Amniotic Membrane Transplantation Induces Apoptosis in T Lymphocytes in Murine Corneas with Experimental Herpetic Stromal Keratitis

Abstract

purpose. To investigate the effect of human amniotic membrane transplantation (AMT) on T-cell immune response in murine corneas with herpetic stromal keratitis (HSK).

methods. Herpes simplex virus (HSV)-1–infected BALB/c mice with necrotizing HSK were treated with AMT. CD3⁺ cell apoptosis was determined in treated corneas and in vitro by flow cytometric analysis using the annexin V/7-AAD system. The effect of interleukin (IL)-2, cyclosporine, rapamycin, or Fas on T-cell survival was measured. Activation phenotype was measured by ³H-thymidine uptake and flow cytometry (CD25, CD69, major histocompatibility complex class II). Cytokine/chemokine secretion from amniotic membrane (AM)-treated corneas or draining lymph node cells was measured. The immune-modulating capacity of long-term AMT treatment and adoptive transfer of AM-treated splenocytes was tested.

results. After AMT, HSK and corneal inflammatory cell infiltration improved, and T-lymphocyte apoptosis occurred. T-cell apoptosis was also induced in vitro, independently of rIL-2, cyclosporine, rapamycin, or Fas. AMT-treated corneas and cultured lymphocytes had reduced IL-2, IL-10, IL-12, CRG-2, and CCL-2 content. Long-term AMT treatment decreased the proliferative response and type 1 helper T-cell cytokine level in draining lymph node cells. The improvement in HSK did not persist. Delayed-type hypersensitivity or HSV-1–specific cytotoxicity was not altered

conclusions. The results suggest that murine HSK improves after AMT through reduced local T-helper cell immune responses by inducing apoptosis in T lymphocytes, independently of passive apoptosis or activation-induced cell death. AM also reduces local T-helper cytokine and chemokine levels but does not result in immune deviation. Immunologic memory against HSV-1 is not affected by AMT, and long-term protection or tolerance is not induced.

Infections of the ocular surface with herpes simplex virus (HSV)-1 can cause immune-mediated chronic inflammation of the cornea, known as herpetic stromal keratitis (HSK), a major cause of unilateral blindness worldwide. Studies in the murine HSK model have revealed that the most critical event in this disease is influx of CD4⁺ T cells. The related T-helper (Th)1-type cytokines interferon (IFN)-γ and interleukin (IL)-2 have also been shown to be involved in the development of HSK. ^{1 2}  

The amniotic membrane (AM), consisting of a basement membrane and an avascular stromal matrix, is the innermost layer of the placenta. ³ It exhibits many biological properties that may be helpful in treating severe inflammatory diseases of the cornea, including the prevention of scarring ^{4 5} and inflammation, ^{6 7} improved neovascularization, ⁸ and better wound healing. ⁹ AM transplantation (AMT) can successfully treat corneal ulcers. ^{10 11} Diverse anti-inflammatory mechanisms of action have been reported in vivo ¹² and in vitro (Bültmann S, et al. IOVS 1999;40:ARVO Abstract 3044) and have included the suppression of mixed lymphocyte reactions. ¹³ Molecules with an immunoregulatory or an immunosuppressive role in T-cell function were previously found in the AM. For example, IL-1 receptor antagonist (IL-1ra) was detected in epithelial and mesenchymal cells. ^{14 15} In addition, IL-10, which is capable of inducing T-cell anergy in T lymphocytes, reducing chemokine production in corneal cells, ¹⁶ and reducing the severity of HSK in mice, ^{17 18} was also found in the AM. ⁸  

Previously, it was reported that human AMT markedly improves the necrotizing HSK lesion in mice ⁷ and humans. ^{19 20 21 22 23} Rapid reduction in inflammatory cell infiltration and rapid decrease in the number of T lymphocytes were observed within 48 hours of AMT. ²⁴ Given that the AM mechanism underlying the disappearance of lymphocytes has not yet been defined, corneas with necrotizing HSK treated with AM have now been investigated to assess the induction of lymphocyte apoptosis and to determine the proinflammatory cytokine and chemokine content. Lymphocytes cocultured with AM in vitro were analyzed with regard to apoptosis and activation properties. 

Given that T-lymphocyte apoptosis may promote immunosuppression and tolerance, ^{25 26} we also investigated whether these phenomena could be found in the HSK model after AMT. 

Materials and Methods

All experimental procedures conformed to the guidelines of the Institutional Animal Care and Use Committee and outlined in the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Mice

Female BALB/c mice (6–8 weeks of age) were purchased from Charles River Wiga (Sulzfeld, Germany) and were used for corneal infection with HSV-1 and as splenocyte and draining lymph node (DLN) cell donors for the in vitro experiments. In addition, splenocytes from naive Fas knockout mice MRL/MpJ-TNFrsf6lpr (C57/lpr), C57/Bl6 (C57; The Jackson Laboratory, Bar Harbor, ME), or BALB/c (Charles River Wiga) were used for selected experiments. 

Corneal Virus Infection and AMT

Before infecting the cornea, BALB/c mice were anesthetized with 2 mg ketamine hydrochloride (2 mg) and mepivacaine hydrochloride (400 ng) intraperitoneally. Then the central cornea was scarified, and a 5-μL HSV-1–KOS strain suspension (1 × 10⁵ plaque-forming units [PFU]/5 μL) was applied on the cornea. The HSV-1–KOS strain had been propagated on Vero cells, stored at −80°C, and quantified as previously described. ²⁷  

On day 14 after corneal infection, mice were chosen for tarsorrhaphy or AMT after severe ulcerating HSK developed. The cornea was covered with AM, with the epithelium placed face up as a temporary patch and secured by tarsorrhaphy using three interrupted 10–0 nylon sutures. Mice that underwent only tarsorrhaphy were used as a control group. 

