Aminoguanidine Ameliorates and l-Arginine Worsens Brain Damage From Intraluminal Middle Cerebral Artery Occlusion
Background and Purpose We studied whether the inducible nitric oxide synthase (iNOS) inhibitor aminoguanidine reduces focal cerebral ischemic damage in a relatively noninvasive stroke model in which the rat middle cerebral artery (MCA) is occluded using an intravascular filament.
Methods In rats anesthetized with halothane, a nylon filament was advanced into the internal carotid artery until its tip occluded the origin of the MCA. The filament was left in place for 2 hours and then withdrawn. Twenty-four hours later, rats received intraperitoneal injections of aminoguanidine (100 mg/kg BID; n=7), aminoguanidine+l-arginine (300 mg/kg QID; n=7), l-arginine alone (n=6), d-arginine alone (n=6), or vehicle (n=10). Drugs were administered for 3 consecutive days. Infarct volume was determined by image analysis in thionin-stained brain sections 4 days after ischemia. iNOS mRNA was detected with the use of reverse transcription polymerase chain reaction.
Results Cerebral ischemia led to iNOS mRNA expression in the affected brain 48 hours after induction of ischemia. Administration of aminoguanidine reduced neocortical infarct volume by 26% (P<.05 versus vehicle, ANOVA and Tukey’s test), a reduction that was antagonized by coadministration of l-arginine (P>.05 versus vehicle). Administration of l-arginine alone, but not d-arginine, enlarged the infarct by 29% (P<.05). Aminoguanidine or l-arginine did not influence the increase in water content in the postischemic brain, indicating that the effect on infarct volume is not related to modulation of ischemic edema.
Conclusions These results demonstrate that cerebral ischemia is also associated with iNOS expression in a minimally invasive model of transient MCA occlusion and that iNOS inhibition reduces focal ischemic damage. The findings support the hypothesis that nitric oxide produced by iNOS contributes to ischemic brain damage and that inhibition of iNOS may be a valuable tool in the management of cerebral ischemia.
We have previously demonstrated that focal cerebral ischemia produced by occlusion of the rat MCA is associated with expression of iNOS in the postischemic brain.1 2 We have also shown that administration of the relatively selective iNOS inhibitor AG, starting 24 hours after MCA occlusion, ameliorates cerebral ischemic damage.3 The protective effect of AG is reversed by l- but not d-arginine and is independent of inhibition of neuronal or endothelial NOS.3 These findings suggest that NO produced by iNOS contributes to the cerebral ischemic damage and that iNOS inhibition may be valuable in the treatment of cerebral ischemia.
The method for induction of focal ischemia used in these earlier studies, developed by Tamura et al,4 involves performing a craniotomy and opening the dura to expose and cauterize the MCA.5 These neurosurgical manipulations are sufficient to produce low levels of iNOS expression in brain even in the absence of cerebral ischemia.2 It is therefore conceivable that the traumaand inflammation resulting from the surgical procedures facilitate or enhance iNOS expression in the ischemic brain. Consequently, the role of iNOS in the tissue damage might be less pronounced or absent in stroke models in which the procedure for MCA occlusion does not result in background iNOS expression. In view of the potential therapeutic importance of iNOS inhibitors, it would be desirable to determine whether iNOS also contributes to tissue damage when focal ischemia is induced by a less traumatic model.
In this study, therefore, we investigated the effect of AG on focal cerebral ischemic damage using a technique in which the rat MCA is occluded without opening the skull. In this model a nylon filament is inserted into the internal carotid artery and is advanced intraluminally to occlude the origin of the MCA.6 7 We found that posttreatment with AG reduces cerebral ischemic damage in this less invasive model as well and that l-arginine, the substrate for NO synthesis, enlarges the size of the stroke. The data provide additional support for the hypothesis that NO produced by iNOS contributes to cerebral ischemic damage and that AG and other iNOS inhibitors might be useful tools in the management of cerebral ischemia in stroke patients.
Materials and Methods
Methods for MCA occlusion with monitoring of physiological parameters, for determination of infarct size and for detection of iNOS mRNA by RT-PCR, have been described in detail in previous publications2 3 5 8 and will only be summarized.
