Activation of Glomerular Basement Membrane–Specific B Cells in the Renal Draining Lymph Node after T Cell–Mediated Glomerular Injury
Julie Robertson*,
Jean Wu*,
Jon Arends*,
Cindy Zhou*,
John McMahon,
Lisa Torres* and
Ya-Huan Lou*
* Department of Diagnostic Sciences, Dental Branch; and Department of Integrative Biology and Pharmacology, Medical School, University of Texas Health Science Center at Houston, Houston, Texas
Address correspondence to: Dr. Ya-Huan Lou, Department of Diagnostic Sciences, Dental Branch, University of Texas Health Science Center at Houston, 6516 M.D. Anderson Blvd, Houston, TX 77030. Phone: 713-500-4059; Fax: 713-500-4416; E-mail: yahuan.lou{at}uth.tmc.edu
Received for publication April 21, 2005.
Accepted for publication July 22, 2005.
Linear binding of IgG to the glomerular basement membrane (GBM)is the hallmark of anti-GBM glomerulonephritis (GN). However,the precise mechanism by which diverse autoantibodies to GBMare induced in GN has not been determined. It was demonstratedpreviously that a single T cell epitope pCol(28–40) derivedfrom collagen IV 3 chain not only induced severe GN in WistarKyoto rats but also triggered a diversified anti-GBM antibodyresponse through "B cell epitope spreading." In this study,an expansion of T and B cells in the renal draining lymph node(RDLN) of diseased animals after glomerular injury was observed.RDLN was demonstrated to be the location of GBM-specific B cellactivation. First, B cells from RDLN of pCol(28–40)-immunizedrats produced in vitro anti-GBM antibodies and antinuclear antibodies.Second, B cells specific to the peptidic B cell epitope in pCol(28–40)were absent among expanding B cells in RDLN. Those findingsprovided a unique opportunity to track activation of diverseGBM-specific B cells in RDLN. Expression of B lymphocyte-inducedmaturation protein-1, which is involved in differentiation ofplasma cells, in B cells of RDLN was detected and further elevatedonly after T cell–mediated prominent glomerular injury(day 19). This was supported by the fact that anti-GBM antibodiesbecame detectable only after day 20. Those results suggest thatT cell–mediated glomerular injury may trigger de novointernal immunization of autoantigens released from damagedGBM, which further leads to activation of a group of GBM-specificB cells in RDLN.
Production of autoantibodies to diverse autoantigens, whichare usually clustered or physically associated, is a characteristicof human autoimmune diseases (1). It is well exemplified inanti–glomerular basement membrane (GBM) glomerulonephritis(GN) or Goodpasture’s syndrome. Autoantibodies from patientswith GN react with conformational B cell epitopes of diverseGBM proteins, which include the NC domain of collagen IV 3 chain(Col43NC1), different chains of type IV collagen, collagen domainsand S7 domains of type IV collagen, and other noncollagen components(2–5). In addition to their pathologic roles in tissuedamage, the presence of anti-GBM autoantibody is one of themost important diagnostic indices for this disease. However,the precise mechanism by which autoantibodies to GBM are inducedin GN has not been determined.
The mechanisms of the B cell tolerance and production of autoantibodieshave been studied at several levels. First, several signal pathwayshave been revealed to participate in B cell selection/tolerance,especially in T-independent responses (6–8). However,those mechanisms cannot explain diversified autoantibody responsesto clustered target antigens as seen in anti-GBM disease. Second,BcR transgenic mice have been a valuable tool in the study ofB cell tolerance to a special autoantigen, because "autoreactive"B cells can be tracked easily (9,10). Those genetically manipulatedanimals have been instrumental in our understanding of B celltolerance and activation. Obviously, mono-specific B cells intransgenic models may not be suitable for study on diversifiedautoantibody responses as seen in anti-GBM disease.
Many studies suggest that B cell tolerance in many cases isT cell dependent. It has been reported in several animal modelsthat a pathogenic T cell epitope triggers production of diverseautoantibodies, or "B cell epitope spreading" (11–14).B cell epitope spreading also has been observed in our rat modelfor T cell–mediated anti-GBM GN (15). A single pathogenicT cell epitope pCol(28–40) derived from collagen 43 chain(Col43) not only induces severe GN but also elicits an autoantibodyresponse to diverse GBM antigens (16–18). It has beenpostulated that the T cells, activated by a pathogenic T cellepitope, may provide help to a population of autoreactive Bcells, which present the same T cell epitope of endogenous autoantigensduring B cell epitope spreading (19,20). However, it has beendifficult to verify this hypothesis in normal animals becauseof lack of a model in which diverse autoreactive B cell populationscan be tracked for analysis.
In this study, we identified the renal draining lymph node (RDLN)as the location for activation of autoreactive B cells afterT cell–mediated glomerular damage in our anti-GBM GN model.Thus, diverse autoreactive B cell populations can be trackedand their activation can be studied at cellular and molecularlevels. Our model provides a unique chance to investigate howthe GBM-specific B cells are activated.
Antigen Preparation and Induction of GN
Peptides were synthesized on an automatic peptide synthesizer,AMS 422 (Gilson, Middleton, WI) using Fmoc chemistry and werepurified by a reverse-phase C18 column on preparative HPLC (Water,Millford, MA). Purified peptides were analyzed by HPLC for purityand by mass spectrometry for the correct sequence. Peptidesthat exceeded 95% purity were dissolved in milli-Q water at1 mM concentration and used for immunization or other investigativepurposes. Sequences of synthesized peptides are listed in Table 1.
