Quiescent innate and adaptive immune responses maintain the long-term integrity of corneal endothelium reconstituted through allogeneic cell injection therapy

Abstract

This study aims to clarify the immunogenicity in acquired and innate immune responses of cultured human corneal endothelial cells (hCECs) applied for cell injection therapy, a newly established modality for corneal endothelium failures. Thirty-four patients with corneal endothelial failure received injection of allogeneic hCEC suspension into anterior chamber. No sign of immunological rejection was observed in all 34 patients during the 5-8 years postoperative follow-up period. Cell injection therapy was successful in 2 patients treated for endothelial failure after penetrating keratoplasty and one patient with Descemet membrane stripping automated endothelial keratoplasty failure. ELISPOT assays performed in allo-mixed lymphocyte reaction to the alloantigen identical to that on the injected hCECs, elicited sparse IFN-γ-specific spots in the peripheral blood mononuclear cells of patients who received hCEC injection. The therapy generated simple and smooth graft-host junctions without wound stress. The injection of C57BL/6 CECs into the anterior chamber of BALB/c mice, which rejected C57BL/6 corneas 6 weeks ago, induced no sign of inflammatory reactions after the second challenge of alloantigen. Collectively, injection of the hCEC cell suspension in the aqueous humor induces immune tolerance that contributes to the survival of the reconstituted endothelium.

Figure 1

Pre- and post-cell injection therapy images (slit-lamp, specular microscopy and Scheimpflug) of a patient with failure of previous keratoplasty. He underwent penetrating keratoplasty (PK) in 2001 for keratoconus in his left eye and received the second PK in May 2014 for corneal graft failure caused by graft rejection. Subsequently, he underwent non-Descemet stripping automated endothelial keratoplasty (nDSAEK) in May 2016 for graft failure. Although the nDSAEK graft was in good condition at 4 months postoperative, graft failure occurred at 9 months postoperative, resulting in the severe impairment of best corrected visual acuity (BCVA). These images were obtained at pre-cell injection in March 2017 (left) and 3 years post-cell injection in March 2020 (right). Scale bar: 100 μm.

Figure 2

CD4+ dependent immune reaction against hCECs. (a) PBMCs from 3 healthy donors (A, B and C) were stimulated with hCECs or PBMCs from other healthy donors (allo-PBMCs) at the ratio of 5:1 or 1:1. After incubation for 2 days, IFN-γ production was analyzed by ELISPOT assay. Stimulation with anti-CD3 antibody was used as the experimental positive control (lower panel). The numbers of detected spots were shown in each panel (TNTC = Too numerous to count). (b) PBMCs from the patients were stimulated with the same lot of hCECs injected at the S/R ratio of 1:5. Seven patients in the first trial group and ten patients in the second trial group were examined. Normalized data with allo-MLR are shown. (Student’s paired t-test *: p < 0.001). Further, Kruskal–Wallis test was used and post‐hoc analysis with Steel’s multiple comparison test were performed. (Upper panel) (Pre): Kruskal–Wallis test among three groups; normal, PBMC (-), cHCEC, p < 0.001, Steel's multiple comparison test; normal versus PBMC (-) p = 0.003, normal versus cHCEC, p = 0.003 (2 weeks): Kruskal–Wallis test among the same three groups, p = 0.027, Steel's multiple comparison test; normal versus PBMC (-) p = 0.033, normal versus cHCEC, p = 0.047. (Lower panel) (Pre): Kruskal–Wallis test among three groups; normal, PBMC (-), cHCEC, p < 0.061, Steel's multiple comparison test; normal versus PBMC (-) p = 0.055, normal versus cHCEC, p = 0.446 (2 weeks): Kruskal–Wallis test among the same three groups, p = 0.103, Steel's multiple comparison test; normal versus PBMC(-) p = 0.064, normal versus cHCEC, p = 0.242.

Figure 3

(a) Donor-reactive DTH 4 weeks after allogeneic corneal penetrating keratoplasty (PK) or cell injection. One group of BALB/c mice received C57BL/6 corneal penetrating keratoplasty (PK, N = 17), and another group of BALB/c mice received mouse primary corneal endothelial cell injection (mpCECs injection, N = 11). Four weeks later, they were ear-challenged with donor-derived splenocytes, and swelling responses were assessed 24 h later. Naive mice were used as a negative control, and mice subcutaneously immunized with donor splenocytes were used as a positive control. *Indicates responses significantly higher than those of the control group (Student’s paired t-test, p < 0.005). Further, Kruskal–Wallis test was used and post-hoc analysis with Steel's multiple comparison test were performed. Kruskal–Wallis test among five groups; negative control, positive control, PK acceptor, PK rejector and mCEC transplantation, p < 0.001, Steel's multiple comparison test; negative versus positive control p = 0.032, negative control versus PK acceptor p = 0.032, negative control versus PK rejector p = 0.031, negative control versus mCEC transplantation p = 0.458. (b) Systemic allo-sensitization was performed by injecting C57BL/6 splenocytes subcutaneously either into pre-PK treated or mpCEC-injected BALB/c mice. The grafted cornea tissues were promptly rejected in all mice in the pre-PK treated mice (N = 8), whereas in the hCEC pre-injected mice, the reconstituted CECs were not rejected, and their corneas remained clear and stable.

Figure 4

Host response after cell injection. (a) Sera from 9 patients in the clinical study were collected on the day before surgery and analyzed using Bio-Plex. (b) Patients’ sera injected with a relatively lower proportion (56%) of mature hCEC SPs (patients 6 and 7) or a relatively higher proportion (> 90%) (patients 14 and 15) were analyzed using Bio-Plex.

Figure 5

Slit-lamp microscopic images (left) and anterior segment optical coherence tomography (AS-OCT) images (right) of the eyes obtained using CASIA and CASIA2 (Tomey Corporation, Nagoya, Japan) at 6 months postoperative: post-cell injection (top, CASIA), post-Descemet membrane endothelial keratoplasty (DMEK) (second column, CASIA 2), post-Descemet’s stripping automated endothelial keratoplasty (DSAEK) (third column, CASIA 2) and post-penetrating keratoplasty (PK) (Bottom, CASIA 2). The postoperative shape of the cornea receiving cell injection is almost identical to that of a healthy cornea, as well as the eye that underwent DMEK.