The first two authors contributed equally to this work.
Nitroxyl inhibits breast tumor growth and angiogenesis†
Article first published online: 12 DEC 2007
Copyright © 2007 Wiley-Liss, Inc.
International Journal of Cancer
Volume 122, Issue 8, pages 1905–1910, 15 April 2008
How to Cite
Norris, A. J., Sartippour, M. R., Lu, M., Park, T., Rao, J. Y., Jackson, M. I., Fukuto, J. M. and Brooks, M. N. (2008), Nitroxyl inhibits breast tumor growth and angiogenesis. Int. J. Cancer, 122: 1905–1910. doi: 10.1002/ijc.23305
Our study shows for the first time that nitroxyl suppresses breast tumor growth in vitro and in vivo. Significantly, we find that this effect is correlated with inhibition of key regulators of glycolysis and angiogenesis.
- Issue published online: 19 FEB 2008
- Article first published online: 12 DEC 2007
- Manuscript Accepted: 5 OCT 2007
- Manuscript Received: 23 AUG 2007
- Angeli's salt;
- breast cancer;
Nitroxyl (HNO) can inhibit the glycolytic enzyme glyceraldehyde 3-phosphate dehydrogenase (GAPDH). Because of the importance of glycolysis in many malignant cells, we thus propose that HNO can adversely affect tumor growth. This hypothesis was tested using in vitro and in vivo models of breast cancer. We report here for the first time that HNO suppresses the proliferation of both estrogen receptor (ER)-positive and ER-negative human breast cancer cell lines, in a dose dependent manner. Mice treated with HNO either injected into the tumor itself or via the intraperitoneal approach had smaller xenograft tumor size. In addition to significantly decreased blood vessel density in the HNO-treated tumors, we observed lower levels of circulating serum vascular endothelial growth factor (VEGF). Accordingly, there was a decrease in total HIF-1α (hypoxia-inducible factor) protein in HNO-treated tumor cells. Further studies showed inhibition of GAPDH activity in HNO-treated human breast cancer cell lines and in HNO-treated tumor tissue derived from xenografts. One explanation for the multiplicity of actions observed after HNO treatment could be the effect from the initial inhibition of GAPDH, providing a potential therapeutic avenue based upon blocking glycolysis resulting in decreased HIF-1α, thus leading to angiogenesis inhibition. Therefore, HNO appears to act via mechanism(s) different from those of existing breast cancer drugs, making it a potential candidate to overcome known and emerging drug resistance pathways. © 2007 Wiley-Liss, Inc.
Nitrogen oxides are reported to be involved in a number of important physiological processes.1 Nitric oxide (NO), for example, is an established and endogenously generated vasorelaxant. Other nitrogen oxides derived from or related to NO have also been proposed to be important in signal transduction pathways and as effector species in immune response. Thus far, the focus has been on NO and more oxidized species such as peroxynitrite, nitrogen dioxide, nitrite and dinitrogen dioxide. Although physiological accessible, species that are reduced relative to NO have received relatively little interest.2 Nitroxyl (HNO) is the 1-electron reduced and protonated congener of NO.3 Recently, HNO has been found to have important therapeutic potential as a treatment for heart failure4 and possibly as a preconditioning agent to protect against ischemia-reperfusion toxicity,5 along with its current use as an antialcoholic drug.6 The chemistry of HNO is surprisingly complex, but it is predicted that the predominant targets for the actions of HNO are thiols or thiolproteins.7 In fact, the utility of HNO as a pharmacological agent for the treatment of alcoholism is due to its ability to react with and modify the thiolprotein aldehyde dehydrogenase, an enzyme critical for the metabolism of ethanol. Other examples of the thiolprotein reactivity of HNO have also been reported.8 Of particular relevance to the studies discussed herein, Lopez et al.9 have recently reported that HNO is capable of potent inhibition of glycolysis, without affecting the thiol status of the cell (i.e., no change in glutathione levels or oxidation status). The suppression of glycolysis observed in this study was a result of the inhibition of glyceraldehyde 3-phosphate dehydrogenase (GAPDH), another classic thiol containing dehydrogenase. Because of the importance of glycolysis in cancer,10, 11 we set out to investigate whether HNO may have any effects on tumor cells. In the current study, we present for the first time in vitro and in vivo data, which suggests that HNO suppresses breast cancer growth, and may be involved in key regulators of glycolysis and tumor angiogenesis.12
Material and methods
Cell culture and hypoxic condition
Human breast cancer cell lines MDA-MB-231 and MCF-7 and human embryonic kidney HEK-293 cells were purchased from American Tissue Type Culture Collection (ATCC, Rockville, MD). The cells were maintained in RPMI-1640 medium (In Vitrogen, Carlsbad, CA) with 10% heat-inactivated FCS (fetal calf serum), at 37°C in 5% CO2 and 95% air. Human umbilical vein endothelial cells (HUVEC) were purchased from Clonetics (San Diego, CA). The cells were plated on tissue culture flasks coated with 1.5% gelatin (Difco, Detroit, MI) in phosphate buffered saline (PBS). They were maintained in endothelial growth media (EGM, endothelial cell growth medium completed with 10 ng/ml human epithelial growth factor, 2% FCS, Clonetics). For experiments with oxygen deprivation, cells were placed in a hypoxic chamber (Billups-Rothenberg, Del Mar, CA) with 95% N2 and 5% CO2.