After 2 days, the AM was removed, and mice were clinically evaluated for signs of HSV keratitis. Corneas were collected and immediately snap-frozen in liquid nitrogen. ⁷  

For long-term treatment, AMT or tarsorrhaphy was renewed every week. At 21 days, AMs were removed, animals were killed, and DLN or spleen cells were used for proliferation assay. Alternatively, AM was removed or tarsorrhaphy was reversed, and the animals were left untreated for 1 week to assess the reappearance rate of HSK lesions. 

Clinical Evaluation

The severity of stromal inflammation was judged with the use of a surgical microscope (Zeiss, Jena, Germany) on a scale of 0 to 4+, denoting corneal opacity with corneal neovascularization, edema, and thinning as follows: 1+, <25%; 2+, <50%; 3+, <75%; 4+, 75% to 100%. ²⁸  

Preparation of Human AM

Human placenta collected at elective cesarean delivery was used for this study. ^{10 29 30} In preparation the AM was flattened, with the epithelium surface facing up, onto nitrocellulose paper (Hybond N+; Amersham, Little Chalfont, Buckinghamshire, UK). AM samples were stored at −80°C in Dulbecco modified Eagle medium/glycerol 1:1 (vol/vol) until use. ⁷ Histopathologic studies and MTT (3–4,5-dimethylthiazolyl-2)-2,5-diphenyl-2H-tetrazoliumbromide) viability assay demonstrated that AM samples treated in this way did not contain viable cells after preparation and thawing (Bauer D, unpublished observation, 2007). ³¹ Immediately before use, AM was thawed, washed three times with sterile RPMI 1640 medium, and cut into approximately 1.5-cm² pieces. AM with the epithelial side facing up was placed in the cell culture plate (24-well), with the lymphocytes cultured on the bottom. AM homogenized in liquid nitrogen and sonicated for 30 seconds was used at a concentration of 10 mg/mL for the in vitro proliferation assay. 

Histology

For light microscopy, eyes were fixed in McDowell solution (4% formaldehyde, 1% glutaraldehyde, 0.13% sucrose, 0.07 M sodium hydroxide, and 0.08 M sodium phosphate, pH 7.2), rinsed in cacodylate buffer, dehydrated with ethanol, and embedded in paraffin. Five-micrometer sections were stained with hematoxylin-eosin. ²⁷  

Immunofluorescence Staining of Corneal Sections

Corneal sections were fixed for 10 minutes in acetone and then washed in PBS. The sections were blocked with 0.05 mg/mL Fc-block (BD Biosciences, Hamburg, Germany) to avoid unspecific staining. Corneal tissue was then incubated with CD3 antibody conjugated to Alexa Fluor 647 (0.005 μg/mL, clone KT3; Serotec, Oxford, UK) at 4°C for 2 hours. 

Afterward, the sections were rinsed with PBS; the nuclei were stained (Hoechst 33342, 1 μg/mL; Sigma, St. Louis, MO) and were washed again. Stained sections were mounted with glycerol-gelatin (with 0.375 μg/mL p-phenyldiamine; Sigma) and coverslipped. Isotype control antibodies conjugated with Alexa Fluor 647 served as controls. ²⁴ Four tarsorrhaphy- or AMT-treated corneas per group were analyzed. 

Confocal Microscopy

Wholemounted corneas were dissected (cryosection; 7 μm) and examined under a confocal laser scanning microscope (LSM 510; Zeiss, Jena, Germany). Images were taken at medium power magnification with a 20× objective. Specimens were scanned with a helium-neon 2 laser at 633-nm or with an enterprise argon ultraviolet (UV) light laser at 364-nm wavelength. The differently colored scans were merged into one picture. 

Cell Culture

Lymphocytes from HSV-1–infected animals on day 14 after infection or from noninfected animals were harvested from the ipsilateral regional lymph nodes (DLN) or the spleens, separated by gradient centrifugation, and pooled. To each well of a 24-well plate, 5 × 10⁶ cells were added. Then, 2 × 10⁷ PFU of UV-inactivated HSV-1 (UV-HSV-1), concanavalin A (ConA, 5 μg/mL; Sigma), or medium (control) was added. To selected wells (in triplicate), 100, 10, or 1 μg/mL cyclosporine (CsA; Sigma) or 0.2, 0.02, or 0.002 μg/mL of rapamycin (Sigma) was added to modify the T-lymphocyte cell responses. 

Cytokine Quantification by Enzyme-Linked Immunosorbent Assay

Individual corneas (n = 8, each group) obtained from tarsorrhaphy- or AMT-treated corneas were excised and frozen at −80°C until use. On the day of the assay, corneas were thawed on ice, minced in 1 mL PBS, sonicated for 30 seconds, and clarified at 10,000g for 10 minutes The homogenate from each cornea was studied to determine the content of IFN-γ, IL-2, IL-4, IL-10, IL-12, CRG-2 (IP-10), and CCL-2 (MCP-1) with the use of commercially available ELISA kits (OptEIA [PharMingen, Hamburg, Germany] or Duoset [R&D Systems GmbH, Wiesbaden-Nordenstadt, Germany]). 

