An Enzymatic Technique to Facilitate Air Separation of the Stroma–Descemet's Membrane Junction

Abstract

Purpose.: To describe an enzymatic technique that facilitates air separation of Descemet's membrane from the corneal stroma.

Methods.: Fresh human corneoscleral tissue was mounted on an artificial anterior chamber. In a control group, air was injected into the stroma. A second group received a stromal injection of 2.5 mg/mL collagenase type 2 in balanced salt solution that was left in the stroma for 1 hour and 15 minutes. A third group received an injection of 2.5 mg/mL collagenase type 2 in balanced salt solution followed 1 hour and 15 minutes later by an injection of air into the stroma. All injections were performed with a 27-gauge needle into the deep stroma without penetrating Descemet's membrane. Anterior segment optical coherence tomography (AS-OCT), histologic examination, and electron microscopy of the junction between the stroma and Descemet's membrane were performed. The trypan blue exclusion and TUNEL assays were used to study endothelial cell viability after collagenase incubation.

Results.: Injection of air or collagenase into the deep corneal stroma did not result in a reproducible separation of the stroma–Descemet's junction. In contrast, the stroma was easily and reproducibly separated from Descemet's membrane with a combination of intrastromal collagenase and air injection. The separation was confirmed by using light and electron microscopy. The cleavage plane seemed to be located between the junction of the posterior stroma and the anterior banded layer of Descemet's membrane. Trypan blue staining demonstrated the viability of endothelial cells after collagenase incubation. TUNEL assay confirmed excellent viability after collagenase+air separation.

Conclusions.: This technique facilitates the separation of Descemet's membrane from the stroma without affecting endothelial cell viability.

A technique that separates Descemet's membrane from the corneal stroma is potentially useful for the advancement of either anterior or posterior lamellar corneal transplantation. It is believed that a Descemet's membrane-to-stroma interface can yield a better final visual outcome than a stroma-to-stroma interface. A reproducible and simple technique for separation of Descemet's membrane from the corneal stroma would be of great importance for Descemet's membrane endothelial keratoplasty (DMEK) and deep anterior lamellar keratoplasty (DALK). 

A critical step for a successful DMEK transplant is the preparation of the donor graft consisting of an intact Descemet's membrane with viable endothelial cells. Tissue procurement and handling of the delicate Descemet's membrane may result in membrane tears and loss of valuable tissue. Different Descemet's membrane harvesting techniques, are reported, and modifications are frequently published. Ignacio et al. ¹ peeled Descemet's membrane from a donor corneoscleral rim mounted with the endothelium facing up on an artificial anterior chamber. Tappin ² peeled Descemet's membrane and used it in corneal transplantation. Lie et al. ³ described a novel harvest and implantation technique using trypan blue and forceps to remove the membrane from the stroma. Price et al. ⁴ reported the SCUBA (submerged corneas using backgrounds away) technique. Busin et al., ⁵ Studeny et al., ⁶ and McCauley et al. ⁷ described variations with a big bubble used to isolate donor Descemet's membranes. Recently, Kruse et al. ⁸ published a modified technique to isolate Descemet's membrane by using a razor blade and two forceps to improve endothelial cell survival. 

The collagenolytic activity of collagenase, obtained from Clostridium histolyticum, is currently approved by the U.S. Food and Drug Administration for the clinical management of Peyronie's disease ⁹ and is currently being evaluated for Dupuytren disease. ^10,11 We present an enzymatic technique using collagenase obtained from Clostridium histolyticum that creates a cleavage plane at the stroma–Descemet's membrane interface and facilitates air separation of Descemet's membrane from the corneal stroma. 

Materials and Methods

Descemet's Membrane: Stroma Separation Techniques

Fresh human corneoscleral tissue, preserved in storage medium (Optisol; Chiron Vision, Irvine, CA) and not suitable for transplantation was obtained from the Mississippi Lions Eyebank (Flowood, MS). The tissue was obtained and managed according to the Declaration of Helsinki. The corneoscleral tissue was mounted on a Barron artificial anterior chamber (Katena Products Inc., Denville, NJ). In a control group (n = 5), air was injected into the deep stroma, although most likely there was some variation in the depth at which it was injected in each eye. Air injection was performed with a 27-gauge needle with the beveled edge facing the Descemet's membrane as described in the big bubble technique. ¹² A small air bubble was left in the artificial chamber that helped to visualize Descemet's folds. A second group (n = 5) received a stromal injection of 0.3 mL of 2.5 mg/mL collagenase type 2 (Worthington, Lakewood, NJ) in commercial balanced salt solution (BSS; Alcon, Forth Worth, TX) also using a 27-gauge, beveled needle. The solution of collagenase was left in the stroma for 1 hour and 15 minutes at room temperature. Finally, a third group (n = 18) received an injection with 0.3 mL of 2.5 mg/mL collagenase (Figs. 1A–C) that was left in the stroma for 1 hour and 15 minutes, at which point an injection of 0.7 mL of air into the deep stroma with a 27-gauge, beveled needle was performed (Fig. 1D). Air was easily dispersed into the cornea stroma (see Supplementary Video S1). After the separation of Descemet's was attempted, corneal tissue was fixed in formaldehyde for paraffin sectioning or in OCT (optimal cutting temperature compound; Tissue-Tek; Sakura Finetek, Torrance, CA) for cryosections. 

