B cells in ANCA-associated vasculitides
von Borstel, Anouk
DOI:
10.33612/diss.93537940
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2019
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
von Borstel, A. (2019). B cells in ANCA-associated vasculitides: from pathogenic players to biomarkers. University of Groningen. https://doi.org/10.33612/diss.93537940
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
Remission Predicts Relapsing
Disease in Granulomatosis
with Polyangiitis Patients
6
aDepartment of Internal Medicine, Division of Nephrology,
University Medical Center Groningen, University of Groningen, Hanzeplein 1, 9713 GZ Groningen, the Netherlands bDepartment of Rheumatology and Clinical
Immunology, University Medical Center Groningen, University of Groningen, Hanzeplein 1, 9713 GZ Groningen, the Netherlands cDepartment of Pathology and Medical
Biology, University Medical Center Groningen, University of Groningen, Hanzeplein 1, 9713 GZ Groningen, the Netherlands
Submitted
Anouk von Borstela, Judith Landb, Wayel H. Abdulahadb,c, Abraham Rutgersb, Coen A. Stegemana, Arjan Diepstrac, Peter Heeringac and Jan-Stephan Sandersa
Abstract
Background Granulomatosis with polyangiitis (GPA) patients are prone to disease
relapses. We aimed to determine whether GPA patients at risk for relapse can be identified by increased plasmablast frequencies.
Methods Ninety GPA patients were monitored for 2.7 years (median: 3.1; range:
0.1-6.3). Circulating B cell subset frequencies, including plasmablasts, were determined by flow cytometry. The plasmablast frequency at inclusion of future-relapsing (F-R) and non-relapsing (N-R) patients was compared and related to relapse-free survival. Additionally, plasmablasts were determined in urine and kidney biopsies from active ANCA-associated vasculitis patients with renal involvement.
Results Within 1.6 years, 30% experienced a relapse. Plasmablast frequency was
increased in F-R- compared to N-R patients (p<0.002) and HCs (p=0.02). Increased plasmablast frequency at inclusion was related to decreased relapse-free survival in GPA patients. Also, 85.4% of patients with increased plasmablast frequency relapsed during follow-up compared to 22.1% of patients with a low plasmablast frequency. No correlations were found between plasmablasts and ANCA levels. Plasmablast frequencies were increased in the urine compared to the circulation, and were also found in kidney biopsies, which may indicate plasmablast migration during active disease.
Conclusions Our data suggests that increased circulating plasmablast levels during
remission is related to higher relapse risk in GPA patients, and therefore might be a potential marker to identify those GPA patients at risk for relapse.
6
Introduction
Anti-neutrophil cytoplasmic autoantibody (ANCA)-associated vasculitides (AAV) are autoimmune diseases involving small- to medium-sized blood vessels69. AAV are classified into granulomatosis with polyangiitis (GPA), microscopic polyangiitis (MPA) and eosinophilic GPA (EGPA)2. In these patients, ANCA directed against proteinase 3 (PR3)172 or myeloperoxidase (MPO)72 are frequently present. PR3-ANCA are detected in the majority of GPA patients, whereas MPO-ANCA are more often detected in MPA and EGPA patients. PR3-ANCA-positive patients have a significantly increased relapse-rate compared to MPO-ANCA-positive patients. Approximately 50% of PR3-ANCA-positive patients experience a relapse within four years of follow-up, compared to 25% of the MPO-ANCA-positive patients190. Predicting relapses remains pivotal for patient care, as relapses are associated with considerable morbidity caused by both disease- and therapy-related damage. So far, no biomarker for the prediction of relapse has been found to be reliable in GPA patients. Combined data of multiple studies has shown that ANCA titers can predict future relapses only to a limited extent68.
Although the pathogenic potential of ANCA is well established (reviewed in 191), the mechanisms that trigger ANCA production and disease relapse are less clear. The clinical benefit observed in B cell depletion trials using rituximab in patients with AAV strongly supports the contention that B cells are key contributors in the AAV immunopathogenesis40,41,149. Since rituximab targets all CD20+ B cells, it also depletes the precursors of ANCA-producing plasmablasts/plasma cells. These plasmablasts/plasma cells can secrete pathogenic autoantibodies, and are crucial for the development of autoimmune pathology192,193. Plasmablasts can circulate for prolonged periods and could be a useful biomarker for assessing disease activity194–196.
