JOURNAL OF VASCULAR SURGERY Volume 24, Number 1 Special Communication 159

effect on the animals as the drugs must be administered by intraperitoneal injection. We can obtain the same inhibition of intimai thickening by introducing virally transduced smooth muscle cells overexpressing the natural inhibitor of MMPs, TIMP-1, without side effects.

The current smooth muscle cell-based approach to gene transfer is not likely to be of clinical utility because it is somewhat cumbersome and requires cell culture. Nevertheless, it provides proof of concept, which can then be exploited further when better direct gene transfer methods come on line.

C o n c l u s i o n

We have been able to show that vascular smooth muscle cells can serve as useful vectors for transferring genes Of interest into blood vessels. Using smooth muscle cells transduced with replica- tion-defective retrovirus, we have been able to show that this approach is usefi,fl for systemic gene therapy, the development of models of vascular disease, and local vascular pharmacology.

Alexander W Clowes, MD University of Washington School of Medicine Seattle, Wash.

REFERENCES

1. Lynch CM, Clowes MM, Osborne WRA, Clowes AW, Miller AD. Long-term expression of human adenosine deaminase in vascular smooth muscle cells of rats: a model for gene therapy. Proc Natl Acad Sci U S A 1992;89:1138-42.

2. Clowes MM, Lynch CM, Miller AD, Miller DG, Osborne WRA, Clowes AW. Long-term biologic response of injured rat carotid artery seeded with smooth muscle cells expressing retrovirally introduced human genes. J Clin Invest 1994;93: 644-51.

3. Geary RL, Clowes AW, Lau S, Vergel S, Dale DC, Osborne WRA. Gene transfer in baboons using prosthetic vascular grafts seeded with retrovirally transduced smooth muscle cells: a model for local and systemic gene therapy. Hum Gene Ther 1994;5:1211-6.

4. Miller AD, Rosman GL Improved retroviral vectors for gene transfer and expression. Biol Techniques 1989;7:980-90.

5. Osborne WRA, Ramesh N, Lau S, Clowes MM, Dale DC, Clowes AW. Gene therapy for long-term expression oferythro- poietin in rats. Proc Nail Acad Sci U S A 1995;92:8055-8.

iNOS OVEREXPESSION TO MODIFY D I S E A S E S T A T E S

Normal human vascular physiology is dictated by an orchestrated array of counterregulatory mediators collaborating to maintain optimal tissue perfusion. One such mediator is endothelium- derived relaxing factor, first described in 1980, eventually identified as nitric oxide (NO) or a derivative thereof in 1987.1 This NO, originating from the endothelium from a resident constitutively expressed NO synthase enzyme (ecNOS or NOS-3), is a potent smooth muscle relaxant and is responsible for maintaining a state of resting vasodilation. NO has also been demonstrated to possess a number of other vasoprotective properties. NO inhibits platelet adhesion and aggregation, and platelets themselves exhibit the capacity to synthesize an independent source of NO. In vitro data suggest that NO may inhibit vascular smooth muscle cell prolifera- tion through cell cycle arrest. Finally, NO may reduce leukocyte adherence and infiltration of the endothelial barrier. All of these functions may be essential defenses against vascular occlusive complications.

The precise role of NO in the pathogenesis of vascular occlusive disorders such as atheroscterosis and restenosis is still under intense investigation, but several studies hint at the relevance of NO depletion in the establishment and progression of these maladaptive states. An early event in the development of an atherosclerotic plaque is the accumulation of lipid-laden macro- phages in the intima with disturbance of the endothelium, z Endothefial NO synthesis may become impaired at these locations. In addition, oxidized lipoproteins have been demonstrated to inactivate NO, thereby creating an environment that lacks NO. Diabetic patients are especially prone to an aggressive form of atherosclerosis. Their predilection may stem in part from progres- sive accumulation of glycosylated products that may also inactivate NO. Indeed, atherosclerotic arteries are less vasoresponsive to agonists such as acetylcholine but are still able to vasodilate in response to authentic NO, suggesting reduced endogenous NO release.

