Culture medium from TNF-a–stimulated mesenchymal stem

cells attenuates allergic conjunctivitis through multiple

antiallergic mechanisms
Wenru Su, MD, PhD,a,b* Qian Wan, MD,a* Jingwen Huang, MD,a* Longhui Han, MD,a Xiaoqing Chen, MD,a,c
Guihua Chen, MD, PhD,b Nancy Olsen, MD,c Song Guo Zheng, MD, PhD,b,c and Dan Liang, MD, PhDa Guangzhou, China,
and Hershey, Pa

Background: The immunomodulatory and anti-inflammatory Key words: Allergy, mesenchymal stem cells, stems cells, B cells,
functions of mesenchymal stem cells (MSCs) have been mast cells, histamine, allergic conjunctivitis
demonstrated in several autoimmune/inflammatory diseases,
but their contribution to allergic conjunctivitis and underlying
antiallergic mechanisms remain elusive. Allergies are increasingly prevalent and have become a major
Objective: We sought to explore the clinical application of world health problem. Among allergic diseases, allergic conjunc-
MSCs to experimental allergic conjunctivitis (EAC) and its tivitis (AC), which includes seasonal and perennial AC, vernal
underlying antiallergic mechanisms. keratoconjunctivitis, and atopic keratoconjunctivitis, is the most
Methods: Culture medium from TNF-a–stimulated, bone prevalent form of mucosal allergy and currently affects more than
marrow–derived MSCs (MSC-CMT) was administered 20% of the US population.1,2 AC is characterized by classical
topically to mice with EAC, and the related allergic symptoms symptoms that include itching, tearing, eyelid swelling, and che-
and biological changes were evaluated. Murine spleen-derived B mosis and redness of the eyes, and AC symptoms significantly
cells, bone marrow–derived mast cells (MCs), and lung vascular affect a patient’s health and quality of life. Topical antihistamines,
endothelial cells were cultured in vitro to investigate the mast cell (MC) stabilizers, or dual-acting antihistamine/MC sta-
antiallergic MSC-CMT mechanisms. bilizers remain the mainstay of AC therapy. Although corticoste-
Results: Topical instillation of MSC-CMT significantly roids are potent antiallergic agents in reducing ocular allergic
attenuated the clinical symptoms of short ragweed pollen– symptoms, especially in patients with severe and chronic AC,
induced EAC, with a significant decrease in inflammatory cell they are associated with various ocular adverse effects, such as
frequency, nuclear factor kB p65 expression, and TNF-a and secondary infection, increased intraocular pressure, and cataract
IL-4 production. In vitro MSC-CMT significantly inhibited the formation.3,4 Additionally, with prolonged use, topical antihista-
activation of MCs and B-cell IgE release and reduced histamine- mines might be irritating to the eye, and many antihistamines are
induced vascular hyperpermeability. During EAC, MSC-CMT known to have anticholinergic effects that can cause ocular drying
treatment also decreased IgE production, histamine release, through muscarinic receptor inhibition.3-6 Therefore investi-
enrichment and activation of MCs, and conjunctival vascular gating safer and more effective treatments is necessary.
hyperpermeability. The MSC-CMT–mediated inhibition of B Mesenchymal stem cells (MSCs; also referred to as mesen-
cells, MCs, and histamine and its antiallergic effects during chymal stromal cells) possess biological functions ranging from
EAC were abrogated when MSCs were pretreated with COX2 tissue repair/regeneration to immunomodulatory actions.7-9
small interfering RNA. Furthermore, different culture conditions or additional treatment
Conclusions: Our findings provide compelling evidence that might regulate or enhance the biological functions of MSCs.
MSC-CMT inhibits EAC through COX2-dependent multiple Because of their multifunctional properties, numerous animal
antiallergic mechanisms and support the use of MSC-CMT as a and clinical trials are examining the use of MSCs to treat a
novel strategy for treating allergic conjunctivitis. (J Allergy Clin wide range of diseases. The antiallergic properties of MSCs
Immunol 2015;136:423-32.) have recently received increasing attention, and several studies
have reported that MSCs suppress allergic responses in experi-
From athe State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun mental mouse models of asthma,10,11 allergic rhinitis,12 and con-
Yat-sen University, Guangzhou; bthe Center for Clinic Immunology, Sun Yat-sen Uni- tact hypersensitivity.13 However, the mechanisms by which
versity Third Affiliated Hospital, Guangzhou; and cthe Division of Rheumatology, MSCs mediate antiallergic responses remain largely elusive.
Department of Medicine, Penn State University Hershey College of Medicine, Hershey. The roles of MSCs in patients with AC have not yet been
*These authors contributed equally to this work.
Supported by the Natural Science Foundation of China (81271051) and the Natural Sci-
explored. In previous studies MSCs have been administered for
ence Foundation of China (81300740). the treatment of ocular surface disorders through various routes,
Disclosure of potential conflict of interest: The authors declare that they have no relevant including intravenous injection, application with a special hollow
conflicts of interest. plastic tube, transplantation with an amniotic membrane, and
Received for publication June 13, 2014; revised December 18, 2014; accepted for pub-
lication December 29, 2014.
subconjunctival injection.14-18 However, in a clinical setting these
Available online February 1, 2015. administration routes are not acceptable for treating patients with
Corresponding author: Dan Liang, MD, PhD, 54 Xianlie South Rd, Guangzhou 510060, AC; therefore the development of a clinically feasible MSC
China. E-mail: liangd631004@[Link]. Or: Song Guo Zheng, MD, PhD, 500 Univer- administration method for the treatment of AC is desirable.
sity Drive, Hershey, PA 17033. E-mail: szheng1@[Link]. Although the exact mechanism of MSC immunomodulatory
0091-6749/$36.00
Ó 2015 American Academy of Allergy, Asthma & Immunology
action remains largely unknown, a large number of studies have
[Link] demonstrated that a variety of soluble factors, such as IL-10,

