Screening anti-inflammatory compounds in injured spinal cord with microarrays: a comparison of bioinformatics analysis approaches.

Inflammatory responses contribute to secondary tissue damage following spinal cord injury (SCI). A potent anti-inflammatory glucocorticoid, methylprednisolone (MP), is the only currently accepted therapy for acute SCI but its efficacy has been questioned. To search for additional anti-inflammatory compounds, we combined microarray analysis with an explanted spinal cord slice culture injury model. We compared gene expression profiles after treatment with MP, acetaminophen, indomethacin, NS398, and combined cytokine inhibitors (IL-1ra and soluble TNFR). Multiple gene filtering methods and statistical clustering analyses were applied to the multi-dimensional data set and results were compared. Our analysis showed a consistent and unique gene expression profile associated with NS398, the selective cyclooxygenase-2 (COX-2) inhibitor, in which the overall effect of these upregulated genes could be interpreted as neuroprotective. In vivo testing demonstrated that NS398 reduced lesion volumes, unlike MP or acetaminophen, consistent with a predicted physiological effect in spinal cord. Combining explanted spinal cultures, microarrays, and flexible clustering algorithms allows us to accelerate selection of compounds for in vivo testing.