Higher activation of TLR9 in plasmacytoid dendritic cells ... - CiteSeerX

Uncorrected Version. Published on July 20, 2009 as DOI:10.1189/jlb.0509314

Meeting Article

Higher activation of TLR9 in plasmacytoid dendritic cells by microbial DNA compared with self-DNA based on CpG-specific recognition of phosphodiester DNA Christoph Coch,*,1 Nicolas Busch,*,1 Vera Wimmenauer,* Evelyn Hartmann,† Markus Janke,* Mona Mohamed Ahmed Abdel-Mottaleb,‡ Alf Lamprecht,‡ Janos Ludwig,* Winfried Barchet,* Martin Schlee,*,1 and Gunther Hartmann*,1,2 *Institute of Clinical Chemistry and Pharmacology, and †Clinic and Policlinic for Otolaryngology/Ear, Nose and Throat Surgery, University Hospital of Bonn, Bonn, Germany; and ‡Laboratory of Pharmaceutical Engineering and Biopharmaceutics, Institute of Pharmacy, University of Bonn, Bonn, Germany RECEIVED MAY 11, 2009; REVISED JUNE 11, 2009; ACCEPTED JUNE 16, 2009. DOI: 10.1189/jlb.0509314

ABSTRACT

Introduction

TLR9 detects DNA in endolysosomal compartments of human B cells and PDC. Recently, the concept of the CpG motif specificity of TLR9-mediated detection, specifically of natural phosphodiester DNA, has been challenged. Unlike in human B cells, CpG specificity of natural phosphodiester DNA recognition in human PDC has not been analyzed in the literature. Here, we found that the induction of IFN-! and TNF-! in human PDC by phosphodiester ODNs containing one or two CG dinucleotides was reduced to a lower level when the CG dinucleotides were methylated and was abolished if the CGs were switched to GCs. Consistent with a high frequency of unmethylated CG dinucleotides, bacterial DNA induced high levels of IFN-! in PDC; IFN-! was reduced but not abolished upon methylation of bacterial DNA. Mammalian DNA containing low numbers of CG dinucleotides, which are frequently methylated, induced IFN-! in PDC consistently but on a much lower level than bacterial DNA. For activation of PDC, phosphodiester ODNs and genomic DNA strictly required complexation with cationic molecules such as the keratinocyte-derived antimicrobial peptide LL37 or a scrambled derivative. In conclusion, we demonstrate that self-DNA complexed to cationic molecules activate PDC and thus, indeed, may function as DAMPs; nevertheless, the preference of PDC for CpG containing DNA provides the basis for the discrimination of microbial from self-DNA even if DNA is presented in the condensed form of a complex. J. Leukoc. Biol. 86: 663– 670; 2009.

TLRs are involved in the innate recognition of foreign material, and their activation leads to innate and adaptive immune responses directed against invading pathogens [1]. Among the four members of the TLR family that are responsible for the detection of nucleic acids in the endolysosomal compartment of immune cells, TLR3, TLR7, and TLR8 detect RNA, and TLR9 detects DNA [1–5]. As DNA oligonucleotides are rapidly degraded by nucleases in cell culture, most studies are performed with phosphorothioatemodified DNA [6 – 8]. However, the phosphorothioate backbone increases the nonspecific protein binding of DNA. Thus, the identification of the optimal sequence motif (CpG motif, GTCGTT, C unmethylated) for TLR9 was performed using natural phosphodiester DNA at high concentrations for the activation of primary human B cells [9]. Much less is known about phosphodiester DNA-induced activation of the second human immune cell subset that expresses TLR9, the PDC. In contrast, activation of human PDC with CpG-containing phosphorothioate oligonucleotides is well-established. As a monomer, phosphorothioate oligonucleotides containing CpG motifs (CpG-B, for example, ODN 2006 [10]) activate PDC (up-regulation of CD86, secretion of TNF-! and IL-6) but induce only marginal amounts of IFN-!. CpG-containing phosphorothioate oligonucleotides designed to form nucleic acid-based nanoparticles spontanously as a result of polyG motifs at both ends and a palindromic sequence in the center (CpG-A, for example, ODN 2216 [11]) stimulate the maximum production of IFN-! in PDC [12]. Similarly, high levels of IFN-! are induced by CpG-B bound to different types of nanoparticles with cationic surface (polystyrol, protamin [13]), suggesting

Abbreviations: BDCA!blood DC antigen, DAMP!damage-associated molecular pattern, DC!dendritic cell, DOTAP!N-[1-(2,3-dioleoyloxy)]-N,N,Ntrimethlammonium propane methylsulfate, HEK!human embryo kidney, ODN!oligodeoxynucleotide, PDC!plasmacytoid dendritic cell(s), TE buffer!Tris-HCL/EDTA buffer

0741-5400/09/0086-663 © Society for Leukocyte Biology

1. These authors contributed equally to this work. 2. Correspondence: Institut fu¨r Klinische Chemie und Pharmakologie, Universita¨tsklinikum Bonn, Sigmund-Freud-Str. 25, D-53127 Bonn, Germany. E-mail: [email protected]

Volume 86, September 2009

Journal of Leukocyte Biology

Copyright 2009 by The Society for Leukocyte Biology.

