Comparative Assessment of Short-Term Tendon-Scleral Postoperative Inflammation and α-Smooth Muscle Actin Expression following Oral and Topical Diclofenac Administration for Strabismus Surgery in Rabbits

Abstract Purpose Wound healing and fibrosis modulation are considered pivotal for the long-term outcome of strabismus surgery. Nonsteroidal anti-inflammatory drugs, including diclofenac sodium, are inflammation suppressive drugs that may modulate wound healing, including postoperative inflammation. This study aimed to compare the effect of oral and 0.1% topical diclofenac sodium on short-term inflammation and α-smooth muscle actin (α-SMA) expression at the tendon-scleral attachment site following strabismus surgery in an experimental rabbit model. Methods Superior rectus recession was performed in 12 eyes of six New Zealand rabbits. Rabbits were divided into three groups: oral diclofenac 2 × 5 mg/kg for three days (group A), 0.1% diclofenac sodium eye drops 3 times/day for three days (group B), and controls (group C). On postoperative day 14, enucleation was performed. Macroscopic adhesion score, microscopic adhesion score, percentage of postoperative inflammation area (Masson’s trichrome staining), and α-SMA (immunohistochemistry staining) were assessed. Data analysis was performed using a semi-quantitative and quantitative assessment with ImageJ. All groups were compared with reciprocal staining intensity (RSI) values to measure α-SMA expression. Results All groups showed no difference in macroscopic (p = 0.13) and microscopic adhesion scores (p = 0.28). The percentage of postoperative inflammation area in group B (12.44% (8.63–18.29)) was significantly lower than group A (26.76% (21.38–37.56) p = 0.03) and group C (27.80% (16.42–36.28), p = 0.04). Comparative RSI analysis found that group B had a significantly lower α-SMA expression than group C (174.08 ± 21.78 vs 212.58 ± 12.06, p = 0.04). Conclusion The results suggest that compared to oral, the administration of topical diclofenac showed a more significant reduction of short-term postoperative inflammation and α-SMA expression at the tendon-scleral attachment site following strabismus surgery.


Introduction
Secondary strabismus is one of the unexpected outcomes following strabismus surgery. 1 It comprises of residual and consecutive strabismus that occurs later in an undefined time frame, although satisfactory ocular alignment results may have been initially achieved. 2,3 Poor wound healing that causes secondary strabismus may arise as under or overcorrection. Under-correction contributes a significant proportion, ranging from 20 to 59% cases. 3 On the other hand, overcorrection accounted for 2-20% of secondary strabismus cases. 3 The true incidence of secondary strabismus is hard to estimate due to the potency of its late occurrence. In 1999, Ludwig explored the results of strabismus surgery that manifested secondary strabismus during 12 years of follow-up. Amorphous scar tissue separating the tendon and its attachment site in the sclera was identified. Stretch scar tissue was noted in 198 muscles (123 patients). 2 Excessive muscle fibrosis following strabismus surgery may result in significant problem. Kersey and Vivian found that extraocular muscle fibrosis implicated to the development of re-deviation for up to 50 prism diopters and re-manifested diplopia. 4 Fibrosis is a form of pathological or aberrant wound healing. The presence of fibrosis reflects the loss of some healthy cell/tissue fractions. Persistent stimulation and overexpression of pro-inflammatory cytokines may activate immune cells, leading to myofibroblast hyperactivation. This pro-fibrotic condition results in excessive extracellular matrix formation. Transforming Growth Factor (TGF)-b is involved in a significant portion of growth factors during fibrosis formation. The TGF-b cascade stimulates a-smooth muscle actin (a-SMA) expression in myofibroblasts.
Excessive a-SMA expression was found to be associated with fibrosis. [5][6][7][8] Several methods to control wound healing and fibrosis following strabismus surgery have been investigated, including the use of amnion membrane 9,10 and the application drugs in the surgical wound site during surgery. [11][12][13][14][15] Meanwhile, modulation of wound healing through prostaglandin inhibition serves as a promising alternative therapeutic target. The proliferation, differentiation, and fusion of cells involved in wound healing are previously known to be modulated by prostaglandins. During the inflammation phase of wound healing, prostaglandin, an arachidonic acid product from the cell walls of macrophages and neutrophils, contributes to inflammatory cell chemotaxis, nitric oxide induction, and increased vascular permeability. It is modulated by cyclooxygenase (COX) enzymes, especially COX-2. Nonsteroidal anti-inflammatory drugs (NSAIDs), including diclofenac, inhibit COX. NSAIDs administration will result in decreased prostaglandins, prostacyclin, and thromboxane synthesis. [16][17][18] Of note, comparison of topical and oral diclofenac administration in reducing postoperative inflammation on extraocular muscle has not been previously studied. The current study was conducted to evaluate the effect of diclofenac on short-term postoperative inflammation following strabismus surgery. Besides, a comparative assessment of its delivery routes, oral versus topical, was performed using semiquantitative and quantitative measures of a-SMA expression.

