Suppression of Type I Collagen Expression by miR-29b Via PI3K, Akt, and Sp1 Pathway, Part II: An In Vivo Investigation

Abstract

Purpose: We investigated the efficacy of miR-29b in inhibiting scar formation in rabbits who undergo glaucoma filtering surgery (GFS).

Methods: Trabeculectomy was performed on 60 rabbits diagnosed with glaucoma. The rabbits were divided into 5 groups: a blank group, single surgery group, positive control group that was treated with intraoperative mitomycin C (MMC), negative control group that was treated twice with empty vector postoperatively, and experimental group that was treated twice with Lentivirus-mediated miR-29b after being subjected to trabeculectomy. The operated eyes were tracked and followed up from postoperative days 1 to 28 (D1–D28). After the surgery, real-time PCR and Western blot analysis were performed on D28.

Results: At 1 week after undergoing GFS, the IOP was significantly lower in the eyes having filtering blebs. No statistically significant difference was found in the four treatment groups. After 21 days, the filtering bleb function score of the experimental group was the highest; however, their IOP was the lowest. On postoperative D28, the mean number of fibroblasts in the experimental group was significantly the lowest. The experimental group had the least collagen content according to Sircol assay. In the experimental group, the level of Col1A1 expression also was reduced in the sclera and conjunctival areas.

Conclusions: A subconjunctival injection of lentivirus-mediated miR-29b lowers postoperative IOP and sustains the function of filtering bleb. It inhibits the proliferation of fibroblasts and reduces collagen deposition by repressing the PI3K/Akt/Sp1 pathway in rabbits subjected to GFS.

This is the second report in which we have proposed a novel rationale for developing miRNA-based strategies, which can prevent scar formation in patients undergoing glaucoma filtering surgery (GFS). To develop a microRNA (miRNA)-based gene-silencing method that tackles in vitro antifibrosis, we identified miRNAs that could be used as candidates by confirming the presence of fibrosis-related genes, which were described in the first report.¹ We found that miR-29b suppresses type I collagen gene by repressing the PI3K/Akt/Sp1 pathway in human tenon's fibroblasts (HTFs). In this study, we used an animal model to evaluate the feasibility and efficacy of an miRNA-based gene-silencing method that is used to tackle fibrosis. 

Bleb scarring is the most common problem encountered while performing trabeculectomy on patients with glaucoma. In these patients, the recurrence of high IOP is witnessed commonly.² Poor prognosis of trabeculectomy is seen in patients who have IOP postoperatively. To attenuate scar formation, mitomycin C (MMC) and 5-fluorouracil (5-FU) are always applied intraoperatively; however, they cause significant postoperative complications that further intensify IOP, causing loss of vision.^3,4 Researchers now are trying to develop an alternative antiscarring method. 

In previous studies, researchers have reported that miRNAs have a pivotal role in organ fibrosis. Moreover, miR-29b directly targets 3′-UTR of PI3Kp85α, Sp1, and Col1A1 mRNAs, which encode profibrotic proteins in various cells.^5–10 Therefore, miR-29b potentially can be used to treat fibrotic diseases that afflict the heart,^5,6 liver,^7,11,12 kidneys,⁹ and other body organs.^10,13–17 In our previous report, we performed a systematic analysis to evaluate the effects of miR-29b on the biological properties of HTFs. We also have proved that an overexpression of miR-29b reduces the expression of fibrosis-related genes in HTFs. In this study, to determine whether miR-29b can inhibit in vivo scar formation effectively, we placed blebs and administered a subconjunctival injection to transfect lentivirus-mediated miR-29b into rabbits that had undergone GFS previously. Then, we performed real-time PCR and Western blot analysis on D28 to evaluate the in vivo antiscar activity of miR-29b. To determine the underlying mechanism through which miR-29b suppresses scar formation, we clinically monitored IOP, filtering blebs, complications, and side effects. 

