INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

CALCITONIN GENE-RELATED PEPTIDE (CGRP), initially discovered as an alternative RNA splicing product from the calcitonin gene (2), is abundant in perivascular nerve fibers surrounding cerebral and some peripheral blood vessels (26, 36, 42). Release of the neuropeptide from trigeminal sensory fibers (25) is an extremely potent vasodilator mechanism (4), especially in the cerebral circulation (44). CGRP may have functional effects in the cerebral circulation under some disease states, including subarachnoid hemorrhage (SAH). Several studies have demonstrated depletion of CGRP in perivascular nerve terminals after SAH (11, 28), which may correlate with the onset of vasospasm.

Transduction of genes with dilator potency to blood vessels is a possible strategy for the prevention and treatment of cardiovascular and cerebrovascular disorders. For example, Ooboshi et al. (30, 32) and others (6, 7) demonstrated that transduction of recombinant endothelial nitric oxide synthase (eNOS) using adenoviral vectors ex vivo and in vivo alters reactivity of carotid and cerebral arteries. Thus we speculated that overexpression of CGRP in blood vessels by adenovirus-mediated gene transfer may also alter vascular tone.

The goal of this study was to determine whether gene transfer of CGRP produces a vasoactive product that has functional effects in cerebral arteries. We chose CGRP because of its potency as a vasodilator and because it is readily diffusible. For this strategy, we constructed a recombinant adenovirus containing the cDNA for human prepro-CGRP. We chose to prepare a virus that produces prepro-CGRP, rather than CGRP, because we anticipated that the precursor would facilitate the required modification and release of the peptide from the cell.

First, we determined whether biologically active CGRP is processed and released in cell cultures following gene transfer with this virus. We examined release of CGRP and accumulation of intracellular cAMP after gene transfer in two mammalian cell lines. Although there is not a consensus regarding receptor-effector coupling mechanisms for the vasodilator response to CGRP, cAMP increases when CGRP receptors are stimulated (22, 39), and there is a correlation between CGRP-induced vasorelaxation and elevation of intracellular cAMP (13). Thus cAMP appears to be an important second messenger of the coupling mechanism, in part by cAMP-mediated activation of ATP-sensitive potassium channels (21). Second, we determined whether perivascular gene transfer using the virus in vivo expresses CGRP in the adventitia and alters cerebral vascular reactivity. Perivascular delivery of transgenes produces overexpression of transgene products in vascular adventitia and perivascular tissues (7, 9, 27, 33, 34). This mode of delivery does not require interruption of blood flow, which is potentially useful for cerebral circulation. Adventitial expression of transduced eNOS can augment NO-mediated vasorelaxation (7). Thus we anticipated that expression of CGRP in adventitia of cerebral blood vessels might augment vasorelaxation.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Adenoviral vectors. A replication-deficient adenovirus encoding the human prepro-CGRP gene (Adprepro-CGRP) expressed from a long terminal repeat of Rous sarcoma virus (RSV-LTR) as a promotor was generated using standard methods by the University of Iowa Gene Transfer Vector Core (Iowa City, IA) (30). Briefly, the entire coding sequence of human prepro-CGRP was cloned by blunt end ligation into pAdRSV4. The resultant plasmid and adenovirus backbone sequences restricted of E1 were transfected into human embryonic kidney (HEK) 293 cells, and plaques were isolated and amplified for analysis of CGRP expression. A recombinant adenovirus encoding nuclear-targeted bacterial -galactosidase expressed from the RSV-LTR promotor (Adgal) was used as a control virus. Recombinant adenoviruses were triple plaque purified, and virus titer was determined by plaque assay on HEK-293 cells. Purified viruses were stored in phosphate-buffered saline (PBS) with 3% sucrose and kept at 80°C until use.

