Corneal Transduction to Inhibit Angiogenesis and Graft Failure

Raghu C. Murthy,1,2 Trevor J. McFarland,1,2 Jon Yoken,1 Sandy Chen,1 Chris Barone,1 Dorthea Burke,1 Yi Zhang,1 Binoy Appukuttan,1 and J. Timothy Stout1

PURPOSE. To test whether lentivirus-mediated expression of an a significant number of corneal transplantations end in rejec- endostatin::kringle-5 (E::K-5) fusion gene has an inhibitory ef- tion and graft failure every year. The need for regrafting a failed fect on neovascularization and failure of corneal transplants. transplant is one of the top two indications for corneal trans- ETHODS A lentiviral vector containing a fusion transgene plantation in many centers in the United States, competing M . 2 comprising the human endostatin gene and the kringle-5 do- with pseudophakic bullous keratopathy in frequency. The main of the human plasminogen gene (E::K-5) was used for major risk factors for rejection are prior corneal transplanta- 3 transduction of corneal buttons ex vivo. The corneal buttons tion, glaucoma, and preoperative corneal vascularization. Pre- were transplanted after overnight incubation in media contain- vention of corneal neovascularization would be a pivotal step ing either lentivirus or PBS. Sixteen rabbits underwent allo- toward inhibiting graft failure and rejection. genic penetrating keratoplasty in one eye. The area of neovas- Endothelial cell migration, proliferation, and modification of cularization from the limbus to within the graft was the extracellular matrix are all involved in angiogenesis. Mole- documented after surgery. RT-PCR was performed to demon- cules that stimulate or inhibit vessel growth modulate angio- 4 strate the presence of transgene mRNA within the graft. His- genesis. A variety of pathologic states can result from a de- topathology was used to analyze neovascularization, inflamma- creased production of inhibitors or an increased production of 4 tion, and rejection morphology. stimulators. For example, an elevated level of vascular endothelial growth factor (VEGF) is associated with the abnormal RESULTS. Less neovascularization was observed in corneas neovascularization in diabetic retinopathy.5 The development treated with the lentivirus E::K-5 fusion vector. Early onset and of a biological agent to combat proangiogenic stimulation profound neovascularization was observed in control eyes. would be a useful tool. Prior attempts to inhibit rejection have E::K-5–treated animals did not have graft failure, whereas five included ex vivo gene therapy in a sheep corneal transplanta- of the six control animals had graft failure, as classified by 6 opacification of the graft. All E::K-5 transduced corneas tested tion model. Donor grafts were transduced with an adenovirus- were positive by RT-PCR for the unique fusion gene sequence. mediated delivery system with a gene encoding IL-10, which Histopathology corroborated a significant increase of blood downregulates some of the steps in the cascade of cell-medi- vessel presence and inflammatory reaction in control com- ated immunity. These animals showed prolonged corneal allo- pared with treated eyes. graft survival. Endostatin, a 20-kDa C-terminal fragment of collagen XVIII CONCLUSIONS. Corneas transduced with a lentivirus containing (Fig. 1) has been shown to be an endogenous inhibitor of an endostatin::kringle-5 fusion