Photochemical Keratodesmos for Bonding Corneal Incisions

Abstract

purpose. To evaluate the immediate and long-term effectiveness of a dye-plus-laser irradiation treatment (photochemical keratodesmos [PKD]) for sealing corneal incisions.

methods. Incisions (3.5 mm) in rabbit corneas were treated on the incision walls with rose bengal dye followed by exposure to 514-nm laser radiation. PKD was evaluated in three groups (n = 3–6) using laser fluences of 115, 153, or 192 J/cm² (180-, 240-, and 300- second exposures, respectively) compared with an untreated group (n = 8). The intraocular pressure at which leakage occurred (IOP_L) during infusion of saline into the anterior chamber was determined. In a long-term study, treated and control corneas were observed weekly for 10 weeks for the appearance of neovascularization, anterior chamber inflammation, iridocorneal adhesion, corneal melting, and scarring.

results. Immediately after treatment, the IOP_L increased with increasing laser fluence, producing IOPs of 230 ± 90, 370 ± 120, and more than 500 mm Hg at 115, 153, and 192 J/cm², respectively, compared with 40 ± 20 mm Hg in control eyes (P < 0.005). No reduction in the IOP_L was observed up to 14 days after surgery. Corneal melting in PKD-treated or control eyes was not observed in the 10-week healing study. Neovascularization, which peaked at 4 weeks but resolved by 8 weeks, was detected around the incision in both PKD-treated and control eyes.

conclusions. Immediate and lasting sealing of corneal incisions was obtained in eyes treated with PKD, using short irradiation times. These results suggest that PKD has potential for improved corneal tissue bonding.

Although the visual prognoses after penetrating keratoplasty and surgical repair of perforating injuries of the anterior segment of the eye have improved significantly over the past 20 years, the use of sutures as the closure method may lead to several potential complications. Exposed sutures cause irritation and inflammation. ^{1 2} Exposed knots may form a nidus for infection from bacteria, viruses, and fungi. ^{3 4 5 6 7 8} Premature loosening of a suture, whether interrupted or running, may cause wound leakage, frequently associated with a flat or shallow anterior chamber and low intraocular pressure (IOP). The use of 10-0 nylon sutures slows the wound-healing process. ⁹ Finally, sutures that are too loose or tight can cause astigmatism, which limits visual acuity. ^{10 11}  

We have developed a method for corneal wound closure that utilizes a light-activated dye to produce covalent cross-links between tissue surfaces. ¹² In our previous study of the application of this method, called photochemical keratodesmos (PKD), we used fresh ex vivo New Zealand rabbit eyes and obtained promising results for the bonding of cataract surgical incisions. In those studies, watertight seals of incisional wounds were achieved by applying a photosensitizing dye, rose bengal (RB), and irradiating the incision site with an argon laser at 514 nm. A significant increase (P < 0.005) in the IOP required to cause leakage (IOP_L) from a sealed incision was seen after PKD treatment when RB and laser irradiation were applied together. Neither dye nor light was effective independently. RB was chosen for these studies because it photosensitizes cross-linking of solubilized collagen on light excitation. Formation of protein–protein cross-links by dye photosensitization has been reported. ^{13 14} In addition, light and photosensitizers have been reported to cause collagen cross-links. ^{15 16} The mechanism is believed to involve light absorption by the photosensitizer (e.g., rose bengal) followed by generation of radicals in the proteins. These radicals then couple to form the covalent cross-links. 

In the present study, we extended our investigations to an in vivo rabbit model. The strength of the seal produced by PKD was assessed relative to unirradiated control rabbit eyes both immediately after surgery and, in a pilot study, at times up to 2 weeks. Healing of the PKD-treated and control incisions was monitored for tissue responses over a 10-week period. 