Gene transfer to cultured cells. Cos-7 cells [African green monkey kidney cells transduced with simian virus 40 (SV40) T-antigen] and NIH-3T3 cells (mouse fibroblast cells) were cultured on 35-mm dishes in Dulbecco's modified Eagle's medium (DMEM, high glucose) supplemented with 10% fetal bovine serum, 100 U/ml penicillin, and 100 µg/ml streptomycin. Cells were seeded at 3 × 10⁴ cells/cm² and infected 18-24 h after seeding when cells were 70% confluent. The medium was replaced with 1 ml of DMEM containing Adgal [500 plaque-forming units (pfu)/cell], Adprepro-CGRP (10, 100, and 500 pfu/cell), or vehicle (20 µl of PBS with 3% sucrose). For 3T3 cells, a noncovalent complex of Adgal or Adprepro-CGRP with poly-L-lysine at a 1:100 ratio of adenovirus particles to poly-L-lysine molecules was used to enhance the efficiency of gene transfer (15). The complex was formed as previously described (15, 40). Cells were incubated for 2 h in a 5% CO₂ humidified atmosphere at 37°C, the infection solution was removed, fresh serum-containing DMEM was added, and the cells were incubated for an additional 72 h. The culture medium was replaced every 24 h. The infection solution and each 24-h culture medium were stored at 80°C for measurement of CGRP-like immunoreactivity.

Gene transfer to rabbits in vivo. All animal procedures were approved by the Animal Care and Use Review Committee at the University of Iowa. Male New Zealand White rabbits weighing 2.4-3.0 kg were anesthetized with 5 mg/kg im of xylazine and 50 mg/kg im of ketamine. A 25-gauge needle was aseptically inserted into the cisterna magna as previously described (23), 250 µl of cerebrospinal fluid (CSF) were withdrawn for measurement of white blood cell count and CGRP-like immunoreactivity, and 250 µl of viral suspension (PBS with 3% sucrose with Adgal or Adprepro-CGRP) or vehicle (PBS with 3% sucrose) were infused. The adenoviral titer in suspension varied from 1 × 10⁹ to 3 × 10¹⁰ pfu/ml.

Five days after injection, rabbits were anesthetized with 50 mg/kg iv of pentobarbital sodium and exsanguinated from the common carotid artery after collection of CSF from the cisterna magna. Basilar arteries were removed and cut into four rings, each 3 mm in length, for assays of -galactosidase activity and cAMP, immunohistochemistry for CGRP, and recording of isometric tension. The brain was also removed for immunohistochemistry for CGRP. Rectal temperature was measured just before injection and before exsanguination. Some rabbits were perfused transcardially with 2% paraformaldehyde and 0.2% glutaraldehyde in PBS without exsanguination, and the whole brain was removed for histochemistry for -galactosidase.

In separate experiments, basilar arterial rings were prepared from additional male New Zealand White rabbits without any intervention to determine effects of synthetic CGRP on intracellular cAMP levels and vascular reactivity.

Chemiluminescent assay for -galactosidase. To determine the optimal adenoviral titer for in vivo gene transfer, -galactosidase activity was assayed from arterial rings of rabbits treated with viral suspension of four different titers (1 × 10⁹, 3 × 10⁹, 1 × 10¹⁰, and 3 × 10¹⁰ pfu/ml) of Adgal, as we described previously (40).

Measurement of CGRP-like immunoreactivity. CGRP-like immunoreactivity in media of cell cultures and CSF was determined to examine release of CGRP following gene transfer in vitro and in vivo. Media of cell cultures and CSF (180 µl) were acidified with 20 µl of 1 N HCl, heated at 95°C for 10 min, and centrifuged at 4°C for 4 min (41). The supernatants were neutralized with 1 N NaOH and immediately assayed for CGRP-like immunoreactivity using a commercially available radioimmunoassay (RIA) kit (Peninsula Laboratories, Belmont, CA). Results are expressed using molarity of mature CGRP.

Measurement of cAMP. Generation of intracellular cAMP in cell culture was determined to ascertain bioactivity of transduced CGRP. Because CGRP receptors are expressed endogenously in baby hamster kidney (BHK)-21 cells (5), this cell line was used as recipient cells. BHK-21 cells on 35-mm wells with 70% confluency were incubated for 15 min in 1 ml of DMEM maintained at 37°C with 1 mmol/l 3-isobutyl-1-methylxanthine (IBMX), a nonselective inhibitor of cyclic nucleotide phosphodiesterase. The medium was then replaced by the media from Cos-7 cells that had been transfected with Adprepro-CGRP (500 pfu/cell), Adgal, or vehicle for 15 min at 37°C. For some BHK-21 cells, 1 µmol/l CGRP-(8-37), an antagonist for CGRP₁ receptors, was added in DMEM for 10 min before replacement of medium. We then removed the culture media, added ice-cold 70% ethanol, agitated wells gently for 10 min, collected the supernatant, and centrifuged it at 2,000 g for 15 min. The supernatant was evaporated to dryness under nitrogen, resuspended in 50 mmol/l sodium acetate buffer (pH 5.8), acetylated, and assayed for cAMP levels by the Biotrak cAMP ¹²⁵I RIA system (Amersham Life Science, Arlington Heights, IL).

