HUMAN GENE THERAPY 10:2551– 2557 (October 10, 1999)Mary Ann Liebert, Inc.

Target Gene Transfer of Tissue Plasminogen Activator toCornea by Electric Pulse Inhibits Intracameral Fibrin

Formation and Corneal Cloudiness

TAIJI SAKAMOTO, YUJI OSHIMA, KAZUNORI NAKAGAWA, TATSURO ISHIBASHI, HAJIME INOMATA, and KATSUO SUEISHI

ABSTRACT

Intracameral fibrin formation, a complication of ocular inflammation and intraocular operations, sometimesresults in glaucoma and/or corneal damage leading to permanent visual loss. We transferred a therapeuticgene to the corneal endothelium in order to use it as a therapeutic organ. A plasmid encoding tissue plas-minogen activator (tPA) was injected into the anterior chamber of rats and electric pulses (EPs) were givensubsequently, which transferred a plasmid gene to a highly selected area of corneal endothelium with no in-flammation. The biologically active tPA was clearly present for 4 days after treatment. Fibrin formation in-duced by YAG laser-generated bleeding in the anterior chamber decreased significantly more in treated eyesthan in control eyes. Corneal opacity was significantly lower in treated eyes than in control eyes and histo-logical damage was not apparent in the treated eyes. This genetic modification allows us to use the cornealendothelium to treat various ocular diseases and could be a new and effective type of pharmacologic genetherapy.

2551

OVERVIEW SUMMARY

Since the corneal endothelium plays an important role notonly in keeping the cornea transparent, but also in modu-lating the homeostasis of the anterior eye, gene transfer tothe corneal endothelium could be an alternative approachto drug therapy for diseases of the anterior eye, such as in-tracameral inflammation, fibrin formation, and glaucoma.To maintain good vision, gene transfer to the central corneashould be avoided, because the long-term effect of genetransfer is unclear. However, there are no such limitationson gene transfer to selected areas of corneal endotheliumwith no inflammation. In the present study, we apply tissueplasminogen activator gene by electric pulse-mediatedtransfer to a selected area of corneal endothelium and showits therapeutic efficacy to clear intracameral fibrin forma-tion after bleeding. This new pharmacologic gene therapymethod opens a new avenue for the treatment for variousocular diseases.

INTRODUCTION

CO RN EA IS A transparent anterior ocular tissue that plays animportant role as a barrier against stimulation from outside

the eye (Waring et al., 1982). In ocular inflam mation or afteran intraocular operation, plasma proteins enter the aqueous hu-mor after the breakdown of the blood–ocular barrier, and pro-teinous exudate can be clinically observed in the anterior cham-ber. Fibrin is a major component of this exudate (Rowland etal., 1985). Fibrin is an important product in the pathogenesisof anterior eye disorders: first, it is an important enhancer ofinflammation and mesenchymal activation and thus excessivefibrin exudate induces inflammation and impairs the tissue(McKay, 1972); second, the intracameral fibrin sometimes ob-structs aqueous humor outflow and subsequently induces glau-coma, which will ultimately damage vision (Lundy et al., 1996).The direct injection of plasminogen activators— tissue plas-minogen activator (tPA) or urokinase-type plasminogen acti-vator (uPA)—has been attempted to resolve intracameral fib-

Departments of Ophthalmology and First Pathology, Faculty of Medicine, Kyushu University, Fukuoka 812-8582 Japan.

rin clots (Lundy et al., 1996; WuDunn, 1997). However, thedirect injection of plasminogen activator sometim es causes un-predictable complications such as cataract formation or en-dophthalmitis and, furthermore, fear of this treatment is not neg-ligible among patients.

Gene transfer technology has opened new possibilities notonly for the treatment of inherited diseases but also for newpharmacological treatments of common diseases (Crystal,1995a,b). Most current gene therapy applications are not in-tended to correct the chromosome, but to insert genes to curepatients by changing the microenvironm ent: the genes cure pa-tients by acting as a drug delivery system. Gene transfer-ori-ented drug therapy has potentially a great advantage comparedwith present drug delivery systems. For example, gene trans-fer-oriented drug delivery can keep the concentration of thera-peutic material rather high for a long period, whereas a highconcentration can be obtained only transiently by a single in-jection of therapeutic materials and repeated administration isneeded. Therefore, the therapeutic effect of genes might be ac-complished with an even lower concentration of drug than isexpected.