Table 1. Peptides pCol(28–40), derivatives used in this study and their immune properties and nephritogenicity
Female Wistar Kyoto (WKY) rats were allowed to acclimate fora minimum of 3 d and were immunized in one hind footpad andat the base of the tail with a peptide (0.125 µmol) emulsifiedin complete Freund adjuvant (CFA). Rats that were immunizedwith CFA alone served as controls. GN was evaluated by measurementof proteinuria/albuminuria and examination of renal histopathology(16). Renal tissues fixed in Bouin’s solution were usedfor hematoxylin and eosin staining (16). Glomerular injury scorewas calculated as the following: [(number of crescentic glomerulix 100) + (number of hypercellular glomeruli x 50)] ÷total number of glomeruli.
Investigation of Renal Lymphatic Drainage
Rats were anesthetized with ketamine/xylazine (50/7 mg/kg).A small dorsal incision was made, and the left kidney was exposedthrough the incision. India ink (tissue culture grade; Sigma,St. Louis, MO) was injected into the external border of therenal cortex using a 20-gauge needle (three sites, 10 µl/site).The animals were allowed to recover and were killed 3 to 12h later. The RDLN, which was stained by the ink, was identified,and the presence of carbon particles in its cortical area wasconfirmed under a dissecting microscope.
Isolation of Kidney-Infiltrating Leukocytes
Renal cortex was sliced, placed in a cold DMEM that contained10% FCS, ground against #100 mesh, and rinsed repeatedly. Thepassing-through materials included intact glomeruli, fragmentedtubules, and single cells. The single cells were separated bya cell constrainer (80 µm) and designated as "interstitialcells." The glomeruli, which were collected from the cell constrainer,were then separated from the tubules by recombination of gravitysediment and low-speed centrifugation (200 rpm for 4 min) (15).The purity of the glomeruli, which was determined under a microscope,was approximately 80 to 85%. The purified glomeruli were incubatedfurther with collagenase IV at 37°C for 30 min with periodicstirring. After the incubation, glomerular fragments were eliminatedby low-speed centrifugation (300 rpm for 5 min), and the singlecells were collected by a higher speed centrifugation (1000rpm for 10 min). Those cells were designated as "glomerularcells." Both interstitial cells and glomerular cells were stainedfor several leukocyte markers, including CD4, CD8, CD11b/c,OX33, and IgG/M, and analyzed with a flow cytometer (FACSCalibur;Becton Dickinson, San Jose, CA).
Flow Cytometry and Immunofluorescence Detection of pCol(28–40)-Specific B Cells
FITC-pCol(28–40) was commercially synthesized throughAhx at the amino terminus of the peptide (Biosynthesis, Lewisville,TX), dissolved in PBS at a concentration of 1 mM (1.8 mg/ml),and kept in the dark. ELISA confirmed that anti-pCol(28–40)antibody bound to FITC-pCol(28–40).
The lymphocytes were adjusted to 5 x 106 cells/ml and fixedin 0.5% paraformaldehyde (PFA) for 5 min. The cells (0.5 x 106cells) were stained with FITC-pCol(28–40) (10 µM)and PE-labeled mAb OX33 (10 µg/ml) on ice for 30 min.After washing with PBS, the cells were analyzed by a flow cytometer.B cells that were specific to pCol(28–40) were identifiedas double positive for both pCol(28–40) and OX33. As plasmacells do not express OX33, they were identified as pCol(28–40)single positive cells. FITC-pCol(28–40) is able to bindto MHC class II (RT.1Bl) on several cell populations such asmacrophages at a lower intensity (18). mAb OX6, which reactswith RT.1Bl, was used to reduce the background.
For immunofluorescence, frozen sections of lymph nodes werebriefly fixed in cold acetone for 10 min and washed in coldPBS. The sections were preincubated with OX6 (10 µg/ml)to prevent possible binding of pCol(28–40) to MHC classII, followed by incubation with FITC-pCol(28–40) and PE-OX33at room temperature. The sections were sealed with a mountingsolution and viewed with a confocal microscope.
In Vitro Antibody Production
Lymphocytes were isolated from either RDLN or inguinal lymphnodes (ILN). The cells were incubated in a serum-free mediumat a cell density of 5 x 106 cell/ml following our previouslypublished method (17). The supernatant of the culture was collected,and IgG concentration was determined by ELISA with a pair ofanti-rat IgG antibodies using normal rat IgG as a standard (SouthernBiotechnology, Birmingham, AL). The supernatant was dialyzedagainst 20 mM ammonium carbonate and lyophilized. The driedpowder was re-dissolved in sterilized PBS (50 mM, pH 7.2) with1 mg/ml BSA as a stabilizer, and the IgG concentration was re-measured.The IgG was kept at 4°C with 0.1% sodium azide for subsequentapplications.
Detection of Antibody Activity
A previously described ELISA method was applied to detect antibodiesto peptides (15). Briefly, plates were coated with 50 µlof peptide of 10 µM in a carbonate buffer (pH 9.5). Seriallydiluted serum (100 to 800) was added to each well in duplicate.The bound rat IgG antibodies were detected with horseradishperoxidase–labeled goat anti-rat IgG (1:10,000; SouthernBiotechnology) using O-phenoldiamine (0.25 mg/ml) as the substrate.The plates were read on an ELISA reader at 490 nm. A similarELISA was used for detection of anti-GBM antibody using digestedGBM proteins (15).