Proliferation assays were performed as previously described.13 Cells were plated onto 48-well plates at 5,000 cells/well for MDA-MB-231 and 10,000 cells/well for MCF-7 cells. The plates were kept in normoxic conditions overnight. Control cells were kept in growth media. In experimental conditions, media was removed and cells were treated with either the HNO donor Angeli's salt (AS) or decomposed Angeli's salt (dAS) solutions for 45 min. Afterwards, the solution was removed and replaced with growth media for the rest of the assay. To prepare AS treatment, the AS was added to PBS at pH 6.5 containing 0.2 mM CaCl2 and used on cells within 5 min. dAS solution was made similarly, but the previous day, to deactivate AS. For hypoxic studies, the plates were placed in a hypoxic chamber during the treatment phase and subsequently thereafter. Six hours after treatment, 1 μCi of [methyl-3H] thymidine was added to each well. Approximately 15 hr later, the cells were lysed and radioactivity counted.
Mouse tumor model
Briefly, severe combined immunodeficient (SCID) Beige CB17 mice purchased from Charles River (Wilmington, MA) were used for experiments performed as previously described.14 About 107 MDA-MB-231 cells were injected subcutaneously into the flank/back of the mice. For Figure 1c, treatments were started 15 days after the inoculation of tumor cells, to allow the formation of palpable nodules. Treated mice received 50 mg/kg of AS 3 days per week in 20 μL as direct injections into the tumors. Control mice received PBS in the same volume injections. For Figure 1d, treatments began the day after the inoculation of tumor cells and administered intraperitoneally as 50 mg/kg AS twice weekly in 20 μL for 28 days. Control mice had PBS injections. In the experiment depicted by Figure 1e, AS treatments were similar but at a lower dose of 17 mg/kg. The control group was given dAS at the same concentration. Four mice were used for each experimental condition in Figures 1c and 1d and 8 mice per group for the study depicted by Figure 1e.
Apoptosis and blood vessel density
Determination of apoptosis was carried out with the TUNEL (TdT-dUTP terminal nick-end labelling) assay as previously described.15 The cells underwent treatments with AS or dAS for 45 min and the apoptotic assay then carried out the next day. Apoptosis was assayed in tumor specimens obtained at the termination time of the respective xenograft studies, when mice were sacrificed. Blood vessel density was measured as previously described.16 Microscopic fields containing the highest density of vWf (von Willebrand factor)-positive vessels, that is, “bursts,” were identified at scanning power and then counted at 400 × magnification. The numbers for the 4 fields were averaged.
ELISA, western analysis and immunohistochemistry
Mouse blood was obtained by cardiac puncture at time of sacrifice, when xenograft experiments were terminated. Blood was allowed to clot and serum supernatant was saved and stored at −80°C. Vascular endothelial growth factor (VEGF, a major angiogenic factor) enzyme linked immunosorbent assay (ELISA) kits were purchased from R&D (Minneapolis, MN) and used according to the manufacturers' protocol.13 For TransBinding HIF-1α ELISA, cells were placed under normoxic or hypoxic conditions at 37°C for 4 hr and then treated with 0–4 mM AS or dAS for 45 min. Nuclear extractions were then collected and 10 μg samples were subjected to ELISA following the manufacturers' instructions (Panomics, Redwood City, CA). HIF-1α activity was measured by absorbance at 450 nm. The experiment was done in duplicates.
