- Email: [email protected]
Beryllium-stimulation does not activate transcription factors in a mouse hybrid macrophage cell line
Toxicology 143 (2000) 249 – 261 www.elsevier.com/locate/toxicol
Beryllium-stimulation does not activate transcription factors in a mouse hybrid macrophage cell line Hironobu Hamada a,*, Richard T. Sawyer a,b, Lori A. Kittle a, Lee S. Newman a,b,c a Di6ision of En6ironmental and Occupational Health Sciences, Department of Medicine, National Jewish Medical and Research Center, 1400 Jackson Street, Den6er, CO 80206, USA b Di6ision of Pulmonary and Critical Care Medicine, Department of Medicine, Uni6ersity of Colorado Health Sciences Center, Den6er, CO USA c Department of Pre6enti6e Medicine and Biometrics, Uni6ersity of Colorado Health Sciences Center, Den6er, CO, USA
Received 14 August 1999; accepted 22 October 1999
Abstract We tested the hypothesis that beryllium (Be) could stimulate H36.12j cell (12j) TNF-a production by transcription factor-mediated pathways similar to those induced by either LPS- or IFN-g stimulation. Unstimulated 12j cells produce constitutive levels of TNF-a (175918 pg/ml, mean9 SEM) detected by ELISA of culture supernatants after 24 h. Beryllium-stimulated (100 mM BeSO4) 12j cell TNF-a (724947 pg/ml) was observed after 24 h while LPS-stimulated (1 mg/ml) TNF-a (5159151 pg/ml) after 6 h. Recombinant-Mu-IFN-g (10 U) stimulated 12j cell TNF-a at lower levels (284 931 pg/ml) while rMu-IFN-g+Be-stimulated 12j cells produced 1195 9225 pg/ml TNF-a. Constitutive levels of transcription factors were observed in unstimulated 12j cell nuclei. In LPS-stimulated 12j cells IkBa was degraded in the cytoplasm and increased levels of NF-kB were found in nuclei after 30 min. After 3 h there were increased levels of AP-1 and CREB, with increased amounts of Fos family, Jun B and Jun D transcription factors. In contrast, Be-stimulation failed to increase the levels of any transcription factor tested, NF-kB, AP-1, AP-2, CREB, C/EBP, Sp-1, Egr-1, Ets, NF-Y or Oct-1, in 12j cells. A pattern of increased transcription factors, similar to that observed for LPS-stimulation, was found in 12j cell nuclei after stimulation with rMu-IFN-g. However, NF-kB was increased at 3 h while AP-1 (Jun B and Jun D) and CREB were increased at 15 h. Co-stimulation of 12j cells with rMu-IFN-g+ Be increased the levels of NF-kB in 12j cell nuclei at 3 h, and the levels of AP-1 and CREB at 15 h, however, only Jun B was increased. Our data show 12j cell TNF-a production was associated with increased levels of transcription factors present in nuclei with disparate kinetics and patterns of expression depending on the trigger. We reject our initial hypothesis and conclude that Be-stimulation signals 12j cell TNF-a synthesis via a transcription factor-independent pathway. Beryllium may induce novel pathways of macrophage cytokine gene regulation. © 2000 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Beryllium; Chronic beryllium disease; Transcription factors; Macrophages; Hybrid macrophages; NF-kB/IkBa; AP-1; CREB; Bzip proteins; Lipopolysaccharide; Interferon-g; Tumor necrosis factor-a; TNF-a promoter * Corresponding author. Tel.: + 1-303-398-1167; fax: + 1-303-398-1851. E-mail address: [email protected] (H. Hamada) 0300-483X/00/$ - see front matter © 2000 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 3 0 0 - 4 8 3 X ( 9 9 ) 0 0 1 8 3 - 3
250
H. Hamada et al. / Toxicology 143 (2000) 249–261
1. Introduction Chronic beryllium disease (CBD) results from exposure to environmental beryllium (Be), which occurs most often in an industrial setting (Newman, 1995). In vitro, Be-stimulated bronchoalveolar lavage (BAL) cells obtained from CBD subjects produce high levels of pro-inflammatory cytokines. These cytokines are believed to participate in the proliferation of Be-sensitized CD4+ T cells and in the pathogenesis of the CBD granuloma (Newman et al., 1996; Fontenot et al., 1998). For example, Be-stimulated CBD/BAL cells produce high levels of IFN-g ( \ 10 ng/ml) and transient levels of IL-2, but do not produce IL-4 or IL-7 (Tinkle et al., 1996, 1997). Be-stimulated CBD/BAL cells also produce high levels of IL-6 and TNF-a (Tinkle and Newman, 1997). Bost et al. (1994) showed that in comparison to normal subject BAL macrophages, isolated CBD/BAL macrophages expressed higher levels of TNF-a mRNA. The mechanism by which Be stimulates CBD/BAL cells to produce cytokines is unknown. At present, there are no in vitro models that use Be-responding cells or cell lines, that would permit the generation of testable hypothesis directly related to answering questions about possible mechanisms of Be-induced cytokine up-regulation in CBD. Our own studies (Bost et al., 1994; Tinkle et al., 1996, 1997; Tinkle and Newman 1997), designed to investigate mechanisms of Be-stimulated cytokine synthesis, have been limited by the availability of specific CBD cell types. However, we have identified a macrophage cell line that mimics the production of Be-induced macrophage TNF-a synthesis by CBD/BAL cells. H36.12j ‘Pixie’ cells (12j) are a hybrid macrophage cell line (Canono and Campbell, 1992; Sawyer et al., 1998). We hypothesized that Be stimulates 12j cell TNF-a production by transcription factor mediated pathways similar to those induced by either LPS- or IFN-g stimulation. We tested this hypothesis by measuring the levels of 12j cell transcription factors, known to be involved in binding the TNF-a promoter region (+1 to −1500 bp) (Tracey, 1998). These included the NF-kB/IkBa family of transcription
factors, bzip transcription factors (AP-1, Fos/Jun family, CREB, C/EBP), AP-2, Sp1, Ets and Egr1. Using 12j cells as a model, we observed disparate transcription factor pathways by which different triggers increase TNF-a synthesis. Our data shows that Be stimulation may induce macrophage cytokine production independent of known nuclear transcription factor regulation. Of interest, co-stimulation with IFN-g plus Be stimulates TNF-a production by an AP-1 mediated mechanism not previously described.
2. Materials and methods
2.1. Chemicals and reagents A filter-sterilized, aqueous stock solution, made in pyrogen-free, sterile water (Abbot Lab. Chicago, IL) of 1 mM BeSO4·4H2O (BE, Fisher Sci., Springfield, NJ) was held at 4°C prior to use. Sterile transfers were made into 12j cell cultures at zero time, to a final concentration of 100 mM. This amount of BeSO4 was selected based on a preliminary dose–response experiment showing that at 48 h, 12j cells stimulated with 100 mM BeSO4 produced 3029 41 pg/ml (mean9SEM) TNF-a, 163 9 2 pg/ml TNF-a at 50 mM BeSO4, 619 2 pg/ml TNF-a at 25 mM BeSO4, and 60 9 2 pg/ml TNF-a at 5 mM BeSO4. Thus, 100 mM BeSO4-stimulation induced maximal TNF-a production by 12j cells. Previous study also shows that maximal TNF-a levels are produced by CBD/BAL cells stimulated with 100 mM BeSO4 (Tinkle and Newman, 1997). 12j cells and culture supernatants were harvested either immediately (zero time) or at the indicated intervals up to 24 h. As controls, 12j cells were unstimulated, stimulated with 1 mg/ml of Salmonella typhimurium lipopolysaccharide (LPS, Sigma Chem., St. Louis, MO) or 10 U of rMuIFN-g (Genzyme, Boston, MA). We tested our beryllium-salt stimulatory reagents using the Limulus amebocyte lysate assay (Associatesof Cape Cod, Woods Hole, MA), which has a limit of detection of 0.025 ng of LPS. All reagents and glassware used in this study were pyrogen free. The
H. Hamada et al. / Toxicology 143 (2000) 249–261
calf serum contained B0.05 ng/ml of LPS (BioWhittiker).
2.2. Macrophage cell lines and culti6ation H36.12j ‘Pixie’ cells are clonally derived, hybrid precursor macrophages, derived from the fusion of drug selected P388D.1 (DBA/2, H2d) macrophages with percoll gradient purified, proteose peptone elicited macrophages obtained from C57Bl/6N (H2b) mice (Canono and Campbell, 1992). 12j cells were negative for the presence of Mycoplasma species epitopes by PCR analysis (Stratagene, LaJolla, CA) and have been deposited with the ATCC (Rockville, MD). For all studies, 12j cells were enumerated by hemocytometer and viability was greater than 90% at the initiation of each experiment. 12j cells were cultivated in Dulbecco’s Modified Eagle’s Medium (BioWhittiker, Walkerville, MD) supplemented with 10% heat inactivated calf serum, 200 mM L-glutamine, 10 000 U/ml penicillin G and 10 000 mg/ml streptomycin sulfate. 12j cells were cultivated at 1×105 cells/200 ml/well in 96 well round bottom culture dishes (Corning, NY) at 37°C in a humidified atmosphere containing 5% CO2. 12j cells were harvested by aspiration and concentrated by microcentrifugation. Culture supernatants were held at −80°C until use.
2.3. Quantification of cytokine protein Cytokine protein was measured with a solidphase sandwich ELISA (R&D Sys., Minneapolis, MN) with a reported sensitivity of 4 pg/ml and data are reported as the mean 9 SEM pg/ml from four wells, using six separate experimental determinations.
2.4. Western immunoblot analysis At the indicated times 1×107 12j cells/condition were harvested and nuclear and cytoplasmic protein extracts prepared by the method of Das and White (1997) and Hagman et al. (1989). Assay (Pierce, Rockford, IL) for total protein indicated a yield of : 500 mg of protein/1 ×107 12j cells. Nuclear and cytoplasmic levels of NF-
251
kB/IkBa were determined by Western immunoblot analysis. Western immunoblots were prepared by transfer of 12j cell extract proteins separated by 10% SDS-PAGE. Immunoblots were incubated with primary anti-transcription factor antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) for NF-kB, p65 or p50, and IkBa. Data obtained using the anti-p50 NF-kB antibody are reported. Identical results were obtained in Western immunoblots probed with the anti-p65 NF-kB antibody (not shown). Binding of antitranscription factor antibody was detected using a second anti-globulin coupled to alkaline phosphatase. Chemifluorescence (ECF Western Blotting System, Amersham, UK) was detected using the Storm 840 PhosphorImager and ImageQuant analysis program (Molecular Dynamics, Sunnyvale, CA).