Direct immunofluorescence study was carried out to detect theanti-GBM antibody that bound to GBM in vivo (15), using FITC-labeledgoat anti-rat IgG or IgM antibodies (1:50; Southern Biotechnology).Indirect immunofluorescence was used to detect antibodies tonative GBM. Rat kidney contains a significant amount of Ig,which will also be stained by anti-rat IgG. Such backgroundis not acceptable in the case for detection of a small quantityof rat antibody to GBM. To minimize the background, we usednormal human kidneys, as we have demonstrated that anti-GBMantibodies in our model also reacted with human GBM (16). Briefly,frozen sections of normal human kidney were incubated with thetesting antibody for 2 h at 20°C and washed extensively.The section was incubated further with FITC-labeled goat anti-ratIgG or IgM antibodies (1:50; Southern Biotechnology) and viewedwith a confocal microscope in a double-blind manner (FV 500;Olympus, Melville, NY). Confocal images were digitally recorded.
Purification of B Cell Population
Lymphocytes were isolated from different lymph nodes and adjustedto 107 cells/ml. B cells were purified with positive magneticbeads sorting following the manufacturer’s instructions(MACS; Miltenyi Biotech Inc., Auburn, CA). Briefly, after preincubationwith normal mouse IgG, the lymphocytes were incubated on icewith magnetic beads coated with mAb OX33 for 30 min and thenadded to the isolation column located in a magnetic field. Thecolumn was washed thoroughly with cold PBS, and B cells wereeluted from the reverse direction in the absence of a magneticfield. The cell fractions were analyzed with flow cytometryafter staining with PE-labeled OX33 and FITC-labeled anti-ratIgG/M. The purified B cells were used immediately for isolationof total RNA.
Reverse Transcription–PCR Detection of B Lymphocyte-Induced Maturation Protein Expression
Total RNA was isolated from lymph nodes or cell fractions atdesignated time points (Ambion, Austin, TX), and cDNA was synthesizedusing 0.1 µg of total RNA through a reverse transcription(RT) reaction (RNA PCR Core Kit; Applied Biosystems, FosterCity, CA). PCR was used to detect B lymphocyte-induced maturationprotein (blimp) mRNA using a pair of primers (5'-CCTTGCTACATGTGTGTCAGTTGG-3',5'-TGCTAGCATGTGTGGAATCTCTGG-3'), which resulted in a 400-bpproduct, under the following conditions: Preheating at 94°Cfor 3 min followed by 30 cycles (94°C for 60 s, 64°Cfor 30 s, 72°C for 60 s; GeneAmp9700; Applied Biosystems)(21). Expression of a housekeeper gene (glyceraldehyde-3-phosphatedehydrogenase) was used as an internal control following a previouspublished RT-PCR method, which resulted in a 238-bp product(22). The products were separated by electrophoresis in 1.5%agarose gel, stained with ethidium bromide, and visualized underultraviolet light illumination. Selected samples were also usedfor real-time PCR following the identical conditions as describedabove. Briefly, after the RT reaction, cDNA concentrations weredetermined by a Microplate Spectrophotometer (SPECTRAmax; MolecularDevices, Sunnyvale, CA). Real-time PCR was carried out at twoconcentrations of the cDNA (1.5 ng/50 µl and 0.75 ng/50µl) in duplicate in an iCycler iQ real-Time PCR detectionsystem using iQ SYBR Green Supermix (Bio-Rad, Hercules, CA).Housekeeper gene glyceraldehyde-3-phosphate dehydrogenase wasused as the standard. Relative abundance of blimp-1 expressionwas calculated following the 2–CT method (23).
Expansion of T and B Cells in RDLN of Rats with Anti-GBM GN
As the first step to investigate how anti-GBM antibodies wereproduced, we searched for the location of anti-GBM antibodyproduction in the immunized WKY rats. Several previous studieshave demonstrated the presence of B cells/extra-lymph node germinalcenter in the kidneys of several types of GN (24,25). We firstexamined the kidney-infiltrating leukocytes in pCol(28–40)-immunizedWKY rats. The infiltrating leukocytes were isolated from eitherthe interstitial tissue or glomeruli of diseased kidneys attwo different stages (days 25 and 40 after immunization) andanalyzed by flow cytometry (Figure 1). Although robust increasesin several leukocyte populations, especially CD4+ and CD11b/c+populations, were observed (data not shown), no significantnumber of OX33+/IgG/M+ cells (B cells) among the infiltratingleukocytes was detected at either stage (Figure 1B). Thus, thekidney was an unlikely location for anti-GBM antibody production.
Figure 1. Flow cytometry demonstrates absence of OX33+ IgG/M+ B cells among the leukocytes infiltrating the diseased kidney of a Wistar Kyoto (WKY) rat that was immunized with pCol(28–40) at day 35. (A) Lymphocytes from the inguinal lymph node (ILN) of the same rat as the positive control show a OX33+IgG/M+ B cell population. (B) Leukocyte populations infiltrating the kidney; the cells are gated on forward scatter/side scatter for analysis; notice absence of OX33+IgG/M+ B cells.
We observed that a special lymph node in close proximity tothe renal hilum, above the renal vein, of each diseased kidneywas consistently enlarged in the immunized WKY rats after 25to 35 d after immunization (Figure 2C). However, enlargementof the lymph nodes at the same location was not observed inCFA-immunized WKY controls (Figure 2B). In rats that were immunizedwith pCol34A, which has altered specificity of T cell epitopeand nonpathogenic (see Table 1), enlargement of the lymph nodewas also not observed. Thus, enlargement of the lymph node wasassociated only with glomerular injury as well as anti-GBM antibody.