Intracellular cAMP in basilar arteries was also determined. Two arterial rings from each rabbit were incubated in 1 ml of Krebs bicarbonate solution (in mmol/l: 118 NaCl, 4.7 KCl, 1.2 KH₂PO₄, 1.2 MgSO₄ · 7H₂O, 11.1 D-glucose, 25.0 NaHCO₃, and 2.54 CaCl₂ · H₂O) maintained at 37°C, and bubbled with 95% O₂-5% CO₂, with 1 mmol/l IBMX for 20 min immediately after removal from the brain. To one of these two rings, 1 µmol/l CGRP-(8-37) was added. Arterial rings were then quickly frozen in liquid nitrogen and stored at 80°C. The rings were thawed, homogenized in ice-cold 10% trichloroacetic acid, and centrifuged at 2,000 g for 15 min to pellet the residual tissue fragments. The supernatant was washed with diethyl ether to remove trichloroacetic acid, evaporated to dryness under nitrogen, and assayed for cAMP levels by the above-mentioned methods. The tissue residue was dissolved in 2 mol/l sodium hydroxide and the protein content determined by the Bio-Rad DC Protein Assay.

In separate experiments, basilar arterial rings from rabbits without any intervention were also incubated in Krebs solution with 1 mmol/l IBMX for 30 min, and then further incubated in Krebs solution with 1 mmol/l IBMX and 0, 1, or 100 nmol/l synthetic CGRP for 15 min. For some rings, 1 µmol/l CGRP-(8-37) was further added in Krebs solution for 15 min before exposure to CGRP (100 nmol/l). The rings were frozen and later assayed for cAMP levels.

Histochemistry. Histochemical staining for detection of -galactosidase was performed as described previously (33) to examine localization of transgene product after gene transfer in vivo. The number of blue-stained nuclei and total nuclei of adventitial cells were counted in consecutive cross sections of the basilar artery as described previously (31).

Measurement of vascular reactivity. Arterial rings were suspended in organ baths filled with Krebs bicarbonate solution maintained at 37°C and bubbled with 95% O₂-5% CO₂ and connected to force transducers. Resting tension of vessels was increased to 500 mg, and the preparations were allowed to equilibrate for 60 min. Concentration-response curves for three different vasocontractive agents, KCl (40 mmol/l), histamine (0.01-3 µmol/l), and serotonin (0.01-3 µmol/l), were then generated. Responses to histamine and serotonin were expressed as percent contraction of response to 40 mmol/l KCl. Concentration-response curves for three different vasodilator agents, IBMX (0.1-3 µmol/l), acetylcholine (0.01-10 µmol/l), and sodium nitroprusside (0.01-10 µmol/l), were also generated following vasocontraction with 0.3-1 µmol/l of histamine (corresponding to 50% of maximal contraction produced by 3 µmol/l histamine). Relaxation was expressed as percent relaxation of contraction produced by histamine. Responses to histamine, serotonin, and IBMX were also examined in the presence of 0.5 µmol/l CGRP-(8-37), which was applied 5 min before contraction or precontraction.

In separate experiments, rings of the basilar artery from rabbits without any intervention were suspended in organ baths, precontracted with 0.3-1 µmol/l of histamine, and assayed for concentration-response curves for synthetic CGRP. In other experiments, the basilar artery was exposed to 10 or 100 nmol/l of synthetic CGRP for 5 min, and then concentration-response curves to histamine were examined.

Statistical analysis. Values are expressed as means ± SE. One-way factorial ANOVA followed by Bonferroni's test was used for comparison of variables among groups. Two-way repeated-measures ANOVA followed by Bonferroni's test was used for comparison of concentration-response curves among groups. P < 0.05 was accepted as statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Gene transfer in vitro. To test whether adenovirus-mediated gene transfer of prepro-CGRP resulted in transgene expression in mammalian cells, levels of CGRP-like immunoreactivity were determined in cell culture media following infection. Both Cos-7 and NIH-3T3 cell lines expressed minimal levels of CGRP-like immunoreactivity after treatment with vehicle or Ad-gal (Fig. 1, A and B). Treatment with Adprepro-CGRP increased levels of CGRP-like immunoreactivity in both Cos-7 and 3T3 cells in a dose-dependent manner (P < 0.0001 vs. vehicle or Adgal), which indicates effective expression of CGRP following gene transfer.

