Overexpression of CDKN1B Inhibits Fibroblast Proliferation in a Rabbit Model of Experimental Glaucoma Filtration Surgery

Home , Apoptosis, CDKN1B, Cyclin D, Cyclin D1, Cyclin E, G1 phase

Glaucoma Overexpression of CDKN1B Inhibits Fibroblast Proliferation in a Rabbit Model of Experimental Glaucoma Filtration Surgery

Jian-gang Yang,1 Ying Deng,1 Ling-xiao Zhou,1 Xiao-yan Li,1 Peng-rui Sun,1 and Nai-xue Sun2

PURPOSE. To investigate the potential antiproliferative effect of scarring of the conjunctiva at the bleb and sclerostomy sites cyclin-dependent kinase inhibitor 1B (CDKN1B) overexpres- after GFS often leads to the failure of the surgery.1–4 It has been sion in a rabbit model of glaucoma filtration surgery (GFS). proved that the proliferation of fibroblasts from the subconjunctival space is the major cause of the scarring process.5,6 METHODS. The recombinant adenovector expressing exogenous CDKN1B was delivered to Tenon’s capsule by subconjunctival Both positive and negative regulators of cell-cycle activity injection during unilateral filtration surgery. The time course of are the critical regulators of cell proliferation and differenti- CDKN1B expression was monitored by immunohistochemistry ation. The positive regulators include cyclins and cyclin- dependent kinases (Cdks), whereas the negative regulators and Western blot analysis. Evaluation of proliferating activity 7–9 was performed by proliferating cell nuclear antigen (PCNA), consisted in cyclin-dependent kinase inhibitors (CKIs). Cyclin-dependent kinase inhibitor 1B (CDKN1B, also known argyrophilic nucleolar organizing region (AgNOR) staining, and Kip1 fibroblast-specific protein 1 (FSP-1). Cyclin-dependent kinase 2 as p27 or p27) is a well-characterized CKI that belongs to (Cdk2) and Cdk4 expression were detected with immunohis- the Cip/Kip family, which is an important negative modulator tochemical analysis. of cell-cycle progression. The protein level of CDKN1B is highest in quiescent cells and declines as cells are stimulated RESULTS. The overexpression of CDKN1B in Tenon’s capsule to reenter the cell cycle.10,11 Once activated, CDKN1B protein was monitored throughout the experimental period. Immuno- regulates cell-cycle progression that specifically inhibits cyclin reactivity to CDKN1B was mainly observed in the nucleus of E/Cdk2, cyclin A/Cdk2, and cyclin D/Cdk4/6 complexes, fibroblasts. The increased expression of CDKN1B in sclera was which are necessary for DNA replication, and thereby detected up to 21 days after viral infection, whereas the level abrogates their catalytic activity, leading to potent arrest at of CDKN1B protein in corneal stroma was not significantly 12–14 the G1/S-phase transition. In S-phase, CDKN1B is phos- increased. The overexpression of CDKN1B induced a signifi- phorylated at Thr-187 by cyclin E/Cdk2 and then ubiquitylated cant decrease in AgNOR number/nucleus and area/nucleus, in proliferating cells by SCFSkp2 (S-phase kinase-associated PCNA staining, FSP-1 positive cells, and the decreased protein 2), relieving the blockade of cell-cycle progres- expressions of Cdk2 and Cdk4, as evidenced by nuclear and sion.15,16 cytoplasmic immunoreactivity to Cdk2 and Cdk4 antibodies in CDKN1B, as one of the most widely distributed CKIs in positive fibroblasts. most human tissues, is expressed both in proliferating and 17 CONCLUSIONS. The persistent overexpression of CDKN1B differentiated cells. In our previous study, we performed a mediated by the recombinant adenovector expressing exoge- rabbit model of experimental GFS that locally applied nous CDKN1B in Tenon’s fibroblasts after GFS may lead to the adenovirus-mediated CDKN1B (Ad-CDKN1B). Ad-CDKN1B inhibition of fibroblast proliferation and the downregulation of could induce an ocular hypotensive response and increase Cdk2 and Cdk4 activity, thereby reducing the severity of scar bleb survival during a 28-day period. However, Ad-CDKN1B did formation and the surgical outcome. (Invest Ophthalmol Vis not result in a variety of toxicities induced by mitomycin C, Sci. 2013;54:343–352) DOI:10.1167/iovs.12-10176 including wound leakage, corneal erosion, and chronic hypotony. The blebs displayed histologic features of marked reduction in total cellularity, consisted of the decreased laucoma filtration surgery (GFS) is the most widely used number of conjunctival epithelial cells and goblet cells, and Gprocedure for maintaining a suitable target intraocular the noted reduction in subconjunctival scar tissue following pressure (IOP) for glaucoma patients. However, excessive Ad-CDKN1B delivery. Ultrastructurally, Tenon’s fibroblasts showed the reduced activation and increased apoptosis at the surgical sites. We speculate that the prolonged wound From the 1Department of Ophthalmology, Affiliated Hospital, healing is associated with the inhibition of subconjunctival Xi’anMedicalUniversity,Xi’an,China;andthe2Department of fibroblast proliferation. Recent studies have suggested a Ophthalmology, Second Affiliated Hospital of Medical College, Xi’an physiologic role of CDKN1B as a regulator of fibroblast growth Jiaotong University, Xi’an, China. in the pathogenesis of proliferative diseases. In vitro experi- Supported by the Key Science and Technology Program of ments have shown that overexpression of CDKN1B efficiently Shanxi Province, China Grant 2009K17-02, and Xi’an Science and blocked fibroblast activity in cultured Tenon’s fibroblasts and Technology Program, China Grant HM1116(5). limited fibroblast proliferation.18 Reduced Cdk2 activity and Submitted for publication May 10, 2012; revised October 1 and the decline in cell proliferation that took place at late time December 3, 2012; accepted December 4, 2012. points after angioplasty correlated with a marked induction of Disclosure: J.-G. Yang,None;Y. Deng,None;L.-X. Zhou, 19–21 None; X.-Y. Li,None;P.-R. Sun,None;N.-X. Sun,None CDKN1B. The nuclear accumulation of CKI was associat- 22,23 Corresponding author: Jian-gang Yang, Department of Ophthal- ed with a quiescent and static phenotype. mology, Affiliated Hospital, Xi’an Medical University, 710077, Xi’an, Thus, we hypothesize that CDKN1B plays an important role China; [email protected]. in modulating Tenon’s fibroblast proliferation after GFS. In the