Figure 2. Enlargement of renal draining lymph node (RDLN) in WKY rats that were immunized with pCol(28–40). (A) India ink injection revealing a stained RDLN. (B) A normal RDLN in a complete Freund adjuvant (CFA)-immunized rat at day 35. (C) An enlarged RDLN in a pCol(28–40)-immunized rat at day 35. Arrows indicate RDLN. Ad, adrenal gland.
Using India ink as a tracer, lymphatic drainage of the renalcortex was investigated. It took 3 h for the ink to stain thefirst lymph node, which was in the identical location of theenlarged lymph node (Figure 2A). Microscopic observation showedaccumulation of carbon particles in the cortex of the lymphnode. Thus, the enlarged lymph node described above was thedraining lymph node to the renal cortex. This lymph node wasdesignated as RDLN. Quantitative analyses on the lymphocytenumbers from RDLN ranged from 31.5 up to 55 x 106 lymphocytes/nodein pCol(28–40)-immunized WKY rats sampled at 35 to 40d, as compared the range of 0.9 to 1.3 x 106/node in CFA controlsat the same time. Flow cytometry further showed a significantexpansion of OX33+ B cells and CD4+ cells in the enlarged RDLN(data not shown). Those results suggested that the enlargedRDLN probably was the location of anti-GBM antibody production.
B Cells in RDLN Produce Anti-GBM
We next asked whether B cells in enlarged RDLN of diseased kidneyswere specific to GBM antigens. Lymphocytes were isolated fromRDLN of WKY rats 35 d after immunization with pCol(28–40)for in vitro antibody production (Table 2). As an internal control,the cells that were isolated from ILN, to which the injectedpCol(28–40) drained, were used. Lymphocytes that wereisolated from RDLN or ILN of the rats that were immunized withpCol34A were used as additional controls. The lymphocytes fromILN of rats that were immunized with pCol(28–40) produceda significant amount of IgG (Table 2). ELISA demonstrated thatthe produced IgG showed strong antibody activity to pCol(28–40)(Table 2). However, the IgG that was produced by ILN cells,even when concentrated up to 50 µg/ml, failed to reactwith GBM by either ELISA or immunofluorescence (Figure 3A).
Figure 3. Immunofluorescence detection of antibodies to native glomerular basement membrane (GBM) that were produced by lymphocytes from ILN (A) or RDLN (B) of a pCol(28–40)-immunized WKY rat, sampled at day 35. Notice that antibodies from RDLN (B) react not only with GBM but also nuclear antigens (arrows).
Conversely, the lymphocytes that were isolated from RDLN ofonly pCol(28–40)-immunized WKY rats produced a significantquantity of IgG (Table 2). The IgG that was produced by RDLNof pCol(28–40)-immunized WKY failed to react with pCol(28–40)as measured by ELISA (Table 2). However, indirect immunofluorescencedemonstrated that the IgG that was produced by RDLN stronglyreacted with native GBM at a concentration as low as 1 µgIgG/ml (Figure 3B). The IgG that was produced by RDLN also showedan unanticipated reactivity with nuclear antigens, suggestingthe production of anti-nuclear antibodies in RDLN as well (Figure 3B).In contrast, only a trace amount of IgG was detected inRDLN from pCol34A-immunized WKY rats (Table 2). We concludedthat RDLN of pCol(28–40)-immunized WKY rats was the locationof autoreactive B cells that were specific to GBM antigens.IgM antibody to GBM was undetectable by indirect immunofluorescenceor ELISA. Although anti-nuclear antibody was produced by RDLNB cells, circulating anti-nuclear antibody was not detectedby indirect immunofluorescence.
Absence of pCol(28–40)-Specific B Cells in RDLN
To exclude any possible connections between GBM- and pCol(28–40)-specificB cells, we next analyzed the specificity of B cells in boththe RDLN and ILN of pCol(28–40)-immunized WKY rats atday 20, when anti-GBM antibodies became detectable by immunofluorescence.FITC-pCol(28–40) was used as a probe to identify pCol(28–40)-specificB cells (Table 1). FITC-pCol(28–40) bound to a population(approximately 12.2%) of cells that were isolated from ILN ofpCol(28–40)-immunized WKY rats and accounted for approximately33.8% of total OX33+ B cells (Figure 4A). We also observed apopulation of OX33–pCol(28–40)+, which probablywas anti-GBM plasma cells, which lost CD45R expression (26).Binding of FITC-pCol(28–40) to the B cells was specific,as nonlabeled pCol(28–40) could inhibit the binding. Incontrast, only 0.1% of cells from RDLN of the same animals werestained as "positive" at day 20, and thus only <0.5% of Bcells might be pCol(28–40)+ cells (Figure 4B). This percentagewas close to the background, which was set by lymphocytes fromILN of control rats that were immunized with CFA alone or fromnormal WKY. B cells that were specific to pCol(28–40)were present in RDLN at a much later stage (day 35; 6.8%). Aslightly higher percentage of pCol(28–40)-specific B cells(7.5%) were also observed in other lymph nodes, such as axillarynodes, which were more distant from ILN. Those results suggestedthat presence of pCol(28–40)-specific B cells in RDLNat a later stage probably was due to secondary migration ofantigen-specific B cells throughout the lymphatic system. Weconcluded that pCol(28–40)-specific B cells were not presentin RDLN.
