Thrombospondin-1 Polymorphisms Influence Risk of Corneal Allograft Rejection

Abstract

Purpose.: We investigated single nucleotide polymorphisms (SNPs) of thrombospondin-1 (TSP-1) on the risk of corneal allograft rejection. The TSP-1 is known to be involved in the immune response of the anterior chamber of the eye, activating TGF-β₂, promoting peripheral and systemic tolerance, and counteracting the proangiogenic activity of VEGF.

Methods.: Three tagging SNPs spanning the TSP-1 region (rs1478604, A>G; rs2228261, C>T; and rs2228262, A>G) were genotyped. Association with risk of rejection was investigated in a group of 378 corneal transplant recipients with risk factors for allograft rejection. Transplant recipients had completed 3-year follow-up.

Results.: The TSP-1 rs1478604 A SNP was associated significantly with an increased risk of corneal allograft rejection (odds ratio [OR], 1.58; 95% confidence interval [CI], 1.02–2.45; P = 0.04) and there was a trend toward the rs1478604, rs2228261, rs2228262 ACA haplotype increasing risk of rejection.

Conclusions.: The results suggest that TSP-1 rs1478604 AA homozygotes may be at increased risk of corneal transplant rejection, especially if they carry the ACA haplotype.

Introduction

The eye is one of a number of sites within the body demonstrating a highly regulated relationship with the immune system. Unlike solid organ transplants, the success of human leukocyte antigen (HLA) unmatched corneal grafts with minimal immunosuppression illustrates the contribution of immune privilege in the anterior chamber to the prevention of allograft rejection. ¹ In addition, the high acceptance of corneal allografts in low-risk keratoconus patients provides further compelling evidence for this immune privilege of corneal allografts. ² Mechanisms underlying ocular immune privilege are well documented, ^3–6 and have been attributed to various factors, including absence of blood and lymphatic vessels in the graft bed in low risk penetrating keratoplasty (PKP), an immunosuppressive ocular microenvironment due to regulatory molecules, such as TGF-β, thrombospondin-1 (TSP-1), α-melanocyte stimulating hormone (α-MSH), vasoactive intestinal peptide (VIP), calcitonin gene-related peptide (CGRP), and cortisol, ^7–9 and anterior chamber–associated immune deviation (ACAID), ^10,11 in which antigen-specific delayed-type hypersensitivity (DTH) is suppressed. However, immune privilege is not absolute and a major cause of corneal transplant failure is immunologic rejection. ¹² Overall, corneal graft survival for all indications is 90% at one year, but this falls to 70% by 5 years, which is lower than for renal transplantation. ¹³ In patients who have suffered at least one rejection episode, 5-year survival is only 50% and this falls to less than 35% at 10 years. ¹⁴ Therefore, reducing the risk of allograft rejection is an important goal. 

Thrombospondins are multi-domain, calcium binding, extracellular glycoproteins, which help maintain an anti-angiogenic environment in the eye. ^15–17 The human TSP-1 gene (THBS1), located on chromosome 15q15, ¹⁸ is the best studied member of the TSP family, ¹⁹ and TSP-1 binds latent TGF-β and promotes its activation. ²⁰ Approximately 80% to 90% of the TGF-β activity in the aqueous humor (AH) of rabbits and humans is due to TGF-β₂. ²¹ The TSP-1 glycoprotein is expressed constitutively by the cornea in mice ²² and humans, ²³ and has been reported consistently in corneal epithelial basement membrane, the corneal endothelium, posterior Descemet's membrane, the trabecular meshwork, lens epithelium, and blood vessels of normal adult mammalian eyes. The TSP-1 glycoprotein also is present in the intraocular fluids and drainage pathway, where it may function in maintaining the antiangiogenic environment and in IOP control, respectively. ²⁴ Regulation of angiogenesis in the eye, cornea and retina, is important in maintaining the integrity of the visual axis. The TSP-1 glycoprotein is thought to inhibit hemangiogenesis through direct effects on vascular endothelial cell migration and survival. ²⁵ Lymphangiogenesis also has a major role in corneal transplant outcome by providing a conduit for foreign antigen to migrate to regional lymph nodes. ²⁶  

