| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* INSERM U 643, Insitut de Transplantation et de Recherche en Transplantation (ITERT) Nantes, France;
Laboratorio de Oncologia Basica y Biologia Molecular, Departamento de Bioquimica, Facultad de Medicina, Montevideo, Uruguay;
Institut for Medical Immunology, Charité Hospital. Berlin, Germany;
Department of Anatomy, Uniformed Services University of the Health Sciences, Bethesda, USA; and
|| INSERM U 408, Faculté de Médecine Xavier Bichat, Paris, France
1Correspondence: INSERM U 643, ITERT, 30 Bv. Jean Monnet, 44093, Nantes, cedex 1, France. E-mail: ianegon{at}nantes.inserm.fr
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: HO-1 • IDO • breast cancer • apoptosis • proliferation
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. This "asocial" cellular behavior is the consequence of a multistep process initiated by nonlethal genetic lesions that subsequently lead to a deregulation of proliferation together with an acquired resistance to apoptosis (2)
. Altered regulation of molecular pathways leading to cellular apoptosis also partly accounts for the resistance of human cancer to several chemotherapeutics agents (3)
and immunotherapy protocols (4)
. Persistent oxidative stress within carcinoma cells may be responsible for activation of growth-promoting signaling pathways and increased resistance to apoptosis (5)
. Mammary carcinogens have been shown to produce reactive oxygen species (ROS), thereby stimulating proliferation of breast epithelial cells (6)
. Furthermore, an endogenous imbalance in antioxidant enzymes is thought to be an important event in mammary carcinogenesis (7)
, and antioxidant molecules have been shown to have anti-proliferative and apoptotic effects on human breast cancer cells (8)
. Hence, understanding the molecular mechanisms leading to sublethal oxidative stress in breast cancer will not only improve current knowledge of tumor biology and carcinogenesis but will also help in the development of therapeutic strategies.
Heme oxygenase (HO) (EC 1.14.99.3), the rate limiting enzyme of heme catabolism, is a powerful antioxidant system (9
, 10)
. Several biological properties of the inducible HO isoform HO-1 have been attributed to its antioxidant effects (11)
, including the inhibition of vascular (12)
and airway smooth muscle cell (13)
proliferation. In cancer, HO-1 has been described as a protumoral molecule because of its anti-apoptotic effects in colon cancer (14)
and hepatoma (15)
in murine models and its proangiogeneic effects in human pancreatic cancer (16)
. In contrast, in human tongue cancer, low HO-1 expression has been associated with an increased risk of developing lymph node metastasis (17)
. Hence, the role of HO-1 in cancer biology is far from being completely understood. Moreover, so far no studies have been devoted to ascertaining the function of HO-1 in breast cancer.
Indoleamine 2,3-dioxygenase (IDO) (EC 1.13.11.42), which contains heme as its sole prosthetic group, catabolizes the essential amino acid tryptophan through the kynurenine pathway (18)
. IDO-dependent tryptophan consumption by malignant cells has been classically described to inhibit their growth (19)
. However, it has recently been suggested that IDO activity could constitute a mechanism by which tumor cells resist immune action and that in vivo inhibition of this enzyme could be a promising anti-tumor adjuvant therapy (20)
.
Here we studied the role of HO-1 and IDO in breast cancer biology using rat and human cell lines. Thus, we show that HO-1 strongly inhibits breast cancer cells proliferation. If HO-1 is expressed, IDO promotes proliferation by inhibiting HO-1; if HO-1 is not expressed, then IDO inhibits proliferation. In vivo expression of these enzymes during N-nitrosomethylurea (NMU) -induced carcinogenesis correlates with our in vitro results. Finally, we describe for the first time, a cross-talk between these enzymes in which HO-1 activity inhibits IDO and vice versa.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Mammary carcinogenesis
Rat breast adenocarcinomas were induced with NMU as described elsewhere (21)
.
