IMMUNOLOGY ORIGINAL ARTICLE

Various members of the Toll-like receptor family contribute

to the innate immune response of human epidermal keratinocytes

Gabriele Köllisch,1,2 Behnam Naderi Summary

Kalali,2 Verena Voelcker,2 Reinhard
Wallich,3 Heidrun Behrendt,1
Johannes Ring,2 Stefan Bauer,4
Thilo Jakob,1,2 Martin Mempel2,5
and Markus Ollert2,5
1
Division of Environmental Dermatology and
Allergy GSF/TUM, GSF National Research
Center for Environment and Health, Neuher-
berg, Germany, Departments of 2Dermatology
and Allergy, Biederstein, 4Medical Microbio-
logy, Immunology and Hygiene, 5Clinical
Research Division of Molecular and Clinical
Allergotoxicology, Technische Universität Mün-
chen, Germany and 3Institute for Immunology,
University of Heidelberg, Heidelberg, Germany
doi:10.1111/j.1365-2567.2005.02122.x
Received 23 September 2004; revised 22
December 2004; accepted 22 December 2004.
Correspondence: Dr Gabriele Köllisch,
Department of Dermatology and Allergy,
Biederstein, TU München, Biedersteiner Str.
29, D-80802 Munich, Germany.
Email: koellisch@gsf.de
Senior author: Dr M. Ollert,
email: ollert@lrz.tum.de
Toll-like receptors (TLRs) are important pattern recognition molecules
that activate the nuclear factor (NF)-jB pathway leading to the produc-
tion of antimicrobial immune mediators. As keratinocytes represent the
first barrier against exogenous pathogens in human skin, we investigated
their complete functional TLR1–10 expression profile. First, reverse tran-
scription–polymerase chain reaction (PCR) analysis revealed a very similar
pattern of TLR mRNA expression when comparing freshly isolated human
epidermis and cultured primary human keratinocytes. Thus, further
experiments were carried out with primary keratinocytes in comparison
with the spontaneously immortalized human keratinocyte cell line HaCaT.
The quantitative expression of TLR1–10 mRNA in real-time PCR of pri-
mary human keratinocytes and HaCaT cells was analysed. Both cell types
constitutively expressed TLR2, TLR3, TLR5, and to a lesser extent TLR10.
TLR4 was only found in HaCaT cells, TLR1 to a higher degree in primary
keratinocytes. In line with this, LPS induced mRNA expression of CD14
and TLR4 only in HaCaT cells. After stimulation with various TLR lig-
ands, the NF-jB-activated chemokine interleukin-8 (IL-8) was measured.
In primary keratinocytes and HaCaT cells the TLR3 ligand poly (I:C) was
the most potent stimulator of IL-8 secretion. The TLR ligands peptidogly-
can, Pam3Cys and flagellin which bind to TLR2, TLR1/TLR2 heterodimer,
and TLR5, respectively, also induced IL-8 secretion, whereas no IL-8 was
induced by LPS, R-848, loxoribine and cytosine guanine dinucleotide-
containing oligodeoxynucleotide. A corresponding pattern was found in
the RelA NF-jB translocation assay after ligand stimulation of primary
keratinocytes. These studies provide substantial evidence for a functional
TLR expression and signalling profile of normal human keratinocytes con-
tributing to the antimicrobial defence barrier of human skin.
Keywords: innate immunity of the skin; HaCaT cells; LPS; NF-jB

major cell population of human epidermis. Within the

Introduction
The human epidermis, the multilayered stratified epi-
thelium of the skin, provides a first line defense barrier
for the host. Representing the outmost compartment of
human skin the epidermis consists of three main resident
cell populations: the keratinocytes, which are of epithelial
origin, the bone marrow-derived Langerhans cells and the
neuroectodermal melanocytes. Keratinocytes represent the
epidermis, keratinocytes are involved in both, physical
and first line immune protection of the host.1–3
Resistance to infection of the skin, has been shown in
recent years, is based on the function of intact innate
immune mechanisms provided by epidermal keratinocytes
such as the production of antimicrobial pore-forming
peptides of the human b-defensin subfamily.4,5 b-Defen-
sins are up-regulated by a multitude of inflammatory and

Abbreviations: CpG-ODN, cytosine guanine dinucleotide-containing oligodeoxynucleotide; LPS, lipopolysaccharide;

LTA, lipoteichoic acid; Pam3Cys, tripalmitoyl-S-glyceryl-cysteine; PGN, peptidoglycan; poly (I:C), poly(inosinic-cytidylic) acid;
RT–PCR, reverse transcriptase–polymerase chain reaction; TLR, Toll-like receptor.