Downloaded from jov.arvojournals.org on 10/01/2021 344 Yang et al. IOVS, January 2013, Vol. 54, No. 1

present study, we used a rabbit model of GFS to analyze the Immunohistochemical Analysis ectopic expression of CDKN1B and attempted to elucidate the potential mechanism of antiproliferative effects from CDKN1B. For immunohistochemical analysis, 5-lm sections were deparaffinized and treated by two rounds of microwave heating for 5 minutes in 10 lM sodium citrate (pH 6). Endogenous peroxidase was blocked by METHODS incubation with 0.3% H2O2. Slides were blocked in 10% normal goat serum in PBS for 30 minutes and then stained for 1 hour with an anti- Construction of Recombinant Adenovirus Vector human proliferating cell nuclear antigen (PCNA) monoclonal antibody, rabbit polyclonal anti–fibroblast-specific protein-1 (FSP-1)/S100A4 Construction of the recombinant adenovirus vectors has been (Dako, Glostrup, Denmark), anti-human CDKN1B monoclonal antibody described in detail elsewhere.17,24 Essentially, full-length human (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), anti-human Cdk2 CDKN1B cDNA was cloned into a shuttle vector containing the human monoclonal antibody, and anti-human Cdk4 monoclonal antibody cytomegalovirus immediate-early promoter and a plasmid that contains (Sigma, St. Louis, MO), respectively. They were subsequently incubated replication adenovirus genome. Adenoviruses were grown and with a biotinylated goat anti-mouse secondary reagent (Sigma) and with propagated in the HEK-293 cell line. The presence of CDKN1B cDNA horseradish peroxidase–labeled streptavidin (Zhongshan Biotechnolo- in the viral genome was further purified using standard protocols. Viral gy, Beijing, China). Bound antibodies were visualized with 0.02% 3,30- particle number of purified recombinant adenoviral vectors was diaminobenzidine tetrahydrochloride. The sections were counter- determined using HPLC quantification and reported as plaque-forming stained with hematoxylin. units (pfu), which was 2.1 3 1012 pfu/mL. Positive immunostaining was defined as nuclear or cytoplasmic staining of Tenon’s cells. The percentages of the positive cells in total CDKN1B Delivery for GFS fibroblasts were calculated by examining four random high-power fields of each slide.28 In addition, immunoreactions were scored on A randomized, controlled, masked-observer study was conducted in a each tissue slide for protein expression based on a semiquantitatively rabbit model of GFS. Adult albino rabbits weighing between 2 and 3 kg scoring system that measured percentage positive fibroblasts (0, none; were anesthetized with an intramuscular injection of ketamine (35–45 1, <10%; 2, 10%–50%; 3, 50%–80%; 4, >80%) and cellular localization mg/kg) and xylazine (5–10 mg/kg). Surgical eyes underwent a (cytoplasm, nucleus, or combination thereof).29 unilateral sclerectomy using a limbal-based triangle scleral flap with a peripheral iridectomy approach, the details of which are described Western Blot Analysis elsewhere.17 17 A total of 96 rabbits were randomly divided into Ad-CDKN1B and Western blotting was carried out as previously described. In brief, the control groups: 48 were used for Western blot analysis (six in each tissues at surgical sites including Tenon’s capsule, corneal stroma, and group) and 48 for morphometric analysis (six in each group). The Ad- superficial scleral stroma were peeled carefully after enucleating, and the conjunctival and corneal epithelium were separated completely CDKN1B group was delivered as 8.0 3 1011 pfu/mL (100 lL volume) from Tenon’s capsule and corneal stroma, respectively. The samples Ad-CDKN1B into Tenon’s capsule by a subconjunctival injection at the were collected by scraping with a spatula in 10 mL of cold medium superolateral quadrant during filtration surgery. Injection of PBS served containing 50 mM 3-(N-morpholino) propanesulfonic acid (pH 7.0), 10 as a control. After surgery, 1% atropine sulfate ophthalmic ointment, mM NaF, 1 mM EDTA, 0.3 M sucrose, 0.4 mM Pefabloc, 102 g/L neomycin sulfate, and dexamethasone were applied to the eye. aprotinin, 2 3 103 g/L leupeptin, and 103 g/L pepstatin A. All procedures in these animal experiments were performed in After centrifugation at 2500g for 2 minutes to remove debris, an accordance with the ARVO Statement for the Use of Animals in equal protein amount of cell lysates was separated on SDS-PAGE and Ophthalmic and Vision Research, and approved by the Committee for transferred onto nitrocellulose membranes. The membranes were Animal Research, Xi’an Medical University. blocked in an ice-cold lysis bufer (10 mM Tris-Cl, 1% SDS, and 1 mM