In the literature, previous reports have established that lentivirus vectors (LVs) persist in dividing and nondividing cells. Compared to other vectors, they have a better capacity to incorporate interested gene fragments within safety limits.¹⁸ Therefore, we used LV to effectively deliver miR-29b in vivo. In this study, we developed an experimental model of GFS to determine whether the basis of new therapeutic strategies involves an efficient delivery of miR-29b. The aim of this study was to address the in vivo antiscarring effect of miR-29b. Therefore, we tried to develop a novel therapeutic strategy to suppress the activation and progression of bleb scarring that afflicts patients who have undergone GFS. 

Materials and Methods

Animal Models

Adult male rabbits (Oryctolagus cuniculus, 2 years old, weighing 2.5–3.5 kg) were used in this study. The procedure was performed in accordance with the animal care guidelines published by the Institute for Laboratory Animal Research (Guide for the Care and Use of Laboratory Animals) and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The IOP of both eyes was recorded with a Tonopen (Medtronic, Minneapolis, MN, USA) after applying oxybuprocaine, a topical anesthesia. Before performing the surgical procedure, the rabbits were further anesthetized using an intravenous injection of pentobarbital sodium (30 mg/kg). For 4 weeks, we developed chronic ocular hypertension in these rabbits by injecting 0.1 mL of 2% methylcellulose into the anterior chamber of their right eye once a week. After administering the injection, IOP was measured with a Tonopen every 3 days. In this research study, we only included eyes in which the IOP was in a stable range of 25 to 35 mm Hg. 

To establish the animal model of scar formation after GFS, we performed trabeculectomy in the superior quadrant of eyes. The trapezoidal sclera flap had dimensions of approximately 4 × 3 × 3 mm³. In the 1.5-mm thickness of trapezoidal sclera flap, we performed combined trabeculectomy on trabeculum tissues, whose dimensions were approximately 2.0 × 2.0 mm². Thereafter, we performed superior iridectomy on the same tissues. The conjunctival flaps were sutured using a 10-0 nylon suture. All surgeries were performed by the same investigator, who was experienced in performing ophthalmic surgeries on animals. The investigator performed these surgeries by viewing the sections under an operating microscope. After the operation, we applied Tobradex eye ointment (Alcon, Fort Worth, TX, USA). In addition, we also instilled in these rabbits Tobradex eyedrops (Alcon) four times a day for 1 week postoperatively. 

Transfection and Groups

While performing functional analysis, we transfected miR-29b into appropriate cells using LV as per manufacturer's instructions. Furthermore, miR-29b LVs (miR-29b-LV) harboring green fluorescent protein (GFP) were constructed and detected by Cyagen Bioscience (Guangzhou, China). A recombinant miR-29b-LV and a negative control lentivirus (NC-LV) were prepared and titrated into 1 × 10⁹ transfection units (TU)/mL. A subconjunctival injection was administered beside the filtering bleb using a sterile microinjector on days 3 and 10 postoperatively. We evaluated GFP expression by conducting standard examinations of enucleated tissues of eyes. 

We randomly divided 60 rabbits (60 eyes) diagnosed with glaucoma (IOP was stable in the range of 25–35 mm Hg) into five groups: (1) no surgery—the blank group that was not provided with any treatment, (2) surgery only—this group was treated only with trabeculectomy, (3) MMC—in this positive control group trabeculectomy was performed using 0.4 mg/mL MMC cotton pat intraoperatively for approximately 3 minutes, (4) empty vector—the negative control group was treated with 25 μL of NC-LV administered through a subconjunctival injection on days 3 and 10 postoperatively, and (5) miR-29b—the experimental group was treated with 25 μL of miR-29b-LV (containing 10 mg/mL polybrene) fluid administered through a subconjunctival injection on days 3 and 10 postoperatively. We included 12 rabbits in each group. The operated eyes were tracked and followed up from postoperative day 1 to day 28 (D1–D28; clinical observation included monitoring of IOP, filtering blebs, and complications). We performed hematoxylin and eosin (HE) staining, Van Gieson staining, real-time PCR, and Western blot analysis on D28 postoperatively. 