2005 Blackwell Publishing Ltd, Immunology, 114, 531–541 531

G. Köllisch et al.
infectious stimuli such as tumour necrosis factor-a
(TNF-a), interleukin 1 (IL-1), bacterial lipopolysaccharide
(LPS), or bacteria such as the skin commensal Staphylo-
coccus epidermidis through signal transduction pathways
utilizing either nuclear factor (NF)-jB transcription or
the mitogen-activated protein (MAP) kinase system.6
More recently, Toll-like receptors (TLRs) have also
been identified on epithelial cells7,8 thus adding a new
component to the overall epithelial antimicrobial barrier.
TLRs are important pattern recognition molecules of the
innate immune system that activate the transcription fac-
tor NF-jB, leading to the production of antimicrobial
and antiviral cytokines, chemokines and peptides.9 TLRs
recognize a wide range of microbial ligands including
LPS, peptidoglycan (PGN), lipoteichoic acid (LTA), bac-
terial lipoproteins, bacterial heat shock proteins, myco-
bacterial phosphoinositol mannosides, bacterial cytosine
guanine dinucleotide-containing oligodeoxynucleotides
(CpG-ODN), viral single- and double-stranded RNA, and
bacterial flagellin. Up to now, 10 different human TLRs
with either single or shared specificity for the various
microbial ligands have been characterized.9,10 All of the
10 members of the TLR family are characterized by the
presence of variable numbers of leucine-rich repeat (LRR)
domains in their extracellular portion and an intracellular
Toll–IL-1 receptor (TIR) domain.10 Recently, it was sug-
gested to group the various TLRs into five subfamilies
according to their amino acid sequence similarities and
their intracellular signalling dependent on the different
adaptor molecules of the TLR pathway.10–12 According to
this subdivision, the TLR subfamilies 2, 3, 4, 5, and 9 can
be categorized.
Receptors of the TLR family represent not only import-
ant factors of host innate immune cells helping to pro-
mote an adaptive immune response, but have also been
demonstrated on cells of the epithelial lineage such as
intestinal epithelial cells7 or human keratinocytes.8,13,14
Of the various members of the TLR family expressed
by human keratinocytes, a functional annotation could
recently be made for keratinocyte TLR2. We demonstra-
ted that TLR2 mediates NF-jB dependent gene activation
of inducible nitric oxide synthetase, cyclooxygenase-2 and
IL-8 by the important skin-associated pathogen Staphylo-
coccus aureus in human keratinocytes.14 Transcription of
these genes was followed by production of increased
amounts of IL-8 protein and NO. In line, the purified
staphylococcal cell wall components LTA and PGN,
known to signal through TLR2, also showed NF-jB acti-
vation in human keratinocytes, thus indicating a crucial
role of the staphylococcal cell wall in the innate immune
stimulation of human keratinocytes. The goal of the pre-
sent study was to characterize further the expression pro-
file of other members of the TLR family in addition to
TLR2 in human keratinocytes by using whole epidermis,
primary keratinocytes cultured from neonatal foreskins,
532
and the immortalized human keratinocyte cell line
HaCaT. We hypothesized that human epidermal keratino-
cytes can respond to multiple TLR ligands of different
sources through expression of various TLR family mem-
bers. We investigated the expression profile of TLR1–10
by quantitative PCR and determined the functional util-
ization of the expressed TLRs by using specific ligands of
each TLR. The usage of the NF-jB pathway in human
keratinocytes after TLR ligation was verified semiquantita-
tively through RelA staining and by quantitative methods
using the NF-jB dependent proinflammatory chemokine
IL-8 as an indicator.14 We provide evidence that human
epidermal keratinocytes express functional TLR family
members of three of the five TLR subfamilies as they are
defined by amino acid sequence similarities.10 This pat-
tern of expression of an evolutionarily conserved class of
innate immune receptors enables human keratinocytes to
respond to a wide range of pathogenic and commensal
micro-organisms, and as such keratinocytes may be
viewed as sentinels of skin homeostasis.
Materials and methods
Cell culture and reagents
Primary human keratinocytes were obtained from neo-
natal foreskins and cultured in keratinocyte serum-free
medium (Gibco/Life Technologies, Eggenstein, Germany)
as described previously.14 The immortalized keratinocyte
cell line HaCaT15 kindly provided by N. Fusenig (German
Cancer Research Center, Heidelberg, Germany), was cul-
tured at the same conditions and in the same medium as
primary cells. For in vitro stimulation assays of human
keratinocytes the following substances were used: TNF-a
(Sigma, Munich, Germany), LPS from Escherichia coli
0127:B8 (Sigma), LPS from E. coli K235 (<0
008% pro-
tein), kindly provided by Dr S. N. Vogel (Bethesda, MD),
repurified as described,16 Pam3Cys (Calbiochem, Bad
Soden, Germany), PGN (Sigma), poly (I:C) and poly
(A:U) (Sigma), recombinant flagellin from Borrelia burg-
dorferi,17 flagellin purified from Salmonella typhimurium
(Invivogen, San Diego, CA), loxoribine (Fluka, Seelze,
Germany), R-848 (Coley Pharmaceuticals, Langen, Ger-
many), CpG-ODN 2006 and non-CpG-ODN 2006K18
(synthesized by TipMolBiol, Berlin, Germany). RNA
digestion of poly (I:C) and poly (A:U) was performed
with RNAse A (Sigma). 20 lg/ml poly (I:C) or poly
(A:U) were incubated over night at room temperature
with RNAse A. The stimulation experiments of the cells
were performed at a final concentration of 20 lg/ml poly
(I:C) or poly (A:U). Sterile water incubated over night
with RNAse A served as negative control. To test, whether
cells could still be activated in the presence of RNAse A,
RNAse A-treated cells were subsequently incubated with
TNF-a. A goat polyclonal IgG antibody against TLR3

2005 Blackwell Publishing Ltd, Immunology, 114, 531–541

(Q18) was obtained from Santa Cruz (Santa Cruz, CA)
and was used for immunofluorescence staining of kera-
tinocytes14 at a 1 : 50 dilution. As secondary antibody,
fluoroscein isothiocynanate (FITC)-conjugated rabbit
anti-goat was used at a 1 : 50 dilution (Sigma).
LPS stimulation of dendritic cells (DC)
Monocyte derived DC were prepared from peripheral
blood of healthy individuals as described.19 In brief,
adherent peripheral blood mononuclear cells (> 90% pure
CD14+ cells) were cultured at 1 · 106 cells/ml in RPMI-
1640 supplemented with 1 mm sodium pyruvate, 0
1 mm
nonessential amino acids, 2 mm l-glutamine, 0
05 mm
2-mercaptoethanol, 100 U/ml penicillin, and 100 lg/ml
streptomycin (all from Life Technologies, Chagrin Falls,
OH) supplemented with 10% fetal bovine serum,
500 U/ml human recombinant granulocyte–macrophage
colony-stimulating factor (Essex Pharma, Munich, Ger-
many) and 500 U/ml human rIL-4 (Promocell, Heidel-
berg, Germany) (complete DC medium) at 37 under 5%
CO2. At day 5 cells (> 95% CD1a+, CD14–) were harves-
ted and recultured in complete DC medium for 24 hr at
37 in the presence or absence of LPS from E. coli
0127:B8 or E. coli K235 (100 ng/ml). The cells were
washed and stained for CD83, CD86, and MHC class II
and analysed by flow cytometry as described previously.20
Reverse transcription–polymerase chain reaction
(RT–PCR) and real-time PCR
For RT–PCR and real-time PCR experiments, cells were
cultured in six-well-plates either unstimulated or stimula-
ted with LPS from E. coli 0127:B8 (Sigma) or LPS from
E. coli K235 (<0
008% protein). Human skin epidermis
was obtained from neonatal foreskins according to the
same protocol used for primary keratinocytes14 except
that after the dispase digest instead of trypsinizing the
epidermal sheets, they were directly taken up in Trizol for
mRNA extraction. RNA and cDNA were prepared as des-
cribed.14 For quantitative real-time PCR of constitutive
TLR expression, primers and Taqman probes were syn-
thesized according to Zarember and Godowski,21 except
for TLR4 and TLR6, where SYBR Green (ABgene, Ham-
burg, Germany) was used with primers as published.14
TLR4, CD14 and MD-2 were quantified after 0, 2, 8 and
24 hr of LPS stimulation. Real-time PCR for CD14 and
MD-2 was also performed with SYBR Green and primers
according to Song et al.8 and Smith et al.22, respectively.
All primers and probes were purchased from MWG Bio-
tech (Ebersberg, Germany). For each real-time PCR reac-
tion 200 ng of reverse transcribed RNA were used and
input cDNA was normalized according to glyceraldehyde-
3-phosphate dehydrogenase (GAPDH) as an internal con-
trol gene. Standard curves were constructed with serial