Material and methods
An animal experiment was performed at the Animal Research Facility, Indonesian Medical Education and Research Institute. A randomized controlled trial was conducted on six New Zealand White rabbits, weighing 2600 grams to 3300 grams, without prior strabismus or identifiable muscular abnormalities. Rabbits were acclimatized for seven days and treated according to the resolution for the use of animals in ophthalmic and vision research published by The Association for Research in Vision and Ophthalmology. 19 Rabbits were housed in an individual cage and given access to maintenance food and water ad libitum. Ethical clearance was obtained from the Ethical Committee Faculty of Medicine Universitas Indonesia (No. KET-8/ UN2.F1/ETIK/PPM.00.02/2021).

Sample size calculation
Sample size calculation was done using the resource equation approach. The approved degree of freedom (DF) ranges from 10 to 20. 20 This approach is deemed adequate for this exploratory study in which determining standard deviation and effect size is difficult. The formula being used was [n ¼ (DF/k) þ1] with k as the number of treatment arms. The minimum number of samples needed in each group was 10/ 3 þ 1 ¼ 4.33 $ 4 samples (eyes) per treatment group by taking into account the ethical principles of 3 R (replacement, reduction, refinement). A total of 6 rabbits (12 eyes) were then randomized to determine 2 rabbits (4 eyes) given oral diclofenac, 2 rabbits (4 eyes) received diclofenac eye drops, and 2 rabbits (4 eyes) as controls.

Surgical procedure and intervention
Superior rectus muscle recession was performed by an experienced strabismologist (APB). The animals were anesthetized with intramuscular injection containing a mixture of ketamine HCl (50 mg/kg) and xylazine (5 mg/kg). Topical anesthesia with 0.5% propacaine hydrochloride eye drop with subsequent 5% povidone-iodine solution as antisepsis was applied prior to surgery. A 180-degree conjunctival peritomy was performed, and the superior rectus muscle was isolated using two muscle hooks. Three-millimeter superior rectus muscle recession was done with 6-0 double-armed polyglactin sutures (Vicryl V R , Ethicon, Piscataway, NY, USA). The conjunctiva was then repositioned and gently closed with 8-0 polyglactin sutures. After the procedure, each rabbit received ofloxacin eye drops 3 times for three days and the assigned treatment concurrently. The rabbits were divided into one of the treatment groups:

Macroscopic evaluation
All eyes were enucleated on day 14 following surgery. Before enucleation, euthanasia was performed by injecting ketamine 300 mg/kg and xylazine 30 mg/kg subcutaneously. A macroscopic adhesion assessment was done immediately after euthanasia by one of the authors (IP) masked to the treatment group. Macroscopic adhesion was scored as previously described: 14 0 ¼ no adhesion; 1 ¼ adhesion separable easily with gentle dissection; 2 ¼ mild-moderate adhesion with freely dissectible plane; 3 ¼ moderate-dense adhesion with difficult dissection; and 4 ¼ non-dissectible adhesion.