Histopathological Observation

All rabbits were killed directly through air embolism. Thereafter, we performed enucleation procedures on these rabbits. The operation sites of enucleated eyes were dissected into blocks that contained the bleb, conjunctiva, Tenon's capsule, and sclera. Thereafter, we performed HE staining, Van Gieson staining, and immunostaining of α-smooth muscle actin (α-SMA) to detect the number of fibroblasts, distribution of myofibroblasts, degree of fibrosis, and collagen deposition, respectively. To determine collagen content, we analyzed homogenates using Sircol assay (Biocolor, Carrickfergus, United Kingdom) according to manufacturer's instructions. In this assay, the tissue is homogenized with pepsin, and only newly formed collagen is measured. This is because only procollagen can be dissolved in pepsin, whereas crosslinked collagen will remain insoluble. Instead, it estimates a portion of procollagen supporting the argument that the diminished collagen is a lack of producing “new collagen.”¹⁹ 

Quantitative RT-PCR

After performing trabeculectomy on the eyes of rabbits, we harvested their tissues from the bleb area, including conjunctiva, Tenon's capsule, and sclera, on postoperative day 28 (D28). In this procedure, total RNA was extracted from the tissues using Trizol (Invitrogen, Wuhan, China). Then, the quantity and quality of the isolated RNAs were determined using a spectrophotometer (Lengguang, Shanghai, China). We synthesized cDNAs using 10 ng of total RNA, and they subsequently were used for carrying out quantitative RT-PCR. This procedure was conducted using SYBR green expression master mix (Applied Biosystems, Inc., Foster City, CA, USA). All experiments were performed in triplicate. The ΔΔCT method (2^−ΔΔCt) was applied to calculate the relative differences between the control and treatment groups, and the results were expressed as fold changes in gene expression. The forward and reverse primers were as follows: Actin K Forward 5′-CATCCTGCGTCTGGACCTGG-3′, Reverse 5′-TAATGTCACGCACGATTTCC-3′; PI3Kp85α – Forward 5′-ATCCTGTAGCGTGGGTAAAT-3′, Reverse 5′-CGAAGGGAGGGTAATAATAAG-3′; SP1 – Forward 5′-GCTGCCGCTCCCAACTT-3′, Reverse 5′-TTGCCCATCAACGGTCTG-3′; and COL1A1 – Forward 5′-GTTGTGCGATGACGTGATCTGTGA-3′, Reverse 5′-TTCTTGGTCGGTGGGTGACTCTG-3′. 

Western Blot Analysis

Samples of conjunctival (including Tenon's capsule) and sclera tissues were collected from the five groups. These samples were washed thrice using ice-cold PBS. Then, they were extracted in cold RIPA lysis buffer (strong), which consisted of 20 mmol/L Tris (pH 7.5), 150 mmol/L NaCl, 1% Triton X-100, 2.5 mmol/L sodium pyrophosphate, 1 mmol/L EDTA, 1% Na₃VO₄, 5 μg/mL leupeptin, and 1 mmol/L phenyl methyl sulfonyl fluoride (PMSF). After centrifugation for 10 minutes at 16099g, we collected the supernatant and used it in Western blot analysis. We visualized the immunoreactive proteins on autoradiograph films using chemiluminescence detection reagents (ECL; GE Healthcare, Laurel, MD, USA). Monoclonal antibodies to PI3Kp85α, phospho-PI3Kp85α, Akt, phospho-Akt, Sp1, and Col1A1 were obtained commercially (Cell Signaling Technologies, Danvers, MA, USA; GE Healthcare). The antibody β-Actin (Sigma-Aldrich Corp., St. Louis, MO, USA) was used as the loading control in all cases. 

Statistical Analysis

All experiments were performed in triplicate. We statistically evaluated the samples using Student's t-test to compare the differences between the treated and blank groups. The intergroup differences were tested by ANOVA. The data were expressed as mean ± SEM, and P < 0.05 was considered to be statistically significant. 