The corneal endothelium maintains the transparency of thecornea by dehydrating the corneal stroma, and modulatescorneal inflammation through fibrinolytic activities (Waring etal., 1982). Since the corneal endothelium faces the aqueous hu-mor directly, it could also affect the aqueous humor. Therefore,changing the characteristics of corneal endothelium by genetransfer for therapeutic purposes might be useful for the ther-apy of anterior chamber diseases, such as anterior chamber fib-rin formation and glaucoma. To date, the corneal endotheliumhas not received much attention as a therapeutic target for anydisease, but gene transfer technology may allow us to use it asa therapeutic modulator (pharmacologic gene therapy).

In the present study, we transferred plasmid tPA cDNA tothe corneal endothelium by electric pulse (EP) and examinedits effect on fibrin formation and corneal opacification after in-tracameral bleeding.

MATERIALS AND METHODS

cDNA expression vector

The expression vector pCMV-tPA, encoding human tPAcDNA (GenBank accession no. E08757) under the control ofthe promoter of the human cytomegalovirus (CMV) immedi-ate-early gene, was obtained from Mitsui Pharmacologic alCompany (Tokyo, Japan). As the control plasmid, pCMV-tPA(–) was constructed according to the following strategy: tPAcDNA was eliminated from pCMV-tPA by XhoI–XbaI diges-tion. Both XhoI and XbaI cutting ends were blunt ended byKlenow enzym e, and ligated to form the circular plasmidpCMV-tPA(–). pCH110, a mammalian expression vector car-rying the lacZ gene with the simian virus 40 (SV40) early pro-moter, was also used to visualize the area of gene transfer.pCH110 was obtained from Pharmacia (Uppsala, Sweden).Plasmids for gene transfer were prepared by a standard equi-librium centrifugation method in a CsCl–ethidium bromide gra-dient. They were dissolved in phosphate buffered saline, pH7.35 (PBS), and stored at 4°C until use.

Gene transfer to corneal endothelium

All animals were used humanely, after obtaining the properapproval of Kyushu University and in strict compliance with theAssociation of Research for Vision and Ophthalmology (ARVO)Statement for the Use of Animals in Ophthalmic and Vision Re-search and the Declaration of Helsinki. Male Brown-Norwayrats or Wistar rats (8 weeks old, 250–280 g; Kyudo, Fukuoka,Japan) were used in the following studies. The Brown-Norwayrats were randomized into four experimental groups: (1) plas-mid pCMV-tPA injection (500 ng/ml) and EP, (2) plasmidpCMV-tPA injection (500 ng/ml) alone, (3) plasmid pCMV-tPA(–) injection (500 ng/ml) and EP, and (4) no treatment. Plas-mid DNA was transferred to the corneal endothelium by the pre-viously described method (Oshima et al., 1998) with somemodification. Briefly, 5 m l of balanced salt solution (BSS) con-taining plasmids was injected with a 30-gauge needle into theanterior chamber at the limbus. Immediately after the injectionof plasmids, a circular stainless steel electrode, measuring 0.5mm in diameter and coated with gold (Oshima et al., 1998), wasput on the surface of the cornea of each eye. A series of eightEPs on the same site with a pulse length range of 50 msec (80-msec intervals) was delivered with a standard square wave elec-troporator (CUY20; Tokiwa, Fukuoka, Japan). To see the dis-tribution of gene transfection, pCA110 plasmid injection andEPs were also applied to Wistar rats, because b -galactosidasecan be clearly visualized as a blue-colored product in the white-colored animals by 5-bromo-4-chloro-3-in dolyl- b -D-galactopy-ranoside (X-Gal) staining (Oshima et al., 1998).

To see the effect of direct injection of protein, human re-combinant tPA was injected into the rat anterior chamberthrough the limbus immediately after the induction of fibrin for-mation by YAG (yttrium–aluminum–garnet) laser. The dose oftPA for each eye (300 ng/eye) was determined according to theclinical protocol of Lundy et al. (1996).

