Expression of MHC molecules by parenchymal cells of the target organ in autoimmune disease decreases the detrimental effect of invading autoreactive T cells

Introduction

Autoreactive T cells are found in healthy people ¹ and non-immunized animals ²; however, most people do not develop autoimmune disease and the induction of autoimmune disease in animals requires specific treatment regimes and the use of particular genetic strains. It appears that autoimmune reactions are normally suppressed and that autoimmune disease results from breakdown of this suppression. Thus, a complete understanding of the pathogenesis of autoimmune uveitis requires knowledge of the suppressive mechanisms that are normally operative, but fail during disease. Additionally, pathogenic T cells may have attributes that are lacking in their non-pathogenic counterparts. One such protective mechanism is the ability of parenchymal cells of the target organ to inhibit autoreactive T cells. Indeed, "intrinsic abnormality of the target organs" has been previously proposed to explain the mechanism(s) by which autoreactive T cells mediate diseases ^3;4.

Experimental autoimmune uveitis (EAU), an autoimmune disease induced in experimental animals ^5–8, has been, for many years, a popular laboratory model of human uveitis and even, to some extent, other T cell-mediated autoimmune diseases. This highly reproducible animal model offers a variety of approaches for studying the etiology and pathogenesis of human autoimmune disease. EAU can be induced either by immunization of susceptible strains of rodents with a defined autoantigen or by adoptive transfer of autoreactive autoantigen-specific T cells.

Active [ImmunizationOK??] EAU

By definition, active EAU is initiated by injection of ocular antigen in an immunogenic form, usually as an emulsion in complete Freund's adjuvant. This elicits a peripheral immune response, and, in susceptible animals, ocular inflammation of the eye. Generally, symptoms of uveitis appear by the 10–13th day post-immunization, persist for a little over a week, then subside. Antigen is taken up and processed into smaller peptide fragments that become complexed to major histocompatibility complex (MHC) molecules expressed on the surface of antigen presenting cells (APCs). For example, when interphotoreceptor retinoid-binding protein (IRBP) is used as the eliciting antigen in the C57BL/6 mouse, a large portion of the immune response is directed against the 20-mer peptide containing residues 1–20 of the protein (GPTHLFQPSLVLDMAKVLLD).

Adoptive Transfer of EAU

Somewhat simpler than active EAU, this begins with the transfer of lymphocytes from already immunized donors to recipients ^9;10. Thus, the initial stages of immunization, including adjuvant effects and the activation of disease-causing T cell subsets, do not occur in the recipients. T cells prepared from the lymph nodes of animals undergoing active EAU are re-stimulated in culture and adoptively transferred to syngeneic animals in which they cause inflammation in the eye and consequent tissue damage. Disease onset is more rapid than in active EA, beginning at day 4 in rats ^5;11 and day 8 in mice ^9;10. Adoptive transfer to a naïve animal of a few million newly activated syngeneic autoreactive T cells can readily induce disease, suggesting the pathogenic role of autoreactive T cells in disease. The mechanism by which an organ-specific autoimmune disease can be adoptively transferred by a few million autoreactive T cells, of which only a fraction enters the autoimmune organ, remains unclear. It is hypothesized that the entry of the pathogenic T cells provokes MHC expression on parenchymal cells and release of chemoattractant factors, which, in turn, recruit inflammatory cells.

Autoreactive T cell lines and clones

To characterize the mechanism by which autoreactive T cells initiate autoimmune disease and to determine the various structural and functional features that distinguish between subsets of autoreactive T cells and other antigen specific, non-pathogenic T cells, enriched T cell populations have been prepared to determine the requirements for T cell activation ¹¹ and the usage of the T cell antigen receptor (TCR) ^12–14 and accessory molecules ^15;16 and to assess the various cell-interacting cytokines produced by these cells ^17–19. The need for the characterization of the structure and function of autoreactive T cells by isolating antigen-specific T cell lines and clones became especially apparent when it is realized that limiting dilution analysis indicates that the number of autoreactive T cells in immunized rodents or in humans suffering from multiple sclerosis rarely exceeds one in ten thousand T cells ^20;21 and that the overwhelming majority of activated T cells associated with disease development are nonspecifically expanded ²².

The availability of established autoreactive T cell lines and clones and cultured parenchymal cells, such as astrocytes and retinal pigment epithelium (RPE) cells, allows us to establish an in vitro system to examine possible in vivo interactions between the two cells. Extensive laboratory studies have provided convincing evidence that the key pathogenic events include: i) the activation and expansion of the pathogenic T cells, ii) the homing of autoreactive T cells to the specific target organ, and iii) the reactivation of autoreactive T cells in the autoimmune organ and their interaction with the parenchymal cells there.

