- Email: [email protected]
Methemoglobin contributes to the growth of human tumor cells
Life Sciences 70 (2002) 907–916
Methemoglobin contributes to the growth of human tumor cells Wu-Nan Wen* Institute of Biochemistry, College of Medicine, National Taiwan University, Taipei, 100, Taiwan Received 19 February 2001; accepted 6 August 2001
Abstract Methemoglobin (metHb) has been reported to be present in areas surrounding solid tumors. The effects of human metHb on the growth of one human hepatocellular carcinoma cell line and one human glioma cell line that simply replicate in Ham’s nutrient mixture F12 (F12) were investigated. MetHb, depending on its concentration, stimulated or inhibited the in vitro growth of both cancer cell lines. The stimulatory or inhibitory effect was due to the release of hemin from metHb, which was recognized by its characteristic light absorption spectrum. The possibility of metHb or hemin acting initially through a 39, 59-cyclic guanosine monophosphate- (cGMP-) or prostaglandin E2- (PGE2-) mediated pathway to enhance cell growth was excluded. Ferric iron derived from the catabolic degradation of hemin increased cell growth, whereas biliverdin (Bv) and its reduction product, bilirubin (Br), decreased cell growth. Hemoglobin oxidized to metHb in conditions found in tumors showing neovascularization and hemorrhage may contribute significantly to increased proliferation of cancerous cells. © 2002 Elsevier Science Inc. All rights reserved. Keywords: Methemoglobin; Tumor growth; Ferric ion
Introduction Magnetic resonance imaging can often identify the presence of metHb in areas surrounding solid tumors [1–4]. Immunological surveillance may play a role in generating metHb in vivo since ferrohemoglobin in solution or contained in red blood cells can be oxidized to metHb by activated inflammatory cells [5–8]. Abnormal vasculature showing leaky walls and hemorrhage is a common feature of malignant tumors [9–11], suggesting that metHb might contribute to the growth of tumor cells. In the metHb molecule, ferriheme is not associated by a covalent bond [8, 12]. Ferriheme has been shown to undergo exchange among human hemoglobins and even between human * Corresponding author. Tel.: 886-2-2312-3456, ext. 8210; fax: 886-2-2391-5295. E-mail address: [email protected] (W.-N. Wen) 0024-3205/02/$ – see front matter © 2002 Elsevier Science Inc. All rights reserved. PII: S 0 0 2 4 - 3 2 0 5 ( 0 1 )0 1 4 6 5 -5
908
W.-N. Wen / Life Sciences 70 (2002) 907–916
hemoglobin and primate albumin [13], suggesting metHb may release hemin under physiological conditions. Hemin tested in a variety of cultured cells, including chicken embryonic fibroblasts, myoblasts and hepatocytes [14], human peripheral blood mononuclear cells [15] and promyelocytic leukemia cells (HL-60) [16], was growth-promoting. The mechanism of this promotion remains unclear in terms of our current understanding, however, all catabolic products of hemin are considered biologically significant [17]. It is difficult to to be certain of the effects of metHb on the growth of tumor cells when they are grown in media supplemented with serum, which is prepared from blood and contains, batch to batch, different amounts of hemoglobin. In the present study, two human cancer cell lines maintained simply in F12 containing no further exogenous growth additive were used to examine the effects of metHb, and particularly to determine how metHb can be growth-stimulating. The significance of the results in the context of tumor growth, vasculature, and hemorrhage is discussed. Materials and methods Reagents Medium-sized Sephadex G-25 was from Amersham Pharmacia Biotech (Taipei, Taiwan). Br and ferric nitrate nonahydrate were from Merck KGaA (Darmstadt, Germany). Bv was obtained from ICN Pharmaceuticals, Inc. (Costa Mesa, CA, USA). Millex-GV filter units were from Millipore, Co. (Bedford, MA, USA). 8-bromo-cGMP, F12, human hemoglobin and pooled plasma haptoglobin (Htb) (claimed to bind 0.5–0.9 mg hemoglobin per mg Htb), hemin, and protoporphyrin IX (ppIX) were purchased from Sigma Chemical Co. (St. Louis, MO, USA). PGE2 was from Cayman Chemical Co. (Ann Arbor, MI, USA) Cell cultures Human hepatocellular carcinoma cell line NTU-Bw, expressing transferrin and albumin, was established from a 70-year-old female Chinese patient [18]. Cell line NTU-G was derived from a clump of pituitary tumor obtained from a 44-year-old Chinese female patient undergoing an operation in National Taiwan University Hospital. NTU-G cells are glioma cells that express glial fibrillary acidic protein [19]. These two cell lines were cultured continuously in F12, which contains 0.834 mg/ml ferrous sulfate heptahydrate, but no ferric salt. It should be noted that the F12 used here for cell culture was not supplemented with any further exogenous growth ingredient such as growth factor, hormone or serum. NTU-Bw and NTU-G cells were kept in a humidified incubator containing 5% CO2 and 95% air at 37 8C. NTU-Bw and NTU-G cells were seeded in triplicate in F12 at a density of 10,000 cells/ 3.5-cm culture dish overnight before being subjected to different treatments. Water-insoluble reagents were dissolved in dimethyl sulfoxide, which was kept below 0.5% to avoid a cytotoxic effect. The addition of light sensitive compounds to culture media was performed under dim light conditions, and treated cells were cultured in the dark. Culture media were replenished every two to three days and cell cultures were terminated by trypsinization 7 to 10 days after seeding, depending on cell density. Growth was estimated by counting cells in a hemocytometer after staining cells with trypan blue. Cell yields from cultures maintained in
W.-N. Wen / Life Sciences 70 (2002) 907–916
909
seeding medium served as growth controls for computing the relative cell growth: 7 and 10 days of culture of NTU-Bw cells in F12 yielded (7.460.6)3104 and (18.562.5)3104 cells, respectively, wheras 7 and 10 days of culture of NTU-G cells in F12 yielded (6.560.4)3104 and (15.661.2)3104 cells, respectively. Data were compared using the Student t-test. Results were considered statistically significant when p , 0.05. MetHb preparation and purification 100 mg human hemoglobin was dissolved in 10 ml of 67 mM sodium phosphate buffer pH 6.7 and centrifuged at 15,000 3g for 15 min at 4 8C. The resulting supernatant was mixed with a 1.25-fold excess of potassium ferricyanide, and was allowed to stand at room temperature for 10 min for conversion of residual ferrohemoglobin to metHb. Ferrocyanide was removed by loading the reaction mixture into a Sephadex G-25 column eluted with 50 mM sodium chloride. The metHb prepared was analyzed by two-dimensional polyacrylamide gel electrophoresis and its purity was estimated to be 99.5% using the conventional nondiamine silver staining method to detect proteins in the gel (data not shown). For cell cultures, metHb in 50 mM sodium chloride was sterilized by filtration using a Millex-GV filter unit and mixed thoroughly with an equal volume of 23 