Chemokine Receptor Expression in Peripheral Blood Monocytes from Patients with Neovascular Age-Related Macular Degeneration

Abstract

Purpose.: Chemokine signaling and monocytes/macrophages were implicated in the pathogenesis of AMD. We tested the association between chemokines involved in monocyte recruitment and AMD.

Methods.: Immunophenotyping for white blood cell (WBC) populations including CD14++CD16- and CD14+CD16+ monocytes, CD19+, CD3+, and CD16+ lymphocytes, and chemokine receptors CCR1, CCR2, CCR5, CX₃CR1, and CXCR4 was performed on peripheral blood from treatment-naïve neovascular AMD (NV-AMD) patients and controls. The mRNA level of chemokine receptors in monocytes was measured with quantitative-PCR. Systemic levels of major chemokine ligands CCL2, CCL5, CCL3, and CXCL10 were evaluated by ELISA. Genotyping was performed for risk SNPs for AMD in the CFH, C3, and HTRA1 genes.

Results.: The percentage of WBC subpopulations tested was similar between NV-AMD patients (n = 18) and controls (n = 20). CD14+CD16+ monocyte subpopulation showed a 3.5-fold increased expression of CCR1 (P = 0.039; t-test) and a 2.2-fold increased expression of CCR2 (P = 0.027) in patients compared with controls. Increased CCR1 and CCR2 expression was correlated with each other in patients (R ² = 0.64, P < 0.0001), but not controls (R ² = 0.02, P = 0.57). Increased mRNA levels of CCR1 (1.6-fold, P = 0.037) and CCR2 (1.6-fold, P = 0.007) were found in monocytes from NV-AMD patients. Chemokine receptor expression was not correlated with the presence of risk SNPs, and was not associated with blood chemokine levels.

Conclusions.: CCR1 and CCR2 are coupregulated on the CD14+CD16+ monocyte population in NV-AMD patients. These data implicate CD14+CD16+ monocytes and chemokine signaling in AMD. Additional investigation is needed to elucidate the role of these monocytes and their potential as a biomarker or therapeutic target for AMD.

Introduction

The major role that inflammation plays in AMD has been recognized in recent years. ^1–9 In addition to complement activation, several lines of evidence implicate white blood cells (WBC), and in particular, macrophages, in the pathogenesis of the non-neovascular (NNV) and neovascular (NV) forms of AMD. Histological studies have identified macrophages in the proximity of drusen ¹⁰ and macrophages were reported to be present in higher amounts in NV-AMD eyes. ¹¹ Macrophages were also shown to modulate the development of choroidal neovascularization (CNV) in a rodent model of laser-induced CNV, ^12–16 and recently, we have reported a specific gene expression pattern in peripheral blood white blood cells from NV-AMD patients compared with controls. ¹⁷  

Chemokines released by injury and inflammatory sites signal to recruit cells that express chemokine-specific receptors. Certain chemokines have been specifically found on monocytes and macrophages, and some of these influence the classification of these two cell types. ^18,19 Mice with perturbed signaling of CCR2 and CX₃CR1, two of the major chemokine receptors involved in monocyte recruitment, demonstrate retinal alterations that recapitulate some of the features of AMD. ^20–23 Furthermore, studies have recently shown that several of the chemokine ligands, including CCL2 (the ligand for CCR2, also known as MCP-1, or monocyte chemoattractant protein 1), are present in higher intraocular concentrations in NV-AMD eyes ²⁴ and urine. ²⁵ High levels of CCL2 have been found in RPE in vivo, ²⁶ and this chemokine may be associated with CNV growth in mice. ²⁷ In addition, the monocyte population specifically driven by CCL2 has been found to cause apoptosis in the RPE. ²⁸  

