Article (9) - 2009 - Arien-zakay Et Al., Exp Neurol 'neuroprotection By Cord Blood Neuronal Progenitors Involves...'

Experimental Neurology 216 (2009) 83–94

Contents lists available at ScienceDirect

Experimental Neurology j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / y e x n r

Neuroprotection by cord blood neural progenitors involves antioxidants, neurotrophic and angiogenic factors☆ Hadar Arien-Zakay a, Shimon Lecht a, Marian M. Bercu a, Rinat Tabakman a, Ron Kohen b, Hanan Galski c, Arnon Nagler c, Philip Lazarovici a,⁎ a b c

Department of Pharmacology and Experimental Therapeutics, School of Pharmacy, Faculty of Medicine, The Hebrew University of Jerusalem, POB 12065, Jerusalem 91120, Israel Department of Pharmacy, School of Pharmacy, Faculty of Medicine, The Hebrew University of Jerusalem, Jerusalem, 91120, Israel Laboratory of Molecular Immunobiology, Division of Hematology and Bone Marrow Transplantation and Cord Blood Bank, Chaim Sheba Medical Center, Tel-Hashomer, Israel

a r t i c l e

i n f o

Article history: Received 5 October 2008 Revised 6 November 2008 Accepted 14 November 2008 Available online 25 November 2008 Keywords: Neuroprotection Oxygen–glucose deprivation Free radicals NGF VEGF FGF-2 PC12 cells Cord blood progenitors

a b s t r a c t Human umbilical cord blood (HUCB) is a valuable source for cell therapy since it confers neuroprotection in stroke animal models. However, the responsible sub-populations remain to be established and the mechanisms involved are unknown. To explore HUCB neuroprotective properties in a PC12 cell-based ischemic neuronal model, we used an HUCB mononuclear-enriched population of collagen-adherent cells, which can be differentiated in vitro into a neuronal phenotype (HUCBNP). Upon co-culture with insultedPC12 cells, HUCBNP conferred ∼ 30% neuroprotection, as evaluated by decreased lactate dehydrogenase and caspase-3 activities. HUCBNP decreased by 95% the level of free radicals in the insulted-PC12 cells, in correlation with the appearance of antioxidants, as measured by changes in the oxidation–reduction potential of the medium using cyclic-voltammetry. An increased level of nerve growth factor (NGF), vascular endothelial growth factor and basic fibroblast growth factor in the co-culture medium was temporally correlated with a -medium neuroprotection effect, which was partially abolished by heat denaturation. HUCBNP-induced neuroprotection was correlated with changes in gene expression of these neurotrophic factors, while blocked by K252a, an antagonist of the TrkA/NGF receptor. These findings indicate that HUCBNP-induced neuroprotection involves antioxidant(s) and neurotrophic factors, which, by paracrine and/ or autocrine interactions between the insulted-PC12 and the HUCBNP cells, conferred neuroprotection. © 2008 Elsevier Inc. All rights reserved.

Introduction Research on neuronal progenitors is driven by their potential use for cellular therapy of neurodegenerative disorders (Raedt and Boon, 2005) and damage to the central nervous system following stroke and trauma (Chopp and Li, 2006). Since the availability of human neuronal stem cells derived from early embryos is extremely limited, progenitors of other origins, such as bone marrow (Sanchez-Ramos, 2002) or human umbilical cord blood (HUCB) (Buzańska et al., 2006), are being considered. HUCB contains multiple populations of pluripotent stem cells and can be considered an alternative to embryonic stem cells. Interestingly, sub-populations of HUCB cells, either hematopoietic (Chen et al., 2005) or mesenchymal-like (Low et al., 2008) cells upon treatment with specific growth factors are able to differentiate into neuron-like cells in culture with functional voltage- and ligand-gated channels (Sun et al., 2005), and thus amenable to treatment of neurologic diseases (Harris et al., 2007; Low et al., 2008; Newman et

☆ This study is part of a PhD thesis to be submitted to The Hebrew University of Jerusalem by H.A.Z. ⁎ Corresponding author. Fax: +972 2 6757490. E-mail address: [email protected] (P. Lazarovici). 0014-4886/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.expneurol.2008.11.006

al., 2004). Since HUCB cells have been proved therapeutically beneficial in animal models of stroke (Chen et al., 2001) and traumatic brain injury (Lu et al., 2002) albeit without evidence of neuronal tissue replacement, the promise of using HUCB-derived stem cells in the treatment of brain disorders relies on a “bystander” function through secretion of growth factors and cytokines inducing neuroprotection and immuno-modulation (Martino and Pluchino, 2006). Using neuronal conditioning medium and nerve growth factor (NGF), we recently established a novel approach for the isolation and in vitro differentiation of a population of collagen-adherent, nestinpositive progenitors from HUCB, operationally defined as neuronallike progenitors (HUCBNP) (Arien-Zakay et al., 2007). We hypothesize that these cells, in the undifferentiated or neuronal-like differentiated phenotype may be useful for neuronal therapy. Therefore, our aims in the present study were to evaluate in vitro the neuroprotective potential of HUCBNP in an ischemic PC12 cell model and to assess some of the cellular mechanisms involved in the neuroprotective effect. Using a special ischemic device (Tabakman et al., 2002) we developed an in vitro model of ischemia based on combined oxygen and glucose deprivation (OGD) insult, followed by reoxygenation (OGD/reoxygenation), thus mimicking the pathophysiological conditions of brain ischemia (Seta et al., 2002; Tabakman et

84

H. Arien-Zakay et al. / Experimental Neurology 216 (2009) 83–94

al., 2005). In this model, the PC12 cells release lactate dehydrogenase and undergo cell death, expressed by activation of caspase-3 and other apo-necrotic markers (Tabakman et al., 2002, 2004a). The model is well suited for investigating cell-induced neuroprotection since the insult-induced cell death is composed of two distinct phases: i. OGD insult and ii. reoxygenation, mimicking the resupply of oxygen to the insulted tissue. Whereas drug-induced neuroprotection is mostly effective upon pre-treatment of the cultures before the OGD insult (Tabakman et al., 2002), investigation and development of alternative methods for cell-induced neuroprotection during the reoxygenation phase (Tabakman et al., 2005) would be more practical and relevant for clinical use. We demonstrate that HUCBNP cells confer 30% neuroprotection against OGD/reoxygenation-induced cell death, an effect associated with a decrease in reactive oxygen species (ROS) in the insulted neuron, the appearance of antioxidants and increased levels of neurotrophins and angiogenic factors NGF, basic fibroblast growth factor (FGF-2) and vascular endothelial growth factors (VEGFs), in the media, in line with the in vivo “bystander” mechanism of stem cell-induced therapeutic effects (Martino and Pluchino, 2006). Materials and methods Cell cultures PC12 cells, kindly provided by Dr. Gordon Guroff, NICHD, NIH, were grown in 25 cm2 flasks in Dulbecco's modified Eagle's medium (DMEM) supplemented with 7% fetal bovine serum (FBS), 7% horse serum, 10,000 U/ml penicillin and 100 μg/ml streptomycin, as previously described (Tabakman et al., 2004b). The medium was replaced every 2 days. NGF-induced differentiation was achieved by treatment with 50 ng/ml mouse β-subunit nerve growth factor (2S-NGF) (Alomone Labs, Jerusalem, Israel) added every 2 days for 7 days (Abu-Raya et al., 2002). Cells were grown at 37 °C, at high humidity, 6% CO2, in a CO2 incubator. In the OGD/reoxygenation insult experiments, an identical number of cells (1 × 106) was plated in 12-well culture plates coated with 200 μg/ml type-I rat tail collagen (BD Biosciences, Bedford, MA, USA). All experiments were carried out in a clean room, according to ISO7 requirements (10,000 particles/m3). The experiments with K252a were performed as previously described (Koizumi et al., 1988). To isolate HUCBNP, HUCB was collected with the written approval of the mothers, according to a protocol approved by Sheba Medical Center's Committee on Ethics of Human Investigation. The mononuclear cells (MNC) were isolated on a sterile Ficol/Hypaque gradient (IsoPrep; Robbins Scientific Corporation, Sunnyvale, CA, USA), washed with phosphate-buffered-saline (PBS) (Invitrogen, Carlsbad, CA, USA) and counted in a hemocytometer. HUCBNP were isolated and cultured as previously described (Arien-Zakay et al., 2007). Briefly, HUCB, MNC and HUCBNP were plated at a density of ∼ 8 × 104/cm2 on collagen-coated wells of 12-well culture plates or on the membrane of Falcon culture inserts (PET track-etched membranes, 1 μm pore size, diameter 12 mm) (Falcon, Becton Dickinson, Franklin Lakes, NJ, USA) coated with 200 μg/ml type-I rat collagen. The HUCBNP undifferentiated cells were grown for up to 14 days in DMEM supplemented with 10% FBS, 2 mM L -glutamine and antibiotics (100 μg/ml streptomycin and 10,000 U/ml penicillin), all purchased from Biological Industries (Beit Haemek, Israel). To initiate neuronal differentiation, the HUCBNP were treated with the above described medium, supplemented with 10% SH-SY5Y neuroblastoma (ATTC, Manassas, VA, USA)-conditioning media (CM) and 10 ng/ml 2S-NGF for 14 days, as previously described (ArienZakay et al., 2007). The medium was replaced every four days. These conditions favor selection of a population of cells, which gradually develop neurite outgrowths and express typical neuronal markers (Arien-Zakay et al., 2007). The cells were grown at 37 °C, at high humidity, 6% CO2, in a CO2 incubator.

