Interleukin-1β converting enzyme subfamily inhibitors prevent induction of CD86 molecules by butyrate through a CREB-dependent mechanism in HL60 cells

Abstract

To investigate the underlying mechanism for induction of CD86 molecules, we analysed the ability of the histone deacetylase (HDAC) inhibitor, sodium butyrate (NaB), to induce CD86 at the transcriptional level in HL60 cells. Our studies showed that the expression of CD86 on the cell surface was increased by 24 hr of NaB treatment, and the enhancement of CD86 mRNA expression was observed by real-time polymerase chain reaction. When we measured NF-κB binding activity, significant activity was induced upon NaB stimulation, which was suppressed by the addition of pyrrolidine dithiocarbamate. Butyrate also induced phosphorylated cAMP response element-binding protein (CREB), which bound to cAMP-responsive elements. Dibutyryl (db) -cAMP induced active CREB and increased the levels of CD86 by 24 hr. These observations indicated that NF-κB and/or CREB are crucial for butyrate-dependent activation of CD86 gene expression. We examined the inhibitory effects of various caspase inhibitors on the expression of CD86 in cells treated with NaB, because NaB also induced apoptosis with slow kinetics. Intriguingly, our results demonstrated that inhibitors of the interleukin-1β converting enzyme subfamily (caspase-1, -4, -5 and -13) blocked the butyrate-induced increase in level of CD86. These inhibitors interfered with CD86 gene transcription in the presence of activated NF-κB, whereas phosphorylated CREB was down-regulated in the reactions where these inhibitors were added to inhibit CD86 gene expression. These results suggested that butyrate not only acetylates histones on the CD86 promoter through the suppression of HDAC activity, but that butyrate also regulates CREB-mediated transcription, possibly through the caspase activities triggered by NaB.

Introduction

As tumours of myeloid and lymphoid lineage share the ontogeny of professional antigen-presenting cells (APC), the capacity of such malignant cells to present endogenously expressed tumour-associated antigens directly to T cells was suggested previously.¹ On the other hand, such tumour cells are known to evade host immune surveillance as a result of their lack of co-stimulatory molecules, which causes tumour development as a result of the inefficient stimulation of tumour-reactive cytotoxic T cells.¹ Elucidating the transcriptional regulation of the critical co-stimulatory molecules is central to understanding the regulation of T-cell-mediated immune responses. Among the several co-stimulatory signals characterized to date, members of the B7 family (B7/CD80 and B7-2/CD86) on APC interact with CD28 and cytotoxic T-lymphocyte antigen-4 (CTLA-4) on T cells, resulting in efficient T-cell sensitization.² Tumour cells generally express major histocompatibility complex class I and II molecules, but CD80 and CD86 are not always expressed on tumour cells; thus these molecules were the target of immunotherapy in acute myeloid leukaemia.^3,4 The previous reports that some stimuli could induce CD86 molecules in tumour cells, and that introduction of CD86 by gene transfer rendered tumour cells immunogenic prompted us to investigate the mechanism underlying regulation of these molecules in tumour cells.^5–7

Sodium butyrate (NaB) induces differentiation as well as apoptosis in several cell types.^8,9 Butyrate can affect gene transcription in a positive or negative manner, depending on the gene.^10,11 The precise mechanisms of action of butyrate in cell differentiation, apoptosis and gene expression are not yet understood. As butyrate inhibits histone deacetylase (HDAC), and hyperacetylation of histones can lead to alterations in chromatin structure, resulting in conditions that favour accessibility of transcription factors to DNA, the transcriptional and other effects of butyrate are often ascribed to its ability to effect histone hyperacetylation.¹² Butyrate has been shown to increase the expression of target genes such as CD80, CD86 and intercellular adhesion molecule-1 (ICAM-1) on leukaemia cell lines, of which the transcription is dependent on the nuclear factor (NF)-κB consensus site within its promoter.^13–15 In cancer therapy, clinical trials showed phenylbutyrate to be effective in the treatment of several cancers, indicating that the regulation of co-stimulatory and adhesion molecules by acetylation/deacetylation is important as the major mechanism.¹⁶ However, different mechanisms, including regulation of transcription factors, and signalling pathways of apoptosis, are also considered to play roles in some of the observed effects of butyrate.

