Catalase Plays a Critical Role in the CSF-independent Survival of Human Macrophages via Regulation of the Expression of BCL-2 Family*

M-colony-stimulating factor (M-CSF)-induced monocyte-derived macrophages (M-MΦ) required continuous presence of M-CSF for their survival, and depletion of M-CSF from the culture induced apoptosis, whereas human alveolar macrophages (A-MΦ) and granulocyte-macrophage (GM)-CSF-induced monocyte-derived macrophages (GM-MΦ) survived even in the absence of CSF. The expression of BCL-2 was higher in M-MΦ, and M-CSF withdrawal down-regulated the expression. The expression of BCL-XL was higher in A-MΦ and GM-MΦ, and the expression was CSF-independent. The expression of MCL-1 and BAX were not different between M-MΦ and GM-MΦ and were CSF-independent. Down-regulation of the expression of BCL-2 and BCL-XL by RNA interference showed the important role of BCL-2 and BCL-XL in the survival of M-MΦ and GM-MΦ, respectively. Human erythrocyte catalase (HEC) and conditioned medium obtained from GM-MΦ or A-MΦ cultured in the absence of GM-CSF prevented the M-MΦ from apoptosis and restored the expression of BCL-2. The activity of the conditioned medium was abrogated by pretreatment with anti-HEC antibody. Anti-HEC antibody also induced the apoptosis of M-MΦ cultured in the presence of M-CSF and GM-MΦ and A-MΦ cultured in the presence or absence of GM-CSF and down-regulated the expression of BCL-2 and BCL-XL in these MΦs. GM-MΦ and A-MΦ, but not M-MΦ, can produce both extracellular catalase and cell-associated catalase in a CSF-independent manner. Intracellular glutathione levels were kept equivalent in these MΦs, both in the presence or absence of CSF. These results indicate a critical role of extracellular catalase in the survival of human macrophages via regulation of the expression of BCL-2 family genes.

Human tissue macrophages (M⌽) 2 play important roles for homeostasis, and in vivo alveolar (A)-M⌽ acquire a strong antioxidant phenotype that contributes to prevention of the oxidant burst in an aerobic environment and can survive for long periods (1). In a previous study, we reported that human A-M⌽ and GM-CSF-induced monocyte-derived macrophages (GM-M⌽) are resistant to hydrogen peroxide (H 2 O 2 ) via their high basal and inducible levels of catalase activity and that M-CSF-induced monocyte-derived macrophages (M-M⌽) are sensitive to low levels of H 2 O 2 with low levels of catalase activity (2). GM-M⌽ is phenotypically identical to A-M⌽, whereas M-M⌽ closely resembles peritoneal M⌽ in respect to morphology, cell surface antigen expression, and several biological functions (2)(3)(4)(5)(6)(7). In vivo A-M⌽ express BCL-2 family proteins such as BCL-2 and BCL-X L that prevent H 2 O 2 -induced apoptosis via inhibition of caspase-3 or -9 activation and cytochrome c release from mitochondria (8 -11). These findings suggest that a high level of catalase activity enables long survival of GM-M⌽ and A-M⌽ with positive regulation of BCL-2 family protein.
In this study, we found that M-M⌽ absolutely require CSF for their survival and express high levels of BCL-2 gene and protein in the presence of M-CSF. In contrast to M-M⌽, GM-M⌽ and A-M⌽ can survive in the absence of CSF via high levels of BCL-X L gene and protein. We further examined the relation between catalase activity and distinct expression of BCL-2 family protein and make clear the roles of CSF in the regulation of catalase activity and BCL-2 family protein in human tissue macrophages under the influence of oxidative stress.
Human A-M⌽ were obtained from healthy volunteers (non-smokers without diseases) by bronchial alveolar lavage method (2,7). All volunteers gave informed consent to the use of their A-M⌽ in part for this study.
