiogenesis through its antimineralocorticoid effect. This is consistent with previous reports13 that have shown that angiostatic steroids are not dependent on antimineralocorticoid or antiglucocorticoid activity.
Amiloride, another potassium-sparing diuretic, has previously been shown to inhibit inflammatory angiogenesis induced by prostaglandin E2 at a dose of 10 mg/kg IP (which resulted in a 7% weight loss, indicating that this was the maximal tolerated dose without significant weight loss).25 SL, unlike amiloride, did not inhibit UPA activity in vitro, indicating that the mechanism of action of SL is different from that of amiloride. In addition, administration of amiloride at 10 mg/kg IP had no effect on angiogenesis in our rabbit cornea neovascularization model induced by bFGF. This might be because angiogenesis in our model is due to direct cytokine stimulation and not mediated through inflammation.15
These experiments hold promise for the potential use of SL as an orally administered drug for the treatment of diseases dependent on angiogenesis. Since SL selectively inhibits both endothelial and smooth muscle cells, it may be particularly suited for treating atherosclerosis and restenosis after angioplasty. In these diseases, both neovascularization of the walls of coronary arteries5 6 7 and smooth muscle cell proliferation26 27 28 are known to occur. Interestingly, van Belle et al26 showed that neointimal thickening after balloon denudation in a rabbit model was inhibited by SL, although the mechanism was unknown. Our finding that SL directly inhibits smooth muscle cell proliferation in vitro may explain this result.
Finally, since SL inhibited VEGF-stimulated BCEC proliferation as well as bFGF, this drug might be useful in the treatment of a variety of diseases dependent on angiogenesis, such as solid tumor growth, macular degeneration, hemangiomas, and psoriasis.
Selected Abbreviations and Acronyms
|BCEC||=||bovine capillary endothelial cell|
|BCS||=||bovine calf serum|
|bFGF||=||basic fibroblast growth factor|
|CAM||=||chick chorioallantoic membrane|
|TPA||=||tissue plasminogen activator|
|UPA||=||urokinase plasminogen activator|
|VEGF||=||vascular endothelial growth factor|
Special thanks to Dr Judah Folkman for review of this manuscript and to Dr Michael Loughnan, Geraldine Jackson, Catherine Butterfield, Dr Klio Chatzistefanou, and Sean Yetman. This study was supported by a grant from Entremed (Rockville, Md). Dr D'Amato is a Howard Hughes Medical Institute physician research fellow.
Reprint requests to Robert D'Amato, MD, PhD, 300 Longwood Ave, Children's Hospital, Ender's 1006, Boston, MA 02115. E-mail damato email@example.com.
- Received February 29, 1996.
- Revision received May 23, 1996.
- Accepted June 13, 1996.
- Copyright © 1996 by American Heart Association
Folkman J. Tumor angiogenesis: a possible control point in tumor growth. Ann Intern Med. 1975;82:96-100.
Folkman J. What is the evidence that tumors are angiogenesis dependent? J Natl Cancer Inst. 1989;82:4-6.
Heistad DH, Armstrong ML. Blood flow through vasa vasorum of coronary arteries in atherosclerotic monkeys. Arteriosclerosis. 1986;6:326-331.
Crum R, Szabo S, Folkman J. A new class of steroids inhibits angiogenesis in the presence of heparin or a heparin fragment. Science. 1985;230:1375-1378.
Boyden SV. The chemotactic effect of mixtures of antibody and antigen on polymorphonuclear leukocytes. J Exp Med. 1962;115:453-466.
D'Amato RJ, Loughnan MS, Flynn E, Folkman J. Thalidomide is an inhibitor of angiogenesis. Proc Natl Acad Sci U S A. 1994;91:4082-4085.
Cook CS, Hauswald C, Oppermann JA, Schoenhard GL. Involvement of cytochrome P-450IIIA in metabolism of potassium canrenoate to an epoxide: mechanism of inhibition of the epoxide formation by spironolactone and its sulfur-containing metabolite. J Pharmacol Exp Ther. 1993;266:1-7.
Reynolds LP, Killilea SD, Redmer DA. Angiogenesis in the female reproductive system. FASEB J. 1992;6:886-892.
Clowes AW, Clowes MM, Fingerle J, Reidy MA. Regulation of smooth muscle cell growth in injured artery. J Cardiovasc Pharmacol. 1989;14(suppl 6):S12-S15.