Antibodies and cell lines
A rabbit antiserum was raised against a mixture of three rat IDO peptides: MPHSQISPAEGSRRILEEY (1-19), LFSFPGGDCDKGFFLVSLMVE (155-176), and VKPSKQKPMGGHKSEEPS (359-376) coupled to keyhole limpet hemocyanin (Neosystem, Strasbourg, France). Purified anti-IDO antibodies reacted with 293 cells transfected with IDO cDNA but not with LacZ-transfected cells, as evaluated by immunohistology and Western blot analysis (data not shown). The anti-rat HO-1 polyclonal antibody was from Stressgen (San Diego, CA, USA). Anti-quinolinate rabbit antibody was described elsewhere (22)
. 24G7 anti-bilirubin monoclonal antibody (23)
was kindly provided by M. Suematsu (Keio University, Kyoto, Japan). Anti-FLAG was from Sigma and anti-tubulin from Calbiochem (San Diego, CA, USA).
The NMU breast cancer cell line, derived from a tumor induced in a rat treated with NMU (24)
, was obtained from the ATCC and grown in RPMI medium supplemented with 10% FCS, 2 mM L-glutamine, 1 mM sodium pyruvate, 1% HEPES buffer, 100 U/mL penicillin, and 0.1 mg/mL streptomycin. 13762 (kindly provided by Jacques Le Pendu, Nantes University, France) and T47D cell lines were culture in identical conditions. MCF-7 were cultured in DMEM complemented as RPMI. Human cell lines were provided by Dominique Heymann from Nantes University.
Immunohistochemical staining
Immediately after animals were killed, mammary tissues were removed, fixed in neutral-buffered formalin, and embedded in paraffin. Tissue sections (5 µm) were applied onto silane-treated slides and used for histopathological diagnosis (stained with hematoxylin-eosin) and IDO and HO-1 evaluation. For immunohistochemical assays, samples were rehydrated and epitopes were unmasked by heating at 60°C in 10 mM sodium citrate pH 6.0, for 60 min. After three washes in PBS, quenching of endogenous peroxidase activity was performed with 3% H2O2 in PBS for 20 min, then anti-IDO antisera (1 µg/mL) or anti-HO-1 antisera (1/1000 dilution) was incubated overnight at 4°C. Primary antibody binding was determined using anti-rabbit immunoglobulin Vectastain Elite ABC kit (Vector, Burlingame, CA, USA), according to manufacturer’s instructions. The reaction was revealed with DAB and slides were counterstained with Mayer’s hematoxylin and mounted. The degree of staining of each tumor was categorized as follows: 0, for negative samples or stained < 5% of tumor extension; 1, for stained samples between 5 and 39%; 2, for stained samples between 40 and 79%, and 3, for tumors with more than 80% of stained cells. The signal intensity was assigned as strong (3)
, moderate (2)
, weak (1)
, and negative (0).
Cytospins
Suspended cells were spun onto glass slides (Shandon, Runcorn, UK). Fixation was done with acetone during 10 min. Immunocytology was performed as for immunohistology, except that revelation was performed with VIP substrate (Vector Laboratories). When using anti-quinolinate antibody, cells were fixed with carbodiimide (22)
and revelation was performed with ABC kit from Vector Laboratories.
Western blot
Total cellular lysates were used in Western blot experiments. After blocking the nitrocellulose membrane for 1 h at 37°C with TTBS (Tween 20-Tris buffered saline) containing 3% BSA, the primary anti-IDO (1 µg/mL) or anti-HO-1 (1:1000) or anti-tubulin (1 µg/mL) or anti-FLAG (5 µg/mL) antibodies or normal rabbit serum at the same dilution were added in TTBS containing 1% BSA. After overnight incubation at 4°C, the membrane was washed in TTBS and incubated for 1 h at room temperature with HRP-conjugated goat anti-rabbit IgG (Pierce, Rockford, IL, USA) diluted 1:3000 in washing buffer. After additional washes, bands were detected by enhanced chemiluminescence (Amersham, Little Chalfont, UK).
IDO and HO-1 enzymatic activity
The tryptophan-degrading activity of NMU cells reflects a multifactored combination of IDO expression and intracellular conditions that post-translationally affect enzyme activity (25)
. Therefore, we measured the rate of disappearance of tryptophan from culture supernatants over time. Tryptophan was assayed using the method of Bloxam and Warren (26)
. Proteins were precipitated with 10% TCA and free tryptophan assayed after conversion to norharman using formaldehyde and FeCl3. The reaction product was measured spectrofluorometrically (excitation 373 nm, emission 452 nm) and compared against a standard preparation of tryptophan. Although kynurenine is a highly specific marker of IDO activity, we did not use it as a quantitative measure because its accumulation is influenced by the degree of conversion to subsequent downstream metabolites (27)
. Where quantitative data were required, the tryptophan depletion assay described above was used, knowing that NMU cells produce kynurenine (not shown) and quinolinate, another metabolite of tryptophan through the IDO pathway.
