- Email: [email protected]
Intracameral Thrombin and the Corneal Endothelium
Intracameral Thrombin and the Corneal Endothelium Mark L. McDermott, M . D . , Henry F. Edelhauser, P h . D . , and Mark J. M a n n i s , M . D . We perfused the endothelia of isolated human corneas mounted in the specular micro scope with BSS Plus containing 1,000-U/ml or 100-U/ml dilutions of two commercially avail able topical thrombin preparations. Corneas perfused with thrombin at 1,000 U/ml showed intracellular and intercellular vacuole forma tion and altered junctional complexes. As list ed on the package inserts, the thrombin prepa rations contained preservatives and other additives that present a significant osmotic load in 1,000-U/ml preparations. Corneas per fused with 100-U/ml thrombin solutions showed a significant attenuation in their deswelling rate but no ultrastructural altera tions. One available thrombin preparation when diluted to 100 U/ml had a glycine con centration associated with previous retinal electroretinography changes. Polyacrylamide gel electrophoresis of one manufacturer's thrombin solution showed multiple high and low molecular weight constituents. Analysis of particulate contamination showed one 100U/ml thrombin preparation to have a large quantity of particulates. Although thrombin may be useful when applied topically as an aid in surgical hemostasis, its use intraocularly presents substantial concern regarding the preparation's purity, additives, contaminants, and adverse effects on ocular tissues. THROMBIN was used as an intraocular hemostatic agent by Savory 1 in 1947 and later by Hughes 2 in 1950. In 1986, Thompson and asso ciates 3 popularized the use of thrombin added to the infusion solution used during closed
Accepted for publication Aug. 2, 1988. From the Departments of Physiology and Ophthal mology, Medical College of Wisconsin, Milwaukee, Wis consin (Drs. McDermott and Edelhauser) and the Department of Ophthalmology, University of California-Davis, Sacramento, California (Dr. Mannis). Dr. Edelhauser is a Research to Prevent Blindness Sen ior Scientific Investigator. Reprint requests to Henry F. Edelhauser, Ph.D., De partment of Physiology, 8701 Watertown Plank Rd., Milwaukee, WI 53226.
414
vitrectomy when cutting fibrovascular mem branes in patients with proliferative diabetic retinopathy. The addition of bovine thrombin (both Thrombinar and Thrombostat) at concen trations of 100 U/ml resulted in a significant reduction in the duration of intraoperative in traocular bleeding. In 1987, Blacharski and Charles 4 reported that addition of thrombin (manufacturer not provided) to BSS Plus at a concentration of 100 U/ml resulted in reduced intraoperative a n d postoperative bleeding in patients w h o had retinopathy of prematurity stage V with total fractional retinal detachment undergoing closed lensectomy/vitrectomy and epiretinal membrane delamination. Mannis and associates 5 recently reported three cases of efficacious use of thrombin (100 U/ml, manu facturer not stated) in diabetic vitrectomy, trabeculectomy for neovascular glaucoma, and keratoplasty in a vascularized recipient bed. See also p. 485.
In addition to clinical efficacy reports, throm bin toxicity studies have also been reported. In 1985, DeBustros, Glaser, and Johnson 6 devel oped a rabbit model to investigate thrombin's effect on vitreous hemorrhage during closed vitrectomy. As part of the study, whole rabbit corneas were incubated for one hour at 37 C in BSS Plus containing 100 U/ml of thrombin (manufacturer not stated). The corneas were then stained with trypan blue or hematoxylin and eosin. No difference in trypan blue stain ing was observed when compared to control corneas incubated in BSS Plus. Light microsco py on hematoxylin and eosin preparations demonstrated no difference in cell number or structure. In that same study, in vivo darkadapted electroretinograms using ganzfeld stimuli were performed on rabbits following vitrectomy using either BSS Plus or BSS Plus supplemented with thrombin (100 U/ml) as an irrigant. Postoperatively, retinal sensitivity was depressed approximately 0.5 log units in the animals that received the BSS Plus contain ing thrombin during vitrectomy. In 1986, Thompson and associates 3 reported a 20% inci-
©AMERICAN JOURNAL OF OPHTHALMOLOGY 106:414-422, OCTOBER, 1988
Vol. 106, No. 4
Intracameral Thrombin
dence of inflammation (mainly sterile hypopyon formation) postoperatively in their series of patients undergoing vitrectomy with thrombin infusion at 100 U/ml. Recently Mannis and associates 5 presented data on the sheep corneal endothelial reaction to 1,000-U/ml and 100-U/ ml concentrations of thrombin (manufacturer unknown) in normal saline. Using combined trypan blue exclusion and alizarin red S stain ing of the endothelial monolayer, no toxicity was observed. We designed this study to evaluate two dif ferent commercial thrombin preparations on the human corneal endothelium using in vitro corneal endothelial perfusion. This technique allows continuous visual monitoring of the en dothelium and the measurement of corneal thickness, which can be correlated to endothe lial ultrastructure.
Material and Methods We tested two commercially available prepa rations of thrombin: Thrombinar (Armour Pharmaceutical) and Thrombostat (ParkeDavis). Both preparations are two-component systems consisting of lyophilized thrombin powder and a liquid diluent (Table 1). Both 1,000-U/ml and 100-U/ml dilutions of the lyoph ilized product were reconstituted using BSS Plus (Alcon Laboratories) rather than the sup plied diluent. Upon reconstitution, the 1,000U/ml concentrations of either manufacturer yielded a yellow-tinged, cloudy translucent so lution. Solutions of 100-U/ml concentration, while definitely clearer, were not crystal clear. After preparation of the solutions, whole
415
human corneas obtained from the Wisconsin Lions Eye Bank that were free of ocular disease were excised from the globe, mounted in the specular microscope, and the endothelia per fused for three hours with these solutions using a modification of the dual-chamber spec ular microscopic corneal perfusion apparatus of Dikstein and Maurice. 7 The donor criteria included a mean ± S.D. age of 69 ± 10 years, death to enucleation time of 2.1 ± 0.9 hours, and enucleation to experiment time of 14.3 ± 5.5 hours. In each experiment a pair of corneas were perfused: one cornea with the thrombin solution and the other with BSS Plus. During the course of the experiment, corneal thickness was measured every 15 minutes using a mi crometer attachment on the specular micro scope. A graph of absolute corneal thickness vs time was plotted for each experiment. For the 100-U/ml Thrombinar perfusions, a regression line for the test and control corneas was con structed. After perfusion, both corneas were fixed for transmission and scanning electron microscopy in 2.7% glutaraldehyde and phos phate buffer (pH 7.2, 330 mOsm) for at least eight hours at 4 C. To investigate the purity of Thrombinar and Thrombostat, freshly reconstituted samples were submitted for sodium dodecyl sulfate polyacrylamide gel electrophoresis. The prote ase inhibitors, 0.1 M leupeptin and 0.1 M phenolmethylsulfonylfluoride, were added to prevent autodegradation of the samples during electrophoresis. Analysis of particulate levels was performed using a Coulter counter config ured for a 5- to 20-μπι particle size. Dilutions of 100 U/ml of either Thrombinar or Thrombostat were prepared in an identical fashion at the same time and analyzed immediately. To mini-
TABLE1 COMPOSITION OF LYOPHILIZED COMPONENT VIALS OF 5,000 U THROMBIN PREPARATIONS CONCENTRATION AT 1,000 U/ML DILUTION
AMOUNT/VIAL COMPOUND
THROMBINAR"
THROMBOSTAT*
Bovine thrombin Calcium Glycine Mannitol Sodium chloride
5,000 U
5,000 U 5 mg 39 mg
— — 18 mg 18 mg
— 8 mg
THROMBINAR
1,000 U
— — 360 mg/dl 360 mg/dl
'Diluent vial contains isotonic NaCI containing 0.02% benzethonium chloride. 'Diluent vial contains "preservative free" sterile water for injection.