2005 Blackwell Publishing Ltd, Immunology, 114, 531–541
Functional TLR expression in human keratinocytes
dilutions of cloned PCR products. The experiments were
performed on an ABI Prism 7700 (Applied Biosystems,
Darmstadt, Germany). Adequate positive controls for all
primers were performed as described previously.14
IL-8 secretion by keratinocytes
Primary keratinocytes and HaCaT cells were seeded in
flat-bottom 96-well-plates at a density of 2–3 · 104 cells
per well and grown to confluency. Stimulation of cells
was performed for 24 hr with TNFa (50 ng/ml) as a pos-
itive control, cell culture medium alone as a negative
control, LPS (100 ng/ml), Pam3Cys (5 lg/ml), PGN
(10 lg/ml), poly (I:C) and poly (A:U) (20 lg/ml each),
recombinant flagellin from B. burgdorferi (10 lg/ml),
native flagellin from S. typhimurium (1 lg/ml), loxoribine
(1 mm), R-848 (1 lg/ml), CpG DNA and non-CpG DNA
(1 lm each). These ligand concentrations were obtained
in preliminary experiments, where dose–response studies
for each TLR ligand were performed to determine an
optimal and physiologically relevant ligand dose. The
concentration of secreted IL-8 in the medium after 24 hr
of TLR ligand stimulation was measured by enzyme-
linked immunosorbent assay (ELISA; R & D, Wiesbaden,
Germany). Statistical analysis of the data was performed
using a standard two-sample t-test with the S-Plus soft-
ware package.
RelA assay
Primary keratinocytes were grown on chamber slides to
approximately 60% confluency and incubated for 4 hr
with medium (negative control), with TNF-a (positive
control) and with the TLR ligands LPS, Pam3Cys, PGN,
poly (I:C), poly (A:U), flagellin, loxoribine, R-848, CpG
and non-CpG DNA at the concentrations indicated
above. Incubation was stopped and cells were stained with
a primary rabbit anti-p65 antibody (Rockland, Gilberts-
ville, PA) at a dilution of 1 : 500 followed by a FITC-
labelled goat anti-rabbit antibody (Sigma) diluted 1 : 50
according to Song et al.8
Results
Expression of Toll-like receptors in human epidermis
and cultured keratinocytes
When analysing the mechanisms of TLR–mediated inter-
action of the important human skin pathogen S. aureus
with cultured human foreskin keratinocytes, we recently
found TLR2- but not TLR4-mediated signal transduction
pathways to be elicited by S. aureus mainly through PGN
and LTA.14 In addition to TLR2, other members of the
TLR family were also expressed in cultured keratinocytes
as shown by RT–PCR.14 To show a relevance of the TLR
533
G. Köllisch et al.
Table 1. Comparison of mRNA expression of TLR 1–10 between
human primary keratinocytes and human epidermis
TLR
RNA source 1 2 3 4 5 6 7 8
Human epidermis + + + – + + – –
Primary keratinocytes + ++ + – + ++ – –
The mRNA expression of TLR1–10 in freshly prepared human epi-
dermis and cultured human primary keratinocytes was analysed by
real-time PCR normalized to GAPDH. The amount of expressed
RNA is indicated by: (–) no detectable expression (+) normal
expression (++) at least 10-fold > normal expression. The PCR
reaction was run in duplicates either with epidermis from three
unrelated donors or with cultured keratinocytes from three different
foreskins.
expression pattern described for cultured human kera-
tinocytes, the mRNA expression of TLR1-TLR10 was ana-
lysed in freshly prepared human epidermis and compared
to cultured primary unstimulated keratinocytes from
human foreskins (Table 1). This analysis revealed a very
similar expression pattern on the mRNA level for TLRs in
whole human epidermis and cultured primary keratino-
cytes. Expression was observed for members of the TLR2
subfamily (TLR1, -2 and -6), for TLR3 and TLR5. In
addition, expression was found for TLR10, whose ligand
still needs to be defined. Based on the demonstrated close
similarities in the TLR expression profile between whole
epidermis and cultured keratinocytes, further quantitative
and functional analyses were carried out with cultured
primary keratinocytes. Likewise, the spontaneously
immortalized human keratinocyte cell line HaCaT, which
shows almost normal differentiation and keratinization in
skin models23 was used throughout all experiments in
comparison to primary keratinocytes.
The expression of TLR mRNA was investigated by
quantitative real-time PCR in primary keratinocytes and
HaCaT cells (Fig. 1). The most pronounced expression on
the mRNA level was found for TLR2 and TLR3 both in
primary keratinocytes and HaCaT cells. TLR4 was not
expressed in primary keratinocytes but in HaCaT cells.
The expression of TLR4 in HaCaT cells was a stable and
constant finding that was seen in two independent cell
line stocks. TLR1 mRNA was detectable in primary kera-
tinocytes at a 62
5-fold higher level than in HaCaT cells.
No detectable signals were observed for the TLR9 sub-
family members TLR7, TLR8, and TLR9 in both cell
types. TLR5 and TLR10 mRNA were expressed at 8
8-fold
and 22
5-fold higher levels in primary keratinocytes as
compared to HaCaT cells. TLR6 mRNA was present only
in primary keratinocytes but not in HaCaT cells. Further-
more, the constitutive expression of the TLR4 cofactors
CD14 and MD-2 was analyzed (Fig. 1). A 39
8-fold
534
PK
1e+5
HaCaT
RNA concentration (fg/ml)
1e+4
9 10
1e+3
– +
– +
1e+2
1
1 2 3 4 5 6 7 8 9 10 CD14 MD-2
TLR
Figure 1. Real-time PCR for TLRs, CD14 and MD-2 in primary
human keratinocytes (PK) and HaCaT cells. Cells were cultured in
six-well-plates. Constitutive TLR expression of TLR1-10, CD14, and
MD-2 mRNA was analysed by quantitative real-time PCR. Three dif-
ferent mRNA preparations of primary keratinocytes, each pooled
from three independent donors, as well as three different cell prepa-
rations of HaCaT cells were used. Columns show the mean ± stand-
ard deviation of the mRNA amount. Real-time PCR was performed
in duplicates.
higher expression of CD14 mRNA was found in HaCaT
cells compared to primary keratinocytes. In contrast,
MD-2 expression in primary keratinocytes exceeded
MD-2 expression in HaCaT cells by 157-fold.
Differential IL-8 induction by various TLR ligands
in cultured human keratinocytes
In order to characterize the functional relevance of TLRs
in epithelial cells of the skin, primary keratinocytes and
HaCaT cells were stimulated with a panel of defined
microbial and synthetic TLR ligands (Table 2). Subse-
quently, the induction of the proinflammatory chemokine
IL-8, which is under regulation of NF-jB, was quantified
by IL-8 ELISA (Fig. 2). The TLR3 ligand poly (I:C), a
synthetic analogue of double-stranded viral RNA, lead to
the strongest activation of primary (Fig. 2a) and HaCaT
(Fig. 2b) keratinocytes, whereas poly (A:U) as non speci-
fic control substance did not induce relevant amounts of
IL-8. RNAse treatment of poly (I:C) completely abolished
the activation of primary keratinocytes and HaCaT cells
demonstrating an RNA-mediated cell stimulatory mech-
anism (Fig. 3). The expression of TLR3 protein, which is
located intracellularly, changed upon keratinocyte activa-
tion with poly (I:C) to a plasma membrane-like pattern
(Fig. 3, insert). The TLR2 ligand PGN was able to activate
both cell types, although HaCaT cells reacted more pro-
nounced to PGN (Fig. 2). On the contrary, Pam3Cys,
another known ligand for TLR2 strongly stimulated pri-
mary keratinocytes, whereas HaCaT cells reacted to a
lower extent (Fig. 2). The bacterial protein flagellin, a

2005 Blackwell Publishing Ltd, Immunology, 114, 531–541
 Functional TLR expression in human keratinocytes

Table 2. Toll-like receptors (TLRs) and TLR-specific ligands used in 10·0

stimulation experiments

TLR Ligand Source

5·0
½ Pam3Cys Synthetic lipopeptide

IL-8 (ng/ml)
2 PGN S. aureus
0·6
3 Poly (I:C) Synthetic analogue of
double-stranded viral RNA
0·4
Poly (A:U)
(non-specific control)
0·2
4 LPS E. coli 0127:B8
LPS (<0
008% protein) E. coli K235
0·0
5 Flagellin B. burgdorferi

