J. Neurochem. (2011) 116, 122–131.
Acute ischemic stroke is a major risk for morbidity and mortality in our aging population. Currently only one drug, the thrombolytic tissue plasminogen activator, is approved by the US Food and Drug Administration to treat stroke. Therefore, there is a need to develop new drugs that promote neuronal survival following stroke. We have synthesized a novel neuroprotective molecule called CNB-001 (a pyrazole derivative of curcumin) that has neurotrophic activity, enhances memory, and blocks cell death in multiple toxicity assays related to ischemic stroke. In this study, we tested the efficacy of CNB-001 in a rigorous rabbit ischemic stroke model and determined the molecular basis of its in vivo activity. CNB-001 has substantial beneficial properties in an in vitro ischemia assay and improves the behavioral outcome of rabbit ischemic stroke even when administered 1 h after the insult, a therapeutic window in this model comparable to tissue plasminogen activator. In addition, we elucidated the protein kinase pathways involved in neuroprotection. CNB-001 maintains the calcium-calmodulin-dependent kinase signaling pathways associated with neurotrophic growth factors that are critical for the maintenance of neuronal function. On the basis of its in vivo efficacy and novel mode of action, we conclude that CNB-001 has a great potential for the treatment of ischemic stroke as well as other CNS pathologies.
brain derived neurotrophic factor
calcium-calmodulin-dependent protein kinase IIα
cyclohexyl bisphenol A
a pyrazole derivative of curcumin
Dulbecco’s Modified Eagle’s Medium
extracellular receptor kinase
fetal calf serum
US Food and Drug Administration
hippocampal cell line
reactive oxygen species
rabbit small clot embolic stroke model
sodium dodecyl sulfate
Ischemic stroke is a major cause of adult death and disability, resulting in over 6 million deaths annually (Ingall 2004). The nerve cell death associated with cerebral ischemia is because of multiple factors resulting from the lack of oxygen, including the loss of ATP, excitotoxicity, oxidative stress, reduced neurotrophic support, and multiple other metabolic stresses (Dirnagl et al. 1999). Therefore, a drug directed against a single molecular target may not be effective in treating the nerve cell death associated with stroke. In addition, drugs that inhibit a single CNS activity with high potency are likely to be toxic because the activity is probably also required for normal brain function. Indeed, there is no effective treatment for stroke that is approved by the US Food and Drug Administration (FDA) except for recombinant tissue-type plasminogen activator (rt-PA). It follows that combinations of drugs or individual drugs that are broadly neuroprotective may be required.
One group of multi-target compounds that may meet the criteria for the treatment of stroke are plant-derived polyphenolics. Our laboratories have recently shown that two flavones, fisetin and baicalein, improve clinical function in a rigorous rabbit embolic stroke model (Lapchak et al. 2007; Maher et al. 2007). Another polyphenolic that has great potential for the treatment of neurological disorders is curcumin (Fig. 1). Curcumin prevents Alzheimer’s disease pathology in animal models (Frautschy et al. 2001) and is effective at reducing the cognitive losses associated with traumatic brain injury (Wu et al. 2006). In addition, curcumin is effective in reducing multiple pathological indices in the rodent following middle cerebral artery occlusion (Thiyagarajan and Sharma 2004; Dohare et al. 2008). At least 10 potentially neuroprotective biological activities have been attributed to curcumin (Cole et al. 2007), making this multifunctional molecule a good candidate for the treatment of stroke.
There are, however, some properties of curcumin that could be improved, including its stability and its lack of neurotrophic activity and its inability to inhibit excitotoxicity. To circumvent some of these problems, we synthesized a library of hybrid molecules between cyclohexyl bisphenol A (CBA), a molecule with neurotrophic activity, and curcumin (Liu et al. 2008). These compounds were then screened in a variety of assays for protection against stroke-associated nerve cell toxicity. A pyrazole derivative, called CNB-001, that lacks the labile dicarbonyl group of curcumin, was identified (Fig. 1). CNB-001 has broad neuroprotective activity in cell culture models of excitotoxicity, oxidative stress and glucose starvation, and has a unique neurotrophic activity that mimics brain derived neurotrophic factor (BDNF) (Liu et al. 2008). In addition, CNB-001 enhances long-term potentiation and memory in a rodent object recognition assay (Maher et al. 2008). As all of these activities are relevant to the neurological problems associated with stroke (Mattson 2008), it was initially asked if CNB-001 meets the criteria for a candidate stroke treatment using a cell culture assay that mimics ischemia and may be predictive of drug efficacy in vivo (Maher et al. 2007). We then determined some of the steps in the neuroprotective pathway, and finally asked if CNB-001 is able to reduce the behavioral deficits in rabbits following an embolic stroke.
