Recent studies revealing active mechanisms of immune privilege in neural tissues have diminished the putative role of passive tolerance. To examine the significance of Ag localization in the retina on immune privilege, the immune responses of transgenic mice expressing high and low levels of β-galactosidase (β-gal) in the photoreceptor cells of the retina were compared with those of normal mice and those of mice expressing moderate levels of β-gal systemically. Immunization with β-gal induced experimental autoimmune uveoretinitis indistinguishable from that induced by known photoreceptor cell autoantigens, including destruction of photoreceptor cells, in transgenic mice with high level retinal expression. Retinal expression had no apparent effect on the immune responses to β-gal, showing that tolerance was not elicited by levels of retinal β-gal sufficient to serve as a target for autoimmune disease. Mice with systemic expression exhibited reduced lymphoproliferative responses following immunization with β-gal and did not develop autoimmune disease. T cells prepared from normal mice immunized with β-gal transferred experimental autoimmune uveoretinitis to the transgenic mice with high level retinal β-gal expression, but no disease was found in mice with systemic transgene expression under these conditions. The results of our experiments are most consistent with sequestration being the primary mechanism of retinal immune privilege. The results also show that β-gal can serve as an immunopathogenic neural autoantigen, and that T cells raised by immunization of normal mice with a foreign Ag can be immunopathogenic in certain transgenic recipients.

Sequestration is the passive immune tolerance attributable to the localization of Ags behind physiologic and anatomic barriers, especially in the nervous system (1). Sequestration was postulated to reduce lymphocytic perusal of the tissue and minimize leakage of Ags from the tissue, resulting in a lack of tolerance. Current studies of immune privilege in neural tissues such as retina or brain focus on elucidation of active mechanisms, and the significance of sequestration is now questioned. In the eye, these mechanisms of retinal immune privilege include Fas ligand expression (2), limited class II MHC expression (3), and regulatory T cells (4). In addition, evidence for lymphatic drainage of the eye (5) and brain (6) has been found, further arguing against sequestration as an important component of immune privilege.

Experimental evidence for active tolerance has been reported in the experimental autoimmune encephalomyelitis (EAE)3 model as well. For example, autologous myelin basic protein (MBP) is less encephalitogenic than heterologous MBP (7), suggesting that there is active tolerance to MBP, and that it is broken by cross-reactive heterologous Ags. Conversely, evidence of immunologic ignorance in the CNS was revealed by studies of mice carrying a transgenic (Tg) TCR specific for MBP (8). Another mechanism of tolerance was revealed by reports of thymic mRNA expression of various autoantigens, including retinal rod photoreceptor cell arrestin (i.e., S-Ag) and MBP (9, 10). Arrestin, an immunopathogenic autoantigen that mediates experimental autoimmune uveoretinitis (EAU) (11), is a member of a large family of proteins with significant sequence homology (12). Tolerance to widely expressed arrestin family members could contribute to tolerance of rod photoreceptor cell arrestin.

Consequently, there is clear need to examine a neural Ag in a less complicated system (13). Tg mice expressing foreign Ags are frequently used to study tolerance and autoimmunity (reviewed in Refs. 14, 15, 16). Testing the influence of the retinal microenvironment on tolerance is possible using Tg mice whose expression of a neo-self-Ag is directed to the retina and elsewhere. We have chosen Tg mice that express β-galactosidase (β-gal) in retinal photoreceptor cells, and asked whether EAU could be mediated by an immune response to the β-gal transgene. An inflammatory, tissue-destructive autoimmune response was found in the Tg mice with high level retinal expression, and there was no evidence for tolerance resulting from that level of retinal expression. Control Tg mice that express β-gal systemically and normal non-Tg mice were included in the studies for comparison. Our results extend previous studies by demonstrating immunopathology mediated by CD4 T cells specific for a foreign transgene expressed in a neural tissue.

The rho-β-gal mice were derived from C57BL/6 2174-31 founders expressing β-gal under the control of a rhodopsin promoter (17). Expression is first seen 7–8 days postnatally and increases linearly with age. The lo-arr-β-gal mice (low level retinal expression) and hi-arr-β-gal mice (high level retinal expression) were derived from CD-1 mice expressing β-gal using two different forms of the arrestin promoter (18, 19, 20). Col-β-gal mice were provided by Dr. B. deCrombrugghe (M. D. Anderson Cancer Center, Houston, TX) as a B6D2F1 mouse expressing β-gal using the −2000 murine α2(I) collagen promoter (21). Col-β-gal mice express modest levels of β-gal in several tissues, including the retina. All Tg mice were backcrossed onto the B10.A background, providing the MHC type (I-Ak) known for immune reactivity to β-gal (22), and the EAU-permissive B10 background (23). The rho-β-gal mice were used after 5–6 backcrosses, the hi-arr-β-gal mice were used after 5–10 backcrosses, and the lo-arr-β-gal and the col-β-gal mice were used after 9–11 backcrosses. B10.A mice purchased from Charles River (Wilmington, MA) and transgene-negative littermates were also used. Mice were maintained under specific pathogen-free conditions on lactose-free chow.

Mice were immunized s.c. with 50–100 μg of β-gal in CFA supplemented with 2.5 mg/ml Mycobacterium tuberculosis H37Ra. One microgram of pertussis toxin (Sigma, St. Louis, MO) was administered i.p. on the same day. Purified β-gal was purchased from Prozyme (San Leandro, CA). A set of 169 synthetic peptides corresponding to the entire length of β-gal, each 17 residues long and offset by seven residues, was purchased from Chiron Technologies (San Diego, CA). To reduce the number of samples required to screen lymph node (LN) cells for proliferative responses, 34 pools containing five peptides each were made (see example in Table III). A synthetic peptide corresponding to residues 461–480 (LRHNPGGPSSAMPLVLSYFQ) of murine interphotoreceptor retinoid-binding protein (IRBP) was used in some experiments (24). IRBP is currently the most well-characterized immunopathogenic retinal autoantigen used in murine EAU studies (25).

Table III.