Monocytes, the systemic precursor of tissue macrophages, have been divided into two major subpopulations: the classical, resting state CD14++CD16- monocytes, which compromise the majority of blood monocytes, and the CD14+CD16+ monocytes, which are the nonclassical, pro-inflammatory monocytes. ²⁹ The CD14+CD16+ monocyte subset have been linked to many different inflammatory disorders, including atherosclerosis, ³⁰ rheumatoid arthritis, ^31,32 idiopathic arthritis, ³³ lymphoma, ³⁴ liver fibrosis, ³⁵ cancer, ³⁶ and asthma. ^29,37 While previous studies have demonstrated recruitment of macrophages from the periphery to the eye, ostensibly from their precursor monocytes, ³⁸ it is unclear if and what monocyte subsets are involved in the pathogenesis of AMD. 

This research was performed to gain additional insight into the role that macrophages/monocytes and other WBC subpopulations may play in AMD. To that end, we have evaluated the number of cells belonging to major peripheral blood mononuclear cells (PBMC) subsets, their chemokine receptor expression pattern, and serum chemokine levels in NV-AMD patients and controls. 

Materials and Methods

Patients

Peripheral blood was drawn from a consecutive group of treatment- naïve NV-AMD (n = 18: 9 females, 9 males; median age: 78.5 years) and age and ethnic background-matched controls (n = 20: 10 females, 10 males; median age: 73 years). Patient demographic data are provided in Supplementary Tables 1 and 2 (see Supplementary Material and Supplementary Tables 1 and 2). All patients were treated in the retina service of the Hadassah–Hebrew University Medical Center. The criteria for inclusion in the study included age over 60 years and a lesion comprised of greater than 50% active CNV secondary to AMD, without any other cause for CNV (such as myopia, trauma, etc.), no subretinal hemorrhage larger than 25% of the lesion size, no other retinal disease, and no major systemic illness (such as cancer or current treatment for cancer, autoimmune disease, congestive heart failure, or uncontrolled diabetes). Controls were over 60 years of age, healthy, without known retinal diseases, or recent eye surgery, and with unremarkable ophthalmoscopy. All patients and controls signed an informed consent form, and the study was approved by the institutional ethics committee. Findings from the ophthalmic exam were collected, and optical coherence tomography (OCT) and fluorescein angiography (FA) images were reviewed. AMD was diagnosed according to the Age-Related Eye Disease Study (AREDs) criteria. ³⁹  

Immunophenotyping

Immunophenotyping was performed by flow cytometric analysis (FACS [Fluorescence Activated Cell Sorter]; BD Biosciences, San Jose, CA) on whole blood from patients and controls to identify CD14++CD16- monocytes, CD14+CD16+ monocytes, CD3+ T lymphocytes, CD19+ B lymphocytes, and CD16+ and CD16++ subsets of lymphocytes. In addition, the CCR1, CCR2, CCR5, CX₃CR1, and CXCR4 chemokine receptor expression was quantified using the following specific antibodies: anti-human CD14-APC; CD16-FITC; CD56-FITC; CD3-PerCP-Cy5.5; CD56-PE; and CD19-APC (eBioscience, San Diego, CA), along with anti-human CCR1-PE; CCR2-PE; CCR5-PE (R&D Systems, Minneapolis, MN); CX₃CR1-PE (MBL International Corp, Woburn, MA); and CXCR4-PE (eBioscience). The signal was normalized to isotype-specific controls and used to eliminate nonspecific binding (eBioscience). FACS was performed as previously described. ^40,41  