663

that particles are required for full signaling of the pathway responsible for IFN-! release. Of note, CpG-A contains two, respectively, five phosphorothioate modifications at the 5" and 3" end (polyG region) of the oligonucleotide, and all other nucleotides are phosphodiester. A complete phosphorothioate backbone abolished the activity of ODN 2216, as nanoparticle formation is hindered; furthermore, the change of the two CG dinucleotides to GC dinucleotides completely eliminated the IFN-!-inducing activity of ODN 2216, demonstrating CpG specificity within the central phosphodiester backbone of ODN 2216 in PDC [12]. In the light of recent evidence that self-DNA, under certain circumstances, is recognized by B cells and PDC and contributes to autoimmune disease [14], the analysis of the effects of natural phosphodiester DNA gains importance. For example, chromatin-antibody immune complexes were shown to activate B cells expressing surface IgM rheumatoid factors (anti-IgG) by a mechanism involving BCR-mediated translocation to endosomes, where self-DNA engages TLR9 [15]. Furthermore, chromatin-containing immune complexes activate DC by dual engagement of FcRIII (mouse) or FcRII (human) and TLR9 [16 –18]. Furthermore, it was reported that TLR9 activation by mammalian self-DNA can be achieved by complexing DNA to cationic lipids, such as cationic lipid DOTAP, or cationic proteins or peptides, such as protamine [19 –21]. As in mammalian self-DNA, the frequency of CG dinucleotides is suppressed to 1:60, and most of the C in CG dinucleotides is methylated; CpG specificity may be less dominant than proposed previously [7] or may even be a result of the deoxyribose sugar recognition, as suggested in recent work [22, 23]. Here, we provide evidence that not only activation of human B cells but also human PDC, despite the requirement of complexation, is highly CG-specific. Stimulation of PDC was highest for phosphodiester ODNs containing unmethylated CG dinucleotides and much lower for ODNs containing methylated CG dinucleotides. We found that mammalian DNA indeed activated PDC but much less than bacterial DNA, containing a high frequency of unmethylated CG dinucleotides, and that the activity of bacterial DNA is reduced upon methylation. Thus, TLR9 in PDC seems capable of detecting self-DNA but still retains a preference for microbial DNA.

MATERIALS AND METHODS

ODNs and DNA Phosphodiester and phosphorothioate ODNs were purchased from Metabion (Martinsried, Germany). Lowercase letters indicate phosphorothioate linkage and capital letters phosphodiester linkage 3" of the base. (M) indicates methylated cytosin: CpG 2216 5"-ggGGGACGATCGTCgggggG-3"; CpG 2080 5"-TCGTCGTTCCCCCCCCCCCC-3"; 2080 GC 5"-TGCTGCTTCCCCCCCCCCCC-3"; CpG 2710 5"-AAAAGTCGTTAAAAAAAAAA-3"; 2710 GC 5"-AAAAGTGCTTAAAAAAAAAA-3"; CpG 2710 methylated 5"AAAAGTC(M)GTTAAAAAAAAAA-3"; ODN C20 (Control) 5"-CCCCCCCCCCCCCCCCCCCC-3". Genomic DNA was purified by extraction with phenol-chloroform-isopropanol (25/24/1) and ethanol precipitation (Escherichia coli, leukocytes, HEK293 cells) or purchased from BioChain (Hayward, CA, USA; leukocytes) and Invivogen (San Diego, CA, USA; endotoxinfree E. coli DNA).

664 Journal of Leukocyte Biology


Preparation, isolation, and culture of cells Peripheral blood buffy coats from healthy human donors were obtained from the local blood bank. PBMC were obtained by Ficoll-Hypaque density gradient centrifugation (Biochrom, Berlin, Germany). PDC were isolated by MACS using the BCDA-4 DC isolation kit from Miltenyi Biotec (Miltenyi Biotec, Bergisch Gladbach, Germany). Briefly, PDC were labeled with antiBDCA-4 antibody coupled to colloidal paramagnetic microbeads and passed through two magnetic separation columns (first LS column, second MS column; Miltenyi Biotec). The purity of isolated PDC (lineage-negative, MHC-II-positive, and CD123-positive cells) using this technique was above 95%. Cells were cultured in 96-well flat-bottom wells in duplicates at a density of 400,000 PBMC/well or 40,000 PDC/well in RPMI 1640 (Biochrom), supplemented with 10% (v/v) heat-inactivated FCS (Invitrogen, Karlsruhe, Germany), 1 mM L-glutamine, 100 U/ml penicillin, and 100 "g/ml streptomycin (all from Sigma-Aldrich, Munich, Germany). PDC were treated additionally with 5 ng/ml IL-3 (Peprotech, London, UK). Chloroquine (5 "g/ml; Sigma-Aldrich) was used to block endosomal maturation and acidification. HEK293 cells were grown in DMEM (Biochrom) with the same supplements as described above.

Isolation of genomic DNA HEK293 cells and PBMCs were harvested and resuspended in TE buffer (pH 8.0). After 2 h incubation (55°C) with Proteinase K (200 "g/ml; Fermentas, St. Leon-Rot, Germany) and SDS (0.5%), two extractions with chloroform-phenol-isopropanol (Roth, Karlsruhe, Germany; 25/ 24/1) were performed. The solution was treated with RNaseA (25 "g/ ml; Qiagen, Hilden, Germany) for 1 h and precipitated with ice-cold ethanol (Roth). DNA was resuspended in TE buffer, and concentration was measured using NanoDrop (Themo Fischer Scientific, Wilmington, DE, USA). The ratio of 280/260 nm absorbance was between 1.8 and 2.0. Endotoxin contamination of DNA samples was below 1 pg/ml, as tested by comparison with a LPS dose response of IL-6 induction in PBMC (LPS from Sigma-Aldrich).

Methylation of genomic DNA DNA was methylated using M.SssI CpG methyltransferase (New England Biolabs, Ipswich, MA, USA). DNA (5 "g) was mixed with 0.5 "l S-adenosylL-methionin, 2 "l NEBuffer II, 20 units M.SssI in a total volume of 20 "l and incubated for 3 h at 37°C, followed by inactivation of methyltransferase by incubating at 65°C for 20 min. Control DNA was treated equally without adding M.SssI CpG methyltransferase.