Methods

Surgical Procedure

New Zealand White rabbits with a body weight of 1 to 2 kg were used. All experiments were performed in compliance with the guidelines of the Subcommittee on Research Animal Care at the Massachusetts General Hospital and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The surgical procedures were performed in animal surgery suites. The rabbits were anesthetized with intramuscular xylazine (5 mg/kg) and ketamine (35 mg/kg). Corneal incisions were performed with a 3.5-mm angled keratome (BD Biosciences, Lincoln Park, NJ). Three sets of experiments were performed. The first set was designed to examine the acute effect of PKD. A non–self-sealing, 3.5-mm incision was made in each cornea, 2 mm anterior to the limbus, angled 45° from the plane of the iris (Fig. 1) . Three interrupted radial sutures of black monofilament 10-0 nylon (Ethilon; Ethicon, Piscataway, NJ) were used to close the keratome incision. The sutures were removed immediately after the treatment and before pressure measurements. A second study was a pilot study in which the repair strength over a 2-week period was evaluated. In this study, a 3.5-mm incision was made 1 mm anterior to the limbus and parallel to the iris (Fig. 1) . This incision was the same as made in our previous ex vivo study. ¹² Sutures were not used because this incision creates a larger, self-sealing surface in the stroma. In a third set of experiments that were designed to examine healing after treatment, a 3.5-mm, non–self-sealing incision was made in a manner identical with that used in study 1, the acute study (Fig. 1) . Three sutures were used, which were removed immediately after the treatment, and the incisions were allowed to heal for 10 weeks. In the survival studies, a fluoroquinolone antibiotic, ofloxacin, was applied after surgery (0.3%, four times per day for 7 days) to reduce the likelihood of postoperative infection. In each study, the eyes were randomized, using a prescribed order for treatments that did not follow an obvious pattern. Thus, the quality of the incision was not related to the treatment group. 

Photochemical Keratodesmos Procedure

A solution of RB (1.5 mM in phosphate-buffered saline [pH 7.2] ∼20 μL) was applied to the walls of the incision with a 27-gauge needle. The dye was distributed on the incision wall as evenly as possible. The dye adheres to the incision surfaces resulting in characteristic red staining. Laser irradiation was performed using a continuous-wave argon-ion laser (Innova 100; Coherent, Palo Alto, CA) at 514 nm, coupled to a 1-mm diameter quartz fiber. A 1-cm diameter spot on the tissue was created, using a combination of 1- and 2-in. focal length S1-UV grade, fused silica biconvex lenses (Esco Products, Oak Ridge, NJ), mounted in a cage assembly (SM1 series; ThorLabs, Newton, NJ). The 1-cm diameter circular spot was sufficient to cover the entire incision, and the optics were adjusted so that the laser light was incident on the cornea at an angle approximately 45° to the plane of the incision (Fig. 1) . The fluence (in joules per square centimeter; energy delivered per area of tissue surface) was controlled by varying the treatment time (seconds) using a constant predetermined irradiance (in watts per square centimeter; rate of energy delivery per area of tissue surface). Results of pilot experiments showed that the irradiances and fluences used in these experiments did not produce visually obvious photothermal tissue damage in the cornea or iris. The irradiance was at least 1000 times lower than that used for laser iridectomy and iridoplasty. ¹⁷ The fluences used in this study, 115, 153, and 192 J/cm², were delivered in 180, 240, and 300 seconds, respectively, when the irradiance was 0.64 W/cm². The cornea was kept moist with phosphate-buffered saline during treatment. 

IOP Measurement

The integrity of the tissue seal was determined at various time points after treatment by measuring the IOP required to cause leakage of aqueous fluid through the incision (IOP_L). The IOP was gradually increased by infusion of saline (1 mL/min) into the anterior chamber. The same procedure that has been described in studies of ex vivo eyes was used. ¹² Briefly, a 27-gauge needle was inserted parallel to the iris, 2 mm above the limbus into clear cornea, diametrically opposite the incision site and positioned above the lens. The needle was connected to both a calibrated blood pressure transducer (Harvard Apparatus, South Natick, MA) and a mini-infuser (model 400; Bard Harvard) through a T-coupler. The signal generated by the transducer-amplifier combination is proportional to the lower limit of IOP. The IOP was increased until either the incision opened and fluid leaked from the anterior chamber or the maximum pressure reading of the transducer (500 mm Hg) was attained. Typical traces are shown in Figure 2 . 

Assessment of Corneal Changes

In study 3, eyes were evaluated during weekly slit lamp observations for corneal melting, scarring, neovascularization, adhesion of the iris to the cornea, anterior chamber inflammation and fibrin formation. These observations were tabulated, but not photographed. Given that the corneal neovascularization was mainly evident on slit lamp biomicroscopy at high magnification, eyes were scored for the extent of neovascularization using a subjective scoring system of 0 to 3, where 0 represents no evidence of neovascularization, 1 represents a single vessel, 3 represents neovascularization of the whole wound, and 2 is intermediate between 1 and 3. 