Induction of intracameral fibrin formation and its evaluation

After gene transfer by EP, intracam eral fibrin formation wasinduced by YAG laser shot at three portions of the iris (2, 6,and 10 o’clock, 2.0 mJ; Zeiss, Oberkochen, Germany). Mod-erate to severe bleeding occurred immediately afterward and aneasily discerned fibrin clot then formed on the surface of theiris and lens. The value of fibrin formation and bleeding in theanterior chamber was graded on day 3 by masked observers,using surgical microscopy, on the basis of the following crite-ria: fibrin clot and/or bleeding occupies more than half of theanterior chamber, 4 1 ; between one-third and one-half of ante-rior chamber, 3 1 ; less than one-third of the anterior chamber,2 1 ; faint fibrin and/or bleeding, 1 1 ; no fibrin or bleeding, 0.The value of corneal cloudiness was also rated according to thefollowing criteria; the corneal area of opacity included almostall of the cornea, 3 1 ; almost half of the cornea, 2 1 ; one-thirdand one-half of the cornea, 2 1 ; less than one-third of the cornea,1 1 ; and clear cornea, 0. A histological study was also per-formed in paraffin-embedd ed sections. The eyes were enucle-ated on day 3 and examined by routine light microscopy and/orby immunohistoc hemistry for fibrinogen (rabbit polyclonalanti-fibrinogen antibody; Dako, Glostrup, Denmark).

SAKAMOTO ET AL.2552

Detection of tPA in the aqueous humor

The aqueous humor was collected with a 30-gauge needlesyringe (3–7 m l/eye) for 14 days and stored frozen at 2 80°Cuntil the experiment. The amount of tPA antigen was measuredby a two-phase ELISA method, using TintElize tPA (Biopool,Ventura, CA). The minimal detectable concentration of tPA was100 pg/ml. The biological activity of plasm inogen activator was

quantified by chromogenic assay, using H-D -But-CHT-Lys-pNA and poly-D-lysine (Biopool). To measure the biologicalactivity specifically caused by tPA, Chromolize tPA (Biopool)was used. Briefly, the bottoms of the wells of a 96-well platewere coated with monoclonal antibody to tPA (specificallybinds to tPA). The activity of plasminogen activity was thenmeasured by chrom ogenic assay. The minimum biological ac-

tPA GENE TRANSFER TO CORNEAL ENDOTHELIUM 2553

FIG. 1. Photography of eye treated with lacZ plasmid and electric pulses on day 3.(A) The eye was injected with plasmid encoding the b -galactosidase gene, electric pulseswere administered, and the eye was enucleated after 3 days and stained with X-Gal. Thecornea shows a positive reaction where the electric probes were placed (arrows). Thecentral cornea is free from gene transfer (arrowheads). (B) The corneal endotheliumshows a positive reaction. Original magnification: 3 80.

FIG. 2. tPA in the aqueous humor of the treated eyes. (A) The amount of tPA was measured by ELISA, and it was highest onday 2. (B) The biological activity of tPA was measured by chrom ogenic assay. It was found in pCMV-tPA-injected and EP-treated eyes on day 2, but it was not detected in any other eyes or at any other time. (C) The fibrinolytic activity profile of theaqueous humor was analyzed by fibrin zymographic assay. Lane 1, no treatment; lane 2, sham operation; lane 3, pCMV-tPA in-jection and EP; lane 4, pCMV-tPA injection alone; lane 5, pCMV-tPA(–) injection and EP; lane 6, pCMV-tPA injection and EP,which is analyzed by gels containing anti-tPA antibody. Fibrinolytic bands are shown at 40 and 112 kDa. Both were neutralizedby anti-tPA antibody. tPA, Tissue plasminogen activator; pCMV-tPA, plasmid containing tPA cDNA; pCMV-tPA(–), controlplasm id; EP, electric pulse.

A

B

A

C

B

tivity detected by these methods was 0.1 IU/ml. The biologicalactivity of plasmingen activator was also analyzed by fibrin zy-mography according to a previously described method (Budenzet al., 1995). Briefly, sodium dodecyl sulfate (SDS)-polyacryl-amide slab gels were prepared with a separation gel of 10%acrylamide and a stacking gel of 4% acrylamide, and then thesamples (4 m l for each lane) were applied. Next, the SDS gelwas placed on the indicator gel after washing and the fibri-nolytic band was observed. The indicator gel was composed of1% agarose containing 0.1% fibrin (Sigma) and plasminogen(7 m g/ml; Sigma), and was stained with Coomassie BrilliantBlue (Nakarai Chemical, Tokyo, Japan). In some experim entsthe samples were analyzed by fibrin gels containing anti-tPAantibody (10-mg/ml gel; American Diagnostica, Greenwich,CT).

