Human Lupus Nephritis and Tubulointerstitial Inflammation in In

In Situ B Cell-Mediated Immune Responses and Tubulointerstitial Inflammation in Human Lupus Nephritis

This information is current as Anthony Chang, Scott G. Henderson, Daniel Brandt, Ni Liu, of September 24, 2021. Riteesha Guttikonda, Christine Hsieh, Natasha Kaverina, Tammy O. Utset, Shane M. Meehan, Richard J. Quigg, Eric Meffre and Marcus R. Clark J Immunol 2011; 186:1849-1860; Prepublished online 27

December 2010; Downloaded from doi: 10.4049/jimmunol.1001983 http://www.jimmunol.org/content/186/3/1849

Supplementary http://www.jimmunol.org/content/suppl/2010/12/27/jimmunol.100198 http://www.jimmunol.org/ Material 3.DC1 References This article cites 85 articles, 21 of which you can access for free at: http://www.jimmunol.org/content/186/3/1849.full#ref-list-1

Why The JI? Submit online. by guest on September 24, 2021 • Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2011 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

In Situ B Cell-Mediated Immune Responses and Tubulointerstitial Inﬂammation in Human Lupus Nephritis

Anthony Chang,*,1 Scott G. Henderson,†,1 Daniel Brandt,† Ni Liu,† Riteesha Guttikonda,† Christine Hsieh,† Natasha Kaverina,† Tammy O. Utset,† Shane M. Meehan,* Richard J. Quigg,‡ Eric Meffre,x and Marcus R. Clark†

The most prevalent severe manifestation of systemic lupus erythematosus is nephritis, which is characterized by immune complex deposition, inflammation, and scarring in glomeruli and the tubulointerstitium. Numerous studies indicated that glomerulonephritis results from a systemic break in B cell tolerance, resulting in the local deposition of immune complexes containing Abs reactive with ubiquitous self-Ags. However, the pathogenesis of systemic lupus erythematosus tubulointerstitial disease is not known. In this article, we demonstrate that in more than half of a cohort of 68 lupus nephritis biopsies, the tubulointerstitial infiltrate was organized into well-circumscribed T:B cell aggregates or germinal centers (GCs) containing follicular dendritic cells. Sampling of the Downloaded from in situ-expressed Ig repertoire revealed that both histological patterns were associated with intrarenal B cell clonal expansion and ongoing somatic hypermutation. However, in the GC histology, the proliferating cells were CD1382CD20+ centroblasts, whereas they were CD138+CD20low/2 plasmablasts in T:B aggregates. The presence of GCs or T:B aggregates was strongly associated with tubular basement membrane immune complexes. These data implicate tertiary lymphoid neogenesis in the pathogenesis of lupus tubulointerstitial inflammation. The Journal of Immunology, 2011, 186: 1849–1860. http://www.jimmunol.org/

he clinical manifestations of systemic lupus erythematosus renal manifestation most clearly related to the central pathogenic (SLE) are myriad and range from mild arthralgias and skin feature of SLE: systemic loss of B cell tolerance (1, 7, 8). T rashes to life-threatening nephritis (1). Patients with focal The presence of serum anti-dsDNA Abs identifies lupus patients or diffuse proliferative lupus nephritis (LN) have a more rapidly at increased risk for GN, whereas increasing titers of anti-dsDNA progressive disease and require more aggressive treatment than do Abs often herald renal flares in individual SLE patients (1). Anti- those with mesangial proliferative or membranous LN (2). Sig- dsDNA Abs have been isolated from SLE renal biopsies (9, 10), nificant morbidity and mortality are consequences of the nephritis and infusion of some murine and human anti-dsDNA Abs can and the cytotoxic therapeutic regimens used to treat it (3). induce GN in mice (11, 12). Furthermore, recent immunoelectron by guest on September 24, 2021 Pathologically, LN is characterized by immune complex de- microscopic studies directly demonstrated the presence of anti- position and inflammation in glomeruli and the tubulointerstitium dsDNA Ab-containing immune complexes in diseased glomeruli that, if left untreated, can result in scarring and irreversible organ (13, 14). These observations led to the hypothesis that anti-dsDNA failure. Of the pathological manifestations of LN, glomerulone- IgG Abs play a central role in the pathogenesis of lupus GN. phritis (GN) is the best studied and the feature most often replicat- However, tubulointerstitial inflammation (TI) is also a com- ed in murine models of autoimmune disease (4–6). GN is also the mon feature of LN (15–17) (C. Hsieh, A. Chang, R. Guttikonda, D. Brandt, T. O. Utset, and M.R. Clark, submitted for publication). On renal biopsy, the presence and degree of TI identify those *Department of Pathology, University of Chicago, Chicago, IL 60637; patients with LN who are at risk for progression to renal failure † Section of Rheumatology, Department of Medicine and Knapp Center for Lupus (15, 16) (C. Hsieh et al., submitted for publication). In contrast, Research, University of Chicago, Chicago, IL 60637; ‡Section of Nephrology, De- partment of Medicine, University of Chicago, Chicago, IL 60637; and xDepartment the National Institutes of Health activity index, which primarily of Rheumatology, Hospital for Special Surgery, Cornell University, New York, NY assesses glomerular inflammation, does not correlate with prog- 10021 nosis (15, 18, 19). Furthermore, the presence of tubulointerstitial 1 A.C. and S.G.H. contributed equally to this work. scarring on renal biopsy is more predictive of subsequent renal Received for publication June 16, 2010. Accepted for publication November 23, failure than is glomerular scarring (15, 16) (C. Hsieh et al., sub- 2010. mitted for publication). TI can occur independently of GN (20, This work was supported by grants from the Lupus Research Institute and the 21), and TI severity does not correlate with titers of anti-dsDNA National Institutes of Health (AI082724, AR055646). Abs (C. Hsieh et al., submitted for publication). These data indicate Address correspondence and reprint requests to Dr. Marcus Clark, Section of Rheu- matology, Department of Medicine, University of Chicago, 5841 S. Maryland Ave- that TI is an important manifestation of LN that might arise from nue, Chicago, IL 60637. E-mail address: [email protected] different pathogenic mechanisms from those implicated in GN. The online version of this article contains supplemental material. Organ-specific inflammation is a defining feature of many au- Abbreviations used in this article: FDC, follicular dendritic cell; FWR, framework toimmune diseases, including Hashimoto’s thyroiditis (22), rheu- region; GC, germinal center; GN, glomerulonephritis; IMGT, international ImMuno- matoid arthritis (RA) (23), Sjo¨gren’s syndrome (24), and multiple GeneTics information system; LCM, laser capture microdissection; LN, lupus nephritis; R, replacement; RA, rheumatoid arthritis; S, silent; SLE, systemic lupus sclerosis (25). In these diseases, infiltrating lymphocytes are often erythematosus; TBM, tubular basement membrane; TBMIC, tubular basement mem- highly organized and resemble lymphoid structures found in sec- brane immune complex; TI, tubulointerstitial inflammation; TLN, tertiary lymphoid ondary lymphoid organs during a normal immune response. This neogenesis. feature is referred to as tertiary lymphoid neogenesis (TLN). Copyright Ó 2011 by The American Association of Immunologists, Inc. 0022-1767/11/$16.00 B cells within these lymphoid structures secrete autoantibodies www.jimmunol.org/cgi/doi/10.4049/jimmunol.1001983 1850 IN SITU IMMUNE RESPONSES IN LUPUS NEPHRITIS

(26, 27) and are required to locally maintain activated T cells Laser capture microdissection of interstitial CD38+ or Ki-67+ (28). Several reports noted that infiltrating T cells are a prom- cells inent feature of LN (29–33), whereas the presence of B cells At the time of procurement, renal biopsies were immediately frozen in OCT was recently noted (34–36). However, the significance of these media (Tissue-Tek, Torrance, CA) and stored at 280˚C. These consisted of lymphocyte populations in the interstitial infiltrate was unclear. 10 kidney biopsies from 10 patients. Patients A, B, C, D, and F were In this article, we demonstrate that in moderate or severe TI the females (23–40 y of age; one white, three African American, and one inflammatory infiltrates are usually organized into structures re- Hispanic) with established diagnoses of SLE from 4 mo to 10 y who were treated with low doses of prednisone prior to kidney biopsy. Patients E, G, miniscent of those observed in secondary lymphoid organs. Most H, and J were females (13–37 y of age; three African American and one commonly, aggregates of T and B cells containing plasmablasts Hispanic) with new-onset SLE and no significant immunosuppressive were observed. However, in some renal biopsies there were ger- therapy prior to renal biopsy. Patient I was a 28-year-old male with hy- minal center (GC)-like structures containing well-organized fol- pertension and anabolic steroid use who had acute kidney injury, hyper- bilirubinemia, and no evidence of SLE. Eight-micrometer-thick frozen licular dendritic cell (FDC) networks and centroblasts. Both tissue sections were placed on positively charged glass slides and immu- lymphoid structures were functional, because they were associated nostained for CD38 (DAKO) or Ki-67, as previously described (40). with in situ B cell clonal expansion and somatic hypermutation. Briefly, cryosections were fixed in acetone at 220˚C for 5 min, washed These findings implicate organ-intrinsic adaptive-immune re- with PBS on ice for 30 s, treated with 0.1% H2O2 for 30 s, and incubated in sponses in the pathogenesis of lupus TI. PBS with 10% normal goat serum for 2 min. Tissue sections were incubated with anti-CD38 or anti-Ki67 Abs for 10 min, rinsed in PBS, incubated with HRP-conjugated horse anti-mouse Abs and 5% goat serum in PBS for 5 min, rinsed in PBS, and incubated in 3,39-diaminobenzidine

