Main

The pathogenesis of perinatal cerebral hypoxic-ischemic injury has been extensively examined by using models of carotid ligation and subsequent exposure to moderate hypoxia in immature rats. Many interrelated mechanisms—including excessive stimulation of excitatory amino acid receptors, intracellular calcium accumulation, lipid peroxidation, and free radical generation—may play a role in the evolution of hypoxic-ischemic injury (1, 2). Recent evidence implicates inflammatory cells, cytokines (e.g., TNF-α and IL-1β), and the lipid mediator platelet-activating factor as mediators of hypoxic-ischemic injury in immature brain. Neutrophils (3) and activated macrophages and microglia accumulate (4, 5), and TNF-α and IL-1β protein and gene expression increase within 6 h after a cerebral hypoxic-ischemic insult in rats at postnatal d 7 (P7) (6, 7). Systemic administration of either a pharmacologic antagonist of the IL-1 receptor (8) or a platelet-activating factor receptor antagonist (9) attenuates neonatal cerebral hypoxic-ischemic injury in P7 rats.

The alkylxanthine phosphodiesterase inhibitor pentoxifylline [3,7-dimethyl-1-(5-oxohexyl)-xanthine]—which improves red cell deformability, decreases platelet and red cell aggregation and lowers fibrinogen levels and plasma viscosity (10)—is already used clinically for treatment of intermittent claudication (11, 12). Recently, attention has focused on the influence of pentoxifylline on mononuclear phagocyte, neutrophil, and endothelial activation and production of inflammatory and thrombogenic mediators. In vitro, pentoxifylline inhibits gene transcription and production of TNF-α by mononuclear phagocytes (13) and microglia (14), and inhibits tissue factor gene transcription and production by activated endothelium (15, 16). In addition, pentoxifylline inhibits production of platelet-activating factor in experimental ischemia-reperfusion injury (17) and inhibits activation of neutrophils by inflammatory mediators such as TNF-α (18–20). Pentoxifylline penetrates the blood-brain barrier rapidly and efficiently after systemic administration (21). In vivo, systemically administered pentoxifylline attenuates brain TNF-α production after closed head injury in rats (22) or brain irradiation in mice (23), and it attenuates experimental allergic encephalomyelitis in rats (24).

The multiple potentially beneficial vascular and anti-inflammatory activities of pentoxifylline led several investigators to test the neuroprotective efficacy of pentoxifylline in mature animal models of cerebral ischemia; results were inconsistent (25–29). None of those reports addressed the efficacy of pentoxifylline in an immature animal model of cerebral ischemic injury. Reports of enhanced effects of pentoxifylline on neonatal neutrophil function in vitro (30, 31), as well as demonstration of neuroprotective efficacy of other anti-inflammatory strategies in immature brain (8, 9, 32), prompted us to test the hypothesis that pentoxifylline would attenuate neonatal hypoxic-ischemic brain damage. We used an in vivo model of neonatal cerebral hypoxia-ischemia with subsequent reperfusion, elicited by unilateral carotid ligation followed by timed exposure to moderate hypoxia in P7 rats. Precise correlations with human brain development are difficult; however, brains of P7 rats are similar to third-trimester human fetuses (and thus premature or full-term human newborns) in terms of cellular proliferation, cortical organization, synapse number, neurotransmitter synthetic enzymes, and electrophysiology (33–35).

We evaluated the neuroprotective efficacy of pentoxifylline by comparing the incidence of liquefactive necrosis and by measuring regional cross-sectional areas in drug-treated and control animals. Clinically, neuroprotective drugs might be used either prophylactically, in patients at known risk of hypoxic-ischemic brain injury (i.e., infants undergoing cardiopulmonary bypass for treatment of congenital heart disease) or as post-insult “rescue” therapy (e.g., after acute perinatal asphyxia). In these experiments, we tested the neuroprotective efficacy of pentoxifylline in both prophylactic and rescue regimens. Our results demonstrated that pentoxifylline treatment initiated either before or immediately after cerebral hypoxia-ischemia attenuated brain damage in P7 rats.

METHODS

Hypoxia-ischemia.