HO activity was assessed biochemically by quantifying bilirubin production as already described (13)
.
Lentiviral vector production
A four-plasmid expression system was used to generate third generation lentiviral vectors by transient transfection. The four plasmids were 1) the vector plasmid (HIV-HO-1) in which the human HO-1 cDNA fused to a Flag sequence (28)
was cloned in the pRRLSIN.cPPT.PGK.WPRE plasmid (29)
. A similar HIV vector encoding GFP (HIV-GFP) was used as a control; 2) the packaging plasmid encoding Gag and Pol; 3) the Rev protein expressor; and 4) the envelope plasmid, VSV-G, for pseudotyping the virion. The vectors were produced by transfection of plasmid DNA into 293T cells using a calcium phosphate method. Transfections were done in 150 mm dishes using 15 µg vector, 12.5 µg of the packaging plasmid, 4 µg of Rev expressor plasmid, and 4 µg of the VSV-G expressor plasmid using calcium phosphate. The medium was changed 7 h later and 48 h after the start of the transfection, the medium was removed and filtered. Virion stocks were quantified by determination of p24 using an ELISA (Perkin-Elmer) and the moi by transducing 293 cells and by FACS analysis.
Cell transduction was performed by incubating NMU cells with HIV-GFP at an moi of 5 (pilot experiments showed that this moi allowed >90% NMU cell transduction) or equivalent amount of p24 for HIV-HO-1 in the presence of 8 mg/mL of polybrene for 48 h.
SiRNA synthesis and transfection
siRNA against IDO and nonspecific was synthesized by Qiagen (Hilden, Germany). They were transfected to NMU cells by electroporation with Amaxa device (Cologne, Germany).
Cell proliferation, apoptosis, and cell cycle studies
Cells were cultured in 96-well plates by sextuplicate. One µCi/mL of [methyl-3H] thymidine was added for the final 8 h of culture and DNA synthesis was measured by scintillation counting. For cell cycle analysis, NMU cells were cultured in 10 cm diameter Petri dishes until confluence. Thereafter, cells were trypsinized and apoptosis was evaluated by analyzing annexin V staining (Immunotech, Paris, France) using a FACScalibur cytometer (Becton Dickinson, San Jose, CA, USA). Alternatively, cells were ethanol-fixed for 45 min at –20°C and treated with 100 µg/mL RNase (DNase-free) and 10 µg/mL PI. DNA content was then analyzed by cytofluorimetry. Caspase activation was quantified by flow cytometry using Apostat reagent (R&D Systems, Lille, France) according to the manufacturer’s instructions.
Measurement of NMU cell oxidation states
NMU cells were stimulated with different treatments for 45 min, and an oxidation-sensitive dye (5-(and 6-)-chloromethyl-2', 7'-dichlorodihydrofluorescein diacetate, acetyl ester (CM-H2DCFDA) (Molecular Probes, Montluçon, France) was added at 2 µM for the last 15 min of incubation. Samples were washed and fluorescence analyzed by cytofluorimetry.
Statistical analysis
Results were expressed as mean ±SD. Statistical significance was evaluated using a 1-way ANOVA test. P < 0.05 was considered significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A, a), whereas no staining was observed with rabbit serum (Fig. 1A, b
). Moreover, specific staining of NMU cells was observed with the anti-bilirubin antibody (Fig. 1A, c
), demonstrating that HO-1 was functionally active. Rat NMU and 13762 cells were positive for HO-1 as analyzed by Western blot (Fig. 1B
). In contrast, human cell lines did not show HO-1 reactivity (Fig. 1B
). As we will discuss below, the absence of HO-1 expression in human cell lines may explain differential biological consequences upon IDO inhibition.