Briefly, blood was collected in vaccutainer tubes containing EDTA as anticoagulant (BD Biosciences) and processed immediately for immunophenotyping; 100 μL of blood were dispensed per tube, washed two times with 2 mL of PBS/0.5% BSA, and centrifuged to recover cells. Flow cytometry tubes (Falcon; BD Biosciences, Bedford, MA) were incubated with antibodies and similar dilutions of isotype controls, to determine the amount of nonspecific binding according to manufacturers' instructions for 30 minutes in the dark on ice. Dilutions of 0.25 μg of anti-CD14-APC; 0.03 μg anti-CD16-FITC; 0.03 μg of anti-CD19-APC; 0.125 μg of CD3-PerCP-Cy5.5; 1 μg of anti-CD56-PE; 0.25 μg of anti-CCR1-PE; 0.125 μg of anti-CCR2-PE; 0.375 μg of anti-CCR5-PE; 0.1 μg of anti-CX₃CR1-PE; and 0.03 μg of CXCR4-per every 100 μL of blood tested were used. Red blood cells were then lysed in a 2-mL flow cytometry lysing solution (FACS; BD Biosciences) for 5 minutes. The remaining cells were washed and filtered through a 60-μm mesh and fluorescence intensities were read on the flow cytometer (LSR II FACS; BD Biosciences). The flow cytometer reading (BD Biosciences) was terminated after a total of 3000 CD14+ monocyte events had been read per each sample for accurate viewing. The total number of cells, percentage of each different cell type, and expression of each of the five chemokine receptors was determined using the cytometry data analysis software and gating technology (FCS Express; De Novo Software, Los Angeles, CA). 

Gating was performed according to standard protocols (BD Biosciences, 2009), including the isolation of the two different subsets of monocytes. ^40–44 The two monocyte subsets, the CD14++CD16- classical subset and the CD14+CD16+ nonclassical subset, were viewed on a CD14/CD16 plot, with a CD56-PE fluorescent staining to remove possible natural killer (NK) cell contamination. After viewing the monocyte and lymphocyte populations on the plot, the vertical line for gating was set based on the isotype controls for the various fluorophores. The horizontal line was set based on the CD56 positive, CD14/DC16 double-negative staining, as well as the isotype control, to differentiate the two monocyte populations via CD16 expression. Thus, any gate placed below the horizontal line would be CD16-negative, thereby, facilitating distinction between CD16 positive and negative monocytes. The monocyte and lymphocyte populations were then viewed again on a forward scatter/side scatter (FSC/SSC) plot to clarify the correct location. Backgating was performed around the “monolymph” location to remove neutrophil influence (Fig. 1). 

Figure 1. 

 

FSS/SSC plots for viewing of the monocyte and lymphocyte population as separate from the granulocyte population. All cell types were placed under the control of the “monolymph” gate as shown (please see “Materials and Methods” section for a detailed description). Unstained, monolymph, and monocyte populations are viewable as labeled.

Following this validation showing appropriate monocyte subset separation, the gates were applied to separate out the CD14++CD16- and the CD14+CD16+ monocytes based on expression of CD16 (Fig. 2). Monocytes that showed CD14++ versus lymphocytes that showed CD16++ were gated separately. For further validation of the purity of the monocyte subsets, a plot was created with CD56 staining against CD14 staining, to view the possible overlap of NK cells against the gating of both subsets of monocytes. The average ± SEM percent of such overlap was 1.79 ± 0.91, and it was similar in patients and controls (P = 0.42, Student's t-test; Fig. 3). Similar exclusion plots also verified that CD3+ T cells and CD19+ B cells did not contaminate the monocyte population. 

Figure 2. 

 

Representative FACS gating plot of white blood cell subtypes used to separate the monocyte subsets. CD14++CD16- and CD14+CD16+ monocytes can be viewed according to the CD14-APC and CD16-FITC expression. CD16+ and CD16++ cells can also be viewed, and were gated according to CD14 and CD16 expression levels as described in the methods section. CD56-PE was used to exclude NK cells on a CD14/CD16 plot. The isotype control was used to set the horizontal and vertical bars as to split the two monocyte subsets and determine the CD16++ lymphocyte population. Populations are identified by color in the legend: Light green population: CD14++CD16- monocyte subset. Dark green population: CD14+CD16+ monocyte subset. Light blue population: CD16++ lymphocytes. Brown population: CD16+ lymphocytes. Red population: CD56+ NK cells, and nonspecific binding.

