Gene Transfer to Cerebral Blood Vessels After Subarachnoid Hemorrhage
Background and Purpose Vasospasm remains a major cause of morbidity and mortality after aneurysmal subarachnoid hemorrhage. One step toward gene therapy to prevent spasm of cerebral vessels is to determine whether subarachnoid blood prevents transgene expression.
Methods Vasospasm was induced in mongrel dogs using the double-hemorrhage intracranial-injection model. Diameter of the basilar artery was assessed by angiography, and profound vasospasm (>50% decrease in diameter) was demonstrated at 4 and 7 days. Recombinant adenovirus expressing nuclear-targeted β-galactosidase (reporter gene) under the control of the cytomegalovirus promoter was injected into the cisterna magna at the same time as (n=9) or 2 days after (n=4) injection of blood for induction of vasospasm. Brains were removed and examined histochemically for expression of nuclear β-galactosidase.
Results At 2 to 7 days, β-galactosidase was expressed in leptomeninges over the brain stem, cortex, cerebral arteries, in small vessels in the cerebrum and brain stem, and in the ependymal lining of the ventricles. Transgene expression was observed in adventitia of blood vessels but not in vascular muscle or endothelium. Transgene expression was observed after simultaneous injection of virus and blood or when virus was injected 2 days after blood.
Conclusions The findings indicate that intracisternal injection of recombinant adenovirus can be used for gene transfer to cerebral blood vessels and overlying meninges, even in the presence of cisternal blood. We speculate that transfer of genes using recombinant viral vectors that encode for enzymes with vasodilator function to cerebral blood vessels and perivascular tissues may be useful for prevention or treatment of cerebral vasospasm after subarachnoid hemorrhage.
Vasospasm remains a major cause of morbidity and mortality in patients with SAH.1 2 Some therapeutic approaches are promising,3 4 5 6 but none has been shown to completely ameliorate vasospasm or the ischemic deficit associated with vasospasm. The present study, which involves gene transfer to cerebral blood vessels, represents a potentially novel approach to treatment of vasospasm after SAH.
Proteins and peptides have been delivered into the brain parenchyma with a variety of vectors that are injected directly into the central nervous system.7 8 9 Transgenes introduced into the CSF or central nervous system by replication-deficient adenovirus also can be expressed in a wide variety of recipient cell types, including neurons, glia, and the ependymal cells lining the ventricles.10 11 12 13 14 Gene transfer to cerebral blood vessels is very difficult, however, because most approaches to vascular gene transfer require that blood flow is stopped.15
The vascular wall has been a target organ for cardiovascular gene therapy in previous studies.16 17 18 19 We have obtained evidence that genes can be transferred into the meninges and adventitia overlying normal cerebral vessels after injection of adenovirus into CSF.20 Gene transfer to intracranial vessels or vessels in spasm after SAH has not been accomplished previously.
In this study, we tested the hypothesis that during vasospasm induced by SAH, adenovirus can be used to transfer genes to cerebral blood vessels. A major concern was that the presence of subarachnoid blood might prevent access of adenovirus to vessels and thus prevent transgene expression. Angiography and studies of vascular reactivity were performed to confirm that vasospasm was induced in dogs.
Materials and Methods
All experiments were performed using procedures approved by the University of Iowa animal care and recombinant DNA committees. Adult mongrel dogs of both sexes (n=21, 15 to 25 kg) were used. Four groups of animals were studied. In group 1 (n=6), the double-hemorrhage model of canine vasospasm was used to demonstrate that vasospasm was produced in the experimental model.21 22 23 In group 2 (n=9), blood and adenovirus expressing a reporter gene were injected simultaneously into the cisterna magna. In group 3 (controls, n=2), control adenovirus containing an irrelevant transgene was injected into the cisterna magna. In group 4 (n=4), adenoviral vectors were injected 2 days after injection of blood to determine whether tissue could be transduced in the presence of subarachnoid blood.