Statistical analysis

All experiments were statistically analyzed, using theWilcoxon rank sum p value. A p value of less than 0.05 wasconsidered to be significant.

RESULTS

Gene transfer to corneal endothelium by electric pulse

Gene transfer. First, in order to see the gene-transfected cells,plasm id pCH110 encoding the lacZ gene was injected into theanterior chamber of Wistar rats from the corneal limbus andeight consecutive EPs were delivered. Our previous studyshowed that consecutive EPs could transfer plasmid cDNA tothe intended area of corneal endothelium efficiently; we per-formed the same procedure with a minor modification (Oshimaet al., 1998). After 3 days of this treatment, b -galactosidase wasobserved in the corneal endothelium within the area where theelectric probe was placed (Fig. 1A and 1B).

tPA gene transfer. Next, pCMV-tPA encoding human tPAcDNA, or the control plasmid pCMV-tPA(–), was injected intothe anterior chamber of Brown-Norway rats to insert tPA cDNAinto the corneal endothelium. Brown-Norway rats were used inthis study because fibrin formation and corneal change can beclearly observed only in pigmented eyes.

An ELISA study disclosed that tPA antigen was detected inall samples and that the amount of tPA was significantly largerin eyes receiving both pCMV-tPA injection and EP treatmentthan in other, control eyes on day 2 (1.61 6 0.54 ng/ml, n 5 7,p , 0.05). The amount of tPA in the treated eyes returned tothe control level (0.50 ng/ml or less) after day 6 (Fig. 2A). Theamount of tPA was less than 0.5 ng/ml in the control eyes (non-treated eyes, eyes receiving pCMV-tPA injection without EP,and eyes receiving pCMV-tPA(–) and EP).

The biological activity of tPA in the aqueous humor was mea-sured by chromogenic assay (level of detection, 0.1 IU/ml ormore). On days 1–4 activity was found only in eyes receivingpCMV-tPA injection and EP treatment (0.56 6 0.24 IU/ml onday 2, p , 0.05), but activity was not detected in any other eyesor at any other time period (Fig. 2B). The fibrin zymographicassay disclosed in all samples the existence of fibrinolytic ac-

tivity, which was mainly composed of tPA and tPA–plasmino-gen activity inhibitor complex because it was neutralized by anti-tPA antibody (Fig. 2B and C and our previous results[Fukushima et al., 1989; Hayashi et al., 1989]). There was nei-ther uncontrollable nor long-lasting bleeding in any of theseeyes. After that period, biologically active tPA began to appearin the eyes treated with pCMV-tPA and EP (0.56 IU/ml on day2). Although both tPA and plasminogen activator inhibitor weredetected in the aqueous humor of control eyes by fibrin zy-mography (Fig. 2C), the fibrinolytic activity was not detectableby choromogenic quantitative assay (0.1 IU/ml or less).

Subsequent to the injection of equivalent amounts of re-combinant tPA into rat eyes after YAG laser-induced bleeding(300 ng/rat eye), uncontrollable bleeding was observed in fourof nine eyes and a large bleeding mass filled the whole ante-rior chamber of two of these four eyes by day 3. In addition,there still remained a few small fibrin clots in other eyes (score1.72 6 0.82). On day 2, tPA activity decreased to a less thandetectable level even in the tPA protein-injected eyes.

Effect on intracameral fibrin deposition

On day 0 (immediately after YAG laser shots) a large fibrinclot admixed with blood filled the anterior chamber. This fib-rin clot gradually became reduced in size in all eyes. The valueof fibrin formation and bleeding in the anterior chamber wasgraded on day 3. The moderate fibrin formation associating withblood clot was observed in the anterior chamber of control eyes:no treatment (score, 3.14 6 0.46; n 5 7), pCMV-tPA injectionalone (score, 2.92 6 0.66; n 5 7), and pCMV-tPA(–) injectionand EP (score, 2.38 6 0.73; n 5 10). On the other hand, itshould be noted that only a few small fragments of fibrin clotwas observed in the eyes treated with pCMV-tPA injection andEP (score, 1.13 6 0.35; n 5 14; p , 0.05) (Fig. 3).