Cytokine production by Ag-stimulated, primed LN cellsa

MiceCytokines in Culture Supernatantb
IFN-γIL-2IL-4IL-6TNF-α
B10.A 107 ± 24.3 0.32 ± 0.01 BDc 1.4 ± 0.09 4.2 ± 0.2 
hi-arr-β-gal 16.3 ± 3.3 BDc BDc 0.43 ± 0.14 3.9 ± 1.3 
col-β-gal 30.8 ± 12.9 0.24 ± 0.07 BDc 0.58 ± 0.11 2.4 ± 1.4 
Naive/Con Ae 109 ± 4 20.1 ± 1.7 BDd 0.9 ± 0.7 5.1 ± 0.4 
MiceCytokines in Culture Supernatantb
IFN-γIL-2IL-4IL-6TNF-α
B10.A 107 ± 24.3 0.32 ± 0.01 BDc 1.4 ± 0.09 4.2 ± 0.2 
hi-arr-β-gal 16.3 ± 3.3 BDc BDc 0.43 ± 0.14 3.9 ± 1.3 
col-β-gal 30.8 ± 12.9 0.24 ± 0.07 BDc 0.58 ± 0.11 2.4 ± 1.4 
Naive/Con Ae 109 ± 4 20.1 ± 1.7 BDd 0.9 ± 0.7 5.1 ± 0.4 
a

Representative of two experiments.

b

Data given in ng/ml; mean ± SD; determined by ELISA.

c

BD, below detection (<180 pg/ml IL-2; <95 pg/ml IL-4).

d

BD, below detection (<24 pg/ml IL-4).

e

Con A-activated B10.A LN cells.

Nine days after priming, draining LNs from several mice were pooled and stimulated with 75 μg/ml β-gal in T-25 flasks in RPMI 1640 with 10% FCS. The T cells were restimulated at 9- to 10-day intervals with irradiated splenocytes and Ag to make T cell lines. No exogenous IL-2 was used to develop the MT-1 line. Clones were derived from the MT-1 line by limiting dilution. A line specific for the 461–480 peptide of IRBP was similarly made in B10 mice primed with the 461–480 peptide. Peptides of β-gal or IRBP were used at 2 μM to stimulate some T cell cultures as specified. Adoptive transfer of EAU was performed with either β-gal-specific or IRBP461–480-specific, T cells collected 2 days postactivation with Ag and splenic APC, and transferred i.p. in 0.5 ml of saline. For proliferation assays, draining LN cells were prepared and dispensed into 96-well plates at 5 × 105 cells/well in RPMI 1640 with 10% FCS. Ag was added as indicated. After 3 days of culture, 1 μCi of [3H]TdR was added, and the cultures were harvested 18 h later. Assays of line cells and clones were performed with 6 × 105 irradiated splenocytes and 4–6 × 104 T cells/well. [3H]TdR was added after 48 h.

The cytokine content of culture supernatants taken from assays as described above was determined by ELISA according to the manufacturer’s suggestions (PharMingen, San Diego, CA). The Ab pairs and the murine cytokine standards were purchased from PharMingen.

Serial dilutions of sera were added to ELISA plates coated with 500 ng/well of β-gal, washed, and then incubated with alkaline phosphatase-conjugated anti-mouse IgG (Sigma). After washing, the wells were incubated with p-nitrophenyl phosphate substrate. Absorbance was measured at 405 nm.

Eyes were fixed overnight in 10% buffered formalin, paraffin embedded, sectioned (5 microns), and stained with hematoxylin and eosin. The slides were examined in a masked fashion according to previously reported criteria (26).

Minor modifications were made to a published protocol (27). Cryostat sections of OCT-embedded tissue were fixed in PBS containing 2% paraformaldehyde and 0.25% glutaraldehyde for 7 min. The sections were washed, covered in X-Gal substrate and incubated at 37°C for 1 h. The sections were counterstained with Nuclear Fast Red for 3 min.

Quantitation of β-gal in the eyes was performed as previously described (17), using purified Escherichia coli β-gal (Prozyme) as the standard, and ONPG (Sigma) as the colorimetric substrate.

Using Tg mice, retinal expression of a new Ag, β-gal, was achieved without the consequences of mechanical inoculation used in previous studies. Expression in tissues from Tg mice was examined by X-Gal staining (Fig. 1). High level expression was found in the photoreceptor cell layer of the retina in the rho-β-gal mice (Fig. 1,C) and hi-arr-β-gal mice (Fig. 1,B); no expression was found in β-gal-negative littermates (Fig. 1,A). A slight stain was found in the lo-arr-β-gal mice only after prolonged incubation in the X-Gal substrate. Col-β-gal mice expressed a small amount of β-gal in the peripheral retina (Fig. 1,D), skin (Fig. 1,E), kidney (Fig. 1 F), and elsewhere. Thymuses from the Tg mice were also examined by X-Gal staining and RT-PCR. No evidence of β-gal or β-gal mRNA was found (data not shown).

FIGURE 1.

X-Gal staining reveals the presence and distribution of β-gal in Tg mice. The entire photoreceptor cell (PC) layer of the retina was heavily stained with blue chromogen in the hi-arr-β-gal (B) and rho-β-gal (C) Tg mice. Normal retina (A) was unstained. The inner nuclear layer (INL) of peripheral retina was stained in the col-β-gal mice (D). Skin (E) and kidney (F) tissue from the col-β-gal mice were also X-Gal-stained. RPE, retinal pigment epithelium; OS, outer segments. Original magnifications: A–C, ×40; D–F, ×10.

FIGURE 1.

X-Gal staining reveals the presence and distribution of β-gal in Tg mice. The entire photoreceptor cell (PC) layer of the retina was heavily stained with blue chromogen in the hi-arr-β-gal (B) and rho-β-gal (C) Tg mice. Normal retina (A) was unstained. The inner nuclear layer (INL) of peripheral retina was stained in the col-β-gal mice (D). Skin (E) and kidney (F) tissue from the col-β-gal mice were also X-Gal-stained. RPE, retinal pigment epithelium; OS, outer segments. Original magnifications: A–C, ×40; D–F, ×10.

