Human β-defensin-3 activates professional antigen-presenting cells via Toll-like receptors 1 and 2

Abstract

There is increasing evidence that innate and adaptive immune responses are intimately linked. This linkage is in part mediated through the recognition of conserved microbial products by Toll-like receptors (TLRs). Detection of microbial products by TLRs can result in induction of inflammatory cytokines and activation of professional antigen-presenting cells, thereby enhancing adaptive immune responses. Here, we show that human β-defensin-3 (hBD-3), an innate antimicrobial peptide, can induce expression of the costimulatory molecules CD80, CD86, and CD40, on monocytes and myeloid dendritic cells in a TLR-dependent manner. Activation of monocytes by hBD-3 is mediated by interaction with TLRs 1 and 2, resulting in signaling that requires myeloid differentiating factor 88 and results in IL-1 receptor-associated kinase-1 phosphorylation. In studies with HEK cells engineered to express various TLRs, we show that activation of NF-κB by hBD-3 depends on the expression of both TLR1 and TLR2. Thus, human TLR signaling is not restricted to recognition of microbial patterns but also can be initiated by host-derived peptides such as hBD-3.

Human β-defensins (hBDs) are antimicrobial peptides that are produced by epithelial cells (1–3). Some hBDs such as hBD-1 are constitutively expressed, whereas others (hBD-2 and hBD-3) are induced by microbial products, inflammatory cytokines, or epidermal growth factor (1, 2, 4–6). Several different functions have been described for hBDs that relate to host defense including direct killing of microorganisms (3, 5, 7) and chemoattraction of immature dendritic cells and memory T cells through interactions with CCR6 and chemoattraction of monocytes through interactions with an undefined receptor (5, 8, 9). The chemotactic properties of hBDs suggest a possible role for these peptides in bridging innate and adaptive immunity. Here, we consider another potential mechanism for hBDs to enhance adaptive immune responses by testing the hypothesis that hBD-3 also can induce the differentiation of professional antigen-presenting cells (APCs).

Innate host defenses can be activated by exposure to conserved microbial structures that interact with Toll-like receptors (TLRs) (10–12). Signaling through TLRs begins at cytoplasmic TIR (Toll/IL-1 receptor) domains and proceeds downstream via myeloid differentiating factor 88 (MyD88)-dependent or -independent pathways and subsequently through multiple kinases, including IL-1 receptor-associated kinase-1 (IRAK-1) (13). TLR signaling leads to activation of the transcription factors NF-κB and AP-1, resulting in the direct induction of inflammatory cytokine and costimulatory molecule expression by professional APCs.

In the present studies, we explore an intersection between the inducible innate host defense molecule, hBD-3, and TLR activation. Only a single previous observation provides indication of this type of interaction, describing the activation of TLR4 by mouse β-defensin-2. The lack of homology between mouse β-defensin-2 and any known human defensin, raises questions surrounding the importance of this observation in human biology. Our studies address this issue by demonstrating that hBD-3 can promote signaling via interaction with TLRs 1 and 2 and by demonstrating that this interaction plays an important role in human APC activation.

Results

hBD-3 Induces Costimulatory Molecule Expression on APCs.

To determine whether hBD-3 could induce expression of costimulatory molecules on APCs, peripheral blood mononuclear cells (PBMCs) were incubated overnight in the presence of recombinant hBD-3 (rhBD-3). Incubation with rhBD-3 resulted in increased surface expression of CD80, CD86, and CD40 on myeloid dendritic cells (mDCs) but not on plasmacytoid dendritic cells (pDCs) (Fig. 1A). Monocyte expression of CD80, CD86, and CD40 was also enhanced (Fig. 1B), but there was no detectable effect of hBD-3 on B lymphocytes (data not shown). Although a modest increase in costimulatory molecule density could be seen on monocytes after 8 h of incubation with rhBD-3, peak induction was observed at 18 h and remained elevated for as long as 3 days (data not shown). Concentrations of rhBD-3 as low as 5 μg/ml (≈1 μM) could induce increases in costimulatory molecule expression and this response peaked at 20 μg/ml (3.8 μM). Purified CD14⁺ monocytes also responded to hBD-3 by increasing surface expression of costimulatory molecules and the magnitude of the response was similar to that seen in PBMC preparations (data not shown). Thus, hBD-3 induces costimulatory molecule expression on selected subsets of professional APCs and mediates its effects on monocytes by directly activating these cells.

Fig. 1.