Figure 3. 

 

Assessment of possible NK cell overlap through gating on a CD14-APC and CD56-PE plot. Horizontal and vertical lines were set for both (A) CD14++CD16- monocytes and (B) CD14+CD16+ monocytes. (C) Isotype control demonstrates specificity of the antibodies and allowed for setting the appropriate vertical and horizontal plot lines. (D) CD19+ B cell and CD3+ T cell exclusion through the same method is also shown.

QPCR

Monocytes were isolated from whole blood, from treatment-naïve NV-AMD patients (n = 13, 7 of which also underwent FACS immunophenotyping), and controls (n = 8, 4 of which also underwent FACS immunophenotyping), using a negative selection method as described. ⁴⁵ Briefly, PBMCs were separated from blood on a gradient (Histopaque-Ficoll; Sigma-Aldrich, Munich, Germany), washed twice at 1500 rpm for 10 minutes to remove platelets, and live cells were counted with a hemocytometer (MSUM Biochem and Biotech, Moorhead, MN) using the Trypan Blue exclusion method. Total blood monocytes, including both subsets of CD14++CD16- and CD14+CD16+ monocytes were isolated using a negative selection kit (EasySep, Stemcell Technologies, Vancouver, Canada) according to manufacturer's instructions. Flow cytometric analysis of the isolated cells validated that the entire monocyte population of blood was represented. We normally obtained 5 × 10⁵ monocytes/mL blood from both NV-AMD patients and controls, which is in accordance with the normal range. RNA was extracted from the isolated monocytes using RNA isolation reagent (TriReagent; Sigma-Aldrich) according to the manufacturer's protocol. Possible remnants of DNA were degraded (DNA-free; Ambion, Austin, TX), the RNA was purified (RNAeasy MinElute Cleanup Kit; QIAGEN, Hilden, Germany), and samples were stored at −80°C until further use. RNA concentration and quality was measured using both a spectrophotometer (NanoDrop; Thermo Scientific, Waltham, MA), and a micro-fluidics based platform and data analyzer (Bioanalyzer; Agilent, Santa Clara, CA), respectively, to ascertain purity and concentration. Samples with an RNA integrity number of 8 or higher were included. 

Reverse transcription of 1 μg RNA to cDNA was performed using a kit (High Capacity Reverse Transcription kit; Applied Biosystems, Carlsbad, CA) according to manufacturer's protocol. TaqMan gene expression assay for the genes CCR1 (Assay ID #Hs000174298_m1); CCR2 (Hs01560352_m1); CCR5 (Hs00152917); CX₃CR1 (Hs01922583_s1); and CXCR4 (Hs00607978_s1), along with the endogenous control genes HPRT1 (Hs99999909_m1) and RPLP0 (HS99999902; Applied Biosystems) were used for quantitative real time PCR (QPCR). Both control genes were previously determined to be appropriate for PBMC and monocyte RNA. ⁴⁶  

Real Time PCR was performed as we have previously described. ¹⁷ Briefly, triplicate reactions containing master mix (TaqMan Master Mix with Fast Enzyme; Applied Biosystems, Inc., Foster City, CA) and appropriate primers (TaqMan Gene Expression Assay; Applied Biosystems, Inc.), were created on 96 well plates, measured via the real-time PCR system (7900 HT FastReal Time System; Applied Biosystems, Inc.), analyzed using spreadsheet software (Excel; Microsoft, Redmond, WA), multiplate calibration software (RQ Manager; Applied Biosystems, Inc.), and a data analysis tool (DataAssist Software; Applied Biosystems, Inc.). ⁴⁷ Expression levels for each gene was calculated according to the 2-Delta Delta C method. ^48–50 The data analysis software (Applied Biosystems, Inc.) used a normalization technique by geometric averaging of multiple endogenous genes. ⁵¹  

