Immunological Homeostasis of the Eye

7/25/2019 Immunological Homeostasis of the Eye

1/18

Immunological homeostasis of the eye

Manabu Mochizuki a,*,1, Sunao Sugita a,b,1, Koju Kamoi a,1

a Department of Ophthalmology & Visual Science, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8519, Japanb Laboratory for Retinal Regeneration, RIKEN Center for Developmental Biology, Japan

a r t i c l e i n f o

Article history:

Available online 26 October 2012

Keywords:

Uveitis

Experimental autoimmune uveitisVogteKoyanagieHarada disease

Human T-cell leukemia virus type 1 uveitis

Ocular pigment epithelial cellsCorneal endothelial cells

Effector T-cells

Th 17 cells

Regulatory T cells

Transforming growth factor

Foxp3

a b s t r a c t

Uveitis is a sight-threatening disease caused by autoimmune or infection-related immune responses.

Studies in experimental autoimmune uveitis and in human diseases imply that activated CD4

T cells,Th1 and Th17 cells, play an effector role in ocular inammation. The eye has a unique regional immune

system to protect vision-related cells and tissues from these effector T cells. The immunological balancebetween the pathogenic CD4 T cells and regional immune system in the eye contributes to the main-

tenance of ocular homeostasis and good vision. Current studies have demonstrated that ocular paren-chymal cells at the inner surface of the blood-ocular barrier, i.e. corneal endothelial (CE) cells, iris

pigment epithelial (PE) cells, ciliary body PE cells, and retinal PE cells, contribute to the regional immunesystem of the eye. Murine ocular resident cells directly suppress activation of bystander T cells andproduction of inammatory cytokines. The ocular resident cells possess distinct properties of immu-

noregulation that are related to disparate anatomical location. CE cells and iris PE cells, which are locatedat the anterior segment of the eye and face the aqueous humor, suppress activation of T cells via cell-to-

cell contact mechanisms, whereas retinal PE cells suppress the activation of T cells via soluble factors. Inaddition to direct immune suppression, the ocular resident cells have another unique immunosup-

pressive property, the induction of CD25Foxp3 Treg cells that also suppress the activation of bystanderT cells. Iris PE cells convert CD8 T cells into Treg cells, while retinal PE cells convert CD4 T cells greatly

and CD8 T cells moderately into Treg cells. CE cells also convert both CD4 T cells and CD8 T cells into

Treg cells. The immunomodulation by ocular resident cells is mediated by various soluble or membrane-bound molecules that include TGF-b TSP-1, B7-2 (CD86), CTLA-2a, PD-L1 (B7-H1), galectin 1, pigment

epithelial-derived factor PEDF), GIRTL, and retinoic acid. Human retinal PE cells also possess similar

immune properties to induce Treg cells. Although there are many issues to be answered, human Tregcells induced by ocular resident cells such as retinal PE cells and related immunosuppressive moleculescan be applied as immune therapy for refractive autoimmune uveitis in humans in the future.

2012 Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

2. T cells and intraocular inflammation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

2.1. Experimental autoimmune uveoretinitis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

2.2. VogteKoyanagieHarada (VKH) disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

2.3. Human T-lymphotropic virus type 1 (HTLV-1) uveitis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

3. Regional defense system in the eye . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17

3.1. Ocular resident cells and their direct immune regulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

3.1.1. Direct T-cell suppression by retinal PE cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

3.1.2. Direct T-cell suppression by iris PE cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

* Corresponding author. Tel.: 81 3 5803 5296; fax: 81 3 5803 0145.

E-mail address: [email protected](M. Mochizuki).1 Percentage of work contributed by each author in the production of the manuscript is as follows: Manabu Mochizuki: 50%; Sunao Sugita: 25%; Koju Kamoi: 25%.

Contents lists available at SciVerse ScienceDirect

Progress in Retinal and Eye Research

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c om / l o c a t e / p r e r

1350-9462/$ e see front matter 2012 Elsevier Ltd. All rights reserved.

http://dx.doi.org/10.1016/j.preteyeres.2012.10.002

Progress in Retinal and Eye Research 33 (2013) 10e27


2/18

3.1.3. Direct T-cell suppression by CE cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

3.2. Regulatory T cells generated by ocular resident cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

3.2.1. Regulatory T cells generated by ocular PE cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

3.2.2. Regulatory T cells generated by CE cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

3.2.3. Regulatory T cells in humans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

4. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

1. Introduction

Uveitis, an intraocular inammatory disorder, is a sight-

threatening disease caused by autoimmune mechanisms or infec-tious agents. Although the pathogenesis of autoimmune uveitisis still controversial, the involvement of auto-antigens or auto-antibodies had been speculated for years prior to the discovery of

a pathogenicretinal auto-antigen by Wacker and colleagues (Wackerand Lipton,1965; Wacker et al.,1977). Since then, theunderstandingof pathogenic mechanisms of autoimmune uveitis has greatly pro-

gressed. Investigations in animal models have shown that activeCD4 T cells (De Kozak et al., 1976; Faure et al., 1977; Salinas-

Carmona et al., 1982; Gregerson and Abrahams, 1983; Mochizukiet al., 1984,1985a; Caspi et al., 1988, 1996; Sonoda et al., 2003;Amadi-Obi et al., 2007), particularly Th1 and Th17 cells (Amadi-Obi

et al., 2007), inltrate the eye and play a central role in the patho-genesis of autoimmune uveitis. In humans, activated CD4 T cellsalso contribute to the immunopathogenic mechanisms of variousdiseases, suchas VogteKoyanagieHarada disease(Sugitaet al.,1996,2006d; Yamaki et al., 2000a, 2000b) and human T-lymphotropic

virus type 1 uveitis (Sagawa et al., 1995;Ono et al., 1997).The eye is protected from invasive pathogens by two systems, an

anatomical barrier and an immunological barrier. The bloodeocular

barrier anatomically blocks harmful pathogens from invading the

eyefrom the peripheralbloodstreamand protects sight-related cellsand tissues. Pigment epithelial (PE) cells of the iris, ciliary body, andretina, vascular endothelial cells of the retina, and corneal endo-

thelial (CE) cells are important components of the bloodeocularbarrier. Once the bloodeocular barrier is disrupted, anotherdefense system, the regional immune system of the eye, suppresses

pathogenic T cells and protects the eye. Because of this uniquefeature, the eye is considered to be an immune-privileged site likethe brain and testes. Streilein and colleagues provided the rstevidence of ocular immune privilege (Streilein et al.,1980; Streilein,

2003). A typical example of ocular immune privilege is anteriorchamber-associated immunedeviation (ACAID) proposed by Kaplanand Streilein (Kaplan, 1977; Kaplan and Streilein, 1978). Ocularresident cells at the inner surface of the bloodeocular barrier, such

as iris PE cells, ciliary body PE cells, retinal PE cells, and CE cells, cansuppress the activation of bystander CD4 T cells (Yoshida et al.,

2000;Ishida et al., 2003;Sugita and Streilein, 2003;Sugita et al.,2004, 2006a, 2009b, 2009c, 2011b, 2012; Futagami et al., 2007;Horie et al., 2009). Furthermore, these ocular resident cells have theunique capacity to convert T cells into regulatory T (Treg) cells,

which suppress the activation of bystander T cells (Sugita et al.,2006b, 2006c, 2007, 2008, 2009a, 2010a, 2011a; Yamada et al.,2010;Horie et al., 2010). Thus, a balance between activated CD4

Tcells in theeye andocularresidentcells maintains thehomeostasis

of the eye.In this reviewarticle, wefocuson (1) the role of activatedCD4 T

cells in uveitis and (2) the role of the regional immune system inthe maintenance of ocular homeostasis, with special attention

given to cellular and molecular mechanisms.

2. T cells and intraocular inammation

2.1. Experimental autoimmune uveoretinitis

The search for pathogenic mechanisms and auto-antigens inautoimmune uveitis was conducted for many years until Wackerand colleagues rst reported a retinal auto-antigen isolated from

water-soluble fractions of bovine retina (Wacker et al., 1977).The antigen was localized to the photoreceptor layer of the retinaand the pineal gland. Immunization of the antigen in complete

Freunds adjuvant at a site away from the eye induced inamma-tion in both eyes and the pineal gland. The antigen was namedretinal soluble antigen (S antigen), and the animal model wasdesignated as experimental autoimmune uveoretinitis (EAU).Later, it was found that EAU is also induced by other retinal anti-

gens, such as interphotoreceptor retinoid-binding protein (IRBP)(Gery et al.,1990), phosducin (Duaet al.,1992), andrecoverin (Geryet al., 1994), and synthetic peptides of these antigens. EAU can beinduced in guinea pigs (De Kozak et al., 1976;Wacker et al., 1977),

rats (De Kozak et al., 1981), mice (Caspi et al., 1988;Iwase et al.,1990), and monkeys (Nussenblatt et al., 1981a). Early studies(Salinas-Carmona et al., 1982; Mochizuki et al., 1985a, 1985b;

Nussenblatt et al., 1981b; Mochizuki and deSmet, 1994) demon-strated that EAU is an organ-specic autoimmune disease medi-

ated by T cells, particularly CD4 T cells. Much later, the successfulinduction of EAU in mice (Caspi et al., 1988) allowed us to dissectthe immunological mechanisms of EAU by using transgenic andknockout (KO) mice.

The development of EAU in mice can be studied with conven-tional induced models or more recently developed spontaneousmodels. In induced EAU models, mice are immunized with a retinalantigen in complete Freunds adjuvant (CFA) (Agarwal and Caspi,

2004). This immunization activates antigen-presenting cells(APCs), such as dendritic cells (DCs) and macrophages, through theinnate immune response and activates retinal antigen-specic T

cells in skin-draining lymph nodes. The spontaneous EAU modelsutilize transgenic mice and do not require active immunizationsand adjuvant. In a representative spontaneous EAU model, trans-

genic mice expressing the neo-self antigen, hen egg lysozyme (HEL)under the retinal-specic IRBP promoter are crossed to the 3A9HEL-specic CD4 T cell receptor (TCR)-transgenic line (Lambeet al., 2007). A more recently developed spontaneous EAU modeluses transgenic mice (R161H) that express a TCR specic to themajor uveogenic epitope of IRBP (Chaon et al., 2012). The sponta-

neous models are much more similar to human uveitis than theinduced models in terms of clinical course. Investigations utilizingthe induced or spontaneous EAU models have revealed the cellularand molecular mechanisms of immune regulation that lead to

ocular immunity and tolerance (Fig. 1).Numerous T cells cannot escapethymic selection due to death

by neglect (the elimination of T cells with no afnity to auto-antigen) or negative selection (the elimination of T cells with high

af

nity to auto-antigen). This central tolerance is controlled by the

M. Mochizuki et al. / Progress in Retinal and Eye Research 33 (2013) 10e27 11


3/18

autoimmune regulator AIRE, which is responsible for the ectopicexpression of tissue-restricted auto-antigens including retinal auto-

antigens in the thymus (Egwuagu et al.,1997; Anderson et al., 2002;Zhang etal., 2003; Ham et al., 2004; Takase et al.,2005; Lambe et al.,2007). Positivelyselected and weaklyauto-reactive T cells can causeinammation. In other words, escaped retinal antigen-specic

CD4 T cells play a central role in EAU models. Natural regulatoryT cells (nTregs) are also generated in the thymus and have poten-tially suppressive effects on autoimmunity (Sakaguchi and

Sakaguchi, 2005). EAU is exacerbated when nTregs are depleted(Avichezer et al., 2003; Grajewski et al., 2006). nTregs and effector Tcells have the same antigenic specicity as they are both educatedand positively selected in the thymus (Sakaguchi and Sakaguchi,2005; Grajewski et al., 2006). Therefore, nTregs target multiple

antigens of the eye and are important for peripheral tolerance andregional tolerance of the eye.

