Impaired expression of indoleamine 2, 3-dioxygenase in monocyte-derived dendritic cells in response to Toll-like receptor-7/8 ligands

Reagents

Recombinant human granulocyte–macrophage colony-stimulating factor (GM-CSF) and IL-4 were purchased from R & D system. The IDO inhibitor 1-methyl-d,l tryptophan was purchased from Aldrich Chemical (Milwaukee, WI). Neutralizing antibodies against TNF-α and IL-1β and IFN-α were purchased from Pharmingen (San Diego, CA). Monoclonal antibody against IDO was purchased from US Biological (Nordic Biosite AS, Taby, Sweden) and monoclonal antibodies against Stat-1 and β actin were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). CD4 Dynabeads and CD14 microbeads were purchased from Invitrogen (San Diego, CA) and Miltenyi Biotech (Auburn, CA), respectively. Unmodified and 2′-modidied RNA oligonucleotides were purchased from Eurogentec or Ambion. The sequences of the used RNAs are:

27S, 5′-GUCCGGGCAGGUCUACUUUTT-3 ′;

19S, 5′-GAGGCAAUCACCAAUAGCATT-3′;

19A, 5′-UGCUAUUGGUGAUUGCCUCTT-3′;

CMV-UL37, 5′-GUAGCAUGUUUUAGGAAUG-3′.

Cell isolation and culture

Peripheral blood mononuclear cells (PBMC) were isolated by density gradient centrifugation (Lymphoprep, Nycomed Pharm, Oslo, Norway) from buffy coats obtained from healthy adult donors. Monocytes were isolated by plastic adherence. Also, they were isolated using CD14 microBeads (Miltenyi Biotec) according to the manufacturer's instructions. Monocyte-derived immature (imoDC) and mature (mmoDC) DC were generated by culturing adherent monocytes in the presence of GM-CSF (25 ng/ml) and IL-4 (50 ng/ml). After 4–5 days in culture, imoDC were stimulated with TNF-α (50 ng/ml) to generate mmoDC. Blood myeloid and plasmacytoid DC were purified using the BDCA-1 or BDCA-4 cell isolation kit, respectively, according to the manufacturer's instructions (Miltenyi Biotec).

Cell stimulation and cytokine determination

Monocytes were plated in 96-well tissue culture dishes (2 × 10⁵ cells/200 µl) in RMPI-1640, supplemented with fetal calf serum (FCS) and antibiotics. The cells were incubated with no stimulus or with RNA molecules. The cultures were incubated for 16 hr at 37°, and the levels of TNF-α, IL-β and IL-6 were determined by enzyme-linked immunosorbent assay (ELISA) of culture supernatants. For protein preparations, cells were plated in 6-, 12- or 24-well plates at concentration of 10⁶/ml, and then transfected with the test molecules for 16 hr. Subsequently, cells were harvested, washed with phosphate-buffered saline (PBS) and suspended in lysis buffer (PBS + 1% NP40) and protease inhibitors. After incubation at 4° for 30 min, samples were centrifuged at 16 000 g for 10 min at 4°. The resultant supernatants (protein extracts) were stored at −20°. In all experiments, isRNAs, R848, LPS and PamCys3 were used at 500 ng/ml, 5 µm, 100 ng/ml, and 10 µg/ml, respectively. Clinical isolate of HCMV (HCMV2006) used in this study was kindly provided by the Dr Halvor Rollag (Department of Microbiology, Rikshospitalet National Hospital, Oslo, Norway). HCMV2006 was added to the monocytes at a multiplicity of infection of 5.

Reverse transcription–polymerase chain reaction (RT–PCR)

Total RNA was extracted using Trizol (Invitrogen), and cDNA was synthesized from 0.5 to 2 µg total RNA using the first strand cDNA synthesis kit and oligo-dT primer according to the manufacturer's instructions. IDO-specific mRNA is amplified using the following primers: IDO forward, 5′-GGAAATAGCAGCTGCTTCTGCA-3′; IDO reverse, 5′-CTCCTCAGGGAGACCAGAGCTT3′; As an internal control β-actin mRNA was also amplified using the following primers: β-actin forward, 5′-ATCTGGCACCACACCTTCTACAATGAGCTGCG-3′; β-actin reverse, 5′-CGTCATACTCCTGCTTGCTGATCCACATCTGC-3′. PCR products were analysed by agarose gels.