OGD/reoxygenation-PC12 cell model and neuroprotection protocol To induce in vitro ischemia, the undifferentiated or NGF-differentiated PC12 cells were cultured on 12-well plates at a density of 3.2×105 cells/ cm2 and introduced into an ischemic device consisting of a chamber maintained at 37 °C by warmed, circulating water with the aid of a heating system, as previously described (Tabakman et al., 2002). The oxygen level within the device was maintained below 1% (representing hypoxic insult) and was measured on-line by an oxygen sensor coupled to an oxygen monitor. On the day of the experiment, the regular highglucose DMEM (4.5 mg/ml) was replaced with glucose-free DMEM supplemented with serum and antibiotics. The cultures were then introduced into the ischemic device for 4.5 h to initiate the ischemic insult. The insult consists of two phases: phase I — deprivation of both oxygen and glucose (OGD insult) for 4.5 h; phase II — OGD insult followed by reoxygenation for 18 h (OGD/reoxygenation). At the end of phase I, glucose was added to a final concentration of 4.5 mg/ml, and the cultures were incubated for an additional 18 h under reoxygenation conditions. Control cultures were maintained under regular oxygen atmospheric conditions (normoxia). The anti-oxidant 4-hydroxy-2,2,6,6tetramethylpiperidine-1-oxyl (Tempol) (Sigma, St. Louis, MO, USA) was added 30 min before exposure to the OGD insult and was present during the OGD insult and the reoxygenation phase (Tabakman et al., 2002). To measure neuroprotection, undifferentiated or NGF-differentiated PC12 cells underwent OGD insult as above (phase I). At the end of the OGD insult and before initiation of reoxygenation, HUCB, MNC, undifferentiated or differentiated HUCBNP were added and present during the reoxygenation period (phase II). The ratio between the PC12 cells and HUCB-derived cells was ∼ 10:1, respectively. The experiments were performed in a single or double chamber coculture system. Using the single chamber approach, HUCBNP were cultured on the surface of the OGD insulted-PC12 monolayer in a feeding layer configuration. In the double chamber approach, HUCB, MNC or HUCBNP grown on inserts were added to OGD-insulted PC12 cells grown on 12-well plates in contact with the medium only. Using both co-culture approaches, the medium was completed to a volume of 1.5 ml/well with the regular high-glucose DMEM-containing PC12 medium and the cells were incubated under normoxic conditions for the reoxygenation period. To evaluate neuroprotection by HUCBderived cells, cell death was measured at the end of the reoxygenation period by the release of lactate dehydrogenase (LDH) into the medium (Tabakman et al., 2002). Total LDH (extracellular + intracellular) was obtained by freezing and thawing the cultures. In the single chamber approach, the total LDH of the PC12 cultures was estimated upon subtracting the total LDH of the separate HUCBNP cultures. In the double chamber approach, the total LDH of the PC12 cultures was estimated after removing the insert containing the HUCB-derived cells. Therefore, in both approaches, total LDH estimates mainly the activity of the insulted-neuron. Basal LDH release was measured in both PC12 and HUCB-derived cell cultures maintained under normoxic conditions. Insult-induced LDH release was expressed as the percentage of total LDH released into the medium. The neuroprotective effect is defined as the percent decrease in LDH release in the presence of HUCB-derived cells or Tempol, compared with that from untreated OGD/reoxygenation-insulted cultures (Tabakman et al., 2002). Each experiment, unless otherwise stated, was performed five times in sextuplicate. Caspase-3 activity (Tabakman et al., 2004a) in the OGD/reoxygenation-insulted PC12 cultures was measured using the double chamber approach, in which the insert containing the HUCBNP was separated at the end of the experiment from the insulted-PC12 cells and caspase-3 activity was estimated separately for each cell type. Basal caspase-3 activity was measured in cultures maintained under normoxic conditions. OGD/ reoxygenation insult-induced caspase-3 activity was expressed as the percentage of the normoxic values of caspase-3. The neuroprotective effect is defined as the percent decrease in caspase-3 activity in the

H. Arien-Zakay et al. / Experimental Neurology 216 (2009) 83–94

insulted-PC12 cells in the presence of HUCBNP or Tempol, compared with that from untreated OGD/reoxygenation-insulted cultures. Each experiment was performed three times in triplicate. To measure the neuroprotective effect of the culture medium, medium from PC12 cells, HUCBNP, insulted-PC12 cells or insultedPC12 cells co-cultured with HUCBNP was collected at 1 or 18 h of reoxygenation. The medium, either untreated or heat-denaturated (boiling for 30 min followed by 1 h cooling on ice), was then filtered through 0.2 μm Millipore™ filters and added to insulted PC12 cells at the end of OGD insult, at a concentration of 15% or 70% of the final culture medium. The culture medium was then completed to a volume of 1.5 ml/well with the regular high-glucose DMEM-containing PC12 medium and the cells were incubated under normoxic conditions during the reoxygenation phase as described above. ROS measurements and cyclic voltammetry (CV) ROS measurements were carried out as described previously (Mills et al., 1998). Dichlorodihydro-fluorescein diacetate (DCHF-DA, Invitrogen) was dissolved in ethanol and diluted with calcium-free PBS buffer to a final concentration of 5 μM. Two days before the experiment, 1 × 106 cells were plated on cover glasses, placed in 12well plates and coated with 200 μg/ml type-I rat tail collagen. Before the OGD insult, the PC12 cells were loaded with the oxidant-sensitive dye DCHF-DA in PBS for 5 min, and the solution was replaced with calcium and glucose-free PBS containing 0.5 mM magnesium. The cells were then exposed to OGD insult, followed by reoxygenation for 30 min, to allow maximal ROS detection, with or without the addition of HUCBNP grown on inserts. At the end of this short reoxygenation period, the inserts were removed and the PC12 cells were washed twice with PBS. The cover glasses were then removed and placed on microscope slides. In parallel, control cultures were loaded with DCHF-DA as described. The cells were then maintained in calcium-free PBS containing glucose and magnesium for 5 h under normoxia. The fluorescence of the intracellular DCHF-DA was quantified by a confocal laser scanning fluorescent microscope (Olympus 300 IX-70, Japan), using excitation and emission wavelengths of 488 and 520 nm, respectively. Images were collected using a 512 pixel format and stored for later analysis. The fluorescent intensity of single cells was measured with the aid of ImagePro Plus (Media Cybernetics, Silver Spring, MD). For each treatment, three different fields were recorded and the relative fluorescence intensity was measured in 15 cells per field. The experiments were repeated 3 times. To measure the presence of antioxidants, media from PC12 cells, HUCBNP, OGD-insulted PC12 cells or OGD-insulted PC12 cells cocultured with HUCBNP was collected at 1- or 18 h of the reoxygenation period and submitted in 0.1 M phosphate buffer (pH 7.4) for CV measurements, using the BAS model CV-1B CV apparatus, modified for a 250 μl cell volume (West Lafayette, IN, USA). CV tracings were recorded at a range of 0–1.3 V and at a rate of 100 mV/s, using an Ag/AgCl reference electrode. A three-electrode system was used throughout the study. The working electrode was a carbon disk (BAS MF-2012) of 3.2 mm diameter. Platinum wire served as the counter electrode. Analysis of CV tracings, namely, determinations of oxidation potential (E1/2) and detector anodic current (Ia), were carried out as previously described (Kohen et al., 2000). A typical cyclic voltammogram (anodic current–voltage relationship) of an antioxidant shows the typical peak currents (Ip) occurring at discrete voltages (Ep). The intensity of these currents is proportional to the amount of antioxidants present in the medium (Kohen et al., 2000). A change of 50 mV in potential was considered significant. Experiments were repeated 3 times in triplicates. Neurotrophic and angiogenic factors ELISA and NGF bioassay Samples of media were collected from the OGD/Reoxygenation insulted-PC12 cells exposed to different treatments. The samples were

85

assayed for factors using the ELISA assay. β-NGF levels were estimated using the Promega's Emax® ImmunoAssay System according to the manufacturer's instructions. The amount of NGF in pg/ml was calculated from a NGF standard curve, with a sensitivity of 8 pg/ml. The anti-NGF polyclonal antibody used in the assay cross-react with rat and human NGF, but shows less than 3% cross-reactivity with other neurotrophic factors at 10 ng/ml. The samples were also analyzed for human β-NGF using a Human β-NGF ELISA Development Kit (PeproTech, NJ, USA). The amount of human NGF in pg/ml was calculated from a human recombinant NGF standard curve, with a sensitivity of 16 pg/ml. The antibody exhibited at 50 ng/ml less than 1% cross-reactivity with rodent β-NGF and other growth factors. The same samples were analyzed for human FGF-basic (FGF-2), using a Human FGF-basic ELISA Development Kit (PeproTech). The amount of human FGF-2 in pg/ml was calculated from a human recombinant FGF-2 standard curve, with a sensitivity of 62 pg/ml. The antibody exhibited at 50 ng/ml 100% cross-reactivity with murine FGF-2, but exhibited less than 1% cross-reactivity with other FGF isoforms. The samples were further evaluated for human VEGF, using a Human VEGF-basic ELISA Development Kit (PeproTech). The amount of human VEGF-A (both the 165 and 121 isoforms) in pg/ml was calculated from a human recombinant VEGF standard curve, with a sensitivity of 63 pg/ml. The antibody exhibited at 50 ng/ml 100% crossreactivity with murine and human VEGF-A (165 and 121), but exhibited less than 1% cross-reactivity with other growth factors. Preparation of plates and solutions and ELISA protocols was according to the manufacturer's instructions. The experiments were repeated 3 times using different medium dilutions. Samples of media collected from OGD/reoxygenation insultedPC12 cells exposed to different treatments were evaluated for induction of neurite outgrowth, using the PC12-based neurotropic bioassay. The assessment of neuronal differentiation was quantified from images acquired with a Nikon Coolpix5000 camera mounted on a Nikon Eclipse TS100 (Japan) microscope, and processed by SigmaScan software. The degree of elongation of the neurite outgrowths (E) (n = 52–88 cells in each treatment) was estimated as previously described (Katzir et al., 2002). RT-PCR Total RNA was isolated and genomic DNA was degraded from the RNA preparations, using the SV total RNA isolation system (Qiagen GmbH, Hilden, Germany). A quantity of 1 μg of total RNA was reverse transcribed using the Reverse Transcription System (Promega, Madison, WI), according to the manufacturer's instructions. PCR was then performed in a final volume of 50 μl containing 5 μg cDNA, 50 pmol of each upstream sense and downstream sense primer, and 25 μl of GoTaq® Green Master Mix (Promega). For uniformity of comparison all PCR experiments were conducted for 30 cycles (for fibroblast growth factor 2 — FGF-2 and for vascular endothelial growth factor — VEGF) and 35 cycles (for NGF), since these conditions were within the range of linearity as predetermined from several trials. To generate various cDNA fragments, a Mastercycler gradient (Eppendorf, Germany) was programmed as follows: denaturation of cyclic parameters at 94 °C for 1 min, annealing at 58 °C (for FGF-2), 58 °C (for VEGF), 60 °C (for NGF) and 65 °C (for β-actin) for 1 min and elongation at 72 °C for 1 min. To identify the angiogenic and neurotrophic factors from PC12 (rat) and HUCBNP (human), the following primers were used, after checking their sequence with the gene bank database to ensure very high homology: FGF-2 (318 bp) sense: CCCAAGCGGCTCTACT-3′ (rat NM_019305, bp 614; human AC_000136, bp552) and antisense: TTTATACTGCCCAGTTCGTTT (rat, bp 911; human, bp 849), VEGF-A121 and VEGF-A165 (360 and 492 bp, respectively) sense: TGCACCCACGACAGAAGGGGA (rat NM_031836, bp 78; human NM_001025366, bp 1109) and antisense: TCACCGCCTTGGCTTGTCACAT (rat, bp 624; human, bp 1709), NGF (271 bp) sense:

86

H. Arien-Zakay et al. / Experimental Neurology 216 (2009) 83–94

Results HUCBNP induce neuroprotection in the OGD/reoxygenation-PC12 cell model

Fig. 1. The neuroprotective effect of HUCB-derived cell populations applied during reoxygenation in a single-chamber system. Undifferentiated PC12 cultures were exposed to OGD followed by 18 h reoxygenation (OGD/Reox) in the absence (Control) or presence of HUCBNP or Tempol. Normoxic cultures were maintained under regular atmospheric conditions. (A) Time course of OGD-induced cell death, resulting in ∼35% cell death at 4.5 h, served as a reference point. The data are the mean ± SD of triplicate experiments performed in sextuplicate. ⁎ vs time 0 p b 0.001. (B) HUCBNP-induced neuroprotection in relation to their differentiation. Cell death (mean ± SD) was calculated as described in Materials and methods, using the LDH assay (n = 3 in sextuplicate). Insert: Micrographs (× 200) at the top of the bars represent typical fields of undifferentiated (left) and differentiated (right) HUCBNP. ⁎ vs normoxia p b 0.001; ⁎⁎ vs OGD/ reoxygenated PC12 cells (Control) p b 0.01.