Increased circulating plasmablasts have been described in several inflammatory disorders, in particular, systemic lupus erythematosus (SLE)195,196, rheumatoid arthritis (RA)197, inflammatory bowel disease198, and IgG4-related disease199. However, whether circulating plasmablasts are associated with smoldering inflammation and disease relapses in GPA has not been investigated yet. Given the pathogenic role of ANCA, detection of plasmablasts, the precursors of (auto)antibody-producing plasma cells could provide an early warning for disease relapse. Thus, assessing the presence and frequency of circulating plasmablasts in GPA patients might be a good prognostic tool to predict disease relapses, and is perhaps more efficient than measuring serum ANCA levels.
This study aimed to elucidate whether circulating plasmablasts are increased in GPA patients and to demonstrate whether they are associated with ANCA levels and risk for relapse. Also, we aimed to assess plasmablast migration to kidneys of active GPA patients with renal involvement. Plasmablast frequencies were determined in GPA patients in remission, with and without future disease relapses, and their relation to
disease-relapses and serum ANCA levels were assessed. Additionally, plasmablasts were investigated in matched urine and blood samples and kidney biopsies from active AAV patients with renal involvement.
Materials and Methods
Study Population
Ninety GPA patients and 48 age-matched healthy controls (HCs; 52.1% male, median: 57.3 years, 40-74) were enrolled in the study between 2010-2013. GPA diagnosis was based on definitions described in the Chapel Hill Consensus Conference2. At inclusion, all patients were in complete remission (Birmingham vasculitis activity score (BVAS) = 0). Patients tested PR3-ANCA-positive at least once during their disease course. None of the patients or controls experienced an infection at the time of sampling. Only patients that were at least ten months post-rituximab treatment were included. Patients that experienced a relapse during follow-up were designated to the future-relapsing (F-R) group and the patients remaining in remission were assigned to the non-relapsing (N-R) group. Diagnosis of disease relapse was based on clinical judgment and had to result in initiation/increase of immunosuppressive treatment. Relapses were monitored in all patients and median time between sampling and diagnosis of relapse was 1.6 years (range: 0.1-3.8), whereas the total follow-up time for N-R patients was 3.2 years (range: 0.1-6.3). The main clinical and laboratory data of the patients are summarized in Table 1. All patients and HCs provided informed consent. All experimental procedures described below were conducted according to the policies of the UMCG and the medical ethics committee of the UMCG, Groningen, the Netherlands, approved the study.
In addition to patients in remission, ten AAV patients (seven GPA and three MPA; Table 2) with clinical signs and symptoms of active vasculitis with renal involvement were included to assess their urine and blood samples simultaneously. After additional diagnostic investigation, renal disease could be confirmed for six AAV patients (three GPA and three MPA). Also, four active GPA patients with renal involvement were included to assess plasma cell infiltration in kidney biopsies (Table 2).
6
Table 1. Clinical data and characteristics of GPA patients. GPA – no relapse during follow-up (N-R) GPA – relapsed during follow-up (F-R) P-Value(N-R vs. F-R) Subjects, n (% male) 63 (41.3) 27 (44.4) 0.7799
Age, mean (range) 60 (26-84) 55 (30-81) 0.2467
cANCA titer, median (range) 1:80 (0-1:640) 1:80 (0-1:640) 0.5812
Creatinine µmol/L, median (range) 71 (13-168) 73 (21-171) 0.1722
CRP mg/L, median (range) 4.9 (0.5-20) 4.9 (0.4-83) 0.5286
Disease duration in years, median (range) 9.1 (0.2-42.1) 11.4 (2.1-28.7) 0.2451 Number of total relapses before inclusion, median
(range)
1 (0-6) 3 (0-10) 0.0002
Lymphocyte Count *106/L, median (range) 1090 (340-2900) 695 (240-1640) 0.0043
IS therapy at time of sampling, n (%) 24 (38.1) 19 (70.4) 0.005
Azathioprine, n (%) 4 (6.4) 8 (29.6) 0.0029
Azathioprine + prednisolone, n (%) 9 (14.3) 6 (22.2) 0.3545
Cyclophosphamide + prednisolone, n (%) 1 (1.6) 0 (0) 0.5103
Mycophenolate mofetil + prednisolone, n (%) 3 (4.8) 4 (14.8) 0.1027
Prednisolone, n (%) 7 (11.1) 1 (3.7) 0.2578
Induction therapy
Azathioprine + prednisone, n (%) Cyclophosphamide + prednisone, n (%) Methotrexate + prednisone, n (%) Mycophenolate mofetil + prednisone, n (%) Cotrimoxazole, n (%) 2 (3.2) 55 (87.3) 2 (3.2) 0 (0) 4 (6.3) 0 (0) 26 (96.3) 0 (0) 1 (3.7) 0 (0) 0.3491 0.1924 0.3491 0.1245 0.1804 No. clinical manifestations baseline, median (range) 3 (1-6) 4 (1-6) 0.0089
Kidney involvement, n (%) 36 (57.1) 19 (70.4) 0.2382
Airway involvement, n (%) 57 (90.5) 26 (96.3) 0.3448
AAV: ANCA-associated vasculitides; cANCA: cytoplasmic anti neutrophil cytoplasmic antibody; CRP: C-reactive protein; F-R: future-relapsing; IS: immunosuppressive; N-R: non-relapsing; No.: number
Table 2. Clinical data and characteristics of AAV patients with active disease and signs of renal involvement.
