Background/Purpose:  Serum receptor activator of NFkB ligand (RANKL) and its natural decoy receptor, osteoprotegerin (OPG), play key roles in osteoclast activation. In a group of patients with active RA, we tested the hypothesis that initiation of TNF blockade would improve osteoclast activation (reflected by the serum OPG:RANKL ratio).  DAS response was correlated with markers of B cell activation and musculoskeletal ultrasound.

Methods: 18 patients with RA failing methotrexate therapy (disease activity score (DAS) > 4.4) were randomized to receive either etanercept or adalimumab as part of a larger NIH sponsored Autoimmunity of Excellence (ACE) study (n=60). Soluble serum RANKL and OPG were measured by ELISA at baseline, 24, and 52 wk (pg/ml). RA disease activity was followed by DAS28 at baseline, 12 and 24 weeks. Peripheral blood (PB) B cells were phenotyped by flow cytometry using the following surface markers: CD19, IgD, CD27, CD24, CD38, CD95, and CD21. High frequency gray scale (GS) and power Doppler (PD) musculoskeletal ultrasound (MSK US) was performed on a subset of subjects (n=8) on the wrists, MCP, and PIP joints of both hands according to EULAR guidelines by a rheumatologist (DT and RT) certified in MSK US at baseline 4, 12, and 24 wk.  GS US and PD US were scored semiquantitatively (on a severity scale from 0-3). 

Results: In the current analysis of the Rochester cohort (n=18), 61% of patients were good DAS responders, 28% moderate responders, and 11% non-responders at 24 weeks.  Good DAS responders had higher baseline RANKL levels than moderate/non-responders (2653 vs. 1218).  Serum RANKL decreased significantly with anti-TNF treatment (2381+3025 at baseline vs. 921+1630 at 52 wk, p=0.03, n=8 longitudinal), as did OPG (150+92 at baseline vs. 82+51 at 52 wk, p=0.01), regardless of DAS response. Interestingly, the OPG:RANKL ratio was very low in these active RA patients (0.76+1.2 compared to historic healthy controls 3.0) and did not change significantly with anti-TNF treatment. On MSK US, the higher the baseline # of GS+ joints the less likely patients were to achieve a good DAS response.  Additionally good DAS responders tended to have a low # of GS+ joints/absent PD on follow-up US (12 and 24 wk) in contrast to moderate/non-responders.  Finally, RA patients at baseline had an expansion of activated memory B cells in the PB (CD95+ on CD27+IgD- memory in RA [n=13] 50.5+19.5% vs. HC [n=14] 29.6+9.5%, p=0.001; also significant for CD95+ on CD27-IgD- memory p<0.0001 and CD21- on CD27+IgD- memory p=0.02). There was a strong negative correlation between OPG:RANKL ratio and B cell activation (R2=-0.62 for correlation with CD95+ on CD27-IgD- memory), suggesting that activated B cells may contribute to RANKL activation.   Longitudinal assessment of B cell subsets in response to anti-TNF is underway. 

Conclusion: In our study, the OPG:RANKL ratio was considerably lower than the ratio reported for control subjects and similar to that observed in patients with multiple myeloma. B cell activation was characteristic of active RA and correlated with RANKL activation.  These results support the notion that OPG:RANKL and B cells have a central role in RA-associated joint destruction.

Supported in part by the University of Rochester ACE U19 AI563262 and P01 AI078907

---

« Back to 2012 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/impact-of-anti-tumor-necrosis-factor-alpha-therapy-in-rheumatoid-arthritis-on-osteoclast-activation-and-b-cells/