For the stroke studies, we used the rabbit small clot embolic stroke model (RSCEM), which is a possible indicator of treatments that show efficacy in human clinical trials, and was used in the development and FDA-approval of tissue plasminogen activator (Zivin et al. 1985; Lapchak 2010b). The RSCEM uses embolization to cause multifocal cerebral ischemia, which results in graded behavioral deficits that can be quantified (Lapchak 2010b). The primary endpoint in the RSCEM is behavior based upon motor function components of the National Institutes of Health Stroke Scale for stroke in humans (Lyden et al. 1999; Hacke et al. 2008). While the RSCEM is a superior translational model to develop and evaluate the behavioral effects of novel pharmacological agents, it is not a model where infarct volume can be measured using conventional techniques because of the presence of widespread ischemic cores in multiple brain areas and the requirement of radioactive tracer to measure clot burden in brain postmortem. Therefore, in the present study we utilized functional analysis as the primary endpoint. As secondary endpoint we measured a series of biochemical markers in ischemic cortical tissue.
It is shown that CNB-001 is effective in both cell culture and animal models of stroke. This protection is at least partially mediated via the maintenance of the MAPK and phosphatidylinositol-3-kinase (PI3K)-Akt kinase pathways, ATP levels, and elevated calcium-calmodulin-dependent protein kinase IIα (CaMKIIα), which are all pathways whose loss or reduction have been previously implicated in stroke pathology. The unique mode of action of CNB-001 is likely a direct result of the novel paradigm used for the synthesis and selection of CNB-001.
Materials and methods
Curcumin was from Sigma/Aldrich (St. Louis, MO, USA). CNB-001 was synthesized by AQ BioPharma Co., Ltd. (Shanghai, China) according to Liu et al. (2008). All other chemicals were from Sigma.
Fetal calf serum (FCS) and dialyzed FCS were from Hyclone (Logan, UT, USA). Dulbecco’s Modified Eagle’s Medium (DMEM) was purchased from Invitrogen (Carlsbad, CA, USA). HT-22 cells (Davis and Maher 1994; Maher and Davis 1996) were grown in DMEM supplemented with 10% FCS and antibiotics. HT-22 cells expressing the BDNF receptor were generously supplied by Dr. Gerald Thiel (Homburg, Germany).
In vitro ischemia assay
In vitro ischemia was done using HT-22 hippocampal neurons according to Maher et al. (Maher et al. 2007). Briefly, cells were seeded onto 96-well microtiter plates at a density of 5 × 103 cells per well. The next day, the medium was replaced with DMEM supplemented with 7.5% dialyzed FCS and the cells were treated with 20 μM iodoacetic acid (IAA) alone or in the presence of the different compounds. After 2 h, the medium in each well was aspirated and replaced with fresh medium without IAA but containing the same compounds. Twenty hours later, the medium in each well was aspirated and replaced with fresh medium containing 5 μg/mL 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide. After 4 h of incubation at 37°C, cells were solubilized with 100 μL of a solution containing 50% dimethylformamide and 20% sodium dodecyl sulfate (SDS; pH 4.7). The absorbance at 570 nm was measured on the following day. Results obtained from the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide assay correlated directly with the extent of cell death as confirmed visually. Controls included compound alone to test for toxicity and compound with no cells to test for interference with the assay chemistry.
Primary cultures of cortical neurons that die reproducibly by excitotoxicity were prepared by combining aspects of two published protocols as described (Schubert and Piasecki 2001). Briefly, BALB/c mouse embryo cortices were minced and treated with 0.1% trypsin for 20 min. After centrifugation, the cells were resuspended in B27 Neurobasal medium (Invitrogen) plus 10% fetal calf serum and were dissociated by repeated pipetting through a 1 mL blue Eppendorf pipette tip. Then the cells were plated at 1 × 105 cells per well in 96-well poly-l-lysine and laminin-coated microtiter plates (Becton Dickinson, Bedford, MA, USA) in B27 Neurobasal plus 10% fetal calf serum. Two days later the medium was aspirated and replaced by serum-free B27 Neurobasal medium plus 10 μg/mL cytosine arabinoside. The cultures were used without media change 11 days after plating and were essentially free of astrocytes. They were exposed to 10 μM glutamate followed by the test compounds. Cell viability was determined 24 h later using the fluorescent live/dead assay.