Stimulation of cells DNA was added into the supernatant of the cells with or without previous complexation to DOTAP (Roth), as recommended by the manufacturer. In brief, DNA and DOTAP were first diluted separately in OptiMEM (Biochrom) and then mixed gently and incubated for 10 min. The ratio of DNA to DOTAP was always kept 1:10. For LL37 complexation, DNA (2 "g) was preincubated for 30 min with LL37 (10 "g; sequence: LLGDFFRK SKEKIGKEFKRIVQRIKDFLRNLVPRTES) or the scrambled control LL37 scr (GLKLRFEFSKIKGEFLKTPEVRFRDIKLKDNRISVQR; both Innovagen, Lund, Sweden) in RPMI in a final volume of 12 "l before addition to cells. All incubations were carried out in duplicates.

ELISA After the indicated time-points, cell culture supernatants were collected and stored at –20°C. The IFN-! module set Bender MedSystems (Graz, Austria; BMS216MST) was used to measure IFN-!. This ELISA detects most isoforms of IFN-! at a detection range of 8 –500 pg/ml. Concentrations of TNF-! and IL-6 were detected using the BD OptEIA TNF and IL-6 ELISA set (BD Biosciences, Heidelberg, Germany). All ELISA procedures were performed according to the manufacturer’s recommendations.

www.jleukbio.org

Coch et al. CpG-specific DNA recognition in PDC

Flow cytometry Flow cytometric data were acquired on a LSR II flow cytometer (BD Biosciences). Human B cells were identified by positive staining with antiCD19-allophycocyanin. Activated B cells were measured by staining with anti-CD86-PE (antibodies obtained from BD Biosciences). Data were analyzed using FlowJo software (Version 8.7.3; Tree Star, Stanford, CA, USA).

Analysis of particle size The size of lipid complexes was analyzed by photon correlation spectroscopy (particle size analyzer, Brookhaven Instruments Corp., Holtsville, NY, USA) at a fixed angle of 90° at 25°C. The DNA/DOTAP liposomal suspension was diluted with distilled water before the analysis. Table 1 shows the means and sd of five repeated measurements.

DNA incorporation to DOTAP complexes DNA binding to DOTAP was determined as described by Marty and colleagues [24]. In brief, ODN (2.55 "M) was incubated with increasing concentrations of DOTAP (5– 40 "M) or without DOTAP in OptiMEM. After 10 min, the maximum UV absorbance for each DOTAP concentration was determined using the UV spectrophotometer UV mini-1240 (Shimadzu, Duisburg, Germany) with a 1-ml cuvette and 1-cm path length. In addition, DNA incorporation was analyzed by the degree of protection of DNA against DNase I-mediated degradation as described previously [25]. Briefly, genomic DNA was complexed with DOTAP and subsequently exposed to DNase I for 15 min. DNase I enzymatic activity was stopped by EDTA and heating. Heparin was added, and DNA was analyzed on a 0.6% agarose gel.

Statistical analysis

published studies, CpG motif-containing oligonucleotides are phosphorothioate-modified to render oligonucleotides more stable. Recent publications suggest that CpG specificity only applies to phosphorothioate-modified oligonucleotides but not to DNA with the natural phosphodiester backbone [22]. In our previous study, the identificaton of the human CpG motif was achieved by using unmodified phosphodiester oligonucleotides; activation with phosphodiester CpG ODN was found to require at least three repeated additions of high concentrations (30 "g/ml) of CpG ODN [9]. To expand on these earlier findings, we examined human B cell activation upon stimulation with repeated additions of high concentrations (30 "g/ml) of a phosphodiester oligonucleotide containing the human CpG motif (2080) on a polyC background and the corresponding polyC control oligonucleotide. The CpG oligonucleotide induced strong activation (CD86 expression) of CD19$ B cells, whereas the polyC control oligonucleotide showed weak activity (Fig. 1). Similar results were obtained with MHC II expression on human B cells (not shown in Fig. 1). These data demonstrate that human B cells do not require complexation or enhanced delivery to respond to CpG oligonucleotides and together with our earlier findings [9], confirm that this recognition is largely CpG-dependent, even with natural phosphodiester DNA.

Activation of human PDC requires complexation of phosphodiester CpG DNA

Statistical significance of differences was determined by the paired twotailed Student’s t-test. Differences were considered statistically significant for P # 0.05. In case of multiple comparisons, the global null-hypothesis test was performed by the nonparametric Kruskal-Wallis test, followed by Mann-Whitney U tests for planned pairwise comparisons. The adjustment for multiple testing was done by applying the Bonferroni-Holm method. Statistical analyses were performed using Microsoft Excel software (Microsoft Corp., Seattle, WA, USA) and GraphPad Prism (Version 4, GraphPad Software, Inc., La Jolla, CA, USA).

RESULTS

Activation of human B cells in response to repeated addition of high concentrations of phosphodiester CpG ODNs Unmodified phosphodiester oligonucleotides are degraded rapidly by nucleases in primary cell culture conditions. In most

Unlike in human B cells, a detailed analysis of the requirement of unmethylated CpG motifs in DNA for TLR9 activation in human PDC has not been performed. Primary human PDC isolated from PBMC were stimulated with repeated additions of high concentrations (3%30 "g/ml, such as in B cells) of CpG 2080 and CpG 2710 (one CpG motif in phosphodiester polyA). Alternatively, CpG 2080 and CpG 2710 were complexed with the cationic lipid DOTAP at an optimal concentration, as determined by dose-response experiments. Activation of PDC was analyzed by the production of IFN-! and TNF-!. Unlike in B cells, in the absence of complexation, activation of PDC was weak and not CpG-dependent (Fig. 2A). In contrast, both CpG oligonucleotides but not the polyC control oligonucleotide (CpG specificity confirmed in subsequent experiments by the appropriate non-CpG control oligonucleo-

TABLE 1. Size of DOTAP-DNA Complexes Measured by Photon Correlation Spectroscopy (n!5)

Complexes with ODN

Complexes with genomic DNA

Controls

www.jleukbio.org

Sample

Mean particle size (nm)