Close modal

Retinas from the Tg mice were assayed with ONPG at 4 wk of age. Rho-β-gal and hi-arr-β-gal retinas contained approximately 58 and 66 ng of β-gal, respectively. These amounts correspond to concentrations of approximately 19 and 22 μg/ml in the retina based on a calculated retinal volume of 3 μl. These are clearly immunologically active levels, well above the level of 0.5 ng/ml that is known to be immunologically significant in hen egg lysozyme-expressing Tg mice, for example (28). No activity was detected in the retinas of lo-arr-β-gal mice by ONPG. Approximately 4 ng of β-gal was found per col-β-gal retina. No β-gal activity was found in thymuses or sera from these Tg mice by ONPG assay.

It is well known that adoptive transfer of T cells specific for Ags associated with retinal photoreceptor cells can induce EAU in rodents (26, 29). Accordingly, the adoptive transfer of EAU by activated, β-gal-specific T cells would demonstrate that the β-gal expressed in the Tg mice can be processed and presented by local APC at levels capable of recognition by T cells. The MT-1 T cell line was made from normal B10.A mice immunized s.c. with β-gal in CFA. Several clones were isolated from the MT-1 line and were found to be responsive to β-gal in vitro in a dose-dependent manner (Fig. 2). EAU was induced in both the rho-β-gal and hi-arr-β-gal mice following adoptive transfer with the MT-1 line (Fig. 3, A and B, and Table I). Three of eight clones examined to date, 3E9, 1G8, and 1H9, also transferred EAU to the hi-arr-β-gal mice (Fig. 3,C and Table I). No EAU was found in the lo-arr-β-gal mice. These results show the dose-dependent efferent immune recognition of β-gal and immunopathogenesis of autoimmune disease in the retina. The histopathology of EAU mediated by β-gal-specific T cells was concentrated on the photoreceptor cells, especially in the later stages of the inflammation, similar to that observed following the adoptive transfer of uveitogenic T cells specific for the 461–480 peptide of IRBP (Fig. 3, G and H), a known uveitogenic retinal protein (23). The col-β-gal mice were not susceptible to adoptive transfer of EAU (Table I), nor did they express evidence of autoimmunity in any of several other tissues using this transfer protocol.

FIGURE 2.

β-gal-induced proliferative responses of the T cell lines and clones. A, Proliferation of two Ag-specific T cell lines and eight clones derived from the MT-1 line stimulated with 75 μg/ml β-gal. B, Proliferation of primed LN cells and the three pathogenic T cell clones.

FIGURE 2.

β-gal-induced proliferative responses of the T cell lines and clones. A, Proliferation of two Ag-specific T cell lines and eight clones derived from the MT-1 line stimulated with 75 μg/ml β-gal. B, Proliferation of primed LN cells and the three pathogenic T cell clones.

Close modal
FIGURE 3.

Histopathology of EAU following adoptive transfer or immunization. A, Severe EAU in rho-β-gal mice 13 days after transfer of 5 × 106 MT-1 cells. B, EAU following transfer of 5 × 106 MT-1 cells into hi-arr-β-gal mice. C, EAU induced in hi-arr-β-gal mice by the 1G8 T cell clone. D, Normal retina from B10.A mouse. E and F, EAU in hi-arr-β-gal mice 34 and 29 days after immunization with 100 μg of β-gal. G and H, EAU induced in B10 mice following adoptive transfer of 4.3 × 106 pathogenic mIRBP T cells; the eyes were harvested at 13 and 7 days post-transfer, respectively. Arrows indicate focal infiltrates and lesions. ONL, outer nuclear layer. Original magnifications: H, ×10; A—-F, ×20; G, ×40.

FIGURE 3.

Histopathology of EAU following adoptive transfer or immunization. A, Severe EAU in rho-β-gal mice 13 days after transfer of 5 × 106 MT-1 cells. B, EAU following transfer of 5 × 106 MT-1 cells into hi-arr-β-gal mice. C, EAU induced in hi-arr-β-gal mice by the 1G8 T cell clone. D, Normal retina from B10.A mouse. E and F, EAU in hi-arr-β-gal mice 34 and 29 days after immunization with 100 μg of β-gal. G and H, EAU induced in B10 mice following adoptive transfer of 4.3 × 106 pathogenic mIRBP T cells; the eyes were harvested at 13 and 7 days post-transfer, respectively. Arrows indicate focal infiltrates and lesions. ONL, outer nuclear layer. Original magnifications: H, ×10; A—-F, ×20; G, ×40.

Close modal
Table I.

Summary of EAU induction in the Tg mice

MiceEAU Inductiona
T cell adoptive transferbImmunization
MT-13E91G81H9
hi-arr-β-gal 2 /5c 1/3 1/3 1/4 2/6d 
rho-β-gal 1 /2 ND ND ND ND 
col-β-gal 0 /12 ND ND ND 0/14 
lo-arr-β-gal 0 /17 ND ND ND 0/16 
Controlse 0 /21 0/3 ND 0/1 0/37 
MiceEAU Inductiona
T cell adoptive transferbImmunization
MT-13E91G81H9
hi-arr-β-gal 2 /5c 1/3 1/3 1/4 2/6d 
rho-β-gal 1 /2 ND ND ND ND 
col-β-gal 0 /12 ND ND ND 0/14 
lo-arr-β-gal 0 /17 ND ND ND 0/16 
Controlse 0 /21 0/3 ND 0/1 0/37 
a

Number of mice with EAU/total mice.

b

Transfer of 5 × 106 activated T cells.

c

p = 0.003 by χ2 compared to controls.

d

p = 0.0003 by χ2 compared to controls.

e

Transgene-negative littermates and normal B10.A mice.

The MT-1 line and the clones were CD4+ and α/β TCR+; there was no particular association with any TCR Vβ usage (data not shown). Cytokine analysis of MT-1 line cells and the pathogenic 3E9 clone shows that TNF-α and substantial levels of IFN-γ are produced, but little or no IL-2 or IL-4 (Table II). Assays of other T cells were used to confirm assay reliability for detection of IL-2, IL-4, and IL-6 in the culture supernatants (data not shown).

Table II.