Urine Analysis Plasma cell Histology
Disease subtype, n (% male) MPA, 3 (33.3) / GPA, 3 (100) GPA, 4 (100)
Age, median (range) 73.5 (59-82.5) 71.1 (59-86.4)
ANCA positive, n (%) 6 (100) 4 (100)
BVAS, median (range) 12 (11-21) 13 (11-15)
Creatinine umol/L, median (range) 165 (94-566) 236.5 (165-566)
CRP mg/L, median (range) 56 (6-85) 22 (6-85)
Proteinuria urine g/L, median (range) 1.08 (0.4-3.57) 2.5 (0.87-3.57*)
IS therapy, n (%) 3 (50) 2 (50)
No. clinical manifestations, median (range) 2 (1-4) 2 (1-2)
BVAS: Birmingham Vasculitis Activity Score; cANCA: cytoplasmic anti-neutrophil cytoplasmic autoantibody; CRP: c-reactive protein; GPA: granulomatosis with polyangiitis; IS: immunosuppressive; MPA: microscopic polyangiitis; No.: number; *not determined for 1 patient
Flow Cytometry Analysis of B cell Subsets
EDTA venous blood was obtained from GPA patients and HCs. Immediately after sampling, blood was washed twice in PBS with 1% BSA (wash buffer). Next, 100 µl cell suspension was stained using anti-human CD19-eFluor450, CD27-APC-eFluor780, CD38-PE-Cy7 (eBioscience, San Diego, USA) or the corresponding isotype controls. Cells were treated with 10x FACS Lysing solution (BD Biosciences, San Jose, USA) and acquired on an LSR-II flow cytometer (BD Biosciences). For all flow cytometry analyses, data were collected for at least 2*105 cells, and analyzed using Kaluza 1.5a software (Beckman Coulter, Brea, USA). Supplementary Figure 1 shows the gating strategy and Figure 1A the representative gating examples for each group. The B cell subset percentages were converted to absolute numbers using the lymphocyte count and the percentage of total B cells. HC lymphocyte counts were not available.
Flow Cytometry Analysis of Plasmablasts in Blood and Urine
Urine and blood samples were collected from ten AAV patients with active disease. Urine samples were prepared as described previously200. Briefly, urine was diluted 1:1 in PBS and centrifuged at 600 x g. The sediment was resuspended in PBS and mononuclear cells (MNCs) were isolated using lymphoprep (Axis-Shield, Oslo, Norway). Next, MNCs were resuspended in wash buffer and stained with anti-human CD19-PerCP-Cy5.5, CD45-BV605, CD27-APC (BioLegend, San Diego, USA), CD3-BUV395 and CD38-BB515 (BD Biosciences) for 15 minutes at room temperature in the dark. Isotype-matched non-specific antibodies were used as negative controls. In parallel, blood samples were labelled with the aforementioned monoclonal antibodies. Afterwards, cells were treated with 10x diluted FACS lysing solution for 10 minutes, washed twice in wash buffer and immediately analyzed. Stained urine and blood samples were acquired on the LSRII and data was analyzed using Kaluza 1.5a software. Figure 3A shows a representative gating example of both blood and urine. Four patients were excluded because no renal involvement was diagnosed and accordingly no B cells were present in the urine. Histology Analysis of Plasmablasts in Kidney Biopsies
Kidney biopsies of four active GPA patients with renal involvement were stained for plasmablast/plasma cell infiltration using a MUM1/IRF4 monoclonal antibody (clone MUM1p, DAKO, Santa Clara, USA) in an automated stainer (Ventana, Roche, Basel, Switzerland).