Materials and Methods (Vector Laboratories, Burlingame, CA) for 5 min. All incubation steps were Downloaded from Patients and renal biopsies performed with the glass slide on an ice pack. Finally, the sections were dehydrated in graded alcohols (95 and 100% EtOH) for 30 s each and cleared The University of Chicago Medical Center Institutional Review Board with two rinses of xylene for 3 min each. Laser capture microdissection approved this study. We reviewed the pathology files at the University of (LCM) using the Arcturus Pixcell II (Molecular Devices, Sunnyvale, CA) Chicago Medical Center for inpatient renal biopsies consistent with LN platform and Capsure HS LCM caps (Molecular Devices) was performed to between 2001 and 2007. Among this group, 68 subjects had biopsies isolate CD38- or Ki-67–staining cells. The pulse power was 70 mW, size containing sufficient material for analysis (six or more glomeruli and spot diameter was 15 mm, pulse duration was 5 ms, and target voltage was http://www.jimmunol.org/ a length $0.5 cm) and that did not display Class I or VI nephritis, as 170 mW. Extraction reservoirs were placed directly on the HS LCM caps. defined by the 2003 International Society of Nephrology/Renal Pathology Society revised LN classification criteria (37), and who, on review of Synthesis of cDNA records, fulfilled American College of Rheumatology revised criteria for 3 the classification of SLE (38). Each diagnostic biopsy sample consisted of Ten microliters of lysis buffer (0.5 PBS, 10 mM DTT, 20 RNAsin at least three tissue cores that were predominantly divided for light mi- [Promega, Madison, WI], and 1 U Prime RNAse Inhibitor [Eppendorf, croscopy, with smaller portions submitted for immunofluorescence and Hamburg, Germany]) was aliquoted directly into the extraction reservoir of the LCM cap. Caps were immediately placed on dry ice and stored at electron microscopy. Using the National Institutes of Health system, the 2 activity and chronicity indices were scored at the time of renal biopsy by 70˚C. cDNA was synthesized within the extraction reservoir using 25 ml one of two renal pathologists (A.C., S.M.M.) (39). Control normal renal RT-PCR buffer [375 ng random hexamer primer (pd(N)6, Amersham Pharmacia Biotech), 2 mM 29-deoxynucleoside 59-triphosphate-Mix tissue was obtained from autopsy studies. The clinical charts were by guest on September 24, 2021 reviewed to collect pertinent clinical data, including age, gender, disease (Promega), 10 mM DTT, 5% v/v Nonidet P-40, 10 U RNAsin, 15 U manifestations, medication history, and serological parameters. Standard Prime RNase Inhibitor, and 125 U Superscript II reverse transcriptase procedures were used to process formalin-fixed, paraffin-embedded 2-mm (Invitrogen)]. The cap was incubated for 60 min at 37˚C. RT-PCR reactions tissue sections for evaluation by light microscopy. For each biopsy, five were collected by centrifugation at 8000 rpm for 60 s into a 500-ml col- distributed H&E stains and three periodic acid-Schiff stains were per- lection tube and heat inactivated at 70˚C for 15 min. formed. Standard procedures for direct immunofluorescence microscopy were applied to all cases with FITC-conjugated Abs reactive with the Amplification of Ig H and L chains following Ags: IgG, IgA, IgM, C3, C1q, fibrinogen, k and l L chains, and g,Igl, and Igk rearrangements were amplified by two rounds of nested albumin (DAKO, Carpinteria, CA). The intensity of immunofluorescence PCR in a 40-ml volume (50 nM primers, 1 mM 29-deoxynucleoside 59- staining was semiquantitatively scored on a scale of 0 to 4+. All biopsies triphosphates, 13 Coral Load PCR buffer [Qiagen, Valencia, CA], 1.25 U demonstrated strong immunofluorescence glomerular staining generally in HotstarPlus Taq polymerase [Qiagen], and 8 ml cDNA [or first-round PCR a full-house (IgG, IgA, IgM, C3, C1q) pattern, which is characteristic of product]). The IgG-specific primers consisted of mixtures complementary LN. Standard procedures for electron microscopy were applied to evaluate to conserved leader and framework regions from H and L chain variable the renal biopsies using a Philips CM10 electron microscope. regions and conserved 39 constant L chain regions (41–43) and the 39g C Standard immunohistochemistry was performed on serial paraffin tissue region (44). It should be noted that the originally published primer se- sections using mAbs to CD3 (Labvision, Fremont, CA), CD20 (DAKO), quences 39Sal1-JH1/2 and 39Sal1-JH4/5 each had a base pair deletion. The CD45 (DAKO), MUM1 (DAKO), and CD138 (DAKO) as primary reagents correct sequences are 39Sal1-JH1/2 59-TGCGAAGTCGACGCTGAGGA- and appropriate HRP-conjugated secondary Abs. Isotype controls for each GACAGTGACCAGG-39 and 39Sal1-JH4/5 59-TGCGAAGTCGACGCTG- primary reagent, using the matched secondary reagent, are provided in AGGAGACGGTGACCAGG-39. Targeted sequences were amplified using Supplemental Fig. 1. Ags were retrieved by boiling slides for 20 min under 50 cycles of PCR under the following conditions: 94˚C for 30 s, 57˚C pressure in 1 mM EDTA (pH 8). Double stains were done using the En- (g/Igk) or 60˚C (Igl) for 30 s, 72˚C for 55 s, with a final extension cycle Vision G/2 Doublestain System (DAKO). The number of positively staining of 72˚C for 7 min. cells was counted without knowledge of the clinical data by one renal pathologist (A.C.). Biopsies with prominent aggregates of B cells were Sequence analysis stained with CD21 (DAKO) and Ki-67 (Labvision), and others were stained with CD4 (Labvision), CD8 (Labvision), CXCL10 (R&D Systems, Min- PCR products were cloned into the pCR4-TOPO TAvector (Invitrogen), and neapolis, MN), CXCL12 (R&D Systems), CXCL13 (R&D Systems), plasmid DNA was purified using QIAprep spin mini-prep kits (Qiagen). CCL21 (R&D Systems), and BAFF (Alexis Biochemicals, San Diego, Multiple bacterial clones were randomly selected for sequencing to ensure CA). The interstitial infiltrate was categorized into three patterns: diffuse proper sampling. The V-QUEST program from the international ImMuno- and scattered, T:B cell aggregates, and ectopic GC. An arbitrary cut-off of GeneTics (IMGT) information system (http://imgt.cines.fr/IMGT_vquest/ 50 cells composed of T and B cells satisfied the designation of T:B cell vquest) was used to identify the V(D)J germline, as well as the CDRs and aggregate pattern. The presence of an FDC (CD21+) network was neces- framework regions (FWRs). Nucleotide and amino acid mutations in the sary to assign the GC pattern. Of note, the T:B cell aggregate always had V-region were identified by alignment with the closest corresponding areas with diffuse inflammation, and the ectopic GC pattern often showed germline using IMGT. In some cases, the corresponding genomic V seg- T:B cell aggregates and diffuse inflammation. Univariate statistical anal- ments were amplified and sequenced to confirm that the observed mutations ysis was performed using the Fisher Exact test and the Mann–Whitney U did not represent allelic polymorphisms. Multiple nucleotide changes in test, with p , 0.05 considered statistically significant. a single codon were scored as a single replacement mutation. Related clones The Journal of Immunology 1851 were defined by similar CDR3 regions, as identified by the junction analysis from glomeruli with only rare circulating CD20+ B cells in glo- software provided by IMGT. Sequences that were out of frame or contained merular capillaries. In 46% of cases (31/68), there were well- mutations that resulted in a nonproductive sequence were excluded from + + ∼ circumscribed aggregates of CD20 B cells or CD138 plasma analysis ( 15% of all sequences; data not shown). For tabulating clonal + frequency, a clone was counted if it was obtained from separate LCM picks cells with CD3 T cells in tubulointerstitium. Most of these T cells or if it differed by three or more nucleotides from other cloned sequences expressed CD4 (data not shown). Finally, in 6% of biopsies (4/68), within the same LCM “pick”. This latter criterion was used to exclude structures consistent with GCs were observed. In these structures, clones that might appear different as a result of PCR error. Ag-driven se- CD20+ B cells preferentially occupied the central zone, whereas lection was calculated using the JAVA applet from Lossos et al. (45) (http:// + + www-stat.stanford.edu/immunoglobulin/). Genealogical trees showing the CD3 CD4 T cells tended to occur peripherally. Staining with the relationships between plasma cells were constructed by analysis of the dendritic cell marker CD21 revealed a central dense reticular pattern of somatic mutations. network characteristic of FDCs. CD138+ plasma cells were rare within the central FDC network but common in the surrounding Results areas. The spatial organizations of B cells, T cells, FDCs, and To examine the possible pathogenic significance of B cells in LN, tingible body macrophages were all consistent with the observed we identified 68 patients, with SLE diagnosed by a rheumatologist, histological structures being bona fide GCs. who had undergone diagnostic renal biopsies for presumed LN Chemokines induce and maintain the spatial organization of (C. Hsieh et al., submitted for publication). The average age of immunocytes within secondary lymphoid organs (46). Therefore, this cohort was 31 y, 85% were female, and 81% were African we examined whether some of these same chemokines were American. At the time of biopsy, the average duration of disease present in SLE renal biopsies (Fig. 2). When biopsies with T:B was 36 mo, the median creatinine was 1.0 mg/dl, and 81% had aggregates or GC phenotype were stained, CXCL12 (46) and Downloaded from detectable Abs to dsDNA. Twenty-five percent had only low-dose BAFF (47) were observed in almost all samples tested with oral prednisone prior to renal biopsy (#20 mg/d). Three of the prevalences of 95% (18 of 19 biopsies) and 100% (4 of 4 bi- biopsies were International Society of Nephrology/Renal Pathol- opsies), respectively. Several other chemokines were also com- ogy Society class II, 22 were class III, 33 were class IV, and 10 monly expressed. Approximately 70% (12/17) of biopsies with the were class V. Tissue sections from each biopsy were examined T:B aggregate or GC phenotypes had detectable staining for

first by immunohistochemistry with Abs to CD45, CD3, CD4, CXCL13 (35), 50% (8/16) had detectable staining for CCL21, and http://www.jimmunol.org/ CD20, and CD138. 42% (8/19) had detectable staining for CXCL10. These chemokines were not detectable in normal renal tissue (data not shown). Ectopic lymphoid structures in LN Interestingly, there was little chemokine staining in renal biopsies Three distinct patterns of B cell infiltration were evident (Fig. 1). In with a diffuse histological phenotype (n = 10; data not shown). 48% of biopsies (33/68), the predominant pattern was one of di- There were no clear differences between the T:B aggregates and ffuse and scattered lymphocytic (CD45+) infiltration, with varying GC phenotypes, except that the GC phenotype was associated degrees of coinfiltration with CD20+ B lymphocytes or CD138+ with more intense chemokine staining. These data suggested that plasma cells. Plasma cells and B cells were invariably excluded similar factors may organize and maintain the T:B and GC-like by guest on September 24, 2021

