Evidence of Long-term Survival of Donor-Derived Cells after Limbal Allograft Transplantation

Abstract

purpose. Severe destruction of the corneal limbus causes conjunctival invasion and subsequent visual loss. Limbal allograft transplantation (LAT) was recently proposed for the treatment of these disorders. However, whether the method functions as a stem cell transplantation of the corneal epithelium remains unclear. This study provided evidence that donor-derived corneal epithelial cells survive long after LAT.

methods. Epithelial cells on the paracentral cornea in patients who have undergone LAT were subjected to fluorescence in situ hybridization (FISH) and polymerase chain reaction restriction fragment length polymorphism (RFLP) analysis. X and Y chromosomes were detected using sex chromosome–specific probes in the FISH analysis, and HLA-DPB1 antigens were examined in the RFLP analysis. Eyes receiving conventional penetrating keratoplasty (PKP) served as controls.

results. Donor-derived epithelial cells were detected in three of five eyes (60.0%) in the FISH analysis and in seven of nine eyes (77.8%) in the RFLP analysis. Among these eyes, one and three eyes in the FISH and RFLP analysis, respectively, had both donor- and recipient-derived cells. In control PKP eyes, none of the eyes in the FISH analysis and one of eight eyes (12.5%) in the RFLP analysis had donor-derived cells.

conclusions. These results suggest that donor-derived cells survive much longer after LAT than those after PKP, and that LAT may function as stem cell transplantation of the corneal epithelium.

The corneal epithelium maintains homeostasis by balancing desquamation from the surface and proliferation of the basal cells plus centripetal movement from the stem cells located at the limbus. ^{1 2} When the corneal limbus is severely damaged (i.e., limbal dysfunction), conjunctival epithelium replaces the corneal epithelium, which causes significant visual deterioration. The surgical prognosis of penetrating keratoplasty (PKP) in eyes with limbal dysfunction remains very poor, because the epithelium on the graft is not supplied by stem cells. ³ Recently, transplantation of the limbal tissues using either autografts ⁴ or eye bank eyes (limbal allograft transplantation: LAT) has been proposed for the treatment of limbal dysfunction. ^{5 6} The concept of LAT is to supply stem cells of the corneal epithelium to reconstruct healthy ocular surface epithelia. With the combination of LAT and subsequent immunosuppression, prognosis of severe ocular surface disorders caused by limbal dysfunction has dramatically improved. ^{5 6} However, despite these encouraging clinical reports, the fate of the graft epithelium remains unclear. It has been shown that after PKP, donor-derived corneal epithelial cells are replaced by host cells within 1 year after surgery. ^{7 8} There is no evidence to date that donor-derived stem cells of the corneal epithelium repopulate the ocular surface after LAT. Conversely, previous reports of an experimental model and a human study indicate that donor-derived cells do not survive long after LAT, ^{9 10} suggesting that LAT may simply function as transplantation of limbal substrates. In the present study, we examined the survival of corneal epithelial cells in LAT recipients using fluorescence in situ hybridization (FISH) and HLA analysis. Correlation between the clinical outcomes and the results of the analysis was also investigated. 

Materials and Methods

Subjects

This study followed the tenets of the Declaration of Helsinki, and informed consent was obtained after explanation of the purpose and possible consequences of the study. Ten eyes of nine LAT recipients with follow-up periods of 300 days or longer were examined in this study. Eyes with severe recurrent conjunctival invasion of the central cornea were excluded. In the FISH analysis, five patients receiving sex-mismatched LAT were examined (male to female in three patients, female to male in two patients). The mean age was 71.6 years (range, 69–77 years), and the mean follow-up period was 445 days (range, 335–691 days). Original diseases of these patients were ocular cicatricial pemphigoid (two eyes), pseudo-ocular cicatricial pemphigoid (one eye), familial subepithelial amyloidosis (one eye), and Stevens–Johnson syndrome (one eye). Four of these five eyes had simultaneous PKP. Nine eyes from eight patients who had undergone PKP without LAT served as controls (Table 1) . 

In RFLP analysis, nine eyes of eight patients with LAT (four men and four women, with a mean age of 58.4 years) were studied. Eight eyes in five patients who underwent PKP (three men and two women, with a mean age of 50.4 years) served as controls. The mean follow-up period was 578 days (range, 326–978 days) and 377 days (range, 182–630 days), respectively. Four and two eyes that received LAT and PKP, respectively, were also examined by FISH analysis. Original disease, surgical procedures, and postoperative conditions are listed in Table 2 . 

