Protection against Murine Cytomegalovirus Retinitis by Adoptive Transfer of Virus-Specific CD8⁺ T Cells

Abstract

purpose. Human cytomegalovirus retinitis, the most common ophthalmic infection of AIDS patients, has been modeled in BALB/c mice infected with murine cytomegalovirus by the supraciliary route. A series of depletion and adoptive transfer studies was performed to determine whether adoptive transfer of T cells protects mice from retinitis caused by murine cytomegalovirus infection after supraciliary inoculation and to determine which subset of T cells is responsible for protection.

methods. BALB/c mice were thymectomized and T cell–depleted by injection of monoclonal antibodies to CD4, CD8, or both. Murine cytomegalovirus (9 × 10² plaque forming units [pfu]) was injected into the supraciliary space. Experimental animals received murine cytomegalovirus-specific T cells or subsets of T cells 2 hours before virus injection, whereas control animals received herpes simplex virus type 1–specific T cells by tail vein injection. Eight days after virus injection, retinal pathology was scored by histopathologic examination of hematoxylin and eosin–stained ocular sections.

results. CD8⁺ T cell depletion was sufficient for development of retinitis after supraciliary injection of murine cytomegalovirus. Adoptive transfer of murine cytomegalovirus-specific T cells, but not herpes simplex virus type 1–specific T cells, provided protection from retinitis. Additionally, separation of the murine cytomegalovirus-specific T cells into CD8⁺ and CD4⁺ subsets before adoptive transfer showed that the CD8⁺ fraction of the adoptive T cells was responsible for protection.

conclusions. These results suggest that adoptive transfer of cytomegalovirus-specific T cells or T cell subsets might be used to treat or prevent cytomegalovirus retinitis in immunosuppressed human patients.

Cytomegalovirus (CMV) retinitis is the most common ophthalmic infection in patients with AIDS. ^{1 2 3 4 5 6} CMV retinitis, which is usually a complication of end-stage AIDS, is a focal progressive necrotizing infection of the retina, which causes reduction in visual acuity and, if left untreated, blindness. ^{3 6} Before the advent of highly active antiretroviral therapy (HAART), CMV retinitis was observed in 29% to 32% of patients with AIDS during the course of their disease. ^{1 2} Use of HAART has decreased the frequency of CMV retinitis, ⁷ and in many cases, CMV retinitis patients on HAART partially recover their immunity to CMV and are able to discontinue anti-CMV therapy. ^{8 9 10 11} However, not all patients can tolerate HAART, and some patients who are on HAART (and whose CD4 cell count increases to over 100 cells/mm³) still develop CMV retinitis. ⁷ For these patients, current CMV retinitis therapies are usually effective; however, disease progression may in some patients be due to a lack of compliance, emergence of drug-resistant virus, and/or cessation of therapy due to drug toxicity. ^{4 5 7} Adoptive transfer of CMV-specific CD8⁺ T cells has been shown to protect bone marrow transplant patients from CMV pneumonia, ^{12 13} and although the pathogenesis of this disease is different from CMV retinitis, one approach to treat CMV retinitis might be adoptive transfer of CMV-specific T cells. 

To explore the usefulness of adoptive transfer of T cells for prevention of CMV retinitis, a series of depletion and reconstitution experiments was conducted using a mouse model of CMV retinitis. Because members of the cytomegalovirus family are extremely species-restricted, murine cytomegalovirus (MCMV) has been used to study the pathogenesis of cytomegalovirus infections of the retina in BALB/c mice. ^{14 15 16 17 18} Injection of a low dose (5 × 10²–1 × 10³ plaque forming units [pfu]) of MCMV into the supraciliary space of one eye of an immunosuppressed or T cell–depleted BALB/c mouse causes a focal necrotizing retinal infection that resembles CMV retinitis in humans. ^{16 18}  

Depletion of CD8⁺ T cells from euthymic mice is sufficient to predispose the animals to retinitis after injection of a low-dose of MCMV via the supraciliary route. ¹⁶ Subsequently, Lu and coworkers determined that protection from MCMV retinitis is provided by adoptive transfer of unfractionated lymph node cells from MCMV-infected mice to thymectomized T cell–depleted mice. ¹⁸ By building on the earlier results of Lu and coworkers, the results of the present study show that the T cell component of the MCMV-specific lymph node cells is responsible for protecting thymectomized T cell–depleted mice from retinitis. Furthermore, adoptive transfer of CD8⁺ T cells alone, but not CD4⁺ T cells alone, protects against MCMV retinitis. 

