A IMMUNOHISTOCHEMICAL ANALYSIS OF A RAT MODEL OF PROLIFERATIVE VITREORETINOPATHY AND A COMPARISON OF THE EXPRESSION OF TGF-β AND PDGF AMONG THE INDUCTION METHODS

Proliferative vitreoretinopathy (PVR) is a serious complication of retinal detachment surgery or ocular trauma. Our previous study indicated that intravitreal co-injection of retinal pigmented epithelial (RPE) -J cells and platelet-rich plasma ( PRP) (not RPE-J cells or PRP alone) in Wistar rat eyes can successfully induce a model of PVR. But which cells are involved in this process and why diff erent induction methods, intravitreal injection of RPE-J cells or/and PRP, induced a diff erent situation remain to be unknown. In this study, immunohistochemistry was performed to identify the main cell types involved in this process. Y e expression levels of transforming growth factor (TGF) -β, platelet-derived growth factor (PDGF)AA and PDGF-BB were tested using enzyme-linked immunosorbent assay (ELISA).Y e results showed that RPE cells, glial cells, fi broblasts and macrophages took part in the pathogenesis of this model. Y e expression levels and durations of TGF-β and PDGF-BB partially explained the diff erent results induced by the diff erent induction methods. Y is provides an experimental proof for attenuation of the experimental PVR by targeting at a specifi c cells or growth factor.


Introduction
Proliferative vitreoretinopathy (PVR) is a serious complication of retinal detachment surgery or ocular trauma ().To study the pathogenesis of PVR and the therapies for this disease, various in vivo models have been developed.However, as one of the most studied animal, rat, their models of PVR are still seldom reported.In our previous research, we have confi rmed that intravitreal co-injection of retinal pigmented epithelial (RPE) -J cells and platelet-rich plasma ( PRP) (not RPE-J cells or PRP alone) in Wistar rat eyes could eff ectively induce a PVR-like condition characterized by the sequential appearance of: )inflammatory cell infiltration, ) extracellular collagen production, and ) formation of epiretinal membranes with or without retinal folds or detachment ().Quantitative detection of the expression level of TGF-β, PDGF-AA, and PDGF-BB in this model indicates that TGF-β and PDGF-BB played an important role in the pathogenesis.But two questions arise..Which cells involved in this process?.Why the induction methods, intravitreal injection of RPE-J cells or/and PRP, induced a diff erent situation ?
In this study, in order to clarify the questions mentioned above, immunohistochemistry was performed to identify the main cell types involved in this process, and a comparative analysis was done among the different induction methods from the angle of the expression levels and durations of growth factors, TGF-β, PDGF-AA and PDGF-BB.

RPE-J cells preparation
RPE-J cells (CRL-, ATCC, Rockville, USA) were defrosted from - o C liquid nitrogen and suspended in Dulbecco's modifi ed Eagle's medium (Gibco, Grand Island, USA).Cell viability was controlled by trypanblue counting.After defrosting, an average of  of RPE-J cells was observed dead.Then, the surviving RPE-J cells were seeded at a density of ×  cells/cm  in six-well culture plates, and maintained in DMEM with ,g/L glucose,  mM L-glutamine, and .mM non-essential amino acids supplemented with  (v/v) fetal bovine serum (Gibco, Grand Island, USA) at  o C,  CO/ air (,).For subsequent passages, cells from confluent cultures were detached by , trypsin digestion and seeded as described above.Just before intravitreal injection, RPE-J cells were collected from six-well culture plates, centri-fuged at  rpm for  minutes, and resuspended in sterile pyrogen-free normal saline (NS) or platelet poor plasma (PPP) at a concentration of ×  /mL.

Preparation of PRP
Wistar rats were anesthetized by intraperitoneal injection of  chloral hydrate ( mg/kg body weight).Blood was collected into ethylene diamine tetraacetie acid vacuum tubes from the tail vein.e samples were centrifuged at g for five minutes to separate PRP from erythrocytes and leukocytes.PRP was transferred to a clean tube and centrifuged at g for  minutes to separate platelets from PPP (,).The platelet number in PRP was counted by an automatic hemocytometer and adjusted to ,×  / mL with PPP.The PRP was conserved at - o C for about  minutes until intravitreal injection.

Animal preparation
After obtaining the approval of the local ethics committee,  normal adult Wistar rats (male or female, age=- weeks, weight=-g, SLACCAS, Shanghai, China) were enrolled in this experiment.All animals were bred, maintained, and sacrifi ced humanely in strict compliance with the policies stated in the statement of Association for Research in Vision and Ophthalmology for the use of animals in ophthalmic and vision research.