General Surgical Procedures
Studies were approved by the Animal Care Committee of the University of Minnesota. Experiments were conducted on 71 male Sprague-Dawley rats (Harland) weighing 300 to 400 g. Under halothane anesthesia (induction, 5%; maintenance, 1%, in 100% oxygen) the left femoral artery was cannulated, and rats were placed on a stereotaxic frame (Kopf). Body temperature was maintained at 37±0.5°C by a thermostatically controlled infrared lamp. The arterial catheter was connected to a pressure transducer for recording of mean arterial pressure and heart rate. Plasma glucose was measured by a glucose analyzer (Beckman). Blood gases were not measured because it was previously determined that arterial Pco2, Po2, and pH are stable during procedures for MCA occlusion in spontaneously breathing rats.5 After completion of the surgical procedures, the arterial catheter was tunneled under the skin and exteriorized at the level of the tail. The catheter was used for recording of arterial pressure and for determination of plasma glucose and hematocrit at different times after MCA occlusion.
MCA Occlusion and Measurement of Infarct Volume
The MCA was occluded by insertion of an intravascular filament according to a technique used extensively in studies of cerebral ischemia.6 7 Under halothane anesthesia a 4-0 nylon monofilament, the tip of which was rounded with the use of an open flame, was inserted centripetally into the external carotid artery and advanced into the internal carotid until a slight resistance was felt. Such resistance was an indication that the filament was wedged into the circle of Willis. Rats were returned to their cages, and their temperature was maintained at 37°C. Two hours after induction of ischemia, rats were reanesthetized and the filament was slowly withdrawn until the tip reached the external carotid artery. Animals were then returned to their cages and closely monitored until they recovered from anesthesia completely.
Ninety-six hours after induction of ischemia, rats were killed and their brains removed. The forebrain was frozen in cooled isopentane (−30°C). Coronal forebrain sections (thickness, 30 μm) were serially cut in a cryostat, collected at 300-μm intervals, and stained with thionin. As described in detail elsewhere,5 infarct volume was determined with the use of an image analyzer (MCID, Imaging Research Inc). Infarct volume in the cerebral cortex was corrected for swelling according to the method of Lin et al,9 as previously described.3 10
Determination of Brain Water Content
The infarcted cortex and the contralateral intact cortex were dissected on a cooled glass plate and placed in preweighed vials.11 The vials were rapidly capped, reweighed, and placed in an oven at 60°C with the cap loosened. Vials were then weighed 7 days later for determination of the dry weight. The percent water content of the tissue was computed from the following formula: dry weight/wet weight×100.
iNOS mRNA was detected in the ischemic brain by RT-PCR, as previously described.2 8 12 A 4-mm-thick coronal brain slice was cut at the level of the optic chiasm, and the infarcted cortex was dissected with the corpus callosum as a ventral landmark. The corresponding region of the contralateral cortex was also dissected. Total RNA was extracted from the tissue according to the method of Chomczynski and Sacchi.13 The integrity of the RNA was determined on denaturing formaldehyde gels. First strand cDNA synthesis was then performed with the use of 0.25, 0.5, and 1.0 μg of total RNA, oligo(dT) primer (BRL), and M-MuLV reverse transcriptase (New England BioLabs) according to manufacturer’s instructions. Aliquots (5 μL each) from the RT reaction were then used for PCR amplification with primer pairs for both iNOS and a ubiquitously expressed control sequence porphobilinogen deaminase. NOS primer sequences were chosen for their ability to distinguish iNOS from the other NOS isoforms and were designed to flank a known intron-exon boundary of the genomic iNOS sequence.8 Therefore, only products corresponding to the iNOS mRNA were amplified by RT-PCR. This eliminates the concern that products might be generated from genomic DNA contaminants in the RNA samples. The iNOS primer pair used was as follows: forward: 5′ ACAACGTGGAGAAAACCCCAGGTG 3′; reverse: 5′ ACAGCTCCGGGCATCGAAGACC 3′. The “hot start” method was used (Stratagene) with the following cycle parameters: 94°C, 60 seconds; 68°C, 60 seconds; 73°C, 45 seconds for five cycles, then 94°C, 60 seconds; 62°C, 60 seconds; 73°C, 45 seconds for 35 cycles. Reaction products were then separated on a 7% acrylamide gel, ethidium stained, and photographed before resolved bands were transferred to nylon filters. The size of the PCR fragment representing iNOS was 557 bp. Southern blot analysis of filters with the use of an internal iNOS cDNA was then performed to confirm the identity of RT-PCR products.2 8 Each set of PCRs included control samples run without RNA or in which the RT step was omitted. The RT-PCR procedure was highly reproducible under the present experimental conditions.