gene demonstrated an inhibi- angiogenesis and tumor growth in a hemangioendothelioma tion of neovascularization and graft failure. E::K-5 gene trans- model in rats.7 Endostatin impedes proliferation and migration duction through a lentiviral vector system may be a useful by downregulating the expression of genes involved in cell adjunct to prevent graft neovascularization and corneal graft growth, antiapoptosis, and angiogenesis, specifically within rejection in high-risk corneal transplants with antecedent re- endothelial cells.8 Endostatin has also been reported to inhibit jection or neovascularization. (Invest Ophthalmol Vis Sci. cell matrix adhesion in endothelial cells and to promote a G1 2003;44:1837–1842) DOI:10.1167/iovs.02-0853 arrest through inhibition of cyclin D1.9,10 Angiostatin, a protein derived from proteolytic cleavage of an internal fragment ore than 30,000 corneal transplantations are performed of plasminogen (Fig. 1), containing up to four kringle domains, Meach year in the United States—more than all heart, inhibits angiogenesis-dependent tumor growth.11,12 Kringle-5 kidney, and liver transplantations.1 Corneal transplantation (or of plasminogen shares 46% to 57% amino acid identity to each penetrating keratoplasty, PK) is one of the most successful of of the four kringle domains of angiostatin and is a more potent such operations in humans, with success exceeding 90%. Still, inhibitor of basic fibroblast growth factor (bFGF)–stimulated angiogenesis than is angiostatin alone.13 Kringle-5 acts specifically on endothelial cells by inhibiting cell migration, although 14 1 the exact mechanism has not been fully elucidated. Reports From the Clayton Gene Therapy Laboratory, Casey Eye Institute, have implied that interactions between angiostatin and integrin Oregon Health Sciences University, Portland, Oregon. ␣ ␤ 2Contributed equally to the work and therefore should be consid- v 3 (an integrin common to immature, newly synthesized ered equivalent senior authors. vessels), and caveolin-1 may be essential components in this 15,16 Supported by the Clayton Foundation for Research, Research to process. The angiostatic fusion protein consisting of Prevent Blindness, and the Heed Ophthalmic Foundation. mouse endostatin and mouse angiostatin has been shown to Submitted for publication August 20, 2002; revised October 1, have a more potent biological effect than either gene product 2002; accepted November 7, 2002. alone in an in vitro cancer model.17 In the current study, the Disclosure: R.C. Murthy, None; T.J. McFarland, None; J. Yo- biologically active domains of human endostatin 18 and human ken, None; S. Chen, None; C. Barone, None; D. Burke, None; Y. kringle-5 were linked to make the fusion protein E::K-5 for the Zhang, None; B. Appukuttan, None; J.T. Stout, None purpose of producing a protein able to inhibit both endothelial The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertise- cell proliferation and migration. ment” in accordance with 18 U.S.C. §1734 solely to indicate this fact. Lentiviral vectors are efficient at transducing a number of Corresponding author: J. Timothy Stout, Casey Eye Institute, 3375 target cells, have a large transgene-carrying capacity, and ex- SW Terwilliger Boulevard, Portland, OR 97201; [email protected]. hibit early-onset expression. As a result of proviral integration

