Vitreous and serum levels of vascular endothelial growth factor and platelet-derived growth factor and their correlation in patients with non-proliferative diabetic retinopathy and clinically significant macula oedema

Authors


Anna Praidou
2nd Department of Ophthalmology
Papageorgiou General Hospital
Aristotle University of Thessaloniki
Thessaloniki 56403
Greece
Tel: + 30 697 261 3156
Fax: + 30 231 069 0417
Email: praidou2003@yahoo.co.uk

Abstract.

Purpose:  To investigate possible correlations between vitreous and/or serum levels of platelet-derived growth factor isoforms (PDGF-AA, -AB and -BB) with parameters associated with non-proliferative diabetic retinopathy (NPDR) and clinically significant macula oedema (CSMO); to compare the results to relevant results regarding vascular endothelial growth factor (VEGF), which is an established growth factor affecting NPDR.

Methods:  Fifteen patients with NPDR, 31 patients with proliferative diabetic retinopathy (PDR) and 15 non-diabetic patients were included in the study. Vitreous and serum samples were obtained during vitrectomy. PDGF-AA, -AB and -BB, as well as VEGF, were measured by enzyme-linked immunosorbent assay (ELISA).

Results:  PDGF-AA, -AB and -BB and VEGF were all expressed in the serum and vitreous of controls and patients with NPDR. The levels of PDGF-AA, PDGF-AB and VEGF in vitreous were significantly increased in the NPDR group compared to controls, while PDGF-BB levels were significantly decreased in the NPDR group compared to controls. The levels of all PDGF isoforms and VEGF in vitreous were significantly increased in the PDR group compared to the NPDR group. No such differences were evident in serum. PDGF-AA and VEGF correlated significantly to the severity of NPDR. PDGF or VEGF in vitreous of NPDR patients did not correlate with retinal photocoagulation (RP) or the serum levels of haemoglobin A1c (HbA1c). There was no correlation between the vitreous and serum levels of VEGF or PDGF in patients with PDR. Only PDGF-AB vitreous levels correlated significantly with PDGF-BB vitreous levels in the NPDR group.

Conclusion:  It appears that in addition to VEGF, almost all PDGF isoforms in the vitreous are also correlated with NPDR and CSMO.

Introduction

Diabetic macular oedema (DMO) is the most frequent cause of severe vision impairment in patients with non-proliferative diabetic retinopathy (NPDR). It can be hypothesized that DMO is related to the breakdown of the blood–retinal barrier (BRB) and subsequent leakage of intraretinal fluid from abnormal retinal capillaries and microaneurysms, the secretion of vasopermeability factors from the retina into the vitreous (Antchliff & Marshall 1999) and vitreoretinal adhesions and traction on the macula (Lewis 2001).

Recently, retinal hypoxia has been implicated in the pathogenesis of DMO. Hypoxia causes increased expression of vascular endothelial growth factor (VEGF), which is a potent inducer of vascular permeability that has been shown to cause leakage from retinal vessels (Ferrara 2004;Frank 2006). Therefore, it is reasonable to hypothesize that VEGF contributes to DMO. Several monoclonal antibodies against VEGF have been introduced recently in ocular pathology (Cunningham et al. 2005; Haritoglou et al. 2006; Nguyen et al. 2006; Soliman et al. 2008).

VEGF has significant homology to platelet-derived growth factor (PDGF) (Ferrara 2004;Young & Miller 2006). PDGF, originally isolated from platelets, is one of the most ubiquitous growth factors that stimulates cellular proliferation and directs cellular movement. Two different PDGF chains exist, designated as PDGF A and PDGF B, giving rise to three PDGF isoforms: PDGF-AA, -BB or -AB (Freyberger et al. 2000). PDGF isoforms have been associated with the stimulation of collagen synthesis (Cassidy et al. 1998), and have been implicated in the pathogenesis of proliferative vitreoretinal disorders (Cassidy et al. 1998). It is an autocrine stimulator of growth in retinal pigment epithelial cells, participating also in retinal wound repair and epiretinal membrane formation (Cassidy et al. 1998). Increased PDGF levels in the vitreous of patients with proliferative diabetic retinopathy (PDR) have been reported previously (Endo et al. 2000; Freyberger et al. 2000; Praidou et al. 2009), but there are no published data regarding the correlation of vitreous and serum levels of PDGF isoforms in NPDR.

