Internalization and coreceptor expression are critical for TLR2-mediated recognition of lipoteichoic acid in human peripheral blood.

Abstract

Lipoteichoic acid (LTA), a ubiquitous cell wall component of Gram-positive bacteria, represents a potent immunostimulatory molecule. Because LTA of a mutant Staphylococcus aureus strain lacking lipoproteins (Deltalgt-LTA) has been described to be immunobiologically inactive despite a lack of ascertained structural differences to wild-type LTA (wt-LTA), we investigated the functional requirements for the recognition of Deltalgt-LTA by human peripheral blood cells. In this study, we demonstrate that Deltalgt-LTA-induced immune activation critically depends on the immobilization of LTA and the presence of human serum components, which, to a lesser degree, was also observed for wt-LTA. Under experimental conditions allowing LTA-mediated stimulation, we found no differences between the immunostimulatory capacity of Deltalgt-LTA and wt-LTA in human blood cells, arguing for a limited contribution of possible lipoprotein contaminants to wt-LTA-mediated immune activation. In contrast to human blood cells, TLR2-transfected human embryonic kidney 293 cells could be activated only by wt-LTA, whereas activation of these cells by Deltalgt-LTA required the additional expression of TLR6 and CD14, suggesting that activation of human embryonic kidney 293 cells expressing solely TLR2 is probably mediated by residual lipoproteins in wt-LTA. Notably, in human peripheral blood, LTA-specific IgG Abs are essential for Deltalgt-LTA-mediated immune activation and appear to induce the phagocytic uptake of Deltalgt-LTA via engagement of FcgammaRII. In this study, we have elucidated a novel mechanism of LTA-induced cytokine induction in human peripheral blood cells that involves uptake of LTA and subsequent intracellular recognition driven by TLR2, TLR6, and CD14.

Full text

Page 1

For Peer Review. Do not distribute. Destroy after use.

1
Internalization and co-receptor expression are critical for TLR2-mediated recognition
of lipoteichoic acid in human peripheral blood 1

Running title: LTA recognition requires uptake, TLR2 and co-receptors

Sebastian Bunk 2,3*, Stefanie Sigel 2*†‡, Daniela Metzdorf *, Omar Sharif †‡, Kathy Triantafilou §,
Martha Triantafilou §, Thomas Hartung *¶, Sylvia Knapp †‡ and Sonja von Aulock *║

* Department of Biochemical Pharmacology, University of Konstanz, Konstanz, Germany
† Research Center for Molecular Medicine of the Austrian Academy of Science, Vienna,
Austria
‡ Department of Medicine I, Division of Infectious Diseases and Tropical Medicine,
Medical University Vienna, Vienna, Austria
§ Infection and Immunity Group, School of Life Sciences, University of Sussex, Brighton,
United Kingdom
¶ CAAT, Johns Hopkins University, Baltimore, MD, USA
║ Zukunftskolleg, University of Konstanz, Konstanz, Germany


Key words: innate immunity, TLR2, Staphylococcus aureus, lipoteichoic acid, internalization

Correspondence to:
Dr. Sebastian Bunk, Department of Biochemical Pharmacology; University of Konstanz,
78457 Konstanz, Germany; e-mail: Sebastian.Bunk@gmx.net
Phone: 0043-1-20100-1724, Fax: 0043-1-20100-5017

