Proliferative vitreoretinopathy (PVR), the most common cause of failure of retinal reattachment surgery, is characterised by the development of fibrocellular membranes on either side of the retina. The formation and gradual contraction of these membranes cause a marked distortion of the retina and result in complex retinal detachments that are difficult to repair. The presence of α-smooth-muscle actin (α-SMA) expressing myofibroblasts in PVR epiretinal membranes has been previously reported.1^–3 Myofibroblasts are found at sites of wound healing and chronic inflammation, and are believed to play a pivotal role in the healing process and in the pathogenesis of fibrosis. By secreting extracellular matrix proteins, and by promoting the contraction of the granulation tissue through the expression of the contractile protein α-SMA, these cells are essential for wound repair.4 Understanding the origin of these cells and the mechanism of recruitment should shed further insight into the basis for the development of PVR.

Fibrocytes are circulating bone marrow-derived mesenchymal progenitors that co-express haematopoietic stem-cell antigens, markers of the monocyte lineage and fibroblast products. These cells represent a very low percentage (0.1–0.5%) of the circulating leucocyte population and display an adherent spindle-shaped morphology when cultured in vitro. Fibrocytes express the haematopoietic stem cell/progenitor marker CD34, the leucocyte common antigen CD45 (a pan-haematopoietic marker) and several markers of the monocyte linage in conjunction with vimentin and collagens.5 6 Human circulating fibrocytes acquire the myofibroblast phenotype under in vitro stimulation with the fibrogenic cytokines transforming growth factor (TGF)-β₁ and endothelin-1.6^–10

Several studies demonstrated that circulating fibrocytes traffic to the sites of tissue injury, differentiate into myofibroblasts and contribute to collagen deposition and fibrosis.6 8^–13 There is increasing evidence that these cells contribute to the new population of myofibroblasts that emerge at the tissue site during wound repair, hypertrophic scars and keloids, airway remodelling in asthma, interstitial pulmonary fibrosis, systemic fibroses, intimal hyperplasia, atherosclerosis, reactive fibrosis in chronic pancreatitis and cystitis, and tumour-induced stromal reaction.9 12 Subsequent studies have revealed that these circulating fibrocytes express the chemokine receptors CCR7 and CXCR4.6 10 Furthermore, CCR7- and CXCR4-positive fibrocytes migrate in response to secondary lymphoid tissue chemokine (CCL21/SLC) and stromal cell-derived factor-1 (CXCL12/SDF-1), respectively, to sites of fibrosis.6 10 11

The roles of fibrocytes in the pathogenesis of PVR and their trafficking into the eye have not yet been investigated. The aim of the present study was, therefore, to determine whether circulating fibrocytes contribute to the myofibroblast population present in the epiretinal membranes from patients with PVR and whether CCL21/CCR7 and CXCL12/CXCR4 chemotactic pathways are expressed in these membranes. To identify fibrocytes that had undergone local differentiation into myofibroblasts, double immunohistochemistry was performed to reveal cells co-expressing both CD34 and α-SMA and both CD45 and α-SMA.

METHODS

Epiretinal membrane specimens

Epiretinal membranes were obtained from 16 eyes undergoing vitreoretinal surgery for the treatment of retinal detachment complicated by PVR. All eyes had had previous vitrectomy for rhegmatogenous retinal detachment. The mean interval between the two interventions was 117.4 (73.5) days (range 43–300 days). Membranes were fixed in 10% formalin solution and embedded in paraffin. The study was approved by the Research Centre, College of Medicine, King Saud University.

Immunohistochemical staining

After deparaffinisation, endogenous peroxidase was abolished with 2% hydrogen peroxide in methanol for 20 min, and nonspecific background staining was blocked by incubating the sections for 5 min in normal swine serum. Antigen retrieval was performed by boiling the sections in 10 mM citrate buffer (pH 6) for 30 min. Subsequently, the sections were incubated with the monoclonal antibodies listed in table 1. Optimal working concentration and incubation time for the antibodies were determined earlier in pilot experiments. For CD34, vimentin, CCR7, CXCR4 and CXCL12 immuno-histochemistry, the sections were incubated for 30 min with goat anti-rabbit or anti-mouse immunoglobulins conjugated to peroxidase-labelled dextran polymer (EnVision⁺; Dako, Carpinteria, CA). For α-SMA, CD45 and CCL21 immunohistochemistry, the sections were incubated for 30 min with the biotinylated secondary antibody and reacted with the avidin-biotinylated peroxidase complex (Dako). The reaction product was visualised by the addition of 0.05% 3-amino-9-ethylcarbazole (Sigma-Aldrich, Bornem, Belgium) and hydrogen peroxide, resulting in bright-red immunoreactive sites or 3, 3′-diaminobenzidine (Dako) and hydrogen peroxide, resulting in brown immunoreactive sites. The slides were faintly counterstained with Harris haematoxylin. Finally, the sections were rinsed with distilled water and coverslipped with glycerol.