concentrated F12. MetHb concentration was determined using Bradford’s method [20]. Globin preparation and purification Globin was prepared from human hemoglobin using the methyl ethyl ketone method as described [21], then purified from a Sephadex G-25 column and filtered for cell cultures as described above for metHb. Identification and quantification of hemin 15.76 mg metHb freshly isolated from a Sephadex G-25 column in 10-ml F12 was distributed to each of four 10-cm dishes. The dishes were then placed in a CO2 incubator set identically to the cell culture conditions. At different incubation times, the contents of the dishes were loaded onto a 134 cm plastic column filled with medium-sized Sephadex G-25. The column was first extensively eluted with 50 mM NaCl till no absorption was detected at 280 nm, then eluted with dimethyl sulfoxide to elute out fractions separately containing protein and the gel-bound hemin. The hemin from the column was identified by its light absorption spectrum scanned from 260 to 700 nm and the spectrum obtained was compared with the spectrum generated by authentic hemin dissolved in dimethyl sulfoxide or metHb dissolved in 50 mM NaCl. The amount of hemin released from metHb was subsequently determined from its light absorption at 403 nm using different amounts of the authentic hemin dissolved in dimethyl sulfoxide for calibration, and later normalized. Results NTU-Bw and NTU-G cells plausibly secret autocrine growth factors, and therefore divide ceaselessly in F12 supplemented with no further growth ingredient. Yet I found that cell growth was increased maximally by 3 fold (p , 0.001) when NTU-Bw and NTU-G cells were cultured in F12 containing 40 and 80 mg/ml purified human metHb, respectively, in comparison with the
910
W.-N. Wen / Life Sciences 70 (2002) 907–916
Fig. 1. Effect of human metHb (empty symbols) or globin (filled symbols) on the growth of NTU-Bw (circles) and NTU-G (rectangles) cells. All values represent the mean of three independent experiments 6 standard deviation (SD). T: SD.
Fig. 2. Effect of Htb on the human metHb-induced growth of NTU-Bw (white columns) and NTU-G (black columns) cells. The cells were cultured in F12 supplemented differently as indicated under the columns. All values represent the mean of two independent experiments 6 SD.
W.-N. Wen / Life Sciences 70 (2002) 907–916
911
cell yields from cells cultured in seeding media (Fig. 1). Fig. 1 also shows that metHb at concentrations higher than 320 mg/ml became growth-inhibiting. Globin tested at concentrations from 1.3 to 320 mg/ml, on the other hand, did not show a significant effect on cell growth (Fig. 1). MetHb as prepared for the aforementioned studies could potentially contaminated with some unidentified potent growth factor responsible for its growth-stimulating effect. However, as shown in Fig. 2, the metHb-induced growth of NTU-Bw and NTU-G cells was found to be eradicable by Htb (columns 2 vs. columns 4 to 6), a molecule that binds to hemoglobin [22], and was recoverable using higher molar ratios of metHb over Htb (columns 7 and 8). The concentrations of Htb tested were not cytotoxic (column 1 vs. column 3). These data together suggest metHb was the key molecule that enhanced cell growth. Human metHb was dissolved in F12 and left in a CO2 incubator where cells were cultured. This metHb released hemin that increased with incubation time. The hemin was isolated by Sephadex G-25 column chromatography for further characterization and quantification (Fig. 3 and inset). Authentic hemin and the free hemin isolated from Sephadex G-25 column showed an identical light absorption spectrum at the wavelengths 260 to 700 nm and a distinct absorption maximum at 403–404 nm. On the other hand, scanning of metHb revealed two noticeable absorption peaks at 405–406 and 276 nm. In a three-day period of incubation (Fig. 3, inset), 2.4% of the theoretical hemin molecules in 15.76 mg human metHb became unbound. Fig. 4 and 5 show that hemin behaved like metHb in F12: there was no effect on cell proliferation at concentrations below 12.5 ng/ml; concentration-dependent growth stimulation
Fig. 3. Characterization of hemin dissociated from human metHb. Dotted and solid lines represent light absorption spectra of authentic hemin and hemin released from metHb, respectively. Dash-dotted line: absorption spectrum of metHb. Inset: hemin isolated and quantitated at different times of incubation (one determination).
912
W.-N. Wen / Life Sciences 70 (2002) 907–916
Fig. 4. Effect of hemin (white circles), Bv (dots), PGE2 (squares) or ppIX (triangles) on the growth of NTU-Bw cells. All values represent the mean of two independent experiments 6 SD.
reached a maximum of a 3.5- and 2.8-fold increase (p , 0.001) at 50 and 100 ng/ml for NTU-Bw and NTU-G cells, respectively; and an inhibition of cell growth occurred at concentrations higher than 800 ng/ml. Fig. 4 and 5 also show that ppIX, which is structurally similar to hemin but possesses no iron, was not growth-stimulating, though it was reported to possibly activate soluble guanylate cyclase (GCase) [23, 24). Bv and Br, the two catabolic products of hemin in human cells, although they possibly function as antioxidants [25], actually inhibited cell proliferation at concentrations that remained growth-stimulatory if hemin were used (Fig. 4 and 5). Carbon monoxide, another product of hemin degradation, is currently believed to be, like nitric oxide, a signal molecule related to cGMP and PGE2 production in different biological systems [26–29]. PGE2 (Fig. 4 and 5) and 8-bromo-cGMP (Fig. 6), a membrane-permeable and nondegradable cGMP analogue, tested at a wide range of concentrations did not reveal an enhancement of cell growth. In contrast, ferric nitrate added to the culture medium maximally increased the growth of NTU-Bw and NTU-G cells by 3 fold at 0.64 mg/ml (Fig. 6). Ferric ion is also a known product of hemin degradation. Discussion Cancers are grim mostly because of uncontrolled cell growth. In this study, I searched for in vitro evidence to support the idea that hemoglobin originating from extravasated red blood
W.-N. Wen / Life Sciences 70 (2002) 907–916
913
Fig. 5. Effect of hemin (white rectangles), Br (circles), PGE2 (black rectangles) or ppIX (triangles) on the growth of NTU-G cells. All values represent the mean of two independent experiments 6 SD.
cells and subsequently oxidized to metHb, a condition found in tumors showing neovascularization and hemorrhage, contributes to increased proliferation of cancerous cells. One human hepatoma cell line, NTU-Bw, and one human glioma cell line, NTU-G, established previously in my own laboratory were particularly suitable for this investigation because they simply replicate in F12 which eliminates the complications arising from the use of serum for cell
Fig. 6. Effect of iron supplement (empty symbols) or 8-Br-cGMP (filled symbols) on the growth of NTU-Bw (ovals) and NTU-G (rectangles) cells. All values represent the mean calculated from the cell culture in triplicate 6 SD.