RT-PCR in Sham-Operated Rats and in Rats Subjected to Cerebral Ischemia
In these experiments we studied iNOS mRNA expression in the cerebral cortex in sham-operated rats and in rats 48 hours after induction of ischemia using the intraluminal filament technique. In sham-operated rats (n=3), the external carotid artery was surgically exposed and prepared for insertion of the filament, but the filament was not inserted. Brain samples for RT-PCR were collected 48 hours later. Rats (n=3) in which RT-PCR was performed 48 hours after MCA occlusion by the intraluminal filament served as positive controls. The 48-hour time interval was selected because it is at this time that iNOS expression is maximal after focal cerebral ischemia produced by the Tamura model.2
Effect of AG and/or Arginine on Infarct Volume and Water Content
The MCA was occluded for 2 hours and treatments were begun 24 hours after induction of ischemia. Drugs were given for 3 consecutive days according to a protocol previously established.3 In one group of rats (n=7), AG (Sigma; 100 mg/kg in 1 mL of saline) was administered intraperitoneally at 10 am and 6 pm. A second group of rats (n=7) was treated with both AG (100 mg/kg at 10 am and 6 pm) and l-arginine (Sigma; 300 mg/kg in 1 mL of saline at 9 am, 10 am, 5 pm, and 9 pm). A third group of rats (n=6) was treated with l-arginine alone (300 mg/kg at 9 am, 10 am, 5 pm, and 9 pm). A fourth group of rats (n=6) was treated with d-arginine alone (300 mg/kg at 9 am, 10 am, 5 pm, and 9 pm), and a fifth group of rats (n=10) was treated with saline (vehicle; 1 mL at 10 am and 6 pm). The pH of the solutions injected was adjusted to 7.0. Arterial pressure, rectal temperature, and plasma glucose were measured daily at 9 am, 1 pm, 5 pm, and 9 pm. Arterial hematocrit was measured before injection and 24, 48, 72, and 96 hours after ischemia. Ninety-six hours after induction of ischemia, rats were killed for determination of infarct size. We have previously demonstrated that treatment with AG according to this protocol inhibits postischemic brain iNOS activity without affecting cNOS activity.3
Water content was measured in the cerebral cortex in sham-operated rats (n=5), vehicle-treated rats (n=9), and rats treated with AG (n=8) or l-arginine (n=7) according to the protocol described above. Measurements were performed 96 hours after induction of ischemia.
Data presented in the text, table, and figures are expressed as mean±SD. Comparisons among multiple groups were statistically evaluated by ANOVA and the Tukey-Kramer modification of Tukey’s test (Systat, Inc). Comparisons between two groups were evaluated by Student’s t test. Differences were considered significant at P<.05.
iNOS mRNA Expression in Sham-Operated Rats and in Rats Subjected to Ischemia
In these experiments we used RT-PCR to study iNOS mRNA expression after MCA occlusion using the intraluminal filament technique. The results are illustrated in Fig 1⇓. In sham-operated rats, no iNOS signal was detected on either side of the cerebral cortex. In contrast, in rats in which the MCA was occluded with the use of the intraluminal filament, iNOS mRNA expression was observed in the postischemic cerebral cortex 48 hours after induction of ischemia.
Effect of AG and/or Arginine on Infarct Size
Arterial pressure and plasma glucose remained stable during the period of study (Fig 2⇓). Differences in arterial pressure and plasma glucose among the treatment groups were small and, with one exception (Fig 2⇓), not statistically significant (P>.05, ANOVA and Tukey’s test). No differences in hematocrit were noted among the different groups during the experimental period (P>.05; data not shown). Rectal temperature rose by 1°C to 1.5°C during the first 36 to 48 hours after ischemia in all groups of rats studied (Fig 2⇓; P<.05). Mild hyperthermia has been previously reported with this model and has been attributed to hypothalamic ischemia or infarction resulting from occlusion of arterial branches arising from the circle of Willis and supplying the hippocampal region.14 Except for two time points (Fig 2⇓), values did not differ in the various treatment groups (P>.05).
In vehicle-treated rats (n=10), MCA occlusion caused reproducible infarcts involving the cerebral cortex and the striatum (Table⇓). The magnitude and regional distribution of the ischemic lesion were similar to those previously reported when this technique was used for MCA occlusion (eg, References 15 and 16). Administration of AG (n=7) reduced infarct size in cerebral cortex (P<.05, ANOVA) but not in the striatum (Table⇓, Fig 3⇓; P>.05). The reduction in stroke size was not observed in rats in which AG was coadministered with l-arginine (n=7) (P>.05).