Lymphocyte cell culture supernatants were obtained from DLN or spleens that had been cocultured with UV-HSV-1 or ConA (5 × 10⁶ cells/mL from HSV-1–infected mice, day 14 after infection). After 24 hours of incubation at 37°C in 5% CO₂, the supernatants were analyzed to determine the cytokine levels, as described previously. ³²  

Proliferation Assay with ³H-Thymidine

DLN and splenic cells obtained from HSV-1–infected mice (day 14 after infection) were used for lymphocyte proliferation assays, as previously described. ³²  

MTT Viability Assay

To determine the effect of AM on the viability of lymphocytes, 5 × 10⁶ cells/mL (DLN cells from HSV-1–infected mice, day 14 after infection) were cocultured with AM. The MTT viability assay measures the conversion of MTT to colored formazan. Some wells were treated with recombinant IL-1α, IL-1β, IL-2, IL-4, IL-10, or supernatants collected from ConA-activated splenocytes. After incubating the DLN cells or splenocytes with AM or medium, 10 μL of 5 mg/mL MTT solution was added to 100 μL of the incubation medium, and cells were incubated in a CO₂ incubator for 2 hours. Then, 100 μL lysis buffer (20% sodium dodecyl sulfate and 50% dimethyl formamide) was added to the cells, which had been incubated overnight at 37°C. Finally, optical density (OD) was measured at 570 nm (reference wavelength, 690 nm) by an ELISA reader (MRX; Dynatech Laboratories, Chantilly, VA). 

Hoechst Staining

After incubation with AM or medium for 24 hours, DLN cells (5 × 10⁶ cells/mL) within the chamber slides (Nalge Nunc International, Rochester, NY) were stained with 1 μg/mL DNA dye bisbenzamide (Hoechst 33324, B-2261; Sigma) in PBS. Nuclei were viewed under a fluorescence microscope (BX40; Olympus, Tokyo, Japan). Apoptotic features of cells included condensed and fragmented highly fluorescent nuclei. 

DNA Laddering Assay

After stimulation with UV-HSV-1 or ConA and cocultivation with medium or AM for 24 hours, DLN or spleen cells at 14 days after infection were centrifuged, and the pellet (5 × 10⁶ cells) was used to isolate DNA with the use of a commercially available kit (DNeasy; Qiagen, Hilden, Germany). The isolated DNA was subjected to electrophoresis in a 2% agarose gel containing 5 μg/mL ethidium bromide. DNA laddering, indicating DNA fragmentation, was visualized by UV light transillumination. 

Flow Cytometry

HSK corneas were excised 12 hours after tarsorrhaphy or AMT and incubated in RPMI at 4°C for 2 hours. Corneas were separated from the underlying endothelium, and eight corneas per group were digested in 80 to 100 U/cornea collagenase type 1 (Sigma-Aldrich, Hilden, Germany) for 2 hours at 37°C. Single-cell suspensions were prepared with the use of a 40-μm cell strainer cap (BD Biosciences) and washed with PBS. The suspension was incubated with CD16/CD32 (FCγIII/II receptor block; BD Biosciences), then stained with CD45 (FITC, clone 30-F11; BD Biosciences), annexin-5 (PE), 7-aminoactinomycin (7-AAD; BD Biosciences), and CD3 (Alexa-Fluor 647, clone KT3; Serotec). Experiments using unstained specimens and isotype controls (BD Biosciences) were performed to exclude unspecific bindings of antibodies. The suspension was analyzed by flow cytometry (FACSCalibur; Becton Dickinson, San Jose, CA), counting 1.5 × 10⁵ cells in each sample. ³³  

Single-cell suspensions from DLN or spleens of mice were collected at 14 days after infection and stained with monoclonal antibodies (BD PharMingen): anti-CD3e (clone 145-2C11), anti-CD25 (IL-2R, clone 7D4), anti-CD69 (very early activation antigen; clone H1.2F3), or anti-major histocompatibility complex (MHC) class II (I-Ad/I-Ed, clone 2G9). Cells were then analyzed with a cytometer (FACSCalibur), counting 2 × 10⁴ cells in each sample. 

Signs of early apoptosis were detected with an apoptosis detection kit (Annexin V-PE Apoptosis Detection Kit; BD PharMingen) to localize membrane phospholipid phosphatidylserine and the vital dye 7-AAD. ³⁴  

Delayed-Type Hypersensitivity Reaction

Delayed-type hypersensitivity (DTH) responsiveness was determined in HSV-1–infected mice (n = 8) using the footpad swelling assay, as described previously. ³² Fifty microliters UV-inactivated virus solution was inoculated in the right hind footpad using a 27-gauge needle, and the injection of 50 μL RPMI in the left hind footpad was taken as a control. Footpad swelling was measured with a micrometer after 24 hours. Results were expressed as specific footpad swelling in millimeters. 