2.5. Electrophoretic mobility shift assay (EMSA) The EMSA was performed using unstimulated, LPS-stimulated, rMu-IFN-g-stimulated, BeSO4stimulated, and rMu-IFN-g + BeSO4-stimulated 12j cell nuclear extracts. In brief, at the indicated intervals, 12j cells were concentrated by centrifugation and suspended at 1× 107 cells in 1 ml of buffer A (10 mM HEPES (Sigma Chemical, St. Louis, MO) pH 7.9; 300 mM sucrose; 10 mM KCl; 1.5 mM MgCl2; 1 mM ditiothreitol (DTT, Sigma) and 26 ml of a 5 mg of PMSF/ml of isopropanol stock, 0.76 mM final). The 12j cells were concentrated by microcentrifugation at 1200 rpm for 3 min, the supernatant decanted and the cell suspended in 900 ml of buffer A plus 100 ml of NP40 (0.5% final, Sigma) and 0.76 mM PMSF. The suspended cells were vortexed for 1 min at the high setting. A 10 ml aliquot was removed and examined microscopically for the formation of cell nuclei. The nuclei were concentrated by microcentrifugation at 14 000 rpm for 15 min. The supernatant was removed using a micropipet, and 100 ml aliquots of the cytoplasmic extract were held at − 80°C until use. The nuclear pellet was suspended in 150 ml of buffer C (20 mM HEPES, pH 7.9; 300 mM sucrose; 420 mM NaCl; 1.5 mM MgCl2; 0.76 mM PMSF). The nuclei were lysed by vortex for 15 min and the debris concentrated
252
H. Hamada et al. / Toxicology 143 (2000) 249–261
by microcentrifugation at 14 000 rpm for 5 min. The nuclear extract was transferred to a tube containing 15 ml of glycerol, mixed and stored at − 80°C until use. Nuclear extract binding reactions were performed as described (Das and White, 1997) in the absence (negative control) and presence of double stranded oligonucleotides (Promega, Madison, WI) containing consensus binding sequences for Sp-1, AP-1, AP-2, NF-kB, Oct-1, CREB, C/EBP, NF-Y, Ets, and Egr-1. Probes were end-labeled with 32P-ATP. Cold competition, control reactions were performed using 32 P-labeled oligonucleotide probe plus 100-fold excess of the unlabeled probe. As an additional control, binding reactions were performed in the absence of nuclear extract but with the addition of recombinant transcription factor (2 ng/reaction, Santa Cruz). Levels (32P counts) of transcription factors were measured using the Storm 840 PhosphorImager and ImageQuant analysis program at the indicated times (0 – 24 h) after stimulation.
2.6. Antibody supershift EMSA The EMSA was performed as above, in the absence and presence of antibodies directed against nuclear transcription factors. Antibodies directed against transcription factors were purchased from Santa Cruz Biotechnologies, Inc.
(Santa Cruz, CA) and used according to the manufacturer’s specifications. Antibodies were added to binding buffer 20 min before the addition of labeled-oligonucleotide probes. We calculated data for anti-transcription factor antibody+ transcription factor(s) that were super-shifted by subtracting the remaining bands after antibody super-shift EMSA from the negative control transcription factor band (formed by a reaction between 32P-oligonucleotide+ nuclear extract, but without added anti-transcription factor antibody). Antibody-transcription factor complexes are delayed in their migration through the EMSA gel, permitting specific identification and quantification using the Storm 840 PhosphorImager.
2.7. Statistical analysis In order to validate our analysis of changes in the levels of cytoplasmic and nuclear transcription factors Western immunoblot, EMSA and EMSA supershift analysis were performed using extracts from three separate experiments. The figures presented, therefore, are representative of these repeated experiments. Analysis of variance was performed using the Tukey Kramer test (Tinkle et al., 1997; Tinkle and Newman, 1997). Positive values for pairs of means were considered significantly different at a P value of 5 0.05.
3. Results
3.1. The kinetics of 12j cell TNF-a production
Fig. 1. The production of TNF-a (mean9 SEM pg/ml) at 6 h (open) and 24 h (black) by unstimulated, Be-stimulated (100 mM), LPS-stimulated (1 mg/ml), IFN-g stimulated (10U) and IFN-g (10U) +Be (100 mM)-stimulated 12j cells as measured by ELISA of culture supernatants. *P5 0.05 at 24 h for TNF-a levels from Be, IFN-g and IFN-g+ Be stimulated 12j cells, and at both 6 and 24 h for LPS-stimulated 12j cells.
TNF-a levels (mean9SEM pg/ml) were measured by ELISA of 12j culture supernatants at 6 and 24 h. Unstimulated 12j cells produced constitutive levels of TNF-a with 8793 pg/ml at 6 h, and 1759 18 pg/ml at 24 h (Fig. 1). LPS stimulated the production of 5159 151 pg/ml of TNFa in 12j cell culture supernatants at 6 h, and 7859150 pg/ml at 24 h. After 6 h of stimulation with BeSO4, IFN-g and IFN-g plus BeSO4, 12j cells did not produce significant amounts of TNFa. After 24 h in culture, Be-stimulated 12j cell culture supernatants had 7249 47 pg/ml, IFN-g
H. Hamada et al. / Toxicology 143 (2000) 249–261
253
Fig. 2. LPS-induced NF-kB/IkBa activation in 12j cells showing a 30 min decrease in cytoplasmic IkBa with an increase in nuclear NF-kB. Western immunoblot analysis of NF-kB levels in the nuclei, and IkBa in the cytoplasm of unstimulated, Be-stimulated (100 mM) and LPS-stimulated (1 mg/ml) 12j cells between 0 and 180 min.
stimulated supernatants had 2849 31 pg/ml and IFN-g plus Be-stimulated supernatants had 11999225 pg/ml TNF-a (P 50.05). The data show that Be, LPS, IFN-g and IFN-g plus Be trigger the production of TNF-a by 12j cells, however, with disparate kinetics of cytokine production. LPS stimulated the early, 6 h, production of TNF-a whereas Be, IFN-g and IFN-g plus Be stimulated the late, 24 h, production of TNF-a by 12j cells.
3.2. LPS-stimulation acti6ates the NF-kB/IkBa family of transcription factors in 12j cells Western immunoblot analysis was used to compare the activation of cytoplasmic NF-kB/IkBa and NF-kB nuclear translocation in unstimulated, Be-stimulated and LPS-stimulated 12j cells. There were no changes in the constitutive levels of NFkB or IkBa in either unstimulated or Be-stimulated 12j cell nuclear or cytoplasmic extracts at the measured times (Fig. 2). LPS activated the NF-kB/IkBa family of transcription factors in 12j cells with the degradation of IkBa in the cytoplasm and the nuclear translocation of NF-kB after 30 min of stimulation. NF-kB levels were elevated in LPS-stimulated 12j cell nuclear extracts from 30 to 180 min. By comparison, Bestimulation failed to alter cytoplasmic or nuclear
NF-kB/IkBa levels in 12j cells at any of the measured intervals.
3.3. LPS-stimulation increases le6els of NF-kB and bzip transcription factors in 12j cells The electrophoretic mobility shift assay (EMSA) was used to measure levels of transcription factors at various times in nuclear extracts of unstimulated, Be-stimulated and LPS-stimulated 12j cells. NF-kB levels were increased in 12j cell nuclei after 30 min of LPS-stimulation, but not in unstimulated or Be-stimulated 12j cell nuclei (Fig. 3). After 3 h, LPS also stimulated an increase in the amounts of both AP-1 and CREB transcription factors in 12j cell nuclei. LPS stimulated an early pattern of transcription factor activation in 12j cell nuclei with increased NF-kB at 30 min, increased bzip transcription factors at 3 h, and TNF-a protein production at 6 h. In contrast, Be did not alter the levels of either NF-kB or bzip transcription factors in 12j cell nuclei at any of the measured times. Antibody super-shift EMSA, using antibodies for Jun and Fos family members of AP-1 transcription factors showed that 3 h after LPS-stimulation only 12j cell nuclei contained increased amounts of Jun B, Jun D and Fos family transcription factors (Fig. 4).
254
H. Hamada et al. / Toxicology 143 (2000) 249–261
3.4. Recombinant murine (rMu) -IFN-g co-stimulation conser6es increased NF-kB and Jun B transcription factors in Be-stimulated 12j cell nuclei The EMSA was used to measure levels of transcription factors at various times, in nuclear extracts of unstimulated, Be, rMu-IFN-g and rMu-IFN-g plus Be-stimulated 12j cells. 12j cell nuclei stimulated with rMu-IFN-g or rMu-IFN-g
plus Be, but not unstimulated or Be-stimulated 12j cell nuclei, contained increased amounts of NF-kB at 3 h (Fig. 5). After 15 h, rMu-IFN-g and rMu-IFN-g plus Be-stimulated 12j cell nuclei also showed increased amounts of AP-1 and CREB transcription factors. Antibody super-shift EMSA, using antibodies for Jun and Fos family members of AP-1 transcription factors showed that at 15 h following rMu-INF-g 12j cell nuclei contained increased
Fig. 3. LPS-stimulation, but not Be-stimulation, increases levels of NF-kB at 30 min, and bzip transcription factors AP-1 and CREB at 3 h, in 12j cell nuclei. EMSA of (A) NF-kB, (B) AP-1 and (C) CREB levels (32P counts) in unstimulated (U =open), Be-stimulated (100 mM BeSO4, B = black) and LPS-stimulated (1 mg/ml; L, dots) 12j cell nuclei at 0, 0.5 and 3 h. N, negative control; C, cold competition control.