Antisense Oligonucleotide (AS) Treatment-2Ј-O-methyl-modified oligoribonucleotide phosphorothioate 18-mer two CpG motifs targeted to the BCL-2 initiation codon (G3139 (BCL AS): 5Ј-TCTCCCAGCGT-* This study was supported in part by grants for Research on Health Sciences Focusing on Drug Innovation from the Japan Health Sciences Foundation and the Ministry of Health, Labor, and Welfare of Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 To whom correspondence should be addressed. GCGCCAT-3Ј), G3139 variant with single base mismatch at each CpG motif (G4126 (BCL missense (MS)): 5Ј-TCTCCCAGCATGTGCCAT-3Ј) (12) and 18-mer to the 5Ј-splice site of BCL-X L (5Ј-BCL-X AS, ACCCAGCCGCCGUUCUCC) (13) and MCL-1 AS (ISIS 20 408, TTG-GCTTTGTGTCCTTGGCG) (14) were synthesized by Proligo France SAS (Paris, France). Oligonucleotides with random sequences were used as negative controls. Cells were treated with oligonucleotides complexed with Lipofectin (5 mg/ml; Invitrogen) cationic lipid delivery agent according to the manufacturer's directions. Assessment of Cell Number and Cell Viability-The number of adherent Mo and M⌽ was determined by counting the liberated intact nuclei from lysed cells stained with 1% (w/v) cetyltrimethyl ammonium bromide (Cetavelone; Wako Pure Chemical Industries, Ltd., Osaka, Japan) in 0.1 M citric acid with 0.05% (w/v) naphthol blue black (Sigma). Cell viability was assessed by the trypan blue dye exclusion test.
Assessment of Apoptosis-DNA fragmentation was detected by immunohistochemical staining using the terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end-labeling method (Apo-tag kit; Oncor Co.) or visualized as DNA ladder formation (5). Cells were preincubated with 1% H 2 O 2 in phosphate-buffered saline for inactivation of endogenous peroxidase for 5 min at room temperature. Incorporation of digoxigenin-conjugated dUTP to the terminal 3Ј-OH of fragmented DNA by exogenous terminal deoxynucleotidyl transferase was carried out at 37°C for 1 h. The reaction products were incubated with horseradish peroxidase-linked anti-digoxigenin antibody at 37°C for 30 min and visualized with the substrate 3-3Ј diaminobenzidine plus 0.6% H 2 O 2 .
Cells were lysed with hypotonic lysis buffer (10 mM Tris-Cl (pH 7.4), 10 mM EDTA (pH 8.0), 0.5% Triton-X), and crude DNA was extracted from the lysed cells by incubation with 40 g/ml RNase and 40 g/ml proteinase K for 1 h at 37°C. The DNA was precipitated with final 50% propanol at Ϫ20°C overnight and washed with 70% (w/v) ethanol. Electrophoresis was performed in 2% agarose at 50 V, and the migrated DNA was visualized by ethidium bromide staining.
Neutralization of Conditioned Medium-Conditioned medium was obtained from M⌽ cultured in medium alone for 48 h at 37°C and incubated for 60 min at 37°C with 10 g/ml rabbit anti-human erythrocyte catalase (HEC) IgG (lot PTC 9301; Athens Research and Technology, Inc., Athens GA) or with normal rabbit IgG (Organon Teknika Co.) as a control.
Measurement of Catalase Activity-Intracellular and extracellular catalase activity was measured according to Aebi's modified method as described previously (2). Purified HEC (5 ϫ 10 4 units/mg, lot 643793; Calbiochem-Nova Biochem) was used for conversion of the catalase activity in samples into real catalase activity, and the activity was expressed as milliunits/ml per 2.5 ϫ 10 5 cells or units/mg protein. Protein was measured using a protein assay kit (Bio-Rad Laboratories).
Measurement of GSH Level-The total GSH level was determined using a BIOXYTECH S. A. kit. Briefly, the cells were lysed with 5% metaphosphoric acid, and the chromophore formation catalyzed at pH 13.4 was measured at 400 nm as described previously (15). The intracellular GSH level was expressed as pmol/mg protein.
Isolation of RNA and Northern Blot Analysis-Isolation of total RNA and Northern blot analysis were performed as described previously (5). Blots were hybridized with cDNA probes against human catalase (kindly donated by Dr. K. Onozaki, Faculty of Pharmaceutical Sciences, Nagoya City University) and ␤-actin (Sigma). All the probes were labeled using a multiprime DNA-labeling system with [␣-32 P]dCTP (New England Nuclear Research Products, Boston, MA). The blots were analyzed using a Fuji BAS 2000 bioimage analyzer (Fuji Photo Film Co., Ltd., Tokyo).
