The KCa3.1 Blocker TRAM-34 Reduces Infarction and Neurological Deficit in a Rat Model of Ischemia/Reperfusion Stroke

Abstract

Microglia and brain infiltrating macrophages significantly contribute to the secondary inflammatory damage in the wake of ischemic stroke. Here, we investigated whether inhibition of KCa3.1 (IKCa1/KCNN4), a calcium-activated K⁺ channel that is involved in microglia and macrophage activation and expression of which increases on microglia in the infarcted area, has beneficial effects in a rat model of ischemic stroke. Using an HPLC/MS assay, we first confirmed that our small molecule KCa3.1 blocker TRAM-34 effectively penetrates into the brain and achieves micromolar plasma and brain concentrations after intraperitoneal injection. Then, we subjected male Wistar rats to 90 minutes of middle cerebral artery occlusion (MCAO) and administered either vehicle or TRAM-34 (10 or 40 mg/kg intraperitoneally twice daily) for 7 days starting 12 hours after reperfusion. Both compound doses reduced infarct area by ∼50% as determined by hematoxylin & eosin staining on day 7 and the higher dose also significantly improved neurological deficit. We further observed a significant reduction in ED1⁺-activated microglia and TUNEL-positive neurons as well as increases in NeuN⁺ neurons in the infarcted hemisphere. Our findings suggest that KCa3.1 blockade constitutes an attractive approach for the treatment of ischemic stroke because it is still effective when initiated 12 hours after the insult.

Introduction

In addition to directly causing neuronal damage, focal ischemic stroke elicits a strong and long-lasting inflammatory response (Weinstein et al, 2010; Yenari et al, 2010). Activated by multiple stimuli, which include hypoxia, neuronal debris, ATP and glutamate, microglia retract their branched processes, round up and transform into ‘reactive’ microglia. Partial breakdown of the blood–brain barrier additionally promotes the infiltration of macrophages, neutrophils, and activated T cells from the blood. In both rodent models of cerebral ischemia and in histopathological studies on human postmortem brain sections activated microglia/macrophages are abundant in the infarcted area and the peri-infarct zone 18 to 96 hours after an ischemic insult (Beschorner et al, 2002; Campanella et al, 2002; Price et al, 2006), and are still present in chronic cystic stages months after a stroke (Beschorner et al, 2002). More recent positron emission tomography imaging in ischemic stroke patients demonstrated microglia activation in the peri-infarct zone on a slightly more delayed time scale: starting at 72 hours and lasting for at least 4 weeks (Price et al, 2006). While microglia can of course exert neuroprotective functions by releasing neurotrophic growth factors such as brain-derived neuroprotective factor or phagocytosing debris and potentially even invading neutrophils (Denes et al, 2007), activated microglia/macrophages are also the main source of inflammatory cytokines such as IL-1β and TNF-α, reactive oxygen species, nitric oxide, and cyclooxygenase-2 reaction products (del Zoppo et al, 2000).

Since interventions aiming at neuroprotection in the acute phase of ischemic stroke have largely failed in the clinic due to the fact that many stroke patients only reach medical attention several hours after the insult, more recent attempts at therapeutic intervention have refocused on inflammation because of its delayed onset. Based on studies demonstrating that immunosuppressive strategies such as general immunosuppression with cyclosporine, neutralization of IL-β or TNF-α, inhibition of microglia activation with minocycline (Yrjanheikki et al, 1998), depletion of γδT cells, or blockade of T-cell infiltration with FTY720 (Shichita et al, 2009) reduce infarct size in rodent models of ischemic stroke, while depletion of regulatory T cells profoundly increases delayed brain damage by augmenting postischemic activation of microglia and invading inflammatory cells (Liesz et al, 2009), several anti-inflammatory or immunosuppressive therapies have entered clinical trials. The most encouraging data suggesting that reducing inflammation can indeed improve stroke outcome, currently come from a phase-2 study with the IL-1 receptor antagonist anakinra (Emsley et al, 2005), which is FDA approved for the treatment of rheumatoid arthritis, and from an open-label study with the microglia inhibitor minocycline (Lampl et al, 2007).

Another attractive pharmacological target for inhibiting brain inflammation after ischemic stroke is the intermediate-conductance calcium-activated K⁺ channel KCa3.1 for which our laboratory designed a small molecule inhibitor, TRAM-34, that blocks this channel with an IC₅₀ of 20 nmol/L and exhibits 200- to 1,500-fold selectivity over other K⁺ channels (Wulff et al, 2000). KCa3.1 is expressed on proliferating fibroblasts (Pena et al, 2000), on dedifferentiated vascular smooth muscle cells (Köhler et al, 2003), and on immune cells including microglia and macrophages (Hanley et al, 2004; Kaushal et al, 2007; Khanna et al, 2001), activated CCR7⁺ T cells and IgD⁺ B cells (Ghanshani et al, 2000; Wulff et al, 2004). In all these cells, KCa3.1 is part of signaling cascades that involve relatively global and prolonged calcium rises during cellular proliferation, cytokine secretion, and volume regulation (see Wulff and Castle, 2010 for a recent review). KCa3.1 channels are voltage independent and only require a small increase in intracellular calcium to open and then maintain a negative membrane potential through K⁺ efflux. KCa3.1 channels, thus, provide the driving force for store-operated inward-rectifier calcium channels like CRAC (calcium release-activated Ca²⁺ channel) or transient receptor potential channels like TRPC1 (Wulff and Castle, 2010).