H. Hamada et al. / Toxicology 143 (2000) 249–261
255
Fig. 4. LPS-stimulation, but not Be-stimulation, increases levels of Jun B, Jun D and Fos family transcription factors at 15 h in 12j cell nuclei. Antibody super-shift EMSA of (A) AP-1 levels (32P counts) in (B) unstimulated (open), Be-stimulated (100 mM BeSO4, black), and LPS-stimulated (1 mg/ml, stripe) 12j cell nuclei at 3 h. Nuclear extract binding reactions were performed using antibodies against: (1) negative control (no antibody); (2) Jun family; (3) c-Jun; (4) Jun B; (5) Jun D; (6) Fos family.
amounts of Jun and Fos family transcription factors (Fig. 6). Recombinant-Mu-IFN-g plus Bestimulated 12j cell nuclei had increased amounts of only Jun B. Thus, the co-stimulation of 12j cells with Be plus IFN-g increased the amounts of both NF-kB and Jun B nuclear transcription factors, not found in 12j cells stimulated with Be alone. These results correlated with the production of higher levels of TNF-a protein at 24 h.
3.5. Be-stimulation does not alter le6els of transcription factors that bind to the mouse TNF-a promoter region in 12j cell nuclei A segment of the mouse TNF-a promoter region, spanning a DNA sequence from +1 to − 800 bp, contains binding sites for NF-kB (kB1 at −587, kB2 at −212, and kB3 at − 97), Sp1 (at −172 and −52), Egr-1 ( −172), Ets ( − 116), CRE (− 106), AP-1 (−66) and AP-2 (−36)
(Tracey, 1998). We used the EMSA to measure the nuclear levels of all these transcription factors, and others, in Be, LPS-, rMu-IFN-g and rMuIFN-g plus Be-stimulated 12j cell nuclei at 0–24 h. Our results are summarized in Table 1. We observed that Be-stimulation failed to increase levels of the transcription factors at any interval in 12j cell nuclei. In comparison, LPS and rMuIFN-g stimulation increased the levels of NF-kB, AP-1 and CREB in 12j cell nuclei. RecombinantMu-IFN-g stimulation also increased levels of Oct-1 transcription factor in 12j cell nuclei. Interestingly, co-stimulation of Be-stimulated 12j cells with rMu-IFN-g conserved the increased levels of NF-kB and bzip transcription factors, especially Jun B, in 12j cell nuclei. Based on the data, Fig. 7 summarizes the various early versus late pathways of TNF-a-related nuclear transcription factor expression and TNFa synthesis in 12j cells triggered by LPS, Be,
256
H. Hamada et al. / Toxicology 143 (2000) 249–261
rMu-IFN-g and rMu-IFN-g +Be. LPS stimulates an early increase in 12j cell NF-kB and bzip transcription factors with an increase in Jun, especially Jun B and Jun D, and Fos family members. TNF-a protein is detected by 6 h. IFN-g stimulates a similar pattern of nuclear transcription factors and TNF-a protein production, but at a later time in comparison to LPS. BeSO4 fails to increase the levels of any transcription factors within 12j cell nuclei, yet TNF-a protein is produced by 24 h. Be-stimulated 12j cells co-stimulated with IFN-g display a similar late pattern of NF-kB activation, but only Jun B increased. This
pattern of nuclear transcription factors was associated with the production of very high TNF-a protein levels by 12j cells.
4. Discussion The mechanism by which the light-weight metal Be stimulates TNF-a synthesis in CBD/BAL cells (Tinkle et al., 1997; Tinkle and Newman, 1997), and in a mouse hybrid macrophage cell line (Sawyer et al., 1998) is unknown. Because 12j cells mimic the response of Be-stimulated CBD/BAL
Fig. 5. rMu-IFN-g and rMu-IFN-g+Be stimulation increase NF-kB levels at 3 h, and AP-1 and CREB levels at 15 h, in the nuclei of 12j cells. EMSA of (A) NF-kB, (B) AP-1 and (C) CREB levels (32P counts) in unstimulated (U = open), Be-stimulated (100 mM BeSO4; B, black), rMu-IFN-g stimulated (10 U; I,dots), and Be (100 mM BeSO4) +rMu-IFN-g (10U) stimulated (IB, stripe) 12j cell nuclei at 0, 3 and 15 h. N, negative control; C, cold competition control.
H. Hamada et al. / Toxicology 143 (2000) 249–261
257
Fig. 6. rMu-IFN-g (Jun B and Jun D) and rMu-IFN-g+ Be stimulation (Jun B only) increase levels of only Jun family transcription factors at 15 h in 12j cell nuclei. Antibody super-shift EMSA of (A) AP-1 levels (32P counts) in (B) unstimulated (open), Be-stimulated (100 mM BeSO4, black), rMu-IFN-g stimulated (10U, dots) and Be (100 mM BeSO4) +rMu-IFN-g (10U) stimulated (stripe) 12j cell nuclei at 15 h. Nuclear extract binding reactions were performed using antibodies against: (1) negative control (no antibody); (2) Jun family; (3) c-Jun; (4) Jun B; (5) Jun D; (6) Fos family.
cells for the production of TNF-a we used them as an experimental tool to test the hypothesis that Be might stimulate TNF-a production via the same nuclear transcription factor pathways by which LPS and IFN-g trigger TNF-a synthesis. We found that LPS, rMu-IFN-g, Be and combined rMu-IFN-g + Be, all trigger 12j TNF-a production, but with different production kinetics. This suggested the possibility that Be might stimulate 12j TNF-a synthesis along pathways known to be elicited by LPS or IFN-g stimulation (Bach et al., 1997; Boehm et al., 1997; Yao et al., 1997; Amura et al., 1998; Leonard and O’Shea, 1998; Stark et al., 1998; Tracey, 1998).
We were able to identify both an early and a late pattern of response in 12j cells. LPS stimulated peak 12j cell TNF-a production early (6 h) whereas IFN-g, Be and IFN-g + Be stimulated peak 12j cell TNF-a late (24 h). LPS stimulated the activation of cytoplasmic NF-kB/IkBa with IkBa degradation in the cytoplasm and nuclear translocation of NF-kB after 30 min followed by an increase in the levels of AP-1 and CREB at 3 h. We found increased levels of both Jun and Fos family members of AP-1 in the nuclei of 12j cells after 3 h. Our result, showing LPS-stimulated NF-kB/IkBa and AP-1 activation in 12j cells, is in keeping with previously published reports using
H. Hamada et al. / Toxicology 143 (2000) 249–261
258
different macrophage models (Yao et al., 1997; Amura et al., 1998; Delpedro et al., 1998; Tracey, 1998). IFN-g stimulated 12j cells up regulated TNF-a protein production late (24 h), with an increase in nuclear NF-kB at 3 h and an increase Table 1 Be-stimulation fails to alter TNF-a promoter region transcription factor levels in 12j cells Transcription factor
NF-kB AP-1 CREB C/EBP AP-2 Sp1 Egr-1 Ets Oct-1 NF-Y
12j cell stimulus Be
LPS
IFN-g
IFN-g+Be
− − − − − − − − − −
+ + + − − − − − − −
+ + + − − − − − + −
+ + + − − − − − − −
Fig. 7. (A) LPS, (B) IFN-g (C) Be and (D) IFN-g+ Be trigger 12j cell TNF-a production associated with NF-kB and AP-1 transcription factor activation (LPS and IFN-g) or independent of transcription factors (Be). Co-stimulation of 12j cells with IFN-g+Be conserves NF-kB and Jun B up regulation.
in AP-1 and CREB at 15 h. We found increased amounts of Jun B and Jun D transcription factors in IFN-g stimulated 12j cell nuclear extracts. An increase in the levels of Jun family members of AP-1 has not been noted previously in other experimental models of IFN-g stimulated macrophage TNF-a synthesis (Bach et al., 1997; Boehm et al., 1997; Leonard and O’Shea, 1998; Stark et al., 1998). It is unknown whether this response to IFN-g stimulation occurs only in 12j cells, or whether this response is shared by the chronic inflammatory exudate macrophages found in the CBD granuloma. Based on the data in this study, it is reasonable to hypothesize that a similar, IFN-g-induced and AP-1 mediated up-regulation of the TNF-a promoter exists in CBD macrophages, but not macrophages from normal subjects that do not produce Be-stimulated TNFa (Tinkle and Newman, 1997). In contrast, Be stimulation of 12j cells failed to alter the levels of any transcription factor tested. Yet, despite a failure to increase the levels of transcription factors Be stimulated the production of TNF-a protein late (24 h). We conclude that Be-stimulated 12j cell TNF-a synthesis cannot be attributed to an increase in nuclear transcription factors known to bind the TNF-a promoter region of DNA. We suggest that Be-stimulated macrophage TNF-a production might occur by a transcription factor-independent pathway. We cannot exclude the possibility that Be induces TNF-a production in 12j cells via a novel nuclear regulating factor or factors. Based on the data we, therefore, reject our initial hypothesis. TNF-a synthesis was not associated with common transcription factor-activation pathways in 12j cells. The combination of IFN-g plus Be stimulated 12j cells to produce high levels of TNF-a. When human CBD/BAL cells are stimulated by similar amounts of Be in vitro, they also produce high levels of both IFN-g and TNF-a (Tinkle et al., 1997; Tinkle and Newman, 1997). It has been hypothesized that Be-stimulated IFN-g produced by Be-sensitized CBD Th1 lymphocytes stimulates the production of CBD macrophage TNF-a (Newman, 1995; Tinkle et al., 1996). Our data suggest that Be may directly stimulate macrophages to produce TNF-a independent of
H. Hamada et al. / Toxicology 143 (2000) 249–261