Methods

Animals

Female euthymic BALB/c mice, 6 to 8 weeks old, were obtained from Taconic (Germantown, NY). Animals were housed on a 12-hour light/dark cycle and given unrestricted access to food and water. Animals were housed in accordance with National Institutes of Health guidelines, and all procedures in this study conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Virus and Virus Titrations

Stocks of MCMV (Smith strain) were prepared from salivary gland homogenates from MCMV-infected BALB/c mice as previously described. ¹³ Virus stocks and virus recovered from tissues were titered in duplicate on Swiss Brown mouse embryo fibroblasts (BioWhittaker, Walkersville, MD). 

Thymectomy and T-Cell Depletion

Thymectomies were performed on 6-week-old mice using a modification of a protocol by Chin. ^{19 20} Thymectomized mice were rested for 1 week before T cell depletion. T cell depletion was accomplished by intravenous injection of 500 μg of anti-CD4 (GK1.5) monoclonal antibody and 150 μg of anti-CD8 (2.43) monoclonal antibody (American Type Culture Collection, Rockville, MD). This protocol has been shown to deplete >94% of CD4⁺ T cells and >99% of CD8⁺ T cells. ²¹ Mice were then rested for 28 days to allow catabolization of the rat anti-mouse T cell antibodies before adoptive transfer. ¹⁸ Non–cross-reactive phycoerythrin (PE)-labeled anti–Ly-3.2 (clone 53-5.8; Pharmingen, San Diego, CA) and fluorescein isothiocyanate [FITC[]–labeled anti-L3T4 (clone RM4-4, Pharmingen) recognizing CD8 and CD4, respectively, were used to determine the efficiency of T cell depletion. 

Ocular Inoculation

Mice were anesthetized by intramuscular injection of a cocktail containing 0.02 ml rompun and 0.03 ml ketamine per 25 g body mass. The left eyes of mice were injected with 9.0 × 10² pfu MCMV in a volume of 2 μl via the supraciliary route as previously described. ¹³ Briefly, a superficial transscleral entry wound was made parallel and just posterior to the limbus by introducing the bevel of a sharp 30-gauge needle into the supraciliary space. Two microliters of virus followed by 3 μl of air was then injected. The injection was judged successful if ophthalmic observation using the dissecting microscope showed a chorioretinal detachment associated with the appearance of air in the supraciliary space immediately after injection. 

Flow Cytometry

Animals were deeply anesthetized before they were killed. Lymph nodes were harvested, and single cell suspensions were made by grinding the tissues between frosted glass slides in 5 ml of cold Hanks’ balanced salt solution (HBSS) and then aspirating the suspension through a 22-gauge needle. Cells were washed 3 times in HBSS and resuspended in FACS buffer (phosphate-buffered saline [PBS] with 3% fetal bovine serum). Samples were incubated with anti-CD32/CD16 (Fc receptor block; Pharmingen) according to the manufacturer’s recommendations for 15 minutes. Cells were then resuspended in FITC-labeled anti-L3T4 (CD4; Pharmingen), PE-labeled anti–Ly-3.2 (CD8; Pharmingen), FITC-labeled anti-CD3 (GIBCO, Gaithersburg, MD), and/or PE-labeled anti-CD4 (GIBCO), as required. After 15 minutes, cells were washed 3 times in FACS buffer and resuspended in FACS buffer for analysis. Gating was set to count large granular cells including lymphocytes and macrophages. Percent positive cells = % positive stained cells minus % positive cells from the same unstained sample. 

Adoptive Transfer

Cells for adoptive transfer were prepared by harvesting the draining lymph nodes from mice 7 days after footpad injection of 5 × 10⁵ pfu of MCMV or herpes simplex virus type-1 (HSV-1). Single cell suspensions were prepared from the lymph nodes as described above for flow cytometry. Cells were washed 2 times in HBSS and 1 time in column buffer (prepared to manufacturer’s specifications). Cells were then resuspended in column buffer, and T cells, CD4⁺ T cells, or CD8⁺ T cells were purified by passing the total cell suspension over the appropriate T cell or T cell subset enrichment column (R&D Systems, Minneapolis, MN) and eluting the desired cell type from the column with column buffer. Eluted cells were then washed 3 times in PBS; viability and cell counts were determined by trypan blue exclusion. Cells were resuspended so that the appropriate number of cells was contained in 100 μl. Adoptive transfers of cells were made via tail vein injection using a 27-gauge needle. 