Particularly in microglia, KCa3.1 has been shown to be involved in respiratory bursting (Khanna et al, 2001), migration (Schilling et al, 2004), proliferation (Maezawa et al, 2010), and lipopolysaccharide or amyloid-β oligomer induced nitric oxide production (Kaushal et al, 2007; Maezawa et al, 2010), as well as in microglia-mediated neuronal killing in cultures (Kaushal et al, 2007) and organotypic hippocampal slices (Maezawa et al, 2010), suggesting that KCa3.1 suppression might be useful for reducing microglia activity in stroke, traumatic brain injury, multiple sclerosis, and Alzheimer's disease. Using intraocular injection of our KCa3.1 blocker TRAM-34, Kaushal et al (2007) tested this hypothesis by demonstrating that KCa3.1 blockade reduced retinal ganglion cell degeneration after optic nerve transection in rats, while scientists at Schering showed that TRAM-34 treats experimental autoimmune encephalomyelitis in mice (Reich et al, 2005). Interestingly, in the nerve transection study KCa3.1 blockade did not prevent microglia from aligning with damaged axons and from phagocytosing damaged neurons but increased the number of surviving retinal ganglion cells presumably by reducing the production and secretion of neurotoxic molecules in the retina (Kaushal et al, 2007). This observation raises the exciting possibility that KCa3.1 blockade, which inhibits lipopolysaccharide-stimulated p38 mitogen-activated protein kinase activation but not nuclear factor κB activation in microglia, might preferentially target microglia activities that are involved in neuronal killing without affecting beneficial functions such as scavenging of debris (Kaushal et al, 2007). In addition, Mauler et al (2004) demonstrated that two structurally different KCa3.1 inhibitors, a triarylmethane and a cyclohexadiene, reduced infarct volume and brain edema after traumatic brain injury caused by acute subdural hematoma in rats. Based on these encouraging findings, we here tested whether pharmacological KCa3.1 blockade has beneficial effects in a rat model of ischemic stroke with reperfusion.

Materials and methods

Middle Cerebral Artery Occlusion With 7 Days of Reperfusion

This study was approved by the University of California, Davis, Animal Use and Care Committee and conducted in accordance with the guidelines of Animal Use and Care of the National Institutes of Health and the University of California, Davis for survival surgery in rodents. Adult male Wistar rats weighing 160 to 180 g were purchased from Charles River (Wilmington, MA, USA), acclimatized to the new vivarium for 5 to 7 days and used for the surgery when they weighed 200 to 230 g. Rats were anesthetized using box induction with 5% isoflurane and then maintained on 0.5 to 1.5% isoflurane in medical grade oxygen via a facemask. To assure consistent reduction of cerebral blood flow (CBF) throughout the procedure, we affixed a small hand-made adapter for the Laser Doppler probe (Moor Instruments, Wilmington, DE, USA) to the surface of the skull. The center of the adapter was 5 mm lateral to the central fusion line and 2.5 mm posterior to bregma. Instant adhesive and dental cement were applied to the base and around the edges of the small plastic adapter to hold the Doppler probe. The adapter with the attached probe remained in place throughout the MCAO surgery to confirm continuous occlusion and later the establishment of reperfusion. Focal cerebral ischemia was then induced by occlusion of the left middle cerebral artery (MCA) according to Zea Longa (Longa et al, 1989; O’Donnell et al, 2004). Briefly, the left common carotid artery was surgically exposed, the external carotid artery was ligated distally from the common carotid artery, and a silicone rubber-coated nylon monofilament with a tip diameter of 0.43±0.02 mm (Doccol Corp., Redlands, CA, USA) was inserted into the external carotid artery and advanced into the internal carotid artery to block the origin of the MCA (when maximum CBF reduction observed). The filament was kept in place for 90 minutes and then withdrawn and removed from the blood vessel to restore blood supply. Rats received TRAM-34 at 10 mg/kg, 40 mg/kg or vehicle (Miglyol 812 neutral oil at 1 μL/g) twice daily intraperitoneally for 7 days starting 12 hours after reperfusion. Neurological deficits were scored according to a 4-score test (Menzies et al, 1992) and a tactile and proprioceptive limb-placing test (De Ryck et al, 1989) as follows: (1) 4-score test (higher score for more severe neurological deficits): 0=no apparent deficit; 1=contralateral forelimb is consistently flexed during suspension by holding the tail; 2=decreasing grip ability on the contralateral forelimb while tail pulled; 3=spontaneous movement in all directions but circling to contralateral side when pulled by the tail; 4=spontaneous contralateral circling or depressed level of consciousness. (2) 14-score limb-placing test (lower score for more severe neurological deficits): proprioception, forward extension, lateral abduction, and adduction were tested with vision or tactile stimuli. For visual limb placing, rats were held and slowly moved forward or lateral toward the top of a table. Normal rats placed both forepaws on the tabletop. Tactile forward and lateral limb placing were tested by lightly contacting the table edge with the dorsal or lateral surface of a rat's paw while avoiding whisker contact and covering the eyes to avoid vision. For proprioceptive hindlimb placing, each rat was pushed along the edge of an elevated platform in order to test proprioceptive hindlimb adduction. The paw was pulled down and away from the platform edge, and the ability to retrieve and place the paw on the table surface upon sudden release was assessed. For each test, limb-placing scores were 0=no placing; 1=incomplete and delayed (>2 seconds) placing; or 2=immediate and complete placing. For each body side, the maximum summed visual limb-placing score was 4 and the maximum summed tactile and proprioceptive limb-placing score, including the platform test, was 10.