Cytokine production by the uveitogenic, β-gal-specific T cellsa

T CellsCytokines in Culture Supernatantb
IFN-γIL-2IL-4IL-6TNF-α
3E9 clone 561 ± 42 BDc BD 1.8 ± 0.2 13.4 ± 2.5 
MT-1 line 1178 ± 375 1.0 ± 0.1 BD 3.4 ± 0.6 11.0 ± 0.3 
T CellsCytokines in Culture Supernatantb
IFN-γIL-2IL-4IL-6TNF-α
3E9 clone 561 ± 42 BDc BD 1.8 ± 0.2 13.4 ± 2.5 
MT-1 line 1178 ± 375 1.0 ± 0.1 BD 3.4 ± 0.6 11.0 ± 0.3 
a

Representative of three experiments.

b

Data given in ng/ml; mean ± SD; determined by ELISA.

c

BD, below detection (<40 pg/ml IL-2; <24 pg/ml IL-4).

Immunizations were performed using a procedure adapted from a protocol designed to induce EAU with IRBP in mice (30). The hi-arr-β-gal mice developed EAU (Fig. 3, E and F), but the col-β-gal and lo-arr-β-gal mice did not (Table I). Clearly, tolerance capable of protecting the mice from actively induced, β-gal-mediated EAU was not present in mice expressing β-gal only in the retina. The lack of disease in the col-β-gal mice has several possible explanations; the amount of β-gal in the col-β-gal retinas may be too low to support a pathogenic response, or there may be tolerance to β-gal resulting from the low, systemic level of expression. Evidence for partial tolerance is shown below.

Evidence for tolerance to β-gal as a self-Ag was also sought from in vitro assays. Draining LN cells from Tg and control mice primed with β-gal were tested in proliferation assays. Based on the dose responses of normal, immunized mice to β-gal (Fig. 4,A), 0.63 μM (75 μg/ml) Ag was used to survey the responses of the various Tg and non-Tg control mice. Both lo- and hi-arr-β-gal mice were indistinguishable from B10.A controls, but col-β-gal mice exhibited a modestly reduced response (p = 0.05 compared with β-gal-negative littermates; Fig. 5). The reduced response of the col-β-gal mice was the result of an upward shift in the dose-response curve, so that the use of high doses of β-gal in the proliferation assays significantly restored their responsiveness, which was lost at low doses (Fig. 4,B). The dose-response curve of the lo-arr-β-gal mice was unaffected (Fig. 4 C). Unprimed mice had no response to β-gal.

FIGURE 4.

In vitro proliferative responses of draining LN cells from β-gal-immunized control and Tg mice. A, Dose response of normal, immunized mice to β-gal. B and C, Dose responses of the col-β-gal mice (B) and the lo-arr-β-gal mice (C). Data are the averages of three, nine, and six mice, respectively.

FIGURE 4.

In vitro proliferative responses of draining LN cells from β-gal-immunized control and Tg mice. A, Dose response of normal, immunized mice to β-gal. B and C, Dose responses of the col-β-gal mice (B) and the lo-arr-β-gal mice (C). Data are the averages of three, nine, and six mice, respectively.

Close modal
FIGURE 5.

Survey of in vitro assays of β-gal-immunized control and Tg mice. A, Scatter plot of individual lymphocyte proliferation assays. Draining LNs were collected 8–10 days after priming and were stimulated with 75 μg/ml (0.63 μM) Ag (+) or without Ag (−). The means (•) and SD (○) are shown for each group. B, ELISA titers of anti-β-gal serum Ab responses of individual Tg and control mice were obtained with serum samples collected 28–31 days after immunization with β-gal. Titers of unprimed Tg and control mice were below detection. The geometric mean (•) and SD (○) are shown.

FIGURE 5.

Survey of in vitro assays of β-gal-immunized control and Tg mice. A, Scatter plot of individual lymphocyte proliferation assays. Draining LNs were collected 8–10 days after priming and were stimulated with 75 μg/ml (0.63 μM) Ag (+) or without Ag (−). The means (•) and SD (○) are shown for each group. B, ELISA titers of anti-β-gal serum Ab responses of individual Tg and control mice were obtained with serum samples collected 28–31 days after immunization with β-gal. Titers of unprimed Tg and control mice were below detection. The geometric mean (•) and SD (○) are shown.

Close modal

Except for a single col-β-gal mouse, similar titers of serum Abs to β-gal (Fig. 5 B), were found in all β-gal-primed mice, revealing little evidence for tolerance in B cell responsiveness. The titers of unprimed control mice, both Tg and normal, were below detection. Even though the hi-arr-β-gal mice expressed β-gal in the retina sufficient to be susceptible to EAU, they did not appear to develop tolerance to β-gal based on that expression.

It has been proposed that tolerance to self-Ags need not be complete if tolerance to the immunodominant epitopes of the self-Ag is induced and maintained, since biologically significant responses are unlikely to be generated to the cryptic sites (31). Using Tg mice with widespread expression of hen egg lysozyme (HEL), it was shown that tolerance was preferentially induced to immunodominant determinants (32). The modest decrease in the responses of the col-β-gal mice described above raised the possibility that T cells specific for subdominant or cryptic sites in β-gal were raised by the adjuvant-promoted response and were responsible for the β-gal responses. The Tg mice described here allow this hypothesis to be easily tested, since the reference response of normal, β-gal-negative mice to β-gal can be determined. The 34 peptide pools described in Materials and Methods were used to screen mice for responses to β-gal. Based on preliminary screening assays showing that the peak response to the immunodominant group 14 peptides was achieved at approximately 1 μM, 2 μM peptide was used in subsequent experiments. A highly reproducible pattern of responses was found (Fig. 6). The patterns of responses of the hi- and lo-arr-β-gal mice were indistinguishable from each other and from those of the non-Tg B10.A control mice, further demonstrating immunological ignorance of the retinal β-gal.

FIGURE 6.

Proliferative responses of β-gal-primed mice to the 34 peptide pools of β-gal. Groups of eight mice were immunized with β-gal. After 8–10 days, draining LN cells were harvested, pooled, and assayed with the peptides (2 μM each). The horizontal bar indicates the response of the unstimulated controls. Results are representative of two or three independent experiments.