Double immunohistochemistry

To identify the phenotype of cells expressing CD34, CD45, CCR7, CXCR4, CCL21, and CXCL12, sequential double immunohistochemistry was performed. Antigen retrieval was performed by boiling the sections in 10 mM citrate buffer (pH 6) for 30 min. The sections were then incubated with the first antibody. For CD34, CCR7, CXCR4 and CXCL12 immunohistochemistry, the sections were incubated for 30 min with goat anti-rabbit or anti-mouse immunoglobulins conjugated to peroxidase-labelled dextran polymer (Envision⁺, Dako). For CD45 and CCL21 immunohistochemistry, the sections were incubated for 30 min with the biotinylated secondary antibody and reacted with the avidin-biotinylated peroxidase complex (Dako). The reaction product was visualised by addition of 0.05% 3-amino-9-ethylcarbazole (sigma) and hydrogen peroxide, resulting in bright-red immunoreactive sites. Subsequently, the slides were incubated with anti-α-SMA monoclonal antibody. The slides were then incubated for 30 min with the biotinylated secondary antibody and reacted with the avidin-biotinylated alkaline phosphatase complex (Dako). The blue reaction product was developed using 5-bromo-4-chloro-3-indoxyl phosphate and nitro blue tetrazolium chloride (Dako) for 30 min. No counterstain was applied.

Omission or substitution of the primary antibody with an irrelevant antibody of the same species and staining with chromogen alone were used as negative controls. Sections from patients with colorectal carcinoma were used as positive controls. The sections from the control patients were obtained from patients treated at the University Hospital, University of Leuven, Belgium, in full compliance with the tenets of the Declaration of Helsinki.

Quantitation

Cells were counted in five random representative fields, using an eyepiece calibrated grid with 40× magnification. With this magnification and calibration, cells present in an area of 0.33×0.22 mm were counted. Counting was performed by two independent observers (AMA and KG). One of them (KG) was unaware of the type of the antibody staining. The numbers are the means of the two counts. Spearman’s rank correlation coefficients were computed to investigate the linear relationship between the variables investigated. Mean cell counts for two independent groups of cell types were compared using the Mann–Whitney test. A p value less than 0.05 indicated statistical significance.

RESULTS

There was no staining in the negative control slides (fig 1A). All PVR membranes showed myofibroblasts expressing α-SMA (fig 1B) and vimentin, with a mean number of 396.7.8 (SD 205.9)/mm² (range 104.7–743.8/mm²). Spindle cells expressing CD34 (fig 1C) were noted in eight (50%) membranes with a mean number of 52.9 (82.9)/mm² (range 0–294.8/mm²). Cells expressing CD45 (fig 1D) were present in 12 (75%) membranes, with a mean number of 130.5 (141.8)/mm² (range 0–482.1/mm²). Although the number of cells expressing CD45 was higher than the number of cells expressing CD34, the difference between the two means was not significant (p = 0.065; Mann–Whitney test). Double immunohistochemistry indicated that all cells expressing CD34 (fig 1E), and CD45 (fig 1F) co-expressed α-SMA.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Proliferative vitreoretinopathy epiretinal membranes. (A) Negative control slide that was treated identically with an irrelevant antibody showing no labelling (original magnifcation ×40). (B) Immunohistochemical staining for α-smooth-muscle actin showing immunoreactivity in myofibroblasts (original magnification ×400). (C) Immunohistochemical staining for CD34 showing cells expressing CD34 (original magnification ×40). (D) Immunohistochemical staining for CD45 showing cells expressing CD45 (original magnification ×40). (E) Double immunohistochemistry for CD34 (red) and α-smooth-muscle actin (blue). Low-power (original magnification ×40) (left) and high-power (original magnification ×100) (right) showing cells expressing only α-smooth-muscle actin (arrowheads) and cells co-expressing CD34 and α-smooth-muscle actin (arrows). No counterstain was applied. (F) Double immunohistochemistry for CD45 (red) and α-smooth-muscle actin (blue) showing cells expressing only α-smooth-muscle actin (arrowheads) and cells co-expressing CD45 and α-smooth-muscle actin (arrows). No counterstain was applied (original magnification ×100). Patient consent has been obtained for publication of this figure.