Immunoblot analysis

Equal amounts of protein extracts were analysed by polyacrylamide gel electrophoresis (PAGE) and then transferred to nitrocellulose membrane by electrophoresis. Subsequent to blocking with 5% dry milk in PBS/Tween (0.1%) for 16 hr at 4°, membranes were probed with antibodies specific to IDO or signal transducer and activator of transcription-1 (STAT-1), followed by horseradish peroxidase-conjugated rabbit or mouse secondary antibodies. Signals were visualized with autoradiography using ECL system. Blots were stripped and reprobed with antibodies against β-actin.

T-cell isolation and activation

CD4⁺ T cells were isolated from PBMC with CD4 Dynabeads® in combination with Detachabead®reagent (Dynal, Oslo, Norway) to >99% purity as determined by flow cytometry. Subsequently freshly isolated T cells were labelled in PBS with 1.5 µm carboxyl flurescein succinimidyl ester (CFSE) for 5 min at 37°. CFSE-labelled T cells (10⁵ cells/well) were washed and cocultured with autologous isRNA-stimulated monocytes (5 × 10⁴/cells/well) along with phyohaemagglutinin (PHA) mitogen (5 µg/ml). As control, unstimulated autologous monocytes were added to T cells as non-suppressive accessory cells. T-cell responses were analysed by flow cytometry after 3 days of culture.

Statistical analysis

anova with Bonferroni t-test was used to compare the data between treated and untreated cells. Data are expressed as means ± SD. P-values of <0.05 were considered to indicate statistical significance.

Immunostimulatory RNAs induce IDO expression in human monocytes

Monocytes are circulating peripheral blood cells that can be differentiated by cytokines into macrophages, dendritic cells as well as into several other cell types such as osteoclasts and microglia-like cells.¹⁷ Furthermore, they represent a major source of cytokines that modulate innate and adaptive immune responses. Because IDO enzyme plays a key role in immune regulation, we first investigated its expression in human monocytes after being exposed to TLR7/8 ligands. Our main prototype for TLR7/8 activator is the sense strand of a siRNA, known as 27S. This sequence efficiently induced the production of inflammatory cytokines and type I IFN in human immune cells.^10,13 Consistent with their expression of TLR8, monocytes responded to 27S as shown by significant production of proinflammatory cytokines such as TNF-α¹³ (data not shown).

To evaluate the effects of 27S molecule on IDO expression, Western blot analysis was performed (Fig. 1a). The data showed that human monocytes do not constitutively express IDO, but a substantial amount of IDO protein is induced in transfected monocytes. To determine the time frame for IDO expression following TLR ligation, a kinetic study was performed. The cells were stimulated with siRNA 27S for 16 hr, followed by cell sampling at various time points. A significant up-regulation of IDO occurred at 16 hr incubation time (Fig. 1b). In some cases, a very weak expression of IDO was seen in untreated monocytes. This is more likely caused by donor variation.

Having shown that monocytes can up-regulate the expression of IDO in response to isRNA 27S, in the next experiments we investigated their response to TLR2 and TLR4 ligands. To generalize our data, the following TLR7/8 ligands and control molecules were also tested: R848, a specific ligand for TLR7/8, a 2′-fluoro uridine modified 27S (27F), a second isRNA (19A), and a non isRNA (19S). Under our experimental conditions, all TLR7/8 ligands induced the expression of IDO (Fig. 1c, shows a representative example). Monocytes did show a significant up-regulation of IDO in response to TLR4 ligand LPS. However, when compared with TLR7/8 ligands, the effect was nearly threefold lesser. TLR2 ligand Pam3Cys exhibited no stimulatory effect. Cells treated with either 27F or the non isRNA 19A did not up-regulate IDO when compared to untreated cells.