According to molecular mimicry theory, antigen-specic T cells,which escape nTreg patrol, are primed in the periphery by micro-

bial stimuli (Shinohara et al., 1990;Singh et al., 1992;Wildner andDiedrichs-Moehring, 2005). Pathogens, such as microbial compo-nents, express pathogen-associated molecular patterns (PAMPs)and interact with innate pattern recognition receptors, such as toll-

like receptors (TLRs) (Kawai and Akira, 2005), on APCs.

TLRs 1e

11 in humans and TLRs 1e

13 in mice have been identi-ed and many of their ligands have been found (Kawai and Akira,2010). TLRs expressed on APCs, such as DCs and macrophages,

play a critical role in linking innate immunity with adaptiveimmunity. The TLReTLR ligand interaction promotes the matura-tion of APCs through the production of pro-inammatory cytokines

and the up-regulation of co-stimulatory and MHC molecules. TheTLReTLR ligand interaction also allows APCs to become efcient inthe presentation of specic antigens to naive T cells (Janeway andMedzhitov, 2002;Akira and Takeda, 2004; Pulendran, 2005). The

TLReTLR ligand signaling is moderated by the adaptor moleculeMyD88, which activates NF-kB and other signaling pathways toproduce inammatory cytokines. Both MyD88-dependent andMyD88-independent pathways participate in the pathogenesis of

autoimmune diseases (Kawai and Akira, 2007). Although

escaped

self-specic cells circulate without activation under healthyconditions, bacterial infection activates these cells and converts

them into pathogenic effector cells through the TLReTLR ligandsystem (Fujimoto et al., 2006,2008). MyD88 KO mice, but not micelacking TLR2, -4, or -9, had impaired EAU development (Su et al.,2005). The redundancy of TLR2, -3, -4, and -9 in the adjuvant

effect is required to induce EAU (Fang et al., 2010).Some TLRs, such as TLR1/2, 2/6, -4, and -5, respond to bacterial

proteins, whereas other TLRs, such as TLR3, -7, -8, and -9, mainly

respond to bacterial and viral nucleic acids. All TLRs have beendetected at the mRNA level in eye tissues, and the most widelyexpressed TLR in the eye is TLR4, which is expressed in the cornea,conjunctiva, uvea, retina, and sclera (Chang et al., 2006).

Ocular resident cells such as RPE cells, microglia, and astrocytes

have the potential to act as APCs (Percopo et al., 1990;Brodericket al., 2002; Jiang et al., 2008, 2009). Retinal astrocytes allowresponder T cells to induce uveitis in mice when activated by TLR3ligand, polyinosinic-polycytidylic acid, TLR4 ligand, lipopolysac-

charide (Jiang et al., 2009), NOD2, and TLR2 ligand (Jiang et al.,2012).

Damage-associated molecular pattern molecules (DAMPs), such

as heat shock proteins and S100 proteins, are released by stressedcells and bind to DAMPs receptors (Rubartelli and Lotze, 2007).

PAMPs and DAMPs act as danger signals that trigger innateimmunity. Induced models of EAU require immunization with

retinal antigen and adjuvant (pertussis toxin and/or CFA, whichcontains heat-killed tuberculosis bacteria). This bacterial compo-nent triggers pattern recognition receptors on innate immune cells

such as DCs and stimulates the auto-pathogenic effector pathway inEAU. Broad-spectrum antibiotic treatment, which depletes thecommensal bacterial ora, inhibits spontaneous EAU, which mightimply the existence of crosstalk between commensal microbiota in

the gut and uveitis in the eye (Chaon et al., 2012) and might also beclinical evidence of the danger theory.

The main function of DCs is to initiate T-cell-mediated immu-

nity. DCs support naive T-cell activation, clonal expansion, anddifferentiation into effector and memory cell (Bousso, 2008).

However, immature DCs in steady state lead to tolerance (Steinman,

Fig.1. CD4 T cell differentiation in EAU: central, peripheral, and regional mechanism. Positively selected retinal antigen-specic CD4 T cells playa central role in ocular immunity

and tolerance. Danger signal (PAMPs and DAMPs) interacts with PPRs on APCs such as DCs, which leads to naive T cell activation, clonal expansion, differentiation into effector T cell.

Proinammatory cytokine and chemokine contribute to the breakdown of RBB. Accordingly, activated/differentiated CD4 T cells migrate and inltrate into the eye, which results in

EAU. Immune cells such as immature DCs, nTregs, and iTregs induce peripheral tolerance. Anergy has a role in tolerizing mechanism. Resident PE cells and CE cells are main players

in regional tolerance of the eye. mTEC: thymic medullary epithelial cell, AIRE: autoimmune regulator, DC: dendritic cell, PAMPs: pathogen-associated molecular patterns, DAMPs:

damage-associated molecular patterns, PRRs: pattern recognition receptors, TLR: toll like receptor, APC: antigen-presenting cell, IL: interleukin, Th: T helper cell, Treg: regulatory T

cell, BOB: bloodeocular barrier, EAU: experimental autoimmune uveoretinitis, PE: pigment epithelium, CE: corneal endothelium, ACAID: anterior chamber associated immune

deviation.

M. Mochizuki et al. / Progress in Retinal and Eye Research 33 (2013) 10e2712


4/18

2001). Immature DCs have the potential to expand nTreg cells inperipheral tolerance. This expansion is mediated by the release of

suppressive cytokines such as IL (interleukin)-10 and transforminggrowth factor (TGF)-b. Therefore, the administration of immatureDCs suppresses induced and spontaneous EAU (Jiang et al., 2003;Kamoi et al., 2011;Klaska et al., 2012). Interestingly, the targeted

delivery of self-antigen to specic DC subsets can promote toler-ance or immunity in EAU, depending on the DC-specic surfacemolecule (Kamoi et al., 2012).

T cells, macrophages, and neutrophils inltrate eye tissues suchas the choroid and retina during the acute phase in EAU ( De Kozaket al.,1978; Dick et al., 1995; McMenamin et al., 1993). Macrophagesplay a critical role in tissue destruction (Pouvreau et al., 1998;

Forrester et al., 1998; Dick, 2000). Macrophages that cross theblooderetinal barrier and inltrate the retina release mediatorssuch as nitric oxide synthase (NOS)-2 and TNF-a that can causesevere retinal damage (Miura-Takeda et al., 2008). The activation ofrecruited non-antigen-specic macrophages and neutrophils has

an effect on the inammatory function of Th1 and Th17 cells (Kerret al., 2008).

The normal retina contains myeloid-derived macrophages and

microglia (Broderick et al., 2000;Dick et al., 1995;Gregerson et al.,

2004;Penfold et al., 1991), which act as a bridge between innateand adaptive immunity by responding to signals from patternrecognition receptors such as TLRs and modulating Th-cell stimu-

lation. However, as an opposite function, the inltration of myeloidcells has the potential to suppress T-cell activation and proliferationvia nitric oxide- and prostaglandin-mediated pathways (Raveneyet al., 2009,2010).

Effector T cells, especially Th1 and Th17 cells, play a pathogenicrole in EAU (Caspi, 2010). Adoptive transfer of polarized Th1 andTh17 cells into a naive recipient induces EAU (Shi et al., 2008;Coxet al., 2008). EAU induced by IRBP and CFA is IL-17 dependent, but

EAU induced with antigen-pulsed DCs requires IFN-g-producingeffector T cells (Luger et al., 2008;Tang et al., 2007). In induced EAUmodels, Th17 is thought to play a central role in the pathogenesis of

ocular inammation. However, an analysis of a spontaneous EAUmodel (R161H mice) crossed with IFN-g KO mice indicated thatTh17 might be less pathogenic than currently believed (Chaon et al.,2012). Taken together, there is a great deal of evidence that Th1 andTh17 cells contribute to the immune-pathogenesis of EAU (Amadi-

Obi et al., 2007). More recently, the roles of Th9 cells expressing IL-9and Th22 cells expressing IL-22 have been examined in EAU. Th9cells in EAU retain IL-9 production for a brief period, but haveunique functions in the ocular immune system (Tan et al., 2010). IL-

22 induces regulatory CD11b APCs to convert pathogenic T cellsintoTregs, so that IL-22 can reduce the severity of EAU and decreasethe numbers of Th1 and Th17 cells (Ke et al., 2011).

gdT cells also participate in ocular inammation.gd T cells can

either upregulate (Spahn et al., 1999;Rajan et al., 2000;Odyniec

et al., 2004;Tagawa et al., 2004) or downregulate (DSouza et al.,

1997; Tsuchiya et al., 2003; Uezu et al., 2004) adaptive immuneresponses. gd T cells have a stronger effect on Th17-type autor-

eactive T cells than Th1-type autoreactive T cells (Cui et al., 2009;Nian et al., 2010). Nonactivated gd T cells have little effect on ab Tcells, whereas activated gd T cells promote the activation ofab T

cells and enhance EAU development (Nian et al., 2011).The mechanism for trafcking self-antigen from the eye is

unknown. However, T cells must be activated in the peripherallymph nodes or spleen before they inltrate the retina and cause

inammation. The removal of the eye-draining lymph node prior tocorneal transplant prevents corneal graft rejection (Kuffov et al.,2008). This result indicates that self-antigen is trafcked throughthe eye-draining lymph node. Other evidence of antigen trafcking

via lymphatic drainage is that Treg expansion in the eye-draining

lymph node down-regulates pathogenic cells and suppresses

ocular inammation in the spontaneous EAU model (Kamoi et al.,2011). There are candidate resident major histocompatibilitycomplex (MHC) II-expressing APCs in the retina that might activate

