Abstract
Purpose.:
To investigate the effect of poly (lactic-co-glycolic acid) (PLGA) implants loaded with mitomycin C (MMC) and with different adjuvant treatments after glaucoma filtration surgery (GFS), in comparison to standard treatments.
Methods.:
Forty-two New Zealand White rabbits underwent bilateral GFS and received different treatments: topical MMC (group 1); topical 5-fluorouracil (5-FU; group 2); PLGA implant (group 3); MMC-loaded and -coated PLGA implant (group 4); MMC-loaded and 5-FU–coated PLGA implant (group 5); subconjunctival bevacizumab (group 6); MMC-loaded PLGA implant and subconjunctival bevacizumab (group 7); and no treatment (right eye of all animals; control group). Intraocular pressure (IOP) and filtering bleb were evaluated on different days after GFS. Histology was performed to examine the conjunctiva, sclerotomy, filtering bleb, and persistence of the implant.
Results.:
The best hypotensive results were achieved in the MMC-loaded and -coated PLGA implant group, which presented the lowest IOP values on days 1, 5, 7, 14, and 28 after GFS. Excluding the implant groups, in which the bleb could not be properly measured, bleb survival was superior to controls in groups 1, 2 and lower in group 6. Group 7 presented greater extension, height, and vascularization of the bleb. Epithelial thinning and lymphoplasmacytic infiltrate were observed in groups 1, 2, 4, 5, and 7. The rates of closure of the sclerotomy and bleb were 100% and 76%, respectively, and implant persistence was 95%.
Conclusions.:
MMC-loaded and -coated implants have optimal surgical results, followed by topical MMC application. In this experimental model, bevacizumab could interact with MMC.
All animal experiments were conducted according to the guidelines of the ARVO Statement for the Use of Animals in Ophthalmic and Visual Research and approved by the Ethics Committee of the Basque Country University. Forty-two New Zealand White female rabbits (weight, 2.5–2.7 kg) were used in the study. Rabbits were anesthetized using a combination of ketamine (40 mg/kg; Ketolar 50 mg/mL; Pfizer, Madrid, Spain) and xylazine (8 mg/kg; Rompun 20 mg/mL; Bayer, Barcelona, Spain), administered by intramuscular injection before surgery.
GFS was performed by the same surgeon in both eyes of each animal, considering the right eye as the control. Briefly, after placing a partial-thickness corneal traction suture, a fornix-based conjunctival flap was performed on the superior lateral quadrant of the eye. A half-thickness, rectangular, 4- × 3-mm scleral flap was then dissected, and a clear corneal paracentesis was carried out. Then, a 2- × 1.5-mm sclerectomy followed by peripheral iridectomy was performed. The scleral flap was closed with two 10-0 nylon sutures (Alcon Surgical, Fort Worth, TX). Finally, one drop of both atropine 1% eye drops (Colircusi Atropina; Alcon-Cusí, El Masnou, Barcelona, Spain) and the combination of tobramycin and dexamethasone (Tobradex; Alcon-Cusí) were instilled.
Implants were obtained by aggregation of PLGA 50:50 microspheres, prepared using the double emulsion/solvent extraction method (w/o/w).
20 Four sets of implants were obtained: drug-free implants, MMC-loaded implants, MMC-loaded and -coated implants, and MMC-loaded and 5-FU–coated implants.
Rabbits were randomly allocated to seven groups (n = 6 eyes per group), in relation to the adjuvant treatment administered to the left eye, described as follows:
-
Group 1: topical intraoperative MMC (surgical sponge soaked in 0.4 mg/mL solution, applied for 5 minutes);
-
Group 2: topical intraoperative 5-FU (surgical sponge soaked in 50 mg/mL solution, applied for 5 minutes);
-
Group 3: PLGA drug-free implant without adjuvant treatment;
-
Group 4: MMC-coated PLGA implant containing MMC microparticles;
-
Group 5: 5-FU-coated PLGA implant containing MMC microparticles;
-
Group 6: postoperative subconjunctival bevacizumab (0.05 mL; Avastin, Roche Farma, Madrid, Spain); and
-
Group 7: noncoated PLGA implant containing MMC microparticles and subconjunctival bevacizumab.
For the control group, the right eye of each animal was treated with trabeculectomy without adjunctive treatment (i.e., all animals in groups 1–7; n = 42 eyes).