SDS–polyacrylamide gel electrophoresis and immunoblotting
HT-22 cells at the same density as used for the cell death assays were untreated or treated with the compounds alone or in the presence of 20 μM IAA for 2 h followed by 2 h recovery in the presence of the compounds. The cells were washed twice in phosphate buffered saline and solubilized in SDS-sample buffer containing 0.1 mM Na3VO4 and 1 mM phenylmethylsulfonyl fluoride, boiled for 5 min and either analyzed immediately or stored frozen at −70°C. Proteins were separated on 10% SDS-polyacrylamide gels and transferred to nitrocellulose. Equal loading and transfer of the samples was confirmed by staining the nitrocellulose with Ponceau-S. Transfers were blocked for 1 h at 22°C with 5% non-fat milk in Tris-buffered saline (TBS)/0.1% Tween 20 and then incubated overnight at 4°C in the primary antibody diluted in 5% bovine serum albumin in TBS/0.05% Tween 20. The primary antibodies used were: phospho-p44/42 MAPK antibody (#9101 and #9106, 1/1000) and phospho-Akt Ser473 (#9271, #4051, 1/1000) from Cell Signaling (Beverly, MA, USA); ORP150 antibody (#10301, 1/100 000) from IBL (Gunma, Japan), anti-CaMKIIα (Zymed Laboratories Inc., South San Francisco, CA, USA, Invitrogen #137300), actin antibody (#A5441, 1/800 000) from Sigma, anti-Akt (#05-591, 1/1000) from Millipore (Bedford, MA, USA), and pan extracellular receptor kinase (ERK) antibody (#E17120, 1/1000) from Transduction Laboratories (San Diego, CA, USA). The transfers were rinsed with TBS/0.05% Tween 20 and incubated for 1 h at 22°C in horseradish peroxidase-goat anti-rabbit or goat anti-mouse (Bio-Rad, Hercules, CA, USA) diluted 1/5000 in 5% non-fat milk in TBS/0.1% Tween 20. The immunoblots were developed with the Super Signal reagent (Pierce, Rockford, IL, USA).
To assay kinase pathways in tissue, groups of animals separate from those used for behavioral analysis were embolized with identical size (100–250 μm) and weights (2 mg) of clots, followed by the drug or vehicle 1 h later. Non-embolized animals received vehicle alone (see below). Six hours following embolization rabbits were killed with 1–1.5 mL of Beuthanasia-D via the marginal ear vein. The brain was rapidly removed from the skull within 30 s and the complete thickness of the parietal/occipital cortex was dissected on ice and then quick frozen using liquid nitrogen as previously described (Lapchak and De Taboada 2010). The identical area was taken from control animals (n = 5 per group). Brain samples were sonicated in phosphate-buffered saline (PBS) containing protease (CompleteMini, Roche, Mannheim, Germany) and phosphatase inhibitors (Phosphatase Inhibitor Cocktail 2, Sigma). Equal amounts of protein were solubilized in SDS-sample buffer and analyzed as for the cell extracts.
Total GSH and ATP
Total intracellular GSH and ATP were determined by chemical and chemiluminescent assays as described (Maher et al. 2007).
The in vitro cell death assays and the biochemical assays were repeated at least three times in triplicate each time and analyzed using Instat software (GraphPad Software, San Diego, CA, USA). The data are presented as the mean ± SD. Statistical analysis was done by anova followed by Bonferroni’s test. p < 0.05 was considered significant.
Rabbit small clot embolism model
Male New Zealand white rabbits weighing 2–2.5 kg were purchased from Rabbit Source Farms, Ramona, CA, USA and were supplied food (alfalfa cubes) and water ad libitum while under quarantine in an enriched environment for at least 5 days prior to experimental use. Surgery was done in a sterile controlled environment with a room temperature between 22.8 and 23.2°C. Institutional Animal Care and Use Committee approved the surgical and treatment procedures used in this study. Care was used throughout the study to minimize pain and discomfort. Per the Institutional Animal Care and Use Committee-approved protocol, rabbits were killed if they were in pain, showed extreme discomfort, or if they were unable to reach food or water.