Results

Analysis of IOP Changes

In Figure 1, we illustrated the pre- to postoperative changes in the IOP of rabbits. The baseline IOP did not show any significant difference in all five groups (P > 0.05, ANOVA). After undergoing trabeculectomy surgery, the subjects of treated groups had remarkably lower IOP. Compared to the blank group, the IOP significantly decreased in the treated groups 2 weeks after the surgery (P < 0.05). However, the postoperative IOP of some surgical eyes gradually increased with time, and it completely returned to the preoperative level. Based on the mean IOP of surgical eyes, we inferred that the single surgery and negative control groups had the fastest increasing rate of IOP. Therefore, on postoperative D21, the IOP of these two groups was not significantly lower than that of the blank group (P > 0.05). At the last follow-up session of the positive control group, there was no significant difference between the IOP values that were measured before and after the operation (P > 0.05). On postoperative D28, the IOP of the experimental group was the lowest (P < 0.05); moreover, it still was significantly lower than the preoperative IOP level (P < 0.01). 

Figure 1

 

Measurement of IOP at each observed time points in the five groups.

Comparison of Filtering Bleb Condition

Under the light of a handheld slit-lamp, we conducted daily examinations to identify any changes at the surgical site and to assess bleb vascularity. The detailed grading and scoring was conducted according to criteria set by the Würzburg bleb classification score^20,21 (WBCS), which included the following items: vascularization, corkscrew vessels, encapsulation, and microcysts. According to the WBCS scores, there was no significant difference among the four operation groups (P > 0.05) on postoperative days D1 and D7. After 14 days, there was a significant difference between the experimental and other operation groups (P < 0.05). The single surgery and negative control groups had similar WBCS scores; moreover, the scores of both groups were lower than that of the MMC group. On postoperative D28, the functional filter was retained only in the MMC and experimental groups. Furthermore, the experimental group scored higher on the WBCS criteria (Fig. 2). 

Figure 2

 

Filtering bleb condition in the postoperative 4 weeks. (A) The blank group—the conjunctiva and sclera in the upper nasal space. (B) Blebs on postoperative D28. (B1) The single surgery group. (B2) The positive control group. (B3) The negative control group. (B4) The experimental group. (C) Filtering bleb score at each time point observed after operation in the four operated groups.

Downregulation of PI3Kp85, Sp1, and Col1A1 mRNAs With miR-29b-LV Transfection In Vivo

In previous studies, researchers had reported that miR-29b directly targeted 3′-UTR of PI3Kp85α, Sp1, and Col1A1 mRNAs in different cells.^5–10 Then, in our previous study, we also found that miR-29b was a potent miRNA regulating collagen, which was produced in HTFs.¹ In this study, we treated the surgical eyes of the experimental group with 25 μL of miR-29b-LV, containing 10 mg/mL of polybrene. This fluid was administered through a subconjunctival injection on days 3 and 10, postoperatively. Thus, we achieved an efficient transfection. The transfection efficiency was optimized according to the fluorescent intensity of reporter's dye using fluorescence microscopy (Fig. 3). To study the in vivo effects of an overexpression of miR-29b on PI3Kp85α, Sp1, and Col1A1 mRNA levels, we harvested the ocular tissues, including conjunctiva, Tenon's capsule, and sclera, from the bleb area on D28 after performing trabeculectomy. Based on our results, we deduce that transfection of miR-29b-LV can significantly inhibit the expression of PI3Kp85α, Sp1, and Col1A1 mRNAs in the rabbit model of GFS (Fig. 4). 

Figure 3

 

The operation sites of enucleated eyes were treated with 25 μL of miR-29b-LV fluid harboring GFP; the fluid was administered by a subconjunctival injection on days 3 and 10 postoperatively. (A) Staining with 4′,6-diamidino-2-phenylendole (DAPI) showed blue cell nucleus at a magnification of ×20. Scale bar: 100 μm. (B) Staining with DAPI was evaluated using micro-image analysis and GFP immunofluorescence staining. Green fluorescence corresponds to GFP-expressing cells at a magnification of ×20. Scale bar: 100 μm.