Half width (nm)

Polydispersity

CpG 2080 2080 GC CpG 2710 2710 GC 2710 methyl C20 DNA E. coli E. coli methyl HEK293 HEK293 methyl PBL Medium alone

916.8 & 26.1 810.2 & 6.9 685.9 & 5.4 738.9 & 276.3 711.0 & 17.1 687.7 & 18.4 620.7 & 19.0 772.9 & 10.4 656.4 & 15.2 759.4 & 19.3 893.6 & 28.0 0.3 & 0.3

447.2 & 38.7 232.3 & 70.8 229.2 & 48.1 276.3 & 39.7 276.0 & 60.9 367.4 & 13.2 313.5 & 14.7 349.9 & 25.5 178.3 & 58.0 330.6 & 29.2 389.1 & 80.7 0.2 & 0.2

0.243 & 0.036 0.111 & 0.058 0.130 & 0.035 0.154 & 0.038 0.176 & 0.049 0.286 & 0.012 0.285 & 0.010 0.210 & 0.029 0.113 & 0.056 0.197 & 0.032 0.232 & 0.063 0.101 & 0.101



665

B 250K

200K

200K

150K

150K

100 80

4.67

100K

100K 50K

50K

80.4 50K

100K

150K

200K

Medium

60

p=0 002 p=0.002

40

0

250K

0 102

FSC

103

104

2

105

CD19

*

Control

20

0 10

0

0 0

cell c count

250K

SSC

SSC

A

Medium

3

10

CD86

4

10

p=0.004

5

10

CpG 2080

Figure 1. B cells within PBMCs are activated by uncomplexed phosControl phodiester CpG DNA. PBMCs were incubated in 96 wells (400,000 0 2500 5000 7500 10000 CpG 2080 cells/well) and stimulated with CpG ODN 2080, C20 DNA (Control; CD86 PE - MFI 30 "g/ml added three times at 0 h; 12 h; 24 h), or medium alone. (A) After 48 h, cells were harvested and gated for lymphocytes and CD19-positive cells. SSC ! Side-scatter; FSC ! forward-scatter. (B) CD19-positive cells were analyzed for CD86 expression. A histogram of one representative donor out of six is shown [CD86 mean fluorescence intensity (MFI) of six donors in three independent experiments is shown as mean values&sem]. *, Adjusted ! niveau for Bonferroni-Holm: P ! !/2.

tides) induced strong PDC activation when complexed with DOTAP (Fig. 2A). Activation of PDC was mediated by a mechanism that requires endosomal maturation, as activation was abolished in the presence of chloroquine (Fig. 2B).

Activation of human PDC by complexed phosphodiester DNA is CpG-specific CpG specificity is confirmed by the loss of activity of a CG containing DNA when CG is switched to GC, or the Cs in CG dinucleotides are methylated. We examined the effect of switching CG to GC and of C methylation on the activation of PDC by CpG 2710 and CpG 2080 complexed with DOTAP over a wide range of concentrations. We found that elimination of the CG (GC CpG 2710) abolished IFN-! induction completely, whereas the methylation of the C (CpG

2710 Methyl) did not abolish but strongly reduced IFN-! induction (Fig. 3A). The induction of IFN-! by CpG 2710 and CpG 2080 was dose-dependent but absent for the corresponding GC control oligonucleotides, even at high concentrations (Fig. 3B). At all concentrations, methylation strongly reduced but did not abolish IFN-! induction (Fig. 3C). The determination of the size of DOTAP-DNA complexes by photon correlation spectroscopy and the analysis of DNA incorporation in lipid complexes confirmed that CpG specificity seen was not a result of CpG-dependent differences in size or DNA content of DOTAP complexes (Tables 1 and 2 and Materials and Methods). Together, these results demonstrate the high degree of CpG specificity of PDC activation by phosphodiester oligonucleotides.

Figure 2. Natural phosphodiester CpG A Medium Medium DNA complexed to DOTAP induces cytoControl Control kine expression in PDCs. (A) PDC were p 2080 - Dotap p CpG cultured in 96 wells (40,000 cells/well) for - Dotap CpG 2080 48 h. Cells were stimulated three times (at CpG 2710 CpG 2710 0 h; 12 h; 24 h) with 30 "g/ml uncomControl Control * p = 0,009 plexed C20 DNA (Control), CpG 2080, or p = 0,004 * + Dotap CpG 2080 p = 0,0152 + Dotap CpG 2080 p = 0,026 CpG 2710 or with 800 ng/ml of the same CpG 2710 CpG 2710 DNA complexed to DOTAP (added once at CpG 2216(3µg/ml) -DotapCpG 2216(3µg/ml) 0 h). Additionally, cells were stimulated 0 20 40 60 80 with medium alone or CpG 2216 (3 "g/ml 0 300 600 900 1200 IFN-alpha (ng/ml) TNF-alpha (pg/ml) added at 0 h). Supernatants were harvested and analyzed for IFN-! and TNF-! expresB Control sion by ELISA. Six donors in three indeControl + Dotap pendent experiments are shown as mean CpG 2080 + Dotap - Chloroquin - Chloroquin CpG 2080 values & sem. *, Adjusted ! niveau for BonCpG 2710 ferroni-Holm: P ! !/2. (B) PDC were isoCpG 2710 lated using MACS separation and cultured Control Control in 96 wells (40,000 cells/well), with or with+ Dotap + Dotap + Chloroquin CpG 2080 out addition of chloroquine (5 "g/ml). CpG 2080 + Chloroquin Cells were incubated with C20 DNA (ConCpG 2710 CpG 2710 trol), CpG 2080, or CpG 2710 (800 ng/ml) 0 5 10 15 0 400 800 1200 complexed to DOTAP. After 24 h, supernaIFN alpha (ng/ml) IFN-alpha TNF-alpha (pg/ml) tants were harvested, and IFN-! and TNF-! expression was analyzed using ELISA. Results from seven donors in three independent experiments are depicted as means & sem.