Topical MMC and 5-FU were applied over the filtration site prior to the dissection of the scleral flap and irrigated thoroughly with 20 mL of balanced salt solution. Implants were placed adjacent to the surgical site, subconjunctivally, just before conjunctival closure. The concentration of MMC employed in the manufacture of the implants, as well as on their coating, was 1.25 mg/mL. The concentration of 5-FU employed as coating was 66.25 mg/mL. Bevacizumab was administered in a single subconjunctival injection at the bleb site, once surgery had concluded. The PLGA implant was prepared by accumulation of microspheres, obtained through the solvent evaporation/extraction method.
Evaluations were performed preoperatively (5 minutes before the surgery) and postoperatively at 0, 1, 5, 7, 14, 21, and 28 days after trabeculectomy. These examinations included measurement of intraocular pressure (IOP) and analysis of the bleb.
IOP was determined by rebound tonometry (Icare VET; Icare Finland, Helsinki, Finland) after instillation of a mixture of oxybuprocaine and tetracaine (Colircusi anestésico doble; Alcon-Cusí). Two sets of six measurements of IOP were performed, yielding an average value for both series. All determinations were performed by the same observer and at the same time of the day (10:00–12:00 AM).
The characteristics of the filtering bleb were analyzed in a blinded way according to the Moorfields bleb grading system (Clarke JC, et al. IOVS 2003;44:ARVO E-Abstract 1201). This classification scores the bleb with respect to a set of reference photographs, taking into account bleb height, extension, and vascularity.
On day 30, animals were killed by means of anesthetic overdose. After the enucleation of both eyes, specimens were fixed in 4% formaldehyde, cryopreserved in 30% sucrose, and included in OCT. Then, they were cut into 14-μm sections on a cryostat and stained with hematoxylin-eosin and Verhoeff (Verhoeff's hematoxylin). Sections were cut taking as a reference the suture points of the trabeculectomy, in such a way that the sections included the operated area and the control area (opposite limbal region). These preparations were examined and photographed using a microscope (Axioskop 2; Carl Zeiss Microscopy, LLC, Thornwood, NY), with particular attention to the characteristics of the conjunctival epithelium, stromal cell infiltration and persistence of the sclerotomy, filtering bleb, and implant.
Statistics were performed using statistical software (SPSS; SPSS Sciences 19.0, Chicago, IL). Normality of variables was analyzed with the Kolmogorov-Smirnov test. IOP values were analyzed using parametric tests (Student's t-test, ANOVA of repeated measurements and Tukey's contrast test). The bleb was analyzed using nonparametric tests (Mann-Whitney U test, Wilcoxon W test, and the Kruskal-Wallis H test). The correlation between IOP and the filtering bleb variables was evaluated using the Spearman correlation test. We also analyzed GFS survival using the Kaplan-Meier method, defining surgical failure as the return to the basal IOP ± 1 mm Hg or the disappearance of the filtering bleb (grade 1 extension bleb) and the log rank test to make comparisons between groups. The level of statistical significance was considered to be P < 0.05.
Control Group.
Study Groups.
Group 1 (topical MMC;
n = 6 eyes) showed improved results with respect to the control group throughout the study period (
Fig. 3,
Table 3) and particularly at 14 (
P = 0.006) and 21 days (
P = 0.005) after surgery. Paradoxically, IOP was higher than control immediately after surgery (
P = 0.016). Mean surgical survival for this group was 21.6 days ± 4.51 (mean ± SE; range, 5–30 days;
Fig. 2,
Table 2).
