Abstract
purpose. The Th2-biased immune system can promote penetrating keratoplasty survival in mice. A series of experiments were performed to determine whether this system could prolong corneal limbal transplant (LT) survival.
methods. BALB/c (H-2d) mice were immunized with 50 μg of keyhole limpet hemocyanin (KLH) in incomplete Freud’s adjuvant. Four weeks later, the corneal epithelium, including the limbal area, was removed, and the mice received LT from B10.D2 (H-2d), C57BL/10 (H-2b), or enhanced green fluorescence protein (EGFP) transgenic (H-2b) donor mice. Immediately thereafter, recipient mice were immunized with 50 μg of KLH or Hanks’ balanced salt solution (HBSS; control) in complete Freund’s adjuvant. The allograft fates were assessed clinically. Lymphocytes of recipients were examined for donor-specific proliferation and for donor-specific cytokine production in vitro.
result. The regenerated epithelia of all C57BL/10 (n = 14) and B10.D2 (n = 18) grafts were rejected swiftly in control mice, whereas 66.6% of C57BL/10 grafts (8/12, P < 0.001) and 62.8% of B10.D2 grafts (22/35, P < 0.001) in the KLH immune group remained significantly clear for 8 weeks. Moreover, EGFP donor epithelial cells were detected from the healthy corneas of KLH-immunized mice. As for the in vitro assay, at 1 week after B10.D2 grafting, lymphocytes from KLH-immunized groups showed neither proliferation nor increased cytokine secretion.
conclusions. The Th2-biased immune system can support LT prolongation irrespective of donor disparity and can suppress corneal neovascularization. This prolongation is not due to induction of donor-specific regulatory cells, but is presumably at least associated with the suppression of allosensitization.
Limbal transplantation (LT) is universally performed to reconstitute the ocular surface in patients who have no corneal epithelium stem cells, such as after a chemical burn or Stevens-Johnson syndrome.
1 2 3 If ocular surface damage is bilateral, allogeneic LT is the best means of supplying healthy corneal epithelial and limbal cells, including stem cells. The renewed allogeneic epithelium has been shown to remain viable only under attentive postoperative care, with comparatively long-term use of immunosuppressants to avoid intense and frequent rejection. However, the local and/or systemic use of corticosteroids or alternative general immunosuppressants is associated with significant complications.
4 5 Even with the use of immunosuppressants, certain patients exhibit intense postoperative corneal allograft rejection.
6 Moreover, Mills et al.
7 recently reported, using the rat model of LT, that immunosuppression could mediate clinical allograft survival, but not donor cell survival on the ocular surface. It is therefore apparent that development of immunologic strategies that can suppress allograft rejection may contribute to ocular surface reconstruction.
Previously, Yao et al.
8 reported that either major histocompatibility complex (MHC) only disparate or minor histocompatibility antigen (minor H) only incompatible LT showed epithelial rejection at a frequency similar to that of MHC
+ minor H disparate LT, unlike penetrating keratoplasty (PKP).
9 Yao et al. concluded therefore that major and minor histocompatibility antigens are both related to corneal epithelial rejection. The immune mechanism in LT differs from that in PKP for several reasons. In the LT model, the donor grafts involve the limbus, which contains Langerhans’ cells, allowing for greater host recognition of the graft, and the corneal limbus of the recipient possesses blood vessels and lymphatics, as well as Langerhans’ cells, allowing for acute allosensitization and swift rejection.
However, as in PKP,
10 11 allogeneic LT rejection appears to be mediated primarily by a delayed-type hypersensitivity (DTH) response, rather than by a cytotoxic T lymphocyte (CTL) response.
12 Moreover, donor-specific DTH suppression by induction of anterior chamber-associated immune deviation (ACAID) could promote survival of both LT
13 and PKP allografts.
14 15 16 Therefore, a DTH-suppressing strategy must be realized to avoid epithelial rejection.
DTH is typically mediated by T-helper type 1 (Th1) CD4
+ T cells, that secrete interferon (IFN)-γ.
17 Th1 cells can be cross-regulated by a different subset of CD4
+ T cells that secrete IL-4 and -10, but not IFN-γ, which are termed Th2. Thus, the cytokines produced by Th2 cells suppress cytokine activation and release by Th1 cells,
17 thereby limiting the ability of the latter to mediate effector responses such as the allodestructive DTH response. Actually, a Th2-biased immune system, in a PKP mice model, enhanced allogeneic graft survival.
