Graefe's Archive for Clinical and Experimental Ophthalmology

, Volume 247, Issue 3, pp 331–334

Effect of intravitreal plasmin on vitreous removal through a 25-gauge cutting system in the rabbit in vivo


    • Department of OphthalmologyRWTH Aachen University
  • Jonathan Prenner
    • Associated Retinal Consultants
  • Motaz Alabdulrazzak
    • Associated Retinal Consultants
  • Wendy Dailey
    • NuVue Technologies
  • Michael Hartzer
    • NuVue Technologies
Retinal Disorders

DOI: 10.1007/s00417-008-1000-7

Cite this article as:
Hermel, M., Prenner, J., Alabdulrazzak, M. et al. Graefes Arch Clin Exp Ophthalmol (2009) 247: 331. doi:10.1007/s00417-008-1000-7



Intravitreal plasmin creates a posterior vitreous detachment, but may also liquefy the vitreous. This study measures the rate of vitreous removal from rabbit eyes after plasmin injection in vivo.


Intravitreal injections of 150 IU hyaluronidase (n = 5), 0.5 activity units (AU, n = 6) or 0.9 AU of streptokinase-activated human plasmin (n = four groups of 6) in 0.1 ml were performed in rabbits, the fellow eyes received 0.1 ml BSS. After 30 min (hyaluronidase), 30 min, 4 h, 12 h or 24 h (0.9 AU plasmin) or 24 h (0.5 AU plasmin), 1 ml of vitreous was removed from each eye without infusion, using a 25-gauge cutter and a standardized protocol. Animals were sacrificed after surgery.


Compared to fellow eyes, the average rate of vitreous removal was increased by hyaluronidase by 68.9 ± 6.3% (p < 0.05) and by 0.5 AU plasmin (24 h) by 26.8 ± 3.3% (p < 0.05). 0.9 AU of plasmin increased removal rates by 0.8 ± 10% (n.s.), 15.4 ± 6.3% (p < 0.05), 40.3 ± 3.1% (p < 0.05), and 71.9 ± 32.4% (p < 0.05) after 30 min, 4 h, 12 h and 24 h incubation respectively. The ratios of removal rates of treated/control eyes in the 0.9 AU groups showed a linear correlation with incubation time (r = 0.783, p < 0.0001).


Intravitreal plasmin increases the rate of vitreous removal in rabbits.


Enzyme administration and dosagePlasminPlasmin drug effectsPlasmin therapeutic useVitrectomyVitrectomy methodsVitreous body drug effectsVitreous body surgery


The objective of pars-plana vitrectomy is a complete removal of the vitreous, as its remnants on the retina may lead to proliferative complications. Mechanical vitreous peeling often proves difficult and traumatic to the retina, particularly in young patients [1]. Plasmin, a protease best known for its fibrinolytic properties, can cleave components of the vitreoretinal juncture [2] and, after intravitreal injection, lead to an atraumatic posterior vitreous disinsertion in vitro and in vivo [35]. Anecdotal clinical reports [68] and studies on pig cadaver eyes [9, 10] suggest that plasmin may also liquefy the vitreous and be of particular benefit in combination with small-gauge vitrectomy systems, as well as a possible component in a future “vitreolytic cocktail”.

This in-vivo study assesses plasmin-induced changes in the rate of vitreous removal from the rabbit eye using a small-gauge vitrectomy system.

Materials and methods

Plasminogen was isolated from human plasma using an affinity chromatography kit (NuVue, Keene, NH, USA) [11], and activated by streptokinase (Behringwerke, Marburg, Germany). Plasmin activity was measured spectrophotometrically using D-Val-Leu-Lys-pNA (Sigma-Aldrich, St. Louis, MO, USA) as substrate [12], at 22°C, and expressed in activity units (AU) previously used by us [3, 4, 13]. Lyophilized hyaluronidase type IV S (Sigma-Aldrich) was reconstituted to 150 IU/0.1 ml BSS.

Animals were used according to the "Principles of Laboratory Animal Care" (NIH publication No. 85-23, revised 1985), the OPRR Public Health Service Policy on the Humane Care and Use of Laboratory Animals (revised 1986) and the U.S. Animal Welfare Act, as amended, as well as the resolution of the ARVO on the use of animals in research. The study was approved by the institutional Animal Care Committee.

Preliminary experiments indicated an inter-individual variability and an age dependency in the rate of vitreous removal. Therefore, a paired approach using fellow eyes as controls, and age-matched animals were used. Based on our experience with vitreous procedures in rabbits, an intravitreal injection of 0.1 ml BSS does not lead to significant functional or anatomical impairment.