In methoxyflurane-anesthetized P7 Sprague-Dawley rats (Charles River Laboratories), the right carotid artery was isolated and double-ligated (<5 min procedure); 1–2 h later, animals were exposed to 8% oxygen (balance nitrogen) for either 3.25 h (first three experiments) or 2.75 h (subsequent two experiments) in glass chambers, partially submerged in a water bath (temperature maintained at 38°C). The hypoxia duration was shortened in two experiments, because in both litters, the mean animal weight was significantly higher than the first three litters [mean (±SD) weight on P7: first three litters 11.8 (±1.2) g; subsequent two litters 15.2 (±1.3) g;p = 0.0001, t test]. Our strategy to compensate for increased animal weight was to reduce the duration of hypoxia exposure, to attempt to maintain a comparable severity of injury in control animals. This approach was based on a report that the severity of hypoxic-ischemic brain injury among P7 rats varies in direct proportion to weight;i.e., the higher the body weight, the greater the severity of brain injury (36). Experiments that used different hypoxia durations were analyzed separately (Data Analysis and Results). In two separate experiments, to test the efficacy of posthypoxic-ischemic pentoxifylline treatment, rats underwent right carotid ligation followed by 2.75 h hypoxia (8% O2) exposure. For 1 h after the end of hypoxia, animals were housed in a warm air incubator (ambient temperature: 35–36°C), before they were returned to the dam.

At higher doses (200 mg/kg), pentoxifylline can induce mild hypothermia (about 2°C core temperature reduction) in mice (37). Because hypothermia can attenuate cerebral ischemic injury, intrahypoxic skin temperature was measured in a subset of pretreated rats, and esophageal temperature was monitored for 8 h after hypoxia in rats from four experiments (40 or 75 mg/kg/dose) (YSI thermometer 43TA with probe 554, Yellow Spring Instruments, Yellow Spring, OH).

For evaluation of histopathologic outcome, rats were decapitated and brains rapidly dissected and frozen under powdered dry ice on P12. All efforts were made to minimize animal suffering and minimize the number of animals used. Experimental protocols were approved by the University of Michigan Committee on Use and Care of Animals.

Drug treatment.

In five independent experiments to test the efficacy of prehypoxic-ischemic treatment, P7 rats received intraperitoneal injections of pentoxifylline (Research Biochemicals International, Natick, MA) immediately before and again immediately after hypoxia. In the first three experiments, with 3.25 h hypoxia duration, P7 rats received pentoxifylline 25, 75, or 150 mg/kg/dose (n = 5, 11, and 5, respectively), or an equal volume of PBS (PBS) vehicle (n = 17), immediately before and again immediately after hypoxia exposure. This initial dose range, derived from previous reports, was chosen to evaluate both safety and efficacy. In the subsequent two experiments, with 2.75-h hypoxia duration, P7 rats received pentoxifylline 40 mg/kg/dose (n = 13), or an equal volume of vehicle (PBS, n = 10), immediately before and again immediately after hypoxia exposure. In the prophylactic treatment regimens, we administered two doses of pentoxifylline, because the drug has a half-life of ≤1 h in rats (21, 38). By giving a dose immediately before the onset of hypoxia-ischemia and again immediately after the end of hypoxia exposure, we expected to have the drug in the circulation and in the CNS both during the episode of cerebral ischemia and during the early postischemic reperfusion phase, which begins immediately after the end of the hypoxia exposure. To test the efficacy of posthypoxic-ischemic treatment, we administered pentoxifylline to the animals (40 mg/kg/dose, n = 8) or vehicle (n = 10) immediately after the end of hypoxia exposure (2.75 h), and again 3 h later.

Neuropathology.

On P12, when brains were removed, ipsilateral liquefactive cerebral infarction was evaluated by an observer unaware of treatment identity (J.D.E.B.). In this model, marked hemispheric atrophy evolves rapidly and is apparent on gross inspection 5 d later (39). All brains were frozen under powered dry ice, and 20-μm coronal frozen brain sections, postfixed over paraformaldehyde vapors, were stained with cresyl violet for assessment of histopathology and morphometry.

Quantitative analysis of injury.

