Administration of adenoviral vectors induces gangrene in acutely ischemic rat hindlimbs: role of capsid protein-induced inflammation

PURPOSE

The initial purpose of this study was to determine the effects of intravascular adenoviral vector-mediated gene transfer of endothelial nitric oxide synthase (AdeNOS) on experimental hindlimb ischemia in the rat. Unexpectedly, administration of AdeNOS immediately after induction of acute limb ischemia led to limb gangrene. We subsequently sought to define the molecular mechanisms responsible for this unusual effect and to devise adenoviral gene transfer strategies to prevent the development of gangrene in acutely ischemic limbs.

METHODS

Phosphate-buffered saline or adenoviral vectors containing the bovine endothelial nitric oxide synthase gene (AdeNOS) or no transgene (Ad-E1) were injected intra-arterially into the hindlimb of a rat under vascular isolation immediately after surgical induction of severe ischemia. Hematoxylin and eosin staining was performed on muscle sections to evaluate inflammation. A separate group of animals was injected with an adenovirus containing a nontranscribable genome, treated with cyclosporine, or received delayed administration of the adenoviral vector. Gene expression after delayed adenoviral gene transfer was assessed with immunohistochemistry, Western blotting, and nitric oxide synthase (NOS) activity assay.

RESULTS

Both AdeNOS and Ad-E1 caused gangrene of the entire hindlimb within 12 days in a dose-dependent manner, at a threshold concentration of 1 x 10(9) plaque-forming unit/mL. Adenoviral delivery was associated with more inflammation and edema compared with phosphate-buffered saline histologically. Inactivation of adenoviral DNA transcription did not affect induction of gangrene. However, gangrene was prevented by concurrent immunosuppression with cyclosporine or delayed administration of the vector. Delayed administration allowed adenoviral gene expression as determined by immunohistochemistry, NOS protein levels, and an assay of NOS enzyme activity.

CONCLUSION

Intra-arterial administration of adenoviral vectors, under vascular isolation, immediately after induction of acute ischemia causes inflammation and subsequent limb gangrene. The inflammatory response is unrelated to the expression of the recombinant transgene or the adenoviral genome and is likely due to the adenoviral capsid proteins. However, administration of cyclosporine or delayed injection of the adenoviral vector is a method that can be used for adenoviral mediated gene transfer in limb ischemia.

Exercise-Induced Hyperemia Unmasks Regional Blood Flow Deficit in Experimental Hindlimb Ischemia

Many experimental models of hindlimb ischemia are characterized by spontaneous and rapid normalization of resting muscle blood flow (BF) rates which complicates the long-term evaluation of angiogenic therapies to reverse limb ischemia. We tested the hypothesis that peroneal nerve stimulation in an ischemic hindlimb would increase the oxygen (O(2)) demand and BF rate, thereby unmasking a severe blood flow deficit that is not apparent at rest.

METHODS

Ischemia was induced in adult rats by ligation of the left common iliac, femoral arteries, and their branches. Peroneal nerves were stimulated to allow measurement of exercise-induced regional BF rates with fluorescent microspheres. Hemodynamics were monitored. Fluorescent microspheres were injected before and after 5 min of nerve stimulation 3, 10, and 24 days postischemia. The tibialis anterior (TA) and gastrocnemius (GC) muscles and skin were harvested and weighed, and fluorescence was measured. BF rate was calculated as milliters per minute per gram of tissue and compared to normal muscle and skin of unoperated control rats. In order to determine the accuracy of BF rate measurements in ischemic muscle when <400 microspheres was delivered per specimen, 3 rats were studied by simultaneous injection of 4 x 10(5) blue and 1 x 10(5) yellow-green fluorescent microspheres. The correlation coefficient between the number of different colored microspheres delivered was measured.

