Content uploaded by Stefano Cavicchini
Author content
All content in this area was uploaded by Stefano Cavicchini
Content may be subject to copyright.
Low Molecular Weight Hyaluronic Acid Increases the
Self-Defense of Skin Epithelium by Induction of

-Defensin 2
via TLR2 and TLR4
Silvia Gariboldi,
1
* Marco Palazzo,
1
* Laura Zanobbio,* Silvia Selleri,* Michele Sommariva,*
Lucia Sfondrini,
†
Stefano Cavicchini,
‡
Andrea Balsari,
†§
and Cristiano Rumio
2
*
In sites of inflammation or tissue injury, hyaluronic acid (HA), ubiquitous in the extracellular matrix, is broken down into low
m.w. HA (LMW-HA) fragments that have been reported to activate immunocompetent cells. We found that LMW-HA induces
activation of keratinocytes, which respond by producing

-defensin 2. This production is mediated by TLR2 and TLR4 activation
and involves a c-Fos-mediated, protein kinase C-dependent signaling pathway. LMW-HA-induced activation of keratinocytes
seems not to be accompanied by an inflammatory response, because no production of IL-8, TNF-
␣
, IL-1

, or IL-6 was observed.
Ex vivo and in vivo treatments of murine skin with LMW-HA showed a release of mouse

-defensin 2 in all layers of the epidermal
compartment. Therefore, the breakdown of extracellular matrix components, for example after injury, stimulates keratinocytes
to release

-defensin 2, which protects cutaneous tissue at a time when it is particularly vulnerable to infection. In addition, our
observation might be important to open new perspectives in the development of possible topical products containing LMW-HA
to improve the release of

-defensins by keratinocytes, thus ameliorating the self-defense of the skin for the protection of cutaneous
tissue from infection by microorganisms. The Journal of Immunology, 2008, 181: 2103–2110.
The stratum corneum of the skin is the first barrier that
pathogenic bacteria have to cross to penetrate into the
organism. Keratinocytes of the epidermis not only have
an important structural role in forming a physical barrier to foreign
Ags and microorganisms, but they also secrete soluble factors with
antibacterial activity, i.e.,

-defensins (1–3).
The release of

-defensins by keratinocytes may be particularly
important in the case of skin lesions when, in the absence of the
stratum corneum, microorganisms may reach the connective tissue
of the dermis. In this context, the tissue damage may induce deg-
radation of the extracellular matrix with the release of hyaluronic
acid (HA)
3
fragments. HA is a high m.w. (HMW) glycosamino-
glycan, which is ubiquitous in the extracellular matrix; it is in-
volved in maintaining the water balance, in the distribution and
transport of plasmatic proteins, and in maintaining an intact matrix
structure. In sites of inflammation or tissue injury, HMW-HA may
be depolymerized in low m.w. (LMW) fragments through the ac-
tivity of oxygen radicals or via enzymatic activity by hyaluroni-
dase,

-glucuronidase, and hexosaminidase. In contrast to the
HMW form, which is biologically inert regarding its ability to
activate immune cells, LMW fragments are able to activate the
innate immune defense, promoting the production of different cy-
tokines. Recently, it has been suggested that LMW-HA may trig-
ger TLR2 and TLR4 in immunocompetent cells, stimulating the
production of chemokines and cytokines by macrophages (4) and
activating dendritic cells and T cells (5, 6).
We have recently shown (7) that murine skin expresses proteins
belonging to the family of TLRs and that the stimulation of murine
skin with their specific ligands, such as LPS, peptidoglycan (PGN),
and flagellin, induces production of

-defensin 2 (DEFB2).
DEFB2, a peptide produced by different epithelial cells, exerts a
strong antimicrobial activity against Gram-negative bacteria and
Candida albicans, together with a good bacteriostatic activity
against Gram-positive bacteria (8). Chronnell et al. (9) showed that
keratinocytes control the proliferation of the microflora that resides
in the pilosebaceous unit by the release of DEFB2. It is possible
that LMW-HA, released following skin injury, might stimulate
TLR2 and TLR4, inducing DEFB2 production.
In the present article we have evaluated the effect of LMW-HA
on the production of DEFB2 and the involvement of TLR2 and
TLR4 in this production through experiments in vitro using a ker-
atinocyte cell line, ex vivo on murine or human skin samples, and
in vivo by the application of LMW-HA on mouse dorsal skin.
Materials and Methods
Cell treatments
NCTC 2544 human keratinocytes (American Type Culture Collection
(ATCC)) were cultured in DMEM Glutamax with 10% FBS and 10 ml/L
penicillin/streptomycin (all from Invitrogen). NHEK (normal human epi-
dermal keratinocyte) primary human keratinocytes (Lonza) were cultured
in KGM keratinocyte medium (Lonza) in a serum-free environment. For
stimulation experiments on NCTC 2544 cells, the FBS concentration was
reduced to 2.5%. For experiments on DEFB2 stimulation, cells were cul-
tured in medium alone or in medium added with LPS from Escherichia coli
(5
g/ml; Sigma-Aldrich) or PGN from E. coli (10
g/ml, InVivoGen) or
*Mucosal Immunity Laboratory, Department of Human Morphology and
†
Institute of
Pathology, Universita` degli Studi di Milano, Milan, Italy;
‡
Unita` Operativa Derma-
tologia, Fondazione Ospedale Maggiore Policlinico, Mangiagalli e Regina Elena Is-
tituto di Ricovero e Cura a Carattere Scientifico, Milan, Italy; and
§
Istituto Nazionale
per lo Studio e la Cura dei Tumori, Milan, Italy
Received for publication December 18, 2007. Accepted for publication May 20, 2008.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact.
1
S.G. and M.P. contributed equally to this work.
2
Address correspondence and reprint requests to Prof. Cristiano Rumio, Faculty of
Pharmacy, Department of Human Morphology, Universita` degli Studi di Milano, Via
Mangiagalli 31, 20133 Milan, Italy. E-mail address: cristiano.rumio@unimi.it
3
Abbreviations used in this paper: HA, hyaluronic acid; ChIP, chromatin immuno-
precipitation; DEFB2,