Overnight culture with hBD-3 results in increased expression of CD80, CD86, and CD40 on mDCs and monocytes, but not on pDCs. PBMCs were cultured overnight in medium alone (RPMI 1640 plus 10% human AB serum) or in medium supplemented with hBD-3 (20 μg/ml). (A) Histograms show expression levels of CD80, CD86, and CD40 on mDCs and pDCs, which were identified by gating on cells that lacked lineage markers (CD14, CD19, CD3, CD56, or CD16) and coexpressed HLA-DR and CD11c (mDCs) or HLA-DR and CD123 (pDCs). (B) The MFI of the indicated markers was measured on CD14⁺ monocytes. Each symbol represents an experiment performed by using cells from a different donor. The mean shifts in surface expression of CD80, CD86, and CD40 were significantly different (P < 0.001, paired t test) following overnight culture with hBD-3. (C) PBMCs were incubated overnight in medium or in medium treated with rhBD-3 or shBD-3 (20 μg/ml), and expression of costimulatory molecules was measured on CD14⁺ cells. Results are representative of five experiments.

Concentrations of endotoxin in the rhBD-3 preparations were <0.5 ng/mg, the lower limit of detection in the assay system. The minimum concentration of endotoxin required to induce comparable expression of CD80 in this system was 20 ng/ml (data not shown), roughly 1,000-fold greater than the possible levels of endotoxin within the rhBD-3 preparations. In addition, hBD-3 that was chemically synthesized [synthetic hBD-3 (shBD-3)] also induced comparable costimulatory molecule expression on monocytes (Fig. 1C) and mDCs (data not shown), suggesting that contamination with microbial products does not account for this effect.

hBD-3 Activation of Monocytes Occurs Through the Signaling Molecules MyD88 and IRAK-1.

We next wanted to identify the pathway by which hBD3 mediated its activity. Previous studies indicate that hBD-3 can interact with the chemokine receptor CCR6, or other G protein-coupled receptors, to promote chemotaxis (8, 9, 14, 15). Alternatively, hBDs might interact with TLRs because murine β-defensin-2 activates mouse dendritic cells via a TLR4-dependent mechanism (16). To explore these possibilities, we evaluated the ability of hBD-3 to induce APC maturation in the presence of a JAK2 inhibitor (AG490; 50 μM; Sigma–Aldrich, St. Louis, MO) that interferes with signaling through CCR6 and other G protein-coupled receptors, or an inhibitor of MyD88 (MyD88I; 50 μM; Imgenex, San Diego, CA), an element required for signaling by most TLRs (17–19). In three experiments (Fig. 2), AG490 had minimal and nonsignificant effects on hBD-3-induced costimulatory molecule expression (mean ± SEM, 14 ± 9% inhibition; P = 0.27) even though the same concentration of AG490 did block induction of CD80 in monocytes stimulated with 20 μM SDF-1 [Δmean fluorescence intensity (MFI), 819 and 209 in cells incubated in the absence or presence of AG490, respectively; data not shown]. Preincubation with MyD88I, however, almost completely inhibited the induction of CD80 expression by hBD-3 (84 ± 6% inhibition; P = 0.008). MyD88I also significantly inhibited CD80 induction by LPS (85% ± 5% inhibition; P = 0.01), but had minimal and nonsignificant effects on costimulatory molecule induction by IFN-α (17 ± 10% inhibition; P = 0.3). Thus, hBD-3 relies on signaling via MyD88- but not Jak2-dependent pathways to induce CD80 expression, suggesting a central role for TLR signaling in this induction of APC differentiation (13).

Fig. 2.

MyD88I inhibits induction of CD80 expression by hBD-3. PBMCs were cultured overnight in medium alone or in medium supplemented with hBD-3 (20 μg/ml), LPS (20 ng/ml), or IFN-α (1,000 units/ml). MyD88I (50 μM) or AG490 (50 μM) was added 18 or 1 h before stimulants, respectively. CD14⁺ cells were evaluated for surface expression of CD80. (A) Expression of CD80 is shown for CD14⁺ cells incubated with hBD-3 (Left) or LPS (Right) in the presence or absence of AG490. (B) Expression of CD80 is shown for CD14⁺ cells incubated with hBD-3, LPS, or IFN-α in the presence or absence of MyD88I. (C) Mean percent inhibition of CD80 induction by AG490 or MyD88I among cells treated with hBD-3, LPS, or IFN-α in four experiments. Error bars represent the standard error of the mean.