)
)
-α

i ne
s

el l i n
trol

48
PGN

CpG
LPS

G
(A:U
(I:C
3Cy

-Cp
TNF

R-8
Salmonella typhimurium

orib
Con

Fl ag
P ol y
Pam
(a)

Non
P ol y
7 R-848, loxoribine Synthetic

Lox
8 R-848 Synthetic 10·0
9 CpG ODN 2006 Synthetic unmethylated
Non-CpG ODN 2006K bacterial DNA
(non-specific control) 5·0

IL-8 (ng/ml)
PGN, peptidoglycan; LPS, lipopolysaccharide; ODN, oligodesoxy-
nucleotide. 0·6

0·4
ligand for TLR5, also stimulated both cell types (Fig. 2).
In primary keratinocytes, native flagellin from S. typhi- 0·2
murium induced a very strong IL-8 stimulation at lower
concentrations than recombinant flagellin from B. burg- 0·0
-α

ine
s

llin
trol

48

G
PGN

CpG
LPS

)
)
dorferi. In HaCaT cells, the stimulatory potential of both

(A:U
(I:C
3Cy

-Cp
TNF

R-8
e
orib
Con

Flag
flagellins was consistently lower than in primary keratino- Pam

Poly

Non
Poly
(b)

cytes (Fig. 4). Other TLR ligands, like LPS, loxoribine,
R-848, CpG and non-CpG DNA (Fig. 2) did not induce
a pronounced or specific IL-8 production as tested for
different concentrations and, in case of primary keratino-
cytes, using cells from various unrelated donors.
NF-jB induction by TLR ligands in cultured human
keratinocytes
For primary keratinocytes, the NF-jB activation and
translocation from the cytoplasm into the nucleus was
monitored in the RelA assay (Fig. 5). Corresponding to
the extent of TLR ligand-induced IL-8 secretion, poly
(I:C), flagellin, and TNFa (positive control) induced the
strongest nuclear translocation of RelA. The TLR2 ligands
PGN and Pam3Cys also stimulated nuclear translocation
of RelA (Fig. 5), although to a lesser extent, whereas LPS
(Fig. 5) and all other ligands induced no nuclear RelA
staining (data not shown).
Role of LPS and TLR4 in cultured human
keratinocytes
HaCaT cells express TLR4 at the mRNA level, but do not
produce IL-8 upon stimulation with LPS. To investigate
this feature, we analysed the possible induction of TLR4,
CD14 and MD-2 mRNA through LPS stimulation by
Lox
Figure 2. IL-8 production of primary human keratinocytes and
HaCaT cells after stimulation with various TLR ligands. Primary ker-
atinocytes (a) and HaCaT cells (b) were cultured in 96-well-plates
and stimulated with the following ligands at indicated final concen-
trations: TNF-a (50 ng/ml), LPS from E. coli 0127:B8 (100 ng/ml),
Pam3Cys (5 lg/ml), PGN (10 lg/ml), poly (I:C) and poly (A:U)
(20 lg/ml each), recombinant flagellin (10 lg/ml), loxoribine
(1 mm), R-848 (1 lg/ml), CpG and non-CpG-ODN (1 lm each).
Normal cell culture medium was used as control. The concentration
of secreted IL-8 in the medium after 24 hr of TLR ligand stimulation
was measured by ELISA. Columns show the mean ± standard devi-
ation of six independent wells. Shown is one representative IL-8
stimulation experiment out of four. Significant differences versus the
control incubation (P < 0
001) were obtained for the following TLR
ligands: TNF-a, Pam3Cys, PGN, poly (I:C), flagellin, CpG and non-
CpG-ODN (primary keratinocytes); TNF-a, PGN, poly (I:C) and
flagellin (HaCaT cells).
real-time PCR at various time points (Fig. 6). In primary
keratinocytes, TLR4 was not inducible by LPS, whereas
in HaCaT cells an LPS-dependent up-regulation of TLR4
was found (Fig. 6a). CD14 was mainly expressed in
HaCaT cells, and a strong up-regulation after LPS stimu-
lation was only seen in HaCaT cells (Fig. 6b). MD-2, on
the other hand, was expressed in primary keratinocytes,
but only marginally in HaCaT cells. LPS stimulation lead
to an insignificant increase of the detectable MD-2 signal

2005 Blackwell Publishing Ltd, Immunology, 114, 531–541 535

G. Köllisch et al.

25·0 PK HaCaT
4·0 4·0
3·5 3·5
10 µm

IL-8 (ng/ml)

IL-8 (ng/ml)
3·0 3·0
2·5 2·5
2·0 2·0
20·0
Isotype control α-TLR3 α-TLR3 1·5 1·5
IL-8 (ng/ml)

+poly(I:C) 1·0 1·0

0·5 0·5
2·5 0·0 0·0

m
trol

trol
-α

-α
ri

ri
orfe

orfe
riu

riu
TNF

TNF
2·0

Con

Con
imu

imu
urgd

urgd
1·5

yph

yph
B. b

B. b
S. t

S. t
1·0
Flagellin Flagellin
0·5
0·0 Figure 4. Analysis of IL-8 production of primary human keratino-
trol

se

e
-α

se
(I : C

(A:U

NAs

NAs
cytes (PK) and HaCaT cells after stimulation with different flagellins.
TNF

RNA

RNA
Con

Poly

Poly

)+R

)+R
Two different flagellin preparations (B. burgdorferi; S. typhimurium)
trol+

-α+
(I : C

(A:U
were tested on primary keratinocytes (a) and HaCaT cells (b).