SAKAMOTO ET AL.2554

FIG. 3. Effect of EP-mediated tPA gene transfer on intra-cameral fibrin formation. Intracam eral fibrin formation was in-duced by YAG laser shots on the iris. Immediately afterward,a large fibrin clot admixed with blood was present, filling theanterior chamber. This fibrin clot gradually shrank and the valueof fibrin formation and bleeding was graded on day 3 (for grad-ing criteria, see text). Moderate fibrin formation associated withblood clot was observed in the anterior chamber of control eyes:eyes receiving no treatment, eyes treated by pCMV-tPA injec-tion alone, and eyes treated by pCMV-tPA(–) injection and EP.Fibrin formation was apparently inhibited by pCMV-tPA andEP treatment (p , 0.01). tPA, Tissue plasminogen activator;pCMV-tPA, plasmid containing tPA cDNA; pCMV-tPA(–),control plasmid; EP, electric pulse.

tPA GENE TRANSFER TO CORNEAL ENDOTHELIUM 2555

FIG. 4. Photographs of anterior eye 3 days after YAG laser-induced intracameral bleeding and treatment. (A) No treatment;(B) pCMV-tPA injection alone; (C) pCMV-tPA(–) injectionand EP; (D) pCMV-tPA injection and EP. The eye that wastreated by pCMV-tPA injection and EP shows less edema anda clearer cornea. tPA, Tissue plasminogen activator; pCMV-tPA, plasmid containing tPA cDNA; pCMV-tPA(–), controlplasm id; EP, electric pulse.

A

C

B

D

FIG. 5. Effect of EP-mediated tPA gene transfer on cornealopacity. Intracameral fibrin formation was induced by YAGlaser shots on the iris. The value of fibrin formation and bleed-ing was graded on day 3 (for grading criteria, see text). Clini-cal corneal cloudiness was graded 3 days after treatment (forgrading, see text). In control eyes, the cornea was edematouson day 3 and showed cloudiness to various extents. Controls:no treatment, pCMV-tPA plasmid injection alone, and pCMV-tPA(–) injection combined with EP. The eyes treated by pCMV-tPA injection and EP showed less edema and a clear cornea(p , 0.01).

FIG. 6. Effect of EP-mediated tPA gene transfer onanterior chamber tissue. Intracameral fibrin formationwas induced by YAG laser shots on the iris. Immedi-ately afterward, a large fibrin clot admixed with bloodwas present, filling the anterior chamber. This fibrinclot gradually shrank, the eye was enucleated on day3, and the histological study was performed. (A) In thiscontrol eye (no treatment), the trabecular meshworkwas filled with proteinaceous exudate (arrow). The irisadhered to the corneal endothelium in association withinflammatory cell infiltrate (double arrows, eye with notreatment after YAG laser shot). (B) In this control eye,the proteinaceous exudate is positive immunohisto-chemically for fibrinogen (arrow). (C and D) In theseeyes treated with pCMV-tPA and EP, the cornea is clearand no proteinaceous exudate was found in the trabec-ular meshwork (double arrows, A). (A and C) Hema-toxylin and eosin; (B and D) avidin–biotin–peroxidasecomplex method for fibrinogen. Original magnifica-tion: (A–D) 3 40. tPA, Tissue plasminogen activator;pCMV-tPA, plasmid containing tPA cDNA; pCMV-tPA(–), control plasm id; EP, electric pulse.