Next, we asked whether a downstream signal of TLR activation, IRAK-1 activation, was also induced by hBD-3. Monocytes, purified by negative selection, were cultured in medium alone or medium supplemented with shBD-3, LPS, or IFN-α. Phosphorylation of IRAK-1 was observed in monocytes after stimulation with shBD-3 or with LPS (1–30 min), but not with IFN-α (Fig. 3). The bands shown for IRAK-1 represent hyperphosphorylated forms of the molecule as has been described previously (20). Thus, hBD-3 induces rapid phosphorylation of IRAK-1, a molecule typically associated with TLR signaling in human monocytes.

Fig. 3.

hBD-3 induces phosphorylation of IRAK-1 in purified monocytes. Purified monocytes (2 × 10⁶ cells per milliliter; ≈90% pure) were cultured in medium alone or in medium supplemented with hBD-3 (20 μg/ml), LPS (20 ng/ml), or IFN-α (1,000 units/ml) for 30 min. Cells lysates were assessed for phosphorylated IRAK-1 by Western blot. Actin was measured as a loading control. Results are representative of three additional experiments.

hBD-3 Requires Expression of TLR1 and -2 for Cellular Activation.

The activity of shBD-3 was tested in HEK293 cell lines engineered to express a single TLR gene (TLR2, -3, -4, -5, -7, -8, or -9) and a reporter gene responsive to NF-κB activation (21). Incubation of shBD-3 with these HEK cell lines did not provide evidence of an interaction between hBD-3 and any individual TLR [supporting information (SI) Fig. 5]. We considered the possibility that TLR heterodimers such as TLR1/2 or TLR2/6 that are required for responsiveness to the lipopeptide Pam3CysSerLys4 (Pam₃CSK₄) or macrophage-activating lipopeptide-2 (MALP-2) (22–25), respectively, might be required for the activity of hBD-3. Therefore, we evaluated the ability of shBD-3 to activate NF-κB in HEK293 cells coexpressing either TLR2 and TLR1, or TLR2 and TLR6. shBD-3 induced phosphorylation of NF-κB p65 in HEK293 TLR2/1 but not in TLR2/6 cells (Fig. 4A). The absence of a response in the HEK293 cells coexpressing TLR2 and TLR6 is consistent with the absence of a response in the cell line expressing TLR2 alone, and confirms that TLR2 is necessary, but not sufficient, to mediate signaling in response to hBD-3. These observations, taken together, indicate that hBD-3 can mediate signaling and cellular activation via a mechanism dependent on TLR2 and TLR1.

Fig. 4.

hBD-3 activates cells via TLR1 and TLR2. (A) HEK293 cells (1 × 10⁶) expressing TLR1 and TLR2 (HEK1/2) or TLR2 and TLR6 (HEK2/6) were incubated with shBD-3 (20 μg/ml), Pam₃CSK₄ (100 ng/ml), or MALP-2 (50 ng/ml) for 30 min. Cell lysates were examined for phospho-NF-κB p65 by Western blot analysis. A representative blot from five experiments is shown. (B) CHO.TLR2 or CHO.TLR4 reporter cells were cultured for 24 h in medium alone or in medium supplemented with hBD-3 (1, 5, and 20 μg), Pam₃CSK₄ (4 μg), or LPS (5 μg/ml). Expression of CD25 was measured by flow cytometry. Data are representative of two experiments. (C) PBMCs were cultured with a monoclonal antibody to TLR1 (GD2.F4; 10 μg/ml), TLR2 (TLR2.5; 10 μg/ml), a combination of GD2.F4 and TLR2.5 (10 μg/ml each) or purified mouse IgG₁ (20 μg/ml) for 1 h before addition of hBD-3 (20 μg/ml), Pam₃CSK₄ (100 ng/ml), or LPS (20 ng/ml). Summary data from four experiments are shown.

We sought to confirm the requirement for TLR2 in the activity of hBD-3 by examining the ability of hBD-3 to induce TNF-α production in bone marrow-derived macrophages (26) from wild-type or TLR2 knockout mice. Interestingly, even cells from wild-type mice did not respond to hBD-3 (SI Fig. 6), suggesting that hBD-3 may not be detected by mouse TLRs. Therefore, we assessed the ability of hBD-3 to activate CHO reporter cell lines engineered to express either human TLR2 or human TLR4. CHO.TLR2 and CHO.TLR4 cells express inducible surface CD25 under the control of an NF-κB-dependent promoter (27). Incubation of CHO.TLR2 cells with Pam₃CSK₄ or with hBD-3 resulted in increased surface expression of CD25 in a dose-dependent manner. In contrast, the CHO.TLR4 cell line responded to LPS, but not to hBD-3. These results provide further evidence that hBD-3 mediates cellular activation via a TLR2-dependent mechanism (Fig. 4B).

