Lipopolysaccharide (LPS; Escherichia coli O111:B4) was purchased from Sigma-Aldrich (St. Louis, USA). Culture multi-wells and pipettes were obtained from Orange Scientific (Graignette, Belgium). Cultured media, fetal bovine serum (FBS) and antibiotics were purchased from Gibco (Invitrogen Corporation, USA). Other reagents were purchased from Sigma-Aldrich unless stated otherwise.

Spinal cord neuronal/glial cultures were prepared from embryonic Sprague-Dawley (SD) rat spinal cords at gestation days 14 - 16 as described in Hung [9] and Tsai [10,11]. Briefly, cells were dissociated with mixtures of papain/protease/deoxyribonuclease I (0.1:0.1:0.03%) and plated in poly-lysine coated multiwell plates (12 × 24 mm). Cells were maintained in a Dulbecco’s Modified Eagle medium (DMEM, Invitrogen) supplemented with 10% FBS (Gibco BRL, Gaithersburg, MD) at 37˚C in a water-saturated atmosphere of 5%CO₂/95% air. Neurons in neuronal/glial cultures were identified by immunoreactivity to β-III tubulin (Covance, USA) and to synaptic vesicle protein 2 (SV2) (Developmental Studies Hybridoma Bank, USA). Oligodendrocytes were identified by immunoreactivity to myelin basic protein (MBP, Chemicon, USA). Microglial cells were isolated from confluent glial cell cultures as described previously [12]. Briefly, floating cells and weakly attached cells on the confluent mixed glial cell layer [13,14] were isolated by gently shaking the flask for 1 - 2 hr at 180 rpm. The resulting cell suspension was transferred to culture dish and allowed to adhere at 37˚C. Unattached cells were removed after 30 min, microglia were isolated as strongly adhering cells. After seeding (1 day), the enriched microglial cultures were 99% positive for Iba-1 (Wako, Japan) or ED-1 (CD68, Serotec, USA) as assessed by immunostaining.

Medium of enriched microglia was replaced with fresh medium. Two days later, the conditioned media was collected, cleared from free-floating cells by centrifugation for 5 min at 1000 × g, and sterile filtered. For fractionation of conditioned medium, MCM was further passed through an Amicon Ultracentrifugal filter with 50 kDa cut-off (Millipore, USA) by centrifugation 3000 g at 4˚C. The resulted two fractions (>50 kDa or <50 kDa) of MCM were saved and frozen until use. For cell survival test, spinal cord neuronal/glial cultures were maintained in the presence or absence of MCM for 5 days and then fixed for immunostaining. For examining the effect of LPS stimulation, cultures were treated with MCM in the presence or absence of LPS (1.2 µg/ml) on the 2nd or 3rd day after seeding. Two days later, cultured medium was collected for Griess assay of released nitrite, an indicator of nitric oxide, and the cells were fixed for immunostaining.

SCI was induced in adult SD rats using the New York University weight-drop device as described previously [15]. Briefly, adult female SD rats (250 ± 20 g) were anesthetized with isoflurane and a laminectomy was performed at thoracic (T) spinal cords. A contusive injury at rat T 9 - 10 was induced by dropping a 10 g rod from a height of 25 mm. Within 30 min after contusive injury, the rats received medium or MCM injection. A 5 ml Exmire microsyringe with a 31-gauge needle was positioned near the midline of the cords, 2 mm rostral and caudal to the contusive center. Twenty microliters of MCM (5 ml each; 4 injections) was intrathecally administered to the thoracic spinal cord (~1 mm ventral to the dura surface) within 30 min after eliciting contusive injury. Following injury and treatment, the incision was closed and sutured. Each rat was then returned to its cage. To avoid urinary tract infections, manual emptying of the urinary bladder was carried out twice daily. All surgical interventions and animal care were performed in accordance with the Laboratory Animal Welfare Act, the Guide for the Care and Use of Laboratory Animals, National Institutes of Health and were approved by the Animal Committee of Taipei Veterans General Hospital, Taiwan.

Rats were allowed to recover for 6 weeks post-injury. This time course allowed for examining the effect of MCM during the initial deficit in the behavioral locomotor assay and during the plateau phase of recovery after trauma. Recovery of hindlimb stepping was assessed while subjects moved freely about an open field. The Basso, Beattie and Bresnehan (BBB) open field score was used to evaluate locomotion in terms of hindlimb functional improvement of the SCI rats as previously described [11]. In this study, behavior analyses of SCI rats were conducted weekly until 6 weeks (at least 6 rats per group). The behavior tests were recorded by a video camera, and two examiners were blind to each behavior evaluation group. BBB score ranges from 0 (no hindlimb movement) to 21 (normal movementcoordinated gait). At 6 weeks following SCI and treatment, the animals were sacrificed for morphological examinations.