TGCTGAACCAATAGCTGCC-3′ (rat XM_227525, bp 354; human NM_002506, bp 334) and antisense: ATCTCCAACCCACACACTGAC-3′ (rat, bp 604; human, bp 584), β-actin (Promega) (285 bp, internal control) sense: TCATGAAGTGTGACGTTGACATCCGT-3′ and antisense: CTTAGAAGCATTTGCGGTGCACGATG-3′. All RT-PCR experiments were routinely controlled by conducting PCR without first performing a reverse transcription reaction. PCR products were analyzed by electrophoresis on agarose gel (2%) containing ethidium bromide for UV visualization and sequenced to prove identity. SH-SY5Y and rat brain-derived mRNAs served as a positive control for validation. The mRNA levels were semi-quantitatively evaluated by densitometric analysis of the RT-PCR products (Zaheer et al., 1995), performed using Quantity One 1-D Analysis Software program, 2003 version (Bio-Rad). The mean ± SD values of the ratio mRNA growth factor band intensity (arbitrary units) divided by mRNA βactin of the same experiment, were calculated from three independent experiments. Statistics Unless otherwise stated, each experiment was repeated 3–7 times, using sextuplicate cultures and the results are presented as the mean ± SD and evaluated using the GraphPad InStat 3 program (GraphPad Software Inc, San Diego, CA). Statistically significant differences between the experimental groups were determined by analysis of variance (ANOVA) followed by Bonferroni or Tukey posttests and were considered significant when p b 0.05.

To evaluate the HUCBNP neuroprotective potential we adjusted the OGD/reoxygenation model using PC12 cells, in accordance with a coculture system in which there is cell–cell interaction (single chamber) or media contact (double chamber). In this system, we investigated the potential neuroprotective effect of HUCBNP upon its addition to the PC12 cells after the OGD insult and during the reoxygenation phase. In this ischemic model, a mild insult was chosen, taking in consideration that the degree of achieved neuroprotection is reciprocally related to insult severity (Tabakman et al., 2002, 2004b). Fig. 1A demonstrates a typical kinetic experiment showing the cell death of undifferentiated PC12 cells in relation to the duration of OGD insult. At 4.5 h of OGD followed by 18 h of reoxygenation insult, mild (∼35%) cell death was observed. Therefore, this time point was chosen for the neuroprotective experiments with HUCBNP to allow measurement of statistically significant and reproducible neuroprotective effects. To explore the neuroprotective effect of HUCBNP we first used a single chamber approach that allows direct contact between the cells, using undifferentiated PC12 cells (Fig 1B). The OGD insult followed by 18 h of reoxygenation induced 36% cell death, whereas under normoxic conditions basal cell death is b3%. A concentration of 0.5 mM Tempol, a stable, cell membrane permeable radical scavenger and superoxide dismutase-mimetic compound, considered as positive control in the experiments, reduced the cell death by 8%, indicating 23% neuroprotection, as previously reported (Tabakman et al., 2002, 2005). We next investigated the ability of undifferentiated and neuronal-like differentiated HUCBNP to induce neuroprotection. HUCBNP were grown as undifferentiated or differentiated phenotypes (Fig. 1B, inserts), as previously described (Arien-Zakay et al., 2007), before co-culturing with insulted-PC12 cells in the single chamber

Fig. 2. The neuroprotective effect of HUCB-derived cell populations applied during reoxygenation in a double-chamber system. (A) Left: Scheme of the double chamber coculture system. 1–3: Micrographs (× 200) of undifferentiated PC12 cells exposed to: 1— normoxia, 2—OGD/reoxygenation, 3—OGD/reoxygenation in the presence of undifferentiated HUCBNP. (B) Cord blood and derived cell populations induced neuroprotection in undifferentiated and NGF-differentiated PC12 cells. Undifferentiated (gray bars) and 7 days, 50 ng/ml NGF-differentiated (white bars) PC12 cultures were exposed to OGD followed by 18 h of reoxygenation (OGD/Reox) in the absence (Control) or presence of 0.1 × 106 cells/insert HUCB, MNC or undifferentiated HUCBNP or 0.5 mM Tempol. Normoxic cultures were maintained under regular atmospheric conditions. Cell death (mean ± SD) was calculated as described in Materials and methods using the LDH assay (n = 3 in sextuplicate). (B) ⁎ vs normoxia p b 0.001; ⁎⁎ vs OGD/reoxygenated PC12 cells (Control) p b 0.05.

H. Arien-Zakay et al. / Experimental Neurology 216 (2009) 83–94

after OGD insult. Differentiated and undifferentiated HUCBNP cells conferred 23% and 17% neuroprotective effects, respectively, indicating that neuroprotection is an intrinsic property of these progenitor cells. Since in the above experiments the neuroprotective effects of HUCBNP were measured upon contact with the OGD/reoxygenation insulted-PC12 cells, we next investigated the relationship between cell–cell contact and the neuroprotective effect. For this purpose we adopted a double-chamber technology (Fig. 2A, left) with separate maintenance of the cells, allowing their contact only by medium (containing soluble factors from both cell types), during the reoxygenation period. In this experimental set-up, by removing the insert with the HUCBNP at the end of reoxygenation phase, it is possible to evaluate by light microscopy the morphological appearance of the insulted PC12 cells. Fig. 2A shows typical light micrographs of normoxic undifferentiated PC12 cells (Fig. 2A-1), insulted undifferentiated PC12 cells (Fig. 2A-2) and insulted undifferentiated PC12 cells after co-culture with undifferentiated HUCBNP (Fig. 2A-3). The neurotoxic effect induced in the OGD/reoxygenation-insulted undifferentiated PC12 cells was characterized by shrinkage, clustering, vacuolization and debris formation (Fig. 2A-2) as compared with normoxic cells (Fig. 2A-1). Repeating the experiment in co-culture with HUCBNP for the reoxygenation period clearly indicates a reduction in the severity of the above pathological changes (Fig. 2A-3). The advantage of the OGD/reoxygenation insult-induced PC12 cell model is the possibility to perform the experimental paradigm with undifferentiated or NGF-differentiated PC12 cells (Abu-Raya et al., 2002; Tabakman et al., 2005). Whereas undifferentiated cells resemble embryonal cromaffin cells, the NGF-differentiated cells are more similar to post-mitotic sympathetic neurons of the central nervous system (Fujita et al., 1989). Therefore, we sought to investigate the neuroprotective effect of HUCBNP on both undifferentiated and NGF-differentiated PC12 cells (Fig. 2B). In the same series of experiments we also evaluated the neuroprotective effect of whole cord blood (HUCB) cells – which was found to confer neuroprotection in an in vivo stroke model (Chen et al., 2001) – and of derived mononuclear fraction (MNC), which has been shown to provide neuroprotection after neuronal hypoxia using a neuroblastoma in vitro model (Hau et al., 2008). As in the single chamber approach, the level of cell death in normoxic undifferentiated and NGF-differentiated PC12 cultures was minimal (b3%), and 4.5 h of OGD insult, followed by 18 h of reoxygenation, induced ∼35% cell death. Under these conditions, Tempol reduced the cell death by 8.5% indicating 25% neuroprotection (Fig. 2B). Treatment of OGD/reoxygenation-insulted PC12 cells cultured with HUCB cells failed to confer neuroprotection, whereas the same number of MNC cells conferred 12% and 11% neuroprotection and undifferentiated HUCBNP conferred 22% and 20% neuroprotection on undifferentiated and NGF-differentiated PC12 cells, respectively (Fig. 2B). Considering the higher neuroprotective effect conferred by HUCBNP compared with MNC, allowing more accurate neuroprotection measurements, in the following experiments we employed only HUCBNP. In addition, since similar levels of neuroprotection were conferred by HUCBNP on undifferentiated and NGF-differentiated PC12 cells, as a matter of convenience, in the following experiments we used only undifferentiated PC12 cells. Since apoptosis plays an important role in OGD/reoxygenationinduced PC12 cell death (Tabakman et al., 2004a), we compared the neuroprotective effect of undifferentiated and differentiated HUCBNP in the same series of experiments, by measuring caspase-3 activation in the PC12 cell monolayer (apoptosis) and LDH release into the media (necrosis) (Fig. 3). Morphological evaluation and LDH release of HUCBNP at different time points during the 18 h of reoxygenation indicated marginal cell death of less than 3% and these cells were able to survive in culture for several weeks (data not shown), supporting the notion that ischemic environments promote survival of neuronal stem cells (Mylotte et al., 2008; Theus et al., 2008). Therefore, we

87

Fig. 3. HUCBNP-induced neuroprotection in relation to their differentiation. The HUCBNP cultured on inserts were introduced during reoxygenation in the doublechamber system. Undifferentiated PC12 cultures were exposed to OGD followed by 18 h of reoxygenation (OGD/Reox) in the absence (Control) or presence of 0.1 × 106 cells/ insert undifferentiated or differentiated HUCBNP, 0.5 mM Tempol or a combination of both. Normoxic cultures were maintained under regular atmospheric conditions. Cell death (mean ± SD) was calculated as described in Materials and methods using the (A) LDH assay (n = 5 in sextuplicate) and (B) caspase-3 (n = 3 in triplicate) assay. ⁎ vs normoxia p b 0.001; ⁎⁎ vs OGD/reoxygenated PC12 cells (Control) p b 0.01; # vs Tempol treatment p b 0.01, ## vs HUCBNP treatment p b 0.05.

assume that LDH release into the co-cultured medium represents a major contribution of OGD/reoxygenation insulted undifferentiated PC12 cells. Cell death measured by caspase-3 activity is attributable only to the insulted-PC12 cells, since it was evaluated in the PC12 cell monolayers after removal of the inserts containing HUCBNP. Upon culture expose to normoxia, the level of cell death estimated by LDH release in undifferentiated PC12 cultures was minimal, and 4.5 h of OGD insult, followed by 18 h of reoxygenation, induced ∼36% cell death (Fig. 3A). Under these conditions, Tempol reduced cell death by 8%, indicating 23% neuroprotection (Fig. 3A). Upon OGD/reoxygenation insult of undifferentiated PC12 cells, in the presence of undifferentiated or differentiated HUCBNP, a neuroprotective effect of ∼28%, measured by LDH release was observed (Fig. 3A). In the following experiment, we investigated the neuroprotective effect upon co-treatment with Tempol, applied with the OGD insult, followed by the addition of HUCBNP in the reoxygenation phase. At 0.1 × 106 cells/inserts of undifferentiated and differentiated HUCBNP and a fixed concentration of 0.5 mM Tempol, neuroprotective effects of 42% and 45%, respectively, were observed (Fig. 3A). Measurements of cell death by caspase-3 activation in the OGD/ reoxygenation insulted undifferentiated PC12 cells (Fig. 3B) revealed a similar neuroprotective trend. The caspase-3 activity in normoxic