The purpose of the present study was to identify possible correlations between vitreous and/or serum levels of PDGF isoforms with various parameters associated with NPDR [i.e. levels of haemoglobin A1c (HbA1c) in serum, prior grid laser photocoagulation] and to compare these to relevant results regarding VEGF.

Materials and Methods

Patients

The study was conducted between 2005 and 2007 in accordance with the tenets of the Declaration of Helsinki. Approval of the Institutional Review Board Ethics Committee of the Medical School of Aristotle University of Thessaloniki was obtained. All patients signed an informed consent after the purpose of the study was explained in detail to each participant.

We categorized our patient population into three district groups. For group A (NPDR), inclusion criteria were one of the following: (i) patients with diabetes mellitus who were candidates for vitreoretinal surgery for clinically significant macular oedema (CSMO); (ii) previously unsuccessful treatment with grid laser photocoagulation; (iii) visual acuity of 20/60–20/200; (iv) no evidence of PDR on fundus fluorescein angiography (FFA) examination; or (v) no apparent vitreoretinal traction in the macula [confirmed at optical coherence tomography (OCT) examination]. Group B (PDR) consisted of patients who underwent pars plana vitrectomy (PPV) for long-standing (> 3 months) or recurrent vitreous or preretinal haemorrhage or tractional retinal detachment (TRD) involving or threatening the macula.

The control group (group C) consisted of patients without diabetes who underwent PPV for an idiopathic macular hole.

Exclusion criteria for group A included a thick and taught posterior hyaloid with evidence of macular traction, macular ischaemia [as defined by an enlarged foveolar avascular zone (FAZ) > 1000 μm] or significant perifoveal capillary loss on FFA (Patel et al. 2006). Exclusion criteria for group B included previous ocular surgery within the past 2 years, present or past history of ocular inflammation, rubeosis iridis or neovascular glaucoma and rhegmatogenous retinal detachment (RRD) (Funatsu et al. 2004). Vitreous samples were obtained from all patients included in the study.

Biomicroscopic findings: diabetic retinopathy assessment

All patients were evaluated by careful biomicroscopic examination using a 90-D fundus non-contact lens (Volk Super Field NC, Mentor, Ohio, USA). Fundus findings were confirmed preoperatively by standardized fundus colour photography and FFA, which was performed using a Topcon TRC-50 EX fundus camera with an IMAGENET system (Tokyo Optical Co. Ltd, Tokyo, Japan).

The severity of diabetic retinopathy was graded according to the modified Early Treatment Diabetic Retinopathy Study (ETDRS) system (Early Treatment Diabetic Retinopathy Study Research Group 1991a, 1991b). All patients with CSMO underwent grid photocoagulation at least 3 months before the surgical intervention (Funatsu et al. 2006). The retinal thickness at the central fovea was measured using OCT (OCT-E, Carl Zeiss, Meditec Dublin, California, USA) and was defined as the distance between the inner retinal surface and the retinal pigment epithelium.

Sample collection

At the beginning of the PPV procedure, 0.7–1.0 ml of undiluted vitreous sample was obtained by aspiration into a 2.5 ml sterile syringe attached to the vitreous cutter (Accurus Ophthalmic Surgical System; Alcon Surgical Laboratory Inc., Irvine, California, USA) with the stopcock of the infusion closed. The samples were transferred in sterilized Corning micro centrifuge tubes (1.5 ml), placed immediately on ice and centrifuged for 5 min at 4 °C at 10 000 rpm. The absence of platelets in supernatants was confirmed by manual phase contrast microscopy. Supernatants without sediment were divided into aliquots and frozen immediately at −80 °C until assayed. Blood samples were collected in sterilized tubes from all patients in the morning prior to scheduled surgery, and were left at room temperature until clotted. Serum was separated by centrifugation at 4 °C for 16 min at 1600 rpm, divided in aliquots and stored at −80 °C until assayed.

