Comparative Analysis of Aqueous Humor Cytokine Levels Between Patients With Exudative Age-Related Macular Degeneration and Normal Controls

Abstract

Purpose.: To investigate the cytokine markers associated with exudative AMD present in aqueous humor. This goal was achieved by comparing the concentrations of more than 500 molecules in aqueous humor, between exudative AMD patients and controls.

Methods.: Aqueous humor samples were acquired from 20 patients with exudative AMD and 20 control subjects. Raybio human antibody array technology was used to simultaneously screen for any difference in the expression of any of 507 molecules. To validate the antibody array result, concentrations of insulin-like growth factor binding protein 2 (IGFBP-2), insulin-like growth factor-1 (IGF-1), and VEGF were measured by ELISA.

Results.: Twenty molecules studied exhibited intergroup differences. Twelve molecules including IGFBP-2, IGFBP-6, IGFBP-7, and glucocorticoid-induced tumor necrosis factor receptor family related gene (GITR) ligand, were detected in high densities in exudative AMD patients. Eight other molecules were present at higher concentrations in control patients. ELISA confirmed that IGFBP-2 levels were higher in patients with exudative AMD (7.47 ± 6.19 ng/mL) in comparison with control subjects (3.07 ± 3.34 ng/mL, P = 0.008). IGF-1 and VEGF levels were also increased in the former group (2.20 ± 0.26 vs. 1.99 ± 0.35 ng/mL, P = 0.040; 122.25 ± 63.24 vs. 86.98 ± 44.41 pg/mL, P = 0.048, respectively).

Conclusions.: The pattern of cytokine expression in the aqueous humor of exudative AMD patients varies from that of normal control subjects. The increased levels of IGFBP-2 and IGF-1 in exudative AMD eyes indicate that the altered expression of IGF-related molecules may be involved in disease pathogenesis and suggests potential biomarkers for exudative AMD.

Introduction

Age-related macular degeneration is one of the leading causes of irreversible blindness in the elderly. Age-related macular degeneration is often grouped into two clinical categories: the “dry” atrophic form and the “wet” exudative form. Exudative AMD is characterized by the presence of choroidal neovascularization, which accounts for approximately 80% of cases with severe visual loss. 

The discovery that VEGF is associated with exudative AMD was followed by the development of anti-VEGF agents that can dramatically improve central visual acuity in patients with exudative AMD. ^1–4 Nevertheless, the pathophysiology of exudative AMD remains to be elucidated. Numerous studies have demonstrated that various cytokines as well as VEGF were found in elevated concentrations in patients with exudative AMD. This group included basic fibroblast growth factor, platelet-derived growth factor, erythropoietin, monocyte chemoattractant protein-1 (MCP-1), interleukin-6 (IL-6), and IL-8. ^5–11 Notably, each study investigated only a few of these molecules. 

Raybio is an antibody-based protein array technology (RayBiotech, Inc., Norcross, GA) developed recently to simultaneously scrutinize 507 cytokines, chemokines, growth factors, and other molecules. It is thought to provide a new tool with which to deepen our understanding of the abstruse molecular pathophysiology of exudative AMD. This technology could even be used to identify cytokine biomarkers for exudative AMD. In this study, we used Raybio technology to compare the aqueous humor cytokine profiles of exudative AMD patients with those of normal control subjects. 