914
W.-N. Wen / Life Sciences 70 (2002) 907–916
cultures. In these culture systems, metHb unambiguously stimulated and inhibited the growth of these two cancer cell lines, depending upon the concentrations of metHb used for cell culture (Fig. 1). Globin has been shown to contain sequences with opioid activity [30, 31] and might therefore also contain sequences with mitogenic activity. This mitogenic activity can be realized when globin in culture medium is hydrolyzed by a specific protease, possibly produced by some transformed cells and secreted into culture medium. However, globin prepared from hemoglobin using the methyl ethyl ketone method and tested in these cancer cell cultures did not result in growth enhancement (Fig. 1). As indicated in Fig. 3, metHb released hemin under cell culture conditions. This finding simply reflects the fact that hemin and globin are not linked by a covalent bond [8, 12]. Hemin is known to be growth-stimulating for a variety of cultured cells [14–16], however, the growth-stimulating mechanism remains unclear in terms of our current knowledge that all catabolic products of heme or hemin are biologically significant [17]. Fig. 4 and 5 show that the growth-stimulating effect of hemin also applies to NTU-Bw and NTU-G cells. Htb plausibly strengthens the liganding between hemin and globin [8], thereby decreasing metHbinduced cell growth (Fig. 2). Hemin was reported to activate particulate GCase [32], and although a number of recent reports have demonstrated that carbon monoxide functions through cGMP as a signal molecule [26, 27, 33], the possibility of hemin acting initially through cGMP to stimulate the proliferation of NTU-Bw and NTU-G cells seems unlikely since: a) a membrane-permeable and nondegradable cGMP analogue, 8-Br-cGMP, did not increase cell growth (Fig. 6); and b) while ppIX possibly activates soluble GCase [23, 24], it did not stimulate cell growth (Fig. 4 and 5). Though carbon monoxide might stimulate cell proliferation through cGMP-independent signal pathways, this seems unlikely in view of the fact that ppIX also possibly converts into carbon monoxide but is not growth-stimulating (Fig. 4 and 5). Hemin and carbon monoxide were previously reported to be responsible for an increase in PGE2 production and release from hypothalamic explants [34, 35]. PGE2 in turn enhances the growth of U937 human myeloid leukemic cells [36]. The results shown in Fig. 4 and 5 indicated that PGE2 in NTU-Bw and NTU-G cells was not growth-stimulating. I therefore also dismiss the possibility that hemin acts through a PGE2-mediated pathway to enhance cell growth. Bv and Br, the two potent antioxidants derived from hemin, though they possibly allow cells to survive better under oxidative stress, they also could not be regarded as growth-stimulating metabolites (Fig. 4 and 5). In the present study, iron derived from heme degradation was the only product that was identified to be growth-stimulatory (Fig. 6). Uptake of iron in greater amounts could potentially generate more hydroxyl radicals and other reactive oxygen species that may lead to cellular damage [37], thereby inhibiting cell growth, unless other molecules such as ferritin which serves as a warehouse where irons are stored and redistributed [38, 39], are promptly induced to reduce the toxic effects of redox-active iron. The growth-inhibitory effect of metHb (Fig 1) and hemin (Fig. 4 and 5) observed in the present study may have been due to the generation of higher amounts of Bv and Br when higher concentrations of metHb and hemin were tested in cell cultures. High concentrations of Bv and Br indeed inhibited the growth of NTU-Bw and NTU-G cells (Fig. 4 and 5). However, Br, the reduction product of Bv, formed in peripheral tissues inside the human body, is
W.-N. Wen / Life Sciences 70 (2002) 907–916
915
transported to the liver for further processing and eventually excretion, indicating that Bv and Br may in fact never reach a local level toxic enough to inhibit tumor growth. This likelihood, along with consideration of the other results obtained in the present study, suggests that hemoglobin derived from extravasated red blood cells in tumors showing neovascularization and hemorrhage may significantly contribute to uncontrolled tumor growth. Acknowledgments This investigation was supported by Grant NSC-88-2314-B-002-201 from the National Science Council, Taiwan. References 1. Zimmerman R A, Bilaniuk LT, Hackney DB, Goldberg HI, Grossman RI. Magnetic resonance imaging of dural venous sinus invasion, occlusion and thrombosis. Acta Radiologica-Supplementum 1986;369:110–112. 2. Spoto GP, Press GA, Hesselink JR, Solomon M. Intracranial ependymoma and subependymoma: MR manifestations. American Journal of Neuroradiology. 1990;11:83–91. 3. Kuroiwa T, Moriwaki K, Ohta T, Tsutsumi A, Tajima S. Extradural extension of primitive neuroectodermal tumor-case report. Neurologia Medico-Chirurgica. 1994;34:379–381. 4. Savci G, Kilicturgay S, Sivri Z, Parlak M, Tuncel E. Solid and papillary epithelial neoplasm of the pancreas: CT and MR findings. European Radiology 1996;6:86–88. 5. Weiss SJ. Neutrophil-mediated methemoglobin formation in the erythrocyte. The role of superoxide and hydrogen peroxide. Journal of Biological Chemistry 1982;257:2947–2953. 6. Vercellotti GM, van Asbeck BS, Jacob HS. Oxygen radical-induced erythrocyte hemolysis by neutrophils: critical role of iron and lactoferrin. Journal of Clinical Investigation 1985;76:956–962. 7. Dallegri F, Ballestrero A, Frumento G, Patrone F. Augmentation of neutrophil-mediated erythrocyte lysis by cells derived in vitro from human monocytes. Blood 1987;70:1743–1749. 8. Balla J, Jaco HS, Balla G, Nath K, a Eaton JW. Endothelial-cell heme uptake from heme proteins: Induction of sensitization and desensitization to oxidant damage. Proceedings of the National Academy of Sciences of USA 1993;90:9285–9289. 9. Grunt TW, Lametschwandtner A, Staindl O. The vascular pattern of basal cell tumors: light microscopy and scanning electron microscopic study on vascular corrosion casts. Microvascular Research 1985;29:371–386. 10. Deewhirst MW, Tso CY, Secomb TW, Gross JF. Morphologic and hemodynamic comparison of tumor and healing normal tissue microvasculature. International Journal of Radiation Oncology Biology Physics 1989;17: 91–99. 11. Shah-Yukich AA, Nelson AC. Characterization of solid tumor microvasculature: a three-dimensional analysis using the polymer casting technique. Laboratory Investigation 1988;58:236–244. 12. Pauling L, Coryell CD. The magnetic properties and structure of hemoglobin, oxyhemoglobin and carbonmonosyhemoglobin. Proceedings of the National Academy of Sciences of USA 1936;22:210–216. 13. Bunn HF, Jandl JH. Exchange of heme among hemoglobins and between hemoglobin and albumin. Journal of Biological Chemistry 1968;243:465–475. 14. Verger C, Sassa S, Kappas A. Growth-promoting effects of iron- and cobalt-protoporphyrins on cultured embryonic cells. Journal Cellular Physiology 1983;116:135–141. 15. Stenzel KH, Rubin AL, Novogrodsky A. Mitogenic and co-mitogenic properties of hemin. Journal of Immunology 1981;127:2469–2473. 16. Palkowski A, Sikorski AF. Effect of hemin on growth and DNA synthesis of HL-60 cells. In Vitro Cellular and Developmental Biology 1993;29A:679–682. 17. Maines MD. The heme oxygenase system: a regulator of second messenger gases. Annual Review of Pharmacology and Toxicology 1997;37:517–554.