Surgical Procedure for LAT

All surgical procedures were performed with patients under retrobulbar anesthesia. After excising inflamed conjunctiva and subconjunctival fibrosis, PKP was performed when the central cornea was opaque. The grafts were preserved in preservation media (Optisol GS, Chiron, Irvine, CA) for several days. Simultaneous cataract extraction and/or intraocular lens implantation was performed in three eyes. Preserved human amniotic membrane was placed on the sclera in four eyes. In eyes receiving both PKP and LAT, the central cornea and limbal tissue were obtained from the same donor eye. To obtain limbal tissue, the scleral portion of the donor rim tissue was excised, and the remaining limbal tissue was trimmed as thin as possible. The grafts were then placed and secured on the limbal area. Various shapes of the limbal grafts were used in this study: ring-shaped (four eyes in FISH analysis, five eyes in RFLP analysis), segmental (one eye in RFLP analysis), and C-shaped (one eye in FISH analysis and three eyes in RFLP analysis). 

After LAT, dexamethasone (Rinderon; Shionogi Pharmaceutical, Osaka, Japan) was administered with a starting dose of 8 mg/d and tapered within 3 weeks, and cyclosporin A (Sandimmune, Sandoz, Basel, Switzerland) was administered intravenously with a starting dose of 3 mg/kg for 1 week, with trough levels of 100 to 150 ng/ml maintained for at least 6 months. Topical antibiotics (0.3% ofloxacin, Tarivid; Santen Pharmaceutical, Osaka, Japan), corticosteroid (0.1% dexamethasone, Sanbetason; Santen Pharmaceutical), autoserum dissolved in physiological saline, and 0.05% cyclosporin A dissolved inα -cyclodextrin were used five times a day. In patients who underwent PKP without LAT, systemic cyclosporin A was not used. Systemic administration of corticosteroid and/or topical cyclosporin A was used in some cases. 

Clinical Sampling

Epithelial cells in the paracentral cornea (approximately 0.5 mm in diameter) in the inferotemporal region were obtained using fine forceps. For the FISH analysis, cells were placed on a glass slide with saline, then dried, and fixed with absolute alcohol. For the RFLP analysis, the samples were placed in microtubes with physiological saline and kept frozen until analysis. 

FISH

Samples slides were treated with 100 mg/ml RNase A in 2× SSC at 37°C for 1 hour, incubated with 0.01 N HCl containing 50 mg/ml pepsin for 10 minutes at 37°C, and fixed with 1% formaldehyde in phosphate-buffered saline (PBS) containing 50 mM MgCl₂. After washing with PBS, cells were dehydrated in an ethanol series and then denatured by incubating in 70% formaldehyde and 2× SSC at 71°C for 2 minutes. 

FISH was performed with sex chromosome–specific dual color probe (CEP X SpectrumOrange/CEP Y SpectrumGreen; Vysis, Downers Grove, IL). Hybridization using each probe was carried out according to the manufacturer’s recommendation. The samples were counterstained with 4′,6-diamimo-2-phenylindole (DAPI). Signals from fluorescein isothiocyanate (FITC), rhodamine, the two color probes, and DAPI were visualized by fluorescence microscope (Nikon, Tokyo, Japan) with triple-band-pass filter. X chromosomes were visualized as red signals, and Y chromosomes as green. Typically, 20 to 50 cells were examined in each sample. 

RFLP Analysis

Genomic DNAs from the sample cells were extracted by the conventional phenolchloroform method. HLA-DPB1 alleles were genotyped by the polymerase chain reaction–restriction fragment length polymorphism (PCR-RFLP) method. ¹¹ A test typing kit (Sumitomo Metal Industries, Tokyo, Japan) was used. In brief, the reaction mixture containing 0.5 mg genomic DNA, 10 μl 10× PCR buffer, 10 μl solution containing dNTPs, mixed primers, 2.5 U Taq polymerase (TakaraTaq; Takara Shuzo, Shiga, Japan) and MilliQ water for a total volume of 100 ml, was covered with 1 drop of mineral oil in a 0.5-ml microcentrifuge tube. PCR amplifications were carried out by 30 cycles of 96°C denaturation (60 seconds), 56°C annealing (60 seconds), and 72°C extension (120 seconds), followed by an additional extension (72°C, 5 minutes), in an automated oil-bath PCR thermal sequencer (model TSR-300; Iwaki Glass, Chiba, Japan). After PCR, 5-μl aliquots of the PCR product with 1.5 μl of gel loading buffer were checked for amplification by 10% acrylamide gel electrophoresis in 0.5× TBE buffer in MUPID (Cosmo Bio, Tokyo, Japan). When the DNA sample was amplified for the target DPB1 loci, 10-μl aliquots of the PCR product were digested with enzyme solution, which contained 2 U restriction endonucleases and reaction buffer solution, for 3 hours to overnight at 37°C. Seven restriction enzymes (Bsp1286I, BssHII, Cfr13I, DdeI, EcoNI, FokI and RsaI) were used. Three milliliters gel loading buffer with proteinase K (Boehringer Mannheim, Mannheim, Germany) was added and incubated for 30 minutes at 37°C. The digested fragments were detected by 10% acrylamide gel electrophoresis, and the allelic types were determined from the RFLP patterns. 