Cells expressing CCR7 (fig 2A) were noted in 11 (68.8%) of the 14 membranes stained for CCR7, with a mean number of 74.6 (63.0)/mm² (range 0–192.8/mm²). Cells expressing CXCR4 (fig 2B) were detected in all membranes with a mean number of 171.8 (119.5)/mm² (range 33.1–427.0/mm²). The mean number of cells expressing CXCR4 was significantly higher than the mean number of cells expressing CCR7 (p = 0.02; Mann–Whitney test). Double immunohistochemistry indicated that all cells expressing CCR7 (fig 2C) and CXCR4 (fig 2D) co-expressed α-SMA.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2 Proliferative vitreoretinopathy epiretinal membranes. (A) Immunohistochemical staining for CCR7 showing cells expressing CCR7 (original magnification ×40). (B) Immunohistochemical staining for CXCR4 showing cells expressing CXCR4 (original magnification ×40). (C) Double immunohistochemistry for CCR7 (red) and α-smooth-muscle actin (blue) showing cells co-expressing CCR7 and α-smooth-muscle actin (arrows). No counterstain was applied (original magnification ×100). (D) Double immunohistochemistry for CXCR4 (red) and α-smooth-muscle actin (blue) showing cells co-expressing CXCR4 and α-smooth-muscle actin (arrows). No counterstain was applied (original magnification ×100). Patient consent has been obtained for publication of this figure.

Immunoreactivity for CCL21 (fig 3A) was noted in large cells in 12 (80%) of the 15 membranes stained for CCL21, with a mean number of 108.4 (80.6)/mm² (range 0–220.4/mm²). Immunoreactivity for CXCL12 (fig 3B) was noted in large cells in 15 (93.8%) membranes with a mean number of 200.5 (126.5)/mm² (range 0–413.2/mm²). In serial sections, the distribution and morphology of cells expressing CCL21 and CXCL12 was similar to the distribution and morphology of myofibroblasts expressing α-SMA. The mean number of cells expressing CXCL12 was significantly higher than the mean number of cells expressing CCL21 (p = 0.024; Mann–Whitney test). Double immunohistochemistry confirmed that all cells expressing CCL21 (fig 3C) and CXCL12 (fig 3D) co-expressed α-SMA.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3 Proliferative vitreoretinopathy epiretinal membranes. (A) Immunohistochemical staining for CCL21 showing cells expressing CCL21 (original magnification ×100). (B) Immunohistochemical staining for CXCL12 showing cells expressing CXCL12 (original magnification ×100). (C) Double immunohistochemistry for CCL21 (red) and α-smooth-muscle actin (blue) showing cells co-expressing CCL21 and α-smooth-muscle actin (arrows). No counterstain was applied (original magnification ×100). (D) Double immunohistochemistry for CXCL12 (red) and α-smooth-muscle actin (blue) showing cells expressing only α-smooth-muscle actin (arrowheads) and cells co-expressing CXCL12 and α-smooth-muscle actin (arrows). No counterstain was applied (original magnification ×100). Patient consent has been obtained for publication of this figure.

Table 2 shows Spearman’s rank correlation coefficients between the number of the studied variables. The number of cells expressing α-SMA correlated significantly with the number of cells expressing CXCL12. The number of cells expressing CD34 correlated significantly with the numbers of cells expressing CD45, CXCL12 and CCL21. Finally, a significant correlation was detected between the numbers of cells expressing CXCR4 and CCR7, and CXCL12 and CCL21, respectively.

DISCUSSION

In the present study, we tested the hypothesis that circulating fibrocytes are contributing to the myofibroblast population present in PVR epiretinal membranes. Using immunohistochemical techniques, we demonstrated for the first time the presence of cells co-expressing both CD34 and α-SMA and both CD45 and α-SMA. The expression of α-SMA is a marker of differentiation of fibrocytes into myofibroblasts,6^–10 and fibrocytes are the only cell population known to express CD34 and CD45 in conjunction with α-SMA.7 8 The ability of expressing α-SMA while retaining CD34 and CD45 expression seems to be a unique property of circulating fibrocytes that have entered a phase of differentiation into α-SMA expressing myofibroblasts.7 The presence of CD34⁺ α-SMA⁺ cells and CD45⁺ α-SMA⁺ cells were also reported in the bronchial mucosa of patients with allergic asthma7 14 and in intimal hyperplasia specimens15 confirming the presence of mature fibrocytes. These cells localised to areas of collagen deposition below the epithelium,7 and their density correlated with the thickness of the basement membrane14 in bronchial biopsies.

Fibrocytes express extracellular matrix proteins, such as collagen type I, collagen type III and fibronectin, which may be involved in connective tissue formation and promotion of fibrosis in vivo.6 7 10 11 13 Fibrocytes have been found to have a number of functions other than promoting fibrosis. They are potent antigen-presenting cells and can elicit the recruitment and activation of T cells.16 Fibrocytes can also induce an angiogenic phenotype in microvascular endothelial cells in vitro and promote of angiogenesis in vivo.17 They can also secrete chemokines, cytokines, matrix metalloproteinase-9 and growth factors which are relevant in mediating fibroproliferation.17^–19 In previous studies, we demonstrated the expression of the chemokine monocyte chemoattractant protein-1,2 connective tissue growth factor, matrix metalloproteinase-93 and vascular endothelial growth factor (unpublished observation) by myofibroblasts in PVR epiretinal membranes.