For initial quantification of the severity of brain injury on P12, bilateral hemispheric cross-sectional areas were measured from coronal sections at the level of the anterior commissure and mid-dorsal hippocampus (Fig. 1), using a computerized video camera-based image analysis system (with NIH Image software). A more detailed additional quantification of injury was performed on animals that had undergone posthypoxic-ischemic rescue treatment, to determine whether neuroprotection was regionally selective. For the latter analysis, bilateral striatal, neocortical, and dorsal hippocampal areas were measured in regularly spaced coronal sections, beginning at the level of the anterior genu of the corpus callosum and continuing to the level of the posterior genu of the corpus callosum. Unilateral cerebral atrophy evolves after unilateral cerebral hypoxia-ischemia in immature rats: thus, inter-hemispheric differences in areas or volumes provide a sensitive measure for quantification of injury or neuroprotection (40, 41).

Figure 1
figure 1

Neuroprotective efficacy of pentoxifylline administration. These four cresyl violet-stained sections are from two concurrently lesioned, littermate P12 rats, evaluated 5 d after right carotid ligation, followed by 2.75 h in 8% O2. A and B are from a vehicle-treated control. C and D are from a rat treated with pentoxifylline (40 mg/kg/dose) immediately before and again after hypoxia. In A and B, note right cortical infarction (arrowhead), striatal atrophy (asterisk), and hippocampal atrophy and pyramidal cell loss (arrow). In C and D, note preservation of tissue integrity in the right cerebral hemisphere. cort = cortex;str = striatum;hip = hippocampus (scale bar = 1 mm).

Full size image

Data analysis.

Microcomputer-based statistical programs [Statview (ABACUS, Berkeley, CA) and Systat (Systat Inc., Evanston, IL)] were used. Mean percent right-sided damage, compared with intact left side {i.e., mean % left-right (L-R) difference in hemispheric or regional areas or volumes [100 · (L-R)/L]} was compared among treatment groups and control rats by parametric and nonparametric tests. Hemisphere area measurements from pretreated animals exposed to either 3.25 or 2.75 h of hypoxia were analyzed separately. As a measure of neuroprotective efficacy, percent protection by pentoxifylline was calculated, based on comparison of the severity of injury in drug- and vehicle-treated animals, using the formula {100*[1-(% damage, drug-treated/% damage, vehicle-treated)]}. As a measure of the range of neuroprotection values observed, the SEM for percent protection was calculated as [100*(SEM, drug-treated/% damage vehicle-treated)], as previously described (9).

RESULTS

Right carotid ligation followed by 2.75–3.25 h of exposure to 8% oxygen resulted in a 75% incidence of ipsilateral liquefactive cerebral infarction in vehicle-treated controls. Histopathologic findings ipsilateral to carotid ligation included pallor, atrophy, and tissue loss in striatum, hippocampus, cortex, and thalamus (Fig. 1). Our first three experiments encompassed a broad dosage range, for evaluation of both safety and efficacy. Pretreatment with 150 mg/kg pentoxifylline resulted in 100% mortality; treatment with 75 mg/kg/dose resulted in approximately 20% mortality but marked neuroprotection in survivors, and there was similar mortality but no apparent neuroprotection at 25 mg/kg/dose (Table 1). These results influenced our subsequent study design, in which the intermediate dose level of 40 mg/kg was added, to evaluate dose-dependence of neuroprotective efficacy. Pretreatment with pentoxifylline (40 or 75 mg/kg/dose) resulted in a decreased incidence of liquefactive cerebral infarction on P12, ipsilateral to carotid ligation (p < 0.001, χ2 trend test) (Table 1 and Fig. 1). Quantification of hemispheric areas at two anatomic levels in rats that had received two serial doses of pentoxifylline (25, 40, or 75 mg/kg/dose immediately before and again immediately after hypoxia) and in control rats confirmed the results of the initial evaluation of tissue infarction (Table 2).