Na2VO4, pH 7.4). The total protein concentration was determined by Argyrophilic Nucleolar Organizing Region the bicinchoninic acid assay method (Pierce, Rockford, IL). The blots Staining were incubated with rabbit monoclonal antibodies to CDKN1B (Santa Cruz Biotechnology, Inc.) for 2 hours at room temperature. b-Actin Argyrophilic nucleolar organizing region (AgNOR) histochemical (Sigma) served as an internal positive control. Reactions were staining was performed using a previously described modified silver- visualized with a suitable secondary antibody conjugated with staining technique.25 In brief, the eyes were enucleated together with horseradish peroxidase using enhanced chemiluminescence reagents. the conjunctiva to preserve the bleb on days 7, 14, 21, and 28 after surgery. Formalin-fixed, paraffin-embedded tissue sections were cut at Statistical Analysis 5 lm, deparaffinized in xylene, rehydrated in graded ethanol, and rinsed in distilled deionized water. Slides were incubated for 25 to 30 Statistical analysis was carried out using commercial statistical software minutes at room temperature in the dark with freshly made AgNOR (SPSS Statistical Program, version 13.0 for Windows; SPSS, Inc., staining solution, consisting of 0.02 g gelatin in 1 mL 1% formic acid Chicago, IL). Group differences were evaluated by one-way ANOVA and 1 g silver nitrate in 2 mL of distilled water. Following AgNOR or Student’s t-test. The independent samples nonparametric test was staining, slides were rinsed with distilled deionized water, dehydrated used to analyze scores. A value of P < 0.05 indicates a statistically with graded ethanol, cleared with xylene, and mounted in synthetic significant difference. medium. Morphometric analysis and quantification of AgNORs were performed by microscopy at 3400 magnification with an image RESULTS analysis system (KS400; Carl Zeiss Microscopy, Jena, Germany). On CDKN1B Protein Expression at the Surgical Sites each slide, AgNORs were counted in 100 randomly selected nuclei of fibroblasts. Mean nucleus area, AgNOR area/cell, AgNOR number/cell, Immunohistochemstry and Western blot analysis were used to and AgNOR ratio (AgNOR area/cell divided by the nucleus area) were observe fibroblast proliferation at surgical sites following Ad- then determined on the basis of averaging the counts within these 100 CDKN1B delivery. Figures 1A and 1B showed the time-course cells as shown by previous reports.26,27 Slides were examined by two data of CDKN1B expression in Tenon’s capsule by Western independent evaluators. blot analysis. The CDKN1B level was higher obtained by Ad-

Downloaded from jov.arvojournals.org on 10/01/2021 IOVS, January 2013, Vol. 54, No. 1 When CDKN1B Inhibits Fibroblast Proliferation 345

FIGURE 1. Overexpression of CDKN1B in Tenon’s capsule. The elevated levels of CDKN1B protein are detected using Western blot throughout the experimental period, which the highest protein level is observed on day 21 (A, B). Immunohistochemistry analysis further conﬁrms the overexpression of CDKN1B, which is localized in the nucleus of Tenon’s ﬁbroblasts (arrows) (Ad-CDKN1B group, C–F; control group, G–J; 3200). There are continuously high scores during the experimental period (K). Means 6 SEM (n ¼ 6). *P < 0.05, **P < 0.01. C, conjunctiva; TC, Tenon’s capsule.

CDKN1B application than that by placebo on day 7 after To better understand the roles of Ad-CDKN1B at surgical surgery (P ¼ 0.001). A further increase of CDKN1B expression sites, we also monitored the time course of CDKN1B levels was recorded on day 14 (P ¼ 0.000). The elevated expression in corneal stroma and superficial scleral stroma CDKN1B protein in Ad-CDKN1B–treated eyes was still (Fig. 2). In scleral stroma, CDKN1B overexpression following detectable after 28 days (P ¼ 0.001). Ad-CDKN1B delivery was detected in comparison with Immunohistochemical staining for CDKN1B in cells was placebo on day 7 (P ¼ 0.016). Furthermore, the elevated scored semiquantitatively based on the number of CDKN1B- protein level was maintained up to day 21 (P ¼ 0.000). positive fibroblasts. In the present study, immunoreactivity to However, no difference of CDKN1B expression induced by Ad- CDKN1B was mainly observed in the nucleus of fibroblasts. CDKN1B and placebo was detected on day 28 (P ¼ 0.781), The scores were significantly elevated on day 7, in which the although the densitometric signal remained relatively high by mean percentage of positive cells with Ad-CDKN1B was Western blot (Figs. 2A, 2C). The immunoreactivity to CDKN1B 85.08% vs. 39.53% of placebo (mean score of immunoactivity, was mainly observed in the nucleus. As expected, the CDKN1B 3.78 vs. 2.11 P ¼ 0.000) (Figs. 1C, 1G, 1K). CDKN1B level induced by Ad-CDKN1B was further increased on day 14. The expression using immunohistochemistry was consistent with mean percentage of positive cells was 76.19% vs. 21.05% of Western blot during the 28-day period (Figs. 2G–J). placebo (mean score, 3.56 vs. 1.83, P ¼ 0.000) (Figs. 1D, 1H, In corneal stroma, CDKN1B expressions were similar by Ad- 1K). The score remained elevated up to day 28, in which the CDKN1B delivery and placebo throughout the experimental mean percentage of Ad-CDKN1B was 72.66% vs. 34.92% of period (Figs. 2B, 2D) using both Western blot and immunohis- placebo (mean score, 3.05 vs. 2.05, P ¼ 0.000) (Figs. 1F, 1I–K). tochemstry. The immunohistochemical staining in fibroblasts

Downloaded from jov.arvojournals.org on 10/01/2021 346 Yang et al. IOVS, January 2013, Vol. 54, No. 1

FIGURE 2. CDKN1B expression in corneal stroma and superficial scleral stroma of surgical sites. There is no increase in protein expression in corneal stroma throughout the experimental period (A, C). The immunostaining in fibroblasts is not observed either in nucleus or cytoplasm (Ad- CDKN1B, E; control, F). CDKN1B overexpression is detected in superficial scleral stroma up to day 21 (B, D). The immunostaining is localized in the nucleus of positive cells (Ad-CDKN1B, G, I; control, H, J; 2003). *P < 0.05, **P < 0.01.