other cytokine signals. IFN-g derived from Bestimulated Th 1 lymphocytes in CBD might, therefore, account for the observed production of high levels of TNF-a in vitro and Be-stimulated CBD macrophage TNF-a may up regulate IFN-g production (Tinkle and Newman, 1997). In the CBD patient’s lung, establishing such a Be-specific, antigen-driven, cytokine amplification loop may contribute to the formation of granulomatous lesions within the lungs. Co-stimulation of 12j cells with IFN-g plus Be induced an increase in nuclear transcription factors in a pattern similar to that observed using IFN-g stimulation alone, however, only Jun B levels were elevated at 15 h. We detected no increase in the levels of c-Jun, Jun-D or Fos family members, or other bzip nuclear proteins, in combined IFN-g plus Be-stimulated 12j cell nuclei. We conclude that the signal provided by IFN-g stimulation up regulated Jun B in 12j cell nuclei, and this signal is conserved in the presence of Be. Our data show that multiple signals can trigger the production of TNF-a by 12j cells. LPS and IFN-g share in common the ability to stimulate macrophage TNF-a synthesis by ligation of macrophage surface receptors. For example, LPS, in combination with serum lipid binding protein, binds surface CD14 with subsequent NF-kB nuclear translocation (Fenton and Golenbock, 1998), whereas ligation of the IFN-g receptor also up regulates NF-kB (Bach et al., 1997; Boehm et al., 1997). In our study, LPS and IFN-g stimulation shared in common the up regulation of NFkB and its nuclear translocation, followed by an up regulation in the AP-1 family, and the CREB, bzip transcription factors, but with different kinetics. This is contrasted by our observation that Be stimulation did not increase the levels of any nuclear transcription factors known to bind the mouse TNF-a promoter region, yet Be stimulation alone up regulated TNF-a production at 24 h. We do not know how a metal salt might up regulate TNF-a synthesis independent of the known transcription factors but we can envisage several possible mechanisms that could be tested experimentally. Be could bind another macrophage surface receptor, like CD14 or the
259
IFN-g receptor, that stimulates TNF-a synthesis by up regulating a novel, unknown, nuclear transcription factor. For example, c-Jun interacts with NF-kB p50/p65 to synergistically enhance LPS-induced TNF-a (Yao et al., 1997) and c-Jun complexes with ATF-2 and other CREB/ATF proteins to promote transactivation of IFN-g (Penix et al., 1996). We have not evaluated the possibility that IFN-g induces an AP-1 complex with other transcription factors to from a novel TNF-a inducing transcription complex in 12j cells, a mechanism that occurs in other systems (Penix et al., 1996; Horvai et al., 1997; Foletta et al., 1998). Beryllium might interact directly with DNA, or DNA-associated proteins that play a role in transcription factor binding to the TNF-a promoter region. Beryllium interacts directly with nuclear acidic proteins, G proteins and protein kinases (Parker and Steven, 1979; Kaser et al., 1980; Cummings et al., 1982; Bigay et al., 1987). Thus, Be bound to DNA may stabilize a constitutively expressed transcription factor, such as AP1, in a manner that allows the promoter to be continuously turned ‘on.’ This possibility suggests experiments designed to determine if Be, like other divalent cations (Laundon and Griffith, 1987), is able to alter the primary structure of DNA. Last, our studies do not rule out the possibility that Be alters TNF-a mRNA processing. We previously measured increased TNF-a mRNA levels in Be-stimulated 12j cells at 24 h (Sawyer et al., 1998) suggesting that mRNA production is accompanied by its accumulation in 12j cells but in the absence of transcription factor activation. It was interesting to find that IFN-g stimulation up regulated of both NF-kB and Jun B in the presence of Be. At present, we can not explain why both Jun D and Oct-1 were not detected in nuclear extracts of IFN-g plus Be-stimulated 12j cells. It is possible that these two transcription factors, up-regulated in IFN-g stimulated 12j cells, may be down-regulated in the presence of Be. Conservation of this late NF-kB/Jun B pathway in 12j cells, in combination with the putative Be-stimulated transcription factor-independent pathway may explain why such high levels of TNF-a protein production occur in the presence of both cytokine and metal-salt. This is important
260
H. Hamada et al. / Toxicology 143 (2000) 249–261
because the combination of IFN-g and Be may be present within pulmonary lesions of CBD patients where Be, once within tissues, persists for prolonged time periods (Newman, 1995). The data from this study can be used to generate and to test the hypothesis that Be stimulates TNF-a in CBD macrophages by a transcription factor-independent mechanism, by a mechanism involving an IFN-g-induced AP-1 up-regulation, or by novel transcription factor complexes. The fact that Be is able to up regulate the synthesis of these pro-inflammatory cytokines by a pathway that differs from those established for other cytokine stimuli such as LPS and IFN-g, may be of emerging importance since the industrial use of, and environmental contamination by, Be continues to increase.
Acknowledgements This research was supported by Grant ES06538-06, NHLBI Specialized Center of Research (SCOR) Grant HL-27353 (LSN), and Grant AI220229. H36.12j cells are named ‘Pixie’ cells posthumously. The author’s wish to thank Dr Sherry D. Flemming and Elizabeth P. Canono (NJMRC, Denver, CO) for advice and technical assistance with 12j cells. Dr James Hagman, Dr William H. Wheat and Dr Kamuda Das (NJMRC, Denver, CO) provided invaluable technical assistance with the EMSA.
References Amura, C.R., Kamei, T., Ito, N., Soares, M.J., Morrison, D.C., 1998. Differential regulation of lipopolysaccharide (LPS) activation pathways in mouse macrophages by LPSbinding proteins. J. Immunol. 161, 2552–2560. Bach, A.E., Aguet, M., Schreiber, R.D., 1997. The IFN-g receptor: a paradigm for cytokine receptor signaling. Ann. Rev. Immunol. 15, 563–591. Bigay, J., Deterre, P., Pfister, C., Chabre, M., 1987. Fluoride complexes of aluminum or beryllium act on G-proteins as reversibly bound analogues of the gamma phosphate of GTP. EMBO J. 6, 2907–2913. Boehm, U., Klamp, T., Groot, M., Howard, J.C., 1997. Cellular responses to interferon-g. Ann. Rev. Immunol. 15, 749 – 795.
Bost, T.W., Riches, D.W.H., Schumacher, B., Carre, P.C., Khan, T.Z., Martinez, J.A., Newman, L.S., 1994. Alveolar macrophages from patients with beryllium disease and sarcoidosis express increaed levels of mRNA for TNF-a and IL-6 but not IL-1b. Am. J. Respir. Cell Mol. Biol. 10, 506 – 513. Canono, B.P., Campbell, P.A., 1992. Production of mouse inflammatory macrophage hybridomas. J. Tissue Cult. Methods 14, 3 – 8. Cummings, B., Kaser, K., Wiggins, G., Ord, M.G., Stocken, L.A., 1982. Beryllium toxicity: the selective inhibition of casine kinase 1. Biochem. J. 208, 141 – 146. Das, K.C., White, C.W., 1997. Activation of NF-kB by antineoplastic agents. J. Biol. Chem. 272, 14914 – 14920. Delpedro, A.D., Barjavel, M.J., Mamdouh, Z., Faure, S., Bakouche, O., 1998. Signal transduction in LPS-activated aged and young monocytes. J. Interferon Cytokine Res. 18, 429 – 437. Fenton, M.J., Golenbock, D.T., 1998. LPS-binding proteins and receptors. J. Leukoc. Biol. 64, 25 – 32. Foletta, V.C., Segal, D.H., Cohen, D.R., 1998. Transcriptional regulation in the immune system: all roads lead to AP-1. J. Leukoc. Biol. 63, 139 – 152. Fontenot, A.P, Kotzin, B.L., Comment, C.E., Newman, L.S., 1998. Expansions of T-cell subsets expressing particular T-cell receptor variable regions in chronic beryllium disease. Am. J. Respir. Cell Mol. Biol. 18, 581 – 589. Hagman, J.D., Doglio, L.T., Hackett, J., Rudin, C.M., Haasch, D., Brinster, R., Storb, U., 1989. Inhibition of immunoglobulin gene rearrangement by the expression of a b2 transgene. J. Exp. Med. 169, 1911 – 1929. Horvai, A.E., Xu, L, Korzus, E., Brard, G., Kalafus, D., Mullen, T.M., Rose, D.W., Rosenfeld, M.G., Glass, C.K., 1997. Nuclear integration of JAK/STAT and Ras/AP-1 signaling by CBP and p300. Proc. Natl. Acad. Sci. USA 94, 1074 – 1079. Kaser, M.R., Ord, M.G., Stoken, L.A., 1980. The effects of beryllium on histone phosphorylation in liver. Biochem. Intern. 1, 148 – 150. Laundon, C.H., Griffith, J.D., 1987. Cationic metals promote sequence-directed DNA bending. Biochemistry 26, 3759 – 3762. Leonard, W.J., O’Shea, J.J., 1998. Jaks and stats: biological implications. Ann. Rev. Immunol. 16, 293 – 322. Newman, L.S., 1995. Current concepts: occupational illness. New Engl. J. Med. 333, 1128 – 1134. Newman, L.S., Lloyd, K., Daniloff, E.A., 1996. The natural history of beryllium sensitization and chronic beryllium disease. Environ. Health Perspect. 104, 937 – 943. Parker, V.H., Steven, C., 1979. Binding of beryllium to nuclear acidic proteins. Chem. Biol. Interact. 26, 167 – 177. Penix, L.A., Sweetser, M.T., Weaver, W.M., Hoeffler, J.P., Kerppola, T.K., Wilson, C.R., 1996. The proximal regulatory element of the interferon-g promoter mediates selective expression in T cells. J. Biol. Chem. 271, 31964 – 31972. Sawyer, R.T., Kittle, L.A., Campbell, P.A., Newman, L.S., 1998. Direct stimulation of TNF-a by beryllium in an
H. Hamada et al. / Toxicology 143 (2000) 249–261 inflammatory hybrid mouse macrophage cell line. Am. J. Respir. Crit. Care Med. 157, A32. Stark, G.R., Kerr, J.M., Williams, B.R.G., Silverman, R.H., Schreiber, R.D., 1998. How cells respond to interferons. Ann. Rev. Biochem. 67, 227–264. Tinkle, S.S., Schwitters, P.W., Newman, L.S., 1996. Cytokine production by bronchoalveolar lavage cells in chronic beryllium disease. Environ. Health Perspect. 104, 969–997. Tinkle, S.S., Kittle, L.A., Schumacher, B.A., Newman, L.S., 1997. Beryllium induces IL-2 and IFN-g in berylliosis. J. Immunol. 158, 518 – 526.
.
261
Tinkle, S.S., Newman, L.S., 1997. Beryllium-stimulated release of tumor necrosis factor-a, interleukin-6, and their soluble receptors in chronic beryllium disease. Am. J. Respir. Crit. Care Med. 156, 1884 – 1891. Tracey, K.J., 1998. Tumor necrosis factor-alpha. In: Thomson, A. (Ed.), The Cytokine Handbook, 3rd ed. Academic Press, NY, pp. 289 – 304. Yao, J., Mackman, N., Edgington, T.S., Fan, S.T., 1997. Lipopolysaccharide induction of the tumor necrosis factora promoter in human monocytic cells. J. Biol. Chem. 272, 17795 – 17801.