Genotyping

DNA was extracted from 300 μL whole blood using a DNA kit (FlexiGene DNA Kit; QIAGEN). Genotyping was performed for the three major risk SNPs for AMD: CFH (rs1061170); HTRA1 (rs11200638); and C3 (rs2231099), using specific predesigned primers, ^52–54 and PCR reaction mixture (ReadyMix PCR; Sigma Aldrich) as we have previously described. ^53,54 PCR products were evaluated on a 1.5% agarose gel to confirm the success of the PCR reaction, using an nucleic acid staining solution (edSafe; INtRON Biotechnology, Kyunggi-do, Korea), followed by automatic sequencing (Macrogen, South Korea). 

ELISA

ELISA was performed on serum to determine the amounts of CCL2 (MCP-1; the major ligand of CCR2); CCL5 (RANTES; a ligand for CCR1, CCR3, and CCR5); CCL3 (MIP-1α; a ligand for CCR1 and CCR5); and CXCL10 (IP-10; a ligand for CXCR3A-B, often upregulated in chronic viral infection ^55,56 ). For CCL2, serum was extracted from blood of both treatment-naïve patients and controls (NV-AMD, n = 30; controls, n = 27). Serum was diluted 1:1 with PBS. A kit was used for CCR2 measurements (eBioscience), according to the manufacturer's instructions, and read on the spectrophotometer (FluoStar; BMG LABTECH GmBH, Ortenberg, Germany) within a half hour of completion. 

For CCL5, CCL3, and CXCL10, the analyte detection system (FlowCytomix MultiAnalyte Detection System; eBioscience) was used, which allows detection of multiple analytes via flow cytometry. Serum was extracted from blood of both patients and controls (NV-AMD, n = 20; controls, n = 20), and the reaction was performed according to the manufacturer's instructions. Serum was diluted 1:1 with PBS. Reading was performed within 12 hours of completion, and measurements were calculated by using flow cytometer software (FlowCytomix Pro 2.4 Software; eBioscience). 

Statistics

All calculations were performed using biostatistics software (InStat; GraphPad, San Diego, CA). Student's t-test was utilized to compare continuous variables between groups while correlation coefficient and ANOVA was calculated to assess correlation of chemokine receptors expression levels with variable factors. Wilcoxon matched-pairs signed ranks test and Spearman's rank correlation were used to compare chemokine expression levels between the groups while accounting for potential age-associated effects. 

Results

Immunophenotyping

Average percentage of the different cell types measured—including CD14++CD16- monocytes, CD14+CD16+ monocytes, CD16++ lymphocytes, CD16+ lymphocytes, CD19+ B cells, and CD3+ T cells—were not different between patients and controls (Fig. 4). Each cell type was then examined individually for the expression of the five different chemokine receptors: CCR1, CCR2, CCR5, CX₃CR1, and CXCR4, using a differently labeled fluorophore (Fig. 5). 

Figure 4. 

 

Average percentage of cell types measured by FACS analysis of PBMCs from NV-AMD patients (n = 18) and controls (n = 20). CD14++ and CD14+CD16+ monocytes, along with CD16++, CD16+, CD19+, and CD3+ lymphocytes were measured. Dark bars: NV-AMD patients. Light bars: Controls. P > 0.05 for each comparison.

Figure 5. 

 

Representative plots for identification of the two monocyte subsets and chemokine receptors. CD14++CD16- monocyte subset shown on CD14-APC plot with double staining using a PE–conjugated antibody for (A) CCR1, (B) CCR2, and (C) CCR5. Upper-left quadrant are cells positive for CD14 only, while lower-right quadrant indicates cells positive for the chemokine receptor only. The upper-right quadrant indicates percent positive CD14++CD16- cells for both CD14 and the respective chemokine receptor (CCR1, CCR2, or CCR5). CD14+CD16+ monocyte subset is also shown on a CD14-APC plot with (D) CCR1, (E) CCR2, and (F) CCR5. (G) Isotype control is shown for both APC and PE fluorophores, which was used to generate the horizontal and vertical plot lines to validate positive double staining, and to ascertain non-specific binding areas, shown in bottom-left quadrant.