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Expression of calcitonin gene-related peptide (CGRP)-like immunoreactivity and cAMP following gene transfer in vitro. A and B: concentration of CGRP-like immunoreactivity in the cell culture medium of Cos-7 (A) and NIH-3T3 (B) cells that were transduced with adenovirus-encoded -galactosidase (Adgal) [500 plaque-forming units (pfu)/cell] or adenoviral prepro-CGRP (Adprepro-CGRP: 10, 100, and 500 pfu/cell). Concentrations in vehicle- and Adgal-treated groups are below sensitivity of measurement. Concentration differs significantly in 500 pfu/cell of Adprepro-CGRP vs. other groups (P < 0.0001) in both cell lines. C: cAMP levels in baby hamster kidney (BHK)-21 cells that were incubated with (CGRP) or without (control) 1 nmol/l synthetic CGRP or with cell culture media of Cos-7 cells treated with vehicle, Adgal, or Adprepro-CGRP. Some cells (hatched bars) were pretreated with 1 µmol/l CGRP-(8-37). * P < 0.0005 vs. control,  P < 0.0005 vs. vehicle, § P < 0.0005 vs. Adgal. Values are means ± SE (n = 6 cell cultures for A, 4 for B, 6 for C).

After adenoviral infection, we transferred culture medium from Cos-7 cells to BHK-21 cells and measured cAMP levels in the recipient cells. Application of medium from Cos-7 cells treated with vehicle and Adgal produced similar low levels of cAMP generation in BHK-21 cells (Fig. 1C). Medium, which contained 12.6 ± 2.4 nmol/l CGRP-like immunoreactiviy, from Cos-7 cells treated with Adprepro-CGRP increased cAMP levels in BHK-21 cells by threefold (P < 0.0005). This increase in cAMP in recipient cells was abolished by pretreatment of the cells with CGRP-(8-37). The findings indicate that medium from Adprepro-CGRP-transduced cells is biologically active in cell culture.

Effects of synthetic CGRP on rabbit basilar artery. Synthetic CGRP produced an increase in cAMP levels in rabbit basilar arteries and relaxation of the arteries (Fig. 2, A and B). For example, 1 nmol/l CGRP increased cAMP levels by 2.3-fold and relaxed the arteries by 49 ± 10% relative to the baseline value. Thus both cAMP levels and vascular reactivity are indicators of bioactivity of CGRP. Pretreatment of the basilar artery with 10 or 100 nmol/l of synthetic CGRP also attenuated vasocontraction to 0.03-0.3 µmol/l of histamine (Fig. 2C).

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Effects of synthetic CGRP on normal basilar artery. A: cAMP levels in artery. * P < 0.05 vs. vehicle,  P < 0.001 vs. vehicle, § P < 0.01 vs. +CGRP-(8-37). B: relaxation of histamine-precontracted arteries to synthetic CGRP. Tension generated by precontraction was 1.05 ± 0.06 g. C: contraction of synthetic CGRP (10 or 100 nmol/l)-pretreated arteries to histamine. * P < 0.01,  P < 0.001 vs. arteries without pretreatment [CGRP()]. Values are means ± SE (n = 6 rabbits for A, 5 for B, 6 for C).

Gene transfer in vivo. We injected Adprepro-CGRP into CSF of rabbits to examine expression of CGRP in basilar arteries and effects on vascular reactivity. All rabbits tolerated the injection of adenoviral vectors without alteration of rectal temperature (39.0 ± 0.1°C before injection and 39.1 ± 0.1°C 5 days after injection) or detectable neurological deficits. Leukocyte count in CSF increased from 3 ± 1 to 249 ± 100 cells/mm³ 5 days after Adgal injection and 193 ± 68 cells/mm³ after Adprepro-CGRP. Leukocyte count did not increase 5 days after injection of vehicle (5 ± 2 cells/mm³).