Pharmacokinetics, Brain Concentrations, and Plasma Protein Binding of TRAM-34

TRAM-34 was synthesized in our laboratory as previously described (Wulff et al, 2000) and its chemical identity and purity checked by ¹H NMR and high pressure liquid chromatography/mass spectrometry (HPLC/MS). For intravenous application, TRAM-34 was dissolved at 5 mg/mL in a mixture of 25% CremophorEL (Sigma-Aldrich, St. Louis, MO, USA) and 75% phosphate-buffered saline and then injected at 10 mg/kg into the tail vein of male Wistar rats. At various time points after the injection, ∼100 to 200 μL of blood was collected from a tail nick into EDTA blood sample collection tubes. For simultaneous determinations of plasma and brain concentrations, TRAM-34 was dissolved in Miglyol 812 neutral oil (caprylic/capric triglyceride; Trade name Neobee M5, Spectrum Chemicals, Gardena, CA, USA) at 10 or 40 mg/mL and injected intraperitoneally at 10 or 40 mg/kg. Blood samples were taken by cardiac puncture under deep isoflurane anesthesia. The right atrium was then cut open and 20 mL of saline slowly injected into the left ventricle to flush the blood out of the circulation. The rats were then sacrificed and brains removed. Plasma was separated by centrifugation and samples stored at −80 °C for pending analysis. Plasma and homogenized brain samples were purified using C18 solid phase extraction cartridges. Elutioned fractions corresponding to TRAM-34 were dried under nitrogen and reconstituted in acetonitrile. LC/MS analysis was performed with a Hewlett-Packard 1100 series HPLC stack (Hewlett-Parkard; now Agilent Technologies, Santa Clara, CA, USA) equipped with a Merck KGaA RT 250-4 LiChrosorb RP-18 column (EMD Chemicals, Gibbstown, NJ, USA) interfaced to a Finnigan LCQ Classic MS (Thermo Electron, now ThermoFisher Scientific, Waltham, MA, USA). The mobile phase consisted of acetonitrile and water, both containing 0.2% formic acid. With a flow rate of 1.0 mL/min, the gradient was ramped from 20/80 to 70/30 in 5 minutes, then to 80/20 over 11 minutes, to 5/95 till 16.5 minutes, and finally back to 80/20 till 38 minutes. With the column temperature maintained at 30 °C, TRAM-34 eluted at 14.4 minutes and was detected by a variable wavelength detector set to 190 nm and the MS in series. Using electrospray ionization/ion trap MS (capillary temperature 270 °C, capillary voltage 1 V, tube lens offset −15 V, positive ion mode), TRAM-34 was quantified by its base peak of 277 m/z (2-chlorotrityl fragment) and concentrations calculated with a 5-point calibration curve from 25 nmol/L to 2.5 μmol/L. Concentrations above 2.5 μmol/L were quantified by their UV absorption at 190 nm. The related compound TRAM-46 (base peak of 261 m/z, 2-fluorotrityl fragment) was used as an internal standard.

The percentage of plasma protein binding for TRAM-34 was determined by ultrafiltration. Rat plasma was spiked with 50 and 100 μmol/L TRAM-34 in 1% dimethylsulfoxide and the sample loaded onto a Microcon YM-100 Centrifugal Filter (Millipore Corp., Bedford, MA, USA) and centrifuged at 14,000 g for 15 minutes at room temperature. The centrifugate (=free TRAM-34) was directly analyzed for TRAM-34 by HPLC-MS. The retentate was collected by inverting the filter into an Eppendorf tube and spinning at 14,000 g for 15 minutes. The retentate then underwent sample preparation as per the above-described procedure for determining total TRAM-34 concentration in plasma. The plasma protein binding of TRAM-34 was found to be 98±0.5% (n=3) and the unbound (=free) fraction 2.0±0.4%.

Assessment of Infarct Area

Rats were euthanized with an overdose of isoflurane. Blood samples for determination of electrolytes, pH, pCO₂, glucose and hemoglobin (I-STAT; Abbott, Princeton, NJ, USA) were drawn from the vena cava and brains quickly removed and sectioned into eight 2-mm thick slices starting from the frontal pole. Slices were then fixed in 10% buffer formalin embedded in paraffin and sectioned at 5 μm. Sections were stained with hematoxylin & eosin (H&E) and scanned (O’Donnell et al, 2004). The resulting jpg images were analyzed in Adobe Photoshop CS3 for infarct area using the Magnetic Lasso tool to outline the area and the Histogram tool to determine the number of pixels in the respective area. Percent infarct for each slice was calculated as (pixels in ipsilateral side/pixels in whole control hemisphere) × 100. Percentage of total infarct area of whole hemisphere was calculated as (summation of pixels in infarct from eight slices/summation of pixels in whole control hemisphere from eight slices) × 100. The degree of brain shrinkage was calculated from the same data.

