Background/Purpose: Serum 14-3-3η is an RA diagnostic marker and higher levels are associated with joint damage and more severe disease. Mechanistically, 14-3-3η upregulates factors involved in joint damage and inflammation, including TNFα. 14-3-3η’s extracellular expression in RA elicits a robust, early auto-antibody response to the native protein, which may confer a protective effect. We investigated the relationship between the 14-3-3η protein as well as its auto-antibodies and response to anti-TNF therapy, in established RA.

Methods: Serum 14-3-3η protein and its auto-antibodies were measured at baseline in 99 RA patients who were refractory to either a standard DMARD or a first TNF-blocker. 14-3-3η protein was measured by ELISA and the auto-antibody levels by Meso-Scale-Discovery electrochemiluminescent platform. The changes in clinical response variables (within 15 weeks of treatment initiation) were determined and median group differences between patients who were 14-3-3η protein positive (> 0.19 ng/ml) and negative or 14-3-3η auto-antibody positive (> 380 U/ml) and negative were tested using the Mann-Whitney U-Test. Regression analyses were used to test whether the 14-3-3η markers were independent predictors of a good EULAR response (DAS28 < 3.2, ΔDAS > 1.2) or DAS remission (DAS-ESR<2.6) and contingency analyses for frequency distribution of biomarker positivity and response were conducted. Spearman correlations were performed to evaluate the relationship between the expression of the 14-3-3η protein and its auto-antibodies.

Results: At baseline, 88 of 99 (89%) patients were positive for 14-3-3η and 11 of 99 (11%) were positive for the 14-3-3η auto-ab. Median (IQR) and mean (SD) 14-3-3η serum levels were 2.1 (0.7 – 15.7) ng/ml and 6.6 (8.1) ng/ml with median and mean auto-ab titres being 216 (172 – 311) U/ml and 248 (103) U/ml. At 15 weeks, median ESR (29 vs. 17, p < 0.02) and CRP (7.3 vs. 2.9, p < 0.03) were significantly higher in 14-3-3η +ve patients. 14-3-3η auto-ab +ve patients achieved a significantly lower DAS (4.2 vs 4.7, p < 0.03), a greater change in DAS (2.3 vs 1.7, p < 0.03), a lower ESR (29 vs. 15, p=0.05), and a lower CRP (6.8 vs. 2.9, p=0.05) than 14-3-3η auto-ab -ve patients. Contingency analysis revealed that 14-3-3η negativity was predictive of achieving a good EULAR response with an LR of 5.0, p<0.03, OR=4.8 (1.3 - 18.1) and RR=3.1 (95%CI, 1.4-7.0). 14-3-3η auto-ab positivity marked a higher likelihood of achieving both a good EULAR response [LR = 5.0, p< 0.03, OR=0.2 (95%CI, 0.1-0.8), RR=3.1 (1.4-7.0)] and DAS remission [LR = 4.0, p< 0.05, OR=0.2 (0.1 - 0.9), RR=3.2 (1.4-8.5)]. The multivariate logistic regression returned both 14-3-3η and its auto-antibodies as significant independent predictors of a good EULAR response (p<0.02). The low Spearman correlation (r = -0.13, p = 0.2) between serum 14-3-3η and its auto-antibodies, supports their combined utility to inform therapy response.

Conclusion: RA patients who are negative for 14-3-3η and positive for its auto-antibodies are more likely to achieve a Good EULAR response and DAS remission with anti-TNF therapy.

ACR Meeting Abstracts - https://acrabstracts.org/abstract/14-3-3%ce%b7-and-its-auto-antibodies-predict-response-to-anti-tnf-therapy/