Table 3 Evolution of IOP (Mean ± SD) Over Time in All Study Groups
Table 3 Evolution of IOP (Mean ± SD) Over Time in All Study Groups
| | Days Postsurgery |
| Pre | 0 | 1 | 5 | 7 | 14 | 21 | 28 |
| Control | 9.39 ± 1.61 | 3.29 ± 3.07 | 6.49 ± 2.73 | 7.56 ± 1.32 | 7.88 ± 1.93 | 9.44 ± 2.05 | 8.59 ± 2.11 | 8.86 ± 1.49 |
| Group 1 (MMC) | 9.50 ± 1.87 | 8.00 ± 5.72 | 5.17 ± 2.04 | 7.50 ± 1.51 | 7.33 ± 1.75 | 6.83 ± 2.31* | 5.67 ± 3.32* | 8.86 ± 1.49 |
| Group 2 (5FU) | 9.17 ± 1.47 | 5.50 ± 5.28 | 7.83 ± 1.94 | 7.00 ± 1.54 | 6.17 ± 0.75* | 7.67 ± 1.63* | 7.33 ± 2.33 | 7.67 ± 1.86 |
| Group 3 (Imp) | 8.50 ± 1.76 | 3.50 ± 3.32 | 5.00 ± 1.41 | 6.00 ± 1.67* | 7.00 ± 1.26 | 7.50 ± 1.04* | 8.33 ± 2.73 | 8.50 ± 1.76 |
| Group 4 (Imp-MMC-MMC) | 8.17 ± 1.60 | 0.67 ± 1.63 | 4.83 ± 1.72 | 5.67 ± 1.36* | 5.17 ± 2.92* | 6.33 ± 2.25* | 7.67 ± 1.75 | 6.50 ± 1.64* |
| Group 5 (Imp-MMC-5FU) | 8.50 ± 0.83 | 0.17 ± 0.40 | 5.67 ± 2.16 | 6.83 ± 1.94 | 6.67 ± 2.06 | 9.67 ± 2.58 | 8.33 ± 3.38 | 7.33 ± 1.50* |
| Group 6 (BVC) | 8.83 ± 0.98 | 2.83 ± 3.35 | 3.83 ± 2.48* | 6.67 ± 0.82 | 8.17 ± 1.47 | 7.67 ± 1.63* | 8.83 ± 0.41 | 6.83 ± 1.47* |
| Group 7 (Imp-MMC-BVC) | 9.17 ± 2.31 | 0.50 ± 0.83 | 6.33 ± 1.86 | 6.50 ± 2.07 | 7.83 ± 2.63 | 8.67 ± 1.21 | 8.33 ± 1.36 | 7.33 ± 1.86* |
In group 2 (topical 5-FU;
n = 6 eyes), IOP was found to decrease in a similar manner (
Fig. 3,
Table 3), although this decrease was significantly different with respect to the control group at 7 (
P = 0.03) and 14 days (
P = 0.049) postop. The only difference with group 1 was found at day 1, in which IOP with 5-FU was also higher than control. Mean surgical survival for this group was 16.6 days ± 4.08 (mean ± SE; range, 5–30 days), which was longer than that for the control group, but shorter than that of the topical MMC group (
Fig. 2,
Table 2).
In group 3 (empty PLGA implant;
n = 6 eyes), IOP was reduced at 1, 5, 7, and 14 days postsurgery, with respect to the control group (
Fig. 3,
Table 3), with this reduction being statistically significant at 5 (
P = 0.012) and 14 (
P = 0.029) days. Surgical survival in this group was similar to that of the control group (13.1 days ± 3.32; mean ± SE; range, 7–30 days;
Fig. 2,
Table 2). These results are demonstrative of an isolated hypotensive effect due to the spacing effect of the implant.
Group 4 (MMC-loaded and -coated PLGA implant;
n = 6 eyes) presented IOP reduction at all stages of the study (
Fig. 3,
Table 3), and significantly at 5 (
P = 0.002); 7 (
P = 0.004); 14 (
P = 0.001); and 28 days (
P = 0.001). The longest survival results are associated with this group (22 days ± 3.89; mean ± SE; range, 7–30 days;
Fig. 2,
Table 2).
Group 5 (MMC-loaded, 5-FU–coated, PLGA implant;
n = 6 eyes) exhibited values inferior to those of the control group throughout the study (
Fig. 3,
Table 3). These differences were statistically significant at 28 days postsurgery (
P = 0.027). Mean surgical survival for this group was 16.5 days ± 4.45 (mean ± SE; range, 1–30 days), which is longer than that of the control group (
Fig. 2,
Table 2).
Group 6 animals (subconjunctival bevacizumab;
n = 6 eyes) experienced IOP reduction after 1, 5, 14, and 28 days of GFS (
Fig. 3,
Table 3), with this reduction being statistically significant at 1 (
P = 0.03); 14 (
P = 0.049); and 28 days (
P = 0.004). Despite this, mean GFS survival for this group was 11.3 ± 3.07 days (mean ± SE; range, 5–21 days); shorter than that of the control group (11.9 days;
Fig. 2,
Table 2).
In group 7 (MMC-loaded PLGA implant and subconjunctival bevacizumab;
n = 6 eyes), we observed IOP reduction throughout the study (
Fig. 3,
Table 3), which was statistically significant only after 28 days (
P = 0.032), as was observed in group 5. Mean survival for this group was 16.6 days ± 4.08 (mean ± SE; range, 5–30 days;
Fig. 2,
Table 2).
Upon comparing the different coadjuvant treatments, we found statistically significant differences between groups. The groups that contributed to these differences were group 4 (MMC-loaded and -coated PLGA implant) and group 8 (control). The most significant results were obtained in group 4, which presented the lowest values of IOP at days 1, 5, 7, 14, and 28. These results were exceeded only by group 2 at day 21 and by group 5 immediately postsurgery.