Figure 4. Absence of pCol(28–40)-specific B cells in RDLN of a pCol(28–40)-immunized WKY rat at day 20. (A) Flow cytometry on lymphocytes of ILN showing OX33+pCol(28–40)– and OX33+pCol(28–40)+ B cell populations. (B) Histogram based on gated OX33+ cells as shown in A; blue line, ILN; green line, RDLN; percentages of pCol(28–40)+ cells among OX33+ cells are indicated. (C and D) Confocal micrographs of immunofluorescence for detecting pCol(28–40)-specific B cells in the germinal center, stained with OX33-PE and FITC-pCol(28–40); notice that FITC-pCol(28–40) staining is absent in RDLN (C) but present in ILN; one germinal center is shown for each.
Absence of pCol(28–40)-specific B cells in the RDLN wasalso confirmed by two-color immunofluorescence (Figure 4, Cand D). RDLN and ILN were sampled from the same rats at day20 after immunization. Both lymph nodes displayed well-developedgerminal centers (GC) with organized OX33+ B cells. Bindingof FITC-pCol(28–40) to the B cells in many GC was observedin ILN (Figure 4D). In contrast, FITC-pCol(28–40) failedto bind to B cells in any GC in RDLN (Figure 4C). Those resultssupported our earlier observation that RDLN of pCol(28–40)-immunizedrats lacked pCol(28–40)-specific B cells.
Timing for Activation of B Cells in RDLN
To estimate when the B cells, specific to either pCol(28–40)or native GBM in ILN or RDLN, were activated, we establisheda time course of antibody response, as well as of glomerularinjury, in pCol(28–40)-immunized WKY rats (Figure 5).Prominent glomerular injury was histologically observed at day20 after immunization and rapidly developed into crescenticlesions/tuft necrosis in 100% of glomeruli after day 30 (Figure 5A).However, our previous study showed that significant proteinuriadeveloped in the immunized rats at approximately 15 to 18 d,suggesting the occurrence of subhistologic damage in the GBM(16).
Figure 5. Time courses of renal pathology and antibody response in WKY rats that were immunized with pCol(28–40). (A) Renal pathology expressed as a glomerular injury score; each dot represents one individual; time for beginning of proteinuria is indicated by an arrow. (B) Circulating antibodies to pCol(28–40) () and GBM (•) as measured by ELISA; antibody to native GBM, eluted from diseased glomeruli, was used as a positive control (); 3 samples for each time point. (C) IgG GBM antibodies, expressed as frequency of the intensity (0 to 4) of direct immunofluorescence on GBM, at different time points; samples (>10) in each group were taken from the indicated time ± 2 d.
IgG-type circulating antibodies to the peptide pCol(28–40)became detectable as early as day 8, and its titer elevatedrapidly after day 12. The circulating antibodies reached a plateauat day 20 and persisted at this level during the entire experimentalperiod (Figure 5B). However, direct immunofluorescence showeda positive GBM-bound IgG antibody in 37% of the rats at day20. At day 25, most (75%) of rats showed positive in the bindingof antibody to GBM. After day 30, the binding of antibody toGBM became extremely intensive in most of the rats (Figure 5C).IgG circulating antibodies to GBM, however, were not detectableby either ELISA or indirect immunofluorescence (Figure 5B).On the basis of the comparisons between antibodies and glomerulardamage, we concluded that production of anti-pCol(28–40)antibodies significantly preceded the glomerular injury, butproduction of anti-GBM antibodies occurred after the glomerularinjury. Thus, there was a significant time gap between productionof anti-pCol(28–40) antibodies in ILN and anti-GBM antibodiesin RDLN. Those data also suggested that that T cell–mediatedglomerular damage was a prerequisite for production of anti-GBMantibodies in the immunized rats.
For confirming the time gap between B cell activation in thetwo lymph nodes, expression of blimp-1, which is associatedwith B cell activation and plasma cell differentiation (21),in RDLN and ILN was investigated by RT-PCR. Expression of blimp-1became detectable in the ILN as early as day 5 and elevatedto a high level from day 10 to day 30. Its expression then declinedto a medium level at day 35 (Figure 6A). In contrast, blimp-1expression in RDLN was only slightly above the background level(lymphocytes from nonimmunized rats) at day 10 and detectableat a medium level at day 19. Its expression in RDLN reacheda high level especially at day 35 (Figure 6A). Real-time quantitativePCR revealed a >40-fold increase in blimp-1 expression levelsin RDLN between days 19 and 35. Using magnetic bead sorting,B cells were purified from RDLN lymphocytes (Figure 6B). TheOX33+ B cell fraction showed a high level of blimp expression,whereas OX33– cells showed no expression of blimp (Figure 6C).Thus, detected blimp expression originated from the B cells.This experiment again demonstrated that B cell activation inRDLN occurred at a much later stage than that in ILN.
Figure 6. Reverse transcription–PCR detection of expression of B lymphocyte-induced maturation protein (blimp-1) in RDLN (RD) or ILN (IL) of pCol(28–40)-immunized WKY or CFA control (A). Expression of housekeeper gene glyceraldehyde-3-phosphate dehydrogenase is also shown. (B) Flow cytometry of OX33– and OX33+ cells after magnetic bead sorting. (C) Blimp-1 expression in sorted OX33+ or OX33– populations.
