Intra-arterial administration of recombinant tissue-type plasminogen activator (rt-PA) causes more intracranial bleeding than does intravenous rt-PA in a transient rat middle cerebral artery occlusion model

Upon arrival, the rats were assigned a number by the animal facility staff and housed individually. They were allowed at least one week to acclimate. The rats were on a 12 hour light/dark diurnal cycle and food and water were provided ad libitum. To conserve test article, rats were assigned to experimental groups in tandem (2 rats to the same group/experimental day) as outlined in Table 1.

All rats were fasted overnight to provide a low consistent plasma glucose concentration before surgery. The rats were anesthetized with isoflurane (5% in O₂) in an induction chamber and orotracheally intubated. For pain relief, all rats received 0.03 mg/kg buprenorphine, subQ (PharmaForce, Inc., Columbus, Ohio). Anesthesia was maintained by spontaneous ventilation of isoflurane (1.5-3% in O₂). The ventral tail artery was cannulated (RPT-037, RenaPulse tubing, Braintree Scientific, Inc., Braintree MA or PE50 tubing) by cutdown. The arterial line was used to continuously monitor blood pressure and heart rate (Ponemah Physiological Monitoring System, DSI, Cleveland, Ohio) and to obtain arterial blood for blood gas determinations (i-Stat, Heska AG, Switzerland). Following arterial catheterization, the rats were artificially ventilated to maintain PaCO₂ between 35 and 45 mmHg. Continued isoflurane anesthesia (1-1.5%) was driven by compressed air supplemented with O₂ to maintain PaO₂ between 95 to 170 mmHg. Body temperature was maintained at or slightly above 37°C using a heat lamp in a feedback circuit with a rectal temperature probe (TCAT-2 Temperature Controller, Physitemp Instruments, Inc., Clifton, NJ). Just before MCAo, reflow, and during the blood flow studies, the isoflurane was reduced to approximately 1% (from 1.5%) to allow for normalization of blood pressure while maintaining an adequate level of anesthesia.

Rats were placed in dorsal recumbency and the common carotid (CCA), external carotid (ECA), internal carotid (ICA), occipital and pterygopalatine arteries were isolated by cutdown. A perivascular blood flow probe (Transonic Systems Inc., Ithaca, NY) was placed around the CCA for continuous blood flow monitoring. Sequentially, the occipital, the external carotid, and the pterygopalatine arteries were occluded leaving the blood flow to the brain via the extra-cranial internal carotid artery (EC-ICA).

Six (6) hours of transient MCAo was induced in the SHR using the snare ligature model first described in mice [14], adapted to the rat [15] and illustrated in Figure 1. The head was immobilized in a specially designed head holder and a skin incision was made between the eye and the external auditory canal. Using a surgical microscope (Zeiss OPMI-6C, Prescott's Inc., Monument, CO), the temporalis muscle was bisected and retracted and a small craniotomy was performed using a dental drill anteromedial to the zygomatic arch-squamosal bone junction. The dura was opened and the MCA was isolated from the arachnoid and pia by blunt dissection proximal to the inferior cerebral vein. The snare ligature was constructed as illustrated in Figure 1A using a strand of 10-0 nylon suture, a piece of 6-0 prolene monofilament and silastic tubing (ID 0.76 mm × OD 1.65 mm × ~1 mm h; Dow Corning Corporation). The MCA was gently pulled into the lumen of the silastic tubing resulting in occlusion of the artery (Figure 1B). MCAo was visually confirmed through the surgical microscope (40 × magnification). A piece of Gelfoam^® was placed in the cranial deficit and the surgical site was closed in two layers. The rat was recovered from anesthesia, extubated and returned to a clean cage. The normal bedding in the cage was replaced with an absorbent cage pad (Recovery Pad ™, Bed O' Cobs/The Andersons, Distributor: Granville Milling Co., Creedmoor, NC). Between the MCAo surgery and the removal of the snare ligature, the rats had free access to gel style food (Hydro Gel and Diet Gel-Recovery; H2O, Portland, Maine). The location of the snare ligature, distal to the lenticulostrate arteries but proximal to the inferior cerebral vein, resulted in nearly pure cortical infarction (Additional file 1, Figure S1).

We chose the rat snare ligature model of MCAo to strictly control the ischemic duration which is not possible in a thromboembolic model. Further, in a thromboembolic model, control groups such as saline and vehicle become de facto permanent occlusion animals and thus a proper control for specific ischemic durations with reflow is not practicable. In addition, the snare ligature model allows for the initiation of local IA dosing to immediately precede reflow. Thus, the drug would be in high concentration during the first pass of blood into the ischemic tissue thereby better mimicking human thrombolytic treatment by IA administration. This is not possible using an intraluminal model where a significant delay between reflow and IA dosing would occur due to exchanging the occluding filament for the dosing catheter. This is in addition to other complications associated with the intraluminal model such as pre-mature reflow [16], ischemia to the hypothalamus resulting in hyperthermia [17], distension of the MCA (personal observations) and the possibility of filament related subarachnoid hemorrhage [16].