A

C

B

D

Effect on corneal opacity and anterior chamber tissue

The EP treatment caused transient corneal cloudiness exactlywhere the electric probe was placed; the cornea gradually re-gained its clearness. In control eyes, the cornea was still ede-matous on day 3 and showed cloudiness to various extents: notreatment (corneal cloudiness score, 2.25 6 0.76; n 5 7),pCMV-tPA injection alone (2.52 6 0.48, n 5 7), and pCMV-tPA(–) injection and EP (2.29 6 0.61, n 5 7). The eyes treatedby pCMV-tPA injection and EP had less edematous or cloudycorneas (1.12 6 0.26, n 5 14, p , 0.05) (Figs. 4 and 5). A his-tological study showed that the trabecular meshwork of controleyes was filled with proteineous exudate that was immunohis-tochemically positive for fibrinogen (Fig. 6A and B). The iriswas partially adhered to the corneal endothelium in associationwith inflammatory cell infiltrate (Fig. 6A, double arrows). Onthe other hand, in the eyes treated with pCMV-tPA and EP, thecornea was intact and no proteineous exudate was found in thetrabecular meshwork (Fig. 6C and D).

DISCUSSION

In the present study, intracameral fibrin formation afterbleeding and subsequent corneal cloudiness were inhibited sig-nificantly by EP-mediated tPA gene transfer to the corneal en-dothelium, without any significant inflammation.

The present treatment is considered to be superior to the di-rect injection of tPA into the eye, owing to the following points.In clinical intracam eral fibrin formation, recombinant tPA wassometimes injected directly into the anterior chamber to clearthe fibrin (25 m g/eye) (Lundy et al., 1996). When we injectedan equivalent amount of recombinant tPA into rat eyes afterYAG laser-induced bleeding (300 ng/rat eye), uncontrollablebleeding was observed in 44% of eyes and a large bleedingmass filled the whole anterior chamber in half of these eyes onday 3. In addition, a few small fibrin clots remained in the othereyes. Since the biological half-life of tPA is short (native tPA,2 min) (Lucore et al., 1989), partly owing due to the degrada-tion of tPA itself and partly because of binding to inhibitors ofplasm inogen activators (Sozka and Olszewski, 1986; Lucore etal., 1989), a large amount of tPA ordinarily is needed to re-solve a large fibrin clot. It should be noted that an excessiveamount of free tPA sometimes has an unwanted effect on thesurrounding tissues, especially after tissue damage in the earlyphase (e.g., local but uncontrollable bleeding). Such massivebleeding in the anterior chamber sometimes impairs the sur-rounding tissue irreversibly (e.g., subsequent corneal damageor glaucoma). On the other hand, no such unwanted phenom -ena were seen in the tPA gene-transferred eyes of the presentstudy. In the present method, the biologically active tPA is con-tinuously synthesized by the gene-transferred cells for at least4 days after tPA gene transfer and therefore a good therapeu-tic effect could be achieved by a relatively low concentrationof tPA in the aqueous humor, which could thus reduce the riskof unwanted complications. In this way, a low concentration oftPA can clear the fibrin clot efficiently, and is less harmful tothe surrounding tissue compared with the direct injection of alarge amount of tPA protein.

The present method of tPA gene transfer to the corneal en-dothelium reduced subsequent corneal cloudiness after bleed-

ing. Histological damage to the corneal endothelium was notapparent in either nontreated or treated eyes in this study. How-ever, focal adhesion of the iris and corneal endothelium asso-ciated with mild inflammation was always found in nontreatedeyes: fibrin is a strong proinflam matory protein and thus theexcessive fibrin caused inflam mation and damaged the sur-rounding tissues, including the corneal endothelium and tra-becular meshworks of nontreated eyes.

The possible merit of the present therapy is that gene trans-fer was accomplished nonvirally by an EP-mediated method.Viral vectors, such as adenovirus vectors, can transfer exoge-nous genes to mammalian cells, including corneal endothelialand trabecular meshwork cells, with enormous efficacy (Bu-denz et al., 1995; Larkin et al., 1996). However, current viralvectors have side effects other than the transferring of exoge-nous genes, e.g., an adenovirus vector also causes immunore-actions (Crystal, 1995a,b; Sakamoto et al., 1998). These risksare still too high, precluding patients and physicians from us-ing viral vectors for the treatment of common eye diseases. Thesimple injection of plasm ids can accomplish effective genetransfer to muscles (Acsadi et al., 1991; Davis et al., 1993), butthis method could not be used to transfer genes to the oculartissue in this study. On the other hand, the injection of plas-mids followed by EP, while not as effective for introducinggenes to the cells in vivo as viral vector-m ediated gene trans-fer, does not have any local or system ic toxicity (Oshima et al.,1998).