Beryllium-stimulation does not activate transcription factors in a mouse hybrid macrophage cell line Hironobu Hamada a,*, Richard T. Sawyer a,b, Lori A. Kittle a, Lee S. Newman a,b,c a Di6ision of En6ironmental and Occupational Health Sciences, Department of Medicine, National Jewish Medical and Research Center, 1400 Jackson Street, Den6er, CO 80206, USA b Di6ision of Pulmonary and Critical Care Medicine, Department of Medicine, Uni6ersity of Colorado Health Sciences Center, Den6er, CO USA c Department of Pre6enti6e Medicine and Biometrics, Uni6ersity of Colorado Health Sciences Center, Den6er, CO, USA
Received 14 August 1999; accepted 22 October 1999
Abstract We tested the hypothesis that beryllium (Be) could stimulate H36.12j cell (12j) TNF-a production by transcription factor-mediated pathways similar to those induced by either LPS- or IFN-g stimulation. Unstimulated 12j cells produce constitutive levels of TNF-a (175918 pg/ml, mean9 SEM) detected by ELISA of culture supernatants after 24 h. Beryllium-stimulated (100 mM BeSO4) 12j cell TNF-a (724947 pg/ml) was observed after 24 h while LPS-stimulated (1 mg/ml) TNF-a (5159151 pg/ml) after 6 h. Recombinant-Mu-IFN-g (10 U) stimulated 12j cell TNF-a at lower levels (284 931 pg/ml) while rMu-IFN-g+Be-stimulated 12j cells produced 1195 9225 pg/ml TNF-a. Constitutive levels of transcription factors were observed in unstimulated 12j cell nuclei. In LPS-stimulated 12j cells IkBa was degraded in the cytoplasm and increased levels of NF-kB were found in nuclei after 30 min. After 3 h there were increased levels of AP-1 and CREB, with increased amounts of Fos family, Jun B and Jun D transcription factors. In contrast, Be-stimulation failed to increase the levels of any transcription factor tested, NF-kB, AP-1, AP-2, CREB, C/EBP, Sp-1, Egr-1, Ets, NF-Y or Oct-1, in 12j cells. A pattern of increased transcription factors, similar to that observed for LPS-stimulation, was found in 12j cell nuclei after stimulation with rMu-IFN-g. However, NF-kB was increased at 3 h while AP-1 (Jun B and Jun D) and CREB were increased at 15 h. Co-stimulation of 12j cells with rMu-IFN-g+ Be increased the levels of NF-kB in 12j cell nuclei at 3 h, and the levels of AP-1 and CREB at 15 h, however, only Jun B was increased. Our data show 12j cell TNF-a production was associated with increased levels of transcription factors present in nuclei with disparate kinetics and patterns of expression depending on the trigger. We reject our initial hypothesis and conclude that Be-stimulation signals 12j cell TNF-a synthesis via a transcription factor-independent pathway. Beryllium may induce novel pathways of macrophage cytokine gene regulation. © 2000 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Beryllium; Chronic beryllium disease; Transcription factors; Macrophages; Hybrid macrophages; NF-kB/IkBa; AP-1; CREB; Bzip proteins; Lipopolysaccharide; Interferon-g; Tumor necrosis factor-a; TNF-a promoter * Corresponding author. Tel.: + 1-303-398-1167; fax: + 1-303-398-1851. E-mail address: [email protected] (H. Hamada) 0300-483X/00/$ - see front matter © 2000 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 3 0 0 - 4 8 3 X ( 9 9 ) 0 0 1 8 3 - 3
250
H. Hamada et al. / Toxicology 143 (2000) 249–261
1. Introduction Chronic beryllium disease (CBD) results from exposure to environmental beryllium (Be), which occurs most often in an industrial setting (Newman, 1995). In vitro, Be-stimulated bronchoalveolar lavage (BAL) cells obtained from CBD subjects produce high levels of pro-inflammatory cytokines. These cytokines are believed to participate in the proliferation of Be-sensitized CD4+ T cells and in the pathogenesis of the CBD granuloma (Newman et al., 1996; Fontenot et al., 1998). For example, Be-stimulated CBD/BAL cells produce high levels of IFN-g ( \ 10 ng/ml) and transient levels of IL-2, but do not produce IL-4 or IL-7 (Tinkle et al., 1996, 1997). Be-stimulated CBD/BAL cells also produce high levels of IL-6 and TNF-a (Tinkle and Newman, 1997). Bost et al. (1994) showed that in comparison to normal subject BAL macrophages, isolated CBD/BAL macrophages expressed higher levels of TNF-a mRNA. The mechanism by which Be stimulates CBD/BAL cells to produce cytokines is unknown. At present, there are no in vitro models that use Be-responding cells or cell lines, that would permit the generation of testable hypothesis directly related to answering questions about possible mechanisms of Be-induced cytokine up-regulation in CBD. Our own studies (Bost et al., 1994; Tinkle et al., 1996, 1997; Tinkle and Newman 1997), designed to investigate mechanisms of Be-stimulated cytokine synthesis, have been limited by the availability of specific CBD cell types. However, we have identified a macrophage cell line that mimics the production of Be-induced macrophage TNF-a synthesis by CBD/BAL cells. H36.12j ‘Pixie’ cells (12j) are a hybrid macrophage cell line (Canono and Campbell, 1992; Sawyer et al., 1998). We hypothesized that Be stimulates 12j cell TNF-a production by transcription factor mediated pathways similar to those induced by either LPS- or IFN-g stimulation. We tested this hypothesis by measuring the levels of 12j cell transcription factors, known to be involved in binding the TNF-a promoter region (+1 to −1500 bp) (Tracey, 1998). These included the NF-kB/IkBa family of transcription
factors, bzip transcription factors (AP-1, Fos/Jun family, CREB, C/EBP), AP-2, Sp1, Ets and Egr1. Using 12j cells as a model, we observed disparate transcription factor pathways by which different triggers increase TNF-a synthesis. Our data shows that Be stimulation may induce macrophage cytokine production independent of known nuclear transcription factor regulation. Of interest, co-stimulation with IFN-g plus Be stimulates TNF-a production by an AP-1 mediated mechanism not previously described.
2. Materials and methods
2.1. Chemicals and reagents A filter-sterilized, aqueous stock solution, made in pyrogen-free, sterile water (Abbot Lab. Chicago, IL) of 1 mM BeSO4·4H2O (BE, Fisher Sci., Springfield, NJ) was held at 4°C prior to use. Sterile transfers were made into 12j cell cultures at zero time, to a final concentration of 100 mM. This amount of BeSO4 was selected based on a preliminary dose–response experiment showing that at 48 h, 12j cells stimulated with 100 mM BeSO4 produced 3029 41 pg/ml (mean9SEM) TNF-a, 163 9 2 pg/ml TNF-a at 50 mM BeSO4, 619 2 pg/ml TNF-a at 25 mM BeSO4, and 60 9 2 pg/ml TNF-a at 5 mM BeSO4. Thus, 100 mM BeSO4-stimulation induced maximal TNF-a production by 12j cells. Previous study also shows that maximal TNF-a levels are produced by CBD/BAL cells stimulated with 100 mM BeSO4 (Tinkle and Newman, 1997). 12j cells and culture supernatants were harvested either immediately (zero time) or at the indicated intervals up to 24 h. As controls, 12j cells were unstimulated, stimulated with 1 mg/ml of Salmonella typhimurium lipopolysaccharide (LPS, Sigma Chem., St. Louis, MO) or 10 U of rMuIFN-g (Genzyme, Boston, MA). We tested our beryllium-salt stimulatory reagents using the Limulus amebocyte lysate assay (Associatesof Cape Cod, Woods Hole, MA), which has a limit of detection of 0.025 ng of LPS. All reagents and glassware used in this study were pyrogen free. The
H. Hamada et al. / Toxicology 143 (2000) 249–261
calf serum contained B0.05 ng/ml of LPS (BioWhittiker).
2.2. Macrophage cell lines and culti6ation H36.12j ‘Pixie’ cells are clonally derived, hybrid precursor macrophages, derived from the fusion of drug selected P388D.1 (DBA/2, H2d) macrophages with percoll gradient purified, proteose peptone elicited macrophages obtained from C57Bl/6N (H2b) mice (Canono and Campbell, 1992). 12j cells were negative for the presence of Mycoplasma species epitopes by PCR analysis (Stratagene, LaJolla, CA) and have been deposited with the ATCC (Rockville, MD). For all studies, 12j cells were enumerated by hemocytometer and viability was greater than 90% at the initiation of each experiment. 12j cells were cultivated in Dulbecco’s Modified Eagle’s Medium (BioWhittiker, Walkerville, MD) supplemented with 10% heat inactivated calf serum, 200 mM L-glutamine, 10 000 U/ml penicillin G and 10 000 mg/ml streptomycin sulfate. 12j cells were cultivated at 1×105 cells/200 ml/well in 96 well round bottom culture dishes (Corning, NY) at 37°C in a humidified atmosphere containing 5% CO2. 12j cells were harvested by aspiration and concentrated by microcentrifugation. Culture supernatants were held at −80°C until use.
2.3. Quantification of cytokine protein Cytokine protein was measured with a solidphase sandwich ELISA (R&D Sys., Minneapolis, MN) with a reported sensitivity of 4 pg/ml and data are reported as the mean 9 SEM pg/ml from four wells, using six separate experimental determinations.
2.4. Western immunoblot analysis At the indicated times 1×107 12j cells/condition were harvested and nuclear and cytoplasmic protein extracts prepared by the method of Das and White (1997) and Hagman et al. (1989). Assay (Pierce, Rockford, IL) for total protein indicated a yield of : 500 mg of protein/1 ×107 12j cells. Nuclear and cytoplasmic levels of NF-
251
kB/IkBa were determined by Western immunoblot analysis. Western immunoblots were prepared by transfer of 12j cell extract proteins separated by 10% SDS-PAGE. Immunoblots were incubated with primary anti-transcription factor antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) for NF-kB, p65 or p50, and IkBa. Data obtained using the anti-p50 NF-kB antibody are reported. Identical results were obtained in Western immunoblots probed with the anti-p65 NF-kB antibody (not shown). Binding of antitranscription factor antibody was detected using a second anti-globulin coupled to alkaline phosphatase. Chemifluorescence (ECF Western Blotting System, Amersham, UK) was detected using the Storm 840 PhosphorImager and ImageQuant analysis program (Molecular Dynamics, Sunnyvale, CA).