In Vivo Cytotoxicity

To determine whether increased HSV-1–specific cell cytotoxicity was present after AMT, an in vivo cytotoxicity assay was performed as described previously. ²⁶ Splenocytes obtained from normal or HSV-1–infected BALB/c mice (collected on day 14 after infection) were separated by Ficoll gradient centrifugation and suspended at 5 × 10⁷ cells/mL in RPMI at 37°C. From the 5 mM stock-CFSE (carboxyfluorescein succinimidyl ester), 1.25 μL/mL was added to the splenocytes obtained from HSV-1–infected mice (CFSE^high population), and a 10-fold dilution of this stock CFSE-solution was added to splenocytes obtained from uninfected BALB/c mice (CFSE^low population). Cell solutions were incubated for 10 minutes at 37°C. The reaction was stopped with cold PBS and washed three times. Then 5 × 10⁷ splenocytes from HSV-1–infected mice (target cells) and 5 × 10⁷ splenocytes from normal mice (reference cells) were injected retro-orbitally into mice with ulcerative HSK after long-term AMT or tarsorrhaphy (day 20 after AMT or T). Spleen or DLN cells were harvested 24 hours later and were analyzed separately by flow cytometry. The number of recovered target cells was enumerated by examining 5000 events. Three mice per group were analyzed, and the percentage killing in each group of mice was determined. ²⁶  

Statistical Analysis

All experiments were performed twice. Mean ± SEM values were calculated. Wherever specified, data were analyzed statistically by Student’s t-test. Fisher protected least significant difference test was used to determine statistical differences between mean values of clinical keratitis scores. P < 0.05 was regarded as significant. 

Results

Reduction of T Lymphocytes in Corneas with HSK Lesions after AMT

Fourteen days after infection, all corneas showed severe HSK with heavy stromal necrosis and ulceration. However, stromal inflammation and ulceration dramatically improved after 2 days in the AMT group compared with the tarsorrhaphy group, in agreement with findings from previous publications. ⁷  

Light microscopy showed that corneas from mice treated for 2 days with tarsorrhaphy still contained many inflammatory cells (Fig. 1A) , whereas the number of cells had strongly decreased in corneas that were treated with AMT (Fig. 1B) . In the tarsorrhaphy group, many infiltrating cells were still viable (Fig. 1C) , whereas many of the inflammatory cells found after AMT showed evidence of apoptotic cell death (e.g., shrunken and condensed nuclei; Fig. 1D ). For the analysis of flatmount sections, laser scanning microscopy was chosen for better detection of inflammatory cell infiltration into the cornea. Results showed that the corneas of the tarsorrhaphy group contained many CD3⁺ cells (Fig. 1E) , whereas corneas of the AMT group contained far fewer CD3⁺ cells (Fig. 1F) . 

Induction of Apoptosis in Lymphocytes by AM in HSV-1–Infected Corneas and In Vitro

We hypothesized that many CD3⁺ T cells in the corneas treated with AMT would be composed of T cells with apoptotic properties, whereas T lymphocytes in corneas with HSK and tarsorrhaphy would be composed mainly of viable T cells. To test this, cell suspensions obtained from HSK corneas treated with tarsorrhaphy or AMT were stained with CD45 (FITC), annexin V (PE), 7-AAD, and CD3 (Alexa Fluor 647). HSK corneas treated with AMT contained a higher CD3⁺dead/viable ratio (percentage of apoptotic CD3⁺ cells: first experiment: tarsorrhaphy, 31.1%; AMT, 37.5%; second experiment: tarsorrhaphy, 46.1%; AMT, 54.1%; third experiment: tarsorrhaphy, 45.9%; AMT, 51% [Figs. 2A 2B ]). 

We next examined whether the induction of apoptosis in T lymphocytes could also be induced in vitro. DNA from murine DLN cells or splenocytes, which were cocultivated with AM, was isolated and used for the DNA laddering assay. Results showed that murine AM-treated DLN (Fig. 3A)or splenic cells (data not shown) had bands typical for apoptosis, and DNA laddering was detected in cells cultured together with medium, UV-HSV-1, or ConA. In contrast, only slight DNA laddering bands were detected when cells had been incubated without AM. 

Although cells obtained from DLN at day 14 after infection that had been cocultured with AM and stained with staining reagent showed typical apoptotic features (e.g., fragmented nuclei), cells cultured without AM showed no such morphology (Figs. 3B 3C) . 

After diverse incubation times with AM (2 or 24 hours), CD3⁺ DLN cells were stained with annexin-V/7-AAD. After 2 hours, typical signs of early apoptosis were observed, such as high annexin-V and low 7-AAD–positive cells. After 24 hours, cells also stained positively for 7-AAD (Fig. 3D) . 

The viability of cells cocultured with AM for various time points was assessed with the MTT test. When DLN cells had been cocultured with AM, a continuous decrease in MTT conversion was observed. After 13 hours of incubation, no viable cells were left in those wells with AM, as shown by MTT conversion assay (Fig. 3E) . The addition of rIL-1α, -1β, -2, -4, -10, or supernatants collected from ConA-activated splenocytes to the lymphocyte and AM culture did not inhibit cell death. Homogenates from HSV-1–infected corneas with severe HSK were also unable to prevent cell death (Fig 3F) . 

To elucidate whether lymphocytes were killed by mechanisms other than apoptosis, DLN cells were incubated with AM, and cell membrane integrity was tested by trypan blue exclusion. Results showed that cell membrane degradation was not induced by AM on DLN cells in vitro (data not shown). To determine whether cells other than lymphocytes were susceptible to apoptosis induction after AM coculture, the murine aneuploid fibrosarcoma cell line L929 was incubated with AM. As detected with the MTT test, cell viability did not decrease (data not shown). Taken together with the results obtained from the other tests, these data suggest that AM induces the apoptosis of T lymphocytes. 

Role of Activation-Induced Cell Death in AM-Induced Apoptosis

To elucidate whether AM-induced apoptosis of lymphocytes could be induced by activation-induced cell death (AICD), DLN cells from HSV-1–infected animals were cocultured with AM. Various concentrations of CsA or rapamycin were used to interfere with the first step (through T-cell receptor) or second step (through IL-2 receptor) of lymphocyte activation, respectively. 