Figure 4

 

Real-time RT-PCR revealed that transfection of miR-29b-LV can significantly downregulate the mRNA level of PI3Kp85α, Sp1, and Col1A1in rabbit model of GFS. Compared to the other three operation groups, the experiment group had significantly lower expression of PI3Kp85α (A), Sp1 (B), and Col1A1 (C). P > 0.05, P < 0.05, P < 0.01 versus blank group by 1-way ANOVA. Data are mean ± SEM.

Regulation of Protein Expression Involved in PI3K/Akt and Col1A1 Pathways Using In Vivo miR-29b-LV Transfection

In our previous study, HTFs were transfected with miR-29b mimic. Consequently, an overexpression of miR-29b in HTFs significantly counteracted the TGFβ1-induced phosphorylation of PI3Kp85α and Akt. However, the total levels of PI3Kp85α and Akt were unaltered in all the treatment groups. 

In this study, we meticulously transfected miR-29b-LV into rabbits who had undergone GFS (Fig. 3). Based on the results of Western blotting analysis, we inferred that all groups were similar as they had the same level of total PI3Kp85α and Akt. However, the expression of phospho-PI3Kp85α and phospho-Akt decreases in various groups occurred in the following order: single surgery, negative control, positive control, experimental, and blank groups (Figs. 5A, 5B). Based on this finding, we further investigated the in vivo effects of miR-29b on Sp1 and Col1A1. Our results indicated that there was a decrease in Col1A1 and Sp1 miRNA, and protein levels when they were transfected with miR-29b-LV in GFS model. This was confirmed by RT-PCR and Western blot analysis (Figs. 4B, 4C, 5C). 

Figure 5

 

In vivo effects of miR-29b in PI3K/Akt and Col1A1 pathways, as determined by protein extraction and Western blot analysis. Beta-actin was used as the loading control. No surgery—the blank group, surgery only—the single surgery group, MMC—the positive control group, empty vector—the negative control group, and miR-29b—the experimental group.

Histologic Effects

On postoperative D28, we treated the surgical area blocks containing the bleb, conjunctiva, tenon, and sclera with HE staining and Van Gieson staining. As shown in Figure 6A, the mean numbers of fibroblasts in the single surgery, positive control, negative control, and experimental groups were 69.8 ± 11.2, 44.4 ± 9.8, 68.2 ± 10.6, and 33.8 ± 10.2, respectively. The mean number of fibroblasts in the experimental group was significantly the lowest (^aP < 0.01, ^bP < 0.05, ^cP < 0.01; Fig. 6A). The results of Van Gieson staining indicated that in the experimental group, which was transfected with miR-29b-LV, the collagen formation was significantly less between the conjunctiva and sclera of the operated area (Fig. 6C4). In the positive control group (Fig. 6C2), we observed some level of collagen formation. In the single surgery and negative control groups, there was abundant deposition of collagen. However, there was no significant difference in the mean macrophage number of these two groups (Figs. 6C1, 6C3). 

Figure 6

 

(A) Number of fibroblast in surgical area in the four operated groups on postoperative D28. ^aP < 0.01 versus single surgery group, ^bP < 0.05 versus MMC group, ^cP < 0.01 versus negative control group. (B) The gap between conjunctiva and sclera of the blank group is detected by Van Gieson staining (×100). (C) The conjunctiva and sclera in the operated area of the four surgical groups is detected by Van Gieson staining (×100). (C1) The single surgery group. (C2) The positive control group. (C3) The negative control group. (C4) The experimental group.