Figure 1.

 

Surgical microscope view of the procedure. A corneoscleral rim is mounted on a Barron artificial chamber (A). A 27-gauge needle is advanced, bevel down, over an air bubble to facilitate visualization of Descemet's membrane (B). A collagenase solution is injected into the corneal stroma (C). Air is then injected into the corneal stroma (D).

Establishing the Appropriate Enzymatic Digestion Time

To better determine and establish the incubation time needed to create a separation cleavage plane using collagenase, human corneoscleral tissue mounted on the artificial chamber received an injection of 0.3 mL of 2.5 mg/mL collagenase in balanced salt solution. Two different collagenase doses were evaluated: 1.25 and 2.5 mg/mL. For each of the above concentrations, collagenase was injected into the stroma for 25 (n = 2), 50 (n = 3), 75 (n = 3), and 100 (n = 3) minutes. An injection of 0.7 mL of air into the deep stroma was then performed as described above. 

Light and Electron Microscopy and Anterior Segment OCT Imaging

Corneal sections were examined by light and transmission electron microscopy. For transmission electron microscopy, corneal tissue was fixed in Carson-Millonig buffered formaldehyde, postfixed in 1% osmium tetroxide, dehydrated through a graded ethanol series, and embedded in epoxy resin (Eponate 912; Ted Pella, Inc., Redwood, CA). Ultrathin (70 nm) sections were collected on copper grids and stained for 1 hour each with uranyl acetate and lead citrate before examination by transmission electron microscope (Leo 912; Carl Zeiss Meditec, Dublin, CA). Corneoscleral tissue after collagenase intrastromal injection, as described above, and after incubation for 75 minutes and air separation was mounted on a Barron artificial anterior chamber (n = 3) and evaluated with AS-OCT (Visante; Carl Zeiss Meditec) to further demonstrate Descemet's membrane separation. 

Trypan Blue Dye Exclusion Test

Three pairs of corneoscleral rims from three different donors were used to evaluate cell viability after Descemet's isolation by mounting three of the six rims on a Barron artificial anterior chamber, with storage medium (Optisol; Chiron Vision) filling the endothelial side of the chamber. Contralateral eyes, without any enzymatic treatment, were used as a control group to evaluate the effect of collagenase incubation and air injection in endothelial cell viability. Collagenase in the same concentration as above was injected into the deep stroma. One hour and 15 minutes later, Descemet's membrane was separated from the stroma as described earlier. Subsequently, Descemet's membrane flat mounts were prepared by placing the air-separated Descemet's membrane, still attached to the stroma at the periphery, on a Barron's punch. An 8.5-mm Descemet's membrane piece was trephined from the digested stroma. Each isolated membrane was cut into four pieces with a razor blade to facilitate flattening the membrane. Small Descemet's membrane sections were placed for 2 minutes on a glass slide and 0.4% trypan blue solution in balanced salt solution, to evaluate cell viability. In the control group, corneas were sectioned in four pieces and stained with trypan blue. The percentage of dead cells was estimated by counting the percentage of blue-stained cells in a 20× magnification in the areas that did not curl, avoiding the edge areas were the tissue was sectioned. 

Tissue Fixation and Sectioning and TUNEL Assay

Corneal sections (7 μm thick), after incubation for 75 minutes and air injection, were cut on a cryostat (HM 505M; Micron GmbH, Walldorf, Germany). Sections were placed on microscope slides (Superfrost Plus; Fisher Scientific, Pittsburgh, PA) and maintained at −85°C until staining was performed. To detect fragmentation of DNA associated with cell death, tissue sections were fixed in acetone at −20°C for 5 minutes, dried at room temperature for 5 minutes, and then placed in balanced salt solution. A fluorescence-based TUNEL assay was performed according to the manufacturer's instructions (ApopTag; Intergen Co, Purchase, NY). 

Results

A 2.5-mg/mL Collagenase Solution Facilitates Air Separation of Descemet's Membrane

Time exposure to collagenase seemed to be an important factor in determining the success of separation. No separation was possible with either 25 or 100 minutes of incubation. Separation was successful in only one of three eyes after 50 minutes' exposure, while complete separation was noted in three of three eyes after 75 minutes of exposure. 