916
W.-N. Wen / Life Sciences 70 (2002) 907–916
18. Wen W, Liu C, Chang M, Huang H, Lee C. Continual growth of two newly established human hepatoma cell lines in basal media supplemented with no further exogenous growth ingredients. In: Murakami H, editor. Trends in Animal Cell Culture Technology. Tokyo: Kodansha Ltd., 1990. pp.75–80. 19. Wen W-N, Chang L. A Novel human glioma cell line from pituitaries. In Vitro Cellular and Developmental Biology 1993;29A:111–113. 20. Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Analytical Biochemistry 1976;72: 248–254. 21. Ascoli F, Fanelli MRR, Antonini E. Preparation and properties of Apohemoglobin and reconstituted hemoglobins. In: Antonini E, Rossi-Bernardi L, Chiancone E, editors. Methods in Enzymology. New York: Academic Press; 1981;76:72–85. 22. Valette I, Waks M, Wejman JC, Arcoleo JP, Greer, J. Haptoglobin heavy and light chains. Structural properties, reassembly, and formation of minicomplex with hemoglobin. Journal of Biological Chemistry 1981;256; 672–679. 23. Ignarro LJ, Wood KS, Wolin MS. Activation of purified soluble guanylate cyclase by protoporphyrin IX. Proceedings of the National Academy of Sciences of USA 1982;79:2870–2873. 24. Ignarro LJ, Ballot B, Wood KS. Regulation of soluble guanylate cyclase Activity by porphyrins and metalloporphyrins. Journal of Biological Chemistry 1984;259:6201–6207. 25. Stocker R, Yamamoto Y, McDonagh AF, Glazer AN, Ames BN. Bilirubin is an antioxidant of possible physiological importance. Science 1987;235:1043–1047. 26. Utz J, Ullrich V. Carbon monoxide relaxes ilial smooth muscle through activation of guanylate cyclase. Biochemical Pharmacology 1991;41:1195–2001. 27. Ingi T, Cheng J, Ronnett GV. Carbon monoxide: an endogenous modulator of the nitric oxide-cyclic GMP signaling system. Neuron 1996;16:835–842. 28. Mancuso C, Pistritto G, Tringali G, Grossman AB, Preziosi P, Navarra P. Evidence that carbon monoxide stimulates prostaglandin endoperoxide synthase activity in rat hypothalamic explants and in primary cultures of rat hypothalamic astrocytes. Molecular Brain Research 1997;45:294–300. 29. Mancuso C, Tringali G, Grossman A, Preziosi P, Navarra P. The generation of nitric oxide and carbon monoxide produces opposite effects on the release of immunoreactive interleukin-1b from the rat hypothalamus in vitro: evidence for the involvement of different signaling pathways. Endocrinology 1998;139:1031–1037. 30. Glamsta EL; Marklund A, Hellman U, Wernstedt C, Terenius L, Nyberg F. Isolation and characterization of a hemoglobin-derived opioid peptide from the human pituitary gland. Regulatory Peptide 1991;34:169–179. 31. Glamsta EL, Meyerson B, Silberring J, Terenius L, Nyberg F. Isolation of a hemoglobin-derived opioid peptide from cerebrospinal fluid of patients with cerebrovascular bleedings. Biochemical and Biophysical Research Communications. 1992;184:1060–1066. 32. Waldman SA, Sinacore MS, Lweicki JA, Chang LY, Murad F. Selective activation of particulate guanylate cyclase by a specific class of porphyrins. Journal of Biological Chemistry 1984;259:4038–4042. 33. Ramos KS, Lin H, McGrath JJ. Modulation of cyclic guanosine monophosphate levels in cultured aortic smooth muscle cells by carbon monoxide. Biochemical Pharmacology 1989;38:1368–1370. 34. Mancuso C, Pistritto G, Tringali G, Grossman A, Preziosi P, Navarra P. Evidence that carbon monoxide stimulates prostaglandin endoperoxide synthase activity in rat hypothalamic explants and in primary cultures of rat hypothalamic astrocytes. Molecular Brain Research 1997;45:294–300. 35. Mancuso C, Tringali G, Grossman A, Preziosi P, Navarra P. The generation of nitric oxide and carbon monoxide produces opposite effects on the release of immunoreactive interleukin-1b from the rat hypothalamus in vitro: evidence for the involvement of different signaling pathways. Endocrinology 1998;139:1031–1037. 36. Jea J-C, Liu T-C, Wang S-Y, Sung Y-J. Nitric oxide enhances the growth of U937 human leukemic cells through a cyclooxygenase-mediated pathway. Journal of Leukocyte Biology 1998;64:451–458. 37. Halliwell B, Gutteridge JMC. Role of free radicals and catalytic metal ions in human disease: an overview. In: Packer L, Glazer AN editors. Methods in Enzymology. New York: Academic Press 1990;186:1–85. 38. Zahringer J, Baliga BS, Munro HN. Novel mechanism for translational control in regulation of ferritin synthesis by iron. Proceedings of the National Academy of Sciences of USA 1976;73:857–861. 39. Aziz N, Munro HN. Both subunits of rat liver ferritin are regulated at a translational level by iron induction. Nucleic Acids Research 1986;14:915–921.