In the next set of experiments, we have tested the effects of 27S on positively selected monocytes. In this respect, monocytes were isolated from PBMC with CD14-beads to >99 purity as determined by flow cytometry (data not shown). Cells stimulated with 27S molecule showed a dramatic up-regulation of IDO protein levels as assessed by Western blots (Fig. 1d). IsRNA-treated monocytes produced a large amount of inflammatory cytokines (e.g. TNF-α, IL-6), but not IFN-α and -γ (data not shown).

Having shown that monocytes can respond to synthetic isRNAs by up-regulating IDO and producing proinflammatory cytokines, next we evaluated whether they respond to viral RNAs. It should be noted that monocytes are a major cytomegalovirus (CMV) target cells in vivo, and are responsible for dissemination of the virus throughout the body during acute and late phase of infection.¹⁸ As shown in Fig. 1(e), transfection of human monocytes with a synthetic viral RNA derived from the UL37 gene or infection with CMV induced IDO. Notably, a huge amount of IDO protein was induced by CMV when compared to mock-treated cells.

Potential mechanism of TLR-induction of IDO in freshly isolated monocytes

In response to TLR, monocytes produce proinflammatory cytokines such as TNF-α, IL-1β and IL-6.¹³ In the present study, we have investigated the involvement of TNF-α in IDO induction by 27S and R848. Neutralizing antibodies against TNF-α reduced IDO induction by approximately 50 ± 5% (P < 0.002) as estimated by densitometric scanning (Fig. 1f). Thus, it seems likely that IDO induction by TLR7/8 signalling in freshly isolated monocytes is mediated in part by TNF-α in an autocrine pathway. Neutralizing antibodies against either IL-1β or IFN-α did not affect IDO induction in monocytes (data not shown).

Expression of IDO in immature monocyte-derived DC

Previous studies have shown that DC can up-regulate the expression of IDO in responses to several signals such as interferon-γ.¹⁴ Therefore, we explored in additional experiments whether monocyte-derived DC up-regulate the IDO expression in response to isRNAs. In these experiments, freshly isolated monocytes were treated with IL-4 and GM-CSF for 5 days. Subsequently, they were treated with 27S molecule for 16 hr. Surprisingly, we could not detect any induction of IDO in imoDC obtained from several donors (Fig. 2a, shows a representative example). To assess whether the IDO protein levels were reflected at mRNA levels, RT–PCR was performed. As shown, a large amount of IDO mRNA was detected in only 27S-treated monocytes, but not in controls or 27-treated imoDCs (Fig. 2a lower panels).

The kinetics of IDO gene silencing in imoDC were measured by coincubating monocytes with IL-4 and GM-CSF for various times points and then treating the cells with R848 in order to trigger IDO expression (Fig. 2b). By one day incubation with IL-4 and GM-CSF prior to treatment with R848, nearly 95% inhibition of IDO gene expression had already occurred in IL-4/G-CSF-treated monocytes compared to their untreated counterparts (day 0). Thus, the induction of a DC maturation programme is more likely to be responsible for IDO gene silencing. It should noted that cotreatment of monocytes with isRNAs along with IL-4 and/or GM-CSF did not block IDO up-regulation (data not shown).

Monocyte-derived DC express STAT-1

Among the transcription factors that activate IDO expression in DC are STAT-1 and IFR-1. We have previously shown that signalling via TLR in PBMC in response to isRNA induced the activation of several transcription factors, including STATs and IRFs.¹⁹ Therefore, we have asked whether a comparable effect occurred when monocytes are activated with isRNAs. Again, a significant induction of IDO was seen in monocytes, but not in imoDC (Fig. 2c). Interestingly, CpG oligo 2006 (TLR9 ligand) also induced IDO expression in monocytes. To examine the expression of STAT-1, the nitrocellulose membrane was stripped and then probed with polyclonal antibodies against STAT-1. As shown, imoDC expressed a huge amount of STAT-1 protein. Monocytes expressed very low levels of STA-1 protein. Collectively, these data suggest that the suppression of IDO expression in human imoDC is not caused by the absence of a key transcription factor such as STAT-1, but it is more likely the result of alteration in chromatin structures, which needs further investigation.