Surgical procedures were done as described previously (Lapchak et al. 2004, 2007). Briefly, rabbits were anesthetized with isoflurane via a face-mask, 5% in 3 L/min at induction, and 2–3% in 3 L/min as a maintenance dose and placed on a QuantumHeat™ heat-pack (Quantum Products, Cleaveland, OH, USA). The right internal carotid artery was exposed, and the external carotid artery and the common carotid artery were ligated. A Becton, Dickinson and Company (B-D) plastic catheter oriented toward the brain was inserted into the common carotid artery positioned toward the brain and secured with ligatures. The incision was closed around the catheter so that the distal end was accessible outside. The catheter was filled with 0.2 mL of heparinized saline (33 units/mL) and plugged with injection caps. The animals were allowed to recover from anesthesia for at least 2–3 h before embolization.
Blood was drawn from one or more donor rabbits and allowed to clot for 3 h at 37°C. The large blood clots were then suspended in PBS (pH 7.4) with 0.1% bovine serum albumin and Polytron-generated fragments were sequentially passed through sizing screens to result in a suspension of small-sized blood clots (100–250 μm), which were suspended in PBS containing 1% bovine serum albumin. The blood clot suspension was then spiked with 57Co-labeled microspheres (Perkin-Elmer Inc., Waltham, MA, USA) to allow for tracking into the brain. An aliquot of the solution was removed for the determination of specific activity. For embolization, rabbits were placed in a restrainer and 1 mL of a clot particle suspension was injected through the carotid catheter, which was then flushed with 5 mL of sterile saline. Rabbits are fully awake during the embolization procedure and they are self-maintaining (i.e. they do not require artificial respiration or other external support). This allows for immediate observation of the effects of embolization on behavior at the time of embolus injection and thereafter. After the embolization process was completed, the catheter was heat ligated close to the neck leaving a small portion protruding from the skin.
Clinical rating score behavioral and quantal dose-response analysis
The use of clinical rating scores is a desirable primary endpoint to use when a novel therapeutic is being tested for further development and to support a clinical trial (Lapchak 2010b). Clinical scores in combination with quantal analysis is a sophisticated statistical analysis method to determine how a large population of stroke ‘patients’, in this case, rabbits, will respond to a treatment. To evaluate the quantitative relationship between clot dose lodged in the brain and behavioral deficits or clinical scores, logistic sigmoidal (S-shaped) quantal analysis curves were fit to dose-response data as originally described by Waud (1972). To construct a quantal analysis curve, a wide range of clot doses (3–7 mg) were injected via the indwelling carotid catheter in order to produce a spectrum of behaviorally normal to abnormal animals, which included death on the continuum of embolization-induced effects (Lapchak 2010b). In the absence of a neuroprotective treatment regimen, small numbers of microclots lodged in the brain vasculature cause no grossly apparent neurologic dysfunction. However, when large numbers of microclots become lodged in the vasculature, they invariably caused encephalopathy because of ischemia, neuronal degeneration, depletion of ATP and cell death (Lapchak and De Taboada 2010). In this study, less than 5% of animals died because of embolization, but in the event that embolization did result in a rapid death, there was a positive correlation with a high clot dose measured in brain and the animal was represented as abnormal (or dead) on the quantal analysis curve if the rabbit received treatment, either vehicle or drug. A 3-tiered clinical scoring scale is used for analysis of embolized rabbits: normal, abnormal or dead. Embolized rabbits are scored as abnormal if they have one or more of the following symptoms: ataxia, leaning, circling, lethargy, nystagmus, loss of balance, loss of limb/facial sensation and occasionally, hind-limb paraplegia. Using a simple dichotomous rating system, with a reproducible composite result and low inter-rater variability (< 5%), each surviving animal is rated as either behaviorally normal or abnormal by an observer naïve to the study design and/or treatment groups. For construction of quantal analysis curves, clot burden is plotted against behavior in order to define the P50 value. A separate quantal curve was generated for each treatment condition and a statistically significant increase in the P50 value (or the amount of blood clots in brain that produce neurologic dysfunction in 50% of a group of animals) compared to a control curve is indicative of neuroprotection or a behavioral improvement in the study population.