Immunohistochemistry

Sections were dewaxed with xylene, rehydrated through an alcohol gradient, and heated with 10 mmol/L Na citrate (pH 6) in a microwave for 15 minutes to retrieve antigenic determinants. After treatment with 1% H₂O₂ to inactivate endogenous peroxidase activity and blocking with 5% goat serum in phosphate-buffered saline, the sections were incubated overnight at 4°C with the primary antibody in phosphate-buffered saline/2% goat serum. The following primary antibodies were used: KCa3.1 (1:500; AV35098, Sigma, St Louis, MO, USA), CD68 (ED1, 1:1,000; Serotec, Raleigh, NC, USA), and NeuN (1:1,000; A60, Millipore, Billerica, MA, USA). The polyclonal anti-KCa3.1 antibody, which recognizes human, rat and mouse KCa3.1 was tested for specificity with spleen and vascular sections from KCa3.1 wild-type and KCa3.1^−/− mice (Si et al, 2006) obtained from the laboratory of Dr Ralf Köhler at the University of Southern Denmark in Odense. Bound primary antibodies were detected with a biotinylated donkey anti-mouse IgG secondary antibody for CD68 and NeuN, or with a biotinylated goat anti-rabbit IgG secondary antibody (both 1:500, Jackson ImmunoResearch, West Grove, PA, USA) for KCa3.1 followed by a horseradish peroxidase-conjugated avidin complex (Vectastain Elite ABC Kit, Vector Laboratories, Burlingame, CA, USA). Peroxidase activity was visualized with 3,3′-diaminobenzidine (DAB Substrate Kit for Peroxidase, Vector Laboratories). Sections were counterstained with hematoxylin (Fisher, Pittsburg, PA, USA), dehydrated and mounted with Permount (Fisher). Apoptosis was assessed with the ApopTag Peroxidase in situ Apoptosis detection kit (Millipore, Billerica, MA, USA) according to the manufacturer's protocol.

ED1 stains CD68, a lysosomal membrane protein, which is mainly found in phagocytosing macrophages and reactive microglia (Damoiseaux et al, 1994). At 1:1,000, the antibody produced no stain on resting microglia. Infiltration of ED1⁺ cells was evaluated according to the method of Lehr et al (1997). Briefly, sections stained for ED1 were photographed and the resulting photos composited into whole-slide images with Photoshop. The Magnetic Lasso tool was used to outline hemisphere borders, brown pixels were selected with the magic wand tool and the number of brown pixels determined with the Histogram tool. The results are reported as brown=ED1⁺-positive pixels per one millimeter square area (pixels/mm²). NeuN is a DNA-binding, neuron-specific protein present in neuronal nuclei, perikarya, and some proximal neuronal processes. Strong nuclear staining suggests proper nuclear regulatory protein function representative of a healthy neuron. Sections stained for NeuN were photographed and the resulting photos composited into whole-slide images with Photoshop. NeuN and TUNEL-positive cells in the infracted hemisphere were counted with the Photoshop CS3 extended count tool.

Statistical Analysis

Statistical analyses of infarct area, neurological deficit scoring and immunohistochemistry were performed with one-way analysis of variance (Origin software) followed by post hoc pairwise comparison of the different groups using Tukey's method, also referred to as honestly significant difference test, as recommended by Schlattmann and Dirnagl (2010) for MCAO studies. P<0.05 was used as the level of significance; ∗P<0.05, ∗∗P<0.01, and ∗∗∗P<0.001. All data with the exception of the pharmacokinetic data in Figure 2, which shows mean±s.d., are given as mean±s.e.m.

Results

Middle Cerebral Artery Occlusion With 7 Days of Reperfusion Induces Substantial Activation of Microglia/Macrophages Expressing KCa3.1

After first performing filament MCAO with reperfusion in both male Sprague-Dawley and male Wistar rats, we chose Wistar rats for our experiments because in agreement with several reports on the differing cerebrovascular anatomy of different rat strains (Aspey et al, 2000; Dittmar et al, 2006) we found infarcts in this strain less variable than in Sprague-Dawley rats. Next, we needed to determine the optimal length of reperfusion for our experiments as we intended to evaluate whether KCa3.1 blockade reduces inflammatory damage. In keeping with previous studies investigating the time course of immune cell infiltration after MCAO in rodents and ischemic stroke in humans (Beschorner et al, 2002; Campanella et al, 2002), 48 hours of reperfusion only induced mild-to-moderate inflammation as measured by the number of ED1⁺ (=CD68⁺)-activated microglia/macrophages (data not shown), while 7 days of reperfusion resulted in a dramatic increase in ED1⁺ cells in the infarcted brain areas as shown in Figure 1, which depicts paraffin-embedded sections from a 90-minute MCAO with 7 days of reperfusion. Staining of serial sections with a polyclonal anti-KCa3.1 antibody, which did not produce any stain on lymphoid and vascular tissues from KCa3.1^−/− mice, revealed strong KCa3.1 expression on cells with the round or ‘ruffled’ shape characteristic of reactive microglia/macrophages in the infarcted areas. While the two stains clearly were on the same cells, they did not strictly colocalize because CD68 is a lysosomal protein (Damoiseaux et al, 1994), whereas the KCa3.1 channel is expressed on the plasma membrane and traffics through the endoplasmic reticulum (ER). As seen on the microvessel in the right panels of Figure 1, KCa3.1 protein was also detectable on vascular endothelial cells in keeping with the known expression of KCa3.1 in vascular endothelium and its role in the endothelium-derived hyperpolarizing factor (EDHF) response (Busse et al, 2002; Si et al, 2006). The abundantly present KCa3.1⁺ microglia in the infarcted area suggests that KCa3.1 inhibition might be of therapeutic benefit for curbing the secondary inflammatory damage in the wake of ischemic stroke.

Figure 1

Activated microglia/macrophages in infarcted brain areas express KCa3.1. Staining of serial paraffin-embedded sections from a 90-minute middle cerebral artery occlusion (MCAO) with 7 days of reperfusion for ED1 and KCa3.1.