T cells to produce cytokines (Jiang et al., 1999). Furthermore, T cellsaccumulate around the 33D1 cells in the retina at the peripheral

margins and around the optic nerve at the onset of EAU (Xu et al.,2007). Another explanation is that CD45 cells derived from thecirculation are sufcient to provideAPC function, even withouteye-resident APCs (Gregerson et al., 2004). Anergy is a tolerizing

mechanism that is induced when peripheral self-antigen presen-tation occurs in the absence of immunogenic co-stimulation(Schwartz, 2003;Appleman and Boussiotis, 2003;van Parijs et al.,1998). Typically, naive T cells encountering antigen in tissue-

draining lymph nodes undergo an initial burst of proliferation fol-lowed by deletion or anergy. The lack of antigen presentation in theeye-draining lymph nodes results in a lack of T-cell anergy, which

leads to ocular inammation (Lambe et al., 2007).Activated retinal antigen-specic T cells are thought to migrate

into the eye from the eye-draining lymph node and/or spleen. Themigration and recruitment of inammatory cells are controlled by

adhesion molecules, chemokines, and chemokine receptors. In EAU,

integrin very late activation antigen-4 (Martn et al., 2005),lymphocyte function-associated antigen 1, intercellular adhesion

molecule 1 (Whitcup et al., 1993), CD44 (Xu et al., 2002), the CXCchemokine receptor (CXCR) 3 (Chen et al., 2004), and CXCR5 (Craneet al., 2006) are molecules responsible for the migration. After themigration, inltrating T cells are controlled by regional mechanisms

of the eye. As nTregs and the effector T cells share the same anti-genic specicity, regional immunity and tolerance that target thesame retinal antigen surely exist together in the eye (Gregerson

et al., 2009).Although EAU is considered to be an animal model for uveitis in

humans, the clinical manifestations and histopathology of EAU arenot exactly the same as those in human disease. Recent research

has shown the critical role of TRL4 in EAU, but single nucleotide

polymorphisms in TLR4 are not important for the development ofuveitis in patients with sarcoidosis (Asukata et al., 2009). Thus,there are limitations to the application of ndings from EAU to

human diseases. However, studies of EAU indicate that CD4 T cellsplay a critical role in ocular immunity and tolerance. Therefore, we

studied the role of CD4 T cells in VogteKoyanagieHarada disease,which is a prototype of autoimmune uveitis, and uveitis associatedwith human T-lymphotropic virus type 1 infection.

2.2. VogteKoyanagieHarada (VKH) disease

VKH disease was rst described by a dermatologist (Vogt, 1906)in a case report of a patient with poliosis, vitiligo, and uveitis. More

precise ocular manifestations were later described by Koyanagi

(1914)and Harada (1926). The disease is a non-traumatic bilateraluveitis that is common in pigmented ethnic groups, such as the

Japanese, but rare in non-pigmented ethnic groups, such as

Caucasians. In the early acute stage of the disease, diffuse choroi-ditis is the primary lesion, as described by Koyanagi (1914) andHarada (1926)and as demonstrated in the histopathology of ocular

lesions reported by Inomata (1989). In this early stage, patientsdevelop bilateral uveitis (Fig. 2A, B) together with systemic symp-toms such as headache, nausea, tinnitus, and hearing disturbance.Another characteristic feature of this disease is depigmentation in

pigment-containing tissues such as sunset glow fundus of the eye(Fig. 2C, D), vitiligo, alopecia, and poliosis (Fig. 3) in the late stage ofthe disease. Becauseof these clinical and histopathological features,VKH disease is considered to be an autoimmune disease against

melanocytes. However, the pathogenic auto-antigen was unknown



5/18

untilYamaki et al. (2000a,2000b)established an animal model ofVKH disease by immunizing animals with tyrosinase, which is

a melanocyte-associated antigen and an enzyme related to melanin

synthesis. Tyrosinase-immunized animals developed bilateral

uveitis 2 weeks after immunization, followed by depigmentation inthe skin and in the choroid resembling sunset glow fundus 8 weeks

after the immunization. More importantly, peripheral CD4 T cells

from patients with VKH disease responded to a tyrosinase peptide.

Fig. 3. Depigmentation in VogteKoyanagieHarada disease (A) Alopecia in a 12-year-old girl 3 months after onset (B) Poliosis of eye lash in the same patient as (A), (C) Vitiligo in

a 70-year-old woman 2 years after onset.

Fig. 2. Ocular fundus photograph of a patient with VogteKoyanagieHarada disease (18-year-old woman) (A) Right eye: 6 days after onset, 6 April 2004 (BCVA 0.2). (B) Left eye: 6

days after onset, 6 April 2004 (BCVA 0.3). (C) Right eye: 11 months after onset, 5 March 2005, exhibiting sunset glow fundus with choroidal neovascularization (BCVA 0.08). (D)

Left eye 11 months after the onset, 5 March 2005, exhibiting sunset glow fundus with chorioretinal atrophy and pigment migration (BCVA 0.8). BCVA: best corrected visual acuity.



6/18

To determine if ocular inltrating lymphocytes are sensitized to thetyrosinase antigen, we assayed the lymphocyte response to

a tyrosinase peptide, tyrosinase450e462, by using T-cell clonesestablished from inltrating cells in the eyes of patients with VKHdisease and other clinical entities of uveitis (Sugita et al., 2006d).CD4 T-cell clones established from VKH disease, but not from

Behets disease and sarcoidosis, responded to the tyrosinase450e462peptide. The T-cell clone from VKH disease did not respond tocontrol antigen, an inuenza virus peptide. The response wasstrictly dependent on the concentration and structure of the

peptide, indicating that CD4 Tcells of VKH disease, but notof othertypes of uveitis, are sensitized to and respond to tyrosinase peptide.However, the molecular mechanisms by which CD4 T cells of VKHpatients are sensitized to tyrosinase are still unknown. We

hypothesized that exogenous pathogens, such as a virus, mighthave structural homology with the tyrosinase peptide. GeneBankdata analysis disclosed only one exogenous antigen, cytomegalo-virus envelope glycoprotein H290e302 (CMV-egH290e302), that

shared structural homology with the tyrosinase450e462peptide. Sixamino acids in the 13-amino acid tyrosinase peptide had homologs,and the sequence was an essential motif recognized by HLA-class II

antigen. VKH disease is signicantly associated with HLA-class II

antigen, HLA-DR4, in many ethnic groups. We further studiedwhether CD4 T cells of VKH patients responded to bothtyrosinase450e462and CMV-egH290e302in culture. Peripheral blood

mononuclear cells from patients with VKH disease, but not frompatients with other types of uveitis or healthy donors, responded toboth tyrosinase450e462 and CMV-egH290e302. More importantly,CD4 T-cell clones established from ocular inltrating cells of VKH

disease reacted to both tyrosinase450e462and CMV-egH290e302andproduced inammatory cytokines.

In the mimicry phenomenon, T cells specic against microbialproducts recognize and react to autologous tissue molecules

(Oldstone, 1998;Barnaba and Sinigaglia, 1997). The phenomenonwas investigated in EAU, and several microbial molecules thatsufciently cross-react with uveitogenic antigens were identied

(Singh et al., 1989a,1989b,1990). T cells stimulated with mimicrypeptides from retinal auto-antigens could penetrate the oculareblood barrier and cause intraocular inammation (Wildner andDiedrichs-Mhring, 2003,2004).

The TLReTLR ligand system plays a crucial role in triggering

innate immunity. From the standpoint of the danger theory, it hasbeen revealed that TLR9 recognizes unmethylated cytidine-phosphate guanosine DNA motifs that are frequently present inviruses including CMV. However, single nucleotide polymorphisms

in the TLR9 gene were not signicantly associated with suscepti-bility to VKH (Ito et al., 2011). Based on studies conducted byourselves and others, one of the most critical aspects of VKHpathogenesis can be explained by molecular mimicry theory but

not by the danger theory, which does not consider molecular

structure.The molecular mechanisms of VKH disease are summarized in

Fig. 4. Following CMV infection, CD4 T cells of the infected patient

are sensitized to CMV-egH290e302 peptide. A tyrosinase450e462peptide expressed on the melanocytes in the eye, meninges, andskin is recognized by the CMV-sensitized CD4 T cells due to the

molecular mimicry between the two peptides. This recognitioncauses the inammatory reaction characteristic of VKH disease.Thus, the pathogenesis of VKH disease highlights the importance ofCD4 T cells in human autoimmune uveitis.

2.3. Human T-lymphotropic virus type 1 (HTLV-1) uveitis

A signicant role for CD4 T cells in human uveitis has been

further highlighted in another type of uveitis associated with a viral

infection, HTLV-1 uveitis. HTLV-1 is a human retrovirus, an RNAvirus that encodes an RNA-dependent DNA polymerase that

translates the viral RNA into a DNA provirus. The provirus is thenintegrated into the T-cell genome. HTLV-1 can cause adult T-cellleukemia and T-cell lymphoma when the integration is monoclonal(Hinuma et al., 1981;Yoshida et al., 1982,1984;Yamaguchi et al.,

1984). If the integration is polyclonal, it causes an inammatorydisorder in the central nervous system, tropical spastic paraparesis(Gessain et al., 1985). The virus is endemic in the Caribbean islands,parts of central Africa and southwestern Japan (Mochizuki et al.,

1996). A causative relationship for HTLV-1 and uveitis wasdemonstrated by a series of sero-epidemiological, virological, andclinical surveys performed in southwestern Japan (Miyakonojo)(Mochizuki et al., 1992a, 1992b; Yamaguchi et al., 1994; Sagawa

et al., 1995; Masuoka et al., 1995; Ono et al., 1995, 1997, 1998).The HTLV-1 seroprevalence in patients with idiopathic uveitis of

unknown etiology was signicantly higher than that of two controlgroups (patients with non-uveitis ocular diseases, such as cataracts

and glaucoma, and patients with uveitis of a dened etiology, suchas Behets disease and sarcoidosis) (Mochizuki et al.,1992a, 1992b;Mochizuki, 2009) (Fig. 5). The seroprevalence in the control groups

Fig. 4. Hypothesis of molecular mechanisms of VogteKoyanagieHarada disease. The

gure is reproduced (Mochizuki, 2009, Fig. 6) with permission. CMV: cytomegalovirus.

Fig. 5. Seroepidemiology of HTLV-1 in an endemic area. The gure is reproduced

(Mochizuki, 2009, Fig. 7) with permission. -: Patients with idiopathic uveitis without

dened etiology,: Patients with uveitis with dened etiologyB: Patients with non-

uveitis ocular diseases.