THROMBOSTAT
1,000 U 100 mg/dl 780 mg/dl
— 160 mg/dl
CONCENTRATION AT 100 U/ML DILUTION THROMBINAR
100 U
THROMBOSTAT
— —
100 U 10 mg/dl 78 mg/dl
36 mg/dl 36 mg/dl
16 mg/dl
—
416
October, 1988
AMERICAN JOURNAL OF OPHTHALMOLOGY
mize inadvertent introduction of particles, the following procedure was used. BSS Plus in 500-ml bottles was reconstituted according to the manufacturer's instructions. Four hundred milliliters of the reconstituted solution were removed and a 20-ml aliquot was counted and recorded as the background particulate load. A small volume of the remaining 100 ml of BSS Plus was withdrawn with a new syringe and introduced into a sealed vial containing 10,000 units of lyophilized thrombin via the rubber septum. The contents were mixed and the solu tion was withdrawn (using the same syringe and needle) and reintroduced into the BSS Plus bottle containing the balance of the 100 ml. A solution of 100 U/ml was thus obtained. Ali quote of these 100-U/ml solutions were ana lyzed using the Coulter counter. For this analy sis the instrument was calibrated using 5-, 10-, and 20-μπι beads, creating a 5- to 20-μπι sized particle window.
Results Corneal endothelial perfusion using both 1,000-U/ml thrombin preparations resulted in an immediate and protracted thinning of the cornea compared to its control (Fig. 1). Observation of the endothelial monolayer through the specular microscope showed inter cellular and intracellular vacuole formation. The thrombin-perfused cells appeared vacuolated with junctional abnormalities compared to the BSS Plus control (Figs. 2 and 3). In a comparision of the p H and osmolality characteristics of various concentrations of
each thrombin preparation, the 1,000-U/ml thrombin solutions had increased osmolality (Table 2). Further studies were performed with 100-U/ ml dilutions of Thrombinar (which lacks gly cine). Using an identical protocol to the one previously described, the endothelia of four pairs of human corneas (one cornea as test and the other cornea as control) were perfused with 100-U/ml dilutions (in BSS Plus) of Thrombinar (pH 7.55, 314 mOsm/kg; Table 2). An average corneal deswelling rate was determined for the BSS Plus and 100-U/ml thrombin groups (Fig. 4). The Thrombinar-perfused corneas snowed a limited ability to deturgesce when compared to BSS Plus (P < .02, paired f-test). The thrombinperfused corneas showed a deswelling rate of about 5 |xm/hour compared to a 19^m/hour deswelling rate for the BSS Plus-perfused cor neas. During perfusion, the endothelium showed slight intracellular darkening and a surface ground glass appearance. Scanning electron micrographs, however, did not show alterations in junctional morphology in the eyes perfused with 100-U/ml Thrombinar. Transmission electron micrographs of Thrombinar-perfused corneas appeared similar to control corneas. Gel electrophoresis results are presented in Figure 5. Because the total protein content for a specified activity varied between the different manufacturer's preparations, dissimilar dilu tions of the thrombin solutions were used. Lane 1 shows the electrophoretic result of a 50-U/ml Thrombostat (Parke-Davis) solution. At the top of the running gel there are aggre gated high molecular weight constituents; within the running gel numerous distinct
800
800
E 3. ! S A—· A - Ì 700-·
A
0)
c
D c
o o
600
It
■
600o,
\
,Ο—O—O—O—O—O—O-
% -
· .• . . · A· · A ·BSS· Plus ·
500
400
700
\ .
•
1
0
1
I
1
1
1
1
1
σ
•Thrombostat 1000 U / m l
1
1
1
1
15 30 45 60 75 90 105120135150165180195
Time (min)
A A BSS Plus O — O T h rombinar 1000 U/ml
500-
o
h
O
400
-O-O
60
120
180
Time (min)
Fig. 1 (McDermott, Edelhauser, and Mannis). Representative plots of human corneal thickness values over time when perfused with a 1,000-U/ml solution of Thrombostat (left) and a solution of 1,000 U/ml of Thrombinar (right) compared to BSS Plus control.
' **s*n"!:"*'""' Λ" ''"':Ύ' ■' -' · ■ ■ " "· ■ ρ ' ^ ^ ^ ^ φ ^ ^ ^ Μ & ^ Φ ^ ,
Fig. 2 (McDermott, Edelhauser, and Mannis). Left, Scanning and transmission electron micrographs of a human cornea perfused with a 1,000-U/ml Thrombostat solution. The cells appear shrunken with " p o r e " formation on scanning electron microscopy (top) (x 500). Transmission electron micrographs show intercellular vacuoles and junctional disruption (middle and bottom) (x 4,000). Right, Scanning and transmission electron micrographs of a human cornea of the companions of figures on the left perfused with BSS Plus. Scanning electron micrographs (top) (X500) and transmission electron micrographs (middle and bottom) (x 5,000) demonstrate a normal-appearing monolayer with intact junctional morphology.
418
October, 1988
AMERICAN JOURNAL OF OPHTHALMOLOGY
~"V-?- * ! 1 > ^
tsßm^
ψ&Ρ·3&
f
J'.. fit
Vit' ■&$■ -i ' : rff-
Λ*
■ —»-<■
.
·\ ν *
".■"I*
'
·+
*.*S »
&ÏYW
Fig. 3 (McDermott, Edelhauser, and Mannis). Left, Scanning and transmission electron micrographs of a human cornea perfused with 1,000-U/ml Thrombinar solution. The cells appear shrunken with creeping of the apical cell membrane over adjacent cells on scanning electron microscopy (top) (x 1,000). Transmission electron micrographs (middle and bottom) (x 6,000) show excessive junctional overlap. Right, Scanning and transmis sion electron micrographs of the companion cornea shown on the left perfused with BSS Plus. Scanning electron micrograph (top) (x 1,000) and transmission electron micrographs (middle and bottom) (x5,000) demonstrate an ultrastructurally normal endothelial monolayer.
Vol. 106, No. 4
Intracameral Thrombin
TABLE 2 pH AND OSMOLALITY OF THROMBIN SOLUTIONS MEAN ± S.D. OSMOLALITY MEAN ± S.D. pH
PREPARATION
Thrombinar 1,000 U/ml* Thrombinar 100 U/mP Thrombostat 1,000 U/ml* BSS Plus
7.86 7.55 7.43 7.46
± ± ± ±
0.9 0.06 0.1 0.05
(MOSM/KG)
407 314 492 302
± ± ± ±
1 5 1 10
»Diluted in BSS Plus.
bands are present as well as a large amount of interband protein. Bands A and B have estimat ed molecular weights of 44.5 kd and 37.5 kd. Lane 2 shows the result obtained with a 200-U/ ml solution of Thrombinar (Armour). Bands A and B are the most prominent, but several faint low molecular weight constituents are also present. Note the distinct lack of interband protein deposition and absence of high molecu lar weight aggregates at the top of the running gel. Since protease inhibitors were present in the thrombin solutions during electrophoresis, autodegradation of the polypeptide chain to produce low molecular weight products is un likely. In an analysis of the 100-U/ml dilutions of Thrombinar and Thrombostat for particulate contamination, as well as the background par ticulate load for the BSS Plus diluent, the 100U/ml Thrombostat preparation had a particu late load 30 times greater than that seen in the Thrombinar solution (Table 3).