Table 1 Reduced incidence of hypoxia-ischemia-induced cerebral infarction by pretreatment with pentoxifylline in P7 rats* * P7 rats underwent right carotid ligation and received two doses of pentoxifylline, or an equal volume of PBS, by intraperitoneal injection, immediately before exposure to 8% O2, and again after the end of exposure to hypoxia. There were 27 rats in the PBS-treated group. In the pentoxifylline-treated group there were 28 rats (5 rats received two 25-mg doses, 13 received two 40-mg doses, and 11 received two 75-mg doses). † Cerebral infarction was evaluated by inspection of the brain on P12, 5 days after lesioning. ‡p < 0.0001, χ2 trend test for effect of pentoxifylline dose on infarction. §p < 0.005, χ2 trend test for effect of pentoxifylline dose on mortality. ∥ 5/5 animals receiving 150 mg/kg died during hypoxia.
Full size table
Table 2 Attenuation of hypoxia-ischemia-induced brain injury by pretreatment with pentoxifylline in P7 rats * Mean percent reduction in right cerebral hemisphere areas, ipsilateral to carotid ligation, in comparison to intact left-sided regions {left-right difference in areas = [100*(L-R)/L]}. Bilateral hemispheric cross-sectional areas on P12 were measured by image analysis at the level of the anterior commissure (anterior) and mid-dorsal hippocampus (posterior). † P7 rats underwent right common carotid artery ligation, and, beginning 1 h later, received intraperitoneal injections of pentoxifylline (25 or 75 mg/kg/dose) immediately before and again immediately after exposure to 8% O2 for 3.25 h. ‡ P7 rats underwent right common carotid artery ligation, and, beginning 1 h later, received intraperitoneal injections of pentoxifylline (40 mg/kg/dose) immediately before and again immediately after exposure to 8% O2 for 2.75 h. §p < 0.02, ANOVA;p < 0.05, post-hoc Fisher LSD test, compared to concurrent controls also exposed to 3.25 h hypoxia. ∥p < 0.005, t test, compared with concurrent controls also exposed to 2.75 h hypoxia.
Full size table

Although hypoxic-ischemic conditions were different in the first three and the subsequent two pretreatment experiments, in both cases, pentoxifylline pretreatment markedly attenuated the resulting brain damage. As a measure of neuroprotective efficacy of pentoxifylline pretreatment, we calculated the percent protection, based on inter-hemisphere area differences. Percent protection (mean ± SEM) by pretreatment with pentoxifylline 40 mg/kg/dose ranged from 72 ± 12% (anterior hemisphere) to 77 ± 8% (posterior hemisphere). Percent protection by pentoxifylline 75 mg/kg/dose ranged from 80 ± 10% (anterior hemisphere) to 84 ± 10% (posterior hemisphere). By posthoc analysis, pentoxifylline 25 mg/kg/dose did not attenuate hypoxic-ischemic brain injury.

Posthypoxic-ischemic rescue treatment with pentoxifylline (40 mg/kg/dose) did not result in a decrease in the incidence of cerebral infarction on P12 (Table 3). Furthermore, analysis of bilateral hemisphere areas indicated no difference in mean hemisphere areas between animals treated with pentoxifylline posthypoxia-ischemia and concurrent controls (Table 4). However, review of the histopathology suggested that cortical infarctions were less extensive in pentoxifylline-treated rats than in controls. To confirm this impression, we performed a morphometric analysis of bilateral cortical, striatal, and hippocampal volumes in animals treated posthypoxia-ischemia and in concurrent controls. This analysis of regional volumes indicated that neocortical damage was less severe in animals treated with pentoxifylline beginning after hypoxia-ischemia than it was in concurrent controls (Table 5). The percent protection (mean ± SEM) of cortical tissue by posthypoxic-ischemic rescue treatment with pentoxifylline was 54 ± 17.9%. In contrast, the severity of striatal and hippocampal damage was not attenuated by pentoxifylline administration post-insult. There were no deaths among the rats that received posthypoxic-ischemic pentoxifylline treatment, or among their littermate controls.