Page 2

For Peer Review. Do not distribute. Destroy after use.
2
Abstract
Lipoteichoic acid (LTA), an ubiquitous cell wall component of Gram-positive bacteria,
represents a potent immunostimulatory molecule. Since LTA of a mutant Staphylococcus
aureus strain lacking lipoproteins (Δlgt-LTA) has been described to be immunobiologically
inactive despite a lack of ascertained structural differences to wild type LTA (wt-LTA), we
investigated the functional requirements for the recognition of Δlgt-LTA by human peripheral
blood cells. Here we demonstrate that Δlgt-LTA-induced immune activation critically
depends on the immobilization of LTA and the presence of human serum components, which
to a lesser degree was also observed for wt-LTA. Under experimental conditions allowing
LTA-mediated stimulation, we found no differences between the immunostimulatory capacity
of ∆lgt-LTA and wt-LTA in human blood cells, arguing for a limited contribution of possible
lipoprotein contaminants to wt-LTA-mediated immune activation. In contrast to human blood
cells, TLR2-transfected HEK293 cells could be activated only by wt-LTA, whereas activation
of these cells by ∆lgt-LTA required the additional expression of TLR6 and CD14, suggesting
that activation of HEK293 expressing solely TLR2 is probably mediated by residual
lipoproteins in wt-LTA. Notably, in human peripheral blood, LTA-specific IgG antibodies are
essential for Δlgt-LTA-mediated immune activation and appear to induce the phagocytic
uptake of Δlgt-LTA via engagement of Fcγ receptor II. Here we have elucidated a novel
mechanism of LTA-induced cytokine induction in human peripheral blood cells that involves
uptake of LTA and subsequent intracellular recognition driven by TLR2, TLR6 and CD14.

Page 3

For Peer Review. Do not distribute. Destroy after use.
interact with LTA, were found to have an important role during the host immune response
3
Introduction
Staphylococcus aureus (S. aureus), a common cause of severe infections acquired in
hospitals, often colonize human skin asymptomatically and can be isolated from the nasal
mucosa of up to 40% of healthy individuals (1). Having overcome skin or mucosal barriers
S. aureus establish an infection that, although mostly restricted to the skin, may involve tissue
or systemic blood bacteraemia associated with sepsis and sometimes death (2). Upon contact
with immune cells, S. aureus initially provoke an innate immune response resulting in
secretion of cytokines and chemokines as well as in up-regulation of phagocytosis and
co-stimulatory molecules able to trigger adaptive immune responses to control bacteraemia
(3-4). This response is set in motion by the recognition of highly conserved bacterial
molecules, called pathogen-associated molecular patterns (PAMPs), via dedicated receptors,
such as TLRs or cytosolic receptors containing a nucleotide-binding oligomerization domain
(NOD) (5-7).
As for other Gram-positive bacteria, the innate immune recognition of S. aureus strongly
depends on TLR2 (8-9). Human TLR2 is expressed on the cell surface and within endosomes
of antigen presenting cells like monocytes, macrophages, dendritic cells and neutrophils (10).
It recognizes various staphylococcal cell wall components including LTA (11), lipoproteins
(12) and peptidoglycan (13) albeit recognition of the latter is controversially discussed (14-
15). This broad range of different bacterial structures described as TLR2 ligands can be
explained by heterodimer formation between TLR2 and other TLRs, such as TLR1 or TLR6
(16-17). Beside TLR2, the mannose-binding lectin (MBL) and CD36, both able to directly
against S. aureus (8, 18-19). Whereas membrane bound CD36 was found to augment
TLR2/6-dependent recognition of S. aureus and its LTA via improving their intracellular
uptake (20), soluble CD36 and MBL were described to enhance LTA-mediated activation by
improving LTA delivery to the TLR2/6 heterodimer (18-19).

Page 4

For Peer Review. Do not distribute. Destroy after use.
4
Despite strong evidence for a major role of LTA in the immune activation by Gram-positive
bacteria (21-23), recent reports suggested that not LTA but lipoproteins are dominant
immunobiologically active structures of S. aureus (24-25). The latter authors demonstrated a
100-fold decreased immunostimulatory capacity of LTA preparations derived from an Δlgt
deletion mutant S. aureus lacking palmitate-labeled lipoproteins compared to LTA from the
respective wild type strain. From this finding and from other observations they concluded
contaminating lipoproteins in the LTA preparations to be responsible for LTA-mediated
immune activation (24-26). Contrarily to the latter investigators (27), we found that Δlgt-LTA
and wt-LTA were equipotent in inducing cytokine release from human whole blood. The
major discrepancy in LTA activation capacity observed in the two groups both using human
whole blood were puzzling and led us to comparatively investigate the specific requirements
for Δlgt- and wt-LTA-mediated cytokine induction in human blood cells in order to find an
explanation.
In this report, we describe that cytokine induction by Δlgt-LTA in human peripheral blood is
critically dependent on surface immobilization of LTA and requires the presence of
LTA-specific antibodies as well as the phagocytic activity of blood cells. We further provide
evidence that recognition of Δlgt-LTA requires the co-expression of TLR2, TLR6 and CD14.
The results presented here not only confirm the equipotent immune activation by Δlgt- and
wt-LTA in human whole blood observed previously (28), but also provide a conclusive
explanation for contradictory findings obtained in previous experiments analyzing the
immunostimulatory capacity of Δlgt-LTA.