www.jleukbio.org


A

B 40000

Medium

CpG 2080 CpG 2710

30000

C t l Control CpG 2710

p = 0,002

*

2710 GC CpG 2710 Methyl 10 20 IFN-alpha (ng/ml)

p=0,04

2710 GC 2080 GC Control

10000 0

0

p=0,01

20000

p = 0,041

C

0

1 2 3 4 5 Oligo-Concentration (µg/ml)

40000 CpG 2710

30000

p=0,05

20000

2710 M 10000 0

Control 0

1 2 3 4 5 Oligo-Concentration (µg/ml)

Figure 3. Elimination of the CpG motif abolishes 1200 1200 and methylation of CpG motif reduces PDC actiCpG 2080 vation. PDC were cultured in 96 wells (40,000 800 800 cells/well). (A) PDC were stimulated with DOTAPCpG 2710 CpG 2710 complexed C20 DNA (Control), CpG 2710, 2710 GC, methylated CpG 2710 (800 ng/ml), or me400 2080 GC 400 dium alone. After 24 h, supernatants were har2710 GC 2710 M Control vested, and IFN-! was analyzed using ELISA. ReControl 0 0 sults from six donors in three experiments are 0 1 2 3 4 5 0 1 2 3 4 5 Oligo-Concentration (µg/ml) depicted (&sem). *, Adjusted ! niveau for BonferOligo-Concentration (µg/ml) roni-Holm: P ! !/2. (B and C) PDC were incubated with increasing concentrations of DOTAP-complexed C20 DNA (Control), CpG 2080, CpG 2710, 2080 GC, 2710 GC, or 2710 M as indicated. After 24 h, supernatants were harvested and analyzed for IFN-! and TNF-! expression by ELISA. Data from four donors are shown as mean values & sem.

Higher and methylation-sensitive activation of human PDC by bacterial DNA compared with self-DNA Bacterial DNA contains CG dinucleotides at the expected statistical frequency (1:16), and most of the CGs are unmethylated, and in vertebrate DNA, the frequency of CG dinucleotides is suppressed (1:60), and there is a high degree of C methylation in CG dinucleotides. Based on the higher frequency of unmethylated CG dinucleotides, TLR9 is thought to preferentially detect foreign microbial DNA and not self-DNA. We isolated bacterial genomic DNA from E. coli and “self” genomic DNA from leukocytes or HEK293 cells. Similar to phosphodiester CpG oligonucleotides, genomic DNA only activated IFN-! production in PDC when complexed to DOTAP (Fig. 4A). With DOTAP complexation, we consistently detected induction of IFN-! by leukocyte- and HEK293 cell-derived DNA, but IFN-! induction by bacterial DNA was much higher (Fig. 4A). The activity of bacterial and self-DNA in PDC was blocked completely by chloroquine, confirming endosomal recognition in PDC (Fig. 4B). Methylation of genomic DNA strongly reduced but did not abolish the IFN-! induction by bacterial DNA (Fig. 4C). Methylation TABLE 2. Analysis of ODN Incorporation in DOTAP by UV Absorption UV absorbance max. DOTAP DOTAP DOTAP DOTAP DOTAP

5 "M 10 "M 20 "M 30 "M 40 "M

www.jleukbio.org

ODN 2710

ODN 2710 GC

ODN 2710 M

0,0072 0,1125 0,2272 0,2598 0,4183

0,0153 0,0774 0,1421 0,3249 0,4632

0,0157 0,0484 0,1482 0,2648 0,3988

did not change the level of DNA incorporation in DOTAP particles, as determined by the degree of protection against DNase I (Fig. 4D). These results show that activation of PDC by genomic DNA depends on the source (microbial higher than self) and the methylation status of the DNA and support the concept of CpG-specific recognition of genomic DNA.

CpG-specific and methylation-sensitive activation of human PDC by phosophodiester DNA complexed with the natural cationic antimicrobial peptide LL37 Although the cationic lipid DOTAP is an artificial complexation reagent, the cationic peptide LL37 is a natural antimicrobial peptide secreted by keratinocytes. It has been proposed that triggering of TLR9 by self-DNA bound to LL37 contributes to the pathogenesis of psoriasis [26]. We were interested whether CpG specificity is maintained in LL37-mediated activation of PDC, whether the source of genomic DNA has an impact on PDC activation, and whether the specific sequence of LL37 or solely its cationic properties are critical for mediating TLR9 activation. Oligonucleotides or genomic DNA were complexed with LL37 or with a peptide containing the same amino acids but in a random order (scrambled LL37, LL37 scr). Similar to DOTAP, we found that CpG ODN 2080 complexed to LL37 strongly induced IFN-! in PDC, and the GC control oligonucleotide had no activity (Fig. 5A). Bacterial DNA complexed to LL37 induced much higher levels of IFN-! than DNA prepared from leukocytes or HEK293 cells (self-DNA). The strong IFN-! induction by bacterial DNA and the lower IFN-! induction by self-DNA were abolished in the presence of chloroquine (Fig. 5B). As for LL37, complexation of bacterial DNA with the scrambled version of LL37 showed strong IFN-! inVolume 86, September 2009


667

A

B

Medium

E.Coli-DNA - Dotap HEK293-DNA

+ Dotap - Chloroquine

L k t DNA Leukocyte-DNA

Medium E.Coli-DNA

HEK293-DNA

+ Dotap

HEK293 -DNA

Leukocyte-DNA eu ocyte

Le koc te DNA Leukocyte-DNA

E.Coli-DNA

E.Coli-DNA + Dotap

C

Medium E.Coli-DNA

+ Dotap + Chloroquine

HEK293-DNA Leukocyte-DNA 0

20 30 IFN-alpha (ng/ml)