In this study, we demonstrated a mechanism of transcriptional regulation of the CD86 gene in HL60 cells by NaB. The transcriptional activity by butyrate was dependent on the activation of NF-κB and/or cAMP response element-binding protein (CREB). Interestingly, caspase inhibitors of the interleukin-1β converting enzyme (ICE) subfamily interfered with CD86 gene transcription in the presence of activated NF-κB, which was dependent on phospho-CREB binding activity.

Materials and methods

Cells and cell culture

The human myelomonocytic leukaemia cell lines (HL60, U937 and THP-1) were obtained from the Cell Resource Center for Biomedical Research, Institute of Development, Aging and Cancer, Tohoku University (Sendai, Japan). NKM-1 cells were obtained from the Institute for Fermentation (Osaka, Japan). All cell lines were cultured in RPMI-1640 medium containing 10% fetal calf serum and 2 mm l-glutamine at a concentration of 5 × 10⁵ cells/ml. Cells were split in logarithmic growth phase by routine passage every 2–3 days.

Reagents and monoclonal antibodies (mAbs)

Sodium butyrate was obtained from Wako Pure Chemical Industries (Osaka, Japan). Pyrrolidine dithiocarbamate (PDTC) and lipopolysaccharide (LPS; from Escherichia coli O55 : B5) were obtained from Sigma Chemical (St Louis, MO). N-Tosyl-l-phenylalanine chloromethyl ketone (TPCK), dibutyryl (db)-cAMP, SB203580 and PD98059 were purchased from Calbiochem (La Jolla, CA). These reagents were divided into aliquots and frozen at − 80°. Fluorochrome-labelled murine mAbs specific for human CD86 (IT2.2; PharMingen, San Diego, CA), CD80 (L307.4; Becton Dickinson, San Jose, CA), CD14 (MY4; Coulter, Hialeah, FL), ICAM-1/CD54 (HA58; PharMingen) as well as murine isotype controls (Caltag Labs, Burlingame, CA) were used for flow cytometry.

Caspase inhibitors

Benzyloxycarbonyl(z)-tyrosine-valine-alanine-aspartic acid-fluoromethyl ketone (YVAD-fmk, caspase-1 inhibitor), z-valine-aspartic acid-valine-alanine-aspartic acid (VDVAD)-fmk (caspase-2 inhibitor), z-aspartic acid-glutamic acid-valine-aspartic acid (DEVD)-fmk (caspase-3 inhibitor), z-leucine-glutamic acid-valine-aspartic acid (LEVD)-fmk (caspase-4 inhibitor), z-tryptophan-glutamic acid-histidine-aspartic acid (WEHD)-fmk (caspase-5 inhibitor), z-leucine-glutamic acid-histidine-aspartic acid (LEHD)-fmk (caspase-9 inhibitor) and z-leucine-glutamic acid-glutamic acid-aspartic acid (LEED)-fmk (caspase-13 inhibitor) were purchased from MBL (Nagoya, Japan); these inhibitors were added at 10–20 μm in dimethyl sulphoxide (DMSO). Cells were pretreated with these inhibitors for 1 hr at 37°. NaB was then added, followed by incubation for 24 hr, after which apoptosis and CD86 induction were evaluated. A DMSO control was also included as a control for the given concentration of each inhibitor.

Expression of CD86 on the cell surface

Cells were treated with NaB for 24 hr in the absence or presence of various inhibitors. Cells were suspended in cold staining buffer. The optimal concentration of phycoerythrin (PE)-conjugated anti-CD86 mAb was added to cells. Cells were incubated in the dark at 4° for 20 min, washed twice, and then analysed on a FACScan (Becton Dickinson). PE-conjugated murine immunoglobulin G2b (IgG2b) of unrelated specificity was used as control, and a total of 15 000 cells/sample were analysed.