Methemoglobin contributes to the growth of human tumor cells Wu-Nan Wen* Institute of Biochemistry, College of Medicine, National Taiwan University, Taipei, 100, Taiwan Received 19 February 2001; accepted 6 August 2001
Abstract Methemoglobin (metHb) has been reported to be present in areas surrounding solid tumors. The effects of human metHb on the growth of one human hepatocellular carcinoma cell line and one human glioma cell line that simply replicate in Ham’s nutrient mixture F12 (F12) were investigated. MetHb, depending on its concentration, stimulated or inhibited the in vitro growth of both cancer cell lines. The stimulatory or inhibitory effect was due to the release of hemin from metHb, which was recognized by its characteristic light absorption spectrum. The possibility of metHb or hemin acting initially through a 39, 59-cyclic guanosine monophosphate- (cGMP-) or prostaglandin E2- (PGE2-) mediated pathway to enhance cell growth was excluded. Ferric iron derived from the catabolic degradation of hemin increased cell growth, whereas biliverdin (Bv) and its reduction product, bilirubin (Br), decreased cell growth. Hemoglobin oxidized to metHb in conditions found in tumors showing neovascularization and hemorrhage may contribute significantly to increased proliferation of cancerous cells. © 2002 Elsevier Science Inc. All rights reserved. Keywords: Methemoglobin; Tumor growth; Ferric ion
Introduction Magnetic resonance imaging can often identify the presence of metHb in areas surrounding solid tumors [1–4]. Immunological surveillance may play a role in generating metHb in vivo since ferrohemoglobin in solution or contained in red blood cells can be oxidized to metHb by activated inflammatory cells [5–8]. Abnormal vasculature showing leaky walls and hemorrhage is a common feature of malignant tumors [9–11], suggesting that metHb might contribute to the growth of tumor cells. In the metHb molecule, ferriheme is not associated by a covalent bond [8, 12]. Ferriheme has been shown to undergo exchange among human hemoglobins and even between human * Corresponding author. Tel.: 886-2-2312-3456, ext. 8210; fax: 886-2-2391-5295. E-mail address: [email protected] (W.-N. Wen) 0024-3205/02/$ – see front matter © 2002 Elsevier Science Inc. All rights reserved. PII: S 0 0 2 4 - 3 2 0 5 ( 0 1 )0 1 4 6 5 -5
908
W.-N. Wen / Life Sciences 70 (2002) 907–916
hemoglobin and primate albumin [13], suggesting metHb may release hemin under physiological conditions. Hemin tested in a variety of cultured cells, including chicken embryonic fibroblasts, myoblasts and hepatocytes [14], human peripheral blood mononuclear cells [15] and promyelocytic leukemia cells (HL-60) [16], was growth-promoting. The mechanism of this promotion remains unclear in terms of our current understanding, however, all catabolic products of hemin are considered biologically significant [17]. It is difficult to to be certain of the effects of metHb on the growth of tumor cells when they are grown in media supplemented with serum, which is prepared from blood and contains, batch to batch, different amounts of hemoglobin. In the present study, two human cancer cell lines maintained simply in F12 containing no further exogenous growth additive were used to examine the effects of metHb, and particularly to determine how metHb can be growth-stimulating. The significance of the results in the context of tumor growth, vasculature, and hemorrhage is discussed. Materials and methods Reagents Medium-sized Sephadex G-25 was from Amersham Pharmacia Biotech (Taipei, Taiwan). Br and ferric nitrate nonahydrate were from Merck KGaA (Darmstadt, Germany). Bv was obtained from ICN Pharmaceuticals, Inc. (Costa Mesa, CA, USA). Millex-GV filter units were from Millipore, Co. (Bedford, MA, USA). 8-bromo-cGMP, F12, human hemoglobin and pooled plasma haptoglobin (Htb) (claimed to bind 0.5–0.9 mg hemoglobin per mg Htb), hemin, and protoporphyrin IX (ppIX) were purchased from Sigma Chemical Co. (St. Louis, MO, USA). PGE2 was from Cayman Chemical Co. (Ann Arbor, MI, USA) Cell cultures Human hepatocellular carcinoma cell line NTU-Bw, expressing transferrin and albumin, was established from a 70-year-old female Chinese patient [18]. Cell line NTU-G was derived from a clump of pituitary tumor obtained from a 44-year-old Chinese female patient undergoing an operation in National Taiwan University Hospital. NTU-G cells are glioma cells that express glial fibrillary acidic protein [19]. These two cell lines were cultured continuously in F12, which contains 0.834 mg/ml ferrous sulfate heptahydrate, but no ferric salt. It should be noted that the F12 used here for cell culture was not supplemented with any further exogenous growth ingredient such as growth factor, hormone or serum. NTU-Bw and NTU-G cells were kept in a humidified incubator containing 5% CO2 and 95% air at 37 8C. NTU-Bw and NTU-G cells were seeded in triplicate in F12 at a density of 10,000 cells/ 3.5-cm culture dish overnight before being subjected to different treatments. Water-insoluble reagents were dissolved in dimethyl sulfoxide, which was kept below 0.5% to avoid a cytotoxic effect. The addition of light sensitive compounds to culture media was performed under dim light conditions, and treated cells were cultured in the dark. Culture media were replenished every two to three days and cell cultures were terminated by trypsinization 7 to 10 days after seeding, depending on cell density. Growth was estimated by counting cells in a hemocytometer after staining cells with trypan blue. Cell yields from cultures maintained in
W.-N. Wen / Life Sciences 70 (2002) 907–916
909
seeding medium served as growth controls for computing the relative cell growth: 7 and 10 days of culture of NTU-Bw cells in F12 yielded (7.460.6)3104 and (18.562.5)3104 cells, respectively, wheras 7 and 10 days of culture of NTU-G cells in F12 yielded (6.560.4)3104 and (15.661.2)3104 cells, respectively. Data were compared using the Student t-test. Results were considered statistically significant when p , 0.05. MetHb preparation and purification 100 mg human hemoglobin was dissolved in 10 ml of 67 mM sodium phosphate buffer pH 6.7 and centrifuged at 15,000 3g for 15 min at 4 8C. The resulting supernatant was mixed with a 1.25-fold excess of potassium ferricyanide, and was allowed to stand at room temperature for 10 min for conversion of residual ferrohemoglobin to metHb. Ferrocyanide was removed by loading the reaction mixture into a Sephadex G-25 column eluted with 50 mM sodium chloride. The metHb prepared was analyzed by two-dimensional polyacrylamide gel electrophoresis and its purity was estimated to be 99.5% using the conventional nondiamine silver staining method to detect proteins in the gel (data not shown). For cell cultures, metHb in 50 mM sodium chloride was sterilized by filtration using a Millex-GV filter unit and mixed thoroughly with an equal volume of 23 concentrated F12. MetHb concentration was determined using Bradford’s