TNF
Con

Poly

Recombinant flagellin from B. burgdorferi was used at a final concen-

Poly

tration of 10 lg/ml, whereas native purified flagellin from S. typhimu-

rium was used at a final concentration of 1 lg/ml. TNF-a (50 ng/ml)
Figure 3. Analysis of TLR3 expression (insert) and of IL-8 produc- served as a positive stimulation control, medium alone (control) as a
tion of primary human keratinocytes after stimulation with the negative stimulation control. The concentration of secreted IL-8 in
dsRNA analogue poly (I:C). Primary keratinocytes were tested for the medium after 24 hr of ligand stimulation was measured by ELISA.
specificity of TLR3 activation by poly (I:C) using RNAse A digest. Columns show the mean ± standard deviation of six independent
Medium (control), poly (I:C), and poly (A:U) were incubated over- wells. Shown is one representative IL-8 stimulation experiment out of
night with RNAse A at a concentration of 20 lg/ml and were subse- three.
quently used for stimulation of keratinocytes. Medium (control),
poly (I:C) and poly (A:U) were also used without prior RNAse treat-
ment. TNF-a (50 ng/ml) was used as another positive stimulus for by S. aureus or its defined ligands, the NK-jB pathway
IL-8 induction in keratinocytes with (TNF-a + RNAse) and without is induced in human keratinocytes.14 In this study,
RNAse treatment. The concentration of secreted IL-8 in the medium
we investigated the expression of TLR1–10 in epidermal
after 24 hr of ligand stimulation was measured by ELISA. Columns
keratinocytes and confirmed the functional relevance
show the mean ± standard deviation of six independent wells.
Shown is one representative IL-8 stimulation experiment out of
of keratinocyte TLR expression through stimulation with
three. Insert: expression of TLR3 in primary keratinocytes under TLR-specific ligands. We provide evidence that human
baseline conditions (a-TLR3) and after 16 h stimulation with poly epidermal keratinocytes and the keratinocyte cell line
(I:C) (a-TLR3 + poly (I:C)) as detected by goat polyclonal antibody HaCaT express several TLR family members. Except
directed against the amino terminus of TLR3. Non-immune goat for TLR7, -8 and -9, all TLRs are expressed in freshly
IgG was used as control. prepared human epidermis, cultured primary foreskin
keratinocytes, and in the permanent keratinocyte cell line
HaCaT. The comparison of the TLR expression profile of
in primary keratinocytes but not in HaCaT cells (Fig. 6c). primary keratinocytes and HaCaT cells revealed widely
All LPS stimulation experiments were also performed comparable data for most of the TLRs, but indicated also
with LPS from E. coli K235 (low in protein content) lead- specific differences for TLR1 and TLR4, which were dif-
ing to identical results (data not shown). As positive sti- ferentially expressed. Furthermore, the orphan receptor
mulation control, human monocyte-derived dendritic TLR10, for which a ligand has yet to be defined, and
cells were stimulated for 24 hr with 100 ng/ml of each TLR6 were found to be expressed either in both cell types
LPS preparation and a strong induction of the activation (TLR10) or in primary keratinocytes (TLR6) by quantita-
markers CD83, CD86, and HLA-DR was found in flow tive real-time PCR, which is known to be more sensitive
cytometry (Fig. 6d). Even after continuous LPS stimula- than conventional RT–PCR.24 In a previous study, we
tion of primary keratinocytes or HaCaT cells over a per- could not detect expression of TLR6 and TLR10 in pri-
iod of 4 days with LPS preparations from the two mary keratinocytes using conventional RT–PCR.14 TLR11,
different sources, no significant amount of IL-8 was pro- which has been described as a receptor that prevents
duced in either cell type at any timepoint (Fig. 7). infection by uropathogenic bacteria in the mouse and is
probably not expressed as a full-length protein in
humans,25 was not investigated in this study.
Discussion
Regarding the TLR2 subfamily (TLR1, -2, -6) func-
We recently showed that TLR2, but not TLR4, is the tional expression on human keratinocytes was found for
specific receptor for S. aureus.14 Upon ligation of TLR2 TLR1 and TLR2, which are both known to recognize

536
2005 Blackwell Publishing Ltd, Immunology, 114, 531–541


10 µm 10 µm
Unstimulated Poly(I:C)
10 µm 10 µm 10 µm
PGN Pam3Cys Flagellin
10 µm 10 µm
IM
TNF-α LPS
Figure 5. RelA assay for NF-jB translocation of primary human
keratinocytes after stimulation with various TLR ligands. For RelA
staining of primary human keratinocytes, cells were incubated for
4 hr with medium (unstimulated) as a negative control, with TNF-a
(50 ng/ml) as a positive control, and with the TLR ligands poly
(I:C), PGN, Pam3Cys, flagellin, and LPS at the same concentrations
used for the IL-8 assay (see legend to Fig. 2). Incubation was
stopped and cells were stained with a primary rabbit anti-p65 anti-
body followed by an FITC-labelled goat anti-rabbit antibody. Cells
which are not reactive to the stimulus are represented by a cytoplas-
mic staining pattern, reactive cells by a nuclear one as denoted by
arrows. Arrowheads point to cells representing a typical fluorescence
pattern for each of the incubations.
bacterial cell wall components. Pam3Cys26 activated IL-8
secretion in primary keratinocytes to a higher degree than
in HaCaT cells. Whereas PGN is recognized by TLR2
alone27 Pam3Cys recognition is mediated by a heterodi-
mer of TLR2 and TLR1
9 Based on the higher extent of
TLR1 mRNA expression in primary keratinocytes as com-
pared to HaCaT cells, it can be speculated that primary
keratinocytes are also able to form a higher number of
TLR1/TLR2 heterodimers than HaCaT cells which could
explain the predominant reactivity of primary keratino-
cytes to Pam3Cys.

2005 Blackwell Publishing Ltd, Immunology, 114, 531–541
Functional TLR expression in human keratinocytes
TLR3 constitutes a separate subfamily within the mam-
malian TLRs that is characterized by intracellular expres-
sion.28 Ligation of TLR3, which senses double-stranded
RNA10 was the most prominent stimulus for IL-8 secre-
tion in primary human keratinocytes. TLR3 activation
was achieved by the TLR3 ligand poly (I:C)29 a synthetic
analogue of double-stranded viral RNA. Incubation of
keratinocytes with poly (I:C) also induced a redistribution
of TLR3 staining to a more plasma membrane-like stain-
ing pattern (see Fig. 3). It will be interesting to differenti-
ate whether the poly (I:C) mediated effects can solely be
attributed to TLR3 or are also dependent on double-
stranded RNA-activated protein kinase, an interferon-c
(IFN-c)-induced protein constitutively expressed in
human keratinocytes.30 Recent findings of Lebre et al.31
underline the importance of double-stranded viral RNA
reactivity in keratinocytes for subsequent IFN-c-mediated
immune reactions. Thus, human keratinocytes might
possibly sense viral infections through their functionally
active TLR3 receptors which enable the initiation of an
immediate innate and delayed adaptive immune response
to viruses infecting the skin.
TLR4 was the first human homologue of Toll to be
described32 and was subsequently characterized as a
receptor for LPS signalling.33 In addition to LPS, TLR4
also recognizes other ligands such as the fusion protein of
respiratory syncytial virus, stress-induced members of the
endogenous heat-shock protein family, and taxol in the
murine system.10 TLR4 is the sole member of the TLR4
subfamily. Controversial results have recently been repor-
ted regarding the role of LPS reactivity mediated by TLR4
in primary human keratinocytes.34,35 Some authors8,36,37
found TLR4 and CD14 to be expressed by primary and
HaCaT keratinocytes and keratinocytes to be activated by
LPS, whereas Kawai et al.13 showed no TLR4 and CD14
expression and no LPS reactivity in keratinocyte cultures.
There is some evidence, however, that TLR4 is expressed
on epidermal keratinocytes in normal human skin in the
absence of CD14 expression.13 On the other hand, TLR4
expression was shown to be dependent on keratinocyte
differentiation.37 The functional consequence of this find-
ing is still unclear. Our results strengthen the finding that
TLR4 is neither constitutively expressed nor functional in
cultured primary keratinocytes. HaCaT cells, on the other
hand, express TLR4 and CD14 at the mRNA level, but
nevertheless do not react to LPS, which might be due to
the lack of MD-2 expression.38 Recent studies clarified the
early molecular events involved in TLR4-mediated LPS
signalling. MD-2 was shown to enable TLR4 binding of
LPS and the formation of stable receptor complexes.
CD14 on the other hand enhances LPS binding to
MD-2.38 Thus, even in the presence of TLR4 expression,
as it is the case in HaCaT cells, no functional LPS signal-
ling can be expected without a sufficient expression of
both, MD-2 and CD14.
537
G. Köllisch et al.

