Serum progranulin as a predictive marker for high activity of antineutrophil cytoplasmic antibody‐associated vasculitis

Abstract Background This study investigated whether serum progranulin could act as a predictive marker for high disease activity of antineutrophil cytoplasmic antibody (ANCA)‐associated vasculitis (AAV). Methods Fifty‐eight AAV patients were included in this study. Clinical and laboratory data were obtained at blood collection. The Short‐Form 36‐Item Health Survey Physical and Mental Component Summaries (SF‐36 PCS and SF‐36 MCS), Birmingham Vasculitis activity score (BVAS), Five‐Factor Score (FFS), and Vasculitis Damage Index (VDI) were assessed as AAV‐specific indices. Whole blood was collected and serum samples were isolated and stored at −80°C. Serum progranulin concentration was quantified by ELISA kits. Results The median age of patients was 63.0 years (19 men). The median BVAS was 11.0, and the median serum progranulin level was 49.0 ng/ml. Serum progranulin was significantly correlated with BVAS, FFS, erythrocyte sedimentation rate, C‐reactive protein level, SF‐36 PCS, haemoglobin, and serum albumin. Severe AAV was arbitrarily defined as the highest tertile of BVAS (BVAS ≥16). When the cut‐offs of serum progranulin were set as 55.16 ng/ml and 43.01 ng/ml for severe AAV, AAV patients with serum progranulin ≥55.16 and 43.01 ng/ml had significantly higher risks of severe AAV than those without (relative risk (RR) 4.167 and 4.524, respectively). Conclusions Progranulin might play an anti‐inflammatory role in AAV pathogenesis and serum progranulin could be used as a predictive marker for high activity of AAV.


| INTRODUC TI ON
Progranulin, a 68.5 kDa protein consisting of 593 amino acids, is known to participate in cell development, cell proliferation, and wound healing. 1 Progranulin is a precursor for several types of granulins and is cleaved into granulins by elastase, proteinase 3, and matrix metalloproteinase present in the extracellular matrix and lysosomes. 2 Progranulin is known to assume both proinflammatory roles and anti-inflammatory roles in the immune response. Binding of progranulin to ephrin type-A receptor 2 could trigger its pro-inflammatory effect by accelerating inflammation and upregulating progranulin expression through the mitogen-activated protein kinase and protein kinase B pathways. 3 Additionally, progranulin augments and strengthens toll-like receptor (TLR 9)-mediated intracellular signalling, which may play a pro-inflammatory role. 3 Conversely, progranulin may exert an anti-inflammatory effect by binding to tumour necrosis factor (TNF) receptor (TNFR)1 and TNFR2 with an affinity comparable to that to TNFα, resulting in inhibition of TNFα binding to its receptor. 4 Furthermore, progranulin may polarise Treg and enhance IL-10 production. 5 The role of progranulin in the inflammatory process of several autoimmune diseases remains controversial: its pro-inflammatory action was reported to exhibit a significant correlation with granulin levels and disease activity in systemic lupus erythematosus (SLE) and systemic sclerosis, whereas its anti-inflammatory action was known to be implicated in rheumatoid arthritis (RA), inflammatory bowel disease, and psoriasis. 4,6 An array of immune cells and pro-inflammatory cytokines are known to participate in the pathogenesis of antineutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV). 7,8 Therefore, progranulin may contribute to inflammation or reflect the inflammatory burden in AAV patients. However, there is no study to date investigating the clinical significance of progranulin in AAV patients. In this context, we investigated here whether serum progranulin could act as a predictive marker for high disease activity in AAV patients.

| Study subjects
We selected 58 AAV patients with consecutive hospital iden-

| Collection of clinical and laboratory data and AAV-specific indices
Data regarding age, sex, AAV subtypes, ANCAs, and organ involvement were collected from patients during the visit for blood sample collection. As laboratory data, white blood cell and platelet counts, erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), haemoglobin, glucose, blood urea nitrogen, serum creatinine, total protein, serum albumin, and complements 3 and 4 were collected. Birmingham vasculitis activity score (BVAS, version 3), 12 five-factor score (FFS), 13 and vasculitis damage index (VDI) 14 were assessed as AAV-specific indices.