Retinitis Scoring

Eyes were fixed in buffered formalin, embedded in paraffin, and sectioned at 6 levels, 200 μm apart. The sections were then stained with hematoxylin and eosin (H&E). Changes in the posterior segment of each section were evaluated microscopically using the following scale, which was developed by Atherton and coworkers ¹⁶ and modified by Bigger and coworkers ²¹ : 0 = normal or injection artifact; 1/2 = mild atypical retinopathy—absence of cytomegaly plus retinal folds involving less than three quarters of the retinal section; 1 = moderate atypical retinopathy—absence of cytomegaly plus retinal folds involving more than three quarters of the retinal section plus photoreceptor atrophy or retinal infiltration by leukocytes involving more than one quarter of the retina; 2 = retinal infection—cytomegaly of retinal cells plus partial-thickness retinal necrosis or full-thickness necrosis extending from the ciliary body, but not beyond a one quarter retinal section from the ciliary body; 3 = necrotizing retinitis—cytomegaly plus full-thickness retinal necrosis existing further than one quarter of a retinal section from the ciliary body or full-thickness retinal necrosis extending from the ciliary body through one quarter of the section; and 4 = severe necrotizing retinitis—cytomegaly with full-thickness necrosis involving the entire retinal section. 

A score of 2 indicates virus infection of the retina adjacent to the site of injection; to differentiate between cytomegalic cells in the retina near the site of injection and the presence of cytomegalic cells and necrosis in the retina remote from the site of injection, a score of 3 or higher was considered positive for retinitis. The highest posterior segment score for each eye was the retinal score, which reflects the focal nature of this infection. Kruskal–Wallis analysis was used to determine whether there was a difference between any of the experimental groups based on the retinal scores. If a significant difference was detected by the Kruskal–Wallis analysis, the Dunn’s test was used to determine which groups were different from each other. ²² Retinitis frequencies were analyzed by two-tailed Fisher’s exact test using the Bonferroni inequality to adjust probability values to account for multiple tests. 

Results

Retinitis in CD8⁺ T Cell Depleted Mice

Atherton and coworkers demonstrated that euthymic mice depleted of both CD4⁺ and CD8⁺ T cells develop retinitis after supraciliary infection with MCMV. ¹⁶ Lu and coworkers demonstrated that thymectomized T cell–depleted mice also develop retinitis after injection of MCMV into the supraciliary space. ¹⁸ Because mice in the present adoptive transfer studies were thymectomized and rested for 28 days to allow catabolization of the injected rat anti-mouse T cell antibodies before adoptive transfer of T cells, ¹⁸ it was necessary to ensure that the prevalence of MCMV retinitis among mice used in these experiments was similar to that observed previously for non-thymectomized T cell–depleted mice. ¹⁶  

To determine whether thymectomized mice depleted of CD4⁺ cells, CD8⁺ cells, or both developed MCMV retinitis, BALB/c mice were thymectomized, depleted of CD4⁺, CD8⁺ T cells, or both by injection of monoclonal antibodies, and rested for 28 days. Control mice were thymectomized and mock depleted by injection of normal rat IgG. The mice were then injected with 9.0 × 10² pfu of MCMV via the supraciliary route. Eight days after virus injection, the mice were killed and eyes were removed for histopathologic examination. As shown in Table 1 , 5 of 10 of the CD8-depleted mice and 5 of 10 of the CD4- and CD8-depleted mice developed retinitis compared with 1 of 10 and 2 of 10 of the mock-depleted and CD4-depleted mice, respectively. Analysis of the histopathologic scores of the retinas (Table 1) by the Kruskal–Wallis test showed that there was a significant difference between the groups (P < 0.03). However, Dunn’s multiple range test was unable to show between which of the groups this difference was found. However, comparison of the retinal scores between the CD8-depleted group and either the mock-depleted group or CD4-depleted group suggests that depletion of CD8⁺ T cells was sufficient to predispose thymectomized mice to retinitis (Table 1 ; Figs. 1 A and 1B). This finding is similar to the results from experiments conducted previously in T cell-depleted euthymic mice. ¹⁶  