Medium
Isotype
LPS E. coli 0127:B8
LPS E. coli K235

8000 CD83
TLR4

0 10 20 30 40 50 60 70 80 90
RNA concentration (fg/ml)

7000
6000
PK
5000
HaCaT

Counts
4000
3000
2000
1000
0
0 2 8 24 100 101 102 103 104
(a) Time (hr) FL2-H

8000 CD14 CD86

0 10 20 30 40 50 60 70 80 90
RNA concentration (fg/ml)

7000
6000 PK
HaCaT
5000
Counts

4000
3000
2000
1000
0 Figure 6. Induction of TLR4, CD14 and MD-2
0 2 8 24 100 101 102 103 104 by LPS in primary keratinocytes (PK) and
(b) Time (hr) FL1-H HaCaT cells. Cells were cultured in six-well-
plates stimulated with LPS (100 ng/ml) from
8000 HLA-DR E. coli 0127:B8. The LPS-inducible expression of
MD-2
0 10 20 30 40 50 60 70 80 90
RNA concentration (fg/ml)

7000 TLR4 (a), CD14 (b) and MD-2 (c) was quanti-
6000 fied after 0, 2, 8 and 24 hr of LPS stimulation.
PK
5000
Real-time PCR was performed as described in
HaCaT
the legend to Fig. 1. The activity of the LPS
Counts

4000
preparations used was demonstrated in human
3000 monocyte-derived dendritic cells (d). After 24 h
2000 incubation with 100 ng/ml LPS (LPS from
1000 E. coli 0127:B8 and LPS from E. coli K235), the
0 cells were stained for the activation markers
0 2 8 24 100 101 102 103 104 CD83, CD86, and MHC class II (HLA-DR) and
(c) Time (hr) (d) FL1-H analysed by flow cytometry.

TLR5, a separate subfamily of the mammalian Toll specific recognition of the bacterial component flagellin
homologues, recognizes the bacterial motor protein flagel- through keratinocyte TLR5 and the flagellin-induced sti-
lin through a conserved site on flagellin required for pro- mulation of epithelial IL-8 production could thus enable
tofilament formation and bacterial motility.22,39 Our data human skin to react to invading flagellated bacteria. Sur-
provide evidence that TLR5 was expressed both in pri- prisingly, our experiments with primary keratinocytes
mary human keratinocytes and in the permanent kera- showed an even more pronounced induction of IL-8 after
tinocyte cell line HaCaT at the mRNA level and the stimulation with flagellin from S. typhimurium, which is
corresponding ligand, flagellin, induced NF-jB transloca- not a common skin pathogen. However, typical skin
tion and IL-8 secretion. HaCaT cells were not as reactive lesions are known to occur during the course of enteric
towards flagellin as primary keratinocytes, which might fever.40 Regarding the preparation of the two flagellins
correspond to the lower level of TLR5 mRNA expression used in our study, it must be noted that S. typhimurium
in the cell line. The finding of TLR5 expression and reac- flagellin was a purified native protein, whereas flagellin
tivity in human epidermal keratinocytes is of particular from B. burgdorferi was a recombinant protein produced
interest as skin infecting bacteria such as B. burgdorferi, in E. coli. Furthermore, B. burgdorferi flagellin has an
which causes migratory erythema during the course of identity score of only 56
4% and 44
4% with S. typhimu-
Lyme disease, are known to produce flagellin.17 The rium flagellin in the conserved flagellin sites recognized

538
2005 Blackwell Publishing Ltd, Immunology, 114, 531–541

 Functional TLR expression in human keratinocytes

4·0 PK infectious process.10 In addition, synthetic ligands with

TLR7 and TLR8 agonistic activity such as imiquimod,
3·5
R848, and loxoribine have been described. In our experi-
3·0 ments, the synthetic immunomodulators R-848 (specifi-
2·5
city for TLR7 and TLR8)42,43 and loxoribine (specificity
IL-8 (ng/ml)

for TLR7)44,45 did not cause any IL-8 secretion or nuclear

2·0 RelA staining, which corresponds well with the lack of
1·5 TLR7 and TLR8 mRNA expression in primary keratino-
cytes and HaCaT cells. However, Schön et al.46 could
1·0
show an apoptotic effect of the R-848 analogue imiqui-
0·5 mod (TLR7 ligand) in tumour-derived keratinocytes, but
the possible involvement of TLR7 was not investigated in
0·0
their study. Interestingly, the TLR7 ligand imiquimod
-α
-α

-α

-α
trol

trol

trol

trol
LPS

LPS

LPS

LPS
TNF
TNF

TNF

TNF
now is increasingly used for the treatment of various skin
Con

Con

Con

Con disorders where keratinocyte transformations occur, such

Day 1 Day 2 Day 3 Day 4 as in genital and common warts, in actinic keratosis, and
more recently also in basal cell carcinoma.47 As we did
4·0
HaCaT not observe a significant basal expression of TLR7 in cul-
3·5 tured normal keratinocytes, corresponding to the lack of
irritation caused by imiquimod on normal human skin,
3·0
the possible induction and up-regulation of TLR7 as a
2·5 target of imiquimod in papillomavirus-infected or trans-
IL-8 (ng/ml)

formed keratinocytes should be investigated. As for TLR9

2·0
and reactivity to immunostimulatory DNA sequences18,48
1·5 recent studies demonstrated activation of keratinocytes by
1·0
CpG-ODN without investigating TLR9 expression in par-
ticular.49 In multiple experiments using cells from differ-
0·5 ent keratinocyte donors, we could not find any specific
0·0 effect of CpG-ODN at concentrations up to 5 lm. Only
-α

-α

-α

-α
trol

trol

trol

trol
LPS

LPS

LPS

LPS

nonspecific activation of primary keratinocytes both by

TNF

TNF

TNF

TNF
Con

Con

Con

Con

CpG-ODN and non-CpG-ODN was obtained, which can

Day 1 Day 2
Figure 7. Analysis of IL-8 production of primary human keratino-
cytes (PK) and HaCaT cells after stimulation with LPS over a time
period of 4 days. Primary keratinocytes and HaCaT cells were incu-
bated with 100 ng/ml LPS from E. coli K235. After 24 hr, the super-
natant was removed for analysis of IL-8 production. The same cells
were incubated with a new preparation of LPS from E. coli K235
(100 ng/ml) for another 24 hr. This procedure was repeated over a
time period of 4 days (96 hr). TNF-a (50 ng/ml) served as a positive
stimulation control, medium alone (control) as a negative stimula-
tion control. The concentration of secreted IL-8 in the medium after
24 hr of LPS stimulation was measured by ELISA. Columns show
the mean ± standard deviation of six independent wells. Shown is
one representative IL-8 stimulation experiment out of three.
by TLR5.41 Therefore, no final conclusions on the differ-
ential activities of the two different flagellins are justified.
However, our data provide substantial evidence for a
strong overall reactivity of human keratinocytes to flagel-
lin from different bacterial sources.
Members of the TLR9 subfamily (TLR7, 8, 9) of mam-
malian Toll homologues primarily sense pathogen-derived
RNA and DNA motifs generated intracellularly in the
not be explained by TLR9-mediated DNA sequence-speci-
Day 3 Day 4
fic effects. Because no TLR9 mRNA expression was
observed with primers optimized for quantitative real-
time PCR, it is not unexpected that no specific CpG-
ODN effect was detectable. The non-specific CpG-ODN
effects could be caused by the polyanion character of the
ODN at higher concentrations. Moreover, it was recently
demonstrated that the TLR9–CpG interaction only occurs
after TLR9 recruitment from the endoplasmatic reticulum
to a tubular lysosomal compartment, a subcellular struc-
ture typically found in professional antigen-presenting
cells but not in epithelial cells.50
Our data illustrate the diversity of TLR-mediated
pattern recognition pathways present in human epider-
mal keratinocytes, which indicates that human skin may
initiate a first line response to a variety of pathogen-
derived components. With few exceptions such as in the
case of TLR4, a strong correlation of the TLR mRNA
expression pattern and the functional reactivity of pri-
mary keratinocytes and HaCaT cells to the various TLR
ligands was found. However, differences observed in the
constitutive and/or functional expression of some TLRs
and TLR cofactor molecules between primary keratino-
cytes and HaCaT cells also demonstrate the importance