No significant difference in expression of the five chemokine receptors was found between CD16++ cells, CD16+ cells, CD19+ cells, CD3+ cells, and the CD14++CD16- monocyte subset of patients and controls. However, when expression of the chemokine receptors was analyzed on the CD14+CD16+ subset of monocytes, a significant difference between AMD patients and controls was found (Fig. 3). Both CCR1 (3.5-fold, P = 0.039, t-test) and CCR2 (2.2-fold, P = 0.027, t-test) were significantly elevated on CD14+CD16+ cells in patients compared with controls (Fig. 6). To exclude age as a potential factor for bias in the comparison of receptor expression between patients and controls, receptor expression was also evaluated using a nonparametric matched-pairs statistical analysis comparing 18 NV-AMD patients with age-matched controls (±8 years). The results validated t-test findings (CCR1: P = 0.009; CCR2: P = 0.039, Wilcoxon matched-pairs signed-ranks test), indicating that age is not a significant factor in CCR1/2 expression on CD14+CD16+ monocytes in this study (see Supplementary Material and Supplementary Table 3). In addition, there was no correlation between CCR1 or CCR2 expression and age in either AMD patients or controls (R ² < 0.2; P > 0.05 Spearman's rank correlation). 

Figure 6. 

 

CCR receptor expression on CD14+CD16+ nonclassical monocytes from NV-AMD patients (n = 18) and controls (n = 20). Receptor expression was measured by FACS Dark bars: NV-AMD patients. Light bars: Controls. * P < 0.05 (t-test).

A direct correlation between the upregulation of CCR1 and the upregulation of CCR2 was found in AMD patients (R ² = 0.64; P < 0.0001), but this direct correlation was not found when compared with the matched controls (R ² = 0.02, P = 0.57). There was a nonsignificant trend toward upregulation of CCR5 in patients versus controls on the CD14+CD16+ monocytes (1.8-fold, P = 0.15, t-test). 

Levels of mRNA for Chemokine Receptors

mRNA levels for all five chemokine receptors (CCR1, CCR2, CCR5, CX₃CR1, and CXCR4) were measured in isolated monocytes (including both CD14+CD16- and CD14+CD16+ populations) using QPCR. Increased mRNA expression levels of CCR1 (1.6-fold, P = 0.037, t-test), and CCR2 (1.6-fold, P = 0.007, t-test) were found in AMD patients as opposed to controls. No correlation was found between age and chemokine receptor mRNA expression (P > 0.05, Spearman's rank correlation). The possible bias from age on chemokine expression was also tested using Wilcoxon's matched-pairs signed-ranks test, which compared eight age-matched NV-AMD patients with eight controls (±7 years; 5 AMD patients were omitted from this analysis to allow pair comparison (see Supplementary Material and Supplementary Table 4). This analysis confirmed increased mRNA levels of CCR2 (P = 0.05), and a trend toward increased CCR1 (P = 0.07) expression in AMD patients (Fig. 7). 

Figure 7. 

 

Normalized QPCR values for each chemokine receptor tested on the total blood monocyte population from NV-AMD patients (n = 14) and controls (n = 9). Dark bars: NV-AMD patients, Light bars: Controls.

Genotyping for Risk SNPs for AMD

Genotyping for major known risk SNPs for AMD in the HTRA1, C3, and CFH genes was performed among patients and controls to assess if these SNPs underlie differential expression of chemokine receptors in the monocyte subsets. A 3-way ANOVA and t-test were performed to view the correlation between SNP genotype and chemokine receptor expression. Results showed no correlation between any of the risk SNPs and the expression levels of the chemokine receptors tested, via both the univariate and multivariate analysis. 