Pharmacokinetics and Brain Permeability of TRAM-34

To reduce brain inflammation the KCa3.1 blocker TRAM-34 should ideally reach pharmacologically active concentrations in the brain. To address this question and to determine TRAM-34's pharmacokinetics in rats, we established an HPLC/MS assay to measure TRAM-34 concentrations in plasma and tissue. After intravenous administration at 10 mg/kg, total TRAM-34 plasma concentrations fell from a peak of 40 μmol/L at 8 minutes after application to 250 nmol/L at 24 hours. This decay in plasma levels was best fitted triexponentially reflecting a three-compartment model with rapid distribution from blood into tissue followed by elimination and slow repartitioning from body fat acting as a deep compartment back into plasma (Figures 2A and 2B). The half-life of TRAM-34 was calculated from the elimination part of the plot and found to be ∼2 hours, which is slightly longer than the 1-hour half-life we previously determined in mice (Toyama et al, 2008). Next, we injected TRAM-34 intraperitoneally at 10 and 40 mg/kg and measured total plasma and brain concentrations at various time points (Figures 2C and 2D). After administration of 10 mg/kg total plasma and brain levels of TRAM-34 initially peaked around 2.5 μmol/L between 30 minutes and 1 hour of application and then rapidly fell to 58±9 nmol/L in plasma and 191±41 nmol/L in homogenized brain tissue within 12 hours. The higher TRAM-34 dose of 40 mg/kg in contrast resulted in a much more protracted absorption from the intraperitoneal space (Figure 2D) and achieved plasma and brain concentrations exceeding 1 μM for 8 hours and only slowly falling to roughly 400 nmol/L at 12 hours. Since TRAM-34 had been previously administered subcutaneously once daily to prevent restenosis after angioplasty in rats (Köhler et al, 2003), we also determined TRAM-34 plasma levels after subcutaneous injection. In comparison with intravenous or intraperitoneal application, TRAM-34 showed very poor bioavailibility after subcutaneous administration and we needed to use 120 mg/kg to achieve plasma peaks of 2.5±1 μmol/L (data not shown). Release was further slow and varied greatly in its kinetics between individual animals, making subcutaneous application unsuitable for short-term in vivo trials despite the obvious convenience of once daily application of a high dose. To determine how much of the total TRAM-34 concentration was free, and thus available for blocking KCa3.1, we also determined TRAM-34's plasma protein binding by ultrafiltration. In keeping with TRAM-34's high lipophilicity, plasma protein binding was high (98%) leaving a free concentration of only 2%.

Figure 2

Pharmacokinetics of TRAM-34 in rats. (A) Total TRAM-34 plasma concentrations in rats (n=3) after intravenous administration at 10 mg/kg. The data are best fitted as triexponential decay in keeping with a three-compartment model. Inset: structure of TRAM-34. (B) Same data as in (A) shown on a logarithmic scale to better visualize the three slopes. Total TRAM-34 plasma (•) and brain (Δ, density assumed as 1 g/mL) concentrations after intraperitoneal injection at 10 mg/kg (C) or 40 mg/kg (D). Each data point in (C) and (D) is from three animals. All values are mean values±s.d., as usual for pharmacokinetic experiments.

Taken together, these results demonstrate that TRAM-34 has reasonably good pharmacokinetics in rats and effectively reaches the brain even when the blood–brain barrier is intact (C_brain/C_plasma=1.2). Based on its plasma half-life of 2 hours and its fast availability after intraperitoneal application, we decided to administer TRAM-34 twice daily intraperitoneally at 12-hour intervals.

KCa3.1 Blockade With TRAM-34 Reduces Infarction and Microglia Activation in Middle Cerebral Artery Occlusion With 7 Days of Reperfusion When Treatment Is Started 2 Hours After Reperfusion

In preliminary experiments, we induced relatively mild infarcts by reducing CBF by only 50% (control: 51.2±8.2% flux reduction, mean±s.d., n=10; TRAM-34: 49.0±7.4% flux reduction, n=11) and then administering TRAM-34 at 10 mg/kg twice daily for 7 days starting 2 hours after reperfusion. Under these conditions, we found a reduction in mean infarct area from 18.8±3.5% of the ipsilateral triphenyltetrazolium chloride-positive hemisphere area (n=10) in controls to 6.6±1.9% in TRAM-34-treated animals (n=11; mean±s.e.m., P=0.007). Supplementary Figure 1 shows an analysis of each affected brain slice from 2 to 16 mm from the frontal pole of the brain and a plot of the mean infarct area in the two groups. To access if the reduction in infarct area by TRAM-34 was accompanied by a reduction in microglia/macrophage activation, we stained formalin-fixed sections from the center of the infarct in the 8- and 10-mm slices from five animals of each group and evaluated ED1⁺ staining according to Lehr et al (1997). We chose this pixel-based method here because it is ideal for evaluating the large numbers of the intensely ED1⁺-stained and irregularly shaped microglia/macrophages. As shown in Supplementary Figure 1, TRAM-34 treatment reduced the amount of activated microglia/macrophages in the infarcted hemisphere (n=5 per group).

KCa3.1 Blockade with TRAM-34 Reduces Infarction in Middle Cerebral Artery Occlusion With 7 Days of Reperfusion When Treatment Is Started 12 hours After Reperfusion

Encouraged by the above-described results suggesting that KCa3.1 blockade can indeed reduce infarction and microglia activation, we then performed a second set of MCAO experiments where we reduced CBF more severely and evaluated infarct area by H&E staining instead of the quicker and more cost-effective triphenyltetrazolium chloride (TTC) method. H&E is more accurate for infarcts older than 48 to 72 hours because the reactive oxygen species produced by the massive numbers of activated microglia present in ‘aged’ infarcts can oxidize TTC leading to an underestimation of the infarct area. However, in order to be able to prepare undamaged 5 μm thin sections from the often brittle and delicate aged infarcts, the 2-mm coronal slices were fixed in 10% formalin for 1 day and the resulting paraffin blocks trimmed until several undamaged section could be obtained from each slice. To better simulate possible treatment conditions in the clinic, we also delayed the start of the TRAM-34 treatment until 12 hours after reperfusion.