Table 3 Effect of post-hypoxic-ischemic administration of pentoxifylline on incidence of cerebral infarction in P7 rats * P7 rats underwent right carotid ligation, followed 1 h later by 2.75 h exposure to 8% O2, and then received two doses of pentoxifylline 40 mg/kg/dose, or an equal volume of PBS, by intraperitoneal injection, immediately after the end of exposure to 8% O2, and again 3 h after the end of exposure to hypoxia. † Cerebral infarction, defined as liquefaction, pallor or atrophy, was evaluated by inspection of the brain on P12, 5 days after lesioning.
Full size table
Table 4 Effect of post-hypoxic-ischemic administration of pentoxifylline on severity of hemisphere damage in P7 rats * Mean percent reduction in right cerebral hemisphere areas, ipsilateral to carotid ligation, in comparison to intact left-sided regions {left-right difference in areas = [100*(L-R)/L]}. Bilateral hemispheric cross-sectional areas on P12 were measured by image analysis at the level of the anterior commissure (anterior) and mid-dorsal hippocampus (posterior). † P7 rats underwent right carotid ligation, followed 1 h later by 2.75-h exposure to 8% O2, and then received two doses of pentoxifylline, or an equal volume of PBS, by intraperitoneal injection, immediately after the end of exposure to 8% O2, and again 3 h after the end of exposure to hypoxia.
Full size table
Table 5 Effect of post-hypoxic-ischemic administration of pentoxifylline on severity of cortical, striatal and hippocampal damage in P7 rats * Mean percent reduction in right cerebral hemisphere volumes, ipsilateral to carotid ligation, in comparison to intact left sided regions {left-right difference in volumes = [100*(L-R)/L]}. Bilateral regional cross-sectional areas on P12 were measured by image analysis of regularly spaced coronal sections from the level of the anterior genu of the corpus callosum to the posterior genu of the corpus callosum). † P7 rats underwent right carotid ligation, followed 1 h later by 2.75 h exposure to 8% O2, and then received two doses of pentoxifylline, or an equal volume of PBS, by intraperitoneal injection, immediately after the end of exposure to 8% O2, and again 3 h after the end of exposure to hypoxia. ‡p < 0.05, Mann-Whitney U test, compared to concurrent controls also exposed to 2.75 h hypoxia.
Full size table

Intrahypoxic skin temperature and posthypoxic esophageal temperature in rats from four experiments did not differ among groups.

DISCUSSION

Treatment of immature rats with the alkylxanthine phosphodiesterase inhibitor pentoxifylline before induction of cerebral hypoxia-ischemia markedly attenuated subsequent ischemic forebrain injury. In contrast, when treatment with pentoxifylline was not initiated until after the end of the hypoxic-ischemic insult, the attenuation of damage was modest and was limited to the neocortex. This probably reflects progression of the injury cascade to the point of irreversibility before drug levels reached an effective tissue concentration. Our results add to the growing body of evidence that anti-inflammatory strategies can substantially attenuate neonatal hypoxic-ischemic brain injury. Other effective anti-inflammatory neuroprotective strategies for neonatal cerebral hypoxia-ischemia include pretreatment with dexamethasone (32), pre- or posthypoxic-ischemic treatment with the platelet-activating factor receptor antagonist BN 52021 (9), pre- or posthypoxic-ischemic treatment with IL-1 receptor antagonist (8, 42), and immune-mediated iatrogenic neutropenia before induction of hypoxia-ischemia (3). A comparison of the reported neuroprotective efficacy of these different anti-inflammatory strategies is presented in Table 6. The efficacy of pentoxifylline pretreatment is similar to some other pretreatment regimens (e.g., platelet-activating factor antagonist, IL-1 receptor antagonist, neutropenia) but is less effective than dexamethasone pretreatment. The neuroprotective efficacy of posthypoxic-ischemic pentoxifylline treatment is limited, compared with platelet-activating factor antagonist or IL-1 receptor antagonist administration.

Table 6 Neuroprotective efficacy of anti-inflammatory therapies in neonatal cerebral hypoxia-ischemia in rats * Citations are listed in descending order of neuroprotective efficacy. † Percent protection was either reported by authors in the cited report, or was calculated, based on comparison of the cited severity of injury in drug- and vehicle-treated animals, using the formula {100*[1-(% damage drug-treated/% damage vehicle-treated)]}. ‡ i.c.v., intracerebroventricular.
Full size table

Pentoxifylline is a cAMP phosphodiesterase inhibitor. Although other phosphodiesterase inhibitors (e.g., rolipram) have been reported to share some of the anti-inflammatory activities of pentoxifylline, we evaluated pentoxifylline both because of the large body of knowledge about its activities in vitro and in vivo, and because it is already approved for clinical use. Administration of pentoxifylline leads to increased intracellular cAMP accumulation in several cell types, including mononuclear phagocytes (43), microglia (14), neutrophils (43), vascular smooth muscle, and endothelium (44). Increased intracellular cAMP affects gene transcription through several transcription factors, such as cAMP response element-binding protein (45). Multiple complementary anti-inflammatory effects of pentoxifylline could be responsible for its neuroprotective efficacy, including inhibition of neutrophil (18) or monocyte/microglial (14) activation, attenuation of inflammatory mediator production (e.g., platelet-activating factor (17) or TNF-α (13), and prevention of endothelial-leukocyte adhesion (46).