423
424 SU ET AL J ALLERGY CLIN IMMUNOL
AUGUST 2015

and tibias with a–Minimum Essential Medium (a-MEM; Invitrogen, Grand

Abbreviations used Island, NY) containing 2% heat-inactivated FBS. Single-suspension murine
AC: Allergic conjunctivitis bone marrow–derived nucleated cells were seeded at a density of 15 3 106
ACED: Autologous serum eye drop cells in 100-mm culture dishes (Corning Costar, Corning, NY) and maintained
BMMSC: Bone marrow–derived mesenchymal stem cell at 378C in a 5% CO2 atmosphere. Nonadherent cells were removed after 48
EAC: Experimental allergic conjunctivitis hours, and adherent cells were maintained for 16 days in a-MEM supple-
HRP: Horseradish peroxidase mented with 20% FBS, 2 mmol/L L-glutamine, 55 mmol/L 2-
IDO: Indoleamine-2,3-dioxygenase mercaptoethanol, 100 U/mL penicillin, and 100 mg/mL streptomycin.
L-NAME: N-nitro-L-arginine methyl ester Colonies formed by adherent cells were passaged once for additional experi-
LVEC: Lung vascular endothelial cell mental use. Flow cytometric analyses demonstrated that bone marrow–
MAPK: Mitogen-activated protein kinase derived mesenchymal stem cells (BMMSCs) expressed high levels of
MC: Mast cell CD29, CD44, and CD90 but did not express the hematopoietic markers
MSC: Mesenchymal stem cell CD34, CD45, or CD11b (eBioscience, San Diego, Calif; data not shown).
MSC-CM: Culture medium from bone marrow–derived mesen- BMMSCs were cultured in 100-mm dishes until subconfluence to obtain
chymal stem cells culture medium from bone marrow–derived mesenchymal stem cells (MSC-
MSC-CMT: Culture medium from TNF-a–stimulated, bone marrow– CM) and TNF-a–stimulated BMMSCs (MSC-CMT). The cells were then
derived mesenchymal stem cells washed twice with HBSS to remove serum and incubated in 10 mL of a-MEM
1-MT: 1-Methyl-tryptophan in the presence or absence of carrier-free recombinant mouse TNF-a (10 ng/
NF-kB: Nuclear factor kB mL) for 48 hours. The supernatants were collected and centrifuged to remove
NO: Nitric oxide cell debris. TNF-a was depleted from MSC-CMT by using anti–TNF-a
OLF: Ophthalmic lavage fluid antibody immobilized on protein G–agarose beads, according to the
PGE2: Prostaglandin E2 manufacturer’s instructions. In brief, 0.2 mg of anti–TNF-a antibody in PBS
PMACI: Phorbol 12-myristate 13-acetate plus calcium ionophore (30 mL) was mixed with a protein G–agarose bead suspension (50% slurry) for
siRNA: Small interfering RNA 1 hour at 48C with intermittent shaking. After centrifugation, beads bound
SRW: Short ragweed with the anti–TNF-a antibody were washed 3 times. MSC-CMT was
STAT: Signal transducer and activator of transcription incubated with anti–TNF-a antibody immobilized on protein G–agarose
beads for 1 hour at 48C. The protein G–agarose beads were removed by using
centrifugation. TNF-a, IL-10, IL-4, IL-13, and IFN-g levels were measured
(see Fig E1 in this article’s Online Repository at [Link]). The
TGF-b, prostaglandin E2 (PGE2), nitric oxide (NO), TNF- resultant supernatants were collected and immediately used for experiments.
stimulated protein 6, and indoleamine-2,3-dioxygenase (IDO), For further information on isolation, culture, and activation of B cells;
are involved in the MSC-mediated immunosuppression of various isolation, culture, and activation of bone marrow–derived MCs; isolation and
immune responses.7-9 These studies suggest that MSC culture culture of lung vascular endothelial cells (LVECs); in vitro and in vivo vascular
permeability assays; transfection; and EAC treatment with MSC-CMT, see the
medium, which contains these soluble secreted factors, might
Methods section in this article’s Online Repository at [Link].
be an effective AC treatment. However, recent data suggest that The murine EAC model was generated, as previously reported.2,20,21 In
resting MSCs have low immunomodulatory activity, whereas brief, a mixture of 50 mg of short ragweed (SRW) pollen (Greer Laboratories,
MSCs stimulated by inflammatory cytokines, such as TNF-a, Lenoir, NC) in 5 mg of Imject Alum (Pierce, Rockford, Ill) was applied
secrete large quantities of soluble mediators and develop their through footpad injection on the first day. The sensitization procedure was
full immunosuppressive potential.7-9 Therefore MSC conditional repeated on day 5 to enhance the allergic reaction. The mice were challenged
medium stimulated by inflammatory cytokines might be an effec- with 1.5 mg of SRW pollen suspended in 10 mL of PBS in both eyes on days 10
tive treatment for AC. to 14. MSC-CMT (10 mL) was topically applied once 30 minutes before the
In this study we investigated the therapeutic effects of culture SRW pollen challenge and 4 times per day on days 10 to 14 in both eyes.
medium from TNF-a–stimulated, bone marrow–derived mesen- PBS was applied to the control groups in the same manner (n 5 6 per group).
For further information on clinical evaluation and immunohistochemistry,
chymal stem cells (MSC-CMT) on experimental allergic
Western blot analyses, ELISAs, and real-time PCR, see the Methods section in
conjunctivitis (EAC) and the underlying cellular and molecular this article’s Online Repository.
mechanisms of this potential therapy.

Statistical analyses
METHODS The Student t test was used to analyze significant differences (SPSS 16.0;
Animals SPSS, Chicago, Ill). A P value of less than .05 was considered significant.
BALB/c mice were supplied by and maintained in the Guangzhou Animal
Testing Center and used between 4 and 6 weeks of age. The studies were
approved by the Institutional Animal Care and Use Committee of the
Zhongshan Ophthalmic Center at Sun Yat-sen University. All procedures RESULTS
involving animal eye studies were conducted in accordance with the Asso- MSC-CMT treatment suppresses EAC
ciation for Research in Vision and Ophthalmology ‘‘Statement for the use of SRW pollen–induced EAC is a well-established mouse model
animals in ophthalmic and vision research.’’ The animals were maintained in
of human AC. MSC-CMT (subjected to TNF-a immunodeple-
specific pathogen-free conditions on a 12-hour light-dark cycle with
controlled temperature (238C 6 28C) and humidity (55% 6 10%). tion) or MSC-CM was applied to the ocular surfaces of mice with
EAC on days 10 to 14 after immunization (day 0) to examine the
therapeutic effect of different MSC culture media on AC (Fig 1,
Isolation and culture of bone marrow–derived A). The effect of MSC-CMT on ocular symptoms in mice was
mesenchymal stem cells and MSC-CMT preparation evaluated after topical SRW pollen challenge. Throughout the
BMMSCs were isolated from BALB/c mice, as previously described.19 study, ocular administration of MSC-CMT, but not MSC-CM,
Briefly, bone marrow cells were flushed from the bone cavities of femurs significantly alleviated all of the symptoms in the mice with
J ALLERGY CLIN IMMUNOL SU ET AL 425
VOLUME 136, NUMBER 2

FIG 1. MSC-CMT treatment attenuates EAC. A, Experimental protocol. B and C, Inflammatory scores of EAC
(Fig 1, B) and scratch times (Fig 1, C) were evaluated at the indicated time points after challenge in different
experimental groups. D, Representative images of ocular symptoms in the indicated experimental group 30
minutes after the last challenge. Red arrow, Lid swelling; white arrow, conjunctival edema. n 5 6 mice for
each group. *P < .05 and **P < .01 between the MSC-CMT and EAC groups. Error bars 5 means 6 SEMs.

EAC compared with the control group (P < .01; Fig 1, B). SRW (MAPK), caspase-1, and signal transducer and activator of
pollen application also induced a significant increase in the transcription (STAT) 3 and STAT6 expression levels were
scratching response compared with normal mice, and MSC- also measured by means of Western blotting. MSC-CMT treat-
CMT administration significantly inhibited the scratch response ment reduced phosphorylation of NF-kB p65, p38 MAPK, and
in mice compared with the control group (P < .01; Fig 1, C). STAT6 and caspase-1 expression but increased STAT3 phos-
Representative images of ocular symptoms demonstrate that phorylation in the conjunctiva of mice with EAC (Fig 2, H-K,
MSC-CMT treatment significantly attenuated eyelid swelling, and see Fig E2, D).
conjunctival edema, and redness (Fig 1, D).

MSC-CMT inhibits B-cell IgE release through a

MSC-CMT treatment attenuates inflammation in COX2-dependent mechanism
the mice with EAC B cells play important roles in the pathophysiology of allergic
Next, we investigated the effect of MSC-CMT treatment on diseases, primarily through their ability to produce IgE anti-
inflammatory profiles in mice with EAC. Histologic analyses of bodies.23 Thus we investigated whether the antiallergic effects
the conjunctival tissue harvested 14 days after immunization of MSC-CMT might involve inhibition of B-cell function. First,
revealed a significantly decreased infiltration of inflammatory the in vivo effects of MSC-CMT on IgE production in mice with
cells and eosinophil accumulation in MSC-CMT–treated mice EAC were determined. As shown in Fig 3, A, SRW application
compared with that seen in control animals (Fig 2, A-C). Next, we significantly increased IgE levels of ophthalmic lavage fluid
determined the effects of MSC-CMT on local inflammatory cyto- (OLF) in mice with EAC compared with those in normal
kine levels in the conjunctival tissue using real-time PCR and mice, whereas MSC-CMT treatment significantly reduced IgE
ELISA. The results demonstrated a significant increase in TNF- levels.
a, IL-4, IL-5, IL-1b, and TGF-b expression in the conjunctiva We next performed in vitro studies to confirm the inhibitory
of mice with EAC. MSC-CMT suppressed the upregulation of effect of MSC-CMT on B cells. Purified mouse splenic B cells
TNF-a, IL-4, IL-5, and IL-1b but increased TGF-b expression were cultured with IL-4 and LPS in different medium for 5
(Fig 2, D-G, and see Fig E2, A-C, in this article’s Online Repos- days. IgE and IgG1 levels in culture supernatants were measured
itory at [Link]). by means of ELISA. A significant inhibition of IgE and IgG1
Nuclear factor kB (NF-kB) is critical for ocular inflamma- release by activated B cells was observed in the presence
tion, including inflammation associated with microbial in- of MSC-CMT but not MSC-CM (Fig 3, B and C). Notably,
fections, allergic eye diseases, and dry eye.22 Thus we MSC-CMT did not affect B-cell viability (Fig 3, D).
examined whether MSC-CMT modulates NF-kB signaling We next examined the mechanisms involved in MSC-CMT–
in EAC. Additionally, p38 mitogen-activated protein kinase mediated inhibition of IgE production by B cells. In these
426 SU ET AL J ALLERGY CLIN IMMUNOL
AUGUST 2015