method [20]. Globin preparation and purification Globin was prepared from human hemoglobin using the methyl ethyl ketone method as described [21], then purified from a Sephadex G-25 column and filtered for cell cultures as described above for metHb. Identification and quantification of hemin 15.76 mg metHb freshly isolated from a Sephadex G-25 column in 10-ml F12 was distributed to each of four 10-cm dishes. The dishes were then placed in a CO2 incubator set identically to the cell culture conditions. At different incubation times, the contents of the dishes were loaded onto a 134 cm plastic column filled with medium-sized Sephadex G-25. The column was first extensively eluted with 50 mM NaCl till no absorption was detected at 280 nm, then eluted with dimethyl sulfoxide to elute out fractions separately containing protein and the gel-bound hemin. The hemin from the column was identified by its light absorption spectrum scanned from 260 to 700 nm and the spectrum obtained was compared with the spectrum generated by authentic hemin dissolved in dimethyl sulfoxide or metHb dissolved in 50 mM NaCl. The amount of hemin released from metHb was subsequently determined from its light absorption at 403 nm using different amounts of the authentic hemin dissolved in dimethyl sulfoxide for calibration, and later normalized. Results NTU-Bw and NTU-G cells plausibly secret autocrine growth factors, and therefore divide ceaselessly in F12 supplemented with no further growth ingredient. Yet I found that cell growth was increased maximally by 3 fold (p , 0.001) when NTU-Bw and NTU-G cells were cultured in F12 containing 40 and 80 mg/ml purified human metHb, respectively, in comparison with the
910
W.-N. Wen / Life Sciences 70 (2002) 907–916
Fig. 1. Effect of human metHb (empty symbols) or globin (filled symbols) on the growth of NTU-Bw (circles) and NTU-G (rectangles) cells. All values represent the mean of three independent experiments 6 standard deviation (SD). T: SD.
Fig. 2. Effect of Htb on the human metHb-induced growth of NTU-Bw (white columns) and NTU-G (black columns) cells. The cells were cultured in F12 supplemented differently as indicated under the columns. All values represent the mean of two independent experiments 6 SD.
W.-N. Wen / Life Sciences 70 (2002) 907–916
911
cell yields from cells cultured in seeding media (Fig. 1). Fig. 1 also shows that metHb at concentrations higher than 320 mg/ml became growth-inhibiting. Globin tested at concentrations from 1.3 to 320 mg/ml, on the other hand, did not show a significant effect on cell growth (Fig. 1). MetHb as prepared for the aforementioned studies could potentially contaminated with some unidentified potent growth factor responsible for its growth-stimulating effect. However, as shown in Fig. 2, the metHb-induced growth of NTU-Bw and NTU-G cells was found to be eradicable by Htb (columns 2 vs. columns 4 to 6), a molecule that binds to hemoglobin [22], and was recoverable using higher molar ratios of metHb over Htb (columns 7 and 8). The concentrations of Htb tested were not cytotoxic (column 1 vs. column 3). These data together suggest metHb was the key molecule that enhanced cell growth. Human metHb was dissolved in F12 and left in a CO2 incubator where cells were cultured. This metHb released hemin that increased with incubation time. The hemin was isolated by Sephadex G-25 column chromatography for further characterization and quantification (Fig. 3 and inset). Authentic hemin and the free hemin isolated from Sephadex G-25 column showed an identical light absorption spectrum at the wavelengths 260 to 700 nm and a distinct absorption maximum at 403–404 nm. On the other hand, scanning of metHb revealed two noticeable absorption peaks at 405–406 and 276 nm. In a three-day period of incubation (Fig. 3, inset), 2.4% of the theoretical hemin molecules in 15.76 mg human metHb became unbound. Fig. 4 and 5 show that hemin behaved like metHb in F12: there was no effect on cell proliferation at concentrations below 12.5 ng/ml; concentration-dependent growth stimulation
Fig. 3. Characterization of hemin dissociated from human metHb. Dotted and solid lines represent light absorption spectra of authentic hemin and hemin released from metHb, respectively. Dash-dotted line: absorption spectrum of metHb. Inset: hemin isolated and quantitated at different times of incubation (one determination).
912
W.-N. Wen / Life Sciences 70 (2002) 907–916
Fig. 4. Effect of hemin (white circles), Bv (dots), PGE2 (squares) or ppIX (triangles) on the growth of NTU-Bw cells. All values represent the mean of two independent experiments 6 SD.
reached a maximum of a 3.5- and 2.8-fold increase (p , 0.001) at 50 and 100 ng/ml for NTU-Bw and NTU-G cells, respectively; and an inhibition of cell growth occurred at concentrations higher than 800 ng/ml. Fig. 4 and 5 also show that ppIX, which is structurally similar to hemin but possesses no iron, was not growth-stimulating, though it was reported to possibly activate soluble guanylate cyclase (GCase) [23, 24). Bv and Br, the two catabolic products of hemin in human cells, although they possibly function as antioxidants [25], actually inhibited cell proliferation at concentrations that remained growth-stimulatory if hemin were used (Fig. 4 and 5). Carbon monoxide, another product of hemin degradation, is currently believed to be, like nitric oxide, a signal molecule related to cGMP and PGE2 production in different biological systems [26–29]. PGE2 (Fig. 4 and 5) and 8-bromo-cGMP (Fig. 6), a membrane-permeable and nondegradable cGMP analogue, tested at a wide range of concentrations did not reveal an enhancement of cell growth. In contrast, ferric nitrate added to the culture medium maximally increased the growth of NTU-Bw and NTU-G cells by 3 fold at 0.64 mg/ml (Fig. 6). Ferric ion is also a known product of hemin degradation. Discussion Cancers are grim mostly because of uncontrolled cell growth. In this study, I searched for in vitro evidence to support the idea that hemoglobin originating from extravasated red blood
W.-N. Wen / Life Sciences 70 (2002) 907–916
913
Fig. 5. Effect of hemin (white rectangles), Br (circles), PGE2 (black rectangles) or ppIX (triangles) on the growth of NTU-G cells. All values represent the mean of two independent experiments 6 SD.
cells and subsequently oxidized to metHb, a condition found in tumors showing neovascularization and hemorrhage, contributes to increased proliferation of cancerous cells. One human hepatoma cell line, NTU-Bw, and one human glioma cell line, NTU-G, established previously in my own laboratory were particularly suitable for this investigation because they simply replicate in F12 which eliminates the complications arising from the use of serum for cell
Fig. 6. Effect of iron supplement (empty symbols) or 8-Br-cGMP (filled symbols) on the growth of NTU-Bw (ovals) and NTU-G (rectangles) cells. All values represent the mean calculated from the cell culture in triplicate 6 SD.