FIGURE 6.

Proliferative responses of β-gal-primed mice to the 34 peptide pools of β-gal. Groups of eight mice were immunized with β-gal. After 8–10 days, draining LN cells were harvested, pooled, and assayed with the peptides (2 μM each). The horizontal bar indicates the response of the unstimulated controls. Results are representative of two or three independent experiments.

Close modal

Conversely, the peptide responses of the β-gal-immunized col-β-gal mice were substantially reduced compared with the response to the whole β-gal molecule, which was only moderately less than that in the control mice. The peptides most stimulatory in the col-β-gal mice, groups 4, 6, 14, 19, 20, 29, and 31, were also the most stimulatory in normal B10.A mice, showing that tolerance was not preferentially elicited by the immunodominant peptides and that responses to cryptic sites cannot explain the ability of the whole molecule to stimulate responses in the col-β-gal mice. There is the possibility that the synthetic peptides failed to detect every epitope. However, the conclusion that the overall peptide response of the col-β-gal mice is substantially reduced without regard for immunodominance remains valid. Preliminary studies show that increased concentrations of these peptide pools revealed higher activity, as already shown for the responsiveness of col-β-gal mice to intact β-gal in Fig. 4 (data not shown).

The differences in the responses of β-gal-immunized col-β-gal mice compared with those of the other mice were potentially due to differences in cytokine production. Ag-stimulated production of cytokines that directly promote T cell proliferation, IL-2 and IL-4, was very low in all β-gal-primed mice and was not substantially different between Tg and non-Tg mice (Table III). Levels of TNF-α and IL-6, which are associated with the pathogenesis of EAU (33) and EAE (34), were not different between any groups of mice. Although some difference in IFN-γ production, also associated with EAU (35), was observed between Tg and non-Tg mice, col-β-gal mice were not different from hi-arr-β-gal mice. Since all mice produced substantial amounts of IFN-γ, the biological significance of the approximately 3- to 4-fold difference between B10.A and col-β-gal mice is uncertain.

The MT-1 line was screened with the peptide groups after 15–20 cycles of β-gal stimulation and propagation. Even though the line was highly selected in vitro, it responded to multiple dominant sites similar to those identified in the previous screening assays of primed LN cells (Fig. 7). Screening the pathogenic T cell clone 3E9 that was derived from the MT-1 line revealed that it responded well to group 14, as did MT-1 cells (Fig. 7). When tested against the individual peptides of this group, peptides 70 and 71 were found to stimulate the proliferation of 3E9 T cells, identifying the overlapping region of these peptides as the site of recognition (Table IV).

FIGURE 7.

Proliferative responses of the MT-1 T cells and the 3E9 clone to the peptide pools of β-gal. The results are representative of three and two separate experiments, respectively. Peptides were used at 2 μM each.

FIGURE 7.

Proliferative responses of the MT-1 T cells and the 3E9 clone to the peptide pools of β-gal. The results are representative of three and two separate experiments, respectively. Peptides were used at 2 μM each.

Close modal
Table IV.

Peptide specificity of the 3E9 T cell clone

PeptideSequence (β-gal)CPM × 10−3
68a HYPNHPLWYTLCDRYGL 
69 LWYTLCDRYGLYVVDEA 
70 DRYGLYVVDEANIETHG 37 
71 VVDEANIETHGMVPMNR 46 
72 IETHGMVPMNRLTDDPR 
No Ag  
β-galb  65 
PeptideSequence (β-gal)CPM × 10−3
68a HYPNHPLWYTLCDRYGL 
69 LWYTLCDRYGLYVVDEA 
70 DRYGLYVVDEANIETHG 37 
71 VVDEANIETHGMVPMNR 46 
72 IETHGMVPMNRLTDDPR 
No Ag  
β-galb  65 
a

Peptides were used at 2 μM.

b

β-Gal was used at 75 μg/ml.

Studies of active mechanisms of immune privilege rely on the experimental induction of immune deviation phenomena, usually the generation of a regulatory T cell, whose activity is then detected and observed as a change in the immune response to subsequent immunization with the Ag in question. Using this strategy, both the eye and brain have been found to be sites where inoculation of autoantigens leads to deviated responses and resistance to the corresponding experimental autoimmune diseases in those organs (36, 37). Since the tolerance induced by this strategy is experimentally induced by the inoculation of Ag, it is difficult to know its significance in the normal animal that has not received the injections of the Ag required to induce measurable differences. As a result, the role of immune deviation in maintaining immune privilege remains uncertain.

The results of the Tg mouse strategy described in this report support the interpretation that sequestration plays an important role in the immune privilege of the retina. The ability to express a new self Ag in the desired tissues and at different levels also provides a model that will allow more critical testing of immune deviation phenomena. Importantly, the use of transgenes allows the critical control of testing the immune response in the absence of the Ag, a condition that can be difficult to achieve with many self-Ags.

Although we found no evidence for β-gal-specific tolerance in the Tg mice that expressed β-gal via the retinal photoreceptor cell promoters, expression of β-gal in other sites can clearly be tolerogenic. Thymic expression of a β-gal transgene induced tolerance in both CD4 and CD8 T cells if β-gal was expressed in the medullary epithelial cells (38, 39). Recognition of β-gal in skin grafts by CD8 cytotoxic T cells has shown that β-gal can be recognized as a minor histocompatibility Ag in the context of class I MHC (40). Elsewhere, expression of a β-gal transgene in mouse brain parenchyma was found using a glial fibrillary acidic protein promoter (41). Immunization of these mice with β-gal led to a mononuclear infiltrate in the areas of expression, but no clinical signs of disease were observed. Interestingly, these mice were not tolerant to β-gal, although their expression level is lower than we have found in the rho- or hi-arr-β-gal mice (unpublished observations). Partial tolerance to β-gal was found in Tg mice, where expression was directed to lymphoid tissues at a concentration of 20 ng/spleen using an Ig heavy chain promoter and μ transmembrane sequence (42). A moderate reduction in Ab response and affinity to β-gal was found following immunization of those Tg mice, revealing some tolerance induction. There was no indication of immunopathology resulting from autoimmunity.