NO production was assessed using the Griess reaction. The production of NO was assayed as the accumulation of nitrite in the medium using Griess reagents (1% sulfanilaminde/0.1% naphthylethylene diamine dihydrochloride/2% phosphoric acid) as described by Tsai et al. [11]. Briefly, the culture medium was mixed with Griess reagents and incubated at room temperature for 10 min. The absorbance of the resultant products was measured at 540 nm. Sodium nitrite (NaNO₂) was used as a standard to calculate nitrogen dioxide () concentrations.

Cultured cells were fixed in 4% paraformaldehyde at room temperature for 30 min. Cells were further permeabilized with 0.2% Triton X-100, blocked with 2% serum and immunostained with primary antibodies and with the respective fluorescently tagged secondary antibodies (Jackson ImmunoResearch Inc.). Primary antibodies included mouse anti-inducible nitric oxide synthase (iNOS) (BD Bioscience, USA), rabbit anti-bIII tubulin (covance, USA) and rabbit anti-MBP (Chemicon, USA). Images of cultured cells were obtained with a fluorescent microscope equipped with fluorescence optics and with a CCD camera. Animals were perfused with 4% paraformaldehyde at 6 weeks postinjury. The spinal cords (approximately 2 mm length with the contused site at the center of the sample) were removed and processed for longitudinal sectioning at a thickness of 20 mm. Immunostaining was performed as described previously [15]. Briefly, tissue sections were incubated with anti-Neurofilament L (NF-L; Chemicon, Temecula, CA) or ED-1 (Serotec, USA) antibodies overnight at 4˚C. The bound antibodies were visualized using the avidin-biotin-peroxidase complex (Vectastain Elite ABC kit; Vector Laboratories Inc., CA, USA) and with appropriate chromagens.

Experimental data were expressed as the mean of independent values ± S.E.M. One-way ANOVA followed by Bonferroni’s t-test was used to determine statistical differences between treatments. A statistically significant difference was accepted at P < 0.05.

Enriched microglial cells were purified from mixed glial cultures. As shown in Figures 1(A) and (B), more than 99% of cells were immunoreactive to ED-1, a microglial marker. MCM was collected from such enriched microglia. Spinal cord neuronal/glial cultures were treated with MCM or PBS, as a vehicle control, for 5 days. Cultures were then processed for immunohistochemical staining.

More neurite connections of bIII were observed in MCM-treated cells (Figure 1(D)) than in control cultures (Figure 1(C)). Concurrently, the density of synaptic vesicle protein 2 (SV2)-IR was higher in MCM-treated cells than in control cells (Figures 1(E)-(F)). Interestingly, MCM treatment also increased the process of MBP (+) oligodendrocytes (Figure 1(G) and (H)). Spinal cord neuronal/glial cultures were stimulated with endotoxin LPS (1.2 µg/ml) in the presence or absence of MCM for 2 days. Cells were then processed for iNOS immunostaining and medium were collected for nitrite detection. As shown in Figure 2, LPS induced increase of iNOS-IR cells as well as nitrite release to the medium. MCM treatment significantly reduced LPS stimulation in spinal cord cell cultures (Figures 2(A)-(D)).

The hindlimb motor function of injured rats receiving various treatments was monitored according to the guideline of BBB Locomotor Rating Scale [16,17]. As shown in Figure 3, the rats receiving MCM > 50 kDa started to show significant difference with BBB score at 2 weeks (w) after SCI, when compared to MCM < 50 kDa group or to the control group that underwent the same procedure, but received medium treatment. Moreover, statistically significant differences in BBB scores (P < 0.01) between MCM > 50 kDa-treated rats and MCM < 50 kDa-treated or vehicle-treated rats were found at 2 w, 3 w, 4 w and 6 w after contusive injury and treatment. Morphological analysis of the spinal cord was conducted at 6 week post-injury. The immunohistochemical results of the injured epicenter from each group were shown in Figure 4. Contusive spinal injury increased ED-1-IR cells but reduced NF-L-IR cells. ED-1-IR cells were activated microglia/macrophage exhibiting a fully functional “amoeboid” phenotype with the ability to perform phagocytosis and generate free radicals. MCM treatment (>50 kDa or <50 kDa in Figures 4(B), (C), (E), & (F)) reduced ED-1-IR cells and preserved NF-L-IR nerve fibers in the epicenter when compared to vehicle-treated SCI rats (Figures 4(A) & (D)). Particularly, the injured epicenter of MCM > 50 kDa-treated rat preserved more

NF-L-IR fibers (Figure 4(F)).