Immunoblot Analysis-Immunoblot analysis was performed as described in our previous report (7). The membrane was incubated overnight at 4°C with 1 mg/ml rabbit anti-HEC antibody (anti-Cat Ab; Athens Research and Technology, Inc.), mouse anti-BCL-2 antibody (Santa Cruz Biotechnology, Santa Cruz, CA), rabbit anti-BCL-xL/S antibody (Santa Cruz Biotechnology), rabbit anti-BCL-X antibody (Transduction Laboratories, Lexington, KY), mouse anti-MCL-1 antibody (Santa Cruz Biotechnology), rabbit anti-BAX antibody (Santa Cruz Biotechnology), or normal rabbit or mouse IgG and then at room temperature for 1 h with horseradish peroxidase-conjugated goat antirabbit IgG or anti-mouse IgG (Santa Cruz Biotechnology). The blots were visualized with Amersham ECL reagent on Hyper ECL film (Amersham Biosciences). Statistical Analyses-Statistical analyses of the data were performed using Student's t-test. p values Ͻ0.01 were considered significant. The results shown are representative of three to seven independent experiments.

Requirement for Continuous Presence of CSF for the Survival of M-M⌽, but not of GM-M⌽ and A-M⌽-M-M⌽, GM-M⌽, and A-M⌽
were cultured with or without CSF, and then cell viability was determined. M-CSF withdrawal from M-M⌽ induced cell death in a timedependent manner; ϳ60% of the cells died by 2 days and almost all of the cells died by 8 days (Fig. 1A). In contrast, GM-CSF withdrawal from GM-M⌽ or A-M⌽ did not induce cell death significantly during the culture period (Fig. 1A). M-M⌽ cultured in the absence of M-CSF changed from spindle shaped to round and then detached from the dish and floated, with a shrunken cytosol and nucleus and with chromatin condensation and apoptotic vacuoles without microvilli (Fig. 1B). No such morphological change was observed in M-M⌽ cultured with M-CSF or in GM-M⌽ cultured with or without GM-CSF (Fig. 1B). A typical ladder pattern of internucleosomal DNA cleavage was detected in DNA of M-M⌽ cultured for 24 h in the absence of M-CSF, whereas no significant DNA fragmentation was detected in DNA of M-M⌽ cultured in the presence of M-CSF or that of GM-M⌽ cultured in the presence or absence of GM-CSF (Fig. 1C). Apoptosis and DNA fragmentation of A-M⌽ were similar to those of GM-M⌽ (data not shown). These findings suggest that M-CSF withdrawal from M-M⌽ induces apoptosis but both GM-M⌽ and A-M⌽ can survive in the absence of CSF.
Different Expression of BCL-2 and BCL-X L Genes in M-M⌽, GM-M⌽, and A-M⌽ and the Opposite Effect of CSF on Expression-BCL-2 family genes play important roles in the apoptosis of many types of cells (10, 16 -20). We therefore examined the expression of BCL-2 family genes and the effect of CSF on their expression in M-M⌽ and GM-M⌽ by RT-PCR. In M-M⌽, the transcript of the BCL-2 gene was stronger than that of the BCL-X L gene, and M-CSF withdrawal decreased the expression of the BCL-2 gene but induced a slight decrease in the expression of the BCL-X L gene ( Fig. 2A). In contrast to M-M⌽, GM-M⌽ expressed strongly the BCL-X L gene but weakly the BCL-2 gene, and GM-CSF withdrawal had no significant effect on the expression of these two genes ( Fig. 2A). In contrast to BCL-2 and BCL-X L , the transcriptional levels of the MCL-1 and BAX genes were not significantly different between M-M⌽ and GM-M⌽ and were not affected by CSF deprivation (Fig. 2A). In accordance with the gene expression, the expression of the BCL-2 protein was higher in M-M⌽ compared with GM-M⌽ and M-CSF withdrawal decreased the expression of BCL-2 protein in M-M⌽ (Fig. 2B). Similarly, the expression of the BCL-X L protein was higher in GM-M⌽ than in M-M⌽, but GM-CSF withdrawal had no significant effect on the expression of BCL-X L protein in GM-M⌽ (Fig. 2B).
The expression of BCL-2 family proteins in A-M⌽ resembles that of GM-M⌽, and the expression was independent of CSF (  (Fig. 3A).
A previous study showed that human CEM T cells can survive in serum-free medium supplemented with a low level of extracellular catalase (21). Therefore, we examined whether extracellular catalase pre-   protein in both GM-M⌽ and A-M⌽ cultured without GM-CSF (Fig.  2B). In accordance with these data, addition of anti-Cat Ab induced cell death of M-M⌽ cultured with M-CSF and GM-M⌽ cultured with or without GM-CSF, and addition of control IgG had no effect on the cell viability of these M⌽s (Fig. 3C). The cell death of these M⌽ induced by anti-Cat Ab was due to apoptosis, as indicated by terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end-labeling staining (data not shown).

