Discussion
In order to avoid the influence of glucocorticoid and any other immunosuppressant on the main outcome measures in this study, all the 40 selected patients with active MPO-AAV were initially diagnosed without any treatment with these agents. Among them, 39 patients (97.5%) had MPA, which was consistent with the fact that MPA with MPO-ANCA was most common AAV in China and other East Asian countries[4]. Also, Age and gender distribution in our study was respectively comparable between the patient and HC group as statistical analysis showed there was no significant difference for them between the 2 groups (Table I).
PAD4 is one of the 5 isoenzymes of PAD (PAD1-4 and PAD6) in human, a family of Ca2+-dependent deiminase enzymes. It is mainly expressed in the nucleus of neutrophil [27, 28]. Li et al. constructed PAD4 knockout mice and stimulated them with chemokines and bacteria. They found chromatin decondensation was inhibited in neutrophils, NETs could not form and the antimicrobial ability of mice was significantly decreased, which indicated PAD4 played an important role in formation of NETs by mediating histone citrullination[29]. Studies shown that PAD4 was also involved in the pathogenesis of both tumor and autoimmune diseases, such as SLE and rheumatoid arthritis (RA) [21, 30-32]. Knight et al. published that MRL/lpr SLE model mice treated by PAD4 inhibitors presented lighter clinical manifestations and disease activity compared to the controls[21]. Moreover, PAD4 in synovial fluid of RA was detected at significantly higher levels than that of osteoarthritis or psoriatic arthritis, which might be account for the generation of autoantigens in RA[33]. In this study, we detected the expression level of PAD4 in neutrophils by FCM, and found both the percentage of PAD4+ neutrophils and the MFI of PAD4+ neutrophils were significantly higher in patient group than in HC group, respectively (Fig.1). This might account for the NETs in patient group was markedly higher than that in HC group (Fig. 2A). Further mover, the percentage of PAD4+ neutrophils and the MFI of PAD4+ neutrophils respectively manifested positive correlation with NETs in patient group (Fig.3A and 3B).
As for NETs, in addition to its basic function of innate immunity against microbial infection[14], it was found to play important roles in pathogenesis of various autoimmune diseases, especially of AAV, by promoting both innate and adaptive immunity pathway. In a vitro experiment, Wang et al. demonstrated NETs could activate the cAP, an innate immune pathway to aggravate inflammation by producing C5a [18]. In AAV, the most common autoantigens (MPO and PR3) were the component of NETs[14], Sangaletti et al. reported that autoantigens in NETs could be processed and present to adaptive immune T cell more effectively by myeloid dendritic cells to induce the production of autoantibodies including MPO-ANCA[34]. MPO-ANCA combined with MPO on neutrophils and promote the neutrophils to discharge granules, to produce reactive oxygen radicals (ROS), to form NETs and finally to damage the tissues and organs by the inflammatorily destroyed vessels[15, 16]. The discharged granules from the neutrophils provoked by MPO-ANCA contained factors, such as properdin, that could boost cAP to present more C5a[35, 36]. In this study, we found serum C5a of patient group was significantly higher than that in HC group, which supported the activated cAP in active MPO-AAV (Fig. 2B). Although Wang et al. reported NETs could activate the cAP[18], our study indicated that, compare to NETs, MPO-ANCA might play more important roles to activate cAP by releasing granules because C5a was significantly associated with MPO-ANCA but not with NETs (Fig. 3C and Fig. 3D). Our result also showed MPO-ANCA had significant positive correlation with NETs (Fig.3E), which might be attributed to that MPO in NETs could be more effectively processed and present as autoantigen by myeloid dendritic cells to activate adaptive immunity[34]. Kusunoki et al. found in their mice model of MPO-AAV that histone citrullination could decrease by inhibiting PAD4 with Cl-amidine, and then serious MPO-ANCA could also be reduced [37].
As discussed above, PAD4 played a key role in NETs formation, which might promote adaptive immunity to facilitate the generation of MPO-ANCA and also might activate cAP to increase C5a. Both MPO-ANCA and C5a might trigger neutrophils to form NETs, to accelerate cAP, to produce ROS and etc. This then could create a self-fueling inflammatory amplification loop, a “vicious cycle”, to aggravate clinical inflammatory damage of MPO-AAV. Actually, Univariable correlation analysis in this study showed that the percentage of PAD4+neutrophils, MFI of PAD4+neutrophils, NETs, C5a and MPO-ANCA was positively correlated with BVAS, respectively (Tab. II). However, multivariable correlation analysis revealed that only the percentage of PAD4+neutrophils and MPO-ANCA was independently correlated with BVAS, respectively (Tab. II). In fact, MPO-ANCA was the hallmark for MPO-AAV as an autoimmune disease. Previous studies had demonstrated MPO-ANCA as a pathogenic autoantibody by experimental research and clinical report[8, 38]. Though PAD4, NETs, C5a and the “vicious cycle” among them might contribute to the production of MPO-ANCA, MPO-ANCA present as an independent factor affecting the vasculitis activity in this clinical study. Meanwhile, PAD4 could have independent influence on BVAS, which should be attributed to its enzymatic function in NETs generation and the NETs pathologic efficacy in MPO-AAV.