Adoptive Transfer of MCMV-Specific T Cells

Lu and coworkers demonstrated that adoptive transfer of MCMV-specific but unfractionated lymph node cells protected T cell–depleted mice from retinitis after supraciliary injection of MCMV. ¹⁸ To determine whether T cells in the adoptively transferred lymph node cell population were responsible for protection, adoptive transfer experiments were performed using purified T cell populations from lymph node cells of MCMV-infected or HSV-1–infected mice. Immunocompetent mice were infected in the footpad with 5 × 10⁵ pfu of MCMV or HSV-1. Seven days after infection, the lymph node cells were harvested, and the T cells were purified by a T cell enrichment column. As shown in Figure 2 , after column enrichment, the resulting T cells were at least 95% pure and had the same CD4⁺-to-CD8⁺ cell ratio as unpurified lymph node cells. MCMV-specific or HSV-1–specific purified T cells were then injected via tail vein into T cell–depleted mice. Within 2 hours of T cell transfer, 9 × 10² pfu of MCMV was injected into the supraciliary space; the injected eyes were harvested 8 days later, fixed, sectioned, stained, and scored for retinal pathology. Only 1 of 10 mice that received an adoptive transfer of MCMV-specific T cells developed retinitis (Table 2) compared with 5 of 10 mice that received HSV-1–specific T cells and 7 of 10 mice injected with PBS alone. These results show that adoptive transfer of MCMV-specific T cells, but not HSV-1–specific T cells, protected susceptible mice from retinitis (P < 0.04). Also, comparison of retinal scores between the groups again showed that the mice receiving MCMV-specific T cells, but not mice receiving HSV-1–specific T cells, had less retinal pathology than the mice receiving PBS (P < 0.04). 

Adoptive Transfer of CD8⁺ T Cells

Because depletion of CD8⁺ T cells has been shown to be sufficient to predispose mice to retinitis, we wished to determine whether adoptive transfer of CD8⁺ T cells alone would restore protection from MCMV retinitis. To determine whether the CD8⁺ T cell subset alone could protect susceptible mice from retinitis, MCMV-specific lymph node cells were fractionated on T cell subset enrichment columns. The purity of the CD4⁺ T cells and of the CD8⁺ T cells was ≥95% (Fig. 3) , and these populations contained less than 5% contaminating cells, which were CD3- (not shown). CD3⁺, CD4⁺, or CD8⁺ T cells were then injected into the tail vein of T cell–depleted mice. Because 5 × 10⁶ MCMV–specific purified T cells protected mice from retinitis after supraciliary infection with MCMV (Table 2) , and because the ratio of CD4⁺ T cells–to–CD8⁺ T cells was approximately 2:1 in unfractionated lymph node cells (Fig. 2) , the same total number of CD4⁺ or CD8⁺ T cells was adoptively transferred (i.e., 3.4 × 10⁶ CD4⁺ T cells or 1.7 × 10⁶ CD8⁺ T cells). Nine hundred plaque-forming units of MCMV was then injected into the supraciliary space; 8 days later, injected eyes were harvested, fixed, sectioned, stained, and scored for retinal pathology. As shown in Table 3 , adoptive transfer of CD8⁺ MCMV–specific T cells decreased retinal pathology compared with the PBS-treated group (P < 0.05). Additionally, the results in Table 3 again show that adoptive transfer of unfractionated MCMV-specific (CD3⁺) T cells protected T cell–depleted mice from MCMV retinitis. 

Discussion

The pathogenesis of cytomegalovirus infection in different tissue sites has been studied using mice inoculated with MCMV by several routes. MCMV infection is modulated by CD4⁺ T cells in the salivary glands and by CD8⁺ T cells in most other tissues, including adrenal glands and lungs. ^{23 24 25} Our present results address the pathogenesis of MCMV in the retina using a semiquantitative histopathologic scoring method to compare immunologically manipulated mice to controls to determine the efficacy of adoptive transfer therapies. Using the histopathologic scores, the incidence of retinitis was compared between the experimental groups as well. When compared with controls, T cell depletion (Table 1) predisposed mice to increased MCMV-induced retinal pathology; however, more importantly, adoptive transfer of MCMV-specific, but not HSV-1–specific, T cells reduced both retinal pathology and the incidence of retinitis in susceptible mice. 