Statistical Analysis

Student’s unpaired t-test was used to compare IOP_L levels and neovascularization scores for treated versus untreated groups. 

Results

Acute Studies

The influence of laser fluence on IOP_L was evaluated in three experimental groups and in a control group. RB was applied to the walls of the incisions, made as shown in Figure 1 , in the eyes to be treated with PKD. The incisions in both PKD and control group corneas were then closed using three interrupted sutures to approximate the incision surfaces. Three PKD groups were treated with laser fluences of 115, 153, or 192 J/ cm² delivered over 180, 240, and 300 seconds, respectively, at a constant irradiance of 0.64 W/cm². Incisions in the control group were sutured but were not treated with RB or laser irradiation, as our prior studies had shown that neither light nor dye treatment alone produces significant bonding. ¹² Each treated group contained three to six eyes, and the untreated group contained eight eyes. The sutures were removed from the experimental corneas immediately after the PKD treatment and from the control corneas, and the IOP_L was measured. The results are shown in Figure 3 . The IOP_L was greater in all PKD-treated corneas than in the control (P < 0.005). The IOP_L increased with laser fluence, reaching 230 ± 95 at 115 J/cm² and 370 ± 125 at 153 J/cm². The IOP_L was greater than 500 mm Hg (the maximum pressure measurable) for eyes treated with 192 J/cm². No signs of thermal damage, such as tissue shrinkage, were observed under the irradiation conditions used. 

Long-Term Studies

In addition to measuring the acute strength of the corneal repair, we determined, in a pilot study, whether the repair strength is maintained for various time intervals after treatment. In a study involving 18 rabbits, incisions were made in a manner identical with that used in our previous ex vivo study (see Fig. 1 ) and the treatment did not employ sutures. ¹² Immediately after treatment, the IOP_L of corneas treated with RB and 40 J/cm² was ∼500 mm Hg (3 eyes). IOP_L measurements made on days 2 (n = 4), 3 (n = 3), 5 (n = 1), 7 (n = 1), and 14 (n = 6) remained greater than 500 mm Hg in the PKD-treated eye. The results of this small study indicate that the seal produced by PKD does not degrade in strength with time. 

During this pilot study, we observed neovascularization around some corneal incisions. To determine whether this effect is a potential deterrent to the use of PKD in corneal surgery, we performed experiments on nine rabbits divided into two groups. One eye in six rabbits was treated with RB plus 115 J/cm² of 514 nm laser light. The other eye served as a control and was treated identically, except it received no dye or laser irradiation. In the other three rabbits one eye was treated with PKD using 192 J/cm² (same irradiation time but using a higher irradiance of 1.07 W/cm²) and the other was untreated (no RB or laser). Three sutures were used to close corneal incisions in both the experimental and control groups and were removed immediately after the PKD or sham treatment. Slit lamp examination was performed weekly for 10 weeks. Documentation of the number of vessels in the corneal incision and subjective grading of scarring, anterior chamber reaction, and iridocorneal adhesions were used to compare PKD-treated and control eyes. In addition, the eyes were photographed weekly with a single lens reflex (SLR) camera. Because of the relatively small number of vessels per cornea and the limited number of strands forming adhesions between the iris and cornea, these photographs did not prove useful for performing additional quantitative analysis. Examples are shown in Figure 4 . 

New blood vessel formation was observed in seven of nine treated corneas and in five of nine control corneas (Fig. 4) . The neovascularization is presented as a function of time over a 10-week period in Figure 5 for the groups treated with 115 J/cm² (n = 6) and 192 J/cm² (n = 3) and the control group (n = 9). Neovascularization was usually observed by 2 weeks, peaked at 3 or 4 weeks and resolved by week 8 in all groups. Although the mean of the response was greater in corneas treated with 115 J/cm² than that in the controls, results in the two groups did not differ at any time point (P > 0.1). The degree of neovascularization in eyes treated with 192 J/cm² PKD closely resembled those with 115 J/cm² but were not tested for significance, because only three corneas were treated at the higher fluence. Iris attachment to the cornea was observed in seven of nine treated corneas and in two of nine control corneas and remained constant over the 10-week study period. Fibrin was present in three of nine treated corneas and in two of nine control corneas a few days after treatment but disappeared by 10 weeks. White, opaque scars that became more transparent with time were observed in six of nine treated and in four of nine control animals. The scar was limited to the incision site. No apparent differences in these features were observed between rabbits treated with PKD using 115 or 192 J/cm². 