Results of the MTT tests showed that cell death induced by AM is not affected by activation through HSV-1 or even ConA. Furthermore, the addition of CsA (Fig. 4A)or rapamycin (Fig. 4B)did not antagonize the cell death-inducing effect of AM. Although CsA or rapamycin decreased the survival of DLN cells compared with controls without treatment, this was not detected in the cells cocultured with AM. DLN cells were almost completely nonviable after 24 hours of AM coculture. CsA or rapamycin decreased the secretion of IL-2 in the DLN cells, and a reduction of IL-2 could also be found when DLN cells had been cocultured with AM. Combined cocultivation with AM and CsA showed a further reduction in IL-2 expression only after stimulation with ConA. A further decrease in IL-2 secretion was not found when rapamycin was used instead of CsA. 

The next experiments were performed to elucidate whether apoptosis induction in lymphocytes from AM is restricted to the Fas and Fas ligand (FasL) system and whether the CD3⁺ DLN-cells would upregulate FasL expression after cocultivation with AM. After 24 hours of cocultivation with AM, CD3⁺-DLN cells were entirely apoptotic, but few cells expressed FasL receptor (Fig. 4C) . 

The induction of lymphocyte apoptosis with AM was also investigated using splenic cells from Fas knockout mice MRL/MpJ-TNFrsf6lpr (C57lpr) lacking a functional Fas receptor and C57/Bl6 (C57) or Balb/c as controls. MTT tests showed that the cells collected from C57lpr mice also underwent cell death when cocultivated with AM and that fewer of these cells survived compared with cells from C57/Bl6 or Balb/c mice (Fig. 4D) . These data suggest that the AM contains a mechanism to induce the apoptosis of T lymphocytes and that this mechanism is independent of AICD. 

Expression of Proinflammatory Cytokines and Chemokines in the Cornea

Corneas collected from mice after tarsorrhaphy or AMT were analyzed for their content of Th1 (IFN-γ, IL-2) and Th2 (IL-4, IL-10) cytokines and of IL-12. Levels of CCL2 and CRG-2 were also determined. Here, the amounts of IL-12, -2, and -10 were decreased (Fig. 5) . Furthermore, the CCL-2 and CRG-2 content had decreased after AMT compared with the tarsorrhaphy group of mice (Fig. 5) . Therefore, AMT impairs part of the adaptive immune response in corneas with ulcerative keratitis. 

Activation Phenotype of Lymphocytes Cocultured with AM

The proliferative response of splenocytes or regional lymph node cells from HSV-1–infected BALB/c mice in the cell culture was determined in the presence of AM in vitro. Experiments showed that cocultivation with AM significantly reduced the proliferative response of medium, HSV-1–specific, and ConA-activated lymphocytes (Fig. 6A) . 

The influence of AM on the secretion of typical cytokines such as IFN-γ, IL-2, IL-4, and IL-10 from DLN lymphocytes into the supernatant was determined. Here, the secretion of IFN-γ, IL-2, and IL-10 was significantly reduced after cocultivation with AM (Fig. 6B) , whereas that of IL-4 was not significantly affected. 

Flow cytometric analysis of CD3⁺ DLN cells showed that the surface molecules CD25 (IL-2R), CD69 (Very Early Activation antigen), and MHC class II were all reduced after 24 hours of cocultivation with AM (Figs. 6C 6D 6E) . After only 4 hours of cocultivation, however, the expression of CD25 and of MHC class II was increased in DLN lymphocytes. 

AMT Improves HSK but Does Not Induce Tolerance or Long-term Protection against Corneal Lesions

To elucidate whether a tolerance mechanism could be associated with AM, we evaluated the long-term course and reappearance of HSK after AMT. Groups of mice with ulcerative HSK were treated for 3 weeks with AMT or tarsorrhaphy. Cells from the DLN and the spleen had significantly reduced proliferative responses after HSV-1–specific or ConA challenge (Figs. 7A 7B) . Analysis of the Th1/Th2 cytokine profile in DLN cells showed that IFN-γ content and IL-2 content were significantly impaired after stimulation with HSV-1 antigen. (Figs. 7C 7D) . We did not find a respective influence on spleen cells or any difference in the DTH response even after extended treatment with AMT compared with tarsorrhaphy (Fig. 7E) . The in vivo cytotoxicity test showed that there was no sign of any specific depletion of HSV-1–specific cells compared with naive cells and confirmed that AMT does not impair memory against HSV-1 (Fig. 7F) . 

Furthermore, the HSK lesions always reappeared after a period of 5 days when the AM was removed (data not shown). These data strongly suggest that tolerance or long-term protection was not induced by AMT. 

Discussion

Previous studies have shown that AMT can be used to treat murine and human HSK, which is a T cell–mediated disease. ^{7 19} We found that AMT strongly reduced the number of CD3⁺ T lymphocytes in the cornea, and increased lymphocyte apoptosis was observed in corneas with HSK and in T cells harvested from these corneas early after AMT. An earlier study reported that AM collected from patients treated with AMT contained varying degrees of inflammatory cells (positively stained with CD14, CD4⁻, and CD8) in the AM stroma. Many of these cells also stained positively with TUNEL, suggesting apoptotic cell death. ³⁵ However, the AM collected from HSK mice did not contain significant numbers of inflammatory cells, suggesting that attraction of cells into the AM may not be the main anti-inflammatory mechanism in the HSK model (data not shown). Our findings suggest that soluble factors from the matrix may be responsible for the observed improving effects on ulcerating HSK and for the immune-modulating action on inflammatory cells. ^{13 36}  