Measurement of PDGF and VEGF levels

PDGF isoforms and VEGF were measured in the vitreous and serum samples by enzyme-linked immunosorbent assay (ELISA) using the Quantikine human PDGF-AΒ assay kit (Quantikine human PDGF-AB; R&D Systems, Minneapolis, Minnesota, USA) (Vieweg et al. 1999; Weibric et al. 2002), Quantikine human PDGF-ΒΒ assay kit, Quantikine PDGF-AA human assay kit and Quantikine human VEGF assay kit (all R&D Systems). The ELISA for VEGF was able to detect two of the four VEGF isoforms – VEGF121 and VEGF165– because these two shorter isoforms are secreted, while the two longer isoforms are cell-associated. Each assay was performed in triplicate, according to the manufacturer’s instructions, using 100-μl aliquots of vitreous or serum samples, diluted accordingly to comply with the detection range of the relevant assay, and eventually to contain the same amount of protein. Data presented are per mg of protein. The optical density was determined at 450 nm and 570 nm using an absorption spectrophotometer (Stat Fax-2100; Awareness Technology Inc., Palm City, Florida, USA).

The standard solution (100 μl for VEGF ELISA, 150 μl for PDGF-AA ELISA, 100 μl for PDGF-AΒ ELISA, 100 μl for PDGF-ΒΒ ELISA) and the sample (100 μl for VEGF ELISA, 75 μl for PDGF-AA ELISA, 100 μl for PDGF-AΒ ELISA, 100 μl for PDGF-ΒΒ ELISA) were added to the wells of a 96-well plate coated with a monoclonal antibody. After incubation, the plate was washed and an enzyme-labelled antibody was added. After further incubation, the plate was washed again and the substrate was added. The reaction was arrested by adding the stop solution after colour had developed, and the optical density was determined at 450 nm and 570 nm using an absorption spectrophotometer (Stat Fax-2100; Awareness Technology Inc.).

A standard curve was plotted from measurements performed with the standard solution (5–1000 pg/ml for VEGF, 2–2000 pg/ml for PDGF-AA, 1.7–2000 pg/ml for PDGF-AΒ and 15–2000 pg/ml for PDGF-ΒΒ) and was used to determine the concentration of VEGF or PDGF-AA in each sample. The level of each factor in the vitreous and serum was within the detection range of the relevant assay, with the minimum detectable concentration being 5.0 pg /ml for VEGF [the intra-assay coefficient of variation (CV) was 5.1% and the inter-assay CV was 6.2%], 2 pg /ml for PDGF-AA (intra-assay CV of 6.3% and inter-assay CV of 6.9%), 1.7 pg /ml for PDGF-AΒ (intra-assay CV of 2.2% and inter-assay CV of 6.5%) and 15 pg/ml for PDGF-ΒΒ (intra-assay CV of 4.5% and inter-assay CV of 7.6%).

Protein content was determined in aliquots of tissue specimens using the standard Bradford assay (Bio-Rad Laboratories, Hercules, California, USA) using bovine serum albumin (Sigma-Aldrich Corporate Offices, St Louis, Missouri, USA) as standard. All data presented were normalized per protein content.

Statistical analysis

All values were expressed as the mean ± standard error (SE). Statistical analysis was performed using the Statistical Package for the Social Sciences (spss, SPSS Inc., Chicago, Illinois, USA). The Kolmogorov–Smirnov test was employed to confirm the assumption of the normality of the variables. The differences between groups were analysed by multivariate analysis of variance (manova). The differences between groups with respect to sex were tested by chi-square test (Table 1). Correlations between groups were studied by Pearson’s correlation coefficient test or Spearman rank correlation test in case of non-continuous variables. A two-tailed p-value of < 0.05 was considered to indicate statistical significance.

Table 1.   Characteristics of diabetic patients and controls included in the study.
 NPDRPDRControls
  1. NPDR, non-proliferative diabetic retinopathy; PDR, proliferative diabetic retinopathy; HbA1c, haemoglobin A1c.