More importantly, the injected gene is expressed by thecorneal endothelium within a highly limited area (Fig. 1). In li-posome-media ted gene transfer or viral vector-mediated genetransfer, an exogenous gene can be introduced to the cornealendothelium; however, it can also be transferred to other tis-sues, such as the trabecular meshwork and anywhere in thecornea (Budenz et al., 1995; Hangai et al., 1998). Even in thecontext of extremely safe methods, the functional changes incells caused by gene transfer cannot be fully predicted at pres-ent and the gene transfer might also possibly change the phys-iological function of the cells. If gene transfer affects the fun-damental function of the corneal endothelium, which is to keepthe cornea clear, it would thus seriously impair the vision. Inaddition, gene insertion into lens epithelial cells might causeunexpected cataracts. Therefore, gene transfer should be strictlylimited to the intended area and should avoid the pupillary zoneof the cornea or lens, to maintain the vision of the patients. Thepresent technique is quite suitable for the treatment of anteriorchamber diseases. In addition, the transient expression of theinserted gene was found to be sufficient for the present pur-pose.

The eye is one of the organs most accessible to gene trans-fer. It is easily visible and treatable by surgeons and the pres-ent method can also be perform ed during cataract surgery,corneal transplantation, and trabeculectomy procedures. Wechose the corneal endothelium as the therapeutic tissue for genetransfer in this study. The modification of the microenviron-ment by the present method represents a new and effective op-tion for the treatment of common ocular diseases. Consequently,genetic modification allows us to use any organ or tissue forthe treatment of various diseases. This study therefore providesimportant information for the further developm ent of pharma-cologic gene therapy for ocular diseases.

SAKAMOTO ET AL.2556

ACKNOWLEDGMENTS

This work was supported in part by Grant-in-Aid 09671804for Scientific Research from the Ministry of Education, Sci-ence, Sports, and Culture of the Japanese Governm ent, theJapan National Society for the Prevention of Blindness (Tokyo,Japan), the Fukuoka Anti-Cancer Association (Fukuoka, Japan),the Kaibara Morikazu Medical Science Promotion Foundation(Fukuoka, Japan), and the Casio Science Promotion Founda-tion (Tokyo, Japan).

REFERENCES

ACSADI, G., DICKSON, G., LOVE, D.R., JANI, A., WALSH, F.S.,GURUSINGHE, A., WOLFF, J.A., and DAVIS, K.E. (1991). Hu-man dystrophin expression in mdx mice after intramuscular injectionof DNA constructs. Nature (London) 352, 815–818.

BECKER, G.J., RABE, F.E., HOGAN, R.D., EVAN, A.P.,FRANKLIN, T.D., HOLDEN, R.W., BANG, N.U., DILLEY, R.S.,BENDICK, P.J., and KLATTE, E.C. (1983). Thrombolytic agents:Effects of local infusion on microvascular physiology and macrovas-cular ultrastructure. Radiology 148, 403–406.

BUDENZ, D.L., BENNETT, J., ALONSO, L., and MAGUIRE, A.(1995). In vivo gene transfer into murine corneal endothelial and tra-becular meshwork cells. Invest. Ophthalmol. Vis. Sci. 36, 2211–2215.

CRYSTAL, R.G. (1995a). Transfer of genes to humans: Early lessonsand obstacles to success. Science 270, 404–410.

CRYSTAL, R.G. (1995b). The gene as the drug. Nature Med. 1, 15–17.DAVIS, H.L., DEMENEIX, B.A., QUANTIN, B., COULOMBE, J.,

and WHALEN. R.G. (1993). Plasmid DNA is superior to viral vec-tors for direct gene transfer into adult mouse skeletal muscle. Hum.Gene Ther. 4, 733–740.

ELNER, V.M., ELNER, S.G., PAVILACK, M.A., TODD, R.F., III,YUE, B.Y., and HUBER, A.R. (1991). Intercellular adhesion mole-cule-1 in human corneal endothelium. Modulation and function. Am.J. Pathol. 138, 525–536.

FUKUSHIMA, M., NAKASHIMA, Y., and SUEISHI, K. (1989). Throm-bin enhances release of tissue plasminogen activator from bovinecorneal endothelial cells. Invest. Ophthalmol. Vis. Sci. 30, 1576–1583.