The effect of l-arginine alone on infarct size was then investigated. l-Arginine (n=6) enlarged the size of the infarct in neocortex but not in the striatum (Table⇑, Fig 3⇑; P<.05). In contrast, d-arginine (n=6) had no effect on infarct size (Table⇑, Fig 3⇑; P>.05), indicating that the effect of arginine on infarct volume is stereospecific.
Effect of AG or l-Arginine on Infarct Water Content
To provide evidence that the changes in infarct volume produced by AG or l-arginine were not due to effects on ischemic brain edema, the water content of the cerebral cortex was measured in rats treated with AG or l-arginine. In vehicle-treated rats (n=9), MCA occlusion increased water content in the ipsilateral cerebral cortex (Fig 4⇓; P<.05), a finding reflecting postischemic edema. In the cortex contralateral to the infarct, water content was not different from that of the cerebral cortex of sham-operated rats (n=5; P>.05). AG or l-arginine did not affect the magnitude of the increase in water content in the infarcted brain (P>.05 from contralateral side or vehicle).
We studied the effect of posttreatment with the iNOS inhibitor AG on the infarct produced by occlusion of the MCA using an intraluminal filament. We had previously shown that 24 to 96 hours after focal cerebral ischemia there is expression of iNOS mRNA and enzymatic activity in the postischemic brain.1 2 iNOS immunoreactivity was observed in inflammatory neutrophils infiltrating the ischemic brain.2 Administration of AG 24 hours after MCA occlusion reduced infarct size, suggesting that iNOS induction plays a role in the development of brain damage.3 The model used for induction of focal ischemia, however, involves surgical exposure of the MCA through a craniotomy and occlusion of the artery by cautery. In this model, iNOS expression was observed even in sham-operated rats in which the artery was exposed and manipulated but not occluded.2 The possibility therefore exists that the trauma and inflammation associated with the procedure for MCA occlusion could have primed and/or enhanced iNOS induction, leading to artificially high levels of iNOS in the ischemic brain. Consequently, the protective effect of iNOS inhibition may occur only in this particular model of focal ischemia in which levels of iNOS expression are amplified by the trauma of the surgical procedure.
In the present study we therefore investigated the effect of AG in a less invasive model of focal ischemia in which the MCA is occluded by intraluminal insertion of a nylon filament. Using RT-PCR to detect iNOS mRNA, we found that iNOS expression occurs only in the postischemic brain but not contralaterally or in the brain of sham-operated rats. We then studied whether AG could also reduce ischemic damage in this model of cerebral ischemia. It was found that treatment with AG, starting 24 hours after induction of ischemia, reduced infarct volume. The protective effect was not observed if AG was administered in conjunction with l-arginine. In addition, administration of l-arginine alone enlarged infarct volume, an effect that was stereospecific because d-arginine did not influence the volume of the infarct. The data, therefore, suggest that AG reduces the size of the ischemic lesion in this relatively nontraumatic model of MCA occlusion as well and that the NO precursor l-arginine enlarges the size of the infarct. The findings of the present study also indicate that iNOS expression also occurs in transient focal ischemia. However, the complete time course of iNOS expression (iNOS mRNA and enzymatic activity) and the cell type(s) in which iNOS is expressed remain to be defined. Studies addressing these important issues are under way.
The effect of AG and arginine cannot be the result of differences in arterial pressure or plasma glucose because these parameters were carefully monitored and were not substantially different in the treatment groups. Differences in rectal temperature were observed at selected time points in some groups, but these small and transient differences are unlikely to be responsible for the profound effects of AG or l-arginine on infarct size. The protection exerted by AG cannot be related to effects on cerebral blood flow because we have previously demonstrated that this drug does not affect resting cerebral blood flow or the reactivity of the cerebral circulation to hypercapnia.3 Similarly, the influence of AG or l-arginine on infarct size cannot be the result of effects on ischemic edema rather than brain damage because treatment with these agents does not affect the ischemia-induced changes in water content in the affected brain. Finally, it is unlikely that iNOS expression in sham-operated rats occurred earlier than 48 hours after ischemia and was missed. This is because in preliminary experiments we found that iNOS is not expressed 6, 12, and 24 hours after sham surgery (F.Z. et al, unpublished data). Therefore, effects related to iNOS expression not secondary to cerebral ischemia can be excluded.