was consistently almost absent in nucleus and cytoplasm (Figs. sites. Histologic profiling revealed mild or moderate fibrotic 2E, 2F). response with extracellular matrix (ECM) deposition in Tenon’s capsule following Ad-CDKN1B delivery (Figs. 3A–D). Argyrophilic Nucleolar Organizing Region The sclerotomy site, in contrast, showed massive subconjunc- Histologic Evaluation of Wound Healing tival scarring, which consisted of dense ECM in placebo (Figs. 3E–H). AgNORs are nucleolar substructures that are associated with The quantitative evaluation of AgNORs was carried out by ribosomal RNA transcription. It has been demonstrated that morphometric analysis, by counting the number occupied by AgNOR count is a good marker of proliferation. In the present the silver-stained structures. On day 7, the mean AgNOR study, the tissues were stained with AgNOR to give an overall number/nucleus was significantly decreased following Ad- impression to determine proliferative fibroblasts at surgical CDKN1B delivery compared with placebo (95% confidence

Downloaded from jov.arvojournals.org on 10/01/2021 IOVS, January 2013, Vol. 54, No. 1 When CDKN1B Inhibits Fibroblast Proliferation 347

FIGURE 3. Histochemical staining for AgNOR at surgical sites. Histologic features show less proliferation (A–D, 3100) and less AgNOR number of cells (3400) in Tenon’s capsule following Ad-CDKN1B delivery in comparison with placebo (E–H). S, sclera.

interval [CI] for Ad-CDKN1B, 1.93–2.84; for placebo, 3.53–4.58 PCNA expression revealed a decrease of PCNA accumula- P ¼ 0.000). The decreased number/nucleus was maintained tion on day 7 after Ad-CDKN1B delivery (41.18% of mean throughout the experimental period, and on day 28, the PCNA-positive cells of Ad-CDKN1B vs. 83.33% of placebo; difference was still significant (95% CI for Ad-CDKN1B, 1.95– mean score, 2.17 vs. 3.50, P ¼ 0.000) (Figs. 4A, 4E, 4I). To day 3.15; for placebo, 2.98–4.16; P ¼ 0.000) (Table). 14, the decrease was still significant (52.38% of Ad-CDKN1B vs. Comparison of the nuclear area of AgNOR between Ad- 87.10% of placebo; mean score, 2.72 vs. 3.72, P ¼ 0.000) (Figs. CDKN1B and placebo revealed that there was a significant 4B, 4F, 4I), and the decreased expression was maintained up to decrease with Ad-CDKN1B delivery on day 7 (95% CI for Ad- day 28 (32.26% of Ad-CDKN1B vs. 55.36% of placebo; mean score, 2.00 vs. 2.72, P ¼ 0.000) (Figs. 4D, 4H, 4I). CDKN1B, 0.77–1.10 lm2; for placebo, 1.13–1.59 lm2; P ¼ 0.004). The decrease was maintained for 28 days (95% CI for Ad-CDKN1B, 0.64–0.93 lm2; for placebo, 1.78–2.25 lm2; P ¼ FSP-1 Expression of Tenon’s Fibroblasts 0.000) (Table). FSP-1 is a member of the S100 family of cytoplasmic calcium- binding proteins, which is expressed on fibroblasts and cells PCNA Expression of Tenon’s Fibroblasts that are transitioning into fibroblasts. It regulates cellular motility through a direct interaction with myosin-IIA. FSP-1 is PCNA is another marker of cell proliferation in fibroblasts that identified by subtractive and differential mRNA hybridization as has a very long half-life. In the present study, PCNA expression a gene expressed in fibroblasts but not in epithelial cells.30 In relative to CDKN1B was significantly reduced compared with the present study, a few FSP-1–positive cells were found placebo throughout the experimental period. This reduction in scattered at the surgical sites with Ad-CDKN1B delivery nuclear immunoreactivity was representative of the inhibition throughout the experimental period, whereas the positive on proliferation and repair mechanics. cells were abundant with placebo.

Downloaded from jov.arvojournals.org on 10/01/2021 348 Yang et al. IOVS, January 2013, Vol. 54, No. 1

FSP-1–positive cells, staining in fibroblast cytoplasm, P revealed a significant decrease on day 7 after Ad-CDKN1B delivery (38.10% of mean positive cells of Ad-CDKN1B vs. 81.82% of placebo; mean score, 2.22 vs. 3.72, P ¼ 0.000) (Figs. 5A, 5E, 5I). To day 14, only a small fraction (29.17%) of 0.32 0.000 0.35 0.000 0.41 0.003 0.30 0.000 fibroblasts was constituted by FSP-1–positive cells with Ad- 6 6 6 6 CDKN1B, but the percentage of positive cells (77.55%) was still high with placebo. The mean score of immunoactivity was 1.94 of Ad-CDKN1B vs. 3.89 of placebo (P ¼ 0.000) (Figs. 5B, 5F, 5I). The low percentage of positive cells was maintained AgNOR Ratio, % over the experimental time course (35.71% of Ad-CDKN1B vs. 0.33 9.78 0.29 7.78 0.22 7.37 0.26 8.35 84.85% of placebo on day 28; mean score, 2.00 vs. 3.83, P ¼ 6 6 6 6 0.000) (Figs. 5D, 5H, 5I).