-Galactosidase activity in basilar arterial rings from rabbits treated with 1 × 10⁹, 3 × 10⁹, 1 × 10¹⁰, and 3 × 10¹⁰ pfu/ml of Adgal was 14 ± 4, 523 ± 56, 413 ± 149, and 3 ± 2 mU/mg protein, respectively (n = 3 for all viral titers), suggesting that injection of 3 × 10⁹ pfu/ml is optimal for adenoviral vectors. Thus we performed the following assays using 250 µl of viral suspension with 3 × 10⁹ pfu/ml of Adgal or AdCGRP for gene transfer in vivo. Because the total volume of freely flowing CSF in the rabbit is >2 ml (23), the final adenoviral titer in CSF was estimated to be <4 × 10⁸ pfu/ml.

Overexpression of CGRP after gene transfer in vivo. Five days after injection of Adprepro-CGRP in vivo, histochemistry for -galactosidase revealed blue staining extensively on the ventral surface of the brain and in ventricles in rabbits treated with Adgal as we previously demonstrated in other species (9, 27, 33). The percentage of blue-stained cells in the adventitia of the basilar artery was 34 ± 3% (n = 4 rabbits).

Release of CGRP into CSF was evaluated by RIA. Levels of CGRP-like immunoreactivity in CSF were similar before and 5 days after injection of Adgal or vehicle (Fig. 3A). Five days after injection of Adprepro-CGRP, however, CGRP-like immunoreactivity was 93-fold greater than baseline values (P < 0.0005).

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Effects of gene transfer in vivo on expression of CGRP and cAMP concentration. Samples were collected 5 days after injection of vehicle, Adgal, or Adprepro-CGRP. A: concentration of CGRP-like immunoreactivity in cerebrospinal fluid before (day 0) and 5 days after injection of vehicle or virus. Concentration differs significantly in Adprepro-CGRP vs. vehicle (P < 0.0005) or Adgal (P < 0.0001). B: cAMP levels in basilar arteries with or without pretreatment by 1 µmol/l CGRP-(8-37). Levels differ significantly in Adprepro-CGRP vs. vehicle (* P < 0.005) or Adgal ( P < 0.02). Values are means ± SE (n = 8 rabbits for A and 6 for B).

Intracellular cAMP was measured in basilar arteries to verify bioactivity of transduced CGRP following gene transfer in vivo. Five days after injection of AdCGRP, but not Adgal, cAMP in basilar arteries was greater than after injection of vehicle (P < 0.01, Fig. 3B). Increased cAMP was restored to baseline levels by pretreatment of the harvested basilar artery with CGRP-(8-37) before assay. Thus gene transfer of prepro-CGRP resulted in overexpression of CGRP in adventitia of basilar arteries, on the ventral brain stem and ventricles, and in CSF, and expressed CGRP was bioactive.

Effects of gene transfer on vascular reactivity. Contraction to 40 mmol/l KCl was similar in arteries treated with vehicle (1.55 ± 0.07 g), Adgal (1.56 ± 0.18 g), and Adprepro-CGRP (1.43 ± 0.12 g). Contractile responses to histamine (0.01-3 µmol/l) were not significantly different in arteries treated with vehicle or Adgal (Fig. 4A). In arteries treated with Adprepro-CGRP, however, contractile responses to histamine were attenuated (P < 0.02 vs. vehicle-treated arteries, and P < 0.01 vs. Adgal-treated arteries). EC₅₀ [log(mol/l)] (6.16 ± 0.11) and maximal contraction (1.40 ± 0.14 g, 98 ± 10% of response to 40 mmol/l KCl) for Adprepro-CGRP were significantly different from those for vehicle [EC₅₀, 6.46 ± 0.07, P < 0.05; maximal response (R_max) 1.91 ± 0.08 g, 123 ± 5%, P < 0.01] or Adgal (EC₅₀, 6.45 ± 0.07, P < 0.05; R_max 1.92 ± 0.12 g, 124 ± 8%, P < 0.005).