| Blood collection and storage
Whole blood samples were collected after patients' consent and serum samples were isolated and stored at −80°C. Clinical and laboratory data and AAV-specific indices were obtained on the same day as that for blood sample collection.

| Estimation of serum progranulin
Serum progranulin concentration was quantified by ELISA kits (R&D Systems) from stored sera according to the manufacturer's instruction.

| Statistical analyses
All statistical analyses were performed using IBM SPSS Statistics for Windows, version 25 (IBM Corp.). Continuous and categorical variables are expressed as medians with interquartile ranges and numbers (percentages), respectively. The Mann-Whitney U test was used to check for significant differences between two groups on continuous variables. The correlation coefficient (r) between two variables was obtained using either the Pearson's correlation analysis or univariable linear regression analysis.
Multivariable linear regression analysis was performed using statistically significant variables identified from the univariable analysis. The odds ratio (OR) was obtained using multivariable logistic regression analysis of variables with p < 0.005 in the univariable logistic regression analysis. The optimal cut-off was extrapolated by performing receiver operator characteristic (ROC) curve analysis, and the maximum sum of sensitivity and specificity was selected. The relative risk (RR) of the cut-off for high disease activity of AAV was analysed using contingency tables and the chi-square test. p-values less than 0.05 were considered statistically significant.

| Characteristics of patients
Clinical and laboratory data and the AAV-specific indices estimated at the time of blood collection are shown in Table 1 (Table 1). However, the median serum progranulin did not show significant differences (p = 0.104) between patients with and without kidney involvement ( Figure S1B).

| Cut-off value of serum progranulin for the cross-sectional severe AAV
Severe AAV was defined as the highest tertile of BVAS (BVAS ≥16) estimated in this study. The optimal cut-off value of serum progranulin for severe AAV was obtained using the ROC curve (area 0.664, 95% confidence interval (CI) 0.517, 0.812, p = 0.037). When the cut-off was set at 55.16 ng/ml, the sensitivity and specificity were 45.5% and 83.3%, respectively. However, when the cut-off was set at 43.01 ng/ml, the sensitivity increased up to 86.4% whereas the specificity decreased to 41.7% (Figure 2A). Total protein (g/dl) 6

| Estimation of relative risk
When AAV patients were divided into two groups based on a cut-

| Linear regression analysis for the crosssectional BVAS
The variables estimated at the time of blood collection for the

| DISCUSS ION
This study was designed to investigate the role of progranulin in AAV pathogenesis. In terms of a TLR9-mediated pro-inflammatory role, granulins, produced via the cleavage of progranulin by extracellular matrix proteases may increase the endocytosis of TLR9 ligands, thereby enhancing TLR9=associated intracellular signalling. 3 This mechanism has been verified in SLE patients, in whom serum progranulin was shown to be significantly elevated when compared with that in healthy controls. 15 TLR9 stimulation has also been reported to accelerate the activation, adhesion and degranulation of polymorphonuclear leucocytes (PMNs) in PR3-ANCA vasculitis. 16 Therefore, serum progranulin might bind to TLR9 with its ligands and thereby aggravate AAV by the activation and degranulation of PMNs.
Conversely, the anti-inflammatory effect of progranulin mediated by F I G U R E 2 A) When the cut-offs of serum progranulin were set as 55.16 ng/ml and 43.01 ng/ml for severe AAV, the sensitivity and specificity were 45.5% and 83.3%, and 86.4% and 41.7%, respectively. B) When the cut-off of serum progranulin was set as 55.16 ng/ml, serum progranulin ≥55.16 ng/ml had a significantly higher risk of severe AAV than those with serum progranulin <55.16 ng/ml (RR 4.167). C) When the cut-off of serum progranulin was set as 43.01 ng/ml, serum progranulin ≥43.01 ng/ml had a significantly higher risk of severe AAV than those with serum progranulin <43.01 ng/ml (RR 4.524). AAV, antineutrophil cytoplasmic antibody-associated vasculitis; RR, relative risk TNFR1 and TNFR2, leads to alleviation of TNF-mediated inflammatory signalling by the competitive inhibition of TNFα binding to its receptors. 3 This mechanism has been demonstrated in mice model for inflammatory arthritis; progranulin was found to bind to TNFRs and prevent inflammation by inhibiting TNF-a-activated intracellular signalling in animal models of inflammatory arthritis. 17 Both TNFR1