RESULTS

The ischemia caused atrophy of the TA and GC muscles. The mean muscle mass of the ischemic TA and GC as a percentage of total body weight decreased over time vs control [TA 0.13 +/- 0.05% vs 0.25 +/- 0.03%, P < 0.05; GC 0.51 +/- 0.27% vs 0.70 +/- 0.07%, P = 0.07 at 24 days (24D)]. Despite clinical evidence of severe hindlimb ischemia in experimental groups, i.e., pressure sores, muscle atrophy, and weakness, resting BF rates were not significantly different from those of control. The BF rate of the TA was of 0.11 ml/min/g after 3D of ischemia, 0.14 ml/min/g after 10D, and 0.13 ml/min/g after 24D. The mean BF rate in normal muscle of unoperated controls was 0.16 ml/min/g (P > 0.05). However, the exercise-induced hyperemia in the skeletal muscle was significantly blunted in all of the ischemic groups. The unoperated control TA had a greater than 10-fold increase in BF to 1.95 ml/min/g in response to exercise while the ischemic TA had no increase in BF at 3D, 2-fold increase at 10D, and a 5-fold increase at 24D. Parallel findings were noted in the GC muscles. There was no significant difference in the BF rate in the skin. The accuracy of this microsphere technique in measuring very low BF rates found in ischemic muscle was supported by the significant correlation coefficient (r = 0.99) comparing two quantities of microspheres injected simultaneously.

CONCLUSION

Despite clinical signs of severe hindlimb ischemia, resting BF rates in the ischemic groups were not significantly decreased. Peroneal nerve stimulation resulted in up to 10-fold increase in BF rate and unmasked a severe deficit in vascular reserve in the ischemic groups. Resting BF rate is not always an accurate reflection of the flow deficit in models of critical limb ischemia, and this model of exercise-induced hindlimb hyperemia may allow better long-term evaluation of angiogenic therapies designed to reverse critical limb ischemia.

Calcium independence of phosphoinositide hydrolysis-induced increase in cyclic AMP accumulation in SK-N-SH human neuroblastoma cells

Previous work has shown that stimulation of muscarinic receptors in various cell lines increases intracellular cyclic AMP (cAMP) levels. This unusual response has been hypothesized to be mediated by stimulation of calcium/calmodulin-sensitive adenylate cyclase, secondary to inositol trisphosphate (IP3)-mediated calcium mobilization. To test this hypothesis, we stimulated muscarinic receptors in SK-N-SH human neuroblastoma cells while blocking the IP3-mediated rise in intracellular calcium concentration using two different methods. Loading cells with the intracellular calcium chelator 1,2-bis(o-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid (BAPTA) abolished the carbachol-mediated intracellular calcium release without abolishing the carbachol-mediated increase in cAMP level. Similarly, in cells preexposed to carbachol, the agonist-induced change in intracellular calcium level was blocked, but the cAMP response was not. Thus, both of these methods failed to block the muscarinic receptor-mediated increase in cAMP level, thereby demonstrating that this cAMP level increase is not mediated by a detectable rise in intracellular calcium concentration.

Differences in the functional responses of two cell lines each expressing Pi-hydrolysis-coupled muscarinic receptors

Fluorescent oxonol dyes were used to measure changes in the membrane potential of two different cell lines each expressing Pi-hydrolysis coupled muscarinic receptors. Both SK-N-SH human neuroblastoma cells and m1-transfected A9 L cells express muscarinic receptors which, when stimulated, elicit a large increase in intracellular calcium, and release of inositol phosphates. Despite the similarity in this second-messenger response, muscarinic stimulation resulted in a hyperpolarization in the transfected A9 L cells whereas a small depolarization was observed in the neuroblastoma cells. The carbachol-mediated hyperpolarization of the transfected A9 L cells could be mimicked by increasing intracellular calcium with the ionophore A23187, suggesting that it may be mediated by calcium-activated potassium channels. Exposure of SK-N-SH cells to A23187, on the other hand, had no effect on the membrane potential. These studies demonstrate that the activation of a second messenger system does not solely dictate the electrophysiological response of a cell, but that other factors such as the expression of ion-channels is critical in the determination of that response.

Agents that stimulate phosphoinositide turnover also elevate cAMP in SK-N-SH human neuroblastoma cells

Stimulation of m1 and of m3 muscarinic receptors has previously been shown to increase intracellular cAMP levels in a variety of cells. Although the mechanism underlying this response is not fully understood, it has been hypothesized to be secondary to the IP3-mediated rise in intracellular calcium. In order to determine whether other means of elevating intracellular calcium also raise cAMP levels, we stimulated SK-N-SH human neuroblastoma cells with bradykinin or with maitotoxin. Both of these agents stimulated phospholipase C, stimulated inositol phosphate release and elevated cAMP levels, thus demonstrating that this cAMP response is not unique to muscarinic receptor stimulation.