Downloaded from tvst.arvojournals.org on 09/29/2021 1838 Murthy et al. IOVS, May 2003, Vol. 44, No. 5

BamHI site and the reverse primer (5ЈCAATGTATCGGATCCTGTC- GAGCTAGC3Ј) containing a BamHI site. This fusion gene encodes 20

amino acids from the human IL-2 secretion signal, amino acids Ala1333

to Lys1516 from the human collagen XVIII gene (endostatin), an 8-ami-

no-acid elastin linker motif VPGVGTAS, and amino acids Pro466 to

Asp566 from the human plasminogen gene. The PCR fragment was digested with BamHI and ligated into the lentiviral vector pHRЈ, under the transcriptional control of the cytomegalovirus (CMV) promoter (Fig. 2). Construction of the E::Kr5 fusion gene was confirmed by direct sequencing of the transgene insert. Replication-deficient E::Kr5 lentivirus particles were prepared as previously described.19 Virus particles were resuspended in a minimal volume of PBS and stored at Ϫ80°C. FIGURE 1. Diagram of the proteins plasminogen and collagen XVIII. The fusion protein E::K-5 was derived from the kringle 5 domain of plasminogen and the cleaved product endostatin 18 of collagen Viral Assay XVIII. The presence of virus particles was confirmed with a quantitative HIV-1 p24 antigen ELISA kit (ZeptoMetrix, Buffalo, NY), according to into the host genome, sustained long-term expression is 18,19 the manufacturer’s instructions. To ensure the infectivity of the lenti- achieved in many tissue types, including nondividing cells. viral reagent, 10, 50, and 100 ␮L of virus was placed into a six-well In a previous study, we confirmed that a lentivirus containing plate of human dermal microvascular endothelial cells (HDMECs) for a reporter gene transduces all cornea cell types efficiently, and 20 minutes at 37°C. Medium 131 (Cascade Biologicals, Portland, OR) expression is persistent in corneal cultures maintained up to 60 20 was then added, and cells were incubated at 37°C5%CO2 for 5 days, days. To study the neovascularization and rejection of cor- with medium changes every other day. On day 5, RNA was isolated neal transplants, we used a rabbit model of allogenic corneal with extraction reagent (TRIzol; Gibco-BRL, Grand Island, NY), and transplantation. standard RT-PCR reactions and analyses were performed. The forward Prior studies have documented a high rate of transplant primer (5ЈTCTGAGGGTCCGCTGAAGCCCGGGG3Ј) and reverse rejection in rabbits with retained graft– host sutures after Ј Ј 21 primer (5 CAAATGAAGGGGCCGCAC3 ) flanked the elastin linker re- PK. In this study we investigated the ability of a lentiviral gion and thus amplified only the fusion transcript. vector to deliver a potentially antiangiogenic fusion gene to donor corneal buttons and to correlate successful transduction and grafting with postoperative neovascularization and Corneal Transduction graft failure. All animal procedures were performed under Institutional Animal Care and Use Committee (IACUC)–approved protocols, which conform to METHODS the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research standard for humane animal care. Sixteen 7-mm trephined Lentivirus Production donor corneas were obtained from eight New Zealand White rabbits. An endostatin-kringle-5 (E::Kr5) fusion cDNA was amplified by PCR Each button was harvested in a sterile surgical fashion and placed in from the E::Kr5 pBlast vector (InvivoGen, San Diego, CA) using the 2-mL of corneal storage medium containing chondroitin sulfate dex- forward primer (5ЈCTGAGGGATCCGGCGAAGGAG3Ј) containing a tran medium supplemented with 100 ␮g/mL gentamicin and 200

FIGURE 2. Plasmid map of the lentiviral vector containing the cytomegalovirus (CMV) promoter, E::K-5 fusion cDNA, internal ribosomal entry site (ires), and the enhanced green ﬂuorescent protein (eGFP) cDNA. Restriction enzymes used for cloning are highlighted (map created with Redasoft Visual Cloning 2000; Reda- soft Corp. Toronto, Ontario, Canada).

Downloaded from tvst.arvojournals.org on 09/29/2021 IOVS, May 2003, Vol. 44, No. 5 Corneal Transduction to Inhibit Angiogenesis and Graft Failure 1839

FIGURE 3. Quantitative analysis of the area of neovascularization (NV). (A) Average area (av) Ϯ SD of NV in postoperative E::K-5– and control-treated corneas at each time point studied. The average total area of rabbit corneas was 143 mm2. Only one subject (treated with eGFP) had a surviving graft at day 36; therefore, there was no deviation. (B) Standard t-test using a one-tailed distribution comparing results obtained with E::K-5 with those obtained with eGFP or PBS. Differences in average area of NV are signiﬁcant at all time points (P ϭ 0.05 or less), with the exception of eGFP at day 9.