Histopathologic evaluation
Enucleated eyes were fixated under 10% buffered-formalin for 24-72 h before histological slide processing. Histological processing was done in a predefined area of superior rectus muscle attachment (2-4 mm wide from the central muscle attachment site extending 2 mm anteriorly and 4 mm posteriorly). The length of the tendon at the scleral attachment site was not quantified. However, we observed that the knot was surrounded by muscle tissue. Masson's trichrome staining was used to assess microscopic adhesion and to quantify postoperative inflammation by evaluating the area of early fibrotic reaction and collagen deposition at the tendon-scleral attachment site. Microscopic adhesion was scored by an anatomic pathologist (ES) who was masked to the trial as follows: 21 0 ¼ No fibrosis; 1 ¼ mild perimuscular fibrotic reaction; thin collagen band; 2 ¼ easily identified thick collagen bands; 3 ¼ dense collagen bands; and 4 ¼ dense and large fibrotic reaction area.
Immunohistochemistry (IHC) staining was done using a specific antibody to a-SMA in muscle tissue (ARG52485, antialpha smooth muscle actin antibody 1A4, Arigobio). Expression of the a-SMA under IHC staining appears as colored cells. Evaluation was carried out by an expert (ES) in the Department of Anatomical Pathology, Faculty of Medicine, Universitas Indonesia. The semi-quantitative assessment was performed under a light microscope by ES as follows: 22 1. Percentage of positive cells: The percentage of positive cells multiplied by staining intensity was calculated and grouped into index values: low (score 1-2), medium (score 3-5), and high (score 6-9). 22 Quantitative assessment using ImageJ analysis Quantitative assessment of postoperative inflammation by the evaluation of early fibrotic reaction and collagen deposition area percentage used microscope photograph of Masson's trichrome staining from each slide using 100 Â magnification in three different visual fields (Figure 1). Digital quantification of postoperative inflammation area was performed using ImageJ software (https://imagej.nih.gov/ij) as previously described by Chen et al. 23 Images were processed using color threshold deconvolution method (threshold is around 110-145) in an area of 800 Â 1000 pixels around the insertion area of the new muscle without including the sclera or artefact area. Processed images to quantify inflammation area are provided in Appendix A (Supplementary file).
For the quantification of a-SMA expression, ImageJ software with immunohistochemisty (IHC) toolbox plugin (https://imagej.nih.gov/ij/plugins/ihc-toolbox/index.html) was used. A representative photo of 100 times magnification was performed using hematoxylin -Diaminobenzidine (H-DAB) color deconvolution, which was then transformed into binary color deconvolution in an area of 800 Â 1000 pixels. The area around the new muscle insertion and thread was selected. Sclera and artefacts were avoided ( Figure 1). a-SMA expression was quantified by calculating reciprocal staining intensity (RSI) using RSI formula (255mean grey value) according to a previously published method. 24,25 Processed images to quantify a-SMA expression in each eye are provided in Appendix A (Supplementary file).

Statistical analysis
The comparative assessment for macroscopic and microscopic adhesion scores was performed using the Kruskal-wallis test.  (a and b) with the implementation of deconvoluted color threshold as described in the text (selected area; yellow box). Image processing of IHC staining using IHC toolbox plugins in ImageJ (c and d).
For the percentage postoperative inflammation area and RSI analysis of a-SMA quantification, the ANOVA with Bonferroni post-hoc test was selected as both values gave normal distribution on the Shapiro-Wilk normality test. Statistical analyses were performed with SPSS V R version 25.0 for windows (SPSS, Inc., Chicago, IL, USA). Graphics were created using GraphPad Prism V R version 8 for windows (GraphPad Software, San Diego, CA, USA). The level of statistical significance was set at 5% (p < 0.05).

Macroscopic and microscopic adhesion
Both macroscopic (p ¼ 0.13) and microscopic (p ¼ 0.28) adhesions were not statistically different in all groups. However, eyes in group B tended to have a lower macroscopic adhesion score (score 0: 2 eyes; Figure 2). The distribution of microscopic adhesion score is shown in Table 1. Of note, no eyes in group B had a score 4.

Percentage of postoperative inflammation area
The percentage of postoperative inflammation area in group B (median 12.44% (8.63-18.29)) was significantly lower than group A (26.76% (21.38-37.56)) and group C (27.80% (16.42-36.28)). The mean difference between eyes in group B and C was 14.12% (standard error 5.17%). There was no difference between group A and C (Figure 3).

A-SMA expression
All enucleated eye preparations in all groups showed strong positive cell percentage (score 3) based on the semi-quantitative assessment. All eyes except one in group A and C group yielded index score 6 ( Table 2). Using RSI analysis, group B demonstrated a statistically lower a-SMA expression than group A and C (RSI group B ¼ 174.08 ± 21.78 vs. group A ¼ 206.50 ± 18.93 vs. group C ¼ 212.58 ± 12.06; one-way ANOVA p ¼ 0.03; post-hoc Bonferroni diclofenac eye drop group vs control group¼ 0.04; Figure 3b). The mean difference in RSI scores between group B and C was -38.49 (std error 12.77). In addition, we found that the higher the intensity of a-SMA expression, the higher the percentage of early fibrotic reaction and collagen deposition area with a significant moderate correlation (r ¼ 0.587; p ¼ 0.045; R2 ¼ 19.7%).