To further detect the degree of fibrosis and collagen deposition, we determined the immunohistochemistry of α-SMA and performed the Sircol assay. Between the conjunctiva and sclera, the expression of α-SMA was negative for the blank group (Fig. 7B: however, in the single surgery (Fig. 7C1) and negative control (Fig. 7C3) groups, α-SMA was highly positive under the blebs. In the positive control group, the expression of α-SMA was weak yet wide in the operation area (Fig. 7C2). In the experimental group, the expression of α-SMA was seldom (Fig. 7C4). We performed the Sircol assay to determine the collagen content in the surgical area of various groups. After 28 days, we found that miR-29b–treated eyes had significantly less collagen deposition compared to that in the other operated groups. Thus, the collagen levels in miR-29b-treated eyes was close to normal (Fig. 7A). As mentioned in methods, in the Sircol assay only procollagen can be dissolved in pepsin, and only newly formed collagen is measured. The results indicate that when miR-29b-LV concentrate is administered through a subconjunctival injection, it inhibits the in vivo formation of new collagen. 

Figure 7

 

(A) Sircol assay of collagen content in the tissue of surgical area. Results are expressed as means ± SD for six rabbits in each group. *P < 0.05 versus blank group, #P > 0.05 versus blank group. (B) The gap between the conjunctiva and sclera of the blank group is determined by α-SMA immunohistochemistry (×100). (C) The conjunctiva and sclera in the bleb area of the four operated groups is determined by α-SMA immunohistochemistry (×100). Representative images showing immunostaining of α-SMA; arrows indicate the presence of myofibroblasts. (C1) the single surgery group, (C2) the positive control group, (C3) the negative control group, and (C4) the experimental group.

Discussion

It is important to note that GFS is a surgical procedure that usually is performed on glaucoma patients, showing poor prognosis of conservative treatment. However, after undergoing GFS, most patients develop bleb scarring, which is the most common cause of surgical failure. In these patients, the IOP increases substantially.²² Intra- and postoperatively applied antimetabolites, such as 5-FU and MMC, have significantly decreased bleb scarring that occurs postoperatively.²³ However, the systemic application of antimetabolites usually causes damage to the cells, which is followed by bleb leakage, persistently low IOP, corneal scarring, scleritis, and endophthalmitis.^24,25 Owing to these undesirable effects, researchers now are trying to develop new therapeutic strategies to improve the success rate of GFS by preventing scar formation and inhibiting the development of fibroblasts in subconjunctival tissue. 

In previous studies, researchers tried several strategies to suppress the formation of fibroblasts and scar tissues; these strategies involved several kinds of biological and other approaches, including the inhibition of TGF-β,^26,27 VEGF,^28,29 ROCK,^30,31 photodynamic therapy,⁴ saratin,³² and so forth. Although some of these agents have produced amazing results by improving the outcome of GFS in a standard rabbit model, very few approaches have been successful in large prospective clinical trials. This indicates that none of these approaches can replace the use of MMC or 5-FU in clinical practice to date. 

MicroRNAs have been considered as critical regulators of fibrotic processes. They are great potential candidates for treating fibrosis in multiple organs of the human body.^33–35 Furthermore, HTFs, which are the major fibrotic cells in Tenon's tissue, induce bleb scarring in patients who have undergone transdifferentiation of myofibroblast-like cells.³⁶ In our previous study,¹ the in vitro studies indicated that levels decrease in myofibroblasts when miR-29b directly targets 3′-UTRs of PI3Kp85a, Sp1, and Col1A1. Moreover, miR-29b regulates the expression of genes in collagen by modulating the phosphorylation of PI3K/Akt/Sp1 signal pathway in HTFs. 

In this study, we used LV-mediated transfection to perform an in vivo gene delivery. Then, we performed RT-PCR and Western blot techniques to identify the expression of the transfected gene and to determine the change in protein levels. Based on our findings, we suggest that in rabbits who have undergone GFS, the transfection of miR-29b is achieved when the lentivirus concentrate is injected in the subconjunctiva that lies beside the filtering bleb. The resultant miR-29b-LV can inhibit the formation of new collagen in these rabbits by suppressing the repression of PI3K/Akt/Sp1 signaling pathway. Based on the findings of our previous in vitro studies and the current in vivo study, we deduced that miR-29b can effectively attenuate collagen deposition provided it is delivered exogenously. Thus, it suppresses the formation of bleb scarring in patients undergoing GFS. For the first time to our knowledge, in vitro and in vivo studies have established that delivering miR-29b can efficiently prevent the scarring of subconjunctival tissue in patients who have undergone GFS. Thus, we have developed a novel therapeutic strategy that can be used while performing glaucoma surgeries in the near future. 