Serum ANCA Levels
6
previously201. ANCA titers >1:20 were considered positive. Serum PR3-ANCA levels could be determined in samples of 51 patients by Phadia ImmunoCAP® 250 analyzer using EliA PR3S (Thermo Fisher Scientific, Waltham, USA).
In vitro PR3-ANCA and IgG
Peripheral blood MNCs (PBMCs) were isolated and cultured as described before23. In short, for 79 GPA patients PBMCs were available and were cultured for 12 days either without/with 3.2 µg/mL CpG-ODN 2006 (Hycult Biotech, Uden, the Netherlands), 100 ng/mL IL-21 (Immunotools, Friesoythe, Germany) and 100 ng/mL BAFF (PeproTech Inc., Rocky Hill, USA). Stimulated and spontaneous PR3-ANCA (RU/mL) production was determined in the supernatant by Phadia ImmunoCAP® 250 analyzer using EliA PR3S (Thermo Fisher Scientific) and total spontaneous and stimulated IgG was assessed by ELISA. Five samples (2 F-R, 3 N-R) were excluded because of a culture infection.
Statistical Analysis
Statistical analysis was performed using GraphPad Prism v7.0 (GraphPad Software, San Diego, USA) and SPSS (IBM Corporation, Chicago, USA). B cell subset frequencies of HCs, F-R and N-R GPA patients were compared using a Kruskal-Wallis test. Individual groups were compared with Dunn's multiple comparison test. The plasmablast counts were compared using the Mann-Whitney U test. Furthermore, a Kaplan-Meier curve of probability of relapse-free survival was plotted for patients with less or more than 2.39% plasmablasts (one deceased patient was excluded) and plasmablast frequency was compared between inclusion samples and samples before relapse of F-R patients using the Wilcoxon signed rank test. For F-R patients the mean difference between inclusion and the last sample taken before relapse was 0.8 ± 0.5 years. Cox regression analysis was performed to determine the relation between log-transformed plasmablasts and immunosuppressive treatment and relapse occurrence.
Correlations between different parameters were determined with the Spearman r correlation. Finally, relapse-free survival was determined using the Log-rank test. P-values<0.05 were considered statistically significant.
Results
Clinical Patient Characteristics
During a median follow-up of 1.6 years, 27 GPA patients experienced a disease relapse (future-relapsing; F-R), whereas 63 patients did not (non-relapsing; N-R). The number of clinical manifestations at first diagnosis was higher in F-R compared to N-R patients
(Table 1). Additionally, F-R patients had experienced more relapses than N-R patients before study inclusion. At inclusion, more F-R patients received immunosuppressive therapy. Induction therapy and other baseline clinical characteristics such as CRP levels and serum creatinine did not differ significantly between both patient groups (Table 1). Increased Frequencies of Circulating Plasmablasts in F-R Patients
The B cell phenotype was determined in 27 F-R and 63 N-R patients in remission and in 48 HCs (Figure 1A). The total B cell frequency was significantly lower in both patient groups compared with HCs (Figure 1B). Studying B cell subsets, the percentage of transitional B cells was decreased in N-R patients compared with HCs (Figure 1B). The naïve B cell frequency was higher in both N-R and F-R patients compared with HCs (Figure 1B) and the percentage of memory B cells was lower in both patients groups as compared with HCs (Figure 1B), which is in line with previously published results46. The circulating plasmablast percentage was significantly increased in F-R patients compared to N-R patients, while N-R patients were similar to HCs (Figure 1C).
The lymphocyte count was significantly decreased in F-R patients (Table 1). Using the lymphocyte count we calculated the absolute numbers of circulating plasmablasts. No differences in numbers of circulating plasmablasts were observed between both patient groups (Figure 1C).