Western blotting was carried out as previously described.13 Cells were placed under hypoxic conditions at 37°C for 4 hr and then treated with 3 mM AS or dAS for 45 min. Approximately 50 μg of protein was loaded per well. Later, the membrane was incubated for 1 hr with a 1:500 dilution of anti human HIF-1α rabbit polyclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA). As control, β-catenin antibody was purchased from BD Biosciences (San Jose, CA). The same HIF-1α antibody was used for immunohistochemical studies of xenograft tumor specimens. Staining intensity was then graded from 0 to 3+.
GAPDH was extracted from tumor tissue as previously described,17 with some modifications. Tumor tissue was obtained at the termination time of the respective xenograft studies, cut (cross sections), weighed and placed in 75 mM phosphate buffer pH 7.4 at a concentration of 20 mg/ml. The specimen was then sonicated on ice for 30-sec intervals until tissue is completely homogenized. Samples were centrifuged at 13,000 rpm for 10 min; the supernatant was then removed and the tube respun as before for 8 min for additional supernatant. For the experiment, each sample tube contained 75 mM phosphate buffer pH 7.4, 100 μl of supernatant and 1.0 mM NAD+. Each reaction was started by the addition of 50 μl of G3P to a final concentration of 1.5 mM in the reaction mixture, with a final volume of 1 ml. For control, 50 μl of phosphate buffer was instead added to the same 1 ml final volume. NADH formation was monitored via absorbance at 340 nm with readings every second for 120 sec on a Shimadzu UV-2501PC (Shimadzu, Duisburg, Germany). For in vitro studies, cells were treated with 3 mM AS or dAS for 45 min, before GAPDH measurement.
Descriptive statistics, such as mean, standard error (SE) and standard deviation (SD), were used to summarize the results. The analysis of variance (ANOVA) test was performed for comparison among the various groups. The Student's t test was used for comparison between only 2 groups. Statistical significance is defined by p < 0.05.
Nitroxyl inhibits breast cancer
HNO inhibited human breast cancer cell proliferation in hypoxic culture. The proliferation results are as follows for MDA-MB-231 cells: 9,888 ± 1,799 cpm in 0.5 mM AS; 8,048 ± 2,016 cpm in 1 mM AS; and 2,523 ± 94 cpm in 3 mM AS (Fig. 1a). MCF-7 cells were similarly suppressed, in a dose dependent manner, as follows: 33,274 ± 1,303 cpm in 0.5 mM AS; 25,284 ± 925 cpm in 1 mM AS; and 9,382 ± 808 cpm in 3 mM AS (Fig. 1b). ANOVA analysis shows that the inhibitory effects are significant on both MDA-MB-231 cells (p = 0.005) and MCF-7 cells (p < 0.0001). These effects were also observed under normoxic conditions (data not shown).
We next examined the effect of HNO on the growth of breast cancer xenografts. About 107 MDA-MB-231 cells were injected subcutaneously into SCID mice and tumor size measured in 3 dimensions. As seen in Figure 1c, treatments were started on day 15 as intratumoral injections of PBS for control mice and 50 mg/kg of AS for experimental mice 3 times weekly. On day 32, the xenograft tumor sizes were as follows: 678 ± 229 mm3 in treated mice versus 1,463 ± 185 mm3 in control mice. For the experiment depicted in Figure 1d, treatments began the day after the inoculation of tumor cells and administered intraperitoneally as 50 mg/kg AS twice weekly. Control mice had PBS injections on the same schedule. Treated xenografts measured 627 ± 77 mm3, in comparison to 1,114 ± 215 mm3 in control xenografts. In another experiment, AS treatments were similar but at a lower dose of 17 mg/kg and the control group was given instead dAS at the same concentration (Fig. 1e). We observed smaller tumor size of 477 ± 119 mm3 in treated versus 694 ± 80 mm3 in control mice on day 28. The inhibitory effects were significant (p < 0.0001) by ANOVA for all of these mouse studies. There was no lung metastasis detected in any of the mice, as determined by lung weight and surface examination of the organ.
HNO increases apoptosis
The TUNEL assay was used to quantitate apoptosis in MDA-MB-231 cells (Fig. 1f). The number of apoptotic cells was counted per high power field. There was a significant increase in the level of apoptosis in cells treated with AS compared to those treated with dAS (p < 0.05) or untreated cells (p < 0.01). Similarly, we saw an increase, albeit smaller, in caspase-9, a component of the early part of the apoptosis cascade, as measured by commercially available ELISA of human breast cancer cells (data not shown). We carried out the TUNEL assay in xenograft tumors obtained from the above mouse studies. HNO increased the amount of apoptotic cells in AS treatment studies (Fig. 1g, p < 0.01).