FIG 2. Treatment with MSC-CMT attenuates inflammation and alters the cytokine profile in mice with EAC.
A, Representative images of hematoxylin and eosin staining of conjunctival samples from mice in the indi-
cated experimental groups. Scale bars represent 200 mm (upper panels) and 600 mm (lower panels). B and C,
Quantification of cellular components (Fig 2, B) and eosinophils (Fig 2, C) in the conjunctivas of mice in the
indicated experimental groups. D-G, TNF-a, IL-4, IL-1b, and TGF-b mRNA expression in conjunctival tissue
was determined by using real-time PCR. H-K, Phosphorylated NF-kB p65, phosphorylated p38 MAPK, phos-
phorylated STAT6, and phosphorylated STAT3 levels in conjunctival tissue was determined by means of
Western blotting. *P < .05 and **P < .01 between the MSC-CMT and EAC groups. Error bars 5 means 6
SEMs.

experiments neutralizing mAbs specific for TGF-b1 or IL-10 and COX2/PGE2 in MSC-CMT–mediated B-cell suppression by
specific inhibitors of IDO (1-methyl-tryptophan [1-MT]), NO (N- BMMSCs, we used small interfering RNA (siRNA) to knock
nitro-L-arginine methyl ester [L-NAME]), or COX2 (NS-398) down COX2 expression in BMMSCs, and the inhibitory capacity
were used. TGF-b1 and IL-10 neutralizing antibodies and of MSC-CMT on B-cell IgE production was significantly
BMMSC pretreatment with 1-MT or L-NAME did not affect decreased in BMMSCs treated with COX2 siRNA but not control
MSC-CMT–mediated inhibitory effects. In contrast, BMMSC siRNA (Fig 3, H, and see Fig E3, B).
pretreatment with NS-398 significantly but incompletely
reversed the inhibitory effect on IgE production (Fig 3, E, and
see Fig E3, A, in this article’s Online Repository at www. EAC attenuation by MSC-CMT is associated with
[Link]). These results suggest that COX2/PGE2 signaling MC inhibition
in BMMSCs plays an important role in MSC-CMT–mediated We next examined whether the antiallergic effects of MSC-
B-cell inhibition. CMT involve inhibition of MC function. We first explored the
To further investigate the role of COX2/PGE2 in MSC-CMT– in vivo effects of MSC-CMTs on MC functions in challenged con-
mediated B-cell inhibition, we first exposed BMMSCs to different junctiva. Toluidine blue staining revealed that the numbers of MCs
concentrations of exogenous TNF-a for 24 hours and then and percentages of degranulated MCs decreased dramatically in
analyzed COX2 expression and PGE2 production using Western the conjunctiva of MSC-CMT–treated mice compared with those
blot analysis and ELISA, respectively. TNF-a treatment led to seen in control mice (Fig 4, A-C). Additionally, MSC-CMT treat-
dose-dependent increases in COX2 expression and PGE2 produc- ment significantly decreased histamine production in mice with
tion by BMMSCs (Fig 3, F and G). To further confirm the role of EAC compared with that seen in control mice (Fig 4, D).
J ALLERGY CLIN IMMUNOL SU ET AL 427
VOLUME 136, NUMBER 2

FIG 3. MSC-CMT inhibits IgE release by B cells through a COX2-dependent mechanism. A, OLF was
collected after the last challenge, and IgE levels were determined by using ELISA. BALB/c splenic B cells
were cultured with different media for 24 hours. B and C, After stimulation with LPS/IL-4 for 5 days, IgE
(Fig 3, B) and IgG1 (Fig 3, C) levels in supernatants were determined by using ELISA. D, Cell viability was
assayed by using trypan blue exclusion. E, MSC-CMT was added with specific neutralizing antibodies for
TGF-b1 or IL-10 (10 mg/mL, isotype used as controls), or MSC-CMT was collected after pretreating BMMSCs
with a specific IDO inhibitor (1-MT, 500 mmol/L), a specific NO inhibitor (L-NAME, 1 mmol/L), or a specific
COX2 inhibitor (NS-398, 1 mmol/L). F and G, Exogenous TNF-a–induced dose-dependent increases in
COX2 expression and PGE2 production in BMMSCs. H, MSC-CMT was collected after pretreating BMMSCs
with COX2 siRNA (dCMT) or control siRNA (cCMT). **P < .01. Error bars 5 means 6 SEMs.

We next performed in vitro studies to confirm the inhibitory ef- Antihistamine effects of MSC-CMT
fect of MSC-CMT on MCs. MCs were cultured with different me- We next determined whether MSC-CMT has an inhibitory
dia for 24 hours and then stimulated with phorbol 12-myristate effect on histamine function. We initially tested whether
13-acetate plus calcium ionophore (PMACI) for an additional MSC-CMT effectively reduced vascular hyperpermeability in
12 hours. Our results demonstrated that MSC-CMT, but not mice with EAC. The Evans blue permeability assay demon-
MSC-CM, significantly inhibited PMACI-stimulated TNF-a strated that MSC-CMT treatment significantly reduced
and IL-4 release by MCs (Fig 4, E, and see Fig E4, A, in this ar- conjunctival vascular hyperpermeability in mice with EAC
ticle’s Online Repository at [Link]). MSC-CMT did compared with that seen in untreated control mice (Fig 5, A
not affect MC viability (see Fig E4, B). and B). We next performed a series of in vitro studies to
Next, we investigated the mechanisms involved in MSC-CMT– confirm the inhibitory effect of MSC-CMT on histamine func-
mediated MC inhibition. The MCs were cultured with MSC-CMT tion. LVECs purified from mouse liver and lung tissue were
for 24 hours and then stimulated with PMACI for another 6 hours. cultured with or without MSC-CMT in a Transwell system
NF-kB activation was subsequently determined by using Western (BD Biosciences, San Jose, Calif) for 24 hours. Cells from
blot analysis. Although PMACI stimulation resulted in increased each culture condition were stimulated with histamine for
NF-kB p65 protein production, MSC-CMT significantly inhibited 30 minutes and subsequently evaluated in the permeability
NF-kB p65 expression in MCs (Fig 4, F). Most agents, including assay. LVECs cultured with MSC-CMT exhibited significantly
PMACI, activate NF-kB through IkB-a phosphorylation, degra- decreased vascular hyperpermeability to streptavidin–horse-
dation, or both. IkB-a degradation exposes a nuclear localization radish peroxidase (HRP) than LVECs stimulated with hista-
signal and leads to NF-kB activation.24 Thus we investigated mine alone (Fig 5, C).
whether MSC-CMT modulates NF-kB activity in MCs by inhib- We next examined the mechanisms regulating these MSC-
iting IkB-a degradation. As expected, MSC-CMT significantly CMT–mediated inhibitory effects. VE-cadherin–VE-cadherin
inhibited IkB-a degradation in MCs (Fig 4, G). Because COX2/ homophilic interactions and VE-cadherin–b-catenin binding are
PGE2 signaling was essential for the MSC-CMT–mediated inhi- crucial to maintaining normal vascular integrity and permeability.
bition of B-cell function, we examined whether COX2 plays a Histamine increases vascular permeability by inducing the
similar role in MSC-CMT–mediated MC inhibition. As shown tyrosine phosphorylation of VE-cadherin and disrupting VE-
in Fig 4, H, and Fig E4, C, MSC-CMT–mediated MC inhibition cadherin–VE-cadherin homophilic interactions.25,26 Thus we
was partially reversed when BMMSCs were pretreated with analyzed total VE-cadherin and phosphorylated VE-cadherin
COX2 siRNA. expression in LVECs under different conditions. LVECs
428 SU ET AL J ALLERGY CLIN IMMUNOL
AUGUST 2015