E.Coli-DNA

HEK293-DNA

+ Dotap Methylated DNA

Leukocyte-DNA 0

50

20 IFN-alpha (ng/ml)

40

HEK293 -DNA Leukocyte-DNA 0

20 IFN-alpha (ng/ml)

40

D

Figure 4. Activation of PDC by genomic E. coli DNA complexed to DOTAP depends on methylation. PDC were cultured in 96 wells (40,000 cells/well). (A) PDC were stimulated with different uncomplexed genomic DNAs (E. oli-, leukocyte-, HEK293-DNA; 8 "g/ml) or the same genomic DNA complexed with DOTAP (800 ng/ml). After 24 h, supernatants were harvested and checked for IFN-!. Data from four independent donors are shown as mean values & sem. (B) As described for A, but additionally, chloroquine (5 "g/ml) was added to block endosomal acification. (C) Isolated genomic DNAs (E. coli-, leukocyte-, HEK293-DNA) were methylated via M.SssI CpG methyltransferase or incubated in buffer without adding enzyme. After complexation to DOTAP, DNA was given to PDC (800 ng/ ml). After 24 h, supernatants were harvested and checked for IFN-! expression. The results from four individual donors are shown as mean values & sem. (D) Genomic E. coli DNA with and without previous methylation and with or without complexation to DOTAP was exposed to DNase I for 15 min. DNase I enzymatic activity was stopped by EDTA and heating. Heparin was added, and DNA was analyzed on a 0.6% agarose gel.

duction, which was decreased when DNA was methylated (Fig. 5C). Together, these data demonstrate that CpG specificity and the preference for microbial over self-DNA are maintained when DNA is complexed to the natural antimicrobial molecule LL37 and that not the specific sequence of LL37 but rather the possibility of complex formation as a result of the cationic charge of the peptide is responsible for mediating the activation of PDC.

of TLR9, provided the molecular basis for this pathway [27]. The progress in understanding the identity and the biological impact of DAMPs and alarmins [28 –30] and the appreciation of self-DNA as a DAMP have triggered a more recent discussion about the degree of CpG dependency of TLR9-mediated immunorecognition, specifically in the context of the natural phosphodiester DNA backbone [22, 23, 31]. Thus, the question of CpG specificity of TLR9-mediated activation is closely linked to the question of whether selfDNA released from damaged cells is capable of alerting the innate immune system of such damage. Here, we add to previous evidence from our group [9] and others [31] that activation of human B cells by phosphodiester DNA is CpGspecific; furthermore, we provide new evidence that activa-

DISCUSSION More than 10 years ago, Arthur Krieg and colleagues [7] established the concept of CpG-specific immunorecognition of DNA in the murine system and later on, the involvement

A

B

Medium

C

Medium LL-37

Control

E.Coli-DNA

Control

- LL-37

E.Coli-DNA

Leukocyte-DNA y

E.Coli-DNA methyl

HEK293-DNA

LL-37 + CpG 2080

+ LL-37

E.Coli-DNA + LL-37

2080 GC

Leukocyte-DNA

+ LL-37scr

HEK293-DNA 0

3 6 9 IFN-alpha(ng/ml)

12

Medium

E.Coli-DNA + LL LL-37 37 Leukocyte-DNA Leukocyte DNA +Chloroquine HEK293-DNA

E.Coli-DNA E.Coli-DNA methyl E.Coli-DNA E.Coli-DNA methyl 0

10

20

30

40

IFN-alpha (ng/ml)

Figure 5. E. coli complexed to the antimicrobial peptide LL37 activates PDC, which were cultured in 96 wells (40,000 cells/well). (A) CpG 2080, 2080 GC, or C20 DNA (8 "g/ml; Control) complexed to LL37 was added to PDC. After 24 h, the supernatant was harvested and checked for IFN-!. Data are depicted as mean values & sem from two donors. (B) Genomic E. coli-, leukocyte-, or HEK293DNA (8 "g/ml), complexed to LL37 or uncomplexed, was added to PDC in the presence or absence of chloroquine (5 "g/ml). Medium and LL37 alone served as controls. After 24 h, IFN-! was analyzed in the supernatants. The results from three donors are shown as mean values & sem. (C) Genomic DNA from E. coli was methylated by using M.SssI CpG methyltransferase. DNA was complexed with LL37 or the scrambled control LL37 scr and added to PDC. After 24 h, IFN-! was analyzed in the supernatants. Results from two donors are depicted. 0


5 10 15 IFN-alpha(ng/ml)