914
W.-N. Wen / Life Sciences 70 (2002) 907–916
cultures. In these culture systems, metHb unambiguously stimulated and inhibited the growth of these two cancer cell lines, depending upon the concentrations of metHb used for cell culture (Fig. 1). Globin has been shown to contain sequences with opioid activity [30, 31] and might therefore also contain sequences with mitogenic activity. This mitogenic activity can be realized when globin in culture medium is hydrolyzed by a specific protease, possibly produced by some transformed cells and secreted into culture medium. However, globin prepared from hemoglobin using the methyl ethyl ketone method and tested in these cancer cell cultures did not result in growth enhancement (Fig. 1). As indicated in Fig. 3, metHb released hemin under cell culture conditions. This finding simply reflects the fact that hemin and globin are not linked by a covalent bond [8, 12]. Hemin is known to be growth-stimulating for a variety of cultured cells [14–16], however, the growth-stimulating mechanism remains unclear in terms of our current knowledge that all catabolic products of heme or hemin are biologically significant [17]. Fig. 4 and 5 show that the growth-stimulating effect of hemin also applies to NTU-Bw and NTU-G cells. Htb plausibly strengthens the liganding between hemin and globin [8], thereby decreasing metHbinduced cell growth (Fig. 2). Hemin was reported to activate particulate GCase [32], and although a number of recent reports have demonstrated that carbon monoxide functions through cGMP as a signal molecule [26, 27, 33], the possibility of hemin acting initially through cGMP to stimulate the proliferation of NTU-Bw and NTU-G cells seems unlikely since: a) a membrane-permeable and nondegradable cGMP analogue, 8-Br-cGMP, did not increase cell growth (Fig. 6); and b) while ppIX possibly activates soluble GCase [23, 24], it did not stimulate cell growth (Fig. 4 and 5). Though carbon monoxide might stimulate cell proliferation through cGMP-independent signal pathways, this seems unlikely in view of the fact that ppIX also possibly converts into carbon monoxide but is not growth-stimulating (Fig. 4 and 5). Hemin and carbon monoxide were previously reported to be responsible for an increase in PGE2 production and release from hypothalamic explants [34, 35]. PGE2 in turn enhances the growth of U937 human myeloid leukemic cells [36]. The results shown in Fig. 4 and 5 indicated that PGE2 in NTU-Bw and NTU-G cells was not growth-stimulating. I therefore also dismiss the possibility that hemin acts through a PGE2-mediated pathway to enhance cell growth. Bv and Br, the two potent antioxidants derived from hemin, though they possibly allow cells to survive better under oxidative stress, they also could not be regarded as growth-stimulating metabolites (Fig. 4 and 5). In the present study, iron derived from heme degradation was the only product that was identified to be growth-stimulatory (Fig. 6). Uptake of iron in greater amounts could potentially generate more hydroxyl radicals and other reactive oxygen species that may lead to cellular damage [37], thereby inhibiting cell growth, unless other molecules such as ferritin which serves as a warehouse where irons are stored and redistributed [38, 39], are promptly induced to reduce the toxic effects of redox-active iron. The growth-inhibitory effect of metHb (Fig 1) and hemin (Fig. 4 and 5) observed in the present study may have been due to the generation of higher amounts of Bv and Br when higher concentrations of metHb and hemin were tested in cell cultures. High concentrations of Bv and Br indeed inhibited the growth of NTU-Bw and NTU-G cells (Fig. 4 and 5). However, Br, the reduction product of Bv, formed in peripheral tissues inside the human body, is
W.-N. Wen / Life Sciences 70 (2002) 907–916
915
transported to the liver for further processing and eventually excretion, indicating that Bv and Br may in fact never reach a local level toxic enough to inhibit tumor growth. This likelihood, along with consideration of the other results obtained in the present study, suggests that hemoglobin derived from extravasated red blood cells in tumors showing neovascularization and hemorrhage may significantly contribute to uncontrolled tumor growth. Acknowledgments This investigation was supported by Grant NSC-88-2314-B-002-201 from the National Science Council, Taiwan. References 1. Zimmerman R A, Bilaniuk LT, Hackney DB, Goldberg HI, Grossman RI. Magnetic resonance imaging of dural venous sinus invasion, occlusion and thrombosis. Acta Radiologica-Supplementum 1986;369:110–112. 2. Spoto GP, Press GA, Hesselink JR, Solomon M. Intracranial ependymoma and subependymoma: MR manifestations. American Journal of Neuroradiology. 1990;11:83–91. 3. Kuroiwa T, Moriwaki K, Ohta T, Tsutsumi A, Tajima S. Extradural extension of primitive neuroectodermal tumor-case report. Neurologia Medico-Chirurgica. 1994;34:379–381. 4. Savci G, Kilicturgay S, Sivri Z, Parlak M, Tuncel E. Solid and papillary epithelial neoplasm of the pancreas: CT and MR findings. European Radiology 1996;6:86–88. 5. Weiss SJ. Neutrophil-mediated methemoglobin formation in the erythrocyte. The role of superoxide and hydrogen peroxide. Journal of Biological Chemistry 1982;257:2947–2953. 6. Vercellotti GM, van Asbeck BS, Jacob HS. Oxygen radical-induced erythrocyte hemolysis by neutrophils: critical role of iron and lactoferrin. Journal of Clinical Investigation 1985;76:956–962. 7. Dallegri F, Ballestrero A, Frumento G, Patrone F. Augmentation of neutrophil-mediated erythrocyte lysis by cells derived in vitro from human monocytes. Blood 1987;70:1743–1749. 8. Balla J, Jaco HS, Balla G, Nath K, a Eaton JW. Endothelial-cell heme uptake from heme proteins: Induction of sensitization and desensitization to oxidant damage. Proceedings of the National Academy of Sciences of USA 1993;90:9285–9289. 9. Grunt TW, Lametschwandtner A, Staindl O. The vascular pattern of basal cell tumors: light microscopy and scanning electron microscopic study on vascular corrosion casts. Microvascular Research 1985;29:371–386. 10. Deewhirst MW, Tso CY, Secomb TW, Gross JF. Morphologic and hemodynamic comparison of tumor and healing normal tissue microvasculature. International Journal of Radiation Oncology Biology Physics 1989;17: 91–99. 11. Shah-Yukich AA, Nelson AC. Characterization of solid tumor microvasculature: a three-dimensional analysis using the polymer casting technique. Laboratory Investigation 1988;58:236–244. 12. Pauling L, Coryell CD. The magnetic properties and structure of hemoglobin, oxyhemoglobin and carbonmonosyhemoglobin. Proceedings of the National Academy of Sciences of USA 1936;22:210–216. 13. Bunn HF, Jandl JH. Exchange of heme among hemoglobins and between hemoglobin and albumin. Journal of Biological Chemistry 1968;243:465–475. 14. Verger C, Sassa S, Kappas A. Growth-promoting effects of iron- and cobalt-protoporphyrins on cultured embryonic cells. Journal Cellular Physiology 1983;116:135–141. 15. Stenzel KH, Rubin AL, Novogrodsky A. Mitogenic and co-mitogenic properties of hemin. Journal of Immunology 1981;127:2469–2473. 16. Palkowski A, Sikorski AF. Effect of hemin on growth and DNA synthesis of HL-60 cells. In Vitro Cellular and Developmental Biology 1993;29A:679–682. 17. Maines MD. The heme oxygenase system: a regulator of second messenger gases. Annual Review of Pharmacology and Toxicology 1997;37:517–554.