FIGURE 1. Lymphoid neogenesis in the tubulointerstitium of LN. Three distinct patterns of lymphoid involvement were observed in LN. In the diffuse pattern (A–D), CD138+ plasma cells (A, B) were scattered throughout the renal interstitium between tubules and around Bowman’s capsule. MUM1+ (C) and CD20+ (D) B cells demonstrated a similar distribution (original magnification 3400). E–J,Three examples of the T:B cell aggregate pattern. Immu- nostaining of sequential histological sections revealed well-circumscribed aggregates of CD3+ Tcells(E, G, I)andCD20+ Bcells(F, H)(E and F, original magnification 3200; G–J, original magnification 3400) or CD138+ plasma cells (J; 3400). In the GC pattern, a dense aggregate of lymphocytes consistent with a GC was evident on H&E-stained tissue sections (K; original magnification 3200). On serial sections, CD20+ B cells (L) were clustered in the middle of the lymphoid aggregate surrounded by CD3+ (M) and CD4+ (N) T cells. Scattered CD138+ plasma cells (O) were observed in the surrounding parenchyma. Staining for CD21 revealed a dense FDC network characteristic of GCs (P–R) [the CD21 staining corresponding to the previous panels is provided first followed (P)by examples of CD21 staining in two additional biopsies with a GC (Q, R)] (original magnification 3400). 1852 IN SITU IMMUNE RESPONSES IN LUPUS NEPHRITIS

FIGURE 2. Chemokine expression in lymphoid neogenesis. Strong BAFF immunohistochemical staining was observed in T:B cell aggregates (A; original magnification 3400) and in GCs (B; original magnification 3200). In T:B aggregates, CXCL12 (C) and CXCL13 (D) was often present, whereas CXCL10 (E) and CCL21 (F) were found in fewer cases and often with less staining intensity. C–F, Original magnification 3400.

histological phenotypes. In contrast, different mechanisms may me- flammation (.25% of the interstitium infiltrated by inflammatory Downloaded from diate the diffuse accumulation of plasma cells and B cells in LN. cells) was more likely to be found in biopsies with the T:B aggre- To begin to determine whether the observed histological struc- gate (29 of 31 biopsies) and GC (4 of 4 biopsies) patterns compared tures were functional (48), we determined whether they were as- with 16 of 33 biopsies with the diffuse pattern (p = 0.00002, Fisher sociated with in situ lymphocyte proliferation. Therefore, biopsies exact test) (Fig. 4A). Detectable tubular basement membrane im- manifesting the diffuse, T:B aggregate and GC patterns were mune complexes (TBMICs) (immunofluorescence) were also in-

stained with Abs specific for the proliferative marker Ki-67. As frequent in biopsies manifesting a diffuse pattern (6/33) (Fig. http://www.jimmunol.org/ demonstrated in Fig. 3, the centers of the GC structures contained 4B,4C) but were a usual feature of biopsies with T:B aggregates numerous small proliferating cells. Ki-67+ cells were also com- or GCs (21/35) (Fig. 4B). This difference was highly significant monly observed in T:B aggregates (14 of 22 biopsies; 64%). In (p = 0.00014). Electron microscopic analysis of the biopsy in Fig. contrast, biopsies with a diffuse histology infrequently had Ki-67+ 4D, which had a T:B aggregate histological pattern, revealed that cells (2 of 11 biopsies; 18%). When present, there was usually no the TBMICs resided within the tubular basement membrane (TBM), more than one Ki-67+ cell per high-power field. Ki-67+ tubular characteristic of lupus interstitial nephritis (Fig. 4E). Additional im- epithelial cells were infrequently observed in biopsies from all munofluorescence and electron microscopy images demonstrating three histological groups (data not shown). TBMICs are provided in Supplemental Fig. 2.

To determine whether the observed proliferating cells were by guest on September 24, 2021 lymphocytes, we performed two-color immunohistochemistry with Characterization of in situ Ig repertoire Abs speciﬁc for Ki-67 and either CD20 or CD3. Ki-67+CD4+ positive cells were infrequently observed in all three histological The presence of lymphoid-like structures and aggregates of pro- patterns (data not shown). In contrast, numerous Ki-67+CD20+ cells liferating plasmablasts in LN suggested that in situ Ag-driven were observed in the GC structures (Fig. 2). Ki-67+CD20+ cells clonal expansion and somatic hypermutation were occurring. To were also observed in the T:B aggregates, but they were a minor test this directly, we used LCM coupled to RT-PCR and sequencing fraction of the Ki-67+ cells (,10%). Most Ki-67+ cells were to characterize the in situ H and L chain repertoire in renal biopsies CD20low or CD202. However, most of these Ki-67+ cells in the T:B from LN patients. We analyzed nine patients. On anti-CD38 Ab aggregates expressed CD138 and, therefore, were plasmablasts. stained fresh-frozen sections we used LCM to sample a GC from In situ lymphocyte organization correlated with the extent of TI one LN patient (patient A) and T:B aggregates from four other LN and speciﬁc pathological features (Fig. 4). Severe interstitial in- patients (patients B–E). For comparison, we sampled the expressed Ig repertoire in three LN patients with a diffuse histology (patients F–H) and one nonlupus patient with idiopathic interstitial nephritis (patient I). We also stained sections with anti-Ki67 Abs and sampled a cluster of proliferating cells in a LN patient with T:B aggregates on biopsy (patient J).

Ig clonal restriction and Ag-driven somatic mutation within the intrarenal GC Patient A was a 27-year-old white female with a 7-mo history of SLE who had been treated with low-dose methotrexate and pre- FIGURE 3. In situ proliferating centroblasts or plasmablasts in LN. dnisone. The dsDNA Ab titer at the time of biopsy was 1:2560, and Staining for Ki-67 demonstrated that small proliferating cells were com- anti-Sm Ab was negative. LCM was used to sample 12 separate monly observed in T:B aggregates and in GCs but were rare in patients areas within a GC and another 12 from the surrounding tubu- with a diffuse histology (upper panels; left to right, original magnification lointerstitium. Each LCM pick sampled one to eight visible CD38+ 3400, 3400, 3200). Double staining for CD20 (red) and Ki-67 (brown) (middle panels; original magnification 3400 [left], 3200 [right]) revealed cells. The distribution and frequency of the most commonly that most proliferating cells in the GC were CD20+. In contrast, in the T:B expressed Ig genes identified from the GC or surrounding aggregates, most Ki-67+ cells were CD20low or CD202. Double staining parenchyma are provided in Fig. 5. A total of 26 g,41l, and 8 k for Ki-67 (blue) and CD138 (brown) (lower panel; original magnification distinct sequences were identified. Because only a few k se- 3400) revealed numerous Ki-67+CD138+ plasmablasts (arrows). quences were isolated, they were excluded from further analysis. The Journal of Immunology 1853 Downloaded from

FIGURE 4. Increased TI and immune complex deposition in biopsies with T:B aggregates or GCs. A, The T:B cell aggregate and GC histological patterns tend to have increased tubulointerstitial infiltration by inflammatory cells compared with the diffuse pattern (p = 0.00002; Fisher exact test). The extent of interstitial inflammation was expressed as the average percentage of tubulointerstitium cross-sectional area involved with CD45+ inflammatory cells. B, TBMIC deposition was more common in biopsies with the T:B aggregate or GC pattern than with the diffuse pattern (p = 0.00014; Fisher exact test). C, IgG direct immunofluorescence staining shows no significant TBM staining in a biopsy with diffuse interstitial inflammation (original magnification 3400). D, Strong granular TBM IgG staining is present in a biopsy with T:B cell aggregates (original magnification 3200). E, Electron microscopy from the same http://www.jimmunol.org/ biopsy as in D reveals numerous electron-dense deposits (arrows) that represent immune complexes within the TBMs (original magnification 32950).