Vascular restenosis after therapeutic manipulation may also result from inadequate NO availability. Blood vessel injury is characterized by endothelial denudation with damage to the underlying smooth muscle cells. Exposed collagen and smooth muscle provide a prothrombogenic surface that invites leukocyte and platelet adhesion. A number of chemotactic and mitogenic factors are then released that facilitate cellular migration and proliferation, resulting in the creation of a neointima. It has been reported that the inducible NOS (iNOS or NOS-2) isoform is expressed transiently, beginning immediately and persisting for 1 to 2 weeks, in arterial smooth muscle cells in response to balloon catheter-induced injury in rat carotid arteries? Unlike NOS-3, NOS-2 expression is detected in cells only after cytokine or mechanical perturbation, and the enzyme is several magnitudes more active in the absence ofagonist stimulation. NOS-2 induction as part of the vascular healing response may provide an alternate source of NO until endothelial integrity can be reestablished. Restenosis may derive from a defect in this response.

Although the role of NO in the establishment of intimal hyperplasia is still unclear, the benefit of NO in preventing this response may be more evident. McNamara et al.4 reported a 39% reduction in intimai hyperplasia in injured rabbit carotid arteries with oral supplementation of the NOS substrata L-arginine. Similarly, Davies et ai.5 attenuated neointima formation by 47% and preserved vasoreactivity in vein grafts that were interposed in the carotid circulation with supplemental L-arginine alimentation. Finally, systemic administration of an NO donor 6 reduced the intimal thickness by 82% and accelerated reendothelialization in injured rat carotid arteries. These results, although indirect and in rodent models, strongly suggest the benefit of increased NO availability for the attenuation of the hyperplastic reaction to vascular injury.

With these observations, the application of NO in the preven- tion ofrestenotic complications and even atherosclerosis may have therapeutic utility. To evaluate this hypothesis, a method of delivering a therapeutic dose of NO is necessary. One such option is systemic administration of NO donors. However, the potential for hypotensive side-effects is real. Also, the complexity of NO biology and all of its purported cytotoxic effects are incentives to localize NO exposure in a site-specific fashion. To achieve this, the best method may be gene transfer ofa NOS gene to sites of vascular injury at the time of therapeutic intervention, such as immediately after angioplasty. Indeed, v o n d e r Leyen et al. 7 examined the efficacy of liposome-mediated NOS-3 gene transfer at reducing intimal hyperplasia in a rat carotid artery injury model. The authors demonstrated reconstitution of NOS-3 activity in the injured

JOU1GNAL OF VASCULAR SURGERY 1 6 0 Special Communication July 1996

vessels to a level comparable to uninjured vessels. The NOS- 3-transfected vessels showed a -70% reduction in neointimal thickness as compared with control injured carotid arteries. These results suggest NOS gene transfer may be beneficial in restenosis.

A very important consideration that must be weighed is gene delivery efficiency by viral or nonviral methods. For clinical applications, vascular gene transfer must be accomplished with a minimal period of flow occlusion, especially in the coronary circulation, where minutes of occlusion can result in significant myocardial ischemia or infarction. With even the most efficient delivery system, that being the adenoviral vector, gene transfer efficiency will be low. For this reason, we hypothesize that gene delivery of the NOS-2 isoform may have some advantages over NOS-3. Low-efficiency transfer o f this extremely active enzyme may still provide sufficient local concentrations of NO to achieve a biologic effect. In addition, NOS-2 activity does not require calcium fluxes and will be maximal in the absence of agonist stimulation.

Cytotoxicity from NOS-2 overexpression in native cells is a potential concern. Also, all NOS enzymes require the cofactor tetrahydrobiopterin (BH4). Certain cells lack constitutive biosyn- thesis of this cofactor and may be unable to support NOS-2 activity, s Because of these concerns, we evaluated the feasibility of NOS-2 gene transfer to vascular target cells. We constructed a retroviral vector carrying the human NOS-2 cDNA and neomycin phosphotransferase (DFGiNOS). 9 Sheep vascular endothelial cells infected with DFGiNOS and selected with neomycin supported mammal NOS-2 activity, producing over 20-fold more nitrite (NO2-) than cells infected with a control retrovirus ( 155.0 + 10.7 n m o l / m g prote in /24 hr vs 5.5 _+ 1.1), without requirements for supplemental BH4. In contrast, rat vascular smooth muscle cells infected with DFGiNOS had no detectable NO 2- accumulation over control cells until BH 4 was provided (37.7 + 2.6 n m o l / m g protein/24 hrwithout BH4,291.8 +_ 10.4withBH4),demonstrat- ing the insufficient cofactor levels in these cells. Interestingly, continuous expression of NOS-2 with NO synthesis did not result in significantly increased cytotoxicity as measured by Slchromium or lactate dehydrogenase release.