Apoptosis was also induced by AM in isolated lymphocytes, indicating that the AM interacts directly with these cells and that the presence of corneal tissue, known to express surface molecules such as FasL in healthy corneas, is not essential. ³⁷ Furthermore, these results indicate that lymphocyte apoptosis induced by AM occurs primarily independently of cytokines, thereby excluding passive apoptosis as a result of a lack of IL-2. AICD was excluded as another possibility for lymphocyte apoptosis from AM. For AICD to be induced in nontransformed T lymphocytes, a combination of two mitogenic stimuli at relatively high doses is required, an antigen or mitogen that acts on the T-cell receptor (TCR) and a lymphokine such as IL-2. It was demonstrated that AICD involved induction of the Fas ligand and, thus, induction of Fas-mediated apoptosis. Hence, pharmacologic interference with the IL-2 pathway in activated T cells, such as by the addition of CsA or rapamycin, typically prevented Fas-dependent lymphocyte depletion. ^{38 39}  

Treatment with CsA or rapamycin did not protect the lymphocytes against AM-mediated cell death in our experiments. Furthermore, FasL was not induced with AM, and splenocytes from Fas knockout mice (C57lpr) were not resistant to AM-induced apoptosis. Consequently, AM-mediated apoptosis in lymphocytes does not appear to be related to AICD. 

Previous studies have shown that the expression of IL-8, Gro-α, and epithelial cell-derived neutrophil attractant (ENA) was suppressed in keratocytes cultured on AM stromal matrix (Bültmann S, et al. IOVS 1999;40:ARVO Abstract 3044). ⁶ We observed that CRG-2/IP-10 and CCL-2/MCP-1 levels were significantly lower in the murine corneas with HSK after AMT. IFN-γ–inducible protein 10 (CRG-2/IP10) is a chemokine that preferentially attracts Th1 lymphocytes through its receptor CXCR3, which is expressed at high levels on these cells. This is in agreement with previous observations that the treatment of HSV-1–infected mice with antibody to IP-10 significantly reduced ICAM-1 and CXCR3 transcript expression, MIP-1α and IP-10 levels, and corneal disease. ⁴⁰ The monocyte chemoattractant protein-1 (CCL-2/MCP-1) plays a crucial role in recruiting monocytes and lymphocytes during inflammatory responses. ⁴¹ It has been shown that CCL-2/MCP-1 depletion resulted in innate immune modulation and increased corneal inflammation as a result of MIP-2 overexpression. ⁴² In fact, we found increased amounts of MIP-2/CXCL2 in murine corneas with HSK and AMT. ²⁴ Neutralizing antibodies to MIP-2 suppressed HSV-1–specific DTH and sharply reduced neutrophil accumulation in the ear pinna. ⁴³  

Recently, it has been shown that AM is able to suppress IL-1 mRNA transcription and IL-1 expression in human limbal epithelial cells. ⁶ In our experiments, we found that the expression of IL-12, a cytokine produced by a variety of immune effector cells and thus leading to a Th1 response in HSK, ^{44 45} was downregulated in the corneas after AMT. This suggests that, in addition to apoptosis induction, other AM mechanisms are in play and affect lymphocyte activation and influx into the cornea. 

Thus, the typical Th1 cytokine IL-2 was found to be suppressed in AMT-treated corneas. In the experimental HSK model, inflammation could be prevented by neutralizing IL-2 or disease remission could be induced. Indeed, neutralizing IL-2 resulted in decreased IFN-γ production, reduced chemotaxis, and loss of PMN viability in the infected cornea. ^{2 46 47} The fact that the anti-inflammatory cytokine IL-10 was also decreased after AMT in our experiments was unexpected because it was previously reported that IL-10 treatment can prevent HSK from developing and this was associated with reduced chemokine production. ^{17 48 49} A previous study showed that human AM can suppress alloreactive T-cell synthesis in Th1- and Th2-type cytokines in vitro. ¹³ Our in vitro experiments showed that AM decreased DLN cell proliferation and production of IFN-γ, IL-2, and IL-10. AM also decreased the activation markers CD25, CD69, and MHC class II on DLN cells 24 hours after HSV-1 antigen- specific and antigen-unspecific activation. However, an increase in CD25⁺ and MHC class II+ surface molecules on DLN cells was present after 4 hours. These findings are consistent with the induction of apoptosis, as others have reported that the generation of inflammatory signals can be an intrinsic component of the apoptotic machinery. ^{50 51}  

Other studies have shown that apoptosis of lymphocytes can also result in immune tolerance. ^{52 53} Mechanisms proposed for the tolerogenic nature of apoptosis included deletion of reactive clones, ²⁶ anergy, ⁵⁴ immune deviation (Th1 to Th2), ⁵⁵ and active regulation by regulatory T cells. ⁵⁶  

Previously we demonstrated that the immune response in the spleen was not affected 2 days after AMT treatment of mice with HSK. ⁵⁷ Given that AM may remain for weeks or even months in the clinical setting ¹⁹ and given that the systemic immune response against HSV-1 is also important to control virus latency, we determined how long-term AM treatment might affect the immune response in the regional and systemic immune system. Proliferation and IFN-γ and IL-2 secretion in DLN cells was reduced. Results show that even prolonged AMT predominantly affected the regional (Th1) immune responses. These findings could, at least in part, be the result of decreased cytokine secretion (e.g., IL-12) in corneas with HSK after AMT, ultimately decreasing proliferation and Th1-cytokine production in the DLN. 

