PURPOSE

Classically, corneal allograft rejection is thought to be a T(H)1-mediated phenomenon. However, T(H)2-mediated allograft rejection has been reported in other transplanted organ systems, including the heart and kidney. We previously reported a form of T(H)2-mediated corneal allograft rejection in a murine model with a T(H)2 immune bias. In this study we sought to determine if there was any evidence for this form of corneal allograft rejection in humans.

DESIGN

Experimental study with an interventional case series.

METHODS

The clinical records of all keratoconus patients undergoing penetrating keratoplasty at the University of Texas, Southwestern Medical Center from 1994 to 1999 were reviewed. Careful attention was paid to a clinical history of atopy. Atopic patients were selected, because these patients have been shown to have a "T(H)2 immune bias." The corneal graft rejection rate in these patients and the number of repeat corneal transplants performed was determined. The experimental group consisted of patients with a clinical history of atopy and keratoconus who had at least one repeat penetrating keratoplasty for an immunologically rejected corneal transplant. Any patient with evidence of primary allograft failure was excluded from this study. Tissue specimens from these patients were embedded in paraffin, serially sectioned, stained with Giemsa stains, and examined histologically. The control group consisted of patients without a clinical history of allergy (and therefore no T(H)2 immune bias) who underwent corneal transplantation for Fuch corneal endothelial dystrophy, or aphakic/pseudophakic bullous keratopathy. Failed grafts from these control patients were also paraffin embedded, serially sectioned, stained, and examined histologically. The human experimental and control corneal specimens were compared with data obtained in a murine model of T(H)2-mediated corneal allograft rejection. Briefly, full-thickness penetrating C57BL/6ByJ corneal allografts were transplanted onto Balb/cByJ and Balb/c-IFN-gamma(tm1Ts) (Balb/c-IFN-gamma knockout) mice. Additionally, full-thickness Balb/cByJ corneal allografts were transplanted onto C57BL/6ByJ and C57BL/6ByJ-IFN-gamma(tm1Ts) mice. Corneal allograft rejection rates and mean rejection times were calculated and compared between wild-type and interferon gamma (IFN-gamma) knockout hosts. The rejected allografts were examined histologically by the same methods used in the human tissue.

RESULTS

There were 84 penetrating keratoplasties performed from 1994 to 1999 for keratoconus. Seven of these 84 patients rejected their corneal grafts. Of the 7 patients who rejected their corneal allografts, 4 had repeat penetrating keratoplasty. Of these 4 repeat corneal allografts, 3 showed eosinophilia when compared with rejected grafts in control patients. Atopic keratoconus patients had a mixed inflammatory cellular infiltrate in the rejected corneal tissue specimen with a significantly greater density of eosinophils (P =.001) compared with patients who did not have a pre-existing T(H)2 bias. The inflammatory infiltrate in these patients without a T(H)2 immune bias was mononuclear. In the murine model, corneal allograft rejection did occur in the absence of IFN-gamma, a critical T(H)1 cytokine in both fully allogeneic donor-host combinations. Histologically, rejection in these ("T(H)2 mice") was characterized by a predominant eosinophilic infiltrate in the rejected graft bed when compared with wild-type animals ("T(H)1 mice") that had a predominantly mononuclear infiltrate in the rejected corneal graft bed.

CONCLUSIONS

Preliminary findings show that corneal allograft rejection in patients with a pre-existing T(H)2 phenotype is similar to what is seen in the murine model of T(H)2-mediated corneal allograft rejection. Based on this small sample, it appears that eosinophils may play a role in corneal allograft rejection in this group of patients. However, further study is necessary to determine the importance of these cells in allograft rejection.