Of the 26 distinct expressed g V sequences identified, 12 were All of the VH3-9*01D3-10*01JH3*02-encoded sequences were cloned from the GC, and 14 were cloned from the surrounding heavily mutated compared with reported germline segments, with tubulointerstitium. Among the 12 GC g sequences, four (33%) all containing a similar core of coding and noncoding mutations arose from a single rearrangement (VH3-9*01D3-10*01JH3*02). This (Fig. 6A). It was possible that some of the apparent mutations in rearrangement was also the predominant expressed H chain de- VH3-9*01D3-10*01JH3*02 represented an unreported allelic poly- by guest on September 24, 2021 tected in the parenchyma surrounding the GC (Fig. 5). Another morphic form of VH3-9. However, these mutations were not ob- three (25%) of the GC g V sequences identified arose from an- served when the corresponding genomic VH3-9 segment from other single rearrangement (VH3-23*01D3-22*01JH3*02). These data patient A was amplified from peripheral blood and sequenced indicated that more than half (58%) of the g-chain repertoire (data not shown). within the GC arose from two unique recombination events. Comparison of the different VH3-9*01D3-10*01JH3*02-expressed sequences isolated from the GC and the surrounding parenchyma suggested that some B cells had sequentially acquired Ig mutations during clonal expansion. The relative genealogy of these mutations can be demonstrated in a clonal tree (Fig. 6B). The accumulation of such hierarchical g mutations, as well as the fact that they were identified in closely adjacent B cells, is consistent with in situ clonal expansion and somatic hypermutation (49). To examine whether Ag was selecting for particular somatic mutations within the GC, we analyzed the type and frequency of mutations in the VH3-9*01D3-10*01JH3*02-encoded V regions (containing CDR1 and CDR2). In the absence of antigenic selection, replacement (R) and silent (S) mutations occur randomly in the CDRs and FWRs. However, if there is Ag-driven selection, R mutations in the CDRs will be overrepresented (50), whereas R mutations in the FWRs will be underrepresented. We used the multinomial method of Lossos et al. (45) to determine the probability that B cells expressing VH3-9*01D3-10*01JH3*02-encoded V regions had undergone Ag-driven selection (Fig. 6C). R mutations were overrepresented (R/S = 8:1; p = 0.028–0.035) in the CDRs and underrepresented in the FWRs (12:8 to 13:7; p = 0.018–0.041). FIGURE 5. Schematic representation showing the most frequently A similar picture was obtained when the l-chain GC repertoire expressed Ig genes found within a GC and the surrounding parenchyma from patient A. Expressed Ig genes observed outside the GC were grouped was analyzed. Of the seven sequences obtained from the GC (Fig. into those that were also found in the GC (GC Related) or those that were 5), two were from one unique recombination (VL2-8*01JL2*01) not represented in the GC (GC Unrelated). The frequency of each (Fig. 7A), and two were from another unique recombination expressed clone relative to the GC or the extra-GC pools of expressed H (VL1-44*01JL3*02), indicating that more than half (57%) of the and L chain Igs is provided. detectable GC l clones arose from one of two unique recom- 1854 IN SITU IMMUNE RESPONSES IN LUPUS NEPHRITIS Downloaded from http://www.jimmunol.org/ by guest on September 24, 2021

FIGURE 6. Evidence of in situ clonal expansion and Ag-selected somatic hypermutation in a intrarenal GC. A, Nucleotide sequences of VH3-9*01 D3-10*01JH3*02 amplified transcripts were aligned with germline. Identical nucleotides are indicated by dashes. Dots indicate unmutated intervening portions of the V regions. Numbers refer to IMGT amino acid positions. Sequences isolated from the GC are marked with an asterisk. B, Genealogical relationships of sequences are illustrated in a clonal tree with a predicted germline clone at the top. C, R/S ratios of clones in FWRs and CDRs (including the FWR 59 to CDR1, which is not shown in A). The p values represent the probability that the observed mutations arose from random chance and had not been selected for affinity to Ag. bination events. Both of these rearrangements were also found in tion of R and S mutations in the CDR (R/S = 3:0; p = 0.021) was the surrounding parenchyma (Fig. 5). consistent with Ag-driven selection (Fig. 7C). A similar trend was Alignment of the different VL2-8*01JL2*01-derived sequences seen in the FWRs (R/S = 3:1; p = 0.155). with predicted germline sequences indicated that all had accu- mulated somatic mutations (Fig. 7A). However, the overall num- Clonal selection in the tubulointerstitium surrounding the GC ber of mutations was less than that observed in the clonally ex- Analysis of the surrounding parenchyma provided further evidence panded GC g sequences (Fig. 6A). All three mutations observed in of clonal selection. Of the 34 distinct l sequences identified in the CDR1 and CDR2 encoded amino acid replacements, again sug- tubulointerstitium, four arose from a single VL2-14*01JL2*01 rear- gesting selection by Ag. Sequencing of a different, nonselected rangement, and three arose from a single VL1-47*01JL2*01 rear- VL2-8*01 containing l V region from patient A confirmed that the rangement. In addition, four distinct interstitial l sequences arose observed mutations did not represent allelic variation (data not from two genomic rearrangements identified in GC-expressed se- shown). Assembly of the different VL2-8*01JL2*01 sequences into quences (Fig. 5). Therefore, 32% (11/34) of the identified expres- a clonal tree revealed a mutational hierarchy suggestive of on- sed l sequences were derived from four rearrangement events. going clonal expansion and somatic hypermutation (Fig. 7B). In Alignment and further analysis of the VL2-14*01JL2*01-related the most related VL2-8*01JL2*01 sequences (Al4b,c), the distribu- clones with the predicted corresponding germline segments The Journal of Immunology 1855 Downloaded from

FIGURE 7. Analysis of a L chain selected in the GC also revealed evidence of Ag-selected somatic hypermutation. A, Nucleotide sequences of VL2-8*01 JL2*01 ampliﬁed transcripts were aligned with germline, as in Fig. 6. B, Genealogical relationships of sequences are illustrated in a clonal tree with a predicted germline clone at the top. The black circle indicates a proposed intermediate clone. C, R/S ratios of clones in FWRs and CDRs. The p values http://www.jimmunol.org/ were calculated as in Fig. 6.

revealed evidence of sequential somatic hypermutation and Ag- VL1-40*01JL3*02 clones were consistent with extensive somatic driven clonal selection (Supplemental Fig. 3). hypermutation and Ag-driven clonal selection (Supplemental Fig. Comparison of the repertoire between the GC and surrounding 4). There was also evidence of clonality in the 31k sequences, tubulointerstitium revealed important interrelationships and sig- with an overall clonality of 11/31 or 35% (data not shown). nificant differences. As described above, the repertoire of expressed Clonality was observed in the Ig repertoire expressed in the three g and l Ig chains in the GC was well represented in the paren- other T:B aggregate biopsies that were sampled (patients C–E). A by guest on September 24, 2021 chyma. However, several clonally expanded and/or prevalent pa- summary of these results is provided in Supplemental Table I. renchymal l Ig chains were absent from the GC (Fig. 5). These Clonally related sequences from three of the four biopsies were observations suggested that the GC can contribute to the paren- heavily somatically mutated (Fig. 8, data not shown). However, in chymal repertoire but that the parenchyma does not necessarily one T:B aggregate (patient C), the observed clonally selected contribute to the GC repertoire. k-chains (9 of 37 distinct sequences) were very similar to the corresponding germline sequences (Supplemental Fig. 5). These In situ Ig expression in T:B aggregates results indicated that T:B aggregates are associated with moderate Patient B was a 40-year-old African American female with a 4-mo clonal restriction. Ongoing somatic hypermutation was observed history of SLE treated with 20 mg/d of oral prednisone. The anti- but was not an invariant feature of the Ig chains being selected in Sm Ab was positive, and the anti-dsDNA Ab titer prior to biopsy the T:B aggregates. was 1:320. LCM was used to obtain 28 independent samples from In the GC histology, 17 of 41 distinct l-chains used VL2-14*01, two different T:B aggregates. A total of 68 distinct g,27l, and 31 and 10 used VL1-44*01. For the four patients with a T:B aggregate k sequences was identified. histology, 33 of 94 were VL2-14*01, and 19 were VL1-44*01. Within the g population, 10 rearrangements were observed Overall, these two V regions were found in ∼55% of identified more than once (24 of 68 sequences, for an overall clonality of l-chains. In the GC and T:B aggregate histologies, the VL2-14*01 35%). The most common rearrangement was observed four times and VL1-44*01 segments were primarily found in expanded clonal (4/68; VH1-3*01D4-23-*01JH4*02). Alignment of these cDNA frag- populations. These results indicated that the overall frequency of ments with the predicted germline sequence revealed that the iden- specific expressed l variable segments was similar between the tified sequences had undergone extensive somatic hypermutation GC and T:B histological patterns. (Fig. 8A). These clones could also be assembled into a simple To determine whether the observed restricted-expressed reper- clonal tree, suggesting ongoing somatic hypermutation (Fig. 8B). toires were a specific feature of the GC and T:B aggregates, we Analysis of the distribution of mutations indicated selection for sampled the expressed Ig repertoire in three LN patients who had replacement mutations in the CDRs and for silent mutations in diffuse B cell infiltration on biopsy (Supplemental Table I; patients the FWRs (Fig. 8C). These findings are consistent with Ag-driven F–H). From patient F, we cloned and compared 37 distinct Ig H clonal selection. chains. Only two arose from the same rearrangement. In the sec- There was also evidence of clonality when the corresponding ond patient (patient G), of 37 sequences, 3 arose from one rear- L chain sequences were examined. Of the 27 distinct l sequences, rangement, 3 arose from another rearrangement, and 2 arose from four rearrangements were found more than once, with one ob- a third rearrangement. In patient H, of 31 sequences, 3 arose from served three times (VL1-40*01JL3*02), for an overall clonality of 13/ one rearrangement, and there were two examples where the same 27 or 48%. The frequency and distribution of mutations in the rearrangement was observed twice. A similar degree of clonality 1856 IN SITU IMMUNE RESPONSES IN LUPUS NEPHRITIS Downloaded from http://www.jimmunol.org/