-defensin 2; HMW, high m.w.; LEF-1, lymphoid enhancer-
binding factor-1; LMW, low m.w.; NHEK, normal human epidermal keratinocyte;
PGN, peptidoglycan; PKC, protein-kinase C.
Copyright © 2008 by The American Association of Immunologists, Inc. 0022-1767/08/$2.00
The Journal of Immunology
www.jimmunol.org
in medium containing different concentrations (0.025–0.5%) of LMW-HA
sodium salt (M
r
⬍200 kDa, obtained by biosynthesis; Soliance).
LMW-HA concentrations were chosen according to literature (10 –13) and
were dependent on the solubility of the sodium salt and on cell viability.
Following preliminary dose-response experiments, 0.2% LMW-HA
was chosen as standard treatment dosage because this is also a classical
concentration of HA that may be found in topical products (14 –16). The
absence of endotoxin contamination in the LMW-HA sodium salt was
initially assessed by the Limulus amebocyte lysate (LAL) test (Sigma-
Aldrich); however, because the LAL assay may be too insensitive to detect
small amounts of LPS and does not detect lipoproteins, additional controls
included both hyaluronidase (bovine testicular hyaluronidase; Sigma-
Aldrich) treatment of LMW-HA (3 h at 37°C with 10 U of hyaluronidase
per 50
l LMW-HA solution) and boiling of the LMW-HA solution to
denature possible protein contaminants. Incubation with a solution con-
taining 0.2% HMW-HA sodium salt or 0.2% LMW-HA plus 0.2%
HMW-HA sodium salts served as additional controls. All experiments
were performed in triplicate.
For evaluation of DEFB2 release, media were collected at different time
points from 30 min to 18 h after the beginning of treatment. For evaluation
of IL-8, TNF-
␣
, IL-1, and IL-6 release in the medium, NCTC 2544 su-
pernatants were collected 18 h after the beginning of treatment. For ex-
periments on TLR blocking, anti-TLR2, anti-TLR4 (Santa Cruz Biotech-
nology), and anti-LEF-1 (lymphoid enhancer-binding factor-1; Oncogene
Research) Abs were used at different concentrations (1–20
g/ml) during
the 18 h of incubation with LPS, PGN or 0.2% LMW-HA. Protein kinase
C (PKC) was inhibited with a 5
M mixture of rottlerin and bisindolyl-
maleimide (Sigma-Aldrich), the respective inhibitors of PKC-
␦
and classic
PKCs (such as
␣
and

); the concentration used was chosen according to
data from the literature. PKC-inhibited cells were treated with HA as re-
ported above and their DEFB2 production was evaluated by PCR analysis
and ELISA. All experiments were performed in triplicate.
To assess NF-
B involvement in LMW-HA-induced pathway, NCTC
2544 cells were incubated for 18 h in medium containing 0.2% LMW-HA
sodium salt in the presence of 5
M NF-
B inhibitor BAY-7082 (Alexis
Biochemicals); at the end of treatment, supernatants were collected and
evaluated for their DEFB2 content by ELISA.
Treatment of human skin
Biopsies of human dorsal skin have been obtained from patients undergo-
ing routine esthetic dermatosurgery; ethical approval was obtained and
patients gave informed consent. Samples of 4 ⫻4-mm size were obtained
and the ipodermis was mechanically removed. The specimens for the ex
vivo assay were incubated in medium alone or in a LMW-HA solution at
different concentrations (0.025–0.5%) for 3 h; at the end of treatment the
supernatants were collected for evaluation of DEFB2 by ELISA while skin
specimen were immediately frozen in liquid nitrogen and successively used
for the antimicrobial assay.
Mouse treatments
C57BL/6 wild-type mice were from Charles River Laboratories; TLR4-
deficient mice were provided by Dr. Besusso, Istituto Nazionale Tumori,
Milan, Italy. Experimental protocols were approved by the Ethics Com-
mittee for Animal Experimentation of the Istituto Nazionale Tumori, Mi-
lan. For studies on DEFB2 production ex vivo, dorsal murine skin frag-
ments (n⫽5/group) were incubated with medium alone, with the specific
ligands for TLR2 (PGN) and TLR4 (LPS), or with 0.2% LMW-HA. The
LMW-HA dosage was chosen according to the previous in vitro experi-
mental data and by referring to the usual HA concentration found in topical
products (14 –16). Each specimen was treated for 18 h at 37°C and pro-
cessed for mRNA extraction.
For studies on DEFB2 production in vivo, the dorsal skin of wild-type
and TLR4-deficient mice was treated with a mixture of wax and rosin to
allow for the complete removal of all hair shafts. On the 8th day after
depilation mice (n⫽5/group) were treated three times with a cream con-
taining 1% LMW-HA or with vehicle only (100 mg/dose). After 24 h of
treatment, dorsal skin samples were collected and processed for mRNA
extraction or fixed in 10% neutral buffered formalin and embedded in par-
affin. Sections were stained for DEFB2 by immunohistochemistry.
PCR analysis
Expression of DEFB2 was investigated in NCTC 2544 cells and in murine
skin by RT-PCR. Total RNA was isolated and converted into cDNA. PCR
was performed using the cMaster PCR enzyme mix (Eppendorf). The PCR
cycle for human DEFB2 included denaturation at 94°C for 5 min followed
by 35 cycles of denaturation at 94°C for 1 min, annealing at 62°C for 1
min, extension at 72°C for 2 min, and a final extension at 72°C for 10 min.
The PCR profile for murine DEFB2 included denaturation at 94°C for 5
min, followed by 35 cycles of denaturation at 94°C for 1 min, annealing at
60°C for 1 min, extension at 72°C for 1 min, and a final extension at 72°C
for 10 min. The following primers (Primm) were used: human DEFB2
forward, 5⬘-TTTGGTGGTATAGGCGATCC-3⬘; human DEFB2 reverse
5⬘-GAGGGAGCCCTTTCTGAATC-3⬘; mouse DEFB2 forward, 5⬘-GCC
ATGAGGACTCTCTGCTC-3⬘; and mouse DEFB2 reverse, 5⬘-AGGGG
TTCTTCTCTGGGAAA-3⬘. The PCR products were visualized on an
ethidium bromide 1% agarose gel.
Immunofluorescence
NCTC 2544 keratinocytes were subcultured on cover glasses and fixed in
methanol. For TLR2 and TLR4 staining, cells were incubated with the
respective primary Abs and secondary Ab donkey anti-goat tetramethyl-
rhodamine isothiocyanate (Molecular Probes). For DEFB2 staining,
cells were incubated with swine serum (DakoCytomation), rabbit anti-
DEFB2 Ab (Alpha Diagnostic International), and then with goat anti-
rabbit FITC Ab (Molecular Probes); nuclei were counterstained with
4⬘,6⬘-diamidino-2-phenylindole.
ELISA
Production of human DEFB2 in NCTC 2544 and NHEK cell supernatants
or in medium from ex vivo experiments on human skin was quantified
using an enzyme immunoassay kit from Phoenix Pharmaceuticals. Con-
centrations of IL-8, TNF-
␣
, IL-1