Flow cytometric analysis of apoptosis

NaB was added to the cells in culture at a final concentration of 1 mm. PDTC (1 μm) for inhibition of NF-κB binding activity, or various caspase inhibitors (20 μm) was applied 1 hr prior to NaB treatment. Cells were cultured in 24-well plates. After treatment with NaB for the indicated times, measurement of phosphatidylserine externalization was performed by staining cells with fluorescein isothiocyanate (FITC)-conjugated annexin V (Trevigen Inc., Gaithersburg, MD) for 20 min at room temperature. Apoptotic cells were analysed on a FACScan.¹⁷

Measurement of intracellular caspase activity

Intracellular caspase activities were measured by using the carboxyfluorescein (FAM) -labelled caspase inhibitor (Intergen, Purchase, NY). Briefly, treated cells were labelled with FAM-YVAD-fmk (caspase-1 activity) or FAM-DEVD-fmk (caspase-3 activity). Cells were analysed by flow cytometry to determine the percentage of intracellular active caspase-1⁺ or -3⁺ cells.

Real-time polymerase chain reaction (PCR)

RNA was extracted from cells stimulated with NaB. The following reverse transcription for cDNA synthesis was performed by the standard method. The primers for the amplification of cDNA were as follows: CD86, 5′-TGGTGCTGCTCCTCTGAAGATTC-3′, 5′-ATCATTCCTGTGGGCTTTTTGTG-3′,300 base pairs (bp); and β-actin, 5′-CTGTCTGGCGGCACCACCAT-3′, 5′-GCAACTAAGTCATAGTCCGC-3′, 254 bp. Real-time PCR analysis was performed using SYBR Green PCR Master Mix (Qiagen, valineencia, CA) by the icycler iq Real time System (Bio-Rad, Hercules, CA). For quantification of mRNA, we calculated the relative quantity of each sample using standard curves. The results were shown as the fold induction relative to medium alone.

Assay for NF-κB and phospho-CREB activity

Activated NF-κB was measured with an NF-κB assay kit specific for the p65 subunit according to the manufacturer's instructions (Active Motif, Carlsbad, CA). Briefly, samples of whole cell extracts (1–10 μg of protein/well) were added to 96-well plates coated with an oligonucleotide containing the NF-κB consensus site (5′-GGGACTTTCC-3′), and incubated for 1 hr at room temperature with mild agitation. After washing three times, NF-κB p65 antibody was added and incubated for 1 hr without agitation followed by addition of horseradish peroxidase-conjugated anti-mouse IgG1. Colorimetric reactions were developed, stopped and measured at 450 nm. For this assay, LPS-treated HL60 cells were used as a positive control.¹⁸ The specificity of binding was also examined using an oligonucleotide containing a wild-type or mutated NF-κB consensus binding site.

Activated phospho-CREB was measured with a CREB assay kit specific for Ser¹³³-phosphorylated CREB (Active Motif). Samples of nuclear cell extracts were added to wells coated with an oligonucleotide containing the CRE site (5′-TGACGTCA-3′). For this assay, the nuclear cell extract of db-cAMP treatment was used as a positive control.¹⁹

Statistical analysis

Data are given as means ± SD. Multiple comparisons were performed by Scheffe's test. P-values less than 0·05 were regarded as significant.

Results

Induction of CD86 and apoptosis in leukaemia cells by butyrate

To determine the effect of butyrate on the expression of CD86 on the cell surface of various leukaemic cells, we tested four leukaemic cell lines for the action of NaB (1 mm) by flow cytometric analysis (Fig. 1a). The expression of CD86 was up-regulated in all cell lines examined after 24 hr. HL60 cells (5 × 10⁵/ml) were incubated for 24 hr in the presence of NaB at various concentrations. A significant increase in CD86 induction was observed at 1 mm NaB (Fig. 1b). In contrast to CD86, NaB failed to induce CD80, another B7 family member (Fig. 1b). Cell viability reduced markedly at higher concentrations than 5 mm (data not shown).

Figure 1.