FIG 4. Attenuation of EAC by MSC-CMT is associated with its inhibitory effects on MCs. A-C, MCs were iden-
tified by using toluidine blue staining, and activated MCs were identified based on their irregular shape (red
arrows) in conjunctivas of normal mice, mice with EAC, and MSC-CMT–treated mice with EAC, as indicated.
Scale bars represent 200 mm (upper panels) and 600 mm (lower panels). D, Histamine levels in OLF were
determined by using EIA (n 5 6). E and H, MCs were cultured with MSC-CMT or MSC-CM for 24 hours
and then stimulated with PMACI for an additional 12 hours. TNF-a levels in supernatants were determined
by means of ELISA. F and G, IkB-a and NF-kB p65 expression levels in BMMSCs were determined by means
of Western blotting. Fig 4, H, MSC-CMT was collected after pretreating BMMSCs with COX2 siRNA (dCMT)
or control siRNA (cCMT). **P < .01. Error bars 5 means 6 SEMs.

cultured with MSC-CMT exhibited significantly increased total COX2 signaling in BMMSCs is essential for
VE-cadherin expression but reduced the phosphorylation of MSC-CMT–mediated EAC inhibition
VE-cadherin compared with that seen in control mice (Fig 5, D Our in vitro studies demonstrated that COX2/PGE2 signaling
and E). COX2 is essential for MSC-CMT–mediated inhibition plays an essential role in the MSC-CMT–mediated inhibition of
of B-cell function; therefore we investigated whether COX2 plays B cells, MCs, and histamine, and therefore we investigated
a similar role in the MSC-CMT–mediated inhibition of histamine. whether COX2 is also implicated in MSC-CMT–mediated EAC
MSC-CMT–mediated vascular hyperpermeability inhibition was attenuation. COX2 siRNAwas used to knock down COX2 expres-
partially reversed when BMMSCs were pretreated with COX2 sion in BMMSCs, followed by stimulation of the cells with TNF-
siRNA (Fig 5, F). a. MSC-CMT derived from BMMSCs treated with COX2 siRNA
J ALLERGY CLIN IMMUNOL SU ET AL 429
VOLUME 136, NUMBER 2

FIG 5. Antihistamine effects of MSC-CMT. A, Conjunctival vascular permeability was evaluated by using the
Evans blue assay (n 5 6). B, Representative images of eyes after Evans blue injection. C and F, LVECs were
cultured with MSC-CMT for 24 hours. After stimulation with histamine for 30 minutes, the vascular perme-
ability assay was performed. D and E, Total VE-cadherin (Fig 5, D) and phosphorylated VE-cadherin (Fig 5, E)
expression was determined by using Western blotting. Fig 5, F, MSC-CMT was collected after pretreating
BMMSCs with COX2 siRNA (dCMT) or control siRNA (cCMT). **P < .01. Error bars 5 means 6 SEMs.

did not suppress EAC in mice (Fig 6, A and B). COX2 knockdown initiate B-cell production of allergen-specific IgE, which binds
in BMMSCs also significantly decreased the inhibitory effects of to the high-affinity IgE receptor on MCs. Allergen cross-linking
MSC-CMT on inflammatory cell infiltration, IgE production, MC with allergen-specific MC IgE leads to the release of preformed
enrichment and activation, histamine production, and TNF-a and granule-stored allergic mediators, such as histamine. These medi-
IL-4 production in mice with EAC (Fig 6, E-G, and see Fig E5 in ators cause early acute inflammatory responses and activation of
this article’s Online Repository at [Link]). Addition- the de novo synthesis of mediators, such as inflammatory cyto-
ally, topical administration of 16,16-dimethyl PGE2 (20-80 ng/ kines, which sustain the late inflammatory response phase.23
mL in PBS) led to a dose-dependent suppression of EAC appear- Therefore TH2 cells, B cells, MCs, and histamine play crucial
ance, but the treatment response was not as significant as the ef- roles in the development of allergic reactions. Several recent
fect mediated by MSC-CMT treatment (Fig 6, H and I). As studies have reported that the intravenous administration of
shown in Fig E6 (in this article’s Online Repository at www. MSCs during the sensitization phase produces prophylactic anti-
[Link]), our results showed that MSC-CMT reduced the allergic effects in patients with allergic rhinitis and asthma.10-12
production of TNF-a and IL-1b by lung epithelial cells stimu- The inhibition of antigen-specific TH2 cell differentiation, which
lated with lipopolysaccharide. shifts the TH2/TH1 balance, has been proposed as the mechanism
underlying these effects.
In contrast to studies demonstrating MSC-mediated inhibition
DISCUSSION of antigen-specific TH2 cell differentiation during the sensitiza-
MSCs have a striking variety of biological functions; therefore tion phase, the administration of topical MSC-CMT during the
therapeutic applications of MSCs for a wide range of diseases effector phase in the present study yielded promising therapeutic
have been studied extensively. AC, a common clinical problem effects in SRW pollen–induced EAC. Our results demonstrated
for both ophthalmologists and allergists, is caused by an allergen- that the topical administration of MSC-CMT during the effector
induced inflammatory response.1,2 Here we demonstrated that phase of EAC reduced IgE production, histamine release, enrich-
MSC-CMT treatment significantly attenuated the clinical symp- ment and activation of MCs, and conjunctival vascular hyperper-
toms of EAC with a significantly decreased inflammatory cell meability. These results were further supported by the inhibition
frequency, decreased TNF-a and IL-4 production, and reduced of MC activation and B-cell IgE release and reduced histamine-
NF-kB expression. Furthermore, we determined that MSC- induced vascular hyperpermeability mediated by MSC-CMT
CMT inhibits B-cell, MC, and histamine function. These in vitro. These findings collectively suggest that the inhibition
inhibitory actions might contribute to the therapeutic effects of of B-cell, MC, and histamine function contributes to the antialler-
MSC-CMT on EAC. To our knowledge, these findings provide gic functions of MSC-CMT.
the first evidence that MSC manipulation is a potential novel strat- In vitro MC degranulation is always induced in a specific buffer
egy for the treatment of AC. but not in the culture medium.27,28 Therefore we were not able to
In response to allergen stimulation during type I allergic evaluate the in vitro role of MSC-CMT in MC degranulation in
reactions, allergen-specific TH2 cells produce cytokines that this study. Instead, we evaluated the roles of MSC-CMT in
430 SU ET AL J ALLERGY CLIN IMMUNOL
AUGUST 2015

FIG 6. COX2 signaling in BMMSCs is essential for MSC-CMT–mediated inhibition of EAC. EAC was induced
as above, and MSC-CMTs collected after pretreating BMMSCs with COX2 siRNA (dCMT) or control siRNA
(cCMT) were administered to mice with EAC. A and B, Inflammatory scores (Fig 6, A) and scratch times (Fig
6, B) were evaluated at the indicated time points after challenge (n 5 6). C-G, IgE production (Fig 5, C), MC
enrichment and activation (Fig 5, D and E), histamine production (Fig 5, F), and TNF-a expression (Fig 5, G) in
EAC were measured as above at the indicated time points after challenge. Topical 16,16-dimethyl PGE2
(dmPGE2; 20-80 ng/mL) was administered on days 10 to 14 in mice with EAC. H and I, Inflammatory scores
(Fig 5, H) and scratch times (Fig 5, I) were evaluated at indicated time points after challenge (n 5 6). **P < .01
between the MSC-CMT and dMSC-CMT groups. ##P < .01 between the indicated groups. DDP < .01 between
the MSC-CMT and PGE2 80 ng/mL groups. Error bars 5 means 6 SEMs.