Discussion

We have developed a method for photochemical sealing of corneal incisions that produces a strong, immediate, water-tight bond. Tissue bonding with PKD is achieved without glues or solders and uses low laser irradiances, which avoids thermal denaturation of the tissue proteins. PKD required relatively short irradiation times to produce strong seals. Incisions sealed with a 180-second irradiation resisted leakage up to an IOP_L of ∼230 mm Hg, which is well above the normal IOP of the human eye of 18 to 21 mm Hg. Even stronger bonding was produced by irradiating 300 seconds. These results suggest that even shorter irradiation times may be effective, making this method practical for corneal repair. 

Using higher irradiances to produce the same fluence in a shorter time seems logical but may produce thermal damage as an unwanted side effect. Our previous study using rabbit eyes ex vivo showed that thermal damage to the cornea resulted when high irradiances and fluence were used. ¹² Thermal damage was not detected during the current in vivo studies using moderate treatment conditions. Welding tissues together using the thermal effect of lasers has been reported frequently. However, attempts to weld corneal or other ocular tissues with lasers have met with mixed success. ^{18 19} Photothermal welding relies on denaturation of collagen molecules by the laser-induced temperature increase in the tissue. On cooling, noncovalent interactions between denatured collagen molecules is believed to be responsible for the strength of the tissue weld. ²⁰ However, the cornea is not well suited to a photothermal approach, because tissue shrinkage, due to the extended range of protein denaturation, accompanies thermal welding. Such shrinkage would cause corneal deformation and reduce visual acuity. 

Our results indicate that the relatively strong bonding produced by PKD in cornea persisted over a 2-week period. By 2 days after incision, even the untreated control rabbit eyes showed high IOP_L. The incisions in the PKD-treated corneas remained sealed and showed no evidence of corneal damage or inhibition of normal healing. 

Effects were produced in the rabbit corneal incision model that do not parallel those in human eyes, as has been found in other studies. ²¹ In the long-term study, neovascularization was observed in both PKD-treated and control eyes, which peaked at 3 to 4 weeks after incision. The neovascularization subsequently resolved with slight residual effect after 10 weeks. Growth of new vessels may be associated in the rabbit eye with corneal trauma near the limbus and iris adhesion. Clear corneal incisions during cataract surgery in human eyes do not lead to neovascularization, ²² highlighting an important difference in the wound-healing response between human and rabbit cornea. One other difference is that patients often receive steroid drugs after surgery, but this was avoided in our study. In most eyes, PKD-treated or control, an opaque scar limited to the wound site was evident after 10 weeks. This scar was a consequence of the surgical incision and seems to have been unaffected by the PKD treatment. Formation of fiber strands from the iris to the cornea in PKD-treated eyes was another frequent finding in the rabbit eye that is not usually seen in the human cornea after a similar incision. Leakage of aqueous humor from the anterior chamber after incision and the resultant decrease in IOP appeared to cause the more flaccid rabbit iris to protrude toward the leakage site and block the wound, where some iris fibers remained adherent after deepening of the anterior chamber, suturing, and PKD treatment. It is likely that the PKD treatment fixed those strands containing collagen fibers, which are ideal targets for PKD. This is consistent with the higher number of observations of iris strands in the PKD-treated than in the control corneas. The shallowing of the rabbit anterior chamber during surgery, compared with the human, may have contributed further to this effect. In addition, the incisions were close to the limbus and perpendicular to the corneal surface, unlike cataract incisions that are more parallel to the iris. The observed fibrin formation and the exaggerated corneal scarring are also typical responses to corneal incisions in rabbit but not in human eyes. 