Presence of anti-GBM antibody is the hallmark of Goodpasture’ssyndrome, or anti-GBM GN. The precise mechanism by which autoantibodiesto diverse GBM antigens are produced remains unclear. We haveestablished a rat model for anti-GBM GN in which the nephritogenicT cell epitope pCol(28–40), derived from Goodpasture’santigen collagen IV 3 chain, not only induces glomerular injurybut also can elicit anti-GBM autoantibodies (15). Our studiesstrongly argue that B cell epitope spreading probably occursduring glomerular damage, which leads to GBM antibody production.In this study, we further identified RDLN as the location ofthe autoreactive B cell activation. Thus, diverse GBM-specificB cell populations can be tracked and studied. This unique featurewill greatly facilitate our investigations on the precise mechanismof GBM antibody production. We believe that our investigationwill also contribute to our understanding of the mechanism ofGBM antibody production as well as B cell epitope spreading.
Activation of GBM-Specific Autoreactive B Cells in RDLN
The most significant finding in this study is identificationof RDLN as the specific location where the autoreactive B cellsto GBM and nuclear antigens are activated. The most direct evidencecame from in vitro antibody production experiments. IgG thatwas produced only by RDLN B cells showed strong antibody activityto native GBM.
Indirect evidence also supported our conclusion. T cell epitopepCol(28–40) also contains a peptide-specific B cell epitope(16) (Table 1). Because antibodies to this B cell epitope donot react with native GBM, this peptide-specific B cell epitopehas served as an excellent internal control, which has allowedus to observe simultaneously activation of two B cell populationswith different specificities. Using synthetic FITC-pCol(28–40),we were able to detect or locate quantitatively pCol(28–40)-specificB cell populations. We first demonstrated, at either day 20or day 35, the presence of a large percentage of pCol(28–40)-specificB cells in the ILN, to which immunized pCol(28–40) drained.In contrast, pCol(28–40)-specific B cells were not presentin RDLN at an earlier stage (day 20), when anti-GBM antibodieshad become detectable. Therefore, the expanding B cell populationsin RDLN and ILN were different in their specificity: Those inILN were pCol(28–40) specific, whereas those in RDLN werenot. Although we still lack a method to detect directly GBM-specificB cells, we are able to conclude that expanding B cells in RDLNwere GBM or nuclear antigen specific. A small percentage ofpCol(28–40)-specific B cells was observed in RDLN at alater stage (day 35). Because a similar percentage of peptide-specificB cells were also seen in distant lymph nodes such as the axillarynode, it is reasonable to conclude that this B cell populationis not specifically expanded in RDLN but probably is a resultof the secondary migration of B cells throughout the lymphaticsystem.
Our time-course study showed that production of anti-GBM antibodiesoccurred after T cell–mediated glomerular injury. In contrast,production of antibodies to pCol(28–40) preceded glomerularinjury. By tracking B cells in RDLN, we were able to determinemore precisely when the GBM-specific B cells were activated.Blimp-1 is a transcription factor that is involved in differentiationof plasma cells (21). Delayed expression of blimp in B cellsfrom RDLN also indicates that GBM-specific B cells were activatedafter glomerular injury. Those results suggest that glomerularinjury may be a prerequisite for activation of GBM-specificB cells.
Our Model Is a Unique Tool for Studying B Cell Epitope Spreading
Production of autoantibodies is characteristic of autoimmunediseases mediated by either an antibody or a T cell mechanism.It has been a hot topic in immunology to understand mechanismsby which autoreactive B cells are activated. Sophisticated BcRtransgenic models have been instructive in our understandingof the mechanism of B cell tolerance, especially in T independentresponses. However, mono-specific B cell populations limit theirvalue for studying the mechanism of a diversified autoantibodyresponse, which is more frequently seen among human autoimmunediseases (1). How to track diverse B cell populations is a keyto determining how autoreactive B cells are activated especiallyin T cell–mediated autoimmune diseases.
Our model provides an excellent tool to address this issue.First, our model for anti-GBM disease is a typical autoimmunedisease model with production of anti-GBM antibody. Second,identification of the location of autoreactive B cell activation,i.e., RDLN, will allow us to study the activation of autoreactiveB cells after T cell–mediated tissue damage at cellularand molecular levels without genetic manipulation. Third, aunique peptide-specific B cell epitope in pCol(28–40)is an excellent internal control. It allows us to compare theB cell response to an immunizing "foreign" B cell epitope versusautoantigens released from T cell–mediated tissue injury.Fourth, we recently found that Lewis rats possess identicalMHC to WKY. Although Lewis rats are able to mount an identicalT cell response to pCol(28–40), they do not develop anti-GBMdisease or anti-GBM antibodies. Thus, Lewis rats will providean essential control for our investigation on the factors beyondthe T cell mechanism that determines B cell epitope spreading.Finally, many models for glomerular injury are mediated by mechanismsother than autoimmunity (27–29). Those models will helpus to answer why only autoreactive T cell–mediated glomerularinjury leads to activation of autoreactive B cells.
Hypothesis: How Antibodies to Diverse GBM Antigens Are Produced
It has been well demonstrated that GBM patients usually produceautoantibodies to multiple GBM epitopes or proteins. On thebasis of our findings in both this and previous articles, weare able to propose a general hypothetical mechanism by whichdiverse GBM antibodies are produced in our model (Figure 7)(30). First, a single pathogenic T cell epitope activates Tcells, which cause glomerular inflammation and GBM damage. Second,damage of GBM not only leads to proteinuria but also resultsin release of GBM antigens, which drain to RDLN. Finally, Bcells specific to GBM antigens are activated in RDLN and beginto produce diverse GBM antibodies, which eventually bind tonative GBM. Thus, a single T cell epitope triggers a full clinicalspectrum of anti-GBM glomerulonephritis: Proteinuria, glomerulardamage, and diverse GBM antibodies.