Experimental SAH Technique (Group 1)
All procedures were performed after dogs were anesthetized with pentobarbital (25 mg/kg IV). The trachea was intubated, and ventilation was controlled with a respirator. End-tidal CO2 was maintained at 40±2 mm Hg, and anesthesia was maintained with room air and 1% halothane. Arterial blood gas levels were measured immediately before angiography was performed. Values (mean±SE) were pH 7.37±0.01, Paco2 39.8±0.7 mm Hg, and Pao2 78±2 mm Hg. There were no significant changes during the experiment.
A catheter was inserted through the femoral artery to the vertebral artery, and angiography of the basilar artery and circle of Willis was performed in the anterior-posterior plane. Under aseptic conditions, a 20-gauge spinal needle was inserted into the cisterna magna. The animal was maintained in a head-down position (superior plane of the parietal bone was tilted forward by 30°) for 30 minutes to facilitate contact of autologous blood with the base of the brain. Two days after the initial injection of blood, dogs were anesthetized again, and venous blood was again injected into the cisterna magna. On days 4 and 7 after intracisternal injection of blood, angiograms were obtained to determine the incidence and degree of chronic cerebral vasospasm. Diameters of the rostral, middle, and caudal thirds of the basilar arteries were measured (using a scaled loupe), and values from days 4 and 7 were compared with values from day 0.
Reactivity and Endothelial Function of Vessels In Vitro
Dogs after double blood injection (n=6) and dogs without SAH (n=3) were anesthetized with 30 mg/kg IV sodium pentobarbital, and brains were removed and placed in cold modified Krebs-Ringer bicarbonate solution with the following composition (mmol/L): 118.3 NaCl, 4.7 KCl, 2.5 CaCl2, 1.2 MgSO4, 25.0 NaHCO3, 0.026 calcium-ethylenediaminetetraacetic acid, and 11.1 glucose. Rings of branches of the basilar artery 2 to 3 mm in length were dissected and mounted on stainless-steel hooks. Tension was recorded with a linear force transducer (Gould Inc). Over 1 hour, the resting force was gradually increased until the optimal tension (1 g) for generating force during isometric contraction was reached. The vessels remained at this resting tension throughout the remainder of the study. Krebs-Ringer solution was circulated over the vessels, and temperature was maintained at 37±0.5°C. Experiments were performed in the presence of 10−5 mol/L indomethacin.
Responses to sodium nitroprusside, bradykinin, and arginine vasopressin were examined during contraction with UTP (10−5 to 10−6 mol/L). Responses to serotonin and UTP were also examined. Studies were performed to compare vessels from normal dogs and dogs that received intracisternal injection of blood.
The adenoviral vectors, based on adenovirus serotype 2, were constructed as described previously.24 These replication-deficient adenoviral vectors have been deleted of sequences in the E1A and E1B regions, which impairs replication in nonpermissive cells. The vector contains an expression cassette encoding the Escherichia coli β-galactosidase gene (lacZ) downstream from a nuclear-targeting signal. The immediate early promoter of CMV was used to drive transcription of lacZ with a simian virus 40 polyadenylation sequence cloned downstream from this reporter, forming the vector Ad/CMV-βgal. The control adenovirus used was Ad/CMV-CFTR, which carried CMV promoter and cDNA for CFTR. The viral titer was determined as infectious units per milliliter, using anti-adenovirus antibody in a titration assay. The ratio of particles (measured by spectrophotometer) to infectious units is approximately 100 in our experiments. The virus was suspended in 3% sucrose and stored at −80°C until used.