|
IDO expression by rat and human breast cancer cell lines
In Fig. 2
we show that NMU cell line expressed IDO at the protein level. Immunocytology experiments showed that 80% of the cells were IDO-positive, with a cytoplasmic staining pattern (Fig. 2A, a
), whereas coincubation of primary antibody with immunizing peptides abolished immunostaining (Fig. 2A, b
). Moreover, a single 42 kDa band was revealed by Western blot analysis (Fig. 2A, c
). The IDO band recognized in NMU cell lysates corresponds to a protein of smaller size than that obtained in rat IDO cDNA-transfected 293 cells. However, coincubation of primary antibody with immunizing peptides abolished this signal, suggesting that the only band detected in NMU cells did indeed correspond to IDO. The smaller size of the IDO protein in NMU cells may be due to alternative splicing or less glycosylation. Nevertheless, the IDO detected in NMU cells showed functional activity as measured by tryptophan consumption in culture media (Fig. 2A, d
). Quinolinate, a tryptophan metabolite produced through the kynurenine pathway, was detected in NMU cells (Fig. 2A, e
). Furthermore, kynurenine production by cell lysates was also detected (data not shown). Taken together, our results show that the NMU cell line expresses functional IDO. Thus, the NMU cell line constitutes a valuable tool to study an eventual role of IDO and HO-1 in determining a malignant phenotype. Enzymatic interactions at the intracellular and microenvironmental level are also possible in this cell line.
|
13762 cells also express IDO, as well as MCF-7 and T47D cell lines (Fig. 2B
), showing that rat and human breast cancer share this phenotypic feature.
HO-1 inhibits rat and human breast cancer cells growth
To evaluate the functional aspects regarding HO-1 expression in breast cancer cells, we first studied its effect on NMU cell growth after pharmacological modulation with protoporphyrins. SnPP is a known inhibitor while CoPP and heme are potent inducers of this enzyme (30
31
32
33)
. [3
H]-Thymidine incorporation was measured after 24 h of treatment (Fig. 3
A). HO-1 inhibition with SnPP at 10 µM led to a small but significant (P<0.05) increase in proliferation; no differences in proliferation were detected at other doses tested. Notably, HO-1 induction with CoPP drastically inhibited thymidine uptake in a dose-dependent manner. A less dramatic but still significant effect was obtained with heme. We next studied the cell cycle by analyzing PI-labeled DNA (Fig. 3B
). SnPP treatment had no effect as compared with NaOH vehicle. In contrast, heme treatment induced an increase in the number of cells in the G0/G1 phase, thus decreasing the percentage of cells in the S + G2/M phase. CoPP treatment had the same effect on the G0/G1 vs. S + G2/M relationship as heme, and CoPP-induced apoptosis could also be detected by this method. Biochemical quantification of HO-1 enzymatic activity confirmed that CoPP and heme increased HO-1 activity, whereas SnPP inhibited it (Fig. 3C
). CoPP-induced HO-1 activity (33)
was quite variable in different experiments but always significantly different from NaOH vehicle, which also explains different degrees of proliferation inhibition (Fig. 3A
vs. 6G
).
|
|
To confirm specificity of protoporphyrins treatment we transduced cells with lentiviral vector coding for HO-1 and studied cell proliferation in NMU cells (Fig. 3D
) but also in rat 13762, and human MCF-7 and T47D breast cancer cells (Fig. 3E
). Again, HO-1 overexpression led to a dose-dependent inhibition of cancer cells growth as compared with control virus.
Several biological properties of HO-1 have been attributed to antioxidant effects (11)
. Moreover, antioxidants have been shown to regulate human breast cancer cell growth (8)
. Thus, the role of antioxidants in NMU growth was investigated as a plausible mechanism by which HO-1 could modulate NMU cell proliferation. Hence, HIV-HO-1 infection of NMU cells inhibited the oxidation of CM-H2DCFDA probe as compared with control virus, showing antioxidant properties of HO-1 in our system (Fig. 3F
). A dose-dependent inhibition of [3
H]-thymidine incorporation was observed after 24 h treatment with the known antioxidants NAC, catalase and the HO-1 product bilirubin (Fig. 3G
).