Page 5

For Peer Review. Do not distribute. Destroy after use.

5
Material and Methods

Stimuli
LTA from S. aureus wild type (SA 113 wt, wt-LTA) and mutant strain SA 113 lgt::ermB (SA
113 Δlgt, lacking the lipoprotein diacylglycerol transferase, Δlgt-LTA), both strains were kind
gifts from A. Peschel, University of Tübingen, Germany, was isolated by butanol extraction
and hydrophobic interaction chromatography as described previously (21). For UV
inactivation and subsequent whole blood stimulation, S. aureus (SA 113 wt) were cultivated
in tryptic soy broth (BD Biosciences, Heidelberg, Germany) for 16 h at 37°C. Harvested
bacteria were adjusted to 108 bacteria/ml and 1 ml per well was irradiated on ice
(UV-Stratalinker, La Jolla, CA) with an energy density of 1 kN/cm2 (3 mWatt/cm2 x 300 s)
for 5 min in a 6-well cell culture plate (Greiner Bio-One, Frickenhausen, Germany). The
inactivation was controlled by growth on blood agar plates (Columbia-blood agar, Heipa
Diagnostika, Eppelheim, Germany) after 24 h at 37°C and 5% CO2.
Other substances were LPS from Salmonella enterica serovar abortus equi and cytochalsin D
from Zygosporium masoni (Sigma, Deisenhofen, Germany), Poly I:C (Invivogen, San Diego,
USA), Pam2Cys-SK4 and Pam3Cys-SK4 (EMC Microcollections, Tübingen, Germany), CpG
2216 (MWG Biotech AG, Ebersberg, Germany), muramyl dipeptides (MDP, Bachem,
Heidelberg, Germany), human IgG antibodies (Endobulin, Baxter, Wien, Austria), control
IgG1 (Avastin, Roche, Grenzach-Wyhlen, Deutschland) and anti-LTA IgG1 (Pagibaximab,
Biosynexus, Gaithersburg, MD, USA).
Human whole blood incubation
Heparinized venous blood was obtained from healthy volunteers after informed consent.
Differential blood cell counts were routinely determined with a Pentra 60 apparatus (ABX
Diagnostics, Montpellier, France) to exclude acute infections. Blood was diluted fivefold with

Page 6

For Peer Review. Do not distribute. Destroy after use.
(Axis-Shield, Oslo, Norway) and were washed free of serum as described above. Monocytes
6
RPMI 1640 (Lonza, Verviers, Belgium) and 500 µl were incubated in the presence of the
different stimuli in polypropylene tubes (Eppendorf, Hamburg, Germany) overnight for 22 h
at 37°C and 5% CO2. For LTA immobilization, 50 µl of wt-LTA, Δlgt-LTA in RPMI 1640 at
different concentrations was preincubated in tubes for up to 1 h at room temperature before
the addition of whole blood and RPMI 1640. In case of LTA immobilization, all other control
stimuli, such as LPS or UV-inactivated S. aureus were also preincubated before use for
stimulation experiments. To assess the role of phagocytosis in LTA-mediated cytokine
induction, diluted human whole blood was first pre-treated with 3 µM cytochalasin D for
30 min in siliconized glass tubes (Vacutainer, Biosciences) and then applied to the
preincubated LTA. The ability of cytochalasin D to inhibit phagocytosis was controlled by
stimulation with MDP, CpG and Poly I:C as well as with LPS and Pam2Cys-SK4,
respectively. After 22 h, blood cells were suspended by gentle shaking and centrifuged at
400 x g for 2 min. The cell-free supernatants were stored at -80°C for cytokine determination.

