Traumatic proliferative vitreoretinopathy (PVR) is observed in more than 70% of patients with open globe injuries involving the posterior segment, and usually progresses to traction retinal detachment.1 Traumatic PVR and traction retinal detachment have been confirmed as major risk factors for poor anatomical and functional outcomes after injury.2–4 Therefore, timely and appropriate vitrectomy can effectively terminate the pathological process and improve functional outcomes after injury.5–8
Proliferative vitreoretinopathy is characterized by the formation of fibrous membranes on or beneath the retina, along with intraretinal degeneration and gliosis.9 In fact, it has been demonstrated that PVR development involves five distinct stages. These include breakdown of the blood–retinal barrier, chemotaxis and cellular migration, cellular proliferation, membrane formation with remodeling of the extracellular matrix, and contraction.10 Moreover, multiple cytokines and chemokines are involved in these stages. In early experimental animal models, the development of traumatic PVR was shown to involve different stages and affect vitrectomy outcomes for eye injuries.8,11,12 Nevertheless, because of the lack of a simultaneous examination of cytokines and chemokines, the molecular mechanisms underlying the inflammatory processes, cell proliferation, and extracellular matrix environment remain unclear. Therefore, it is essential to examine the clinical course of PVR while keeping both chemokines and cytokine involvement in mind.
In a clinical setting, the determination of the timing of irreversible intraretinal changes is important to elucidate the ideal timing for vitrectomy. In fact, the ideal timing for vitrectomy in patients with posterior segment involvement is unclear.13 Therefore, clinical observations, pathological studies of chemokines and cytokines, and examinations of intraretinal changes are necessary; otherwise, it may be difficult to reflect the clinical condition using animal models.
In this study, we aimed to evaluate epiretinal and subretinal membranes over the clinical course of PVR after open globe injury. We assessed the variations in cellular components and extracellular matrix components, including the retinal pigment epithelium, glial cells, α-smooth muscle actin (α-SMA), vimentin, and collagen I. In addition, proliferative indices for Ki-67 and proliferating cell nuclear antigen (PCNA) were detected, and the grade of inflammation during the clinical course and changes in chronic inflammation caused by variations in macrophages and the retinal wound healing process were investigated.
In the experiment, each subject underwent continuous observation of the same eye over time, whereas the membranes harvested from the individual patients' eyes with PVR were observed during vitreous surgery. Vitrectomy was not planned ahead of time for any patient. Therefore, the investigation of membranes was conducted by dividing them into groups based on the time stage after injury. Thus, data from the same period could be from different individuals. Through the analysis of these data series with continuous time intervals after injury, we can gain insight into the PVR process over time and obtain knowledge to determine the optimal timing for vitrectomy.
This study was performed with the approval of the Institutional Review Board of Peking University Third Hospital Medical Science Research Ethics Committee (reference no. IRB00006761-2015190) and was conducted in accordance with the tenets of the Declaration of Helsinki. Informed consent was obtained routinely from all patients before participation in this study.
Patients and Tissues
Twenty-one samples of epiretinal and/or subretinal membranes were obtained from 21 patients with open globe injuries and traumatic PVR. The patients were divided into three groups according to the interval between injury and vitrectomy: Group A, ≤28 days (n = 6); Group B, 29 days to 120 days (n = 9); and Group C, >120 days (n = 6). No patient reported a history of an intraocular surgical procedure or systemic disease before the traumatic event. Samples with obvious bleeding were excluded. Using horizontal scissors, the epiretinal and subretinal membranes were dissected from the retinal surfaces of patients with PVR who were undergoing pars plana vitrectomy. Because the severity of traumatic PVR is closely associated with the wound site,11 it was noted whether specimens were harvested from sites adjacent to or further away from the wound site. Clinical features associated with the patients' proliferative membranes are presented in Table 1.