We studied apoptosis induction by quantifying caspases activation by flow cytometry (Fig. 4
). HIV-HO-1induced caspases activation in NMU cells as well as in 13762 and MCF-7 cells.
|
Thus, HO-1 inhibits proliferation and induces apoptosis in rat and human breast cancer cells. Antioxidant effect of bilirubin is at least one mechanism by which cell growth is inhibited upon HO-1 overexpression.
IDO inhibition prevents rat breast cancer cells growth but increases that of human cell lines
To study the role of IDO in the growth of the NMU cell line, cell culture was performed in the presence of the IDO specific inhibitor 1-MT (34)
at usual concentrations (35)
. This treatment inhibited NMU [3
H]-thymidine incorporation by 40% in a dose-dependent manner (Fig. 5
A). Cotreatment with 1-MT and different kynurenine pathway metabolites (kynurenine, 3-OH kynurenine, or 3-OH anthranilic acid) at 100 µM was unable to restore NMU growth (not shown), suggesting these molecules do not play a role in determining cell growth, at least when acting separately. Rat 13762 cells also showed a significant but less profound inhibition of proliferation in the presence of 1-MT (25%). In contrast, human cell lines increased proliferation when IDO was inhibited (Fig. 5A
). As we will discuss below, differential expression of HO-1 between rat and human cells may explain different proliferation responses upon IDO inhibition. To discard a simply toxic effect of 1-MT on NMU cells, we down-regulated IDO protein with a specific siRNA (Fig. 5B
). More than 50% of IDO protein was diminished upon this treatment. NMU cells proliferation was inhibited by 40% when IDO was inhibited by this approach.
|
We next analyzed cell cycle and apoptosis induction after 1-MT treatment in NMU cells. This treatment induced apoptosis, as assessed by PI DNA staining (Fig. 5C
), annexin V binding (Fig. 5D
), and caspase activation (Fig. 5E
). Furthermore, the G0G1/SG2M ratio was higher in 1-MT-treated cells than in controls, indicating that IDO inhibition also affected cell cycle progression. Thus, since IDO inhibition impairs NMU cell growth through apoptosis induction and cell cycle arrest, IDO favors NMU growth. As we will show later, the tumor growth-promoting effect of IDO could be explained by HO-1 inhibition.
HO-1 and IDO are cross-regulated in NMU cells
Given that HO-1 activity degrades heme and heme is an indispensable prosthetic group for IDO enzymatic activity (18)
, we speculated that HO-1-provoked heme depletion could inhibit IDO activity. After a 24 h cell treatment with protoporphyrins to 100% confluent cells, we quantified tryptophan consumption in culture media as an indicator of IDO activity.
HO-1 inhibition with SnPP significantly increased tryptophan consumption (Fig. 6
A). In sharp contrast, treatment with 50 µM CoPP drastically inhibited IDO activity by 75% (Fig. 6A
). These observations could not be explained by treatment-induced cytotoxicity, because at 50 µM CoPP-induced apoptosis was 
30% (Fig. 4B
). In agreement, HIV-HO-1 infected cells also diminished tryptophan consumption compared with control virus (data not shown). Treatment with 50 µM heme did not affect tryptophan consumption. Furthermore, neither bilirubin nor NAC changed this parameter. The results obtained with these stimuli suggest that HO-1-dependent IDO inhibition was not mediated by antioxidant mechanisms. Since IDO protein levels were unaffected by protoporphyrin treatment, as determined by Western blot analysis (Fig. 6B
), we concluded that protoporphyrins-induced IDO modulation acts at a post-translational level. In agreement with this, CoPP-induced IDO inhibition was reversed by coincubation with 5 µM heme (Fig. 6C
). This suggests that heme depletion is at least partly responsible for HO-1-induced IDO inhibition.
We also evaluated the impact of IDO activity on HO-1 expression. To study this, NMU cells were treated with 1-MT or IDO siRNA. HO-1 protein expression analysis by Western blot showed that IDO inhibition increased HO-1 protein levels (Fig. 6D
). Furthermore, HO-1 activity was increased upon diminishing IDO protein levels with a specific siRNA (Fig. 6E
). In keeping with this, the inhibition of proliferation by 1-MT was completely reversed by cotreating NMU cells with the HO-1 inhibitor SnPP (Fig. 6F
). We therefore concluded that IDO inhibition diminishes NMU cell proliferation by increasing HO-1 activity. 1-MT treatment did not increase HO-1 protein levels in human cell lines (not shown). Thus, this observation may explain differential responses in terms of proliferation upon IDO inhibition between human and rat cell lines. Taken together, the results shown in Fig. 6D, E, F
suggest that IDO inhibits HO-1 expression and activity. Together, these results show that IDO and HO-1 are cross-inhibited in NMU cells. Heme consumption seems to be a possible mechanism of HO-1-induced IDO post-translational inhibition, but the mechanisms leading to IDO-induced HO-1 inhibition remain to be determined.