Male (%)6041.946.6
Patients (n)9137
Eyes (n)10148
Female (%)405853.3
Patients (n)6188
Eyes (n)81910
Age (years)66.7 ± 1.668.7 ± 1.468.3 ± 1.5
Duration of diabetes (years)16.8 ± 0.618.3 ± 1.4
HbA1c (%)7.6 ± 0.097.7 ± 0.2< 6.4

Results

Classification of patients

Sixty-nine eyes were included in our study: 18 eyes (15 patients) in group A (NPDR), 33 eyes (31 patients) in group B (PDR) and 18 eyes (15 patients) in group C (control). Characteristics of patients included in the study are shown in Table 1. There were no significant differences between the diabetic patients and controls with respect to sex and age [p = 0.578, p = 0.734 (group A versus group C), p = 0.689, p = 0.840 (group B versus group C)]. There were no significant differences between the two diabetic groups with respect to sex, age, duration of diabetic disease and levels of HbA1c in serum (all p > 0.05).

Classification according to ETDRS severity scale regarding baseline diabetic disease revealed that there were six eyes with 43 ETDRS grade, four eyes with 47, eight eyes with 53, six eyes with 71, 11 eyes with 75, nine eyes with 81 and seven eyes with 85 (Table 2). All patients in group A had clinically significant macular oedema (CSMO) and had been treated unsuccessfully with laser photocoagulation previously.

Table 2.   Number of diabetic eyes per Early Treatment for Diabetic Retinopathy Severity (ETDRS) grade.
ETDRS grade43475371758185
Εyes (n)64861197

Effect of age, sex and presence of other diseases on the levels of VEGF and PDGFs in vitreous and serum of patients with NPDR

In patients with NPDR and CSMO, there was no significant correlation between vitreous or serum levels of any of the PDGF isoforms or VEGF and sex, age, duration of diabetes, existence of hypertension, degree of renal function or treatment for hypertensive or hyperglycaemia (results not shown). In all subsequent analysis of the results, patients with NPDR were grouped irrespective of differences in any of the above parameters.

Effect of NPDR on the vitreous and serum levels of PDGF and VEGF

As expected, VEGF in vitreous was significantly increased in the NPDR group (418 ± 32 pg/mg protein) compared to controls (55 ± 7 pg/mg protein) (p < 0.0001) (Fig. 1). PDGF-AA and -AB levels in vitreous were also increased significantly in the NPDR group (100 ± 13pg/mg protein and 52 ± 7 pg/mg protein, respectively) compared to controls (42 ± 11 pg/mg protein, 25 ± 8 pg/mg protein and 39 ± 5 pg/mg protein, respectively) (p < 0.009 and p < 0.047, respectively) (Fig. 1). However, PDGF-BB levels in vitreous were significantly lower in the NPDR group (22 ± 2.6 pg/mg protein) compared to controls (39 ± 3.8 pg/mg protein) (p < 0.049) (Fig. 1).

Figure 1.

 Concentration of vascular endothelial growth factor and platelet-derived growth factor isoforms in vitreous of patients with non-proliferative diabetic retinopathy (= 18) and controls (= 18). Mean values ± standard errors of triplicate determinations for n eyes are shown (***p < 0.0001, **p < 0.009, *p < 0.05).

There were no significant differences in serum levels of VEGF between the NPDR group (115 ± 4 pg/mg protein) and controls (154 ± 23 pg/mg protein) (p = 0.129) (results not shown). Similarly, there were no significant differences in the serum levels of PDGF-AA, PDGF-AB and PDGF-BB between the NPDR group and controls [788 ± 70 pg/mg protein versus 936 ± 184 pg/mg protein (p = 0.484), 4033 ± 427 pg/mg protein versus 5114 ± 434 pg/mg protein (p = 0.153) and 940 ± 69 pg/mg protein versus 1007 ± 101 pg/mg protein (p = 0.656), respectively] (results not shown).