Thirty-five SPF New Zealand White rabbits (Covance, Kalamazoo, MI, USA) were randomly allocated into six groups (Table 1). Before procedures, the animals were anesthetized with an intramuscular injection of ketamine (35 mg/kg) and xylazine (5 mg/kg). A 30-gauge pars-plana intravitreal injection of 0.1 ml of the test solution into the left eye of each rabbit, and of 0.1 ml BSS into the right eye was performed. The injection sites were compressed to prevent reflux. In the 30-min surgery groups, the animals were re-dosed again with 17 mg/kg ketamine + 2.5 mg/kg xylazine 20 min after injection; in the other groups, Maxitrol ointment (Alcon) was applied to both eyes.
Table 1

Times required to remove 1000 μl of rabbit vitreous using a 25-gauge cutter and a standardized protocol in vivo (mean±SD)



Incubation time

Time to remove 1000 μl of vitreous [s]

Enzyme-treated eyes

Paired control eyes

Wilcoxon p-value

Hyaluronidase, 150 IU


30 min

133 ± 10.37

223.2 ± 18.2

p = 0.0431

Plasmin, 0.9 AU


30 min

221.5 ± 47.3

221.2 ± 44


Plasmin, 0.9 AU


4 h

242.2 ± 30.8

279.2 ± 34.7

p = 0.0277

Plasmin, 0.9 AU


12 h

212.7 ± 71.1

283 ± 43.7

p = 0.0464

Plasmin, 0.9 AU


24 h

160.2 ± 20.6

271.6 ± 34.7

p = 0.0277

Plasmin, 0.5 AU


24 h

187.2 ± 34

236.8 ± 40.2

p = 0.0277

Surgeries were started 30 min (hyaluronidase group), 30 min, 4 h, 12 h or 24 h (plasmin groups) after the intravitreal injections. One 25-gauge pars-plana sclerotomy was created in the nasal circumference, 1 mm from the limbus. Alcon Accurus vitrectomy tubing was calibrated with markings to show volumes of 150, 300, 450, 600, 750, 900 and 1000 μl. A vitrectomy (Alcon MVS-XX) was performed with a 25-gauge pneumatic cutter (DORC, Kingston, NH, USA). During surgery, the times at which the liquid reached each marking were recorded. Flow rates through the system were reproducible to a range of about 2 s for 1000 μl of test liquids with a wide range of viscosities.

To avoid infusion of liquid or air which would dilute the vitreous sample leading to artifacts, a collapse of the eye was allowed as the vitreous was being removed. To remove sufficient vitreous volume without engaging adjacent structures, thus avoiding artifacts by measuring merely central liquefied pockets, the cutter was moved slowly in a constant, circular clockwise (right eyes) or counterclockwise (left eyes) motion approximately at half distance between the retina and the lens, timed to 7 s per quadrant. In this way, 1000 μl of vitreous volume could be removed from each eye without damaging the lens or retina. Visualization was obtained by means of an EIBOS wide-angle system (Möller-Wedel, Wedel, Germany) attached to a Zeiss operating microscope. A suction of 200 mmHg and a low cutting rate (1/s) were chosen to prevent incarceration of the vitrector tip, while minimizing artifacts caused by mechanical breakup of the vitreous structure.

At the end of surgery, the animal was turned onto the other side and the fellow eye was operated in identical fashion. In each rabbit, the eye to be operated first was chosen at random. The same cutter–tubing combination was used for each pair of eyes. Before and after each surgery, the system was rinsed thoroughly until the passage time of 1000 μl BSS returned to 20 s, and filled with air. After both eyes were operated, rabbits were sacrificed.

Statistical evaluation of removal rates was performed using the Wilcoxon paired signed rank test (StatView), time-effect relationships using linear regression and correlation analysis (Fisher’s test, StatView), and intergroup comparison using ANOVA with the Tukey post-test (SigmaStat).


The mean times required to remove 1 ml of vitreous were significantly decreased 30 min after hyaluronidase injection, and by intravitreal plasmin in all groups with incubation times of 4 h or more (Table 1).

For further analyses, data were expressed as ratios of the vitreous removal rates for the treated vs the fellow BSS-injected eye. Analysis of variance showed a significant difference between groups (p < 0.001). 150 IU hyaluronidase displayed a significantly stronger effect after 30 min than 0.9 AU plasmin after 30 min, and 4 h hyaluronidase was also more effective than 0.5 AU plasmin after 24h (p < 0.05). Intravitreal injection of 0.9 AU of plasmin with 24 h incubation increased removal ratios significantly more than 0.9 AU plasmin after 30 min and 4 h, as well as 0.5 AU plasmin after 24 h (p < 0.05). 0.9 AU plasmin incubated for 12 h increased removal ratios more than after 30 min incubation (p < 0.05). Other intergroup comparisons did not reveal significant differences. In particular, no difference was seen between the hyaluronidase-treated group and the group treated with 0.9 plasmin for 24 h (difference of means = 0.00087, p = 1.0), demonstrating that plasmin was capable of a similar increase in vitreous removal rates as hyaluronidase, although with a different time-course. Results from the groups treated with 0.9 AU of plasmin showed a significant correlation between incubation time and removal ratios (r = 0.783, p < 0.0001, 95% CI: 0.555–0.920). The increase of vitreous removal ratios obtained as result of intravitreal injection of 0.9 AU plasmin was thus dependent on incubation time (Fig. 1).
Fig. 1