Materials and Methods

Preparation of Patient Samples

Twenty exudative AMD patients who visited Seoul National University Bundang Hospital from October 2011 to June 2012 were enrolled in the study. The inclusion criteria for exudative AMD patients were as follows: older than 60 years; first diagnosed with exudative AMD as confirmed with fluorescein angiography and optical coherence tomography; and no history of treatments such as intravitreal anti-VEGF injection, photodynamic therapy, or photocoagulation. The exclusion criteria included ocular diseases other than cataract and exudative AMD, a history of intraocular surgery other than cataract surgery, cataract surgery performed less than 6 months prior to enrollment, and any significant systemic disease other than hypertension and diabetes mellitus. Patients with any history of ocular inflammation including uveitis were also excluded. Undiluted aqueous humor samples (100–150 μL) were acquired through anterior chamber paracentesis immediately prior to the intravitreal anti-VEGF injection. Twenty age- and sex-matched controls were recruited during the same period—10 cataract patients and 10 idiopathic epiretinal membrane patients. None of the cataract patients presented abnormal retinal findings. Undiluted aqueous humor samples (100–150 μL) were obtained prior to surgery. All aqueous humor samples were immediately stored at −70°C until analysis. Informed consent was obtained from all participants, and the study was approved by the Institutional Review Board of Seoul National University Bundang Hospital and adhered to the tenets of the Declaration of Helsinki. 

Membrane-Based Human Antibody Array

In this study, we used Raybio Biotin Label-based Human Antibody Array I (Catalog No: AAH-BLM-1-2; RayBiotech, Inc.) to compare the differences in protein expression levels in the aqueous humor of exudative AMD patients versus controls. This antibody array membrane is marked with 507 specific antibodies (provided in the public domain by http://www.raybiotech.com/L-series-507-label-based-human-array-1-membrane-2.html) toward cytokines, chemokines, growth factors, angiogenic factors, adipokines, proteases, soluble adhesion molecules, soluble receptors, and other proteins in cell culture supernate and serum to provide the researcher with a broad, panoramic view of aqueous humor composition. The assay was performed according to the manufacturer's instructions. At first, the concentrations of 40 aqueous humor samples were measured using the bicinchoninic acid assay. Protein concentrations in the exudative AMD group ranged from 0.24 to 1.22 μg/μL; those of the control group ranged from 0.23 to 2.05 μg/μL. A volume of aqueous humor corresponding to 20 μg was separated from each sample. Aqueous humor samples from each group (400 μg per group) were mixed and dialyzed with 2 L PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 2 mM KH₂PO₄, pH 8.0) overnight at 4°C. After an internal control was added to the dialyzed samples, 3.5 μL of biotin-labeling reagent was added to the samples and incubated at room temperature (RT) for 30 minutes. The samples were diluted with 10 mL blocking buffer after free biotin was eliminated using a spin column. The array membranes were blocked with 8 mL of blocking buffer at RT for 30 minutes, then exposed to biotin-labeled samples overnight at 4°C. The membranes were cleaned thrice with 20 mL of wash buffers I and II at RT for 5 minutes with shaking and then incubated with 8 mL of 400-fold horseradish peroxidase–conjugated streptavidin overnight at 4°C. After the membranes were cleaned according to the same protocol, 4 mL of mixed detection buffer were added to the membranes so that the entire surfaced was covered. The entire system was then left at RT for 2 minutes. Finally, several films were acquired after exposing the membranes to radiographic film (Kodak Industrex Processor Model; Carestream Health, Inc., Rochester, NY). The two films that would be most suitable for comparing spot signal intensity were selected. 

Comparison of the Two Groups

Differences in spot signal intensity were evaluated meticulously with the naked eye. When an intergroup difference was identified, the densities were measured quantitatively using chemiluminescent image analysis (Bio-Rad Quantity 4.6.7; Bio-Rad Laboratories, Inc., Hercules, CA) after the two films were converted to computer files using commercial software (Bio-Rad ChemiDoc XRS Systems; Bio-Rad Laboratories, Inc.). 

Protein Quantification and Statistical Analysis

The antibody microarray results were validated with target molecules showing differences between AMD patients and controls as well as with VEGF, a growth factor known to be present in elevated levels in eyes with exudative AMD. The concentrations of IGFBP-2, IGF-1, and VEGF were measured in all 40 aqueous samples using commercially available ELISA kits (Quantikine Elisa kits, catalog number DGB200, DG100, and DVE00, respectively; R&D Systems, Minneapolis, MN). The concentration of each sample was measured twice using 2 μL of aqueous humor; these values were then averaged. All ELISAs were performed according to the manufacturer's protocols. An independent t-test was used to determine if there was a significant difference between groups. Commercially available software (SPSS for Windows, Ver. 18.0; SPSS, Inc., Chicago, IL) was used for the statistical analysis. A P value < 0.05 was considered significant. 