The TSP-1 glycoprotein has been shown to be a major activator of TGF-β under in vitro and in vivo conditions, ^20,27 and TGF-β₂ can exert inhibitory activity on T lymphocytes, pertinent as investigation of rejected human corneal allografts have shown intense stromal inflammatory infiltrates of T cells, macrophages, MHC II+ dendritic cells (DCs), and other effector cells. ^28,29 Also, TGF-β₂ functions as a suppressor of DC maturation, promoting the generation of phenotypically and functionally immature DCs, resident antigen presenting cells (APCs) found in human corneal stroma. ^30,31 In mice, it has been demonstrated that ocular immune privilege, a phenomenon known to be dependent on active TGF-β, was lost in TSP-1–deficient mice and led to irreversible destruction of the retina. ³² In humans, active, ³³ but not total TGF-β₂ ³⁴ is reduced during endothelial immune reactions following PKP, suggesting that active TGF-β₂ has a protective effect on corneal grafts. Furthermore, TGF-β₂ is statistically significantly increased in the AH of patients with keratoconus compared to patients with various other corneal diseases, ³⁵ which may explain partially the excellent prognosis of keratoconic patients. Given the crucial role of TSP-1 in activating TGF-β₂, and its role in maintaining corneal avascularity through suppression of hem- and lymphangiogenesis, TSP1 was selected as a candidate gene for investigation in corneal transplant outcome. 

Methods

Study Subjects

We included in the present study 378 patients participating in the Corneal Transplantation Follow-up Study II (CTFS II), a prospective multicenter clinical study investigating the role of human leukocyte antigen (HLA) Class II matching, against a background of HLA Class I matching, on allograft rejection in high risk grafts (International Randomized Control Trial Number ISRCTN25094892, NIHR CRN Portfolio Study: 9871). The CTFS II was approved by South-West Multi-Centre Research Ethics Committee (MREC 97/6/8 main study and MREC/01/6/77 for the candidate gene association study). Local research ethics committee approval was obtained by participating surgeons. Patients with risk factors for allograft rejection were identified by ophthalmologists and registered for participation in CTFS II after giving informed consent. The study adhered to the Declaration of Helsinki. Patients were recruited from England, Scotland, and Ireland, and patient selection criteria were established from a previous study ³⁶ that showed the following conditions increased the risk of immunologic rejection of corneal grafts: vascularized cornea, previous corneal graft rejection, regraft, active inflammatory disease, bullous keratopathy, herpes infection, uveitis. Transplant and follow-up data were collected through NHS Blood and Transplant (NHSBT) ocular tissue transplant audit. Patients in the present study were split into two groups: graft survivors (GS) had remained rejection-free for three years and graft failures (GF) had suffered rejection episodes within three years of transplantation, or had failed owing to irreversible rejection/endothelial decompensation. Graft failures due to definite nonimmunologic causes, such as primary graft failure, acquired infection, or recurrence of original disease were excluded. All patients received penetrating keratoplasties (PK) and had completed three years of follow-up. 

DNA Collection

Corneal transplant recipient DNA used for HLA tissue typing was collected from NHSBT Histocompatibility and Immunogenetics Laboratories located within the United Kingdom. 

Genotyping of TSP-1 Single Nucleotide Polymorphisms (SNPs)

Induced heteroduplex generator (IHG)–based genotyping was utilized for TSP-1. ³⁷ Three TSP-1 tagging SNPs (rs1478604, A>G; rs2228261, C>T; and rs2228262, A>G) were selected using SNP tagger ³⁸ from the Centre d'Etude du Polymorphisms Human from Utah (CEU) population of the HapMap project, ³⁹ using Haploview software ⁴⁰ (Table 1). Primer3 was utilized to select primers for all 3 SNPs (Table 2). ⁴¹ The TSP-1 IHG sequences were designed that mimicked the genomic sequence of each SNP except for the addition of a poly-adenine insert adjacent to the polymorphic site (Table 3). RNA structure version 4.4 ⁴² was used to visualize 2-dimensional models of heteroduplex formation. 