Also, Cox-regression analysis demonstrated a significant relation between (log transformed) plasmablasts and relapse, whereas only a trend between the use of immunosuppressive medication and relapse was observed (Supplementary Table 1). Increased Circulating Plasmablast Frequency in GPA Patients is Related to Decreased Relapse-free Survival
Subsequently, we investigated whether an increased plasmablast frequency was related to future disease relapse. We divided the GPA patients into two groups based on the median plasmablast frequency of the F-R patients: one group consisted of patients with less than 2.39% plasmablasts and the other group consisted of patients with 2.39% or more plasmablasts. Significantly more patients with a high percentage of plasmablasts experienced a future disease relapse than patients with a low plasmablast frequency (Figure 2A). The percentage of patients (77.9%) who remained relapse-free during follow-up was significantly higher in GPA patients with <2.39% circulating plasmablasts compared to patients with ≥2.39% plasmablasts (14.6%).
Moreover, during follow-up the circulating plasmablast frequency in F-R patients tended to decrease 1-3 months prior to relapse compared to remission samples (Figure 2B).
6
HC N-R 52.25% 25.76% 1.64% 20.35% 79.14% 12% 8.1% 0.76% F-R 74.39% 8.65% 13.09% 3.87% CD27 CD38 A B C H C s N -R F-R 0 2 4 6 8 10 12 14 16 18 20 22 24 B c el ls (% ) ** ** * H C s N -R F-R 0 5 10 15 20 25 30 35 Tr an si tio na l B c el ls (% ) * H C s N -R F-R 20 30 40 50 60 70 80 90 10 0 Na iv e B c el ls (% ) ** * ** * H C s N -R F-R 0 10 20 30 40 50 60 M em or y B c el ls (% ) ** ** H C s N -R F-R 0 1 2 3 4 5 9 12 15 18 Pl as m ab la st s (% ) ** N -R F-R 0. 0 0. 5 1. 0 1. 5 2. 0 2. 5 3. 0 3. 5 4. 0 4. 5 5. 0 Pl as m ab la st s (*1 0 6/L ) Figur e 1. Plasmablast fr equenc y is incr eased in F-R pa tien ts . A. Based on CD27 and CD38 expr ession f our subsets w er e distinguished: naiv e B cells as CD27 -CD38 -/dim , tr ansitional B c ells as CD27 -CD38 hi, memor y B c ells as CD27 +CD38 -/dim and plasmablasts as CD27 +CD38 hi in HCs , N-R pa tien ts and F-R pa tien ts . B. T he pr opor tions of B c ells , tr ansitional B c ells , naïv e B cells and memor
y B c ells ar e depic ted f or HCs , N-R pa tien ts and F-R pa tien ts . Red bars r epr esen t the median v alue . C. T he plasmablast fr equenc y is expr essed as a per cen tage within t otal B c ells . T he absolut e plasmablast c oun t w as calcula
ted using the lymphoc
yt e c oun t and plasmablast fr equenc y in F-R and N-R pa tien ts and is g iv en as c oun t*10 6/L per ipher al blood . *p<0.05; **p<0.01; ***p<0.001.
Plasmablasts (%) during remission 0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80 90 100 <2.39% Plasmablasts ≥2.39% Plasmablasts ** **
Time to Relapse (months)
R el ap se -f re e su rv iv al ( % )
Remission Last sample
before relapse 0 1 2 3 4 5
6 Plasmablasts (%) change over time#
% w ith in B c el ls (% ) A B <2.39% plasmablasts 69 61 58 50 28 6 5 3 ≥2.39% plasmablasts 20 17 14 10 5 2 2 2
Figure 2. Increased plasmablast frequency is related to decreased relapse-free survival. A. Percentage relapse-free survival is depicted in patients in remission with < or ≥2.39% circulating plasmablasts (i.e. median value of F-R patients). Hazard ratio of 8.8 (95% CI: 3.35-23.2). In the table the number of subjects at risk are given for each time point. B. The plasmablast frequency is expressed as a percentage within B cells during remission and 1-3 months before relapse (n=15). #p<0.10; ****p<0.0001
Plasmablasts are Present in the Kidney and Increased in the Urine of Active AAV Patients with Renal Involvement
As the circulating plasmablast frequency seemed to decrease in patients before relapse, we hypothesized that plasmablasts migrate to sites of inflammation. To explore this, we analyzed the plasmablast frequency in both the circulation and urine of ten active AAV patients (Figure 3A). None of the four patients without active renal involvement presented with B cells, or plasmablasts, in the urine. Whilst in the six patients with active renal involvement B cells could be detected, and contained an increased proportion of plasmablasts compared to the circulation (Figure 3B). In addition, we stained four kidney biopsies to detect plasma cell infiltration. Figure 3C demonstrates plasma cell infiltration in the kidneys of four active GPA patients. For two of these patients, also the urinary plasmablast frequency was determined and this frequency increased compared to the circulation (Figure 3B). These findings support our hypothesis of plasmablast migration, and suggest that plasmablasts migrate from the circulation to inflamed kidneys during active disease.