Lyophilized rt-PA (Alteplase, recombinant, Genentech, South San Francisco, CA, USA) was solubilized at an initial concentration of 5 mg/mL using the provided diluent (Water for Injection, USP (WFI) and further diluted as needed to 1 mg/mL using saline for injection. Unused portions (1 and 5 mg/mL) were divided into 1-3 mL aliquots, snap frozen and stored at -80°C [18].

Concentrations of rt-PA higher than 5 mg/mL were not feasible because of osmolality (the 5 mg/mL solution was 793 mOsmol/L) and solubility limitations (particularly important for IA administration). The rate of test article infusion IA or IV was 0.033-0.038 mL/min. The experimental groups and dosing details are provided in Table 1. A timeline of the experiment is provided in Figure 2.

Twenty four hours after onset of MCAo, the rats were deeply anesthetized with isoflurane (5% in O₂) and the brain was perfused in situ for approximately 90 seconds with heparinized saline (10 U/mL). The brain was removed, placed in ice cold saline and examined for gross superficial hemorrhage. Digital photographs of the brain surface were obtained.

Sequential coronal sections were obtained every 2 mm throughout the neocortex (8-9 sections/brain). The sections were stained with 1-2% 2, 3, 5 triphenyl tetrazolium chloride (TTC) [22] at 37°C in the dark to effect (~15 min incubation time). The sections were fixed in 10% buffered formalin for 24 hours. Each fixed brain section was digitally photographed with a ruled standard.

The digital photographs of the fixed TTC stained brain sections were imported into an image analysis program (Image-Pro Plus v4.5, Media Cybernetics, Inc., Bethesda, MD) for infarct volume measurement. The infarct volume, reported in mm³, was determined by the indirect method.

Two complete 10 μm sections, spaced 250 μm apart, were obtained from each original 2 mm TTC stained section and processed for H&E histopathology (16-18; 10 μm sections/rat brain). Each section was evaluated for Bleeding Score, according to the following definitions (Examples are shown in Figure 4):

5.5 = Hemorrhage to non ischemic brain tissue

The overall score for the rat brain was the highest score of the evaluated sections.

The rats were assessed for neurological function twice during the study, once prior to the induction of anesthesia to place the catheter for test article administration and again just before euthanasia. The first assessment was to screen for stroke presence; any rat not displaying a score of at least 2 was eliminated from the study. The second assessment was performed just before euthanasia to determine the final functional recovery of the rat; only the second score was analyzed as an experimental variable.

The rats were assessed for neurological function using a Modified Bederson Score [23] in an unblinded manner by RCC and GMT with the following definitions:

Score 0: No apparent neurological deficits

Score 1: Body torsion present

Score 2: Body torsion with right side weakness

Score 3: Body torsion, right side weakness with circling behavior

Score 4: Unresponsiveness - euthanasia < 3 hours after dosing

Score 5. Seizure Activity - immediate euthanasia or death before scheduled termination due to ICH.

Animals were excluded from the study if one of the following occurred: breakage of the MCA during occlusion or recanalization; breakage of a side branch of the MCA causing bleeding at the surgical site; Modified Bederson Score less than 2 prior to start of the reflow surgery; reflow cannot be visually confirmed; excessive bruising of the brain during occlusion surgery; observation of air bubbles or particles in the ICA catheter during infusions; surgical complications requiring euthanasia, or a plasma glucose concentration > 200 mg/dL prior to MCAo.

Statistical analysis of the infarct volume and physiological variables between the experimental groups was accomplished by ANOVA followed by Tukey-Kramer HSD multiple comparison test (JMP statistical software, SAS Institute, Inc., Cary, NC).

Statistical analysis of the Bleeding Score and the Modified Bederson Score data was performed using the Kruskal-Wallis one way ANOVA test for non-parametric data followed by Newman-Keuls multiple comparison test using the GBstat statistical package (Dynamic Microsystem, Inc., Silver Spring, MD).

Analysis of covariance (ANCOVA) with the dependent variables of infarct volume, Bleeding Score and Modified Bederson Score, respectively, with the independent variables of dose and route was performed using SAS^® software version 9.2 (SAS Institute, Inc. Cary, NC).

To determine the number of rats in each group, a power analysis of the vehicle group from a contemporary pilot study was performed using infarct volume as the test variable. To observe a statistically significant difference of 30% between the means of the experimental groups at an α level of 0.05 and a β level of 0.8, an n of at least 5 animals would be required. This is consistent with Brint et al [24] and with our past experience using the SHR in stroke studies. All of the data are presented as mean ± SEM.

The infarct volume and Bleeding Score were determined in a blinded fashion by CPT in the laboratory of JCL (Case Western Reserve University, Cleveland, Ohio).