HO-1-induced NMU growth inhibition is partly mediated by IDO inhibition through heme consumption
We next set out to understand the functional consequences of this cross-regulation. Given our findings that IDO seemed to stimulate NMU cells proliferation (Fig. 5)
and that IDO activity could be inhibited through HO-1-induced heme consumption, we speculated that this mechanism could also be responsible for HO-1-dependent inhibition of NMU cell line growth. We thus studied proliferation of cells treated with CoPP and heme. At the concentrations used, CoPP induced 50% growth inhibition, whereas heme had no effect (Fig. 6G
). Cotreatment with these protoporphyrins partially abolished CoPP-induced growth inhibition, suggesting that HO-1-provoked heme consumption prevented NMU growth (Fig. 6G
). As shown in Fig. 6C
, addition of heme to CoPP-treated cells abolished CoPP-induced IDO inhibition, suggesting that recovery of proliferation after this double treatment was due to an increase in IDO activity. To further determine the role of IDO in this effect, 1-MT was added to cultures treated with CoPP and heme. As expected, in the presence of 1-MT the recovery of proliferation observed with heme in CoPP-treated cells was lost (Fig. 6H
). Thus, one mechanism by which HO-1 induction decreases NMU cell growth is through IDO inhibition due to heme consumption.
HO-1and IDO expression during NMU-induced mammary carcinogenesis
To analyze the in vivo relevance of the mechanisms analyzed here, we next studied HO-1 and IDO expression by immunohistology in five mammary glands from normal untreated rats and in three NMU-induced breast carcinomas. Semiquantification staining is shown in Table 1
and representative tissue sections are shown in Fig. 7
. We found mammary glands from normal rats to be positive for HO-1. Most epithelial cells (>95%) were strongly stained with the anti-HO-1 specific antibody (Table 1
and Fig. 7A, C
). In contrast, in NMU-induced breast carcinomas there was a clear down-regulation of HO-1 expression, as few epithelial cells (<5%) were weakly positive (Fig. 7G, I
). Some stromal cells, which were morphologically similar to macrophages, were positive in tumor samples. A benign adenoma showed no reactivity with the anti-HO-1 antibody (Fig. 7E
), suggesting that HO-1 is down-regulated since early stages in rat mammary carcinogenesis.
|
|
Next we analyzed IDO expression in the same tissues. In normal breast epithelium from untreated rats, IDO protein had a staining pattern similar to that of HO-1, with most (>95%) epithelial cells being moderately positive (Table 1
and Fig. 7B, D
). NMU-induced breast carcinomas showed a less extensive staining area than normal mammary gland, as 30–40% of cancer cells were positive, although intensity did not diminish (Fig. 7H, J
).