In agreement with our previous data, ⁵⁷ we found that DTH responses after AMT were comparable with those after tarsorrhaphy. In addition, the in vivo cytotoxicity assay showed no evidence for a HSV-1–specific depletion of splenocytes after AMT from HSV-1–infected mice. Thus, even long-term AMT did not affect immunologic memory against HSV-1. Accordingly, immune-mediated HSK recurred shortly after the end of long-term application of AMT. 

We also assessed whether adoptive transfer of AM-induced apoptotic splenocytes could induce signs of tolerance induction or immune suppression. Our observations revealed that a transient reduction in antigen-specific and spontaneous proliferation of DLN or splenocytes was induced, but the HSK lesions had not improved (data not shown). As a consequence, tolerance induction by regulatory cells or by systemic cell depletion does not appear to be the primary mechanism underlying the improvement in HSK after AMT. 

We did find some evidence that innate immunity was stimulated (e.g., CXCL2/MIP-2). ²⁴ Given that virus infections are known to elicit immunologic “danger” signals, these signals might prevent the acquisition of long-term tolerance or long-term protection in experimental HSK. 

It is critical that apoptotic cells be removed during tissue remodeling or while inflammation is resolving to preserve normal tissue structure and function. ⁵⁸ Engulfment of apoptotic cells by professional and nonprofessional phagocytes sets up an anti-inflammatory environment within the tissue that is mediated in part by the autocrine/paracrine action of TGF-β. ^{58 59} Others have found that innate immune responses can also be induced by apoptotic cells, as CXCL2/MIP-2 is secreted from peritoneal macrophages after ingestion of apoptotic T cells that also support leukocyte infiltration to the site of apoptosis in some situations. ⁶⁰  

The mechanism underlying T-cell apoptosis induction by AM is still not known and most likely involves multiple factors from AM. ⁴ Furthermore, it has been shown that AM can also be antiapoptotic, as in corneal epithelial cells. ⁶¹ Diverse cell signals that may originate from extrinsic inducers, such as toxins, hormones, growth factors, nitric oxide, or cytokines, might control apoptosis. ^{62 63} Alternatively, apoptosis could also be induced by intracellular cell signaling, such as through glycocorticoids, radiation, nutrient deprivation, heat, hypoxia, viral infection, and increased intracellular calcium concentrations. ⁶⁴ Previously, it was reported that amniotic epithelial cells produce many immunomodulatory factors, including TNF-α, FasL, tumor necrosis factor-related apoptosis-inducing ligand (TRAIL), TGF-β, macrophage migration-inhibitory factor (MIF), and other factors that may induce apoptosis in T and B cells. ⁶⁵ Future studies must define the molecules that induce these immune modulatory events after AMT in HSK mice. 

 Supported by the Ernst and Berta Grimmke Foundation and by Deutsche Forschungsgemeinschaft Grants Ba 2248/1-1, Ba 2248/1-2, and He 1877/12-1.

 Submitted for publication October 21, 2008; revised December 19, 2008, and January 19, 2009; accepted May 13, 2009.

 Disclosure: D. Bauer, None; S. Wasmuth, None; M. Hennig, None; H. Baehler, None; K.-P. Steuhl, None; A. Heiligenhaus, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Improvement in HSK after treatment with AMT. Influence of AMT on inflammatory cells in murine corneas with experimental ulcerative HSK. (A) Murine corneas with HSK and tarsorrhaphy (T; 48 hours). (B) Murine cornea with HSK and AMT (48 hours). Although severe cell infiltration was found in corneas treated with tarsorrhaphy, it was markedly reduced after AMT. (C) Cornea with tarsorrhaphy. Cornea contained many inflammatory cells, primarily PMNs. (D) Cornea with AMT. Cornea contained many nonviable inflammatory cells with apoptotic traits, such as shrunken and condensed nuclei (arrows). (E) Many CD3⁺ cells were present within the area of corneal ulceration in the stroma of murine corneas with HSK and tarsorrhaphy. Flatmount section (original magnification, 200×). (F) Murine cornea with AMT. Corneas after AMT treatment contained far fewer CD3⁺ cells within the area of ulceration after 48 hours. (G) Immunofluorescence staining of the central corneal stroma (flatmount sections) against CD3 (Alexa Fluor 647) and stain after tarsorrhaphy or AMT (48 hours). Data are expressed as mean ± SEM (*P < 0.05).

Figure 2.

 

Induction of T-lymphocyte apoptosis in vivo. Cornea specimens obtained from tarsorrhaphy (T) or AMT corneas (eight corneas were pooled for each group after 12 hours) were stained with CD45 (FITC), Annexin V (PE), 7-AAD, and CD3 (Alexa Fluor 647). CD45⁺CD3⁺ cells were detected in cornea samples treated with tarsorrhaphy (T) or AMT. More CD45⁺CD3⁺ cells after AMT had also stained positively for annexin V and the nuclear dye 7-AAD, indicating cell death (early apoptosis, annexin V+/7-AAD−; late apoptosis, annexin V+/7-AAD+). Tarsorrhaphy, 45.9%; AMT, 51%. Viable cells with intact membrane did not stain by annexin V and excluded 7-AAD. Tarsorrhaphy, 52.9%; AMT, 46.7%.

Figure 3.