7/18

increased with patient age, which was in accordance with the agedistribution of HTLV-1 seroprevalence in healthy blood donors. In

contrast, the HTLV-1 seroprevalence in patients with idiopathicuveitis was extremely high, even in young patients (Fig. 5). Asimilar result was found in Kurume, Japan, an area in which HTLV-1

is less endemic. The odds ratio of idiopathic uveitis for HTLV-1infection was estimated to be 11.0 in Miyakonojo and 7.8 in Kur-ume (Mochizuki et al., 1992b). Uveitis in HTLV-1 carriers is char-acterized by intermediate uveitis with vitreous opacities ( Fig. 6)

(Yoshimura et al., 1993; Takahashi et al., 2000). We comparedocular manifestations of idiopathic uveitis to seropositivity forHTLV-1 (Yoshimura et al., 1993). The HTLV-1-seropositive group

had a signicantly higher incidence ofoaters, vitreous opacities,retinal vasculitis, and intermediate uveitis than the seronegativepatients (P


8/18

The recent discovery of Tregs has helped elucidate the inam-matory mechanisms in HTLV-1-infected hosts. Tregs suppress the

CTL response in HTLV-1-infected patients, and reduced Foxp3expression and Treg function was found in CD4CD25 T cells frompatients with HTLV-1-associated myelopathy/tropical spastic par-aparesis (Hayashi et al., 2008;Michaelsson et al., 2008;Oh et al.,

2006;Ramirez et al., 2010;Yamano et al., 2005). Overexpressionof HTLV-1 Tax reduced Foxp3 expression and inhibited thesuppressive function of Treg cells in vitro (Yamano et al., 2005).

Signicantly decreased numbers of CD4CD25Foxp3 Treg cells

were observed in transgenic mice expressing HTLV-1 Tax (Ohsugiand Kumasaka, 2011). In addition, a newly identied virusprotein, HTLV-1 basic leucine-zipper factor (HBZ), has the potentialto suppress Tax-mediated transcription activation of the long

terminal repeat (Yasunaga and Matsuoka, 2011). CD4Foxp3 Tregcells in HBZ transgenic mice were functionally impaired ( Satouet al., 2011). Taken together, these results indicate that HTLV-1Tax and HBZ must be responsible for immune disturbances in

HTLV-1-infected patients; therefore, these molecules are beinginvestigated as target molecules for the treatment of HTLV-1 uveitis(Kamoi and Mochizuki, 2012). HTLV-1 uveitis is caused by the

alternation of immune status and the cytokine production from

HTLV-1-infected nonmalignant cells (Kamoi and Mochizuki, 2012).Many collaborators with immune cells around infect CD4 T cellscontribute to the pathogenesis of HTLV-1 uveitis, which highlighs

the signicant role played by CD4 T cells in human uveitis.

3. Regional defense system in the eye

Under conditions of intraocular inammation in experimentalanimals and humans, activated CD4 T cells inltrate the eye andcause immune responses and inammation, which damage vision-related cells and tissues. However, the eye has a unique immune

system to protect important cells and tissues from activatedeffector CD4 T cells. Streilein and colleagues provided the rstevidence that ocular parenchymal cells participate in ocular

immune privilege (Streilein, 2003). They demonstrated a phenom-enon termed anterior-chamber-associated immune deviation(ACAID), in which tumor cells inoculated in the anterior chamber ofthe eye induce antigen-specic suppression of the cellular immuneresponse but not the humoral immune response. The immune

deviation arises when antigens are placed not only in the anteriorchamber (Streilein, 1987), but also in the vitreous cavity (Jiang andStreilein, 1991) or subretinal space (Wenkel and Streilein, 1998).ACAID arises because intraocular APCs capture eye-derived antigen,

migrate directly into the bloodstream through the trabecularmeshwork, and trafc to the spleen.

Several lines of evidence support this primary route of migra-tion of eye-derived APCs in ACAID. First, 24e48 h after mice receive

an anterior-chamber injection of antigen, their blood contains APCs

that induce ACAID when injected intravenously into naive mice(Wilbanks and Streilein, 1991). Second, APCs from the iris andciliary body of mice that received an antigen injection in the

anterior chamber induce ACAID when injected intravenously intonaive mice (Wilbanks et al., 1991). Third, conventional APCs treatedin vitro with TGF-b migrate preferentially to the spleen when

injected intravenously into mice (Wilbanks and Streilein, 1992).Antigen-bearing eye-derived cells (F4/80 positive) come to restselectively in the marginal zone of the spleen. Then, CD4 and CD8

T cells that are specic for eye-derived antigen accumulate in the

marginal zone of the spleen. The eye-derived APCs and B cellsin thespleen contribute to the activation of these T cells. Eventually,antigen-specic Tregs that mediate ACAID emerge from these cellclusters in the spleen (Niederkorn and Streilein, 1983; Wilbanks

and Streilein, 1990).

The CD4 population of Tregs is known as afferent becausethese cells suppress the initial activation and differentiation of

naive T cells into Th1-type effector cells. The CD8 population ofTregs is known as efferent because these cells inhibit Th1-mediated immunity, such as delayed-type hypersensitivity. TheCD8 Tregs of ACAID act in the periphery, including the eye,

whereas afferent ACAID CD4 Tregs act in secondary lymphoidorgans (Streilein et al., 2002). Eventually, activated antigen-specicCD4 and CD8 T cells that differentiate into the antigen-specicregulatory T cells of ACAID can protect a graft from immune

rejection after transplantation.Tissue-based ocular mechanisms of immune privilege that

contribute to the ocular defense systems have been identied(Forrester et al., 2008; Forrester, 2009). An important molecule

contributing to immune privilege in the eye is Fas Ligand (FasL). Anabundance of FasL is constitutively expressed throughout the eye(Ferguson and Grifth, 2006). FasL induces apoptosis of any acti-

vated Fas cells that enter the eye, and FasL-induced cell death can

lead to tolerance and blocks the growth of blood vessels that candamage the eye (Ferguson and Grifth, 1997; Dick et al., 1999;Gregory et al., 2002, 2005; Ferguson and Grifth, 2006). TNF-

related apoptosis-inducing ligand is also expressed on ocular

tissues and participates in ocular immune privilege (Ferguson andGrifth, 2007; Wosik et al., 2007). Additionally, immune regula-tion provided by ocular resident cells is a critical component of the

ocular defense system.As described above, ACAID is eye-derived antigen specic

immune regulation. In addition to this, ocular resident cells (PEcells of iris, ciliary body and retina, CE cells) contribute to the

immune privilege of the eye by another immune regulation whichis not specic to eye-derived antigen. The ocular resident cells arecapable of directly suppressing T cells activated by anti-CD3 anti-body, but not by eye-derived antigens. The following chapters

highlight the molecular mechanisms of the immune regulation byocular resident cells.

3.1. Ocular resident cells and their direct immune regulation

It is obvious that immune regulation might work much moreefciently at the gate of the bloodeocular barrier rather than after

activated effector CD4 T cells enter the eye. Table 1summarizes

the immunoregulatory properties of ocular resident cells located atthe inner surface of the bloodeocular barrier. Ocular resident cells,such as PE cells of the iris, ciliary body, and retina and CE cells, haveimmunomodulatory capacity and contribute to ocular immune

privilege (Yoshida et al., 2000;Streilein, 2003;Ishida et al., 2003;Sugita, 2009; Mochizuki, 2009; Hori et al., 2010). For instance,murine retinal PE cells suppress Th1 cells (Sugita et al., 2006a),Th17 cells (Sugita et al., 2011b), CD8 T cells (Sugita et al., 2004,

2008), B cells (Sugita et al., 2010b), and macrophages (Zamiri et al.,

2006), and they upregulate Foxp3

regulatory T cells (Sugita et al.,2008, 2009a). PE cellsof the iris,ciliary body, and retina and CE cellsutilize distinct molecular mechanisms related to their anatomical

locations and microenvironments to mediate immunomodulation.When ocular PE cellsisolated from the iris, ciliary body, or retina

of B57BL/6 mice were co-cultured with bystander T cells in the

presence of anti-CD3 antibody (Sugita and Streilein, 2003), T-cellactivation was signicantly suppressed, indicating that all ocular PEcells are capable of suppressing the activation of target T cells(Fig. 7). However, when transwell membranes were inserted

between ocular PE cells and target T cells, the iris PE cells failed tosuppress the activation of target T cells (Fig. 7), indicating that cell-to-cell contact is necessary for the suppression of responder T cellsby iris PE cells. In contrast, retinal PE cells suppressed T-cell

proliferation even after transwell membranes were inserted



9/18

between retinal PE cells and target T cells (Fig. 8), indicating thatthe suppressive activity of retinal PE cells is mainly mediated by

soluble factors. Ciliary body PE cells had an intermediate property,retaining a modest but signicant capacity to suppress T-cell acti-vation across the membrane (Fig. 7). These observations are inaccordance with the disparate microenvironments in which these

ocular PE cells are located. Namely, iris PE cells are located in theanterior segment of the eye and face the aqueous humor. In such anenvironment, soluble factors would be readily diluted by theaqueous humor, and cell-to-cell contact would be most efcient. On

the other hand, retinal PE cells are located in the cell-rich micro-environment between the choroid and the neural retina. In sucha cell-dense microenvironment, soluble factors would be muchmore efcient than cell-to-cell contact.

3.1.1. Direct T-cell suppression by retinal PE cells

To identify soluble factors participating in the immunoregula-tion by retinal PE cells, we performed a GeneChip microarray assay

with PE cells of the iris, ciliary body, and retina of C57BL/6 mice(Futagami et al., 2007). Several eye-derived immunoregulatory

genes were expressed at high levels in ocular PE cells. Amongthem, two molecules were identied as key soluble factors forimmunoregulation by retinal PE cells, thrombospondin-1 (TSP-1)and TGF-b.

TGF-bis an intense immunosuppressive factor that is expressedin ocular PE cells at the gene level and the protein level. A TGF-

b bioassay using MvILu cells demonstrated that retinal PE cellsproduce and secrete a soluble form of TGF-b(Sugita et al., 2006a).