Results

The study population comprised 20 exudative AMD patients and 20 healthy controls. There was no significant difference in age or sex between the groups (Table 1). The rates of diabetic mellitus and hypertension were also similar in both groups (Table 1). 

Table 1

 

Comparison of Patients With Exudative AMD and Control Group

	Exudative AMD Group, n = 20	Control Group, n = 20	P Value
Age, y, mean ± SD	73.1 ± 7.9	72.7 ± 5.9	0.964*
Sex, M:F	7:13	7:13	1.000†
Diabetic mellitus, n			0.677†
Yes	4	3
No	16	17
Hypertension, n			0.490†
Yes	5	7
No	15	13

*   P value was calculated by independent t-test.

†   P value was calculated by χ² test.

Various cytokines, chemokines, growth factors, and receptors were detected in the aqueous humor samples from both groups: 164 molecules in the control group and 176 molecules in the exudative AMD group (Fig. 1). In the control group, IGFBP-related protein 1 (IGFBP-rp1/IGFBP-7, number 15 in Fig. 1), IL-20 receptor beta, and secreted protein acidic and rich in cysteine (SPARC) were detected at higher levels than observed in the positive control (number 1 in Fig. 1). In the exudative AMD group, IGFBP-rp1, lipocaline-2 (number 18 in Fig. 1), SPARC, and tissue inhibitors of metalloproteinase-1 (TIMP-1, number 20 in Fig. 1) were detected at higher densities than observed in the positive control. 

Figure 1

 

Expression levels of 507 proteins in the aqueous humor of patients with exudative AMD and the control group. The mixed aqueous humor samples in each group (20 exudative AMD patients, 20 age- and sex-matched controls) were assayed using the Raybio human antibody array kit. The molecules with densities that varied between groups are presented in order.

Twelve molecules were detected at higher levels in the exudative AMD group as compared to the control group. These included proliferation-induction ligand (APRIL), chordin-like 1, glucocorticoid induced tumor necrosis factor receptor family related gene (GITR) ligand, growth hormone receptor, IGFBP-2, IGFBP-6, IGFBP-rp1, IL-4 receptor, lipocalin-2, thrombospondin, TIMP-1, and TIMP-2 (Fig. 2). The density ratios between groups were 2.61, 9.92, 1.48, 4.84, 9.02, 4.25, 3.31, 3.44, 3.02, 8.67, 2.29, and 3.31, respectively. These ratios were measured using chemiluminescent image analysis and calibrated according to the density of the positive control in each group (Fig. 3). Eight molecules were detected at higher density in the healthy control group as compared with the exudative AMD group. These included comprised C-C motif chemokine receptor 7 (CCR7), CCR8, CCR9, C-X-C motif chemokine ligand 14 (CXCL14), C-X-C motif chemokine receptor 1 (CXCR1), CXCR2, eotaxin-3, and IL-1 receptor 4 (Fig. 2). These density ratios were 0.14, 0.12, 0.15, 0.04, 0.20, 0.12, 0.31, and 0.49 (Fig. 3). 

Figure 2

 

A list of the molecules with significantly different levels of expression in the aqueous humor of patients with exudative AMD versus control subjects. Twelve molecules exhibited increased expression in patients with exudative AMD; eight molecules were expressed at higher levels in the control group.

Figure 3

 

The ratios of 20 molecules expressed at different levels patients with exudative AMD versus controls. These values are expressed on a log scale. The ratio ranged from −4.64 (CXCL14/BRAK, equivalent to 0.04) to 3.31 (chordin-like 1, equivalent to 9.92).