Histology and Immunohistochemistry
Membrane tissues were harvested during vitreoretinal surgery, placed on a cellulose acetate membrane, and fixed in 4% formalin. Then, serial sections with 4 μm thicknesses were used for histology and immunohistochemistry. The sections were stained with hematoxylin and eosin, periodic acid–Schiff, and Masson's trichrome stains using standard procedures. The sections were blocked in 2% normal goat serum in phosphate-buffered saline for 30 minutes to block nonspecific binding and incubated with diluted antibodies for 60 minutes at 45°C. After washing with phosphate-buffered saline 3 times, the samples were incubated with system-labeled horseradish peroxidase antimouse or antirabbit secondary antibody (Dako, Denmark) at room temperature for 20 minutes. Thereafter, the sections were stained in diaminobenzidine, counterstained in Mayer's hematoxylin, and dehydrated in alcohol and xylene. Phosphate-buffered saline was used as a negative control. Specimen staining and scoring were performed under a microscope. The following primary mouse monoclonal antibodies were used: anti-CD68 (Leica; NCL-CD68; 1:100), anti-CD34 (Leica; NCL-L-END; 1:100), anti–α-SMA (Cell Marque; 202M-96; 1:100), anticytokeratin (anti-CK; Cell Marque; 313M-16; 1:100), antivimentin (Invitrogen; 18-0052; 1:100), anti–glial fibrillary acidic protein (GFAP; EPI; AC-0002; 1:100), anti-Ki-67 (Origene; UM800033; 1:100), and anti-PCNA (Invitrogen; 18-0110; 1:100). In addition, rabbit anticollagen I monoclonal antibody (EPI; AC-0223; 1:100) was used.
Anticytokeratin antibody was used to identify retinal pigment epithelium cells, anti-CD68 antibody to identify macrophages, anti-GFAP antibody to identify glial cells (Müller cells, microglia, and fibrous astrocytes), anti–α-SMA antibody to identify smooth muscle cells and myofibroblasts, antivimentin antibody to identify fibroblasts and other cells of mesenchymal origin, and anticollagen I antibody to identify collagen type I. Anti–Ki-67 antibody and anti-PCNA antibody were used to identify proliferative cells.
After immunohistochemical staining, the scores were evaluated by an investigator blinded to the study protocol. The sections were scored by the percentage of positive cells and the staining intensity. The sections were considered to exhibit positive staining if yellow-brown stains were observed in the cell plasma. The intensity of the plasma staining16 was graded as follows: 0, no color; 1, faint yellow color; 2, yellow color; and 3, yellow-brown color. The percentage of positive cells scores were assigned as follows: 0 (<5% positive cells), 1 (5%–25% positive cells), 2 (25%–50% positive cells), 3 (50%–75% positive cells), and 4 (>75% positive cells). The average weighted score for each section was then evaluated by multiplying the staining intensity and percentage of positive cells scores.
In addition, the proliferative index was defined by the number of Ki-67-stained and PCNA-stained nuclei. Sections of each specimen were sampled by taking representative digital photographs of 4 to 10 fields, depending on the size of the membrane. Photographs were obtained at 400× magnification using a Nikon imaging system. Approximately, 90 to 100 cells per field were subsequently counted, with several fields counted for each specimen.17 The proliferative index was expressed as a percentage of the total number of nuclei (an index of 1 is equivalent to 1%) in each field. The mean value (±SEM) for each specimen was obtained by tallying all fields.
To evaluate the presence of inflammation, we counted the number of inflammatory cells, including plasma cells, lymphocytes, and macrophages, in 4 to 10 randomly selected fields (depending on the size of the membrane) and determined the average proportion of all cells under a microscope (magnification, ×400). If the inflammatory cell ratio was <50%, the specimen was considered to exhibit a moderate grade of inflammation, whereas a ratio of >50% indicated remarkable inflammation.
All data were analyzed using SPSS version 17.0 (SPSS, Inc, Chicago, IL). Fisher's exact tests and a one-way analysis of variance were performed. P-values < 0.05 were considered statistically significant.
Dynamic Changes in Retinal Damage and the Main Constituents of Membranes
Fundus photography was performed during vitrectomy. Retinal and membrane tissues were processed for pathological analysis through hematoxylin and eosin, periodic acid–Schiff, and Masson's trichrome staining; and CD34 expression was investigated based on the consecutive time points of specimen collection, namely days 21, 30, 50, 120, and 180 (Figures 1 and 2). Generally, changes in retinal characteristics worsened with time after the injury, based on observations made during surgeries. Structures of the retina became impacted gradually over the postinjury time course, including a decrease in neuronal cells, structural destruction, and an increase in the components of fibrotic tissues. The results suggest that the cornerstone of the remarkable transition in retinal substance occurred at around 30 days after injury (Figure 1E). The collected membranes were initially loosened and multicellular, and gradually became condensed with fibrotic and vascular components. The evolution of the retinal substance and membranes of the PVR paralleled the clinical manifestations of the retinas.
