The experimental protocols used in this study were reviewed and approved by the Institutional Animal Care and Use Committee of the Samsung Biomedical Research Institute, Seoul, Korea. This study was also performed in accordance with institutional and National Institutes of Health guidelines for laboratory animal care.

Pregnant Sprague-Dawley female rats (Daihan Biolink, Seoul, Korea) were used for the in vitro experiments. Each experiment was repeated two times. For each experiment, four pregnant rats were sacrificed. Living embryos were excised by caesarian section performed under anesthesia with halothane. Each pregnant rat had between 10 to 12 embryos. Cortical neurons were prepared from rat embryos. Brains were removed from rat embryos at embryonic day (E)17, and the cortical area was dissected. Cortices were carefully collected and dissected free from the meninges. The rationale for choosing E17 is that generation of cortical cell types occurs in temporally distinct phases. The ventricular zone arises first, and cells from this area develop mainly into neurons. Ventricular zone neurogenesis peaks at E14 and recedes at E17 in rats. In contrast, cells originating from the subventricular zone at late embryonic days and early postnatal days (rat E17 to postnatal day [P]14) are destined predominantly for glial lineages29,³⁰⁾.

The culture methods were similar to those described previously31,³²⁾. Briefly, cortical tissues from E17 rat embryos were incubated in Dulbecco's phosphate-buffered saline (D-PBS; JBI, Daegu, Korea) containing 0.1% trypsin (Gibco, Rockville, MD, USA) for 15 minutes in a 37℃ water bath. It contained 40 µg/mL deoxyribonuclease I (DNase I; Sigma-Aldrich Co., St. Louis, MO, USA). Digestion was terminated by the addition of 20% fetal bovine serum. The tissue pieces were triturated into single cells by gentle pipetting through a fire-polished Pasteur pipette. Dissociated cells were suspended in neurobasal medium (Gibco) containing 2% B-27 supplement (Gibco). This serum-free combination medium is widely available for primary cultures of neurons and neuronal cell lines³³⁾. Cells were seeded in 24-well culture plates pretreated overnight at 37℃ with 100 µg/mL poly-D-lysine and 4 µg/mL laminin. The final cell density was approximately 2×10⁵ cells/cm² in 2 mL/well of the culture medium. Cells were incubated at 37℃ in a 5% CO₂/95% air atmosphere for 10 days. Culture medium was changed every 2 days in vitro and studies were performed at in vitro days 10 to 11. Each of the experimental groups was tested in four culture media (n=4), and 10 microscopic areas were counted for each medium tested.

L-carnitine (Sigma-Aldrich Co.) was applied to cultures just before they were exposed to OGD, and levels of L-carnitine were maintained during OGD. L-carnitine was also applied to cultures before reoxygenation because the culture medium was changed. L-carnitine was diluted with distilled water and applied to cells at concentrations of 1, 10, or 100 µM.

OGD was performed as described in detail elsewhere34,³⁵⁾. Exposure of rat cortical cultures to OGD was performed at in vitro day 10. The initial culture medium, Dulbecco's Modified Eagle's Medium (DMEM), was replaced by a solution containing (mmol/L): NaCl 130, KCl 5.4, CaCl₂ 1.8, NaHCO₃ 26, and MgSO₄ 0.8, NaH₂PO₄ 1.18, and bubbled with 95% N₂/5% CO₂ for OGD cells (OGD solution). OGD cells were transferred to an anaerobic chamber containing 95% N₂/5% CO₂, humidified at 37℃, and maintained at a constant pressure of 0.015 bar. Cells were exposed to OGD for 60 minutes. OGD was terminated by replacing the exposure medium with serum-free DMEM containing 2% B27 supplement without antioxidants (Gibco) and were returned to the normoxic incubator. Control cultures in a solution identical to the OGD solution but containing glucose (33 mmol/L; control solution) were kept in the normoxic incubator for the same time period as the OGD experiment, and then the incubation solution was replaced with reperfusion buffer and cultures were returned to the normoxic incubator. To explore the neuroprotective effect of L-carnitine, cultures were treated with L-carnitine (1, 10, and 100 µM) during OGD and reoxygenation. Cell viability and cell death were evaluated 24 hours after OGD and ROS detection was evaluated 3 hours after OGD.