2005 Blackwell Publishing Ltd, Immunology, 114, 531–541 539

G. Köllisch et al.
of using primary cells in addition to immortalized cell
lines, when investigating epithelial TLR expression and
stimulation patterns, and possible functional conse-
quences thereof. This study provides evidence for a TLR
expression and response profile of normal human
keratinocytes which extends beyond the role of TLR2
described for responses to S. aureus.14 The variety in
TLR expression may even indicate a role for human
keratinocytes as sentinels of skin homeostasis. Further
studies will have to elucidate the particular role and sig-
nalling response mediated by the various TLRs expressed
by human epidermal keratinocytes and the relative
importance of these TLRs for the innate cutaneous
immune response.
Acknowledgements
We thank G. Lipford from Coley Pharmaceuticals for
providing R-848. We also acknowledge the skilful tech-
nical assistance of G. Roth, B. Heuser, B. Dorn and
B. Maar. This work was supported in part by grant
UW-S15T03 (Sub-project 3b) from the German National
Genome Research Network (NGFN) to T.J and M.O.,
by grant 01GC0104 (Sub-projects 2, 3 and 4) from the
German Federal Ministery of Science and Education
(BMBF) to T.J., S.B., M.M., and M.O., and by KKF grants
870160 and 8760133 from the Medical School (Klinikum
rechts der Isar) of the Technische Universität München
(TUM) to M.M and M.O., respectively.
References
1 Kim J, Modlin RL. Innate Immunity and the skin. In: Freedberg,
IM, Eisen, AZ, Wolff, K, Austen, KF, Goldsmith, LA, Katz, SI,
eds. Fitzpatrick’s Dermatology in General Medicine, Vol. 1. New
York: McGraw-Hill, 2003:247–52.
2 Kupper TS, Fuhlbrigge RC. Immune surveillance in the skin:
mechanisms and clinical consequences. Nat Rev Immunol 2004;
4:211–22.
3 Stingl G, Maurer D, Hauser C, Wolff K. The skin: an immuno-
logic barrier. In: Freedberg, IM, Eisen, AZ, Wolff, K, Austen, KF,
Goldsmith, LA, Katz, SI, eds. Fitzpatrick’s Dermatology in Gen-
eral Medicine, Vol. 1. New York: McGraw-Hill, 2003:253–73.
4 Dinulos JG, Mentele L, Fredericks LP, Dale BA, Darmstadt GL.
Keratinocyte expression of human beta defensin 2 following
bacterial infection: role in cutaneous host defense. Clin Diagn
Lab Immunol 2003; 10:161–6.
5 Harder J, Bartels J, Christophers E, Schroder JM. A peptide anti-
biotic from human skin. Nature 1997; 387:861.
6 Chung WO, Dale BA. Innate immune response of oral and fore-
skin keratinocytes: utilization of different signaling pathways by
various bacterial species. Infect Immun 2004; 72:352–8.
7 Cario E, Brown D, McKee M, Lynch-Devaney K, Gerken G,
Podolsky DK. Commensal-associated molecular patterns induce
selective Toll-like receptor-trafficking from apical membrane to
cytoplasmic compartments in polarized intestinal epithelium.
Am J Pathol 2002; 160:165–73.
540
8 Song PI, Park YM, Abraham T, Harten B, Zivony A, Neparidze
N, Armstrong CA, Ansel JC. Human keratinocytes express func-
tional CD14 and toll-like receptor 4. J Invest Dermatol 2002;
119:424–32.
9 Akira S, Hemmi H. Recognition of pathogen-associated molecu-
lar patterns by TLR family. Immunol Lett 2003; 85:85–95.
10 Wagner H. The immunobiology of the TLR9 subfamily. Trends
Immunol 2004; 25:381–6.
11 Beutler B, Rehli M. Evolution of the TIR, tolls and TLRs: func-
tional inferences from computational biology. Curr Top Micro-
biol Immunol 2002; 270:1–21.
12 Takeda K, Kaisho T, Akira S. Toll-like receptors. Annu Rev
Immunol 2003; 21:335–76.
13 Kawai K, Shimura H, Minagawa M, Ito A, Tomiyama K, Ito M.
Expression of functional Toll-like receptor 2 on human epider-
mal keratinocytes. J Dermatol Sci 2002; 30:185–94.
14 Mempel M, Voelcker V, Köllisch G et al. Toll-like receptor
expression in human keratinocytes: Nuclear factor jB-controlled
gene activation by Staphylococcus aureus is TLR2- but not
TLR4- or platelet activating factor receptor (PAFR)-dependent.
J Invest Dermatol 2003; 121:1389–96.
15 Boukamp P, Petrussevska RT, Breitkreutz D, Hornung J, Mark-
ham A, Fusenig NE. Normal keratinization in a spontaneously
immortalized aneuploid human keratinocyte cell line. J Cell Biol
1988; 106:761–71.
16 McIntire FC, Sievert HW, Barlow RA, Finley RA, Lee AY.
Chemical, physical, and biological properties of a lipopoly-
saccharide from Escherichia coli K-235. Biochemistry 1967;
6:2363–73.
17 Wallich R, Moter SE, Simon MM, Ebnet K, Heiberger A,
Kramer MD. The Borrelia burgdorferi flagellum-associated
41-kilodalton antigen (flagellin): molecular cloning, expres-
sion, and amplification of the gene. Infect Immun 1990; 58:
1711–9.
18 Bauer S, Kirschning CJ, Hacker H, Redecke V, Hausmann S,
Akira S, Wagner H, Lipford GB. Human TLR9 confers respon-
siveness to bacterial DNA via species-specific CpG motif recog-
nition. Proc Natl Acad Sci USA 2001; 98:9237–42.
19 Sallusto F, Lanzavecchia A. Efficient presentation of soluble anti-
gen by cultured human dendritic cells is maintained by granulo-
cyte/macrophage colony-stimulating factor plus interleukin 4
and downregulated by tumor necrosis factor. J Exp Med 1994;
179:1109–18.
20 Jakob T, Udey MC. Regulation of E-cadherin-mediated adhesion
in Langerhans cell-like dendritic cells by inflammatory mediators
that mobilize Langerhans cells in vivo. J Immunol 1998; 160:
4067–73.
21 Zarember KA, Godowski PJ. Tissue expression of human Toll-
like receptors and differential regulation of Toll-like receptor
mRNAs in leukocytes in response to microbes, their products,
and cytokines. J Immunol 2002; 168:554–61.
22 Smith MF Jr, Mitchell A, Li G, Ding S, Fitzmaurice AM, Ryan
K, Crowe S, Goldberg JB. Toll-like receptor (TLR) 2 and TLR5,
but not TLR4, are required for Helicobacter pylori-induced
NF-?B activation and chemokine expression by epithelial cells.
J Biol Chem 2003; 278:32552–60.
23 Breitkreutz D, Schoop VM, Mirancea N, Baur M, Stark HJ,
Fusenig NE. Epidermal differentiation and basement membrane
formation of HaCaT cells in surface transplants. Eur J Cell Biol
1998; 75:273–86.

