From that analysis it was possible to identify an impaired binding capacity of sHLA-G circulating molecules to LILRB1 in RA patients, suggesting an impaired functionality of these molecules regarding this receptor. The first important finding was that circulating sHLA-G levels were increased in late RA patients. This is in agreement with the results from Rizzo et al., which showed that in early untreated RA patients, detectable levels of sHLA-G in plasma could be observed in all subjects, as compared to a minority of healthy controls, and that those levels increased upon antiRA treatment in those patients. This increase could reflect an attempt of the immune system to counterbalance the autoimmune process. However, in our RA patient cohort, no noticeable correlation between plasma sHLA-G levels and disease activity parameters was observed. Furthermore, in our study, sHLA-G levels did not intracellular colocalized differ with respect to anti-RA treatment. The differences in mean sHLA-G levels might be partially explained by different protocols of the HLA-G measurement: while the study from Verbruggen et al. used a two-step ELISA that included the depletion of classic HLA-I and HLA-E and detection by a pan-HLA-I antibody, our assay and also that one used in the work from Rizzo et al. skipped the depletion step and used a more direct detection strategy by an anti-HLA-G antibody. Importantly, the study of Rizzo et al. used the antibody MEM-G/09 as capture antibody where we were using the HLA-G specific antibody G233 and for these two antibodies discrepancies in sHLA-G concentration readouts have been previously reported. Other explanations for those discrepancies might include sample composition and differences in treatment regimens. The most intriguing observation from our study was that circulating sHLA-G molecules are not recognized by their cognate LILRB1 receptor in a substantial number of late RA patients. This observation was possible due to the application of a newly implemented Luminex assay based on the use of microspheres being coated with the HLA-G specific antibody, which has also been used for the quantitative determination of sHLA-G molecules by ELISA. However, the bound sHLA-G molecules were detected by a chimeric LILRB1 receptor. Thus, using this method it is possible to quantify specifically the presence of HLA-G molecules, which can be recognized by LILRB1. The conditions under which HLA-G molecules form dimers are not yet fully understood. It is known that dimerization occurs after passing through the Golgi apparatus, and in solid tumors it has been evidenced that HLA-G dimerization is enhanced by environmental factors such as interferon-�� or��. Moreover, superior long-term immunosuppressive effects of HLA-G dimers over monomers were already documented in a model of collagen-induced arthritis. A higher abundance of HLA-G monomers or even other non-classical HLA-G-like structures could explain why the sHLA-G levels observed in certain patients did not correlate with LILRB1 recognition. Of course it cannot be ruled out that the antibody G233 favors to bind to certain HLA-G structure e.g. monomers, dimers or HLA-G-like structures. However, this antibody was used as capture reagent in both assays, which should allow a correlation. Thus, a lack of correlation can only be attributed to differences in the binding of the detection reagents.