PMACI-stimulated MC activation. MSC-CMT significantly in- mediated EAC immunosuppression pathways. We demon-
hibited MC inflammatory mediator production by modulating strated that the addition of specific neutralizing antibodies
the NF-kB signaling pathway in a COX2-dependent manner. for IL-10 and TGF-b1 and a specific inhibitor for IDO or
MSC-CMT treatment in vivo significantly reduced histamine pro- NO did not affect MSC-CMT–mediated inhibition of IgE
duction and the enrichment and activation of MCs. Together with release by B cells, but pretreatment of MSCs with NS398, a
the aforementioned BMMSC-mediated inhibition of MC degran- specific inhibitor of COX2/PGE2, significantly reversed these
ulation through a COX2-dependent mechanism,29 these data indi- inhibitory effects on B-cell IgE release. Furthermore, COX2
cate that MSC-CMT inhibits MC function in vitro and in vivo. knockdown in BMMSCs also significantly decreased the inhib-
Although the exact mechanism of MSC-mediated immuno- itory effects of MSC-CMT on B-cell IgE release, PMACI-
modulatory functions remains elusive, many studies have stimulated activation of MCs, and histamine-induced vascular
demonstrated that COX2/PGE2 signaling plays important roles hyperpermeability. Additionally, MSCs pretreated with COX2
in MSC-mediated immunosuppression in various immune cells, siRNA before injection into mice lost their suppressive effects
including T lymphocytes, natural killer cells, dendritic cells, on EAC, including the effects on IgE and histamine production
MCs, and macrophages.29-33 In the present study we focused and MC infiltration and activation. These results support previ-
on B cells, MCs, and histamine as major players in MSC- ous findings demonstrating a dominant role of COX2 signaling
J ALLERGY CLIN IMMUNOL SU ET AL 431
VOLUME 136, NUMBER 2

in the MSC-CMT–mediated attenuation of EAC and the under- Recently, autologous serum eye drops (ACEDs) have been
lying immunosuppressive effects on B cells, MCs, and shown to possess potential benefits for ocular surface disorders.
histamine. Several studies reported that the therapeutic benefits of ACEDs in
PGE2 exerts its biological functions through 4 subtypes of patients with persistent corneal epithelial defects and dry eyes are
prostaglandin E receptors, EP1-EP4. The therapeutic effects of mediated by their trophic properties that promote wound heal-
PGE2 or its analogs on allergic inflammation have been ing.44,45 However, to our knowledge, no articles have reported
previously described34-36; however, conflicting results with that ACEDs possess immunomodulatory effects. Thus ACEDs
respect to its efficacy have been obtained, depending on the are different from MSC-CMT, which possesses not only trophic
receptor subtype.37-42 For example, PGE2 suppresses allergic properties but also immunomodulatory properties. In addition,
inflammation mediated through EP3 or EP237-39 while triggering the use of amniotic membranes has also been shown to have ther-
inflammatory responses through EP4 or EP1.40-42 These discrep- apeutic benefits for severe ocular surface disorders, such as ocular
ancies might limit the clinical application of PGE2 and its analogs chemical burns, corneal ulcers, persistent corneal epithelial de-
in the treatment of allergic diseases. In the present study we fects, and corneal limbal stem cell deficiency,46 as well as anti-
demonstrated that 16,16-dimethyl PGE2 topical administration inflammatory and trophic properties on the ocular surface. Of
partially ameliorated EAC in a dose-dependent manner; however, note, the combination of amniotic membrane treatment and
the suppressive effect was not as significant as the effect observed MSCs has shown better therapeutic effects than amniotic mem-
in animals treated with topical MSC-CMT. This finding might be brane treatment alone in patients with ocular chemical burns,
attributed to the actions of multiple factors, rather than PGE2 skin defects, and articular cartilage defects.47,48 Thus for some se-
alone, that promote MSC-CMT–mediated inhibition of EAC. vere ocular disorders, the combination of amniotic membrane
Furthermore, Heo et al43 recently reported that topical MSC- treatment and MSCs or MSC-CMT might be a promising therapy.
CMT enhances cutaneous wound healing by soluble factors. In summary, for the first time, our study demonstrates that MSC
These findings indicate that MSC-CMT, which possesses both manipulation can be used to treat AC. Our results show that MSC-
immunosuppressive and trophic activities, suppresses the inflam- CMT inhibits IgE release from B cells and MC activation through
matory phase of EAC and contributes to the repair of injured COX2-dependent mechanisms. Furthermore, we provide the first
tissues, which might provide additional lasting treatment benefits evidence showing that MSCs possess antihistamine properties.
without multiple dosing by using a single pharmacologic drug. These findings provide compelling evidence that MSC manipu-
The current mainstay of AC therapy includes topical MC lation inhibits allergic reactions and supports the use of MSCs as
stabilizers and antihistamines and has demonstrated variable part of a novel strategy for treating AC. MSC-CMT obtained from
and limited clinical success,3-6 possibly because other factors in donor BMMSCs and preserved as eye drops might be applied
addition to MCs and histamine, such as T cells, B cells, macro- either therapeutically or prophylactically to patients with AC
phages, platelets, dendritic cells, and neutrophils, also play because of its multifunctional capacities. The prophylactic
important roles in AC. Therefore the simultaneous targeting of application of MSC-CMT for AC, especially seasonal allergic
multiple inflammatory signaling mediators, cellular components, conjunctivitis and vernal keratoconjunctivitis, might reduce the
or both might represent a promising modality for the treatment of occurrence rate of AC rather than only alleviating the symptoms.
this type of allergic disease. Our results demonstrated that in
contrast to MC stabilizers and antihistamine drugs, MSC-CMT Clinical implications: MSC-CMT might be a novel therapeutic
simultaneously acts on multiple targets, including B cells, MCs, option for treating AC and ocular surface inflammatory
and histamine. Furthermore, MSCs possess multifunctional diseases.
properties in addition to potent immunosuppressive and anti-
inflammatory functions through their interactions with and/or
inhibition of different subtypes of T cells, natural killer cells, den- REFERENCES
dritic cells, neutrophils, and macrophages through release of sol- 1. Ono SJ, Abelson MB. Allergic conjunctivitis: update on pathophysiology and pros-
uble factors. Taken together, these unique properties suggest that pects for future treatment. J Allergy Clin Immunol 2005;115:118-22.
MSC-CMT might function as a multidrug dispensary or ‘‘drug 2. Li DQ, Zhang L, Pflugfelder SC, De Paiva CS, Zhang X, Zhao G, et al. Short
ragweed pollen triggers allergic inflammation through Toll-like receptor 4-
store’’ with multifunctional capacities superior to a single immu-
dependent thymic stromal lymphopoietin/OX40 ligand/OX40 signaling pathways.
nosuppressive or antiallergic drug. Future studies should address J Allergy Clin Immunol 2011;128:1318-25.e2.
the detailed method of MSC-CMT preparation and MSC-CMT 3. La Rosa M, Lionetti E, Reibaldi M, Russo A, Longo A, Leonardi S, et al. Allergic
concentration to optimize MSC-CMT therapeutic effects at both conjunctivitis: a comprehensive review of the literature. Ital J Pediatr 2013;39:18.
physiologic and pharmacologic levels. 4. Azari AA, Barney NP. Conjunctivitis: a systematic review of diagnosis and treat-
ment. JAMA 2013;310:1721-9.
Previous studies have used intravenous infusion, subconjunc- 5. Bielory L. Allergic conjunctivitis: the evolution of therapeutic options. Allergy
tival injection, amniotic membranes as a carrier, or a special Asthma Proc 2012;33:129-39.
hollow plastic tube to administer MSCs to the ocular surface.14-18 6. Leonardi S, Marchese G, Marseglia GL, La Rosa M. Montelukast in allergic dis-
However, these methods are not appropriate for clinical AC treat- eases beyond asthma. Allergy Asthma Proc 2007;28:287-91.
7. Le BK, Mougiakakos D. Multipotent mesenchymal stromal cells and the innate
ment. By contrast, topical drug instillation is a common and easy
immune system. Nat Rev Immunol 2012;12:383-96.
clinical administration route for the treatment of ocular surface 8. Salem HK, Thiemermann C. Mesenchymal stromal cells: current understanding
disorders. Thus we used topical MSC-CMT instillation to exploit and clinical status. Stem Cells 2010;28:585-96.
the beneficial properties of MSCs in this study. Our results 9. Ma S, Xie N, Li W, Yuan B, Shi Y, Wang Y. Immunobiology of mesenchymal stem
demonstrate that topical MSC-CMT application significantly cells. Cell Death Differ 2014;21:216-25.
10. Nemeth K, Keane-Myers A, Brown JM, Metcalfe DD, Gorham JD, Bundoc VG,
attenuated EAC. This finding suggests that topical MSC-CMT et al. Bone marrow stromal cells use TGF-beta to suppress allergic responses in
application is a promising alternative for the clinical application a mouse model of ragweed-induced asthma. Proc Natl Acad Sci U S A 2010;
of MSCs to ocular surface disorders. 107:5652-7.
432 SU ET AL J ALLERGY CLIN IMMUNOL
AUGUST 2015

