Biomarkers In Patients With Systemic Lupus Erythematosus Using Proteomic Methods

PART one:Proteomic profiling of cerebrospinal fluid identifies biomarkers for neuropsychiatric systemic lupus erythematosusObjective:To identify neuropsychiatric systemic lupus erythematosus (NPSLE) specific cerebrospinal fluid (CSF) biomarkers and construct a decision tree model for NPSLE classification. Methods:we compared the proteomic profile of CSF from39patients with neuropsychiatric systemic lupus erythematosus (NPSLE),10SLE patients without neuropsychiatric symptoms (non-NPSLE) and17patients with scoliosis, using matrix assisted laser desorption/ionization time of flight mass spectrometry (MALDI-TOF-MS) combined with weak cationic exchange magnetic beads. We developed a decision tree model for NPSLE classification with biomarker patterns software, and tested the model with other independent CNS diseases (18samples). Results:We identified25mass-to-charge (m/z) peaks with statistically significant (p<0.05) differences between NPSLE (before treatment) patients and control group1(non-NPSLE). Of them, the m/z peaks at3237,4312,4514,4550,6688,6378,6822,8349,8595,11744and16230were used to construct a decision tree model to distinguish NPSLE from non-NPSLE patients with a sensitivity of92.1%and a specificity of100.0%based on learning data, with a specificity of85.7%based on external blind testing data. There were18m/z peaks with statistically significant (p<0.05) differences between before and after treatment in the same group of NPSLE patients. Of them4550,6822and8595decreased, while in a refractory case, they did not decrease. Conclusion: Our data suggested a potential application of MALDI-TOF-MS combined with weak cationic exchange magnetic beads to profile CSF proteome, several potential CSF biomarkers were indicated to detect, diagnose and monitor NPSLE, a decision tree model for NPSLE classification was set up with high sensitivity and specificity. PART two:Identifying one protein in the proteomic signature as ubiquitinBackground and Objective:Ubiquitin is a protein containing76amino acid protein in eukaryotic cells. The main function is to mark proteins need to be broken down,then be hydrolyzed. A number of studies identified CSF by MALDI-TOF-MS that mass to charge ratio8.6kda protein peak corresponding to the ubiquitin. We undertook this study to prove the presence of m/z peak of8.6kda in CSF profile as ubiquitin.Methods:CSF was immunoprecipitated with rabbit anti(Hu)-ubiquitin antibody, after immunoprecipitation, the proteins of supernatant and agarose beads were detected byMALDI-TOF-MS and western blot, respectively. Results:The m/z peak of8.6kda in CSF profile was validated to be ubiquitin protein by both immunoprecipitation combining with MALDI-TOF-MS&WCX magnetic beads, immunoprecipitation combining with western blot and anti-ubiquitin ELISA. Conclusion:CSF protein mass spectra peaks corresponding to charge ratio8.6kda protein ubiquitin protein, neuropsychiatric systemic lupus erythematosus (NPSLE) patients with cerebrospinal fluid ubiquitin protein (ubiquitin) elevated, suggesting that the ubiquitin protein (ubiquitin) may be biomarkers of NPSLE。 PART three:Proteomics profiles of Methylprednisolone Pulse Therapy in the treatment of Systemic Lupus Erythematosus flaresObjective:To identify serum biomarkers of systemic lupus erythematosus (SLE) before and after methylprednisolone pulse therapy and construct a decision tree model for NPSLE classification. Methods:we compared the proteomic profile of serum from38patients with SLE flares before and14days after methylprednisolone pulse therapy, using matrix assisted laser desorption/ionization time of flight mass spectrometry (MALDI-TOF-MS) combined with weak cationic exchange magnetic beads. Results:Before the treatment, there were8discriminating peaks between effective group and ineffective group, and we used them to construct a decision tree model to distinguish effective from ineffective patients with a sensitivity of92.6%and a specificity of100%based on learning data. After14days of there were25discriminating peaks between effective group and ineffective group, and we used them to construct a decision tree model to distinguish effective from ineffective patients with a sensitivity of96.3%and a specificity of100%based on learning data. And we found18discriminating peaks between group before treatment and group after treatment, and construct a decision tree model to distinguish effective from ineffective patients with a sensitivity of84.2%and a specificity of89.5%based on learning data. Conclusion:Our data suggested a potential application of MALDI-TOF-MS combined with weak cationic exchange magnetic beads to profile serum proteome, several potential biomarkers were indicated to diagnosis and management of SLE, a decision tree model for NPSLE classification was set up with high sensitivity and specificity.