For test substance administration, rabbits were placed in a Plexiglas restrainer (Plaslabs Inc., Lansing, MI, USA) for the duration of the treatment. For all experiments in this study, rabbits were randomly allocated into treatment groups before the embolization procedure, with concealment of the randomization guaranteed by using an independent party. The randomization sequence was not revealed until all postmortem analyses were complete.
Power analysis of quantal dose-response curves indicates that a sample size of approximately 20 animals are required per group (Lapchak et al. 2004; Lapchak and Araujo 2007). The behavioral data are presented as P50 (Mean ± SEM) in mg clot for the number of rabbits in each group (n). P50 values were calculated using an iterative curve fitting algorithm as described previously (Lapchak et al. 2004). They were analyzed for significance using the t-test with SigmaStat 3.5 (Systat Software Inc., San Jose, CA, USA) using the Bonferonni correction when required.
CNB-001 was both designed and selected for its ability to have multiple biological activities (Liu et al. 2008). Selection was based upon a variety of neuroprotection assays representing distinct aspects of stress and toxicity that occur in stroke. More importantly, CNB-001 has a unique neurotrophic activity similar to BDNF (Liu et al. 2008). In addition, we have recently described a cell culture chemical ischemia assay that may have predictive power for screening drugs for stroke (Lapchak et al. 2007; Maher et al. 2007). This in vitro model mimics the loss of energy metabolism and ATP found in stroke, and specifically the rabbit model that was used in this study (Lapchak and De Taboada 2010). In our hands, the chemical ischemia assay is much more reproducible than the commonly used oxygen-glucose deprivation assay and therefore more appropriate to study protein kinase signaling pathways. In this assay, IAA, an irreversible inhibitor of the glycolytic enzyme glyceraldehyde 3-phosphate dehydrogenase, is used to deplete ATP levels in hippocampal HT-22 neurons. Using this in vitro model, it was asked if CNB-001 and/or its precursors, curcumin and CBA, had neuroprotective properties. Figure 2(a and b) shows that CNB-001 is significantly better than either curcumin or CBA in protecting cells from in vitro ischemia. When the three test compounds were added both during and at the end of a 2 h exposure to 20 μM IAA, there was a significant increase in viability with all three compounds, but the EC50 for CNB-001 was approximately 10-fold lower than that for the other compounds and the maximal level of protection was significantly greater (Fig. 2a). If the compounds were added only after the 2 h exposure to IAA, there was still significant protection by CNB-001 at 0.5–1 μM (Fig. 2b) and by curcumin at 10 μM, but CBA was no longer effective (Fig. 2b).
Previously, it was shown that some compounds that protect from in vitro ischemia do so by preventing the loss of GSH and/or ATP (Maher et al. 2007). While CNB-001 had no effect on intracellular GSH levels (data not shown), it did partially maintain ATP levels at 1 μM, its optimal dose in the in vitro ischemia assay (Fig. 2c). Neither curcumin or CBA maintained ATP levels at this concentration. Finally, excitotoxicity is an aspect of stroke pathology and a well studied drug target (Lapchak et al. 2007). The excitotoxicity assay depends upon the opening by low concentrations of glutamate of ionotropic glutamate receptors on E14 cortical neurons that have been in culture for 11 days resulting in rapid cell death. CNB-001 is effective at preventing excitotoxicity at 10 μM, but not at 1 μM, while curcumin is ineffective at both concentrations (Fig. 2d). These in vitro criteria for stroke efficacy suggest that CNB-001 may be neuroprotective in animal stroke.
Motor function is improved by CNB-001 following embolic strokes
To determine if CNB-001 can improve behavioral function or reduce the clinical deficits associated with embolic stroke, we used a rabbit model that has been described and reviewed in detail (Lapchak 2010b). A clinical rating system and quantal analysis technique were used to determine how rabbits subjected to embolic strokes respond to CNB-001 treatment. A clinical endpoint of motor function is a valid assay to use when drugs are being developed for clinical trials (Lapchak 2010b). Based upon previous studies with polyphenolic natural products, CNB-001 was administered subcutaneously at a dose of 100 mg/kg (Lapchak et al. 2007; Maher et al. 2007). In this study, we only administered the drug post-embolization to parallel clinical treatment regimens. An initial exploratory study used a 5-min post-embolization treatment time, which in itself has little opportunity to be useful clinically, but is useful during early stages of drug development to ensure that the dose and route of administration of the drug can produce a neuroprotective signal. There was a significant improvement in behavior following stroke when CNB-001 was administered at this early time-point. The P50 value for the CNB-001-treated group was increased by 157% compared with control (not shown). At this dose there was no overt behavioral response or toxicity with the drug within the time frame of the study. Behavioral analysis was done 24 h after treatment, and was done by an observer naïve of the study design who simply rated the rabbits as completely normal by all criteria or abnormal (see Methods), an all or nothing ‘quantal’ scale.