88

H. Arien-Zakay et al. / Experimental Neurology 216 (2009) 83–94

The HUCBNP-induced neuroprotective effect is associated with the suppression of reactive-oxygen species (ROS) and the appearance of antioxidant activity

Fig. 4. Relationship between cell death and insult duration conferred by undifferentiated HUCBNP from different donors. Undifferentiated PC12 cells were exposed to 4– 7 h of OGD followed by 18 h reoxygenation in the absence (OGD/reoxygenated, gray bars) or presence of 0.1 × 106 undifferentiated HUCBNP/insert (black bars) or 0.5 mM Tempol (white bars). Cell death (mean ± SD) was calculated as described in Materials and methods, using the LDH assay. Each time point represents experiments performed in sextuplicate using HUCBNP from different donors. ⁎ vs OGD/reoxygenated PC12 cells p b 0.05.

cultures was considered 100%. The caspase-3 activity in OGD/ reoxygenation insulted undifferentiated PC12 cultures was 245% compared with that in the normoxic cultures. Tempol treatment reduced the caspase-3 activity by 78% compared with that of the OGD/ reoxygenation-insulted cells (Control), indicating 22% neuroprotection (Fig. 3B). Upon OGD/reoxygenation insult of undifferentiated PC12 cells, in the presence of undifferentiated and differentiated HUCBNP, neuroprotective effects of 16% and 26%, respectively, were observed (Fig. 3B). Upon co-treatment with Tempol and undifferentiated or differentiated HUCBNP, a neuroprotective effect of 31% and 33% was observed. These neuroprotective effects, measured by LDH release or caspase-3 activity may suggest a nearly additive neuroprotective effect between the progenitors and the antioxidant. The additive neuroprotective effect observed upon treatment with Tempol and HUCBNP may suggest that the neuroprotective effect of the mediumderived antioxidants is not maximal and addition of an exogenous antioxidant may further increase the neuroprotective effect. Furthermore, the similarity in the neuroprotective effect of HUCBNP upon application in a single- or double-chamber is indicative of HUCBNPderived soluble factor(s) contributing to the neuroprotection. To investigate the relationship between HUCBNP-induced neuroprotection and the potency of the OGD/reoxygenation insult (insult duration), experiments were performed in a double chamber by exposing the undifferentiated PC12 cultures to an OGD insult of 4–7 h duration. Following the OGD exposure, the cultures were transferred for 18 h reoxygenation in the presence or absence of undifferentiatedHUCBNP or Tempol, and cell death was estimated by LDH release (Fig. 4). It is evident that upon exposure of PC12 cells to OGD insults of up to 5 h duration, causing cell death of up to 50%, the presence of HUCBNP conferred a neuroprotective effect of between 21% and 39%. Upon applying stronger insults (N50% cell death), the neuroprotective effect dropped to 10% or was undetectable (Fig. 4). A similar relationship was obtained with differentiated HUCBNP (data not shown). In view of the similarity between the neuroprotective effect conferred by undifferentiated and differentiated HUCBNP, as a matter of convenience, in the following experiments we used only undifferentiated HUCBNP. The relationship between HUCBNP-induced neuroprotection and insult duration is also expressed under Tempol treatment, although Tempolinduced neuroprotection (16%–33%) is more evident at mild insults, causing lower than 40% cell death (Fig. 4). This analysis, using HUCBNP derived from HUCB of different donors, emphasize the reproducibility and consistency of the HUCBNP-induced neuroprotective effect.

To explore the possibility that HUCBNP confer neuroprotection by affecting the oxidative state of the insulted-PC12 cells in the early period of the reoxygenation phase, the PC12 cells were loaded with fluorescent redox-sensitive dye, and the fluorescence of the cells was followed using confocal microscopy. The PC12 cells were exposed to OGD insult following 30 min reoxygenation, in the presence or absence of HUCBNP (Fig. 5). It is evident that the OGD/reoxygenation insult increased by 65-fold the fluorescence intensity of the dyeloaded cells, whereas the presence of HUCBNP suppressed the PC12 cells-fluorescence intensity, indicating that the ROS level was significantly reduced, and was similar to the level measured in cells under normoxic conditions (Figs. 5A, B). Since the ROS levels of the insulted-cells may be regulated by the overall intracellular and/or medium antioxidant capacity (Genestra, 2007; Liu et al., 2002), we investigated, using cyclic voltammetry, the total redox potential activity (antioxidant activity) in the cell medium, reflecting antioxidants produced and released by both the insultedPC12 cells and/or the HUCBNP. We placed in the cyclic voltammeter culture medium derived from a co-culture of 0.7 × 106 PC12 cells/ml and 0.07 × 106 HUCBNP cells/ml, grown under different ischemic and normoxic conditions. The medium derived from insulted-PC12 cells in the presence or absence of HUCBNP was collected at 1, 4, and 18 h during the reoxygenation period. The voltammogram of the medium derived from PC12 cells exposed to OGD followed by 1 h reoxygenation insult, in the presence of HUCBNP, shows two small peak currents of 17.5 nA at Ep 240 mV and 8.5 nA at Ep 350 mV, resembling the anodic current–voltage relationship of the native low molecular weight antioxidants ascorbate and uric acid (Glantz et al., 2005) (Figs.

Fig. 5. The antioxidant effect of HUCBNP on OGD/reoxygenation-induced accumulation of ROS in PC12 cells. Undifferentiated PC12 cells were loaded with DCHF-DA, as described in Materials and methods. (A) Cell images were captured by confocal microscopy from cells under normoxia or OGD followed by 30 min reoxygenation in the absence (OGD/Reox) or presence (OGD/Reox-HUCBNP) of undifferentiated HUCBNP during the reoxygenation period. (B) The ROS level is expressed as the mean ± SD percentage of normoxia values for 45 cells per treatment (n = 3). ⁎ vs Normoxia p b 0.001; ⁎⁎ vs OGD/Reox p b 0.001.

H. Arien-Zakay et al. / Experimental Neurology 216 (2009) 83–94

89

The voltammogram from normoxic PC12 cells and HUCBNP lacks the above mentioned peak potentials, indicative of the absence of antioxidant(s) in the medium (Fig. 6, Normoxia). Neuroprotective effect of medium derived from PC12-HUCBNP co-cultures exposed to OGD/reoxygenation insult The presence of antioxidant(s) activity in the medium of the PC12/ HUCBNP experimental model called for an experiment to verify its neuroprotective effect. We, therefore, collected media from OGD/ reoxygenation insulted-PC12 cells in the presence (Fig. 7, PC12-OGD/ HUCBNP) or absence (Fig. 7, PC12-OGD) of HUCBNP after 1 (white bars) and 18 (gray bars) h of reoxygenation. The media were evaluated for neuroprotection in separate OGD/reoxygenation experiments with undifferentiated PC12 cells only. The OGD/reoxygenation insult induced 39% cell death, whereas under normoxic conditions cell death was b3%. The medium of HUCBNP and PC12 cultures grown under normoxia, for 1 or 18 h served as control and did not affect OGD/ reoxygenation-induced cell death. After 1 or 18 h of reoxygenation, a final concentration of 15% of added medium derived from insultedPC12 cells in the presence or absence of HUCBNP, did not reduce cell death. After 1 h of reoxygenation, using a concentration of 70% of the original co-culture medium (Fig. 7, PC12-OGD/HUCBNP), cell death was reduced by 14%, representing 36% neuroprotection. This effect is comparable to 1 h of reoxygenation treatment with 70% of the original insulted PC12 medium alone (Fig. 7, PC12-OGD), which did not affect cell death. Upon boiling of the original co-culture medium (stripped bars), treatment with 70% heat-denaturated medium induced a reduction of 6% in cell death, indicating 15% remaining neuroprotection (Fig. 7). Therefore, the heat-denaturation reduced by 58% the neuroprotective effect of the original co-culture medium. After 18 h of reoxygenation, using a concentration of 70% of the medium in the presence or absence of HUCBNP, a similar neuroprotective effect was observed. When 15% of the original culture medium was included, no protective effect was measured, indicating the requirement for an appropriate concentration of neuroprotectant(s), which is present at a 70% concentration of culture medium and reduced by heat-denaturation. The most plausible explanation of these results relies on the possibility that the cross-talk between the insulted PC12 cells and the HUCBNP, induces an increase in medium-neuroprotectant(s), which is Fig. 6. Relationship between anodic peak currents (Ip) and the peak potential (Ep) of the cell medium, as analyzed by cyclic voltammetry. The medium was collected from undifferentiated PC12 and HUCBNP cells under normoxia (black bars) or from OGD insulted-PC12 cells, followed by reoxygenation (OGD/Reox) for 1, 4, and 18 h in the presence (gray bars) or absence (white bars) of HUCBNP. The peak potentials of the three major anodic currents, representing three different antioxidant groups of compounds calculated under the different conditions, are expressed as the mean ± SD (nA, n = 3). ⁎ vs OGD/Reox p b 0.01, ⁎⁎ vs OGD/Reox-HUCBNP 1 h p b 0.05.

6A, B). In addition, a large peak current of 587 nA at Ep 770 mV was measured (Fig. 6C), presumably representing the redox potential activity of culture medium proteins and/or certain unknown antioxidant(s) (Kohen et al., 2000). In the absence of HUCBNP, these peak currents were not present, with the exception of a small peak current at Ep 770 mV (Fig. 6). Therefore, it may be concluded that the Ips at 240 and 350 mV are specifically induced by the presence of HUCBNP in the experimental system. Also, the results suggest that in the absence of HUCBNP, the insulted-PC12 cells medium generated the Ip level at 770 mV, which was enhanced by the presence of HUCBNP. In the presence of HUCBNP, time course analysis of the medium clearly shows a maximal current intensity (level) at 240 and 350 mV after 1 h of reoxygenation. Thereafter, this dropped to different intensities, whereas the peak current intensity at Ep 770 mV remained stable for 18 h. At 770 mV the Ip level of PC12 cells exposed to OGD insult in the absence of HUCBNP, gradually increased during the reoxygenation period, suggesting an accumulation of PC12-derived antioxidant(s).

Fig. 7. Neuroprotective effects of cell medium from the OGD/reoxygenation paradigm. The medium was collected from undifferentiated PC12 (PC12-NOR) and HUCBNP (HUCBNP-NOR) cells under normoxia or OGD insulted-PC12 cells followed by 1 h (white bars) and 18 h (gray bars) of reoxygenation in the absence (PC12-OGD) or presence (PC12-OGD/HUCBNP) of HUCBNP. Samples of PC12-OGD/HUCBNP collected at 1 h of reoxygenation were also denaturated (stripped bars, n = 3). The medium, at a final concentration of either 15% or 70%, was then applied to OGD-insulted PC12 cells (OGD/ Reox, black bar) for an 18 h of reoxygenation period, to measure its neuroprotective effect, using the LDH assay (mean ± SD, n = 5). ⁎ vs OGD/Reox p b 0.01, ⁎⁎ vs PC12-OGD 70% at the respective time points p b 0.01, # vs non-denaturated PC12-OGD/HUCBNP 70% p b 0.05.