419
Discussion Intraocular use of products not designated as such or evaluated for intraocular toxicity is potentially dangerous. While topically applied thrombin may be a useful adjunct to control surface bleeding during ocular surgery on a closed globe, it is important to realize that no manufacturer has isolated thrombin for use in any way other than topically. As such, consid erations as to purity, contaminants, pH, osmolality, and vehicles used are based on topical use only. At this time, no thrombin preparation has been approved by the Food and Drug Ad ministration for intracameral use. We found that high-concentration solutions (that is, 1,000 U/ml of either Thrombinar or Thrombostat) resulted in endothelial morphologic alterations characterized by intercellular vacuole forma tion and altered apical junctions. These findings are identical to the endothelial chang es reported in hyperosmotic media 810 and in clude intercellular space dilatation, intracellular vacuole formation, and lysis of intercellular bridges. All of these conditions at one time or another were observed in corneas perfused with thrombin. It appears likely, therefore, that the changes observed when perfusing with 1,000-U/ml thrombin solution are a result of the solution's hypertonicity. Our data suggest that the addi tives in the vial of lyophilized thrombin are the osmotic agents (Tables 1 and 2). A previous study by Mannis and associates 5 showed that
20
E
0
Fig. 4 (McDermott, Edelhauser, and Mannis). Changes in human corneal thickness when paired cor neas are perfused with either Thrombinar 100 U/ml (open circles) or BSS Plus (filled triangles). The solid lines represent first order re gression lines. Note the significant flattening of the Thrombinar's deswelling curve compared to its BSS Plus-perfused companion cor-
co CO CD
-20 +
c υ
-40
σ α> e
-60
ι_ O
υ
-80 -100
Mean±SEM n=4 A BSS Plus Control y=-(19.28±2.60)x + 10.81 OArmour Thrombin 100/ml y=-(5.24±2.68)*x + 6.25 p LO.02 60
120
time(min)
180
October, 1988
AMERICAN JOURNAL OF OPHTHALMOLOGY
420
2
1
3
92.5Kd ( Phosphorulase B) .ÊÊfaMaÊL·
1
1 *
A *B r*
66.2Kd ( Bovine Serum Albumin) 45.0Kd(0valbumin)
Λ
^gjy|^&
!
mm»
.,,..*
Ξ
Ì
31. OKd (Carbonic Anhydrase)
2l.5Kd (SoybeanTrypsin Inhibitor) 1 l4.4Kd(Lysozyme)
Fig. 5 (McDermott, Edelhauser, and Mannis). Results of 12% polyacrylamide gel electrophoresis. Lane 1, Thrombostat (ParW-Davis) at an activity of 50 U/ml; Lane 2, Thrombinar (Armour) at an activity of 200 U/ml; Lane 3, molecular weight standards. Bands A and B in Lanes 1 and 2 may represent thrombin.
sheep corneal endothelium bathed in 1,000-U/ ml thrombin (in normal saline) appeared nor mal with alizarin red S and trypan blue stain ing. However, this is in distinct contrast to the effects on human corneal endothelium where bathing in normal saline or hypertonicity may result in ultrastructural alterations. 8 The alterations observed may be reversible and not cause permanent loss of endothelial function. However, imposition of an osmotic stress on an eye that may have a preexisting endotheliopathy, such as that seen in diabetes mellitus as reported by Schultz and associ ates, 11 could result in endothelial decomposi tion. This is especially important since vitrectomy in diabetic patients using solutions containing thrombin for intraocular hemostasis has been reported. 3 · 5 In these studies, however, a lower dose (100 U/ml) was used so osmotic effects would be less. Even when using a lower concentration of thrombin (100 U/ml), the presence of certain
additives may prove toxic. Thrombostat con tains large quantities of the amino acid glycine, and a 100-U/ml dilution of Thrombostat (the dose used in all cited studies) contains 78 mg/dl of glycine. The presence of glycine at this con centration is especially important in light of the recent study by Creel, Wang, and Wong12 linkTABLE 3 PARTICULATE ANALYSIS OF 100 U/ML THROMBIN SOLUTIONS PARTICLES/ML* SOLUTION
RANGE
MEAN
S.D.
Thrombinar 100 U/ml Diluent for Thrombinar' Thrombostat 100 U/ml Diluent for Thrombostaf
662-706 154-180 20,830-21,540 254-320
694 171 21,221 281
15 10 270 20
"Using 5- to 20-μηη window setting, 100-/*m aperture. *BSS Plus.
Vol. 106, No. 4
Intracameral Thrombin
ing transient blindness (visual acuity of count ing fingers at 1 meter) with increases in serum glycine levels during postoperative bladder ir rigation after prostatic resection. In their pa tients a graying of vision was temporally asso ciated with loss of oscillatory potentials and 30-Hz flicker following electroretinography with serum glycine levels in excess of 30 mg/dl. Clearly, intracameral use of this preparation presents a much higher local concentration of glycine (Table 1). Perfusion of a 100-U/ml thrombin solution results in a 78-mg/dl glycine level, at least 2.5 times the serum level associat ed with visual symptoms. Therefore, a toxic effect is possible on the neurosensory retina (via inhibitory neurotransmission). Data on long-term toxicity or reversibility of exposure to glycine has not been reported to date. The other thrombin preparation, Thrombinar, does not contain glycine. However, perfusion of this preparation at 100 U/ml appears to be responsible for a significant (P < .02) loss in the ability of the cornea to deturgesce when perfused with BSS Plus. The BSS Plus-perfused corneas deturgesced at a significantly greater rate of 19 |xm/hour as com pared to the thrombin-perfused corneas' deturgescence rate of 5 μιτι/hour. The reversi bility and long-term effect of exposure at this concentration is indeterminant. It appears, however, that short-term irrigation (less than one hour) has little effect on corneal deswelling, but longer durations of perfusion may result in a cumulative attenuation in endothelial function as reflected by the significantly different rates of deswelling. This could be of particular importance if the thrombin is left in the eye after completion of the surgical proce dure. In addition to the tonicity and presence of additives, the protein composition of a specific activity thrombin solution must be considered. The activity in units of a thrombin solution is determined by comparing its ability to clot a plasma extract or purified fibrinogen solution to that of a National Bureau of Standards lot of thrombin. This in no way informs the user of the thrombin product what constituents and concentrations are present. The specific activity is only a meas.ure of the ability to cause clot ting. Since all thrombin solutions are labeled in terms of activity, interlot variability of extrane ous proteins, particulates, and the like can exist. Therefore, even if the surgeon uses the same 100-U/ml dilution of a particular thrombin solution, there is no guarantee that the same solution (in terms of contaminants and particu
421