GM-M⌽ and A-M⌽, but Not M-M⌽, Can Produce Enough Levels of Extracellular Catalase in the Absence of CSF-
The above results suggest that GM-M⌽ and A-M⌽, but not M-M⌽, can produce extracellular catalase in a CSF-independent manner and that the catalase supports the survival of those M⌽s via maintenance of the expression of BCL-2 family genes. To examine this possibility, we measured catalase enzyme activity and catalase protein levels in the CMs of M-M⌽, GM-M⌽, and A-M⌽ cultured for 48 h in the presence or absence of CSF. The extracellular catalase activities in the CMs obtained from GM-M⌽ and A-M⌽ cultured both in the presence or absence of GM-CSF were not significantly different, and the levels were ϳ4-fold higher (ϳ240 milliunits/ml/well) than that of M-CSF-treated M-M⌽ (ϳ60 milliunits/ml/ well) (Fig. 4A). In contrast to GM-M⌽ or A-M⌽, the extracellular catalase activity in CM of M-M⌽ was dependent on CSF and that in CM of M-CSF-withdrawn M-M⌽ was significantly lower than that in CM of M-CSF-treated M-M⌽, ϳ20 milliunits/ml/well (Fig. 4A). Western blot analysis of the extracellular catalase of CMs using anti-HEC antibody showed similar results to the data of the enzyme activity (Fig. 4A).
Next, we examined the cell-associated levels of catalase activity in these M⌽s (Fig. 4B). The levels of catalase activity in GM-M⌽ and A-M⌽ lysates was ϳ5 units/mg protein both in the presence or absence of CSF, whereas that in M-M⌽ in the presence of M-CSF was ϳ1 unit/ mg. The catalase activity in M-CSF-withdrawn M-M⌽ lysates (ϳ250 milliunits/mg protein) was ϳ4-fold lower than that in M-CSF-treated M-M⌽ lysates. In agreement with the measurements of enzyme activity, similar results were observed in Western blot analyses (Fig. 4B).
To further confirm the distinction between the regulation of extracellular and cell-associated catalase activity by CSF in M-M⌽ versus GM-M⌽, we examined the levels of transcription of the catalase gene in these M⌽s cultured with or without CSF (Fig. 4C). The level of the transcript of the catalase gene in M-CSF-withdrawn M-M⌽ was ϳ3-fold lower than that in M-CSF-treated M-M⌽. In contrast, the level of the catalase transcript in GM-CSF-withdrawn GM-M⌽ and A-M⌽ was similar to that in GM-CSF-treated GM-M⌽ and A-M⌽, and the level was 5-fold higher than that in M-CSF-treated M-M⌽.
Thus, the difference of total extracellular plus intracellular catalase activity and the difference in the levels of catalase gene expression between CSF-withdrawn M-M⌽ and GM-M⌽ reached ϳ15-20and ϳ15-fold, respectively. The results indicate that extracellular catalase activity is regulated at the transcription levels by CSF-dependent M-M⌽ but CSF-independent GM-M⌽ and A-M⌽.