Adoptive transfer of MCMV-specific CD8⁺ T cells was sufficient to reduce retinal pathology in susceptible mice (Table 3) , although it was not sufficient to reduce the frequency of retinitis. However, because the retinitis frequency is derived from the scores for retinal pathology, comparisons of the extent of retinal pathology better denote protection against MCMV disease than differences in the frequency of retinitis. 

Adoptive transfer of CD4⁺ T cells did not significantly reduce retinal pathology. Previous T cell–depletion studies conducted in the MCMV retinitis model demonstrated that, although a small proportion of CD4⁺ T cell–depleted mice develop retinitis, CD8⁺ T cells are critical for preventing retinitis. ¹⁶ In the present studies and in previous studies, small, non–statistically significant differences were found between CD4⁺ T cell–depleted and mock-depleted mice after supraciliary infection with MCMV. If the sample sizes in these experiments were larger, perhaps these small differences might be statistically significant. In human patients with AIDS, CD4⁺ T cell counts below 50/mm³ correlate with development of CMV retinitis. ^{26 27} The importance of CD4⁺ T cells in human CMV retinitis has recently been supported by a series of reports describing the recovery of CD4⁺ T cells in AIDS patients with CMV retinitis undergoing HAART; as CD4⁺ T cell counts increase, these patients are able to discontinue anti-CMV therapies without the progression of retinitis. ^{8 9 10 11} Komanduri and coworkers showed that HAART allowed CMV retinitis patients to recover CMV-specific CD4⁺ T cells. ¹¹  

The reduced numbers of CD4⁺ T cells in AIDS patients correlate with reduced numbers of CD8⁺ T cells and loss of natural killer (NK) cell cytotoxicity. ^{28 29 30 31 32} Low CD8⁺ T cell counts have been shown to correlate with the development of CMV retinitis. A recent report by Lowder and coworkers demonstrated that patients with CD8⁺ T cell numbers below 520/mm³ were at risk of developing CMV retinitis. ²⁹ The importance of NK cells in CMV retinitis has been less well studied. Natural killer cells lose cytolytic activity in stage IV AIDS patients, ^{28 32} and no studies have been reported in which NK cytolytic activity was correlated with susceptibility to CMV retinitis. Because of the close correlation between the reduction of CD4⁺ T cells, CD8⁺ T cells, and NK cell cytotoxicity in patients with stage IV AIDS, it is difficult to determine which, if not all, of these lymphocytes play critical roles in the protection of humans from CMV retinitis. Using the mouse model, our laboratory has shown in the present studies and in previous studies that NK cells and CD8⁺ T cells are critical for protecting mice from MCMV retinitis and that CD4⁺ T cells may also contribute to that protection. So, although monitoring CD4⁺ T cell levels predicts the risk of CMV retinitis, loss of CD4⁺ T cell activity, CD8⁺ T-cell activity, and/or NK cell activity may all contribute to susceptibility to CMV retinitis. 

The results presented herein show that adoptive transfer of MCMV-specific T cells and of MCMV-specific CD8⁺ T cells protected T cell–depleted mice from MCMV retinitis. Adoptive transfer of CMV-specific CD8⁺ T cell clones has been shown to prevent CMV pneumonia in human patients after bone marrow transplantation. ^{12 13} Although the pathogenesis of CMV pneumonia is different from CMV retinitis, the results of these adoptive transfer studies in mice suggest that, while CMV retinitis in immunosuppressed transplant patients is a rare event, patients receiving adoptive transfer therapy of CMV-specific CD8⁺ T cells to prevent CMV pneumonia may also be protected from ocular CMV infection. Furthermore, these experiments suggest that adoptive therapy using CMV-specific T cells might be used to protect immunocompromised patients from retinitis, especially those patients with severe reactions or resistance to current anti-CMV therapies, anti-HIV therapies, or both. 

 Supported by Grant EY09169 from the National Institutes of Health.

 Submitted for publication December 28, 1998; revised May 27, 1999; accepted June 23, 1999.

 Commercial relationships policy: N.

 

Table 1.

 

T Cell Depletion Predisposes Mice to Retinitis

Group	Retinal Score* , ‡	Retinitis Frequency, † , §
Mock-depleted	3, 2, 2, 2, 2 2, 1, 1, 1, 0	1/10
CD4-depleted	3, 3, 2, 2, 1 1, 1, 1, 1, 0	2/10
CD8-depleted	3, 3, 3, 3, 3 2, 2, 2, 2, 1	5/10
CD4- and CD8-depleted	3, 3, 3, 3, 3 2, 2, 2, 2, 1	5/10

*  Maximum posterior segment score for each injected eye.