Results

Clinical Outcomes and Complications

Although clarity of the central cornea improved in all eyes receiving LAT, various postoperative complications such as persistent epithelial defects or partial conjunctival invasion were noted. Eyes with conventional PKP were free from these complications in all but one eye, which had a persistent epithelial defect (Table 2) . 

FISH Analysis

Three of five samples contained epithelial cells with donor-derived chromosomes (Table 1) . Figure 1 A shows the fluorescence micrograph of a 69-year-old man with ocular cicatricial pemphigoid who had LAT from a female donor (case 3). All cells in the sample, which were collected at 99 weeks after surgery, showed two red fluorescence signals in each cell, indicating that these cells were of donor origin. Two other samples (cases 1 and 4) also showed the presence of donor-derived cells. In case 1, 14 of 15 cells collected showed both X and Y chromosomes (donor origin, Fig. 1B ), and only one cell had two X chromosome signals (recipient origin). The remaining two cases had only host-derived cells, which corresponded to the two cases that had either postoperative conjunctival invasion or corneal neovascularization. Cells from the nine PKP cases were all of host origin (Table 3 ; Fig. 1C ). 

RFLP Analysis

In nine samples from LAT eyes, seven expressed HLA-DPB1 antigens differing from those of recipients. The HLA combination was identical with the donor’s in four eyes, with the recipient’s in two eyes, and a mixture of donor’s and recipient’s in three eyes (Table 2 3) . In contrast, six of seven samples from PKP eyes showed only cells with DPB1 antigens identical with those of the recipient. One sample from a 79-year-old woman obtained 58 weeks after PKP showed HLA DPB1 of 0401/0402, a combination of antigens of the donor (0401/0901) and of the recipient (0402/0402). 

Discussion

In the present study, we showed that donor-derived corneal epithelium survived much longer in patients who underwent LAT plus PKP than in those who underwent PKP alone. Combined results from both FISH and RFLP analysis revealed that 7 of 10 eyes had donor-derived cells, of which three eyes had both donor- and recipient-derived cells (Table 3) . Both FISH and RFLP analyses were shown to be very sensitive ^{12 13} and allowed the determination of the origins of cells from small samples. Results of the two-analysis method were in good accordance. In four eyes examined by both FISH and RFLP analysis, three showed the same results in the survival of donor cells. One eye showed only recipient cells in the FISH analysis and a mixture of both donor and recipient in RFLP analysis. Therefore, it is highly likely that these results reflected the real status of the graft epithelial cells. 

In contrast to the high rate of donor cell survival after LAT, all but one PKP eye had only recipient-derived cells. The result is in good accordance with previous reports, which showed that the donor graft epithelial cells are gradually replaced by recipient cells within 1 year after surgery. Kinoshita et al. ⁷ studied the survival of donor corneal epithelium using sex–chromatin analysis in rabbits after lamellar keratoplasty and found that donor cells survived only up to 12 weeks after surgery. ⁷ Similar results were reported in human eyes using the FISH analysis. ⁸  

The stem cells of the corneal epithelium are located in the limbal area, and they constantly produce transient amplifying cells. The basal cells have a high capacity of proliferation, but their life span is limited to several months. The corneal epithelial cells, other than basal cells, are terminally differentiated cells, and their life span is only approximately 1 week. All conventional keratoplasties, including PKP, lamellar keratoplasty, and keratoepithelioplasty, are transplantations of transient amplifying cells and terminally differentiated cells. ¹⁴ Therefore, it is understandable that donor epithelium is depleted within several months after surgery. The results of the present study in which donor epithelial cells were detected up to 30 months after LAT strongly suggest that LAT functions as a stem cell transplantation. 