Figure 7. Schematic diagram describes the hypothetic mechanism by which antibodies to diverse GBM antigens are produced after immunization with a single nephritogenic T cell epitope in our model. Bold arrows indicate the steps that are based on the findings described in this article.
Acknowledgments
This study was supported by the National Institutes of Health(NIH) grant RO1 DK60029 (Y.H.L). J.R. is a graduate studentsupported by NIH T32DE015355. L.T. is supported by NIH DK60029-03S1.C.Z. is a postdoctoral trainee supported by NIH T32 HD07324.Histology support was provided by the Histology Core Laboratory,Dental Branch, the University of Texas Health Science Centerat Houston. Confocal microscopic images were provided by theCore Facilities at M.D. Anderson Cancer Center, University ofTexas.
Footnotes
Published online ahead of print. Publication date availableat www.jasn.org.
Boitard C: The spectrum of autoantibody in organ-specific autoimmune diseases. In: B Cells and Autoantibody Production in Autoimmune Diseases. New York, Chapman & Hall, 1996
, pp 7
–80
Wieslander J, Bygren P, Heinegard D: Isolation of the specific glomerular basement membrane antigen involved in Goodpasture syndrome.
Proc Natl Acad Sci U S A 81
: 1544
–1548, 1984[Abstract/Free Full Text]
Hellmark T, Johansson C, Wieslander J: Characterization of anti-GBM antibodies involved in Goodpasture’s syndrome.
Kidney Int 46
: 823
–829, 1994[Medline]
Kefalides NA, Ohno N, Wilson CB: Heterogeneity of antibodies in Goodpasture syndrome reacting with type IV collagen.
Kidney Int 43
: 85
–93, 1993[Medline]
Savige JA, Gallicchio M, Georgiou T, Davies DJ: Diverse target antigens recognized by circulating antibodies in anti-neutrophil cytoplasm antibody-associated renal vasculitis.
Clin Exp Immunol 82
: 238
–243, 1990[Medline]
Fagarasan S, Honjo T: T-Independent immune response: New aspects of B cell biology.
Science 290
: 89
–92, 2000[Abstract/Free Full Text]
Danzer CP, Collins BE, Blixt O, Paulson JC, Nitschke L: Transitional and marginal zone B cells have a high proportion of unmasked CD22: Implications for BCR signaling.
Int Immunol 15
: 1137
–1147, 2003[Abstract/Free Full Text]
Mecklenbrauker I, Saijo K, Zheng NY, Leitges M, Tarakhovsky A: Protein kinase Cdelta controls self-antigen-induced B-cell tolerance.
Nature 416
: 860
–865, 2002[CrossRef][Medline]
Goodnow CC: Transgenic mice and analysis of B-cell tolerance.
Annu Rev Immunol 10
: 489
–518, 1992[CrossRef][Medline]
Chen C, Nagy Z, Radic MZ, Hardy RR, Huszar D, Camper SA, Weigert M: The site and stage of anti-DNA B-cell deletion.
Nature 373
: 252
–255, 1995[CrossRef][Medline]
Lou YH, Tung KS: T cell peptide of a self-protein elicits autoantibody to the protein antigen. Implications for specificity and pathogenetic role of antibody in autoimmunity.
J Immunol 151
: 5790
–5799, 1993[Abstract]
Milich DR, McLachlan A, Raney AK, Houghten R, Thornton GB, Maruyama T, Hughes JL, Jones JE: Autoantibody production in hepatitis B e antigen transgenic mice elicited with self T-cell epitope and inhibited with nonself peptide.
Proc Natl Acad Sci U S A 88
: 4348
–4352, 1994
James JA, Gross T, Scofield RH, Harley JB: Immunoglobulin epitope spreading and autoimmune disease after peptide immunization: Sm B/B'-derived PPPGMRPP and PPPGIRGP induce spliceosome autoimmunity.
J Exp Med 181
: 453
–461, 1995[Abstract/Free Full Text]
Topfer F, Gordon T, McCluskey J: Intra- and intermolecular spreading of autoimmunity involving the nuclear self-antigens La (SS-B) and Ro (SS-A).
Proc Natl Acad Sci U S A 92
: 875
–979, 1995[Abstract/Free Full Text]
Wu J, Arends J, Borillo J, Zhou C, Merszei J, McMahon J, Lou YH: A self T cell epitope induces Autoantibody response: Mechanism for production of antibodies to diverse glomerular basement membrane antigens.
J Immunol 172
: 4567
–4574, 2004[Abstract/Free Full Text]
Wu J, Borillo J, Glass II WF, Hicks J, Ou CN, Lou YH: T cell epitope of glomerular basement membrane antigen induces severe glomerulonephritis.
Kidney Int 64
: 1292
–1301, 2003[CrossRef][Medline]
Wu J, Hicks J, Borillo J, Glass II WF, Lou YH: CD4+ T cells specific to a glomerular basement membrane antigen mediate glomerulonephritis.
J Clin Invest 109
: 517
–524, 2002[CrossRef][Medline]
Robertson J, Wu J, Arends J, Glass Ii W, Southwood S, Sette A, Lou YH: Characterization of the T-cell epitope that causes anti-GBM glomerulonephritis.