FIGURE 8. Clonal selection in a T:B aggregate. A, Nucleotide sequences of VH1-3*01D4-23*01JH4*02 amplified transcripts were aligned with predicted by guest on September 24, 2021 germline sequences as in Fig. 6. B, Genealogical relationships of sequences are illustrated in a clonal tree with predicted germline clone at top. C, R/S ratios of clones in FWRs and CDRs. The p values were calculated as in Fig. 6. was observed in a nonlupus patient (patient I) who had idiopathic mutations, with unique mutations observed in some clones (Fig. 9A). acute interstitial nephritis with diffuse B cell infiltration (Sup- These differences, which allowed construction of a clonal tree, sug- plemental Table I). In that patient, we identified 34 unique se- gest ongoing somatic hypermutation (Fig. 9B). Interestingly, analysis quences that arose from 29 unique recombination events. Four re- of the R/S distribution (Fig. 9C) in the CDR1- and CDR2-containing combinations were observed twice, whereas none was observed regions revealed no evidence that these mutations had undergone Ag three or more times. Therefore, a modest degree of clonal restric- selection. A similar pattern of mutations was observed in those ex- tion can be observed in biopsies with a diffuse pattern of B cell pressed clones arising from the VH4-34*01D1-1-*01JH4*02 recombi- infiltration. However, this degree of clonality does not seem to be nation (data not shown). a specific feature of lupus interstitial nephritis. In general, we observed more clonal restriction in biopsies with Discussion more organized lymphocytic infiltrates. However, there were Interstitial inflammation is a prominent feature of human LN that, exceptions; one biopsy with T:B aggregates had a similar degree of independently of glomerular involvement, identifies patients at clonality (patient E) as that observed in one of the biopsies risk for subsequent renal failure (15, 16) (C. Hsieh et al., submitted manifesting a diffuse morphology (patient H). These observations for publication). In this study, we demonstrated that, in more than indicated that the correlation between the degree of lymphocyte half of our patient cohort, the interstitial infiltrate was organized organization and in situ selection is imperfect. It might be expec- into lymphoid-like structures competent to select for B cells ted that evidence of in situ proliferation may more directly iden- expressing a highly restricted Ig repertoire. The presence of tify B cells that are being selected in situ. Therefore, we used LCM lymphoid-like structures strongly correlated with detectable TBMICs. to capture an aggregate of Ki-67+ cells in a patient with T:B ag- These observations suggested that in LN, GCs and T:B aggregates gregates on histology (Patient J). select for cells that locally secrete pathogenic Abs in the tubu- We cloned 18 unique Ig H chains from a single Ki-67+ focus. Of lointerstitium. these, 11 arose from two unique recombinations (patient J). These The two histological patterns, GC and T:B aggregates, seemed two expressed H chains used the same Vh (VH4-34*01) segment but to reflect different underlying states of B cell selection. Many as- different D (DH1-1*01 and DH5-18*01) and J (VH4*02 and VH4*02) pects of clonal selection in the intrarenal GC were typical of those regions. Analysis of the most common H chain recombination observed in GCs residing in secondary lymphoid structures (51– (VH4-34*01D5-18-*01JH4*02), constituting 6 of 18 clones, is pro- 53). As has been reported for rodent (54) and human (55) GCs, vided in Fig. 9. The aligned sequences contained several common only a few clones accounted for a majority of the sampled rep- The Journal of Immunology 1857

A Downloaded from http://www.jimmunol.org/

B C by guest on September 24, 2021

+ FIGURE 9. Clonal selection in a Ki-67 foci. A, Nucleotide sequences of VH4-34*01D5-18*01JH4*02 amplified transcripts were aligned with predicted germline sequences as in Fig. 6. B, Genealogical relationships of sequences are illustrated in a clonal tree with predicted germline clone at top. C, R/S ratios of clones in FWRs and CDRs. The p values were calculated as in Fig. 6. ertoire. Furthermore, most, if not all, of the observed predominant strating that GCs are open structures that allow B cells to enter clones had undergone somatic hypermutation. However, analysis freely and scan for Ag (52, 56, 57). Our results are also very of the frequency and distribution of mutations in nonselected similar to those obtained when human lymph nodes from normal (singly occurring) GC-expressed Ig chains indicated that most had volunteers were characterized (55). not undergone Ag selection (data not shown). This latter obser- In contrast to the proliferating centroblasts observed in the vation is consistent with elegant in vivo imaging studies demon- GCs, plasmablasts predominated in those patients with T:B aggre- 1858 IN SITU IMMUNE RESPONSES IN LUPUS NEPHRITIS gates. These foci of T cells and plasmablasts are reminiscent of MRL/Mplpr/lpr mice have diffuse or perivascular intrarenal lym- the extrafollicular B cell responses that were recently described in phocytic infiltrations. The in situ organization of B and T cells some murine models of autoimmunity (58). In MRL/Mplpr/lpr mice into lymphoid-like structures could be a unique feature of human and in MRL/Mplpr/lpr mice expressing AM14 (a rheumatoid factor LN. Ab), the production of autoantibodies in secondary lymphoid or- We observed a high frequency of L chains containing V2-14*01 gans preferentially occurs in aggregates of plasmablasts residing and V1-44*01 in GC and T:B aggregates. This could reflect a re- outside follicles (59–61). Selection and somatic hypermutation quirement for these segments to encode certain autoreactive spe- might occur in these sites through T-dependent and TLR-dependent cificities (clonal convergence). However, V2-14*01 and V1-44*01 are mechanisms (47, 62). However, to our knowledge, we are the first highly represented in the mature B cell peripheral repertoire of to demonstrate in humans the existence of functional extrafollicular normal individuals and patients with SLE (75). Therefore, the high plasmablast aggregates in an organ targeted by an autoimmune frequencies noted in this study may reflect the repertoire of pre- disease. Furthermore, our data indicated that such plasmablast cursor B cell populations. foci are a usual feature of LN complicated by severe TI. The different histological patterns of involvement in LN were Direct sampling of the proliferating cells in a patient with T:B reminiscent of those observed in the synovial tissue of RA patients aggregates revealed a degree of clonal restriction similar to that (24, 76, 77). In both cases, histological features ranged from observed in an intrarenal GC. These data suggested that Ki-67 diffuse lymphocyte infiltration to fully formed GCs. Sampling the expression directly identifies cells undergoing in situ selection. Ig repertoire in RA synovial GCs (78, 79) and total RA synovium The Ki-67+ cells occurred within areas of T:B aggregates, but not (80, 81) also revealed evidence of clonal restriction and Ag-driven all T:B aggregates contained Ki-67+ cells. This relative discor- somatic hypermutation. However, unlike human LN, TLN in RA Downloaded from dance may explain the variable clonality observed when aggre- has not been related to specific, prognostically important patho- gates, identified by CD38 staining, were sampled for in situ Ig logical features (C. Hsieh et al., submitted for publication). expression. These data suggested that there might not be tight Dense B cell aggregates are present in half of allograft biopsies, correlations between the immunohistological characteristics of and occasional GCs (in 11% of biopsies) may be associated with lymphocyte aggregates and their function in selection. Ab-mediated rejection (82). TLN is a usual feature in kidneys that

In the T:B aggregates, we observed selection for germline- have been removed for terminal rejection (83). In contrast, the http://www.jimmunol.org/ encoded and highly mutated Abs. Most Abs that were mutated frequency and significance of TLN in other immune-mediated appeared to have undergone Ag-driven selection. A notable ex- renal diseases are less clear. Prominent B cell aggregates were re- ception was the expressed IgH cloned from the foci of Ki-67+ cells. ported in IgA and membranous nephropathy (84, 85). However, However, current methods of determining Ag-driven selection GCs have not been reported in either nephropathy, and the pres- cannot assess antigenic pressure on the CDR3 because several ence of lymphoid aggregates has not been associated with specific processes determine diversity in this region. Therefore, if anti- pathological features. A GC was reported in 1 of 16 reported cases genic specificity is determined primarily by the CDR3, then se- of anti-neutrophil cytoplasmic Ab-associated nephritis (34), where- lection might not be apparent for the other regions. This pos- as we did not observe any GCs in 32 cases that we examined (86). by guest on September 24, 2021 sibility might be applicable to VH4-34*01D5-18-*01JH4*02, because These observations suggested that TLN might be much more the CDR3 contains three arginines that are known to confer DNA common in LN than in some other autoimmune nephritides. binding (63). In our study, about half of LN renal biopsies had tubulointer- Regardless of apparent differences in B cell populations usually stitial infiltrates that were organized into higher-order lymphoid selected in each histological type, GC and T:B histological patterns structures. However, this is probably an underestimation of the true were strongly associated with TBMICs. This suggests that both prevalence of GCs and T:B aggregates in LN. This is because lymphoid structures select for cells secreting Abs that form in situ a diagnostic renal biopsy represents only a small fraction of the immune complexes with locally available Ags. It is possible that entire kidney, so sampling error always remains an important the selecting Ags are renal specific, because such Abs have been consideration. Even with this limitation, our data clearly demon- detected in the peripheral serum of patients with LN (64). However, strated that the detection of lymphoid structures on diagnostic it is not clear whether these autoantibodies are produced in situ in biopsy is highly predictive of specific pathological features. the kidney or in conventional lymphoid structures. Based on our studies, we propose that human LN arises from at Our studies focused on in situ Ab secretion; however, it is likely least two distinct pathogenic processes. The deposition of proin- that resident B cells are also contributing to local inflammation by flammatory immune complexes in glomeruli likely arises from presenting MHC class II-restricted Ags to neighboring T cells (65). a breach in systemic tolerance. In contrast, our observations The importance of B cells for maintaining systemic pathogenic demonstrated that interstitial nephritis is associated with in situ T cells was demonstrated in MRL/Mplpr/lpr mice (66–68) and in tolerance diatheses. The relative importance of each immunolog- SLE patients treated with rituximab (69, 70). B cells are also re- ical process is unclear. Furthermore, any interdependence between quired for maintaining T cell infiltrates in the synovium of patients the two processes is not known. However, observations that in- with RA (28). The close proximity of T and B cells in intrarenal terstitial nephritis determines renal survivorship (15, 16) (C. Hsieh GCs and T:B aggregates, as well as the attendant clonal expansion et al., submitted for publication) indicate that intrinsic immuno- of B cells expressing isotype-switched Abs, suggest that these logical processes contribute to disease severity. Identification of B cells are productively interacting with coresident T cells. the in situ Ags and factors promoting local B cell selection and It is unclear whether the available murine models of SLE fully expansion in the interstitium should yield important biomarkers mimic the in situ adaptive-immune responses associated with and could lead to novel therapeutic strategies in LN. human lupus interstitial nephritis. Lymphocytic infiltrates in the kidneys of NZB/NZW and MRL/Mplpr/lpr mice contain B cells and/ Acknowledgments or plasma cells, and the Abs that they express display a broad re- We thank Dr. Martin Weigert for critical evaluation of this work. We also pertoire of specificities (71–74). However, it is not known whether thank Dr. Maria Tretiakova for expert assistance with LCM, Dr. Linda murine models manifest functional intrarenal GCs or TB aggre- Wagner-Weiner for identifying patients, and Sarah Powers for careful read- gates. Rather, available evidence suggests that the NZB/NZW and ing of the manuscript. The Journal of Immunology 1859

