Retinoblastoma ( Rb ) promotes senescence of corneal endothelial cells by inhibiting the activation of E2F2

Background Senescence-like changes occur in aging Corneal endothelial cells (CECs), and these changes are associated with decreased vision and age-related corneal diseases. such as Fuchs endothelial dystrophy (FED), chronic corneal allograft dysfunction (CCAD). Such changes have also been shown to cultured cell in vitro after passaging. Therefore, studying the mechanism of CEC senescence would aid in the development of anti-senescence treatment, which would benet FED and CCAD. The tumor suppressor retinoblastoma (RB) gene product pRB triggers senescent growth arrest when inactivated. In this study, we used siRNA treatment to evaluate whether RB knockdown could suppress CEC senescence, and we investigated relevant molecular mechanisms. Methods RCECs were obtained and were cultured with or without siRb. Senescent cells were detected with a β-galactosidase senescence staining kit. The gene p21, which is associated with a senescent phenotype, was measured by RT-PCR. The morphology and migration of cultured RCECs were examined by phase-contrast microscopy. ZO-1 and N-Cadherin, which are involved in pump and barrier functions, were assessed by immunouorescence. Cell cycle assessment was performed using a ow cytometer (BD FACSCalibur).


Background
Corneal endothelial cells (CECs) form a single layer situated in the back of the cornea. [1,2] These cells have a pivotal function in maintaining corneal transparency by performing pump and barrier functions. [3] Because of cell-cell contacts and a lack of relevant growth factors, CECs are arrested in the G1 phase, and their proliferation ceases. [4] The relative number of CECs is approximately 3500-4000 cells/mm 2 at birth, and the number gradually decreases with age or following surgical trauma. [5] When the cell density falls to 500-1000 cells/mm 2 , loss of corneal endothelial function results in visual loss. [6,7] Cell aging is a special form of persistent cell cycle arrest, and it is regulated by a number of biological events. [8] Senescence-like changes take place in aging CECs and are associated with decreased vision and age-related corneal diseases. [2] Fuchs endothelial dystrophy (FED) [9] manifests as corneal epithelial edema and blindness, and it results in upregulated expression of senescence-related proteins. [10,11] Chronic corneal allograft dysfunction (CCAD) [9] has become the main problem of clinical penetrating keratoplasty, and it leads to corneal allograft failure. Previous results have indicated that cellular senescence is associated with the pathogenesis of chronic graft failure. [12,13] Tissue engineering to produce corneas has become a popular research area in recent years, but the availability of seed cells has greatly limited this development. [14][15][16] Several research groups have successfully cultured corneal endothelial cells. However, maintaining cultured CECs with the correct morphology and function is another problem. Moreover, as the cells are passaged, the cell density decreases, which was found to be associated with cellular senescence. [17][18][19] Therefore, analyzing the aging mechanism of CECs will aid in the development of anti-senescence treatments that will bene t FED and CCAD suppresses cellular senescence.
Cellular senescence is a special, irreversible state of cell cycle arrest. It is characterized by a series of features, including upregulated expression of aging-related proteins (e.g., p16 INK4a and p21), and it exhibits a "senescence-associated secretory phenotype" (SASP), which includes the release of proin ammatory cytokines and chemokines. [20][21][22]

Cell culture
All rabbits were placed in restraint devices under general anesthesia, followed by a lethal ear venous injection with 100 mg/kg of sodic pentobarbital. the eyeballs were excised and incubated in D-PBS containing 2% penicillin-streptomycin. Each rabbit corneal endothelium with Descemet's membrane was dissected after the cornea was separated under a stereoscopic dissecting light microscope. Corneas were washed three times with D-PBS containing 1% penicillin-streptomycin. The corneal endothelium, which was attached to Descemet's membrane, was incubated overnight at 37 °C in DMEM containing 10% fetal bovine serum (FBS; Gibco-Life Technologies) and 1% penicillin-streptomycin (Gibco, Grand Island, NY, USA). The next day, the tissue was washed with D-PBS and digested at 37 °C for approximately [10][11][12][13][14][15] minutes in a Trypsin-EDTA solution. The resulting RCECs were suspended and cultured in DMEM containing 10% FBS and 1% penicillin-streptomycin for 5 to 7 days until they were con uent, and then the cells were passaged at a 1:3 ratio using Trypsin-EDTA solution.

Cell cycle analysis by propidium iodide (PI) staining
Cells transfected with siRNAs were maintained for 48 hours, and then they were suspended and xed with 70% ethyl alcohol overnight at -20 °C or at 4 °C for 1 hour for further experiments. Brie y, cells were centrifuged and washed with PBS. Subsequently, the cells were resuspended in PBS with propidium iodide (PI; 50 µg/ml, Sigma) and RNaseA (100 µg/ml, Beyotime), and then they were treated for 20 minutes at room temperature. Cell cycle analysis was performed immediately thereafter using a ow cytometer (BD FACSCalibur).