Induction of CD86 by butyrate in leukaemia cell lines. (a)Induction of CD86 in various cell lines. Leukaemia cell lines (HL60, U937, THP-1 and NKM-1) were cultured for 24 hr with 1 mm NaB. Cells were harvested and stained with PE-conjugated anti-CD86 mAb. (b)Dose response of NaB in CD86 induction. HL60 cells were cultured with various concentrations of NaB. Cells were stained with PE-conjugated anti-CD86 or anti-CD80 mAb. The percentage of positive cells was measured by flow cytometry. The results are expressed as means ± SD of three separate experiments. *P < 0·01 versus medium alone.

To investigate the time–course of CD86 and apoptosis induction, HL60 cells were treated with NaB (1 mm). Cells were harvested every day up to 3 days and stained with PE-conjugated anti-CD86 mAb and FITC-conjugated annexin V. Representative kinetics of surface expression analysis are shown in Fig. 2(a). In cultures without stimulation, frequency of surface staining for CD86 was usually low (15·1 ± 1·5%). Induction of CD86⁺ cells was enhanced by NaB treatment. The percentage of CD86⁺ cells increased up to 48 hr after stimulation and subsequently decreased. NaB induced apoptosis with slower kinetics than that of CD86 induction, so that at 24 hr of treatment with NaB, only 7·5 ± 1·2% of the cells showed apoptotic signs. On increasing the time of treatment, this percentage increased progressively up to 72 hr, when 28·5 ± 2·7% of cells were apoptotic (Fig. 2a). We examined activation of the caspase cascade during NaB treatment. Caspase-3 activation in HL60 cells was slightly augmented at 24 hr after stimulation with NaB (Fig. 2b). Furthermore, active caspase-1⁺ cells were significantly increased in NaB stimulation (Fig. 2c), suggesting the involvement of ICE subfamily caspases in the process of NaB treatment.

Figure 2.

Butyrate induces CD86 and apoptosis with different kinetics. (a) Time kinetics of induction of CD86 and apoptosis. HL60 cells were cultured in the absence (○, □) or presence (•, ▪) of 1 mm NaB. Cells were harvested every 24 hr and stained for 20 min with PE-conjugated anti-CD86 mAb (○, •) and for 20 min with FITC-conjugated annexin V (□, ▪). The percentage of positive cells was measured by flow cytometry. The results are expressed as means ± SD of three separate experiments. (b,c) Activation of caspase cascade. Cells were cultured with or without 1 mm NaB for 24 hr, and intracellular active caspase-3⁺ cells (b)and caspase-1⁺ cells (c)were analysed by flow cytometry.

Effects of various inhibitors on induction of CD86⁺ cells by butyrate

Prior to stimulation, HL60 cells were pretreated for 1 hr with PDTC (1 μm), a potent inhibitor of NF-κB.²⁰ NF-κB binding activity of cells treated with NaB for 24 hr was indeed inhibited by PDTC pretreatment (data not shown). Treatment with PDTC resulted in reduction of the percentages of NaB-induced CD86⁺ or ICAM-1⁺ cells (Fig. 3), indicating that NF-κB activation is partially required for expression of both CD86 and ICAM-1 on HL60 cells. A similar result was obtained by TPCK treatment (data not shown). However, NF-κB activated by LPS had only small effects on expression of these molecules compared with NaB treatment (Fig. 3). We next examined the influence of the specific extracellular signal-regulated kinase 1/2 (ERK1/2) inhibitor PD98059 (PD) and p38 mitogen-activated protein kinase (MAPK) inhibitor SB203580 (SB) on NaB-induced CD86 and ICAM-1 (Fig. 3). CD86 expression was slightly inhibited in the presence of each inhibitor, whereas ICAM-1 was significantly inhibited by PD98059, but not SB203580 treatment. These observations demonstrated that the observed changes in MAP kinase signalling by butyrate are indeed specific for each target gene.

Figure 3.

Butyrate-induced CD86 is partially inhibited by PDTC. HL60 cells were cultured with PDTC (1 μm), PD98059 (PD, 50 μm), or SB203580 (SB, 40 μm) for 1 hr. After 24 hr of stimulation with NaB (1 mm) or LPS (100 ng/ml), cells were stained with PE-conjugated anti-CD86 or anti-ICAM-1 mAb. Percentages of CD86⁺ or ICAM-1⁺ cells are expressed as means ± SD of three separate experiments. *P < 0·01 versus NaB treatment.