20

www.jleukbio.org


tion of human PDC, despite the requirement of complex formation of DNA for such activation, shows a preference for DNA containing unmethylated CG dinucleotides. Although elimination of CG dinucleotides abolished the activity, methylation of the C in CG dinucleotides only partially decreased the activity in PDC. Consistent with these findings, genomic microbial DNA (E. coli DNA containing high numbers of unmethylated CG dinucleotides) was highly active, whereas genomic self-DNA containing methylated CG dinucleotides was recognized by TLR9 but on a much lower level. The biological and possibly pathological impact of self-DNA recognition by TLR9 is highlighted by consistent IFN-! induction in PDC by self-DNA, in complex with the natural antimicrobial peptide LL37, the only human cathelicidin, which is a family of cationic antimicrobial peptides induced in damaged skin that provides a rapid broad-spectrum defense against infections by acting as natural antibiotics [32]. LL37 is found at surfaces such as skin and mucosa [33, 34] and is inducible upon stress or stimulation. Although the specific amino acid sequence of LL37 may be responsible for other biological and antimicrobial effects, we found that the specific sequence of LL37 has no major impact on DNA complexation and TLR9 activation. In the literature, CpG specificity has been questioned specifically for phosphodiester DNA complexed to cationic polymers. It was reasoned that the high local concentration of DNA in the complexed form in the endosomal compartment may overcome the requirement of CpG motifs in DNA [20, 21], that in this situation, the plain presence of the desoxyribose sugar backbone is the critical determinant of TLR9 activity, and furthermore, that CpG motifs in DNA are only required if the backbone is phosphorothioate DNA [22]. Our results do not support such conclusions, specifically the conclusion that CpG dependency of TLR9 activation is restricted to the artificial phosphorothioate backbone. However, notably, our findings do not contradict most of the results of Haas and colleagues [22]. The data presented in their work clearly demonstrate that base-free phosphodiester 2" deoxyribose homopolymers exhibit weak binding to TLR9, that this binding is strongly enhanced by adding any bases to this sugar backbone and is even more enhanced if CG dinucleotides are present, and furthermore, that these binding properties are reflected by the respective biological activity to induce IL-6 in DC [22]. Based on our findings, it is not the requirement of CG dinucleotides but the requirement of an unmethylated status of these CG dinucleotides for TLR9 activation that may have been overestimated in the past. Although full activation in our hands required the presence of unmethylated CG dinucleotides, partial activation in the presence of methylated CG dinucleotides as present in mammalian DNA (although at low frequency) explains why self-DNA indeed may function as DAMP, which is detected by TLR9. In the case of substantial damage, partial PDC activation by abundant self-DNA may be sufficient to activate the appropriate innate responses. Furthermore, autoimmune disease may be initiated by microbial DNA in the context of an infection. www.jleukbio.org

Subsequently, autoimmunity may be propagated by self-DNA released from cells that die in the course of such infection and later, die from autoimmune aggression. In this context, it is interesting to note that in B cells exposed to type I IFNs, CpG specificity is reduced and thus, self-DNA becomes more active [31].

ACKNOWLEDGMENTS This study was supported by grants Bundesministerium fu¨r Bildung und Forschung Biofuture 0311896, SFB 670, SFB 704, KFO115, and KFO177 to G. H. This work is part of the theses of N. B. at the University of Bonn. The authors thank Bastian Lu¨denbach for technical assistance.

REFERENCES

1. Akira, S., Uematsu, S., Takeuchi, O. (2006) Pathogen recognition and innate immunity. Cell 124, 783– 801. 2. Wagner, H. (2004) The immunobiology of the TLR9 subfamily. Trends Immunol. 25, 381–386. 3. Lamphier, M. S., Sirois, C. M., Verma, A., Golenbock, D. T., Latz, E. (2006) TLR9 and the recognition of self and non-self nucleic acids. Ann. N. Y. Acad. Sci. 1082, 31– 43. 4. Barchet, W., Wimmenauer, V., Schlee, M., Hartmann, G. (2008) Accessing the therapeutic potential of immunostimulatory nucleic acids. Curr. Opin. Immunol. 20, 389 –395. 5. Schlee, M., Hornung, V., Hartmann, G. (2006) siRNA and isRNA: two edges of one sword. Mol. Ther. 14, 463– 470. 6. Bauer, S., Kirschning, C. J., Hacker, H., Redecke, V., Hausmann, S., Akira, S., Wagner, H., Lipford, G. B. (2001) Human TLR9 confers responsiveness to bacterial DNA via species-specific CpG motif recognition. Proc. Natl. Acad. Sci. USA 98, 9237–9242. 7. Krieg, A. M., Yi, A. K., Matson, S., Waldschmidt, T. J., Bishop, G. A., Teasdale, R., Koretzky, G. A., Klinman, D. M. (1995) CpG motifs in bacterial DNA trigger direct B-cell activation. Nature 374, 546 –549. 8. Verthelyi, D., Klinman, D. M. (2003) Immunoregulatory activity of CpG oligonucleotides in humans and nonhuman primates. Clin. Immunol. 109, 64 –71. 9. Hartmann, G., Krieg, A. M. (2000) Mechanism and function of a newly identified CpG DNA motif in human primary B cells. J. Immunol. 164, 944 –953. 10. Hartmann, G., Weeratna, R. D., Ballas, Z. K., Payette, P., Blackwell, S., Suparto, I., Rasmussen, W. L., Waldschmidt, M., Sajuthi, D., Purcell, R. H., Davis, H. L., Krieg, A. M. (2000) Delineation of a CpG phosphorothioate oligodeoxynucleotide for activating primate immune responses in vitro and in vivo. J. Immunol. 164, 1617–1624. 11. Krug, A., Rothenfusser, S., Hornung, V., Jahrsdorfer, B., Blackwell, S., Ballas, Z. K., Endres, S., Krieg, A. M., Hartmann, G. (2001) Identification of CpG oligonucleotide sequences with high induction of IFN-!/# in plasmacytoid dendritic cells. Eur. J. Immunol. 31, 2154 –2163. 12. Kerkmann, M., Costa, L. T., Richter, C., Rothenfusser, S., Battiany, J., Hornung, V., Johnson, J., Englert, S., Ketterer, T., Heckl, W., Thalhammer, S., Endres, S., Hartmann, G. (2005) Spontaneous formation of nucleic acid-based nanoparticles is responsible for high interferon-! induction by CpG-A in plasmacytoid dendritic cells. J. Biol. Chem. 280, 8086 – 8093. 13. Kerkmann, M., Lochmann, D., Weyermann, J., Marschner, A., Poeck, H., Wagner, M., Battiany, J., Zimmer, A., Endres, S., Hartmann, G. (2006) Immunostimulatory properties of CpG-oligonucleotides are enhanced by the use of protamine nanoparticles. Oligonucleotides 16, 313–322. 14. Krieg, A. M., Vollmer, J. (2007) Toll-like receptors 7, 8, and 9: linking innate immunity to autoimmunity. Immunol. Rev. 220, 251–269. 15. Leadbetter, E. A., Rifkin, I. R., Hohlbaum, A. M., Beaudette, B. C., Shlomchik, M. J., Marshak-Rothstein, A. (2002) Chromatin-IgG complexes activate B cells by dual engagement of IgM and Toll-like receptors. Nature 416, 603– 607. 16. Boule, M. W., Broughton, C., Mackay, F., Akira, S., Marshak-Rothstein, A., Rifkin, I. R. (2004) Toll-like receptor 9-dependent and -independent dendritic cell activation by chromatin-immunoglobulin G complexes. J. Exp. Med. 199, 1631–1640. 17. Means, T. K., Latz, E., Hayashi, F., Murali, M. R., Golenbock, D. T., Luster, A. D. (2005) Human lupus autoantibody-DNA complexes activate DCs through cooperation of CD32 and TLR9. J. Clin. Invest. 115, 407– 417.