In the follow-up study, we used a 1 h post-embolization treatment time to determine if CNB-001 would improve clinical function. As shown in Fig. 3, CNB-001 improved behavioral function following embolization measured by a significant increase in the P50 values defined as the amount of blood clots in the brain that produce neurologic dysfunction in 50% of a group of animals and a shift in the curve to the right. Table 1 presents the raw data for the 1 h administration study that was used to construct the sigmoidal quantal analysis curve. CNB-001 significantly (p < 0.05) reduced stroke-induced behavioral deficits and increased the P50 value by 233% compared with the control group. The CNB-001-induced improvement in behavior is directly correlated with an increase in the population of rabbits that are behaviorally ‘normal’ following embolization. This is shown in columnar fashion in Table 1. The graph and table show that a larger population of rabbits with higher levels of clot deposited in the brain (clot burdens) are behaviorally ‘normal’ with CNB-001 treatment compared to vehicle-treated control animals with similar levels of clot burden. It should be noted that clot burden is measured postmortem in both groups. As shown in Table 1, there are some instances where there was no equivalent clot burden in the two groups. Nevertheless, using the statistical quantal analysis technique, the drug effect for the overall rabbit population can be calculated and represented by a P50 value.
|Clot weight||Rating score||Clot weight||Rating score|
|P50 = 0.84 ± 0.38 mg||P50 = 2.80 ± 0.59 mg|
There were no statistical differences (p > 0.05) between the P50 value measured in the 5 min and 1 h CNB-001 treatment groups, but both were significantly different from control (p < 0.05). Very few compounds work in the rabbit ischemia model 1 h post-trauma. An exception is rt-PA which has a similar therapeutic window in this assay (Lapchak 2010b).
Mode of action
To obtain an initial understanding of how CNB-001 functions in vivo, we examined a variety of signaling pathways previously implicated in stroke. Brain extracts were taken at 6 h post-embolism from control, stroked and stroked + CNB-001 treated animals. As the Ras-ERK cascade has been implicated in nerve cell survival in ischemia (Mehta et al. 2007), we initially examined this pathway. A significant decrease in ERK phosphorylation in the stroked animals relative to control animals was observed, and this decrease was largely prevented in the presence of CNB-001 (Fig. 4). PI3K-Akt signaling is also implicated in nerve cell survival both in vitro and in vivo (Zhao et al. 2006; Liu et al. 2010). Consistent with these observations, phosphorylation of Akt on Ser473 was largely eliminated in the stroked animals but was maintained in the presence of CNB-001 (Fig. 4). CNB-001 also has BDNF-like activity (Liu et al. 2008), and BDNF causes the activation of the ERK and Akt pathways (Rossler et al. 2004). As the expression of CaMKIIα is also increased by BDNF (Schratt et al. 2004), we assayed CaMKIIα expression in CNB-001-treated animals. Figure 4 shows that CNB-001 increased CaMKIIα expression. Finally, the endoplasmic reticulum-associated chaperone ORP150 has been linked to the PI3K/Akt and BDNF signaling pathways (Tamatani et al. 2001), and the over-expression of ORP150 reduces infarct volume in a mouse stroke model (Tamatani et al. 2001). ORP150 expression was decreased in the stroked rabbit brains and this decrease was again prevented by CNB-001 (Fig. 4).
In order to study the mechanisms underlying these in vivo effects of CNB-001, we turned to the in vitro model of ischemia. Similar to the in vivo results, cells treated for 2 h with 20 μM IAA and allowed to recover for 2 h also showed decreases in both ERK and Akt phosphorylation (Fig. 5a and b). Importantly, treatment with CNB-001 at concentrations ranging from 0.5 μM to 2.5 μM prevented the decreases in ERK and Akt phosphorylation in a dose-dependent manner. In contrast, a similar concentration of CBA or curcumin (2.5 μM) was ineffective. In the HT-22 cell line used for in vitro ischemia, however, ORP150 is expressed at an extremely high level and there was no consistent change in ORP150 levels (not shown).