lates) is present. Herein we found that two thrombin preparations can contain different amounts of protein with presumed nonthrombin activity. Bands A and B are present in the preparations of both manufacturers, and presumably thrombin activity exists in one or both of those bands. This is in agreement with previous reports of multiple bovine thrombin components. 1314 Band B, with a mo lecular weight estimate of 37.5 kd, is close to the reported molecular weight of 39 kd for bovine thrombin. 14 It appears, therefore, that identical activity solutions of thrombin by different manufactur ers have different electrophoretic patterns and amounts of nonthrombin protein. These substances may possess activity in their own right and serve as immunogenic stimuli. More over, during preparation of the crude thrombin product, a variety of particulate matter may be introduced and subsequently perfused intraocularly. Our Coulter counter results showed that certain preparations, even when diluted, con tained large quantities of particulates. When using Thrombostat, irrigation during vitrectomy with 100 ml of the 100-U/ml solution would allow over two million particles to enter the eye. It is entirely possible, therefore, that the hypopyon formation as reported by Thompson and associates 3 was caused by the relatively crude purity of the commercial thrombin preparation. Therefore, one needs to consider whether the intraocular milieu is the appropriate chamber for this product. Even when using an ultra-pure crystalline thrombin preparation one must consider its potential interactions. Thrombin is a potent mitogen for fibroblasts, is chemotactic for neutrophils, and causes smooth muscle contrac tion. Upon exposure to thrombin solutions, vascular endothelium undergoes permeability increases, platelets aggregate with the subse quent generation of eicosanoids, and fibronectin and laminin undergo enzymatic cleavage. Thrombin has been shown to bind to neural tissue causing an increase in cellular cGMP concentration. 15 It is apparent that thrombin is a molecule with enzymatic and receptor activity; however, it is not known whether the intact human corneal endothelium has thrombin receptors. In cultured cell lines good evidence exists for thrombin receptors. Isaacs and associates 16 demonstrated a specific 40,500-dalton, surface membrane-bound, α-thrombin receptor in cul tured bovine corneal endothelium. This recep tor is saturable and specific, but as yet does not
422
AMERICAN JOURNAL OF OPHTHALMOLOGY
h a v e a d e m o n s t r a b l e p h y s i o l o g i c effect o n thrombin-receptor coupling. Cellular r e c e p t o r d e n s i t y in t h e b o v i n e c o r n e al e n d o t h e l i u m m o d e l is a b o u t 100,000 r e c e p tors p e r cell a n d s h o w s s a t u r a t i o n at 2.0 to 2.5 U/ml of t h r o m b i n . In further s t u d i e s of this receptor, Savion a n d associates 1 ' s h o w e d t h a t b o v i n e c o r n e a l e n d o t h e l i u m cells actively inter nalize the r e c e p t o r - a - t h r o m b i n c o m p l e x via a d s o r p t i v e e n d o c y t o s i s . In this m o d e l , internalization of t h e covalent t h r o m b i n r e c e p t o r c o m p l e x r e s u l t s in a n i n c r e a s e d n u m b e r of cellular r e c e p t o r site d e n s i t y . It a p p e a r s , t h e r e f o r e , n o t w i t h s t a n d i n g t h e differences b e t w e e n species a n d t h e differences involved in cell c u l t u r e , that a h u m a n corneal e n d o t h e l i a l r e c e p t o r could exist. C u r r e n t s t u d ies u s i n g 1 2 5 I-human t h r o m b i n to e v a l u a t e h u m a n c o r n e a l e n d o t h e l i a l b i n d i n g are u n d e r w a y . As s u c h , t h e possibility of its p r e s e n c e merits careful c o n s i d e r a t i o n before i n t r o d u c tion of t h r o m b i n i n t r a c a m e r a l l y . This is e s p e cially w o r r i s o m e w i t h t h e i n t r o d u c t i o n of t h r o m b i n into t h e diabetic eye w i t h a n incom petent blood-aqueous barrier and neovascularization. In this s e t t i n g , a m u l t i t u d e of interac tions m a y occur in a d d i t i o n to t h e o b v i o u s c o n v e r s i o n of fibrinogen to fibrin. While m a n y of o u r c o n c e r n s could be v i e w e d as p o t e n t i a l p r o b l e m s , t h e r e are definite differences be t w e e n the two m a n u f a c t u r e r ' s t h r o m b i n p r o d u c t s . M o r e o v e r , u s i n g a sensitive t e c h n i q u e , n a m e l y in vitro corneal p e r f u s i o n , a 100-U/ml thrombin solution, the dose most commonly u s e d , h a s b e e n s h o w n to d e c r e a s e t h e d e s w e l l i n g rate of t h e p e r f u s e d h u m a n c o r n e a during prolonged perfusion.
ACKNOWLEDGMENT
Sally S. T w i n i n g , P h . D . , a n d Patricia M. Wil s o n , M . S . , D e p a r t m e n t of B i o c h e m i s t r y , M e d i cal College of W i s c o n s i n , p e r f o r m e d t h e gel electrophoresis studies.
References 1. Savory, W.: Some uses of thrombin and fibrino gen in ophthalmic surgery. Trans. Ophthalmol. Soc. U.K. 67:323, 1947. 2. Hughes, W. L.: Use of thrombin in the anterior
October, 1988
chamber to control hemorrhage. Acta XVI Concilium Ophthalmologicum 12:99, 1950. 3. Thompson, J. T., Glaser, B. M., Michels, R. G., and DeBustros, S.: The use of thrombin to control hemorrhage during vitrectomy. Ophthalmology 93:279, 1986. 4. Blacharski, P. A., and Charles, S. T.: Thrombin infusion to control bleeding during vitrectomy for stage V retinopathy of prematurity. Arch. Ophthal mol. 105:203, 1987. 5. Mannis, M. S., Sweet, E., Landers, M. B., and Lewis, R. A.: Uses of thrombin in ocular surgery. Effect on the corneal endothelium. Arch. Ophthal mol. 106:251, 1988. 6. DeBustros, S., Glaser, B. M., and Johnson, M. A.: Thrombin infusion for the control of intraocu lar bleeding during vitreous surgery. Arch. Ophthal mol. 103:837, 1985. 7. Dikstein, S., and Maurice, D. M.: The metabolic bases to the fluid pump in the cornea. J. Physiol. 221:29, 1972. 8. Edelhauser, H. F., Hanneken, A. M., Pederson, H. J., and Van Horn, D. L.: Osmotic toler ance of rabbit and human corneal endothelium. Arch. Ophthalmol. 99:1281, 1981. 9. Van Horn, D. L., Schultz, R. O., and Edelhauser, H. F.: Corneal cryopreservation. Altera tions in endothelial intercellular spaces. Am. J. Oph thalmol. 63:454, 1969. 10. Sherrard, E. S.: Characterization of changes observed in corneal endothelium with the specular microscope. Invest. Ophthalmol. Vis. Sci. 17:322, 1978. 11. Schultz, R. O., Matsuda, M., Yee, R. W., Edelhauser, H. F., and Schultz, K. J.: Corneal endo thelial changes in type I and type II diabetes mellitus. Am. J. Ophthalmol. 98:401, 1984. 12. Creel, D. J., Wang, J. M., and Wong, K. C : Transient blindness associated with transurethral re section of the prostate. Arch. Ophthalmol. 105:1537, 1987. 13. Rosenberg, R. D., and Waugh, D. F.: Multiple bovine thrombin components. J. Biol. Chem. 245:5049, 1970. 14. Mann, K. G., Yip, R., Heldebrant, C. M., and Fass, D. N.: Multiple active forms of thrombin. J. Biol. Chem. 248:1868, 1973. 15. Walz, D. A., Fenton, J. W., and Shuman, M. A. (eds.): Bioregulatory functions of thrombin. Ann. N.Y. Acad. Sci. 485:5, 1986. 16. Isaacs, J. D., Savion, N., Gospodarowicz, D., Fenton, J. W., and Shuman, M. A.: Covalent binding of thrombin to specific sites on corneal endothelial cells. Biochemistry 20:398, 1981. 17. Savion, N., Isaacs, J. D., Gospodarowicz, D., and Shuman, M. A.: Internalization and degradation of thrombin and up regulation of thrombin binding sites in corneal endothelial cells. J. Biol. Chem. 256:4514, 1981.