11. Goodwin M, Sueblinvong V, Eisenhauer P, Ziats NP, Leclair L, Poynter ME, et al. of immature DCs: central role of MSC-derived prostaglandin E2. Blood 2009;113:
Bone marrow-derived mesenchymal stromal cells inhibit th2-mediated allergic air- 6576-83.
ways inflammation in mice. Stem Cells 2011;29:1137-48. 31. Aggarwal S, Pittenger MF. Human mesenchymal stem cells modulate allogeneic
12. Cho KS, Park HK, Park HY, Jung JS, Jeon SG, Kim YK, et al. IFATS collection: immune cell responses. Blood 2005;105:1815-22.
Immunomodulatory effects of adipose tissue-derived stem cells in an allergic 32. Kim HS, Shin TH, Lee BC, Yu KR, Seo Y, Lee S, et al. Human umbilical cord
rhinitis mouse model. Stem Cells 2009;27:259-65. blood mesenchymal stem cells reduce colitis in mice by activating NOD2 signaling
13. Su WR, Zhang QZ, Shi SH, Nguyen AL, Le AD. Human gingiva-derived mesen- to COX2. Gastroenterology 2013;145:1392-403, e1-8.
chymal stromal cells attenuate contact hypersensitivity via prostaglandin E2- 33. Nemeth K, Leelahavanichkul A, Yuen PS, Mayer B, Parmelee A, Doi K, et al.
dependent mechanisms. Stem Cells 2011;29:1849-60. Bone marrow stromal cells attenuate sepsis via prostaglandin E(2)-dependent re-
14. Ye J, Yao K, Kim JC. Mesenchymal stem cell transplantation in a rabbit corneal programming of host macrophages to increase their interleukin-10 production.
alkali burn model: engraftment and involvement in wound healing. Eye (Lond) Nat Med 2009;15:42-9.
2006;20:482-90. 34. Rheins LA, Barnes L, Amornsiripanitch S, Collins CE, Nordlund JJ. Suppression
15. Yao L, Li ZR, Su WR, Li YP, Lin ML, Zhang WX, et al. Role of mesenchymal of the cutaneous immune response following topical application of the prosta-
stem cells on cornea wound healing induced by acute alkali burn. PLoS One glandin PGE2. Cell Immunol 1987;106:33-42.
2012;7:e30842. 35. Pavord ID, Wong CS, Williams J, Tattersfield AE. Effect of inhaled prostaglandin
16. Ma Y, Xu Y, Xiao Z, Yang W, Zhang C, Song E, et al. Reconstruction of chemi- E2 on allergen-induced asthma. Am Rev Respir Dis 1993;148:87-90.
cally burned rat corneal surface by bone marrow-derived human mesenchymal 36. Sturm EM, Schratl P, Schuligoi R, Konya V, Sturm GJ, Lippe IT, et al. Prosta-
stem cells. Stem Cells 2006;24:315-21. glandin E2 inhibits eosinophil trafficking through E-prostanoid 2 receptors.
17. Jiang TS, Cai L, Ji WY, Hui YN, Wang YS, Hu D, et al. Reconstruction of the J Immunol 2008;181:7273-83.
corneal epithelium with induced marrow mesenchymal stem cells in rats. Mol 37. Ueta M, Matsuoka T, Narumiya S, Kinoshita S. Prostaglandin E receptor subtype
Vis 2010;16:1304-16. EP3 in conjunctival epithelium regulates late-phase reaction of experimental
18. Oh JY, Kim MK, Shin MS, Lee HJ, Ko JH, Wee WR, et al. The anti-inflammatory allergic conjunctivitis. J Allergy Clin Immunol 2009;123:466-71.
and anti-angiogenic role of mesenchymal stem cells in corneal wound healing 38. Kunikata T, Yamane H, Segi E, Matsuoka T, Sugimoto Y, Tanaka S, et al. Suppres-
following chemical injury. Stem Cells 2008;26:1047-55. sion of allergic inflammation by the prostaglandin E receptor subtype EP3. Nat Im-
19. Akiyama K, Chen C, Wang D, Xu X, Qu C, Yamaza T, et al. Mesenchymal-stem- munol 2005;6:524-31.
cell-induced immunoregulation involves FAS-ligand-/FAS-mediated T cell 39. Zaslona Z, Okunishi K, Bourdonnay E, Domingo-Gonzalez R, Moore BB, Lukacs
apoptosis. Cell Stem Cell 2012;10:544-55. NW, et al. Prostaglandin E2 suppresses allergic sensitization and lung inflamma-
20. Merayo-Lloves J, Zhao TZ, Dutt JE, Foster CS. A new murine model of allergic tion by targeting the E prostanoid 2 receptor on T cells. J Allergy Clin Immunol
conjunctivitis and effectiveness of nedocromil sodium. J Allergy Clin Immunol 2014;133:379-87.e1.
1996;97:1129-40. 40. Krause P, Bruckner M, Uermosi C, Singer E, Groettrup M, Legler DF. Prosta-
21. Magone MT, Chan CC, Rizzo LV, Kozhich AT, Whitcup SM. A novel murine glandin E(2) enhances T-cell proliferation by inducing the costimulatory molecules
model of allergic conjunctivitis. Clin Immunol Immunopathol 1998;87:75-84. OX40L, CD70, and 4-1BBL on dendritic cells. Blood 2009;113:2451-60.
22. Lan W, Petznick A, Heryati S, Rifada M, Tong L. Nuclear factor-kappaB: central 41. Kabashima K, Sakata D, Nagamachi M, Miyachi Y, Inaba K, Narumiya S. Prosta-
regulator in ocular surface inflammation and diseases. Ocul Surf 2012;10:137-48. glandin E2-EP4 signaling initiates skin immune responses by promoting migration
23. Holgate ST, Polosa R. Treatment strategies for allergy and asthma. Nat Rev Immu- and maturation of Langerhans cells. Nat Med 2003;9:744-9.
nol 2008;8:218-30. 42. Yao C, Sakata D, Esaki Y, Li Y, Matsuoka T, Kuroiwa K, et al. Prostaglandin E2-
24. Yamamoto Y, Yin MJ, Lin KM, Gaynor RB. Sulindac inhibits activation of the NF- EP4 signaling promotes immune inflammation through Th1 cell differentiation and
kappaB pathway. J Biol Chem 1999;274:27307-14. Th17 cell expansion. Nat Med 2009;15:633-40.
25. Winter MC, Shasby SS, Ries DR, Shasby DM. Histamine selectively interrupts 43. Heo SC, Jeon ES, Lee IH, Kim HS, Kim MB, Kim JH. Tumor necrosis factor-
VE-cadherin adhesion independently of capacitive calcium entry. Am J Physiol alpha-activated human adipose tissue-derived mesenchymal stem cells accelerate
Lung Cell Mol Physiol 2004;287:L816-23. cutaneous wound healing through paracrine mechanisms. J Invest Dermatol
26. Guo M, Breslin JW, Wu MH, Gottardi CJ, Yuan SY. VE-cadherin and beta-catenin 2011;131:1559-67.
binding dynamics during histamine-induced endothelial hyperpermeability. Am J 44. Young AL, Cheng AC, Ng HK, Cheng LL, Leung GY, Lam DS. The use of autologous
Physiol Cell Physiol 2008;294:C977-84. serum tears in persistent corneal epithelial defects. Eye (Lond) 2004;18:609-14.
27. Gri G, Piconese S, Frossi B, Manfroi V, Merluzzi S, Tripodo C, et al. 45. Pan Q, Angelina A, Zambrano A, Marrone M, Stark WJ, Heflin T, et al. Autolo-
CD41CD251 regulatory T cells suppress mast cell degranulation and allergic re- gous serum eye drops for dry eye. Cochrane Database Syst Rev 2013;8:CD009327.
sponses through OX40-OX40L interaction. Immunity 2008;29:771-81. 46. Meller D, Pauklin M, Thomasen H, Westekemper H, Steuhl KP. Amniotic mem-
28. Kuehn HS, Radinger M, Gilfillan AM. Measuring mast cell mediator release. Curr brane transplantation in the human eye. Dtsch Arztebl Int 2011;108:243-8.
Protoc Immunol 2010;Chapter 7:Unit7.38. 47. Kim SS, Song CK, Shon SK, Lee KY, Kim CH, Lee MJ, et al. Effects of human
29. Brown JM, Nemeth K, Kushnir-Sukhov NM, Metcalfe DD, Mezey E. Bone amniotic membrane grafts combined with marrow mesenchymal stem cells on
marrow stromal cells inhibit mast cell function via a COX2-dependent mechanism. healing of full-thickness skin defects in rabbits. Cell Tissue Res 2009;336:59-66.
Clin Exp Allergy 2011;41:526-34. 48. Liu PF, Guo L, Zhao DW, Zhang ZJ, Kang K, Zhu RP, et al. Study of human acel-
30. Spaggiari GM, Abdelrazik H, Becchetti F, Moretta L. MSCs inhibit monocyte- lular amniotic membrane loading bone marrow mesenchymal stem cells in repair
derived DC maturation and function by selectively interfering with the generation of articular cartilage defect in rabbits. Genet Mol Res 2014;13:7992-8001.
J ALLERGY CLIN IMMUNOL SU ET AL 432.e1
VOLUME 136, NUMBER 2