Tissue damage is caused by activated autoreactive T cells

Even in transgenic mice in which the majority of T cells express the same autoantigen-recognizing TCR, the incidence of spontaneous autoimmunity is very low. Only newly activated autoreactive T cells can adoptively transfer autoimmune disease, as it is not provoked by adoptive transfer of 100 times as many of the same T cells not subjected to an additional antigenic stimulation before administration.

Trafficking of activated T cells via the blood through the vascular endothelium and the blood eye barrier (BBB) into the eye

Like the central nervous system (CNS), the vertebrate eye was thought to be inaccessible to lymphocytes and thus exempt from immunologic threat, the blood-brain-barrier (BBB) or blood-eye-barrier (BEB) serving to protect their integrity (6). However, once activated, antigen-specific T cells can penetrate these barriers ^23;24. Although the mechanism by which these T cells cause disease is unknown, the pathogenic T cells can interact with parenchymal cells expressing the appropriate MHC molecules ^17;25. Most studies suggest that penetrating ability depends on the activation status, rather than the antigen specificity, of the T cell ²³. In addition, the entry of autoreactive T cells into the CNS is influenced by the accessory/adhesion molecules it expresses ^26;27. It remains to be determined whether pathogenic T cells have an increased ability to penetrate the BBB or BEB or are more capable of escaping apoptosis in the CNS than nonpathogenic T cells.

Possible mechanisms by which autoreactive T cells cause tissue damage

A number of hypotheses have been proposed, e.g., pathogenic T cells may destroy their target tissue directly by a cytotoxic effect ^28–30 or indirectly by releasing inflammatory cytokines ^31–35. Given that T cell recognition is restricted by MHC molecules and the dominant cell population able to express MHC molecules in the CNS is glial cells, the ability of autoreactive T cells to interact with glial cells and the outcome of the T cell-glial cell interaction may be highly relevant to the pathogenesis of this disease.

Expression of MHC molecules by parenchymal cells in the autoimmune organ

One of the seminal findings of immunology is that the recognition of T cells is strictly restricted by antigens encoded by the MHC ³⁶. Based on this dogma, it is believed that only those parenchymal cells in the autoimmune organ capable of expressing MHC molecules can directly interact with the invading autoreactive T cells.

Under physiological conditions, parenchymal cells of the autoimmune organ do not express appreciably levels of MHC molecules ³⁷, but do so during disease ³⁸ or when they are activated in vitro by pro-inflammatory cytokines, such as IFN-γ³⁹. Aberrant appearance of MHC molecules by parenchyma cells of the autoimmune organ, such as astrocytes in the CNS and RPE cells in the eye, is therefore believed to alter disease susceptibility ^40;41. These cells, which can be induced to express MHC molecules and thus support T cell activation ^28;42 in the CNS ^3;42, could be likely candidates for auto-attack ²⁸.

The role of MHC molecules expressed by parenchymal cells of the autoimmune organ

The questions arise why parenchymal cells of the autoimmune organ retain the ability to express MHC molecules, but only do so during disease, and whether the appearance of MHC molecules in the autoimmune organ augments the activation of invading autoreactive T cells and thus exacerbates disease or, alternatively, restricts the intensification of local inflammation. It is speculated that, like many biological responses, this is a two-edge sword; whereas the expression of modest levels of MHC molecules inhibits the activation of invading T cells, over-expression of these molecules promotes activation.

Extensive studies have been performed in laboratories, including ours, on one of the major parenchymal cells in the CNS g the astrocyte. Under physiological conditions, astrocytes do not express appreciable levels of MHC molecules, but can be readily induced to do so by exposure to pro-inflammatory cytokines, such as IFN-γ Immunologists are particularly keen to link the MHC molecule with T cell activation. Thus, in early studies, researchers tried to find evidence to support their hypothesis that the expression of MHC molecules might render astrocytes able to stimulate autoreactive T cells ⁴².

Our recent studies have compared a series of immune functions in RPE and astrocytes. We have found that these two cells are extremely alike in terms of their ability to interact with autoreactive T cells.

RPE cells contribute to the immune privileged status of the eye as part of the BEB ⁴³ by secretion of immunosuppressive factors inside the eye ^44–47 and by expression of Fas-ligand (FasL) on their cell surface ^48–50. RPE cells may also assist in the development of intraocular inflammation ^46;51–53 and can respond to a variety of inflammatory cytokines ^51;54;54 and produce a myriad of molecules that can induce inflammation. For example, RPE cells can produce cytokines, such as TNF-α, IL-15 ⁵⁵, and nitric oxide (NO) ⁵⁶, and express cell surface MHC molecules and costimulatory molecules ^57;58. In addition, RPE cells also express a number of uveitogenic antigens, such as soluble retinal antigen (S-antigen) and IRBP, and could therefore become targets for uveitogenic T cells, leading to autoimmune reactions in the eye. While there is strong evidence that RPE cells can express MHC II molecules after activation ^58;59, the role of these molecules in the eye remains unclear.