18 We have used this strategy in an effort to modify LT in mice. Our results indicate that mice with immune systems heavily biased toward Th2 responses accept allogeneic LT at a higher rate than do normal mice. Furthermore, this system is equally effective between a MHC-matched and MHC-mismatched donor–recipient combination. Our results indicate the presence of another suppressive mechanism in the Th2-biased immune system.
Seven to 10-week-old male BALB/c (H-2d) and C57BL/10 (H-2b) mice were purchased from SLC (Osaka, Japan). Mice comprising same-aged male B10.D2 (H-2d) and enhanced green fluorescence protein (EGFP) transgenic mice (H-2b, C57BL/6 background) were purchased from The Jackson Laboratory (Bar Harbor, ME). All animals were treated in accordance with the ARVO Statement on the Use of Animals in Ophthalmic and Vision Research. All experiments were approved by the Committee for Animal Research, Kyoto Prefectural University of Medicine. In the following experiments, all transplant recipients were BALB/c mice.
Keyhole limpet hemocyanin (KLH) of Megathura crenulata (Calbiochem, La Jolla, CA), was used for immunization to induce a Th2-type response. BALB/c and C57BL/6 mice received intraperitoneal (IP) injections of 50 μg KLH emulsified in incomplete Freund’s adjuvant (IFA; Difco Laboratories, Detroit, MI) 28 days before transplantation. Control mice received Hanks’ balanced salt solution (HBSS) plus IFA. In a secondary immunization, mice that received an orthotopic LT in one eye also received, in the nape of the neck immediately thereafter, an injection of KLH (50 μg) or HBSS alone in complete Freund’s adjuvant (CFA; Difco Laboratories).
Before all surgical procedures, each animal was deeply anesthetized with an intramuscular injection of 3 to 4 mg ketamine and 0.1 mg xylazine. The limbal transplantation technique, referred to in previous reports,
8 was modified as follows: Cornea without scleral tissue was excised along the limbus of the donor eye. The donor corneal endothelial cell layer was then completely peeled off with a 23-gauge needle. A lenticule longer than the host limbal graft bed was roundly and spirally cut into a section of around 1.0 × 10.0 mm. Under anesthesia, the BALB/c host corneal epithelium was scraped off completely, and both adjacent limbal and conjunctival tissues were dissected circumferentially. One continuous corneal lenticule was secured around the recipient corneal limbus with five interrupted 11-0 nylon sutures (Sharppoint; Vanguard, Houston, TX;
Fig. 1 ) Antibiotic ointment was applied to the corneal surface, and the lids were sealed for 72 hours with an 8-0 nylon-suture tarsorrhaphy (Alcon Surgical, Dallas, TX). Grafted eyes with technical difficulties (limbal perforation, anterior chamber hemorrhage during or after surgery, or suture loosening and lenticule separation during the observation period) were excluded from the study. In all sets of graft-surviving experiments, the results of daily performing experiments (5–12 limbal transplants) were combined and compared.
Grafts were evaluated daily by slit lamp biomicroscopy. Methylene blue (0.25%) staining was used to evaluate corneal reepithelialization by donor corneal epithelium proceeding from limbus to central cornea. At each time point grafts were scored for opacification and neovascularization (NV). A previously described scoring system
8 was used to grade the degree of opacification from 0 to 4+: 0, clear cornea; 1, lenticular and regional corneal epithelial edema, opacity, or clearly visible iris vessels; 2, diffuse epithelial edema, corneal opacity or both, obscuring iris vessels; 3, diffuse epithelial edema, corneal opacity or both, iris vessels not visible; 4, anterior chamber invisible due to epithelial edema, corneal opacity, or both. Grafts with an opacity score of 2+ or more were considered rejected (immunologic failure), regardless of opacity score at 4 weeks (since some grafts had only transient opacification). Corneal NV was graded from 0 to 4 in each quadrant and totaled (altogether, 0–16): 0, none; 1, invasion to the limbal graft but not to the graft; 2, invasion to the graft, but not to the central 2 mm; 3, invasion to the central 2 mm, but not reaching the center; and 4, invasion to the central cornea.
