Mechanisms of immune privilege in the anterior segment of the eye: what we learn from corneal transplantation

The cornea contains no blood or lymph vessels; therefore, considerable time passes after corneal transplantation before antigen recognition occurs in the regional lymph nodes and effector cells reach the graft. In addition, the corneal epithelial cells, keratocytes, and endothelial cells, do not express MHC class II molecules and also express only low levels of MHC class I molecules [5, 6]. This means that the main targets of a rejection reaction are not MHC antigens, but the minor H antigens in the corneal allografts [7]. When corneal endothelial cells were simultaneously exposed to interferon (IFN)-γ and tumor necrosis factor (TNF)-α, they expressed MHC class II molecules but not the class II transactivator gene or invariant chain [8]. In addition, MHC class II-positive dendritic cells and macrophages are absent in normal cornea, except in the peripheral region of the cornea [5]. There are small numbers of bone marrow-derived cells that still do not express MHC class II antigens in normal cornea, but controversy remains over whether they play the role of mature antigen-presenting cells (APCs), and it is thought that there are no donor-derived mature APCs in the donor cornea [9]. Therefore, the role of antigen presentation in the regional lymph nodes after corneal transplantation is carried out mainly by host-derived APCs rather than donor-derived APCs, and it is thought that recognition of donor antigens occurs in indirect fashion, by host CD4⁺ T cells via donor antigen-bearing host MHC class II molecules on host APCs [7, 10]. The main effector mechanism of rejection reactions in corneal transplantation is thought to be delayed-type hypersensitivity (DTH), rather than cytotoxic T cells (CTL), with the reaction targeting the donor minor H antigens rather than the donor MHC molecules [7, 10]. Thus, it can be surmised that the post-transplantation immune response is weaker in the case of corneal transplants compared with other organ grafts because of not only the anatomical characteristics of the corneal tissue but also its low antigenicity and the above-described mechanisms of antigen presentation and recognition.

Anterior chamber-associated immune deviation refers to a phenomenon in which antigen-specific systemic immunological tolerance is induced to an antigen that has been introduced to the anterior chamber in rodent eyes [1, 2]. ACAID is a phenomenon in which antibody responses are preserved while cellular responses such as DTH and CTL are suppressed. It is known that similar phenomenon occurs even in relation to an antigen that has been introduced the vitreous cavity or the subretinal space of the eye [1]. ACAID is induced in relation to various kinds of antigens, including allo-transplantation antigens, soluble protein antigens, viral antigens, and tumor antigens, and it has been demonstrated to be involved in various events such as acceptance of corneal transplants, autoimmune uveitis, acute retinal necrosis (ARN) in a fellow eye that experienced herpes virus infection in the anterior segment, or progression of intraocular malignant melanoma, in mouse models [1]. Induction of ACAID to autoantigens is reportedly induced by inoculating the anterior chamber with retinal autoantigens followed by a demonstration that such mice are subsequently resistant to the induction of experimental autoimmune uveitis by immunization with the same autoantigen [11]. In several areas, the findings from studies in the mouse model can be extrapolated to understanding the pathogenesis in human patients. For example, patients with ARN develop an ACAID-like response to viral antigens in the intraocular compartment that disappears as the disease resolves [12‐14].

The eye and the spleen are involved in the induction of ACAID (Fig. 1). Transforming growth factor (TGF)-β2 and thrombospondin (TSP)-1 in the anterior chamber are involved in the induction of APC mediators of ACAID, and eye-derived APCs such as macrophages that express F4/80 molecule pass across the trabecular meshwork, enter the bloodstream, and reach the spleen [15‐18]. The eye-derived APCs that reach the marginal zone in the spleen produce TGF-β, macrophage inflammatory protein 2, and CXC-chemokine ligand 2 and attract natural killer T (NKT) cells [19]. Then CD1d molecules that are expressed on the surface of the APCs bind with receptors that are expressed on the surface of the NKT cells, thereby presenting the antigens [20]. NKT cells produce TGF-β, IL-10, RANTES, CC-chemokine ligand 5, and TSP-1 [21, 22]. When marginal zone B cells are also present in this environment rich in immunomodulatory factors, clusters comprised of these three cell types form [23, 24]. When CD4+ and/or CD8+ T cells that are attracted to those clusters recognize the antigens being presented by the eye-derived APCs and the marginal zone B cells, they differentiate into ACAID-inducing regulatory T cells (ACAID-Treg). At that time, Qa-1 molecule on the surface of the marginal zone B cells function as antigen-presenting molecule. CD4+ ACAID-Treg inhibit the differentiation for Th1 cells in secondary lymph tissues such as lymph nodes, while CD8+ ACAID-Treg inhibit the function of effector T cells (Th1 and Th2) in the local site [25, 26]. It was recently reported that thymocytes and splenic γδ T cells are also necessary for induction of ACAID, and it can be understood that immune privilege in the eye is sustained through the cooperation of various cells from organs other than the eye itself (Fig. 1) [1, 2, 27].