Expression of Cellular and Extracellular Matrix Constituents in Proliferative Vitreoretinopathy Specimens
Specimens were examined between 9 days and 365 days after open globe injury. Epiretinal membranes from seven eyes and subretinal membranes from 14 eyes were obtained. Cytokeratin and CD68 expressions decreased over time, whereas GFAP expression increased over time. The expression of α-SMA and collagen I reached a peak by Day 120, whereas the expression of vimentin increased over time (Figure 3). The mean (SEM) staining scores for Groups A (≤28 days), B (29–120 days), and C (≥120 days) are presented in Tables 2 and 3.
Proliferative Indices for Ki-67 and Proliferating Cell Nuclear Antigen in Proliferative Vitreoretinopathy Specimens
Ki-67/PCNA expression was assessed in the cellular nuclei of all membranes. The mean proliferative index for Ki-67, determined by counting the cells positive for Ki-67, was 22.2 ± 2.60 in Group A (≤28 days), 20.50 ± 1.18 in Group B (29–120 days), and 14.45 ± 1.27 (≥120 days) in Group C; these values were significantly different according to the multivariate analysis (analysis of variance, P = 0.044). The univariate analysis indicated no significant difference between any 2 groups (Group B vs. Group A, P = 0.924.; Group C vs. Group A, P = 0.053; Group B vs. Group C, P = 0.066). The mean proliferative index for PCNA, determined by counting the cells positive for PCNA, was 14.97 ± 3.25 in Group A, 31.49 ± 4.85 in Group B, and 10.83 ±1.78 in Group C; these values were significantly different according to the multivariate analysis (analysis of variance, P = 0.001). The univariate analysis showed a significant difference in PCNA between Groups B and A, and between Groups B and C (Group B vs. Group A, P = 0.002; Group B vs. Group C, P < 0.001), although no significant difference was observed between Groups C and A (P = 0.357; Figure 4).
Inflammation Status of Proliferative Vitreoretinopathy Specimens
We assessed grades of inflammation for all specimens and determined the relationship between inflammation and the clinical time course of PVR. The inflammatory reactions exhibited by proliferative membranes varied. The numbers of specimens with remarkable and moderate inflammation were 4 (66.7%) and 2 (33.33%) in Group A (≤28 days), 5 (55.6%) and 4 (44.4%) in Group B (29–120 days), and zero and 6 (100%) in Group C; differences among the groups were significant (Fisher's exact test, P = 0.049). Inflammatory expression tended to be more moderate between Group A and Group C.
Profiles of Retinal Changes Away From the Wound Site
Retinal detachment was observed in all 21 eyes, whereas subretinal hemorrhage was noted in 14 (66.7%). On day 30 after injury, high GFAP expression was observed with remarkable disorganization of the retinal tissues. Masson's trichrome staining showed rare collagen fibers. On day 120 after injury, lower GFAP expression was observed in the retina along with complete disorganization of the retinal tissues. By this time, Masson's trichrome staining indicated the presence of mature collagen fibers (Figure 5).
In this study, all patients with open globe injuries involving the posterior segment were confirmed to exhibit Grade C PVR and total retinal detachment during vitrectomy. All retinal detachments occurred immediately after injury, as confirmed by B-scans. We assessed the dynamic progression of PVR in terms of inflammatory reactions and cell proliferation in the eyes of patients who underwent vitrectomy, using previous experimental animal model–based studies for comparisons.8,11,12 In addition, we simultaneously assessed pathological changes in the retina and the stages of the membranes representing the development of PVR. Finally, we compared the features of the PVR membranes and retinas as observed during vitrectomy and at different time points after injury. Clinical information regarding the later stages of PVR is scarce in previous studies.18,19
With respect to clinical observations during vitrectomy, retinal fold formation was initially observed in PVR foci at approximately 30 days after injury. At that time, the retina was soft and could be flattened easily. Thereafter, the retina became opaque and thick and was characterized by edema and loss of elasticity, followed by stiffening and shortening.