Functionalized congener approach to muscarinic antagonists: analogues of pirenzepine

The M1-selective muscarinic receptor antagonist pirenzepine 6H-pyrido[2,3-b][1,4]benzodiazepin-6-one) was derivatized to explore points of attachment of functionalized side chains for the synthesis of receptor probes and ligands for affinity chromatography. The analogues prepared were evaluated in competitive binding assays versus [3H]-N-methylscopolamine at four muscarinic receptor subtypes (m1AChR-m4AChR) in membranes from rat heart tissue and transfected A9L cells. 9-(Hydroxymethyl)pirenzepine, 8-(methylthio)pirenzepine, and a series of 8-aminosulfonyl derivatives were synthesized. Several 5-substituted analogues of pirenzepine also were prepared. An alternate series of analogues substituted on the 4-position of the piperazine ring was prepared by reaction of 4-desmethylpirenzepine with various electrophiles. An N-chloroethyl analogue of pirenzepine was shown to form a reactive aziridine species in aqueous buffer yet failed to affinity label muscarinic receptors. Within a series of aminoalkyl analogues, the affinity increased as the length of the alkyl chain increased. Shorter chain analogues were generally much less potent than pirenzepine, and longer analogues (7-10 carbons) were roughly as potent as pirenzepine at m1 receptors, but were nonselective. Depending on the methylene chain length, acylation or alkyl substitution of the terminal amine also influenced the affinity at muscarinic receptors.

Muscarinic receptor binding and activation of second messengers by substituted N-methyl-N-[4-(1-azacycloalkyl)-2-butynyl]acetamides

A series of substituted azacycloalkyl analogues of the muscarinic agonist UH 5 (N-methyl-N-[4-(1-pyrrolidinyl)-2-butynyl]acetamide, 1a) were synthesized and evaluated pharmacologically. These compounds were developed as intermediates for further derivatization leading to functionalized congeners of 1a. The compounds were synthesized by using a Mannich-type condensation of N-acetyl-N-methylpropargylamine to various substituted saturated azaheterocycles. The compounds were screened at a single concentration in competitive binding assays in rat cerebral cortical membranes against either [3H]N-methylscopolamine (at 100 microM) or [3H]oxotremorine-M (at 1 microM) labels. Candidates were then selected for further evaluation of their effect on phosphoinositide (PI) turnover in membranes from A9L cells transfected with cDNA of either m1-muscarinic cholinergic receptors (m1AChRs) or m3AChRs. The analogues were also tested for the inhibition of adenylate cyclase in NG108-15 cells expressing m4AChRs. The azetidine analogue of 1a had a Ki value of 12 nM for the inhibition of [3H]oxotremorine-M binding in rat brain and had an agonist potency at m1-,m3-, and m4AChRs comparable to 1a. The substituted 5- and 6-member ring analogues generally had lower binding affinities and were less potent than 1a in stimulating PI turnover. Several compounds were moderately effective in inhibiting cyclic AMP production in NG108-15 cells.

Muscarinic receptor subtype selectivity of novel heterocyclic QNB analogues

In an effort at synthesizing centrally-active subtype-selective antimuscarinic agents, we derivatized QNB (quinuclidinyl benzilate), a potent muscarinic antagonist, by replacing one of the phenyl groups with less lipophilic heterocyclic moieties. The displacement of [3H]-N-methyl scopolamine binding by these novel compounds to membranes from cells expressing m1-m4 receptor subtypes was determined. Most of the novel 4-bromo-QNB analogues were potent and slightly selective for m1 receptors. The 2-thienyl derivative was the most potent, exhibiting a 2-fold greater potency than BrQNB at m1 receptors, and a 4-fold greater potency at m2 receptors. This compound was also considerably less lipophilic than BrQNB as determined from its retention time on C18 reverse phase HPLC. This compound may therefore be useful both for pharmacological studies and as a candidate for a radioiodinated SPECT imaging agent for ml muscarinic receptors in human brain.