␮g/mL streptomycin sulfate (Optisol-GS; Chiron, Irvine, CA). The me- 30, and 40. Fresh corneal tissue was placed in either extraction reagent dium was spiked with 50 ␮L E::K-5 lentivirus, 50 ␮L enhanced green (TRIzol; Gibco-BRL) for RT-PCR or formalin for histopathologic study. fluorescent protein (eGFP) lentivirus, or 50 ␮L PBS. These buttons were incubated for 18 hours at 37°C. RESULTS Corneal Transplantation in Rabbits A fusion cDNA coding for the 20 amino acids of human IL-2 secretion signal, human endostatin (derived from the 184 General anesthesia was induced by mask administration of isoflurane. amino acids of the carboxyl terminus of collagen XVIII), an Heparin and viscoelastic were instilled by paracentesis into the ante- 8-amino-acid elastin linker sequence, and the 101 amino acids rior chamber (AC). A 7-mm Hessburg-Barron trephine (Jedmed Instru- of human plasminogen (corresponding to the fifth kringle ments, St. Louis, MO) was used to remove the host corneal button. A domain) was successfully cloned into the pHRЈ CMV-IRES-eGFP 7-mm–trephined corneal button previously treated with EK-5, eGFP, or vector. Quantitative ELISA confirmed presence of virus parti- PBS was rinsed in balanced salt solution (BSS, Alcon Laboratories, Fort cles and based on standard curve interpolation, 230 pg/mL of Worth, TX) and grafted with 16 interrupted 7-0 nylon sutures. Subcon- antigen was detected, correlating to approximately 108 virus junctival injections (0 0.1 mL) of enrofloxacin (23 mg/mL, Baytril; particles/mL (Zeptometrix). Transcription of the fusion gene Bayer Animal Health, Shawnee Mission, KS) and 40 mg/mL triamcino- was verified by RT-PCR of transduced HDMECs (data not lone (Kenalog; Westwood Squibb Pharmaceuticals, Buffalo, NY) were shown). given. After surgery, all animals received a single dose of topical Postoperative corneal neovascularization was significantly atropine (1%) and a single dose of carprofen (Rimadyl; Pfizer, New lower in E::K-5 lentivirus–treated eyes than in either eGFP York, NY) at 2.5 mg/kg administered subcutaneously, as well as 1 drop lentivirus– or PBS-treated control eyes on postoperative days 5, tobramycin twice daily for 5 days and buprenorphine (Buprenex; 12, 14, 16, 18, 24, 28, and 36 (Fig. 3). Reckitt Benckiser Pharmacy, Hull, UK) at 0.1 mg/kg subcutaneously as All PBS- and all eGFP lentivirus–treated corneas exhibited necessary. No topical steroid drops were administered after surgery. neovascular arborization into the graft bed (Fig. 4B–D, 4G). Vascularization in all control animals was relatively unchanged after postoperative day 9. None of the 10 E::K-5–treated cor- Measurement of Neovascularization and neas had new vessels extending into the graft (Fig. 4E, 4F). Evaluation of Graft Rejection Three of 3 PBS-treated and 2 of 3 eGFP-treated corneas exhibited corneal opacification and graft failure (Figs. 4B, 4D, 4G), Neovascularization was followed by slit lamp examinations on postop- whereas none of the 10 E::K-5–treated grafts completely opac- erative days 5, 9, 12, 14, 16, 24, 28, and 36. Measurements of neovas- ified or failed by postoperative day 39 (Fig. 4F). PBS-treated cularization were made with a portable slit lamp by a single masked control corneas exhibited graft failure by postoperative days 14 observer, an ophthalmologist. Vessel growth onto the clear cornea was and 18, and eGFP-treated corneas by day 24, with the excep- noted in millimeters and number of clock hours. Neovascularization tion of one eGFP-treated control, which remained stable for the was quantified by calculating the wedge-shaped area of vessel growth duration of the experiment (day 36). All five grafts tested by with the formula: A ϭ (clock hours/12)[␲(R2 Ϫ R22], where R is the RT-PCR for the presence of fusion gene transcripts were pos- total radius and R2 is the radius from the center to the border of vessel itive on postoperative days 30 and 40 (Fig. 5A, 5B). All control growth. Graft rejection was evaluated by portable slit lamp. Graft and nonsurgical eyes were negative, according to fusion gene failure was judged by the presence of persistent corneal graft edema RT-PCR. with opacification of 100% of the graft. Serial photographs of the Analysis of serial sections revealed more neovascularization cornea were taken. Animals were killed on postoperative days 9, 21, and basophilic inflammatory infiltrates in control eyes than in