HANGAI, M., TANIHARA, H., HONDA, Y., and KANEDA, Y.(1998). Introduction of DNA into the rat and primate trabecularmeshwork by fusogenic liposomes. Invest. Ophthalmol. Vis. Sci. 39,509–516.

HAYASHI, K., NAKASHIMA, Y., SUEISHI, K., and TANAKA, K.(1989). Fibrinolytic activity and localization of plasminogen activa-tor in bovine vitreous body and aqueous humor. Jpn. J. Ophthalmol.33, 66–75.

LARKIN, D.F., ORAL, H.B., RING, C.J., LEMOINE, N.R., andGEORGE, A.J. (1996). Adenovirus-m ediated gene delivery to thecorneal endothelium. Transplantation 61, 363–370.

LUCORE, C.L., FUJII, S., and SOBEL, B.E. (1989). Dependence of

fibrinolytic activity on the concentration of free rather than total tis-sue-type plasminogen activator in plasma after pharmacologic ad-ministration. Circulation 79, 1204–1213.

LUNDY, D.C., SIDORI, P., WINARKO, T., MINCKLER, D., andHEUER, D.K. (1996). Intracameral tissue plasminogen activator af-ter glaucoma surgery. Indications, effectiveness, and complications.Ophthalmology 103, 274–282.

MCKAY, D.G. (1972). Participation of components of the blood coag-ulation system in the inflammatory response. Am. J. Pathol. 67,181–210.

OSHIMA, Y., SAKAMOTO, T., YAMANAKA, I., NISHI, T.,ISHIBASHI, T., and INOMATA, H. (1998). Targeted gene transfer tocorneal endothelium in vivo by electric pulse. Gene Ther. 5, 1340–1346.

RAMSBY, M.L., and KREUTZER, D.L. (1994). Fibrin induction ofinterleukin-8 expression in corneal endothelial cells in vitro. Invest.Ophthalmol. Vis. Sci. 35, 3980–3990.

ROWLAND, F.N., DONOVAN, M.J., GILLIES, C., OROURKE, J.,and KREUZER, D.L. (1985). Fibrin: Mediator of in vivo or in vitroinjury and inflammation. Curr. Eye Res. 4, 537–553.

SAKAMOTO, T., TAKAHIRA, K., SANUI, H., KOHNO, T., and IN-OMATA, H. (1993). Intercellular adhesion molecule-1 on rat cornealendothelium in experimental uveitis. Exp. Eye Res. 56, 241–246.

SAKAMOTO, T., UENO, H., GOTO, Y., OSHIMA, Y., YA-MANAKA, I., ISHIBASHI, T., and INOMATA, H. (1998). Retinalfunctional change caused by adenoviral vector-mediate d transfectionof lacZ gene. Hum. Gene Ther. 9, 789–799.

SOZKA, T., and OLSZEWSKI, K. (1986). Plasminogen activators andtheir inhibitors in hormal, hyperplastic and carcinomatous human en-dothelium. Thromb. Res. 42, 835–846.

VAN DE WERF, F., BERGMANN, S.R., FOX, K.A., DE GEEST, H.,HOYNG, C.F., SOBEL, B.E., and COLLEN, D. (1984). Coronarythrombolysis with intravenously administered human tissue-typeplasminogen activator produces by recombinant DNA technology.Circulation 69, 605–610.

WARING, G.O., III, BOURNE, W.M., EDELHAUSE, H.F., andKENYON, K.R. (1982). The corneal endothelium. Normal andpathologic structure and function. Ophthalmology 89, 531–590.

WUDUNN, D. (1997). Intracameral urokinase for dissolution of fibrin orblood clots after glaucoma surgery. Am. J. Ophthalmol. 124, 693–695.

Address reprint requests to:Dr. T. Sakamoto

Department of OphthalmologyFaculty of Medicine

Kyushu University3-1-1 Maidashi, Higashi-ku, Fukuoka 812-8582, Japan

E-mail: tsakamot@ eye.med.kyushu -u.ac.jp

Received for publication February 9, 1999; accepted after re-vision July 27, 1999.

tPA GENE TRANSFER TO CORNEAL ENDOTHELIUM 2557