Accepted for publication Aug. 2, 1988. From the Departments of Physiology and Ophthal mology, Medical College of Wisconsin, Milwaukee, Wis consin (Drs. McDermott and Edelhauser) and the Department of Ophthalmology, University of California-Davis, Sacramento, California (Dr. Mannis). Dr. Edelhauser is a Research to Prevent Blindness Sen ior Scientific Investigator. Reprint requests to Henry F. Edelhauser, Ph.D., De partment of Physiology, 8701 Watertown Plank Rd., Milwaukee, WI 53226.
414
vitrectomy when cutting fibrovascular mem branes in patients with proliferative diabetic retinopathy. The addition of bovine thrombin (both Thrombinar and Thrombostat) at concen trations of 100 U/ml resulted in a significant reduction in the duration of intraoperative in traocular bleeding. In 1987, Blacharski and Charles 4 reported that addition of thrombin (manufacturer not provided) to BSS Plus at a concentration of 100 U/ml resulted in reduced intraoperative a n d postoperative bleeding in patients w h o had retinopathy of prematurity stage V with total fractional retinal detachment undergoing closed lensectomy/vitrectomy and epiretinal membrane delamination. Mannis and associates 5 recently reported three cases of efficacious use of thrombin (100 U/ml, manu facturer not stated) in diabetic vitrectomy, trabeculectomy for neovascular glaucoma, and keratoplasty in a vascularized recipient bed. See also p. 485.
In addition to clinical efficacy reports, throm bin toxicity studies have also been reported. In 1985, DeBustros, Glaser, and Johnson 6 devel oped a rabbit model to investigate thrombin's effect on vitreous hemorrhage during closed vitrectomy. As part of the study, whole rabbit corneas were incubated for one hour at 37 C in BSS Plus containing 100 U/ml of thrombin (manufacturer not stated). The corneas were then stained with trypan blue or hematoxylin and eosin. No difference in trypan blue stain ing was observed when compared to control corneas incubated in BSS Plus. Light microsco py on hematoxylin and eosin preparations demonstrated no difference in cell number or structure. In that same study, in vivo darkadapted electroretinograms using ganzfeld stimuli were performed on rabbits following vitrectomy using either BSS Plus or BSS Plus supplemented with thrombin (100 U/ml) as an irrigant. Postoperatively, retinal sensitivity was depressed approximately 0.5 log units in the animals that received the BSS Plus contain ing thrombin during vitrectomy. In 1986, Thompson and associates 3 reported a 20% inci-
©AMERICAN JOURNAL OF OPHTHALMOLOGY 106:414-422, OCTOBER, 1988
Vol. 106, No. 4
Intracameral Thrombin
dence of inflammation (mainly sterile hypopyon formation) postoperatively in their series of patients undergoing vitrectomy with thrombin infusion at 100 U/ml. Recently Mannis and associates 5 presented data on the sheep corneal endothelial reaction to 1,000-U/ml and 100-U/ ml concentrations of thrombin (manufacturer unknown) in normal saline. Using combined trypan blue exclusion and alizarin red S stain ing of the endothelial monolayer, no toxicity was observed. We designed this study to evaluate two dif ferent commercial thrombin preparations on the human corneal endothelium using in vitro corneal endothelial perfusion. This technique allows continuous visual monitoring of the en dothelium and the measurement of corneal thickness, which can be correlated to endothe lial ultrastructure.
Material and Methods We tested two commercially available prepa rations of thrombin: Thrombinar (Armour Pharmaceutical) and Thrombostat (ParkeDavis). Both preparations are two-component systems consisting of lyophilized thrombin powder and a liquid diluent (Table 1). Both 1,000-U/ml and 100-U/ml dilutions of the lyoph ilized product were reconstituted using BSS Plus (Alcon Laboratories) rather than the sup plied diluent. Upon reconstitution, the 1,000U/ml concentrations of either manufacturer yielded a yellow-tinged, cloudy translucent so lution. Solutions of 100-U/ml concentration, while definitely clearer, were not crystal clear. After preparation of the solutions, whole
415
human corneas obtained from the Wisconsin Lions Eye Bank that were free of ocular disease were excised from the globe, mounted in the specular microscope, and the endothelia per fused for three hours with these solutions using a modification of the dual-chamber spec ular microscopic corneal perfusion apparatus of Dikstein and Maurice. 7 The donor criteria included a mean ± S.D. age of 69 ± 10 years, death to enucleation time of 2.1 ± 0.9 hours, and enucleation to experiment time of 14.3 ± 5.5 hours. In each experiment a pair of corneas were perfused: one cornea with the thrombin solution and the other with BSS Plus. During the course of the experiment, corneal thickness was measured every 15 minutes using a mi crometer attachment on the specular micro scope. A graph of absolute corneal thickness vs time was plotted for each experiment. For the 100-U/ml Thrombinar perfusions, a regression line for the test and control corneas was con structed. After perfusion, both corneas were fixed for transmission and scanning electron microscopy in 2.7% glutaraldehyde and phos phate buffer (pH 7.2, 330 mOsm) for at least eight hours at 4 C. To investigate the purity of Thrombinar and Thrombostat, freshly reconstituted samples were submitted for sodium dodecyl sulfate polyacrylamide gel electrophoresis. The prote ase inhibitors, 0.1 M leupeptin and 0.1 M phenolmethylsulfonylfluoride, were added to prevent autodegradation of the samples during electrophoresis. Analysis of particulate levels was performed using a Coulter counter config ured for a 5- to 20-μπι particle size. Dilutions of 100 U/ml of either Thrombinar or Thrombostat were prepared in an identical fashion at the same time and analyzed immediately. To mini-
TABLE1 COMPOSITION OF LYOPHILIZED COMPONENT VIALS OF 5,000 U THROMBIN PREPARATIONS CONCENTRATION AT 1,000 U/ML DILUTION
AMOUNT/VIAL COMPOUND
THROMBINAR"
THROMBOSTAT*
Bovine thrombin Calcium Glycine Mannitol Sodium chloride
5,000 U
5,000 U 5 mg 39 mg
— — 18 mg 18 mg
— 8 mg
THROMBINAR
1,000 U
— — 360 mg/dl 360 mg/dl
'Diluent vial contains isotonic NaCI containing 0.02% benzethonium chloride. 'Diluent vial contains "preservative free" sterile water for injection.