Kidney Int 68
: 1061
–1070, 2005[CrossRef][Medline]
Gordon T, Topfer F, Keech C, Reynolds P, Chen W, Rischmueller M, McCluskey J: How does autoimmunity to La and Ro initiate and spread?
Autoimmunity 18
: 87
–92, 1994[Medline]
Tung KS, Lou YH, Garza KM, Teuscher C: Autoimmune ovarian disease: Mechanism of disease induction and prevention.
Curr Opin Immunol 9
: 839
–847, 1998
Turner CA, Mack DH Jr, Davis MM: Blimp-1, a novel zinc finger-containing protein that can drive the maturation of B lymphocytes into immunoglobulin-secreting cells.
Cell 77
: 297
–306, 1994[CrossRef][Medline]
Hohnel K, Dietz R, Willenbrock R: Quantification of low abundance natriuretic peptide receptor mRNA in rat tissues.
Clin Chem Lab Med 37
: 805
–812, 1999[Medline]
Livak KJ, Schmittgen TD: Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) method.
Methods 25
: 402
–408, 2001[CrossRef][Medline]
Peterson KS, Huang JF, Zhu J, D’Agati V, Liu X, Miller N, Erlander MG, Jackson MR, Winchester RJ: Characterization of heterogeneity in the molecular pathogenesis of lupus nephritis from transcriptional profiles of laser-captured glomeruli.
J Clin Invest 113
: 1722
–1733, 2004[CrossRef][Medline]
Lenda DM, Stanley ER, Kelley VR: Negative role of colony-stimulating factor-1 in macrophage, T cell, and B cell mediated autoimmune disease in MRL-Fas(lpr) mice.
J Immunol 173
: 4744
–4754, 2004[Abstract/Free Full Text]
Lin KI, Tunyaplin C, Calame K: Transcriptional regulatory cascades controlling plasma cell differentiation.
Immunol Rev 194
: 19
–28, 2003[CrossRef][Medline]
Fujigaki Y, Yousif Y, Morioka T, Batsford S, Vogt A, Hishida A, Miyasaka M: Glomerular injury induced by cationic 70-kD staphylococcal protein; specific immune response is not involved in early phase in rats.
J Pathol 184
: 436
–445, 1998[CrossRef][Medline]
Fujihara CK, De Nucci G, Zatz R: Chronic nitric oxide synthase inhibition aggravates glomerular injury in rats with subtotal nephrectomy.
J Am Soc Nephrol 5
: 1498
–1507, 1995[Abstract]
Hruby ZW, Lowry RP, Forbes RD, Blais D: Mechanisms of glomerular injury in experimental immune nephritis. II. Effect of a platelet activating factor receptor antagonist in an autologous nephritis model.
Arch Immunol Ther Exp 39
: 575
–586, 1991
Lou YH: Anti-GBM glomerulonephritis: A T cell mediated autoimmune disease?
Arch Immunol Ther Exp 52
: 96
–103, 2004
Received for publication April 21, 2005.
Accepted for publication July 22, 2005.
Related Article
This Month’s Highlights
J. Am. Soc. Nephrol. 2005 16: 3137-3138.
[Full Text][PDF]
This article has been cited by other articles:
P. Shah, A. Tramontano, and S. P. Makker Intramolecular Epitope Spreading in Heymann Nephritis
J. Am. Soc. Nephrol.,
December 1, 2007;
18(12):
3060 - 3066.
[Abstract][Full Text][PDF]
L. Chen, T. Hellmark, V. Pedchenko, B. G. Hudson, C. D. Pusey, J. W. Fox, and W. K. Bolton A Nephritogenic Peptide Induces Intermolecular Epitope Spreading on Collagen IV in Experimental Autoimmune Glomerulonephritis
J. Am. Soc. Nephrol.,
November 1, 2006;
17(11):
3076 - 3081.
[Abstract][Full Text][PDF]
Top
Abstract
Introduction
3 chain not only induced severe GN in Wistar Kyoto rats but also triggered a diversified anti-GBM antibody response through "B cell epitope spreading." In this study, an expansion of T and B cells in the renal draining lymph node (RDLN) of diseased animals after glomerular injury was observed. RDLN was demonstrated to be the location of GBM-specific B cell activation. First, B cells from RDLN of pCol(28–40)-immunized rats produced in vitro anti-GBM antibodies and antinuclear antibodies. Second, B cells specific to the peptidic B cell epitope in pCol(28–40) were absent among expanding B cells in RDLN. Those findings provided a unique opportunity to track activation of diverse GBM-specific B cells in RDLN. Expression of B lymphocyte-induced maturation protein-1, which is involved in differentiation of plasma cells, in B cells of RDLN was detected and further elevated only after T cell–mediated prominent glomerular injury (day 19). This was supported by the fact that anti-GBM antibodies became detectable only after day 20. Those results suggest that T cell–mediated glomerular injury may trigger de novo internal immunization of autoantigens released from damaged GBM, which further leads to activation of a group of GBM-specific B cells in RDLN. 
) and GBM (•) as measured by ELISA; antibody to native GBM, eluted from diseased glomeruli, was used as a positive control (
); 
3 samples for each time point. (C) IgG GBM antibodies, expressed as frequency of the intensity (0 to 4) of direct immunofluorescence on GBM, at different time points; samples (>10) in each group were taken from the indicated time ± 2 d.