Ad-CDKN1B PBS Cdk2 and Cdk4 Expression of Tenon’s Capsule Fibroblasts

P Cdk2 and Cdk4 expression has been largely implicated in induction of cell proliferation. Immunohistochemical analysis revealed that Cdk2 expressions relative to CDKN1B were

2 consistently present, but at a much lower level of Cdk2 over m 1.06 0.000 6.16 0.65 0.000 5.80 0.96 0.000 5.86 0.86 0.000 6.16

l the entire time course. Cdk2 expression was significantly 6 6 6 6 decreased following Ad-DKN1B delivery on day 7, and the mean percentage of Cdk2-positive fibroblasts was 53.65% of Ad-CDKN1B vs. 78.14% of placebo (mean score, 2.67 vs. 3.44, P ¼ 0.000). Interestingly, both nuclear and cytoplasmic immunoreactivity to Cdk2 were observed in positive cells of Ad-CDKN1B, whereas both intense cytoplasmic immunoreac- 0.61 51.45 0.46 45.19 0.26 48.09 0.46 44.62 tivity and nuclear staining were shown in fibroblasts of placebo 6 6 6 6 (Figs. 6A, 6E, 6I). Then Cdk2 expression showed a relatively small reduction on following days (Figs. 6B, 6C, 6F–J). On day Ad-CDKN1B PBS 28, the significant difference remained between Ad-CDKN1B and placebo (46.82% of mean Cdk2-positive cells of Ad- CDKN1B vs. 71.68% of placebo; mean score, 2.44 vs. 3.28, P ¼ P 0.000), as evidenced by nuclear and cytoplasmic immunoreactivity to Cdk2 antibodies in positive cells between Ad- CDKN1B and placebo (Figs. 6D, 6H, 6I). Similar to Cdk2, the expression of Cdk4 following Ad- 0.12 0.000 20.08 0.15 0.000 17.67 0.11 0.000 15.51 0.12 0.000 15.65 CDKN1B delivery was significantly lower than that of placebo 6 6 6 6 during the 28-day period (Fig. 7). The mean percentage of Cdk4-positive fibroblasts was 42.11% of Ad-CDKN1B and 66.67% of placebo on day 7 (mean score, 2.33 vs. 2.83, P ¼ 0.005). Both nuclear and cytoplasmic immunoreactivity to AgNOR Number Nucleus Area, Cdk4 were observed in positive cells in the two groups (Figs. 0.16 3.81 0.14 4.40 0.11 4.57 0.13 4.06 7A, 7E, 7I). Then Cdk4 expression was maintained at the lower 6 6 6 6 level on following days (Figs. 7B–E). The mean percentage of Cdk4-positive cells was 51.72% of Ad-CDKN1B vs. 89.79% of

Ad-CDKN1B PBS placebo at day 28 (mean score, 2.50 vs. 3.72, P ¼ 0.000), during which the nuclear and cytoplasmic immunostainings were still observed in Cdk4-positive cells (Figs. 7D, 7H, 7I). P

DISCUSSION 2

m The ﬁltration failure due to excessive postoperative scarring is 0.11 0.000 2.58 0.08 0.000 2.65 0.12 0.000 2.85 0.11 0.004 2.39 l the major challenge to the control of IOP and arrest of 6 6 6 6 glaucomatous damage after GFS. The human Tenon’s ﬁbro- blasts are considered to play a principal role in modulating the proliferation, migration, and synthesis of the ECM after the

6). ﬁltration surgery. Thus, inhibiting the proliferation of Tenon’s AgNOR Area, ¼ ﬁbroblasts during the wound healing is an important antiscar- n 0.07 2.02 0.09 1.86 0.07 1.78 0.08 1.36 ring strategy.1–4,17,18 6 6 6 6 The importance of CDKN1B in regulating cell proliferation SEM ( has been demonstrated in varied animal models. It has been 6 Ad-CDKN1B PBS

Comparison of the Mean AgNOR Area and Number between Ad-CDKN1B and Placebo-Treated Eyes shown that there was markedly decreased expression of

. CDKN1B in the corneal endothelium cells after injury in a rat

Means model of connexin43 knockdown,31 whereas the adenovirus- ABLE 28 0.79 21 1.02 T 14 1.03 Day 7 0.94 mediated overexpression of CDKN1B inhibited vascular

Downloaded from jov.arvojournals.org on 10/01/2021 IOVS, January 2013, Vol. 54, No. 1 When CDKN1B Inhibits Fibroblast Proliferation 349

FIGURE 4. Downregulation for PCNA in Tenon’s capsule. The serial panels show the decreased expression for PCNA in Tenon’s cells following Ad- CDKN1B delivery (A–D) in comparison with placebo (E–H, 3200). (I) Immunoactivity scores for PCNA. Means 6 SEM (n ¼ 6). *P < 0.05, **P < 0.01.

FIGURE 5. Expression for FSP-1 in Tenon’s fibroblasts. FSP-1–positive cells are decreased in Ad-CDKN1B. The number of FSP-1–positive fibroblasts was significantly suppressed by Ad-CDKN1B (A–D) compared with placebo (arrows, E–H, 3200). (I) Immunoactivity scores for FSP-1. Means 6 SEM (n ¼ 6). *P < 0.05, **P < 0.01.

Downloaded from jov.arvojournals.org on 10/01/2021 350 Yang et al. IOVS, January 2013, Vol. 54, No. 1

FIGURE 6. Downregulation for Cdk2 in Tenon’s capsule. Cdk2 expression decreases treated with Ad-CDKN1B (A–D) in comparison with placebo (arrows, E–H, 3200). (I) The immunoactivity scores for Cdk2. Means 6 SEM (n ¼ 6). *P < 0.05, **P < 0.01.

FIGURE 7. Cdk4 expression at surgical sites. The decreased expression for Cdk4 is observed with Ad-CDKN1B (A–D) in comparison with placebo (arrows, E–H, 3200). (I) The quantitative scores for Cdk4. Means 6 SEM (n ¼ 6). *P < 0.05, **P < 0.01.