2.5. Electrophoretic mobility shift assay (EMSA) The EMSA was performed using unstimulated, LPS-stimulated, rMu-IFN-g-stimulated, BeSO4stimulated, and rMu-IFN-g + BeSO4-stimulated 12j cell nuclear extracts. In brief, at the indicated intervals, 12j cells were concentrated by centrifugation and suspended at 1× 107 cells in 1 ml of buffer A (10 mM HEPES (Sigma Chemical, St. Louis, MO) pH 7.9; 300 mM sucrose; 10 mM KCl; 1.5 mM MgCl2; 1 mM ditiothreitol (DTT, Sigma) and 26 ml of a 5 mg of PMSF/ml of isopropanol stock, 0.76 mM final). The 12j cells were concentrated by microcentrifugation at 1200 rpm for 3 min, the supernatant decanted and the cell suspended in 900 ml of buffer A plus 100 ml of NP40 (0.5% final, Sigma) and 0.76 mM PMSF. The suspended cells were vortexed for 1 min at the high setting. A 10 ml aliquot was removed and examined microscopically for the formation of cell nuclei. The nuclei were concentrated by microcentrifugation at 14 000 rpm for 15 min. The supernatant was removed using a micropipet, and 100 ml aliquots of the cytoplasmic extract were held at − 80°C until use. The nuclear pellet was suspended in 150 ml of buffer C (20 mM HEPES, pH 7.9; 300 mM sucrose; 420 mM NaCl; 1.5 mM MgCl2; 0.76 mM PMSF). The nuclei were lysed by vortex for 15 min and the debris concentrated
252
H. Hamada et al. / Toxicology 143 (2000) 249–261
by microcentrifugation at 14 000 rpm for 5 min. The nuclear extract was transferred to a tube containing 15 ml of glycerol, mixed and stored at − 80°C until use. Nuclear extract binding reactions were performed as described (Das and White, 1997) in the absence (negative control) and presence of double stranded oligonucleotides (Promega, Madison, WI) containing consensus binding sequences for Sp-1, AP-1, AP-2, NF-kB, Oct-1, CREB, C/EBP, NF-Y, Ets, and Egr-1. Probes were end-labeled with 32P-ATP. Cold competition, control reactions were performed using 32 P-labeled oligonucleotide probe plus 100-fold excess of the unlabeled probe. As an additional control, binding reactions were performed in the absence of nuclear extract but with the addition of recombinant transcription factor (2 ng/reaction, Santa Cruz). Levels (32P counts) of transcription factors were measured using the Storm 840 PhosphorImager and ImageQuant analysis program at the indicated times (0 – 24 h) after stimulation.
2.6. Antibody supershift EMSA The EMSA was performed as above, in the absence and presence of antibodies directed against nuclear transcription factors. Antibodies directed against transcription factors were purchased from Santa Cruz Biotechnologies, Inc.
(Santa Cruz, CA) and used according to the manufacturer’s specifications. Antibodies were added to binding buffer 20 min before the addition of labeled-oligonucleotide probes. We calculated data for anti-transcription factor antibody+ transcription factor(s) that were super-shifted by subtracting the remaining bands after antibody super-shift EMSA from the negative control transcription factor band (formed by a reaction between 32P-oligonucleotide+ nuclear extract, but without added anti-transcription factor antibody). Antibody-transcription factor complexes are delayed in their migration through the EMSA gel, permitting specific identification and quantification using the Storm 840 PhosphorImager.
2.7. Statistical analysis In order to validate our analysis of changes in the levels of cytoplasmic and nuclear transcription factors Western immunoblot, EMSA and EMSA supershift analysis were performed using extracts from three separate experiments. The figures presented, therefore, are representative of these repeated experiments. Analysis of variance was performed using the Tukey Kramer test (Tinkle et al., 1997; Tinkle and Newman, 1997). Positive values for pairs of means were considered significantly different at a P value of 5 0.05.
3. Results
3.1. The kinetics of 12j cell TNF-a production
Fig. 1. The production of TNF-a (mean9 SEM pg/ml) at 6 h (open) and 24 h (black) by unstimulated, Be-stimulated (100 mM), LPS-stimulated (1 mg/ml), IFN-g stimulated (10U) and IFN-g (10U) +Be (100 mM)-stimulated 12j cells as measured by ELISA of culture supernatants. *P5 0.05 at 24 h for TNF-a levels from Be, IFN-g and IFN-g+ Be stimulated 12j cells, and at both 6 and 24 h for LPS-stimulated 12j cells.
TNF-a levels (mean9SEM pg/ml) were measured by ELISA of 12j culture supernatants at 6 and 24 h. Unstimulated 12j cells produced constitutive levels of TNF-a with 8793 pg/ml at 6 h, and 1759 18 pg/ml at 24 h (Fig. 1). LPS stimulated the production of 5159 151 pg/ml of TNFa in 12j cell culture supernatants at 6 h, and 7859150 pg/ml at 24 h. After 6 h of stimulation with BeSO4, IFN-g and IFN-g plus BeSO4, 12j cells did not produce significant amounts of TNFa. After 24 h in culture, Be-stimulated 12j cell culture supernatants had 7249 47 pg/ml, IFN-g
H. Hamada et al. / Toxicology 143 (2000) 249–261
253
Fig. 2. LPS-induced NF-kB/IkBa activation in 12j cells showing a 30 min decrease in cytoplasmic IkBa with an increase in nuclear NF-kB. Western immunoblot analysis of NF-kB levels in the nuclei, and IkBa in the cytoplasm of unstimulated, Be-stimulated (100 mM) and LPS-stimulated (1 mg/ml) 12j cells between 0 and 180 min.
stimulated supernatants had 2849 31 pg/ml and IFN-g plus Be-stimulated supernatants had 11999225 pg/ml TNF-a (P 50.05). The data show that Be, LPS, IFN-g and IFN-g plus Be trigger the production of TNF-a by 12j cells, however, with disparate kinetics of cytokine production. LPS stimulated the early, 6 h, production of TNF-a whereas Be, IFN-g and IFN-g plus Be stimulated the late, 24 h, production of TNF-a by 12j cells.
3.2. LPS-stimulation acti6ates the NF-kB/IkBa family of transcription factors in 12j cells Western immunoblot analysis was used to compare the activation of cytoplasmic NF-kB/IkBa and NF-kB nuclear translocation in unstimulated, Be-stimulated and LPS-stimulated 12j cells. There were no changes in the constitutive levels of NFkB or IkBa in either unstimulated or Be-stimulated 12j cell nuclear or cytoplasmic extracts at the measured times (Fig. 2). LPS activated the NF-kB/IkBa family of transcription factors in 12j cells with the degradation of IkBa in the cytoplasm and the nuclear translocation of NF-kB after 30 min of stimulation. NF-kB levels were elevated in LPS-stimulated 12j cell nuclear extracts from 30 to 180 min. By comparison, Bestimulation failed to alter cytoplasmic or nuclear
NF-kB/IkBa levels in 12j cells at any of the measured intervals.
3.3. LPS-stimulation increases le6els of NF-kB and bzip transcription factors in 12j cells The electrophoretic mobility shift assay (EMSA) was used to measure levels of transcription factors at various times in nuclear extracts of unstimulated, Be-stimulated and LPS-stimulated 12j cells. NF-kB levels were increased in 12j cell nuclei after 30 min of LPS-stimulation, but not in unstimulated or Be-stimulated 12j cell nuclei (Fig. 3). After 3 h, LPS also stimulated an increase in the amounts of both AP-1 and CREB transcription factors in 12j cell nuclei. LPS stimulated an early pattern of transcription factor activation in 12j cell nuclei with increased NF-kB at 30 min, increased bzip transcription factors at 3 h, and TNF-a protein production at 6 h. In contrast, Be did not alter the levels of either NF-kB or bzip transcription factors in 12j cell nuclei at any of the measured times. Antibody super-shift EMSA, using antibodies for Jun and Fos family members of AP-1 transcription factors showed that 3 h after LPS-stimulation only 12j cell nuclei contained increased amounts of Jun B, Jun D and Fos family transcription factors (Fig. 4).
254
H. Hamada et al. / Toxicology 143 (2000) 249–261
3.4. Recombinant murine (rMu) -IFN-g co-stimulation conser6es increased NF-kB and Jun B transcription factors in Be-stimulated 12j cell nuclei The EMSA was used to measure levels of transcription factors at various times, in nuclear extracts of unstimulated, Be, rMu-IFN-g and rMu-IFN-g plus Be-stimulated 12j cells. 12j cell nuclei stimulated with rMu-IFN-g or rMu-IFN-g
plus Be, but not unstimulated or Be-stimulated 12j cell nuclei, contained increased amounts of NF-kB at 3 h (Fig. 5). After 15 h, rMu-IFN-g and rMu-IFN-g plus Be-stimulated 12j cell nuclei also showed increased amounts of AP-1 and CREB transcription factors. Antibody super-shift EMSA, using antibodies for Jun and Fos family members of AP-1 transcription factors showed that at 15 h following rMu-INF-g 12j cell nuclei contained increased
Fig. 3. LPS-stimulation, but not Be-stimulation, increases levels of NF-kB at 30 min, and bzip transcription factors AP-1 and CREB at 3 h, in 12j cell nuclei. EMSA of (A) NF-kB, (B) AP-1 and (C) CREB levels (32P counts) in unstimulated (U =open), Be-stimulated (100 mM BeSO4, B = black) and LPS-stimulated (1 mg/ml; L, dots) 12j cell nuclei at 0, 0.5 and 3 h. N, negative control; C, cold competition control.