In Vivo Gene Transfer (Groups 2 Through 4)
Five mL of autologous venous blood was injected slowly (over 10 minutes) into the cisterna magna, immediately followed by injection of 1 mL of Ad/CMV-βgal (1.2 to 3.8×1010 infectious units) in group 2. Two dogs received 1 mL of control adenovirus (Ad/CMV-CFTR, 1.4×1010 infectious units, group 3). In group 4, 1 mL of Ad/CMV-βgal (2.6×1010 infectious units) was injected into the cisterna magna 2 days after injection of 5 mL of autologous venous blood. In all groups, a volume of CSF equal to the that of injectate was removed before injection of blood and viral suspension to avoid an increase in intracranial pressure. Rectal temperature was measured daily, and aspirin was given if temperature was >38.0°C. Animals that were injected with adenovirus were not studied with angiography. The animals were killed on days 2, 4, 6, or 7. Brains were removed, cut into serial sections, and stained for β-galactosidase.
X-Gal Histochemical Staining
Brain sections were placed in plastic vials and fixed for 10 minutes in fixative (PBS, pH 7.4, 2% paraformaldehyde, and 0.025% glutaraldehyde) at room temperature. The samples were rinsed well in PBS and then immersed in X-Gal stain [PBS: 20 mmol/L K4Fe(CN)6·3H2O, 20 mmol/L K3Fe(CN)6, 2 mmol/L MgCl2, and 1 mg/mL in DMSO of X-Gal stain; Gibco BRL Life Technologies, Inc] for 2 hours at room temperature. After a rinsing in PBS, the samples were fixed for approximately 16 hours in 4% formaldehyde. The samples were embedded in paraffin and thin-sectioned (6 to 8 μm) with a microtome, then counterstained with nuclear fast red. The sections were evaluated to determine whether β-galactosidase was expressed in blood vessels and brain tissue (blue staining). The basilar, anterior, middle, and posterior cerebral arteries; cerebral hemisphere (including lateral ventricles and subarachnoid space between gyri); mid pons; and medulla were cut into sections. The blood vessels (adventitia, muscle, endothelium) were examined under a microscope.
In group 1, baseline diameter of the basilar artery was compared with diameters measured at days 4 and 7 using the paired-samples t test with Bonferroni corrections for multiple comparisons. A value of P<.05 was considered significant. Reactivity of vessels in vitro was compared between normal dogs and dogs with SAH using an unpaired t test.
In groups 2 through 4, five tissue sections from each vessel or region of cerebrum were analyzed. β-Galactosidase staining was estimated with a three-point scale: 0, no stain; +, mild staining (approximately 1% to 25% of nuclei stained blue); ++, moderate staining (approximately 26% to 75% of nuclei stained blue); and +++, diffuse staining (>75% of nuclei stained blue).
Effects of Blood in the CSF (Group 1)
After injection of blood in the cisterna magna, approximately one half of the dogs appeared normal. Some dogs exhibited conjunctival hemorrhage or lethargy and ate little food. All dogs appeared to be pain-free throughout the experiment.
Basilar artery diameter decreased approximately 50% at days 4 and 7 after injection of blood, indicating profound vasospasm (Fig 1⇓, Table 1⇓). Vasospasm was greater by day 7 (compared with day 4) in all dogs.
Contraction to UTP (Fig 2⇓) and relaxation to bradykinin and sodium nitroprusside (Fig 3⇓) were similar in normal dogs and dogs with vasospasm. Arginine vasopressin relaxed the basilar artery of normal dogs, and the response was impaired after SAH (P<.05, Fig 3⇓). Serotonin had little effect on normal vessels and caused contraction in vessels exposed to SAH (Fig 2⇓).
In Vivo Gene Transfer (Group 2)
Systemic effects of CSF replacement with blood and Ad/CMV-βgal were similar to those in group 1 (which received intracisternal blood but not virus), except that temperature increased from 38.1±0.2°C at baseline to 40.0±0.3°C at day 1. The temperature always returned to normal after administration of 325 mg aspirin PO. We did not attempt to quantify the inflammatory response, but we did not detect differences in the inflammatory responses in the subarachnoid space in dogs that received intracisternal blood alone compared with dogs that received intracisternal blood plus recombinant adenovirus.