Serum-free medium used for cultures was composed of RPMI 1640 medium, 10 mM HEPES, 0.1 mM nonessential amino acids, 1 mM sodium pyruvate, 100 U/mL penicillin, 100 μg/mL streptomycin (all from BioWhitaker, Walkersville, MD) and 1 × 10
−5 M 2-mercaptoethanol (Sigma-Aldrich, St. Louis, MO), supplemented with 0.1% bovine serum albumin (Sigma-Aldrich), and ITS
+ culture supplement (1 μg/mL iron-free transferrin, 10 ng/mL linoleic acid, 0.3 ng/mL Na
2Se, and 0.2 μg/mL Fe (NO
3)
3; Collaborative Biomedical Products, Bedford, MA).
19
We used KLH-immunized and control mice that had clear grafts for 1 week after LT. Spleens and graft-site–draining lymph nodes of three animals in each group were collected and pooled. Cells therefrom were then pressed through nylon mesh to produce a single-cell suspension. Red blood cells were lysed. Lymphocytes were then washed and used for adopted transfer and for cultures. In cell cultures, the lymphocytes were resuspended at 4 × 105 in 96-well plates and stimulated with irradiated (2000 rad) B10.D2 splenocytes. In another set of experiment, spleen cells of KLH in IFA preimmunized recipient were stimulated with 50 μg/mL of KLH antigen. Cells were cultured in serum-free medium at 37°C in an atmosphere of 5% CO2.
Cultures similar to those just described were established and sustained for 120 hours. Reconstituted 20 μL XTT (sodium 3′-[1-(phenylaminocarbonyl)-3,4-tetrazolium]-bis (4-methoxy-6-nitro)benzene sulfonic acid hydrate; Sigma-Aldrich) was added to each well according to the manufacturer’s instructions. After a 4-hour incubation, absorbance at a wavelength of 450 nm was measured. Absorption spectra of tetrazolium reagents were measured with a scanning spectrophotometer and were expressed as optical density.
Cultures similar to those in the proliferation assay were established and sustained for 48 or 72 hours. At each time point, supernatants were collected and analyzed for IFN-γ, IL-4, and IL-10 contents using ELISA kits according to the manufacturer’s instructions (PharMingen, San Diego, CA).
We constructed Kaplan-Meier survival curves and used the Breslow-Gehan-Wilcoxon test to compare the probability of allograft survival. Statistical analyses were performed with a Mann-Whitney test used for comparison of NV score, and Student’s t-test for proliferation response. P < 0.05 were deemed significant.
The mouse PKP model in our previous report
18 revealed that minor H–only incompatible corneal allografts display more enhanced survival than do MHC
+ minor H disparate allografts in KLH-immunized BALB/c recipients. Initially, the fates of minor H–only incompatible B10.D2 donor limbal grafts were examined and compared between KLH immune (
n = 35) and control BALB/c mice (
n = 18).
As for recovery from epithelial scraping, the entire corneal surface was reepithelialized at 3 to 4 days after LT in both KLH-immunized and control mice, with no statistically significant (data not shown) difference. These host reepithelialization periods were almost the same as in a previous report.
8
The appearances of syngeneic (
n = 10) and allogeneic LT were virtually identical at 3 days after LT, whether judged clinically or histologically. In control mice, many epithelia remained slightly cloudy at 3 days and became opacified thereafter (a typical rejected cornea at day 7 is shown in
Fig. 2A ). In contrast, KLH-immunized mice significantly recovered a smooth corneal surface with no opacity at 5 days after LT (data not shown). Of the corneas in KLH-immunized mice, 90.4% appeared absolutely clear on day 7
(Fig. 2B) . Some mice were randomly selected and killed on day 7, and the enucleated eyes were examined histologically. Donor lenticule epithelia of control mice contained copious amounts of infiltrating cells in both the limbal
(Fig. 2C) and central areas
(Fig. 2E) , as did the host cornea subepithelium. In contrast, KLH-immunized epithelium contained few invading lymphocytes, and the epithelial cells were aligned in both the limbal
(Fig. 2D) and central areas of the cornea
(Fig. 2F) .
The rate and frequency of epithelial rejection, shown in
Figure 3A , showed that all corneas of control mice were rejected swiftly (
n = 18), within 10 days (mean survival time = 7.1 ± 0.4 days). In contrast, KLH-immunized mice showed significantly improved epithelial survival, maintaining clear corneas until 56 days after LT (22/35 corneas, 62.8%,
P < 0.001). In 56 days after LT, corneal donor graft and host graft were indistinguishably clear. Clinical observations show less NV invasion into the donor lenticule in eyes of KLH-immunized mice. Because inflammatory corneal NV, which is associated with the new development of corneal lymphatics, leads to accelerated allograft rejection, early development of corneal NV was compared. At 5 days after surgery, significant NV suppression could be discerned in KLH-immunized mice
(Fig. 3B) . During the observation periods, there was no NV invasion over the limbal graft in KLH-immunized mice. In addition, the fates of clear B10.D2 corneas in KLH-immunized mice were virtually identical with syngeneic LT graft.