In the current study, the hi-arr-β-gal mice expressed at a much higher concentration in the retina, but did not induce tolerance by the assays we used. Since the expression of a substantial level of β-gal in the retina was not tolerogenic, we conclude that the environment of the retina sequestered the Ag, maintaining tolerance and ocular immune privilege until challenged with a bolus of activated T cells resulting from adoptive transfer or immunization in the presence of CFA and pertussis toxin.

The consequences of Tg expression in the CNS have been examined in a few other reports (43). In these cases the foreign transgenes have been targets for CD8 CTL activity, unlike our CD4-mediated pathology. There have been a number of reports of the activity of CD8 T cells to transgenic Ags leading to autoimmune disease; most of these demonstrate CTL responses to lymphocytic choriomeningitis virus (LCMV) or influenza hemagglutinin Ags expressed in islets in diabetes models. CD4 T cell-mediated pathology was demonstrated by Lo et al. (44), using the hemagglutinin Ag in islet cells. Forster et al. (45) showed that expression of Tag using a rat insulin promoter has a limited ability to induce tolerance in CD4 T cells. However, it is not clear whether it is the thymic expression associated with islet cell promoters or pancreatic expression that is responsible for the observed tolerance in that study. The thymic expression commonly found in islet Ag systems complicates the interpretations of peripheral tolerance. Also, the level of expression in the islet cell models is generally extremely low compared with the high level of retinal expression in our system, which is not further complicated by thymic expression.

In a different model, Goodnow et al. (46) used a double transgenic mouse with the anti-hen egg lysozyme (HEL) TCR Tg and HEL expression directed to the thyroid or islet cells, or systemically. They found extensive infiltrates in the thyroid and pancreas, but not pathology. There have also been numerous demonstrations of tolerance induction by transgenic protein Ags. In contrast, what we have shown is a lack of tolerance to a very well-expressed transgene in the retina and the CD4-mediated immunopathology directed to it.

Clear evidence for partial tolerance at the T cell level was found in the col-β-gal mice following immunization, possibly due to the low level of peripheral β-gal expression, since the substantial levels of retinal expression in the rho- and hi-arr-β-gal mice did not lead to tolerance. The col-β-gal mice gave no sign of expression in the thymus at 4 or more weeks of age. Splenocytes or thymocytes from col-β-gal mice had no ability, in the absence of exogenous β-gal, to stimulate MT-1 or 3E9 cells when used as APC in proliferation assays (data not shown). The much lower expression levels in the col-β-gal mice did not provide convincing targets for autoimmune disease in the retina or at other tissue sites of expression. In preliminary results, a trace level of infiltration of the peripheral retina was found in one of eight col-β-gal mice that were given 500-rad irradiation before transfer with larger numbers (107) of activated MT-1 cells (data not shown). There was no sign of infiltrates in other X-Gal-positive tissues from these mice.

The lack of susceptibility of the col-β-gal mice to adoptive transfer disease in the retina or at the other systemic sites of expression has several possible explanations. While their systemic expression level was sufficient to reduce the in vitro lymphoproliferative response after priming with the whole molecule, indicating that the presence of β-gal was recognized by the immune system, it did not support a pathogenic response. Since it is thought that activated T cells have reduced Ag recognition requirements, it would have been reasonable to predict a response by the transferred, activated T cells. There is the possibility that these mice possess regulatory cells such as the CD4+25+ T cells shown to be potent regulators of autoimmunity in other systems (47, 48). There is also the possibility that the activated T cells used in the adoptive transfer acquired altered migratory routes due to contact with Ag in nonlymphoid tissues, resulting in decreased pathogenicity in the eye and elsewhere. Additional Tg mice are being acquired to further study this question.

While the incidence of EAU in the rho- and hi-arr-β-gal mice was low, its presence is highly significant. We have observed a very large number of rodent eyes over many years, and have not seen spontaneous uveitis resembling EAU. The expression of the transgene itself has not resulted in the appearance of histologic abnormalities or any inflammatory conditions. As shown, immunization of many normal or transgene-negative littermates with β-gal or their transfer with β-gal-specific T cells has also never led to the appearance of any histopathology.

Apparently there is nothing special about the known immunopathogenic autoantigens of the retina, since the E. coli enzyme β-gal can act as such an Ag when present in the retina as a transgene. A substantial level of expression in a sequestered site may be the most important requirements for immunopathogenic autoantigen status. Furthermore, the results show that CD4 T cells raised to a foreign, nonautoantigen in a normal mouse can be immunopathogenic in recipients expressing that Ag.

Our results are most consistent with the interpretation that sequestration is the primary mechanism of retinal immune privilege, and that active mechanisms of ocular immune privilege augment sequestration. These active mechanisms, which are readily induced and detected, may be required to maintain privilege following traumatic or inflammatory injury. The barriers provided by the retinal vascular endothelium and retinal pigment epithelium prevent central tolerance to retinal proteins expressed even at high levels, and peripheral tolerance is limited due to minimal lymphocytic perusal of normal, uninflamed retina and the virtual absence of class II MHC in the retina.

We thank Jing Xiao and Quang Nguyen for expert technical assistance, and Rudy Padua for the histology preparations.

1

This work was supported by National Institutes of Health Grant EY11542, Research to Prevent Blindness, Inc., The Anna Heilmaier Foundation, and the Minnesota Lions and Lionesses Clubs.

3

Abbreviations used in this paper: EAE, experimental autoimmune encephalomyelitis; MBP, myelin basic protein; Tg, transgenic; EAU, experimental autoimmune uveoretinitis; β-gal, β-galactosidase; IRBP, interphotoreceptor retinoid-binding protein; ONPG, o-nitrophenyl β-d-galactopyranoside; X-Gal, 5-bromo-4-chloro-3-indolyl β-d-galactoside; HEL, hen egg lysozyme; LN, lymph node.