Alternative wound and incision-closure methods include tissue adhesives such as cyanoacrylate and fibrin glues, among many others. ^{19 23 24 25} Cyanoacrylate glue has been shown to be an effective alternative to sutures in certain traumatic and infectious ophthalmic wounds in rabbits. However, this glue is irritating to tissue, difficult to apply, and does not give a permanent solution. Fibrin glue has been commercially available in Europe for several years. In ophthalmology, it has been used to seal crystalline lens perforations and to secure scleral reinforcement grafts. Fibrin glue is composed of human fibrinogen and thrombin, is absorbed in a few days, and promotes collagen cross-linking, thus, promoting wound healing. However, such bioadhesives are accompanied by problems associated with application, toxicity, and ultimate rejection of the material. Additional problems include isolation of the donor fibrinogen, which must be autologous or carefully screened for infectious agents. Photochemical methods employing a UV-absorbing chemical or photoactivated “solder” also been tested experimentally. ^{26 27 28 29 30} The solder, composed of a light-absorbing dye and a protein such as fibrinogen, fills the gap between the surfaces and does not permit bonding directly between the incision walls. 

PKD may offer certain advantages over the methods currently used to attach corneal tissue and to close incisions in a variety of surgical procedures. This study demonstrates the potential of PKD to seal full-thickness corneal incisions in a matter of minutes, with immediate and lasting strength. Additional studies evaluating the histologic, pachymetric, and microscopic findings are needed for a better understanding of the mechanism of action of PKD in the cornea. However, our studies suggest that the treatment is unlikely to hinder the normal healing response and, in the rabbit cornea, ultimately produces only a scar at the repair site. The incision model represents a simple corneal repair and was chosen for this reason to evaluate in vivo efficacy of photochemical tissue bonding in the cornea. The demonstration of tissue bonding in this model opens up further applications in corneal surgery in which PKD could provide benefit. 

 Supported by the U.S. Department of Defense Medical Free Electron Laser program, Contract F49620-01-1-0014, and the U.S. Department of Education, Contract DOE-FG02-91ER61228.

 Submitted for publication September 25, 2003; revised March 5, 2004; accepted March 19, 2004.

 Disclosure: C.E. Proaño, None; L. Mulroy, None; E. Jones, None; D.T. Azar, None; R.W. Redmond, None; I.E. Kochevar, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Schematic drawing of a cornea showing the sites of incisions and irradiation. In the first experiment (acute study) and third experiment (10-week study) a 3.5-mm incision was made 2 mm anterior to the limbus, angled 45° from the plane of the iris. In the second experiment (2-week study) the 3.5 mm incision was made 1 mm anterior to the limbus and parallel to the iris. The filled area indicates the laser beam irradiating incision 1,3 at a 45° angle.

Figure 2.

 

Typical traces of IOP with infusion time after treatment for PKD-treated and control eyes in the same rabbit. (○) Control eye; (•) PKD-treated (115 J/cm²) eye, which did not leak at the incision site until a pressure greater than 500 mm Hg, the upper measurement limit of the pressure transducer, was achieved.

Figure 3.

 

Relationship between laser fluence and IOP_L for corneal incisions treated with PKD and for control, unirradiated eyes. Immediately after treatment or sham treatment, eyes were infused with saline and the pressure measured at which leakage through the incision occurred. When the IOP_L was greater than 500 mm Hg, the upper limit of the pressure transducer, a value of 500 mm Hg was used for analysis. The number of corneas per group was 8, 4, 3, and 6 for fluences of 0, 115, 153, and 192 J/cm², respectively. *P < 0.0004 compared with the untreated control.

Figure 4.

 

Neovascularization after PKD or sham treatment. (A) Sham- and (B) PKD-treated (115 J/cm²) corneas of one rabbit 3 weeks after surgery. (C) Sham- and (D) PKD-treated (115 J/cm²) corneas of one rabbit 9 weeks after surgery.

Figure 5.

 

Neovascularization over a 10-week period after PKD or sham treatment. Assessment of neovascularization was made using a scale: 0, no evidence of neovascularization; 1, a single vessel; 2, a moderate number of vessels; and 3, neovascularization of the whole wound. Control, unirradiated (n = 9); PKD 115 J/cm² (n = 6); PKD 192 J/cm², (n = 3). The mean scores for the two PKD-treated groups were the same at weeks 2, 5, 8, and 10. Error bars, SD. P > 0.10 for comparison between the control and PKD-treated groups at all time points.

The authors thank Samir A. Melki and Bilal Khan for helpful advice and Béatrice Aveline, Arthur Foubert, and Christopher Amann for assistance.