Intravitreal injection of RPE-J cells and PRP
All Wistar rats were divided into four groups ( rats per group in Group , , , and  rats in Group ).Group , ,  and Group  received an intravitreal injection of NS, RPE-J cells, PRP, and RPE-J cells +PRP, respectively.Before intravitreal injection, Wistar rats were anesthetized by intraperitoneal injection of  chloral hydrate (mg/kg body weight).Surgical procedures and postsurgery care were performed as described previously ().Briefl y, eyes were gently protruded using a rubber circle and subsequently covered with , ofl oxacin eye ointment (Xingqi, Shenyang, China) to simulate a preset lens.
en, a self-sealing wound tunnel was constructed using a ,-cm, -gauge needle one mm posterior to the corneal limbus.After the vitreous cavity collapsed because of the outfl ow of vitreous fl uid, a blunt -gauge Hamilton syringe was introduced through the sclera into the vitreous cavity under a surgical microscope (SM-J, Eder, Shanghai, China), μL NS containing RPE-J cells (,×  ) or μL PRP containing platelets (×  ) and μL PPP containing RPEJ cells (,×  ) and platelets (×  ) was injected into the left eye.The right eyes served as a control eye, and were injected with μL NS.

Immunohistochemical staining
Fourteen and  days after intravitreal injection, the Wistar rats (four rats per time-point) in Group  were sacrifi ced with a fatal dose of  chloral hydrate.eir eyes were enucleated, and fi xed in  formaldehyde solution at a room temperature.ereafter, they were embedded in paraffi n, and cut into μm-thick sections.After deparaffinisation, endogenous peroxidase was abolished with  hydrogen peroxide in methanol for  min, and non-specifi c background staining was blocked by incubating the sections for  min in normal goat serum.After microwave antigen retrieval, the sections were incubated with the monoclonal mouse anti-rat antibodies listed in Table : Cytokeratin is expressed in all epithelial tissue, and cytokeratin- antibody was used to specifi cally identify proliferating RPE here.Glial fibrillary acidic protein (GFAP) Ab- directed at GFAP was used to identify glial cells.Vimentin is a marker of cytoskeletal intermediate fi laments, while vimentin Ab- was used to identify fibroblasts in this study.ED antibody is a marker for monocytes, macrophages, and some dendritic subpopulations which was used to identify macrophages here.Optimal working concentration and incubation time for the antibodies were determined earlier in pilot experiments.All the sections were incubated for  min with the biotinylated anti-mouse secondary antibody and reacted with the avidin-biotinylated peroxidase complex (mrbiotech, Shanghai, China).The reaction product was visualised by the addition of , -diaminobenzidine (mrbiotech, Shanghai, China) and hydrogen peroxide, resulting in brown immunoreactive sites.e slides were faintly counterstained with Harris haematoxylin.Finally, the sections were rinsed with distilled water and coverslipped with glycerol.Control sections were incubated with Tris-buffered saline (pH ,-,) replacing the primary antibody.A upright microscope (Zeiss Axioplan  imaging, Carl Zeiss, Germany) was used to study the stained sections.
Enzyme-linked immunosorbent assay (ELISA) ree, , , , and  days after intravitreal injection, the Wistar rats ( four rats per time-point) were sacrifi ced with an overdose of  chloral hydrate.e reti-nas were extracted and grinded into homogenate after enucleation of the globe by removing the anterior segment with eye scissors.TGF-β, PDGF-AA and PDGF-BB were measured using a capture sandwich kit with biotinylated affinity purified mouse monoclonal antibodies to rat TGF-β, PDGF-AA or PDGF-BB (Senxiong, Shanghai, China).Briefly, a flat-bottom ELISA plate (Costar -well) was coated with mouse anti-rat TGF-β, PDGF-AA or PDGF-BB antibody, μL of standard preparation (or sample) was added in the wells and incubated at  o C for two hours.After washing six times, μL of biotinylated mouse anti-rat TGF-β, PDGF-AA or PDGF-BB was added and incubated at  o C in the dark for one hour, washed, and μL of horseradish peroxidase labeled streptavidin was added and incubated at  o C for one hour.The wells were washed six times again, and incubated with μL of substrate solution for - minutes.Finally, μL of stop buff er was added to each well.Absorbance at  nm was measured using a microplate reader (Molecular Devices, Sunnyvale, USA).e samples for the detection of TGF-β needed activation using HCl and NaOH just before the conventional procedure.Sensitivity by ELI-SA was  pg/mL with an intra-assay variability of .

Statistical analysis
All data were expressed as mean±SD.Analysis of variance (ANOVA) test was used to determine the significance of the difference in a multiple comparison.The differences were considered significant at p-values less than ,.The software packages used were SPSS (version , Chicago, USA).and RPE-J cells.It was found that the expression levels of TGF-β were far higher than that of PDGF-AA and PDGF-BB, and the expression levels of PDGF-BB were significantly higher than that of PDGF-AA.

Immunohistochemical analysis of eyes  and  days after co-injection of RPE -J cells and PRP As shown in
The expression levels of TGF-β in the RPE -J cells and PRP co-injection group were signifi cantly higher than that of NS, PRP or RPE-J cells injection groups at ,, , and  days (about ,-, times).In this co-injection group, the peak of expression level of TGF-β appeared at  days (,±,pg/ML) and there was a signifi cant decrease at  days.In PRP injection group, the expression level of TGF-β presented a downward trend as a whole, although there was a transient increase at  days.As for the expression level of TGF-β of NS and RPE-J cells injection group, they were little changed with time (Figure A and B).
Interestingly, there was no significant difference of the expression levels of PDGF-AA among the different treatment groups, they all gradually decreased from  days to  days ( Figure C and D).
The expression level of PDGF-BB was another indicator of concern.At , , , and  days after intravitreal injection, the expression levels of PDGF-BB of the RPE-J cells and PRP co-injection group were also significantly higher than that of NS, PRP or RPE-J cells injection groups (about ,-, times).As TGF-β, a similar trend was seen in PDGF-BB.It decreased at  days after achieving its peak at  days, while in the NS, PRP or RPE-J cells injection groups, the trends of PDGF-BB were overall decline, too.