2005 Blackwell Publishing Ltd, Immunology, 114, 531–541

24 Dager H, Donninger H, Hutchinson P, Ghildyal R, Bardin P.
Rhinovirus detection: comparison of real-time and conventional
PCR. J Virol Methods 2004; 117:113–21.
25 Zhang D, Zhang G, Hayden MS, Greenblatt MB, Bussey C,
Flavell RA, Gosh S. A Toll-like receptor that prevents infection
by uropathogenic bacteria. Science, Year?; 303:1522–6.
26 Takeuchi O, Sato S, Horiuchi T, Hoshino K, Takeda K, Dong Z,
Modlin RL, Akira S. Role of Toll-like receptor 1 in mediating
immune response to microbial lipoproteins. J Immunol 2002;
169:10–4.
27 Schwandner R, Dziarski R, Wesche H, Rothe M, Kirschning CJ.
Peptidoglycan- and lipoteichoic acid-induced cell activation is
mediated by toll-like receptor 2. J Biol Chem 1999; 274:17406–9.
28 Funami K, Matsumoto M, Oshiumi H, Akazawa T, Yamamoto
A, Seya T. The cytoplasmic ‘linker region’ in Toll-like receptor 3
controls receptor localization and signaling. Int Immunol 2004;
16:1143–54.
29 Alexopoulou L, Holt AC, Medzhitov R, Flavell RA. Recognition
of double-stranded RNA and activation of NF-jB by Toll-like
receptor 3. Nature 2001; 413:732–8.
30 Kuyama M, Nakanishi G, Arata J, Iwatsuki K, Fujimoto W.
Expression of double-stranded RNA-activated protein kinase in
keratinocytes and keratinocytic neoplasia. J Dermatol 2003;
30:579–89.
31 Lebre MC, Antons JC, Kalinski P, Schuitemaker JH, van Capel
TM, Kapsenberg ML, De Jong EC. Double-stranded RNA-
exposed human keratinocytes promote Th1 responses by indu-
cing a Type-1 polarized phenotype in dendritic cells. role of
keratinocyte-derived tumor necrosis factor alpha, type I interfer-
ons, and interleukin-18. J Invest Dermatol 2003; 120:990–7.
32 Medzhitov R, Preston-Hurlburt P, Janeway CA Jr. A human
homologue of the Drosophila Toll protein signals activation of
adaptive immunity. Nature 1997; 388:394–7.
33 Poltorak A, He X, Smirnova I et al. Defective LPS signaling
in C3H/HeJ and C57BL/10ScCr mice: mutations in Tlr4 gene.
Science 1998; 282:2085–8.
34 Kawai K. Expression of functional Toll-like receptors on cul-
tured human epidermal keratinocytes. J Invest Dermatol 2003;
121:217.
35 Song PI, Armstrong CA, Ansel JC. Human normal keratinocytes
express biologically functional CD14 and Toll-like receptors.
J Invest Dermatol 2003; 121:217–8.
36 Pivarcsi A, Bodai L, Rethi B et al. Expression and function of
Toll-like receptors 2 and 4 in human keratinocytes. Int Immunol
2003; 15:721–30.
37 Pivarcsi A, Koreck A, Bodai L et al. Differentiation-regulated
expression of Toll-like receptors 2 and 4 in HaCaT keratino-
cytes. Arch Dermatol Res 2004; 296:120–4.

2005 Blackwell Publishing Ltd, Immunology, 114, 531–541
Functional TLR expression in human keratinocytes
38 Visintin A, Latz E, Monks BB, Espevik T, Golenbock DT.
Lysines 128 and 132 enable LPS binding to MD-2, leading to
Toll-like receptor 4 aggregation and signal transduction. J Biol
Chem 2003; 278:48313–20.
39 Hayashi F, Smith KD, Ozinsky A et al. The innate immune
response to bacterial flagellin is mediated by Toll-like receptor 5.
Nature 2001; 410:1099–103.
40 Blume JE, Levine EG, Heymann WR. Bacterial Diseases. In:
Bologna, JL, Lorizzo, JL, Rapini, RP, Horn, TH, Mascaro, JM,
Mancini, AJ, Salasche, SJ, Saurat, J-H, Stingl, G, eds. Dermatol-
ogy, Vol. 1. London: Mosby, 2003:1117–44.
41 Smith KD, Andersen-Nissen E, Hayashi F, Strobe K, Bergman
MA, Barrett SL, Cookson BT, Aderem A. Toll-like receptor 5
recognizes a conserved site on flagellin required for protofila-
ment formation and bacterial motility. Nat Immunol 2003;
4:1247–53.
42 Hemmi H, Kaisho T, Takeuchi O et al. Small anti-viral com-
pounds activate immune cells via the TLR7 MyD88-dependent
signaling pathway. Nat Immunol 2002; 3:196–200.
43 Jurk M, Heil F, Vollmer J, Schetter C, Krieg AM, Wagner H,
Lipford G, Bauer S. Human TLR7 or TLR8 independently confer
responsiveness to the antiviral compound R-848. Nat Immunol
2002; 3:499.
44 Heil F, Ahmad-Nejad P, Hemmi H et al. The Toll-like recep-
tor 7 (TLR7)-specific stimulus loxoribine uncovers a strong
relationship within the TLR7, 8 and 9 subfamily. Eur J Immunol
2003; 33:2987–97.
45 Lee J, Chuang TH, Redecke V, She L, Pitha PM, Carson DA,
Raz E, Cottam HB. Molecular basis for the immunostimulatory
activity of guanine nucleoside analogs: activation of Toll-like
receptor 7. Proc Natl Acad Sci USA 2003; 100:6646–51.
46 Schön M, Bong A, Drewniok C et al. Tumor-selective induction
of apoptosis and the small-molecule immune response modifier
imiquimod. J Natl Cancer Inst 2003; 95:1138–49.
47 Tyring S, Conant M, Marini M, Van Der Meijden W, Washenik
K. Imiquimod; an international update on therapeutic uses in
dermatology. Int J Dermatol 2002; 41:810–6.
48 Hemmi H, Takeuchi O, Kawai T et al. A Toll-like receptor
recognizes bacterial DNA. Nature 2000; 408:740–5.
49 Mirmohammadsadegh A, Maschke J, Basner-Tschakarjan E,
Bar A, Hengge UR. Induction of acute phase response genes in
keratinocytes following exposure to oligodeoxynucleotides. J Mol
Med 2002; 80:377–83.
50 Latz E, Schoenemeyer A, Visintin A et al. TLR9 signals after
translocating from the ER to CpG DNA in the lysosome.
Nat Immunol 2004; 5:190–8.
541