METHODS permeability was tested 8 hours later by the addition of 7.5 mL of streptavidin-
HRP (1.5 mg/mL, R&D Systems) to the upper chamber. The monolayers were
Reagents and antibodies
stimulated with histamine (100 mmol/L) for 30 minutes before addition of
LPS, Evans blue dye, histamine, phorbol 12-myristate 13-acetate,
streptavidin-HRP. Medium (50 mL) in the lower chamber was collected 5 mi-
calcium ionophore (A23187), NS398, L-NAME, 1-MT, and protein G–
nutes after addition of streptavidin-HRP and assayed for HRP activity by add-
agarose beads were purchased from Sigma (St Louis, Mo). Recombinant
ing 100 mL of TMB substrate. Color development was detected with a
mouse IL-4, TNF-a, and IL-3 were from PeproTech (Rocky Hill, NJ).
microplate reader at 450 nm.
a-MEM and RPMI 1640 medium were obtained from Invitrogen (Grand
In vivo vascular permeability assays were performed by administering a
Island, NY). The following antibodies were used: anti-COX2, anti–
0.1% Evans blue dye solution in PBS (12 mL/kg) through the tail vein after
VE-cadherin, and anti–TNF-a (Abcam, Cambridge, Mass); anti–NF-kB
the first challenge. Mice were killed 60 minutes after injection. The eyelid
p65, anti–phosphorylated NF-kBp65, anti–IkB-a, anti–p38 MAPKs, anti–
and conjunctival tissue were then removed and collected. Evans blue dye
phosphorylated p38 MAPK, anti-STAT3, anti–phosphorylated STAT3, anti-
was extracted after incubation for 24 hours. The extract was measured at
STAT6, anti–phosphorylated STAT6, anti–caspase-1, and anti–b-actin (Cell
620 nm with a microplate reader.
Signaling Technology, Danvers, Mass) and neutralizing antibodies specific
for mouse IL-10 or TGF-b1 or an isotype-matched mAb (R&D Systems,
Minneapolis, Minn). Transfection
COX2 siRNA and fluorescein-conjugated control siRNA were purchased
from Santa Cruz Biotechnology (Santa Cruz, Calif). Three distinct siRNAs
Isolation, culture, and activation of B cells
with the same specificity were used to recognize and exclude off-target effects.
BALB/c splenic B cells were purified (purity >95%, as determined by
siRNA transfection was performed according to the manufacturer’s in-
means of flow cytometric analysis of B220 cell-surface expression, data not
structions. In brief, BMMSCs were washed once with 2 mL of siRNA
shown) with a B-cell isolation kit (Miltenyi Biotec, Auburn, Calif), according
Transfection Medium. The medium was then aspirated. For each transfection,
to the manufacturer’s instructions. Purified B cells were cultured in RPMI
0.8 mL of siRNATransfection Medium was added to each tube containing the
1640 containing 10% FBS, L-glutamine, and 55 mmol/L 2-mercaptoethanol.
siRNA Transfection Reagent mixture (Solution A 1 Solution B). After gentle
For B-cell activation, B cells (1 3 106/mL) were stimulated by the addition
mixing, the mixture was overlaid onto the washed cells. Cells were then
of IL-4 (50 ng/mL) and LPS (10 mg/mL) to culture medium for 5 days. Day
incubated for 5 to 7 hours at 378C in a CO2 incubator. Next, 1 mL of normal
6 supernatants were collected for IgE and IgG1 analysis.
growth medium containing twice the normal concentration of serum and an-
tibiotics (23 normal growth medium) was added without removing the trans-
Isolation, culture, and activation of bone marrow– fection mixture. If toxicity occurred, the transfection mixture was removed
derived MCs and replaced with 13 normal growth medium before adding 23 normal
Bone marrow–derived MCs were derived from BALB/c mice, as previ- growth medium. The cells were then incubated for an additional 18 to 24
ously reported.E1,E2 In brief, BALB/c mice (Jackson Laboratory, Bar Harbor, hours, and the medium was aspirated and replaced with fresh 13 normal
Me) were killed by means of CO2 asphyxiation and bathed in 70% ethanol. growth medium. BMMSCs were used 24 to 72 hours after the addition of fresh
The femurs were then removed with scissors and forceps and placed in 10 medium.
mL of 378C culture medium. After cutting the ends of the femurs, the bone
marrow was flushed from the interior of each femur with a syringe and needle EAC treatment with MSC-CMT
into a well containing 5 mL of medium. The murine EAC model was generated, as previously reported.E4,E5 In
Cells were cultured in RPMI 1640 medium supplemented with 10% FBS, brief, EAC was induced by using the following protocol: a mixture of 50 mg
100 U/mL penicillin, 100 mg/mL streptomycin, 25 mmol/L HEPES, 1.0 of SRW pollen (Greer Laboratories) in 5 mg of Imject Alum (Pierce) was
mmol/L sodium pyruvate, nonessential amino acids (BioSource International, applied through footpad injection on the first day. The sensitization procedure
Camarillo, Calif), 0.0035% 2-mercaptoethanol, and 30 ng/mL recombinant was repeated on day 5 to enhance the allergic reaction. The mice were chal-
mouse IL-3 (PeproTech). After 4 weeks, MC purity was evaluated by means of lenged with 1.5 mg of SRW pollen suspended in 10 mL of PBS in both eyes
toluidine blue staining, and CD117 (c-Kit) surface staining was evaluated by on days 10 to 14. MSC-CMT (10 mL) was topically applied once 30 minutes
means of flow cytometry. The purity of the MCs used in this study was greater before the SRW pollen challenge and 4 times per day on days 10 to 14 in both
than 95% (data not shown). The MCs were used after 4 to 6 weeks of culture at eyes. PBS was applied to the control groups in the same manner (n 5 6 per
378C and 5% CO2. For MC activation, MCs (2 3 105/mL) were stimulated group).
with phorbol 12-myristate 13-acetate (50 nmol/L) and calcium ionophore A slit lamp was used to examine the eyes throughout the course of the study.
(A23187, 1 mg/mL; PMACI) for 16 hours. After 20 minutes of challenge, blinded researchers recorded and evaluated the
clinical reaction according to a published system.E4 In addition, allergic symp-
Isolation and culture of LVECs toms were graded by using a previously published system.E4 Conjunctival
LVECs were isolated from lung tissue of BALB/c mice by using CD45 and edema, lid swelling, tearing, and conjunctival redness were graded from
CD31 microbeads (Miltenyi Biotec). In brief, blood cells were removed from 0 to 4 based on the criteria. The clinical score was the sum of the 4 parameters.
the lung tissue of BALB/c mice by using PBS perfusion before excision. The The scratching times of each animal 15 minutes after the 30-minute challenge
lung tissue was minced and digested with collagenase II (Worthington, were also counted by observers who were blind to the treatment protocol. The
Freehold, NJ). Cell suspensions were filtered through a 40-mm cell strainer scratching response was defined as rapid movements of the hind paws that
(Corning Costar). Magnetic labeling and separation were performed accord- were precisely directed toward the eye. OLF was collected after the last
ing to the manufacturer’s instructions. After total lung cell suspensions were SRW pollen exposure. PBS (10 mL) was applied to the eye by using a micro-
depleted of CD451 cells, positive selection for CD311 cells was performed. pipette without touching the eye. After 2 or 3 forced blinks, OLF was
The enrichment and purity of the endothelial cells were analyzed by using collected. The lavage was repeated 5 times in each eye. The OLF was centri-
flow cytometry (purity >95%). LVECs were then cultured in EGM-2 Single- fuged at 400g for 10 minutes, and the supernatant was separated for further
Quots (Lonza, Walkersville, Md). analysis. One hour after the last challenge, the mice were killed, and the
eyes were removed for further analysis.
In vitro and in vivo vascular permeability assays
In vitro vascular permeability assays were performed, as described pre- Immunohistochemistry
viously.E3 In brief, LVECs (2 3 105) were grown in a Transwell plate (BD Bio- Hematoxylin and eosin staining was performed on paraffin-embedded
sciences) in 500 mL of medium until a monolayer formed. LVEC monolayer sections for histologic evaluation. For semiquantification, positive signals in at
432.e2 SU ET AL J ALLERGY CLIN IMMUNOL
AUGUST 2015