We adapted an ex vivo organ culture system of arterial injury l° to assess the biologic efficacy of human NOS-2 gene transfer for preventing myointimal hyperplasia. Porcine femoral arterial seg- ments were infected with the DFGiNOS retrovirus or a control retrovirus 5 days after balloon-catheter injury and then cultured for another 9 days. By the fifth day after surgery, near maximal smooth muscle proliferation is occurring, 1° a favorable environ- ment for retroviral infection. Vessels infected with DFGiNOS demonstrated a threefold increase in NO synthesis and a 15-fold increase in cGMP release over uninjured or control manipulated vessels. All early experiment had revealed a dependence of NOS-2 activity in these arterial segments on exogenous BH4, which was subsequently provided to all the experimental groups. Immuno- histochemical analysis revealed only a 1% to 2% gene transfer efficiency, as would be expected with a retroviral vector. Balloon injury of the arterial segments resulted in a 53% increase in total myointimal thickness. Vessels expressing NOS-2, however, had a complete abrogation of this hyperplastic response and resembled uninjured controls; the administration ofa NOS inhibitor reversed this protective effect.

Despite very low gene transfer efficiency, NOS 2 delivery to injured arterial segments successfully reduced intimal hyperplasia. The requirement for BH 4 is consistent with preferential delivery of the NOS-2 gene to the smooth muscle cells. Ongoing studies are aimed at determining whether the cofactor will be required for

NOS-2 gene transfer in vivo and whether NOS-2 transfer will be of benefit in models of restenosis in large animals.

Timothy R. Billiar, MD Edith Tzeng, MD Larry L. Shears II, MD University of Httsburgh Medical Center Pittsburgh, Pa.

R E F E R E N C E S

1. Billiar T K Nitric oxide: novel biology with clinical relevance. Ann Surg 1995;221:339-49.

2. Ross R. The pathogenesis ofatherosclerosis: a prospective for the 1990s. Nature 1993;362:801-9.

3. Hansson GK, Geng Y, Holm ], Hardhammer P, Wenmnalm A, Jennische E. Arterial smooth muscle cells express nitric oxide synthase in response to endothelial injury. ] Exp Med 1994; 180:733-8.

4. McNamara DB, Bedi B, Aurora H, et al. L-arginine inhibits balloon catheter-induced intimal hyperplasia. Biochem Bio- phys Res Commun 1993;I93:291-6.

5. Davies MG, Kim JH, Dalen H, Malchoul RG, Svendsen E, Hagen PO. Reduction of experimental vein graft intimal hyperplasia and preservation of nitric oxide-mediated relax- ation by the nitric oxide precursor L-arginine. Surgery 1994; 116:557-68.

6. Guo J, Milhoan KA, Tuan RS, Leffer AM. Beneficial effect of SPM-S 185, a cysteinecontaining nitric oxide donor, in rat carotid artery intimal injury. Circ Res 1994;75:77-84.

7. v o n d e r Leyen HE, Gibbons GH, Morishita R, et al. Gene therapy inhibiting neointimal vascular lesion: in vivo transfer of endothelial cell nitric oxide synthase gene. Proc Natl Acad Sei U S A 1995;92:1137-41.

8. Tzeng E, Bllliar TR, Robbins PD, Loftus M, Stuehr DJ. Expression of human inducible nitric oxide synthase in a tetrahydrobiopterin (BH4)-deficient cell line: BH 4 promotes assembly of enzyme subunits into an active dimer. Proc Natl Acad Sci U S A (in press).

9. Tzeng E, Shears LL, Robbins PD, et al. Vascular gene transfer o f the human inducible nitric oxide synthase: characterization of activity and effects on myointimal hyperplasia. Mol Med (Submitted).

10. Takeshita S, Gai D, Leclere G, et aft. Increased gene expression after liposome-mediated arterial gene transfer associated with intimal smooth muscle cell proliferation: in vitro and in vivo findings in a rabbit model o f vascular injury. J Clin Invest 1994;93:652-61.

D I R E C T T R A N S F E R - -

V I R A L A P P R O A C H E S

A D E N O V I R A L VECTORS: P R O M I S E S A N D P I T F A L L S

Human adenovirus, representing over 40 different serotypes, are 36 kb double-stranded DNA viruses. The molecular structure and regulation of this virus has been well studied over the years. The virus was first used in the 1980's for introducing foreign genes into cells. More recently, the E1 region of the virus that expresses transcriptional factors necessary to induce appropriate synthesis o f the more than 70 viral proteins needed for viral replication has been removed, supplied in trans, and replaced with therapeutic gene sequences. Over the past 5 years, multiple studies performed by different laboratories have demonstrated that recombinant ade-