Characteristics of Alloreactive T Cells from KLH-Immune Mice That Accepted Limbal Grafts
For these studies, splenocytes of BALB/c mice that received an intraperitoneal injection of KLH (50 μg) in IFA (50 μL) were stimulated in vitro with KLH. The supernatants of these cultures contained significant quantities of IL-4 (1254.5 ± 43.7 pg/mL) and IL-10 (2831.3 ± 126.5 pg/mL), but little IL-2 (31.4 ± 4.3 pg/mL) or IFN-γ (98.4 ± 15.6 pg/mL), indicating that the mice had mounted a Th2-type response to KLH. Our next goal was to determine whether LT-acceptant mice with Th2-biased immune systems respond to immunization with alloantigens by developing alloantigen-specific Th2 responses. One week after LT, ipsilateral cervical lymph nodes and spleens were removed, pressed through nylon mesh to produce a single-cell suspension and pooled. T cells were stimulated in vitro with x-irradiated B10.D2 spleen cells. Naive mice (negative) and mice immunized with 10 million B10.D2 splenocytes subcutaneously 1 week before were used as the control. The results of these experiments with lymph nodes are presented in
Figure 4 . The proliferation response at 96 hours indicates that recipients with limbal allografts did not prime in KLH-immunized mice
(Fig. 4A) . In contrast, T cells from control mice that received HBSS rather than KLH at the time of LT proliferated more vigorously. Supernatants were removed from 24-, 48-, and 72-hour cultures and assayed for IFN-γ content to evaluate Th1 activation, and IL-4 and -10 content for Th2
(Fig. 4B) . T cells stimulated with allogeneic spleen cells produced significant amounts of IFN-γ. IFN-γ production by lymphocytes from both KLH-immunized and control mice was similar. As in Th1 cytokine production, IL-4 and -10 production by lymphocytes from both KLH-immunized and control mice was low. Of importance, donor-specific Th2 cytokine production by lymphocytes of KLH-immunized mice was less than that produced by lymphocytes of control mice. Similar results were observed in donor-specific cytokine production in spleen cells. We conclude from these results that LT graft survival in KLH-immunized mice is not due to donor-specific Th2 cells.
From another standpoint, if rejection-suppressing regulatory cells, including donor-specific Th2 cells, are present in the KLH-immunized recipient, their corneal acceptance can be adoptively transferred.
18 Because only KLH-immunized mice, not control mice, could promote a clear graft for more than 2 weeks, we prepared B10.D2-grafted KLH-immunized mice with corneas that had remained perfectly healthy for 2 weeks. The mice were killed, and their cervical lymph nodes and spleens were removed. Furthermore, cervical lymph nodes and spleen cells were mixed, and single cell suspensions were prepared, pooled, and injected intravenously (one donor equivalent per recipient) into naive BALB/c mice. Immediately thereafter, the mice received B10.D2 limbal grafts (
n = 7). In control, we used BALB/c recipients that were not immunized with KLH and did not receive spleen or lymph cells of recipient mice (
n = 8). The fate of LT
(Fig. 5) , shows that acceptor lymphocytes did not regulate limbal graft rejection. This result indicates that graft acceptance for 2 weeks in KLH-immunized mice is not due to regulatory cells, unlike the PKP model in the previous report.
18
A previous report
18 revealed that the Th2-biased immune system promotes B10.D2 PKP allograft survival better than C57BL/6 allograft survival, concluding that the Th2-biased immune system can influence indirect presentation of allosensitization in the PKP model. However, because the presence of donor-derived Langerhans’ cells in the B10.D2 limbal graft causes direct alloantigen presentation to host T cells (direct recognition), the Th2-biased immune system may have influenced direct presentation of allosensitization and promoted graft survival in this LT model. We therefore performed total disparate (MHC
+ minor H disparate) C57BL/10 LT and observed the results clinically. To our surprise, KLH-immunized mice showed significantly improved epithelial survival, more than half of them maintaining clear corneas until 42 days after LT (
n = 12,
P < 0.02), though control mice rejected all C57BL/10 limbal grafts within 14 days (
n = 14;
Fig. 6 ). We conclude that the Th2-immune system can promote LT survival in both minor H only and fully disparate combinations.