Downloaded from jov.arvojournals.org on 10/01/2021 IOVS, January 2013, Vol. 54, No. 1 When CDKN1B Inhibits Fibroblast Proliferation 351

smooth muscle cells proliferation,32 and led to cell growth cytoplasm.44,45 Our observation reveals weak or moderate arrest in human renal carcinoma cells.33,34 Wang et al.18 nuclear and cytoplasmic immunoreactivity to Cdk2 and Cdk4 suggested that siRNA-mediated gene silencing of Skp2 induced induced by Ad-CDKN1B, suggesting that CDKN1B inhibits the increased CDKN1B protein levels, and subsequently inhibited nuclear and cytoplasmic expression of Cdk2 and Cdk4, thereby the proliferation of rabbit Tenon’s fibroblasts (rTFs) after GFS. arresting DNA and centrosome duplication. Thus, it is presumed that the decreased cell proliferation was PCNA is used to determine the positive number of associated with elevated CDKN1B activity in conjunctival fibroblasts that are actively proliferating (i.e., growth fraction epithelial cells and rTFs, and CDKN1B may be a significant and S-phase index).46 Our data suggest a lower score of prognostic factor determining the proliferation. positive cells of PCNA relative to CDKN1B throughout the In the present study, we distinguished Tenon’s fibroblasts experimental period. This reduction in nuclear immunoreac- stained with AgNOR, PCNA, Cdk2, and Cdk4 based on tivity indicates the inhibition on proliferation and repair morphologic features and localization. The positively stained mechanics. The Cip/Kip family, including CDKN1A and fibroblasts were further confirmed by FPS-1 staining experi- CDKN1B, share a common N-terminal domain for binding to ments. We have identified that Ad-CDKN1B induces an cyclin-Cdk complexes. The CDKN1A also binds to PCNA increase of the CDKN1B protein level in Tenon’s fibroblasts through a separate C-terminal domain affecting DNA replica- after sclerectomy in rabbits. The sustained overexpression of tion and repair, thereby forming cyclin-Cdk-CDKN1A-PCNA CDKN1B via adenovirus-mediated gene transfer leads to a quaternary complexes.40 The CDKN1B also contains unique C- persistence of inhibition of fibroblast proliferation throughout terminal domains whose functions are unknown, but shares the time course after surgery. very little primary sequence similarity with CDKN1A. It has been demonstrated that two pertinent factors In conclusion, this study demonstrates that the sustained influence CDKN1B function: intracellular concentration and overexpression of CDKN1B mediated by Ad-CDKN1B in subcellular localization.35 CDKN1B activity is intimately Tenon’s fibroblasts in sclerectomy rabbits leads to the associated with its nuclear localization in positive Tenon’s inhibition of fibroblast proliferation, and the downregulation fibroblasts of surgical sites, as observed in our study upon Ad- of Cdk2 and Cdk4 activity after GFS. The results indicate that CDKN1B treatment. The nuclear localization of CDKN1B the CDKN1B/Cdks pathway in the model of GFS may regulate protein indicates the antiproliferative effects because the the cellular proliferation in a coordinated manner. nuclear CDKN1B inhibits cyclin E-Cdk2.36 Furthermore, the nuclear export of CDKN1B has been shown to be dependent Acknowledgments on the S10 phosphorylation of CDKN1B, and the phosphorylation regulates the reentry of quiescent cells into the cell The authors thank Yu-mei Tian and Jian-xin Liu (Research Center cycle.35,37 for Neuroscience, Xi’an Jiaotong University, Xi’an, China) for Wen et al.38 showed that adenovirus-mediated overexpres- technical assistance. The 293 cell line was a kind gift from Zeng- sion of CDKN1A (also named p21Cip1 or p21), another Cip/Kip hui Teng (Fourth Military Medical University, Xi’an, China). family member, in Tenon’s fibroblasts inhibited cell proliferation, whereas no expression of CDKN1A protein was observed References in scleral fibroblasts or cornea. Our data suggest that adenovirus-mediated overexpression of CDKN1B in superficial 1. Grisanti S, Szurman P, Warga M, et al. Decorin modulates wound scleral stroma was observed in the most time course, although healing in experimental glaucoma filtration surgery: a pilot study. the expression level was markedly lower compared with that Invest Ophthalmol Vis Sci. 2005;46:191–196. in Tenon’s fibroblasts. These results indicate that Ad-CDKN1B 2. Perkins TW, Faha B, Ni M, et al. Adenovirus-mediated gene therapy may be delivered to the surrounding region of scleral flap, using human p21WAF-1/Cip-1 to prevent wound healing in a rabbit where it inhibits fibroblast proliferation. Yoshida et al.39 model of glaucoma filtration surgery. Arch Ophthalmol. 2002;120: suggested that the disappearance of CDKN1B was well 941–949. correlated with cell proliferation in the corneal epithelium 3. Honjo M, Tanihara H, Kameda T, Kawaji T, Yoshimura N, Araie M. after injury. Cdk4 and CDKN1B regulate proliferation in Potential role of rho-associated protein kinase inhibitor Y-27632 in corneal endothelial cells by regulating the cell-cycle progres- glaucoma filtration surgery. Invest Ophthalmol Vis Sci. 2007;48: sion. In our in vivo study, no evidence of CDKN1B 5549–5557. overexpression in cornea fibroblasts was found using the 4. Johnson KTM, Rödicker F, Heise K, et al. Adenoviral p53 gene adenoviral vector application. Unlike Tenon’s capsule and transfer inhibits human Tenon’s capsule fibroblast proliferation. Br sclera, the subcellular localization of CDKN1B in corneal J Ophthalmol. 2005;89:508–512. fibroblasts was almost undetectable, in both the nucleus and 5. Andrew CB, Aziza A, Richard SM, et al. Effect of diabetes mellitus the cytoplasm. Thus, our results confirm the CDKN1B-specific and hyperglycemia on the proliferation of human Tenon’s capsule inhibition of fibroproliferation in situ after sclerectomy. fibroblasts: implications for wound healing after glaucoma drainage It is currently clear that the deficiency of CDKN1B may surgery. Wound Repair Regen. 2005;13:295–302. induce Cdk2 and Cdk4 activity when cells reenter the cell 6. Shao T, Li X, Ge J. Target drug delivery system as a new scarring cycle. In the present study, Ad-CDKN1B–induced expression of modulation after glaucoma filtration surgery. Diagn Pathol. 2011; CDKN1B is accompanied by inhibition of Cdk2 and Cdk4 6:64. 40–43 activity, and leads to the cell-cycle arrest in the G1-phase. 7. Said TK, Moraes RCB, Singh U, Kittrell FS, Medina D. Cyclin- This is consistent with the ability of CDKN1B to bind to and dependent kinase (cdk) inhibitors/cdk4/cdk2 complexes in early form ternary complexes with cyclin-Cdks. The interaction stages of mouse mammary preneoplasia. Cell Growth Differ. 2001; between CDKN1B and the cyclin-Cdk complexes inhibits the 12:285–295. activities of G1/S- and S-phase Cdks, including cyclin E/Cdk2, 8. Andres´ V. Control of vascular cell proliferation and migration by cyclin A/Cdk2, and cyclin D1/Cdk4,34,40 which inhibits cell- cyclin-dependent kinase signalling: new perspectives and thera- cycle arrest due to the very low or no phosphorylation of the peutic potential. Cardiovasc Res. 2004;63:11–21. pRb protein.44 The cyclin-Cdk complexes, which are active 9. Marra DE, Simoncini T, Liao JK. Inhibition of vascular smooth within the nucleus, sometimes can act in both the nucleus and muscle cell proliferation by sodium salicylate mediated by the cytoplasm. The nucleocytoplasmic shuttling means that upregulation of p21Waf1 and p27Kip1. Circulation. 2000;102: they directly phosphorylate in both the nucleus and the 2124–2130.