, and IL-6 in NCTC 2544 supernatants
were evaluated using ELISA kits obtained from GE Healthcare (IL-8),
Endogen (TNF-
␣
), and MedSystem Diagnostic (both IL-1

and IL-6) and
conducted according to the manufacturers’ instructions.
Western blot analysis
NF-
B and c-Fos nuclear expressions were evaluated in proteins extracted
from treated and untreated NCTC 2544 cells using a nuclear/cytosol fraction-
ation kit (MBL International). Nuclear proteins (15
g) were fractionated on
a 8% acrylamide slab gel containing 0.1% SDS and transferred onto a nitro-
cellulose filter by electroblotting. For detection of NF-
B, the filter with nu-
clear extracts was incubated with mouse Ab to NF-
B p65 (catalog no. sc-
8008, Santa Cruz Biotechnology), whereas for detection of the c-Fos signal the
filter was incubated with rabbit anti-c-Fos Ab (catalog no. sc-253, Santa Cruz
Biotechnology). Total cellular extract was used for evaluation of IkB expres-
sion using a rabbit anti-IkB Ab (catalog no. ab32518, Abcam). To normalize
protein expression of different samples,

-actin protein expression was used as
a reference loading control for total cellular extracts by the use of a mouse
anti-

-actin Ab (catalog no. AC-15, Abcam), while lamin B protein expression
served as nuclear protein control using a goat anti-lamin B Ab (catalog no.
sc-6216, Santa Cruz Biotechnology).
Chromatin immunoprecipitation (ChIP) assay
ChIP assay was performed on NCTC 2544 cells treated for 18 h with 0.2%
LMW-HA or left untreated to confirm the involvement of c-Fos in the
LMW-HA-induced DEFB2 expression. Briefly, 10,000 cells/sample were
cross-linked with 1% formaldehyde for 10 min, lysed, and sonicated using
a Fisher Scientific 550 sonic dismembrator. An aliquot of the sonicated
cells was amplified as input control. Sonicated cells were immunoprecipi-
tated with specific Ab to human c-Fos (catalog no. sc-8047, Santa Cruz
Biotechnology). After reversing the cross-linking, PCR was performed
with the following primers that amplify part of the DEFB2 promoter: 5⬘-
GAGGAATTTTCTGGTCCCAAG-3⬘(forward) and 5⬘-CCATGAGGGTC
TTGTATCTCCTC-3⬘(reverse). The PCR cycle included denaturation at
95°C for 5 min followed by 40 cycles of denaturation at 95°C for 1 min,
annealing at 60°C for 1 min, extension at 72°C for 1 min 30 s, and a final
extension at 72°C for 10 min. The PCR products were visualized on an
ethidium bromide agarose gel and normalized to input control.
Immunohistochemistry
We used an immunohistochemical technique to localize DEFB2 protein in
sections of vehicle- and LMW-HA cream-treated mouse skin samples. Par-
affin sections of 4-
m thickness were placed on silanized slides, deparaf-
finized, and rehydrated. After Ag retrieval by enzymatic treatment with a
0.5% trypsin solution for 20 min at 37°C, the quenching of endogenous
peroxidases was performed in 0.3% H
2
O
2
in PBS for 30 min at room
temperature. To avoid nonspecific binding, the slides were blocked with
rabbit serum diluted 1/50 in PBS/BSA 1% for 30 min at room temperature.
Mouse skin sections were then incubated with goat anti-DEFB2 Ab (1/100
in PBS; catalog no. sc-10858 from Santa Cruz Biotechnology) for1hat
room temperature, rinsed in PBS, and incubated with rabbit anti-goat sec-
ondary Ab (1/100 in PBS) for1hatroom temperature. After incubation
2104 IMMUNOMODULATORY ACTIVITY OF HYALURONAN FRAGMENTS IN SKIN
with goat peroxidase-antiperoxidase (1/100 in PBS) for1hatroom tem-
perature, the staining was revealed using diaminobenzidine. Then mouse
skin samples were washed, dehydrated through an ascending series of eth-
anols, and mounted with Entellan. The same immunohistochemical anal-
ysis was performed on vehicle and LMW-HA cream-treated TLR4
⫺/⫺
mice skin samples.
Antimicrobial assay
Biopsies of human dorsal skin treated with LMW-HA or medium alone and
frozen in liquid nitrogen were pulverized with a pestle in liquid nitrogen
and proteins were extracted under gentle agitation for2hin5%acetic acid
with the addition of protease inhibitors (0.02 mM PMSF, 2 ng/ml pepstatin,
and 2 ng/ml leupeptin). The soluble proteins in the supernatant were dried
under vacuum and resuspended in 0.01% acetic acid. Protein concentra-
tions were determined by bicinchoninic assay (BCA; Pierce
Biotechnology).
For examination of the activity of antimicrobial peptides produced by
skin biopsies, we first incubated 10
4
bacteria/ml (OD
600
) of mid-logarith-
mic-phase E. coli ATCC 4157 (International PBI) in phosphate buffer (pH
7.4) in a final volume of 100
l. The E. coli suspension was mixed with 30
g of the skin protein extracts or with 10
g of DEFB2 recombinant
peptide (positive control) (Millipore) or with 10
l of 0.01% acetic acid
and protease inhibitor mixture (negative control) and incubated at 37°C for
120 min. At the end of the incubation, 10
l of these bacteria suspensions
and of 1/10 and 1/100 dilutions were plated in triplicate in petri dishes
containing tryptic soy broth (International PBI) and incubated overnight at
37°C. At the end of incubation, the number of CFU for each sample was
determined.
Results
Human keratinocytes produce