MCAo surgery was performed on 56 rats, of which 42 were assigned to the different experimental groups. Exclusions were for breakage of the MCA or a side branch (3 rats), bleeding complications prior to craniotomy (3 rats), bruising of brain tissue (2 rats), occlusion device malfunction (2 rats), clot in the ICA just distal to the pterygopalatine artery bifurcation (1 rat) and technical difficulties (3 rats).

All of the physiological variables were within normal range throughout the MCAo and reflow surgeries (Table 3). In all groups, the plasma glucose values prior to MCAo were significantly lower than prior to reflow; the former values reflect fasting the rats overnight and the latter values indicate that the rats were eating the gel style food between the two surgeries.

The Bleeding Score for the experimental groups are shown in Figure 5. Photographs of the gross brain and a TTC stained section of the brain at the level of the basal ganglia and in some cases, at the level of the anterior hippocampus are presented for the animals treated with saline (Figure 6), 10 or 30 mg/kg rt-PA IA (Figures 7 and 8, respectively), and 10 or 30 mg/kg IV (Figures 9 and 10, respectively).

The Bleeding Severity Score of the saline group (1.7 ± 0.5) and the groups treated IA with 1, 5 and 10 mg/kg rtPA were low (2.2 ± 0.4, 2.4 ± 0.3 and 2.5 ± 0.2, respectively) and similar (Figures 5, 6 and 7). Markedly more severe bleeding was observed with 30 mg/kg rtPA delivered IA (4.9 ± 0.4) compared with all other groups except 5 mg/kg IA rt-PA. One rat treated with 30 mg/kg rt-PA IA had a massive space-occupying hemorrhage and did not survive to the scheduled end of the experiment (# 61715, Figure 8). Three rats in this group had left brain stem hemorrhage (# 61713, 61723, and 61721 Figure 8, left panels, respectively; thin arrows) resulting in hypothalamic damage (Figure 8, right panels, respectively; thick arrows). Two of these rats developed seizure activity after dosing that necessitated euthanasia.

An increase in bleeding score occurred with increased doses of rtPA IV (Figure 5). The 30 mg/kg IV rtPA dose showed a higher Bleeding Score (3.1 ± 0.7) as compared to the 10 mg/kg rtPA IV dose (1.0 ± 0.3), and space-occupying hemorrhage occurred in 2 of 5 rats (61366 and 61368, Figure 10; right panel, respectively; thin arrows). Four of 5 rats were euthanized approximately 3 hours after terminating anesthesia after agent administration because of unresponsiveness. None of the rats in this group had brain stem hemorrhage, seizure activity or hypothalamic damage.

The 10 mg/kg IA rt-PA dose caused significantly more bleeding than did 10 mg/kg IV (Figures 5, 7 and 9). Indeed, rats that received only 1 or 5 mg/kg IA rt-PA had significantly more bleeding than rats receiving 10 mg/kg rt-PA IV. Similarly, 30 mg/kg IA rt-PA caused significantly more bleeding than did 30 mg/kg IV (Figures 5, 8 and 10). The Bleeding Score in rats receiving 30 mg/kg rt-PA IV was not significantly different that in rats receiving 10 mg/kg rt-PA IA. Multivariate analysis revealed a statistically significant effect of dose and route on Bleeding Score (Table 4).

The infarct volume after 6 hours of ligature ischemia followed by 18 hours reflow was not significantly different for all treatments (Figure 11), although there was a trend towards a larger infarct volume using the highest dose (30 mg/kg) of rt-PA either IA or IV. Saline treatment (252 ± 41 mm³) and treatment with 1 mg/kg IA (264 ± 19 mm³), 5 mg/kg IA (247 ± 23 mm³) and 10 mg/kg IA (240 ± 8 mm³) had very similar infarct volumes, whereas infarct volume after treatment with 30 mg/kg rtPA IA tended to be larger (306 ± 35 mm³). Infarct volumes after IV rtPA at 10 and 30 mg/kg (298 ± 37 and 338 ± 48 mm³, respectively) tended to be larger than saline control. Multivariate analysis showed no statically significant effects of dose or route on infarct volume (Table 4).

The Modified Bederson Score, reflecting neurological function 24 hours following MCAO (Figure 12) was nearly identical in rats treated with saline (2.4 ± 0.24), with IA rt-PA at 1 mg/kg (2.25 ± 0.22), 5 mg/kg (2.29 ± 0.22) and 10 mg/kg (2.38 ± 0.24) and with IV rtPA at 10 mg/kg (2.00 ± 0.00). In general, the rats were alert and responsive to external stimuli and were eating and drinking. They displayed body torsion and right side weakness but, in most cases, circling was not present. In contrast, rats exposed to 30 mg/kg rt-PA IA (4.30 ± 0.66) and 30 mg/kg rt-PA IV (3.60 ± 0.40) showed more severe neurological dysfunction. Multivariate analysis showed a significant effect of dose and route on the Modified Bederson Score (Table 4).