†  A retinal score of ≥3 was positive for retinitis.

‡  Difference in experimental groups detected by Kruskal–Wallis analysis: P < 0.03, n = 10 mice per group; Dunn’s test did not determine which groups were different.

§  No statistical difference in retinitis frequency detected between CD4- and CD8-depleted group and PBS-treated group; two-tailed Fisher’s exact test with Bonferroni inequality: P < 0.282.

Figure 1.

 

Photomicrographs of the retina of a CD8⁺ T cell–depleted mouse (retinal score = 3; A) or a mock-depleted mouse (retinal score = 0.5; B) 8 days after supraciliary inoculation with 9.0 × 10² pfu of MCMV. Arrows indicate virus-infected cytomegalic cells and associated retinal necrosis. Hematoxylin and eosin; magnification,× 79.

Figure 2.

 

Dual parameter flow cytometry histograms of cells from the draining lymph nodes of mice 7 days after injection of 5 × 10⁵ pfu of MCMV via the footpad (A) and the same lymph node cells after passage through a T cell purification column (B). Flow cytometer gating was set to include large granular lymphocytes. FITC-labeled anti-CD4 is shown on the horizontal axis, and PE-labeled anti-CD8 is shown on the vertical axis. Note that the CD4⁺–CD8⁺ ratio is approximately 2:1 before and after purification.

Table 2.

 

Adoptive Transfer of MCMV-Specific T Cells Protects Mice from Retinitis

Group	Retinal Score* , ‡	Retinitis Frequency, † , §
PBS-injected	3, 3, 3, 3, 3 3, 3, 2, 1, 1	7/10
HSV-1–specific T cells	3, 3, 3, 3, 3 2, 2, 2, 1, 0.5	5/10
MCMV-specific T cells	3, 2, 2, 2, 1 1, 1, 1, 1, 1	1/10

*  Maximum posterior segment score for each injected eye.

†  A retinal score of ≥3 was positive for retinitis.

‡  Difference in experimental groups detected by Kruskal-Wallis analysis: P < 0.04, n = 10 mice per group. MCMV-specific T cell group was significantly different than PBS treated group by retinal score: Dunn’s test, P < 0.05.

§  MCMV-specific T cell group was significantly different from PBS-treated group by retinitis frequency: two-tailed Fisher’s exact test with Bonferroni inequality, P < 0.04.

Figure 3.

 

Single parameter flow cytometry histograms of cells from the draining lymph nodes of mice 7 days after injection of 5 × 10⁵ pfu of MCMV via the footpad after passage of the lymph nodes cells through T cell subset purification columns for CD8⁺ T cells (A) or CD4⁺ T cells (B). Cells were stained with PE-labeled anti-CD8 and PE-labeled anti-CD4, respectively. The purity of the CD4⁺ T cells was 95%; the purity of the CD8⁺ T cells was 96%.

Table 3.

 

Adoptive Transfer of CD8+ T Cells Protects Mice against MCMV Retinitis

Adoptive Transfer	Retinal Score* , ‡	Retinitis Frequency, † , §
PBS	4, 4, 4, 3, 3 3, 3, 3, 2, 0.5	8/10
CD4+ cells	3, 3, 3, 2, 2 2, 1, 0.5, 0.5	3/9
CD8+ cells	3, 3, 1, 1, 1 0.5, 0.5, 0.5	2/8
CD3+ cells	2, 1, 1, 1, 1, 1 0.5, 0.5, 0.5, 0.5	0/10

*  Maximum posterior segment score for each injected eye.

†  A retinal score of ≥3 was positive for retinitis.

‡  Difference in experimental groups detected by Kruskal-Wallis analysis: P < 0.004, n = 8–10 mice per group. CD8⁺ cells and CD3⁺ cell groups were significantly different than PBS-treated group: Dunn’s test; P < 0.05.

§  CD3⁺ cell group was significantly different than PBS-treated group by retinitis frequency: two-tailed Fisher’s exact test with Bonferroni inequality, P < 0.002.

The authors thank William Morgan (Department of Cellular and Structural Biology) and Susan Hilsenbeck (Department of Medicine) for their assistance with data analysis.