Serum Levels of Chemokines

To assess if chemokine receptor expression in monocytes correlates with serum levels of their ligands, we measured levels of CCL2 (MCP-1) by ELISA for both patients and controls. Average (±SEM) serum levels of CCL2 were similar in patients (286.1 ± 20.7 pg/mL) and controls (286.7 ± 51.4 pg/mL; P = 0.37, Student's t-test). To exclude bias introduced by age, CCL2 levels were also analyzed using the Wilcoxon matched-pairs signed-ranks test, by pairing 24 NV-AMD patients with age-matched controls (±6 years, P = 0.41). The levels of CCL5 (RANTES); CCL3 (MIP1-α); and CXCL10 (IP-10) were measured via an analyte detection tool (eBioscience), for both patients and controls. No significant difference was found between serum levels of any of the proteins tested in patients (CCL5: Average ± SEM = 2251.6 ± 254 pg/mL; CCL3 = 254.7 ± 150.9 pg/mL; CXCL10 = 231.7 ± 18.9) and controls (CCL5 = 2529.5 ± 207.2 pg/mL, P = 0.40; CCL3 = 295.1 ± 183.1 pg/mL, P = 0.87; CXCL10 = 229.95 ± 49.7 pg/mL, P = 0.97, t-test). To remove the possible bias of age in our study, CCL5, CCL3, and CXCL10 levels were analyzed using the Wilcoxon matched-pairs signed-ranks test, by pairing NV-AMD patients (CCL5: n = 18, CCL3: n = 16, CXCL10: n = 18) with age-matched controls (±6 years, CCL5: P = 0.60, CCL3: P = 0.90, CXCL10: P = 0.28; see Supplementary Material and Supplementary Tables 5 and 6). Presence of age-associated chemokine levels was also evaluated by assessing the correlation between chemokine levels and age according to the ELISA data. There was no correlation found between expression of any ligand tested and age (P > 0.05, Spearman's rank correlation). 

Clinical Findings

Chemokine receptor levels on the different cell subsets were correlated with clinical parameters and demographics. No correlation was found with any clinical parameter which was evaluated, such as initial and final visual acuity, number of intravitreal anti-VEGF injections administered, lesion thickness as determined by OCT, lesion type, and size as determined by FA, and the change in lesion thickness following anti-VEGF treatment (data not shown). There was a nonsignificant trend for correlation between CXCR4 levels on CD14+CD16+ monocytes and the number of bevacizumab injections which was provided (R ² = 0.2, P = 0.058). 

Discussion

This study shows that CCR1 and CCR2 are coupregulated on the CD14+CD16+ monocyte subset of cells in the peripheral blood of NV-AMD patients. This chemokine receptor upregulation is not associated with or impeded by the white blood cell sub-population count, the major risk SNPs for AMD, or the serum CCL2, CCL5, CCL3, or CXCL10 levels. Another chemokine receptor, CCR5, also showed a trend toward increased expression on CD14+CD16+ monocytes from NV-AMD patients. 

Previous studies suggested that macrophages are involved in AMD and that they may have a variable and even contradictory role in the disease. While perturbation of CCR2 signaling was shown to lead to accumulation of debris or inflammatory cells in the retina, ^20,57,58 macrophage depletion can either enhance or suppress the development of CNV. ^13,59,60 This contradicting role of macrophages may be explained by the polarization of macrophages into specialized subsets having diverse effects. ^18,19,61,62  