Preparation and stimulation of human peripheral blood mononuclear cells and monocytes
PBMC of healthy volunteers were prepared with CPTTM Cell Preparation Tubes
(BD Biosciences, Heidelberg, Germany). After centrifugation at 1600 x g for 20 min, PBMC
were collected and washed four times at 300 x g for 10 min with RPMI 1640 containing
2.5 IU/ml Liquemin (Hoffmann-La Roche) to remove serum residues. For experiments
employing isolated monocytes or monocyte-depleted PBMC, PBMC were isolated from
heparinized human whole blood using LymphoprepTM as described by the manufacturer
were isolated from PBMC by positive or negative magnetic isolation (Miltenyi Biotech,
Bergisch Gladbach, Germany). Briefly, PBMC were incubated with biotinylated anti-CD14
antibody (positive selection) or with a cocktail of biotinylated antibodies against CD3, CD7,
CD16, CD19, CD56 and CD123 (negative selection) for 10 min on ice. Then cells were

Page 7

For Peer Review. Do not distribute. Destroy after use.
(FCS, GIBCO Invitrogen) and 1% Penicillin/Streptomycin (GIBCO Invitrogen). After 48 h, the
7
incubated with magnetic bead-conjugated anti-biotin antibodies for 15 min and applied onto
MACS LS columns. In case of positive selection, PBMC passing the column were collected
as monocyte-depleted PBMC, and the CD14+ monocytes were collected by eluting the bound
cell fraction. In case of negatively selected monocytes, the cells passing the column without
binding were collected. Prior to stimulation, PBMC and monocytes were supplemented with
4% autologous serum unless stated otherwise. Serum-free PBMC were used for stimulation
either with or without addition of different concentrations of autologous serum, different
human antibodies or IgG-depleted autologous serum. Stimulation of PBMC or monocytes and
the storage of supernatants were performed according to the conditions described for whole
blood in a total volume of 220 µl. PBMC were added at a density of 5 x 105cells/tube and
monocytes were used at a density corresponding to the monocyte count in PBMC samples.
For Fcγ receptor blockade experiments, PBMC were preincubated with neutralizing
antibodies against CD16 (clone 3G8, BD Bioscience), CD32 (clone AT10, Abcam,
Cambridge, UK) or CD64 (clone 10.1, eBioscience, Hatfield, UK) or with the respective
isotype control mouse IgG1 antibody (eBioscience) for 30 min at 37°C prior to stimulation.

Stimulation of transfected HEK293 cells
Human embryonic kidney (HEK) 293 cells stably transfected either with human TLR2 alone
or a combination of human TLR2, TLR6 and CD14 as well as wild type HEK293 cells were
seeded at 2 x 105 cells per well in 24-well cell culture plates (BD Bioscience) in Dulbecco’s modified
Eagle medium (GIBCO Invitrogen, Darmstadt, Germany) supplemented with 10% fetal calf serum
medium was replaced by FCS-free medium containing different concentrations of wt-LTA, Δlgt-LTA
or Pam3Cys-SK4. Cell-free supernatants were collected 22h after stimulation and stored at -80°C for
ELISA experiments.

Page 8

For Peer Review. Do not distribute. Destroy after use.