View larger version (34K): [in this window] [in a new window]   Fig. 4.   Effects of gene transfer in vivo on vasocontraction in vitro. Responses of rabbit basilar arteries to (A, B) histamine and (C, D) serotonin, with (B, D) or without (A, C) pretreatment by 0.5 µmol/l CGRP-(8-37). Values are means ± SE (n = 9 rabbits for A, 6 for B, 7 for C, 5 for D). Responses to histamine (A) differ significantly in Adprepro-CGRP vs. vehicle (P < 0.02) and Adgal (P < 0.01). Responses to serotonin (C) differ significantly in Adprepro-CGRP vs. vehicle (P < 0.01) and Adgal (P < 0.02).

Contractile responses to serotonin (0.01-3 µmol/l) were not significantly different in arteries treated with vehicle or Adgal (Fig. 4C). In arteries treated with Adprepro-CGRP, contractile responses to serotonin were attenuated (P < 0.01 vs. vehicle-treated arteries, and P < 0.02 vs. Adgal-treated arteries). EC₅₀ (6.91 ± 0.09) and maximal contraction (0.79 ± 0.12 g, 55 ± 9% of response to 40 mmol/l KCl) for Adprepro-CGRP were significantly different from those for vehicle (EC₅₀, 7.23 ± 0.15, P < 0.05; R_max 1.24 ± 0.03 g, 80 ± 2%, P < 0.01) or Adgal (EC₅₀, 7.40 ± 0.12, P < 0.05; R_max 1.18 ± 0.11 g, 76 ± 8%, P < 0.05).

In arteries treated with Adprepro-CGRP, responses to histamine and serotonin were restored to those of Adgal-treated arteries by pretreatment with CGRP-(8-37) (Fig. 4, B and D). Concentration-contraction curves to histamine and serotonin were similar when contraction was expressed using absolute values for tension. These data indicate that after gene transfer of CGRP in vivo, contractile responses of the basilar artery are attenuated by an effect on CGRP₁ receptors.

Relaxation to IBMX (0.1-3 µmol/l) was significantly augmented in arteries treated with Adprepro-CGRP compared with vehicle (P < 0.0005) or Adgal (P < 0.0005) (Fig. 5A). EC₅₀ for Adprepro-CGRP (6.28 ± 0.05) was significantly different from that for vehicle (6.05 ± 0.03, P < 0.005) or Adgal (6.08 ± 0.06, P < 0.02). Maximal relaxation for Adprepro-CGRP (94 ± 4%) tended to be greater than that for vehicle (83 ± 4%, P < 0.1) or Adgal (79 ± 8%, P < 0.1). Augmented relaxation to IBMX was blocked by CGRP-(8-37) (Fig. 5B). Responses to acetylcholine and nitroprusside were virtually identical among the three groups (Fig. 5, C and D). Concentration-relaxation curves were similar when precontraction and relaxation were expressed using absolute values for tension. Thus gene transfer of prepro-CGRP enhanced relaxation of basilar arteries to an inhibitor of phosphodiesterase but not to other vasorelaxant stimuli.

View larger version (32K): [in this window] [in a new window]   Fig. 5.   Effects of gene transfer in vivo on vasorelaxation in vitro. Responses of histamine-precontracted rabbit basilar arteries to IBMX (A, B), ACh (C), and sodium nitroprusside (SNP) (D). Arteries were pretreated with 0.5 µmol/l CGRP-(8-37) in B. Values are means ± SE (n = 7 rabbits for A, C, D and 5 for B). Responses to IBMX (A) differ significantly in Adprepro-CGRP vs. vehicle (P < 0.0005) and Adgal (P < 0.0005). Tension of precontraction for A was 0.94 ± 0.11 g, and tension for other agents was similar.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study, we have described successful overexpression of vasoactive CGRP after gene transfer of prepro-CGRP. There are three major new findings. First, treatment with Adprepro-CGRP greatly increased levels of CGRP-like immunoreactivity in mammalian cell lines and enhanced levels of cAMP. Thus the recombinant adenoviral vector produces biologically active CGRP, which is released from the cell. Second, gene transfer with Adprepro-CGRP in vivo increased CGRP-like immunoreactivity in the CSF and increased cAMP in the basilar artery. Thus biologically active CGRP can be expressed in vivo by adenovirus-mediated gene transfer. Third, gene transfer of prepro-CGRP in vivo altered reactivity of the basilar artery in vitro with inhibition of contraction and selective augmentation of relaxation.