Downloaded from jov.arvojournals.org on 10/01/2021 352 Yang et al. IOVS, January 2013, Vol. 54, No. 1

10. Rodier G, Montagnoli A, Marcotullio LD, et al. p27 cytoplasmic 29. Wasielewski RV, Mengel M, Wiese B, Rudiger¨ T, Muller-Hermelink¨ localization is regulated by phosphorylation on Ser10 and is not a HK, Kreipe H. Tissue array technology for testing interlaboratory prerequisite for its proteolysis. EMBO J. 2001;20:6672–6682. and interobserver reproducibility of immunohistochemical estro- 11. Frenquelli M, Muzio M, Scielzo M, et al. MicroRNA and gen receptor analysis in a large multicenter trial. Am J Clin Pathol. proliferation control in chronic lymphocytic leukemia: functional 2002;118:675–682. relationship between miR-221/222 cluster and p27. Blood. 2010; 30. Cheng J, Wang Y, Liang A, Jia L, Du J. FSP-1 silencing in bone 115:3949–3959. marrow cells suppresses neointima formation in vein graft. Circ 12. Uchida T, Nakamura T, Hashimoto N, et al. Deletion of Cdkn1b Res. 2012;110:230–240. ameliorates hyperglycemia by maintaining compensatory hyperin- 31. Nakano Y, Oyamada M, Dai P, Nakagami T, Kinoshita S, Takamatsu sulinemia in diabetic mice. Nat Med. 2005;11:175–182. T. Connexin43 knockdown accelerates wound healing but inhibits 13. Gerard´ C, Goldbeter A. Temporal self-organization of the cyclin/ mesenchymal transition after corneal endothelial injury in vivo. Cdk network driving the mammalian cell cycle. Proc Natl Acad Sci Invest Ophthalmol Vis Sci. 2008;49:93–104. USA. 2009;106:21643–21648. 32. Chen D, Krasinski K, Sylvester A, Chen J, Nisen PD, Andres´ V. 14. James MK, Ray A, Leznova D, Blain SW. Differential modification of Downregulation of cyclin-dependent kinase 2 activity and cyclin A p27Kip1 controls its cyclin D-cdk4 inhibitory activity. Mol Cell Biol. promoter activity in vascular smooth muscle cells by p27(KIP1), an 2008;28:498–510. inhibitor of neointima formation in the rat carotid artery. J Clin 15. Timmerbeul I, Garrett-Engele CM, Kossatz U, et al. Testing the Invest. 1997;99:2334–2341. importance of p27 degradation by the SCFskp2 pathway in murine 33. Katner AL, Hoang QB, Gootam P,et al. Induction of cell cycle arrest models of lung and colon cancer. Proc Natl Acad Sci U S A. 2006; and apoptosis in human prostate carcinoma cells by a recombinant 103:14009–14014. adenovirus expressing p27(Kip1). Prostate. 2002;53:77–87. 16. Knight JS, Sharma N, Robertson ES. SCFSkp2 Complex targeted by 34. Katner AL, Gootam P, Hoang QBL, Gnarra JR, Rayford WA. Epstein-Barr virus essential nuclear antigen. Mol Cell Biol. 2005;25: Recombinant adenovirus expressing P27(KIP1) induces cell cycle 1749–1763. arrest and apoptosis in human 786–0 renal carcinoma cells. J Urol. 17. Yang JG, Sun NX, Cui LJ, Wang XH, Feng ZH. Adenovirus-mediated 2002;68:766–773. delivery of p27KIP1 to prevent wound healing after experimental 35. Tsai S, Hollenbeck ST, Ryer EJ, et al. TGF-beta through Smad3 glaucoma filtration surgery. Acta Pharmacol Sin. 2009;30:413– signaling stimulates vascular smooth muscle cell proliferation and 423. neointimal formation. Am J Physiol Heart Circ Physiol. 2009;297: 18. Wang F, Qi LX, Su Y, Yan QH, Teng Y. Inhibition of cell proliferation H540–H549. of Tenon’s capsule fibroblast by S-phase kinase-interacting protein 36.VigliettoG,MottiML,FuscoA.Understandingp27(kip1) 2 targeting SiRNA through increasing p27 protein level. Invest deregulation in cancer: down-regulation or mislocalization. Cell Ophthalmol Vis Sci. 2010;51:1475–1482. Cycle. 2002;1:394–400. 19. Park KW, Kim DH, You HJ, et al. Activated forkhead transcription 37. Smitherman M, Lee K, Swanger J, Kapur R, Clurman BE. factor inhibits neointimal hyperplasia after angioplasty through Characterization and targeted disruption of murine Nup50, a induction of p27. . 