Previously, we showed that CaMKII is activated by CNB-001 (Maher et al. 2008). Thus, we asked what role CaMKII activity plays in the neuroprotective effects of CNB-001. Not only did KN62, a CaMKII inhibitor, reduce the protection by CNB-001 in the in vitro ischemia model (Fig. 6a) but it also blocked the effects of CNB-001 on both ERK and Akt phosphorylation (Fig. 5). These data show that CaMKII regulates the ERK and Akt pathways in this model and is a more proximal target of CNB-001.
Finally, CNB-001 was selected from a library of curcumin derivatives because of its neurotrophic activity (Liu et al. 2008). As CNB-001 rescues HT-22 cells from ischemia, and both ERK and Akt activation and CaMKIIα expression are downstream pathways activated by neurotrophic molecules such as BDNF, it was asked if BDNF is able to rescue HT-22 cells expressing the BDNF receptor, TrkA, in the in vitro ischemia model. Figure 6(b) shows that BDNF rescues cells transfected with wild type TrkA (TK4), but not cells containing a mutated non-functional receptor (TK1) and this rescue is blocked by KN62. The rescue with BDNF is not, however, as complete as with CNB-001, suggesting that CNB-001 also functions by other mechanisms that remain to be defined.
The above data show that CNB-001, a synthetic derivative of curcumin with neurotrophic activity, inhibits nerve cell death in a stroke-related cell culture model of ischemia and can also improve clinical function in rabbits following small clot embolic strokes. Importantly, CNB-001 is equally effective when administered either immediately after embolization or 1 h later, traits shared with only a few potential stroke therapeutics, including the single FDA approved drug, rt-PA (Lapchak 2010b).
CNB-001 is a broadly neuroprotective compound as defined by cell culture assays for excitotoxicity, glucose starvation, loss of trophic factors, and oxidative stress (Liu et al. 2008). CNB-001 does not have any significant antioxidant activity itself (Liu et al. 2008), ruling out a direct antioxidant effect of the molecule. In addition, CNB-001 activates CaMKII and enhances long-term potentiation and memory in rodents (Maher et al. 2008), and increases the expression of CaMKIIα in animals (Fig. 4). Here, we show both in vivo and in vitro that the ERK and Akt phosphorylation pathways as well as ORP150 expression are also involved in stroke pathology and the CNB-001 mediated protection.
Oxidative glutamate toxicity (oxytosis) and in vitro ischemia are two cell culture assays employing the hippocampal nerve cell line HT-22 that were used to determine if CNB-001 had the potential to be effective in the embolic rabbit stroke model; the effect of CNB-001 on oxytosis has been published (Liu et al. 2008). Both assays have been used to screen a large number of flavones to predict potential efficacy for stroke in animals (Maher et al. 2007); both utilize neurotoxic conditions that are similar to those observed in ischemic stroke (Sigalov et al. 2000; Tan et al. 2001). In the oxytosis assay, glutamate blocks cystine uptake, leading to GSH depletion, lipoxygenase activation, reactive oxygen species (ROS) production, and cyclic guanosine monophosphate dependent Ca2+ influx resulting in cell death. We have previously shown that two compounds that block this pathway by either maintaining GSH levels (fisetin) or inhibiting lipoxygenases (baicalein) are effective in the rabbit stroke model (Lapchak et al. 2007; Maher et al. 2007). CNB-001 inhibits the oxytosis pathway at a step between the depletion of GSH and mitochondrial ROS production (not shown).
In the chemical ischemia assay, IAA inhibits the glycolytic enzyme glyceraldehyde 3-phosphate dehydrogenase, resulting in a loss of Akt and ERK phosphorylation, ATP depletion and oxidative stress as determined by increased ROS and lipid peroxidation (Taylor et al. 1996; Maher et al. 2007). The loss of ATP in ischemia results in cell death by necrosis (Vanlangenakker et al. 2008). This assay results in a toxicity profile very similar to the oxygen-glucose deprivation (Taylor et al. 1996), but is much more reproducible for pathway analysis and drug screening. Although CNB-001 is not a direct antioxidant, it is protective in both chemical ischemia in vitro and embolic stroke. What is the molecular basis for this protection?
