Protein array autoantibody profiles to determine diagnostic markers for neuropsychiatric systemic lupus erythematosus

The aim was to investigate the association between autoantibodies (autoAbs) and neuropsychiatric (NP) involvement in patients with SLE and to evaluate whether any autoAb or a combination of these autoAbs could indicate the underlying pathogenic process.

Using a multiplexed protein array for 94 antigens, we compared the serum autoAb profiles of 69 NPSLE patients, 203 SLE patients without NP involvement (non-NPSLE) and 51 healthy controls. Furthermore, we compared the profiles of NPSLE patients with clinical inflammatory (n = 38) and ischaemic (n = 31) NP involvement.

In total, 75 IgG and 47 IgM autoAbs were associated with SLE patients in comparison with healthy controls. Comparing NPSLE with non-NPSLE and healthy control sera, 9 IgG (amyloid, cardiolipin, glycoprotein 2, glycoprotein 210, heparin, heparan sulphate, histone H2A, prothrombin protein and vimentin) and 12 IgM (amyloid, cardiolipin, centromere protein A, collagen II, histones H2A and H2B, heparan sulphate, heparin, mitochondrial 2, nuclear Mi-2, nucleoporin 62 and vimentin) autoAbs were present at significantly different levels in NPSLE. The combination of IgG autoAbs against heparan sulphate, histone H2B and vimentin could differentiate NPSLE from non-NPSLE (area under the curve 0.845, 99.97% CI: 0.756, 0.933; P < 0.0001). Compared with non-NPSLE, four IgG and seven IgM autoAbs were significantly associated with inflammatory NPSLE. In ischaemic NPSLE, three IgG and three IgM autoAbs were significantly different from non-NPSLE patients.

In our cohort, the presence of high levels of anti-heparan sulphate and anti-histone H2B combined with low levels of anti-vimentin IgG autoAbs is highly suggestive of NPSLE. These results need to be validated in external cohorts.