8
FRET analysis
FRET is a non-invasive imaging technique used to determine molecular proximity based on
nonradiative transfer of energy from the excited state of donor molecules to an appropriate
acceptor (29). This energy transfer is inversely proportional to the sixth power of the distance,
between donor and acceptor (30-31). In this study, FRET was measured using a method as
previously described (30-31). Briefly, human monocytes on microchamber culture slides
(Lab-tek, Gibco) were stimulated with 10 µg/ml wt-LTA or Δlgt-LTA or left untreated and
then were labeled with 100 µl of a mixture of donor (Cy3) and acceptor (Cy5) conjugated
antibody. The cells were rinsed twice in PBS/0.02% BSA and fixed with 4% formaldehyde
for 15 min to prevent potential protein reorganization. Cells were imaged on a Carl Zeiss, Inc.
LSM510 confocal microscope (with an Axiovert 200 fluorescent microscope) using a 1.4 NA
63x Zeiss objective. The images were analysed using LSM 2.5 image analysis software (Carl
Zeiss, Inc.). Cy3 and Cy5 were detected using the appropriate filter sets. Using typical
exposure times for image acquisition (less than 5 s), no fluorescence was observed from a
Cy3-labelled specimen using the Cy5 filters, nor was Cy5 fluorescence detected using the
Cy3 filter sets.
For calculation, the energy transfer between donors and acceptor conjugated antibodies was
detected as an increase in donor fluorescence (dequenching) after complete photobleaching of
the acceptor molecule and calculated according to the formula: E(%) x 100 = 10,000 x [(Cy3
postbleach – Cy3 pre-bleach)/Cy3 postbleach]. The scaling factor of 10,000 was used in order
to expand E to the scale of the 12-bit images.
Identification of LTA-binding proteins
For the identification of LTA-interaction components from human serum, the wells of a
6-well cell culture plate (Greiner Bio-One) were incubated with either 10 µg Δlgt-LTA/well
or with only PBS, over night at 4°C and afterwards washed twice with PBS. Then 1 ml/well

Page 9

For Peer Review. Do not distribute. Destroy after use.
based on commercially available pairs of antibodies and standards. Antibody pairs against
9
of 5% human serum in PBS was added for 1 h at room temperature, again followed by two
washing steps with PBS. Surface-bound proteins from six wells were pooled in a total volume
of 100 µl PBS containing 1% SDS and analyzed by SDS-PAGE and silver staining. The two
protein bands at 25 kDa and 50 kDa, found only in the wells preincubated with Δlgt-LTA and
not in the PBS-controls, were excised from the gel and, after in-gel tryptic digestion, were
identified using MALDI-TOF mass spectrometry at the core facility of the Biomedical Centre
at the Ludwig-Maximillians-University Munich, Germany.

Depletion of immunoglobulin G from human serum and Western blot
IgG was removed from human serum using HiTrapTM protein A columns (1 ml, GE
Healthcare, Munich, Germany). According to the manufacturer’s protocol, 1 ml of 50% serum
in PBS was applied to the column and the flow-through was collected. In order to decrease
remaining IgG residues, three columns were connected in series. The removal of IgG was
confirmed by Western blot analysis. IgG-depleted serum-samples were applied to 12% SDS-
PAGE and blotted to nitrocellulose membranes (Pall Corporation, Dreieich, Germany). IgG
was detected by immunoblotting with horse radish peroxidase-conjugated polyclonal rabbit
anti-human-IgG antibodies (DakoCytomation, Glostrup, Denmark) and enhanced
chemiluminescence detection using an LAS-3000 imaging system (Fuji).

Cytokine measurement
Cytokines released by human whole blood were measured by in-house sandwich ELISA
human IL-1β and IL-6 were purchased from R&D Systems and against human TNF, IL-8 and
IFN-γ from Endogen (Perbio Science, Bonn, Germany). Recombinant standards for IL-1β,
IL-6, TNF and IFN-γ were obtained from the NIBSC, Herts, UK and rIL-8 was obtained from
PeproTech (Tebu, Frankfurt, Germany). Assays were carried out in flat-bottom, ultrasorbent

Page 10

For Peer Review. Do not distribute. Destroy after use.
analysis of two groups of data, the Mann Whitney test was used. In the figures *, ** and ***
10
96-well plates (MaxiSorp, Nunc, Wiesbaden, Germany). Binding of biotinylated secondary
antibody was quantified using streptavidin-conjugated horseradish peroxidase (Biosource,
CA) and TMB (Sigma) was used as substrate. The reaction was stopped with 1 M sulphuric
acid and the absorption was measured in an ELISA reader at 450 nm with a reference
wavelength of 690 nm. Cytokine levels are given per ml of blood.