Effects of caspase inhibitors of ICE subfamily on CD86 induction by butyrate

When we examined the inhibitory effects of a variety of caspase inhibitors on CD86 induced by NaB, CD86 was found to be significantly down-regulated by pretreatment for 1 hr with 20 μm of ICE subfamily inhibitors, YVAD-fmk (caspase-1), LEVD-fmk (caspase-4), WEHD-fmk (caspase-5), and LEED-fmk (caspase-13) (Fig. 4b). YVAD-fmk inhibited NaB-induced CD86 at doses of 10–20 μm, whereas the inhibitory ability of LEED-fmk was high (>20-fold) compared with YVAD-fmk (Fig. 4a). On the other hand, inhibitors of members of other subfamilies, VDVAD-fmk (caspase-2), DEVD-fmk (caspase-3), or LEHD-fmk(caspase-9), had no effect on CD86 induction in HL60 cells by NaB (Fig. 4c). Caspases were activated by treatment with NaB for 24 hr but apoptotic signs were detected in only a small proportion of cells (Fig. 2).

Figure 4.

Caspase inhibitors of the ICE subfamily prevent butyrate-induced CD86. (a)Dose response of caspase inhibitors. HL60 cells were incubated for 1 hr with the indicated concentrations of YVAD-fmk (caspase-1 inhibitor) or LEED-fmk (caspase-13 inhibitor) prior to stimulation with NaB (1 mm) for 24 hr. A DMSO control was also included as a control for the 20 μm of inhibitor. Cells were harvested and stained with PE-conjugated anti-CD86 mAb. Results are representative of three separate experiments. (b, c) Inhibition of CD86 by caspase inhibitors. One hour after addition of caspase inhibitors of ICE subfamily (caspase-1, -4, -5 and -13) (b), or caspase inhibitors of other subfamilies (caspase-2, -3 and -9) (c)at 20 μm, cells were stimulated with NaB (1 mm) for 24 hr and stained. NaB-induced CD86⁺ population is expressed as 1·0, and the effects of caspase inhibitors are shown as ratios relative to this value. The results are expressed as means ± SD of three separate experiments. *P < 0·01 versus NaB treatment.

Transcriptional inhibition of CD86 by ICE subfamily inhibitors

To confirm whether the down-regulation of CD86 by ICE subfamily inhibitors was regulated transcriptionally or post-transcriptionally, the transcripts of this gene were quantitatively assayed by real-time PCR (Fig. 5). CD86 mRNA increased (>sevenfold) after 24 hr of treatment with NaB or db-cAMP compared with the control. Pretreatment with ICE subfamily inhibitor (caspase-13) for 1 hr completely inhibited the increase of CD86 mRNA expression by NaB, and PDTC also showed an apparent effect. No significant effect was observed in pretreatment with caspase-3 inhibitor. Thus, the down-regulation of CD86 by ICE subfamily inhibitor was observed at the mRNA level, indicating that the alterations in CD86 mRNA level are well correlated with changes in cell-surface expression detected by flow cytometry (Figs 3 and 4).

Figure 5.

Caspase-13 inhibitor prevents butyrate-induced CD86 at the transcriptional level. HL60 cells were preincubated for 1 hr in the absence or presence of PDTC (1 μm), DEVD-fmk (20 μm), or LEED-fmk (20 μm), and were then treated for 4 hr with NaB (1 mm) or db-cAMP (200 μm). Total cellular RNA from cells was subjected to real-time PCR using specific primers. Transcripts of CD86 were quantified by relative standard curves. The representative fold induction relative to control is shown.