A macroscopic examination in the collagenase+air group showed complete Descemet's membrane separation (Fig. 2A, side view, and Fig. 2B, front view) in 15 eyes. The separation extended to the peripheral cornea in most of the eyes. Central separation was noted in only four. Air bubbles were noted between the Descemet's membrane and the stroma. Separation of Descemet's membrane was noted in only one eye of the air-only group and in none of the collagenase-only group (Table 1). 

Figure 2.

 

Side view of the air bubble formed with stars delineating Descemet's membrane (A). Posterior view of the elevated Descemet's membrane and presence of air bubbles between the stroma and Descemet's membrane (B).

Table 1.

 

Summary of Methods Used for Separation

	Eyes, n	Successful Descemet's Detachment, n (%)	Histologic Appearance of Endothelial cells
Air injection	5	1 (20)	Normal, no endothelial cell damage
Collagenase injection	5	0 (0)	Normal, no endothelial cell damage
Collagenase/air injection	18	15 (83.3)	Normal, no endothelial cell damage

Histologic Confirmation of Descemet's Membrane and Stroma Separation

Histologic examination demonstrated normal corneal lamellae and the presence of air in the deep stroma close to Descemet's membrane in the air-injection group (Fig. 3A). Complete stroma–Descemet's membrane separation was not noted at high magnification (Fig. 4B). In the collagenase-only group, light microscopy showed collagenolysis at low magnification (Fig. 3B) with no evident stroma–Descemet's membrane separation. High magnification showed small clefts of separation at the stroma–Descemet's membrane junction (Fig. 4C). In contrast, the collagenase+air group showed evident Descemet's membrane separation (Figs. 3C). Areas suggestive of collagenolysis were evident at high magnification and a continuous separation cleft was noted (Fig. 4D). Anterior segment OCT was used to demonstrate the injection and location of collagenase in the stroma (Fig 5A) and the stroma–Descemet's membrane separation after collagenase incubation for 75 minutes and air injection (Figs 5B, 5C). Electron microscopy further confirmed the separation of Descemet's membrane from the stroma. Figure 6 (right) shows a perfect separation between the corneal stroma and the anterior banded layer of Descemet's membrane. However, although light microscopy suggested perfect separation, in some higher magnification EM sections, stromal attachments to Descemet's membrane were noted (Fig. 6, left). 

Figure 3.

 

Hematoxylin eosin staining demonstrating attempted separation of Descemet's membrane. Air infiltration into the corneal stroma after air injection (A). Loss of corneal lamellae architecture was noted after injection of collagenase and incubation for 75 minutes (B). Collagenase injection followed by intrastromal air injection produced air infiltration and clear separation of Descemet's membrane (C).

Figure 4.

 

Light microscopy evaluation of a cleavage plane after attempted separation of Descemet's membrane. Normal control showed no separation at the Descemet's membrane–stroma junction (A). Injection of stromal air showed no separation of the junction (B). Collagenase digestion for 75 minutes created areas of cleavage at the junction (★) (C). Collagenase incubation followed by air injection produced a cleavage plane at the junction. Arrow: the beginning of the cleavage plane (D).

Figure 5.

 

OCT confirmation of Descemet's membrane separation. A high-magnification image confirms the injection of collagenase into the stroma (A). A total Descemet's membrane detachment is imaged (B). A partial detachment is shown (C).

Figure 6.

 

Illustrative electron photomicrograph showing the separation of stroma from Descemet's membrane after collagenase incubation and air injection. Left microphotograph shows a very thin layer of stroma attached to the inner layer of Descemet's membrane (A). Full separation with no stroma attached to Descemet's membrane (B).

Cell Viability and TUNEL Assay

Evaluation of three separated Descemet's flat mounts revealed excellent endothelial cell survival after collagenase digestion. The trypan blue exclusion test showed a mean cell viability of 87.7% (range, 82.1%–93.9%). Compared with a mean viability of 91.2% (86.5%–92.7%) in the control untreated group (P = 0.1; not statistically significant; paired t-test [GraphPad, San Diego, CA]) (Fig.7A). To further explore the viability of the cells after collagenase incubation and air separation, a TUNEL assay was performed in three different eyes and further confirmed the excellent viability of the cells. It was surprising that no staining was noted in many sections studied, with very rare single-nuclei staining noted in 2 of 21 sections evaluated (Fig. 7B). Positive controls were run at the same time by an experienced laboratory investigator (Fig. 7C). 

Figure 7.