THROMBOSTAT
1,000 U 100 mg/dl 780 mg/dl
— 160 mg/dl
CONCENTRATION AT 100 U/ML DILUTION THROMBINAR
100 U
THROMBOSTAT
— —
100 U 10 mg/dl 78 mg/dl
36 mg/dl 36 mg/dl
16 mg/dl
—
416
October, 1988
AMERICAN JOURNAL OF OPHTHALMOLOGY
mize inadvertent introduction of particles, the following procedure was used. BSS Plus in 500-ml bottles was reconstituted according to the manufacturer's instructions. Four hundred milliliters of the reconstituted solution were removed and a 20-ml aliquot was counted and recorded as the background particulate load. A small volume of the remaining 100 ml of BSS Plus was withdrawn with a new syringe and introduced into a sealed vial containing 10,000 units of lyophilized thrombin via the rubber septum. The contents were mixed and the solu tion was withdrawn (using the same syringe and needle) and reintroduced into the BSS Plus bottle containing the balance of the 100 ml. A solution of 100 U/ml was thus obtained. Ali quote of these 100-U/ml solutions were ana lyzed using the Coulter counter. For this analy sis the instrument was calibrated using 5-, 10-, and 20-μπι beads, creating a 5- to 20-μπι sized particle window.
Results Corneal endothelial perfusion using both 1,000-U/ml thrombin preparations resulted in an immediate and protracted thinning of the cornea compared to its control (Fig. 1). Observation of the endothelial monolayer through the specular microscope showed inter cellular and intracellular vacuole formation. The thrombin-perfused cells appeared vacuolated with junctional abnormalities compared to the BSS Plus control (Figs. 2 and 3). In a comparision of the p H and osmolality characteristics of various concentrations of
each thrombin preparation, the 1,000-U/ml thrombin solutions had increased osmolality (Table 2). Further studies were performed with 100-U/ ml dilutions of Thrombinar (which lacks gly cine). Using an identical protocol to the one previously described, the endothelia of four pairs of human corneas (one cornea as test and the other cornea as control) were perfused with 100-U/ml dilutions (in BSS Plus) of Thrombinar (pH 7.55, 314 mOsm/kg; Table 2). An average corneal deswelling rate was determined for the BSS Plus and 100-U/ml thrombin groups (Fig. 4). The Thrombinar-perfused corneas snowed a limited ability to deturgesce when compared to BSS Plus (P < .02, paired f-test). The thrombinperfused corneas showed a deswelling rate of about 5 |xm/hour compared to a 19^m/hour deswelling rate for the BSS Plus-perfused cor neas. During perfusion, the endothelium showed slight intracellular darkening and a surface ground glass appearance. Scanning electron micrographs, however, did not show alterations in junctional morphology in the eyes perfused with 100-U/ml Thrombinar. Transmission electron micrographs of Thrombinar-perfused corneas appeared similar to control corneas. Gel electrophoresis results are presented in Figure 5. Because the total protein content for a specified activity varied between the different manufacturer's preparations, dissimilar dilu tions of the thrombin solutions were used. Lane 1 shows the electrophoretic result of a 50-U/ml Thrombostat (Parke-Davis) solution. At the top of the running gel there are aggre gated high molecular weight constituents; within the running gel numerous distinct
800
800
E 3. ! S A—· A - Ì 700-·
A
0)
c
D c
o o
600
It
■
600o,
\
,Ο—O—O—O—O—O—O-
% -
· .• . . · A· · A ·BSS· Plus ·
500
400
700
\ .
•
1
0
1
I
1
1
1
1
1
σ
•Thrombostat 1000 U / m l
1
1
1
1
15 30 45 60 75 90 105120135150165180195
Time (min)
A A BSS Plus O — O T h rombinar 1000 U/ml
500-
o
h
O
400
-O-O
60
120
180
Time (min)
Fig. 1 (McDermott, Edelhauser, and Mannis). Representative plots of human corneal thickness values over time when perfused with a 1,000-U/ml solution of Thrombostat (left) and a solution of 1,000 U/ml of Thrombinar (right) compared to BSS Plus control.
' **s*n"!:"*'""' Λ" ''"':Ύ' ■' -' · ■ ■ " "· ■ ρ ' ^ ^ ^ ^ φ ^ ^ ^ Μ & ^ Φ ^ ,
Fig. 2 (McDermott, Edelhauser, and Mannis). Left, Scanning and transmission electron micrographs of a human cornea perfused with a 1,000-U/ml Thrombostat solution. The cells appear shrunken with " p o r e " formation on scanning electron microscopy (top) (x 500). Transmission electron micrographs show intercellular vacuoles and junctional disruption (middle and bottom) (x 4,000). Right, Scanning and transmission electron micrographs of a human cornea of the companions of figures on the left perfused with BSS Plus. Scanning electron micrographs (top) (X500) and transmission electron micrographs (middle and bottom) (x 5,000) demonstrate a normal-appearing monolayer with intact junctional morphology.
418
October, 1988
AMERICAN JOURNAL OF OPHTHALMOLOGY
~"V-?- * ! 1 > ^
tsßm^
ψ&Ρ·3&
f
J'.. fit
Vit' ■&$■ -i ' : rff-
Λ*
■ —»-<■
.
·\ ν *
".■"I*
'
·+
*.*S »
&ÏYW
Fig. 3 (McDermott, Edelhauser, and Mannis). Left, Scanning and transmission electron micrographs of a human cornea perfused with 1,000-U/ml Thrombinar solution. The cells appear shrunken with creeping of the apical cell membrane over adjacent cells on scanning electron microscopy (top) (x 1,000). Transmission electron micrographs (middle and bottom) (x 6,000) show excessive junctional overlap. Right, Scanning and transmis sion electron micrographs of the companion cornea shown on the left perfused with BSS Plus. Scanning electron micrograph (top) (x 1,000) and transmission electron micrographs (middle and bottom) (x5,000) demonstrate an ultrastructurally normal endothelial monolayer.
Vol. 106, No. 4
Intracameral Thrombin
TABLE 2 pH AND OSMOLALITY OF THROMBIN SOLUTIONS MEAN ± S.D. OSMOLALITY MEAN ± S.D. pH
PREPARATION
Thrombinar 1,000 U/ml* Thrombinar 100 U/mP Thrombostat 1,000 U/ml* BSS Plus
7.86 7.55 7.43 7.46
± ± ± ±
0.9 0.06 0.1 0.05
(MOSM/KG)
407 314 492 302
± ± ± ±
1 5 1 10
»Diluted in BSS Plus.
bands are present as well as a large amount of interband protein. Bands A and B have estimat ed molecular weights of 44.5 kd and 37.5 kd. Lane 2 shows the result obtained with a 200-U/ ml solution of Thrombinar (Armour). Bands A and B are the most prominent, but several faint low molecular weight constituents are also present. Note the distinct lack of interband protein deposition and absence of high molecu lar weight aggregates at the top of the running gel. Since protease inhibitors were present in the thrombin solutions during electrophoresis, autodegradation of the polypeptide chain to produce low molecular weight products is un likely. In an analysis of the 100-U/ml dilutions of Thrombinar and Thrombostat for particulate contamination, as well as the background par ticulate load for the BSS Plus diluent, the 100U/ml Thrombostat preparation had a particu late load 30 times greater than that seen in the Thrombinar solution (Table 3).