Control Group.
Study Groups.
Similarly, we observed a decreasing bleb grade pattern in all study groups (
Fig. 4). Mean surgical survival, as function of bleb extension, was longer than that of the control group in topical MMC group (22 days ± 3.26; mean ± SE; range, 14–30 days) and topical 5-FU group (14.1 days ± 3.03; mean ± SE; range, 7–30 days), and shorter in subconjunctival bevacizumab group (11.1 days ± 3.48; mean ± SE; range, 5–30 days). In the other groups, survival could not be determined due to the bias induced by the spacing effect of the implant.
Filtering bleb vascularization presented a decreasing pattern in all groups, which was not statistically different to that of the controls, with the exception of the implant groups, with higher hyperemia (
P < 0.05 in all cases, from day 1) and of subconjunctival bevacizumab group, which was less hyperemic on day 1 (
P = 0.003;
Fig. 4).
Group comparison revealed a larger bleb size in the groups with implant. Of these, the MMC-loaded implant and subconjunctival bevacizumab groups presented larger values of height, extension, and vascularization of the bleb conjunctiva.
Analysis of correlation revealed a negative correlation between IOP and the area and height of the filtering bleb. In the case of area, this correlation was significant at day 28 (rs = −0.259; P = 0.03). In the case of height, significance was found on days 5 (rs = −0.262; P = 0.017); 14 (rs = −0.28; P = 0.032); and 28 (rs = −0.329; P = 0.005).
Similar GFS survival periods were found independent of the criterion of surgical survival that was employed (IOP or bleb grade). Thus, once we had excluded the implant groups, due to the bias introduced in bleb grading, we found better results in group 1, followed by groups 2, 8, and 6, respectively.
In the present study, we have used the albino rabbit, since it is the most typically used animal in experimental studies of GFS. In addition, this animal is easy to manage and house and is inexpensive. Given its efficient wound healing characteristics, it allows a shortening of follow-up times and would be equivalent to a model of high-risk surgical failure, with accompanying antimitotic treatment.
21
The standard in wound healing studies after GFS consists in the intervention of a healthy eye, without prior induction of glaucoma. In this way, the iatrogenic effects produced by the induction of glaucoma are avoided. However, by intervening normotensive eyes, IOP measurement as a parameter during follow-up can be invalidated.
22 In fact, in models similar to ours, some authors reject IOP as a criterion of surgical functionality, considering only filtering bleb,
13,23–26 while others claim that the IOP would be higher than the bleb measurement.
18,26 In the present study, we found that both parameters are efficacious, presenting—as expected—an inverse correlation. It should be recognized, however, that in measuring bleb size, we excluded rabbits with implants, since the anatomical effect of the implant was indistinguishable from that of the filtering bleb.
The surgical technique employed herein has been used previously and it is very similar to what we routinely use in daily clinical practice.
8,13 Some other authors tend to use other techniques such as posterior-lip sclerotomy
7,27 or perform sclerotomy with the placing of a cannula draining into the subconjunctival space.
23,28,29 Upon comparing the different techniques, Esson concluded that sclerotomy with cannula induces lower synthesis of wound-healing mediators than posterior-lip sclerotomy.
30 This study demonstrates that the GFS experimental model described herein can efficaciously reduce IOP, and facilitate the formation of the filtering bleb, with similar surgical survival rates, independent of the failure criterion employed. We believe that the influence of anesthesia on IOP values has been minimal in the study, since the examinations were performed under topical anesthesia. However, the immediate postoperative determination could be influenced by the anesthesia, since it has been reported that the effect of ketamine-xylazine mixture remains 30 minutes after the anesthetic induction.
31
The doses and administration routes of 5-FU, MMC, and bevacizumab have been adopted from other studies.
6,23,32–34 In the case of bevacizumab, we have chosen to study its administration in a single dose,
6 in order to be able to compare its effects with those of other treatments. Other authors have recommended its repeated administration, due to its short half-life.
25
All of the evaluated treatments exhibit a hypotensive effect that was superior to that found in the control group. Excluding the implant-containing groups, bleb formation was also higher in the experimental groups than in the control group. The bevacizumab group is the only exception, since surgical survival in this group was shorter than that of the control group, independently of the failure criterion employed.
Topical MMC was surpassed in reducing IOP only by implants that were loaded and covered with MMC. The initial elevation of IOP could be due to postoperative inflammation, whose peak occurs in the first week.