In this study, the cellular components of the membranes showed changes over time from the early cellular stage to the late paucicellular stage. Recent studies have indicated that the early proliferation of retinal pigment epithelium cells and loss of cell–cell contact result in an epithelial–mesenchymal transition.20,21 In comparison, the expression of GFAP, which was used to identify glial cells, was upregulated in the membranes of tissues with advanced PVR. These findings are consistent with clinical observations, wherein the formation of a glial scar-like membrane was noted on both sides of a detached retina.
Previous studies demonstrated that fibroblasts appeared at a much earlier stage because of the occurrence of wound healing outside the retina.1 Thereafter, the number of myofibroblasts increased and the number of fibroblasts gradually decreased.22,23 In the present study, the expression of the contractile protein α-SMA—a myofibroblast marker—showed a stable increase from the early stages until 4 months after injury and decreased thereafter. This finding suggests that the greatest morphological changes in the retina occurred after 1 month. Thereafter, until 3 months to 4 months, the morphological changes were continuously advanced in the retinas, such as remodeling and thickening between the retinal folds and membrane. The expression of vimentin—a fibroblast marker—showed a significant increase from after 1 month until 4 months after injury. Clinical observations demonstrated that retinal detachment and the wound healing process continued until vitrectomy. Moreover, the active proliferative phase can be assumed to range from 1 month to 4 months after injury, according to the expression of Ki-67 and PCNA. In fact, the expressions of Ki-67 and PCNA were still detected to some extent after 4 months. All these findings suggest that cell proliferation continues until vitrectomy. In addition, the proliferation associated with inflammatory processes may also indicate that the process of PVR did not cease until vitrectomy. This can be explained by the low level of trauma-related inflammation observed until 4 months after injury and the nonsignificant difference in CD68 expression between the first and fourth months after injury.
Previous studies have primarily focused on proliferation of the vitreous and epiretinal membranes, whereas intraretinal changes during the PVR process have not been well examined.24 In fact, the pathological process of retinal changes caused by PVR should be a major factor in clinical decision making.25 Attachment of the retina before retinal gliosis formation could be necessary to prevent the PVR process and restore visual function.7,26,27 However, the specific phase that leads to irreversible retinal changes remains unclear. In this study, we noted that the inner and outer nuclear layers could be observed in the detached retina even at 30 days after injury. Moreover, a significantly greater amount of retinal changes was observed at the site of membrane adherence. With the progression of PVR, disorganization of the retinal tissue followed by a decrease in the number of cells and loosening of the structures were observed. In addition, the main component of the epiretinal membrane was found to be collagen, which fused with the severely disorganized retina and could not be clearly distinguished between the retina and membrane. These pathological changes were consistent with the clinical features, including retinal edema and thickening, increases in opacity and stiffness, loss of elasticity, and lack of a photocoagulation reaction. Moreover, the high expressions of GFAP and collagen I in the membranes were correlated with intraretinal pathological changes. These findings suggest that vitrectomy should be performed no later than when irreversible retinal changes occur.
This study has certain limitations. One limitation was the relatively small sample size. Moreover, additional assessments at shorter intervals are required for specimens obtained 29 days to 120 days after injury. Furthermore, variations in GFAP and collagen expressions within the retina and changes in PCNA expression should be assessed in further studies.
In conclusion, our observations based on PVR membranes and clinical findings show that PVR progression in eyes with open globe injuries, including inflammation, proliferation, and scar formation, does not cease until vitrectomy is performed to reattach the retina, although the most active stage of proliferation was around 29 days after injury. With respect to the retinal substance, it progressed from proliferation to irreversible effects as PVR advanced over time, although we were unable to determine the time at which damage to the retina could become irreversible from 29 days to 120 days after injury. The present observational results and analysis can serve as fundamental evidence and an important reference for determining the proper timing for vitrectomy.
1. Winthrop SR, Cleary PE, Minckler DS, Ryan SJ. Penetrating eye injuries: a histopathological review. Br J Ophthalmol 1980;64:809–817.
2. Cardillo JA, Stout JT, LaBree L, et al. Post-traumatic proliferative vitreoretinopathy
. The epidemiologic profile, onset, risk factors, and visual outcome. Ophthalmology 1997;104:1166–1173.
3. AM El-Asrar, Al-Amro SA, Khan NM, Kangave D. Visual outcome and prognostic factors after vitrectomy
for posterior segment foreign bodies. Eur J Ophthalmol 2000;10:304–311.
4. Andreoli MT, Andreoli CM. Surgical rehabilitation of the open globe injury
patient. Am J Ophthalmol 2012;153:856–860.