90

H. Arien-Zakay et al. / Experimental Neurology 216 (2009) 83–94

Table 1 Measurement by ELISA and bioassay of NGF in cell culture medium upon OGD/reoxygenation in the presence or absence of HUCBNP Treatment

ELISA (pg/ml)

Bioassay (E)

Cross-reactive Ab Normoxia OGD/Reox

1h 4h 18 h

Human specific Ab

PC12

PC12/HUCBNP

PC12

PC12/HUCBNP

PC12

PC12/HUCBNP

ND 32 ± 13 43 ± 15⁎ 54 ± 24⁎

ND# 8 ± 15 149 ± 47⁎,⁎⁎ 80 ± 57⁎

ND ND ND ND

ND# 25 ± 5 167 ± 52⁎,⁎⁎ 38 ± 13⁎,⁎⁎

1.3 ± 0.1 NT 2.5 ± 0.14⁎ NT

1.0 ± 0.1 NT 3.2 ± 0.2⁎,⁎⁎ NT

Ab—antibody; Normoxia — PC12 or HUCBNP (#) cultures grown under regular conditions; OGD/Reox — PC12 cultures were exposed to OGD followed by 1, 4 or 18 h of reoxygenation in the presence or absence of HUCBNP co-cultured in the double chamber. At the end of OGD insult the NGF of the media of the experimental groups was measured. The ELISA values represent the mean ± SD of NGF immune reactivity, calculated from the respective NGF standard curve. In parallel experiments, the neurotropic activity of the media (mean ± S.E.M.) was evaluated by neurite outgrowth elongation (E) after 48 h of treatment, using a bioassay. ⁎ vs Normoxia p b 0.05; ⁎⁎ vs OGD/Reox in the absence of HUCBNP p b 0.05. ND—not detected; NT—not tested.

partially heat sensitive. These experiments show a correlation between the neuroprotective effect of HUCBNP observed after 1 h of reoxygenation, the decreased level of ROS in PC12 cells (Fig. 5) and the increased antioxidant levels measured in the co-culture medium (Fig. 6), which most probably are resistant to heat-denaturation, accounting for the remaining neuroprotective effect. Measurement of neurotrophic and angiogenic factors by ELISA and NGF bioassay of the medium derived from PC12-HUCBNP co-cultures exposed to OGD/reoxygenation insult Considering the emerging novel neuroprotective role of neurotrophic and angiogenic factors under ischemic conditions (Alzheimer and Werner, 2002; Sofroniew et al., 2001; Sun and Guo, 2005), it was important to evaluate their presence in the medium derived from OGD/reoxygenation insulted-PC12 cells co-cultured with HUCBNP. Therefore, the medium was collected from PC12 cultures exposed to OGD insult followed by 1, 4, and 18 h of reoxygenation in the presence or absence of HUCBNP. Media collected from normoxic PC12 cultures and HUCBNP cultures served as control. The growth factor level in the media of the experimental groups was measured using ELISA; NGF bioactivity was also evaluated by a neurite outgrowth bioassay. Furthermore, using an antibody selective for human NGF, as well as an antibody cross-reacting with both human and rat NGF, we attempted to identify the origin (PC12 cells or HUCBNP) of the NGF in the medium (Table 1). Low levels of NGF (32– 54 pg/ml) were detected in the medium derived from the PC12 cells exposed to OGD/reoxygenation insult for up to 18 h. A 3.5-fold increase in the level of NGF, to ∼ 150 pg/ml, was detected in the medium derived from the insulted-PC12 cells after 4 h of reoxygenation in the presence of HUCBNP. Using a human-specific anti-NGF antibody, a similar increase in the level of NGF in the medium after 4 h of reoxygenation was found. These findings indicate that HUCBNP respond to an ischemic environment by releasing NGF into the medium. After 18 h of reoxygenation, the amount of NGF in the medium collected from the insulted-PC12 cells in the presence of HUCBNP decreased, most probably reflecting degradation of the growth factor in the medium and/or its reduced synthesis and release by the cells. NGF was not detected in the medium of normoxic cultures of PC12 and HUCBNP. To confirm that the immune-reactivity indeed represents NGF biological activity, we analyzed by bioassay the medium of cells exposed to 4 h of reoxygenation (Table 1). The medium derived from OGD/reoxygenation insulted-PC12 cells in the presence of HUCBNP effectively induced neurite outgrowth elongation, indicative of a NGF/TrkA receptor-mediated effect. Since NGF was found to suppress insult(Kirkland et al., 2007) or mitogen-induced (Mills et al., 1998) ROS levels, it is plausible that the NGF detected in the medium may have contributed to the reduction in ROS level (Fig. 5) in relation to the measured neuroprotective effect (Fig. 7). Selective ELISAs also measured the level of the typical angiogenic factors VEGF and FGF-

2 in the above medium (Table 2). A quantity of 204 pg/ml VEGF was detected in the medium derived from PC12 cells under normoxic conditions, indicating constitutive release, as previously reported (Middeke et al., 2002; Sarker et al., 1999). In the medium derived from the OGD/reoxygenation insulted-PC12 cells exposed to different reoxygenation periods, a gradual increase in the VEGF level was measured (Table 2). In the medium derived from the insulted-PC12 cells in the presence of HUCBNP, a twofold increase in the level of VEGF, to ∼420 pg/ml, was detected after 1 h of reoxygenation. This increase, to a lower extent was also observed after 4 h, reminiscent of the increased NGF level (Table 1). After 18 h of reoxygenation, the amount of VEGF in the medium collected from the insulted-PC12 cells in the presence or absence of HUCBNP was similar. Very low levels of VEGF (23 pg/ml) were detected in the medium of normoxic cultures of HUCBNP, similar to those previously reported for human CD34+ (Majka et al., 2001). A quantity of 165 pg/ml FGF-2 was detected in the medium derived from PC12 cells under normoxic conditions, confirming the previously reported detection of FGF-2 protein in PC12 cells (Grothe and Meisinger, 1997; Meisinger et al., 1996). In the medium derived from the insulted-PC12 cells exposed to 1 h of reoxygenation, a 3.5-fold increase in FGF-2 level was measured, indicating an ischemic effect on FGF-2 release, which gradually decreased to 316 pg/ml after 18 h of reoxygenation (Table 2). In the medium derived from the OGD/reoxygenation insultedPC12 cells in the presence of HUCBNP, a slight increase in the FGF-2 level was measured, reaching a twofold increase after 18 h of reoxygenation. Very low levels of FGF-2 (25 pg/ml) were detected in the medium of normoxic cultures of HUCBNP, similar to the levels previously reported for human CD34+ (Majka et al., 2001). The increased levels of VEGF and FGF-2 during the reoxygenation period in the medium of the insulted-PC12 cells in the presence of HUCBNP may explain the neuroprotective properties of the medium. This is in line with the concept that VEGF (Greenberg and Jin, 2005) and FGF-2 (Ikeda et al., 2005) are neuroprotective in an ischemic environment.

Table 2 Measurement by ELISA of angiogenic factors in cell culture medium upon OGD/ reoxygenation in the presence or absence of HUCBNP Treatment Normoxia OGD/Reox

VEGF (pg/ml)

1h 4h 18 h

FGF-2 (pg/ml)

PC12

PC12/HUCBNP

PC12

PC12/HUCBNP

204 ± 29 217 ± 39 327 ± 37⁎ 427 ± 62⁎

23 ± 17# 420 ± 62⁎,⁎⁎ 495 ± 87⁎,⁎⁎ 449 ± 47⁎

165 ± 40 587 ± 97⁎ 557 ± 87⁎ 316 ± 50⁎

25 ± 16# 625 ± 54⁎ 772 ± 38⁎,⁎⁎ 653 ± 101⁎,⁎⁎

Normoxia — PC12 or HUCBNP (#) cultures grown under regular conditions; OGD/Reox — PC12 cultures were exposed to OGD followed by 1, 4 or 18 h of reoxygenation in the presence or absence of HUCBNP co-cultured in the double chamber. At the end of OGD insult the angiogenic factors in the media of the experimental groups was measured. The ELISA values represent the mean ± SD of immune reactivity, calculated from the respective standard curves. ⁎ vs Normoxia p b 0.01; ⁎⁎ vs OGD/Reox in the absence of HUCBNP p b 0.01.

H. Arien-Zakay et al. / Experimental Neurology 216 (2009) 83–94

Fig. 8. Differential regulation of mRNA transcripts for angiogenic and neurotrophic factors in undifferentiated PC12 (A) and HUCBNP (B) cells under ischemic conditions. (A) mRNA transcripts for NGF, VEGF and FGF-2 in insulted-PC12 cells compared with normoxic conditions. PC12 cells were exposed to 4.5 h of OGD insult (OGD, gray bars) or 4.5 h of OGD followed by 18 h of reoxygenation (OGD/Reox, black bars) or left under normoxic conditions (Normoxia, white bars). (B) Insulted-PC12 cell-derived medium induced regulation of mRNA transcripts of NGF, VEGF and FGF-2 in HUCBNP. Undifferentiated HUCBNP were treated with their control medium (Control, striped bars), medium derived from normoxia PC12 cells (Normoxic-PC12 medium, white bars) or medium derived from PC12 cells exposed to 4.5 h of OGD insult (OGD-PC12 medium, gray bars). Total RNA was extracted, equal amounts of cDNA from each treatment were applied for RT-PCR using specific primers for each growth factor, and carried out as described in Materials and methods. The products were separated by electrophoresis, stained and photographed. Left: Typical mRNA transcripts of NGF, VEGF isoforms 165 and 121, FGF-2 and β-actin are presented. Right: Analysis of growth factor transcript levels expressed as the mean ± SD of the ratio growth factor/β-actin obtained from three different experiments. (A) ⁎ vs Normoxia p b 0.05, ⁎⁎ vs OGD p b 0.01. (B) ⁎ vs Control p b 0.01, ⁎⁎ vs Normoxic-PC12 medium p b 0.05.