Table 1

 

TSP-1 SNP Features

SNP rs Number	Region	Chromosomal Position	Protein Position	Function	1000 Genomes Allele Frequency, GBR
rs1478604	5′UTR	15:39873321	–		A 69%: G 31%
rs2228261	Coding	15:39880358	470	Synonymous	C 88%: T 12%
rs2228262	Coding	15:39882178	700	Missense	A 86%: G 14%

Table 2

 

TSP-1 SNP Primer Sequences

TSP-1 SNPs	Primer Sequence (5′–3′)	Oligo Length, bp	Anneal Tm	N of PCR Cycles
rs1478604	F CGTCCGGAGTAGAGGTTGCT	20	60°C	32
	R GGAGGAATCCCCAGGAATG	19
rs2228261	F TGTGACATGTGGTGATGGTG	20	60°C	32
	R GCAGGCTTTGGTCTCCCGCG	20
rs2228262	F TGGCAATGGCATCATCTG	18	60°C	32
	R CACACAAAGGACCTGGCTCT	20

Table 3

 

IHG Sequences

TSP-1 SNPs	Oligo Length, bp	IHG Sequence
rs1478604	135	CGTCCGGAGTAGAGGTTGCTCCTGGAGAGCGAAA[A]CAGGAGCCCTGAACTCGCAGGCCAGCT CGGGCGCAGCGGCTGGCAAGGCGGAGGAGCCGCGCGCTTTTAAAGGGGCGCT CGCATTCCTGGGGATTCCTCC
rs2228261	112	TGTGACATGTGGTGATGGTGTGATCACAAGGATCCGGCTCTGCAACTCTCCCAGCCCCCAGAT GAAAAA[C]GGGAAACCCTGTGAAGGCGAAGCGCGGGAGACCAAAGCCTGC
rs2228262	122	TGGCAATGGCATCATCTGCGGGGAGGACACAGACCTGGATGGCTGGCCCAAAA[A]TGAGAAC CTGGTGTGCGTGGCCAATGCGACTTACCACTGCAAAAAGGTAGAGCCAGGTCCTTTGTGTG

PCR Parameters

The PCR amplifications were carried out in a total reaction volume of 20 μl. Primers were from Sigma-Aldrich (St. Louis, MO, USA). The PCR parameters were: 0.5 μM each of forward and reverse primers, 1.5 mM MgCl_2, 200 μM of each dNTP, 1X Taq polymerase buffer, comprising 67 mM tris-HCL, pH 8.8, 16 mM (NH₄)₂SO_4, 0.2 units Taq polymerase (Abgene, Epsom, Surrey, UK), and either 50 ng of genomic DNA or diluted IHG reagent. The PCR parameters were optimized for a Biometra II thermal cycler (Biometra, Ltd., Goettingen, Germany): initial denaturation at 94°C for 5 minutes; 32 cycles of 94°C for 1 minute (denaturation), 60°C for 1 minute, 72°C for 1 minute; final extension at 72°C for 5 minutes. 

SNP Genotyping Using IHG Analysis

The SNP genotyping of corneal transplant recipient DNA was carried out using IHG methodology as published. ³⁷ The IHG sequences were synthesized as single oligonucleotides (MWG, Ebersberg, Germany). Genomic DNA and IHG reagents were amplified separately, using the same PCR primers and cycling conditions. Synthetic oligonucleotides containing specific polymorphisms were utilized to confirm banding patterns (see Fig.). 

Figure

 

Homozygotes and heterozygotes for the 3 TSP-1 SNPs.

Statistical Analysis

Genotype and allele distributions were compared between GS and GF groups using Yates corrected χ² analysis using OpenEpi version 2.3.1. ⁴³ Results were expressed as odds ratio (OR) with 95% confidence intervals (95% CI). Where OR could not be calculated, Fisher's exact test was used. The Hardy-Weinberg equilibrium was tested for each locus, within the GS and GF groups (HWE P), with 1 degree of freedom. ^44,45 Haplotypes were inferred using PHASE, version 2.0.2. ^46,47 Bonferroni correction for multiple comparisons ⁴⁸ was applied to extended haplotypes, using the formula P _c = kP (P _c represents the corrected probability, and k = the number of comparisons performed in individual genes), with the restriction that P _c ≤ 1. Arlequin version 2.000 ⁴⁹ was used to determine pairwise linkage disequilibrium (LD) between paired polymorphisms using an extension of the Fisher's exact probability test through contingency tables. The LD was expressed as Lewontin's D' coefficient. The level of significance was set at the 5% level. Based on sample size and minor allele frequencies of SNPs investigated, our study had 80% power, with 95% confidence to detect differences in allele frequencies between GS and GF with minimum ORs of (rs1478604) ≥ 1.55, (rs2228261) ≥ 1.83, and (rs2228262) ≥ 1.77, respectively. 