-defensin 2 in response to the
TLR2 agonist PGN and the TLR4 agonist LPS
The human keratinocyte NCTC 2544 cell line was found to ex-
press TLR2 and TLR4 proteins (Fig. 1A), as previously observed
on other keratinocyte cell lines and primary keratinocytes (17, 18).
An intense production of DEFB2, assessed by PCR and immuno-
fluorescence analysis, was observed when cells were stimulated
with the TLR2- and TLR4-specific bacterial ligands PGN and LPS,
respectively, for 18 h (Fig. 1B).
LMW-HA treatment induces keratinocytes to release DEFB2 but
not proinflammatory mediators via activation of TLR2 and TLR4
To evaluate whether LMW-HA might be able to induce expression
of DEFB2 in keratinocytes, NCTC 2544 cells were incubated for
18 h with a medium containing 0.025– 0.5% LMW-HA; LPS-
treated cells were used as positive control (19). Cells treated with
0.2% of LMW-HA showed a marked production of DEFB2 as
revealed by immunofluorescence and PCR analysis (Fig. 2, Aand
B). Quantification of DEFB2 secretion at different time points of
treatment (30 min, 1 h, 6 h, and 18 h) and with different LMW-HA
concentrations (0.025– 0.5%), assessed by ELISA, showed that the
levels of secreted DEFB2 significantly increased with a progres-
sive accumulation of the peptide in the culture medium over time
(p⬍0.005 vs control) and depending on the concentration of
LMW-HA in the culture medium (Fig. 2, Cand D). Additional
experiments assessed DEFB2 production following incubation of
NCTC 2544 with a 0.2% HMW-HA solution or a 0.2% LMW-HA
plus 0.2% HMW-HA solution. ELISA showed that HMW-HA was
FIGURE 1. TLR and DEFB2 expression in human keratinocytes. A,
Cells were analyzed by immunofluorescence analysis for TLR2 and
TLR4 expression (CTRL, negative control by omission of primary an-
tibody). B, Following treatment with PGN or LPS, cells were analyzed
for their DEFB2 expression in comparison with untreated cells (CTRL).
Bars, 10
m.
FIGURE 2. LMW-HA-induced
DEFB2 expression in human keratin-
ocytes. A–D, Secretion of DEFB2 by
NCTC 2544 human keratinocytes fol-
lowing treatment with LMW-HA has
been shown by immunofluorescence
(A; with 0.2% LMW-HA solution;
bars, 10
m.), PCR (B; with 0.2%
LMW-HA solution), and ELISA on
cells supernatants at different time
points after the beginning of treat-
ment (C; with 0.2% LMW-HA solu-
tion) and with different LMW-HA
concentrations (D). E, Additional ex-
periments assessed the production of
DEFB2 following treatment with
HMW-HA or after digestion or boil-
ing of the 0.2% LMW-HA solution.
F, Secretion of DEFB2 by NHEK pri-
mary keratinocytes following treat-
ment with 0.2% LMW-HA was addi-
tionally analyzed. LPS was used as
positive control. ⴱ,p⬍0.005 vs con-
trol (CTRL).
2105The Journal of Immunology
unable to induce DEFB2 production, while the presence of
HMW-HA does not inhibit the activity of LMW-HA (Fig. 2E).
Additional in vitro experiments demonstrated that the observed
DEFB2 production was not due to the presence of LPS or lipopro-
teins contaminants, because digestion of our LMW-HA solution
with hyaluronidase resulted in complete loss of induction of
DEFB2, whereas boiling the same solution to denature possible
protein contaminants did not affect activity of LMW-HA, as shown
by ELISA (Fig. 2E).
We then treated NHEK primary keratinocytes for 18 h with
0.2% LMW-HA or with medium alone and evaluated their DEFB2
production by ELISA; our results confirmed the data previously
obtained on the keratinocyte cell line, because LMW-HA in-
creased DEFB2 production by NHEK with respect to control cells
(Fig. 2F)(p⬍0.005 vs control).
Trying to understand whether the observed DEFB2 production
following LMW-HA treatment had to be ascribed to the stimula-
tion of TLR2 and TLR4, cells were incubated for 18 h with 0.2%
LMW-HA and an Ab for TLR2 and/or TLR4 blocking. Keratino-
cytes incubated with PGN or LPS and the relative specific anti-
TLR Abs served as control. Cells were analyzed by immunofluo-
rescence analysis for their DEFB2 production; in cells incubated
with one of the TLR-blocking Abs (20
g/ml), the HA-induced
DEFB2 production was substantially diminished in comparison to
that of cells treated with HA alone (Fig. 3A). Keratinocytes incu-
bated with HA and both Abs, anti-TLR2 and anti-TLR4, showed
an almost complete inhibition of DEFB2 production (Fig. 3A). An
unrelated anti-LEF-1 Ab (20
g/ml) showed no interference in the
LMW-HA-induced DEFB2 production (Fig. 3B). Immunofluores-
cence results were confirmed via ELISA by the dosage of DEFB2
peptide released into the cell medium using different concentra-
tions of TLR-blocking Abs (1–20
g/ml), which showed progres-
sive inhibition of the TLR-driven DEFB2 production depending on
the concentrations of the blocking Abs (Fig. 3C).
To assess whether HA induces the release of other molecules
of innate immune response and proinflammatory mediators, we
evaluated IL-8, TNF-
␣
, IL-1

, and IL-6 production in treated
keratinocytes. ELISA test revealed that the levels of these mol-
ecules in HA-treated cells were comparable to basal production
after 18 h of treatment (Fig. 4). LPS-treated cells served as
positive control (20, 21).
FIGURE 3. Involvement of TLR2 and TLR4 in HA-induced DEFB2
production. A, Blocking of TLR2 and TLR4 by specific Abs inhibited
the LMW-HA-induced DEFB2 production as shown by immunofluo-
rescence analysis. B, An unrelated anti-LEF-1 Ab served as control.
Bars, 10
m. C, ELISA performed on cell supernatants with different
Ab concentrations confirmed the results previously obtained by
immunolocalization.
FIGURE 4. Activity of LMW-HA
on proinflammatory mediators. ELISA
on cell supernatants revealed no pro-
duction of IL-8 (A), TNF-
␣
(B), IL-1

(C), and IL-6 (D) by 0.2% LMW-HA-
treated keratinocytes. CTRL, Control.
2106 IMMUNOMODULATORY ACTIVITY OF HYALURONAN FRAGMENTS IN SKIN
Ex vivo treatment of human skin and ex vivo/in vivo treatment
of murine skin with LMW-HA induce DEFB2 production
After assessment of DEFB2 production by human keratinocytes
following LMW-HA treatment, we evaluated the production of the
same antimicrobial peptide in an ex vivo model of human skin.
Skin biopsies were treated in vitro for 3 h with a 0.2% hyaluronan
solution or with LPS used as positive control for DEFB2 produc-
tion (19). An ELISA on human skin supernatants (Fig. 5A) re-
vealed an intense DEFB2 production in samples treated with 0.2%
LMW-HA, while untreated samples resulted negatively for
DEFB2 production ( p⬍0.05 vs control).
After our assessments on the ability of LMW-HA to induce
DEFB2 production in human keratinocyte cell cultures and in hu-
man skin biopsies ex vivo, we thought it was important to develop
an in vivo model also to assess whether our findings obtained in
very simplified models might be transposable to a more complex
system such as an in vivo model; in fact, this would allow us to
think about a possible use of LMW-HA for the stimulation of skin
innate immunity and to test whether the administration of
LMW-HA by topical route might be as efficacious as using it in the
previously tested culture medium. For this aim, the use of an an-
imal model was essential; in any case, in mouse the ortholog of
human DEFB2 does not exist. Thus, we have decided to analyze
the production of murine mDEFB2 because, analogously to human
hDEFB2, mDEFB2 is the mostly inducible peptide in the family of
murine

