Anticardiolipin antibodies in proliferative diabetic retinopathy: An additional risk factor

Abstract

Purpose

To report the prevalence of anticardiolipin antibodies in patients with proliferative diabetic retinopathy (PDR) having high-risk criteria (HRC).

Methods

Diabetic patients having PDR with HRC and diabetics free of retinopathy were compared for the presence of anticardiolipin antibodies.

Results

Among the 34 patients, 6 (17.7%) of diabetics having PDR with HRC were positive for anticardiolipin antibodies. There was no significant association of aCL antibodies with sex or type of diabetes. Using Pearson’s correlation test, no significant associations of aCL antibodies with duration of diabetes or age of patients were found. All patients who were positive for anticardiolipin antibodies had PDR with HRC. The difference was statistically significant.

Conclusion

Presence of anticardiolipin antibodies may represent an additional risk factor for PDR.

1. Introduction

Diabetic retinopathy (DR) is a major cause of blindness in the working population of the Western world; the prevalence of which is strongly related to the duration of diabetes and the glycemic control (Leske et al., 2006). Intensive diabetic management, with the goal of achieving near-normal glycemic control may prevent and/or delay the onset of DR. There is now ample evidence that the development of microangiopathy is a multifactorial process in which genetic, metabolic and growth factors may play an important role (Geiusti, 2004). Leske et al. (2006) have suggested a possible low risk of severe proliferative diabetic retinopathy (PDR) among persons of African origin than in whites. In addition, intravitreal vascular endothelial growth factor (VEGF) but not hepatocytes growth factor (HGF) has been found to be related to PDR activity (Simo et al., 2006).

Antiphospholipid antibodies (aPA) are autoantibodies that target one or more phospholipids (PL) or PL-binding proteins present on cellular membranes (McIntyre et al., 1997). They have been described in patients having lupus erythematosus (Vogel et al., 1991), recurrent abortion (Peluso and Morrone, 2007), ocular ischemia (Hughes, 1993), ischemic heart disease (Foley Nalan et al., 1991), atherosclerosis, stroke and transient ischemic attack (Toschi et al., 1998), solid organ grafts (McIntyre and Wagenknecht, 2001), and diabetes (Geiusti, 2004).

Antiphospholipid antibodies, including anticardiolipin antibodies (aCL), are detected in many conditions, but only those found in association with autoimmune disease require the presence of phospholipid binding serum protein beta-2 glycoprotein I (B2GPI) (Iverson et al., 2004). B2GPI is composed of five homologous domains numbered I–V from the N-terminus. Domains I–V are designated short consensus repeats (complement control protein repeats or “sushi domains”) and are composed of approximately 60 amino acids that contain a motif characterized by four conserved cysteine residues which form two internal disulfide bridges (Kato and Enjoji, 1991). The modified fifth domain differs from domains I–IV in that it contains 82 amino acids resulting in a C-terminal loop with an additional disulfide bond (Hunt et al., 1993). Understanding anticardiolipin autoantibodies has been the focus of significant interest due to their accepted role in the pathology of antiphospholipid syndrome (APS). Antiphospholipid syndrome is an autoimmune disorder associated with arterial and venous thromboses and recurrent fetal loss (Derksen and de Groot, 2004). Anticardiolipin antibodies are not directed against cardiolipin, but rather the complex of cardiolipin and the plasma protein cofactor beta-2 glycoprotein I (B2GPI). At present, B2GPI is considered the true antigenic target for aCL antibodies (Passam and Krilis, 2004). The presence of immunological markers in the blood vessel walls has been documented for some time (Burkholder, 1965). The presence of autoantibodies to endothelial cell surface antigens may initiate vascular injury (Bordron et al., 1998). Recent evidence suggests that DR may be due to an autoimmune process (Kastelan et al., 2007).