Effects of ICE subfamily inhibitors on the levels of transcription factors

We investigated whether transcriptional regulation of CD86 by ICE subfamily inhibitors is associated with activity of transcription factors such as NF-κB and CREB. For this, HL60 cells were pretreated with each caspase inhibitor (20 μm) for 1 hr. After incubation with NaB (1 mm) for 24 hr, the whole cell extracts were prepared and then analysed for NF-κB activity using NF-κB consensus oligonucleotide. The nuclear cell extracts were also prepared to examine phospho-CREB activity. The results shown in Fig. 6(a) indicated that NF-κB binding activity of cells treated with NaB for 24 hr was increased to a level similar to that in LPS (1 μg/ml)-treated cells. By 2 hr of NaB stimulation, NF-κB binding activity was slightly increased (data not shown). The induction of active NF-κB by NaB was inhibited in the presence of excess wild-type (wt) oligonucleotide, suggesting the specificity of binding of NF-κB to the consensus binding sequence. On the other hand, phospho-CREB activity was also increased by NaB treatment for 24 hr, which was similarly up-regulated by db-cAMP (200 μm) as a positive control (Fig. 6b).

Figure 6.

Caspase-13 inhibitor prevents phospho-CREB activity, not NF-κB activity. (a, b) HL60 cells pretreated with inhibitor of caspase-3 or -13 (20 μm) for 1 hr were incubated with NaB (1 mm) for 24 hr. Cells treated with LPS (1 μg/ml) for NF-κB activity or cells treated with db-cAMP (200 μm) for CREB activity were used as positive controls, respectively. Whole cell extracts (a, NF-κB) or nuclear cell extracts (b, CREB) were added to wells coated with oligonucleotide containing the respective consensus sites, and incubated for 1 hr. Cell extracts were also incubated in the presence of a 20-fold excess of wild-type (wt) oligonucleotide. Anti-p65 antibody (a) or anti-Ser¹³³-phospho-CREB antibody (b) was added and incubated for 1 hr followed by the addition of horseradish peroxidase-conjugated anti-mouse IgG. Colorimetric reactions were measured at 450 nm. All samples were assayed in triplicate, and the results are expressed as means ± SD.

Although neither caspase inhibitor, LEED-fmk for caspase-13 and DEVD-fmk for caspase-3, had any effect on down-regulation of NF-κB binding activity, LEED-fmk, but not DEVD-fmk, markedly inhibited CREB activity (Fig. 6a,b). Furthermore, we confirmed that addition of db-cAMP had no effect on NF-κB activity, which up-regulated CREB activity, and db-cAMP effectively induced CD86⁺ cells in HL60 cells (Fig. 7a,b). The db-cAMP-induced CD86 was inhibited by caspase-1 and -13 inhibitors. These results suggested that caspase-13 (ICE subfamily) inhibitor prevents binding activity of phospho-CREB and subsequently inhibits the transcriptional activity of the CD86 gene.

Figure 7.

db-cAMP up-regulates phospho-CREB activity and induces CD86. (a)Activation of CREB by db-cAMP. HL60 cells were cultured with or without db-cAMP (200 μm) for 24 hr. Whole cell extracts (NF-κB) or nuclear cell extracts (CREB) were prepared. Anti-p65 antibody or anti-Ser¹³³-phospho-CREB antibody was added and colorimetric reactions were measured at 450 nm. All samples were assayed in triplicate, and the results are expressed as means ± SD. (b)CD86 induction by db-cAMP. Cells were cultured with db-cAMP (200 μm) in the absence or presence of caspase-1 (20 μm) or caspase-13 (20 μm) inhibitor for 24 hr, harvested and stained with PE-conjugated anti-CD86 mAb. Cells were analysed to quantify the CD86⁺ cells. The results are expressed as means ± SD of three separate experiments. *P < 0·01 versus db-cAMP treatment.