Discussion
The present study indicates that RPE cells, glial cells, fibroblasts and macrophages took part in the pathogenesis of this rat model of PVR.Meanwhile, the expression levels and durations of TGF-β and PDGF-BB partially explained the different results induced by the different induction meth-ods, intravitreal injection of RPE-J cells or/and PRP.
Our clinical and histopathologic observations show that intravitreal co-injection of PRP and RPE-J cells in Wistar rats can successfully induced a PVR-like condition, but the cellular and molecular mechanisims underlying this process were unknown.It is known that RPE cells, fibrous astrocytes, fibroblasts, myofibroblasts and macrophages are the principal cells involved in periretinal proliferation (,).en, are these cells involved in this PVR-like condition, too?Our results of immunohistochemistry showed that RPE cells, glial cells, fibroblasts and macrophages are all the contributors to this pathogenesis, which is more relevant to the human disease, especially human PVR associated with intraocular bleeding.Furthermore, these cells appeared with certain regularity: gradually increased for RPE cells, first increased and then decreased for other cells.is provides an experimental proof for attenuation of the experimental PVR by targeting at a specific cells at a particular time point.
Our previous study also indicates that PRP alone could induce a PVR-like condition characterized by low ratio, slow advancement, and low severity, while RPE-J cells alone could never induce a Wistar rat model of PVR.How to explain this phenomenon?We hypothesize that the diff erent expression levels and durations of growth factors played a key role, because PRP has been shown to be rich in growth factors, which are able to enhance endothelial cell migration and proliferation (, ).PRP and RPE-J cells co-injection may have synergistic eff ects. is view was confi rmed by the fact that the expression levels of TGF-β and PDGF-BB in PRP and RPE-J cells co-injection group were higher than that of other groups, and their durations were also longer, which ensured higher rates of adult model.However, the low expression levels and short durations of TGF-β and PDGF-BB in PRP or RPE-J cells injection group resulted in an opposite outcome.
Growth factors, such as TGF-β, PDGF, hepatocyte growth factor, basic fibroblast growth factor, or interleukin-, are believed to play an important role in promoting the events that contribute to PVR(-).Moreover, there is a cooperative interaction between TGF-β and PDGF.The increase of growth factors favor the recruitment of effector cells, such as RPE cells, glial cells, macrophages, and fibroblast, so as to promote the inflammatory reaction during the first phases of PVR. is is followed by a proliferative process (,).In such a process, PDGF-AA seemes not to play any role.is is because PDGF-AA is selective more than PDGF-BB, which binds PDGF β receptor only.e expression of PDGF β receptor in this model may be low, and further characterization is required to do.Nevertheless, this maybe provides another experimental proof for attenuation of the experimental PVR by targeting at a specifi c growth factor or its receptor.

Conclusion
RPE cells, glial cells, fi broblasts and macrophages took part in the pathogenesis of this rat model of PVR. e diff erences of expression levels and durations of TGF-β and PDGF-BB partially explain the diff erent results induced by the diff erent induction methods, intravitreal injection of RPE-J cells or/and PRP.Intravitreal co-injection of RPE-J cells and PRP in Wistar rat eyes could eff ectively induce a model of PVR which off ers a foundation for the study of the pathogenesis and therapies for PVR.

List of Abbreviations
Figure ., there was no staining in the negative control slides (Figure A).At  days, a small amount of RPE cells expressing cytokeratin- migrated from pigment epithelium layer to the ganglion cell layer, and began to proliferate (Figure B).At  days, numerous proliferating RPE had distributed in the detached retina and proliferative membrane in vitreous cavity (Figure C).At  days, there were a number of glial cells expressing GFAP distributing in the epiretinal membrance, and they significantly reduced at  days (Figure D and E).e same tendency was seen in the fibroblasts which express the vimentin.At  days, a great quantity of fi broblasts distributed at the retinal surface and most of them had established con-Monoclonal antibodies used in this study nections with the inner retina (Figure F), however, at  days, only a few of fi broblasts still distributed at the retinal surface, and the connections with the inner retina was seldom seen (Figure G).In addition, a lot of ED-positive cells, i.e., macrophages, distributed in the epiretinal membrance at  days(Figure H), and they were still present at the retinal surface at  day, despite the small quantity (Figure I).The expression levels and durations of TGF-β, PDGF-AA or PDGF-BB in diff erent treatment groups Figure .shows the growth factors, TGF-β, PDGF-AA and PDGF-BB expressed in the eyes of Wistar rats received intravitreal injection of NS, PRP or/