In summary, HO-1 and IDO expression during NMU-induced rat breast carcinogenesis are down-regulated during rat mammary malignant transformation. Thus, in normal mammary gland both enzymes are expressed in most epithelial cells, whereas in malignant tissue HO-1 is negative and one-third of cancer cells maintain IDO expression.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The role of HO-1 in cancer biology is only just beginning to be studied in detail. Here we show that protoporphyrin-induced HO-1 and HO-1 gene transfer inhibited rat and human breast cancer cell proliferation (Fig. 3)
. This was shown to be the consequence of cell cycle arrest and apoptosis (Figs. 3
, 4)
. Most papers have reported a protumoral effect of this enzyme through anti-apoptotic (14
, 15
, 36
, 37)
or proangiogenic functions (16)
. However, low levels of HO-1 expression have been associated with lymph node metastasis and unfavorable histological differentiation in tongue cancer (17)
. The present study is the first description of HO-1 as an anti-tumor molecule acting through anti-proliferative and proapoptotic mechanisms. Although unexpected, we believe our results to be logical. First of all, the role of HO-1 had never been studied in breast cancer. It is well known that molecules such as iNOS have opposing roles regarding apoptosis in cancer cells depending on the molecular context (38)
. Moreover, HO-1 has anti-proliferative (13
, 39)
and proapoptotic (40
, 41)
functions in other cell types such as vascular (39)
and airway (13)
smooth muscle as well as Leydig cells (40)
. Many of the effects of HO-1 are mediated through antioxidant mechanisms (11)
. Thus, although HO-1 has been described as an anti-apoptotic molecule in cancer cells through its induction of p21 (36
, 42)
, p21 can also act as a proapoptotic molecule when induced by antioxidants in human breast cancer cells (43)
. Here we show that different antioxidant compounds inhibit NMU cell line proliferation, including the HO-1 end product bilirubin (Fig. 3F
). This is a plausible pathway by which HO-1 inhibits NMU cells growth. Thus, it is feasible that HO-1 has a direct effect on cell growth that is independent of IDO, since bilirubin and NAC affected NMU proliferation without inhibiting IDO (Fig. 6A, B
). However, antioxidant-independent mechanisms of cell growth modulation could also be operative. Accordingly, as mentioned above, HO-1-induced IDO inhibition is able to modulate cell proliferation independently of antioxidant mechanisms. Furthermore, CO might also have proapoptotic effects in NMU cells through IDO inhibition or otherwise. Recently CO has been shown to induce apoptosis in Jurkat cells by inhibiting ERK MAP kinase activation (44)
. In breast cancer cells, such inhibition leads to cell growth arrest (45)
. Consequently, this pathway might also explain our results showing HO-1 as an anti-proliferative molecule in rat breast cancer. Finally, the down-regulation of HO-1 in NMU-induced breast tumors (Fig. 7)
shows a correlation with the proposed anti-proliferative effect of HO-1. In keeping with this, neoplastic transformation of mammary epithelial cells is associated with decreased apoptotic cell death in the NMU model (46)
. HO-1 down-regulation in rat breast cancer may therefore be implicated in the reduction in apoptosis, but definitive proof of this mechanism needs further study. The fact that human cell lines do not express HO-1 is also consistent with the anti-tumoral role proposed for this enzyme. Nevertheless, when overexpressed, HO-1 was an anti-proliferative and proapoptotic molecule in both rat and human cell lines (Figs. 3
and 4)
.
Until now, no functional studies of the role of IDO in breast cancer biology have been performed. IDO-dependent tryptophan consumption by malignant cells is known to inhibit their proliferation (19)
. More recently, IDO expression by tumor cells (20)
or dendritic cells (DCs) in tumor draining lymph nodes (47)
has been described as an immunosuppression mechanism. We show here for the first time that IDO activity can stimulate cancer cell growth in vitro (Fig. 5)
. The increased proliferation of NMU cells induced by IDO activity is at least partially dependent on HO-1 inhibition, since diminution in NMU proliferation determined by 1-MT was reversible with the HO-1 inhibitor SnPP (Fig. 6F
). In the case of human cell lines, which do not express HO-1 even after 1-MT treatment (data not shown), IDO inhibited cell growth since proliferation was increased upon IDO inhibition (Fig. 5A
). Thus, IDO seems to be a proproliferative molecule in breast cancer only if the anti-proliferative HO-1 is coexpressed. Hence, by inhibiting HO-1, IDO is able to increase proliferation. Otherwise, IDO leads to inhibition of proliferation as was classically described. A similar dual role for IDO could also be operating in vivo in NMU-induced rat breast carcinogenesis. Thus, in normal mammary gland where HO-1 and IDO are coexpressed, the latter could regulate anti-proliferative effects of HO-1. At the carcinoma stage, however, where HO-1 is mostly negative, IDO may lead to decreased proliferation, so it is down-regulated compared with normal mammary gland (Table 1)
. Nevertheless, IDO expression may also constitute a mechanism by which tumor cells resist immune action (20
, 47)
. The physiological role for IDO expression by normal mammary gland needs further investigation. Since IDO can be induced by estrogens (48)
and inhibited by progesterone (49)
, it is tempting to speculate that IDO, as well as HO-1, could participate at the physiological regulation of mammary proliferation.