Cultures were stained with acridine orange (AO) and propidium iodide (PI) to quantify cell survival of primary cortical neurons. Untreated and L-carnitine-treated cells were washed with phosphate buffered saline (PBS) and stained for 10 minutes at 37℃ with a mixture of AO (15 µg/mL; Sigma-Aldrich Co.) and PI (4.6 µg/mL; Sigma-Aldrich Co.). AO is a nucleic acid-selective fluorescent cationic dye that is used for cell cycle determination. It is cell-permeable and interacts with both DNA and RNA. PI is an intercalating agent and a fluorescent molecule used to stain DNA. It can be used to differentiate necrotic, apoptotic, and normal cells. Neuronal injury curtails AO staining and facilitates PI penetration and interaction with DNA to yield a bright red fluorescent complex. Viability was expressed as the ratio of cells that stained with PI to those stained with AO (PI/AO). The concentrations of PI/AO were modified to visualize living and dead cells simultaneously³⁶⁾. Immediately after staining, cells were examined directly with an E600 fluorescence confocal microscope (Nikon, Tokyo, Japan) and pictures were taken. Cells stained with AO and PI were counted in random microscopic fields. Ten images were selected randomly from each sample, and the mean number of cells stained with AO and PI was determined. Cells were counted using a magnification of ×400.

The 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) cell proliferation assay is a standard colorimetric assay for measuring cellular growth and viability. It can also be used to determine the cytotoxicity of potential medicinal agents and other toxic materials. The yellow, water-soluble tetrazolium MTT is reduced to a blue, insoluble formazan product in mitochondria of viable cells by succininate-tetrazolium dehydrogenase. After incubation, a solubilization solution is then added to dissolve the insoluble purple formazan product and the amount of dye solubilized is quantitated by measuring the optical density at 570 nm.

After OGD, the cultured cells were washed twice with D-PBS and a 10 µL MTT (Roche Diagnostics Co., Indianapolis, IN, USA) solution (5 mg/mL PBS; Molecular Probes Inc., Eugene, OR, USA) was added. After a 3-hour incubation (37℃, 5% CO₂), the MTT solution was aspirated and replaced with 200 µL dimethylsulfoxide (DMSO; Sigma-Aldrich Co.) to dissolve the formazan crystals produced by MTT reduction. The reaction solutions were transferred to 96-well culture microplates and the absorbance was read at 570 nm on an automated microplate reader using SOFTmax program ver. 5.2 (Molecular Device Co., Sunnyvale, CA, USA) to evaluate the degree of cell activation in response to various doses of L-carnitine. Results are expressed as the mean percent increase over background levels based on two experiments.

Cell injury was confirmed by measuring the amount of LDH released into the extracellular fluid by damaged or destroyed cells 24 hours after OGD reoxygenation in the following groups: control, OGD, and pretreatment with L-carnitine followed by OGD exposure. A series of control experiments showed that the specific efflux of LDH induced by OGD reoxygenation was linearly proportional to the number of neurons damaged or destroyed, and that no specific LDH efflux occurred when pure cultured cortical glial cells were similarly exposed to OGD. Cytotoxicity (%) was calculated as follows: (measured LDH activity-LDH activity of control) / (full kill LDH activity-LDH activity of control)×100. "Full kill" means that a control culture was lysed totally for 10 minutes with 1% Triton X-100 (Sigma-Aldrich Co.). After neuronal cell cultures were exposed to OGD reoxygenation for 24 hours, 100 µL of culture medium aspirated from each well was transferred to a 96-well plate, and was then incubated with a 100 µL reaction mixture in the dark for 10 minutes at room temperature. The amount of LDH released was determined by measuring the optical density at 490 nm.

Molecular oxygen is necessary to sustain life, but ROS produced as a metabolic product due to oxygen consumption are paradoxically highly toxic to cells. Oxidative stress was assessed by direct measurement of ROS. Intracellular ROS levels were monitored using the fluorescent dye 2',7'-dichlorodihydrofluorescein diacetate (DCF-DA; Sigma-Aldrich Co.). DCF-DA, which is oxidized to fluorescent DCF by hydroperoxide, was used to measure relative levels of cellular peroxides³⁷⁾. After treatment, cultured cells were washed twice with D-PBS, and incubated with 10 µM DCF-DA in medium for 30 minutes. The cells were then washed with D-PBS. Fluorescence and differential interference contrast images were taken using an Axiovert 200 inverted microscope (Carl Zeiss, Jena, Germany). The intensity of green fluorescence was quantified using image analyzer software. Ten images were taken of each sample, and the mean ROS level of each sample was calculated.

Terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick end labeling (TUNEL) is a common method for detecting DNA fragmentation produced by apoptotic signaling cascades. The assay relies on the presence of nicks in the DNA which can be identified by terminal deoxynucleotidyl transferase, an enzyme that will catalyze the addition of dUTPs that are secondarily labeled with a marker. This assay can also be used to label cells undergoing necrosis or cells that have suffered from severe DNA damage. The number of apoptotic OGD-stimulated neurons was determined by immunofluorescent TUNEL performed using the Apoptag fluorescein in situ detection kit (Chemicon International, Temecula, CA, USA), which detects the 3'-OH region of cleaved DNA during apoptosis. After OGD, cells were washed with D-PBS and fixed on chamber slides with 4% paraformaldehyde at 4℃ for 30 minutes. Cells were incubated in a permeabilization solution (0.2% Triton X-100 and 0.5% bovine serum albumin in D-PBS, pH 7.4) for 10 minutes. TdT reaction reagent was added to the cells for 1 hour at 37℃ and Avidinfluorescein isothiocynate II solution for 30 minutes. PI was used to counterstain stained cells. Slides were washed in PBS, mounted with Vectashield mounting solution with 4,6-diamidino-2-phenylindole (H-1200; Vector Laboratories Inc., Burlingame, CA, USA), and visualized with a Nikon E600 fluorescent microscope using an excitation wavelength of 460 to 490 nm. Apoptotic TUNEL-positive cells were counted under a fluorescence microscope with the viewer blinded to the identity of the treatment group. For each experiment, 10 random fields at 400× magnification were observed. The number of TUNEL-positive cells was initially expressed as the ratio of the number of green to red cells (TUNEL/PI). TUNEL/PI double staining flow cytometry was performed to assess the apoptotic rate and to determine the phase of the apoptotic cells³⁸⁾. The number of TUNEL-positive cells was expressed as the number of green cells divided by the number of total (red and green) cells.

All data are given as the mean±standard error of the mean. The statistical significance of differences between treatment groups was determined by one-way analysis of variance using the Tukey-Kramer multiple comparison test, and T-tests using Student-Newman-Keuls comparisons. All statistical analyses were performed using GraphPad Instat ver. 3.05 (GraphPad Software Inc., La Jolla, CA, USA). P values <0.05 were assumed to be significant.

To examine whether L-carnitine can reduce neuronal injury after OGD, cultured primary cortical neurons were evaluated by AO and PI staining. Fig. 1 shows the cell numbers stained with the fluorescent dyes AO and PI in each group. The mean number of cells that stained with AO and PI were 171.50±17.36 and 2.00±0.68 in the control group, respectively, and 86.50±5.88 and 74.67±7.95 in the OGD group, respectively. In the OGD groups treated with L-carnitine, the number of cells that stained with AO and PI were 92.83±4.45 and 86.83±3.55 in the OGD group treated with 1 µM L-carnitine, respectively, 129.20±2.94 and 77.00±1.08 in the OGD group treated with 10 µM L-carnitine, respectively, and 138.00±2.16 and 81.67±2.37 in the OGD group treated with 100 µM L-carnitine, respectively. As shown in Fig. 1, the number of cells that stained with AO in the OGD groups treated with L-carnitine was proportional to the dose of L-carnitine. At lower doses of L-carnitine, especially 1 µM, the number of cells stained with AO was not increased, whereas at higher doses of L-carnitine, the number of cells that stained with AO increased. However, the differences in staining between groups were not statistically significant. The viability of neuronal cells was evaluated by calculating the PI/AO ratio, which indicates the number of dying cells. The results are shown in Fig. 2. The percentage of dying cells was significantly lower in the groups treated with L-carnitine, especially at concentrations of 10 µM and 100 µM (66.53±2.08% in the OGD group treated with 10µM L-carnitine, 55.80±0.09% in the OGD group treated with 100 µM L-carnitine vs. 85.51±5.22% in the untreated OGD group; P<0.05). However, the PI/AO ratio was not significantly different between the untreated OGD group and the group treated with 1 µM L-carnitine (93.82±2.42% in the OGD group treated with 1 µM L-carnitine vs. 85.51±5.22% in the untreated OGD group). Fig. 3 shows the morphology of neuronal cells after OGD treatment. Higher doses of L-carnitine reduced cellular death compared to lower doses of L-carnitine.