Antiphospholipids antibodies, including anticardiolipin (aCL) antibodies, are autoantibodies that target one or more phospholipids (PL) or PL-binding proteins present on cellular membranes (McIntyre et al., 1997), which may be associated with arterial and/or venous thrombosis (Hughes, 1993). Immunological mechanisms may play a role in the pathogenesis of diabetic microangiopathy via immune complex deposition. Antiphospholipid antibodies directed against endothelial antigens may be responsible for initiating vascular injury, these could be a marker of endothelial dysfunction (Geiusti, 2004). There is a recognized association between aCL antibodies presence and vascular occlusive disease (Kalogeropoulos et al., 1998).

The purpose of this study was to determine the prevalence of anticardiolipin antibodies in patients with proliferative diabetic retinopathy (PDR) with high-risk criteria (HRC) and to investigate whether aCL antibodies presence may be relevant to the pathogenesis DR.

2. Patients and methods

The study was approved by the human ethics committee and all patients signed informed consent before entering the study. All applicable institutional and governmental regulations concerning the ethical use of human volunteers were followed during this research.

Two hundred consecutive diabetic patients attending Mansoura Ophthalmology Center from November, 2006 to February, 2007 were screened for signs of PDR with HRC. Patients were classified according to the National Diabetes Data Group (National Diabetes Data Group, 1979). Retinal vascular morphology was evaluated by colour fundus photography. High-risk criteria were defined according to the Diabetic Retinopathy Study Research Group criteria (Diabetic Retinopathy Study Research Group, 1981). Exclusion criteria were poor glycemic control (HbA1c > 9%), borderline hypertension (>140/90), hypertriglyceridaemia (>1.9 mmol/L), or hypercholesterolaemia (>5.6 mmol/L), smoking and hazy media.

Thirty-four diabetics were enrolled in the study. Thirty-six comparable diabetic patients free of diabetic retinopathy were selected as control group. The groups were comparable with regard to age, sex and duration of disease. All 70 patients underwent ophthalmologic examination including visual acuity, slit lamp examination, fundus contact lens examination and fluorescein angiography. Five milliliters of venous blood was withdrawn from each patient. The samples were centrifuged and separated sera were stored in deep freezer at −70 °C until the time of assay. Serum samples were examined for anticardiolipin antibodies. IgG and IgM were determined using Enzyme-Linked Immuno-Sorbent Assay (ELISA) technique.

Anticardiolipin antibodies (IgG and IgM) were determined by using ELISA. Patients were considered positive for IgG aCL antibodies when it was >10 GPL unit/ml and negative when the titer <10 GPL unit/ml. Patients were considered positive for IgM aCL antibodies when it was >7 MPL unit/ml and negative when the titer was <7 MPL units/ml. For descriptive statistics of the quantitative variables, the mean, range and standard deviation were used to describe central tendency and dispersion. Differences between the means of the independent samples were analyzed by the Student’s t tests. Pearson’s correlation and Fisher’s exact probability tests were used to compare study and control groups and to test association between aCL and other factors such as age, sex, duration of the disease, type of diabetes, and presence of PDR with HRC. All tests were considered significant if (P < 0.05).

3. Results

The study included 70 diabetic patients (33 males, 37 females, age range 22–68 years). Thirty patients were type 1 diabetics while 40 patients were type 2 diabetics. Thirty-four patients had PDR with HRC while 36 were free from diabetic retinopathy. Both groups showed no statistical difference regarding age of patients, duration of diabetes, sex or type of diabetes (Table 1).

Table 1.

Characteristics of patients having PDR with HRC and control diabetics free of diabetic retinopathy.

	Diabetic patients free of diabetic retinopathy	PDR patients with HRC	P value
Age (years)	44.9 ± 15.1	46.5 ± 13.8a	0.3
Sex (M/F)	15(42)/21(58)	18(53)/16(47)b	0.4
Duration (years)
Mean ± SD	13.1 ± 2.7	14 ± 3.1	0.4
Median	14	13
Range	8–20	9–18
Type of diabetes (1/2)	12(33%)/24(67%)	18(53%)/16(47%)	0.1
Total number (%)	36(51.4%)	34(48.6%)

^aMean ± SD.

^bNumber of patients (percentage of total).