Results

The demographics of the corneal transplant recipients are shown (Table 4). Indications for grafting showed a preponderance of regrafts, previous ocular surgery (primarily pseudophakic bullous keratopathy), and dystrophies (primarily Fuch's endothelial dystrophy with other risk factors for rejection). The dystrophy group contained more than twice the number of females than males, which is typical for this indication. The highest number of rejection episodes occurred in the regraft and injuries group. 

Table 4

 

Indications for Corneal Transplant

Indication	n	Age at Transplantation		Sex		% Rejection
		Mean	Range	Male	Female
Ectasias	5	64	30–84	4	1	20
Dystrophies	54	64	26–88	16	38	16
Previous ocular surgery	93	64	24–92	47	46	31
Infection	49	64	25–91	23	26	16
Injury	5	64	22–69	5	0	60
Ulcerative keratitis	7	65	33–86	3	4	0
Regrafts	129	64	21–89	68	61	50
Opacification	9	65	36–85	3	6	11
Miscellaneous	8	67	34–92	3	5	13

High Risk Corneal Allograft Rejection Is Associated With the rs1478604 A>G Polymorphism

A significant association between the rs1478604 A>G SNP was observed with a preponderance of the AA genotype 54% in the GF group compared to 43% in the GS group, which relates to an increased chance of graft rejection in AA homozygotes (OR, 1.58; 95% CI, 1.02–2.45; P = 0.04, Table 5). There were no significant associations for rs2228261 or r2228262. All SNPs investigated were in Hardy-Weinberg equilibrium apart from rs2228261 in the GS group; however, total allele frequencies for this SNP were identical to those from the 1000 Genomes GBR subpopulation (Tables 1, 5). 

Table 5

 

Genotype and Allele Frequencies

SNP	Genotype/Allele	GS (%)	GF (%)	Total Freq	OR	CI	P
		n = 230	n = 148
rs1478604	A A	98 (43)	80 (54)		1.58	1.02–2.45	0.04*
	A G	107 (46)	56 (38)		0.7	0.45–1.09	0.12
	G G	25 (11)	12 (8)		0.72	0.33–1.56	0.48
	A	303 (66)	216 (73)	69	1.4	1–1.95	0.04*
	G	157 (34)	80 (27)	31	0.71	0.51–1.00	0.04*
	HWE P	−0.6	−0.62
rs2228261	C C	176 (76)	116 (78)		1.11	0.66–1.88	0.76
	C T	54 (24)	32 (22)		0.9	0.53–1.52	0.76
	C	406 (88)	264 (89)	88	1.1	0.67–1.79	0.78
	T	54 (12)	32 (11)	12	0.91	0.56–1.48	0.78
	HWE P	−0.04	−0.14
		n = 219	n = 140
rs2228262	A A	161 (73)	108 (77)		1.22	0.72–2.06	0.51
	A G	52 (24)	31 (22)		0.91	0.53–1.56	0.82
	G G	6 (3)	1 (<1)				0.25†
	A	374 (85)	247 (88)	86	1.28	0.80–2.06	0.33
	G	64 (15)	33 (12)	14	0.78	0.49–1.25	0.33
	HWE P	−0.47	−0.44

 Total Freq, total study allele frequency, that is, GS + GF; P, probability.

*  Significant P value.

†  Fisher's exact test.