6
Blood Urine CD38 CD27 1.37% 17.02% B A C 50 µM 50 µM 50 µM 50 µM Donor 1 Donor 2 Donor 3 Donor 4 Blood Urine 0 2 4 6 8 10 12 14 16 18 % within B cells (% ) Plasmablasts (%) *Figure 3. Indication that plasmablasts migrate from the circulation to the kidney in active AAV patients with renal involvement. A. Using CD27 and CD38, the plasmablast subset could be distinguished as CD27+CD38hi in both urine (right) and peripheral blood (left). B.
Circulating and urine plasmablast percentage in active patients with renal involvement (n=4). C. Immunohistochemistry for MUM1/IRF4, showing presence of plasma cells in formalin fixed paraffin embedded renal biopsy tissue samples of four patients. *p<0.05
Plasmablast Frequency is not Related to ANCA Levels in GPA
Plasmablasts are the precursors of antibody-producing plasma cells, and since we found evidence that the plasmablast frequency is related to future relapses, we examined whether plasmablast numbers and percentages correlated with ANCA and total IgG levels. Such a correlation could indicate that the increased plasmablast percentages are disease specific. No significant correlation was found between the percentage of circulating plasmablasts and ANCA titers (Figure 4A). Furthermore, no positive correlation existed between serum PR3-ANCA levels, and the percentage or number (data not shown) of circulating plasmablasts (Figure 4B). Similarly, plasmablast frequency did not correlate with either spontaneous in vitro produced (Figure 4C) or stimulated
in vitro produced PR3-ANCA (Figure 4D), and neither did total in vitro spontaneous or
stimulated IgG (data not shown).
0 2 4 12 13 14 15 16
Sp. r=0.019 p=0.4295 Plasmablasts & ANCA titer
0 1:20 1:40 1:80 1:160 1:320 1:640 Plasmablasts (%) AN CA T ite r 0.0 2.5 5.0 7.5 10.0 12.5 15.0 0 100 200 300 400 500 Sp. r=-0.1357 p=0.1712 Plasmablasts & serum PR3-ANCA
Plasmablasts (%) PR 3-A N C A (R U /m L) 0 2 4 0 1 2 3 4 5 12 13 9.5 10.0 Sp. r=0.1269 p=0.1406 Plasmablasts & spontaneous PR3-ANCA
Plasmablasts (%) PR 3-AN C A (R U /m L) 0 2 4 0 20 40 60 80 100 120 12 13 240 250 Sp. r=0.041 p=0.363 Plasmablasts & stimulated PR3-ANCA
Plasmablasts (%) PR 3-AN C A (R U /m L) B A C D
Figure 4. Plasmablast frequency does not correlate with ANCA levels. The plasmablast
frequency correlated to A. ANCA titer, B. Serum PR3-ANCA levels (IU/mL), C. spontaneous in vitro produced PR3-ANCA (RU/mL), and D. stimulated in vitro produced PR3-ANCA (RU/mL). In each Figure, the Spearman r and p-value are given.
6
Discussion
In this study, we demonstrate an increase in the frequency of circulating plasmablasts in GPA patients with frequent relapses, and their association with an increased relapse-risk. We found that the circulating plasmablast frequency tended to decrease with impending relapse and detected plasmablasts/cells in the kidneys and urine of active patients with renal involvement, which might indicate migration of plasmablasts to inflamed organs.