Disclosures 28. Takemura, S., P. A. Klimiuk, A. Braun, J. J. Goronzy, and C. M. Weyand. 2001. T cell activation in rheumatoid synovium is B cell dependent. J. Immunol. 167: The authors have no financial conflicts of interest. 4710–4718. 29. Brunati, C., B. Brando, R. Confalonieri, L. S. Belli, M. G. Lavagni, and L. Minetti. 1986. Immunophenotyping of mononuclear cell infiltrates associated References with renal disease. Clin. Nephrol. 26: 15–20. 1. Lahita, R. G. 2004. Systemic Lupus Erythematosus. Elsevier Academic Press, 30. Castiglione, A., A. Bucci, G. Fellin, G. d’Amico, and R. C. Atkins. 1988. The San Diego. relationship of infiltrating renal leucocytes to disease activity in lupus and cry- 2. Bono, L., J. S. Cameron, and J. A. Hicks. 1999. The very long-term prognosis oglobulinaemic glomerulonephritis. Nephron 50: 14–23. and complications of lupus nephritis and its treatment. QJM 92: 211–218. 31. Boucher, A., D. Droz, E. Adafer, and L. H. Noe¨l. 1986. Characterization of 3. Waldman, M., and G. B. Appel. 2006. Update on the treatment of lupus ne- mononuclear cell subsets in renal cellular interstitial infiltrates. Kidney Int. 29: phritis. Kidney Int. 70: 1403–1412. 1043–1049. 4. Davidson, A., and C. Aranow. 2010. Lupus nephritis: lessons from murine 32. D’Agati, V. D., G. B. Appel, D. Estes, D. M. Knowles, II, and C. L. Pirani. 1986. models. Nat Rev Rheumatol 6: 13–20. Monoclonal antibody identification of infiltrating mononuclear leukocytes in 5. Li, L., and C. Mohan. 2007. Genetic basis of murine lupus nephritis. Semin. lupus nephritis. Kidney Int. 30: 573–581. Nephrol. 27: 12–21. 33. Alexopoulos, E., D. Seron, R. B. Hartley, and J. S. Cameron. 1990. Lupus ne- 6. Hicks, J., and D. C. Bullard. 2006. Review of autoimmune (lupus-like) glo- phritis: correlation of interstitial cells with glomerular function. Kidney Int. 37: merulonephritis in murine models. Ultrastruct. Pathol. 30: 345–359. 100–109. 7. Tan, E. M., A. S. Cohen, J. F. Fries, A. T. Masi, D. J. McShane, N. F. Rothfield, 34. Steinmetz, O. M., J. Velden, U. Kneissler, M. Marx, A. Klein, U. Helmchen, J. G. Schaller, N. Talal, and R. J. Winchester. 1982. The 1982 revised criteria for R. A. K. Stahl, and U. Panzer. 2008. Analysis and classification of B-cell the classification of systemic lupus erythematosus. Arthritis Rheum. 25: 1271– infiltrates in lupus and ANCA-associated nephritis. Kidney Int. 74: 448–457. 1277. 35.Lee,H.-T.,Y.M.Shiao,T.-H.Wu,W.-S.Chen,Y.-H.Hsu,S.-F.Tsai,and 8. Plotz, P. H. 2003. The autoantibody repertoire: searching for order. Nat. Rev. C.-Y. Tsai. 2010. Serum BLC/CXCL13 concentrations and renal expression Immunol. 3: 73–78. of CXCL13/CXCR5 in patients with systemic lupus erythematosus and lu- 9. Winfield, J. B., I. Faiferman, and D. Koffler. 1977. Avidity of anti-DNA anti- pus nephritis. J. Rheumatol. 37: 45–52. Downloaded from bodies in serum and IgG glomerular eluates from patients with systemic lupus 36. Hutloff, A., K. Bu¨chner, K. Reiter, H. J. Baelde, M. Odendahl, A. Jacobi, erythematosus. Association of high avidity antinative DNA antibody with glo- T. Do¨rner, and R. A. Kroczek. 2004. Involvement of inducible costimulator in merulonephritis. J. Clin. Invest. 59: 90–96. the exaggerated memory B cell and plasma cell generation in systemic lupus 10. Ebling, F., and B. H. Hahn. 1980. Restricted subpopulations of DNA antibodies erythematosus. Arthritis Rheum. 50: 3211–3220. in kidneys of mice with systemic lupus. Comparison of antibodies in serum and 37. Weening, J. J., V. D. D’Agati, M. M. Schwartz, S. V. Seshan, C. E. Alpers, renal eluates. Arthritis Rheum. 23: 392–403. G. B. Appel, J. E. Balow, J. A. Bruijn, T. Cook, F. Ferrario, et al. 2004. The 11. Madaio, M. P., J. Carlson, J. Cataldo, A. Ucci, P. Migliorini, and O. Pankewycz. classification of glomerulonephritis in systemic lupus erythematosus revisited. J. 1987. Murine monoclonal anti-DNA antibodies bind directly to glomerular Am. Soc. Nephrol. 15: 241–250.