After 2 to 7 days, β-galactosidase was expressed in the basilar artery; small arteries on the brain stem; and the anterior, middle, and posterior cerebral arteries (Table 2⇓, Fig 4⇓). Transgene expression was also observed in ependymal lining of the lateral ventricle, floor of the fourth ventricle, pia-arachnoid over the brain stem, and in tissue within the subarachnoid space between the hemisphere gyri. On gross examination, there was diffuse staining of the basal surface of the brain, lining of the ventricles, and the cerebral arteries in all dogs. This diffuse staining represented gene transfer primarily within the pia-arachnoid. On microscopic examination, approximately two thirds of four sections from each dog of basilar, anterior, middle, and posterior cerebral artery had some expression of β-galactosidase in the adventitia. Virtually all sections of pons, medulla, ependymal lining of ventricle, and hemisphere sections had some expression of β-galactosidase.
In large cerebral vessels, there was blue staining in the nucleus of cells in adventitia and pia-arachnoid surrounding the vessels (Fig 4⇑). Occasionally, a cell within the muscularis layers expressed β-galactosidase, but generally there was minimal transduction of vascular muscle. No endothelial cells expressed transgene.
Injection of Control Virus (Group 3)
Cerebrum and vessels from control dogs that received control adenovirus (Ad/CMV-CFTR) were negative for β-galactosidase activity (Table 2⇑). Systemic effects of treatment were similar to those of group 2.
Injection of Virus After SAH (Group 4)
Dogs that received Ad/CMV-βgal 2 days after injection of blood were killed 7 days after the first injection of blood. Patterns of transgene expression were similar to those in group 2. Expression of β-galactosidase was observed in the basilar artery; small arteries on the brain stem; and the anterior, middle, and posterior cerebral arteries. Expression was also found in ependymal lining of the lateral ventricle and pia-arachnoid over the brain stem (Table 3⇓). Microscopically, there was transgene expression in basilar, anterior, middle, and posterior cerebral artery sections in the adventitial layer.
The major new finding in this study is that gene transfer to cerebral blood vessels and perivascular tissue can be accomplished after injection of adenovirus into the cisterna magna in the presence of subarachnoid blood. The findings suggest that an adenoviral vector can be used to transiently transfect cerebral blood vessels after SAH. The time frame of transgene expression from this viral construct may be appropriate for gene therapy, in which a short burst of expression is desirable. This study is the first to report gene transfer to cerebral blood vessels and perivascular tissues during spasm from SAH.
Consideration of the Model
The canine double-hemorrhage model of cerebral vasospasm was used in this study because spasm is produced consistently.21 22 23 Vascular reactivity was studied to ensure that vessels exposed to subarachnoid blood reacted similarly to vessels in other studies after SAH. Other studies have reported similar effects of SAH on responses to vasopressin, sodium nitroprusside, bradykinin, and serotonin,21 25 although responses to sodium nitroprusside and bradykinin may be impaired in a different model.22 26 We therefore confirmed by angiography that vasospasm was induced using the canine double-hemorrhage model, demonstrated that vascular responses were similar to those in previous studies, and used the same model for gene transfer to vessels in spasm.
The major finding in this study was that the presence of subarachnoid blood does not appear to interfere with adenovirus-mediated gene transfer to vessels in spasm. In most animals, there were thrombi in close proximity to vessels, and these thrombi did not seem to interfere with gene transfer.
This model has several limitations. First, the target of gene transfer is not cerebral tissue, but gene transfer nevertheless occurred to the surface of brain, as well as to blood vessels. Therefore, gene transfer is not specific for cerebral blood vessels, and all cells exposed to the subarachnoid space become potential targets of the adenovirus. Second, transfer of the reporter gene to endothelium was not observed. However, this may not detract from the ability to transfer therapeutic genes to vessels. X-Gal staining likely underestimates (up to 1/10) the number of cells that have been transduced.24 Also, transduction of adventitial cells and perivascular cells may be sufficient to alter vessel function. Periadventitial delivery of drugs and bioactive substances, including growth hormone and antisense oligonucleotide, can alter vascular function in vivo.27 28 29 Therefore, it is likely that transduction of adventitial cells and perivascular cells that encode enzymes, which produce highly diffusible substances, will affect functions of underlying vascular smooth muscles. Thus, perivascular tissue may be a useful target when genes that express diffusible substances are transferred.