Downloaded from tvst.arvojournals.org on 09/29/2021 1840 Murthy et al. IOVS, May 2003, Vol. 44, No. 5

FIGURE 4. Representative photographs of neovascularization and graft failure from corneal buttons treated with either E::K-5 or control (eGFP or PBS). (A) Postoperative day (PD)0 graft treated with E::K-5. (B) PD14, eGFP-treated control showing infiltration of new vessels and corneal opacification and (C) dense neovascularization. (D) PD15, PBS-treated control showing opacified graft with vessel growth spanning the donor–host border. (E) PD18, E::K-5–treated graft showing absence of neovascularization extending across the graft border and clarity of the graft. (F) PD40, E::K-5–treated graft showing minimal if any neovascularization. (G) PD40, PBS-treated control showing opaque graft with new vessels extending into the graft.

E::K-5–treated or nonsurgical eyes. Histopathologic study of a tion of the recipient corneal bed, the graft–host interface, or site of retained suture, often the location of an inflammatory subsequently of the graft itself. The development of new blood infiltrate, was void of inflammatory cells in the examined vessels extending into the graft is associated with high levels of E::K-5-treated cornea (Fig. 6C). inflammatory cells, plasma proteins, and cytokines within the graft and is often a presage of rejection and failure. Believing DISCUSSION that corneal neovascularization promotes rejection, investiga- tors have long sought medical or surgical approaches to abort The success of corneal transplantation has expanded the indi- the process. cations for this surgery and has increased the number of ker- We describe a successful approach to inhibiting the devel- atoplasties performed annually. Despite advances in corneal opment of post-PK neovascularization in a rabbit model. This transplantation, such as enhanced storage media and improved approach is based on the ability of lentiviral vectors to trans- donor requirements, graft rejection remains a major prob- duce corneal tissues ex vivo, with genes known to be antian- lem.22 A major risk factor for graft rejection is neovasculariza- giogenic in animal models of tumor angiogenesis. We tested a

FIGURE 5. In vivo RT-PCR demonstrating transcription of E::K-5 in corneal buttons. Only the right eyes received corneal transplants; left eyes were not altered. (A, top and middle) Day 30: lanes 1, 3, 5, 7, and 9: E::K-5–positive right-eye grafts; lanes 2, 4, 6, 8, and 10: left-eye control; lane 11: RT-negative control; lane 12: PCR-negative control; lane 13: E::K-5 cDNA–positive control. (A, bottom) ␤-Actin mRNA controls for lanes 1–8.(B) Day 40 harvest: lane 1: E::K-5 RT-PCR–positive button; lane 2: contralateral control; lanes 3 and 4: ␤-actin control. (A, B) Lane M: molecular weight markers.

Downloaded from tvst.arvojournals.org on 09/29/2021 IOVS, May 2003, Vol. 44, No. 5 Corneal Transduction to Inhibit Angiogenesis and Graft Failure 1841

FIGURE 6. Corneal histopathology at postoperative day (PD)40 after E::K-5 and control treatments. (A) Naive cornea (normal). (B, C, D) E::L-5–treated buttons showing tissue devoid of vessel proliferation with only a slightly elevated presence of immune inﬁltrates (basophils). (E, F, G) eGFP-treated control samples showing an increased presence of immune cells and many blood vessels. (H, I) PBS-treated control samples, with appearance similar to that of the eGFP sections, with increased immune cell presence and vessels apparent in the stroma.