Determination of anti-LTA antibody serum titer
The amount of LTA-binding IgG was determined in human serum obtained from 57 healthy
volunteers. Flat-bottom 96-well plates (MaxiSorp; Nunc) were coated overnight with
10 µg/ml Δlgt-LTA in PBS at 4°C. Then, wells were blocked with 200 µl of 3% BSA in PBS
for 2 h at room temperature and washed with PBS before addition of 4% serum in PBS and
further incubation for 1 h at room temperature. After washing, horse radish
peroxidase-conjugated polyclonal rabbit anti-human-IgG antibody was added to each well for
another 30 min. Enzymatic activity was detected with 3,3’,5,5’,-tetramethybenzidine substrate
(TMB, Sigma) and stopped with 1 M sulphuric acid. The absorption was measured in an
ELISA reader at 450 nm with a reference wavelength of 690 nm.

Statistics
Statistical analysis was performed using the Graph Pad Prism 3 Program (Graph Pad
Software, San Diego, USA). Three or more groups of data were compared by
repeated-measure analysis of variance (ANOVA) followed by Dunn’s post test. For statistical
represent p values <0.05, <0.01 and <0.001, respectively.

Page 11

For Peer Review. Do not distribute. Destroy after use.
LTA in the supernatant was transferred to another tube, and then incubated for 30 min,
11
Results

Immobilized but not soluble LTA of mutant ∆lgt S. aureus induces cytokine release in human
whole blood
There are conflicting reports concerning the immunostimulatory capacity of LTA isolated
from a ∆lgt mutant S. aureus strain that lacks lipoproteins (24, 27-28). Since we previously
reported an increased immune activation by immobilized wt-LTA (32-33), we analyzed
whether immobilization also augments ∆lgt-LTA-mediated immune activation. We
determined the cytokine induction in human whole blood upon stimulation with ∆lgt-LTA
and wt-LTA, either used in an immobilized or non-immobilized form. Immobilization was
achieved by preincubating LTA in polypropylene tubes prior to the addition of diluted blood,
whereas for non-immobilizing conditions LTA was directly added to the blood.
As shown in Fig. 1A, whole blood stimulation with immobilized ∆lgt-LTA led to a potent
induction of IL-1β and TNF, which was comparable to wt-LTA and LPS (Fig. 1A). In
contrast, without immobilization the ∆lgt-LTA-induced cytokine release from whole blood
was fully abrogated, whereas LPS-mediated cytokine induction was unaffected. Notably,
wt-LTA induced also cytokine release when used without prior immobilization albeit at a
significant lower level.
To confirm that LTA does adhere to the polypropylene tubes, we analyzed the
cytokine-inducing capacity of only the LTA bound during the immobilization procedure. For
this, LTA was incubated in polypropylene tubes for 30 seconds and the residual, unbound
followed by another transfer and 30 min incubation, but without subsequent removal of the
supernatant. Even after incubation in the vial for only 30 sec, the bound ∆lgt-LTA and
wt-LTA were able to induce IL-1β and TNF, albeit in low amounts, which was strongly
enhanced after 30 min preincubation. Interestingly, the cytokine-inducing capacity of bound

Page 12

For Peer Review. Do not distribute. Destroy after use.
(dequenching) of Cy3-labelled donors, such as TLR2, TLR4 or MHC class I, after complete
12
LTA was not diminished in tubes incubated with supernatants from previous incubations
(Fig. 1B). No decrease in cytokine induction by the LTA containing supernatant was observed
even when it was transferred up to six times (unpublished observations), indicating that only a
small proportion of LTA binds to the tube thereby enabled to induce cytokine release.