1
Silverstein, A. M..
1990
.
The Early History of Ocular Immunopathology
Elsevier, Amsterdam.
2
Griffith, T. S., T. Brunner, S. M. Fletcher, D. R. Green, T. A. Ferguson.
1995
. Fas ligand-induced apoptosis as a mechanism of immune privilege.
Science
270
:
1189
3
Forrester, J. V., P. G. McMenamin, I. Holthouse, L. Lumsden, J. Liversidge.
1994
. Localization and characterization of major histocompatibility complex class II-positive cells in the posterior segment of the eye: implications for induction of autoimmune uveoretinitis.
Invest. Ophthalmol. Vis. Sci.
35
:
64
4
Streilein, J. W., M. Takeuchi, A. W. Taylor.
1997
. Immune privilege, T-cell tolerance, and tissue-restricted autoimmunity.
Hum. Immunol.
52
:
138
5
Egan, R. M., C. Yorkey, R. Black, W. K. Loh, J. L. Stevens, J. G. Woodward.
1996
. Peptide-specific T cell clonal expansion in vivo following immunization in the eye, an immune-privileged site.
J. Immunol.
157
:
2262
6
Cserr, H. F., P. M. Knopf.
1992
. Cervical lymphatics, the blood-brain barrier and the immunoreactivity of the brain: a new view.
Immunol. Today
13
:
507
7
Happ, M. P., E. Heber-Katz.
1988
. Differences in the repertoire of the Lewis rat T cell response to self and non-self myelin basic proteins.
J. Exp. Med.
167
:
502
8
Goverman, J., A. Woods, L. Larson, L. P. Weiner, L. Hood, D. M. Zaller.
1993
. Transgenic mice that express a myelin basic protein-specific T cell receptor develop spontaneous autoimmunity.
Cell
72
:
551
9
Mathisen, P. M., S. Pease, J. Garvey, L. Hood, C. Readhead.
1993
. Identification of an embryonic isoform of myelin basic protein that is expressed widely in the mouse embryo.
Proc. Natl. Acad. Sci. USA
90
:
10125
10
Egwuagu, C. E., P. Charukamnoetkanok, I. Gery.
1997
. Thymic expression of autoantigens correlates with resistance to autoimmune disease.
J. Immunol.
159
:
3109
11
Wacker, W. B., L. A. Donoso, C. M. Kalsow, J. A. Yankeelov, Jr, D. T. Organisciak.
1977
. Experimental allergic uveitis: isolation, characterization, and localization of a soluble uveitopathogenic antigen from bovine retina.
J. Immunol.
119
:
1949
12
Craft, C. M., D. H. Whitmore.
1995
. The arrestin superfamily: cone arrestins are a fourth family.
FEBS Lett.
362
:
247
13
Gregerson, D. S..
1998
. Immune privilege in the retina.
Ocular Immunol. Inflam.
6
:
257
14
Fazekas de St. Groth, B., M. C. Cook, A. L. Smith.
1997
. The role of T cells in the regulation of B cell tolerance.
Int. Rev. Immunol.
15
:
73
15
Goverman, J., T. Brabb, A. Paez, C. Harrington, P. von Dassow.
1997
. Initiation and regulation of CNS autoimmunity.
Crit. Rev. Immunol.
17
:
469
16
Kruisbeek, A. M., D. Amsen.
1996
. Mechanisms underlying T-cell tolerance.
Curr. Opin. Immunol.
8
:
233
17
Zack, D. J., J. Bennett, Y. Wang, C. Davenport, B. Klaunberg, J. Gearhart, J. Nathans.
1991
. Unusual topography of bovine rhodopsin promoter-lacZ fusion gene expression in transgenic mouse retinas.
Neuron
6
:
187
18
Kikuchi, T., K. Raju, M. L. Breitman, T. Shinohara.
1993
. The proximal promoter of the mouse arrestin gene directs gene expression in photoreceptor cells and contains an evolutionarily conserved retinal factor-binding site.
Mol. Cell. Biol.
13
:
4400
19
Kikuchi, T., E. Wawrousek, T. Lee, S. Di Camillo, K. Kusuzaki, T. Shinohara.
1995
. Position or number of photoreceptor conserved element (PCE1) in arrestin promoter may be altered in molecular evolution.
Invest. Ophthalmol. Vis. Sci.
36
:
5772
(Abstr.).
20
Sunayashiki-Kusuzaki, K., T. Kikuchi, E. F. Wawrousek, T. Shinohara.
1997
. Arrestin and phosducin are expressed in a small number of brain cells.
Mol. Brain Res.
52
:
112
21
Niederreither, K., R. N. D’Souza, B. deCrombrugghe.
1992
. Minimal DNA sequences that control the cell lineage-specific expression of the proα2 (I) collagen promoter in transgenic mice.
J. Cell Biol.
119
:
1361
22
Shivakumar, S., E. E. Sercarz, U. Krzych.
1989
. The molecular context of determinants within the priming antigen establishes a hierarchy of T cell induction: T cell specificities induced by peptides of β-galactosidase vs. the whole antigen.
Eur. J. Immunol.
19
:
681
23
Caspi, R. R., B. G. Grubbs, C. C. Chan, G. J. Chader, B. Wiggert.
1992
. Genetic control of susceptibility to experimental autoimmune uveoretinitis in the mouse model: concomitant regulation by MHC and non-MHC genes.
J. Immunol.
148
:
2384
24
Silver, P. B., L. V. Rizzo, C. C. Chan, L. A. Donoso, B. Wiggert, R. R. Caspi.
1993
. Identification of a putative epitope in the IRBP molecule that is uveitogenic for mice of the H-2b haplotype.
Invest. Ophthalmol. Vis. Sci.
34
:
1482
(Abstr.).
25
Caspi, R. R., C. C. Chan, W. C. Leake, M. Higuchi, B. Wiggert, G. J. Chader.
1990
. Experimental autoimmune uveoretinitis in mice: induction by a single eliciting event and dependence on quantitative parameters of immunization.
J. Autoimmun.
3
:
237
26
Gregerson, D. S., W. F. Obritsch, S. P. Fling, J. D. Cameron.