The HUCBNP-induced neuroprotective effect is correlated with changes in gene expression of neurotrophic factors Neurotrophic factors appear to play important autocrine and paracrine roles in the response to injury or disease that subserves neuroprotection and neural repair (Sofroniew et al., 2001; Storkebaum et al., 2004). This concept and the presence of NGF, VEGF and FGF-2 in the media of the experimental system prompted us to explore the changes in the mRNA levels for these factors in the PC12 cells under normoxia, OGD insult and OGD followed by 18 h of reoxygenation insult (Fig. 8A). Under normoxic conditions, NGF mRNA was expressed at a very low level, as previously reported (Gill et al., 1998), whereas it was increased by 3.7- and 16.2-fold upon OGD and OGD/reoxygenation, respectively, as found in the adult rat brain following cerebral ischemia (Lindvall et al., 1992). In PC12 cells VEGF is encoded by two transcripts representing the well investigated VEGFA165 as well as its isoform VEGF-A121, known to be elevated by hypoxia (Levy et al., 1995; Middeke et al., 2002). Indeed, we observed their presence under normoxic conditions and a 1.2-fold increase upon OGD insult. However, under OGD/reoxygenation their expression was dramatically reduced. The results of our study also indicate moderate levels of FGF-2 transcripts in PC12 cells under normoxia, which were slightly decreased by OGD insult, reminiscent of hypoxia effects in certain neuronal cells (Khaliq et al., 1995), and were not further modulated by OGD/reoxygenation. Since embryonic neuronal progenitor cells express cytokines, angiogenic and neurotrophic factors which regulate their growth,

91

Fig. 9. The modulator effect of K252a on HUCBNP-induced differential regulation of mRNA transcripts of angiogenic and neurotrophic factors in PC12 cells exposed to OGD/ reoxygenation insult (OGD/Reox). PC12 cells were exposed to 4.5 h of OGD followed by 18 h of reoxygenation in the absence (Control) or presence of undifferentiated HUCBNP (HUCBNP), 100 nM K252a (K252a) or a combination of both (HUCBNP/K252a). Total RNA was extracted, equal amounts of cDNA from each treatment were applied for RTPCR, using specific primers for each growth factor, and carried out as described in Materials and methods. The products were separated by electrophoresis, stained and photographed. Typical mRNA transcripts of NGF, VEGF isoforms 165 and 121, FGF-2 and β-actin are presented. Analysis of growth factor transcript levels expressed as the mean ± SD of the ratio growth factor/β-actin obtained from three different experiments is presented in Table 3.

survival, and/or differentiation (Barnabé-Heider and Miller, 2003; Deleyrolle et al., 2006), and their expression in HUCBNP has not been explored, we studied the reciprocal expression of the same factors in HUCBNP grown for 18 h in the presence or absence of media from OGD-insulted PC12 cells (Fig. 8B). The levels of NGF, the two isoforms of VEGF and FGF-2 mRNA transcripts in HUCBNP under normoxic conditions were undetectable, slightly increased upon treatment with normoxic PC12 medium and significantly upregulated upon treatment with OGD-insulted PC12 cell medium. We then evaluated the mRNA expression of the above factors in PC12 cultures exposed to OGD insult followed by 18 h of reoxygenation, co-cultured in the double chamber in the presence or absence of HUCBNP, as previously described in the neuroprotective experiments (Figs. 1–4), and/or upon inclusion of the NGF/TrkA receptor antagonist K252a (Lazarovici et al., 1997) during the reoxygenation phase (Fig. 9, Table 3). The purpose of this experiment was to determine whether there was any indication of cross-talk between the insulted-PC12 cells and the HUCBNP-neuroprotective cells at the level of neurotrophic factors gene expression. It is evident that the presence of HUCBNP differentially modulated the expression of NGF compared with that of the VEGF and FGF-2 mRNA transcripts. The high level of NGF mRNA expression upon OGD/reoxygenation insult was reduced by 3.5-fold in the presence of HUCBNP. However, a significant up-regulation of both VEGF isoforms and FGF-2 transcripts was measured in the presence of HUCBNP. These modulations in neurotrophic factors mRNA expression

Table 3 Effect of HUCBNP on mRNA expression of neurotrophic and angiogenic growth factors in insulted-PC12 cells Growth factor NGF VEGF165 VEGF121 FGF-2

Normoxia 0.17 ± 0.19 2.54 ± 0.03 1.67 ± 0.02 1.24 ± 0.41

OGD/reoxygenation Control

HUCBNP

HUCBNP/K252a

K252a

2.75 ± 0.27 0.09 ± 0.02 0.18 ± 0.03 0.28 ± 0.48

0.79 ± 0.57⁎ 1.79 ± 0.35⁎ 1.32 ± 0.26⁎ 1.56 ± 0.26⁎

1.66 ± 0.25⁎/⁎⁎ 0.48 ± 0.08⁎⁎ 0.43 ± 0.07⁎⁎ 0.89 ± 0.12⁎⁎

1.93 ± 0.32⁎⁎ 1.29 ± 0.23⁎ 1.10 ± 0.20⁎ 1.48 ± 0.08⁎

Normoxia — PC12 cultures grown under regular conditions; Control — PC12 cultures exposed to OGD followed by 18 h of reoxygenation; HUCBNP — co-cultured in the double chamber at the end of OGD insult, for 18 h reoxygenation; HUCBNP/K252a — same as HUCBNP with the addition of 100 nM K252a; K252a — same as control with the addition of 100 nM K252a. The values represent the mean ± SD of mRNA expression of growth factors relative to β-actin (n = 3). ⁎ vs control p b 0.01; ⁎⁎ vs HUCBNP p b 0.05.

92

H. Arien-Zakay et al. / Experimental Neurology 216 (2009) 83–94

induced in the OGD/reoxygenation insulted-PC12 cells in the presence of HUCBNP resembled their levels in PC12 cells under normoxic conditions (Fig. 9, Table 3). Since PC12 cells (Vaudry et al., 2002) and HUCBNP (Arien-Zakay et al., 2007) express NGF/trkA receptor and NGF was detected in the coculture medium, we hypothesized that modulation of the mRNA of the above neurotrophic factors by HUCBNP is mediated, at least in part, by trkA. To assess trkA involvement, we included in the double chamber media 100 nM K252a, a concentration non toxic to PC12 or HUCBNP. K252a, a mixed agonist/antagonist compound, is known to inhibit NGF responses in the presence of NGF and to partially mimic its effects in its absence (Lazarovici et al., 1997). K252a selectively blocked mRNA expression of the HUCBNP-induced decrease in NGF and of the increase in VEGFs and FGF-2 (Fig. 9, Table 3). K252a treatment of the OGD-insulted PC12 cells in the absence of HUCBNP, indicated upregulation of the VEGFs and FGF-2 mRNAs, suggesting the regulatory role of trkA receptors in the expression of these growth factors. The similarity between the effect of K252a alone and HUCBNP on modulation of the gene expression of these growth factors may be explained by K252a activation of several neurotrophic pathways (Lazarovici et al., 1997; Roux et al., 2002). Cumulatively, these findings suggest that the above changes in neurotrophic factors gene expression are mediated in part by NGF/trkA receptors, which may be activated by the increase in NGF in the co-cultured medium. Discussion In this study, using the OGD/reoxygenation-induced PC12 ischemic model, we demonstrate the neuroprotective effect of HUCBNP cocultured with the insulted neuronal cells. This model, using undifferentiated (dopaminergic) or NGF-differentiated (adrenergic) PC12 cells was used in the past to measure the neuroprotective effects of the anti-Parkinsonian drug rasagiline (Azilect™) (Abu-Raya et al., 2002), the antioxidants carnosine, homocarnosine and Tempol (Tabakman et al., 2002) and NGF (Tabakman et al., 2004b). We have now adjusted this model to measure cell–cell-induced neuroprotection, using a double chamber approach. The ability to remove the HUCBNP from the model enabled measurement of apoptotic cell death by caspase-3 activity and necrotic cell death by LDH release from the OGD/reoxygenation insulted-PC12 cells. In this pharmacologically validated model (Tabakman et al., 2005), the HUCBNP conferred neuroprotection upon mild insult (20–50% cell death), independently of their differentiation phenotype and of their contact with the insulted-PC12 cells, indicating soluble factor-mediated neuroprotection. The neuroprotective effect of these soluble factors is mediated in part by protein(s) since this effect was partially abolished by heat denaturation. HUCBNP-induced neuroprotection was correlated with suppression of ROS in the OGD/reoxygenation insulted-PC12 cells and the appearance of antioxidant(s) activity and neurotrophic and angiogenic factors such as NGF, VEGF and FGF-2 in the medium. ELISA assays with different species specificities directed towards NGF clearly indicated that at least NGF in the medium is derived from both OGD/reoxygenation insulted-PC12 and neuroprotective-HUCBNP cells. Furthermore, the co-culture medium proved to be neuroprotective and induced up-regulation of VEGF-A isoforms 165 and 121 and of FGF-2 mRNA expression in the OGD/reoxygenation insulted-PC12 cells. These effects on angiogenic factors gene expression may also be attributed to co-culture medium soluble factors, as evident from the inhibitory effect of K252a, emphasizing the important role of NGF in this process. Based on these findings we propose that under normoxic conditions, the PC12 cells express a physiological redox potential and relatively high levels of mRNA transcripts of VEGF (Levy et al., 1995) and FGF-2 (Meisinger et al., 1996) (Figs. 5 and 8A). Most probably, the ability of these cells to express a high level of angiogenic factors transcripts is related to their pheochromocytoma tumor origin

and transformative properties (Claffey et al., 1992). Upon OGD/ reoxygenation insult, the level of ROS increased as expected (Liu et al., 2003), the expression of VEGF (Levy et al., 1995) and FGF-2 mRNAs transcripts highly and moderately decreased, respectively, while the level of the NGF mRNA transcript was markedly elevated. However, using ELISA specific for these growth factors and measurement at different time points during the reoxygenation period, we found that NGF was constantly released during reoxygenation (Table 1), VEGF slowly accumulated during 18 h of reoxygenation while FGF-2 peaked after 1 h and gradually decreased to moderate level at 18 h of reoxygenation (Table 2). The lack of correlation between mRNA and the protein levels of the angiogenic factors VEGF and FGF-2 may be explained by the emerging concept that angiogenic factors, like neurotrophins, are accumulated in specific intracellular vesicles, which release their content by receptor-regulated secretion (Italiano et al., 2008; Sarker et al., 1999). Based on present results, it is tempting to propose that ROS and neurotrophic factors, released into the medium from the OGD/ reoxygenation insulted-PC12 cells provide an activator signal to the HUCBNP in relation to neuroprotection. This proposal is based on several findings: i. the medium collected from HUCBNP grown in the absence of PC12 cells did not confer neuroprotection (Fig. 7); ii. HUCBNP express NGF/TrkA receptors (Arien-Zakay et al., 2007), therefore the presence of NGF in the co-culture medium is presumably activating these cells, as well documented for many hematopoietic cells expressing NGF receptors (Tabakman et al., 2004c); the presence of VEGF and FGF-2 detected in the medium is also presumably activating the HUCBNP; iii. OGD-insulted PC12 culture medium induced an increase in angiogenic factors mRNA expression in the HUCBNP (Fig. 8B), resembling hypoxia-induced chemokine expression by HUCB (Hau et al., 2008) and the ability of MNC (Liu and Hwang, 2005; Majka et al., 2001) and HUCB-derived neuronal progenitors (Neuhoff et al., 2007) to secrete numerous growth factors, cytokines and chemokines. A similar activating signal, resulting in the enhanced release of cytokines, was observed in a feeding layer configuration, characterized by the interaction of cord blood hematopoietic-derived stem cells with a stromal cell line (De Angeli et al., 2004). We propose that the increased mRNA levels of VEGF and FGF-2 (Fig. 8B) in HUCBNP are accompanied by their respective synthesis and release into the medium, similar to NGF (Table 1). Future experiments using humanselective ELISA for these angiogenic factors should confirm this possibility. Upon OGD/reoxygenation insult in the presence of HUCBNP, the redox potential of PC12 cells reverted to the normoxic level (Fig. 5) and the redox potential of the medium indicated the appearance of unidentified antioxidant(s) (Fig. 6). Furthermore, the level of NGF, VEGF and FGF-2 increased in the medium (Tables 1 and 2), indicating their release by one or both cell types. The PC12-gene expression profile of the neurotrophic factors tended to return to the normoxic level: the VEGFs mRNA level was greatly increased, FGF-2 moderately increased and NGF moderately decreased (Fig. 9, Table 3). The HUCBNP-induced neuroprotective effect may be explained by the paracrine and/or autocrine release of angiogenic factors such as VEGF (Lee et al., 2007) and FGF-2 (Song et al., 2002), which are well known to afford neuroprotection in brain stroke models. Furthermore, an important role may be attributed to NGF, which is known to be released by autocrine regulation of PC12 cells (Gill et al., 1998) and to confer neuroprotection upon treatment of the OGD-insulted PC12 cells (Boniece and Wagner, 1993), a phenomenon coincidentally related to attenuation of MAPK isoforms (Tabakman et al., 2004b). Besides its neuroprotective effect in this in vitro ischemic model and in an in vivo stroke model (Andsberg et al., 2002), NGF is also known for its ability to suppress ROS by rapid activation of antioxidants defenses in neurons (Kirkland et al., 2007). To get an insight into the chemical nature of the neuroprotective compounds we used a heat-denaturation approach, as used to characterize the free radical scavenging