Discussion

CD86 belongs to the B7 family of co-stimulatory molecules expressed on APC. Binding to the T-cell molecules CD28 and CTLA-4 generates co-stimulatory and inhibitory signals in T cells, respectively.² The cytoplasmic tail of CD86 spans three exons, suggesting potential signalling capacity of the CD86 molecule in itself through alternate splicing other than signal delivery.²¹ An understanding of the molecular mechanisms of this regulated expression is central to the understanding of the regulation of T-cell-mediated cytotoxicity. In the present study, we explored the mechanism responsible for CD86 promoter stimulation by butyrate at the transcriptional level. We demonstrated that NF-κB and/or CREB were crucial for butyrate-dependent activation of CD86 gene expression (Fig. 6a,b). Previously, characterization of the CD86 promoter revealed a complex control region that can be triggered by multiple activation pathways.¹⁴ Intriguingly, CD86, which is able to activate T cells via NF-κB,¹⁸ is itself transcriptionally regulated by NF-κB through binding. The representative triggering of CD86 transcription by CD40/CD40L includes activation of various transcription factors in APC, such as NF-κB, CREB, AP-1, NFAT, STAT-3 and STAT-6.^22–25 The main transcriptional activator involved in induction of CD86 gene expression is probably NF-κB, because NF-κB is known to play an important role in many of the biological effects exerted by CD40.²⁶ However, additional factors are also involved in the alternative pathways of induction. CREB is a dimeric transcription factor, which binds to cAMP-responsive elements and is involved in cAMP-signalling pathways.¹⁹ That is, elevated intracellular cAMP level induces the expression of various genes through protein kinase A-mediated phosphorylation of CREB. As the CD86 promoter sequence has a binding site for CREB, db-cAMP induced phosphorylated CREB, and subsequently increased CD86 expression on HL60 cells by 24 hr (Fig. 7a,b). The inducibility of CD86 by db-cAMP may explain the reasons why LPS had only slight activity of CD86 induction, despite its active NF-κB-inducing capacity (Figs 3 and 6a). Moreover, the activity of several inducible transcription factors is regulated through their association with cellular co-activators. Among these, interactions of NF-κB and phospho-CREB with the CREB-binding protein (CBP) appear to be necessary to optimize the CD86 transcriptional activity.²⁷ As CBP is present in limiting amounts in the nucleus, a difference in capacity between NF-κB and CREB for binding to CBP is probably important to exhibit the transcriptional activities of these factors induced by butyrate.

The involvement of signal transduction pathways in mediating the effects of butyrate on cells has been reported. The MAPK signalling cascade comprising the ERK, c-Jun N-terminal kinase (JNK) and p38 may be a potential target of butyrate action for CD86 induction because it was shown that ERK and p38 signal transduction pathways play critical roles in butyrate-induced erythroid differentiation of K562 human leukaemia cells.²⁸ Moreover, the p38 MAPK pathway was shown to contribute to NF-κB-mediated transactivation,²⁹ whereas the ERK1/2 pathway was shown to phosphorylate serine¹³³ of CREB.³⁰ Thus, it is of interest to determine whether CD86 gene expression is affected by inhibition of these MAPK activities. The availability of specific inhibitory drugs for the p38 MAPK (SB203580) and ERK1/2 (PD98059) pathways prompted us to investigate the respective roles of these MAPKs in induction of CD86.^31,32 Unexpectedly, we observed only partial inhibitory effects of these drugs on the up-regulation of CD86 expression in HL60 cells triggered by butyrate (Fig. 3). In contrast to CD86, the up-regulation of cell-surface ICAM-1 expression by butyrate was significantly prevented by treatment with PD98059. These observations indicated that when butyrate is selected as an inducing agent, other signalling pathways are involved in CD86 induction, but ERK1/2 action may pertain to ICAM-1 gene expression. The role of the JNK MAPK remains to be clarified, because no specific inhibitor of this pathway is available.

Butyrate is thought to inhibit the enzymatic activity of HDAC through serine–threonine protein phosphatase as a molecular intermediate and to enhance the transcriptional activities of several genes.^6,33 It has been suggested that histone acetylation promotes transcription, probably by relaxing specific segments of DNA and facilitating the binding of transcription factors.³⁴ Furthermore, non-histone proteins, such as transcription factors, have been reported to serve as substrates for histone acetyltransferase (HAT) activity.³⁵ Several transcription factors bind co-activator CBP, which contains HAT domains and has strong HAT activity.³⁶ CBP can recruit p300/CBP-associated factor (p/CAF), which also has potent HAT activity, indicating that complexes with multiple HAT activities can be formed.³⁷ CREB requires the HAT activity of CBP but not that of p/CAF, whereas NF-κB utilizes the HAT activity of p/CAF but not that of CBP.^37,38 Thus, it is possible that butyrate not only acetylates histones on the CD86 promoter through suppression of HDAC activity, but that butyrate also supports NF-κB- and/or CREB-mediated transcription, possibly through the acetylation by each co-activator.