least 5 random high-power fields were visualized and quantified with Image RESULTS
Pro-Plus 5.1 (Media Cybernetics cell numbers, Silver Spring, Md).
MSC-CMT reduces inflammatory cytokine release
by lung epithelial cells
Western blot analysis and ELISA LA-4, a murine lung epithelial cell line, was used to examine
Cell lysates or mouse conjunctiva lysates (50-100 mg of total protein) were
the effects of MSC-CMT on epithelial cells. As shown in Fig E6,
resolved on polyacrylamide-SDS gels and electroblotted onto nitrocellulose
our results showed that MSC-CMT reduced the production of
membranes (Bio-Rad Laboratories, Hercules, Calif). After blocking with 5%
nonfat dry milk/TBS, the membranes were incubated with antibodies against TNF-a and IL-1b by lung epithelial cells stimulated with
VE-cadherin, COX2, p38 MAPK, STAT3, STAT6, caspase-1, NF-kB p65, or lipopolysaccharide.
IkB-a, followed by incubation with an HRP-conjugated secondary antibody.
The signals were visualized with an enhanced chemiluminescence substrate REFERENCES
(Pierce). The blots were then reprobed with b-actin antibody. E1. Su W, Fan H, Chen M, Wang J, Brand D, He X, et al. Induced CD4(1) forkhead
Concentrations of IgE, IgG1, TNF-a, and IL-4 in the supernatants or OLF box protein-positive T cells inhibit mast cell function and established contact hy-
were determined by using ELISA (eBioscience). Concentrations of PGE2 and persensitivity through TGF-beta1. J Allergy Clin Immunol 2012;130:444-52.e7.
E2. Jensen BM, Swindle EJ, Iwaki S, Gilfillan AM. Generation, isolation, and main-
histamine in the supernatants or OLF were determined by using ELISA kits
tenance of rodent mast cells and mast cell lines. Curr Protoc Immunol 2006;
(Cayman Chemical, Ann Arbor, Mich). Chapter 3:Unit 3.23.
E3. Chuang YC, Lei HY, Liu HS, Lin YS, Fu TF, Yeh TM. Macrophage migration
Real-time PCR inhibitory factor induced by dengue virus infection increases vascular perme-
Total RNA was extracted from the tissue lysates with an RNeasy Mini Kit ability. Cytokine 2011;54:222-31.
E4. Merayo-Lloves J, Zhao TZ, Dutt JE, Foster CS. A new murine model of allergic
(Qiagen, Valencia, Calif), and cDNA was generated by using an Omniscript
conjunctivitis and effectiveness of nedocromil sodium. J Allergy Clin Immunol
RT kit (Qiagen). TNF-a and IL-4 mRNA expression were quantified by using 1996;97:1129-40.
the ABsolute SYBR Green ROX mix (Thermo Fisher Scientific, Waltham, E5. Li DQ, Zhang L, Pflugfelder SC, De Paiva CS, Zhang X, Zhao G, et al. Short
Mass). The samples were performed in triplicate, and the relative expression ragweed pollen triggers allergic inflammation through Toll-like receptor 4-depen-
of TNF-a and IL-4 was determined by normalizing the expression of each dent thymic stromal lymphopoietin/OX40 ligand/OX40 signaling pathways.
target gene to b-actin by using the 22DDCt method. J Allergy Clin Immunol 2011;128:1318-25.e2.
J ALLERGY CLIN IMMUNOL SU ET AL 432.e3
VOLUME 136, NUMBER 2

FIG E1. TNF-a, IL-10, IL-4, IL-13 and IFN-g levels in MSC-CMT. A, After
TNF-a depletion with antibody, the TNF-a level was detected by using
ELISA in MSC-CMT. B, IL-10, IL-4, IL-13, and IFN-g levels were measured
by using ELISA in MSC-CMT. **P < .01. Error bars 5 means 6 SEMs.
432.e4 SU ET AL J ALLERGY CLIN IMMUNOL
AUGUST 2015

FIG E2. Treatment with MSC-CMT reduces TNF-a, IL-4, and IL-5 production
and caspase-1 activation in mice with EAC. A-C, TNF-a (Fig E2, A), IL-4 (Fig
E2, B), and IL-5 (Fig E2, C) protein expression in conjunctival tissue was
determined by means of ELISA. D, Caspase-1 expression in conjunctival tis-
sue was determined by means of Western blotting. **P < .01. Error bars 5
means 6 SEMs.
J ALLERGY CLIN IMMUNOL SU ET AL 432.e5
VOLUME 136, NUMBER 2

FIG E3. MSC-CMT inhibits IgG1 release by B cells through a COX2-

dependent mechanism. B cells were cultured with different media for 24
hours. After stimulation with LPS/IL-4 for 5 days, IgG1 levels in supernatants
were determined by using ELISA. A, MSC-CMT was added with specific
neutralizing antibodies for TGF-b1 or IL-10 (10 mg/mL, isotype used as con-
trols), or MSC-CMT was collected after pretreating BMMSCs with a specific
IDO inhibitor (1-MT, 500 mmol/L), a specific NO inhibitor (L-NAME, 1 mmol/
L), or a specific COX2 inhibitor (NS-398, 1 mmol/L). B, MSC-CMT was
collected after pretreating BMMSCs with COX2 siRNA (dCMT) or control
siRNA (cCMT). **P < .01. Error bars 5 means 6 SEMs.
432.e6 SU ET AL J ALLERGY CLIN IMMUNOL
AUGUST 2015

FIG E4. MSC-CMT inhibits IL-4 release by MCs through a COX2-dependent mechanism. A and C, MCs were
cultured with MSC-CMT or MSC-CM for 24 hours and then stimulated with PMACI for an additional 12
hours. IL-4 levels in supernatants were determined by using ELISA. B, Cell viability was assayed by using
trypan blue exclusion. Fig E4, C, MSC-CMT was collected after pretreating BMMSCs with COX2 siRNA
(dCMT) or control siRNA (cCMT). **P < .01. Error bars 5 means 6 SEMs.
J ALLERGY CLIN IMMUNOL SU ET AL 432.e7
VOLUME 136, NUMBER 2

FIG E5. COX2 signaling in BMMSCs is essential for MSC-CMT–mediated

inhibition of inflammatory cell infiltration and IL-4 production. MSC-CMTs
collected after pretreating BMMSCs with COX2 siRNA (dCMT) or control
siRNA (cCMT) were administered to mice with EAC. Inflammatory cell infil-
tration (A) and IL-4 expression (B) in EAC were measured as above at the
indicated time points after challenge. **P < .01. Error bars 5 means 6
SEMs.
432.e8 SU ET AL J ALLERGY CLIN IMMUNOL
AUGUST 2015

FIG E6. MSC-CMT reduces inflammatory cytokine release by lung epithelial

cells. LA-4, a murine lung epithelial cell line, was cultured with MSC-CMT or
MSC-CM for 24 hours and then stimulated with LPS for an additional 12
hours. TNF-a (A) and IL-1b (B) levels in the supernatants were determined
by using ELISA. **P < .01. Error bars 5 means 6 SEMs.