To evaluate the requirement for TLRs 1 and 2 in the hBD-3-mediated activation of primary human monocytes, antagonistic antibodies to each of these receptors (GD2.F4 and/or TLR2.5; 10 μg/ml) (28, 29) were preincubated with PBMCs for 1 h before overnight exposure to shBD-3, Pam₃CSK₄, or LPS. The shBD-3-induced shift in CD80 expression was partially inhibited by GD2.F4 or by TLR2.5 (mean percent inhibition ± SEM, 53 ± 6%, n = 4; and 57 ± 6%, n = 4, respectively). The combination of these antibodies, however, almost completely blocked the induction of CD80 by shBD-3 (85 ± 6% inhibition; n = 4; P = 0.014) (Fig. 4C). Preincubation of cells with the combination of anti-TLR1 and anti-TLR2 antibodies also interfered with CD80 induction by Pam₃CSK₄ (74 ± 10% inhibition; n = 4), but not by LPS (20 ± 2% inhibition; n = 4). An isotype- matched control antibody (IgG₁; 20 μg/ml) inhibited the induction of CD80 by ≈20% for each of the three ligands tested. These findings confirm the roles of both TLR1 and TLR2 in the human monocyte response to hBD-3.

It was important to confirm that these effects of hBD-3 were not mediated by microbial lipopeptide or LPS contaminants. Therefore, we boiled the shBD-3 and found that boiled shBD-3 lost its ability to activate monocytes, whereas boiled PAM₃CSK₄ retained activity under the same conditions (SI Fig. 7). Moreover, a shBD-2 molecule that was obtained from the same source as shBD-3 and purified under similar conditions failed to induce monocyte activation (SI Fig. 7). Thus, the TLR agonist activity of hBD-3 is not a consequence of bacterial product contamination.

Discussion

TLRs are thought to discriminate between self and nonself by recognizing highly conserved microbial patterns (10–12) known as pathogen-associated molecular patterns (PAMPS). Our observations demonstrate that an innate host defense element, hBD-3, also can be recognized by TLRs and through this interaction, can promote cellular differentiation. As expression of hBD-3 is induced by a variety of microbial danger signals, this new pathway may represent a means to amplify host immune responses to microbial invasion. A precedent for such a pathway can be found in mice wherein murine β-defensin-2 is thought to activate dendritic cells via a TLR4-dependent mechanism (16). Importantly, our findings in the human system are clearly distinguished from this earlier work because there is no homology between murine β-defensin-2 and hBD-3, and hBD-3 engages different TLRs. Thus, our studies provide evidence for a functional interaction between an inducible human innate defense element and TLRs.

We provide evidence that hBD-3 activates cells in a TLR1- and TLR2-dependent manner. Heterodimerization of TLR1 and TLR2, or TLR6 and TLR2, can increase the range of motifs that can be recognized by these receptors (22–25). The expression profile of TLRs on professional APCs varies by cell type (10, 30) and our finding that hBD-3 induces activation of monocytes and mDCs, but not pDCs or B cells, is consistent with the expected pattern of TLR2 and TLR1 expression on these cells (30, 31). Furthermore, the requirement for MyD88 signaling and the rapid induction of IRAK-1 phosphorylation by hBD-3 provide independent confirmation that TLR signaling mediates these effects. Notably, in preliminary NMR analyses, we have found that the characteristic amide peaks of shBD-3 are consistent with the published characterization of the native form, suggesting that the biological activities of the synthetic molecule are biologically relevant (32).