Nitroxyl suppresses angiogenesis
The xenograft tumors were further analyzed for the extent of angiogenesis. The histological slides were reviewed by a Board-certified pathologist (J.Y.R). HNO significantly decreased the overall blood vessel density in mouse tumors (Fig. 2a, p < 0.005). The treated xenografts from 2 other mouse experiments also had significantly lower blood vessel density, in comparison to control ones (data not shown). We also observed lower levels of circulating serum VEGF (vascular endothelial growth factor, a major angiogenic factor) in the mice (Fig. 2b), although the data did not reach statistical significance. In vitro experiments with human breast cancer cell culture revealed that AS decreased the levels of VEGF (data not shown).
HNO decreases HIF-1α levels
Western analysis of multiple cell lysates demonstrated that HIF-1α level was decreased with AS treatment, compared to dAS (Fig. 2c). The cells studied included human breast cancer cancer cells MDA-MB-231 and MCF-7, human umbilical endothelial cells HUVEC and human embryonic kidney HEK293 cells. Furthermore, the inhibition was dose dependent, as seen in HIF-1α activity measured by ELISA (Fig. 2d). In hypoxic conditions, HIF-1α levels were as follows: 0.32 ± 0.08 with 4 mM AS, 0.57 ± 0.12 with 2 mM AS and 0.91 ± 0.25 with 1 mM AS, compared to 1.75 ± 0.02 without treatment. We observed HIF-1α inhibition as well under normoxia: 0.53 ± 0.02 with 4 mM AS, 0.69 ± 0.01 with 2 mM AS and 0.81 ± 0.09 with 1 mM AS, compared to 0.97 ± 0.01 without treatment. In immunohistochemical studies, the treated xenografts had lower intensity HIF-1α staining (Fig. 2e), although the data did not reach statistical significance.
Nitroxyl inhibits GAPDH
HNO inhibited GAPDH activity in vitro in both MDA-MB-231 and MCF-7 breast cancer cells (Fig. 2f). Analysis of xenograft tumors treated with AS yielded GAPDH activity of (2.61 ± 0.47) × 10−4 units, in comparison to (3.54 ± 0.45) × 10−4 units in control tumors exposed to dAS (p < 0.0001, Fig. 2g). We also observed inhibition of human GAPDH activity by HNO in cell-free chemical reactions (data not shown).
In this report, we show for the first time that HNO inhibits the proliferation of 2 commonly used human breast cancer cell lines: highly aggressive estrogen receptor (ER)-negative MDA-MB-231 and less aggressive ER-positive MCF-7. This inhibition was dose dependent and also observed under both hypoxic and normoxic conditions. We then expanded these studies to an in vivo mouse xenograft model, where we observed that mice treated with AS either injected into the tumor itself or via the intraperitoneal approach had smaller xenograft tumor size. This effect was dose dependent and remained significant even when the dosage was reduced 3-fold. These results are consistent with a previous observation that the HNO-donor AS was cytotoxic towards pheochromocytoma cells.18 As a HNO donor, AS has a limited half-life, 2.1 min at 37°C, making prolonged HNO delivery difficult and consequently increasing the effective dosage required in our study.19 Although AS has a short half life, HNO (released from AS) itself does not have an inherent half life, but its existence is entirely dependent upon the surrounding chemical environment. Multiple studies have supported the conclusion that indeed HNO is the active species when AS is used.20, 21
Cellular proliferation and apoptosis both contribute to the final size of a tumor, therefore we ascertained whether HNO induced apoptosis in addition to its inhibitory effects on cell growth. Our results showed that HNO increased the amount of apoptotic cells in vitro and in xenograft derived tumors from treated mice. It is possible that the apoptosis measured herein was a direct result of HNO inducing a cell death response, but it may be that this is a secondary effect due to additional actions of HNO such as those discussed below.
The process of angiogenesis plays a critical role in the pathogenesis of breast cancer.12 As oxidants and nitrogen oxides are thought to influence vascularization, we analyzed the effects of AS-derived HNO on angiogenesis as manifested by tumor blood vessel density in the xenograft tissues. In addition to significantly decreased blood vessel density in the AS treated tumors, we observed lower levels of circulating serum VEGF in the mice, although the latter data did not reach statistical significance.