Although the clinical success of LAT is remarkable, previous studies failed to show the long-term survival of donor-derived cells after LAT. Swift et al. ¹⁰ studied donor cell survival in rabbit limbal dysfunction models using sex–chromatin analysis. They showed that the number of epithelial cells positive for the Barr body (female origin) was not significantly different in female rabbits receiving LAT from male donors than in control animals. The most probable reason that we found more donor-derived cells than Swift et al. is the difference in postoperative treatment. Relatively strong postoperative immunosuppression was used in our study, whereas topical steroid was used only in a subgroup of animals. ¹⁰  

Interestingly, in Swift et al., more donor-derived epithelial cells were observed in eyes receiving topical steroid than in those that did not. We believe that intense care against immunologic rejection is a key to longer survival of donor-derived epithelial cells. In human studies, Williams et al. ⁹ investigated the survival of donor-derived epithelial cells in an LAT patient using short tandem–repeat DNA polymorphism. They reported that donor-derived cells were detected at the 12th postoperative week but disappeared by the 20th week, despite strong systemic immunosuppression. One possible explanation is that they placed limbal grafts segmentally (from 4 to 8 o’clock and 10 to 2 o’clock). Host cells may therefore have invaded the central cornea through nasal and temporal openings of the limbal grafts. In our experience, a ring-shaped graft blocks conjunctival invasion to the cornea more efficiently than do segmentally placed grafts. 

Interestingly, some of the samples in our series had both donor and recipient cells. It is not clear whether this coexistence of donor and recipient cells is a stable condition or represents slow but continuous replacement of donor-to-recipient cells. Recent studies show that persistence of donor-derived cells after organ transplantation is a relatively common phenomenon when molecular analysis is used (microchimerism). ^{12 13 15} Whether the chimeric distribution on the cornea contributes to the graft acceptance in LAT, as is postulated in liver transplantation, ¹⁵ is unclear. The limbal area of eyes undergoing LAT was totally covered by fibrous tissues before surgery that was completely excised during surgery. Therefore, recipient limbal stem cells are unlikely to survive, and recipient-derived cells found in the samples were presumably of host conjunctival epithelium origin. Phenotypic change of the conjunctival to corneal epithelium has been observed when limbal cells are totally destroyed. ¹⁶ We did not find any differences in clinical outcomes, including the development of postoperative complications between eyes with or without donor-derived epithelial cells. More studies are needed to elucidate the clinical significance in donor cell survival. 

In conclusion, the present study indicates that donor-derived corneal epithelium survived for up to 30 months after LAT, which is significantly longer than conventional keratoplasty. LAT is likely to function as a stem cell transplantation of the corneal epithelium. 

 Reprint requests: Jun Shimazaki, Department of Ophthalmology, Tokyo Dental College, 5-11-13 Sugano, Ichikawa, Chiba, 272-3518 Japan.

 Supported by grants from the Japanese Ministry of Health and Welfare and in part by a Grant-in-Aid for Research supported by Tokyo Dental College.

 Submitted for publication July 7, 1998; revised January 22, 1999; accepted March 3, 1999.

 Proprietary interest category: N.

 

Table 1.

 

Table 1.

 

Patient Profile and Results of FISH Analysis

Case No.	Age/Sex	Donor Gender	Original Disease	LAT	Lenticle Shape	Other Surgical Methods	Follow-up (Days)	Postoperative Complication	FISH Result	Origin
1	73F	M	P-OCP	+	R	PKP, ICCE, AMT	348		XY;14, XX;1	Mix
2	77F	M	OCP	+	R	PKP, ECCE, IOL, AMT	349	PED, Conj	XX	Recipient
3	69M	F	OCP	+	R	PKP, AMT	691	PED	XX	Donor
4	70F	M	FSA	+	C	PKP	504	PED	XY	Donor
5	69M	F	SJS	+	R	PEA, IOL, AMT	335	Conj	XY	Recipient
6	58M	F	BK	−		PKP	460		XY	Recipient
7	30F	M	Scar	−		PKP	375		XX	Recipient
8	24M	F	Scar	−		PKP	362		XY	Recipient
9	84F	M	BK	−		PKP, IOL, ASR, AV	365		XX	Recipient
10	58F	M	Scar	−		PKP	429		XX	Recipient
11	30M	F	LK	−		PKP, IOL	277		XY	Recipient
12	81F	M	BK	−		PKP, ACIOL, ASR, AV	294		XX	Recipient
13*	15M	F	KC	−		PKP	399		XY	Recipient
14*	15M	F	KC	−		PKP	630		XY	Recipient

 C, C-shaped; R, ring-shaped; FSA, familial subepithelial amyloidosis of the cornea; OCP, ocular cicatricial pemphigoid; P-OCP, pseudo-OCP; SJS, Stevens–Johnson syndrome; BK, bullous keratopathy; LK, lipid keratopathy; KC, keratoconus; ICCE, intracapsular cataract extraction; AMT, amniotic membrane transplantation; PEA, phacoemulcification and aspiration of the lens; IOL, posterior chamber intraocular lens implantation; ASR, anterior segment reconstruction; AV, anterior vitrectomy; ACIOL, anterior chamber intraocular lens implantation; PED, persistent epithelial defect of the cornea; Conj, conjunctivalization.