Discussion
Our study found that topical but not oral diclofenac administration for the first three days following strabismus surgery significantly reduced postoperative inflammation and a-SMA expression. This reflects the ability of topical NSAID to reduce myofibroblast activity in extraocular muscle. As previously studied, diclofenac eye drops can reach posterior sclera and be deposited locally. However, data on its concentrations in the sclera or extraocular muscles at various time points following its administration have not been reported. 26 Besides, we argued that topical diclofenac might reach therapeutic level that capable of suppressing local short-term inflammation and fibrotic reaction in extraocular muscle, even though the exact diclofenac level was not measured. The optimum route of delivery, whether topical or oral, that would result in the greatest potency of diclofenac has been evaluated using the knee as the target organ. Oral diclofenac administration showed a higher COX-2 inhibition than topical. 27 This may be due to the presence of mechanical barriers that consists of bone and integumentary tissue that alter topical diclofenac absorption in the knee. Besides, differences in vascular caliber between various organs potentially contribute to different drug level. 27 However, such mechanical barriers are not exist in ocular surface and this may, theoretically, explain the greater effect of topical  diclofenac in our study in contrast to previous study using the knee. Moreover, mechanical sweeping to clear off local debris or contaminants with eye drop administration may give more benefit to topical over the oral route in enhancing wound healing. 28 Compared to topical steroids, diclofenac demonstrated an excellent anti-inflammatory effect without increased risk of raised intraocular pressure. Following strabismus surgery, there was no significant difference between topical diclofenac and steroid in reducing postoperative inflammation with conjunctival edema and erythema as parameters. 29 Another study by Kim et al. tried to compare pain relief and conjunctival injection with the administration of topical diclofenac, fluorometholone, or dexamethasone. Topical diclofenac effectively reduced pain and conjunctival injection without intraocular pressure increase that was observed in the other two steroid groups. 30 Based on those findings, diclofenac administration may be preferred to reduce postoperative inflammation. However, our study expanded the exploration of its benefit toward fibrosis modulation as excessive inflammatory response and muscle fibrosis are known to be closely related to some extents, 6,8,31 which has never been explored in extraocular muscle tissue.
The presence of early fibrotic reaction and collagen deposition, which might arguably lead to long-term fibrosis, would contribute to a mechanical restriction that may complicate re-surgery if it is performed, particularly in complex cases. Adhesion involving muscle/tendon, sclera, and conjunctiva has been observed intraoperatively. 32,33 Several experiments to prevent excessive fibrosis using rabbits have been reported prior to our study. Amnion membrane placed in the scleral bed or surrounding the recessed/resected muscle showed inconsistent results. Amnion membrane could dampen inflammation, thus reducing fibrosis. 9,10 However, the occurrence of xenograft rejection had been also reported and associated with more severe inflammation and fibrosis. 34 With regard to antifibrotic agents, mitomycin-C, 5-fluorouracil, infliximab, and tranilast eye drop showed promising results. [11][12][13][14][15] Of note, these antifibrotic drugs may not always be readily available, limiting their utility in clinical settings. Hence, our study using diclofenac, which is relatively more affordable, may serve as an alternative to reduce postoperative inflammation following strabismus surgery.
During inflammation phase following muscle injury, components of danger-associated molecular patterns will trigger the recruitment of inflammatory cells and the secretion of IL-4, IL-5, IL-10, and IL-13. These cytokines would trigger macrophage phenotype in promoting fibrosis. 35 Given its anti-inflammatory effect, NSAIDs administration during this phase showed varying results. Blomgran et al. found that parecoxib (COX-2 selective) did not decrease inflammatory markers based on their assessment on CD45, CD11b, CD68, CCR7, CD163, CD206, CD3, and CD4 cells even though muscle healing was well achieved. 36 In contrast, another study in mice found that diclofenac reduced macrophages and neutrophils concentrations in the paratenon but not in the tendon core area. However, there was no difference in tissue regeneration between diclofenac-treated group and controls based on the collagen appearance under Masson's trichrome staining. 