Our studies have shown that, depending on their state of activation, RPE cells can either inhibit or activate IRBP-specific T cells. In contrast to peripheral APCs, which elicit full activation (proliferation and cytokine release) of autoreactive T cells, RPE cells elicit only partial activation ( $\text{TNF}-\underset{.}{\alpha }$ and IFÑ-γ production, but not proliferation) ⁶⁰.

MHC-expressing astrocytes and RPEs partially activate autoreactive T cells and drive these T cells into a refractory phase

The mere presence of MHC molecules and appropriate antigen is not sufficient to induce T cell activation and the presence of costimulatory molecules on MHC II-expressing cells is crucial ^61;62. MHC II-expressing cells that lack accessory molecules may not only fail to function effectively as APCs, but can also result in unresponsiveness of T cells ⁶². Conceivably, too low a density of accessory molecules on astrocytes/RPE cells may result in inhibition of T cell activation. We are currently investigating this possibility.

The availablity of MHC molecules in the autoimmune organ may cause the invading T cell to be activated. However, T cell biology studies tell us that two biological features of T cells have a closer relationship with the pathogenic activity of the autoreactive T cell. Firstly, the T cell can be activated to various degrees ^63–66. So-called “partial activation” means that the T cells are activated, but only some of the activation-related T cell functions are turned on. Given that the damaging effect of autoreactive T cells is more closely correlated to the degree of activation than the number of T cells, partially activated T cells have only limited pathogenic activity, possibly because they produce lower than pathogenic amount of damaging factors and are less cytotoxic. Secondly and more importantly, both fully and partially activated T cells enter a refractory phase. T cells are cycling cells and, once activated, can only been re-activated after a lag-period. For both rat and mouse T cells, the duration of this cycle is approximately 5–7 days. Thus, immediately after entry into the autoimmune organ or before massive inflammation has been initiated, the expression of MHC molecules allows the parenchymal cell to interact with the invading T cells. This interaction renders the invading T cells partially activated and they then enter a refractory phase; as a result, when professional APCs arrive at the peak of the inflammation, the refractory T cells cannot be reactivated. Such a consequence is clearly OK?? beneficial, since, if they remained non-activated, the invading T cells would undergo much stronger activation following arrival of the infiltrating inflammatory cells containing a large number of professional APCs. In this sense, the ability to express MHC molecules gives the parenchymal cell a protective capability, restricting the intensity of inflammation. This assumption has been tested in in vitro assays. Thus, we have examined whether the interaction of T cells and astrocytes affects T cell responsiveness to subsequent antigenic challenge by first treating T cells with autoantigen in the presence or absence of astrocytes/RPE, then assessing their response to professional APCs. The results showed that pretreatment of T cells with astrocytes or RPE greatly decreases the ability of the T cells to respond to subsequent antigenic challenge ^18;60.

This assumption is also supported by the results of an in vitro experiment comparing the antigen-presenting activity of IFN-γ-activated astrocytes with that of professional APCs. We have observed that, although astrocytes can activate autoreactive T cells, they are only 5–10% as effective as professional APCs ¹⁷. More importantly, T cells exposed to astrocytes expressing maximal levels of MHC molecules produce only part of their cytokine repertoire compared to the same T cells stimulated by professional APCs. These observations led to the conclusion that, unlike professional APCs, MHC-expressing activated astrocytes can evoke only some of the functional properties of a T cell population ¹⁷.

Tissue damage provoked by invasion of autoreactive T cells appears to involve cascading responses in which the generation of cytokines and the recruitment of inflammatory cells reciprocally stimulate each other. ClearlyOK??, regulatory mechanisms are needed to control the intensity of inflammation and avoid tissue damage. It is hypothesized that the entry of autoreactive T cells elicits the release of cytokines or chemokines, which then cause massive infiltration of inflammatory cells. Among the infiltrating cells are tissue-damaging cells, such as NK cells and macrophages, and others with antigen-presenting activity, such as dendritic cells and macrophages, resulting in further activation of the invading autoreactive T cells by the infiltrating dendritic cells, leading to augmented infiltration and cascading response.