Other effects of pentoxifylline, independent of its anti-inflammatory actions, could also contribute to its neuroprotective efficacy. Cerebrovascular resistance is regulated by cAMP phosphodiesterases (47). Pentoxifylline reportedly increases cerebral blood flow in humans who have cerebrovascular disease (48). Yet, under controlled laboratory conditions in mature dogs or rats undergoing global cerebral ischemia or ischemia-reperfusion, there is no increase in intra-ischemic or post-reperfusion regional cerebral blood flow with pentoxifylline treatment (27, 49). Despite these reports, it is possible that pentoxifylline might have the opposite effect on intra-ischemic or post-reperfusion cerebral blood flow in the immature brain. We have not yet evaluated the effect of pentoxifylline administration on either intra- or posthypoxic-ischemic cerebral blood flow in P7 rats; thus we cannot rule out the possibility that pentoxifylline's beneficial effect in P7 rats is due to increased intrahypoxic cerebral blood flow. Pentoxifylline could also have beneficial CNS parenchymal effects independent of effects on cerebral blood flow, perhaps mediated by preservation of CNS cAMP levels (50).

Variable neuroprotective efficacy of pentoxifylline therapy in mature animal models of cerebral ischemia has been reported (25–29). Possible explanations for inconsistencies in neuroprotective efficacy among studies include difference in drug dosage and timing, differences in injury models, and differences in outcome measures. This neonatal stroke model includes both a period of focal cerebral hypoxia-ischemia and a subsequent period of normoxia with reperfusion (51). Anti-inflammatory strategies, such as pentoxifylline treatment, may be more neuroprotective in the setting of cerebral ischemia with reperfusion than in permanent cerebral ischemia (52).

The cause of the high intra-hypoxic mortality, noted at 150 mg/kg/dose, is unknown. In mice, single pentoxifylline doses of 200 mg/kg are well tolerated, whereas apparent neurotoxicity and death follow a 400 mg/kg dose (37). P7 rats that died became apneic, with no visible seizures preceding apnea. Assisted ventilation to prevent apnea is not feasible in P7 rats because of their small size. Alternatively, increased mortality could have resulted from hypotension; blood pressure cannot be monitored because of their small size. Similar mortality was noted in experiments to evaluate the effect of pentoxifylline on N-methyl-D-aspartate neurotoxicity in P7 rats (F. S. Silverstein, personal communication); thus these detrimental effects are unlikely to be directly related to an interaction with systemic hypoxemia. Clinical trials of pentoxifylline as an adjunctive therapy in critically ill neonates and adults have not raised concerns about adverse side effects of pentoxifylline. Results of recent clinical trials suggest that pentoxifylline treatment may decrease mortality and morbidity associated with neonatal bacterial sepsis (53) and may attenuate multi-organ failure after cardiothoracic surgery in adults (54). Thus pentoxifylline might be suitable for clinical trial as a human neuroprotective agent in carefully selected patients at high risk of cerebral hypoxic or ischemic injury.

In conclusion, our results add to the growing body of evidence supporting the hypothesis that inflammatory mediators play a pathogenetic role in hypoxic-ischemic injury to the immature brain. Several strategies to interrupt the inflammatory cascade have proven effective in decreasing the severity of neonatal hypoxic-ischemic brain injury in experimental animals. These therapies have not yet reached clinical practice. Pretreatment with pentoxifylline may offer an effective means to decrease the incidence and severity of hypoxic-ischemic injury to the immature brain. Prophylactic treatment may be feasible in infants identified as having a high risk of cerebral hypoxia-ischemia—for example, infants undergoing heart-lung bypass for surgical repair of congenital cardiac defects or extracorporeal membrane oxygenation.