Correlation of VEGF and PDGF in vitreous with the severity of NPDR

The levels of VEGF in the vitreous correlated significantly with the severity of NPDR (= 0.762, p = 0.017) (Fig. 2A). PDGF-AA also correlated significantly with the severity of NPDR (= 0.683, p = 0.042) (Fig. 2B). The vitreous levels of PDGF-AB and PDGF-AB did not correlate significantly with the severity of NPDR (= 0.077, p = 0.845 and = 0.416, p = 0.265, respectively).

Figure 2.

 (A) Vascular endothelial growth factor and Early Treatment Diabetic Retinopathy Study (ETDRS) severity scale. ETDRS severity scale: 1 = 43 (six eyes), 2 = 47 (four eyes), 3 = 53 (eight eyes). (B) Platelet-derived growth factor-AA and ETDRS severity scale. ETDRS severity scale: 1 = 43 (six eyes), 2 = 47 (four eyes), 3 = 53 (eight eyes).

Effect of grid laser retinal photocoagulation on the vitreous levels of VEGF and PDGF in patients with NPDR

There were no significant correlations between the vitreous levels of VEGF, PDGF-AA, PDGF-AB and PDGF-BB and the extent of laser retinal photocoagulation (RP) that was performed before vitrectomy in patients with NPDR [= 0.274 (p = 0.476), = 0.130 (p = 0.740), = 0.087 (p = 0.824) and = 0.409 (p = 0.274), respectively] (results not shown).

Correlation of VEGF and PDGF with HbA1c in patients with NPDR

There were no significant correlations between the levels of VEGF or PDGF isoforms in the vitreous or serum of patients with NPDR and the levels of HbA1c (results not shown).

Correlation between vitreous and serum levels of VEGF or PDGF in patients with NPDR

There was no correlation between vitreous and serum levels of VEGF [= 0.207 (p = 0.592)], PDGF-AA [= 0.006 (p = 0.987)], PDGF-AB [= 0.472 (p = 0.200)] or PDGF-BB [= 0.039 (p = 0.921)] in patients with NPDR (results not shown).

Correlations between vitreous concentrations of VEGF and PDGF isoforms in patients with NPDR

A strong correlation between vitreous levels of PDGF-AB and PDGF-BB was found only in the NPDR group [= 0.695 (p = 0.038)] (Fig. 3).

Figure 3.

 Correlations between vitreous levels of platelet-derived growth factor (PDGF)-AB and PDGF-BB in patients with non-proliferative diabetic retinopathy (= 18).

No significant correlations were found between vitreous levels of VEGF and any of the PDGF isoforms [PDGF-AA: = 0.453 (p = 0.221); PDGF-AB: = 0.150 (p = 0.701); PDGF-BB: = 0.160 (p = 0.681)] (results not shown). There were also no significant correlations between vitreous levels of PDGF-AA and PDGF-AB (= 0.433, p = 0.245) or PDGF-AA and PDGF-BB (= 0.285, p = 0.457) in the NPDR group (results not shown). There was no correlation between any of the growth factors measured in the serum of patients with NPDR (results not shown).

Comparison between vitreous or serum levels of VEGF and PDGF isoforms in PDR and NPDR

As expected, VEGF in vitreous was significantly increased in the PDR group (2632 ± 330 pg/mg protein) compared to the NPDR group (418 ± 32 pg/mg protein) (p < 0.0001) (Fig. 4). PDGF-AA, -AB and -BB levels in vitreous were also significantly increased in the PDR group (337 ± 47 pg/mg protein, 101 ± 10 pg/mg protein and 79 ± 8 pg/mg protein, respectively) compared to the NPDR group (100 ± 13 pg/mg protein, 52 ± 7 pg/mg protein and 22 ± 2.6 pg/mg protein, respectively) (p < 0.009, p < 0.018 and p < 0.001, respectively) (Fig. 4).

Figure 4.

 Concentration of vascular endothelial growth factor and platelet-derived growth factor isoforms in vitreous of patients with proliferative diabetic retinopathy (= 33) and non-proliferative diabetic retinopathy (= 18). Mean values ± standard errors of triplicate determinations for n eyes are shown (***p < 0.0001, *p < 0.05).