H. Hamada et al. / Toxicology 143 (2000) 249–261
255
Fig. 4. LPS-stimulation, but not Be-stimulation, increases levels of Jun B, Jun D and Fos family transcription factors at 15 h in 12j cell nuclei. Antibody super-shift EMSA of (A) AP-1 levels (32P counts) in (B) unstimulated (open), Be-stimulated (100 mM BeSO4, black), and LPS-stimulated (1 mg/ml, stripe) 12j cell nuclei at 3 h. Nuclear extract binding reactions were performed using antibodies against: (1) negative control (no antibody); (2) Jun family; (3) c-Jun; (4) Jun B; (5) Jun D; (6) Fos family.
amounts of Jun and Fos family transcription factors (Fig. 6). Recombinant-Mu-IFN-g plus Bestimulated 12j cell nuclei had increased amounts of only Jun B. Thus, the co-stimulation of 12j cells with Be plus IFN-g increased the amounts of both NF-kB and Jun B nuclear transcription factors, not found in 12j cells stimulated with Be alone. These results correlated with the production of higher levels of TNF-a protein at 24 h.
3.5. Be-stimulation does not alter le6els of transcription factors that bind to the mouse TNF-a promoter region in 12j cell nuclei A segment of the mouse TNF-a promoter region, spanning a DNA sequence from +1 to − 800 bp, contains binding sites for NF-kB (kB1 at −587, kB2 at −212, and kB3 at − 97), Sp1 (at −172 and −52), Egr-1 ( −172), Ets ( − 116), CRE (− 106), AP-1 (−66) and AP-2 (−36)
(Tracey, 1998). We used the EMSA to measure the nuclear levels of all these transcription factors, and others, in Be, LPS-, rMu-IFN-g and rMuIFN-g plus Be-stimulated 12j cell nuclei at 0–24 h. Our results are summarized in Table 1. We observed that Be-stimulation failed to increase levels of the transcription factors at any interval in 12j cell nuclei. In comparison, LPS and rMuIFN-g stimulation increased the levels of NF-kB, AP-1 and CREB in 12j cell nuclei. RecombinantMu-IFN-g stimulation also increased levels of Oct-1 transcription factor in 12j cell nuclei. Interestingly, co-stimulation of Be-stimulated 12j cells with rMu-IFN-g conserved the increased levels of NF-kB and bzip transcription factors, especially Jun B, in 12j cell nuclei. Based on the data, Fig. 7 summarizes the various early versus late pathways of TNF-a-related nuclear transcription factor expression and TNFa synthesis in 12j cells triggered by LPS, Be,
256
H. Hamada et al. / Toxicology 143 (2000) 249–261
rMu-IFN-g and rMu-IFN-g +Be. LPS stimulates an early increase in 12j cell NF-kB and bzip transcription factors with an increase in Jun, especially Jun B and Jun D, and Fos family members. TNF-a protein is detected by 6 h. IFN-g stimulates a similar pattern of nuclear transcription factors and TNF-a protein production, but at a later time in comparison to LPS. BeSO4 fails to increase the levels of any transcription factors within 12j cell nuclei, yet TNF-a protein is produced by 24 h. Be-stimulated 12j cells co-stimulated with IFN-g display a similar late pattern of NF-kB activation, but only Jun B increased. This
pattern of nuclear transcription factors was associated with the production of very high TNF-a protein levels by 12j cells.
4. Discussion The mechanism by which the light-weight metal Be stimulates TNF-a synthesis in CBD/BAL cells (Tinkle et al., 1997; Tinkle and Newman, 1997), and in a mouse hybrid macrophage cell line (Sawyer et al., 1998) is unknown. Because 12j cells mimic the response of Be-stimulated CBD/BAL
Fig. 5. rMu-IFN-g and rMu-IFN-g+Be stimulation increase NF-kB levels at 3 h, and AP-1 and CREB levels at 15 h, in the nuclei of 12j cells. EMSA of (A) NF-kB, (B) AP-1 and (C) CREB levels (32P counts) in unstimulated (U = open), Be-stimulated (100 mM BeSO4; B, black), rMu-IFN-g stimulated (10 U; I,dots), and Be (100 mM BeSO4) +rMu-IFN-g (10U) stimulated (IB, stripe) 12j cell nuclei at 0, 3 and 15 h. N, negative control; C, cold competition control.
H. Hamada et al. / Toxicology 143 (2000) 249–261
257
Fig. 6. rMu-IFN-g (Jun B and Jun D) and rMu-IFN-g+ Be stimulation (Jun B only) increase levels of only Jun family transcription factors at 15 h in 12j cell nuclei. Antibody super-shift EMSA of (A) AP-1 levels (32P counts) in (B) unstimulated (open), Be-stimulated (100 mM BeSO4, black), rMu-IFN-g stimulated (10U, dots) and Be (100 mM BeSO4) +rMu-IFN-g (10U) stimulated (stripe) 12j cell nuclei at 15 h. Nuclear extract binding reactions were performed using antibodies against: (1) negative control (no antibody); (2) Jun family; (3) c-Jun; (4) Jun B; (5) Jun D; (6) Fos family.
cells for the production of TNF-a we used them as an experimental tool to test the hypothesis that Be might stimulate TNF-a production via the same nuclear transcription factor pathways by which LPS and IFN-g trigger TNF-a synthesis. We found that LPS, rMu-IFN-g, Be and combined rMu-IFN-g + Be, all trigger 12j TNF-a production, but with different production kinetics. This suggested the possibility that Be might stimulate 12j TNF-a synthesis along pathways known to be elicited by LPS or IFN-g stimulation (Bach et al., 1997; Boehm et al., 1997; Yao et al., 1997; Amura et al., 1998; Leonard and O’Shea, 1998; Stark et al., 1998; Tracey, 1998).
We were able to identify both an early and a late pattern of response in 12j cells. LPS stimulated peak 12j cell TNF-a production early (6 h) whereas IFN-g, Be and IFN-g + Be stimulated peak 12j cell TNF-a late (24 h). LPS stimulated the activation of cytoplasmic NF-kB/IkBa with IkBa degradation in the cytoplasm and nuclear translocation of NF-kB after 30 min followed by an increase in the levels of AP-1 and CREB at 3 h. We found increased levels of both Jun and Fos family members of AP-1 in the nuclei of 12j cells after 3 h. Our result, showing LPS-stimulated NF-kB/IkBa and AP-1 activation in 12j cells, is in keeping with previously published reports using
H. Hamada et al. / Toxicology 143 (2000) 249–261
258
different macrophage models (Yao et al., 1997; Amura et al., 1998; Delpedro et al., 1998; Tracey, 1998). IFN-g stimulated 12j cells up regulated TNF-a protein production late (24 h), with an increase in nuclear NF-kB at 3 h and an increase Table 1 Be-stimulation fails to alter TNF-a promoter region transcription factor levels in 12j cells Transcription factor
NF-kB AP-1 CREB C/EBP AP-2 Sp1 Egr-1 Ets Oct-1 NF-Y
12j cell stimulus Be
LPS
IFN-g
IFN-g+Be
− − − − − − − − − −
+ + + − − − − − − −
+ + + − − − − − + −
+ + + − − − − − − −
Fig. 7. (A) LPS, (B) IFN-g (C) Be and (D) IFN-g+ Be trigger 12j cell TNF-a production associated with NF-kB and AP-1 transcription factor activation (LPS and IFN-g) or independent of transcription factors (Be). Co-stimulation of 12j cells with IFN-g+Be conserves NF-kB and Jun B up regulation.
in AP-1 and CREB at 15 h. We found increased amounts of Jun B and Jun D transcription factors in IFN-g stimulated 12j cell nuclear extracts. An increase in the levels of Jun family members of AP-1 has not been noted previously in other experimental models of IFN-g stimulated macrophage TNF-a synthesis (Bach et al., 1997; Boehm et al., 1997; Leonard and O’Shea, 1998; Stark et al., 1998). It is unknown whether this response to IFN-g stimulation occurs only in 12j cells, or whether this response is shared by the chronic inflammatory exudate macrophages found in the CBD granuloma. Based on the data in this study, it is reasonable to hypothesize that a similar, IFN-g-induced and AP-1 mediated up-regulation of the TNF-a promoter exists in CBD macrophages, but not macrophages from normal subjects that do not produce Be-stimulated TNFa (Tinkle and Newman, 1997). In contrast, Be stimulation of 12j cells failed to alter the levels of any transcription factor tested. Yet, despite a failure to increase the levels of transcription factors Be stimulated the production of TNF-a protein late (24 h). We conclude that Be-stimulated 12j cell TNF-a synthesis cannot be attributed to an increase in nuclear transcription factors known to bind the TNF-a promoter region of DNA. We suggest that Be-stimulated macrophage TNF-a production might occur by a transcription factor-independent pathway. We cannot exclude the possibility that Be induces TNF-a production in 12j cells via a novel nuclear regulating factor or factors. Based on the data we, therefore, reject our initial hypothesis. TNF-a synthesis was not associated with common transcription factor-activation pathways in 12j cells. The combination of IFN-g plus Be stimulated 12j cells to produce high levels of TNF-a. When human CBD/BAL cells are stimulated by similar amounts of Be in vitro, they also produce high levels of both IFN-g and TNF-a (Tinkle et al., 1997; Tinkle and Newman, 1997). It has been hypothesized that Be-stimulated IFN-g produced by Be-sensitized CBD Th1 lymphocytes stimulates the production of CBD macrophage TNF-a (Newman, 1995; Tinkle et al., 1996). Our data suggest that Be may directly stimulate macrophages to produce TNF-a independent of
H. Hamada et al. / Toxicology 143 (2000) 249–261