We focused on the fact that IGFBP-2, IGFBP-6, and IGFBP-rp1 were increased in exudative AMD group. They are members of IGF family, consisting of IGF-1, 2, and various IGFBPs, which are known to regulate various cellular processes such as proliferation, differentiation, and neovascularization. ¹² There is little evidence that IGF family was involved in the pathophysiology of exudative AMD, although IFG-1, the core protein of IGF family, was suggested to induce the proliferation of choriocapillary endothelial cells in exudative AMD ¹³ and to be involved in the increase of VEGF secretion in retinal pigment epithelial cells. ¹⁴ IGFBP-2, showing the second-strongest association with AMD in our experiment, was known to be secreted by RPE cells ¹⁵ as well as to directly enhance VEGF gene promoter activity and subsequent angiogenesis. ¹⁶ On this background, we decided to validate IGF-1 and IGFBP-2. VEGF was also selected because of its renowned association with exudative AMD. ^17–20 As measured by ELISA, the mean concentration of IGFBP-2 was 7.47 ± 6.19 ng/mL in patients with exudative AMD versus 3.07 ± 3.34 ng/mL in the control group (P = 0.008). The mean concentrations in AMD patients and controls were 2.20 ± 0.26 vs. 1.99 ± 0.35 ng/mL for IGF-1 (P = 0.040) and 122.25 ± 63.24 vs. 86.98 ± 44.41 pg/mL for VEGF (P = 0.048; Fig. 4, Table 2). 

Figure 4

 

The concentrations of IGFBP-2, IGF-1, and VEGF as measured by ELISA. The concentrations of three molecules as measured using ELISA in patients with exudative AMD (n = 20) and controls (n = 20). The density of IGFBP-2 expression was higher in the exudative AMD group (P = 0.008, independent t-test). The density of IGF-1 was also measured as higher in the exudative AMD group (P = 0.040), although the ratio (1.11) was much smaller than that observed for IGFBP-1 (2.43). VEGF levels differed as well (P = 0.048). The absolute concentration of VEGF was much lower than that of IGF-1 or IGFBP-2.

Table 2. 

 

Aqueous Humor Concentrations of IGFBP-2, IGF-1, and VEGF as Measured by ELISA

	Exudative AMD Group, n = 20	Control Group, n = 20	Ratio of Exudative AMD/Control	P Value*
IGFBP-2, ng/mL	7.47 ± 6.19	3.07 ± 3.34	2.43	0.008
IGF-1, ng/mL	2.20 ± 0.26	1.99 ± 0.35	1.11	0.040
VEGF, pg/mL	122.25 ± 63.24	86.98 ± 44.41	1.42	0.048

 Results obtained from patients with AMD and controls (mean ± SD).

*   P value was calculated by independent t-test.

Discussion

Exudative AMD leads to neovascularization under the macula via a cascade of angiogenic and inflammatory responses after damage to the outer retinal cells and retinal pigment epithelium. VEGFs are known to be the most potent regulators of angiogenesis related to the pathogenesis of exudative AMD. Clinical studies demonstrated the value of intravitreal anti-VEGF injections in the treatment of exudative AMD. ^2,3 However, the curative value leaves substantial room for improvement: approximately 70% of patients with exudative AMD increased by less than 15 Early Treatment Diabetic Retinopathy Study letters from baseline visual acuity after repetitive intravitreal anti-VEGF agent injections. ²¹ Numerous studies have identified molecules other than VEGF that play an important role in the pathogenesis of exudative AMD. ^5–11 However, most studies limited their analysis to specific targeted cytokines. Recently, advanced technologies such as Multiple reaction monitoring (MRM) mass spectrometry have been used to investigate aqueous humor composition in patients with exudative AMD. ²² MRM allows for the rapid detection of hundreds of proteins without any specific antibodies, but the technique is expensive and the results are difficult to analyze. 