TA B L E 2 Linear regression analysis of continuous variables for simultaneous BVAS
and TNFR2 are known to be expressed at higher levels in AAV patients than in healthy controls, with the extents of their expression being significantly correlated with AAV activity based on BVAS. 18 Therefore, the binding of serum progranulin to TNFR1 and TNFR2 could be speculated to improve AAV by competitive inhibition of TNF-a binding to its receptors.
Therefore, consistent with the higher serum progranulin in SLE patients (in whom progranulin exhibited a pro-inflammatory TLRmediated action) than in healthy controls, serum progranulin in RA patients (in whom progranulin assumes an anti-inflammatory role through TNFR1 and TNFR2) was expected to be lower than that in healthy controls. However, serum progranulin in RA patients was also reported to be significantly higher than that in healthy controls. 19 In this case, the increased serum progranulin was explained as an epiphenomenon that acts as negative feedback to alleviate inflammation, 20 i.e. elevated serum progranulin may act as a direct proinflammatory factor accelerating inflammation, but could also act as an indirect anti-inflammatory factor for alleviating inflammation.
In the present study, serum progranulin was significantly cor- therefore, this result could be explained by invoking the concept of negative feedback. As the activity of AAV increases, the production and secretion of progranulin are promoted, thereby alleviating inflammation via TNFRs. 18 We estimated the optimum cut-off value of serum progranulin for predicting severe AAV and found two cut-offs, one with high sensitivity and the other with high specificity. We demonstrated that the two cut-off values showed a significant RR for the Abbreviations: AAV, ANCA-associated vasculitis; ANCA, antineutrophil cytoplasmic antibody; BVAS, Birmingham vasculitis activity score; C, cytoplasmic; CI, confidence interval; CRP, C-reactive protein; ESR, erythrocyte sedimentation rate; Hb, haemoglobin; LUC, large unstained cell; MPO, myeloperoxidase; P, perinuclear; PLT, platelet; PR3, proteinase 3; WBC, white blood cell.
cross-sectional severe AAV. Next, multivariable logistic regression analysis was performed to obtain an independent predictor of severe AAV, but both cut-offs could not act as an independent predictor. This might be because of the indirect role of progranulin in negative feedback, as opposed to a direct pro-inflammatory role.
Nevertheless, we believe that this study may provide valuable information on the potential application of serum progranulin as a biomarker for detecting the cross-sectional AAV activity and predicting the risk for severe AAV.
To the best of our knowledge, this is the first study to investigate the clinical implication of serum progranulin in AAV patients.
However, the number of patients included in this study was too small to obtain generalised conclusions that could be applied to AAV patients in real clinical settings. A future study with a larger cohort of AAV patients will not only validate these results but also provide more reliable information on the clinical implications of changes in serum progranulin observed in AAV patients.
In conclusion, progranulin might play an anti-inflammatory role in AAV pathogenesis and serum progranulin could be used as a predictive marker for high activity of AAV.

ACK N OWLED G M ENT
We would like to appreciate Dr In Kyung Min for the statistical ad-

vice. (Biostatistics Collaboration Unit Department of Biomedical
Systems Informatics Yonsei University College of Medicine, Seoul, Republic of Korea).