other cytokine signals. IFN-g derived from Bestimulated Th 1 lymphocytes in CBD might, therefore, account for the observed production of high levels of TNF-a in vitro and Be-stimulated CBD macrophage TNF-a may up regulate IFN-g production (Tinkle and Newman, 1997). In the CBD patient’s lung, establishing such a Be-specific, antigen-driven, cytokine amplification loop may contribute to the formation of granulomatous lesions within the lungs. Co-stimulation of 12j cells with IFN-g plus Be induced an increase in nuclear transcription factors in a pattern similar to that observed using IFN-g stimulation alone, however, only Jun B levels were elevated at 15 h. We detected no increase in the levels of c-Jun, Jun-D or Fos family members, or other bzip nuclear proteins, in combined IFN-g plus Be-stimulated 12j cell nuclei. We conclude that the signal provided by IFN-g stimulation up regulated Jun B in 12j cell nuclei, and this signal is conserved in the presence of Be. Our data show that multiple signals can trigger the production of TNF-a by 12j cells. LPS and IFN-g share in common the ability to stimulate macrophage TNF-a synthesis by ligation of macrophage surface receptors. For example, LPS, in combination with serum lipid binding protein, binds surface CD14 with subsequent NF-kB nuclear translocation (Fenton and Golenbock, 1998), whereas ligation of the IFN-g receptor also up regulates NF-kB (Bach et al., 1997; Boehm et al., 1997). In our study, LPS and IFN-g stimulation shared in common the up regulation of NFkB and its nuclear translocation, followed by an up regulation in the AP-1 family, and the CREB, bzip transcription factors, but with different kinetics. This is contrasted by our observation that Be stimulation did not increase the levels of any nuclear transcription factors known to bind the mouse TNF-a promoter region, yet Be stimulation alone up regulated TNF-a production at 24 h. We do not know how a metal salt might up regulate TNF-a synthesis independent of the known transcription factors but we can envisage several possible mechanisms that could be tested experimentally. Be could bind another macrophage surface receptor, like CD14 or the
259
IFN-g receptor, that stimulates TNF-a synthesis by up regulating a novel, unknown, nuclear transcription factor. For example, c-Jun interacts with NF-kB p50/p65 to synergistically enhance LPS-induced TNF-a (Yao et al., 1997) and c-Jun complexes with ATF-2 and other CREB/ATF proteins to promote transactivation of IFN-g (Penix et al., 1996). We have not evaluated the possibility that IFN-g induces an AP-1 complex with other transcription factors to from a novel TNF-a inducing transcription complex in 12j cells, a mechanism that occurs in other systems (Penix et al., 1996; Horvai et al., 1997; Foletta et al., 1998). Beryllium might interact directly with DNA, or DNA-associated proteins that play a role in transcription factor binding to the TNF-a promoter region. Beryllium interacts directly with nuclear acidic proteins, G proteins and protein kinases (Parker and Steven, 1979; Kaser et al., 1980; Cummings et al., 1982; Bigay et al., 1987). Thus, Be bound to DNA may stabilize a constitutively expressed transcription factor, such as AP1, in a manner that allows the promoter to be continuously turned ‘on.’ This possibility suggests experiments designed to determine if Be, like other divalent cations (Laundon and Griffith, 1987), is able to alter the primary structure of DNA. Last, our studies do not rule out the possibility that Be alters TNF-a mRNA processing. We previously measured increased TNF-a mRNA levels in Be-stimulated 12j cells at 24 h (Sawyer et al., 1998) suggesting that mRNA production is accompanied by its accumulation in 12j cells but in the absence of transcription factor activation. It was interesting to find that IFN-g stimulation up regulated of both NF-kB and Jun B in the presence of Be. At present, we can not explain why both Jun D and Oct-1 were not detected in nuclear extracts of IFN-g plus Be-stimulated 12j cells. It is possible that these two transcription factors, up-regulated in IFN-g stimulated 12j cells, may be down-regulated in the presence of Be. Conservation of this late NF-kB/Jun B pathway in 12j cells, in combination with the putative Be-stimulated transcription factor-independent pathway may explain why such high levels of TNF-a protein production occur in the presence of both cytokine and metal-salt. This is important
260
H. Hamada et al. / Toxicology 143 (2000) 249–261
because the combination of IFN-g and Be may be present within pulmonary lesions of CBD patients where Be, once within tissues, persists for prolonged time periods (Newman, 1995). The data from this study can be used to generate and to test the hypothesis that Be stimulates TNF-a in CBD macrophages by a transcription factor-independent mechanism, by a mechanism involving an IFN-g-induced AP-1 up-regulation, or by novel transcription factor complexes. The fact that Be is able to up regulate the synthesis of these pro-inflammatory cytokines by a pathway that differs from those established for other cytokine stimuli such as LPS and IFN-g, may be of emerging importance since the industrial use of, and environmental contamination by, Be continues to increase.
Acknowledgements This research was supported by Grant ES06538-06, NHLBI Specialized Center of Research (SCOR) Grant HL-27353 (LSN), and Grant AI220229. H36.12j cells are named ‘Pixie’ cells posthumously. The author’s wish to thank Dr Sherry D. Flemming and Elizabeth P. Canono (NJMRC, Denver, CO) for advice and technical assistance with 12j cells. Dr James Hagman, Dr William H. Wheat and Dr Kamuda Das (NJMRC, Denver, CO) provided invaluable technical assistance with the EMSA.
References Amura, C.R., Kamei, T., Ito, N., Soares, M.J., Morrison, D.C., 1998. Differential regulation of lipopolysaccharide (LPS) activation pathways in mouse macrophages by LPSbinding proteins. J. Immunol. 161, 2552–2560. Bach, A.E., Aguet, M., Schreiber, R.D., 1997. The IFN-g receptor: a paradigm for cytokine receptor signaling. Ann. Rev. Immunol. 15, 563–591. Bigay, J., Deterre, P., Pfister, C., Chabre, M., 1987. Fluoride complexes of aluminum or beryllium act on G-proteins as reversibly bound analogues of the gamma phosphate of GTP. EMBO J. 6, 2907–2913. Boehm, U., Klamp, T., Groot, M., Howard, J.C., 1997. Cellular responses to interferon-g. Ann. Rev. Immunol. 15, 749 – 795.
Bost, T.W., Riches, D.W.H., Schumacher, B., Carre, P.C., Khan, T.Z., Martinez, J.A., Newman, L.S., 1994. Alveolar macrophages from patients with beryllium disease and sarcoidosis express increaed levels of mRNA for TNF-a and IL-6 but not IL-1b. Am. J. Respir. Cell Mol. Biol. 10, 506 – 513. Canono, B.P., Campbell, P.A., 1992. Production of mouse inflammatory macrophage hybridomas. J. Tissue Cult. Methods 14, 3 – 8. Cummings, B., Kaser, K., Wiggins, G., Ord, M.G., Stocken, L.A., 1982. Beryllium toxicity: the selective inhibition of casine kinase 1. Biochem. J. 208, 141 – 146. Das, K.C., White, C.W., 1997. Activation of NF-kB by antineoplastic agents. J. Biol. Chem. 272, 14914 – 14920. Delpedro, A.D., Barjavel, M.J., Mamdouh, Z., Faure, S., Bakouche, O., 1998. Signal transduction in LPS-activated aged and young monocytes. J. Interferon Cytokine Res. 18, 429 – 437. Fenton, M.J., Golenbock, D.T., 1998. LPS-binding proteins and receptors. J. Leukoc. Biol. 64, 25 – 32. Foletta, V.C., Segal, D.H., Cohen, D.R., 1998. Transcriptional regulation in the immune system: all roads lead to AP-1. J. Leukoc. Biol. 63, 139 – 152. Fontenot, A.P, Kotzin, B.L., Comment, C.E., Newman, L.S., 1998. Expansions of T-cell subsets expressing particular T-cell receptor variable regions in chronic beryllium disease. Am. J. Respir. Cell Mol. Biol. 18, 581 – 589. Hagman, J.D., Doglio, L.T., Hackett, J., Rudin, C.M., Haasch, D., Brinster, R., Storb, U., 1989. Inhibition of immunoglobulin gene rearrangement by the expression of a b2 transgene. J. Exp. Med. 169, 1911 – 1929. Horvai, A.E., Xu, L, Korzus, E., Brard, G., Kalafus, D., Mullen, T.M., Rose, D.W., Rosenfeld, M.G., Glass, C.K., 1997. Nuclear integration of JAK/STAT and Ras/AP-1 signaling by CBP and p300. Proc. Natl. Acad. Sci. USA 94, 1074 – 1079. Kaser, M.R., Ord, M.G., Stoken, L.A., 1980. The effects of beryllium on histone phosphorylation in liver. Biochem. Intern. 1, 148 – 150. Laundon, C.H., Griffith, J.D., 1987. Cationic metals promote sequence-directed DNA bending. Biochemistry 26, 3759 – 3762. Leonard, W.J., O’Shea, J.J., 1998. Jaks and stats: biological implications. Ann. Rev. Immunol. 16, 293 – 322. Newman, L.S., 1995. Current concepts: occupational illness. New Engl. J. Med. 333, 1128 – 1134. Newman, L.S., Lloyd, K., Daniloff, E.A., 1996. The natural history of beryllium sensitization and chronic beryllium disease. Environ. Health Perspect. 104, 937 – 943. Parker, V.H., Steven, C., 1979. Binding of beryllium to nuclear acidic proteins. Chem. Biol. Interact. 26, 167 – 177. Penix, L.A., Sweetser, M.T., Weaver, W.M., Hoeffler, J.P., Kerppola, T.K., Wilson, C.R., 1996. The proximal regulatory element of the interferon-g promoter mediates selective expression in T cells. J. Biol. Chem. 271, 31964 – 31972. Sawyer, R.T., Kittle, L.A., Campbell, P.A., Newman, L.S., 1998. Direct stimulation of TNF-a by beryllium in an
H. Hamada et al. / Toxicology 143 (2000) 249–261 inflammatory hybrid mouse macrophage cell line. Am. J. Respir. Crit. Care Med. 157, A32. Stark, G.R., Kerr, J.M., Williams, B.R.G., Silverman, R.H., Schreiber, R.D., 1998. How cells respond to interferons. Ann. Rev. Biochem. 67, 227–264. Tinkle, S.S., Schwitters, P.W., Newman, L.S., 1996. Cytokine production by bronchoalveolar lavage cells in chronic beryllium disease. Environ. Health Perspect. 104, 969–997. Tinkle, S.S., Kittle, L.A., Schumacher, B.A., Newman, L.S., 1997. Beryllium induces IL-2 and IFN-g in berylliosis. J. Immunol. 158, 518 – 526.
.
261
Tinkle, S.S., Newman, L.S., 1997. Beryllium-stimulated release of tumor necrosis factor-a, interleukin-6, and their soluble receptors in chronic beryllium disease. Am. J. Respir. Crit. Care Med. 156, 1884 – 1891. Tracey, K.J., 1998. Tumor necrosis factor-alpha. In: Thomson, A. (Ed.), The Cytokine Handbook, 3rd ed. Academic Press, NY, pp. 289 – 304. Yao, J., Mackman, N., Edgington, T.S., Fan, S.T., 1997. Lipopolysaccharide induction of the tumor necrosis factora promoter in human monocytic cells. J. Biol. Chem. 272, 17795 – 17801.