In the present work we also studied IDO/HO-1 interactions. We hypothesized that HO-1 activity could interfere with that of IDO. As IDO contains heme as its sole prosthetic group (18)
, we speculated that the consumption of this molecule upon HO-1 activity might inhibit IDO enzymatic activity. It has already been established that de novo synthesis of heme is required for IDO activity (50)
and that CO binding to the heme component of IDO weakens its affinity for L- tryptophan (51)
. However, this is the first time that HO-1 activity is shown to inhibit IDO. We have shown that HO-1 inhibits IDO activity at a post-translational level through heme consumption and that inhibition of IDO decreases NMU cell growth (Fig. 5
, Fig. 6A-C
). The expression and activity of nitric oxide synthase (52
, 53)
, cyclooxygenase-2 (54)
, and NAD(P)H oxidase (55)
, three enzymes requiring heme in their catalytic site, is decreased after HO-1-induced heme depletion. Furthermore, heme starvation may also explain that the effects of HO-1 in cancer cells are not reproduced by HO-1-derived metabolites (37)
. The antioxidant effects of HO-1 activity do not seem to regulate IDO activity, as neither bilirubin nor NAC were able to modulate tryptophan consumption. It now remains to be determined whether HO-1-derived CO or free iron are able to regulate IDO activity.
We also found that IDO activity seems to inhibit HO-1 expression and activity (Fig. 6D-F
). Whether tryptophan metabolites and/or tryptophan starvation are responsible for IDO-mediated regulation of HO-1 expression is unclear. Recently, a tryptophan metabolite produced through the kynurenines pathway was reported to enhance HO-1 expression in a macrophage cell line (56)
. The latter study was the first to report any interaction between the IDO and HO-1 pathways. Depending on the molecular context in each cell type, this interaction could occur in various ways. Furthermore, a tryptophan metabolite may not have the same effect as IDO activity, since other metabolites are produced and tryptophan starvation may also determine the final biological effect (57)
.
In conclusion, in the present study we have investigated the functional aspects regarding these enzymes in breast cancer (Fig. 8
). Notably, we found that HO-1 has anti-tumor functions whereas IDO stimulates cancer growth if HO-1 is coexpressed. This is the first report that HO-1 inhibits IDO activity at a post-translational level, in part through heme starvation. We showed that the IDO pathway can inhibit HO-1 expression. HO-1/IDO interaction could also operate in the immune system. Functional consequences in this setting need to be studied.
|
|
ACKNOWLEDGMENTS |
|---|
Received for publication February 16, 2005. Accepted for publication July 26, 2005.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Remy, P. Blancou, L. Tesson, V. Tardif, R. Brion, P. J. Royer, R. Motterlini, R. Foresti, M. Painchaut, S. Pogu, et al. Carbon Monoxide Inhibits TLR-Induced Dendritic Cell Immunogenicity J. Immunol., February 15, 2009; 182(4): 1877 - 1884. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C.-W. Lin, S.-C. Shen, W.-C. Hou, L.-Y. Yang, and Y.-C. Chen Heme oxygenase-1 inhibits breast cancer invasion via suppressing the expression of matrix metalloproteinase-9 Mol. Cancer Ther., May 1, 2008; 7(5): 1195 - 1206. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. A. Rushworth and D. J. MacEwan HO-1 underlies resistance of AML cells to TNF-induced apoptosis Blood, April 1, 2008; 111(7): 3793 - 3801. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Chabannes, M. Hill, E. Merieau, J. Rossignol, R. Brion, J. P. Soulillou, I. Anegon, and M. C. Cuturi A role for heme oxygenase-1 in the immunosuppressive effect of adult rat and human mesenchymal stem cells Blood, November 15, 2007; 110(10): 3691 - 3694. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Was, T. Cichon, R. Smolarczyk, D. Rudnicka, M. Stopa, C. Chevalier, J. J. Leger, B. Lackowska, A. Grochot, K. Bojkowska, et al. Overexpression of Heme Oxygenase-1 in Murine Melanoma: Increased Proliferation and Viability of Tumor Cells, Decreased Survival of Mice Am. J. Pathol., December 1, 2006; 169(6): 2181 - 2198. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