H. Arien-Zakay et al. / Experimental Neurology 216 (2009) 83–94

molecules in brain homogenates (Mori et al., 1994). Our results using heat-denaturated medium (Fig. 7) indicate that both heat-sensitive (probably neurotrophic factors and/or anti-oxidant proteins) and heat-insensitive (probably low molecular weight anti-oxidants) compounds are implicated. The mechanisms by which mediumsoluble compounds protect neurons against ischemia may represent an additive effect of antioxidants and growth factors, as suggested by the increased neuroprotective effect under combined treatment with HUCBNP and Tempol (Fig. 3). The neuroprotective effect of HUCBNP on the OGD/reoxygenation insulted-PC12 cells is supported by similar neuroprotective effects induced by the whole HUCB-derived MNC, from which the present HUCBNP were isolated, on primary rat cortical neuronal cultures exposed to hypoxia and on hippocampal slice cultures exposed to OGD (Vendrame et al., 2005), as well as on differentiated neuroblastoma SH-SY5Y cells exposed to hypoxia in vitro (Hau et al., 2008). In the present study, the MNC neuroprotective effect on OGD/reoxygenation insulted-undifferentiated or differentiated PC12 cells was ∼50% of the HUCBNP neuroprotective effect at the same cell density. The decreased activity of MNC may be explained by physiological differences between the PC12 cells and cortical, hippocampal and neuroblastoma neurons and the different condition used to generate ischemic insults. Despite the controversy of the precise origin of HUCB-derived neuronal-like progenitors and the cellular differentiation process driving the progenitors towards a neuronal phenotype (Low et al., 2008) all studies with HUCB-derived progenitors suggest their usefulness as a source for central nervous system transplantation therapy (Newman et al., 2004). Therefore, based on their neuroprotective effect, the present HUCB population, characterized by collagenadherence and nestin expression, operationally defined as HUCBNP, both in the undifferentiated and neuronal differentiated phenotype, may provide a useful source for neural cell therapy. Although the neuroprotective effect of HUCBNP in animal models of stroke remains to be investigated, here we propose a concept of an in vitro neuroprotective effect of HUCBNP. To date, it is well known that upon intravenous administration of whole HUCB, the injected cells are able to survive, migrate and improve the functional recovery of the brain after stroke (Chen et al., 2001). Since these cells are widely available and have been used clinically, they, MNC or derived HUCBNP populations, conferring neuroprotection in vitro (Fig. 2), may represent an excellent, defined source of cells for the treatment of ischemic damage. As these cells selectively home to lesioned brain areas (Meier et al., 2006), the release of neurotrophic and angiogenic factors (Alexanian et al., 2008; Cho et al., 2006) and/or antioxidants in vivo, as found in the present study in vitro, would be highly specific and locally limited to the damaged brain area. In conclusion, future experiments by injection into the systemic circulation or brain implantation of the HUCB derived-mononuclear fraction or HUCBNP might improve the neuronal outcome by the release of neurotrophic and angiogenic factors and antioxidants, supporting a “bystander” neuroprotective strategy (Martino and Pluchino, 2006). Finally, HUCB progenitors-induced neuroprotection in ischemic neuronal cultures will continue to provide simple models, compared with animal models of stroke, to further elucidate the cellular and molecular mechanisms of stem cell-induced neuroprotection.

References Abu-Raya, S., Tabakman, R., Blaugrund, E., Trembovler, V., Lazarovici, P., 2002. Neuroprotective and neurotoxic effects of monoamine oxidase-B inhibitors and derived metabolites under ischemia in PC12 cells. Eur. J. Pharmacol. 434, 109–116. Alexanian, A.R., Maiman, D.J., Kurpad, S.N., Gennarelli, T.A., 2008. In vitro and in vivo characterization of neurally modified mesenchymal stem cells induced by epigenetic modifiers and neural stem cell environment. Stem Cells Dev. 17, 1123–1130. Alzheimer, C., Werner, S., 2002. Fibroblast growth factors and neuroprotection. Adv. Exp. Med. Biol. 513, 335–351.

93

Andsberg, G., Kokaia, Z., Klein, R.L., Muzyczka, N., Lindvall, O., Mandel, R.J., 2002. Neuropathological and behavioral consequences of adeno-associated viral vectormediated continuous intrastriatal neurotrophin delivery in a focal ischemia model in rats. Neurobiol. Dis. 9, 187–204. Arien-Zakay, H., Nagler, A., Galski, H., Lazarovici, P., 2007. Neuronal conditioning medium and nerve growth factor induce neuronal differentiation of collagenadherent progenitors derived from human umbilical cord blood. J. Mol. Neurosci. 32, 179–191. Barnabé-Heider, F., Miller, F.D., 2003. Endogenously produced neurotrophins regulate survival and differentiation of cortical progenitors via distinct signaling pathways. J. Neurosci. 23, 5149–5160. Boniece, I.R., Wagner, J.A., 1993. Growth factors protect PC12 cells against ischemia by a mechanism that is independent of PKA, PKC, and protein synthesis. J. Neurosci. 13, 4220–4228. Buzańska, L., Jurga, M., Domańska-Janik, K., 2006. Neuronal differentiation of human umbilical cord blood neural stem-like cell line. Neurodegener. Dis. 3, 19–26. Chen, J., Sanberg, P.R., Li, Y., Wang, L., Lu, M., Willing, A.E., Sanchez-Ramos, J., Chopp, M., 2001. Intravenous administration of human umbilical cord blood reduces behavioral deficits after stroke in rats. Stroke 32, 2682–2688. Chen, N., Hudson, J.E., Walczak, P., Misiuta, I., Garbuzova-Davis, S., Jiang, L., SanchezRamos, J., Sanberg, P.R., Zigova, T., Willing, A.E., 2005. Human umbilical cord blood progenitors: the potential of these hematopoietic cells to become neural. Stem Cells 23, 1560–1570. Cho, S.W., Gwak, S.J., Kang, S.W., Bhang, S.H., Won Song, K.W., Yang, Y.S., Choi, C.Y., Kim, B.S., 2006. Enhancement of angiogenic efficacy of human cord blood cell transplantation. Tissue Eng. 12, 1651–1661. Chopp, M., Li, Y., 2006. Transplantation of bone marrow stromal cells for treatment of central nervous system diseases. Adv. Exp. Med. Biol. 585, 49–64. Claffey, K.P., Wilkison, W.O., Spiegelman, B.M., 1992. Vascular endothelial growth factor. Regulation by cell differentiation and activated second messenger pathways. J. Biol. Chem. 267, 16317–16322. De Angeli, S., Di Liddo, R., Buoro, S., Toniolo, L., Conconi, M.T., Belloni, A.S., Parnigotto, P.P., Nussdorfer, G.G., 2004. New immortalized human stromal cell lines enhancing in vitro expansion of cord blood hematopoietic stem cells. Int. J. Mol. Med. 13, 363–371. Deleyrolle, L., Marchal-Victorion, S., Dromard, C., Fritz, V., Saunier, M., Sabourin, J.C., Tran Van Ba, C., Privat, A., Hugnot, J.P., 2006. Exogenous and fibroblast growth factor 2/epidermal growth factor-regulated endogenous cytokines regulate neural precursor cell growth and differentiation. Stem Cells 24, 748–762. Fujita, K., Lazarovici, P., Guroff, G., 1989. Regulation of the differentiation of PC12 pheochromocytoma cells. Environ. Health Perspect. 80, 127–142. Genestra, M., 2007. Oxyl radicals, redox-sensitive signalling cascades and antioxidants. Cell Signal. 19, 1807–1819. Gill, J.S., Schenone, A.E., Podratz, J.L., Windebank, A.J., 1998. Autocrine regulation of neurite outgrowth from PC12 cells by nerve growth factor. Brain Res. Mol. Brain Res. 57, 123–131. Glantz, L., Avramovich, A., Trembovler, V., Gurvitz, V., Kohen, R., Eidelman, L.A., Shohami, E., 2005. Ischemic preconditioning increases antioxidants in the brain and peripheral organs after cerebral ischemia. Exp. Neurol. 192, 117–124. Greenberg, D.A., Jin, K., 2005. From angiogenesis to neuropathology. Nature 438, 954–959. Grothe, C., Meisinger, C., 1997. The multifunctionality of FGF-2 in the adrenal medulla. Anat. Embryol. (Berl.) 195, 103–111. Harris, D.T., Badowski, M., Ahmad, N., Gaballa, M.A., 2007. The potential of cord blood stem cells for use in regenerative medicine. Expert Opin. Biol. Ther. 7, 1311–1322. Hau, S., Reich, D.M., Scholz, M., Naumann, W., Emmrich, F., Kamprad, M., Boltze, J., 2008. Evidence for neuroprotective properties of human umbilical cord blood cells after neuronal hypoxia in vitro. BMC Neurosci. 9, 1–15. Ikeda, N., Nonoguchi, N., Zhao, M.Z., Watanabe, T., Kajimoto, Y., Furutama, D., Kimura, F., Dezawa, M., Coffin, R.S., Otsuki, Y., Kuroiwa, T., Miyatake, S., 2005. Bone marrow stromal cells that enhanced fibroblast growth factor-2 secretion by herpes simplex virus vector improve neurological outcome after transient focal cerebral ischemia in rats. Stroke 36, 2725–2730. Italiano Jr, J.E., Richardson, J.L., Patel-Hett, S., Battinelli, E., Zaslavsky, A., Short, S., Ryeom, S., Folkman, J., Klement, G.L., 2008. Angiogenesis is regulated by a novel mechanism: pro- and antiangiogenic proteins are organized into separate platelet alpha granules and differentially released. Blood 111, 1227–1233. Katzir, I., Shani, J., Regev, K., Shabashov, D., Lazarovici, P., 2002. A quantitative bioassay for nerve growth factor, using PC12 clones expressing different levels of trkA receptors. J. Mol. Neurosci. 18, 251–264. Khaliq, A., Patel, B., Jarvis-Evans, J., Moriarty, P., McLeod, D., Boulton, M., 1995. Oxygen modulates production of bFGF and TGF-beta by retinal cells in vitro. Exp. Eye Res. 60, 415–423. Kirkland, R.A., Saavedra, G.M., Franklin, J.L., 2007. Rapid activation of antioxidant defenses by nerve growth factor suppresses reactive oxygen species during neuronal apoptosis: evidence for a role in cytochrome c redistribution. J. Neurosci. 27, 11315–11326. Kohen, R., Vellaichamy, E., Hrbac, J., Gati, I., Tirosh, O., 2000. Quantification of the overall reactive oxygen species scavenging capacity of biological fluids and tissues. Free Radic. Biol. Med. 28, 871–879. Koizumi, S., Contreras, M.L., Matsuda, Y., Hama, T., Lazarovici, P., Guroff, G., 1988. K-252a: a specific inhibitor of the action of nerve growth factor on PC 12 cells. J. Neurosci. 8, 715–721. Lazarovici, P., Matsuda, Y., Kaplan, D., Guroff, G., 1997. The protein kinase inhibitors K252a and staurosporine as modifiers of neurotrophin receptor signal transduction. In: Gutman, Y., Lazarovici, P. (Eds.), Toxin and Signal Transduction. . Cell. Mol. Mech. Toxin Action. Harwood Academic Pub., Amsterdam, pp. 69–94.

