High mobility group box 1 levels in large vessel vasculitis are not associated with disease activity but are influenced by age and statins

Introduction Takayasu arteritis (TA) and giant cell arteritis (GCA) are large vessel vasculitides (LVV) that usually present as granulomatous inflammation in arterial walls. High mobility group box 1 (HMGB1) is a nuclear protein that acts as an alarmin when released by dying or activated cells. This study aims to evaluate whether serum HMGB1 can be used as a biomarker in LVV. Methods Twenty-nine consecutive TA patients with 29 healthy controls (HC) were evaluated in a cross-sectional study. Eighteen consecutive GCA patients with 16 HC were evaluated at the onset of disease and some of them during follow-up. Serum HMGB1 levels were measured by enzyme-linked immunosorbent assay. Results In GCA patients at disease onset mean serum HMGB1 levels did not differ from HC (5.74 ± 4.19 ng/ml vs. 4.17 ± 3.14 ng/ml; p = 0.230). No differences in HMGB1 levels were found between GCA patients with and without polymyalgia rheumatica (p = 0.167), ischemic manifestations (p = 0.873), systemic manifestations (p = 0.474) or relapsing disease (p = 0.608). During follow-up, no significant fluctuations on serum HMGB1 levels were observed from baseline to 3 months (n = 13) (p = 0.075), 12 months (n = 6) (p = 0.093) and at the first relapse (n = 4) (p = 0.202). Serum HMGB1 levels did not differ between TA patients and HC [1.19 (0.45–2.10) ng/ml vs. 1.46 (0.89–3.34) ng/ml; p = 0.181] and no difference was found between TA patients with active disease and in remission [1.31 (0.63–2.16) ng/ml vs. 0.75 (0.39–2.05) ng/ml; p = 0.281]. HMGB1 levels were significantly lower in 16 TA patients on statins compared with 13 patients without statins [0.59 (0.29–1.46) ng/ml vs. 1.93 (0.88–3.34) ng/ml; p = 0.019]. Age was independently associated with higher HMGB1 levels regardless of LVV or control status. Conclusions Patients with TA and GCA present similar serum HMGB1 levels compared with HC. Serum HMGB1 is not useful to discriminate between active disease and remission. In TA, use of statins was associated with lower HMGB1 levels. HMGB1 is not a biomarker for LVV.

Age was independently associated with higher HMGB1 levels regardless of LVV or control status. 30 Conclusions: Patients with TA and GCA present similar serum HMGB1 levels compared with HC. Serum HMGB1 is 31 not useful to discriminate between active disease and remission. In TA, use of statins was associated with lower 32 HMGB1 levels. HMGB1 is not a biomarker for LVV.

117
In the 18 GCA patients, blood samples were collected at 118 disease onset prior to glucocorticoid therapy and follow-up 119 samples were obtained from 13 patients at 3 months and 120 from six patients at 12 months. Blood samples were col-121 lected from 29 TA patients as a cross-sectional evaluation. or Wilcoxon's test were used to analyze longitudinal data.

236
Higher serum HMGB1 levels observed in GCA com-237 pared with TA seems to be an effect of aging, since 238 serum HMGB1 levels were also higher in GCA controls 239 than in TA controls [2.98 (1.70-6.23) ng/ml vs. 1.46 240 (0.89-3.34) ng/ml; p = 0.019]. A weak correlation was 241 found between serum HMGB1 levels and age in all study 242 participants (rho = 0.244; p = 0.019) while in a linear re-243 gression model, age was independently associated with   . Thus, we 278 evaluated associations between disease activity in LVV and 279 serum HMGB1 levels. Unfortunately, no difference could 280 be found between patients with active disease and remis-281 sion or between patients with LVV and HC.

282
On the other hand, GCA patients at disease onset and 283 prior to therapy presented serum HMGB1 levels that 284 were similar to those of HC, and no association could be 285 found between HMGB1 and acute phase reactants, dis-286 ease manifestations or disease relapse. Moreover, during 287 follow-up no significant fluctuations in serum HMGB1 288 levels were observed in GCA patients. Novel biomarkers 289 in GCA would help to recognize active disease in pa-290 tients with signs and symptoms of GCA but normal 291 acute phase reactants. However, serum HMGB1 levels 292 were not increased in patients with active disease.

293
Serum HMGB1 levels were significantly higher in 294 GCA patients than in TA patients, and even though the 295 ROC analysis showed that a cutoff value of 2.17 ng/ml 296 in HMGB1 levels would help to differentiate GCA from 297 TA, we believe that it is unlikely that in clinical practice 298 it would replace the 50-year-old cutoff point used to dif-299 ferentiate both entities [1]. Furthermore, GCA controls 300 had higher serum HMGB1 than TA controls. These 301 findings indicate that serum HMGB1 levels increase dur-302 ing aging and may be influenced by the burden of  The role of statins in GCA has still to be determined.   377 Authors' contributions 378 AWSS contributed to the study design, performed laboratory tests, 379 conducted the statistical analysis, and drafted the manuscript. KSMG 380 contributed to the study design, evaluated the study participants, collected 381 data from medical records, and revised the manuscript. EB contributed to 382 the study design, collected data from patients' medical records, helped with 383 the interpretation of results, and revised the manuscript. FAGP evaluated the 384 study participants, collected data from medical records, helped with the 385 interpretation of data and revised the manuscript. ACDO evaluated the 386 study participants, collected data from medical records, helped with the 387 interpretation of data and revised the manuscript. EIS contributed to the 388 study design, helped with the interpretation of results, and revised the 389 manuscript. LECA contributed to the study design, helped with the 390 interpretation of results, and revised the manuscript. MB contributed to 391 the study design, interpretation of data and revised the manuscript. JW 392 contributed to the study design, performed laboratory tests, helped with 393 the interpretation of data and revised the manuscript. CGMK conceived the 394 study, contributed to the study design, interpretation of data and revised the 395 manuscript. All authors read and approved the manuscript.