Thiol Derivatives Act as Additive Effectors That Rescue the Cell Death of M⌽-The above data suggest that extracellular catalase plays a major role in M⌽ survival. However, several reports have demonstrated that thiol proteins and thiol compounds such as GSH, adult T cell leukemiaderived factor, and L-cysteine play important roles in the survival of medium with M-CSF, CM pretreated with Anti-Cat Ab (IgG), or rabbit IgG as a control. Cell number and viability of M⌽ were assessed as described in Fig. 1. Values are expressed as the means of triplicate cultures Ϯ S.D. C, M-M⌽ and GM-M⌽ were cultured for 48 h in medium with or without CSF plus Anti-Cat Ab or rabbit IgG. Cell number and viability of M⌽ were assessed as described in Fig. 1. Values are expressed as the means of triplicate cultures Ϯ S.D.  (22,23,24). Thus, thiol derivatives may help M⌽ survival. The level of intracellular GSH, however, was almost the same in M-M⌽ and GM-M⌽ cultured with or without CSF or cultured with catalase (Fig. 5A), and this level (ϳ300 pmol/10 5 cells) was similar to that in A-M⌽ (data not shown). Diamide, which can bind the SH groups of reduced thiols and oxidize them (22), induced cell death of M-M⌽ and GM-M⌽, but the levels of cell death were not very high (ϳ20%) and no significant difference was observed between the effects of diamide on these two M⌽s (Fig. 5B). 40 M diamide and 10 g/ml anti-Cat Ab showed synergistic effects on the cell viability of M-M⌽ cultured with M-CSF and GM-M⌽ cultured without CSF, causing reduction of their viability to Ͻ10% at 48 h (Fig. 5B). These results suggest that, in contrast to that of catalase, The relative amounts of catalase protein in cells were measured using NIH image software, and the expression levels were corrected relative to those of ␤-actin (photo-stim ulating luminescence (PSL), A/mm 2 ). C, mRNA levels of catalase and the ␤-actin gene were examined in total RNA preparations (10 g/lane) from these M⌽ at 3 h of cultivation by Northern blot analysis. The relative transcript levels of catalase mRNA in cells were measured using NIH image software, and the expression levels were corrected relative to those of ␤-actin (photo-stimulating luminescence, A/mm 2 ). the levels of thiol derivatives were constant in these M⌽s with CSF withdrawal-induced oxidative stress and that thiol derivatives have only a minor effect and cannot support the full survival of M⌽.

BCL-2 AS and BCL-X L AS Dominantly Induce Cell Death of M-M⌽ and GM-M⌽, Respectively-M-M⌽
and GM-M⌽ were treated with BCL-2 AS targeted to BCL-2 initiation codon (G3139) and 5Ј-BCL-X AS targeted to the downstream alternative 5Ј-splice site of exon 2 of the BCL-X gene (12,13). As controls, these M⌽s were treated with BCL-2 MS (G4126, variant G3139), control oligonucleotides with random sequence (oligonucleotide MS), or Lipofectin alone. As shown in Fig. 6, BCL-2 AS (G3139) down-regulated the expression of BCL-2 protein in M-M⌽ to 10 -20% of that of control cells at 2 days after the oligonucleotide treatment and induced cell death in a dose-dependent manner. The cell viability markedly decreased to ϳ15% of that of control cells at 7 days of cultivation. In GM-M⌽, however, treatment with G3139 even in the high dose such as 10 M induced only ϳ10% cell death, in agreement with the low expression of BCL-2 protein in this M⌽ (Fig. 7).
In contrast, treatment of GM-M⌽ with 5Ј-BCL-X AS down-regulated the expression of BCL-X L protein to 20% of that of control cells treated with oligonucleotide MS at 2 days after the oligonucleotide treatment and induced the cell death in a dose-dependent manner (Fig.  7). The cell viability markedly decreased to ϳ25% of that of control cells at 7 days of cultivation. As shown in Fig. 6

DISCUSSION
The present study showed that extracellular catalase has a novel role in the prevention of apoptosis in human M⌽ through the dominant expression of BCL-2 in M-M⌽ and BCL-X L in GM-M⌽ and that the regulation of catalase production is CSF-dependent in M-M⌽ but CSFindependent in GM-M⌽ and A-M⌽. Recently, H 2 O 2 has been shown to enhance oxidative damage and apoptosis in C2-ceramide-pretreated HL-60 cells via a mechanism in which C2-ceramide inhibits catalase activity by increasing casapase-3-dependent proteolysis of catalase and down-regulation of catalase mRNA (25). Overexpression of catalase inhibits oxidation-mediated apoptosis through phosphorylated BCL-2, a reduced form of BCL-2 and BAX interaction (20). Thus, constant high activity of extracellular catalase plays an important role in the prevention of ceramide-and caspase-3-induced apoptosis in CSF independent of GM-M⌽ and A-M⌽ and CSF dependent of M-M⌽ (11,26).