94

H. Arien-Zakay et al. / Experimental Neurology 216 (2009) 83–94

Lee, H.J., Kim, K.S., Park, I.H., Kim, S.U., 2007. Human neural stem cells over-expressing VEGF provide neuroprotection, angiogenesis and functional recovery in mouse stroke model. PLoS ONE 2, e156. Levy, A.P., Levy, N.S., Wegner, S., Goldberg, M.A., 1995. Transcriptional regulation of the rat vascular endothelial growth factor gene by hypoxia. J. Biol. Chem. 270, 13333–13340. Lindvall, O., Ernfors, P., Bengzon, J., Kokaia, Z., Smith, M.L., Siesjö, B.K., Persson, H., 1992. Differential regulation of mRNAs for nerve growth factor, brain-derived neurotrophic factor, and neurotrophin 3 in the adult rat brain following cerebral ischemia and hypoglycemic coma. Proc. Natl. Acad. Sci. U. S. A. 89, 648–652. Liu, C.H., Hwang, S.M., 2005. Cytokine interactions in mesenchymal stem cells from cord blood. Cytokine 32, 270–279. Liu, P.K., Robertson, C.S., Valadka, A., 2002. The association between neuronal nitric oxide synthase and neuronal sensitivity in the brain after brain injury. Ann. N.Y. Acad. Sci. 962, 226–241. Liu, Y., Song, X.D., Liu, W., Zhang, T.Y., Zuo, J., 2003. Glucose deprivation induces mitochondrial dysfunction and oxidative stress in PC12 cell line. J. Cell. Mol. Med. 7, 49–56. Low, C.B., Liou, Y.C., Tang, B.L., 2008. Neural differentiation and potential use of stem cells from the human umbilical cord for central nervous system transplantation therapy. J. Neurosci. Res. 86, 1670–1679. Lu, D., Sanberg, P.R., Mahmood, A., Li, Y., Wang, L., Sanchez-Ramos, J., Chopp, M., 2002. Intravenous administration of human umbilical cord blood reduces neurological deficit in the rat after traumatic brain injury. Cell Transplant 11, 275–281. Majka, M., Janowska-Wieczorek, A., Ratajczak, J., Ehrenman, K., Pietrzkowski, Z., Kowalska, M.A., Gewirtz, A.M., Emerson, S.G., Ratajczak, M.Z., 2001. Numerous growth factors, cytokines, and chemokines are secreted by human CD34(+) cells, myeloblasts, erythroblasts, and megakaryoblasts and regulate normal hematopoiesis in an autocrine/paracrine manner. Blood 97, 3075–3085. Martino, G., Pluchino, S., 2006. The therapeutic potential of neural stem cells. Nat. Rev. Neurosci. 7, 395–406. Meier, C., Middelanis, J., Wasielewski, B., Neuhoff, S., Roth-Haerer, A., Gantert, M., Dinse, H.R., Dermietzel, R., Jensen, A., 2006. Spastic paresis after perinatal brain damage in rats is reduced by human cord blood mononuclear cells. Pediatr. Res. 59, 244–249. Meisinger, C., Zeschnigk, C., Grothe, C., 1996. In vivo and in vitro effect of glucocorticoids on fibroblast growth factor (FGF)-2 and FGF receptor 1 expression. J. Biol. Chem. 271, 16520–16525. Middeke, M., Hoffmann, S., Hassan, I., Wunderlich, A., Hofbauer, L.C., Zielke, A., 2002. In vitro and in vivo angiogenesis in PC12 pheochromocytoma cells is mediated by vascular endothelial growth factor. Exp. Clin. Endocrinol. Diabetes 110, 386–392. Mills, E.M., Takeda, K., Yu, Z.X., Ferrans, V., Katagiri, Y., Jiang, H., Lavigne, M.C., Leto, T.L., Guroff, G., 1998. Nerve growth factor treatment prevents the increase in superoxide produced by epidermal growth factor in PC12 cells. J. Biol. Chem. 273, 22165–22168. Mori, A., Liu, J., Wang, X., Kawai, M., 1994. Free radical scavenging by brain homogenate: implication to free radical damage and antioxidant defense in brain. Neurochem. Int. 24, 201–207. Mylotte, L.A., Duffy, A.M., Murphy, M., O'Brien, T., Samali, A., Barry, F., Szegezdi, E., 2008. Metabolic flexibility permits mesenchymal stem cell survival in an ischemic environment. Stem Cells 26, 1325–1336. Neuhoff, S., Moers, J., Rieks, M., Grunwald, T., Jensen, A., Dermietzel, R., Meier, C., 2007. Proliferation, differentiation, and cytokine secretion of human umbilical cord blood-derived mononuclear cells in vitro. Exp. Hematol. 35, 1119–1131. Newman, M.B., Emerich, D.F., Borlongan, C.V., Sanberg, C.D., Sanberg, P.R., 2004. Use of human umbilical cord blood (HUCB) cells to repair the damaged brain. Curr. Neurovasc. Res. 1, 269–281.

Raedt, R., Boon, P., 2005. Cell therapy for neurological disorders: a comprehensive review. Acta Neurol. Belg. S105, 158–170. Roux, P.P., Dorval, G., Boudreau, M., Angers-Loustau, A., Morris, S.J., Makkerh, J., Barker, P.A., 2002. K252a and CEP1347 are neuroprotective compounds that inhibit mixed-lineage kinase-3 and induce activation of Akt and ERK. J Biol Chem. 277, 49473–49480. Sanchez-Ramos, J.R., 2002. Neural cells derived from adult bone marrow and umbilical cord blood. J. Neurosci. Res. 69, 880–893. Sarker, K.P., Yamahata, H., Nakata, M., Arisato, T., Nakajima, T., Kitajima, I., Maruyama, I., 1999. Recombinant thrombomodulin inhibits thrombin-induced vascular endothelial growth factor production in neuronal cells. Haemostasis 29, 343–352. Seta, K.A., Spicer, Z., Yuan, Y., Lu, G., Millhorn, D.E., 2002. Responding to hypoxia: lessons from a model cell line. Sci. STKE 2002, RE11. Sofroniew, M.V., Howe, C.L., Mobley, W.C., 2001. Nerve growth factor signaling, neuroprotection, and neural repair. Annu. Rev. Neurosci. 24, 1217–1281. Song, B.W., Vinters, H.V., Wu, D., Pardridge, W.M., 2002. Enhanced neuroprotective effects of basic fibroblast growth factor in regional brain ischemia after conjugation to a blood-brain barrier delivery vector. J. Pharmacol. Exp. Ther. 301, 605–610. Storkebaum, E., Lambrechts, D., Carmeliet, P., 2004. VEGF: once regarded as a specific angiogenic factor, now implicated in neuroprotection. Bioessays 26, 943–954. Sun, F.Y., Guo, X., 2005. Molecular and cellular mechanisms of neuroprotection by vascular endothelial growth factor. J. Neurosci. Res. 79, 180–184. Sun, W., Buzanska, L., Domanska-Janik, K., Salvi, R.J., Stachowiak, M.K., 2005. Voltagesensitive and ligand-gated channels in differentiating neural stem-like cells derived from the nonhematopoietic fraction of human umbilical cord blood. Stem Cells 23, 931–945. Tabakman, R., Lazarovici, P., Kohen, R., 2002. Neuroprotective effects of carnosine and homocarnosine on pheochromocytoma PC12 cells exposed to ischemia. J. Neurosci. Res. 68, 463–469. Tabakman, R., Jiang, H., Levine, R.A., Kohen, R., Lazarovici, P., 2004a. Apoptotic characteristics of cell death and the neuroprotective effect of homocarnosine on pheochromocytoma PC12 cells exposed to ischemia. J. Neurosci. Res. 75, 499–507. Tabakman, R., Jiang, H., Schaefer, E., Levine, R.A., Lazarovici, P., 2004b. Nerve growth factor pretreatment attenuates oxygen and glucose deprivation-induced c-Jun amino-terminal kinase 1 and stress-activated kinases p38alpha and p38beta activation and confers neuroprotection in the pheochromocytoma PC12 model. J. Mol. Neurosci. 22, 237–250. Tabakman, R., Lecht, S., Sephanova, S., Arien-Zakay, H., Lazarovici, P., 2004c. Interactions between the cells of the immune and nervous system: neurotrophins as neuroprotection mediators in CNS injury. Prog. Brain. Res. 146, 387–401. Tabakman, R., Jiang, H., Shahar, I., Arien-Zakay, H., Levine, R.A., Lazarovici, P., 2005. Neuroprotection by NGF in the PC12 in vitro OGD model: involvement of mitogenactivated protein kinases and gene expression. Ann. N.Y. Acad. Sci. 1053, 84–96. Theus, M.H., Wei, L., Cui, L., Francis, K., Hu, X., Keogh, C., Yu, S.P., 2008. In vitro hypoxic preconditioning of embryonic stem cells as a strategy of promoting cell survival and functional benefits after transplantation into the ischemic rat brain. Exp. Neurol. 210, 656–670. Vaudry, D., Stork, P.J., Lazarovici, P., Eiden, L.E., 2002. Signaling pathways for PC12 cell differentiation: making the right connections. Science 296, 1648–1649. Vendrame, M., Gemma, C., de Mesquita, D., Collier, L., Bickford, P.C., Sanberg, C.D., Sanberg, P.R., Pennypacker, K.R., Willing, A.E., 2005. Anti-inflammatory effects of human cord blood cells in a rat model of stroke. Stem Cells Dev. 14, 595–604. Zaheer, A., Zhong, W., Lim, R., 1995. Expression of mRNAs of multiple growth factors and receptors by neuronal cell lines: detection with RT-PCR. Neurochem. Res. 20, 1457–1463.