We used Raybio antibody-based protein array technology to screen the concentration changes of various molecules found in the aqueous humor in patients with exudative AMD. This technology can scrutinize 507 cytokines, chemokines, growth factors, and other molecules simultaneously. It is cheaper than MRM, and the results are easier to analyze. Few studies have used this approach to investigate protein levels in aqueous humor in patients with exudative AMD. ²³ We compared the concentrations of various molecules in aqueous humor samples from exudative AMD patients and controls. 

The levels of IGFBP-2, IGFBP-6, and IGFBP-rp1 were increased in patients with exudative AMD. These proteins are members of the IGF family, which includes IGF-1, -2, and various IGFBPs. IGFs are important to regulate various cellular processes such as proliferation, differentiation, apoptosis, and neovascularization. ¹² The actions of the IGFs are modulated by a group of high-affinity IGFBPs that includes proteins IGFBP-1 through IGFBP-6. The IGFBPs transport IGFs within the bloodstream and across intact capillary membranes, localize the IGFs to specific cell types and tissues, control IGF interaction with cell surface receptors, and modulate the subsequent physiological effects. ²⁴  

IGFs and IGFBPs are expressed in the eye and were thought to function via autocrine or paracrine pathways. ^25–27 IGF-1 is an angiogenic factor that induces the proliferation of retinal endothelial cells. ²⁸ When IGF-1 was injected intravitreally in the rabbit, the blood-retinal barrier was rapidly disrupted, leading to neovascularization of the optic disc. ²⁹ This proliferation often progressed to tractional retinal detachment, similar to the changes observed in proliferative diabetic retinopathy. ²⁹ The production of IGF is thought to be strongly stimulated by retinal ischemia. ³⁰ IGFBPs regulate IGF-1 action in the retinal endothelium ²⁸ ; vitreal IGFBP concentrations are elevated in ischemic versus nonischemic retina. ³⁰  

IGF is also involved in the pathophysiology of exudative AMD. IGF-1 directly induces the proliferation of bovine choriocapillary endothelial cells, the building blocks of choroidal neovascularization in exudative AMD. ¹³ In vitro studies have shown that IGF-1 also plays an indirect role by increasing VEGF secretion in retinal pigment epithelial cells. ¹⁴ However, the functions of IGFBPs in exudative AMD have remained largely unknown. Our results confirm that IGFBPs are involved in the pathogenesis of exudative AMD. IGFBP-2 is a 34-kDa glycoprotein that is generally known to inhibit the effects of IGF ^31–33 and is secreted by RPE cells. ¹⁵ In exudative AMD, IGFBP-2 would modulate choroidal neovascularization by inhibiting IGF, though it is difficult to explain the observed increase in IGFBP-2, but not IGF-1 levels in patients with exudative AMD. This discrepancy might be explained by the independent function of IGFBP-2 irrespective of the IGF axis. ³⁴ IGFBP-2 has been shown to enhance mitogenesis in uterine endometrial epithelial cells and osteosarcoma cells in the absence of IGFs ^35,36 and IGFBP-2 directly enhanced VEGF gene promoter activity and subsequent angiogenesis in a neuroblastoma cell line. ¹⁶ It may tentatively be concluded that, whereas IGFBP-2 inhibits the IGFs by sequestration in the context of the IGF system, IGFBP-2 plays an independent role in the pathophysiology of exudative AMD. 

IGFBP-rp1, a 36-kDa glycoprotein, also known as IGFBP-7 and mac25, is one of the low-affinity IGFBP-rps (IGFBP-rp1 through IGFBP-rp10). It is known to be involved in developmental processes and tumor growth. ³⁷ IGFBP-rp1 has also been shown to inhibit the stimulatory effect of VEGF on retinal endothelial cells in vitro. ³⁸ The elevation of IGFBP-rp1 levels in the aqueous humor of patients with exudative AMD indicates that increased levels of angiogenic factors such as VEGF and IGF-1 work in a complex interplay with antiangiogenic factors to modulate angiogenesis. The increased levels of thrombospondin, an antiangiogenic factor, observed in the AMD group study can be understood similarly. ³⁹  