419
Discussion Intraocular use of products not designated as such or evaluated for intraocular toxicity is potentially dangerous. While topically applied thrombin may be a useful adjunct to control surface bleeding during ocular surgery on a closed globe, it is important to realize that no manufacturer has isolated thrombin for use in any way other than topically. As such, consid erations as to purity, contaminants, pH, osmolality, and vehicles used are based on topical use only. At this time, no thrombin preparation has been approved by the Food and Drug Ad ministration for intracameral use. We found that high-concentration solutions (that is, 1,000 U/ml of either Thrombinar or Thrombostat) resulted in endothelial morphologic alterations characterized by intercellular vacuole forma tion and altered apical junctions. These findings are identical to the endothelial chang es reported in hyperosmotic media 810 and in clude intercellular space dilatation, intracellular vacuole formation, and lysis of intercellular bridges. All of these conditions at one time or another were observed in corneas perfused with thrombin. It appears likely, therefore, that the changes observed when perfusing with 1,000-U/ml thrombin solution are a result of the solution's hypertonicity. Our data suggest that the addi tives in the vial of lyophilized thrombin are the osmotic agents (Tables 1 and 2). A previous study by Mannis and associates 5 showed that
20
E
0
Fig. 4 (McDermott, Edelhauser, and Mannis). Changes in human corneal thickness when paired cor neas are perfused with either Thrombinar 100 U/ml (open circles) or BSS Plus (filled triangles). The solid lines represent first order re gression lines. Note the significant flattening of the Thrombinar's deswelling curve compared to its BSS Plus-perfused companion cor-
co CO CD
-20 +
c υ
-40
σ α> e
-60
ι_ O
υ
-80 -100
Mean±SEM n=4 A BSS Plus Control y=-(19.28±2.60)x + 10.81 OArmour Thrombin 100/ml y=-(5.24±2.68)*x + 6.25 p LO.02 60
120
time(min)
180
October, 1988
AMERICAN JOURNAL OF OPHTHALMOLOGY
420
2
1
3
92.5Kd ( Phosphorulase B) .ÊÊfaMaÊL·
1
1 *
A *B r*
66.2Kd ( Bovine Serum Albumin) 45.0Kd(0valbumin)
Λ
^gjy|^&
!
mm»
.,,..*
Ξ
Ì
31. OKd (Carbonic Anhydrase)
2l.5Kd (SoybeanTrypsin Inhibitor) 1 l4.4Kd(Lysozyme)
Fig. 5 (McDermott, Edelhauser, and Mannis). Results of 12% polyacrylamide gel electrophoresis. Lane 1, Thrombostat (ParW-Davis) at an activity of 50 U/ml; Lane 2, Thrombinar (Armour) at an activity of 200 U/ml; Lane 3, molecular weight standards. Bands A and B in Lanes 1 and 2 may represent thrombin.
sheep corneal endothelium bathed in 1,000-U/ ml thrombin (in normal saline) appeared nor mal with alizarin red S and trypan blue stain ing. However, this is in distinct contrast to the effects on human corneal endothelium where bathing in normal saline or hypertonicity may result in ultrastructural alterations. 8 The alterations observed may be reversible and not cause permanent loss of endothelial function. However, imposition of an osmotic stress on an eye that may have a preexisting endotheliopathy, such as that seen in diabetes mellitus as reported by Schultz and associ ates, 11 could result in endothelial decomposi tion. This is especially important since vitrectomy in diabetic patients using solutions containing thrombin for intraocular hemostasis has been reported. 3 · 5 In these studies, however, a lower dose (100 U/ml) was used so osmotic effects would be less. Even when using a lower concentration of thrombin (100 U/ml), the presence of certain
additives may prove toxic. Thrombostat con tains large quantities of the amino acid glycine, and a 100-U/ml dilution of Thrombostat (the dose used in all cited studies) contains 78 mg/dl of glycine. The presence of glycine at this con centration is especially important in light of the recent study by Creel, Wang, and Wong12 linkTABLE 3 PARTICULATE ANALYSIS OF 100 U/ML THROMBIN SOLUTIONS PARTICLES/ML* SOLUTION
RANGE
MEAN
S.D.
Thrombinar 100 U/ml Diluent for Thrombinar' Thrombostat 100 U/ml Diluent for Thrombostaf
662-706 154-180 20,830-21,540 254-320
694 171 21,221 281
15 10 270 20
"Using 5- to 20-μηη window setting, 100-/*m aperture. *BSS Plus.
Vol. 106, No. 4
Intracameral Thrombin
ing transient blindness (visual acuity of count ing fingers at 1 meter) with increases in serum glycine levels during postoperative bladder ir rigation after prostatic resection. In their pa tients a graying of vision was temporally asso ciated with loss of oscillatory potentials and 30-Hz flicker following electroretinography with serum glycine levels in excess of 30 mg/dl. Clearly, intracameral use of this preparation presents a much higher local concentration of glycine (Table 1). Perfusion of a 100-U/ml thrombin solution results in a 78-mg/dl glycine level, at least 2.5 times the serum level associat ed with visual symptoms. Therefore, a toxic effect is possible on the neurosensory retina (via inhibitory neurotransmission). Data on long-term toxicity or reversibility of exposure to glycine has not been reported to date. The other thrombin preparation, Thrombinar, does not contain glycine. However, perfusion of this preparation at 100 U/ml appears to be responsible for a significant (P < .02) loss in the ability of the cornea to deturgesce when perfused with BSS Plus. The BSS Plus-perfused corneas deturgesced at a significantly greater rate of 19 |xm/hour as com pared to the thrombin-perfused corneas' deturgescence rate of 5 μιτι/hour. The reversi bility and long-term effect of exposure at this concentration is indeterminant. It appears, however, that short-term irrigation (less than one hour) has little effect on corneal deswelling, but longer durations of perfusion may result in a cumulative attenuation in endothelial function as reflected by the significantly different rates of deswelling. This could be of particular importance if the thrombin is left in the eye after completion of the surgical proce dure. In addition to the tonicity and presence of additives, the protein composition of a specific activity thrombin solution must be considered. The activity in units of a thrombin solution is determined by comparing its ability to clot a plasma extract or purified fibrinogen solution to that of a National Bureau of Standards lot of thrombin. This in no way informs the user of the thrombin product what constituents and concentrations are present. The specific activity is only a meas.ure of the ability to cause clot ting. Since all thrombin solutions are labeled in terms of activity, interlot variability of extrane ous proteins, particulates, and the like can exist. Therefore, even if the surgeon uses the same 100-U/ml dilution of a particular thrombin solution, there is no guarantee that the same solution (in terms of contaminants and particu
421