There were no significant differences in serum levels of VEGF between the PDR group (153 ± 18 pg/mg protein) and the NPDR group (115 ± 4 pg/mg protein) (p = 0.055) (results not shown). Similarly, there were no significant differences in the serum levels of PDGF-AA, PDGF-AB and PDGF-BB between the PDR group (931 ± 96 pg/mg protein, 5146 ± 446 pg/mg protein and 1194 ± 75 pg/mg protein, respectively) and the NPDR group (788 ± 70 pg/mg protein, 4033 ± 427 pg/mg protein and 940 ± 69 pg/mg protein, respectively) (p = 0.309, p = 0.210 and p = 0.090, respectively) (results not shown).

Comparative diagrams between PDR, NPDR and controls regarding VEGF and each PDGF isoform

VEGF in vitreous was significantly increased in the PDR group (2632 ± 330 pg/mg protein) compared to controls (55 ± 7 pg/mg protein) (p < 0.001), as described previously (Praidou et al. 2009) (Fig. 5). PDGF-AA, -AB and -BB levels in vitreous were also significantly increased in the PDR group (337 ± 47 pg/mg protein, 101 ± 10 pg/mg protein and 79 ± 8 pg/mg protein, respectively) compared to controls (42 ± 11 pg/mg protein, 25 ± 8 pg/mg protein and 39 ± 5 pg/mg protein, respectively) (p < 0.001), as described previously (Praidou et al. 2009) (Figs 6–8, respectively).

Figure 5.

 Concentration of vascular endothelial growth factor in vitreous of patients with proliferative diabetic retinopathy (= 33), non-proliferative diabetic retinopathy (= 18) and controls (= 18). Mean values ± standard errors of triplicate determinations for n eyes are shown (***p < 0.0001).

Figure 6.

 Concentration of platelet-derived growth factor-AA in vitreous of patients with proliferative diabetic retinopathy (= 33), non-proliferative diabetic retinopathy (= 18) and controls (= 18). Mean values ± standard errors of triplicate determinations for n eyes are shown (***p < 0.0001, *p = 0.009).

Figure 7.

 Concentration of platelet-derived growth factor-AB in vitreous of patients with proliferative diabetic retinopathy (= 33), non-proliferative diabetic retinopathy (= 18) and controls (= 18). Mean values ± standard errors of triplicate determinations for n eyes are shown (***p < 0.0001, **p = 0.018, *p = 0.047).

Figure 8.

 Concentration of platelet-derived growth factor-BB in vitreous of patients with proliferative diabetic retinopathy (= 33), non-proliferative diabetic retinopathy (= 18) and controls (= 18). Mean values ± standard errors of triplicate determinations for n eyes are shown (***p < 0.0001, *p < 0.05).

Discussion

In this study we present evidence that, in addition to VEGF, PDGF isoforms are also associated with NPDR pathology, and that vitreous levels of PDGF isoforms correlate differentially to specific parameters associated with NPDR. The results regarding PDGF isoforms are presented in comparison to the well-established involvement of VEGF in NPDR (Funatsu et al. 2002, 2003, 2005, 2006; Patel et al. 2006).

There was no significant correlation between vitreous or serum levels of any of the PDGF isoforms or VEGF and specific characteristics of the patients with NPDR and CSMO, such as sex, age, duration of diabetes, existence of hypertension, renal dysfunction, or treatment for hypertensive or hyperglycaemia.

The levels of VEGF in vitreous in the NPDR group were significantly increased compared to controls, in accordance with previous reports in the literature (Funatsu et al. 2002, 2003, 2005, 2006; Patel et al. 2006), confirming the involvement of VEGF in increased vascular permeability. The levels of VEGF in the serum of patients with NPDR were similar to controls, suggesting that increased levels of VEGF in vitreous in the NPDR group might be attributed to an intraocular synthesis.