-defensins (22, 23). We started our assays on the murine
model by developing an ex vivo experiment on mouse skin to
assess whether LMW-HA had an analogous DEFB2-increasing ef-
fect on the murine mDEFB2 peptide; Fig. 5Bshows by PCR anal-
ysis that LMW-HA induces a marked increase in mDEFB2 tran-
scription, analogously to the previously observed effects on human
samples.
Following the evaluation of DEFB2 production by hyaluronan-
treated skin ex vivo, we evaluated the effects of in vivo treatment
of mice with a cream formulation containing 1% of the previously
tested LMW-HA. Following depilation of dorsal skin, mice (n⫽
5/group) were treated three times in 1 day with the cream contain-
ing LMW-HA or with vehicle cream only (100 mg/dose). After
24 h of treatment, dorsal skin samples were collected and DEFB2
production by cutaneous compartment was evaluated by both PCR
analysis and immunohistochemistry. PCR analysis showed an in-
tense transcription of the DEFB2 gene in hyaluronan cream-treated
animals (Fig. 5C); immunohistochemistry showed that the antimi-
crobial peptide is present in all of the layers of the epidermal com-
partment (Fig. 5D). The same experiment performed on TLR4-
deficient mice showed a massive reduction of DEFB2 induction by
LMW-HA (Fig. 5D).
Extracts of human skin biopsies treated with LMW-HA exert
increased antimicrobial activity
Given the observed increase in DEFB2 production induced by
LMW-HA treatment in both keratinocytes and human or murine
skin, we aimed to evaluate the antimicrobial activity in protein
extracts of skin biopsies treated with different concentrations
(0.025– 0.5%) of LMW-HA. Therefore, we prepared protein ex-
tracts from LMW-HA-treated and untreated human skin biopsies
and tested the antibacterial activity of these extracts against E. coli
ATCC 4157, a bacterial strain that is reported to be sensitive to
DEFB2 (24). We evaluated the number of CFU in the different
samples (n⫽5/treatment) after overnight incubation of bacteria
previously mixed with the protein extracts; as shown by Table I,
the extract from untreated skin induces a slight inhibition of bac-
terial CFU formation, whereas the skin treated with increasing
concentrations of LMW-HA exerts a progressively higher antimi-
crobial activity with respect to control untreated skin.
TLR-driven DEFB2 production following hyaluronan treatment
involves a c-Fos-mediated, PKC-dependent pathway
We analyzed the pathway of DEFB2 production following TLRs
activation by LMW-HA because TLRs activation may induce dif-
ferent intracellular signaling cascades. First of all, we further in-
vestigated the involvement of the TLR pathway by analyzing the
Table I. Evaluation of the percentage inhibition of E. coli CFU
formation following incubation of bacteria with extracts of human skin
samples treated with LMW-HA
Skin Sample
Percentage Inhibition of
CFU Formation vs 0.01%
Acetic Acid-Treated Skin (%)
Control 22.7
0.025% LMW-HA 36.8
0.05% LMW-HA 49.9
0.1% LMW-HA 61.6
0.2% LMW-HA 73.5
0.5% LMW-HA 97.7
FIGURE 5. Ex vivo and in vivo activity of LMW-
HA. A, Ex vivo experiments performed on human skin
samples revealed an intense secretion of human DEFB2
in the supernatant following LMW-HA treatment as
shown by ELISA (ⴱ,p⬍0.05 vs CTRL). B, Murine
skin samples treated with 0.2% LMW-HA in vitro
showed induction of murine DEFB2 transcription by
PCR analysis. Cand D, The same DEFB2 stimulation
was observed following in vivo treatment of murine skin
with a formulation containing LMW-HA as assessed by
PCR (C) and immunohistochemistry on wild-type mice,
while TLR4-deficient animals showed a massive reduc-
tion of DEFB2 induction (D). Bars, 20
m. CTRL, Con-
trol; KO, knockout.
2107The Journal of Immunology
contribution of PKC, a molecule known to be involved in the TLR
pathway (25), in the observed DEFB2 production. We used rot-
tlerin and bisindolylmaleimide, two specific inhibitors of PKC
␦
,
␣
, and

(26); the DEFB2 production of inhibited cells fol-
lowing LMW-HA treatment was significantly lower than the
one of uninhibited cells ( p⬍0.005 LMW-HA vs LMW-HA
plus PKC inhibitors) as shown by both PCR analysis and
ELISA (Fig. 6, Aand B).
Among the transcription factors activated by TLR2 and 4,
NF-
B is certainly one of the most frequently involved. Therefore,
we aimed to determine whether this transcription factor was re-
sponsible for the LMW-HA-induced DEFB2 expression; hence,
we determined by Western blot analysis the presence of the NF-
B
p65 protein in the nuclear compartment of HA-treated cells, which
indicates its activation by translocation from the cytosol to the
nucleus. As shown in Fig. 6C, after 18 h of LMW-HA treatment
nuclear NF-
B protein expression was comparable to basal levels,
indicating a complete absence of NF-
B activation; this result sug-
gests that NF-
B seems not to be involved in the observed DEFB2
production. Because inactivated NF-
B dimers are sequestered in
the cytosol of cells via noncovalent interactions with a class of
inhibitor proteins called IkBs, the absence of NF-
B involvement
in this pathway was confirmed by Western blot analysis for IkB in
the total protein extract; no alterations in the expression of this
protein was found in HA-treated cells compared with untreated
cells (Fig. 6C). LPS-treated cells served as a positive control (27).
Additional experiments to exclude the involvement of NF-
B were
performed. Keratinocytes were cultured with 0.2% LMW-HA as
previously described in presence of BAY-7082, an inhibitor of
NF-
B. ELISA for DEFB2 on the collected supernatants showed
a production of the antimicrobial peptide that is comparable to the
one obtained without the NF-
B inhibitor (Fig. 6D).
Because genomic analysis of DEFB2 revealed a promoter re-
gion containing several putative transcription factor binding sites,
including AP-1 (28), we evaluated the nuclear translocation of
c-Fos, one of the members of the AP-1 dimer. Western blot anal-
ysis showed that after 18 h of treatment keratinocytes present a
marked increase in c-Fos nuclear expression, which might indicate
its involvement in the HA-driven DEFB2 production (Fig. 6C). In
addition, given the supposed involvement of c-Fos observed by
Western blot analysis, we chose to determine whether c-Fos binds
to and potentially regulates the DEFB2 promoter in vitro. There-
fore, keratinocytes were cultured with 0.2% LMW-HA or medium
alone, as previously described. ChIP was then performed with an
Ab specific to c-Fos to examine binding to the DEFB2 promoter in
human keratinocytes. The DEFB2 binding region was successfully
amplified from our cells, with notably stronger c-Fos binding in the
LMW-HA-treated keratinocytes than in untreated cells (Fig. 6E).
These data show that c-Fos is the transcription factor involved in
the increase of DEFB2 production following LMW-HA treatment
of keratinocytes.
Discussion
Our data show that LMW-HA is an efficient inducer of DEFB2
production; this finding, observed in a keratinocyte cell line, has
been confirmed by experiments in primary keratinocytes and in ex
vivo and in vivo samples of both murine and human skin.
Because, in the event of skin injury, the degradation of the ex-
tracellular matrix releases LMW-HA fragments in the dermis (29),
our data might explain the intense presence of DEFB2 peptide
observed by Schmidt et al. (30) and Butmarc et al. (31) after injury
to skin, even in the absence of infection. These studies on the
increase of human