The concentration of GITR ligand was also increased in aqueous humor samples from eyes with exudative AMD. The GITR ligand is mainly expressed in antigen-presenting cells and endothelial cells and activates GITR. The GITR/GITR ligand system is known to participate in inflammation. ⁴⁰ Kim et al. reported that GITR ligand was expressed constitutively in the RPE and at high levels in photoreceptor inner segments. Furthermore, the stimulation of proinflammatory cytokines upregulated the concentration of GITR ligand. ⁴¹ Inflammation is considered to play a role in the pathogenesis of AMD, although many studies have focused on dysfunctions of the complement system associated with genetic polymorphisms. ⁴² Further investigation of GITR/GITR ligand might widen our understanding of the role of inflammation in the pathogenesis of exudative AMD. 

VEGF levels in the exudative AMD and control groups differed substantially when measured using ELISA, but not when measured using Raybio technology. It is well known that VEGF levels are increased in the aqueous humor of patients with exudative AMD. Many authors have reported increased VEGF levels in the aqueous humor of eyes with exudative AMD. In addition, intravitreal injections of anti-VEGF agents decrease the expression of VEGF in aqueous humor. ^17–20 The inability of Raybio technology to detect this change in expression could be explained by any or all of the following reasons. First, VEGF concentrations in the aqueous humor are in the range of pg/mL, which is lower than the concentrations of other molecules with differing expression in AMD versus control samples. Thus, the low concentration of VEGF may have affected the discrimination power of the Raybio microarray test. Second, the pooling of aqueous humor samples might have had a more pronounced effect on the sensitivity of the Raybio technology than on the ELISA, which was performed on individual samples. Notably, because the volume of samples required for use of the Raybio technology was much larger than that required for the ELISA, the pooling of multiple aqueous humor samples was inevitable. Third, our interpretation of the results was limited, in that the ELISA was more accurate than the Raybio technology in terms of quantifying the concentration differences of various cytokines. For example, the difference in IGF-1 levels was detected by ELISA but not by Raybio technology, perhaps because the difference in IGF-1 levels (2.20 ± 0.26 vs. 1.99 ± 0.35 ng/mL, P = 0.040) was not as great in magnitude as the difference in IGFBP-2 levels (7.47 ± 6.19 vs. 3.07 ± 3.34 ng/mL, P = 0.008), although both differences were statistically meaningful. Nevertheless, it remains noteworthy that Raybio technology can be used to investigate numerous cytokines simultaneously, which facilitates the investigation of all pathways involved in the pathogenesis of exudative AMD. 

One limitation of our study is the fact that certain significant differences detected by Raybio technology were not confirmed by ELISA. Furthermore, molecules that are well known to be expressed at higher levels in patients with exudative AMD such as VEGF and MCP-1⁶ were not detected using Raybio technology. Nevertheless, Raybio technology is considered as a useful screening tool that can be used to discover molecules previously unknown to be involved in the pathogenesis of exudative AMD. The use of this technology here highlighted the importance of the IGF axis and other molecules previously considered to play no role in the pathophysiology of exudative AMD. 

In conclusion, we used the Raybio antibody array to show that various molecules, especially IGFBP-2, are present at high levels in the aqueous humor of eyes with exudative AMD. Further studies will be needed to clarify the role of IGFBP-2 in exudative AMD. 

Acknowledgments

Supported by a research grant of Seoul National University Bundang Hospital (02-2005-028); National Research Foundation of Korea (NRF) grants funded by the Ministry of Education, Science, and Technology (2009-0072603, 2012R1A1A2008943); and grants from the Korea Health Technology R&D Project, Ministry of Health & Welfare, Republic of Korea (Grant No. A111161). 

Disclosure: D.M. Cha, None; S.J. Woo, None; H.-J. Kim, None; C. Lee, None; K.H. Park, None 

Footnotes

 DMC and SJW contributed equally to the work presented here and should therefore be regarded as equivalent authors.