916
W.-N. Wen / Life Sciences 70 (2002) 907–916
18. Wen W, Liu C, Chang M, Huang H, Lee C. Continual growth of two newly established human hepatoma cell lines in basal media supplemented with no further exogenous growth ingredients. In: Murakami H, editor. Trends in Animal Cell Culture Technology. Tokyo: Kodansha Ltd., 1990. pp.75–80. 19. Wen W-N, Chang L. A Novel human glioma cell line from pituitaries. In Vitro Cellular and Developmental Biology 1993;29A:111–113. 20. Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Analytical Biochemistry 1976;72: 248–254. 21. Ascoli F, Fanelli MRR, Antonini E. Preparation and properties of Apohemoglobin and reconstituted hemoglobins. In: Antonini E, Rossi-Bernardi L, Chiancone E, editors. Methods in Enzymology. New York: Academic Press; 1981;76:72–85. 22. Valette I, Waks M, Wejman JC, Arcoleo JP, Greer, J. Haptoglobin heavy and light chains. Structural properties, reassembly, and formation of minicomplex with hemoglobin. Journal of Biological Chemistry 1981;256; 672–679. 23. Ignarro LJ, Wood KS, Wolin MS. Activation of purified soluble guanylate cyclase by protoporphyrin IX. Proceedings of the National Academy of Sciences of USA 1982;79:2870–2873. 24. Ignarro LJ, Ballot B, Wood KS. Regulation of soluble guanylate cyclase Activity by porphyrins and metalloporphyrins. Journal of Biological Chemistry 1984;259:6201–6207. 25. Stocker R, Yamamoto Y, McDonagh AF, Glazer AN, Ames BN. Bilirubin is an antioxidant of possible physiological importance. Science 1987;235:1043–1047. 26. Utz J, Ullrich V. Carbon monoxide relaxes ilial smooth muscle through activation of guanylate cyclase. Biochemical Pharmacology 1991;41:1195–2001. 27. Ingi T, Cheng J, Ronnett GV. Carbon monoxide: an endogenous modulator of the nitric oxide-cyclic GMP signaling system. Neuron 1996;16:835–842. 28. Mancuso C, Pistritto G, Tringali G, Grossman AB, Preziosi P, Navarra P. Evidence that carbon monoxide stimulates prostaglandin endoperoxide synthase activity in rat hypothalamic explants and in primary cultures of rat hypothalamic astrocytes. Molecular Brain Research 1997;45:294–300. 29. Mancuso C, Tringali G, Grossman A, Preziosi P, Navarra P. The generation of nitric oxide and carbon monoxide produces opposite effects on the release of immunoreactive interleukin-1b from the rat hypothalamus in vitro: evidence for the involvement of different signaling pathways. Endocrinology 1998;139:1031–1037. 30. Glamsta EL; Marklund A, Hellman U, Wernstedt C, Terenius L, Nyberg F. Isolation and characterization of a hemoglobin-derived opioid peptide from the human pituitary gland. Regulatory Peptide 1991;34:169–179. 31. Glamsta EL, Meyerson B, Silberring J, Terenius L, Nyberg F. Isolation of a hemoglobin-derived opioid peptide from cerebrospinal fluid of patients with cerebrovascular bleedings. Biochemical and Biophysical Research Communications. 1992;184:1060–1066. 32. Waldman SA, Sinacore MS, Lweicki JA, Chang LY, Murad F. Selective activation of particulate guanylate cyclase by a specific class of porphyrins. Journal of Biological Chemistry 1984;259:4038–4042. 33. Ramos KS, Lin H, McGrath JJ. Modulation of cyclic guanosine monophosphate levels in cultured aortic smooth muscle cells by carbon monoxide. Biochemical Pharmacology 1989;38:1368–1370. 34. Mancuso C, Pistritto G, Tringali G, Grossman A, Preziosi P, Navarra P. Evidence that carbon monoxide stimulates prostaglandin endoperoxide synthase activity in rat hypothalamic explants and in primary cultures of rat hypothalamic astrocytes. Molecular Brain Research 1997;45:294–300. 35. Mancuso C, Tringali G, Grossman A, Preziosi P, Navarra P. The generation of nitric oxide and carbon monoxide produces opposite effects on the release of immunoreactive interleukin-1b from the rat hypothalamus in vitro: evidence for the involvement of different signaling pathways. Endocrinology 1998;139:1031–1037. 36. Jea J-C, Liu T-C, Wang S-Y, Sung Y-J. Nitric oxide enhances the growth of U937 human leukemic cells through a cyclooxygenase-mediated pathway. Journal of Leukocyte Biology 1998;64:451–458. 37. Halliwell B, Gutteridge JMC. Role of free radicals and catalytic metal ions in human disease: an overview. In: Packer L, Glazer AN editors. Methods in Enzymology. New York: Academic Press 1990;186:1–85. 38. Zahringer J, Baliga BS, Munro HN. Novel mechanism for translational control in regulation of ferritin synthesis by iron. Proceedings of the National Academy of Sciences of USA 1976;73:857–861. 39. Aziz N, Munro HN. Both subunits of rat liver ferritin are regulated at a translational level by iron induction. Nucleic Acids Research 1986;14:915–921.