TLR7/8 signalling does not induce IDO expression in mature monocyte-derived DCs

Recent reports reveal that IDO expression is up-regulated in response to cytokines and IFN-γ. Also, its expression can be induced by ligands (e.g. CD28, CTLA4) or antibodies that bind and cross-link surface molecules expressed by antigen-presenting cells (APC) such as CD80/CD86.^20,21 Also, regulatory T cells (T reg) have been shown to induce IDO expression in target cells via cell surface expression of CTLA4.²² To evaluate TLR7/8 signalling on IDO expression in DC, monocytes were differentiated into mature DC (mmoDC) and then treated with 27S molecule, R848 or IFN-γ(Fig. 3a). Again, TLR7/8 ligands such as R848 and 27S did not induce IDO expression in mmoDC but IFN-γ did.

To assess whether TLR7/8-induced IDO pathway is functional in imoDC, monocytes were differentiated into imoDC and mmDC and then IDO expression was investigated by Western blots (Fig. 3b). Similar to mmDC, imoDC did not up-regulate IDO in response to R848 ligand, whereas addition of IFN-γ to cultures had a strong effect on IDO expression. It should be noted that the incubation of imoDC or mmoDC with TLR7/8 ligands induced cytokines production, including TNF-α and IFN-α, thus confirming the functionality of the TLR7/8 signalling pathways in these cells.¹⁰

Having demonstrated that the expression of IDO in imoDC or mmoDC was not under the TLR7/8 signalling pathways, in the next experiments we have tested the effects of TLR7/8 ligands on IDO expression in fleshly isolated blood myeloid (m) and plasmacytoid (p) DCs. Although a small up-regulation of IDO expression was seen in mDC treated with either R848 or 27S, only IFN-γ significantly up-regulated the expression of IDO in mDC compared to untreated cells (Figs 3b, P < 0.002). Also, pDC responded only to IFN γ, but the effect was very low (Fig. 3c).

2′-O-modified isRNAs suppress isRNA-induced IDO expression in monocytes

As shown in Fig. 1(c), the 27F molecule did not induce IDO expression. Also, our previous data showed that most of the genes, if not all, that are induced by isRNA can be abrogated by 2′-modifications of isRNAs.¹⁹ In the next experiments, we have evaluated whether 2′-modified isRNAs would suppress the production of TNF-α and/or IDO expression elicited by their unmodified version. When 2′-modified isRNA were cocomplexed with unmodified isRNA 27S using DOTAP, the effects on IDO and TNF-α expression were significantly reduced, with 2′-O-methyl-bearing RNA (27M) being the most effective inhibitors (Fig. 4a, b). Indeed, TNF-α and IDO expression were completely inhibited. These observations suggest that 2′-O-methyl modified RNAs can compete with isRNA to bind TLR7/8. It should be noted that most, if not all, the downstream signalling events activated by various isRNA sequences are inhibited by inhibitory 27M molecule at a very low molar ratio, indicating that 27M has a high affinity to TLR7/8 and can function as a general TLR7/8 antagonist (Sioud M. unpublished data).

Effects of IDO⁺ monocytes on T-cell activation

As induction of IDO activity has recently been described as a major T-cell inhibitory mechanism in APC/T-cell interaction.¹⁵ we next investigated the effects of IDO positive monocytes on autologous T cell activation in response to the mitogen PHA. In these experiments, T cells were grown in X-vivo-15 medium (Cambrex, Verviers, Belgium) without serum supplementation. Proliferation of T cells was assessed by flow cytometry as CFSE^1ow in percentages of CD4⁺ cells. As shown in Fig. 5, coculture of T cells with IDO⁺ monocytes did reduce T-cell activation with PHA by 50% (± 10%, P < 0.05), whereas IDO negative monocytes did not. The addition of 1-methyl-tryptophan (1MT), a specific inhibitor of IDO activity, to the coculture suppressed the ability of IDO⁺ monocytes to suppress T-cell proliferation, suggesting that IDO is involved in suppressing T-cell response. Collectively, these findings support the notion that IDO up-regulation in monocytes may regulate T-cell function in vivo.