 

Cells remained viable after Descemet's separation. Trypan blue staining of an isolated flat mount at 20× magnification (A). TUNEL staining showed no nuclear staining in an isolated piece of Descemet's membrane. (★) Area of separation from the stroma. DAPI-stained keratocyte nuclei can be seen (B). Positive control: TUNEL-stained apoptotic nuclei (red); DAPI-stained live nuclei (blue) (C).

Discussion

Corneal haze continues to be a limiting factor in achieving the best potential visual acuity after corneal procedures. It has been hypothesized, but not completely validated, that haze formation can be prevented or decreased in the case of lamellar corneal transplantation if the tissue interface is located between the stroma–Descemet's membrane junction and not between the stroma–stroma junction. However, confocal studies show that interface haze resolves over time. There also seems to be a stromal thickness effect with any irregularities in stromal thickness (as are inevitable with stromal dissection) that alters either the anterior surface contour in DALK or the posterior surface contour in DSEK, with increasing aberrations as a result. 

Early surgical techniques for Descemet's membrane separation from the corneal stroma during DMEK ^{1 –4,8} have relied on the injection of air ^12,13 or balanced saline solution, ¹⁴ the viscoelastic effect, ¹⁵ or just mechanical force. Similar to DALK, a hypothetical advantage of DMEK over DSEK is a better final best corrected visual acuity due to less haze and scar formation if the interface is between Descemet's membrane and the stroma. 

We were unable to find the success rate of all experienced surgeons and centers in preparing Descemet's membrane grafts, to compare to our separation rate. Ignacio et al. ¹ reported a 100% success in peeling Descemet's membrane from stroma, whereas Busin et al. ⁵ reported a 95% rate of success with pneumatic separation. The technique described in this article may increase the success rate of separation for the novice surgeon or eye bank technician preparing corneal tissue. Further work is needed to standardize the time and concentration of intrastromal collagenase. In our study, at least 1 hour and 15 minutes of collagenase incubation at room temperature was needed to allow air separation of Descemet's membrane. 

We do not know why collagenase did not perforate Descemet's membrane in any of the studied eyes or whether microperforations were present. It was difficult to determine whether air was also escaping into the anterior chamber, since after air injection there was no visualization of the anterior chamber. It is possible that in cases in which we were unable to separate Descemet's membrane from the stroma, unintended microperforations were created with the 27-gauge needle during collagenase injection into the stroma or possibly as a result of Descemet's membrane collagenolysis. Although the action of collagenase 2 is not specific against a specific collagen type, clinical observations show that Descemet's membrane is more resistant to perforation, as noted with infectious corneal ulcers where a Descemetocele is present. 

We could not separate Descemet's membrane from the deep stroma if the incubation time was either too short or too long. Our current hypothesis is that a short incubation time does not allow enough collagenolysis and cleavage of the Descemet's–stroma junction. In contrast, longer exposure time (i.e., 100 minutes) may digest the anterior corneal stroma and allow air to leak through the anterior corneal surface. Once air leaks through the anterior corneal stroma, no pressure gradient is formed that forces Descemet's membrane posteriorly into the anterior chamber. We hypothesize that excessive anterior stromal digestion will not allow enough pressure gradient to form and force Descemet's membrane into the anterior chamber. Most likely, enzymatic activity will be enhanced at body temperature and a shorter incubation time will be needed if someday this method gains clinical application. 

Another intriguing finding was the difference in trypan blue staining and the almost absent TUNEL-positive cells found in histology sections. We ascribe this difference to the sensitivity of endothelial cells to manipulation during flat mount preparation. Descemet's membrane, as previously noted by many other investigators, tends to curl immediately after being separated from the stroma. Although the number of eyes studied with trypan blue was small, we believe that our technique does not cause any damage to endothelial cells, as shown by TUNEL assay and that the percentage of dead cells found with trypan blue in isolated membranes and in control eyes was mainly due to Descemet's membrane manipulation after isolation. 

Collagenase is a commonly used enzyme for the isolation of various cell types, including corneal epithelium and keratocytes, and does not need an inhibitor. Some of these isolation techniques require overnight incubation with collagenase. A great advantage of using collagenase is that once the stroma is separated from Descemet's membrane, the activity of collagenase becomes irrelevant. In the case of DMEK, once the membrane is isolated, placing it on balanced salt solution or preservation medium will inactivate the enzyme. 

In conclusion, we believe that our enzymatic technique would be a reliable technique for separating an intact Descemet's membrane for DMEK. This technique has the potential to be used for DALK, but several additional studies must be conducted. The use of collagenase in the recipient stroma as intended for DALK may be challenging due to collagenolysis in the recipient bed and the theoretical increased risk for keratoectasia. 

Supplementary Materials

Movie sv01, WMV - Movie sv01, WMV 

Footnotes

 Disclosure: E.M. Espana, None; B. Huang, None; J. Fratkin, None; J. Henegar, None