We further investigated whether TGF-b signaling is necessary forimmunosuppression by retinal PE cells by using dominant-negativeTGF-b receptor II (TGF-b RII) transgenic mice, in which TGF-

b signals do not enter T cells through their TGF-b receptors. The

proliferative response of activated T cells from wild-type mice wasgreatly suppressed by retinal PE cells, whereas the activation of T

cells from dominant-negative TGF-bRII mice was enhanced ratherthan suppressed by retinal PE cells. Thus, TGF-b signals are essential

Fig. 7. Immunosuppressive properties of murine ocular pigment epithelial cells (iris, ciliary body, retina) on activation of bystander T cells. The gure is reproduced (Sugita and

Streilein, 2003, Fig. 1) with permission. A. Iris pigment epithelial cells (IPE): IPE greatly suppressed activation of bystander T cells, and the suppression completely disappeared

by the use of transwell cell-insert mem branes. B. Ciliary doby pigment epithelial cells (CBPE): CBPE also greatly suppressed activation of bystander T cells, and the suppression was

moderately disturbed by the use of transwell cell-insert membranes. C. Retinal pigment epithelial cells (RPE): RPE also greatly suppressed activation of bystander T cells, but the

suppression was minimally affected by the use of transwell cell-insert membranes.

Table 1

Summary of immunosuppressive properties of the ocular resident cells.

Ocular resident cells Use of cell contact Secretion of soluble factors Induction of Treg cells Candidate molecules Representative reference

Retinal PE () () () (CD4 & CD8) TGFb Sugita et al., 2009a

TSP-1 Futagami et al., 2007

PD-L1 (B7-H1) Sugita et al., 2009b

Galectin-1 Ishida et al., 2003

PEDF Zamiri et al., 2006

CTLA-2a Sugita et al., 2008Retinoic acid Kawazoe et al., 2012

Iris PE () () () (CD8) TGFb Sugita et al., 2006b

TSP-1 Futagami et al., 2007CD86 (B7-2) Sugita and Streilein, 2003

Ciliary body PE () () () TGFb Sugita et al., 2006a

TSP-1 Futagami et al., 2007

Cornea endothelium () () () (CD4 & CD8) TGFb Yamada et al., 2010

PD-L1 (B7-H1) Sugita et al., 2009c

GITRL Hori et al., 2010

CTLA-2a Sugita et al., 2011a

TSP-1, thrombospondin-1; PEDF, pigment epithelial-derived factor; CTLA-2a, cytotoxic T lymphocyte antigen-2 alpha, glucocorticoid-induced tumor necrosis factor receptor

family-related protein ligand (GITRL).



10/18

for T-cell suppression by retinal PE cells. Under normal conditions,TGF-bis constitutively expressed on ocular PE cells in a latent form,which is converted to the active form by TSP-1. TSP-1 is expressed

on all three types of ocular PE cells at the gene level and the proteinlevel, but its expression level on retinal PE cells is particularly high(Futagami et al., 2007). The function of TSP-1 was tested by pre-treating retinal PE cells with antibodies to various soluble factors.

Only the anti-TSP-1 antibody neutralized the suppressive activity of

retinal PE cells, indicating that TSP-1 is necessary for T-cellsuppression by retinal PE cells. Furthermore, the addition ofrecombinant TSP-1 to the culture medium of retinal PE cells

enhanced the secretion of active TGF-b, but not total TGF-b, byretinal PE cells (Futagami et al., 2007).

Another co-stimulatory factor, programmed cell death ligand(PD-L1 or B7-H1), is involved in T-cell suppression by retinal PE

cells (Usui et al., 2008; Sugita et al., 2009b). PD-L1 is widelyexpressed on the thymus, spleen, heart, endothelium, epithelium,tumors, and immunocytes such as T cells, B cells, DCs, and mono-cytes (Dong et al., 1999;Yamazaki et al., 2002). In the eye, PD-L1 is

constitutively expressed on CE cells and provides a negative co-stimulatory signal for effector T cells that helps to inhibit cornealallograft rejection (Hori et al., 2006). In addition to CE cells, human

retinal PE cell lines constitutively express PD-L1 co-stimulatorymolecules, and retinal PE cells can suppress programmed cell death1 (PD-1)-expressing human T cells (Usui et al., 2008). We furtherexamined whether murine retinal PE cells can suppress the acti-vation of bystander T cells during inammatory conditions (Sugita

et al., 2009b) by using IFNg-pretreated retinal PE cells. Althoughprimary murine retinal PE cells did not express PD-L1, the expres-sion of PD-L1 was induced on the surface of IFNg-pretreated retinalPE cells. The IFNg-pretreated retinal PE cells suppressed activation

of PD-1 bystander T cells, but they failed to suppress bystander Tcells from PD-1 KO mice. Anti-PD-L1 antibody neutralized thesuppression of bystander T cells by IFNg-pretreated retinal PE cells.

Thus, PD-L1 molecules are expressed on retinal PE cells in thepresence of inammatory cytokine IFNg, and the interaction

between PD-L1-expressing retinal PE cells and IFNg-producing Th1cells provides negative co-stimulatory signals, thereby suppressing

bystander Th1-type cells.Based on these ndings, the following model has been proposed

to explain the molecular mechanisms of direct suppression ofbystander T cells by retinal PE cells. TSP-1 molecules expressed on

retinal PE cells convert latent TGF-b to active TGF-b, and thensoluble TGF-bis secreted from retinal PE cells. The soluble form ofactive TGF-b provides negative signals to activated CD4 T cellsthrough their TGF-b receptors and suppresses the activation of

bystander T cells. The interaction between PD-L1 molecules onretinal PE cells and PD-1 molecules on IFNg-producing bystander Tcells provides a negative co-stimulatory signal to bystander T cellsunder inammatory conditions.

3.1.2. Direct T-cell suppression by iris PE cells

To identify unique regulatory molecules on iris PE cells that

participate in immunosuppression via cell-contact-dependentmechanisms, the gene expression of co-stimulatory moleculeswas compared among PE cells of the iris, ciliary body, and retina

(Sugita and Streilein, 2003). Among the tested molecules, onlyCD86 (B7-2) molecules were greatly expressed on iris PE cells butnot expressed on retinal PE cells and weakly expressed on ciliarybody PE cells. The expression of B7-2 on iris PE cells was also

demonstrated at the protein level by ow cytometry. B7-2 mole-cules function as co-stimulatory molecules on antigen-presentingcells. To investigate the function of B7-2 molecules on iris PE

cells, we used B7-2 KO mice (Sugita et al., 2004). Iris PE cells fromwild-type mice strongly suppressed T-cell activation, while iris PEcells from B7-2 KO mice failed to suppress T-cell activation, indi-cating that B7-2 molecules expressed on iris PE cells are essential

for direct bystander T-cell suppression by iris PE cells. Becausecytotoxic T lymphocyte-associated antigen-4 (CTLA-4) is a ligandfor the B7-2 molecule, we further performed immunohistochem-ical studies to determine if CTLA-4 molecules were expressed on

activated T cells when T cells were co-cultured with iris PE cells.CTLA-4 was clearly expressed on the cell surface of anti-CD3-

stimulated T cells that contacted iris PE cells (Sugita et al., 2004).Immunohistochemical studies also demonstrated that B7-2 mole-cules were expressed on the cell surface of iris PE cells that con-tacted anti-CD3-stimulated T cells. As for the function of CTLA-4,the activation of bystander T cells from CTLA-4/ wild-type micewas greatly suppressed by iris PE cells, while T cells from CTLA-4/

KO mice were not suppressed by iris PE cells, indicating that the

interaction between B7-2 (CD86) molecules and CTLA-4 moleculesplays a key role in the direct immunosuppressive activity of iris PEcells on activated T cells.

Although B7-2 molecules and CTLA-4 molecules play a key role

in the immunosuppressive activity of iris PE cells, these moleculesare co-stimulatory molecules, not primary immunosuppressivemolecules. Like the posterior segment of the eye, the key immu-

nosuppressive factor in the anterior segment is TGF-b. TGF-b isexpressed on ocular PE cells of the iris, ciliary body, and retina.Ocular tissues (iris, ciliary body, and retina) express TGF-b at themRNA level and the protein level (Sugita et al., 2006b). In addition,cultured iris PE cells were positively stained with anti-TGF-b2

antibody (cell surface and intracellular). The function of TGF-bwasinvestigated by using transgenic dominant-negative TGF-b RIImice, in which TGF-b signals do not enter T cells through T-cellreceptors (Sugita et al., 2006b). The activation of bystander T cells

from dominant-negative TGF-bRII mice was not suppressed by irisPE cells while that from wild-type mice was greatly suppressed byiris PE cells, indicating that TGF-b signals from iris PE cells tobystander T cells are essential.

3.1.3. Direct T-cell suppression by CE cellsCE cells are in contact with the aqueous humor as a part of the

inner surface of the anterior segment of the eye, like iris PE cells.

Because human CE cells lack regenerative capacities, they haveunique mechanisms that provide protection from immune-mediated cell damage. Human CE cells inhibit IL-2 production by

T cells (Kawashima et al., 1994) and constitutively express CD95ligand (Fas L), which induces the apoptosis of T cells expressing CD95 (Fas) molecules (Grifth et al., 1995; Yamagami et al., 1997).Another immune mechanism, the interaction between co-

stimulatory molecules PD-L1 and PD-1, was more recently discov-ered.Hori et al. (2006) demonstrated that PD-L1 co-stimulatorymolecules expressed on CE cells provide negative co-stimulationfor effecter T cells to inhibit corneal allograft rejection. We

Fig. 8. Experimental procedures to identify induction of regulatory T (Treg) cells by

ocular pigment epithelial cells. The gure is reproduced (Mochizuki, 2009, Fig. 29)

with permission.



11/18

hypothesized that this mechanism to might also protect CE cellsfrom activated effector T cells inltrating the anterior segment of

the eye with uveitis. To examine this hypothesis, we studied thesuppressive activity of CE cell lines from human primary CE cells

and allogeneic activated T cells or T-cell clones established frominltrating cells in the eyes of uveitis patients as target cells ( Sugitaet al., 2009c). Human CE cells suppressed both in vitro proliferationand IFNgproduction by activated allogeneic CD4 T cells and T-cell

clones from uveitis patients via a cell contact-dependent mecha-nism, like iris PE cells. Human CE cells constitutively expressed co-stimulatory PD-L1 and PD-L2 molecules, and their expression wasenhanced by IFNg. Human CE cells efciently inhibited the prolif-

eration of PD-1 Th1 cells and activated CD4 T cell lines and clonesestablished from patients with ocular inammatory diseases suchas Behets disease, sarcoidosis, VKH disease, and corneal endo-theliitis. A neutralizing antibody for PD-L1, but not PD-L2, blocked

the suppressive effect of human CE cells on Th1 cells (Sugita et al.,2009c). Thus, CE cells suppress Th1-inltrating effector CD4 Tcellsby a cell-contact mechanism via the PD-L1/PD-1 interaction andparticipate in the immune system in the anterior segment of the

eye.