fusion gene that combines the human endostatin gene and the growth in solid tumors in a rat hemangioendothelioma model.7 fifth kringle element of the human plasminogen gene as an Kringle-5 is a specific inhibitor of endothelial cell proliferation inhibitor of new blood vessel growth. These moieties are in bovine capillary cells and prevents migration of endothelial preceded by the IL-2 secretory signal and are fused by an cells in a number of cell cultures stimulated by bFGF.14 En- elastin linker motif. Once transduced, corneal tissues demon- dostatin and angiostatin proteins given together have been strated a significantly decreased propensity for postoperative shown to cause more regression of blood vessels and ovarian neovascularization. cancer cells in an experimental model, than when adminis- Treatment of corneal buttons with E::K-5 by lentiviral vec- tered individually.23,24 Similarly, these two proteins, when tor prevented new vessel growth onto the donor graft in all given together in an Fc-angiostatin plus Fc-endostatin combi- treated corneas. Histologic study revealed a marked decrease in nation, resulted in a significant reduction of tumor size in inflammation in E::K-5–treated corneas, including the areas Ripa1-Tag2 mice; given separately, they had a markedly re- around retained sutures, a commonly inflamed area. Further- duced effect.25 We wanted to test a potentially bifunctional more, there was no evidence of graft failure, as measured by protein that combines the antiangiogenic domains of collagen persistent corneal edema and corneal opacification in XVIII and plasminogen, as an inhibitor of corneal angiogenesis. E::K-5–treated corneas, whereas five of six control corneas Although the lentiviral vectors used are replication defec- exhibited evidence of opacification and failure. tive, questions about the safety of genetic transduction are Corneas expressing E::K5 are less vascularized and less important and appropriate. Transduction of corneal tissue ex inflamed than control corneas, however, the mechanism of vivo is an attractive alternative to somatic in vivo therapy. ␣ ␤ inhibition is not clear. Disruption of v 3 integrin distribution, Inducible control of transgene expression may also be benefi- expression of caveolin-1, or levels of cyclin D1 or changes in cial. Tetracycline-regulated promoters have been used in the the levels of tyrosine phosphorylation of focal adhesion kinase past as a system for controlling gene expression by exogenous and paxillin may independently or in combination have a role administration of tetracycline and should be evaluated.26,27 in this result.9,10,15,16 Our data suggest that ex vivo lentiviral transduction of donor The ability of the fusion protein E::K-5 to retard new blood corneal tissue with a fusion antiangiogenic gene, before PK, vessel growth may be augmented by its potentially bifunctional may increase the likelihood of long-term graft survival and be a structure. Endostatin has been shown to inhibit new vessel useful surgical adjunct.

Downloaded from tvst.arvojournals.org on 09/29/2021 1842 Murthy et al. IOVS, May 2003, Vol. 44, No. 5