PDGF-AA and -AB levels were also significantly increased in the vitreous of patients with NPDR compared to controls, indicating that PDGF is also involved in the pathology of NPDR. However, PDGF-BB vitreous levels were significantly decreased in NPDR compared to controls. A reduction in PDGF-BB expression in diabetes reveals yet another inequality between the vascular endothelial cells and pericytes of the retinal microvasculature during diabetes (Cox et al. 2003). Therefore, in early diabetes endothelial cells may be sustained while pericytes are deprived of their principal growth factor (Kodama et al. 2001; Cox et al. 2003). Diabetes has been shown to upregulate the expression of VEGF in the retina (Cai & Boulton 2002), a major survival factor for endothelial cells, but increased PDGF has been demonstrated only in proliferative fibrovascular tissue from epiretinal membranes (Cox et al. 2003) removed from patients with end-stage PDR or in vitreous from similar patients in whom the growth factor may have derived from haemorrhage or vascular leakage (Freyberger et al. 2000). This inequality may contribute to the differential loss of pericytes compared to endothelial cells in DR. In PDR both VEGF and PDGF-BB were increased compared to controls.

There was no significant correlation between the levels of VEGF and each of the PDGF isoforms in the vitreous of patients with NPDR. Only PDGF-AB vitreous levels correlated significantly to the PDGF-BB levels in vitreous. In contrast, PDGF isoforms in serum did not differ between NPDR and controls, indicating that the high vitreous concentrations of PDGF isoforms observed in NPDR patients should be attributed to intraocular synthesis and not to serum diffusion. However, it must be pointed out that increased PDGF and especially PDGF-BB serum levels have been reported in diabetic patients compared to controls (Nakashima et al. 1992; Harrison et al. 1994). Finally, there were no correlations between vitreous and serum levels of VEGF or PDGF in patients with NPDR, indicating that serum levels of VEGF and/or PDGF may not be a useful predictor of the severity of NPDR.

With respect to the severity of NPDR (Early Treatment Diabetic Retinopathy Study Research Group 1991a, 1991b), we found significant correlation to the levels of VEGF and PDGF-AA in the vitreous. This might imply that these particular factors are specifically related to the progression of NPDR.

We found no significant correlation between VEGF (Funatsu et al. 2002, 2005, 2006) or any of the PDGF isoforms in vitreous and the extent of laser photocoagulation that was performed before vitrectomy.

The levels of VEGF or PDGF isoforms in the vitreous or in serum did not correlate significantly to HbA1c levels in serum (Funatsu et al. 2005, 2006). The lack of correlation between levels of VEGF or PDGF and HbA1c may be attributed to the fact that the NPDR patients participating in the study underwent strict glycaemic control before surgery. Therefore, HbA1c levels at the time of operation may not necessarily reflect their long-term glycaemic control, making it difficult to assess the real relationship between hyperglycaemia and vitreous levels of PDGF or VEGF.

The levels of VEGF in vitreous in the PDR group were significantly increased compared to NPDR with CSMO, in accordance with previous reports (Ogata et al. 2002). Nevertheless, in one paper (Patel et al. 2006) the exact opposite result was found: VEGF in vitreous in the NPDR group with CSMO was significantly increased compared to PDR (Patel et al. 2006). The levels of VEGF in the serum of patients with PDR were similar to those found in the NPDR group. PDGF-AA, -AB and -BB levels were also significantly increased in the vitreous of patients with PDR compared to the NPDR group. PDGF isoforms in serum did not differ between the PDR and NPDR groups.

In conclusion, our findings indicate that, in addition to VEGF, the levels of PDGF isoforms in the vitreous (but not in the serum) also correlate with the pathology of DR, revealing the complex pathogenesis of this disease. Downregulation of PDGF isoforms, supplementary to anti-VEGF treatment – a popular treatment for complicated DR cases nowadays (Adamis et al. 2006; Jorge et al. 2006; Andreoli & Miller 2007; Nagpal et al. 2007; Minnella et al. 2008; Tonello et al. 2008)– may offer an alternative, synergistic target for the treatment of DR.

Acknowledgements

This research project (PENED) is co-financed by the EU European Social Fund (75%) and the Greek Ministry of Development – General Secretariat of Research and Technology (25%). The last two authors contributed equally to the project and share senior authorship.

Ancillary