Trend Toward the ACA Haplotype Associating With Risk of Rejection

Using PHASE analysis, we identified four frequent haplotypes (>1%, Table 6). Haplotypes with a frequency of <1% were grouped as “others.” The three locus haplotypes rs1478604, rs2228261, rs2228262 ACA showed an increase in the GF group 71% compared to the GS group 64% (P = 0.06), but was not significant. However, as ACA was the only haplotype with an A allele at rs1478604, and the A allele was significantly associated with risk of rejection (P = 0.04), it follows that patients carrying this haplotype may be at increased risk of graft rejection. Pairwise LD between all 3 TSP-1 SNPs revealed a high level of LD (rs1478604 and rs2228261, D' 0.95; rs1478604 and rs2228262, D' 0.89; and rs2228261 and rs2228262, D' 1.00). The G allele of rs2228262 (N700S) associated with the G allele of rs1478604. The T allele of rs2228261 associated with the G allele of rs1478604 and the A allele of rs2228262. Haplotype frequencies were slightly different compared to the original CEU panel used to select the tag SNPs, and despite there being a high level of LD, SNPs were not 100% correlated; that is, SNP rs2228261 T allele would always associate with the A allele of rs2228262, but a C allele at rs2228261 could associate with either a G or A allele at rs2228262. 

Table 6

 

TSP-1 Haplotype Frequencies

Haplotype	GS, n = 219 (%)	GF, n = 140 (%)	OR	95% CI	P	P c
A C A	282 (64)	200 (71)	1.38	0.99–1.94	0.06	NS
G C G	59 (13)	31 (11)	0.80	0.49–1.30	0.343	NS
G T A	52 (12)	29 (10)	0.86	0.52–1.42	0.531	NS
G C A	38 (9)	17 (6)	0.68	0.36–1.27	0.200	NS
Others	7 (2)	3 (1)

 TSP-1 rs1478604, rs2228261, and rs2228262 inferred haplotypes from PHASE.

Discussion

Allograft rejection remains a leading cause of corneal transplant failure. ⁶ With increasing evidence of the importance of genetic factors and gene-environment interactions in the etiology of eye-related disorders, including allograft rejection, identification of genetic variation associated with disease risk may help provide insight into the mechanisms of disease pathogenesis, and help identify novel targets for preventative and therapeutic interventions. The results of this study showed that risk of rejection in high risk corneal transplantation is associated with the rs1478604 A>G polymorphism, with the A allele in the 5′ untranslated region (5′UTR) conferring additional risk. The 5′UTR is a regulatory region of DNA situated at the 5′ end of all protein-coding genes that is translated into mRNA, but not translated into protein. Accumulating evidence indicates that it is variation within these noncoding sequences that produces phenotypic variation between individuals ⁵⁰ ; however, functionality of these TSP-1 SNPs has yet to be confirmed and warrants further investigation depending on favorable ethical approval. There also was a trend toward the haplotype ACA associating with increased risk of rejection (P = 0.06), but this was not significant. As with all gene association studies, however, replication of our results in a larger cohort of similar ethnicity is required, exemplified by the rare GG genotype of rs2228262 found in our study, but not in a South Indian cohort. ⁵¹  