Importantly, our studies included numerous controls to demonstrate that the activity of hBD-3 was not a consequence of endotoxin or lipopeptide contamination. First, our recombinant molecule did not contain detectable endotoxin. Moreover, neither the recombinant nor the shBD-3 induced activation of TLR4-expressing HEK293 cell lines. Finally, the antagonistic anti-TLR1/2 antibodies inhibited the activation of APCs by shBD-3 but had no effect on LPS activity. Thus, there is clear evidence that the effects of our hBD-3 reagents were not mediated by endotoxin contamination. Furthermore, shBD-3 did not activate HEK293 cells expressing a variety of TLRs including TLR2, TLR3, TLR4, TLR5, TLR7, TLR8, TLR9 or TLR2 and TLR6. Thus, the only remaining possibility is that hBD-3 is contaminated specifically with a TLR1 and TLR2 agonists to the exclusion of all other known TLR agonists and that this same contaminant is present in two independent sources of hBD-3. To us, this seems implausible. Nevertheless, we have added additional controls to ensure that this is not the case. First, we have boiled the shBD-3 molecule. Boiled hBD-3 lost its ability to induce costimulatory molecules on APCs. In contrast, boiled PAM₃CSK₄ maintained its ability to activate APCs. Moreover, shBD-2, derived from the same source and synthesized under similar conditions, did not mediate activation of APCs. Finally, preliminary analyses have indicated that linear hBD-3 made by recombinant techniques does not activate APCs. Together, these data provide compelling evidence that the hBD-3 activity we report here is not a consequence of bacterial contaminants.

The induction of critical costimulatory molecules (CD80, CD86, and CD40) on the surface of professional APCs by hBD-3 may represent a key physiological event that bridges innate and adaptive immunity at mucosal surfaces. Because hBDs can both recruit and activate professional APCs, they may be attractive candidates for study as vaccine adjuvants. Thus, further studies exploring the novel immunoregulatory properties of hBDs will be important to ascertain the full potential of these molecules to induce or otherwise regulate host defenses.

Materials and Methods

Reagents.

rhBD-3 was generated with a histidine tag fusion construct, generated by PCR and cloned into pET-30c (a gift from J. Schroder, Kiel University, Kiel, Germany) (1). Recombinant hBD-3 was tested for identity, purity, and biological activity by migration in acid urea–polyacrylamide gels, Western analysis with native peptides, N-terminal amino acid sequencing, matrix-assisted laser desorption ionization time-of-flight mass spectrometry, and killing of Escherichia coli ML35p, respectively. The level of endotoxin contamination of rhBD-3 was determined by the Charles River Laboratories (Charleston, SC), using the Limulus amebocyte lysate assay. shBD-3 and shBD-2 were purchased from Peptides International (Louisville, KY). LPS from E. coli was purchased from Sigma–Aldrich, IFN-α2a (IFN-α) was obtained from PBL (Piscataway, NJ), Pam₃CSK₄ was purchased from EMC Microcollection (Tubingen, Germany), and MALP-2 was obtained from Alexis Biochemicals (Braunschweig, Germany). Antibodies to TLR1 (GD2.F4), TLR2 (TLR-2.5), and control mouse IgG₁ were purchased from Invivogen (San Diego, CA). Inhibitors used in these studies included AG490 (50 μM; Sigma–Aldrich) and MyD88 inhibitor peptide set (MyD88I; 50 μM; Imgenex).

Primary Cells and Culture Conditions.

Studies were performed with informed consent and according to policies of the Institutional Review Board at University Hospitals of Cleveland. Blood from healthy donors was drawn into heparin-coated tubes, and PBMCs were isolated over a Ficoll-Hypaque cushion, plated at 3 × 10⁶ cells per milliliter, and incubated overnight with rhBD-3, shBD-3, IFN-α, LPS, or medium alone (RPMI 1640 supplemented with antibiotics, 2 mM l-glutamine and 10% human AB serum). PBMCs were cultured with AG490, GD2.F4, or TLR2.5 for 1 h before addition of hBD-3, LPS, Pam₃CSK₄, or IFN-α. MyD88I was added to cultures 24 h before addition of hBD-3, LPS, or IFN-α. For cultures using purified monocytes, CD14⁺ cells (85–90% purity) were obtained by magnetic bead negative selection (Miltenyi Biotec, Bergisch Gladbach, Germany).

HEK293 Cell Transfectants.

Human embryonic kidney cells (HEK293) expressing TLR1/2 or TLR2/6 (Invivogen) were cultured in DMEM (Cambrex, Walkersville, MD) containing 4.5 g/liter l-glucose (Sigma–Aldrich), 10% fetal bovine serum (Summit, Ft. Collins, CO), 10 μg/ml blasticidin (Invivogen), and 100 μg/ml normicin (Invivogen). Cells were split by using trypsin Versene (Cambrex) after reaching a level of ≈75–85% confluence.

Flow Cytometry.