Following corneal transplantation, ACAID is induced by the mechanism described above after the transplantation antigens on the endothelial surface of the cornea are taken up by the eye-derived APCs in the anterior chamber and transported to the spleen [1, 2]. Induction of ACAID leads to inhibition of the allo-antigen specific DTH and results in long-term survival of the graft [10, 28, 29]. ACAID cannot be induced in the case of infiltration of lymph vessels into the cornea, suturing, inflammation, trauma, or neurotomy in the cornea. In this situation, the eye is said to be at high risk of rejection, and in animal models of corneal transplantation, the rejection rate within 3 weeks after transplantation is 100% [3, 28]. Clinically, as well, rejection reactions readily occur in patients presenting this condition.

As the anatomical and cellular barriers, or ACAID, are not absolute, innate and adaptive immune cells and molecules can still access the eye. In response to threats to vision, the eye has soluble and cell surface immunomodulatory factors that act within the oculi to suppress cells and molecules that mediate innate and adaptive immune inflammation [1]. This intraocular milieu is called the immune suppressive microenvironment. The functions of the various cells and factors that manage natural immunity and acquired immunity are inhibited by the various factors that are expressed in the anterior segment and are shown in Table 1 [1, 2, 15, 16, 30‐39]. Among those factors, α-melanocyte-stimulating hormone, vasoactive intestinal peptide, calcitonin gene-related peptide, TGF-β2, and TSP-1 regulate the functions of macrophages and dendritic cells. TGF-β2 and TSP-1 are essential factors for the induction of ACAID as described in the previous section. As shown in Table 1, various immunomodulatory factors are expressed in corneal endothelial cells and iris–ciliary body. We recently elucidated that the inhibitory costimulatory signaling molecules of the B7 family are involved in immune suppression in the cornea. Those molecules are introduced below.

B7-H1 (PD-L1) was identified as a new B7 family molecule that binds to programmed death (PD)-1 on the surface of activated T cells and sends inhibitory signals to the T cells [40]. On the basis of the results generated in various animal models of autoimmune diseases, it has been reported that B7-H1 is involved in the induction of peripheral immune tolerance [40, 41]. In ocular tissues, B7-H1 is constitutively expressed in endothelial cells of the cornea, some stromal cells, iris–ciliary body, and the neural retina. The rejection reaction after corneal transplantation is intensified by blockade of B7-H1 or PD-1 with antibodies [39]. B7-H1 expressed in the cornea induces apoptosis of PD-1-expressing T cells, and deletion of effector T cells in the cornea results in inhibition in the effector phase of the rejection reaction [39]. It is highly interesting that the T-cell apoptosis mediated by B7-H1 has been only observed in immune privileged tissues or sites such as tumors, liver, and cornea so far [41, 42]. Corneal endothelial cells also constitutively express Fas ligand, and apoptosis of effector T cells is induced via Fas [35‐37]. It is unclear whether, in this state of the effector T cells having been eliminated from the eye, B7-H1 and Fas ligand interact on the surface of the corneal endothelial cells (Fig. 2). On the other hand, the previously described ACAID is induced even if the PD-1/B7-H1 pathway is blocked. That is, the induction of ACAID-Treg by the eye-derived APCs and spleen cells is not dependent on the PD-1/B7-H1 pathway.

The culture system of corneal tissue and T cells in vitro has been established and shown that B7-H1 expressed in the cornea shows local immunosuppressive activity [39]. This system permits complete elimination of any involvement of the secondary lymphatic organs and makes it possible to isolate and analyze only the effector phase of the rejection reaction when the corneal endothelial cells has been damaged by effector T cells. The results showed that B7-H1 expressed in the corneal cells not only inhibited corneal endothelial damage by allo-reactive T cells but also inhibited bystander damage caused by activated T cells that are specific to third party antigen. In addition, PD-1 on the surface of the T cells was up-regulated as a result of contact with the corneal cells, thus accelerating apoptosis. Moreover, B7-H1 expression on epithelial cells was induced in the presence of inflammatory cytokine such as IFN-γ, and T-cell apoptosis mediated by B7-H1 and PD-1 was induced in each of the epithelium, stroma, and endothelium of the cornea. (Figure 3).

As described above, the PD-1/B7-H1 pathway is more involved in interactions between the effector T cells and the corneal cells within the eye than in the immune responses in the secondary lymphatic organs [39]. Thus, these molecules contribute to maintenance of the local immune suppressive microenvironment in the eye.

B7-H3 was recently identified as a new B7 family molecule, but its receptor has not yet been identified [43]. There have been conflicting reports, based on different experimental systems, as to whether B7-H3 is a costimulatory signal molecule that sends positive signals or negative signals to T cells [43]. This is currently a source of controversy. In the eye, B7-H3 is constitutively expressed in corneal endothelial cells and the iris–ciliary body, and that a rejection is induced after corneal transplantation in experimental hosts to which anti-B7-H3 blocking antibody had been intraperitoneally administered [44]. It is of interest that ACAID is not induced in such animals administered with anti-B7-H3 antibody. That is, B7-H3 is involved in induction of ACAID-Treg that are dependent on the spleen and mediated by eye-derived antigen-presenting cells, and it thus plays a different role from that of B7-H1 [44]. Also, the possibility that B7-H3, like the previously described B7-H1, plays a role in eliminating or inhibiting effector immune cells within the cornea remains to be established.