antigens and form immune deposits. J. Immunol. 138: 2883–2889. 38. Hochberg, M. C. 1997. Updating the American College of Rheumatology re- http://www.jimmunol.org/ 12. Ehrenstein, M. R., D. R. Katz, M. H. Griffiths, L. Papadaki, T. H. Winkler, vised criteria for the classification of systemic lupus erythematosus. Arthritis J. R. Kalden, and D. A. Isenberg. 1995. Human IgG anti-DNA antibodies deposit Rheum. 40: 1725–1730. in kidneys and induce proteinuria in SCID mice. Kidney Int. 48: 705–711. 39. Austin, H. A., III, D. T. Boumpas, E. M. Vaughan, and J. E. Balow. 1994. 13. Kalaaji, M., K. A. Fenton, E. S. Mortensen, R. Olsen, G. Sturfelt, P. Alm, and Predicting renal outcomes in severe lupus nephritis: contributions of clinical and O. P. Rekvig. 2007. Glomerular apoptotic nucleosomes are central target histologic data. Kidney Int. 45: 544–550. structures for nephritogenic antibodies in human SLE nephritis. Kidney Int. 71: 40. Burgoon, M. P., K. M. Keays, G. P. Owens, A. M. Ritchie, P. R. Rai, C. D. Cool, 664–672. and D. H. Gilden. 2005. Laser-capture microdissection of plasma cells from 14. van Bavel, C. C., J. van der Vlag, and J. H. Berden. 2007. Glomerular binding of subacute sclerosing panencephalitis brain reveals intrathecal disease-relevant anti-dsDNA autoantibodies: the dispute resolved? Kidney Int. 71: 600–601. antibodies. Proc. Natl. Acad. Sci. USA 102: 7245–7250. 15. Esdaile, J. M., C. Levinton, W. Federgreen, J. P. Hayslett, and M. Kashgarian. 1989. 41. Meffre, E., E. Davis, C. Schiff, C. Cunningham-Rundles, L. B. Ivashkiv, The clinical and renal biopsy predictors of long-term outcome in lupus nephritis: L. M. Staudt, J. W. Young, and M. C. Nussenzweig. 2000. Circulating human a study of 87 patients and review of the literature. Q. J. Med. 72: 779–833. B cells that express surrogate light chains and edited receptors. Nat. Immunol. 1: by guest on September 24, 2021 16. Schwartz, M. M., J. S. Fennell, and E. J. Lewis. 1982. Pathologic changes in the 207–213. renal tubule in systemic lupus erythematosus. Hum. Pathol. 13: 534–547. 42. Wardemann, H., S. Yurasov, A. Schaefer, J. W. Young, E. Meffre, and 17. Yu, F., L. H. Wu, Y. Tan, L.-H. Li, C.-L. Wang, W. K. Wang, Z. Qu, M. H. Chen, M. C. Nussenzweig. 2003. Predominant autoantibody production by early human J. J. Gao, Z. Y. Li, et al. 2010. Tubulointerstitial lesions of patients with lupus B cell precursors. Science 301: 1374–1377. nephritis classified by the 2003 International Society of Nephrology and Renal 43. Yurasov, S., H. Wardemann, J. Hammersen, M. Tsuiji, E. Meffre, V. Pascual, and Pathology Society system. Kidney Int. 77: 820–829. M. C. Nussenzweig. 2005. Defective B cell tolerance checkpoints in systemic 18. Neumann, K., D. J. Wallace, C. Azen, S. Nessim, M. Fichman, A. L. Metzger, lupus erythematosus. J. Exp. Med. 201: 703–711. and J. R. Klinenberg. 1995. Lupus in the 1980s: III. Influence of clinical vari- 44. Tiller, T., M. Tsuiji, S. Yurasov, K. Velinzon, M. C. Nussenzweig, and ables, biopsy, and treatment on the outcome in 150 patients with lupus nephritis H. Wardemann. 2007. Autoreactivity in human IgG+ memory B cells. Immunity seen at a single center. Semin. Arthritis Rheum. 25: 47–55. 26: 205–213. 19. Austin, H. A., III, L. R. Muenz, K. M. Joyce, T. A. Antonovych, M. E. Kullick, 45. Lossos, I. S., R. Tibshirani, B. Narasimhan, and R. Levy. 2000. The inference of J. H. Klippel, J. L. Decker, and J. E. Balow. 1983. Prognostic factors in lupus antigen selection on Ig genes. J. Immunol. 165: 5122–5126. nephritis. Contribution of renal histologic data. Am. J. Med. 75: 382–391. 46. Cyster, J. G. 2005. Chemokines, sphingosine-1-phosphate, and cell migration in 20. Mori, Y., N. Kishimoto, H. Yamahara, Y. Kijima, A. Nose, Y. Uchiyama-Tanaka, secondary lymphoid organs. Annu. Rev. Immunol. 23: 127–159. M. Fukui, T. Kitamura, T. Tokoro, H. Masaki, et al. 2005. Predominant tubu- 47. Schneider, P. 2005. The role of APRIL and BAFF in lymphocyte activation. lointerstitial nephritis in a patient with systemic lupus nephritis. Clin. Exp. Curr. Opin. Immunol. 17: 282–289. Nephrol. 9: 79–84. 48. Vinuesa, C. G., I. Sanz, and M. C. Cook. 2009. Dysregulation of germinal 21. Singh, A. K., A. Ucci, and N. E. Madias. 1996. Predominant tubulointerstitial centres in autoimmune disease. Nat. Rev. Immunol. 9: 845–857. lupus nephritis. Am. J. Kidney Dis. 27: 273–278. 49. Kleinstein, S. H., Y. Louzoun, and M. J. Shlomchik. 2003. Estimating hyper- 22. Soderstrom, N., and A. Biorklund. 1974. Organization of the invading lymphoid mutation rates from clonal tree data. J. Immunol. 171: 4639–4649. tissue in human lymphoid thyroiditis. Scand. J. Immunol. 3: 295–301. 50. Bose, B., and S. Sinha. 2005. Problems in using statistical analysis of re- 23. Takemura, S., A. Braun, C. Crowson, P. J. Kurtin, R. H. Cofield, W. M. O’Fallon, placement and silent mutations in antibody genes for determining antigen-driven J. J. Goronzy, and C. M. Weyand. 2001. Lymphoid neogenesis in rheumatoid affinity selection. Immunology 116: 172–183. synovitis. J. Immunol. 167: 1072–1080. 51. MacLennan, I. C. 1994. Germinal centers. Annu. Rev. Immunol. 12: 117–139. 24. Barone, F., M. Bombardieri, A. Manzo, M. C. Blades, P. R. Morgan, 52. Allen, C. D., T. Okada, and J. G. Cyster. 2007. Germinal-center organization and S. J. Challacombe, G. Valesini, and C. Pitzalis. 2005. Association of CXCL13 cellular dynamics. Immunity 27: 190–202. and CCL21 expression with the progressive organization of lymphoid-like 53. Allen, C. D., T. Okada, H. L. Tang, and J. G. Cyster. 2007. Imaging of germinal structures in Sjo¨gren’s syndrome. Arthritis Rheum. 52: 1773–1784. center selection events during affinity maturation. Science 315: 528–531. 25. Serafini, B., B. Rosicarelli, R. Magliozzi, E. Stigliano, and F. Aloisi. 2004. 54. Liu, Y. J., J. Zhang, P. J. Lane, E. Y. Chan, and I. C. MacLennan. 1991. Sites of Detection of ectopic B-cell follicles with germinal centers in the meninges of specific B cell activation in primary and secondary responses to T cell-dependent patients with secondary progressive multiple sclerosis. Brain Pathol. 14: 164– and T cell-independent antigens. Eur. J. Immunol. 21: 2951–2962. 174. 55. Ku¨ppers, R., M. Zhao, M. L. Hansmann, and K. Rajewsky. 1993. Tracing B cell 26. Armengol, M. P., M. Juan, A. Lucas-Martıń, M. T. Fernańdez-Figueras, development in human germinal centres by molecular analysis of single cells D. Jaraquemada, T. Gallart, and R. Pujol-Borrell. 2001. Thyroid autoimmune picked from histological sections. EMBO J. 12: 4955–4967. disease: demonstration of thyroid antigen-specific B cells and recombination- 56. Hauser, A. E., T. Junt, T. R. Mempel, M. W. Sneddon, S. H. Kleinstein, activating gene expression in chemokine-containing active intrathyroidal ger- S. E. Henrickson, U. H. von Andrian, M. J. Shlomchik, and A. M. Haberman. minal centers. Am. J. Pathol. 159: 861–873. 2007. Definition of germinal-center B cell migration in vivo reveals predominant 27. Salomonsson, S., M. V. Jonsson, K. Skarstein, K. A. Brokstad, P. Hjelmstro¨m, intrazonal circulation patterns. Immunity 26: 655–667. M. Wahren-Herlenius, and R. Jonsson. 2003. Cellular basis of ectopic germinal 57. Schwickert, T. A., R. L. Lindquist, G. Shakhar, G. Livshits, D. Skokos, center formation and autoantibody production in the target organ of patients with M. H. Kosco-Vilbois, M. L. Dustin, and M. C. Nussenzweig. 2007. In vivo im- Sjo¨gren’s syndrome. Arthritis Rheum. 48: 3187–3201. aging of germinal centres reveals a dynamic open structure. Nature 446: 83–87. 1860 IN SITU IMMUNE RESPONSES IN LUPUS NEPHRITIS

58. Shlomchik, M. J. 2008. Sites and stages of autoreactive B cell activation and 73. Cassese, G., S. Lindenau, B. de Boer, S. Arce, A. Hauser, G. Riemekasten, regulation. Immunity 28: 18–28. C. Berek, F. Hiepe, V. Krenn, A. Radbruch, and R. A. Manz. 2001. Inflamed 59. Jacobson, B. A., T. L. Rothstein, and A. Marshak-Rothstein. 1997. Unique site of kidneys of NZB/W mice are a major site for the homeostasis of plasma cells. IgG2a and rheumatoid factor production in MRL/lpr mice. Immunol. Rev. 156: Eur. J. Immunol. 31: 2726–2732. 103–110. 74. Sekine, H., H. Watanabe, and G. S. Gilkeson. 2004. Enrichment of anti- 60. William, J., C. Euler, S. Christensen, and M. J. Shlomchik. 2002. Evolution of glomerular antigen antibody-producing cells in the kidneys of MRL/MpJ-Fas autoantibody responses via somatic hypermutation outside of germinal centers. (lpr) mice. J. Immunol. 172: 3913–3921. Science 297: 2066–2070. 75. Lee, J., Y. J. Cho, and P. E. Lipsky. 2004. The V(lambda)-J(lambda) repertoire of 61. Mandik-Nayak, L., S. Seo, A. Eaton-Bassiri, D. Allman, R. R. Hardy, and patients with systemic lupus erythematosus manifests characteristics of the J. Erikson. 2000. Functional consequences of the developmental arrest and follic- natural antibody repertoire. Arthritis Rheum. 50: 2604–2614. ular exclusion of anti-double-stranded DNA B cells. J. Immunol. 164: 1161–1168. 76. Schro¨der, A. E., A. Greiner, C. Seyfert, and C. Berek. 1996. Differentiation of 62. Herlands, R. A., S. R. Christensen, R. A. Sweet, U. Hershberg, and B cells in the nonlymphoid tissue of the synovial membrane of patients with M. J. Shlomchik. 2008. T cell-independent and toll-like receptor-dependent rheumatoid arthritis. Proc. Natl. Acad. Sci. USA 93: 221–225. antigen-driven activation of autoreactive B cells. Immunity 29: 249–260. 77. Goronzy, J. J., and C. M. Weyand. 2005. Rheumatoid arthritis. Immunol. Rev. 63. Radic, M. Z., and M. Weigert. 1994. Genetic and structural evidence for antigen 204: 55–73. selection of anti-DNA antibodies. Annu. Rev. Immunol. 12: 487–520. 78. Kim, H. J., V. Krenn, G. Steinhauser, and C. Berek. 1999. Plasma cell de- 64. Li, Q. Z., C. Xie, T. Wu, M. Mackay, C. Aranow, C. Putterman, and C. Mohan. velopment in synovial germinal centers in patients with rheumatoid and reactive 2005. Identification of autoantibody clusters that best predict lupus disease ac- arthritis. J. Immunol. 162: 3053–3062. tivity using glomerular proteome arrays [Published erratum appears in 2006 J. 79. Humby, F., M. Bombardieri, A. Manzo, S. Kelly, M. C. Blades, B. Kirkham, Clin. Invest. 116: 548.] J. Clin. Invest. 115: 3428–3439. J. Spencer, and C. Pitzalis. 2009. Ectopic lymphoid structures support ongoing 65. Clark, M. R., D. Massenburg, K. Siemasko, P. Hou, and M. Zhang. 2004. B-cell production of class-switched autoantibodies in rheumatoid synovium. PLoS antigen receptor signaling requirements for targeting antigen to the MHC class II Med. 6: e1. presentation pathway. Curr. Opin. Immunol. 16: 382–387. 80. Olee, T., E. W. Lu, D.-F. Huang, R. W. Soto-Gil, M. Deftos, F. Kozin, 66. Kemp, J. D., J. S. Cowdery, A. D. Steinberg, and R. K. Gershon. 1982. Genetic D. A. Carson, and P. P. Chen. 1992. Genetic analysis of self-associating im- control of autoimmune disease: interactions between xid and Ipr. J. Immunol. munoglobulin G rheumatoid factors from two rheumatoid synovia implicates an 128: 388–392. antigen-driven response. J. Exp. Med. 175: 831–842. Downloaded from 67. Shlomchik, M. J., M. P. Madaio, D. Ni, M. Trounstein, and D. Huszar. 1994. The 81. Lee, S. K., S. L. Bridges, Jr., W. J. Koopman, and H. W. Schroeder, Jr. 1992. The role of B cells in lpr/lpr-induced autoimmunity. J. Exp. Med. 180: 1295–1306. immunoglobulin kappa light chain repertoire expressed in the synovium of 68. Chan, O. T., M. P. Madaio, and M. J. Shlomchik. 1999. The central and multiple a patient with rheumatoid arthritis. Arthritis Rheum. 35: 905–913. roles of B cells in lupus pathogenesis. Immunol. Rev. 169: 107–121. 82. Zarkhin, V., N. Kambham, L. Li, S. Kwok, S. C. Hsieh, O. Salvatierra, and 69. Sfikakis, P.P.,J. N. Boletis, S. Lionaki, V.Vigklis, K. G. Fragiadaki, A. Iniotaki, and M. M. Sarwal. 2008. Characterization of intra-graft B cells during renal allograft H. M. Moutsopoulos. 2005. Remission of proliferative lupus nephritis following rejection. Kidney Int. 74: 664–673. B cell depletion therapy is preceded by down-regulation of the T cell costimulatory 83. Thaunat, O., N. Patey, G. Caligiuri, C. Gautreau, M. Mamani-Matsuda,