Real-time polymerase chain reaction (RT-PCR)
RT-PCR was performed as previously described. [29] In brief, total RNA was extracted by TRIzol reagent (Invitrogen, #15596026)from each sample 4 days after siRNA transfection, and cDNA was obtained using a PrimeScript™ RT reagent kit (Takara, #RR037B) according to the manufacturer's protocol. All RT-PCR was performed with SYBR GREEN reagents with a CFX96 Real-Time System (BIO-RAD, CFX96 optics module). The primer sequences used for RT-PCR analysis in our study are shown in Table 1. The resulting expression values were analyzed using the 2 −ΔΔCT method.

β-Galactosidase Senescence Staining
Senescent cells were detected with a cell senescence β-galactosidase staining kit (Beyotime, C0602) according to the protocol provided by the manufacturer. First, each cell sample was washed with PBS once, and then it was xed with β-galactosidase xing solution buffer for 10 minutes. After that, the cells were washed three times with PBS for 3 minutes and incubated at 37 °C overnight with a staining buffer containing X-gal. Samples were visualized the next day using a uorescence inverted microscope (Olympus, IX73).

Immuno uorescent Staining
Immuno uorescent staining was used to assess the expression of ZO-1 and N-cadherin, which are involved in pump and barrier functions. Brie y, con uent RCECs were xed with a 4% paraformaldehyde solution, which was followed by blocking with 1% bovine serum albumin (BSA) in PBS for 1 hour; then, cells were permeabilized with 0.2% Triton X-100 in PBS for 10 minutes.

Statistical Analysis
We used Prism 6 software for statistical analysis. All results are shown as the means ± SD, and all data were analyzed via Student's t-test. P values < 0.05 were considered statistically signi cant.

Corneal endothelial cell senescence and senescenceassociated secretion of factors throughout the passaging of cells
We initially compared the size and morphology of rabbit corneal endothelial cells at passages 1-5. As the passage number increased, the RCECs gradually became larger, the cell morphology gradually shifted from being hexagon to irregular, and the cell density showed a decline ( Fig. 1a and b). The senescence marker SA-β-Gal, which was detected with a senescence detection kit, was used to evaluate the senescence of RCECs; the percentage of SA-β-gal-positive RCECs increased with cell passage (Fig. 1c and  d). Furthermore, RT-PCR experiments were performed to measure the expression of the senescencerelated gene p21 and to measure secreted factors associated with SASP. Compared with P1, the expression of the senescence-related gene p21 increased at the mRNA level (Fig. 1e). Moreover, the expression of 7 secreted factors (CCL2, CXCL10, TNF, IL-1α, IL-1β, IL-6, and IL-8) was signi cantly upregulated in P3 (Fig. 1f). These data show that corneal endothelial cells gradually senesce throughout culture.

Effect of siRB on the senescence of RCECs
To investigate the effect of RB on RCEC senescence, we used siRNAs to knock down RB in P3 cells, which is a passage of cells that exhibits qualities of cellular senescence but is in a good cellular state. RB mRNA levels were downregulated following siRNA treatment, as shown by graphs, and two RB siRNAs signi cantly reduced the level of RB compared with the negative control (Fig. 2a). SA-β-Gal staining was next used to assess corneal endothelium senescence, and our results showed that the percentage of SAβ-Gal-positive RCECs was reduced upon RB knockdown ( Fig. 2b and c). A RT-PCR assay was used to measure the expression level of the senescence-related gene p21. Our results showed that the expression of the senescence-related gene p21 was reduced at the transcription level in the sample with siRb ( Fig. 2d), and the expression of 5 secreted factors (CCL2, IL-1α, IL-1, IL-6, and IL-8) was also signi cantly decreased (Fig. 2e). These results suggested that siRb inhibits RCEC senescence.

Effect of siRB on cell proliferation and functional phenotype of RCECs
First, we wanted to know whether reducing RB might result in cell proliferation in RCECs. Cell cycle results proved that the numbers of cells at the G2/M phase and at the S phase were signi cantly increased following knockdown of RB and 72 subsequent hours of culture (Fig. 3a). The potential effect of RB on wound restoration was evaluated by a migration assay, and the results showed that the wound restoration capacity of cells following RB knockdown was also not signi cantly different from that of NC ( Fig. 3b and c). N-Cadherin and ZO-1, which play roles in pump and barrier functions, were assessed by immuno uorescence, and their expression was the same as that of the control (Fig. 3d). These results suggest that siRb did not in uence cell proliferation or the functional phenotype of RCECs.