35 The subsequent IOP reduction could be caused by the maximum effect of the drug in this experimental model. We found no evidence of corneal complications (opacity and neovascularization), hypotony, or endophthalmitis, in contrast to that reported by other authors.
18,33 The histological alterations observed in the conjunctiva are very similar to those reported by Sherwood, despite the difference in the time of analysis (2 weeks vs. 1 month in our study).
28
Topical 5-FU achieved a hypotensive effect that was lower than that of MMC administered topically or via implant, but higher than that associated with the isolated implant and bevacizumab. These results corroborate those of other authors in terms of hypotensive effects,
13,15 but differ in terms of bleb formation. Thus, Einmahl et al.
13 reported an 83% persistence of bleb 1 month after injection of poly(ortho esters) loaded with 5-FU. As in the case of MMC, we did not find any complications such as corneal edema
13,15 or avascular filtering bleb.
33
Subconjunctival bevacizumab treatment achieved reduced IOP in comparison with controls. However, this reduction was somewhat irregular and consequently, surgical survival was not improved. In addition, we found no improvement with bevacizumab in terms of filtering bleb formation. These findings contrast with those of other authors who have reported a longer life of the bleb, in the absence of IOP differences, when administered either alone
36 or repeatedly.
25 As reported previously,
25,37 the blebs of this group were found to be less hyperemic at the early postoperative stage. Histologically, we found more substantial conjunctival alterations than those reported by Memarzadeh et al.,
25 despite the fact that these authors performed multiple postoperative injections.
The use of drug-free PLGA implants was also associated with reduced IOP, corroborating the spacing effect of the implant, as has been previously reported.
17 Our implants led to hyperemia throughout the study period, without associated complications arising, in line with the findings of Cui et al. for PLA.
15 We did not observe implant encapsulation in any cases of this group, as might occur with other experimental groups. Other authors have already reported the presence of multinucleated gigantic cells,
10,13,16,27 which likely represent a foreign body response to the implants.
The mixed MMC application (encapsulated and covering the implant) was found to be the treatment with highest hypotensive efficacy. Given our previous in vitro results, the release of MMC from the microspheres starts from the first week (data not shown). Thus, it is likely that the MMC that covers the microspheres acts during the earlier stages, and is substituted by that liberated by the microspheres at a later stage. As detected in the other groups involving drug-loaded implants, we observed a correlation between IOP and the degree of persistence of the implant. This may be due to the encapsulation of the implant, which in turn would lead to inhibition of MMC diffusion to tissue, resulting in increased IOP.
MMC implants coated with 5-FU exhibit improved IOP levels with respect to the control group. However, these improvements are not as large as those associated with the MMC-loaded and -coated implants group and the topical MMC group. Lu obtained similar results upon comparing the effect of PLGA microparticles loaded with 5-FU and covered with MMC versus topical MMC, despite the fact that the MMC dose in these studies was different.
18 The similarity between this group and the topical 5-FU group would indicate the absence of an additive effect between MMC and 5-FU in our study. Curiously, in this group, we observed a higher rate of implant degradation at the end of the study in comparison with the other implant groups. This finding has been previously reported by others regarding PLGA
16 and polyorthoesters.
13
Finally, the group with MMC implants and bevacizumab exhibited reduced IOP, but to a lesser extent than that observed with the other implant groups, indicating that there may be a certain interaction between these drugs. In fact, Hilgert reported a higher degree of fibrosis upon combining both drugs, in comparison to using MMC alone (Hilgert CR, et al.
IOVS 2011;52:ARVO E-Abstract 634). This interaction can be seen histologically as a higher degree of implant encapsulation in this group. In contrast, other authors have reported a synergistic effect of combining topical MMC and subconjunctival bevacizumab in humans.
38 However, in our study, we cannot exclude a possible role of PLGA in the interaction between the two drugs.
In conclusion, in the present animal model, MMC-loaded and -coated PLGA implants provide optimal hypotensive results after GFS, indicating that this methodology may provide improved surgical outcomes in humans.
Supported in part by the Ministry of Science and Innovation through PROFIT program, Grant dex-580000-2008-41, Grupos Consolidados Gobierno Vasco (GOBE) IT437-10, and the Ministry of Science and Innovation through Apply Research, Grant PPT-010000-2009-30. The authors alone are responsible for the content and writing of the paper.
Disclosure: I. Rodríguez-Agirretxe, None; S.C. Vega, Bioftalmik SL (E); R. Rezola, None; E. Vecino, None; J. Mendicute, None; T. Suarez-Cortes, Bioftalmik SL (E); A. Acera, Bioftalmik SL (E)