Table 2 summarizes aCL positivity. Six patients were positive for aCL antibodies. Fisher’s exact probability test showed no significant associations of aCL antibodies with sex or type of diabetes. Using Pearson’s correlation test, no significant associations of aCL antibodies with duration of diabetes or age of patients were found. Six of 70 diabetic patients (8.6%) were positive for aCL antibodies. All of them had IgG antibodies while two (2.9%) had IgM antibodies. 17.7% of patients having PDR with HRC were positive for aCL. All patients with aCL positivity had PDR with HRC. Most frequent isotype IgG was more common in type 1 than type 2. Five patients were type 1 diabetics (two of them had both IgG and IgM antibodies and three had IgG only) while one patient was type 2 diabetic (IgG antibodies). None of diabetic retinopathy-free patients was positive for aCL antibodies.

Table 2.

Prevalence of aCL antibodies in patients having PDR with HRC and control diabetics free of diabetic retinopathy.

	Type 1 diabetics free of retinopathy	Type 1 diabetics having PDR with HRC	Type 2 diabetics free of retinopathy	Type 2 diabetics having PDR with HRC
IgG	0 (0%)a	5 (27.8%)	0 (0%)	1 (6.3%)
IgM	0 (0%)	2 (11.1%)	0 (0%)	0 (0%)
Total number	12	18	24	16

^aNumber of patients (percentage of total).

The presence of aCL was detected in 6 (17.7%) patients (three males and three females) having PDR with HRC. Fisher’s exact probability test was statistically significant (P = 0.01). In type 1 diabetics, aCL was found in 27.8% of PDR patients with HRC. On the other hand, aCL was found in only 6.3% of type 2 diabetics having PDR with HRC (Table 3).

Table 3.

The frequency of aCL antibodies in the studied patients.

	No retinopathy	PDR with HRC	P valuea
Type 1	0/12 (0%)	5/18 (27.8%)	0.12
Type 2	0/24 (0%)	1/16 (6.3%)	0.4
Total	0/36 (0%)	6/34 (17.7%)	0.01

^aFisher’s exact test.

4. Discussion

The vascular endothelium is a major regulator of haemostasis. Vascular damage and endothelial cell dysfunction occur early in the course of diabetic microangiopathy (Barnett, 1991). The net effect of these changes is the conversion of endothelium from a thromboresistant to a thrombogenic surface (Gargiulo et al., 1997). Autoantibodies to endothelial cell surface antigens initiate vascular injury (Bordron et al., 1998). The molecular mechanism by which aCL antibodies activate vascular endothelial cells is not known. Numerous studies suggest that exogenous B2GPI, a circulating glycoprotein, is necessary, and aCL recognize a complex antigen that includes B2GPI and anionic phospholipids (McNeil et al., 1990). It is likely that the interaction of circulating B2GPI with endothelial cell anionic phospholipid induces formation of a neoepitope that confers recognition specificity for aCL. Association between the presence of the IgG isotype and thrombosis in patients with APS has been demonstrated previously (Cabiedes et al., 1995).

Although B2GPI has been shown to be necessary for the anticardiolipin-mediated pathophysiological effect, a suggestion that an endothelial cell membrane protein acts as a cofactor cannot be excluded (Simantov et al., 1995). However, there has been no reliable predictor of thrombotic events in patients with aCL antibodies. Anticardiolipin antibodies are found in the immunoglobulin classes IgG, IgM and/or IgA. The determination of IgA antibodies seems to have a greater importance in the African-Caribbean population (Molina et al., 1997). In the present study IgG and IgM were studied. Horbach et al. (1996) demonstrated the relationship between IgM, aCL antibodies and venous thrombosis. Other study found a significant association between IgG isotype and thrombosis, especially arterial but not venous (Ogawa et al., 2000). In addition, aCL antibodies may be associated with occlusive ocular disorders. The most represented feature of posterior involvement in patient with abnormal aCL antibodies was retinal vasculitis (60%) (Miserocchi et al., 2002). A statistically significant prevalence of aCL antibodies has been found in patients with retinal vascular occlusive disorders in the absence of main accepted risk factors for retinal thrombosis (Cobo-Soriano et al., 2000). A simultaneous bilateral central retinal vein occlusion was found to be associated with anticardiolipin antibodies in a leukamic patient (Al-Abdulla et al., 2001). The incidence of aCL IgG was higher in patients with Behcet’s disease with non-occlusive thrombosis than in patients with retinal occlusive events or in patients without occlusions or thrombosis (Ermakova et al., 2002).