∆lgt-LTA-mediated cytokine release depends on the presence of monocytes and requires the
expression of TLR6 and CD14 in addition to TLR2
To determine which type of blood cells respond to stimulation with ∆lgt-LTA, we compared
cytokine release from PBMC, isolated monocytes and monocyte-depleted PBMC. Stimulation
with ∆lgt-LTA, wt-LTA or LPS led to a comparable release of IL-1β and TNF from PBMC
and from monocytes obtained by negative isolation (Fig. 2A). Furthermore, PBMC depleted
of CD14-expressing monocytes showed strongly reduced cytokine induction suggesting that
monocytes represent the main ∆lgt-LTA-responsive cell type in human peripheral blood.
Interestingly, monocytes enriched with anti-CD14 antibodies (positive isolation) showed
markedly diminished cytokine production after stimulation with ∆lgt-LTA but not wt-LTA or
LPS, indicating a role for membrane bound CD14 in the recognition of ∆lgt-LTA.
LTA is believed to induce cytokines via TLR2 engagement, but recent studies demonstrated
the inability of Δlgt-LTA to activate TLR2-transfected cells (24-25, 27). In order to analyze if
Δlgt-LTA engages TLR2 independent of its ability to activate cells, we employed FRET
analysis to determine Δlgt-LTA-mediated recruitment of TLR2 to lipid rafts of human
monocytes (34). For this purpose we measured the increase in fluorescence intensity
photobleaching of Cy5-labelled acceptors, such as the lipid raft markers CD14 or GM1
ganglioside. Following Δlgt-LTA incubation increased donor fluorescence of TLR2 was
observed after bleaching the acceptors CD14 or GM1, which was not observed for
unstimulated monocytes (Table 1). In contrast, TLR4 and MHC class I showed unchanged

Page 13

For Peer Review. Do not distribute. Destroy after use.
these data demonstrate the important role of TLR2 and the co-receptors TLR6 as well as
13
donor fluorescence in the presence of Δlgt-LTA, thus demonstrating that Δlgt-LTA
specifically recruits TLR2 to the lipid raft. Comparable results were observed for wt-LTA,
suggesting that both LTAs engage the TLR2 receptor. The important role of TLR2 in
Δlgt-LTA and wt-LTA recognition was further confirmed in TLR2 knockout mice, which
upon stimulation with both LTA showed significantly reduced cytokine induction compared
to wt-mice (Fig. S1, supplemental data).
To determine the innate immune receptors required for recognition of Δlgt-LTA and wt-LTA,
we analyzed the activation of HEK293 cells stably transfected with human TLR2,
TLR2/CD14 or TLR2/TLR6/CD14. In line with previous reports (27), Δlgt-LTA did not
activate TLR2-HEK293, whereas these cells showed pronounced IL-8 release after
stimulation with wt-LTA or Pam3Cys (Fig. 2B). The same discrepancy between Δlgt-LTA
and wt-LTA was observed for HEK293 cells co-expressing TLR2 and CD14. Interestingly,
the simultaneous expression of TLR2, CD14 and TLR6 enabled HEK293 cells to effectively
recognize Δlgt-LTA. We observed a significant release of IL-8 by TLR2/6/CD14-HEK293
cells upon stimulation with 10 µg/ml Δlgt-LTA. Whereas TLR6 expression was pivotal for
Δlgt-LTA recognition, almost no effect was observed for Pam3Cys and wt-LTA, respectively.
Notably, for HEK293 experiments Δlgt-LTA could not be immobilized prior to the
stimulation, a circumstance that might have contributed to the lower potency of Δlgt-LTA
compared to wt-LTA. As indicated for isolated monocytes, CD14 was also required for Δlgt-
LTA recognition by HEK293 cells, since HEK293 cells transiently transfected with TLR2 and
TLR6 did not respond to Δlgt-LTA in the absence of CD14 (data not shown). Taken together,
CD14 in the recognition of Δlgt-LTA by human immune cells.