Gene Transfer to Vessels
Several investigators have accomplished adenovirus-mediated gene transfer to endothelium and smooth muscle cells in both intact and balloon-injured arteries by an intravascular approach in vivo.18 30 31 However, efficient gene transfer seems to require increased intraluminal pressure that may cause damage to the vessels.32 Our study differed from those studies in that we approached the vessel from the adventitia, rather than from the luminal side. This approach also avoids the need to stop flow in the vessel and allows longer contact of the virus with the vessel wall.
Direct In Vivo Gene Transfer to Central Nervous System
The central nervous system is relatively inaccessible to circulating proteins and peptides because of the presence of a blood-brain barrier. Several years ago, an alternative method to deliver therapeutic protein/peptides directly into the central nervous system was developed. This approach involves the use of several vectors for gene transfer to produce proteins in the CSF and brain.7 8 9 33
Replication-deficient adenoviruses, which are promising vectors for gene transfer to the CNS, can deliver genes to many cell types. Injection of adenovirus carrying β-galactosidase into brain parenchyma results in extensive expression of the reporter gene in neurons or glia in the injection site and in ependymal cells of the ventricles.10 12 13 34 35 36 The transgene was also expressed in ependyma after intraventricular injection, but effects on cerebral vessels were not described.11 14 34 Recently, we injected adenoviral vectors in the cisterna magna or lateral ventricle of rats and observed that genes can be transferred to the adventitia and perivascular tissues overlying cerebral blood vessels.20 However, it was not clear whether subarachnoid blood would interfere with access of viral vectors to the vessels and expression of transgene.
The present study indicates that genes can be transferred to perivascular tissue and adventitia of cerebral blood vessels during blood-induced vasospasm with an adenoviral vector injected into the cisterna magna. Transfer of genes with therapeutic potential, such as genes that encode enzymes that produce vasodilatation, may be useful. The vasoactive substance may need to be highly diffusible because gene transduction does not occur consistently in vascular muscle. Substances such as nitric oxide synthase, which produces nitric oxide (which is highly diffusible), or calcitonin gene–related peptide may exert vasomotor effects when they are released from perivascular tissue or adventitia. The similar time frame for onset of vasospasm after SAH (3 to 10 days) and onset of protein synthesis after gene transfer makes gene transfer to the vessel wall an attractive alternative in the search for effective prevention or treatment of cerebral vasospasm.
Selected Abbreviations and Acronyms
|CFTR||=||cystic fibrosis transmembrane conductance regulator|
This research was supported by National Institutes of Health grants NS-24621, HL-16066, HL-14388, AG-10269, and HD-33531 and research funds from the Veterans Administration. B.L. Davidson is a fellow of the Roy J. Carver Charitable Trust, and M.J. Welsh is an investigator of the Howard Hughes Medical Institute. The authors thank Richard D. Anderson and Lisa DeBerg for preparation of the adenovirus, William Barnhart for assistance with angiography, Pamela Tompkins for assistance with X-Gal staining and thin sectioning, Donna G. Martin and Kristen Rummelhart for help with studies of vessel reactivity, Dr James Torner for assistance with statistical analysis, Dr Frank Faraci for review of this manuscript, and Dr Alan E. Smith, Genzyme, for the gift of Ad/CMV-βgal and Ad/CMV-CFTR.
- Received September 9, 1996.
- Revision received December 27, 1996.
- Accepted January 16, 1997.
- Copyright © 1997 by American Heart Association
Kiwak KJ, Heros RC. Cerebral vasospasm after subarachnoid hemorrhage. Trend Neurosci. 1987;10:89-92.