-defensin 2 expression in wound healing are
further supported by studies made by Sørensen et al. (32, 33), who
found that the growth factors important in wound healing, insulin-
like growth factor I and TGF-
␣
, induce the expression of several
antimicrobial peptides/polypeptides in human keratinocytes such
as human cationic antimicrobial protein hCAP-18/LL-37, human

-defensin 3, neutrophil-gelatinase-associated lipocalin, and secre-
tory leukocyte protease inhibitor, defining a host defense role for
growth factors in wound healing. Thus, in the case of skin injury,
both extracellular matrix components and growth factors may con-
tribute to the production of an array of protective antimicrobial
peptides.
Microbial CFU assay showed a progressive inhibitory activity
of extracts from skin treated with increasing concentrations of
FIGURE 6. Analysis of the signaling pathway activated by LMW-HA
treatment. Aand B, Inhibition of PKC reduced DEFB2 production follow-
ing LMW-HA treatment of keratinocytes as assessed by PCR analysis (A)
and ELISA (B)(ⴱ,p⬍0.005 LMW-HA vs LMW-HA plus PKC inhibi-
tors). C, Western blot analysis showed an increase of nuclear translocation
of c-Fos following LMW-HA-treatment, whereas no difference has been
noted regarding NF-
B and IkB with respect to control (CTRL) cells;

-actin expression (for total cellular extracts) and lamin B expression (for
nuclear extracts) were used as loading controls. D, Treatment of keratin-
ocytes with LMW-HA and the NF-
B inhibitor BAY-7082 showed no
inhibition of DEFB2 production. E, In addition, ChIP analysis showed a
stronger c-Fos binding to the DEFB2 promoter region in the LMW-HA-
treated keratinocytes than in untreated (CTRL) cells.
2108 IMMUNOMODULATORY ACTIVITY OF HYALURONAN FRAGMENTS IN SKIN
LMW-HA and it is plausible that DEFB2 greatly contributed to
this antimicrobial activity, although other antimicrobial peptides
can be produced by the skin (34). The stimulation of defensin
expression in sites of skin injury may decrease the possible bac-
terial contamination by pathogens and may even facilitate tissue
repair by providing innate immunity through the chemotactic ac-
tivity of these antimicrobial peptides (35).
Blocking experiments using Abs against TLR2 and TLR4
clearly show that LMW-HA-induced DEFB2 production involves
the activation of these two receptors. Although other surface-ex-
pressed receptors, such as CD44 and S1P receptors, are reported to
bind LMW-HA, the results of our experiments with specific block-
ing Abs have excluded their involvement in this signaling because
they completely blocked LMW-HA-induced DEFB2 production.
TLRs, in addition to the recognition of conserved microbial prod-
ucts that signal the presence of an infection, have been found to
detect endogenous ligands that signal danger conditions such as
degradation products of macromolecules and products of proteo-
lytic cascades (35). The previously reported (5, 6) ligand activity
of LMW-HA on TLR2 and TLR4 may be linked to the structure of
HA, which consists of a repeating disaccharide with features of
“pathogen-associated molecular patterns” (29). Because HA is
usually present in the dermis in the form of a HMW polymer, the
presence of important amounts of LMW fragments due to a tissue
injury is perceived by the epithelial compartment as a danger sig-
nal for the organism (36), which responds with DEFB2 release to
fight an eventual microorganism invasion.
It is noteworthy that proinflammatory mediators and chemo-
kines such as TNF-
␣
, IL-1

, IL-6, and IL-8, which frequently are
produced following TLR activation, were not detected in the su-
pernatants of LMW-HA-activated keratinocytes (37). Accord-
ingly, Uehara et al. (38) observed in different epithelial cell lines
that the stimulation of TLRs with synthetic ligands induced
DEFB2 expression without concomitant IL-6, IL-8, and MCP-1
secretion. It has also been reported that DEFB2 mRNA expression
in gingival epithelial cells and tissue, induced by several natural
stimuli, is regulated differently from that of IL-8 (39). The TLR
signaling pathway may activate a variety of transcription factors,
including NF-
B (p50/p65) and AP-1 (c-Fos/c-Jun), that lead to
the activation of different genes. Previous genomic analysis of hu-
man DEFB2 revealed a promoter region containing several puta-
tive transcription factor binding sites, including AP-1 (40). More-
over, cloning of the human DEFB2 gene has revealed that it is
unique among defensin genes, having three binding sites for the
transcription factor NF-
B (41). Western blot analysis of nuclear
extracts showed a marked increase of c-Fos but not NF-
B nuclear
translocation in cells treated with LMW-HA. Additional experi-
ments confirmed that the use of an inhibitor of NF-
B did not
interfere with LMW-HA-induced DEFB2 production; moreover,
ChIP analysis demonstrated the ability of c-Fos to link the DEFB2
promoter of keratinocytes following LMW-HA treatment. There-
fore, our data implicate a role for the AP-1 transcription factor
family in DEFB2 regulation in response to LMW-HA treatment of
keratinocytes and suggest that members of the NF-
B transcription
factor family are not required for DEFB2 regulation. Clearly, this
does not exclude the possibility that NF-
B may be involved in
DEFB2 regulation in other epithelial cell lines or in the presence of
other stimulants.
Our immunohistochemistry experiments on murine LMW-HA-
treated skin also showed mDEFB2 expression in hair follicles,
indicating that the administration of exogenous LMW-HA by top-
ical application is able to stimulate not only the easily accessible
epidermis keratinocytes but also the less reachable hair canals. The
induction of DEFB2 in the infundibulum and the sebaceous duct of
the hair follicle, which represent frequent ways of access for in-
vading microorganisms, suggests a possible use of HA for topical
application as an antimicrobial peptide stimulator in controlled mi-
crobial-related hair pathologies. To our knowledge, this is the first
time that LMW-HA has been proposed as active topical agent.
In conclusion, although the skin immune defense has been clas-
sically ascribed only to the immune cells, which reside in a matrix
typically considered inert, our data show that the matrix itself,
through the activity of its components’ breakdown products, may
actively stimulate the production of DEFB2 by the overlying epi-
dermis. Our data, which indicate a production of DEFB2 following
LMW-HA treatment, may clarify a physiological mechanism that
is probably present in the skin in the presence of endogenous dan-
ger signals to avoid an eventual bacterial infection, although this
assessment of ours is certainly only speculative at the moment,
based on data of the literature. In addition, the observation that the
topical application of LMW-HA on the skin is able to induce an
antibacterial response suggests the possible use of this natural mol-
ecule as a physiological stimulator of cutaneous antibacterial
activity.
Disclosures
The authors have no financial conflict of interest.
References
1. Kanda, N., and S. Watanabe. 2007. Histamine enhances the production of human