Our results show that the activity of several signaling pathways implicated in neuroprotection are decreased in the cerebral cortex of stroked rabbits and that this decrease is partially attenuated by CNB-001. Specifically, CNB-001 prevents decreases in both ERK and Akt phosphorylation and maintains the levels of CaMKIIα and the pro-survival chaperone ORP150. Using our in vitro ischemia model, we were able to show that at least part of the neuroprotective action of CNB-001 is mediated by CaMKII, consistent with our previous observation that CNB-001 activates this kinase (Maher et al. 2008). CaMKII activates both ERK (Li et al. 2009) and Akt (Lahair et al. 2006) in nerve cell lines and elevates CaMKIIα protein expression (Schratt et al. 2004); ORP150 protein expression is also associated with elevated BDNF levels in the brain (Tamatani et al. 2001). Our results are in agreement with these observations because we show that the CaMKII inhibitor KN62 prevents both the maintenance of ERK and Akt phosphorylation by CNB-001 in the in vitro ischemia model as well as cell rescue by CNB-001. However, as KN62 only partially blocks the neuroprotective effects of CNB-001, it is likely that other pathways are involved. Like CNB-001, the neurotrophic molecule BDNF is also able to activate both the ERK and Akt pathways and prevent cell death in the in vitro ischemia assay (Fig. 6b).
One of the other pathways may involve ORP150, an endoplasmic reticulum resident chaperone whose over-expression was implicated in neuroprotection by a BDNF-mediated mechanism in a mouse stroke model (Tamatani et al. 2001). ORP150 is decreased at 6 h following embolic stroke in the rabbits and this decrease is prevented by CNB-001. There is evidence that ORP150 plays a role in Akt phosphorylation as in endothelial cells ORP150 silencing results in inhibition of Akt phosphorylation (Sanson et al. 2009) and phosphorylation of Akt is altered in the liver of mice expressing an antisense version of ORP150 (Nakatani et al. 2005). Thus, it is possible that the decrease in ORP150 levels contributes to the decrease in Akt phosphorylation in the rabbits. Although unlikely, we have not ruled out the possibility that CNB-001 has vascular effects that that may contribute to stroke outcome.
In conclusion, we have shown that the neurotrophic compound CNB-001 is neuroprotective in both in vitro and can significantly reduce motor function deficits in vivo, with a therapeutic window similar to that of rt-PA. The beneficial effects of CNB-001 correlates with its ability to maintain levels of ATP, and phosphorylated ERK and Akt, which are dramatically reduced following ischemia. As CNB-001 activates CaMKII and a CaMKII inhibitor reduces protection by CNB-001 as well as ERK and Akt phosphorylation, it is likely that CaMKII is a more proximal target for CNB-001. Because of its reasonable pharmacological properties (Liu et al. 2008), and its unique ability to mimic some activities of neurotrophic growth factors such as BDNF, CNB-001 is a viable lead compound for the development of drugs to treat ischemic stroke and other neurological conditions. Based upon a correlative analysis hypothesis described by Lapchak (2010b), a 1 h effective therapeutic window in the RSCEM was calculated to represent approximately 2.4–3 h in stroke patients, a time frame in which the majority of patients arrive at the hospital. We recognize that there are potential obstacles to the widespread use of a drug with a short therapeutic window. However, with an estimated therapeutic window of approximately 3 h, it will be possible to conduct clinical trials with the drug within the hyperacute phase following a stroke as defined by Saver and colleagues (see Ferguson et al. 2004). Moreover, as we further advance the development of CNB-001 toward a clinical trial, we envision using of state-of the-art trial design with a logistic regression model adjusting for baseline National Institutes of Health Stroke Scale (stroke severity), age, and presence or absence of diabetes. The use of shift analysis clinical trial design will allow us to gauge change in outcome distributions over the full range of measured outcomes, incorporating both benefit and harm (Saver and Gornbein 2009).
This work was supported by National Institutes of Health National Institute of Neurological Disorders and Stroke [Grant 1 R01 NS060864-01] (DS, PAL, PM) and National Institutes of Health National Institute of Neurological Disorders and Stroke [Grant 1 U01 NS60685-01] (PAL). The authors state that there are no actual or potential conflicts of interest.