37 We hypothesized that diclofenac may modulate muscle tendon inflammation through different machrophage subtypes being stimulated. Macrophages can be divided into two subtypes with two distinct characteristics: M1 and M2. M1 macrophage plays a significant role during the inflammation phase. Although the muscle tendon from the surgical site was not explicitly addressed in previous studies, results obtained by evaluating the role of macrophage polarization on muscle regeneration might partly enlighten the anti-inflammatory drug modulation properties of the underlying wound healing process, including extrapolation to our observation at the tendonscleral attachment site. Prolonged administration of NSAID  may potentially alter muscle regeneration as M2 macrophage affect myoblast activity. 6,8,16,31 Compared to M1, M2 macrophages have more anti-inflammatory properties. After day 3 of muscle injury, it is known that M2 macrophage population will increase over time. 38,39 Thus, NSAID administration during inflammation phase seems adequate to modulate short-term postoperative inflammation. Furthermore, this may partially explain the reduction of inflammation with topical diclofenac that arguably reaches a desirable antiinflammatory effect during the inflammation phase following extraocular muscle injury, as may occur in our study. Of note, our results are based on the evaluation on day 14 following surgery, in which coagulative properties during the initial wound healing phase take into account and may be different or have no effect on the long-term impact of fibrosis formation. The comparative assessment of postoperative inflammation and a-SMA expression in this study used ImageJ with the hope of better and more reliable results than manual microscopic assessment. This method can minimized subjectivity possessed by manual interpretation. 24 However, similar with previous study, different results from manual and ImageJ interpretation were also observed in our study. This may happen due to the inclusion of intensity readings in stromal areas, blood vessel walls consisting of myofibroblast, or even stained inflammatory cells under ImageJ analysis but not with manual interpretation. 40 Manual interpretation would only analyze stained myofibroblast contained in muscle tissue. Moreover, the previous study that we adopted for IHC interpretation used tumor tissue instead of muscle, 22 potentially affecting the agreement between manual and ImageJ readings in our study. The use of software or artificial intelligence in quantifying histology preparations is relatively new and is not readily become a standard; thus, further validation studies are needed.
Several limitations were encountered in this study. A thorough inflammation assessment was not performed as our evaluation was conducted on day 14 following strabismus surgery without serial evaluation at different stages of wound healing process. Assessment of subcellular and cytokines involved and their relationship to inflammation and fibrosis might provide a better picture of inflammation-fibrosis pathophysiology. Of note, the selection of day 14 was based on the evidence of concurrent complete muscle and scleral attachment 41 and the peak of a-SMA expression 42 from previous studies, reflecting a proper comparative assessment of those two parameters. However, the long-term outcome of tensile strength concerning the progressed wound healing phase and fibrosis formation might be different and need to be evaluated in the upcoming study. Besides, tissue angiogenesis was not quantified. Theoretically, angiogenesis may affect muscle healing, although the molecular relationship is still under investigation. 43 Lastly, all comparative assessments were performed without multiple observers which may affect the reliability of data interpretation. The recommendation for the clinical use of diclofenac eye drop following strabismus surgery still warrants further investigation, particularly in comparison to steroids, as our study did not compare diclofenac and steroids. Further study is required to elicit the best anti-inflammatory and anti-fibrotic regimen following strabismus surgery and consider the use of artificial intelligence to quantify long-term muscle fibrosis formation and a-SMA expression, including human histological sections obtained from reoperated cases, to better delineate the histological changes with the treatment.

Conclusions
Our study is the first to report the effect of topical diclofenac on reducing short-term postoperative inflammation in extraocular muscle following strabismus surgery by quantification of its effect using image-processing software analysis on histologic sections. Whether our finding on the initial postoperative inflammation and early fibrotic reaction would impact the ultimate fibrosis was not evaluated. We found no statistically significant difference in microscopic and macroscopic adhesion between topical diclofenac, oral diclofenac, and control groups. However, diclofenac eye drop administration significantly reduced the postoperative inflammation from histologic sections and a-SMA expression based on the quantitative assessment using ImageJ. Our study results reflect the potency of diclofenac eye drop to minimize short-term postopertive inflammation and raise the possibility of being prescribed following strabismus surgery, especially in complex cases with a high probability of repeat surgery.