In this study we showed that both HMGB1 levels and anti-HMGB1 levels are increased in SLE patients, and are related to SLE disease activity scores and serological measures of disease activity [24
]. Recent studies showed increased levels of serum HMGB1 and anti-HMGB1 in several autoimmune diseases including SLE. However, the relation between levels of HMGB1 and anti-HMGB1 antibodies has not been evaluated in a large group of SLE patients. We measured serum HMGB1 using two methods, ELISA and Western blot. The ELISA kit used in this study has been used to detect serum HMGB1 levels mainly in other chronic inflammatory diseases [30
]. Urbonaviciute et al
. showed that there was a discrepancy in serum/plasma HMGB1 results obtained from their in-house developed ELISA and Western blot in SLE patients due to possible interference of HMGB1-binding antibodies and unidentified serum proteins with HMGB1 [32
]. Indeed, we could show that IgG depletion decreased the amount of HMGB1 in sera from lupus patients, suggesting that at least part of HMGB1 is complexed with anti-HMGB1 antibodies. In this study prevalence of serum HMGB1 in patients obtained by ELISA were low in accordance with Ma et al.
, who also detected serum HMGB1 levels by the same kit [33
]. To avoid possible interference of immune-complex formation or other serum proteins in ELISA, we assessed HMGB1 levels in all SLE and HC using Western blot. HMGB1 levels were significantly increased in all SLE patients compared to HC and HMGB1 levels were highest in patients with active disease.
Our Western blot results are in line with Li et al.
, who also found elevated levels of serum HMGB1 in Chinese SLE patients [29
]. Increased expression of HMGB1 has also been reported in other autoimmune diseases particularly in RA. It has been shown that HMGB1 is increased in synovial tissues where inflammation exists; however, HMGB1 levels in sera of RA patients were not detectable [32
]. It is speculated that in RA activated inflammatory cells such as macrophages in the synovial tissue are the source of HMGB1 [34
]. In SLE, the increased HMGB1 levels found in serum might also be the product of activated inflammatory cells but might be a product of uncleared apoptotic cells as well. Overall, our results and previous findings indicate that HMGB1, as a pro-inflammatory cytokine, could be an important player in the inflammatory processes in SLE and might be specific for the disease.
In accordance with Li et al.
, HMGB1 levels were correlated positively with SLEDAI. The study by Li et al.
, did not demonstrate an association of HMGB1 with specific organ involvement [29
]. HMGB1 has been implicated in the pathogenesis of kidney diseases in patients with lupus nephritis and antineitrophil cytoplasmic antibody-associated vasculitis with renal involvement, and also has been shown to promote granulomatous nephritis in an adenine-fed rat model [31
]. In this model it was shown that granuloma formation was associated with high HMGB1 expression and up-regulation of HMGB1 receptors such as RAGE and TLR4. Therefore, we investigated if levels of HMGB1 are related to kidney involvement in SLE. Serum HMGB1 levels were higher in patients with renal involvement compared to patients without renal involvement. In chronic kidney disease, HMGB1 has been shown to correlate with renal function [37
]. This could not be demonstrated in our SLE patients as no correlation could be demonstrated between the presence of HMGB1 and renal function parameters, such as serum creatinine, and eGRF. However, we could observe a positive correlation between serum HMGB1 levels and proteinuria. Such association has been shown in a murine adenovirus accelerated SLE model [42
]. Interestingly, antagonizing HMGB1 by administration of anti-HMGB1 mAb IA-4 in this model markedly inhibited proteinuria and led to kidney protection. Combined, all these data indicate that HMGB1 might be an important molecule involved in renal pathology.
Besides the role of HMGB1 as a pro-inflammatory protein, it has been reported that antibodies against HMGB1 occur in SLE patients. In line with published findings of Hayashi et al.
, anti-HMGB1 antibodies were significantly elevated in all SLE patients compared to healthy controls [28
]. Moreover, our results indicate, similarly to HMGB1, that anti-HMGB1 antibodies correlate positively with SLEDAI in SLE patients suggesting a pathological role of these antibodies. Until now, no reports are available on the relation between HMGB1 and anti-HMGB1 antibodies. Our results showed a positive correlation between HMGB1 and anti-HMGB1 antibodies in SLE patients. The pathogenic or neutralizing role of these antibodies is still debated. The interaction between anti-HMGB1 antibodies and HMGB1 in the pathogenesis of SLE needs further study. In addition, we found that levels of anti-HMGB1 antibodies and anti-dsDNA antibodies are correlated suggesting that antibodies against HMGB1 and nucleosomes might be produced as a result of, possibly, concurrent with B cell activation by HMGB1 and nucleosomes.
In summary, this study shows that serum HMGB1 is significantly increased in SLE patients, in particular those with renal involvement, and correlates with proteinuria, concurrent with anti-HMGB1 antibodies were detected. HMGB1 and anti-HMGB1 antibodies may serve as a new biomarker to be used in the diagnosis and assessment of disease activity in SLE patients, as well to predict the disease flares. However, the differential role of HMGB1, anti-HMGB1 and their complexes in the pathogenesis of SLE, in particular in patients with renal involvement needs further investigation. Therapeutic targeting of HMGB1 might open new therapeutic opportunities in SLE.