As shown in Fig. 4, exposure to OGD resulted in decreased cell viability compared to the control group as assessed using the MTT assay. Relative to the 100.00±0.00% viability of the control group, the cell viability was 23.24±1.93% in the untreated OGD group, 20.15±1.45% in the OGD group treated with 1 µM L-carnitine, 29.04±1.02% in the OGD group treated with 10 µM L-carnitine, and 35.86±2.54% in the OGD group treated with 100 µM L-carnitine. Thus cells treated with L-carnitine prior to OGD group were significantly more viable than cells not pre-treated with L-carnitine. Treatment of neuronal cells with L-carnitine followed by OGD exposure resulted in dose-dependent protection. In particular, the increase in cell viability in the OGD group treated with 100 µM L-carnitine was significantly greater than that of the untreated OGD group (P<0.05).

LDH leakage corresponds to cytotoxicity and membrane damage. OGD caused cell death, as shown by the increased LDH efflux to the medium. The LDH leakage results confirmed the MTT assay results. The percentage cytotoxicity based on LDH measurements was 46.04±1.70% in the untreated OGD group, 35.18±1.28% in the OGD group treated with 1 µM L-carnitine, 35.42±2.44% in the OGD group treated with 10 µM L-carnitine, and 30.09±1.59% in the OGD group treated with 100 µM L-carnitine (Fig. 5). There was a significant decrease in LDH leakage in all OGD groups treated with L-carnitine compared to the untreated OGD group.

To investigate the involvement of free radical-mediated oxidative stress during OGD injury, we measured ROS levels in neuronal cell cultures. ROS formation can lead to protein oxidation, lipid peroxidation, DNA damage, and eventually cell death. Intracellular ROS levels were monitored using the fluorescent dye DCF-DA. Fig. 6A shows the accumulation of ROS in neuronal cells exposed to OGD. Treatment with L-carnitine inhibited ROS accumulation in neuronal cells after OGD. A significant decrease in ROS formation was found in the OGD groups treated with 10 µM or 100 µM L-carnitine compared to the control group (P<0.05) (Fig. 6A). ROS levels were 29.89±20.35 in the control group and 8,049.71±770.07 in the OGD group that did not receive L-carnitine treatment, while the ROS level was 7,837.50±654.90 in the OGD group treated with 1 µM L-carnitine, 4,838.00±294.82 in the OGD group treated with 10 µM L-carnitine, and 3,770.00±419.98 in the OGD group treated with 100 µM L-carnitine (Fig. 6B).

Neuronal viability and apoptosis were also assessed using the TUNEL assay. Fig. 7 shows TUNEL-positive cells after treatment of cells with OGD. The number of TUNEL-positive (green) cells in OGD groups treated with L-carnitine decreased in a dose-dependent manner compared to the number of TUNEL-positive cells in the OGD group that was not treated (Fig. 8). However, the differences between groups were not significant. The mean number of cells that stained with PI and were TUNEL-positive was 150.50±14.44 and 0.63±0.24 in the control group, respectively, and 100.70±5.90 and 7.60±0.64 in the untreated OGD group, respectively. In the OGD groups treated with L-carnitine, the mean number of cells that stained with PI and were TUNEL-positive were 110.70±5.49 and 6.30±1.03 in the OGD group treated with 1 µM L-carnitine, respectively, 112.10±7.58 and 8.56±0.82 in the OGD group treated with 10 µM L-carnitine, respectively, and 119.10 7.63 and 4.90±0.72 in the OGD group treated with 100 µM L-carnitine, respectively. The incidence of TUNEL-positive cells was expressed as the ratio of green cells to red cells (TUNEL/PI). TUNEL/PI was 0.48±0.22% in the control group, 7.72±0.65% in the untreated OGD group, 5.76±0.91% in the OGD group treated with 1 µM L-carnitine, 7.90±0.88% in the OGD group treated with 10 µM L-carnitine, and 4.12±0.59% in the OGD group treated with 100 µM L-carnitine (Fig. 9). There was a significant decrease in the TUNEL/PI ratio in the OGD groups treated with 100 µM L-carnitine compared to the untreated OGD group (P<0.05). The differences in the TUNEL/PI ratio were not significantly different among the other groups. The results indicated that a high dose of L-carnitine treatment reduces apoptosis of cortical cells after OGD.