*  Right and left eyes of the same patient.

Table 2.

 

Table 2.

 

Results of HLA DPB1 Analysis

Case No.	Age/Sex	Original Disease	LAT	Lenticle Shape	Other Surgical Methods	Follow-up (Days)	Postoperative Complication	Donor Cornea	Recipient PBL	Graft Epithelium	Status
1	73F	P-OCP	+	R	PKP, ICCE, AMT	378		0101/0402	0402/0901	0402/0901/0101	Mix
2	77F	OCP	+	R	PKP, ECCE, IOL, AMT	349	PED, Conj	NA	0301/0501	0301/0501/0402	Mix
3	69M	OCP	+	R	PKP, AMT	326	PED, Conj	NA	0501/0501	0301/0401	Donor
4*	70F	FSA	+	C	PKP	750	PED	0201/0301	0501/1401	0201/0301	Donor
5*	70F	FSA	+	C	PKP	888		NA	0501/1401	0401/1401	Donor
6	45F	SJS	+	R	PKP	474	PED	NA	0401/0501	0401/0501	Recipient
7	58M	BK	+	S	PKP, ECCE, IOL	978	ED	0401/0401	0501/1701	0501/1701	Recipient
8	55M	CB	+	C	PKP, ECCE, IOL	651	PED, Conj	0402/1701	0201/0301	0201/1701/0301	Mix
9	20M	CB	+	R	PKP, AMT	404	Conj, Rej	0301/0401	0202/0501	0301/0401	Donor
10*	15M	KC	−		PKP	399		0301/0401	0402/0501	0402/0501	Recipient
11*	15M	KC	−		PKP	630		NA	0402/0501	0402/0501	Recipient
12*	75M	Scar	−		PKP	381		0301/0401	0201/0402	0201/0402	Recipient
13*	75M	Scar	−		PKP	182		0101/0201	0201/0402	0201/0402	Recipient
14	25F	KC	−		PKP	306		0402/1301	0201/0501	0201/0501	Recipient
15*	79F	BK	−		PKP	410		0401/0901	0402/0402	0401/0402	Mix
16*	79F	BK	−		PKP, ECCE, IOL, AV	509	PED	0402/0501	0402/0402	0402/0402	Recipient
17	58M	BK	−		PKP, ACIOL	202		0101/0201	0501/1701	0501/1701	Recipient

*  Different eyes from the same patient.

 PBL, peripheral blood lymphocyte; C, C-shaped; R, ring-shaped; S, segmental; FSA, familial subepithelial amyloidosis of the cornea; OCP, ocular cicatricial pemphigoid; P-OCP, pseudo-OCP; SJS, Stevens–Johnson syndrome; BK, bullous keratopathy; CB, chemical burn; KC, keratoconus; PKP, penetrating keratoplasty; ICCE, intracapsular cataract extraction; AMT, amniotic membrane transplantation; PEA, phacoemulcification and aspiration of the lens; IOL, posterior chamber intraocular lens implantation; ASR, anterior segment reconstruction; AV, anterior vitrectomy; ACIOL, anterior chamber intraocular lens implantation; PED, persistent epithelial defect of the cornea; Conj, conjunctivalization.

Figure 1.

 

(A) FISH analysis of a male patient with ocular cicatricial pemphigoid (OCP; case 3, Table 1 ) receiving LAT from a female donor. Two X chromosomes (red signal, arrows) are found in each cell, suggesting that these cells were of donor origin. (B) FISH picture of a female patient with pseudo-OCP (case 1, Table 1 ) who underwent LAT from a male donor. Fourteen of 15 cells were of donor (male) origin, and only 1 cell was of recipient (female) origin. One of the cells with male origin is shown. Green (arrow) and red (arrowhead) signals are found in a single cell. (C) FISH picture of a female patient with corneal scar (case 7, Table 1 ) who underwent PKP from a male donor. Only recipient-derived cells (two red signals, arrows) were found. Original magnification, (A) ×200; (B, C)× 100.

Table 3.

 

Table 3.

 

Summary of the Results of FISH and RFLP Analyses

Surgery	Origin of Epithelial Cells
		Donor Only	Recipient Only	Mixture
FISH
LAT	2	2	1
PKP	0	9	0
RFLP
LAT	4	2	3
PKP	0	7	1

The authors thank Mitsubishi Kagaku Bio-Clinical Laboratories for technical and advisory support.