Administration of AG reduced the size of the ischemic lesion in the cerebral cortex but not in the striatum. This finding is in agreement with previous observations indicating that neuroprotective treatments are less efficacious or even ineffective in the striatum (eg, see References 5 and 17 through 19). Perhaps the paucity of collateral anastomosis in the striatal circulation predisposes this region to intense ischemia and irretrievable tissue damage (see References 5 and 20 for discussion). In addition, poor collateral flow may limit the delivery into the striatum of blood-borne neuroprotective agents.
We have previously demonstrated that AG, at the doses used in the present study, attenuates iNOS enzymatic activity in the postischemic brain without affecting cNOS activity.3 It is therefore likely that the reduction in infarct size produced by AG is related to iNOS inhibition in this model as well. Although in theory AG could also act by downregulating iNOS expression, we have no experimental data to support or dismiss this possibility. The hypothesis that AG acts by inhibiting iNOS enzymatic activity is supported by the observation that l-arginine, the precursor of NO, counteracts the protective effect of AG. However, we cannot rule out the possibility that l-arginine and AG influence infarct size by distinct mechanisms (see below) and that their opposing effects cancel each other out.
It was also found that l-arginine enlarged the size of the infarct while d-arginine did not. This finding is consistent with the hypothesis that l-arginine increases tissue damage by enhancing NO production by iNOS. Unlike the constitutive isoforms of NOS, enzymes that are activated only during increases in intracellular calcium, iNOS is continuously active.21 Thus, NO production by iNOS depends greatly on the availability of substrates and cofactors, including l-arginine (see Reference 22 for a review). It is therefore conceivable that administration of l-arginine increased the output of NO by iNOS. Another possibility is that the effect of l-arginine on infarct size is not related to NO synthesis. l-Arginine is endowed with other biological actions that could potentially influence cerebral ischemic damage. These include stimulation of insulin, glucagon, growth hormone, and catecholamine secretion.23 Arginine could also be converted to ornithine and serve as a substrate for the synthesis of polyamines24 or could be decarboxylated to form agmatine, a putative neurotransmitter.25 Effects of l-arginine on hormonal secretion are unlikely because administration of this amino acid did not affect body temperature, arterial pressure, and plasma glucose, parameters that would be altered by hypersecretion of these hormones. However, it remains to be established whether effects on the synthesis of polyamines or agmatine could contribute to the worsening of cerebral ischemic damage observed with l-arginine administration. Irrespective of the potential mechanisms, the finding that l-arginine worsens ischemic damage suggests that administration of this amino acid to patients with ischemic stroke, in the setting of either oral or parenteral nutrition, may not be desirable.
At variance with the results of the present study, we previously found that l-arginine does not affect infarct size in focal cerebral ischemia produced by permanent MCA occlusion in spontaneously hypertensive rats.3 The reasons for this discrepancy are unclear at the present time. One possibility is that the deleterious effects of l-arginine are more marked in transient ischemia than in permanent ischemia. Alternatively, differences in the rat strain (spontaneously hypertensive rats versus Sprague-Dawley rats) could also play a role. Further studies are needed to address this issue.
MCA occlusion by an intraluminal filament is associated with hyperthermia, a condition that aggravates cerebral ischemic damage (eg, Reference 26). This effect may be related to occlusion of small arteries arising from the circle of Willis and vascularizing the hypothalamus and its thermoregulatory centers.14 It has been proposed that the hyperthermia associated with this model may obliterate the effect of neuroprotective drugs.14 For example, the free radical scavenger dimethylthiourea fails to ameliorate focal cerebral ischemic damage produced by the intraluminal filament technique.27 In the present study, instead, it was found that the protective effect of AG occurs despite hyperthermia. This observation indicates that the protective effect of AG is robust and occurs even in experimental paradigms in which other neuroprotective agents lose their effectiveness. It also suggests that the reduction in infarct size afforded by AG could have been greater in the absence of hyperthermia.