In vitro and in vivo studies demonstrated that the expression of CD34 and CD45 is downregulated while fibrocytes undergo phenotypic differentiation into α-SMA expressing myofibroblasts.8 10 These data may explain why many myofibroblasts in PVR epiretinal membranes did not express CD34 and CD45 antigens, as they could be fibrocytes that had already completed their differentiation, and this may explain the variations in the numbers of cells expressing CD34 and CD45 in these membranes. Aiba and Tagami20 demonstrated that in cutaneous wounds, especially in the expanding margins of keloids or inflammatory scars, CD34 expression gradually decreased over time while there was an increase in the expression of prolyl-4-hydroxylase, the enzyme required for the synthesis of new collagen. In the present study, the number of cells expressing CD45 was higher than the number of cells expressing CD34. Our results are consistent with prior in vitro studies by Phillips et al,10 who showed that 90% of newly purified human fibrocytes expressed high levels of CD34 and CD45. After 3 weeks in culture, almost 30% of fibrocytes were found to stain positive for CD45, and only 5% of fibrocytes were found to stain positive for CD34.

Another aim of the present study was to determine whether the CCL21/CCR7 and CXCL12/CXCR4 chemokine axes are expressed in PVR epiretinal membranes. We demonstrated the presence of CCR7⁺ α-SMA⁺ and CXCR4⁺ α-SMA⁺ cells and demonstrated that CCL21 and CXCL12 proteins were specifically localised in myofibroblasts. Similarly, Orimo et al21 demonstrated using immunohistochemical techniques that carcinoma-associated α-SMA expressing myofibroblasts were also positive for CXCL12. They also demonstrated that CXCL12 released by carcinoma-associated myofibroblasts enhanced tumour growth, acting through the cognate receptor, CXCR4, which is expressed by carcinoma cells and promoted angiogenesis by recruiting endothelial progenitor cells. Our finding of CCL21 expression by myofibroblasts in PVR epiretinal membranes, however, was not previously reported. Our analysis indicated that the number of cells expressing CD34 in PVR membranes correlated significantly with the number of cells expressing CCL21 and CXCL12. Our findings, thereby, confirm the contribution of CCL21 and CXCL12 to the trafficking of circulating fibrocytes to PVR membranes. Several studies demonstrated important roles of CCL21/CCR7 and CXCL12/CXCR4 signalling pathways in the recruitment of circulating fibrocytes to sites of tissue injury and fibrosis. Fibrocytes express CCR7, the receptor for CCL21, and CXCR4, the receptor for CXCL12.6 10 11 In vitro studies showed directed chemotaxis of cultured fibrocytes in response to the ligand of CCR7, CCL216 and in response to the ligand of CXCR4, CXCL12.10 In vivo studies demonstrated that the CCL21/CCR7 pathway strongly contributes to the trafficking of fibrocytes to sites of cutaneous tissue injury,6 and into the kidney in a murine model of renal fibrosis.11 Blockade of CCL21/CCR7 signalling by anti-CCL21 antibodies reduced the severity of fibrocyte infiltration and the extent of fibrosis in a murine model of renal fibrosis, with the concomitant decrease in renal expression of TGF-β₁ and collagen type I.11 In another study, CXCR4⁺ fibrocytes migrated in response to CXCL12 and trafficked to the lungs in a murine model of bleomycin-induced pulmonary fibrosis.10 Furthermore, treatment of bleomycin-exposed animals with specific neutralising anti-CXCL12 antibodies inhibited infiltration of CXCR4⁺ fibrocytes, attenuated lung fibrosis and reduced α-SMA staining.10 More recently, Mehrad et al22 demonstrated increased expression of the fibrocyte-attracting chemokine, CXCL12, in the lungs and plasma of patients with fibrotic interstitial lung disease and marked expansion of circulating fibrocyte pool in these patients.

In PVR epiretinal membranes, the numbers of cells expressing CXCL12 and CXCR4 were significantly higher than the numbers of cells expressing CCL21 and CCR7. Thus, our data suggest that although two chemotactic signalling pathways are expressed in PVR membranes, the CXCL12/CXCR4 chemokine axis appears to be predominant for the recruitment of fibrocytes into the eye in patients with PVR. Taken together, these findings suggest that circulating fibrocytes may function as precursors of myofibroblasts in PVR epiretinal membranes and may contribute to the genesis of proliferative vitreoretinal disorders. Thus, novel strategies aimed at preventing PVR might address the participation of fibrocytes and the modulation of their activity. Inhibition of CXCL12/CXCR4 and CCL21/CCR7 signalling might provide a new therapeutic approach to prevent PVR.