-defensin-2 in human keratinocytes. Am. J. Physiol. 293: C1916 –C1923.
2. Braff, M. H., and R. L. Gallo. 2006. Antimicrobial peptides: an essential com-
ponent of the skin defensive barrier. Curr. Top. Microbiol. Immnunol. 306:
91–110.
3. Donnarumma, G., I. Paoletti, E. Buommino, M. Orlando, M. A. Tufano, and
A. Baroni. 2004. Malassezia furfur induces the expression of

-defensin-2 in
human keratinocytes in a protein kinase C-dependent manner. Arch. Dermatol.
Res. 295: 474 – 481.
4. Jiang, D., J. Liang, J. Fan, S. Yu, S. Chen, Y. Luo, G. D. Prestwich,
M. M. Mascarenhas, H. G. Garg, D. A. Quinn, et al. 2005. Regulation of lung
injury and repair by Toll-like receptors and hyaluronan. Nat. Med. 11:
1173–1179.
5. Termeer, C., F. Benedix, J. Sleeman, C. Fieber, U. Voith, T. Ahrens, K. Miyake,
M. Freudenberg, C. Galanos, and J. C. Simon. 2002. Oligosaccharides of hya-
luronan activate dendritic cells via Toll-like receptor 4. J. Exp. Med. 195:
99 –111.
6. Scheibner, K. A., M. A. Lutz, S. Boodoo, M. J. Fenton, J. D. Powell, and
M. R. Horton. 2006. Hyaluronan fragments act as an endogenous danger signal
by engaging TLR2. J. Immunol. 177: 1272–1281.
7. Selleri, S., F. Arnaboldi, M. Palazzo, S. Gariboldi, L. Zanobbio, E. Opizzi,
Y. F. Shirai, A. Balsari, and C. Rumio. 2007. Toll-like receptor agonists regulate

-defensin 2 release in hair follicle. Br. J. Dermatol. 156: 1172–1177.
8. Schroder, J. M., and J. Harder. 1999. Human

-defensin 2. Int. J. Biochem. Cell
Biol. 31: 645– 651.
9. Chronnell, C. M., L. R. Ghali, R. S. Ali, A. G. Quinn, D. B. Holland, J. J. Bull,
W. J. Cunliffe, I. A. McKay, M. P. Philpott, and S. Muller-Rover. 2001. Human

defensin-1 and -2 expression in human pilosebaceous units: upregulation in
acne vulgaris lesions. J. Invest. Dermatol. 117: 1120 –1125.
10. Sikkink, C. J., M. M. Reijnen, P. Falk, H. van Goor, and L. Holmdahl. 2005.
Influence of monocyte-like cells on the fibrinolytic activity of peritoneal me-
sothelial cells and the effect of sodium hyaluronate. Fertil. Steril. 84 (Suppl. 2):
1072–1077.
11. Zou, L., X. Zou, L. Chen, H. Li, T. Mygind, M. Kassem, and C. Bu¨nger. 2007.
Effect of hyaluronan on osteogenic differentiation of porcine bone marrow stro-
mal cells in vitro. J. Orthop. Res. 26: 713–720.
12. Santangelo, K. S., A. L. Johnson, A. S. Ruppert, and A. L. Bertone. Effects of
hyaluronan treatment on lipopolysaccharide-challenged fibroblast-like synovial
cells. 2007. Arthritis Res. Ther. 9: R1.
13. Rosines, E., H. J. Schmidt, and S. K. Nigam. 2007. The effect of hyaluronic acid
size and concentration on branching morphogenesis and tubule differentiation in
developing kidney culture systems: potential applications to engineering of renal
tissues. Biomaterials 28: 4806 – 4817.
14. Trabucchi, E., S. Pallotta, M. Morini, F. Corsi, R. Franceschini, A. Casiraghi,
A. Pravettoni, D. Foschi, and P. Minghetti. 2002. Low molecular weight hyal-
uronic acid prevents oxygen free radical damage to granulation tissue during
wound healing. Int. J. Tissue React. 24: 65–71.
15. Lee, J. H., J. Y. Jung, and D. Bang. 2007. The efficacy of topical 0.2% hyaluronic
acid gel on recurrent oral ulcers: comparison between recurrent aphthous ulcers
and the oral ulcers of Behc¸et’s disease. J. Eur. Acad. Dermatol. Venereol. 22:
590 –595.
16. Nolan, A., C. Baillie, J. Badminton, M. Rudralingham, and R. A. Seymour. 2006.
The efficacy of topical hyaluronic acid in the management of recurrent aphthous
ulceration. J. Oral. Pathol. Med. 35: 461– 465.
2109The Journal of Immunology
17. Pivarcsi, A., A. Koreck, L. Bodai, M. Szell, C. Szeg, N. Belso,
A. Kenderessy-Szabo, Z. Bata-Csorgo, and A. Dobozy. 2004. Differentiation-
regulated expression of Toll-like receptors 2 and 4 in HaCaT keratinocytes. Arch.
Dermatol. Res. 296: 120 –124.
18. Lebre, M. C., A. M. van der Aar, L. van Baarsen, T. M. van Capel,
J. H. Schuitemaker, M. L. Kapsenberg, and E. C. de Jong. 2007. Human kera-
tinocytes express functional Toll-like receptor 3, 4, 5 and 9. J. Invest. Dermatol.
127: 331–341.
19. Chadebech, P., D. Goidin, C. Jacquet, J. Viac, D. Schmitt, and M. J. Staquet.
2003. Use of human reconstructed epidermis to analyze the regulation of