Rb may induce RCEC senescence by inhibiting E2F2 activation
RB promotes cellular senescence; thus, we tested whether knocking down RB was associated with the activation of E2F. The tumor suppressor RB gene product pRB triggers senescent growth arrest when inactivated. When inactivated, E2F is tightly bound to pRB and prevents activation of E2F and unregulated DNA synthesis. RT-PCR experiments were employed to measure the expression of E2F2 and E2F4, and the results showed that knocking down RB upregulated E2F2 (Fig. 4a), suggesting that RB may promote the senescence of corneal endothelial cells by inhibiting the activation of E2F2.

Discussion
As experimental and clinical evidence has shown, Fuchs endothelial dystrophy (FED), chronic corneal allograft failure, and cell density decreases when cells are cultured; in other words, cells exhibit characteristics of senescence. [9,12,30] The characteristics of cellular senescence include a senescencelike morphology with enlarged cell bodies, an irregular shape, and an increased proportion of multinucleated cells. Our team tried to expand HCECs in vitro under standard culture conditions, but it was di cult to maintain the morphology and function after 2-3 passages. After that, our research group attempted to apply conditioned medium to promote cell proliferation and maintain morphology in RCECs.
Treatment with conditioned medium resulted in an increase in cell proliferation, and cells exhibited hexagonal morphology; further, cell pump function evaluation revealed complete functionality. However, there was no signi cant improvement in human corneal endothelial cells. The greatest challenge is still cells losing density after several cell passages because of cellular senescence. However, the underlying mechanisms of senescence in CE cells are largely unclear. Cellular senescence is identi ed by positive SA-β-gal staining, high expression of cyclin-dependent kinase inhibitors such as p16 INK4a , p21 CIP1 , p27 KIP1 or p53 14 , enlarged cell size, and the acquisition of an associated secretory phenotype (SASP).
Cellular senescence can be stimulated by inducers, such as H 2 O 2 and TGF-β. [31,32] It can also be produced by several days of culture. [9] In our study, rabbit corneal endothelial cells were cultured in standard medium for 5 to 7 days and then were passaged at a 1:3 ratio using a trypsin-EDTA solution. As the passage number increased, RCECs gradually became larger, the cell morphology gradually shifted to an irregular shape from a round shape, the cell density declined, and the percentage of SA-β-gal-positive cells was upregulated as the cell passage increased. Furthermore, compared with P1, the expression of 7 cytokines and chemokines (CCL2, CXCL10, TNF, IL-1α, IL-1β, IL-6, and IL-8) was signi cantly upregulated in P3. Moreover, the expression of the senescence-related gene p21 also increased at the mRNA level. We used siRNAs to knock down RB, and our results showed weak positive SA-β-gal staining and low expression of p21 CIP1 following RB knockdown. We also found that cell size was smaller than that of the NC cells. Cellular senescence is a special, irreversible state of cycle arrest, and knocking down RB signi cantly increased the cell cycle but had no in uence on the migratory ability or functional phenotype.
In addition, the RB gene product is pRB, which is a tumor suppressor gene. [26] When inactivated, E2F is tightly bound to pRB, which prevents activation of E2F, disrupts regulation of DNA synthesis, and triggers senescent growth arrest. [1] Therefore, we assessed the expression levels of E2F2 and E2F4 after knocking down RB and found that E2F2 expression was signi cantly upregulated at the mRNA level, which explained the effect of Rb knockdown on the cell cycle and cell senescence.
Certainly, there are many limitations in our research. First, because human corneal endothelial cells are di cult to obtain and culture, all the studies in this paper were on rabbit corneal endothelial cells. In addition, as the proteins studied in our study used rabbit antigens, it was di cult to use antibodies to assess the corresponding protein levels with WB; therefore, the results in this paper are mostly limited to the mRNA level. Furthermore, the mechanism of the effect of Rb on cell senescence still needs further study, but our work provides a theoretical basis of an anti-senescent treatment for corneal endothelial cells.

Conclusions
We demonstrated that as the cells increased in passage number, the number of senescent RCECs, the levels of the senescence-related genes p21, and the levels of senescence-associated secretory factors increased. Retinoblastoma (Rb) promoted the senescence of corneal endothelial cells by inhibiting the activation of E2F2.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.  qPCR measured the level of p21 and the SASP factors. (*P < 0.05, ***P < 0.001, and ****P < 0.0001; n =3).  Rb may induce RCEC senescence by inhibiting E2F2 activation. (A) RT -PCR was used to measure the level of E2F. (***P < 0.001; n =3).