Relationship between incubation time and the effect of 0.9 AU plasmin injected into rabbit eyes on the vitreous removal through the 25-gauge vitrectomy system, expressed as ratios of vitreous removal rates of treated/paired control eyes, mean ± SD. Solid line: linear regression, r2 = 0.613; dashed lines: 95% CI. See Results section for significance values of intergroup comparisons


The human vitreous is a non-homogeneous structure, which given its small volume makes traditional measurement of its viscosity, elasticity, plasticity and thixotropy technically challenging [14]. Thus, alternative methods of measuring vitreous biomechanical properties of cadaver eyes have been developed, using dynamic light scattering [10], NMR imaging [15] or core vitrectomy and measurement of wet weight reduction of eyes [9]. However, these approaches may be biased by biochemical changes and autolytic processes [1618], potentially affecting the reaction of the vitreous to enzyme treatment. Thus, it is essential that results obtained in cadaver eyes are confirmed in vivo, particularly when long incubation times are required.

We developed a method to study vitreous removal rate from the rabbit eye by a standardized 25-gauge system which, while not aimed to reproduce real-life surgery, permits in vivo measurement of vitreous removal through a small-gauge system, and the use of longer incubation times. This approach was verified by the use of hyaluronidase, which is known to degrade rabbit vitreous [9, 15, 19], as a positive control. Although hyaluronidase does partially liquefy the vitreous, it is unable to reproducibly create a posterior vitreous detachment and facilitate the removal of the posterior hyaloid, a goal that plasmin has been shown to achieve [35].

Plasmin significantly increased the vitrous removal rates through by system at all incubation times longer than 30 minutes. The effect of 0.9 AU of plasmin after 24 h was comparable to the liquefaction induced by 150 IU of hyaluronidase after 30 min. The effect of plasmin appeared time-dependent, showing correlation between incubation time and ratios of vitreous removal rates of enzyme-treated vs control eyes.

Treatment with 0.5 AU of plasmin for 24 h resulted in a significantly slower rate of vitreous removal than with 0.9 AU. Although further doses have to be studied to confirm a dose dependency, similar findings were made using microplasmin in pig cadaver eyes [10]. In another study, a reduction of the wet weight of enucleated pig eyes injected with plasmin (3 and 30 CU) was seen when vitreous was removed by core vitrectomy after 1 or 3 h, albeit without dose dependency [9]. This suggests plasmin may degrade porcine vitreous. There are also anecdotal reports regarding liquefaction of human vitreous after plasmin injection [68]

The mechanism responsible for vitreous changes following plasmin injection is not well-understood. Plasmin activity in the rabbit vitreous decreases to about 15% of original activity within 2 hours after injection, and is undetectable at 24 hours [3]. The enzyme was reported to generate posterior vitreous detachment within 30–60 min [35], but its effects in our experiments follow a much slower time-course. As a result, plasmin is unlikely to be directly responsible for the increased vitreous removal rates seen at 12 and 24 hours. In addition, fibrillar collagen (Type I, II, and III) and hyaluronate are not directly cleaved by plasmin [20]. However, plasmin activates members of the matrix metalloproteinase (MMP) family [21, 22].

MMP-1 (collagenase-1) and MMP-2 (gelatinase A) are normally present as an inactive proenzyme in the rabbit and human vitreous and are activated by plasmin cleavage [21, 22]. MMP-1 degrades vitreous collagen type II [21]; MMP-2 degrades vitreous collagen type V/XI and type IX, but not type II [23]. Following activation by plasmin, the combination of these MMPs could result in the degradation of much of the vitreous collagen [2126]. We hypothesize that degradation of vitreous collagen by activated MMPs is likely to explain the increased vitreous removal rates following plasmin injection.

The main benefit of intravitreal plasmin is currently seen in its ability to generate vitreoretinal separation, but improved vitreous removal may prove to be an added benefit, shortening surgery, reducing surgical trauma and contributing to an expanded spectrum of indications for small-gauge vitrectomy systems.

Copyright information

© Springer-Verlag 2008