2005;25:742–747. Arterioscler Thromb Vasc Biol p27Kip1-interacting component of the nuclear pore complex. Mol 20. Sedding DG, Seay U, Fink L, et al. Mechanosensitive p27Kip1 Cell Biol. 2000;20:5631–5642. regulation and cell cycle entry in vascular smooth muscle cells. 38. Wen SF, Chen Z, Nery J, Faha B. Characterization of adenovirus p21 . 2003;108:616–622. Circulation gene transfer, biodistribution, and immune response after local 21. Tsutsui T, Hesabi B, Moons DS, et al. Targeted disruption of CDK4 ocular delivery in New Zealand white rabbits. Exp Eye Res. 2003; Kip1 delays cell cycle entry with enhanced p27 activity. Mol Cell 77:355–365. Biol. 1999;19:7011–7019. 39. Yoshida K, Nakayama K, Nagahama H, et al. Involvement of 22. Fredersdorf S, Burns J, Milne AM, et al. High level expression of p27KIP1 degradation by Skp2 in the regulation of proliferation in kip1 p27 and cyclin D1 in some human breast cancer cells: Inverse response to wounding of corneal epithelium. Invest Ophthalmol kip1 correlation between the expression of p27 and degree of Vis Sci. 2002;43:364–370. malignancy in human breast and colorectal cancers. Proc Natl 40. Pei XH, Xiong Y. Biochemical and cellular mechanisms of Acad Sci U S A. 1997;94:6380–6385. mammalian CDK inhibitors: a few unresolved issues. Oncogene. 23. Masuda TA, Inoue H, Sonoda H, et al. Clinical and biological 2005;24:2787–2795. significance of S-phase kinase-associated protein 2 (Skp2) gene expression in gastric carcinoma: modulation of malignant pheno- 41. Tikoo R, Osterhout DJ, Casaccia BP, Seth P, Koff A, Chao MV. Kip1 type by Skp2 overexpression, possibly via p27 proteolysis. Cancer Ectopic expression of p27 in oligodendrocyte progenitor cells Res. 2002;62:3819–3825. results in cell-cycle growth arrest. J Neurobiol. 1998;36:431–40. 24. Bryant P, Zheng Q, Pumiglia K. Focal adhesion kinase controls 42. Braun-Dullaeus RC, Mann MJ, Ziegler A, von der Leyen HE, Dzau cellular levels of p27/Kip1 and p21/Cip1 through Skp2-dependent VJ.Anovelroleforthecyclin-dependentkinaseinhibitor and -independent mechanisms. Mol Cell Biol. 2006;26:4201–4213. p27(Kip1) in angiotensin II-stimulated vascular smooth muscle cell hypertrophy. J Clin Invest. 1999;104:815–823. 25. Plotton D, Menager M, Jeannesson P, Himber G, Pigeon F, Adnet J. Improvement in the staining and visualization of the argyrophilic 43. Motti ML, Califano D, Baldassarre G, et al. Reduced E-cadherin proteins of the nucleolar organizer regions at the optical level. expression contributes to the loss of p27kip1-mediated mechanism Histochem J. 1986;18:5–14. of contact inhibition in thyroid anaplastic carcinomas. Carcino- 26. Yamada Y, Yoshimi N, Hirose Y, et al. Sequential analysis of genesis. 2005;26:1021–1034. morphological and biological properties of -catenin-accumulated 44. Surjit M, Kumar R, Mishra RN, et al. The severe acute respiratory crypts, provable premalignant lesions independent of aberrant crypt syndrome coronavirus nucleocapsid protein is phosphorylated and foci in rat colon carcinogenesis. Cancer Res. 2001;61:1874–1878. localizes in the cytoplasm by 14-3-3-mediated translocation. J Virol. 27. Löhr CV, Teifke JP, Failing K, Weiss E. Characterization of the 2005;79:11476–11486. proliferation state in canine mammary tumors by the standardized 45. Jackman M, Kubota Y, Elzen ND, Hagting A, Pines J. Cyclin A- and AgNOR method with postfixation and immunohistologic detection cyclin E-Cdk complexes shuttle between the nucleus and the of Ki-67 and PCNA. Vet Pathol. 1997;34:212–221. cytoplasm. Mol Biol Cell. 2002;13:1030–1045. 28. de Graaf R, Dammers R, Vainas T, Hoeks APG, Tordoir JHM. 46. Su Y, Qu Y, Jiang C, Liu L, Shan Y, Wang F. KLF6SV1 siRNA inhibits Detection of cell-cycle regulators in failed arteriovenous fistulas for proliferation of human lens epithelial cells. Mol Vis. 2012;18:601– haemodialysis. Nephrol Dial Transplant. 2003;18:814–818. 605.

Downloaded from jov.arvojournals.org on 10/01/2021