∆lgt-LTA-induced cytokine release by human PBMC is abrogated after cytochalasin D
treatment and in the absence of human serum

Page 14

For Peer Review. Do not distribute. Destroy after use.
at low amounts (Fig. 3B). The cytokine induction in PBMC by ∆lgt-LTA and wt-LTA peaked
14
Recognition of LTA was described to be amplified by CD36-mediated phagocytosis and by
the presence of serum MBL via a mechanism that requires intracellular uptake of LTA (18,
20). These findings prompted us to investigate the influence of phagocytosis and human
serum components on LTA-induced cytokine release in human blood cells. Using
cytochalasin D to inhibit phagocytic activity we analyzed the cytokine inducing capacity of
∆lgt-LTA and wt-LTA in human whole blood. For control purposes, stimulations were carried
out with different ligands known to activate immune receptors located either intracellular or
extracellular. Surprisingly, in the presence of cytochalasin D the induction of IL-1β and TNF
by ∆lgt-LTA was completely abrogated and in case of wt-LTA strongly attenuated, which
was also observed for the intracellular receptor ligands CpG and MDP as well as for Poly I:C
(Fig 3A). In contrast, recognition of the extracellular TLR ligands LPS and Pam2Cys was not
affected by cytochalasin D treatment suggesting a pivotal role of phagocytosis for
LTA-mediated cytokine induction. To investigate the role of serum components we
determined lgt-LTA and wt-LTA-mediated cytokine release from serum-free PBMC
supplemented with increasing concentrations of autologous human serum. For immobilized
but also non-immobilized (not shown) ∆lgt-LTA we detected no cytokine release from PBMC
in the absence of human serum, which in case of immobilized ∆lgt-LTA could be regained by
the addition of autologous human serum (Fig. 3B). Already at concentrations of 1% serum
significant cytokine induction by ∆lgt-LTA was observed. For wt-LTA a comparable
enhancement of cytokine release in the presence of human serum was found, but in contrast to
∆lgt-LTA, wt-LTA was already able to induce cytokines under serum-free conditions, albeit
at a concentration of 4% serum and the amounts of released IL-1β and TNF were comparable
between both LTA underlining their equal immunostimulatory capacity.

Page 15

For Peer Review. Do not distribute. Destroy after use.
LTA-specific IgG antibodies augment ∆lgt-LTA-mediated cytokine induction in PBMC in a
15
Immobilized ∆lgt-LTA interacts with apolipoprotein A1 and cationic immune globulins from
human serum
To identify serum components that possibly interact with immobilized LTA resulting in
immune activation, cell culture plate wells pre-treated with or without ∆lgt-LTA were
incubated with 4% serum from different human donors and bound components were analyzed
by SDS-PAGE. As shown for two exemplary donors in Fig. 4, the SDS-PAGE profiles of
serum components derived from wells with immobilized ∆lgt-LTA showed two abundant
protein bands (25 kDa and 50 kDa) that were absent in serum samples obtained from control
wells without LTA. The two protein bands were observed in the SDS-PAGE profiles of six
tested donors, albeit the band intensity among these donors was slightly different (data not
shown). Using tandem mass spectrometry analysis of peptides derived from the 25 kDa
protein band we identified two different human proteins, i.e. apolipoprotein A1 and
immunoglobulin kappa light chain, whereas the analysis of the 50 kDa protein band revealed
only the heavy chain of immunoglobulin (Table 2). In addition to the complete
immunoglobulin light chain sequences with a predicted molecular mass of 23 kDa, a Mascot
database search also revealed highest identification scores for the variable domains of light
chains derived from anti-DNA and anti-cardiolipin antibodies. Interestingly, DNA and
cardiolipin share putative epitopes consisting of phosphodiester groups separated by three
adjacent carbon atoms together with the polyglycerolphoshate backbone of LTA (35) arguing
for a specific interaction between the detected antibodies and LTA.

CD32-dependent manner
To analyze the influence of apolipoprotein A1 and specific immunoglobulins on the
recognition of ∆lgt-LTA we performed supplementation experiments using serum-free
PBMC. The addition of up to 10 µg/ml apolipoprotein A1 was unable to restore cytokine