Gene transfer in vitro. CGRP is biosynthesized as a precursor propeptide that is processed posttranslationally before secretion. The processing of prepro-CGRP begins in the rough endoplasmic reticulum with the removal of the signal peptide, then continues in the secretory granules with cleavage at basic residues and COOH-terminal amidation. We prepared an adenovirus that expresses a precursor propeptide, which had a potential disadvantage that the propeptide may not be processed correctly.

We demonstrated an enormous increase in CGRP-like immunoreactivity in cell culture media after gene transfer. This RIA has no cross-reaction to calcitonin or other peptides but may cross-react with pro-CGRP. We measured cAMP to assess bioactivity of the viral product, because cAMP is coupled to CGRP₁ receptors (22, 39) and is a second messenger in CGRP-induced physiological actions (12, 43). Enhanced accumulation of cAMP in the recipient cells after gene transfer suggested effective processing of the propeptide to bioactive CGRP. Thus an important finding of this study is that adenovirus-mediated gene transfer of a precursor propeptide can overexpress a bioactive product.

CGRP-like immunoreactivity in the medium of Cos-7 cells treated with Adprepro-CGRP was >10 nmol/l. Medium from cells treated with Adprepro-CGRP produced a significant increase in cAMP, which indicates that the peptide was biologically active, but cAMP levels in the recipient cells were smaller after application of this medium than after application of 1 nmol/l synthetic CGRP (Fig. 1C). Thus it is possible that biologically inactive CGRP (perhaps unprocessed pro-CGRP) was also detected by radioimmunoassay in the medium of Cos-7 cells. We speculated that because Cos-7 cells have relatively low activity of the amidation enzyme peptidyl-glycine -amidating monooxygenase, which is required for CGRP bioactivity (38), amidation may be less in Cos-7 cells than in other tissues. These findings nevertheless provided evidence in tissue culture that the recombinant adenovirus produces an active product and led us to examine effects of the virus in vivo.

Gene transfer in vivo. Injection of an adenoviral vector into CSF, which results in overexpression of transgene products (-galactosidase and eNOS) in the adventitia of cerebral arteries and perivascular tissue, is a useful approach for gene transfer in vivo (7, 9, 27, 33). From histochemical detection of -galactosidase, we estimate that about one-third of the adventitial cells may be infected after gene transfer of CGRP. Adventitia consists primarily of fibroblasts, and findings in fibroblast cell culture suggest that adventitial fibroblast cells are efficiently transduced by Adprepro-CGRP. Measurement of cAMP and vascular reactivity demonstrated that vasoactive CGRP was overexpressed in the artery. Interestingly, a large amount of CGRP is released in CSF after gene transfer. Concordant with the finding of extensive histochemical staining for -galactosidase, CGRP appears to be overexpressed extensively on the ventral regions of the brain stem surface as well as the adventitia of the arteries after gene transfer of prepro-CGRP. The transgene product on the brain surface may secrete CGRP into CSF and potentially alter vascular function.

Increases in cAMP in the basilar artery after Adprepro-CGRP demonstrated bioactivity of the transduced CGRP, as cAMP increased greater than twofold after gene transfer (Fig. 3B). The increase in cAMP in the basilar artery after gene transfer, which was associated with a concentration of 0.6 ± 0.1 nmol/l CGRP in CSF, was similar to or slightly greater than the response to 1 nmol/l synthetic CGRP (Fig. 2A). Thus overexpression of CGRP in the basilar artery, in addition to increases in CGRP in CSF, probably contributed importantly to responses of the artery in vitro. We speculate that responses of the arteries in vivo may be altered even more, because the arteries are exposed to CGRP present in the CSF in addition to CGRP that is present in the artery. Nevertheless, it is important to note that increases in cAMP in the basilar artery after gene transfer were appropriate for the concentration of CGRP that was achieved and provided clear evidence for bioactivity after gene transfer in vivo.

Effects of gene transfer on vascular reactivity. In carotid arteries after endothelial denudation, gene transfer of eNOS ex vivo enhances vasorelaxation to A-23187 (30). Perivascular delivery of eNOS by adenovirus-mediated gene transfer in vivo transfects adventitial fibroblasts and alters reactivity of the basilar artery (7). Thus expression of eNOS in adventitial cells alters vascular function.