Our studies have shown that MHC molecule expression by parenchymal cells of the autoimmune organ plays a regulatory role in autoreactive T cell activation and thus inflammation formation and tissue damage. This is because the activation of autoreactive T cells by the parenchymal MHC-expressing cells is only “partial” and the production of pro-inflammatory cytokines is lower than pathogenic levels; furthermore, this pre-activation renders the invading T cell refractory when potent professional APCs become available during the later event of inflammation. In short, the expression of MHC molecules by the parenchymal cell of the autoimmune organ induces the invading T cell to become anergic after producing limited amounts of pro-inflammatory cytokines. It is also likely that the expression of MHC class I molecules protects glial cells from NK cell cytotoxic effects, as MHC-negative target cells are more vulnerable to cytolysis by NK cells ⁶⁷. Indeed, studies have shown that, among the cells infiltrating during inflammation, a significant proportion possess an NK-like phenotype and cytotoxic activity ^68;69.

Thus, the expression of MHC class II molecules by parenchymal cell of the autoimmune organ is probably more beneficial than detrimental to the host in terms of preventing the full activation and expansion of potentially pathogenic T cells. Nevertheless, the production of excessive amounts of cytokines by T cells that are partially activated may also facilitate disease progression.

Our studies using a number of in vitro assay systems have also shown that RPE cells and astrocytes are functionally very much alike, in that they express MHC molecules upon exposure to IFN-γand stimulate autoreactive T cells to release TNF-α⁶⁰.

Reciprocal interaction between autoreactive T cells and parenchymal cells of the eye

We have previously shown that autoreactive T cells vary greatly in disease-inducing capacity ^70;71. Unfortunately, this is not always reflected by differences in the fine specificity of the cell response or the cytokine-producing pattern of the maximally activated T cells. Because of this, researchers are searching for other cellular and molecular features showing a better correlation with the pathogenic nature of the cells. For example, studies in our laboratory have shown that the degree to which an autoreactive T cell is activated and its pattern of cytokine production are not innate to the cell and are not solely determined by the type of TCR ligand that induces T cell activation, as the source of the APCs and the dose of antigen available are also important ⁷². Given that the major MHC-expressing cells, such as glial cells in the CNS and astrocytes and RPE cells in the eye, may differ from professional APCs in the periphery in terms of antigen processing or accessory molecule expression, it is of interest to know whether activation of autoreactive T cells inside the autoimmune organ differs from T cell activation in the periphery, particularly in the presence of suboptimal doses of antigen, assuming that optimal in vitro doses would not always be available in vivo. It is likely, for example, that autoreactive T cell subsets capable of responding to limited antigen doses may pose a greater threat in vivo than other T cells with the same antigenic specificity, but activated only by larger doses of antigen.

Autoreactive T cell subsets differ greatly in their ability to interact with parenchymal cells ^73;74. This finding appears to be consistent with the previous observation that not all MBP-reactive T cells produce a similar degree of tissue damage in the CNS ^29;30. It remains to be determined whether the ability of T cells to interactive with parenchymal cells of the autoimmune organ correlates with their pathogenic activity and whether pathogenic T cells have an enhanced or decreased ability to interact with parenchyma cell of the organ.

Summary

Although the physiological role of the expression of MHC molecules by parenchymal cells of the autoimmune organ is still poorly understood, it seems implausible that this expression of MHC molecules during the genesis of autoimmune disease favors the reactivation of invading autoreactive T cell. Since disease is always preceded by inflammation of the diseased organ, a reaction that recruits large numbers of peripheral APCs which may cause stronger activation of the invading T cell, it is assumed that expression of limited amounts of MHC molecules by parenchymal cells should render the invading autoreactive T cells unresponsive to the infiltrating APCs by promoting their entry into a refractory phase of the cell cycle.

By their ability to express variable amounts of MHC molecules, parenchymal cells of the autoimmune organ, such as astrocytes and RPE cells, have the ability to control the degree of T cell activation in the organ. Thus, T cells entering the autoimmune organ and in contact with cells expressing low levels of MHC class II molecules downregulate their TCR and become minimally activated and hyporesponsive. These observations support the premise that the primary role of MHC-expressing cells in the autoimmune organ is to diminish or block full T cell activation in the organ and thereby prevent the release of harmful cytokines. It is possible that, during massive T cell infiltration or infection, the local cells then express higher level of MHC molecules, promoting greater T cell activation, with the accompanying production of proinflammatory cytokines.

Circumstantial evidence indicates that interactions between autoreactive T cells and the parenchymal cells of the autoimmune organ are highly versatile. For example, only activated encephalitogenic T cells are able to penetrate the BBB ²³ and cause EAE ²⁸. The levels of expressed MHC antigens on parenchymal cells determine not only the activation of invading autoreactive T cells, but also the survival of these parenchymal cells faced with the cytolytic activity of autoaggressive T cells ²⁸. In addition, the availability of T cell-specific antigen is critical for the cell-cell interaction ²⁸ and for the persistence of the invading T cell in the organ ^24;75. Further studies should provide a better understanding of disease pathogenesis.