-de-
fensin hBD-1, hBD-2, and hBD-3 expression in response to LPS. Cell Biol.
Toxicol. 19: 313–324.
20. Kis, K., L. Bodai, H. Polyanka, K. Eder, A. Pivarcsi, E. Duda, G. Soos,
Z. Bata-Csorgo, and L. Kemeny. 2006. Budesonide, but not tacrolimus, affects
the immune functions of normal human keratinocytes. Int. Immunopharmacol. 6:
358 –368.
21. Ballanger, F., I. Tenaud, C. Volteau, A. Khammari, and B. Dre´no. 2008. Anti-
inflammatory effects of lithium gluconate on keratinocytes: a possible explana-
tion for efficiency in seborrhoeic dermatitis. Arch. Dermatol. Res. 300: 215–223.
22. Morrison, G. M., D. J. Davidson, and J. R. Dorin. 1999. A novel mouse

defensin, Defb 2, which is upregulated in the airways by lipopolysaccharide.
FEBS Lett. 442: 112–116.
23. Ikeda, A., Y. Nakanishi, T. Sakimoto, J. Shoji, M. Sawa, and N. Nemoto. 2006.
Expression of

defensins in ocular surface tissue of experimentally developed
allergic conjunctivitis mouse model. Jpn. J. Ophthalmol. 50: 1– 6.
24. Harder, J., J. Bartels, E. Christophers, and J. M. Schro¨der. 1997. A peptide an-
tibiotic from human skin. Nature 387: 861.
25. Kim, D. C., S. H. Kim, M. W. Jeong, N. I. Baek, and K. I. Kim. 2005. Effect of
rottlerin, a PKC
␦
inhibitor, on TLR-4 dependent activation of murine microglia.
Biochem. Biophys. Res. Commun. 337: 110 –115.
26. Aksoy, E., M. Goldman, and F. Willems. 2004. Proteine kinase C
⑀
: a new target
to control inflammation and immune-mediated disorders. Int. J. Biochem. Cell
Biol. 36: 183–188.
27. Raghav, S. K., B. Gupta, A. Shrivastava, and H. R. Das. 2007. Inhibition of
lipopolysaccharide-inducible nitric oxide synthase and IL-1

through suppres-
sion of NF-
B activation by 3-(1⬘-1⬘-dimethyl-allyl)-6-hydroxy-7-methoxy-cou-
marin isolated from Ruta graveolens L. Eur. J. Pharmacol. 560: 69 – 80.
28. Lu, Z., K. Kim, M. A. Suico, T. Shuto, J. Li, and H. Kai. 2004. MEF up-regulates
human

-defensin 2 expression in epithelial cells. FEBS Lett. 561: 117–121.
29. Jiang, D., J. Liang, and P. W. Noble. 2007. Hyaluronan in tissue and repair. Annu.
Rev. Cell. Dev. Biol. 23: 435– 461.
30. Schmid, P., O. Grenet, J. Medina, S. D. Chibout, C. Osborne, and D. A. Cox.
2001. An intrinsic antibiotic mechanism in wounds and tissue engineered skin.
J. Invest. Dermatol. 116: 471– 472.
31. Butmarc, J., T. Yufit, P. Carson, and V. Falanga. 2004. Human

-defensin-2
expression is increased in chronic wounds. Wound Repair Regen. 12: 439– 443.
32. Sørensen, O. E., J. B. Cowland, K. Theilgaard-Mo¨ nch, L. Liu, T. Ganz, and
N. Borregaard. 2003. Wound healing and expression of antimicrobial peptides/
polypeptides in human keratinocytes, a consequence of common growth factors.
J. Immunol. 170: 5583–5589.
33. Sørensen, O. E., D. R. Thapa, K. M. Roupe´, E. V. Valore, U. Sjo¨bring,
A. A. Roberts, A. Schmidtchen, and T. Ganz. 2006. Injury-induced innate im-
mune response in human skin mediated by transactivation of the epidermal
growth factor receptor. J. Clin. Invest. 116: 1878 –1885.
34. Namjoshi, S., R. Caccetta, and H. A. Benson. 2007. Skin peptides: biological
activity and therapeutic opportunities. J. Pharm. Sci. 97: 2524 –2542.
35. Goodarzi, H., J. Trowbridge, and R. L. Gallo. 2007 Innate immunity: a cutaneous
perspective. Clin. Rev. Allergy Immunol. 33: 15–26.
36. Matzinger, P. 2004. The danger model: a renewed sense of self. Science 296:
301–305.
37. Hajishengallis, G., H. Sojar, R. J. Genco, and E. DeNardin. 2004. Intracellular
signaling and cytokine induction upon interactions of Porphyromonas gingivalis
fimbriae with pattern-recognition receptors. Immunol. Invest. 33: 157–172.
38. Uehara, A., Y. Fujimoto, K. Fukase, and H. Takada. 2007. Various human epi-
thelial cells express functional Toll-like receptors, NOD1 and NOD2 to produce
anti-microbial peptides, but not proinflammatory cytokines. Mol. Immunol. 44:
3100 –3111.
39. Krisanaprakornkit, S., J. R. Kimball, A. Weinberg, R. P. Darveau,
B. W. Bainbridge, and B. A. Dale. 2000. Inducible expression of human

-de-
fensin 2 by Fusobacterium nucleatum in oral epithelial cells: multiple signaling
pathways and role of commensal bacteria in innate immunity and the epithelial
barrier. Infect. Immun. 68: 2907–2915.
40. Harder, J., U. Meyer-Hoffert, L. M. Teran, L. Schwichtenberg, J. Bartels,
S. Maune, and J. M. Schroder. 2000. Mucoid Pseudomonas aeruginosa, TNF-
␣
,
and IL-1

, but not IL-6, induce human

-defensin-2 in respiratory epithelia.
Am. J. Respir. Cell Mol. Biol. 22: 714 –721.
41. Liu, L., L. Wang, H. P. Jia, C. Zhao, H. H. Heng, B. C. Schutte, P. B. McCray,
Jr., and Ganz, T. 1998. Structure and mapping of the human

-defensin HBD-2
gene and its expression at sites of inflammation. Gene 222: 237–244.
2110 IMMUNOMODULATORY ACTIVITY OF HYALURONAN FRAGMENTS IN SKIN