In previous studies, researchers have developed several strategies to investigate the functional role of miR-29b in fibrosis and to authenticate miR-29b target genes. However, in these previous studies, an in vitro experimental study design was adopted. For example, Ogava et al.⁷ used human stellate cells to prove that IFNα upregulates miR-29b, which is a negative regulator of type I collagen. They facilitated the interaction between the 3′UTRs of Col1A1 and SP1. Luna et al.³⁷ and Villarreal et al.³⁸ performed similar research studies to determine whether miR-29b changes the expression of genes involved in the synthesis and deposition of extracellular matrix (ECM) in human trabecular meshwork cells (HTM) under different conditions. Based on the findings of previous research studies, a handful of in vivo studies have been conducted recently. Furthermore, Dawson et al.⁶ recently reported that they injected adeno-associated virus in an animal model representing congestive heart failure (CHF)-related atrial fibrillation (AF). They concluded that miR-29 probably has a pivotal role in atrial fibrotic remodeling, so it may be a valuable biomarker and/or therapeutic target. 

Rabbit (Oryctolagus cuniculus) is the only lagomorph animal whose genome is sequenced.³⁹ Li and his team^40–42 have performed many studies in which they have identified rabbit miRNAs using SOLiD and Solexa platforms. Subsequently, they have created a comprehensive collection of miRNAs that were expressed in rabbits. Based on these experimental findings, we also developed a rabbit model of GFS. For this purpose, we chose rabbit as the experimental organism of our study. While performing GFS, we implemented gene therapy approaches that were focused on antiscarring strategies. However, these approaches required long-term transgene expression in subconjunctival tissues. Therefore, we achieved this prerequisite condition in a rabbit model using a single subconjunctival injection of miR-29b-LV concentrate. Although 28 days are not enough to be considered as a sustained duration of time, the expression of GFP gene was ample and readily visualized in vivo in the enucleated tissues of operated sites (Fig. 3). Although it is challenging to identify a therapeutic transgene, we cannot underestimate the clinical benefits of gene therapy in inhibiting bleb scarring.⁴³ 

This study has some limitations. To the best of our knowledge, in previous studies, researchers had not performed the transfection of lentivirus-mediated microRNA in rabbits who had undergone GFS. Therefore, we had to determine the transduction time based on the characteristics of the fibrotic process: after undergoing GFS, there is proliferation of fibroblasts in rabbits. Therefore, the conjunctival wound of such rabbits gets repaired within 9 to 14 days postoperatively.^44,45 Furthermore, we were also unable to determine the best frequencies and doses of subconjunctival injection. In previous studies, researchers have reported that the in vivo transfection elicited by lentiviruses has low efficiency, so it is an area of concern and we need to further assess whether gene therapy is a suitable approach.^46–48 Therefore, in this study, to maximize the level of transfection, LV was applied in high titer and used in undiluted form. However, in future studies, we must carefully weigh other factors, such as the choice of promoter, the quality of vector preparation, and infection doses. Furthermore, in all observation processes, we have assumed that there were no shedding of vector particles in tears and aqueous humor. These assumptions are well-established when the observation period is not long, but future studies must investigate whether time-dependent changes have an impact on the efficiency of transfection. 

Conclusions

In conclusion, to our knowledge ours is the first report to describe LV-mediated transfection of microRNA, which was used to perform gene delivery in the GFS animal model. Based on the findings of our previous report,¹ we deduced that miR29b-based therapeutic strategies can be used for attenuating bleb scarring in patients who undergo GFS. 

Acknowledgments

Supported by National Natural Science Foundation of China Grant 81170843. 

Disclosure: J. Yu, None; H. Luo, None; N. Li, None; X. Duan, None