Depending on the inhibitory effect on HDAC, it seems that butyrate induces apoptosis through caspase-3 activation, which is highly synergistic with the release of mitochondrial cytochrome c in Jurkat lymphoid and colorectal cancer cell lines.³⁹ However, butyrate induced apoptosis with slower kinetics in HL60 cells than in these cell lines, so that after 24 hr of treatment only about 7% of the cells showed apoptotic signs. On increasing the time of treatment, this percentage increased progressively up to 72 hr (Fig. 2a). The degradation of the inhibitor of NF-κB (IκBα) through the activation of 26S proteasome by butyrate could be responsible for the delay with which butyrate-induced apoptosis occurs in HL60 cells.⁴⁰ An alternative explanation for the delay on apoptosis is that CREB plays a role in the regulation of bcl-2 expression, which may serve to regulate cytochrome c release.⁴¹ The consequent activation of these transcription factors controlled a survival network of its target genes in HL60 cells, which conferred resistance of cells to apoptosis. Thus, butyrate may be a useful tool for elucidating the mechanism of the cytoplasmic pathway of apoptosis, because it induces apoptosis and alters gene expression, at similar concentrations and in diverse types of cells.

Our studies showed a link between the expression of CD86 and apoptosis in cells treated with an inhibitor of HDAC. The processing of caspase-1 was observed in cells treated with NaB, whereas NaB induced late the processing of caspase-3 (Fig. 2b,c).⁴² Certain caspases are not only degradative enzymes but also highly regulated signalling molecules that control critical biological processes via specific limited proteolysis.⁴³ To ascertain the caspase action on CD86 expression, we first examined the effects of caspase inhibitors on the expression of CD86 in NaB-stimulated HL60 cells. The caspase family has been divided into three groups based on substrate specificity. The caspases belonging to the same family have overlapping substrate specificities, suggesting at least partially overlapping functions.⁴³ Our results demonstrated that inhibitors of the ICE subfamily (caspase-1, -4, -5 and -13) block NaB-induced increase of CD86 (Fig. 4b). In contrast, members of other subfamilies (caspase-2, -3 and -9) did not have prominent roles in CD86 induction. The inhibitory effect was considered to be a result of reduced transcription of the CD86 gene (Fig. 5), suggesting that butyrate signalling for transcriptional activation of the CD86 gene is inhibited by ICE subfamily inhibitors. Our results also showed that these inhibitors interfere with CD86 gene transcription in the presence of activated NF-κB (Fig. 6a). Therefore, we examined whether the activities of other transcription factors are affected by ICE subfamily inhibitors during NaB stimulation. Phosphorylated CREB was down-regulated in the reactions where LEED-fmk was added to inhibit CD86 gene expression (Fig. 6b). Down-regulation of active CREB was specific because there was no decrease in the presence of the DEVD-fmk. These results suggested that specific protein cleavage by the ICE subfamily inhibitor-sensitive proteases modifies phosphorylated CREB and enhances the transcriptional activity of the CD86 gene.

Several powerful methods are now available to elicit cytotoxic T-cell responses to tumour-associated antigens in vitro and in vivo. Absence of co-stimulation remains a candidate as the mechanism by which developing tumours bearing such antigens fail to be rejected. Our results indicated a novel mechanism of CD86 regulation in HL60 cells. Therefore, butyrate is useful to elucidate the mechanism of regulation of co-stimulation because of a pleiotropic agent with multiple effects including additional regulatory mechanisms.

Abbreviations

APC

antigen-presenting cells

CBP

CREB-binding protein

CREB

cAMP response element-binding protein

HAT

histone acetyltransferase

HDAC

histone deacetylase

ICE

interleukin-1β converting enzyme

NaB

sodium butyrate

PDTC

pyrrolidine dithiocarbamate