3.2. Regulatory T cells generated by ocular resident cells

In addition to direct immune suppression, ocular resident cells(iris PE cells, ciliary body PE cells, retinal PE cells, and CE cells) can

induce the conversion of activated T cells to regulatory T (Treg)cells. Ocular resident cell-induced Treg cells can greatly suppressthe activation of T cells, thereby down-regulating the immune andinammatory responses in the eye and maintaining ocular

homeostasis. By understanding the molecular mechanisms of Treg-

cell generation by ocular resident cells, we can identify candidatecells and molecules for new immunotherapies to treat ocular

inammatory diseases.We performed an in vitro study to determine if ocular resident

cells can generate Treg cells from activated T cells as follows (Sugita

et al., 2006c) (Fig. 8). Naive T cells (CD4 or CD8) were isolatedfrom C57BL/6 mice and cultured with PE cells of the iris, ciliarybody, and retina in the presence of a low concentration of anti-CD3antibody (0.01 mg/mL). The T cells exposed to ocular PE cells were

harvested, gamma-irradiated, and added to responder CD3 T cellsthat were harvested from C57BL/6 mice. The responder T cells andocular PE cell-exposed T cells were co-cultured in the presence ofanti-CD3 antibody (1.0 mg/mL), and then the proliferation and

cytokine production by responder T cells were assessed. Thesuppression of responder T cells would indicate that the ocular PEcells generated Treg cells. As described below, the mode of actionand molecular mechanisms of Treg-cell generation by ocular resi-

dent cells differ depending on the microenvironment of the ocularresident cells.

Fig. 9. Induction of regulatory T cells by murine ocular pigment epithelial cells. The

gure is reproduced (Sugita et al., 2008, Fig. 1) with permission.

Fig. 10. Molecular mechanisms of regulatory T (Treg) cells by iris pigment epithelial

cells. PE: pigment epithelial cells, TCR: T cell receptor, CTLA: cytotoxic T lymphocyte-

associated antigen, TSP: thrombospondin, mTGFb: membrane bound transforming

growth factor b, TGFbR: transforming growth factorb receptor.

Fig. 11. Molecular mechanisms of regulatory T (Treg) cells by retinal pigment epithelial

cells. PE: pigment epithelial cells, TCR: T cell receptor, CTLA: cytotoxic T lymphocyte-

associated antigen, TSP: thrombospondin, sTGFb: soluble transforming growth factor

b, TGFbR: transforming growth factorb receptor, CathL: cathepsin L.



12/18

3.2.1. Regulatory T cells generated by ocular PE cells

Ocular PE cells from the iris and retina are able to generate Treg

cells (Fig. 9) (Sugita et al., 2008), but the mode of action andmolecular mechanisms of these two types of ocular PE cells aredifferent. Our studies have revealed the molecular mechanismsresponsible for the generation of Treg cells by iris PE cells (Fig. 10)(Sugita et al., 2006b, 2006c; 2007, 2008; 2010a). Iris PE cells

primarily convert CD8 T cells, but not CD4 T cells, into Treg cells.

Iris PE cells provide negative signals to generate CD8 T cells viacell-contact-dependent mechanisms, including interactionsbetween costimulatory B7-2 (CD86) molecules on iris PE cells andCTLA-4 on CD8 T cells. TSP-1 expressed on iris PE cells converts

latent TGF-b to the active membrane-bound form. The binding ofactive TGF-b to its receptors on CD8 T cells signals these cells to

convert into Treg cells. The iris PE cell-exposed CD8 T cells, whichexpress CD25 and Foxp3, suppress the proliferation of activated

responder T cells and produce suppressive cytokines such as IL-10and TGF-b. The iris PE cell-exposed CD8 T cells express highlevels of co-stimulatory PD-L1 molecules on their cell surface andsuppress the activation of IFNg-producing Th1 cells that express

PD-1, indicating that negative PD-L1 signals are also required forthe induction of Treg cells by iris PE cells. Thus, iris PE cells have thecapacity to convert CD8 T cells, but not CD4 T cells, intoCD25Foxp3 Treg cells that suppress the proliferation of activated

T cells and produce suppressive cytokines.Unlike iris PE cells, retinal PE cells convert both CD4 T cells and

CD8 T cells into Treg cells, but the conversion occurs morefrequently in CD4 T cells than in CD8 T cells (Sugita et al., 2008).

In addition, retinal PE cells use soluble factors, but not cell-contactmechanisms, to generate Treg cells. Our studies have revealed themolecular mechanisms of Treg induction by retinal PE cells (Fig. 11)(Sugita et al., 2008,2009a,2009b;Sugita, 2009). The proliferation

of CD4 responder T cells was greatly suppressed by retinal PE cell-exposed CD4 T cells, while the proliferation of CD8 responder Tcells was weakly but signicantly suppressed by retinal PE cell-

exposed CD8 T cells. The induction of Treg cells by retinal PE

cells was not inuenced by a cell-culture insert membrane whenretinal PE cells and responder T cells were co-cultured, indicatingthat the generation of Treg cells by retinal PE cells is mediated by

soluble factors. Signaling for Treg generation is provided by inter-actions between TGF-breceptors on CD4 T cells and soluble TGF-

bsecreted by retinal PE cells. GeneChip analysis of iris, ciliary body,and retinal PE cells revealed that CTLA-2a, a cysteine protease

inhibitor (cathepsin L inhibitor), is a unique molecule that isconstitutively expressed on retinal PE cells but not on iris or ciliarybody PE cells (Sugita et al., 2008). Recombinant CTLA-2a sup-pressed the proliferation of responder T cells, and siRNA CTLA-2ablocked the ability of retinal PE cells to generate Treg cells (Sugita

Fig. 12.Molecular mechanisms of regulatory T (Treg) cells by corneal endothelial cells.

CE: corneal endothelial cells, TCR: T cell receptor, CTLA: cytotoxic T lymphocyte-

associated antigen, TSP: thrombospondin, mTGFb: membrane bound transforming

growth factor b, TGFbR: transforming growth factor b receptor, CathL: cathepsin L.

Fig. 13. Capacity of retinal pigment epithelial cells (RPE) to convert T cells into regulators. The gure and legends are reproduced (Horie et al., 2010, Fig. 1) with permission. (A)

Puried human CD4 T cells were cultured with supernatants from human RPE cells (ARPE-19) for 24 h in the presence of anti-CD3 antibody (0.1 mg/mL), harvested, X-irradiated,

and used as Treg cells (RPE-induced T reg cells: black bar). The h uman RPE-induced Treg cells were added to cultures containing naive responder T cells (T resp, open bar) plus anti-

CD3. (B) Puried murine CD4 T cells from a spleen were cultured with supernatants of primary cultured RPE cells for 24 h in the presence of anti-mouse CD3 antibody, harvested,

X-irradiated, and used as Treg cells. The murine RPE-induced Treg cells were added to cultures containing target responder T cells.



13/18

et al., 2008), indicating that CTLA-2ais an important soluble factorof retinal PE cells for the generation of Treg cells.

Recently, we showed that retinal PE cells produce retinoic acid,thereby enabling bystander T cells to be converted into Treg cellsthrough the production of TGF-b, a candidate immunosuppressivemolecule for Treg induction (Kawazoe et al., 2012). Indeed, the

conversion of T cells into Treg cells in the eye requires retinoic acid(Zhou et al., 2012). Similar to iris PE cell-induced Treg cells, retinalPE-induced Treg cells also express CD25high and Foxp3 (Fig. 11).

3.2.2. Regulatory T cells generated by CE cells

We have studied the capacity of human and murine CE cell linesto generate Treg cells and determined the corresponding molecularmechanisms (Fig. 12) (Sugita et al., 2009c, 2011a; Yamada et al.,

2010). Cultured human CE cells produce membrane-bound activeTGFb2 that is essential to suppress CD8 T-cell activation via directcell contact. Human CE cells also express co-stimulatory moleculessuch as PD-L1 and PD-L2 and secrete TSP-1, but only membrane-

bound TGFb2 is actually delivered to the CD8 T cells. Treg gener-ation by human CE cells was signicantly abolished by a cell-culture insert membrane, indicating that the conversion of CD8

T cells into Treg cells requires cell-to-cell contact. Anti-TGFb2 orTGFb2 siRNA blocked the ability of human CE cells to generate Treg

cells, indicating that Treg generation is mediated by membrane-bound TGFb2. The phenotype of human CE-generated Treg cells isCD25high and Foxp3, and these cells produce soluble forms ofTGFb1 but not TGFb2.

Like human CE cells, murine CE cells can induce Treg cells(Sugita et al., 2011a). Murine CE cells can convert both CD4 T cellsand CD8 T cells into Treg cells in vitro. CE-exposed CD4 T cellsexpress CD25 and intracellular Foxp3. Depletion of the CD25 cell

population from murine CE-exposed CD4 T cells blocked thesuppression of responder T cells. In the presence of a transwellmembrane, murine CE-exposed CD4 T cells did not express Foxp3and failed to suppress responder T cells, indicating that direct cell

contact is necessary for the induction of Treg cells. Both human andmurine CE cells express TGF-b, whichis important for the inductionof Treg cells. In addition, murine CE cells express CTLA-2a,a cysteine proteinase inhibitor, on their cell surface. Over-

expression of CTLA-2aon murine CE cells promotes TGF-bexpres-sion. CE-exposed Treg cells failed to suppress responder T cellswhen the murine CE cells were pretreated with CTLA-2asiRNA or

Fig.14. Capacity of RPE-induced Treg cells to suppress intraocular T cells. The gure is reproduced with permission (Horie et al., 2010, Fig. 5). Human RPE (ARPE-19) were pretreated

with recombinant TGFb2, and the supernatant of TGFb-pretreated RPE was harvested. Puried human CD4 T cells were cultured with the supernatants from TGFb-pretreated RPE

cells, X-irradiated, and used as RPE-induced human Treg cells. Target T-cell clones were established from in ltrating cells in the aqueous humor of patients with active uveitis:

sarcoidosis, Vogte

Koyanagie

Harada disease, acute anterior uveitis, Behet disease, acute retinal necrosis and cytomegalovirus retinitis.



14/18

antibody to CTLA-2a, indicating that murine CE cells produce CTLA-2aon their surface, thereby converting bystander CD4 T cells into

Treg cells by TGF-bpromotion.Our studies have shown that ocular resident cells have the

potential to inhibit T-cell activation and promote Treg-cell gener-ation. Ocular Treg conversion in vivowas conrmed by Caspi and

colleagues for the rst time (Zhou et al., 2012). They also showedthat the recognition of retinal Ag is required for Treg conversionand that retinoic acid is critical for the conversion as there is

a limited amount of TGF-b in the eye. Treg conversion is impaired

under inammatory conditions. Primed and non-converted T cellsdo not express effector functions in the eye. The contribution ofocular resident cells to the phenomenon in the eye needs to beelucidated to gain a more detailed understanding of tissue-based

mechanisms in ocular defense.