Although aCL antibodies have been associated with a large number of diseases, there are only a few studies of these antibodies in DR patients. In the present study aCL antibodies were examined in patients having PDR with high-risk criteria. Both type 1 and type 2 diabetics were studied because they differ in their auto antibodies (Wroblewski et al., 1998) and in their immunogenetics (Schranz and Lemmark, 1998) and they do not share genetic susceptibility loci (Elbein et al., 1997). In their study, Gargiulo et al. (1999) found that the prevalence of IgG aCL antibodies in diabetic population (type 1 and type 2) was found to be 8.9% while the prevalence of IgM aCL was found to be 5.9%. In their normal population, the prevalence of IgG aCL was found to range from 4% to 4.9% (Ahmed et al., 1999; Merkel et al., 1996; Diez et al., 1993) and the prevalence of IgM aCL was found to be 2.8% (Diez et al., 1993).

The levels of IgM aCL antibodies has been found to be more prevalent in type 1 diabetics with vascular complications than in type 1 diabetics without complications (Ahmed et al., 1999). The incidence of antiphospholipid antibodies has been found to be equally high in type 1 and type 2 diabetics with retinopathy (Simo et al., 2006). The autoantibodies to endothelial cell surface antigens that initiate vascular injury (Bordron et al., 1998; Jones et al., 1992), can be associated with arterial and/or venous thrombosis (Hughes, 1993). The most plausible explanation for endothelial proliferation and neovessels formation in the retina is ischemia of its inner layers secondary to closure of parts of retinal capillary bed (Michelson, 1948; Wise, 1956). The ischemic retina produces a new vessel-stimulating factor, capable of acting locally and diffusing through the vitreous to other areas of the retina, to the optic disc, and into the anterior chamber (Wise, 1956; Patz, 1982). The possibility that these antibodies may be secondary to neovessels formation should be considered in further studies.

An immunological basis for diabetic microangiopathy has been suggested by studies showing an increased deposition of immunoglobulins and antigen–antibody complexes in the basement membrane of small blood vessels (Bloodworth, 1968). Mimura et al. found that the production or existence of glutamic acid decarboxylase autoantibodies may contribute to the prevention of retinopathy (Mimura et al., 2004). Diabetic retinopathy has been suggested to be an autoimmune disease based on the finding of antipericyte and antiendothelial cell autoantibodies in the circulation of diabetic patients. More evidence implicates the presence of autoimmune mechanisms in the proliferative stage of this disease: elevated levels of tumor necrosis factor-alpha, interleukin-8 and soluble interleukin-2 receptor in the serum of diabetic patients, increased vitreous concentration of the interleukin-6 and interleukin-8 in patients with PDR (Kastelan et al., 2007).

In the present study, there is significant increase in the aCL in patients having PDR with HRC suggesting that aCL could represent an additional risk factor for proliferative diabetic retinopathy. The pathological shift of vascular endothelium from antithrombotic to a prothrombotic state may be more pronounced in subjects with anticardiolipin positivity and/or high circulating immune complexes concentrations since they possess the ability not only to induce platelet activation and aggregation but also to activate the complement system via the classical pathway. Therefore, a potential synergism between generation of antibodies, haemostatic alterations and endothelial stress has been suggested (Geiusti, 2004).

Further immunochemical investigations are needed to clarify whether this autoimmune reaction is the cause, consequence or aggravating factor of the disease. With the understanding of new cellular and molecular mechanism, a new generation of therapeutic strategies may be envisioned.