MacDonald RL, Weir BKA. A review of hemoglobin and the pathogenesis of cerebral vasospasm. Stroke. 1991;22:971-982.
Rosenberg MB, Friedmann T, Robertson RC, Tuszynski M, Wolff JA, Breakefield XO, Gage FH. Grafting genetically modified cells to the damaged brain: restorative effects of NGF expression. Science. 1988;242:1575-1578.
Culver KW, Ram Z, Wallbridge S, Ishii H, Oldfield EH, Blaese RM. In vivo gene transfer with retroviral vector-produced cells for treatment of experimental brain tumors. Science. 1992;256:1550-1552.
Lal B, Cahan MA, Couraud P-O, Goldstein GW. Development of endogenous β-galactosidase and aurofluorescence in rat brain microvessels: implication for cell tracking and gene transfer studies. J Histochem Cytochem. 1994;42:953-956.
Heistad DD, Faraci FM. Gene therapy for cerebral vascular disease. Stroke. 1996;27:1688-1693.
Yao SN, Wilson JM, Nabel EG, Kurachi S, Hachiya HL, Kurachi K. Expression of human factor IX in rat capillary endothelial cells: toward somatic gene therapy for hemophilia B. Proc Natl Acad Sci U S A. 1991;88:8101-8105.
Dichek D, Nerville RF, Zwiebel JA, Freeman SM, Leon MB, Anderson WF. Seeding of intravascular stents with genetically engineered endothelial cells. Circulation. 1989;80:1347-1353.
Lemarchand P, Jones M, Yamada I, Crystal RG. In vivo gene transfer and expression in normal uninjured blood vessels using replication-deficient recombinant adenovirus vectors. Circ Res. 1993;72:1132-1138.
Morishita R, Gibbons GH, Kaneda Y, Ogihara T, Dzau VJ. Novel in vitro gene transfer method for study of local modulators in vascular smooth muscle cells. Hypertension. 1993;21:894-899.
Ooboshi H, Welsh MJ, Rios CD, Davidson BL, Heistad DD. Adenovirus-mediated gene transfer in vivo to cerebral blood vessels and perivascular tissue. Circ Res. 1995;77:7-13.
Katusic ZS, Milde JH, Cosentino F, Mitrovic BS. Subarachnoid hemorrhage and endothelial l-arginine pathway in small brain stem arteries in dogs. Stroke. 1993;24:392-399.
Hatake K, Wakabayashi I, Kakishita E, Hishida S. Impairment of endothelium-dependent relaxation in human basilar artery after subarachnoid hemorrhage. Stroke. 1992;23:1111-1117.
Sobey CG, Heistad DD, Faraci FM. Effect of subarachnoid hemorrhage on dilatation of rat basilar artery in vivo. Am J Physiol. 1996;40:H126-H132.
Edelman ER, Nugent MA, Karnovsky MJ. Perivascular and intravenous administration of basic fibroblast growth factor: vascular and solid organ deposition. Proc Natl Acad Sci U S A. 1993;90:1513-1517.
Villa AE, Guzman LA, Chen W, Golomb G, Levy RJ, Topol EJ. Local delivery of dexamethasone for prevention of neointimal proliferation in a rat model of balloon angioplasty. J Clin Invest. 1994;93:1243-1249.
Lee SW, Trapnell BC, Rade JJ, Virmani R, Dichek DA. In vitro adenoviral vector–mediated gene transfer into balloon-injured rat carotid arteries. Circ Res. 1993;73:797-807.
Schulick AH, Dong G, Newman KD, Virmani R, Dichek DA. Endothelium-specific in vivo gene transfer. Circ Res. 1995;77:475-485.
Rome JJ, Shayani V, Flugelman MY, Newman KD, Farb A, Virmani R, Dichek D. Anatomic barriers influence the distribution of in vivo gene transfer into the arterial wall: modeling with microscopic tracer particles and verification with a recombinant adenoviral vector. Arterioscler Thromb. 1994;14:148-161.