Perivascular application of exogenous CGRP in a cranial window in vivo dilates the basilar artery (21), and release of CGRP by perivascular nerves in adventitia affects vascular tone (16). Thus CGRP appears to diffuse from the outer surface of the artery into the smooth muscle cells to produce relaxation. These findings suggested that vascular reactivity might be altered by transduction of CGRP in the adventitia. Because the increase in cAMP in the basilar artery after gene transfer of prepro-CGRP was similar with the response to 1 nmol/l synthetic CGRP, which relaxed the basilar artery by 50% (Fig. 2B), we anticipated that vascular reactivity might be altered by gene transfer of prepro-CGRP.

The present findings indicate that contraction of the basilar artery to histamine and serotonin in vitro was attenuated after gene transfer in vivo. The attenuation was due to an effect on CGRP₁ receptors, because pretreatment with CGRP-(8-37), an antagonist for CGRP₁ receptors, restored the response to normal. We speculate that two mechanisms may account for altered vascular reactivity after gene transfer of CGRP. First, as indicated by findings in fibroblasts in cell culture, basal release of CGRP from fibroblasts in adventitia may attenuate vasoconstrictor responses. Second, adventitial fibroblasts may have receptors for histamine or serotonin, and histamine and serotonin thus may stimulate release of CGRP. A similar mechanism has been described for activation of fibroblasts by bradykinin following gene transfer of eNOS (7). KCl also may release CGRP from adventitia, but depolarization from KCl may prevent hyperpolarization from CGRP. Thus Adprepro-CGRP may not alter responses to KCl.

In addition, relaxation of the basilar artery to IBMX was augmented after gene transfer also by an effect on CGRP₁ receptors. IBMX inhibits phosphodiesterase and prevents phosphodiesterases from breaking down both cAMP and cGMP (3) and thereby relaxes cerebral blood vessels (35). Binding of CGRP to receptors on vascular endothelial and smooth muscle cells produces the accumulation of both cAMP and cGMP (17, 22) and appears to augment vasorelaxation to IBMX. The present findings that vasorelaxation to acetylcholine and nitroprusside were not altered, indicates preservation of both endothelial and smooth muscle function after gene transfer. Thus gene transfer of prepro-CGRP to the basilar artery in vivo produced augmentation of relaxation to IBMX and attenuation of constriction to histamine and serotonin without alteration of endothelial function.

Release of CGRP from sensory nerves produces inflammation (10) with local edema and neutrophil accumulation by an increase in microvascular permeability (18). Transduced CGRP after gene transfer in vivo thus has the potential to affect neuronal and vascular function by a local inflammatory response. In this and previous studies (9, 33), leukocytes were observed in the vascular adventitia, surrounding leptomeninges, and subarachnoid space after injection of adenoviral vectors in the cisterna magna. In this study, we found however that leukocyte counts in CSF were similar after injection of Adgal and Adprepro-CGRP, which suggests that the inflammatory response is the result of primarily the virus rather than the transduced CGRP. A key point is that reactivity of arteries from rabbits treated with vehicle and Adgal is virtually identical, so that inflammation after viral injection does not appear to acount for alteration of vascular function.

Implications. We speculate that alteration of vascular reactivity after gene transfer of prepro-CGRP may become a strategy for cardiovascular and cerebrovascular diseases associated with decreased arterial diameter. In the cerebral circulation, this approach may be useful in the prevention of vasospasm after subarachnoid hemorhage. There is a >90% reduction of CGRP in the middle cerebral arteries of patients who die following SAH (11), which leads us to speculate that gene transfer of CGRP after SAH may produce a beneficial effect.

Intrathecal administration of CGRP increased arterial diameter in animals with vasospasm after SAH (19, 29). Intravenous administration of CGRP to patients with SAH improved neurological outcome despite hypotension (14, 20). We suggest that gene transfer of prepro-CGRP may have important advantages compared with administration of an exogenous peptide. In vivo synthesis of neurotrophic factors may be more effective than administration of exogenous neurotrophins (1, 8). By gene transfer, implantable devices or repeated injections are not needed to deliver the gene product, and peptide concentrations may remain relatively stable compared with repeated injection. Local gene transfer into CSF may avoid systemic effects of CGRP, including hypotension. Currently available adenoviral vectors have important disadvantages, especially induction of inflammation. Recent advances with adenoviral and other vectors, however, may enable greater expression of transgene products and minimal inflammation (24, 37) and increase application of gene transfer to clinical medicine.