Angiogenesis and more specifically VEGF gene expression, has been shown to be regulated by HIF-1α (hypoxia-inducible factor).22 We thus performed Western analysis of the HIF-1α protein from AS treated tumor cells and found a decrease in total HIF-1α protein. Furthermore, the inhibition was dose dependent, as seen in HIF-1α activity measured by ELISA. Currently, there are 3 HIF-1 prolyl hydroxylases (HPH1-3, respectively) and 1 HIF-1 asparaginyl hydroxylase (factor inhibiting HIF, or FIH) that have been identified thus far.23–26 These enzymes are all members of the 2-oxoglutarate-dependent family of dioxygenases and have an absolute requirement for oxygen, ferrous iron and 2-oxoglutarate (2-OG). Since prolyl hydroxylases utilize O2 as a cosubstrate for proline hydroxylation, it functions as an O2 sensor. As a result, HIF-1α transcription factor is an important player in the cellular response to hypoxia. NO has been shown to modify this class of enzymes, as well as modulate HIF-1α activity.23 The observation that HNO decreased HIF-1α protein level may suggest that there is/are thiol dependent sensor(s) either on prolyl hydroxylase or on HIF-1α itself, affecting HIF-1α protein stability. On the other hand, HNO may act by directly decreasing HIF-1α synthesis. HNO may enhance the degradation of HIF-1α through interactions with these critical thiols, by decreasing synthesis of HIF-1α protein or by acting indirectly as discussed below.
As mentioned previously, HNO is a potent inhibitor of enzymes with active site sulfhydryl groups.27 GAPDH possesses a catalytically important thiolate residue and both the yeast and rabbit enzymes have been shown to be inhibited by HNO. As GAPDH is a crucial glycolytic enzyme allowing cancers to satisfy their energetic demands, we examined its inhibition by HNO. We observed inhibition of human GAPDH activity in HNO treated human breast cancer cell lines and, more relevantly, in HNO treated tumor tissue derived from xenografts.
We think that the inhibition of GAPDH observed in our in vivo model may provide insight into the mechanism of antitumorigenicity by HNO. It has been shown that the normoxic stabilization of HIF-1α protein, which occurs in tumor cells is a result of the reversible inhibition of prolyl hydroxylases by glycolytic end products such as pyruvate.28 Cancer cells are known to utilize the glycolytic pathway even during normoxia.29 This phenomenon allows tumor cells, which have a greater degree of reliance upon glycolysis than benign cells, to maintain significant levels of HIF-1α protein in both hypoxic and normoxic microenvironments. A higher steady state level of HIF-1α protein promotes angiogenesis in part by inducing VEGF production, which recruits endothelial cells to initiate vascular vessel formation. If the burgeoning angiogenesis is disrupted, cellular energy reserves cannot keep pace with tumor growth, resulting in cytostasis and cell death. Thus, although the metabolites were not measured herein, inhibition of GAPDH in our system likely led to a decrease in available glycolytic end products pyruvate and lactate. We hypothesize that this probable decrease of stabilizing cytosolic α-oxoacids may result in lowered HIF-1α protein and activity in tumors from HNO treated mice. A consequence of this decreased HIF-1α activity would be lower VEGF production by tumor tissues and an overall decrease in angiogenesis, which were indeed observed in the current study. The increase in apoptotic cells noted in tumor tissues could have resulted from the lack of glycolytic metabolism, blood supply, or a direct action of HNO itself.30, 31
In conclusion, we report here that the nitrogen oxide HNO inhibits the proliferation of cultured breast cancer cells and decreases tumor mass in a mouse xenograft model. Additionally, HNO treatment resulted in inhibition of human GAPDH activity, decreased levels of HIF-1α protein and activity, lower VEGF production, decreased tumor angiogenesis and an increase in apoptotic cells. We propose that one explanation for the multiplicity of actions observed after HNO treatment could have resulted from the initial inhibition of GAPDH, providing a potential therapeutic avenue based upon blocking glycolysis. Further investigations into these antitumor effects as well as the synthesis of more biologically available HNO donors may lead to a better understanding of HNO's action. So far, our observations do provide intriguing possibilities regarding the cancer therapeutic potential of HNO.
- 1Nitric oxide: biology and pathobiology. San Diego: Academic Press, 2000. 1003 p..