Previous studies of AGE in RA have been cross-sectional in nature and in general conducted in patients with disease of several to many years duration. Thus, it was important to examine a cohort of patients with early synovitis to determine whether AGE-damaged proteins were present near disease onset, in order to determine whether the generation of AGEs could be linked to an ongoing pathogenic process or whether they are merely the consequence of damage in the past. The amount of AGE-damaged proteins present in an individual reflects not only the synthesis of the AGE, which can take weeks to months, but also the clearance of the AGE-damage protein either through normal mechanisms for the protein in question or by specific receptors for AGE, one of which is termed receptor for AGE (RAGE) [20
]. The present study shows that the formation of AGE-damaged IgG in this early synovitis cohort is linked to the inflammatory response rather than to hyperglycemia. The latter is key in the AGE associated with diabetes [3
]. We did not measure hemoglobin A1c in the present study, but in a previous study it was not found to correlate with IgG-AGE in patients with longstanding RA [14
]. Thus, it appears that AGE formation associated with arthritis occurs as a result of oxidative stress. Recent studies have shown that high levels of RAGE expression, which is induced by AGE, can contribute to the inflammatory cycle by the induction of proinflammatory cytokines, via the nuclear factor-κB pathway [21
]. No studies have yet examined RAGE expression in the synovium of patients with early synovitis. RAGE appears to play a central role in the arthralgia associated with AGE-modified β2
-microglobulin, the latter being a consequence of long-term dialysis [22
From our studies, the presence of IgG damaged by AGE is not disease specific but occurs in all forms of arthritis, and importantly at an early stage of disease. Approximately 15–30% of the patients had elevated levels of IgG-AGE at their initial visit, and the levels were seen to fluctuate over time because there was no correlation between the amounts of IgG-AGE at the initial and the year 1 follow-up visits. The frequency and levels of IgG-AGE were slightly higher than we previously observed in patients with longstanding disease [14
]. Elevations in pentosidine (a specific type of AGE) were previously found to correlate with increased clinical disease activity in RA and with RF levels [6
]. However, in the present study no clinical parameter of joint disease was found to correlate with levels of IgG-AGE.
Our previous studies found a subset of RA patients with longstanding disease to have antibodies to IgG-AGE [14
]. In the present study we extend this observation to patients with early synovitis. Consistent with previous studies, the ability to make anti-IgG-AGE antibodies was strongly linked to RF-positive status. A mechanism that may account for this is that RF-positive B-cells might play a role in antigen presentation of the IgG-AGE and in induction of the immune response to AGEs. This would be analogous to the process of epitope spreading. Recent studies employing mass spectrometry have shown that much of the AGE on IgG is located within the Fab region [17
], which would not interfere with the binding of the Fc portion of IgG to the RF.
In contrast to the IgG-AGE elevations, which were not found to be disease specific, the anti-IgG-AGE response was much more specific to RA, and in particular to RF-positive RA. Indeed, at the initial visit, whereas a small percentage of patients with UA or spondyloarthropathy were found to have anti-IgG-AGE antibodies, this immune response appeared to be largely transient in these groups only, with the loss of this specificity by the year 1 follow-up visit. In contrast, in the RF-positive RA group, not only did approximately 20% gain this specificity at the year 1 follow-up visit (versus 0–3% in the other groups) but also a much higher proportion maintained this specificity at follow up than in any other group. Thus, the anti-IgG-AGE response is much more specific to RF-positive RA.
In the overall cohort, the anti-IgG-AGE response was associated with the swollen joint count, which probably reflects the increased joint count seen in RA, because within RA there was no significant correlation noted in these patients with early synovitis. In our previous study in patients with longstanding disease [14
], there was a significant correlation between the anti-IgG-AGE response and swollen joint count, whereas the RF response was not a useful biomarker for current disease activity. In the latter study the anti-IgG-AGE response was also linked to a physician-assessed disease activity score (visual analogue score). The latter type of evaluation was not conducted with the early synovitis cohort. Nonetheless, it does appear that, with respect to current disease activity, the level of anti-IgG-AGE antibodies represents a useful biomarker. RFs detected in other early synovitis cohorts have been shown to be predictive of long-term radiologic damage [23
], and thus it will be of interest to follow this cohort over time to determine whether the anti-IgG-AGE antibodies will also be predictive of more severe disease in this subset of RA patients.