The following monoclonal antibodies were used in these studies: anti-CD14 FITC, anti-HLA-DR allophycocyanin, anti-CD80 phycoerytherin (PE), anti-CD40 PE, or CD86 PE, anti-CD11c PE, anti-CD123 PE, and appropriate isotype control monoclonal antibodies (BD Pharmingen, San Diego, CA). Cells were stained for 10 min at room temperature, washed (PBS with 1% BSA and 0.1% sodium azide), fixed in 1% formaldehyde, and analyzed by using a dual-laser flow cytometer (FACSCalibur; BD Biosciences, San Jose, CA) and CellQuest software (BD Biosciences). The TLR blocking experiments were analyzed by using an LSRII instrument (BD Biosciences).

Western Blots.

Purified monocytes (2.0 × 10⁶ cells per milliliter) or HEK293 cells (incubated overnight; 1 × 10⁶ cells per milliliter) were stimulated with rhBD-3 or shBD-3, LPS, Pam₃CSK₄, MALP-2, or IFN-α; samples were then collected at 1, 5, 15, and 30 min and lysed in SDS loading buffer [62.5 mM Tris·HCl/2% (wt/vol) SDS/10% glycerol/50 mM DTT/0.01% (wt/vol) bromophenol blue]. Lysates were heated at 95°C for 5 min. Samples were resolved by SDS/PAGE and transferred to nitrocellulose membranes. Phosphorylated proteins were detected by using primary antibodies to p-IRAK-1(Thr-209) or p-NF-κB p65 (Ser-536) (Cell Signaling Technology, Beverly, MA). Antibodies to actin (Santa Cruz, Santa Cruz, CA) were used to assess loading. Secondary reagents included anti-mouse or anti-rabbit HRP-conjugated antibodies (Cell Signaling Technology). Proteins were detected by chemiluminescence (Western Lightning; PerkinElmer Life Sciences, Boston, MA) and were visualized after exposure of x-ray film.

TLR Ligand Screening.

Analysis of the activation of TLRs 2-, 3-, 4-, 5-, 7-, 8-, and 9-transfected HEK293 cells was performed by Invivogen. Activation of NF-κB was monitored by using an NF-κB inducible reporter construct–secretory alkaline phosphatase activity (Invivogen) and was reported as OD readings.

Murine Macrophage Studies.

Bone marrow-derived macrophages were obtained from female C57BL/6 or TLR2^−/− mice. After 7-day incubation in DMEM/10% FBS (D10F) supplemented with 20% LADMAC conditioned medium, cells were plated (10⁵ cells per well in 96-well plates) in D10F overnight and stimulated the next day with various TLR agonists. Tissue culture plates were centrifuged after 12.5 h, and supernatants were frozen at −80°C until examined for TNF-α by ELISA (BD Biosciences).

CHO Cell Experiments.

CHO cells were cultured in Ham's F-12 medium (10% FBS and 10 μg/ml ciprofloxacin). The CHO.TLR2 and CHO.TLR4 cell lines (27) contain a reporter gene construct under control of a region from the human E-selectin promoter containing NF-κB binding sites. Activation of this reporter gene results in increased surface expression of CD25. The cells were plated at 50,000 per well in a 24-well plate. After an overnight incubation, the medium was replaced with fresh medium or with medium supplemented with hBD-3, Pam₃CSK₄, or LPS. After overnight stimulation, expression of CD25 was determined by flow cytometric analysis.

Statistical Methods.

The MFI of cell surface molecules was compared between cells cultured in medium alone or in medium supplemented with hBD-3 using a paired t test. Differences in the induction of costimulatory molecules by hBD-3 in the presence or absence of AG490 or MyD88I also were assessed by paired t tests. Kruskal–Wallis and Mann–Whitney U tests were used to compare differences in the induction of costimulatory molecule expression by hBD-3 in the presence of various anti-TLR or isotype control immunoglobulins.

Online SI.

The online SI includes SI Fig. 5, describing TLR screening results for HEK 293-TLR lines; SI Fig. 6, describing lack of shBD-3 responsiveness of mouse macrophages; and SI Fig. 7, describing abrogation of shBD-3 activity after heat treatment.

Abbreviations

hBD

human β-defensin

rhBD

recombinant human β-defensin

shBD

synthetic human β-defensin

APC

antigen-presenting cell

TLR

Toll-like receptor

MyD88

myeloid differentiating factor 88

IRAK-1

IL-1 receptor-associated kinase-1

PBMC

peripheral blood mononuclear cell

mDC

myeloid dendritic cell

pDC

plasmacytoid dendritic cell

MFI

mean fluorescence intensity

Pam₃CSK₄

Pam3CysSerLys4

MALP-2

macrophage-activating lipopeptide-2

PE

phycoerytherin.