As in our previous experiments, male Wistar rats were subjected to 90 minutes of MCAO with 7 days of reperfusion and then treated with either TRAM-34 at 10 or 40 mg/kg or vehicle twice daily starting 12 hours after successful reperfusion. CBF reduction was 67.3±9.6% in the controls (n=8), 71.4±5.8% in the high-dose TRAM-34 group (n=8), and 67.9±5.9% in the low-dose TRAM-34 group (n=6; values are mean values±s.d.) as shown in Supplementary Figure 2, which also depicts a representative example of a laser Doppler recording during MCAO surgery and reperfusion. Table 1 shows that there were no differences with respect to plasma concentrations of Na⁺, K⁺, Cl⁻, HCO₃⁻, glucose, blood urea nitrogen, hemoglobin, pH, pCO₂ or hematocrit in venous blood samples taken after the surgery and at the time of sacrifice between animals subjected to 90 minutes of MCAO and treated with either vehicle or TRAM-34. (Please note that pCO₂ directly after surgery was elevated in all groups due to the respiratory depression from ∼2 hours of isoflurane anesthesia.) We also measured TRAM-34 concentration at the time of sacrifice, which was 12 hours after the last application, and found an average plasma concentration of 91±46 nmol/L (n=6) for the 10 mg/kg group and 662±416 nmol/L (n=7; mean±s.d.) for the 40 mg/kg group, while no TRAM-34 was detectable in vehicle-treated animals. Chronic TRAM-34 administration per se had no measurable effects on the hematology, blood chemistry, necropsy, or serology of rats as shown in the Supplementary Toxicity data from a 28-day and a 6-month toxicity study with TRAM-34.

Table 1

Physiological parameters

After reperfusion	MCAO+ vehicle	MCAO+ TRAM-34 (10 mg/kg)	MCAO+ TRAM-34 (40 mg/kg)
Na+ (mmol/L)	133.2±3.0	134.4±2.3	135.5±2.3
K+ (mmol/L)	6.4±1.7	6.3±1.1	5.7±0.5
Cl− (mmol/L)	97.4±1.9	99.0±2.9	99.7±4.0
pH	7.29±0.13	7.34±0.1	7.4±0.13
pCO2 (mm Hg)	78.9±27.8	66.6±28.3	60.8±33.4
HCO3− (mmol/L)	36.0±4.1	33.8±6.4	32.5±5.7
BUN (mg/dL)	22.4±6.2	12.8±7.2	22.2±4.0
Glucose (mg/dL)	203.4±53.2	179.2±48.2	171.8±54.3
Hemoglobin (g/dL)	13.0±0.6	14.3±1.5	14.5±1.4
Day 7 after MCAO
Na+ (mmol/L)	136.9±2.3	137.7±4.1	138.8±3.5
K+ (mmol/L)	4.6±0.7	4.8±1.1	4.2±0.3
Cl− (mmol/L)	102.3±0.8	103.8±4.3	103.4±2.4
pH	7.4±0.06	7.3±0.06	7.4±0.07
pCO2 (mm Hg)	53.8±6.1	58.1±12.7	49.3±7.3
HCO3− (mmol/L)	30.1±0.4	29.7±2.5	29.2±1.3
BUN (mg/dL)	14.3±1.6	18.7±6.8	19.6±4.7
Glucose (mg/dL)	200.3±29.5	185.2±41.2	201.1±40.1
Hemoglobin (g/dL)	12.4±0.9	12.7±2.1	13.3±1.5

MCAO, middle cerebral artery occlusion.

Physiological parameters (±s.d.) measured by I-STAT in venous blood samples drawn after reperfusion from the tail vein or at sacrifice on day 7 after MCAO surgery from the vena cava. No significant differences were found between MCAO+vehicle (n=8), MCAO+TRAM-34 (10 mg/kg; n=6), and MCAO+TRAM-34 (40 mg/kg; n=8). pCO₂ was elevated in all groups due to respiratory depression by isoflurane anesthesia. (pCO₂ in tail vein samples from awake male rats is 49.9±3.9 mm Hg (n=5); 20 min of isoflurane anesthesia raises pCO₂ in tail vein samples to 57±6.7 mm Hg (n=5) in our hands.)

As shown in Figure 3, treatment with TRAM-34 resulted in a significant reduction in H&E defined lesion area with the mean infarct size (Figure 3B) being reduced from 22.6±3.6% in the controls (n=8) to 11.3±2.8% in rats treated with 10 mg/kg TRAM-34 (n=6, mean±s.e.m., P=0.039) and to 8.1±1.9% in rats treated with 40 mg/kg TRAM-34 (n=8; P=0.004). The treatment also tended to reduce brain shrinkage (Figure 3C). However, the results were only statistically significant with 40 mg/kg TRAM-34 (P=0.013), but not for the 10 mg/kg group (P=0.11).

Figure 3

Effect of TRAM-34 on infarct area in rats subjected to 90 minutes of middle cerebral artery occlusion (MCAO) with 7 days of reperfusion. (A) Hematoxylin & eosin (H&E) defined lesion areas in brain slices from 2 to 16 mm from the frontal pole from rats treated for 7 days starting 12 hours after reperfusion with vehicle (n=8), 10 mg/kg TRAM-34 (n=6), or 40 mg/kg (n=8). (B) Total hemisphere infarct area in the three groups. (C) Percentage of hemisphere shrinkage. All values are mean±s.e.m.