lates) is present. Herein we found that two thrombin preparations can contain different amounts of protein with presumed nonthrombin activity. Bands A and B are present in the preparations of both manufacturers, and presumably thrombin activity exists in one or both of those bands. This is in agreement with previous reports of multiple bovine thrombin components. 1314 Band B, with a mo lecular weight estimate of 37.5 kd, is close to the reported molecular weight of 39 kd for bovine thrombin. 14 It appears, therefore, that identical activity solutions of thrombin by different manufactur ers have different electrophoretic patterns and amounts of nonthrombin protein. These substances may possess activity in their own right and serve as immunogenic stimuli. More over, during preparation of the crude thrombin product, a variety of particulate matter may be introduced and subsequently perfused intraocularly. Our Coulter counter results showed that certain preparations, even when diluted, con tained large quantities of particulates. When using Thrombostat, irrigation during vitrectomy with 100 ml of the 100-U/ml solution would allow over two million particles to enter the eye. It is entirely possible, therefore, that the hypopyon formation as reported by Thompson and associates 3 was caused by the relatively crude purity of the commercial thrombin preparation. Therefore, one needs to consider whether the intraocular milieu is the appropriate chamber for this product. Even when using an ultra-pure crystalline thrombin preparation one must consider its potential interactions. Thrombin is a potent mitogen for fibroblasts, is chemotactic for neutrophils, and causes smooth muscle contrac tion. Upon exposure to thrombin solutions, vascular endothelium undergoes permeability increases, platelets aggregate with the subse quent generation of eicosanoids, and fibronectin and laminin undergo enzymatic cleavage. Thrombin has been shown to bind to neural tissue causing an increase in cellular cGMP concentration. 15 It is apparent that thrombin is a molecule with enzymatic and receptor activity; however, it is not known whether the intact human corneal endothelium has thrombin receptors. In cultured cell lines good evidence exists for thrombin receptors. Isaacs and associates 16 demonstrated a specific 40,500-dalton, surface membrane-bound, α-thrombin receptor in cul tured bovine corneal endothelium. This recep tor is saturable and specific, but as yet does not
422
AMERICAN JOURNAL OF OPHTHALMOLOGY
h a v e a d e m o n s t r a b l e p h y s i o l o g i c effect o n thrombin-receptor coupling. Cellular r e c e p t o r d e n s i t y in t h e b o v i n e c o r n e al e n d o t h e l i u m m o d e l is a b o u t 100,000 r e c e p tors p e r cell a n d s h o w s s a t u r a t i o n at 2.0 to 2.5 U/ml of t h r o m b i n . In further s t u d i e s of this receptor, Savion a n d associates 1 ' s h o w e d t h a t b o v i n e c o r n e a l e n d o t h e l i u m cells actively inter nalize the r e c e p t o r - a - t h r o m b i n c o m p l e x via a d s o r p t i v e e n d o c y t o s i s . In this m o d e l , internalization of t h e covalent t h r o m b i n r e c e p t o r c o m p l e x r e s u l t s in a n i n c r e a s e d n u m b e r of cellular r e c e p t o r site d e n s i t y . It a p p e a r s , t h e r e f o r e , n o t w i t h s t a n d i n g t h e differences b e t w e e n species a n d t h e differences involved in cell c u l t u r e , that a h u m a n corneal e n d o t h e l i a l r e c e p t o r could exist. C u r r e n t s t u d ies u s i n g 1 2 5 I-human t h r o m b i n to e v a l u a t e h u m a n c o r n e a l e n d o t h e l i a l b i n d i n g are u n d e r w a y . As s u c h , t h e possibility of its p r e s e n c e merits careful c o n s i d e r a t i o n before i n t r o d u c tion of t h r o m b i n i n t r a c a m e r a l l y . This is e s p e cially w o r r i s o m e w i t h t h e i n t r o d u c t i o n of t h r o m b i n into t h e diabetic eye w i t h a n incom petent blood-aqueous barrier and neovascularization. In this s e t t i n g , a m u l t i t u d e of interac tions m a y occur in a d d i t i o n to t h e o b v i o u s c o n v e r s i o n of fibrinogen to fibrin. While m a n y of o u r c o n c e r n s could be v i e w e d as p o t e n t i a l p r o b l e m s , t h e r e are definite differences be t w e e n the two m a n u f a c t u r e r ' s t h r o m b i n p r o d u c t s . M o r e o v e r , u s i n g a sensitive t e c h n i q u e , n a m e l y in vitro corneal p e r f u s i o n , a 100-U/ml thrombin solution, the dose most commonly u s e d , h a s b e e n s h o w n to d e c r e a s e t h e d e s w e l l i n g rate of t h e p e r f u s e d h u m a n c o r n e a during prolonged perfusion.
ACKNOWLEDGMENT
Sally S. T w i n i n g , P h . D . , a n d Patricia M. Wil s o n , M . S . , D e p a r t m e n t of B i o c h e m i s t r y , M e d i cal College of W i s c o n s i n , p e r f o r m e d t h e gel electrophoresis studies.
References 1. Savory, W.: Some uses of thrombin and fibrino gen in ophthalmic surgery. Trans. Ophthalmol. Soc. U.K. 67:323, 1947. 2. Hughes, W. L.: Use of thrombin in the anterior
October, 1988
chamber to control hemorrhage. Acta XVI Concilium Ophthalmologicum 12:99, 1950. 3. Thompson, J. T., Glaser, B. M., Michels, R. G., and DeBustros, S.: The use of thrombin to control hemorrhage during vitrectomy. Ophthalmology 93:279, 1986. 4. Blacharski, P. A., and Charles, S. T.: Thrombin infusion to control bleeding during vitrectomy for stage V retinopathy of prematurity. Arch. Ophthal mol. 105:203, 1987. 5. Mannis, M. S., Sweet, E., Landers, M. B., and Lewis, R. A.: Uses of thrombin in ocular surgery. Effect on the corneal endothelium. Arch. Ophthal mol. 106:251, 1988. 6. DeBustros, S., Glaser, B. M., and Johnson, M. A.: Thrombin infusion for the control of intraocu lar bleeding during vitreous surgery. Arch. Ophthal mol. 103:837, 1985. 7. Dikstein, S., and Maurice, D. M.: The metabolic bases to the fluid pump in the cornea. J. Physiol. 221:29, 1972. 8. Edelhauser, H. F., Hanneken, A. M., Pederson, H. J., and Van Horn, D. L.: Osmotic toler ance of rabbit and human corneal endothelium. Arch. Ophthalmol. 99:1281, 1981. 9. Van Horn, D. L., Schultz, R. O., and Edelhauser, H. F.: Corneal cryopreservation. Altera tions in endothelial intercellular spaces. Am. J. Oph thalmol. 63:454, 1969. 10. Sherrard, E. S.: Characterization of changes observed in corneal endothelium with the specular microscope. Invest. Ophthalmol. Vis. Sci. 17:322, 1978. 11. Schultz, R. O., Matsuda, M., Yee, R. W., Edelhauser, H. F., and Schultz, K. J.: Corneal endo thelial changes in type I and type II diabetes mellitus. Am. J. Ophthalmol. 98:401, 1984. 12. Creel, D. J., Wang, J. M., and Wong, K. C : Transient blindness associated with transurethral re section of the prostate. Arch. Ophthalmol. 105:1537, 1987. 13. Rosenberg, R. D., and Waugh, D. F.: Multiple bovine thrombin components. J. Biol. Chem. 245:5049, 1970. 14. Mann, K. G., Yip, R., Heldebrant, C. M., and Fass, D. N.: Multiple active forms of thrombin. J. Biol. Chem. 248:1868, 1973. 15. Walz, D. A., Fenton, J. W., and Shuman, M. A. (eds.): Bioregulatory functions of thrombin. Ann. N.Y. Acad. Sci. 485:5, 1986. 16. Isaacs, J. D., Savion, N., Gospodarowicz, D., Fenton, J. W., and Shuman, M. A.: Covalent binding of thrombin to specific sites on corneal endothelial cells. Biochemistry 20:398, 1981. 17. Savion, N., Isaacs, J. D., Gospodarowicz, D., and Shuman, M. A.: Internalization and degradation of thrombin and up regulation of thrombin binding sites in corneal endothelial cells. J. Biol. Chem. 256:4514, 1981.