References 15. Tarui T, Miles LA, Takada Y. Specific interaction of angiostatin with ␣ ␤ integrin v 3 in endothelial cells. J Biol Chem. 2001;276:39562– 1. Council on Scientific Affairs. Report of the organ transplant panel: 39568. corneal transplantation. JAMA 1988;259:719–722. 16. Liu J, Razani B, Tang S, Terman BI, Ware JA, Lisanti MP. Angiogen- 2. Flowers CW, Chang KY, McLeod SD, et al. Changing indications esis activators and inhibitors differentially regulate caveolin-1 ex- for penetrating keratoplasty, 1989–1993. Cornea. 1995;14:583– pression and caveolae formation in vascular endothelial cells: an- 588. giogenesis inhibitors block vascular endothelial growth factor- 3. Price FW Jr, Whitson WE, Johns S, Gonzales JS. Risk factors for induced down-regulation of caveolin-1. J Biol Chem. 1999;274: corneal graft failure. J Refractive Surg. 1996;12:134–143. 15781–15785. 17. Hampl M, Tanaka T, Albert PS, Lee J, Ferrari N, Fine HA. Thera- 4. Carmeliet P, Jain RK. Angiogenesis in cancer and other diseases peutic effects of viral vector-mediated antiangiogenic gene transfer (Review). Nature. 2000;407:249–257. in malignant ascites. Hum Gene Ther. 2001;12:1713–1729. 5. Adamis AP, Miller JW, Bernal MT, et al. Increased vascular endo- 18. Akkina RK, Walton RM, Chen ML, Li QX, Planelles V, Chen IS. High thelial growth factor (VEGF) levels in the vitreous of eyes with efficiency gene transfer into cd34 (ϩ) cells with a human immu- proliferative diabetic retinopathy. Am J Ophthalmol. 1994;118: nodeficiency virus type 1-based retroviral vector pseudotyped 445–450. with vesicular stomatitis virus envelope glycoprotein G. J Virol. 6. Klebe S, Sykes PJ, Coster DJ, et al. Prolongation of sheep corneal 1996;70:2581–2585. allograft survival by ex vivo transfer of the gene encoding inter- 19. Naldini L, Blomer U, Gallay P, et al. In vivo gene delivery and stable leukin-10. Transplantation. 2001;71:1214–1220. transduction of nondividing cells by a lentiviral vector. Science. 7. O’Reilly MS, Boehm T, Shing Y, et al. Endostatin: an endogenous 1996;272:263–267. inhibitor of angiogenesis and tumor growth. Cell. 1997;88:277– 20. Wang X, Appukuttan B, Ott S, et al. Efficient and sustained trans- 285. gene expression in human corneal cells mediated by a lentiviral 8. Shichiri M, Hirata Y. Antiangiogenesis signals by endostatin. FASEB vector. Gene Ther. 2000;7:196–200. J. 2001;15:1044–1053. 21. Khodadoust A. Penetrating keratoplasty in the rabbit. Am J Oph- 9. Dixelius J, Cross M, Matsumoto T, Sasaki T, Timpl R, Claesson- thalmol. 1968;66:899–905. Welsh L. Endostatin regulates endothelial cell adhesion and cy- 22. Wilson SE, Kaufmann HE. Graft failure after penetrating kerato- toskeletal organization. Cancer Res. 2002;62:1944–1947. plasty. Surv Ophthalmol. 1990;34:325–356. 10. Hanai J, Dhanabal M, Karumanchi SA, et al. Endostatin causes G1 23. Yokoyama Y, Dhanabal M, Griffioen AW, Sukhatme VP, Ramakrish- arrest of endothelial cells through inhibition of cyclin D1. J Biol nan S. Synergy between angiostatin and endostatin: inhibition of Chem. 2002;277:16464–16469. ovarian cancer growth. Cancer Res. 2000;60:2190–2196. 24. Scappaticci FA, Smith R, Pathak A, et al. Combination angiostatin 11. O’Reilly MS, Holmgren L, Shing Y, et al. Angiostatin: a novel and endostatin gene transfer induces synergistic antiangiogenic angiogenesis inhibitor that mediates the suppression of metastases activity in vitro and antitumor efficacy in leukemia and solid Cell by a Lewis lung carcinoma. . 1994;79:315–328. tumors in mice. Mol Ther. 2000;3:186–196. 12. O’Reilly MS Angiostatin: an endogenous inhibitor of angiogenesis 25. Bergers G, Javaherian K, Lo KM, Folkman J, Hanahan D. Effects of and of tumor growth. EXS. 1997;79:273–294. angiogenesis inhibitors on multistage carcinogenesis in mice. Sci- 13. Cao Y, Chen A, An SSA, et al. Kringle-5 of plasminogen is a novel ence. 1999;284:808–812. inhibitor of endothelial cell growth. J Biol Chem. 1997;272: 26. Kafri T, Praag HV, Gage FH, Verma M. Lentiviral vectors: regulated 22924–22928. gene expression. Mol Ther. 2000;1:516–521. 14. Ji W-R, Barrientos LG, Llinas M, et al. Selective inhibition by 27. Apparailly F, Millet V, Noel D, Jacquet C, Sany J, Jorgensen C. kringle-5 of human plasminogen on endothelial cell migration, an Tetracycline-inducible interleukin-10 gene transfer mediated by an important process in angiogenesis. Biochem Biophys Res Com- adeno-associated virus: application to experimental arthritis. Hum mun. 1998;247:414–419. Gene Ther. 2002;13:1179–1188.

Downloaded from tvst.arvojournals.org on 09/29/2021