669

18. Rifkin, I. R., Leadbetter, E. A., Busconi, L., Viglianti, G., Marshak-Rothstein, A. (2005) Toll-like receptors, endogenous ligands, and systemic autoimmune disease. Immunol. Rev. 204, 27– 42. 19. Scheel, B., Teufel, R., Probst, J., Carralot, J. P., Geginat, J., Radsak, M., Jarrossay, D., Wagner, H., Jung, G., Rammensee, H. G., Hoerr, I., Pascolo, S. (2005) Toll-like receptor-dependent activation of several human blood cell types by protamine-condensed mRNA. Eur. J. Immunol. 35, 1557–1566. 20. Yasuda, K., Rutz, M., Schlatter, B., Metzger, J., Luppa, P. B., Schmitz, F., Haas, T., Heit, A., Bauer, S., Wagner, H. (2006) CpG motif-independent activation of TLR9 upon endosomal translocation of “natural” phosphodiester DNA. Eur. J. Immunol. 36, 431– 436. 21. Yasuda, K., Yu, P., Kirschning, C. J., Schlatter, B., Schmitz, F., Heit, A., Bauer, S., Hochrein, H., Wagner, H. (2005) Endosomal translocation of vertebrate DNA activates dendritic cells via TLR9-dependent and -independent pathways. J. Immunol. 174, 6129 – 6136. 22. Haas, T., Metzger, J., Schmitz, F., Heit, A., Muller, T., Latz, E., Wagner, H. (2008) The DNA sugar backbone 2" deoxyribose determines Toll-like receptor 9 activation. Immunity 28, 315–323. 23. Wagner, H. (2008) The sweetness of the DNA backbone drives Toll-like receptor 9. Curr. Opin. Immunol. 20, 396 – 400. 24. Marty, R., N’soukpoe´-Kossi, C. N., Charbonneau, D., Weinert, C. M., Kreplak, L., Tajmir-Riahi, H. A. (2009) Structural analysis of DNA complexation with cationic lipids. Nucleic Acids Res. 37, 849 – 857. 25. Moret, I., Esteban Peris, J., Guillem, V. M., Benet, M., Revert, F., Dasi, F., Crespo, A., Alino, S. F. (2001) Stability of PEI-DNA and DOTAP-DNA complexes: effect of alkaline pH, heparin and serum. J. Control. Release 76, 169 –181. 26. Lande, R., Gregorio, J., Facchinetti, V., Chatterjee, B., Wang, Y. H., Homey, B., Cao, W., Wang, Y. H., Su, B., Nestle, F. O., Zal, T., Mellman, I., Schroder, J. M., Liu, Y. J., Gilliet, M. (2007) Plasmacytoid dendritic



27. 28.

29. 30. 31.

32. 33.

34.

cells sense self-DNA coupled with antimicrobial peptide. Nature 449, 564 – 569. Hemmi, H., Takeuchi, O., Kawai, T., Kaisho, T., Sato, S., Sanjo, H., Matsumoto, M., Hoshino, K., Wagner, H., Takeda, K., Akira, S. (2000) A Toll-like receptor recognizes bacterial DNA. Nature 408, 740 –745. Lotze, M. T., Zeh, H. J., Rubartelli, A., Sparvero, L. J., Amoscato, A. A., Washburn, N. R., Devera, M. E., Liang, X., Tor, M., Billiar, T. (2007) The grateful dead: damage-associated molecular pattern molecules and reduction/oxidation regulate immunity. Immunol. Rev. 220, 60 – 81. Oppenheim, J. J., Tewary, P., de la Rosa, G., Yang, D. (2007) Alarmins initiate host defense. Adv. Exp. Med. Biol. 601, 185–194. Bianchi, M. E. (2007) DAMPs, PAMPs and alarmins: all we need to know about danger. J. Leukoc. Biol. 81, 1–5. Uccellini, M. B., Busconi, L., Green, N. M., Busto, P., Christensen, S. R., Shlomchik, M. J., Marshak-Rothstein, A., Viglianti, G. A. (2008) Autoreactive B cells discriminate CpG-rich and CpG-poor DNA and this response is modulated by IFN-!. J. Immunol. 181, 5875–5884. Zanetti, M. (2004) Cathelicidins, multifunctional peptides of the innate immunity. J. Leukoc. Biol. 75, 39 – 48. Paulsen, F., Pufe, T., Conradi, L., Varoga, D., Tsokos, M., Papendieck, J., Petersen, W. (2002) Antimicrobial peptides are expressed and produced in healthy and inflamed human synovial membranes. J. Pathol. 198, 369 – 377. Schauber, J., Rieger, D., Weiler, F., Wehkamp, J., Eck, M., Fellermann, K., Scheppach, W., Gallo, R. L., Stange, E. F. (2006) Heterogeneous expression of human cathelicidin hCAP18/LL-37 in inflammatory bowel diseases. Eur. J. Gastroenterol. Hepatol. 18, 615– 621.

KEY WORDS: IFN-! ! innate immune response

www.jleukbio.org