1986
. S-antigen-specific rat T cell lines recognize peptide fragments of S-antigen and mediate experimental autoimmune uveoretinitis and pinealitis.
J. Immunol.
136
:
2875
27
Bout, A., D. Valerio, B. J. Scholte.
1993
. In vivo transfer and expression of the lacZ gene in the mouse lung.
Exp. Lung Res.
19
:
193
28
Goodnow, C. C..
1992
. Transgenic mice and analysis of B-cell tolerance.
Annu. Rev. Immunol.
10
:
489
29
Rizzo, L. V., P. Silver, B. Wiggert, F. Hakim, R. T. Gazzinelli, C.-C. Chan, R. R. Caspi.
1996
. Establishment and characterization of a murine CD4+ T cell line and clone that induce experimental autoimmune uveoretinitis in B10.
A mice. J. Immunol.
156
:
1654
30
Caspi, R. R., F. G. Roberge, C. C. Chan, B. Wiggert, G. J. Chader, L. A. Rozenszajn, Z. Lando, R. B. Nussenblatt.
1988
. A new model of autoimmune disease: experimental autoimmune uveoretinitis induced in mice with two different retinal antigens.
J. Immunol.
140
:
1490
31
Sercarz, E. E., P. V. Lehmann, A. Ametani, G. Benichou, A. Miller, K. Moudgil.
1993
. Dominance and crypticity of T cell antigenic determinants.
Annu. Rev. Immunol.
11
:
729
32
Cibotti, R., J. M. Kanellopoulos, J.-P. Cabaniols, O. Halle-Panenko, K. Kosmatopolulos, E. Sercarz, P. Kourilsky.
1992
. Tolerance to a self-protein involves its immunodominant but does not involve its subdominant determinants.
Proc. Natl. Acad. Sci. USA
89
:
416
33
Dick, A. D., P. G. McMenamin, H. Korner, B. J. Scallon, J. Ghrayeb, J. V. Forrester, J. D. Sedgwick.
1996
. Inhibition of tumor necrosis factor activity minimizes target organ damage in experimental autoimmune uveoretinitis despite quantitatively normal activated T cell traffic to the retina.
Eur. J. Immunol.
26
:
1018
34
Mendel, I., A. Katz, N. Kozak, A. Bennun, M. Revel.
1998
. Interleukin-6 functions in autoimmune encephalomyelitis: a study in gene-targeted mice.
Eur. J. Immunol.
28
:
1727
35
Sun, B., L. V. Rizzo, S. H. Sun, C. C. Chan, B. Wiggert, R. L. Wilder, R. R. Caspi.
1997
. Genetic susceptibility to experimental autoimmune uveitis involves more than a predisposition to generate a T helper-1-like or a T helper-2-like response.
J. Immunol.
159
:
1004
36
Hara, Y., R. R. Caspi, B. Wiggert, C. C. Chan, G. A. Wilbanks, J. W. Streilein.
1992
. Suppression of experimental autoimmune uveitis in mice by induction of anterior chamber-associated immune deviation with interphotoreceptor retinoid-binding protein.
J. Immunol.
148
:
1685
37
Harling-Berg, C. J., P. M. Knopf, H. F. Cserr.
1991
. Myelin basic protein infused into cerebrospinal fluid suppresses experimental autoimmune encephalomyelitis.
J. Neuroimmunol.
35
:
45
38
Oukka, M., E. Colucci-Guyon, P. L. Tran, M. Cohen-Tannoudji, C. Babinet, V. Lotteau, K. Kosmatopoulos.
1996
. CD4 T cell tolerance to nuclear proteins induced by medullary thymic epithelium.
Immunity
4
:
545
39
Oukka, M., M. Cohen-Tannoudji, Y. Tanaka, C. Babinet, K. Kosmatopoulos.
1996
. Medullary thymic epithelial cells induce tolerance to intracellular proteins.
J. Immunol.
156
:
968
40
Oukka, M., M. Galou, Y. Belkaid, V. Tricotet, N. Riche, M. Reynes, K. Kosmatopoulos.
1997
. Major histocompatibility complex class I presentation of exogenously acquired minor alloantigens initiates skin allograft rejection.
Eur. J. Immunol.
27
:
3499
41
Borrow, P., J. L. Cornell, M. D. Ruppe, L. Mucke.
1995
. Immunization-induced inflammatory infiltration of the central nervous system in transgenic mice expressing a microbial antigen in astrocytes.
J. Neuroimmunol.
61
:
133
42
Theopold, U., G. Kohler.
1990
. Partial tolerance in β-galactosidase-transgenic mice.
Eur. J. Immunol.
20
:
1311
43
Evans, C. F., M. S. Horwitz, M. V. Hobbs, M. B. A. Oldstone.
1996
. Viral infection of transgenic mice expressing a viral protein in oligodendrocytes leads to chronic central nervous system autoimmune disease.
J. Exp. Med.
184
:
2371
44
Degermann, S., C. Reilly, B. Scott, L. Ogata, H. Vonboehmer, D. Lo.
1994
. On the various manifestations of spontaneous autoimmune diabetes in rodent models.
Eur. J. Immunol.
24
:
3155
45
Forster, I., R. Hirose, J. M. Arbeit, B. E. Clausen, D. Hanahan.
1995
. Limited capacity for tolerization of CD4(+) T cells specific for a pancreatic β cell neo-atigen.
Immunity
2
:
573
46
Akkaraju, S., W. Y. Ho, D. Leong, K. Canaan, M. M. Davis, C. C. Goodnow.
1997
. A range of Cd4 T cell tolerance: partial inactivation to organ-specific antigen allows nondestructive thyroiditis or insulitis.
Immunity
7
:
255
47
Suri-Payer, E., A. Z. Amar, A. M. Thornton, E. M. Shevach.
1998
. CD4+CD25+ T cells inhibit both the induction and effector function of autoreactive T cells and represent a unique lineage of immunoregulatory cells.
J. Immunol.
160
:
1212
48
Sakaguchi, S., N. Sakaguchi, M. Asano, M. Itoh, M. Toda.
1995
. Immunologic self-tolerance maintained by activated T cells expressing IL-2 receptor α-chains (CD25).
J. Immunol.
155
:
1151