In contrast to the other inflammatory disorders in which the CD14+CD16+ subset of activated circulating monocytes, a precursor of tissue macrophages, plays a role, ⁶³ limited data exist with respect to the involvement of monocytes in AMD. CD14+CD16+ monocytes have been specifically classified as pro-inflammatory, and are characterized by high production of TNF-α, ⁶⁴ and very low expression of anti-inflammatory cytokine IL-10. ⁶⁵ Furthermore, monocyte-derived macrophages may contribute to angiogenesis in NV-AMD, depending on the growth factors and chemokines present, as well as the environment. ^66–68 It was shown that hypoxia can induce a monocyte-to-macrophage transition, thus contributing to a pro-angiogenic environment, including expression of VEGF. ⁶⁹ Shantsila and colleagues recently showed that the CD14+CD16+CCR2+ subset of the CD14+CD16+ monocytes that represent a small but potent part of that population are pro-angiogenic, secreting large amounts of factors such as VEGF. ⁶¹ These monocytes driven by CCL2 may also cause apoptosis in the RPE. ²⁸  

Unlike our findings, Mo and colleagues, ⁷⁰ previously reported increased sera levels of CXCL10 (IP-10) in NV-AMD patients compared with controls. Both technical and biological factors may underlie these conflicting data. For example, the mean age of the control group was higher in our study compared with Mo's report. IP-10 expression increases with age, ⁷¹ thus, the smaller age difference between the AMD and control groups in our study may underlie the conflicting data between our study and that of Mo and colleagues. Furthermore, all AMD patients in our study were treatment naïve while Mo and colleagues may have included treated patients. Potentially, treatment may interact with chemokine levels as well. Larger group studies may be required to completely ascertain this variable. 

The CD14+CD16+ monocytes are recruited to sites of inflammation typically by CX₃CR1/L1 and CCR2/L2. ^61,72 Previous studies have showed recruitment of inflammatory cells to the retina from the periphery, ¹¹ and the presence of a retinal phenotype in mice with perturb CCR2 and CX3CR1 expression. ²² Thus, it is conceivable that chemokine receptors, including CCR1 and CCR2, and potentially CCR5 (which shares ligands with CCR1 and CCR2 ⁷³ ) and CX3CR1, are involved in the monocyte recruitment process to the retina in the context of AMD. While our finding of increased prevalence of CD14+CD16+ monocytes expressing CCR1 and CCR2 in NV-AMD can be an epiphenomena, it is also possible that such pro-angiogenic and pro-inflammatory monocytes are involved in the progression of the disease. 

CCR1 is present on both monocytes ⁷⁴ and lymphocytes, and its major ligands are MIP-1α and RANTES, which also bind to CCR5. ⁷⁵ The rd mouse strain that exhibits retinal degeneration at an early age shows high CCR1 expression. It has been hypothesized that CCR1-mediated cellular responses lead to apoptosis of photoreceptors. ⁷⁶ CCR1 has also been associated with multiple sclerosis and organ transplant rejection, usually caused by inflammatory recruitment of leukocytes. ^77–79 Therefore, while increased CCR1 levels in monocytes from AMD patients may reflect monocyte inflammation and recruitment in the disease; this molecule may also be involved directly in the pathogenesis of AMD. 

The involvement of the systemic innate immune response in AMD has been suggested by the identification of risk SNPs in complement genes as well as by evidence for complement activation in serum from AMD patients. ^{1–3,52,80,81} Our data implicate another component of the systemic innate immune system, blood monocytes, in NV-AMD. Thus, the involvement of both cellular and humoral innate immune responses in AMD is reflected systemically. Further research is required to provide a comprehensive understanding of the roles of chemokine receptors and their ligands in AMD, as well as their carrier monocytes. It should also be investigated whether these cells and molecules may represent a new therapeutic target or biomarker for AMD. 

Supplementary Materials

pdf S1 

Acknowledgments

The authors thank Eitan Fibach (Department of Hematology–Hadassah-Hebrew University Medical Center) for his astute and valuable help with our FACS experiments and analysis. 

Footnotes

 Supported by grants from the Israel Science Foundation (ISF), the Israeli Ministry of Health (IC), and the Israeli Ministry of Immigrant Absorption for returning Scientists (TB-C).

Footnotes

 Disclosure: M. Grunin, None; T. Burstyn-Cohen, None; S. Hagbi-Levi, None; A. Peled, None; I. Chowers, None