KCa3.1 Blockade With TRAM-34 Reduces Neurological Deficit, Microglia Activation and Neuronal Death

Using both a 4-score neurological evaluation scale shown to correlate well with infarct sizes in the frontoparietal cortex (Menzies et al, 1992) and a 14-score tactile and proprioceptive limb-placing test (De Ryck et al, 1989), rats were evaluated for neurological deficit 12 hours after MCAO and then every 24 hours for 7 days. The combination of both tests was chosen since filament MCAO in rats induces infarction not only in the major MCA territory, the lateral and parietal cortex, but also in the underlying striatum. Rats subjected to MCAO exhibited an average deficit score of 3 in the 4-score system (Figure 4A) and a score of 4 in the tactile and proprioceptive 14-score system (Figure 4B) 12 hours after MCAO. These scores slowly improved in vehicle-treated animals to an average of 2 in the 4-score and an average of 8 in the 14-score system by postsurgery day 7 probably reflecting the resolution of edema and partial compensation. (Please note that a normal rat has a score of 0 in the 4-score and a score of 14 in the 14-score system.) TRAM-34 treatment with 40 mg/kg started at 12 hours after reperfusion started to significantly improve neurological deficit in both the 4-score and the 14-score systems from day 5 or day 4 on and on day 7-treated rats displayed a score of 0.5 in the 4-score and of 12 in the 14-score system (Figures 4A and 4B). The lower TRAM-34 dose of 10 mg/kg only significantly improved neurological deficit in the more grade 14-score system and despite showing a positive trend toward improvement failed to significantly reduce deficit in the 4-score system. Interestingly, the deficit that high-dose TRAM-34-treated animals consistently failed to recover was forelimb placement without vision in keeping with the fact that their infarcts were mostly restricted to the striatum. Vehicle-treated animals in contrast exhibited both forelimb and hindlimb impairment on day 7, indicating infarction in both the frontoparietal cortex and the striatum.

Figure 4

Effect of TRAM-34 on neurological deficit. (A) Neurological deficit in the 4-score system (normal rat=0). Scores at 12 hours after reperfusion are 3.1±0.20 in vehicle-treated animals (n=8), 3.3±0.50 in the 10 mg/kg TRAM-34 group (n=6), and 3.0±0.80 in the 40 mg/kg TRAM-34 group (n=8, P=0.98) and are not significantly different between the groups. Scores at 196 hours (=7 days) after reperfusion are 1.5±0.30 in vehicle-treated animals (n=8), 0.8±0.10 in the 10 mg/kg TRAM-34 group (n=6, P=0.24), and 0.4±0.01 in the 40 mg/kg TRAM-34 group (n=8, P=0.02). (B) Neurological deficit in the 14-score system (normal rat=14). Scores at 12 hours after reperfusion are 4.3±0.94 in vehicle-treated animals, 4.8±0.63 in the 10 mg/kg TRAM-34 group, and 3.4±0.46 in the 40 mg/kg TRAM-34 group and are not significantly different between the groups. Scores at 196 hours (=7 days) after reperfusion are 7.5±0.46 in vehicle-treated animals, 9.8±0.01 in the 10 mg/kg TRAM-34 group (P=0.004), and 11±0.38 in the 40 mg/kg TRAM-34 group (P<0.001). All values are mean±s.e.m.

To determine if the delayed TRAM-34 application (12 hours after MCAO) also reduced microglia/macrophage activation similar to what we had previously seen when treatment was started 2 hours after reperfusion (Supplementary Figure 1), we stained sections from the center of the infarct in the 8- and 10-mm slices from all animals in the vehicle, low-dose and high-dose TRAM-34 group for ED1⁺ microglia and determined the ED1⁺ area according to the pixel-based method by Lehr et al (1997). While the delayed administration of 10 mg/kg TRAM-34 did not result in a significant reduction in microglia activation, the higher TRAM-34 dose of 40 mg/kg reduced the ED1⁺ area from 3,770.6±594.2 to 1,632.6±363.75 pixels/mm² (P=0.03) in the infracted hemisphere (Figure 5). This reduction in microglia activation in brains from rats treated with 40 mg/kg TRAM-34 was accompanied by an increase in the average number of surviving NeuN⁺ neurons (P=0.052) in the infarcted hemisphere from sections from coronal slices 8 and 10 of all animals. An analysis of TUNEL-positive cells in both the cortex and striatum further revealed that the 40 mg/kg dose of TRAM-34 significantly reduced the number of apoptotic cells in both locations (P=0.0067 for cortex and P=0.033 for striatum). The lower dose while showing a trend toward reducing the number of TUNEL-positive cells did not reach statistical significance (P=0.143 for cortex and P=0.230 for striatum).

Figure 5

Effect of TRAM-34 on microglia activation and neuronal survival 7 days after middle cerebral artery occlusion (MCAO). (A) ED1⁺ area (pixels/mm²) in the infarcted hemisphere from the 8- and 10-mm slices from all vehicle and TRAM-34-treated animals. (B) Number of surviving NeuN⁺ neurons in the infarcted hemisphere from the 8- and 10-mm slices from all vehicle and TRAM-34-treated animals. (C) Number of apoptotic TUNEL⁺ cells in the cortex and striatum in the infarcted hemisphere from all vehicle and TRAM-34-treated animals. All values are mean±s.e.m.