 

Induction of T-lymphocyte apoptosis in vitro. (A) Agarose gels containing electrophoretically separated low-molecular–weight DNA fractions from untreated (Med), UV-HSV-1, or ConA-activated lymph node cells. Samples collected from cells that had previously been cocultured with human AM showed short DNA bands, indicating DNA fractionation. In the control without AM, these bands were markedly less distinctive. Samples on chamber slides were collected 24 hours after treatment with AM or medium. Slides were then stained, and were viewed under a fluorescence microscope. (B) Lymphocytes with medium. (C) Lymphocytes with AM. (D) FACScan analysis with annexin V/7-AAD staining from T lymphocytes cocultured with medium or AM. Results showed that most CD3⁺ DLN cells stained positively for annexin-V and 7-AAD after 24 hours of AM cocultivation. (E) The vitality of DLN cells at different points in time after AM-homogenate treatment was analyzed by MTT conversion experiments. Results indicate that MTT conversion was strongly decreased when cells were cocultured with AM. Control versus AM-treated cells. *P < 0.05. (F) Influence of rIL-2 on AM-mediated cell apoptosis. Recombinant cytokines were not able to rescue cells from apoptotic cell death induced by AM. The same results were found when IL-1α, IL-1β, IL-4, IL-10, supernatants from ConA-activated splenocytes, and homogenate of a cornea with HSK on cocultivation of DLN cells were used instead of rIL-2. Control versus AM-treated cells with or without recombinant cytokine or supernatants. *P < 0.05. AM − AM+recombinant cytokine or supernatants: statistically not significant.

Figure 4.

 

Influence of AICD on AM-mediated cell death in lymphocytes. DLN cells were collected from HSV-1–infected mice and used for cocultivation experiments with AM and (A) cyclosporine (CsA) or (B) rapamycin (Rap) at different concentrations. The vitality of cells was tested by the MTT conversion assay (top). IL-2 content was measured by ELISA (bottom). Results show that neither CsA nor Rap treatment could protect cells against AM-induced cell death, and DLN cells produced less IL-2 when treated with AM. When AM and CsA were cocultured, the amount of IL-2 was lower only when cells were stimulated with ConA. CsA or rapamycin did not reduce IL-2 content in the other settings, indicating that AM rapidly induced apoptosis in T lymphocytes. Data are the mean ± SEM. Control versus treated cells: *P < 0.05. (C) Expression of FasL on DLN cells after cocultivation with AM was assessed by flow cytometry. Results show that CD3⁺DLN cells do not express FasL after 24 hours of cocultivation with AM. (D) Functional deletion of Fas in C57lpr mice (knockout mice strain used: MRL/MpJ-TNFrsf6lpr) did not result in resistance of splenic cells to AM-mediated apoptosis.

Figure 5.

 

Immune modulation in the cornea with HSK after AMT. Effect of human AMT on the expression of the cytokine and chemokine profile. Compared with the control group with tarsorrhaphy (T), AMT-treated mice showed a decreased amount of IL-2, IL-10, CRG-2 (IP-10), CCL2 (MCP-1), and IL-12 in the corneas. *P < 0.05. IFN-γ and IL-4: not statistically significant (P > 0.05).

Figure 6.

 

Activation phenotype of lymphocytes cocultured with amniotic membrane. (A) Proliferative response was analyzed by the uptake of ³H-thymidine. Cells that were cocultured with AM homogenate (shown) or pieces of AM (data not shown) demonstrated decreased uptake of ³H-thymidine. *P < 0.05. (B) Cytokine secretion of DLN cells cocultured with amniotic membrane. Compared with the control cells with AM decreased amounts of IFN-γ, IL-2, and IL-10 were produced. *P < 0.05. Surface molecule expression of CD25 (C), CD69 (D), and MHC class II (E) on CD3⁺ cells was analyzed by flow cytometry 4 hours and 24 hours after cocultivation. CD3⁺ DLN cells cocultivated with AM contained fewer CD25⁺, CD69⁺, and MHC class II⁺ cells, even when cells were cocultured with UV-HSV-1 (data not shown) or ConA. After 4 hours of AM cocultivation, a slight increase in CD25⁺ (C) and MHC class II⁺ cells (E) was observed.

Figure 7.

 

Influence of long-term AMT treatment in mice with HSK on the regional and systemic immune response. (A, B) Proliferative response to HSV-1-antigen or ConA in DLN or splenic cells after long-term (21-day) treatment with tarsorrhaphy (T) and AMT. The proliferative response was significantly decreased in the (A) DLN or (B) spleen after long-term AMT. Data are the mean ± SEM. *P < 0.05. (C, D) Cytokine production (IL-2, IFN-γ, IL-4, or IL-10) in response to HSV-1 antigen or ConA in supernatants of DLN or splenic cells after long-term (21-day) tarsorrhaphy (C) or AMT (D). After long-term AMT, a decrease in IFN-γcould be found in the DLN. The reduction in IFN-γafter AMT in the spleen was lower than the reduction found in the DLN. Data are mean ± SEM. *P < 0.05. (E) DTH response of mice with ulcerative HSK treated with long-term tarsorrhaphy or AMT. HSV-1–specific DTH 21 days after AMT or tarsorrhaphy indicates no effect. Data are mean ± SEM. Difference between tarsorrhaphy and AMT was not statistically significant (P > 0.05). (F) In vivo cytotoxicity of HSV-1–specific cells by AMT. Numbers of AMT mice (n = 5) were compared with tarsorrhaphy mice (n = 5). Percentage killing shown in the figure represents the average of mice in the group. Difference between tarsorrhaphy and AMT was not statistically significant (P > 0.05).