We have demonstrated that AG or l-arginine can modulate the size of the infarct even when they are administered 24 hours after induction of the ischemia. These findings suggest that the extent of brain damage is not set a few hours after the ischemic event but that the processes that ultimately lead to the tissue damage can be influenced long after the induction of the ischemia. In support of this hypothesis, morphological studies have demonstrated that after focal cerebral ischemia there is evolution of the tissue damage for days after the ischemic event.15 28 Furthermore, MCA occlusion may initiate genetically driven death programs in the ischemic penumbra that produce apoptotic neuronal death over several days.29 30 It is therefore likely that while in the ischemic core acute cerebral ischemia rapidly produces irreversible tissue damage, in more peripheral regions the process of neuronal death is slower and can be modulated, resulting in larger or smaller infarcts. Our findings therefore support the notion that even 24 hours after cerebral ischemia the brain damage is still in evolution and as such can be targeted by specific therapeutic interventions. On the other hand, the brain damage can also be worsened by factors that enhance these delayed pathogenic processes.
In conclusion, we have demonstrated that transient occlusion of the rat MCA with the use of an intraluminal filament induces iNOS expression in the postischemic brain. Administration of the iNOS inhibitor AG reduces, while the NO precursor l-arginine enlarges, the size of the infarct. These data strengthen previous evidence by demonstrating that cerebral ischemia is followed by iNOS expression in a relatively noninvasive method for MCA occlusion. Furthermore, the findings provide additional support for the hypothesis that NO produced by iNOS contributes to ischemic brain damage and that inhibition of iNOS may be a valuable tool in the management of cerebral ischemia.
Selected Abbreviations and Acronyms
|cNOS||=||constitutive nitric oxide synthase|
|iNOS||=||inducible nitric oxide synthase|
|MCA||=||middle cerebral artery|
|NOS||=||nitric oxide synthase|
|RT-PCR||=||reverse transcription polymerase chain reaction|
This study was supported by a grant from the American Heart Association and the National Institutes of Health (NS-34179). Dr Iadecola is an Established Investigator of the American Heart Association.
- Received August 28, 1995.
- Revision received October 2, 1995.
- Accepted October 16, 1995.
- Copyright © 1996 by American Heart Association
Iadecola C, Xu X, Zhang F, El-Fakahany EE, Ross ME. Marked induction of calcium-independent nitric oxide synthase activity after focal cerebral ischemia. J Cereb Blood Flow Metab. 1995;14:52-59.
Iadecola C, Zhang F, Xu X. Inhibition of inducible nitric oxide synthase ameliorates cerebral ischemic damage. Am J Physiol. 1995;268:R286-R292.
Zea Longa E, Weinstein PR, Carlson S, Cummins R. Reversible middle cerebral artery occlusion without craniectomy in rat. Stroke. 1989;20:84-91.
Nagasawa H, Kogure K. Correlation between cerebral blood flow and histologic changes in a new rat model of middle cerebral artery occlusion. Stroke. 1989;20:1037-1043.
Ross ME, Iadecola C. Nitric oxide synthase expression in cerebral ischemia: neurochemical, immunocytochemical and molecular approaches. In: Packer L, ed. Methods in Enzymology. Orlando, Fla: Academic Press. In press.
Lin TN, He YY, Wu G, Khan M, Hsu CY. Effect of brain edema on infarct volume in a focal cerebral ischemia model in rats. Stroke. 1993;24:117-121.
Kawasaki ES, Clark SS, Coyne MY, Smith SD, Champlin R, Witte ON, McCormick FP. Diagnosis of chronic myeloid and acute lymphocytic leukemia by detection of leukemia-specific mRNA sequence amplified in vitro. Proc Natl Acad Sci U S A. 1988;85:5698-5702.
Ginsberg MD, Busto R. Rodent models of cerebral ischemia. Stroke. 1989;20:1627-1642.
Morris SJ, Billiar TR. New insights into the regulation of inducible nitric oxide synthesis. Am J Physiol.. 1994;266:E829-E839.
Barbul A. Physiology and pharmacology of arginine. In: Moncada S, Higgs EA, ed. Nitric Oxide From L-Arginine: A Bioregulatory System. New York, NY: Elsevier Science Publishing Co; 1990:317-329.
Morgan DM. Polyamines, arginine and nitric oxide. Biochem Soc Trans.. 1994;22:879-883.
Li G, Regutanan S, Barrow CJ, Eshraghi J, Cooper R, Reis DJ. Agmatine: an endogenous clonidine-displacing substance in the brain. Science. 1994;263:966-969.
Linnick MD, Zobrist RH, Hatfield MD. Evidence supporting a role for programmed cell death in focal cerebral ischemia in rats. Stroke. 1993;24:2002-2009.