|
Cytokine multiplex analysis utilizing the Luminex platform measuring up to 80 analytes has become routine in nonclinical safety assessment of biotherapeutics. On September 14th Lovelace Biomedical is presenting an e-chalk talk with the American Association of Pharmaceutical Scientists (www.aaps.org) that is free to ACT members. The presentation will explore the analytical strengths and weaknesses of the platform, specifically the pre-analytical, analytical, and post analytical factors that impact the ability to correlate cytokine multiplex data with adverse toxicology findings.
Multiplex analysis generates large cytokine data sets; allowing for a broad view of cytokine expression patterns that may be associated with key toxicity findings including cytokine storm, allergy, unintended immunosuppression, dysregulation of T helper cell subset activity and antibody response, acute severe inflammation, and chronic inflammation. Despite the analytical advantages of small sample volume and large number of analytes in a single panel measured with acceptable precision and accuracy the analysis of cytokine data sets is often performed with standard statistical analysis such as ANOVA with post hoc analysis to detect statistically significant differences between individual cytokines. This analytical approach does not statistically infer or correlate cytokine response to toxicology findings. An alternative approach is presented to address this issue, the rationale that what is needed is a simple, easily implemented approach to correlate changes in cytokine concentration, with cytokines organized by functional group.
The proposed approach is to group key cytokines by activity, subgroup by opposing function, selecting cytokines already shown to be increased in patients with cytokine storm, allergy, autoimmune disease, and lastly the cytokines that are modulated by the test article. Proposed data analysis approach is to compute group median concentration by interval, then group median ratio of change over interval, follow by ratio of change in cytokine groups of opposing action. The cytokine data organized in this way would be used in a modified odds ratio calculation and calculation of the upper and lower 95% confidence interval for the ratios. Those cytokines or cytokine tiers above or below the 95% confidence interval limits would be identified and plotted against the number of toxicology findings or adverse events over the study intervals to associate changes in cytokine concentrations and toxicity.
John Farmer, PhD
|