3.2.3. Regulatory T cells in humans

Our discussion of immune regulation by the direct immuno-

suppressive effects of ocular resident cells and their capacity togenerate Treg cells is based on studies performed with murine cells.It is important to determine if human ocular resident cells also have

the capacity to convert bystander T cells into Treg cells. For this

purpose, we determined if human retinal PE cells, ARPE-19 cells,could generate Treg cells in vitro (Horie et al., 2010). After incu-bation with anti-CD3 antibody, puried CD4 T cells exposed to

supernatants of human retinal PE cells (ARPE-19 cells) were har-vested, X-irradiated, and added to secondary cultures containingtarget responder T cells, which were established from allogeneic Tcells from the peripheral blood mononuclear cells of healthy

volunteers. Unlike the corresponding studies with murine cells,human CD4 T cells exposed to supernatants of human retinal PEcells did not suppress the activation of responder T cells (Fig. 13)(Horie et al., 2010). When retinal PE cells are treated with

recombinant TGFbin vitro, their production of immunosuppressivefactors such as TGFb and TSP-1 is greatly upregulated (Futagamiet al., 2007). Therefore, we pretreated ARPE-19 cells with

recombinant human TGFb2 (Horie et al., 2010). The TGFb2-pretreated human retinal PE cells produced large amounts ofimmunosuppressive factors (TGFb1, TGFb2, TSP-1, and PGE2) andsecreted them into the supernatant. Furthermore, we conrmedthat CD4 T cells exposed to supernatants from rTGFb2-pretreated

retinal PE cells secreted signicant amounts of the active form ofTGFb1, IL-10, or IFNg. The CD4 T cells exposed to supernatantsfrom rTGFb2-pretreated retinal PE cells suppressed the activationof pan T cells, B cells, CD4 T cells, and Th1 cells in vitro. More

importantly, the CD4 T cells exposed to supernatants from humanrTGFb2-pretreated retinal PE cells (RPE-induced human Treg cells)functionally suppressed the activation of CD4 T-cell clonesestablished from the ocular inltrating cells of patients with ocular

inammatory diseases (sarcoidosis, VKH disease, Behets disease,

acute anterior uveitis, acute retinal necrosis, and cytomegalovirusretinitis) (Fig. 14) (Horie et al., 2010). Thus, RPE-induced humanTreg cells can suppress activated T cells in the eyes of patients with

autoimmune uveitis as well as infectious uveitis. This capacity ofRPE-induced human Treg cells was completely blocked by pre-treating retinal PE cells with TGFb siRNA or anti-TGFbantibody. The

RPE-induced human Treg cells expressed Foxp3, CD25, and CD152(CTLA-4), and the CD4CD25 Treg cells greatly suppressedresponder bystander T cells, whereas CD4CD25 Treg cells did not.These in vitro assays indicate that supernatants of TGFb-exposed

human retinal PE cells can convert CD4 T cells into Treg cells,thereby suppressing T-cell activation.

The current investigations imply the possibility of using Tregcells established from humans for the treatment of intraocular

in

ammatory diseases with non-infectious etiology. However, the

safety and efcacy of human Treg cells in human diseases must rst

be comprehensively studied.

4. Conclusion

Uveitis is a sight-threatening disease caused by autoimmune orinfection-related immune responses. Studies of EAU and human

diseases such as Vogte

Koyanagie

Harada disease and HTLV-1uveitis imply that activated CD4 T cells, Th1 and Th17 cells, playan effector role in ocular inammation. The eye has a uniqueregional immune system that protects vision-related cells andtissues from these effector T cells. The immunological balance

between the pathogenic CD4 T cells and regional immune systemin the eye helps to maintain ocular homeostasis and good vision.Ocular resident cells at the inner surface of the ocularebloodbarrier (CE cells, iris PE cells, ciliary body PE cells, and retinal PE

cells) contribute to the regional immune system of the eye. Murineocular resident cells directly suppress the activation of bystander Tcells and the production of inammatory cytokines. The ocularresident cells possess distinct properties of immunoregulation that

are related to disparate anatomical locations. CE cells and iris PE

cells, which are located at the anterior segment of the eye and facethe aqueous humor, suppress T-cell activation via cell-to-cell

contact mechanisms, whereas retinal PE cells suppress T-cellsactivation via soluble factors. In addition to direct immunesuppression, the ocular resident cells have another unique immu-nosuppressive property, the induction of CD25Foxp3 Treg cellsthat suppress the activation of bystander T cells. Iris PE cells convertCD8 T cells into Treg cells, while retinal PE cells convert CD4 T

cells and, to a lesser extent, CD8 T cells into Treg cells. CE cellsconvert both CD4 T cells and CD8 T cells into Treg cells. Immu-nomodulation by ocular resident cells is mediated by soluble ormembrane-bound molecules including TGFb, TSP-1, B7-2 (CD86),

CTLA-2a, PD-L1 (B7-H1), galectin 1, pigment epithelial-derivedfactor, GITRL (glucocorticoid-induced tumor necrosis factorreceptor family-related protein ligand), and retinoic acid (Table 1).The Treg-cell induction properties of human and murine retinal PE

cells are similar. Although safety and efcacy must be addressed, itappears that human Treg cells induced by ocular resident cells suchas retinal PE cells and related immunosuppressive molecules arepromising for application in immune therapy for refractive auto-

immune uveitis in humans.

References

Agarwal, R.K., Caspi, R.R., 2004. Rodent models of experimental autoimmuneuveitis. Methods Mol. Med. 102, 395e419.

Akira, S., Takeda, K., 2004. Toll-like receptor signalling. Nat. Rev. Immunol. 4, 499e511.

Amadi-Obi, A., Yu, C.R., Liu, X., Mahdi, R.M., Clarke, G.L., Nussenblatt, R.B., Gery, I.,Lee, Y.S., Egwuagu, C.E., 2007. TH17 cells contribute to uveitis and scleritis and

are expanded by IL-2 and inhibited by IL-27/STATI. Nat. Med. 13, 711e

718.Anderson, M.S., Venanzi, E.S., Klein, L., Chen, Z., Berzins, S.P., Turley, S.J., von

Boehmer, H., Bronson, R., Dierich, A., Benoist, C., Mathis, D., 2002. Projection ofan immunological self shadow within the thymus by the aire protein. Science298, 1395e1401.

Appleman, L.J., Boussiotis, V.A., 2003. T cell anergy and costimulation. Immunol.Rev. 192, 161e180.

Asukata, Y., Ota, M., Meguro, A., Katsuyama, Y., Ishihara, M., Namba, K., Kitaichi, N.,Morimoto, S., Kaburaki, T., Ando, Y., Takenaka, S., Inoko, H., Ohno, S., Mizuki, N.,2009. Lack of association between toll-like receptor 4 gene polymorphisms andsarcoidosis-related uveitis in Japan. Mol. Vis. 15, 2673e2682.

Avichezer, D., Grajewski, R.S., Chan, C.C., Mattapallil, M.J., Silver, P.B., Raber, J.A.,Liou, G.I., Wiggert, B., Lewis, G.M., Donoso, L.A., Caspi, R.R., 2003. An immu-nologically privileged retinal antigen elicits tolerance: major role for centralselection mechanisms. J. Exp. Med. 198, 1665e1676.

Barnaba, V., Sinigaglia, F., 1997. Molecular mimicry and T cell-mediated autoim-mune disease. J. Exp. Med. 185, 1529e1531.

Bousso, P., 2008. T-cell activation by dendritic cells in the lymph node: lessons from

the movies. Nat. Rev. Immunol. 8, 675e

684.



15/18

Boxus, M., Twizere, J.C., Legros, S., Dewulf, J.F., Kettmann, R., Willems, L., 2008. TheHTLV-1 tax interactome. Retrovirology 14 (5), 76.

Broderick, C., Duncan, L., Taylor, N., Dick, A.D., 2000. IFN-{gamma} and LPS-mediated IL-10-dependent suppression of retinal microglial activation. Invest.Ophthalmol. Vis. Sci. 41, 2613e2622.

Broderick, C., Hoek, R.M., Forrester, J.V., Liversidge, J., Sedgwick, J.D., Dick, A.D.,2002. Constitutive retinal CD200 expression regulates resident microglia andactivation state of inammatory cells during experimental autoimmuneuveoretinitis. Am. J. Pathol. 161, 1669e1677.

Caspi, R.R., Roberge, F.G., Chan, C.C., Wiggert, B., Chader, G.J., Rozenszajn, L.A.,

Lando, Z., Nussenblatt, R.B., 1988. A new model of autoimmune disease.Experimental autoimmune uveoretinitis induced in mice with two differentretinal antigens. J. Immunol. 140, 1490e1495.

Caspi, R.R., Silver, P.B., Chan, C.C., Sun, B., Agarwal, R.K., Wells, J., Oddo, S., Fujino, Y.,Najaan, F., Wilder, R.L., 1996. Genetic susceptibility to experimental autoim-mune uveoretinitis in the rat is associated with an elevated Th 1 response.J. Immunol. 157, 2668e2675.

Caspi, R.R., 2010. A look at autoimmunity and inammation in the eye. J. Clin. Invest.120, 3073e3083.

Chang, J.H., McCluskey, P.J., Wakeeld, D., 2006. Toll-like receptors in ocularimmunity and the immunopathogenesis of inammatory eye disease. Br. J.Ophthalmol. 90, 103e108.

Chaon, B.C., Horai, R., Chen, J., Zrate-Blads, C., Villasmil, R., Chan, C.C., Caspi, R.R.,2012. The role of interleukin-17A in a spontaneous model of autoimmuneuveitis elicited by retina-specic T cells. Invest. Ophthalmol. Vis. Sci.. ARVO E-Abstract 6307.

Chen, J., Vistica, B.P., Takase, H., Ham, D.I., Fariss, R.N., Wawrousek, E.F., Chan, C.C.,DeMartino, J.A., Farber, J.M., Gery, I., 2004. A unique pattern of up- and down-regulation of chemokine receptor CXCR3 on inammation-inducing Th1 cells.Eur. J. Immunol. 34, 2885

e2894.

Cox, C.A., Shi, G., Yin, H., Vistica, B.P., Wawrousek, E.F., Chan, C.C., Gery, I., 20 08. BothTh1 and Th17 are immunopathogenic but differ in other key biological activi-ties. J. Immunol. 180, 7414e7422.