molecule CD40 ligand: an open-label trial. Arthritis Rheum. 52: 501–513. Y. Mekki, M. C. Dieu-Nosjean, G. Eberl, R. Ecochard, J. B. Michel, et al. 2010. http://www.jimmunol.org/ 70. van Vollenhoven, R. F., I. Gunnarsson, E. Welin-Henriksson, B. Sundelin, Chronic rejection triggers the development of an aggressive intragraft immune A. Osterborg, S. H. Jacobson, and L. Klareskog. 2004. Biopsy-verified response response through recapitulation of lymphoid organogenesis. J. Immunol. 185: of severe lupus nephritis to treatment with rituximab (anti-CD20 monoclonal 717–728. antibody) plus cyclophosphamide after biopsy-documented failure to respond to 84. Cohen, C. D., N. Calvaresi, S. Armelloni, H. Schmid, A. Henger, U. Ott, cyclophosphamide alone. Scand. J. Rheumatol. 33: 423–427. M. P. Rastaldi, and M. Kretzler. 2005. CD20-positive infiltrates in human 71. Adalid-Peralta, L., A. Mathian, T. Tran, L. Delbos, I. Durand-Gasselin, membranous glomerulonephritis. J. Nephrol. 18: 328–333. D. Berrebi, M. Peuchmaur, J. Couderc, D. Emilie, and S. Koutouzov. 2008. 85. Heller, F., M. T. Lindenmeyer, C. D. Cohen, U. Brandt, D. Draganovici, Leukocytes and the kidney contribute to interstitial inflammation in lupus ne- M. Fischereder, M. Kretzler, H. J. Anders, T. Sitter, I. Mosberger, et al. 2007. phritis. Kidney Int. 73: 172–180. The contribution of B cells to renal interstitial inflammation. Am. J. Pathol. 170: 72. Odegard, J. M., L. D. DiPlacido, L. Greenwald, M. Kashgarian, D. H. Kono, 457–468. C. Dong, R. A. Flavell, and J. Craft. 2009. ICOS controls effector function but 86. Henriksen, K. J., M. R. Clark, R. J. Quigg, and A. Chang. 2007. Accumulation of

not trafﬁcking receptor expression of kidney-inﬁltrating effector T cells in mu- interstitial B lymphocytes and plasma cells in pauci-immune ANCA-associated by guest on September 24, 2021 rine lupus. J. Immunol. 182: 4076–4084. crescentic glomerulonephritis. J. Am. Soc. Nephrol. 18: 219A (Abstr.). SUPPLEMENTAL FIGURE LEGENDS

Supplemental Figure S1. Control antibody stainings. Indicated species matched control primary antibodies and appropriate HRP-conjugated secondary antibodies were used to stain renal biopsies from six different patients with lupus nephritis.

Immunohistochemistry was performed as described in Materials and Methods.

Supplemental Figure S2. TBM immune complex deposition. (A) Segmental granular

IgG staining of the tubular basement membranes reveals the presence of immune deposits. (B) Numerous discrete electron dense deposits (arrows) are observed by electron microscopy from the same patient in A. (C) Granular C1q staining of the tubular basement membranes along with IgG (not shown) characteristic for TBM immune deposits of lupus nephritis. (D) Numerous discrete electron dense deposits (arrows) within the tubular basement membrane are seen by electron microscopy from the patient in C. (E) Granular IgG TBM staining indicates the presence of TBM immune complexes.

(F) Electron microscopy confirms the IF findings (of E) with many TBM electron dense deposits (arrows).

Supplemental Figure S3. Evidence for in situ clonal expansion and somatic hypermutation in the predominant light chain expressed outside the GC (VL2-14*01JL2*01).

(A) Nucleotide sequences of amplified transcripts aligned with germline. (B)

Genealogical relationships between sequences are illustrated in a clonal tree. (C) R/S ratios of clones in FRW and CDR regions with the corresponding p-values.

Supplemental Figure S4. In situ clonal expansion and somatic hypermutation in light chain cloned from a T:B aggregate (Patient B). (A) Nucleotides of cloned VL1-40*01JL3*02 aligned to predicted corresponding germline sequence. (B) Genealogical relationships between sequences are illustrated in a clonal tree. (C) R/S ratios of clones in FRW and

CDR regions with the corresponding p-values.

Supplemental Figure S5. Clonal expansion but not somatic hypermutation in the predominant expressed light chain (VK1-33*01JH3*01) in a T:B aggregate from patient C.

(A) Nucleotide sequences of amplified transcripts were aligned with germline. (B)

Genealogical relationships of sequences are illustrated in a clonal tree. (C) R/S ratios of clones in FRW and CDR regions with the corresponding p-values.

SUPPLEMENTAL TABLE 1. Summary of observed unique expressed immunoglobulin heavy and light chains sampled from CD38+ intrarenal T:B aggregates from four patients

(Patients B-E), diffuse infiltrates from four patients (Patients F-H) and diffuse infiltrates from one non-lupus patient with idiopathic interstitial nephritis (Patient I). Sampling of expressed immunoglobulin was also done from a Ki-67+ focus in a patient (Patient J) with T:B aggregates on histology.

V-Region D-Region J-Region Frequency V-Region D-Region J-Region Frequency Patient B Patient E Lambda Lambda VL1-40*01 JL3*02 3/27 VL1-44*01 JL1*01 3/17 VL1-44*01 JL3*01 2/27 VL2-11*01 JL1*01 2/17 VL2-14*01 JL1*01 2/27 VL2-11*01 JL2*01 2/17 VL2-14*01 JL1*01 2/27 VL2-11*01 JL2*01 2/17 VL2-14*01 JL2*01 4/27 Kappa Kappa VK1-6*01 JK1*01 3/26 VK1-9*01 JK4*01 2/31 VK3-11*01 JK4*01 2/26 VK1-33*01 JK3*01 2/31 VK3-20*01 JK2*01 2/26 VK1-39*01 JK4*01 2/31 Heavy VK1-39*01 JK5*01 2/31 VH3-23*01 D3-10*01 JH4*02 2/22 Heavy VH3-30*02 D5-18*01 JH3*02 2/22 VH1-2*02 D3-22*01 JH3*02 3/68 Patient F VH1-2*02 D3-22*01 JH5*02 2/68 Heavy VH1-3*01 D4-23*01 JH4*02 4/68 VH3-9*01 D5-24*01 JH5*02 2/37 VH3-9*01 D2-8*02 JH6*03 2/68 Patient G VH3-9*01 D4-17*01 JH3*01 2/68 Heavy VH3-15*01 D6-6*01 JH3*02 2/68 VH3-23*01 D3-10*01 JH4*02 2/37 VH3-23*04 D2-15*01 JH6*02 2/68 VH3-53*01 D3-10*01 JH4*02 3/37 VH3-30*04 D3-9*01 JH6*02 3/68 VH5-51*01 D6-19*01 JH4*02 3/37 VH3-30*04 D3-22*01 JH4*02 2/68 Patient H VH3-74*03 D3-22*01 JH6*03 2/68 Heavy Patient C VH1-46*01 D2-2*01 JH6*02 2/31 Lambda VH4-34*01 D3-22*01 JH3*02 3/31 VL1-44*01 JL1*01 2/16 VH4-39*01 D3-10*01 JH4*02 2/31 VL1-47*01 JL2*01 2/16 Patient I Kappa Heavy VK1-5*03 JK1*01 2/37 VH3-23*01 D1-26*01 JH4*02 2/34 VK1-33*01 JK4*01 12/37 VH4-34*01 D1-14*01 JH5*02 2/34 VK1-33*01 JK4*01 3/37 VH4-61*01 D2-21*02 JH4*02 2/34 VK1-39*01 JK1*01 2/37 VH5-51*03 D5-24*01 JH5*02 2/34 VK3-11*01 JK5*01 2/37 Patient J Patient D Heavy Lambda VH4-34*01 D1-1*01 JH4*02 5/18 VL1-44*01 JL2*01 2/24 VH4-34*01 D5-18*01 JH4*02 6/18 VL1-44*01 JL3*02 2/24 VL1-44*01 JL3*02 2/24 VL2-14*01 JL2*01 5/24 VL2-14*01 JL2*01 3/24 VL2-14*01 JL3*02 2/24 VL2-14*01 JL3*02 2/24 Kappa VK1-5*01 JK1*01 2/16 VK3-11*01 JK5*01 2/16 Heavy VH7-4-1*02 D1-26*01 JH6*02 2/24 VH7-4-1*02 D5-12*01 JH6*02 3/24