GM-CSF and M-CSF stimulate catalase induction during the differentiation of Mo into M⌽, but GM-CSF alone establishes a CSF-independent autoregulatory system of catalase production in M⌽. We previously showed that human Mo-derived GM-M⌽ resembles human A-M⌽ in several respects (2,(5)(6)(7). In this study, we showed that GM-M⌽ and A-M⌽ have a similar phenotype of the resistance to apoptosis via CSF-independent expression of catalase and BCL-X L . Con-  Fig. 1. Values are expressed as the means of triplicate cultures Ϯ S.D. Western blot analysis of BCL-2, BCL-X L/S , MCL-1, and ␤-actin proteins in cell lysates of M⌽ probed using BCL-2, BCL-X (antibody from Santa Cruz Biotechnology for X L/S band, from Sinal transduction Labo for X L band only), and ␤-actin antibody was performed as described in Fig. 4. The relative amounts of these proteins in cells were measured using NIH image software, and the expression levels were shown as photo-stimulating luminescence (PSL). DECEMBER 16, 2005 • VOLUME 280 • NUMBER 50 sistent with our present study, A-M⌽ from human smokers express higher levels of p21 CIP1/WAF1 and BCL-X L , but not BCL-2, and the former two molecules may reduce apoptosis (9). The autoregulatory mechanism of catalase induction in GM-M⌽ and A-M⌽ is not yet understood but may have a strong correlation with endogenously generated low levels of H 2 O 2 (8), because we previously showed that GM-M⌽ and A-M⌽, but not M-M⌽, can increase catalase expression at both the protein and mRNA levels when stimulated with H 2 O 2 (2). Adequate low levels of H 2 O 2 may keep catalase activity constant to prevent CSF deprivation-induced apoptosis in GM-M⌽ and A-M⌽.

Catalase Inhibits CSF-depleted M⌽ Death via BCL-2 and BCL-X L
Thiol derivatives such as GSH, adult T cell leukemia-derived factor, and L-cysteine play an important role in the survival of lymphocytes or neuronal cells in the absence of growth factors (22)(23)(24). In our study, however, the activity of thiol derivatives was not significantly different between M-M⌽ and GM-M⌽ before and after CSF deprivation, and they had only a partial effect on M⌽ survival. The reason for the differences between our study and previous studies might have been the differences in the experimental conditions (CSF-withdrawal versus serum depletion) or the cell type examined (M⌽ versus lymphocytes/ neuronal cells). Compared with M⌽, lymphocytes or neuronal cells may produce lower levels of catalase so that the thiol derivatives play a dominant role in their survival (21)(22)(23)(24).
We demonstrated that catalase regulates apoptosis through the expression of BCL-2 and BCL-X L in human M⌽ used in the present study. In fact, down-regulation of these proteins by RNA interference treatment induced the cell death of the M⌽s. Proapoptotic BCL-2 family protein, BAX, can induce apoptosis with permeabilization of mitochondrial membranes and cytochrome c release (10,18,27). In a recent study, induction of apoptosis of TF-1 cells by GM-CSF withdrawal is shown to be related to down-regulation of the MCL-1 gene, and over-expression of MCL-1 is shown to delay apoptosis (16). In our study, however, BAX and MCL-1 were not associated with the regulation of apoptosis of M⌽, because both mRNA and protein levels of these genes were not significantly changed in M-M⌽ and GM-M⌽ after CSF deprivation. Moreover, we showed that down-regulation of the expression of MCL-1 protein by MCL-1 As treatment did not affect the viability of the M⌽s.
Our interesting finding is that BCL-2 and BCL-X L are differently expressed in M⌽ during the Mo differentiation into M⌽ in the presence of M-CSF and GM-CSF; the expression of BCL-2 is dominant in M-M⌽ and that of BCL-X L is dominant in GM-M⌽ or A-M⌽. In accordance with such different expression patterns, we found by RNA interference experiments that BCL-2 and BCL-X L play a critical role for the survival of M-M⌽ and GM-M⌽, respectively. Similar differential expression of BCL-2 and BCL-X L is also observed during the selection and maturation of mouse thymocytes toward splenic T cells (17,19). At present, we do not know the mechanisms that control the different induction of BCL-2 and BCL-X L in M-M⌽ and GM-M⌽ or A-M⌽, respectively. The distal promoter region of the BCL-X L gene responds to very low levels of H 2 O 2 at exon 1B-1D in rodent cardiac myocytes, and the expression of BCL-X L protein is increased by H 2 O 2 treatment (28). Thus, GM-M⌽ and A-M⌽, which possess high catalase activity even in the absence of CSF, limit the levels of endogenously generated H 2 O 2 to low levels that are suitable for BCL-X L expression.
In conclusion, the present study is the first to reveal that CSF is a critical regulator of extracellular catalase activity that maintains selective expression of BCL-2 family genes and prevents tissue-specific M⌽ from apoptosis. These distinct patterns of CSF-induced regulation of catalase activity may greatly contribute to the oxidant stress-induced selection of tissue M⌽s suitable for their respective microenvironments.