So far multiple possible predictors of (future) disease activity have been proposed, such as platelet count202 and lung involvement at diagnosis203. However, those markers are not ideal since they only identify active disease or focus on a subpopulation of patients. Currently, ANCA levels are the best biomarker available to identify approaching disease relapses68. A disadvantage of ANCA levels as biomarker is that it is not reliable in all patients, which highlights the need for a better biomarker of future disease activity. Here, we found that GPA patients with increased frequencies of circulating plasmablasts during remission were more likely to relapse in the (near) future.
Plasmablasts are the precursors of (auto)antibody-producing plasma cells204,205 and may therefore play a major role in GPA as they might be precursors of ANCA-producing plasma cells. In other autoimmune diseases such as SLE196,206, IgG4-related disease199 and anti-PLA2R1 related membranous nephropathy207 the plasmablast frequency was related to disease activity, as were absolute plasmablast numbers in SLE206 and IgG4-related disease199. Additionally, plasmablast frequencies and numbers correlated significantly with (auto)antibody levels in IgG4-related disease199 and SLE206, although one study in IgG4-related disease could not confirm this correlation208. Interestingly, in RA patients sorted plasmablasts were found to produce anti-citrullinated protein antibodies (ACPA) with and without in vitro stimulation. The authors did not study whether the circulating plasmablasts correlated to serum ACPA levels197. In this study we did not find a correlation between plasmablasts and ANCA titers, serum PR3-ANCA levels or spontaneous and stimulated in vitro produced PR3-ANCA levels. A possible explanation might be that plasmablasts first need to differentiate into plasma cells and migrate to the bone marrow in order to produce ANCA. Another explanation might be that an increase in plasmablasts is not disease specific but a consequence of smoldering inflammation. Previously, we have found that in vitro PR3-ANCA production is not related to approaching relapses21, and this might also explain the lack of correlation between plasmablasts and in vitro ANCA production.
We demonstrated increased plasmablast frequencies in the kidney and urine of renal active AAV patients compared to the circulation, which might indicate plasmablast migration from the circulation to the inflamed kidney. Flow cytometry analysis from blood samples showed a different B cell subset distribution than those from the urine samples (Figure 3A), which suggests that there is active migration to the kidney instead of
leakage of plasmablasts into the urine. In murine lupus mouse models plasma cells were found to be present in the inflamed kidney209,210. Also, in renal biopsies of SLE patients CXCR4+ plasma cells and plasmablasts have been detected211. In Kawasaki disease, IgA plasma cell infiltration has been demonstrated in the inflamed kidney as well as in other affected organs (e.g. arteries, lungs)212. To our knowledge, we are the first to describe plasma cell infiltration into the inflamed kidney. A recent article by Brix et al.213 described B cell infiltrates in kidney biopsies from active ANCA-associated glomerulonephritis patients and also found an association of ectopic lymphoid structures with end stage renal disease during follow-up. Identification of plasmablasts by immunohistochemistry remains a challenge since plasmablasts do not express distinct phenotypic markers. Plasmablasts resemble plasma cells regarding their marker expression and transcription factors, both express CD27206, CD38214, CD138215, HLA-DR206, BLIMP-1216 and have decreased CD20206 and CD19206 expression. Therefore, only expert pathologists should perform and analyze histologic plasma cell stainings. Although the MUM1/IRF4 protein is also expressed by other B cell subsets, it has the highest expression in plasmablasts and plasma cells217 and therefore staining the MUM1/IRF4 protein is a suitable method to identify these cells.
An important limitation of the current study is the assessment of total plasmablasts rather than PR3-specific plasmablasts. Additional limitations include the limited number of patients in the study as well as the study design as it was performed in a single center. Future research should aim to include more patients to assess the extent of plasmablast infiltration in the kidney and the frequency of urinary plasmablasts.
In conclusion, in GPA patients an increased plasmablast frequency, but not number, during remission might be associated with future relapses of disease. Whether plasmablasts are disease specific in GPA remains to be elucidated. Additional prospective studies in larger GPA patient cohorts are needed to fully establish whether the circulating plasmablast frequency during remission constitutes a novel prognostic marker of disease activity.
Funding
Research leading to these results has received funding from the Jan Kornelis de Cock foundation. JSS is supported by personal grants from the Dutch Kidney Foundation (grant no. 13OKJ39) and the Dutch Organization for Scientific Research (Clinical Fellow grant no. 907-14-542). WHA and PH are supported by the European Union’s Horizon 2020 research and innovation program project RELENT (grant no. 668036).