The TSP-1 glycoprotein interacts with a variety of factors in a synergistic way, playing a crucial role in many stages of the inflammatory response. Mounting evidence has qualified TSP-1 as a key antiangiogenic factor in ocular immune privilege. The TSP-1 glycoprotein can inhibit angiogenic responses by interacting with VEGF directly or indirectly by engaging several TSP-1 receptors, such as CD36, a heavily glycosylated transmembrane scavenger receptor, or CD47. ^16,25 Tethering of TSP-1 by APCs via CD36 or CD47 is paramount for TGF-β activation and function. ⁵² The TSP-1 activation of TGF-β is thought to inhibit hemangiogenesis through direct effects on endothelial cell migration and survival, for example, by inducing vascular endothelial cell apoptosis through binding CD36. ^25,53 Lymphangiogenesis is induced mainly by ligation of VEGF-C and VEGF-D to their high affinity receptor VEGFR3 on lymphatic vascular endothelium, and under inflammatory conditions, macrophages are known to be the primary source of VEGF-C. ⁵⁴ Ligation of TSP-1 with CD36 on inflammatory macrophages, regulates the expression of the main lymphangiogenic growth factor VEGF-C; therefore, limiting the amount of lymphangiogenesis. ¹⁵ In an animal model, 6-month-old TSP-1–deficient mice developed increased spontaneous corneal lymphangiogenesis compared to wild type mice and, similarly, in a model of inflammation-induced corneal neovascularization, young TSP-1–deficient mice developed exacerbated lymphangiogenesis. ¹⁵ Studies of TSP-1 knockout heterozygous and homozygous mice have shown that TSP-1 deficiency significantly decreases the availability of bioactive TGF-β, and leads to enhanced IL-17A expression from inflammatory CD4+ cells detected in the lacrimal gland, ⁵⁵ and that inflammation during experimental autoimmune uveitis in TSP-1–deficient mice leads to irreversible destruction of the retina. ³² A study investigating TSP-1 in corneal transplantation in mice found that TSP-1 is a potent suppressor of immune rejection. Investigators concluded that APC-derived TSP-1 suppresses their capacity to allosensitize T cells, and this regulation stems from their resistance to taking on a phenotypically and functionally mature form, and, therefore, making them less likely to migrate to regional lymph nodes, thereby suppressing DTH, the main effector mechanism of rejection. ⁵⁶ However, the concept of ACAID, in which antigen-bearing APCs migrate from the eye, is not proven in humans and other animal studies have demonstrated that cells do not need to leave the ocular microenvironment for antigen to induce a reduced DTH. Instead, antigen may travel from the eye to the spleen, lymph nodes of the head and neck, and mesenteric lymph nodes in a soluble form through blood and lymph. ^57,58 Given that resident corneal cells express TSP-1 and it is present in intraocular fluids, ^23,24 however, it is likely that TSP-1 is expressed by many cell types, utilizing multiple mechanisms to inhibit angiogenesis and to help maintain immunologic tolerance in the transplantation setting. 

The Arf-inducing transcription factor Dmp-1 encodes a transcriptional activator of TSP-1. Factor Dmp1 directly binds to the genomic sequence of TSP-1 promoter/enhancer regions and TSP1 genes are down-regulated in Dmp-1-knockout mice. ⁵⁹ The Dmp1 promoter is repressed by NF-κB p65 subset ⁶⁰ and, conversely, blocking activation of NF-κB upregulates TSP-1 expression in rat granulation tissue. ⁶¹ Given that Dmp1 is a transcriptional activator of TSP-1 and is repressed by NF-κB p65, it may be pertinent to investigate whether there are any SNPs in the TSP-1 promoter or downstream region, which may alter Dmp1 binding propensity and analyze LD with the ACA haplotype. Expression of cytokines, such as TNF-α, which we have studied previously (submitted for publication), and IL-6 are known to be regulated by NF-κB. This linking of TSP-1–mediated mechanisms to these fundamental transcriptional pathways warrants further investigation. 

In conclusion, we presented novel genotype and haplotype data for the TSP1 gene and have shown association with corneal allograft outcome. The TSP-1 glycoprotein acts as a potent inhibitor of angiogenesis, and its expression in the normal cornea and up-regulation in models of wound healing suggest that it is essential for the maintenance of equilibrium between a pro- and antiangiogenic environment. Targeting TSP-1–mediated TGF-β activation may provide a new therapeutic approach for treatment of corneal neovascularization in high risk patients and increase graft survival rates. An alternative strategy for promoting graft survival could be targeted upregulation of TSP-1 in APCs. In a murine study, local application of TSP-1 successfully inhibited inflammation-induced corneal lymphangiogenesis ¹⁵ ; thus, highlighting the potential therapeutic effects of antilymphangiogenic TSP-1. However, an understanding of the factors and elements involved in the regulation of a particular gene is of paramount importance when designing molecular therapies or when attempting to modulate the expression of a gene. 

Acknowledgments

The authors thank the contributing ophthalmic surgeons, NHS Blood and Transplant for transplant and follow-up data collection, and Departments of Histocompatibility and Immunogenetics throughout the United Kingdom for corneal transplant recipient DNA extraction. 

Supported by the National Eye Research Centre. 

Disclosure: H.L. Winton, None; J.L. Bidwell, None; W.J. Armitage, None