Discussion

The calcium-activated K⁺ channel KCa3.1 has an important role in several microglia functions such as respiratory burst (Khanna et al, 2001), migration (Schilling et al, 2004), and microglia-mediated neuronal killing in vitro and in vivo (Kaushal et al, 2007; Maezawa et al, 2010). Based on these observations, we reasoned that KCa3.1 blockers should reduce brain inflammation in the wake of ischemic stroke and here provide proof-of-concept evidence that KCa3.1 blockade with the small molecule TRAM-34 dose-dependently reduces infarction, neurological deficit, microglia/macrophage activation, and neuronal death in a rat MCAO model with 7 days of reperfusion.

We assume that the effects of TRAM-34 in our MCAO experiments were primarily mediated through inhibition of KCa3.1 on microglia because the majority of cells with activated microglia/macrophage morphology in ischemic brain have been described to arise from resident microglia and not from hematogenous macrophages infiltrating the brain (Schilling et al, 2003). In lipopolysaccharide-activated cultured rat microglia, TRAM-34 has previously been reported to greatly reduce neurotoxic activity through inhibition of p38 mitogen-activated protein kinase activation, reduction of iNOS induction, and suppression of nitric oxide production (Kaushal et al, 2007). TRAM-34 has further been found to inhibit amyloid-β oligomer induced microglia-mediated neuronal killing in organotypic brain slices as well as damage to postsynaptic elements (Maezawa et al, 2010). Based on these results, we hypothesize that the lipophilic TRAM-34, which effectively crosses the blood–brain barrier as shown in our study, exerts similar effects in vivo on activated microglia in the infarcted brain area. However, since KCa3.1 is also expressed on macrophages (Hanley et al, 2004; Toyama et al, 2008) and activated T cells (Ghanshani et al, 2000), we cannot exclude that inhibition of brain infiltrating macrophages and T cells contributed to the effect of TRAM-34. We further cannot rule out that systemic immunosuppression contributed to the effects of TRAM-34 since MCAO also leads to an increase in secretion of the inflammatory cytokines TNF-α, INF-γ, IL-6, and IL-2 in lymph nodes and spleen (Offner et al, 2006). It should further be mentioned here that KCa3.1 is also expressed on the MCA endothelium where it mediates the EDHF response together with the small-conductance Ca²⁺-activated K⁺ channel KCa2.3 (McNeish et al, 2006) and that the EDHF response appears to be increased 24 hours after reperfusion (Cipolla and Godfrey, 2010; Marrelli et al, 1999). While KCa3.1 blockade may of course have an effect on EDHF, any detrimental effects of blocking EDHF are likely to be outweighed by microglia suppressive effects. In this respect it is also noteworthy, that a recent study in open-chest anesthetized dogs reported that while KCa3.1 channels seem to modestly contribute to changes in coronary vascular resistance, TRAM-34 did not alter reactive hyperemic responses after brief coronary artery occlusion (Kurian et al, 2011). But it is of course a limitation for our study that we did not perform continuous blood pressure measurements during the entire lengths of the study; and we, therefore, cannot entirely exclude that blood pressure changes positively or negatively affected our experiments.

We are aware of the fact that our study solely relies on pharmacological evidence for proposing KCa3.1 as a potential therapeutic target for reducing inflammatory damage in stroke. Future studies using genetic manipulation of KCa3.1 would be expected to further clarify the role of KCa3.1 in the pathophysiology of ischemic stroke and will also need to address the question of whether KCa3.1 blockade indeed preferentially affects detrimental microglia functions and not phagocytosis of apoptotic and necrotic neurons as observed by Kaushal et al (2007) after optic nerve transection in rats.

In general, KCa3.1 seems to be relatively safe as a therapeutic target. Two independently generated KCa3.1^−/− mice (Begenisich et al, 2004; Si et al, 2006) were both viable, of normal appearance, produced normal litter sizes, did not show any gross abnormalities in any of their major organs and exhibited rather mild phenotypes: impaired volume regulation in erythrocytes and lymphocytes (Begenisich et al, 2004) and a reduced EDHF response together with an 8- to 14-mm Hg increase in blood pressure (Si et al, 2006; Brähler et al, 2009). Pharmacological blockade of KCa3.1 also seems to be safe and well tolerated. As mentioned in the Introduction, TRAM-34 exhibits an excellent selectivity over other ion channels (Wulff et al, 2000) and was ‘clean’ in a Hit Profiling screen on 32 neuronal receptors and transporters (Toyama et al, 2008). Daily administration of TRAM-34 did further not induce any toxicity in a 28-day toxicity study in mice (Toyama et al, 2008) or in a 28-day or a 6-month toxicity study in rats (see Supplementary Toxicity data for this paper). There have also been no reports about toxicity for the structurally related KCa3.1 blocker ICA-17043, which was developed by Icagen Inc. (Durham, NC, USA) and which has been in clinical trials for both sickle cell anemia and asthma (Wulff and Castle, 2010). ICA-17043 was found to be both effective and safe in phase-2 clinical trials (Ataga et al, 2008), but the phase-3 trials were stopped in 2007 apparently due to a lack of efficacy in reducing sickling crises. We would like to point out here that long-term pharmacological KCa3.1 blockade with TRAM-34 in mice (Toyama et al, 2008) or dose-escalating studies with ICA-17043 in 28 otherwise healthy patients with sickle cell disease did not increase blood pressure or lead to electrocardiogram changes (Ataga et al, 2008; Ataga and Stocker, 2009). TRAM-34 has further been found not to delay influenza virus clearance in rats in contrast to dexamethasone (Toyama et al, 2008), demonstrating that KCa3.1 blockers are relatively mild immunosuppressants and are not likely to increase the risk of infections.

In conclusion, we here propose KCa3.1 blockade as a novel therapeutic principle to reduce detrimental brain inflammation after ischemic stroke and possibly other neurological diseases with an inflammatory component.