5. Cleary PE, Ryan SJ. Vitrectomy
in penetrating eye injury. Results of a controlled trial of vitrectomy
in an experimental posterior penetrating eye injury in the rhesus monkey. Arch Ophthalmol 1981;99:287–292.
6. Entezari M, Rabei HM, Badalabadi MM, Mohebbi M. Visual outcome and ocular survival in open-globe injuries. Injury 2006;37:633–637.
7. Nashed A, Saikia P, Herrmann WA, et al. The outcome of early surgical repair with vitrectomy
and silicone oil in open-globe injuries with retinal detachment. Am J Ophthalmol 2011;151:522–528.
8. Kuhn F. The timing of reconstruction in severe mechanical trauma. Ophthalmic Res 2014;51:67–72.
9. Garweg JG, Tappeiner C, Halberstadt M. Pathophysiology of proliferative vitreoretinopathy
in retinal detachment. Surv Ophthalmol 2013;58:321–329.
10. Wilkins RB, Kulwin DR. Wound healing. Ophthalmology 1979;86:507–510.
11. Cleary PE, Ryan SJ. Histology of wound, vitreous, and retina in experimental posterior penetrating eye injury in the rhesus monkey. Am J Ophthalmol 1979;88:221–231.
12. Ramsay RC, Cantrill HL, Knobloch WH. Vitrectomy
for double penetrating ocular injuries. Am J Ophthalmol 1985;100:586–589.
13. Agrawal R, Shah M, Mireskandari K, Yong GK. Controversies in ocular trauma classification and management: review. Int Ophthalmol 2013;33:435–445.
14. Kuhn F, Morris R, Witherspoon CD. Birmingham Eye Trauma Terminology (BETT): terminology and classification of mechanical eye injuries. Ophthalmol Clin North Am 2002;15:139–143.
15. Machemer R, Aaberg TM, Freeman HM, et al. An updated classification of retinal detachment with proliferative vitreoretinopathy
. Am J Ophthalmol 1991;112:159–165.
16. Hussein MR. Analysis of Bcl-2 protein expression in choroidal melanomas. J Clin Pathol 2005;58:486–489.
17. Zhang X, Barile G, Chang S, et al. Apoptosis and cell proliferation in proliferative retinal disorders: PCNA, Ki-67, caspase-3, and PARP expression. Curr Eye Res 2005;30:395–403.
18. Faulborn J, Topping TM. Proliferations in the vitreous cavity after perforating injuries. A histopathological study. Albrecht Von Graefes Arch Klin Exp Ophthalmol 1978;205:157–166.
19. Esmaeli B, Elner SG, Schork MA, Elner VM. Visual outcome and ocular survival after penetrating trauma. A clinicopathologic study. Ophthalmology 1995;102:393–400.
20. Ishikawa K, He S, Terasaki H, et al. Resveratrol inhibits epithelial-mesenchymal transition of retinal pigment epithelium and development of proliferative vitreoretinopathy
. Sci Rep 2015;5:16386.
21. Tamiya S, Liu L, Kaplan HJ. Epithelial-mesenchymal transition and proliferation of retinal pigment epithelial cells initiated upon loss of cell-cell contact. Invest Ophthalmol Vis Sci 2010;51:2755–2763.
22. Wiedemann P, Ryan SJ, Novak P, Sorgente N. Vitreous stimulates proliferation of fibroblasts and retinal pigment epithelial cells. Exp Eye Res 1985;41:619–628.
23. Feist RM Jr, King JL, Morris R, et al. Myofibroblast and extracellular matrix origins in proliferative vitreoretinopathy
. Graefes Arch Clin Exp Ophthalmol 2014;252:347–357.
24. Wickham L, Charteris DG. Glial cell changes of the human retina in proliferative vitreoretinopathy
. Dev Ophthalmol 2009;44:37–45.
25. Kolomeyer AM, Grigorian RA, Mostafavi D, et al. 360° retinectomy for the treatment of complex retinal detachment. Retina 2011;31:266–274.
26. Stryjewski TP, Andreoli CM, Eliott D. Retinal detachment after open globe injury
. Ophthalmology 2014;121:327–333.
27. Reed DC, Juhn AT, Rayess N, et al. Outcomes of retinal detachment repair after posterior open globe injury
. Retina 2016;36:758–763.