For observations of the presence of donor-derived epithelial cells in long-term clear corneas of KLH-immunized mice that received allogeneic LT, BALB/c mice received limbal grafts of EGFP transgenic mice (C57BL/6 background;
n = 15). At an appropriate time, the mice were killed and their eyes enucleated, frozen, embedded, sectioned, and observed by fluorescein microscopy to detect EGFP-derived donor cells. The grafted eyes with rejected corneas showed no positive fluorescence on day 7 after surgery (
n = 2;
Fig. 7A ). In contrast, donor-derived epithelial cells remained at the central area of the cornea on day 7 (
n = 5)
(Fig. 7B) and on day 14 (
n = 6;
Fig. 7C ) after surgery in KLH-immunized mice, with the fluorescein intensity being similar to that in normal corneas of EGFP transgenic mice (data not shown). Two sets of experiments were combined, and representative photographs are shown in
Figure 7 . Positive fluorescence in the epithelial layer was also detected 4 weeks after surgery in KLH-immunized mice (
n = 2, data not shown), demonstrating the long-term survival of donor epithelial cells on the ocular surface.
Most limbal grafts that fail in both humans and mice do so because of immune-associated rejection, though even the most immunogenetically disparate corneal grafts placed orthotopically can exhibit prolonged, often indefinite, survival. In comparison with LT, the extraordinary success of PKP can be attributed to various features of the normal cornea and anterior segment that in the aggregate account for their immune-privileged state,
20 21 including the avascularity of the stroma, the absence of corneal lymphatics, and the rarity of indigenous professional antigen-presenting Langerhans’ cells or macrophages in the normal graft bed. Because of the factors responsible for ocular immune privilege, it has been found that minor H antigens, rather than antigens encoded within the MHC, are the most important initiators of alloimmunity after PKP.
9 All peptides derived from minor H antigen processing are loaded onto self-MHC molecules on recipient antigen-presenting cells and presented to T cells by the so-called indirect pathway of allorecognition in the PKP model. The Th2-biased immune system was therefore established to induce donor minor H–specific Th2 type response to suppress Th1-mediated immune rejection, because mice that have mounted Th2-type responses to one peptide antigen often display Th2 responses to subsequent immunizations with different antigens.
22 In fact, donor-specific Th2 cells were induced and donor-specific allorejection suppressed in a PKP model mice.
18
In contrast, less immune privilege is exhibited in LT, which involves the presence of Langerhans’ cells in both donor and recipient limbus. Donor Langerhans’ cells can present alloantigens directly to recipient T cells by the so-called direct pathway of allorecognition. In cases of MHC-disparate combination between donor and recipient, MHC alloantigen is presented without the contribution of recipient antigen-presenting cells. In cases of MHC-matched combination, even minor H can be presented directly to recipient T cells by donor Langerhans cells, because both donor and recipient share the same MHC molecule. Theoretically, it is quite natural for minor H antigen presentation in LT to be influenced by the Th2-biased immune system. However, the fact is that neither positive donor-specific Th2 response nor typical donor-specific regulation was detected. Moreover, both MHC disparate and MHC-matched allogeneic LT showed similar graft survival in this system.
The Th2-biased immune system could suppress neovascular invasion at around 5 days after LT, before the initial rejection reaction appeared. Because corneal NV probably plays an important role in facilitating swift antigen presentation and effector elements in the inflamed cornea and is associated with fulminating graft rejection, suppressed initial NV induction should contribute to LT graft survival. Immunologically, alloantigen is presented and alloreactive T cells proliferate at around 3 to 5 days in the case of skin transplantation. NV suppression in the Th2-biased immune system is therefore presumably not mediated by alloreactive T cells, but by anti-inflammatory factors produced in the Th2 response. This NV suppression may be one mechanism to enhance graft survival in this system.
Current prophylactic and therapeutic regimens for LT rejection are associated with significant complications. Moreover, although immunosuppression can mediate clinical allograft survival, it is suggested that donor cells do not survive on the rat ocular surface indefinitely.
7 The Th2-biased immune system unexpectedly promoted donor epithelial cell survival for more than 4 weeks. Hence, the Th2-biased immune system or a modified system can be an effective immune therapy in LT.
The authors thank Junji Hamuro for helpful discussions of the experiments.