Crane, I.J., Xu, H., Wallace, C., Manivannan, A., Mack, M., Liversidge, J., Marquez, G.,Sharp, P.F., Forrester, J.V., 2006. Involvement of CCR5 in the passage of Th1-typecells across the blooderetina barrier in experimental autoimmune uveitis.J. Leukoc. Biol. 79, 435e443.

Cui, Y., Shao, H., Lan, C., Nian, H., OBrien, R.L., Born, W.K., Kaplan, H.J., Sun, D., 2009.Major role of gamma delta T cells in the generation of IL-17 uveitogenic T cells.J. Immunol. 183, 560e567.

De Kozak, Y., Usui, M., Faure, J.P., 1976. Experimental autoimmune uveoretinitis.Ultrastructure of chorioretinal lesions induced in guinea pigs by immunizationagainst the outer rods of the bovine retina. Arch. Ophthalmol. 36, 231e248.

De Kozak, Y., Thillaye, B., Renard, G., Faure, J.P., 1978. Hyperacute form of experi-mental autoimmune uveo-retinitis in Lewis rats; electron microscopic study.Albrecht Von Graefes Arch. Klin. Exp. Ophthalmol. 208, 135e142.

De Kozak, Y., Sakai, J., Thillaye, B., Faure, J.P., 1981. S antigen-induced experimental

autoimmune uveoretinitis in rats. Curr. Eye Res. 1, 327e

337.Dick, A.D., Ford, A.L., Forrester, J.V., Sedgwick, J.D., 1995. Flow cytometric identi -cation of a minority population of MHC class II positive cells in the normal ratretina distinct from CD45lowCD11b/c CD4 low parenchymal microglia. Br. J.Ophthalmol. 79, 834e840.

Dick, A.D., Siepmann, K., Dees, C., Duncan, L., Broderick, C., Liversidge, J.,Forrester, J.V., 1999. FaseFas ligand-mediated apoptosis within aqueous duringidiopathic acute anterior uveitis. Invest. Ophthalmol. Vis. Sci. 40, 2258e2267.

Dick, A.D., 2000. Immune mechanisms of uveitis: insights into disease pathogenesisand treatment. Int. Ophthalmol. Clin. 40, 1e18.

Dong, H., Zhu, G., Tamada, K., Chen, L., 1999. B7-H1, a third member of the B7 family,co-stimulates T-cell proliferation and interleukin-10 secretion. Nat. Med. 5,1365e1369.

DSouza, C.D., Cooper, A.M., Frank, A.A., Mazzaccaro, R.J., Bloom, B.R., Orme, I.M.,1997. An anti-inammatory role for gamma delta T lymphocytes in acquiredimmunity toMycobacterium tuberculosis. J. Immunol. 158, 1217e1221.

Dua, H.S., Lee, R.H., Lolley, R.N., Barrett, J.A., Abrams, M., Forrester, J.V., Donoso, L.A.,1992. Induction of experimental autoimmune uveitis by the retinal photore-ceptor cell protein, phosducin. Curr. Eye Res. 11 (Suppl.), 107e111.

Ego, T., Ariumi, Y., Shimotohno, K., 20 02. The interaction of HTLV-1 Tax with HDAC1negatively regulates the viral gene expression. Oncogene 21, 7241e7246.

Egwuagu, C.E., Charukamnoetkanok, P., Gery, I., 1997. Thymic expression of auto-antigens correlates with resistance to autoimmune disease. J. Immunol. 159,3109e3112.

Faure, J.P., de Kozak, Y., Dorey, C., Tuyen, V.V., 1977. Activite de differentes prepa-rations antigeniques de la retine dans linduction de luveo-retinite autoim-mune experimentale chez le cobaye. Arch. Ophthalmol. (Paris) 37, 47e60.

Fang, J., Fang, D., Silver, P.B., Wen, F., Li, B., Ren, X., Lin, Q., Caspi, R.R., Su, S.B., 2010.The role of TLR2, TRL3, TRL4, and TRL9 signaling in the pathogenesis of auto-immune disease in a retinal autoimmunity model. Invest. Ophthalmol. Vis. Sci.51, 3092e3099.

Ferguson, T.A., Grifth, T.S., 1997. A vision of cell death: insights into immuneprivilege. Immunol. Rev. 156, 167e184.

Ferguson, T.A., Grifth, T.S., 2006. A vision of cell death: Fas ligand and immuneprivilege 10 years later. Immunol. Rev. 213, 228e238.

Ferguson, T.A., Grifth, T.S., 2007. The role of Fas ligand and TNF-related apoptosis-inducing ligand (TRAIL) in the ocular immune response. Chem. Immunol.Allergy 92, 140e154.

Forrester, J.V., Huitinga, I., Lumsden, L., Dijkstra, C.D., 1998. Marrow-derived acti-vated macrophages are required during the effector phase of experimentalautoimmune uveoretinitis in rats. Curr. Eye Res. 17, 426e437.

Forrester, J.V., Xu, H., Lambe, T., Cornall, R., 2008. Immune privilege or privilegedimmunity? Mucosal Immunol. 1, 372e381.

Forrester, J.V., 2009. Privilege revisited: an evaluation of the eye s defence mecha-nisms. Eye 23, 756e766.

Fujimoto, C., Yu, C.R., Shi, G., Vistica, B.P., Wawrousek, E.F., Klinman, D.M., Chan, C.C.,Egwuagu, C.E.,Gery, I.,2006. Pertussistoxinis superior toTLR ligandsin enhancingpathogenic autoimmunity, targeted at a neo-self antigen, by triggering robust

expansionof Th1cellsand theircytokineproduction.J. Immunol.177, 6896e

6903.Fujimoto, C., Shi, G., Gery, I., 2008. Microbial products trigger autoimmune ocular

inammation. Ophthalmic Res. 40, 193e199.Futagami, Y., Sugita, S., Vega, J., Ishida, K., Takase, H., Maruyama, K., Aburatani, H.,

Mochizuki, M., 2007. Role of thrombospondin-1 in T cell response to ocularpigment epithelial cells. J. Immunol. 178, 6994e7005.

Gery, I., Wiggert, B., Redmond, T.M., Kuwabara, T., Crawford, M.A., Vistica, B.P.,Chader, G.J., 1990. Uveoretinitis and pinealitis induced by immunization withinterphotoreceptor retinoid-binding protein. Invest. Ophthalmol. Vis. Sci. 27,1296e1300.

Gery, I., Chanaud III, N.P., Anglade, E., 1994. Recoverin is highly uveitogenic in Lewisrats. Invest. Ophthalmol. Vis. Sci. 35, 3342e3345.

Gessain, A., Barin, F., Vernant, J.C., Gout, O., Maurs, L., Calender, A., de Th, G., 1985.Antibodies to human T-lymphotropic virus type I in patients with tropicalspastic paraparesis. Lancet II 40, 407e410.

Gregerson, D.S., Abrahams, I.W., 1983. Immunologic and biochemical properties ofseveral retinal proteins bound by antibodies in sera from animals with exper-imental autoimmune uveitis and uveitis patients. J. Immunol. 131, 259e264.

Gregerson, D.S., Sam, T.N., McPherson, S.W., 2004. The antigen-presenting activityof fresh, adult parenchymal microglia and perivascular cells from retina.J. Immunol. 172, 6587e6597.

Gregerson, D.S., Heuss, N.D., Lehmann, U., McPherson, S.W., 2009. Peripheralinduction of tolerance by retinal antigen expression. J. Immunol. 183, 814e822.

Gregory, M.S., Repp, A.C., Holhbaum, A.M., Saff, R.R., Marshak-Rothstein, A.,Ksander, B.R., 2002. Membrane Fas ligand activates innate immunity andterminates ocular immune privilege. J. Immunol. 169, 2727e2735.

Gregory, M.S., Koh, S., Huang, E., Saff, R.R., Marshak-Rothstein, A., Mukai, S.,Ksander, B.R., 2005. A novel treatment for ocular tumors using membrane FasLvesicles to activate innate immunity and terminate immune privilege. Invest.Ophthalmol. Vis. Sci. 46, 2495e2502.

Grajewski, R.S., Silver, P.B., Agarwal, R.K., Su, S.B., Chan, C.C., Liou, G.I., Caspi, R.R.,2006. Endogenous IRBP can be dispensable for generation of naturalCD4CD25 regulatory T cells that protect from IRBP-induced retinal auto-immunity. J. Exp. Med. 203, 851e856.

Grifth, T.S., Brunner, T., Fletcher, S.M., Green, D.R., Ferguson, T.A., 1995. Fas ligand-induced apoptosis as a mechanism of immune privilege. Science270,1158e1159.

Ham, D.I., Kim, S.J., Chen, J., Vistica, B.P., Fariss, R.N., Lee, R.S., Wawrousek, E.F.,

Takase, H., Yu, C.R., Egwuagu, C.E., Chan, C.C., Gery, I., 2004. Central immuno-tolerance in transgenic mice expressing a foreign antigen under control of therhodopsin promoter. Invest. Ophthalmol. Vis. Sci. 45, 857e862.

Harada, E., 1926. Beitrag zur klinishen Kenntniss von nichteitriger Choroiditis (Cho-roiditis diffusa acuta). Nippon Ganka Gakkai Zasshi 30, 356e378 (in Japanese).

Hayashi, D., Kubota, R., Takenouchi, N., Tanaka, Y., Hirano, R., Takashima, H.,Osame, M., Izumo, S., Arimura, K., 2008. Reduced Foxp3 expression withincreased cytomegalovirus-specic CTL in HTL V-I-associated myelopathy.J. Neuroimmunol. 200, 115e124.

Hinuma, Y., Nagata, K., Hanaoka, M., Nakai, M., Matsumoto, T., Kinoshita, K.,Shirakawa, S., Miyoshi, I., 1981. Adult T-cell leukemia: antigen in an ATL cell lineand detection of antibodies to the antigen in human sera. Proc. Natl. Acad. Sci.U. S. A. 78, 6476e6480.

Hori, J., Wang, M., Miyashita, M., 2006. B7-H1-induced apoptosis as a mechanism ofimmune privilege of corneal allografts. J. Immunol. 177, 5928e5935.

Hori, J., Taniguchi, H., Wang, M., Oshima, M., Azuma, M., 2010. GITR ligand-mediated local expansion of regulatory T cells and immune privilege ofcorneal allografts. Invest. Ophthalmol. Vis. Sci. 51, 6556e6565.

Horie, S., Sugita, S., Futagami, Y., Kawazoe, Y., Kamoi, K., Shirato, S., Mochizuki, M.,

2009. Human iris pigment epithelium suppresses activation of bystander T cellsvia TGF-T

Documents

Immunological Homeostasis of the Eye